Information

In C. elegans, which neuron has the largest span, and why is it this large?

In C. elegans, which neuron has the largest span, and why is it this large?



We are searching data for your request:

Forums and discussions:
Manuals and reference books:
Data from registers:
Wait the end of the search in all databases.
Upon completion, a link will appear to access the found materials.

The model organism C. elegans is about 1 mm in length. This is quite small. In fact, some C. elegans neurons span >25% of the length of its body (ref.).

This observation leads me to the following question: Which neuron in C. elegans has the largest span, and why is it this large?


I am a member of OpenWorm.org, an open-source project that aims to simulate a C. elegans cell-by-cell. I didn't know the answer to this myself off the top of my head, but I have access to a number of folks who have been working hard on C. elegans anatomy for the last couple of years so I went ahead and asked them this question on our discuss mailing list. Thanks to Chris Grove and Tim Busbice for providing the answer that I am reporting below:

Chris Grove:

Finding the neuron with the largest span partly depends on your definition of "length" in this context. In terms of largest amount of total axon/dendrite length, this would most probably be the PVDL/PVDR sensory neurons. They produce a large number of branches of dendrites that act to perceive mechano and thermo-sensation. See the WormAtlas page here.

Tim Busbice:

If by "length" one means, absolute longest extension from head to tail, that's a tougher call, as there are several that extend from the head to the tail. An example of these are the AVDL/AVDR command interneurons, that function as touch modulators for backward locomotion induced by head-touch, and also drive backward movement of the animal along with driver cell AVA neuron, AVE and A-type motor neurons.

You can find a lot of info on C. elegans anatomy on WormAtlas and on an OpenWorm data spreadsheet in which we are consolidating detailed anatomy data from different sources.

Also I would advise checking out the OpenWorm Browser - a web-based 3D viewer that you can use to explore the C. elegans anatomy cell-by-cell (the 3D model was developed by Chris Grove, see VirtualWorm for details). For instance you can easily search for those "longest" neurons mentioned above! Screenshot of one of them selected (AVDL) below:

Hope this helped!


RESULTS

Unc-64 Encodes a Syntaxin Homolog

unc-64 was first described by Brenner (1974) as a mutant with locomotory defects. unc-64 was suspected to encode a component regulating synaptic transmission because it shares a common phenotype with many other synaptic mutants in C. elegans: resistance to the acetylcholinesterase inhibitor aldicarb ( Alfonso et al., 1993 , 1994 Nonet et al., 1993 , 1997 , 1998 Nguyen et al., 1995 Miller et al., 1996 Iwasaki et al., 1997 ). We isolated a C. elegans gene with strong similarity to rat syntaxin 1 that mapped to a region of chromosome III on the physical map, roughly corresponding to the genetic map position of the unc-64 gene (see Figure ​ Figure1 1 and MATERIALS AND METHODS for details). We tested whether the syntaxin-like gene was encoded by unc-64 using germline transformation ( Mello et al., 1991 ). unc-64(e246) mutants harboring a genomic clone of the C. elegans syntaxin gene were phenotypically rescued to wild-type, as determined by both behavioral and aldicarb resistance assays. Additionally, sequencing of the coding region of the candidate syntaxin gene from unc-64 mutant alleles in each case revealed a molecular lesion that could account for the genetic defect (Figure ​ (Figure1A). 1 A). Thus, we conclude that unc-64 encodes a C. elegans syntaxin homolog.

Three Distinct Forms of Syntaxin Are Encoded by unc-64

We isolated cDNA clones derived from the unc-64 syntaxin locus. Several complete cDNA clones representing one transcript from the C. elegans syntaxin locus were isolated from an embryonic cDNA library, and partial cDNAs representing two additional transcripts were isolated from first-strand cDNA using the PCR. Each transcript is predicted to encode a product with 63% or 64% identity to human syntaxin 1A. The three transcripts differ only in the last of eight coding exons and thus encode products with identical cytoplasmic domains but different transmembrane anchor domains (Figure ​ (Figure1, 1 , A and B). We refer to these three products as the syntaxin A, B, and C products of the unc-64 locus. At least three distinct transcripts of approximately 1.1 kb, 2.6 kb, and 3.8 kb are present on Northern blots of mixed-staged RNA (Figure ​ (Figure1C). 1 C). The size of the full-length syntaxin B cDNA isolated from the cDNA library is of the appropriate size to correspond with the smaller transcript. The larger transcripts are of an appropriate length to represent the syntaxin A and C cDNAs if they utilize the same polyadenylation site. The splicing of the transmembrane domain of syntaxin in C. elegans is reminiscent of the splicing pattern observed for the rat syntaxin 2 and the mouse syntaxin 3 genes ( Bennett et al., 1993 Ibaraki et al., 1995 ).

Unc-64 Syntaxin Is Expressed in the Nervous System and Secretory Tissues

The C. elegans syntaxin sequences are most similar to the vertebrate neuronal-specific syntaxin 1 molecules and hence are likely to be expressed in the nematode nervous system. To examine the expression pattern, antisera were raised against bacterially expressed C. elegans syntaxin fusion protein. An affinity-purified anti-syntaxin antiserum was incubated with fixed whole adult animals, and the antibodies were detected using FITC-conjugated secondary antisera. As expected, syntaxin was detected in the nervous system (Figure ​ (Figure2). 2 ). Syntaxin immunoreactivity was very strong and distributed uniformly along the major process bundles: the nerve ring (Figure ​ (Figure2A), 2 A), ventral cord (Figure ​ (Figure2B), 2 B), and dorsal cord (Figure ​ (Figure2A). 2 A). Syntaxin immunoreactivity also was detected in the vast majority of neuronal cell bodies as well as in commissural and dendritic processes (Figure ​ (Figure2B). 2 B). By contrast, the vesicle-associated markers synaptobrevin, synaptotagmin, and RAB-3 are restricted to the synaptic-rich regions of the nervous system where they show a punctate, rather than uniform, appearance ( Nonet et al., 1993 , 1997 , 1998 ). Syntaxin was also detected in the pharyngeal nervous system (Figure ​ (Figure2A). 2 A). Furthermore, syntaxin was detected outside the nervous system. Immunoreactivity was strong in the uv1 cells of the vulva, in the intestine (Figure ​ (Figure2C), 2 C), and in the spermatheca (Figure ​ (Figure2D). 2 D). Detectable levels of syntaxin protein were not observed in other tissues. Our protein expression data suggest that unc-64 encodes the homolog of the vertebrate syntaxin 1 gene ( Bennett et al., 1992 , 1993 ).

Expression of unc-64 syntaxin. (A𠄽) Whole adult wild-type hermaphrodite worms fixed and stained with α-UNC-64 primary antibodies and visualized with FITC-conjugated antibodies. (A) Lateral view of the head region showing immunoreactivity in the nerve ring, dorsal cord, and pharyngeal nervous system. (B) Ventral view of the midbody region showing expression in ventral cord axons, neuronal cell bodies, and in commissural and sublateral processes. (C) A lateral view of the midbody showing UNC-64 immunoreactivity on the basolateral surface of the intestine. (D) A lateral view of the midbody showing immunoreactivity in the hermaphrodite spermatheca. The male vas deferens also stains. (E) A lateral view of the vulva viewed using differential interference contrast microscopy. (F) Same plane of focus as panel E showing fluorescence from the UNC-64 A::GFP product in the uv1 cells of the vulva. (G–H) Whole adult wild-type animals fixed and stained for LacZ gene activity expressed under the control of the unc-64 promoter. LacZ activity assayed using X-gal as a substrate. The lacZ gene contains a nuclear localization signal. (G) Lateral view of the head region showing expression of unc-64 ::lacZ in neuronal and intestinal nuclei. (H) Lateral view of the midbody of an adult hermaphrodite showing unc-64::lacZ expression in intestinal nuclei, neuronal nuclei, and the spermatheca. Scale bar, in all panels, 20 μm.

As it was not feasible using immunohistochemistry to determine which neurons expressed syntaxin, the expression pattern of the SNARE gene was examined using a translational lacZ fusion. A genomic syntaxin fragment (Figure ​ (Figure1C) 1 C) was inserted into a plasmid containing a nuclear-localized lacZ reporter gene ( Fire et al., 1990 ). As expected, syntaxin was expressed in the vast majority of all neurons (Figure ​ (Figure2, 2 , G and H). In addition, syntaxin expression was high in intestinal nuclei, in the spermatheca, and in the uv1 cells of the vulva (Figure ​ (Figure2H). 2 H). In summary, C. elegans syntaxin is expressed in both secretory and neuronal tissues. In neurons, the protein is not specifically associated or concentrated at synaptic release sites rather, it is ubiquitously distributed.

The three identified products deriving from the C. elegans syntaxin locus differ only in those sequences that are predicted to lie in the lipid bilayer and act as transmembrane anchors. Because production of antibodies capable of distinguishing these products did not seem feasible, we engineered constructs that tagged each product by appending the GFP-coding region to the last codon of each predicted transmembrane domain. The constructs were introduced into wild-type animals by germline transformation, and the expression pattern of the GFP-tagged syntaxins was examined. Syntaxin A and C were expressed in neurons and in five nonneuronal tissues. By contrast, expression of syntaxin B was limited to the five nonneuronal cell types. In neurons, syntaxin A and C were found ubiquitously on neuronal cell bodies, axons, and dendrites. Also stained were the uv1 secretory cells of the vulva, the excretory gland cells, the distal tip cells, the spermatheca, and the intestinal muscle (Figure ​ (Figure2, 2 , E and F). In the uv1 cells, GFP fluorescence was concentrated subcellularly (Figure ​ (Figure2, 2 , E and F). The expression pattern from the GFP, lacZ, and antibody studies were consistent, except that GFP reporter expression was observed in the intestine only with syntaxin C fusions. This expression was limited to the posterior half of the intestine. It is possible that sequences that direct gut expression are absent from the GFP clones since sequences further upstream of the initiation codon are present in the lacZ constructs. In summary, the alternative UNC-64 syntaxin products are expressed in distinct but overlapping neuronal and nonneuronal secretory cells.

Isolation of Additional unc-64 Syntaxin Mutants

Additional alleles of the unc-64 gene were isolated in three independent screens. md130 and md1259 were isolated in a screen for aldicarb- resistant mutants occurring spontaneously in a strain with high levels of Tc1 transposition ( Miller et al., 1996 ). js21 was isolated in a general screen for aldicarb-resistant mutants using EMS as a mutagen (M.L.N., unpublished data). Finally, two lethal alleles, js115 and js116, were isolated in a noncomplementation screen (see MATERIALS AND METHODS). Molecular characterization revealed that the e246 and js21 lesions are missense mutations that result in alanine-to-valine substitutions at codons 241 and 248, respectively. These codons are in the H3 domain of syntaxin, a region thought to assume an amphipathic helical structure capable of forming coiled-coil interactions with other proteins ( Hardwick and Pelham, 1992 Calakos et al., 1994 Fasshauer et al., 1997 ). The point mutations in md130 and md1259 disrupt splicing sites (see MATERIALS AND METHODS for details). The js116 mutation results in premature termination in the transmembrane domain of the syntaxin A product, while the js115 lesion results in premature termination at codon 71. Immunohistochemical examination of whole fixed animals revealed that syntaxin product remained in the nervous system of all viable mutants, although the level of syntaxin was reproducibly lower in the md1259 mutant. UNC-64 staining was not detected in js115. All available evidence suggests that js115 represents the null phenotype (see MATERIALS AND METHODS).

Behavioral Defects of unc-64 Syntaxin Mutants

All six unc-64 mutants exhibit locomotory abnormalities. The lethal mutants exhibit the most severe motor defects. js115 animals are virtually completely paralyzed and rarely maintain a sinusoidal posture, although they occasionally will make slow head movements (Figure ​ (Figure3). 3 ). js116 animals exhibit very slow motor movements (Figure ​ (Figure3) 3 ) and tend to adopt a coiled position. Both js115 and js116 animals arrest development just after hatching and eventually die as L1 larvae. Behaviorally, the viable js21 and e246 mutants are very lethargic, although they retain the ability to move briefly when prodded. Locomotory abnormalities were quantified by examining the rate of movements of animals imaged using a charge-coupled device camera (Table ​ (Table1). 1 ). Despite exhibiting severe locomotory defects, other behaviors of the hypomorphic mutants were only weakly affected (Table ​ (Table1). 1 ). The rate of pharyngeal pumping was relatively normal in the hypomorphic mutants e246 and js21, although pumping was completely abolished in the js115 null mutant. The motor defecation program was only slightly abnormal in hypomorphs, but completely abolished in js115. It is interesting to note that, compared with other synaptic mutants with similarly severe locomotory defects (e.g., unc-18 and unc-13), the unc-64(e246) mutant has much milder deficiencies in other behaviors. unc-64(e246) mutants also exhibit a tendency to form enduring dauer larvae ( Iwasaki et al., 1997 ), a developmental decision regulated by sensory input ( Bargmann and Horvitz, 1991 ). Two other mutants, md130 and md1259, also exhibit milder behavioral defects, but minimal analysis was performed on these alleles due to the difficulty in determining, with certainty, the resulting gene products. Finally, at the morphological level, muscle, the intestine, the pharynx, the cuticle, and the organization of the nervous system appear normal in both the hypomorphs and the null mutant. In summary, a variety of genetic lesions in the syntaxin gene result in behavioral defects of varying severity. Presumably, all these phenotypes represent manifestations of an underlying impairment of synaptic transmission.

Phenotype of lethal unc-64 syntaxin mutants. Bright field images of first larval stage animals of wild-type, unc-64(js115), and unc-64(js116) animals on an E. coli bacterial lawn at the indicated time intervals (wild-type in seconds and mutants in minutes). Scale bar, 100 μm.

Table 1

Behavioral defects unc-64 mutants

StrainLocomotion Pharyngeal pumps/minDefecation
Speed (μm/sec)Efficiency a Cycle time (s)EMC b
N2 c 161 ±�0.63250 ±�42.0 ±𠂓.398%
e24613 ±𠂘0.18209 ±�39.3 ±𠂓.387%
js2151 ±�0.43234 ±�48.8 ±𠂙.280%
js115 d 0NA0.2 ±𠂐.6NoneNA
js116 d 0NA3 ±𠂒NoneNA

Data shown as mean ± SD. A minimum of 18 animals were analyzed. NA, Not applicable. 

Late Arrest of unc-64 Syntaxin Mutants Is Not the Result of Maternal Contribution

In D. melanogaster, mosaic clones that lack syntaxin cannot be isolated, suggesting the molecule is required in all cells ( Schulze and Bellen, 1996 ). Furthermore, syntaxin is also required early in development during cellularization of the Drosophila embryo ( Burgess et al., 1997 ). As our most severe mutant arrests development after completing embryogenesis, we examined whether unc-64 syntaxin mutants completed embryogenesis using maternally derived products. We constructed a strain homozygous for js115, which also harbored an extrachromosomal array bearing a visible marker (rol-6), a muscle-specific unc-54::GFP construct, and the wild-type unc-64 syntaxin gene. These animals segregate both wild-type and mutant progeny as the array is not transmitted to all progeny. We examined the progeny of nine mosaic animals lacking syntaxin product in the germline. These animals were identified on the basis of their failure to produce viable progeny. Nearly 99% (n = 740) of the progeny of these mosaic animals completed embryogenesis and arrested as first larval stage animals with a phenotype indistinguishable from the js115 progeny of heterozygous mothers. The secreted cuticle of these animals appeared normal, containing well organized alae (Figure ​ (Figure4). 4 ). Most of the remaining 1% of progeny arrested late in embryonic development as threefold embryos. None of the progeny of the mosaic animals we examined expressed GFP, confirming that the unc-64(+) array was absent from the germline. We also examined viable mosaic animals that lacked unc-64 expression in muscle. Three mosaics that lacked GFP in muscle anterior of the vulva and 11 mosaics that lacked GFP in both dorsal body wall muscle quadrants all moved normally, suggesting that syntaxin is not required in muscle. Additionally, most of the severely uncoordinated mosaic animals we examined retained GFP expression in muscle, consistent with a neuronal site of action for syntaxin.

Cuticle of L1 js115 syntaxin mutant animals. Lateral view of the cuticle structures (alae) of the L1 larvae of wild-type and homozygous unc-64(js115) animals segregating from a mosaic animal lacking unc-64(+) in the germline. Arrowheads point to the alae (cuticular specializations) observed in L1 larvae. Scale bar, 20 μm.

In separate experiments, we injected antisense RNA into the germline of wild-type animals in an attempt to inactivate the unc-64 gene by RNA-mediated interference ( Guo and Kemphues, 1995 Rocheleau et al., 1997 ). This technique phenocopies the null phenotype of most genes that provide maternal contributions to the embryo, but is much less efficacious in phenocopying late-acting zygotic genes. Our injections failed to confer the js115 phenotype or an aldicarb-resistant phenotype. In summary, we find no evidence for either a maternal contribution of unc-64 or for a requirement for zygotic syntaxin in the early embryo.

Synaptic Transmission Defects in unc-64 Mutants

To assess neurotransmission in unc-64 mutants, we began by examining the effect of cholinergic pharmacological agents. We quantified the sensitivity of unc-64 mutants to aldicarb, a potentiator of released acetylcholine, and to levamisole, an agonist of nicotinic acetylcholine receptors in the nematode ( Lewis et al., 1980a , b ). All of the viable unc-64 mutants examined were resistant to high levels of aldicarb (Figure ​ (Figure5A), 5 A), as was the acetylcholine receptor mutant unc-29 ( Fleming et al., 1997 ). This strongly suggests that cholinergic transmission is reduced in these mutants. However, unlike unc-29, the unc-64 syntaxin mutants were sensitive to levamisole (Figure ​ (Figure5B), 5 B), suggesting an intact postsynaptic apparatus and thus a presynaptic defect. Finally, examination of the lethal mutations revealed that aldicarb enhanced js116 movement but failed to stimulate movement of js115 animals, providing evidence that cholinergic synaptic transmission may be completely abolished in the null mutant.

Pharmacological properties of unc-64 mutants. (A) Aldicarb sensitivity of unc-64 mutants and the wild type. Shown is the percentage of adult animals paralyzed after a 4-h exposure to various concentrations of aldicarb on plates seeded with E. coli wild-type (•), js21(□), e246 (Δ), and unc-29 (○). (B) Levamisole sensitivity of unc-64 mutants and the wild type. Shown is the percentage of adult animals paralyzed after exposure to 100 μM levamisole at various times. Wild-type (•), js21(□), e246 (Δ), and unc-29 (○).

To provide direct evidence of synaptic transmission defects in C. elegans syntaxin mutants, we used an extracellular recording technique developed by Raizen and Avery (1994) . This method permits the detection of postsynaptic potentials in pharyngeal muscle resulting from the action of both excitatory and inhibitory motor neurons. MC is an excitatory motor neuron thought to stimulate pharyngeal muscle contraction ( Raizen et al., 1995 ). Its activation is visualized as a single depolarizing transient preceding many pumps. In unc-64 syntaxin mutants, MC neuronal function appeared defective. Specifically, we observed a series of multiple excitatory postsynaptic potentials before initiation of pharyngeal pumps (arrows, Figure ​ Figure6), 6 ), which may reflect nonsynchronous transmitter release from MC. A similar phenotype has previously been observed in aex-3, rab-3, and snb-1 mutants ( Iwasaki et al., 1997 Nonet et al., 1997 , 1998 ).

Pharyngeal recordings from wild-type and unc-64 mutant animals. Shown are characteristic electrophysiological recordings from the wild-type strain N2 (A), unc-64(e246) (B), unc-64(js21) (C), unc-64(md130) (D), aex-3(y255) (E), snb-1(js17) (F), and unc-64(js21) snb-1(js17) (G). Arrows indicate MC-induced transients, and filled circles indicate M3-induced transients. All traces are millivolts versus time.

Our analysis concentrated on the examination of inhibitory postsynaptic potentials (IPSPs) produced by the motor neuron M3, a glutaminergic neuron that regulates the duration of pharyngeal muscle contractions ( Avery, 1993 Dent et al., 1997 ). In wild-type animals, M3 transients were regularly spaced and initiated 28 ± 5 ms after depolarization of the pharynx. In unc-64 (e246) animals, the first detectable IPSP often was greatly delayed (102 ± 43 ms Figure ​ Figure6). 6 ). This is most easily illustrated by examining the distribution of IPSPs as a function of time after the initiation of the pharyngeal pump (Figure ​ (Figure7, 7 , A and B). While IPSPs occurred relatively synchronously in wild-type, they were broadly distributed in e246 mutants. However, normalizing the time of the IPSP relative to the first IPSP within the pump reveals that, although the initial IPSP was variably delayed in the mutant, subsequent IPSPs were temporally synchronized (Figure ​ (Figure7, 7 , C and D). A second difference between wild-type and mutant was that the amplitude of M3 IPSPs was relatively constant in wild-type but increased substantially in e246. While in wild-type, the initial IPSP was slightly larger than the final IPSP, in e246 there was a threefold increase in IPSP amplitude by the end of the pump (Figure ​ (Figure7, 7 , C and D). js21 exhibited a similar but less severe defect (our unpublished data). These particular defects in M3 transmission are specific to the syntaxin helical mutants. The splicing mutants md130 and md1259 do not exhibit these alterations but instead resemble aex-3(y255), rab-3(y250), and snb-1(js17) in that their IPSPs were consistently decreased in amplitude yet initiated without delay after contraction of the pharynx (Figure ​ (Figure6). 6 ). In summary, lesions in the H3 domain of syntaxin affect both the kinetics of transmitter release and the amplitude of postsynaptic potentials.

Analysis of M3 transmission in wild-type and unc-64(e246) animals. The amplitude of the first ([triof]), second (○), third (▪), fourth (Δ), fifth (•), and six (□) M3-induced IPSPs within each pump (see Figure ​ Figure6) 6 ) is plotted as a function of time of occurrence. M3 amplitudes are normalized relative to the mean R-phase amplitude for each record. (A and B) Time of occurrence is plotted relative to initiation of pharyngeal muscle excitation. (C and D) Time of occurrence is normalized to the first M3 IPSP of the same pump. (A) Each group of M3 IPSPs cluster showing the high temporal synchrony of M3 transmitter release in wild-type (pump duration 148 ± 23 ms [mean ± S.D.], N = 130 pumps). (B) In unc-64(e246) temporal synchrony is lost as the timings of each group of M3 IPSPs are distributed over a larger time span. The first M3 IPSP occurs from 22 to 212 ms after muscle excitation (pump duration 209 ± 44 ms [mean ± S.D.], N = 130 pumps). (C) Aligning the IPSPs by timing of the first M3 IPSP again shows temporal clustering in wild-type. Mean IPSP spacings are ֱst to 2nd] 34 ms, ֲnd to 3rd] 31 ms, ֳrd to 4th] 32 ms, ִth to 5th] 32 ms, and ֵth to 6th] 36 ms. The downward sloping linear regression reveals that subsequent wild-type IPSP amplitudes tend to decrease slightly. (D) Similar alignment to the first M3 IPSP in e246 reveals that subsequent IPSPs are, in fact, regularly spaced at ֱst to 2nd] 41 ms, ֲnd to 3rd] 42 ms, ֳrd to 4th] 48 ms, ִth to 5th] 42 ms, and ֵth to 6th] 43-ms mean intervals. The upward sloping regression line demonstrates that mutant IPSP amplitudes increase up to threefold with successive M3 firings.

Genetic Interactions between Syntaxin and Synaptobrevin Mutants

Syntaxin interacts tightly with synaptobrevin and SNAP-25 to form a stable ternary complex. This complex is proposed to represent an intermediate in the synaptic vesicle fusion cycle ( Sollner et al., 1993 ). In each of these proteins, the domains participating in the interaction are predicted to contain α-helices capable of assembling into coiled-coil structures ( Hardwick and Pelham, 1992 Fasshauer et al., 1997 ). Lesions that lie on the hydrophobic face of the presumed coiled helix region exist in both C. elegans syntaxin and synaptobrevin ( Nonet et al., 1998 ). We constructed double mutants between the synaptobrevin mutants snb-1(js17), snb-1(js44), and snb-1(md247) and the two missense helical domain syntaxin mutants js21 and e246 (Figure ​ (Figure8). 8 ). When analyzing the phenotypes of all six double mutants, we observed strong synergistic effects with certain allelic combinations. The size of young adult animals (Figure ​ (Figure9) 9 ) and the rates of pharyngeal pumping (Table ​ (Table2) 2 ) were drastically reduced in certain double-mutant combinations. For example, when the weakest synaptobrevin mutant, js44, was combined with the weakest syntaxin mutant, js21, the result was a more severe phenotype than when js44 was paired with the stronger e246 syntaxin lesion. Conversely, snb-1(js17) interacted strongly with e246, but not with js21. Neither syntaxin allele showed interactions with snb-1(md247), whose lesion is outside the helical domain and alters the transmembrane domain of synaptobrevin ( Nonet et al., 1998 ). EPG analysis provided no additional insight, as even the healthier double mutants displayed more severe EPG phenotypes than the singles and failed to exhibit any significant M3 activity (Figure ​ (Figure6). 6 ). This was not unexpected, as this assay examining M3 activity is very sensitive and is able to detect defects in mutants that exhibit only very mild behavioral changes ( Nonet et al., 1997 ).

Mutations in snb-1 and unc-64 on the hydrophobic faces of α-helices proposed to mediate interactions between synaptobrevin and syntaxin. Two model amphipathic α-helices composed of a repeating seven-amino acid pattern. The residues are labeled a through g. Residues in a and d positions are usually hydrophobic and are proposed to mediate interactions between binding partners. Below, a portion of the sequences of human, fly, yeast, and worm syntaxin and synaptobrevin ( Archer et al., 1990 DiAntonio et al., 1993 Gerst et al., 1992 ) are aligned. Hydrophobic residues in the a and d positions of the predicted α-helices are labeled. The site of lesions in C. elegans syntaxin and synaptobrevin ( Nonet et al., 1998 ) are also labeled. The two sequences are oriented assuming they will interact in a parallel manner ( Hanson et al., 1997 Lin and Scheller, 1997 ).

Phenotypes of unc-64 snb-1 double- mutant animals. Bright field photographs of young adult animals singly and double mutant for unc-64 syntaxin and snb-1 synaptobrevin mutations. Animals were isolated as L4 larvae and then incubated 1 d before examination. Specific combinations of alleles cause severe behavioral, growth, and pharyngeal pumping defects.

Table 2

Pharyngeal pumping in unc-64 snb-1 double mutants

Pharyngeal pumping (pumps per min) of animals of genotype: unc-64 (horizontal) snb-1 (vertical). 


A Caenorhabditis elegans Model System for Amylopathy Study

We describe methods to study aspects of amylopathies in the worm C. elegans. We show how to construct worms expressing human A&beta42 in neurons and how to test their function in behavioral assays. We further show how to obtain primary neuronal cultures that can be used for pharmacological testing.

Abstract

Amylopathy is a term that describes abnormal synthesis and accumulation of amyloid beta (A&beta) in tissues with time. A&beta is a hallmark of Alzheimer's disease (AD) and is found in Lewy body dementia, inclusion body myositis and cerebral amyloid angiopathy 1-4 . Amylopathies progressively develop with time. For this reason simple organisms with short lifespans may help to elucidate molecular aspects of these conditions. Here, we describe experimental protocols to study A&beta-mediated neurodegeneration using the worm Caenorhabditis elegans. Thus, we construct transgenic worms by injecting DNA encoding human A&beta42 into the syncytial gonads of adult hermaphrodites. Transformant lines are stabilized by a mutagenesis-induced integration. Nematodes are age synchronized by collecting and seeding their eggs. The function of neurons expressing A&beta42 is tested in opportune behavioral assays (chemotaxis assays). Primary neuronal cultures obtained from embryos are used to complement behavioral data and to test the neuroprotective effects of anti-apoptotic compounds.

Introduction

Amyloid beta (A&beta) is a peptide of 36-43 amino acids that is formed after sequential cleavage of the amyloid precursor protein (APP) by &beta and &gamma secretases 1 . The &gamma secretase processes the C-terminal end of the A&beta peptide and is responsible for its variable lengths 5 . The most common forms of A&beta are A&beta40 and A&beta42, the latter being more commonly associated to pathologic conditions such as AD 5 . At high concentrations A&beta form &beta-sheets that aggregate to form amyloid fibrils 6 . Fibrils deposits are the main component of senile plaques surrounding neurons. Both plaques and diffusible, non-plaque A&beta oligomers, are thought to constitute the underlying pathogenic forms of A&beta.

Laboratory study of neuronal amylopathies is complicated by the fact that these conditions progress with time. Therefore, it is important to develop genetically tractable animal models-complementary to mice-with short life span. These models can be used to elucidate specific aspects of amylopathies-typically cellular and molecular-and by virtue of their simplicity, help to capture the essence of the problem. The worm Caenorhabditis elegans falls is this category. It has a short life span,

20 days and in addition basic cellular processes including regulation of gene expression, protein trafficking, neuronal connectivity, synaptogenesis, cell signaling, and death are similar to mammalian 7 . Unique features of the worm include powerful genetics and lack of a vessel system, which enables to study neuronal damage independently of vascular damage. On the other hand, the lack of a brain limits the use of C. elegans to studying many aspects of neurodegeneration. In addition, the reproduction and identification of anatomical distributions of lesions cannot be performed in this organism. Other limitations include the difficulty to assess both differences in gene expression profiles and impairment of complex behavior and memory function. Here we describe methods to generate C. elegans models of amylopathies.

Subscription Required. Please recommend JoVE to your librarian.

Protocol

1. Construction of Transgenic Worms

  1. Transformation.
    1. Prepare injection pads. Place a drop of hot, 2% agarose dissolved in water, onto a glass coverslip. Quickly place a second coverslip on the drop and lightly tap it. After the agarose is solidified, slide coverslips apart, and bake the coverslip-pad in a vacuum oven at 80 °C O/N.
    2. Pull pipettes. We use a Sutter P-97 puller to pull 1/0.5 mm O.D./I.D. borosilicate capillaries with filament. Pipettes are forged with closed tip which is broken open at a later stage.
    3. Prepare injection mix. Make the injection mixture containing the DNA of interest (20 ng/&mul per construct) plus empty plasmid DNA to a final concentration of 150 ng/&mul. Centrifuge the injection mixture to remove any contaminant. This step is crucial because contaminants reduce transformation efficiency. Transfer the top 5 &mul (debris and other contaminants collect in the bottom) to a fresh tube to be used in injection.
    4. Load injection mix by capillarity. The bottom of the pipette is immersed in the injection mix. Typically 0.5 &mul are sufficient to inject 100 worms.
    5. Load injection pipette. Insert the pipette onto the holder of the micromanipulator and break it open by rubbing its tip against the edge of a cover slip mounted on a glass slide or alternatively against debris on the agarose pad. This is a crucial step as too large tips damage the worm and too small tips are easily clogged. The quality of the broken tip can be judged by the shape and most importantly by the flow rate. The flow rate can be assessed by the size of the bubbles flowing from the open tip immersed in a drop of halocarbon 700 oil in response to an injection pressure.
    6. Transfer worms to injection pad. Place a drop of 700 halocarbon oil on an injection pad and transfer several (1 or 2 for beginners) worms to the oil. Use a worm pick to push the worms down onto the pad until they adhere to the agarose. Worms should be oriented in rows with ventral sides facing the same direction. Avoid touching worm's head. If after several attempts the worms fail to adhere to the pad, replace with a fresh pad or increase agarose thickness and/or concentration.
    7. Insert the pipette into the worm. By moving the stage, position the worm under the pipette. Position the pipette in the central core of the gonad because its cytoplasm is shared by many germ cell nuclei. This increases the likelihood to deliver injected DNA to many progeny. The pipette should lie almost parallel to the worm (

    1. Place 40 healthy, well-fed, L4 worms into a fresh plate
    2. Irradiate with &gamma-ray with 4,000 rads for 40 min. Transfer irradiated worms (P0) in fresh, OP50 seeded plates (4 worms/plate). Worms can alternatively be irradiated with a dose of 300 J/m 2 UVs.
    3. Transfer 10-20 F1 transformants from each plate to individual plates, label the plates with the corresponding P0 origin.
    4. Single out 2-4 F2 transformants for each F1 to separate plates. Check the F3 progeny for 100% transmission of the transformation marker. Typically, 1-3% of progeny from an irradiated worm will have an integration event.
    5. Make stocks of three or more independent transformant lines.

    1. Age-synchronization.
      1. Grow worms in standard 10 cm NGM plates + OP50 E. coli until a large population of gravid adults is reached (3-5 days).
      2. Collect the worms in 50 ml Falcon tubes by suspending them in 1 ml M9 buffer.
      3. Add 5 ml M9 buffer and centrifuge at 450 x g for 3 min. Discard supernatant. Repeat 3-4x.
      4. At the end of the last centrifugation, remove the supernatant and add 10 volumes of basic hypochlorite solution (0.5 M NaOH, 1% hypochlorite freshly mixed) to the pelleted worms. Incubate at RT for

      1. Prepare several pieces of agar roughly 0.5x0.5x0.5 cm. We deposit the agar in a 10-cm plate and we cut the agar chunks from there.
      2. Soak the agar chunk in a solution containing the desired attractant (at saturating, near-saturating concentrations) for 2 hr. Typical attractants are lysine (0.5 M), biotin (0.2 M).
      3. Deposit an agar chunk in a 10 cm test plate in which the location of a test spot and a control spot have been marked (Figure 1A). Allow equilibration and formation of a gradient O/N (Figure 1B). Prepare 5 plates for a single experiment.
      4. Prior the experiment add 10 &mul of 20 mM NaN3 (anesthetic) to each spot.
      5. Place 20 age-synchronized worms in the center of the plate. Place the plate in the incubator at 20 °C.
      6. After 1 hr, count the animals on the test/control spots and calculate the chemotaxis index, (C.I.) as follow: where N, Ntest and NCnt., indicate the total number of animals, the number of animals in the test spot and the number of animals in the control spot.

      3. Primary Embryonic Cell Culture

      1. Lyse worms as described in age-synchronization.
      2. Stop the lysis reaction by adding the same volume of sterile egg buffer and centrifuge at 450 x g for 5 min. Gently discard supernatant being careful to not lose pelleted eggs. Repeat 2-3x or until supernatant is clear.
      3. Resuspend pelleted eggs (and carcasses) in 2 ml sterile egg buffer and add 2 ml of sterile 60% sucrose in egg buffer. Mix this solution until eggs are completely resuspended (as they tend to form clumps under centrifugation) by hand or by vortexing.
      4. Centrifuge at 450 x g for 15 min.
      5. Carefully transfer supernatant (containing the eggs) to a sterile tube. Discard pellet which contains carcasses and other by-products of lysis.
      6. Remove residual sucrose by resuspending the eggs in egg buffer and centrifuging at 450 x g for 5 min. Gently collect and discard the supernatant. Repeat 3x.
      7. Under a laminar hood resuspend pelleted eggs in sterile egg buffer containing 1 U/ml chitinase at RT to digest the eggshells. After 30 min start to monitor the reaction (under an inverted cell culture microscope). Each batch of chitinase has a slightly different activity. Typically digestion is completed in 1 hr.
      8. When roughly 70-80% eggshells are digested by chitinase add CM-15 (L-15 cell culture medium containing 10% fetal bovine serum, 50 U/ml penicillin, and 50 &mug/ml streptomycin). Dissociate cells using a syringe with a 27 gauge. Filter the cell suspension with a 5.0 &mum filter to remove intact embryos, clumps of cells and larvae.
      9. Pellet the dissociated cell suspension by centrifugation at 450 x g for 15 min. Remove the supernatant and resuspend the pellet in CM-15 cell culture medium.
      10. Plate dissociated cells on glass cover slips previously coated with peanut lectin (0.1 mg/ml) dissolved in water. Note: cells must adhere to the substrate in order to differentiate.
      11. Cells can be maintained at RT (16-20 °C) in air for more than 2 weeks.

      Subscription Required. Please recommend JoVE to your librarian.

      Representative Results

      With our protocols we study the effects of human A&beta42 oligomer on neuronal function 8 . A fragment encoding human A&beta42 and the artificial signal peptide coding sequence of Fire vector pPD50.52 was amplified from construct PCL12 9 using primers that introduced a Sma 1 restriction endonuclease site at the ends. The fragment was then inserted into a construct containing a 2,481-bp flp-6 promoter sequence in the pPD95.75 Fire vector between the unique Sma 1 site 10 . Using the transformation techniques described in protocol 1 we constructed a transgenic worm expressing A&beta42 in the ASE neurons (FDX(ses25) strain) 8 . To mark positive transformants we used the Pgcy-5::GFP reporter which specifically drives GFP expression in the ASE right (ASER) neuron 11 . Worms stick to the pad because the dry agarose absorbs their water. Therefore it is crucial that the animals are placed onto the injection pad and injected relatively quickly, because otherwise they will desiccate and die. The percentage of F1 progeny that carries a transmissible extrachromosomal array may vary. Typical values are in the 3-7% range. It is important that the injection mix (1.1.3) contains non-encoding DNA sharing sequence homology with the transgenic DNA (usually empty vector) because the DNAs undergo homologous recombination with each other. However, if overexpression of the transgene is a problem the injection mix should be supplemented with 50-100 ng/&mul of genomic DNA digested with Sca 1. ASE neurons detect water soluble attractants such as biotin and therefore their function can be assessed in behavioral assays (chemotaxis assay, protocol 2 and Figures 1A and 1B) 12 . In a typical experiment, we tested seven day old worms expressing Pgcy-5::GFP alone (DA1262 strain) or with A&beta42 (FDX(ses25)) for chemotaxis to biotin. In young worms (3-4 day old) the effects of A&beta42 expression are modest but already detectable (

      10% decrease in chemotaxis index, see ref. 8 ). Representative results of this experiment are shown in Figures 1C and 1D. Most DA1262 worms were found in, or nearby, the attractant spot (Figure 1C). By contrast only a few worms expressing A&beta42 could find the attractant spot (Figure 1D). We tested 100 animals/genotype distributed in 5 test plates/genotype obtaining a chemotaxis index for biotin 0.68±0.09 and 0.12±0.04 for DA1262 and FDX(ses25), respectively. Active worms were individually picked and transferred to the center of the test plate. It is important to not only quickly, but also gently transfer the worms because otherwise they may remain inactive for several minutes and fail to track the attractant. At the end of the experiment we suggest monitor worms activity by looking at their paths. If only a few tracks are visible we usually discard the plate.

      GFP fluorescence in the ASER neurons of FDX(ses25) worms disappears within the first eleven days of life (not shown). This suggests that these cells undergo apoptosis due to the presence of A&beta42. Therefore we determined whether a broad-spectrum inhibitor of apoptosis such as caspase inhibitor N-(2-Quinolyl)valyl-aspartyl-(2,6-difluorophenoxy)methyl Ketone (Q-VD-OPh) 13 could stop the loss of ASER cells. To this end we employed cultured primary ASER neurons from embryos 14 , which were prepared as described in protocol 3. Representative images of an ASER neuron in a young (4 day old) FDX(ses25) worm along with images of neurons in culture are shown in Figures 2A-C. Cultured ASER cells were short-lived (Figure 2D). As expected incubation with 0.3 &mug/ml Q-VD-OPh freshly supplemented daily, completely stopped the loss of fluorescent ASER neurons. When working with primary neurons in culture it is important to maintain an opportune cell density. This parameter mainly depends on the number of worms used to extract the eggs. We measure cell density with a standard hemocytometer and take particular care to maintain cell density constant from cell culture to cell culture. For pharmacological experiments such as those described here we worked with

      200,000 cells/cm 2 which was obtained by harvesting four confluent 10 cm plates. Cells were plated in 12 wells of a 24-well plate. For optimal results it is also important to dissociate the embryonic cells before seeding as they tend to form clumps. We use a syringe with a 27 gauge needle and gentry aspirate the suspension back and forth a couple of times. This is generally sufficient to dissociate most of the cells (especially in the beginning we suggest to check the suspension under a microscope).


      Figure 1. Chemotaxis assay. A. Representative chemotaxis plate. Attractant (biotin) and control spots are marked with an hollow and a filled circle. B. A chemotaxis plate loaded with a 0.5-cm diameter piece of agar used to establish a gradient of biotin. The piece of agar was cut from a 10-cm plate using the top side of a glass Pasteur pipette. C. Representative distribution of worms around the attractant spot. In this example the majority of DA1262 worms were able to localize the source of attractant (biotin). D. As in C. for FDX(ses25) worms. These biotin-insensitive worms exhibited a scattered distribution around the plate and only few were found near the attractant spot.


      Figure 2. Culture of C. elegans embryonic cells. A. Fluorescence microscopy image (left picture) and bright light (right picture) taken from a FDX(ses25) transgenic worm head. This worm expresses GFP in the ASER neuron driven by the gcy-5 promoter. B. Bright light image of a culture of FDX(ses25) embryonic cells. Scale bar is 5 &mum. C. Fluorescence microscopy image of a cultured FDX(ses25) ASER neuron. Images were taken with an Olympus BX61 microscope equipped with a digital camera. D. Representative experiment testing the viability of cultured, age-synchronized, ASER neurons. Cells were obtained from DA1262 embryos (hollow circles) or FDX(ses25) embryos maintained in the absence/presence of 300 ng/ml Q-VD-Oph (hollow and filled squares, respectively). The disappearance of GFP fluorescence was used as a measure of a neuron's viability. The experiment started with

      300 fluorescent ASER neurons. Viability was calculated as 100*(number of fluorescent cells at day X divided by number of fluorescent cells at day 1). Click here to view larger figure.

      Subscription Required. Please recommend JoVE to your librarian.

      Discussion

      Here we describe a combined approach, to study cellular and molecular aspects of amylopathies using C. elegans. The advantages of this approach include: 1) low cost. C.elegans is maintained in normal Petri dish seeded with bacteria, at room temperature. 2) Powerful genetics. Transgenic animals can be obtained in few months and a wide array of promoter sequences is available to drive expression of the desired gene in specific neurons. 3) Simple, well-characterized, nervous system. C. elegans possesses a remarkably simple nervous system (302 neurons). This simplicity has afforded extensive characterization of the worm's nervous system including cell lineage, specific function/role of each neuron and its synaptic connections. The limitations of C. elegans include the small size of the cells, which hinders the application of standard biochemical techniques such as immunohistochemistry and a thick skin (cuticle) which insulates neurons form the external environment. Therefore, pharmacological approaches may be of limited efficacy in C. elegans. Cultures of primary cells represent a valid strategy to partially ameliorate this problem.

      The critical steps in these experimental protocols include starting with large quantities of worms and monitoring the preparations in order to not lose eggs, embryos etc. It is also crucial to learn how to handle the animals during injection. Keeping worms in the agarose pad too long, punching them with a large pipette or injecting too much DNA mix can irreversibly damage them. Transformation efficiency, reproducibility and transgene expression may vary. Efficiency is related to the purity of the injection mix and its composition (presence of non-coding DNA). Extrachromosomal arrays vary from animal to animal therefore it is critical to establish at least 2-3, independent lines per transgene. On the other hand, the addition of digested genomic DNA to the mix represents a valid strategy to reduce transgenic over-expression. Chemotaxis assays are relatively trouble-free. However it is crucial to maintain consistency in the concentration gradient in order to avoid false results. Use pieces of agar of the same size-for example cut them with a Pasteur pipette-and maintain the equilibration time constant. If any, remove excess solution with a cotton swab.

      In conclusion, here we provide an example of how a simple, genetically tractable organism can be exploited to investigate molecular aspects of amylopathies. The same experimental techniques could be applied to the study of other neuronal proteins and to the generation of new animal models of disease.

      Subscription Required. Please recommend JoVE to your librarian.

      Disclosures

      The authors declare that they have no competing financial interests.

      Acknowledgments

      We thank Dr. Shuang Liu for critical reading of the manuscript. The PCL12 construct was a gift form Dr. Christopher D. Link. This work was supported by two National Science Foundation grants (0842708 and 1026958) and an AHA grant (09GRNT2250529) to FS.


      Materials and Methods

      Subjects

      All of the measurements were made on 4-d-old adult C. elegans. A total of 2814 C. elegans were used in these experiments. Of these, 2301 were wild-type N2. To assess the role of chemosensory cues in the effects of isolation, 40 osm-6(p811) worms were used. In this strain, the mutation results in disrupted structure of sensory cilia that causes defects in all types of chemosensory behaviors (Dusenbery and Barr, 1980 Peckol et al., 1999). To assess the role of glr-1 in the effects of isolation 124 of glr-1(n2461) and 157 of the glr-1, rescue strain kp537 was used (in these experiments, n values were high because the rescue strain is not integrated and responses were more variable than in wild-type worms). To assess the role of mec-4(e1611) and egl-4(n477) in the effects of isolation, 37 mec-4 worms and 42 egl-4 worms were reared in either isolation or colonies and tested. For confocal imaging, two genetically engineered strains were used: GLR-1::GFP and pmec-7::SNB-1::GFP. GLR-1::GFP worms (n = 65) contain chimeric receptors made up of GLR-1 (a homolog of a kainate/AMPA-type glutamate receptor expressed on the interneurons of the tap withdrawal circuit) tagged with GFP, which has been used to visualize synapses in C. elegans (Rongo and Kaplan, 1999). In pmec-7::SNB-1::GFP (n = 48), we looked at GFP expression of the gene for C. elegans synaptobrevin (snb-1), a protein associated with synaptic vesicles that plays a role in regulating vesicle fusion at the synaptic terminal. SNB-1::GFP was expressed under the control of the mec-7 promoter, which targets this GFP expression to the six mechanosensory neurons of the tap withdrawal circuit (Nonet, 1999).

      Procedure and apparatus

      All plates were 60 mm in diameter and 15 mm deep, filled with 10 ml of nematode growth medium agar, and streaked with one to two drops of E. coli (strain OP50) to produce a circular lawn ∼10 mm in diameter (Brenner, 1974). For behavioral testing, a platinum wire pick was used to transfer adult worms (96-100 h after eggs were laid) from the plates where they were grown to a test plate without any E. coli.

      The test plate, in a plate holder attached to a micromanipulator, rested on the stage of the stereomicroscope (Wild M3Z Leica, Nussloch, Germany). A video camera (Digital 5100 Panasonic, Secaucus, NJ) was attached to the stereomicroscope and also to a video cassette recorder (Panasonic AG1960) and television monitor (PM-1271A NEC, Tokyo, Japan). The date and time were displayed on the monitor via a time-date generator (Panasonic WJ-810) and videotaped. After 2 unrecorded minutes, the stage lamp was turned on, and a single tap was delivered to the side of the plate via a copper rod driven by a Grass Instruments (Quincy, MA) S88 stimulator.

      Rearing condition

      Colony-raised worms were produced by allowing five adult hermaphrodites to lay between 30 and 40 eggs on a Petri plate with E. coli, then removing the adults so that only the eggs remained. Isolate-raised plates were produced by allowing one adult hermaphrodite to lay one egg on a Petri plate with E. coli, then removing the adult and any extra eggs so that only one egg remained. To protect worms from extraneous stimulation, all plates were then placed in foam- or cotton-lined boxes and placed in a 20 ± 0.2°C incubator. Plates were not touched until testing (96-100 h later).

      Conditioned plates

      In one experiment, isolate and colony worms were raised on plates that had contained colonies of worms for 4 d before the experiment. All of these “conditioning” worms were removed from the plates before the egg-laying for the experiment. This was done to provide the chemical stimuli of groups of worms for the isolate-raised worms.

      Larval stimulation

      In experiments with a stimulated isolate-raised condition, worms were treated as described above for the isolate-raised and colony-raised conditions however, in the third larval stage of development (“L3” ∼36 h of age), they were given a brief period of mechanosensory stimulation. (L3 was chosen as the approximate mid-point in larval development. In ongoing studies, we are investigating whether there is a critical period for rescue of the tap response. However, as mentioned, the isolate effect on behavior is very fragile and is easily disrupted by extraneous stimulation in the incubator at any time during the 4 d period, and we are hypothesizing that there will not be a critical period for the rescue of the behavior.) To deliver the same number of stimulations to all worms at exactly the same age, a cardboard box containing plates of L3 isolate and colony worms was removed from the incubator, and 30 mechanosensory stimuli were delivered to all plates simultaneously by dropping the box onto a table from a height of 5 cm 30 times at a 10 s interstimulus interval (ISI). After this stimulation, the box containing the stimulated isolate-raised and stimulated colony-raised worms was returned to the incubator until testing at 4 d of age.

      Scoring and analysis

      Responses to tap were videotaped and later scored using stop-frame video analysis. For each worm videotaped, a tracing was made of the length and width of the worm. Worm length was calculated by measuring the worm from the tip of the head to the tip of the tail. Worm width was calculated by measuring the width of the worm at the midpoint of its body length. A “reversal” was defined as a worm swimming backward in response to the tap within 1 s after the tap.

      Reversals were traced onto acetate sheets and later scanned (ScanJet 3c Hewlett-Packard, Palo Alto, CA) into a Macintosh (Performa 6200CD Apple Computers, Cupertino, CA) computer using DeskScan II software. Tracings were measured using NIH Image software, and all data analyses were performed in Statview 4.5 (Abacus Concepts, Berkeley, CA). Because differences were found in the length of worms raised under different conditions, all of the response to tap data were reported standardized to a percentage of worm length by dividing the distance each worm reversed by that worm's length and multiplying by 100.

      For spontaneous reversals, 52 colony and 44 isolate worms were each placed alone on agar-filled Petri plates and filmed for 6 min. Backward movements of the body during the 6 min period were counted as spontaneous reversals and scored in the same manner as reversals.

      Because the magnitude of the response to tap is affected by a number of factors including temperature, humidity, and age of the agar plates used, a separate N2 group was run on the same day and under the same conditions as the experimental groups for every experiment. Across experiments, there was some variability in the raw scores of the N2 worms as a result of changes in environmental conditions, but the effect of isolation on the response was consistent (isolates significantly smaller than colony). Because we found that the colony/isolate effect was easily disrupted (if the worms experienced even very low levels of vibration during development the effect was lost), we depended on the N2 worms as an indicator of this disruption. If the N2 worms did not show the colony/isolate effect (resulting from a variety of factors including the incubator being bumped, the plates being contaminated, or the worms being sick), the data from that experiment were thrown out, and the experiment was rerun.

      When doing statistical analysis on mutant strains of worms (see Figs. 1a, 3a-d), comparisons were only made within a strain to directly test our experimental manipulation (i.e., colony vs isolate) and not between strains. We did not make statistical comparisons between strains, because mutations are made on a large number of different genetic backgrounds and may differ from our wild-type worms in a number of ways. Therefore, differences between strains in baseline measures were not relevant for this paper.

      Effects of isolation on tap withdrawal response. a, Mean tap response magnitude (Resp. Magnitude) for wild-type (N2) worms and a chemosensory mutant (osm-6). The isolate-raised worms of both strains showed significantly smaller responses to tap than did the colony-raised worms (n = 20 per group). b, Effects of isolation on reversals to heat. Mean reversal magnitude to tap and to heat probe for isolate-raised and colony-raised worms (n = 20 per group). Isolate-raised worms had significantly smaller responses to tap than colony worms. There were no differences between the isolate-raised and colony-raised worms' responses to the heat probe. c, Effects of stimulation on isolated worms. Mean tap response magnitude for colony-raised, isolate-raised, L3 stimulated colony-raised worms, and L3 stimulated isolate-raised worms (L3 worms were raised in colony or isolation but were given brief stimulation during the third larval stage). Brief mechanosensory stimulation in L3 reversed the effects of isolation on behavior. Stimulated isolate-raised responses to tap were larger than the isolate-raised worms and not different from colony-raised or stimulated-colony raised worms. The asterisks represent statistically significant pairs. Error bars represent SEM. WL, Worm length.

      Heat probe experiment

      A heated probe (a scalpel blade heated in an alcohol flame until it glowed red) was moved manually into the path of a worm during forward locomotion so that the probe was perpendicular to the body axis of the worm and contacted neither the worm nor the agar surface. The reversal response to the probe was filmed and scored in the same way as the reversals to tap.

      Egg counting experiment

      Colony and isolate reared plates were set up as described above. Plates were taken from the incubator at either 66 or 70 h of age, and the number of worms (for colony plates) and the number of eggs per plate were counted and recorded using a manual counter.

      Confocal GFP imaging

      Worms were mounted onto 14 × 14 mm three-square slides (one worm per slide Erie Scientific, Erie, PA) using 12 μl of 2,3-butanedione monoxime for paralysis mixed with Sephadex beads (G-50 medium) to prevent worms from being crushed on the microscope stage during imaging. Images were obtained using a Nikon (Tokyo, Japan) Optiphot-2 microscope with an MRC 600 confocal system (Bio-Rad, Hercules, CA) equipped with a krypton/argon laser. GFP was excited using a 488 nm wavelength laser setting with the emitted light collected by passing through a ∼510-550 nm bandpass filter. The images collected from the MRC 600 were captured in a 768 × 512 pixel field of view with the optical sections collected at 0.5 μm intervals. Worms were imaged using a 60× magnification oil lens.

      GLR-1::GFP was expressed along the ventral cord. Therefore, images were collected along the posterior portion of the ventral cord from the tail to the vulva. Images of GLR-1::GFP expression were composed of 10-15 optical sections for each ventral nerve cord segment. Each resultant stack of images was then compiled into a single projection image. These projection images were used for the analyses. The GFP expression in the ventral nerve cord is uniform in thickness. Therefore, the length of GFP expression was measured (Rose et al., 2003).

      GFP expressed in the pmec-7::SNB-1::GFP strain was captured in a single image stack each composed of ∼12-18 optical sections. Because the GFP in these worms was quite faint, the microscope was set to maximal sensitivity for all worms, and intensity was not measured. Worms with no detectable GFP were discarded (∼13.5% of the worms distributed evenly across groups). Because the SNB-1::GFP appeared as one to three clustered bright spots, area measurements were used to quantify GFP expression (Rose et al., 2003).

      Analysis of confocal images GLR-1::GFP.

      Collected projection images were coded and viewed in NIH Image 1.61. A researcher blind to the condition of the worm measured the number and length of GFP expression in the image. Measurements were entered in Statview 4.5 for statistical analysis.

      pmec-7::SNB-1::GFP. Projection files were coded before analysis so that all measurements were made by a researcher blind to the condition of the worm. Single projection images were opened in NIH Image (as above) and viewed as a binary image with the threshold adjusted to allow for measuring of faint images. Area measurements for each region of GFP expression were calculated by outlining the edge of the region and using the area measure function. Area measurements were entered into Statview 4.5 for statistical analysis.

      Final figures for GFP images were generated using Adobe Photoshop 7.0 (Adobe Systems, San Jose, CA).

      Statistical analysis

      In experiments in which only two groups were run, unpaired t tests were used to determine significance. If multiple groups were run, an ANOVA with Fisher's planned least significant difference (PLSD) planned comparisons were used. For all experiments, α was set to p ≤ 0.05.


      The authors would like to thank Ben King, PhD, and Joel Graber, PhD, for bioinformatics advice and George Sutphin, PhD, for constructive criticism of the manuscript. This work was supported by grants from the National Institute on Aging of the National Institutes of Health (R21AG056743) and by the Ellison Medical Foundation (AG-NS-1087-13), both to AN Rogers. In addition, this project was supported by Institutional Development Awards (IDeA) from the National Institute of General Medical Sciences of the National Institutes of Health (grant numbers P20GM0103423 and P20GM104318, respectively). Some strains were provided by the CGC, which is funded by the NIH Office of Research Infrastructure Programs (P40 OD010440).

      Author Contributions

      JA Rollins: data curation, formal analysis, methodology, writing—original draft, writing—review and editing.

      P Kapahi: conceptualization.

      AN Rogers: conceptualization, funding acquisition, writing—review and editing.


      Discussion

      C. elegans dietary choice behavior

      Here, we describe novel behavioral paradigms in C. elegans, food seeking and food preference. We have identified five worm bacterial foods to establish a range of food quality as measured by the food's ability to support growth (see Fig. S1 in supplementary material). Remarkably, an animal as simple as C. elegans can exhibit dietary choice(Fig. 1). Worms preferred high quality food, i.e. that better supported growth. This choice developed with time, suggesting that animals needed to try the food to make a decision. Using eat mutants, we showed that this choice requires food assessment via feeding. This is similar to mammals, which are also capable of selecting food that better supports growth, and try foods before making a choice(Osborne and Mendel, 1918 Young, 1941).

      AIY neurons function to extend food-seeking periods. Trajectory on mediocre food, E. coli DA837, of (A) a ttx-3 mutant and (B) an animal whose AIY neurons have been killed. Compare to wild-type in Fig. 4C. ttx-3 mutant trajectories did not span the whole lawn and there were far fewer long straight roaming events. Trajectories of AIY worms also had fewer straight long movements than wild-type controls. (C,D) Movement duration distribution of (C) wild type, ttx-3, osm-6, osm-6ttx-3 and (D) AIY-ablated animals, all tested on E. coli DA837 food. N=10 for WT, 10 for ttx-3, 6 for osm-6, 6 for osm-6ttx-3, 10 for AIY ablations and 8 for ttx-3p::GFP controls. (E) ttx-3 was defective in the food preference behavior if bacterial foods were located at a small distance from each other. By 3 h, all ttx-3 worms found food,but there was no preference in the harder arrangement. In contrast to ttx-3, osm-6 animals took longer to discriminate between good and bad food, but they finally managed to make the right choice even if foods were located at a distance. Values are means ± s.e.m. * Different from the wild type (P<0.01) † different from ttx-3 (P<0.01 Student's t-test). (F) Biased food preference for E. coli HB101 over B. megaterium of mutants and animals with laser-ablated neurons. The fraction of animals that reached the central colony of good food, E. coli HB101, was determined. ttx-3 mutants and AIY-ablated animals performed worse than controls. In laser ablation experiments, worms were counted after 20 h. For tests on mutants, the number of assays is 18 for WT, 15-17 for ttx-3 alleles and 6-15 for various mutants tested. For laser ablations, number of worms found in the center and the total number of worms tested is indicated next to the bars. Values are means ± s.e.m. * Different from the wild type (P<0.01 Student's t-test) † Different from the ablation control(P<0.01 χ 2 test of independence).

      AIY neurons function to extend food-seeking periods. Trajectory on mediocre food, E. coli DA837, of (A) a ttx-3 mutant and (B) an animal whose AIY neurons have been killed. Compare to wild-type in Fig. 4C. ttx-3 mutant trajectories did not span the whole lawn and there were far fewer long straight roaming events. Trajectories of AIY worms also had fewer straight long movements than wild-type controls. (C,D) Movement duration distribution of (C) wild type, ttx-3, osm-6, osm-6ttx-3 and (D) AIY-ablated animals, all tested on E. coli DA837 food. N=10 for WT, 10 for ttx-3, 6 for osm-6, 6 for osm-6ttx-3, 10 for AIY ablations and 8 for ttx-3p::GFP controls. (E) ttx-3 was defective in the food preference behavior if bacterial foods were located at a small distance from each other. By 3 h, all ttx-3 worms found food,but there was no preference in the harder arrangement. In contrast to ttx-3, osm-6 animals took longer to discriminate between good and bad food, but they finally managed to make the right choice even if foods were located at a distance. Values are means ± s.e.m. * Different from the wild type (P<0.01) † different from ttx-3 (P<0.01 Student's t-test). (F) Biased food preference for E. coli HB101 over B. megaterium of mutants and animals with laser-ablated neurons. The fraction of animals that reached the central colony of good food, E. coli HB101, was determined. ttx-3 mutants and AIY-ablated animals performed worse than controls. In laser ablation experiments, worms were counted after 20 h. For tests on mutants, the number of assays is 18 for WT, 15-17 for ttx-3 alleles and 6-15 for various mutants tested. For laser ablations, number of worms found in the center and the total number of worms tested is indicated next to the bars. Values are means ± s.e.m. * Different from the wild type (P<0.01 Student's t-test) † Different from the ablation control(P<0.01 χ 2 test of independence).

      Next, we showed that C. elegans left hard-to-eat bacteria(Fig. 2), and, like food preference, this behavior required food quality assessment. Previously, it was generally thought that once C. elegans finds food, it stays there and eats until death or until the source is exhausted, although Lipton et al.(2004) have found recently that adult males leave food in search of hermaphrodites. Leaving experiments showed that even after food is found, the animal could decide to stay in the food or leave, and this decision was based on the assessment of whether the food was good or bad. Leaving behavior is a compromise: on the one hand, the worm risks losing the food that has already been found and ending up in an adverse environment, or, on the other hand, there is a chance of finding even better food.

      Leaving behavior that we describe here is somewhat related to the known phenomenon of adaptation to a volatile or soluble attractant. Upon extended exposure to an odor (typically, 1 h or more) in the absence of food, odortaxis to this odor dwindles (Colbert and Bargmann, 1995 Colbert and Bargmann, 1997). Likewise, attraction to a soluble chemical switches to avoidance after 3-4 h exposure(Saeki et al., 2001) in the absence of food. In view of our results, adaptation is an increase of a food-seeking behavior because of the lack of reinforcement. Consistent with this, Nuttley et al. (2002)have shown that in the presence of food, chemoattraction is suppressed. Also,aerotaxis fades in the presence of food(Gray et al., 2004). And, if animals are conditioned to an odor or taste in the presence of food, no adaptation occurs. If worms are adapted to the stimulus in the absence of food, but then briefly exposed to food, chemoattraction robustly revives(Nuttley et al., 2002 Saeki et al., 2001). These and our data suggest the food feeds back on behavior after it is eaten and acts as a reinforcer in C. elegans.

      Effect of previous dietary experience

      If bacterial food was switched from good to mediocre, C. elegansappetitive behavior was increased compared to animals that had not experienced good food (Fig. 3). Previous experience of good food made worms more risk-loving, more willing to explore.

      In the leaving experiment (Fig. 3C), the time course of the effect of experience could be observed: the enhancement of leaving was not high at the very beginning, but reached a maximum after 0.5-1 h. This indicates that time was needed to assess new conditions, followed by comparison and output. If a worm is taken off food, there is a period of about half an hour of area-restricted search with frequent reversals and turns, followed by active `running', when reversals and turns are suppressed (Gray et al.,2005 Hills et al.,2004). This time, however, is much shorter than that required to deplete fat stores, which is about 6 h(McKay et al., 2003). Probably for worms, which feed continuously throughout life to support a 3-day life cycle, even brief food deprivation or brief decline in food quality is an alert signal that motivates them to explore the environment.

      One might propose a trivial explanation for the results in Fig. 3: well-fed worms are simply healthier and explore the environment more actively than unfed ones,which strive to save energy. We think this is unlikely, because in the continuous presence of good food, both leaving behavior(Fig. 2) and exploratory activity (Fig. 4) were suppressed, while in the continuous presence of poor food, worms were very active. Therefore, it was the switch from good to bad that causes an increase in exploratory behavior.

      At least two other explanations are possible. First, experiencing different quality foods changes worms' satiation (or hunger) status, and that, in turn,affects their food choice and leaving behaviors. More hungry animals tend to accept any first food they encounter, thus their leaving behavior and food choice is less pronounced. Less hungry (more satiated) animals, on the contrary, tend to be more particular about food and their exploratory behavior is more active. Another explanation invokes memory: worms may learn that previous conditions were associated with a different satiation status and compare those with the new conditions. These two mechanisms are not mutually exclusive and may function in parallel.

      In humans, powerful diet preferences can form, especially for nutritious fat- and carbohydrate-rich foods, such as sodas, desserts, pizza, etc. These feeding habits are hard to change. Dieting often results in `food craving',`carbohydrate craving', and `binge eating', which eventually lead to even further increase of food consumption(Capaldi, 1996). These phenotypes are analogous to the increased food-seeking behavior when the food is switched from good to bad in C. elegans. (Of course, the time scale has to be normalized to the life span.) These behaviors are adaptive in the wild, where good food is usually scarce, but in developed countries, where high quality food is always easily available, they contribute to overeating and obesity.

      The C. elegans food-seeking strategy

      C. elegans locomotion, in particular the equilibrium between roaming, rapid straight movement, and dwelling, slow movement with frequent reversals and stops, was affected by the food source(Fig. 4). On poor food,straight rapid movement, called roaming, was drastically increased, while dwelling predominated on high quality food.

      In previous studies, it has been shown that the speed of C. elegans locomotion is increased in the absence of food, and reversals and turns are suppressed after about 30 min in the absence of food(Gray et al., 2005 Hills et al., 2004). Here, we showed that this also happened on food, if the food was hard to eat. This suggests that it is not the mere presence of food that is decisive in regulating C. elegans locomotion, but food quality.

      We identified mutants defective in food preference behavior and two of them, osm-6 and ttx-3 were also defective in roaming. TTX-3 is a transcription factor required for the differentiation of the AIY thermosensory interneuron. The defects of the ttx-3 mutant were partially reproduced by killing AIY. The latter also caused a decrease in duration of roaming periods, suggesting that AIY functions to suppress the roaming-to-dwelling transition and to extend the food-seeking periods. This is consistent with, and extends, the results of other reports, which demonstrated that AIY suppresses reversals and turns(Gray et al., 2005 Tsalik and Hobert, 2003).


      Robustness of Optimization Results to Small Variations of Parameters

      To determine the robustness of the wire-minimized solution, we explored several aspects of the cost function and assessed their impact on the ability to predict neuronal layout.

      First, we analyze the sensitivity of the wire-minimized layout to the normalization coefficient α and the exponent ζ. As mentioned, our cost formulation accounts for multiple synapses on a given neurite by normalizing connection weights by the average number of synapses per neurite (α = 29.3). We test how the predicted layout changes by varying α between 1 and 45. Because the choice of the quadratic form of the cost function may seem arbitrary, we also varied the power of wire length in the cost function, ζ in Eqs. 2 and 3 between values of 1 and 4. As argued previously, the wiring cost is likely to scale supralinearly (ζ > 1) with distance between neurons (26). If so, the minimization problem is convex and can be efficiently solved numerically. The lowest mean deviation, 9.71%, is achieved by using the cost function with normalization coefficient ≈27 and exponent ≈2 (see General Power-Law Cost Function in Supporting Text and Fig. 6, which are published as supporting information on the PNAS web site). Interestingly, these values are close to those chosen from biological considerations and validate the quadratic cost function.

      Second, we test the importance of synaptic multiplicity between neurons. Instead of a wire dedicated to each synapse between cells (Fig. 1 C Inset), we use a single wire to connect a given pair of neurons regardless of the number of synapses (Fig. 1 D Inset). In other words, we minimize the quadratic cost function with a binary connection matrix (only 0 or 1 elements in the matrix A from Eq. 2 ). Using ζ = 2, the lowest mean deviation between predicted and actual position (9.82%) is higher than the result from a synapse-number weighted cost function and was found at α = 8. In the actual worm, the average number of synaptic partners (as opposed to individual synapses) per neurite is 12.2, close to the optimal value of α obtained from the binary connection matrix.

      To summarize, we find that various reasonable cost functions predict neuronal placement incomparably better than the random one. Although mean deviations vary somewhat between different cost functions, they are not far from the best known solution. Thus the wire length minimization approach is rather robust. Because the quadratic cost function can be solved exactly and is reasonably close to the best-known solution, it may serve as the reference predicted layout. Although the predicted placement is only approximately correct, we recall that the problem was solved in one dimension. Such dimensionality reduction may introduce errors on the order of the inverse aspect ratio of the worm, just under 10%. Because the mean deviations we report approach this range, wiring optimization results are encouraging.


      Results

      Alternative splicing and promoter of the nfi-1 gene

      The C. elegans nuclear factor I gene (nfi-1) was first identified by the C. elegans sequencing consortium by its homology to mammalian NFI genes [14]. Only a single NFI gene is present in the C. elegans genome (Fig. ​ (Fig.1A) 1A ) while vertebrates possess 4 NFI genes. To confirm the structure of the nfi-1 gene, we obtained cDNAs from the Kohara and TIGR libraries and made primers within predicted exons for production of cDNAs from RNA of adult worms and purified eggs. These cDNAs confirmed the presence of 14 exons and showed alternative splicing of exons 2 and 13 (Fig. ​ (Fig.1A). 1A ). The differential splicing pattern of nfi-1 indicates that different isoforms may be expressed in different cell types. Primary transcripts from most C. elegans genes are trans-spliced to SL1 leader sequences and detection of SL1-linked transcripts is frequently used to assess the initially transcribed exons of genes [37]. We confirmed the start site of the gene using PCR with an SL1 primer for SL1-containing transcripts and a primer in exon 2.

      A) Alternatively spliced products of the nfi-1 gene. The nfi-1 gene is shown as a line with exons as numbered solid boxes and alternatively spliced exons as gray boxes. Arrows show the direction of transcription. Below the line are cDNAs from the Kohara and TIGR libraries the vertical bar indicates the 5' end of the cDNA. In yK213C10 the letters a and b above exon 2 denote that it is alternatively spliced generating 2a and 2b and the arrow on the right denote undetermined sequence in the cDNA. The asterisk (*) on yk42f10 denotes an alternative 3'splice acceptor site in exon 13 used in yk42f10 but not in CEESQ09. Below are depictions of cDNAs obtained by RT-PCR from total and polyA+ RNA of whole worms or isolated eggs using nfi-1 exon-specific primers (RT-PCR nfi-1 primers). Lastly we show cDNA obtained by PCR using SL1 primer and nfi-1 exon-specific primers (RT-PCR SL1-linked products). Some of the alternatively spliced cDNAs have been described previously [14]. B) Comparison of CeNFI-DBD with the consensus mouse NFI DBD.The sequence of the CeNFI-DBD (top line) is aligned to a consensus sequence from the 4 mouse NFI-DBDs (bottom line). Gaps in the mouse consensus indicate residues that are not identical between the 4 mouse NFI-DBDs. The dash in the 6 th aligned row indicates a single insertion in the CeNFI-DBD sequence needed to align it with the mouse consensus. Dark gray boxes show identical residues in CeNFI-DBD and mouse NFI-DBDs, light gray boxes show residues not identical but similar between CeNFI and mouse NFI-DBDs, unboxed residues with black letters show residues that are not similar between the CeNFI and mouse DBDs, and gray letters in the CeNFI-DBD above gaps in the mouse consensus indicate positions where the 4 mouse genes are not identical. 151 of 190 residues are identical in all 4 mouse NFI-DBDs while 60 of these 151 residues are different in the CeNFI-DBD and 27 of these 60 differences are non-conservative substitutions. The alignment was done in Macvector 6.5.3 using the ClustalW similarity matrix.

      DNA binding by CeNFI is indistinguishable from that of human NFI-C

      The predicted DNA-binding domain of nfi-1 (CeNFI-DBD) shares homology with the vertebrate NFI proteins, however 60 of 151 residues that are completely conserved among the 4 mouse NFI proteins are changed in CeNFI (Fig. ​ (Fig.1B). 1B ). To assess the DNA-binding activity and specificity of CeNFI, the DNA-binding domain (DBD) of CeNFI was cloned in frame with a 6 histidine-tag, expressed in E. coli, partially purified by nickel-affinity chromatography, and was used for in vitro DNA binding assays. The DNA-binding activity of CeNFI-H6 was indistinguishable from that of the human hNFI-C220H6, with both proteins binding the NFI-site oligo 2.6 (Fig. ​ (Fig.2A) 2A ) but not the C2 oligo containing a point mutation that prevents vertebrate NFI binding [19,38]. To ask if wild-type C.elegans contains a protein with similar DNA binding properties as CeNFI-H6, extracts were prepared from a mixed population of worms and tested for binding to the 2.6 and C2 sites. As expected, proteins were detected that bound to the 2.6 but not the C2 site, confirming the binding specificity of CeNFI (Fig. ​ (Fig.2B). 2B ). The specificity of the native and recombinant CeNFI proteins was also measured by competition of binding of the 2.6 oligo with various unlabeled oligonucleotides and was indistinguishable from the specificity of hNFI-C220-H6 (data not shown). Thus despite the differences between CeNFI and vertebrate NFIs in conserved residues of their DNA-binding domains, their DNA-binding activities are indistinguishable.

      A) Specific NFI DNA-binding activity of recombinant CeNFI-H6. Partially purified recombinant H6-tagged CeNFI (containing parts of exons 2𠄸) and human NFI-C220 were incubated with a duplex oligonucleotide containing an NFI binding site (2.6, even lanes) or an oligo with a single point mutation that abolishes NFI binding (C2, odd lanes) and analyzed on a 6.5% non-denaturing polyacrylamide gel. Lanes 1 and 2, crude E. coli extract (neg. control) lanes 3 and 4,

      5 ng partially purified CeNFI-H6 lanes 5 and 6,

      40 ng partially purified CeNFI-H6 lanes 7 and 8,

      5 ng purified human NFI-C220H6. See bottom of panel for sequences of oligonucleotides. B) Specific NFI DNA-binding activity in worm extracts. Nuclear extracts of a mixed population of C. elegans were prepared and used in a gel mobility shift assay with an oligonucleotide that contains an NFI-binding site (lanes 1 & 2, 2.6) or the same oligo with a single point mutation that abolishes NFI binding (lane 3, C2). Lane 1, no extract Lanes 2 and 3, C. elegans extract (

      10 μg). See A for sequences of oligonucleotides.

      Expression of nfi-1 in embryo and adults

      In the mouse, the 4 NFI genes are expressed in a complex overlapping pattern during embryogenesis and in adult tissues [13]. To assess the expression pattern of nfi-1 transcripts in C. elegans, a digoxin-labeled antisense probe from the 3' end of the nfi-1 transcript was used for in-situ hybridization to fixed embryos and whole worms [39]. nfi-1 transcripts are present in the one-cell (not shown) and two-cell stages (Fig. ​ (Fig.3a) 3a ) prior to the onset of zygotic transcription [40,41], indicating that the transcript is maternally inherited. Expression continues in most cells of the embryo throughout early and mid-embryogenesis (Fig. 3b𠄽 ) but decreases after gastrulation and no expression is seen in L1 larvae (Fig. 3f–i and data not shown). As expected for a maternally inherited transcript, expression of nfi-1 reappears in the adult gonad (Fig. 3j–m ). Expression of nfi-1 transcripts are also seen in the cytoplasm of gut cells (Fig. 3j–m ). No signal is seen using control sense probes (Fig. 3n–o and data not shown). This in situ expression pattern indicates that nfi-1 could function both early in embryogenesis and in adult worms.

      Expression of endogenous nfi-1 mRNA in embryo and adults. N2 worms were fixed and hybridized with digoxigenin (DIG)-labeled antisense nfi-1 probe (from plasmid yk42f10) and bound probe was detected using alkaline phosphatase-conjugated anti-DIG antibodies. Panels : a, 2 cell embryo b, 4 cell embryo c, 24 cell embryo d, beginning gastrulation e, mid-gastrulation f, late gastrulation g, comma stage h, 1.5-fold stage and i, 2-fold stage. Embryos in a-d and f-i are positive for staining. We are currently investigating the apparent loss of signal at mid-gastrulation (panel e). Panels j-m, antisense probe, adults with exposed internal organs: mature oocytes intestine gonadal germ cells n-o, control sense probe, no specific staining is seen. Right panels k, m, o are two-fold magnifications of those on the left.

      Isolation of a null allele of nfi-1

      To assess the role of nfi-1 in worms, we isolated a nfi-1 mutant using a reverse genetic approach based on the insertion-excision of the transposon Tc1 [42]. Worms carrying an

      2 kb deletion in nfi-1 were isolated by PCR screening and sib-selection (Fig. 4A,B ). Sequence analysis of the nfi-1 transposon excision allele (designated nfi-1(qa524)) showed loss of the genomic region corresponding to nucleotides 14754� of cosmid ZK1290. This eliminates the first 6 exons of nfi-1 including sequences encoding the DNA-binding domain. RT-PCR shows the absence of nfi-1 transcripts in mutant worms (Fig. ​ (Fig.4C). 4C ). A gel mobility shift assay, using nuclear extracts prepared from mix-stage populations of C. elegans wild type and nfi-1 mutant worms confirmed the loss of CeNFI DNA-binding activity in the mutant worms (Fig ​ (Fig4D). 4D ). Thus, this mutation in the nfi-1 gene is a null allele. The survival of worms homozygous for the null allele shows that nfi-1 is not essential for worm survival. The nfi-1 mutant allele was backcrossed 12 times to the wild-type N2 strain to remove unwanted mutations prior to assessment of the phenotype.

      A) Deletion in C. elegans nfi-1 gene. The relative locations of confirmed exons (boxes) are shown. Square brackets indicate the location of the NFI-1 DNA-binding domain and the region deleted in nfi-1 mutant. The arrows indicate the position of the Tc1 insertion in an intron of the nfi-1 gene in strain NL747 pk240 and locations of PCR primers used in the screening for the deletion. B) Single-worm PCR reactions on N2 worm and nfi-1 homozygous mutant isolated by sib-selection. The arrows indicate a 1007 bp and 2969 bp PCR products corresponding to the nfi-1 mutant (qa524) and wild type alleles respectively. Lane 2 is 1 kb DNA ladder. C) RT-PCR on N2 worms and nfi-1 homozygous mutants. The arrows indicate a 480 bp and 320 bp RT-PCR products amplified using total RNA obtained from N2 worms (lane 1). nfi-1 mutants show loss of nfi-1 transcripts (lane 2). Lane (3) is 100 bp DNA ladder. D) Loss of NFI DNA-binding activity in extracts of nfi-1 mutants. Nuclear extracts of a mixed population of N2 worms and nfi-1 mutants were prepared and used in a gel mobility shift assay with an oligonucleotide (2.6) that contains an NFI-binding site (lanes 3, 5) or the same oligo with a single point mutation that abolishes NFI binding (C2) (lane 4, 6). See Fig. 2A for sequences of oligonucleotides. Extract of nfi-1 mutants show loss of NFI DNA-binding activity (lanes 5, 6). Lanes 1 & 2, no extract.

      Phenotype of nfi-1 mutant worms

      Locomotion defect

      Loss of nfi-1 results in a body movement defect (Unc, uncoordinated). nfi-1 mutant animals are fairly active and healthy, but are sluggish and flaccid at rest and slightly longer and thinner than N2 worms (Fig. 5A,D ). While wild type worms usually move in straight long lines, nfi-1 mutant worms often change direction abruptly (Fig. 5B,E ). The nfi-1mutant worms produce less regular tracks on the bacterial lawn with higher amplitude, and sometimes are slightly coiled when compared to wild-type worms (Fig. 5C,F ). This phenotype is more severe in older adults.

      Locomotion in nfi-1 mutants. Single young adults were spotted in the center of fresh plates and left for 10 min. (A, D) Photographs of N2 worms and nfi-1 mutants (B, E) Track patterns of N2 worms and nfi-1 mutants (C, F) Track patterns of N2 worms and nfi-1 mutants with higher magnification. Note less regular tracks in nfi-1 mutant vs. N2 worms.

      Egg-laying defect (Egl)

      17�% of older nfi-1 mutants have a "bag of worms" phenotype, where the mother is unable to lay fertilized eggs and fills with hatched progeny (Fig. ​ (Fig.6A). 6A ). Appearance and severity of this egg-laying phenotype correlates with the progressive locomotion defect, as young adult nfi-1 mutant worms do not bag. In young adults, serotonin stimulated egg-laying in both the wild type and nfi-1 mutant animals to a similar extent (data not shown), indicating that the postsynaptic response to serotonin is normal and the contractile apparatus for egg-laying is intact in nfi-1 mutant worms.

      A) Egg-laying defect in nfi-1 mutants and transgenic rescue. Bagging was measured in wild-type N2, nfi-1, N2 worms carrying the transgenic array qaEx507(N2 qaEx507) and nfi-1 worms carrying this array (nfi-1 qaEx507). Bars represent % of bagging as the mean of 3𠄴 independent experiments and error bars show the standard deviation. 30� worms of each genotype were scored in each independent experiment. N2 and N2 qaEx507showed υ% bagging. The nfi-1 mutant worms showed

      30% bagging while the rescued nfi-1 qaEx507showed υ% bagging. B) Shortened life span in nfi-1 mutants and transgenic rescue. Survival curves for the strains described above N2 (n = 57), nfi-1 (n = 58), N2 qaEx507 (n = 52) and nfi-1 qaEx507 (n = 31) are shown. Kaplan-Meier analysis (SPSS11 software) was use to determine median, percentile and p values (log rank test) and Excel was used to construct survival curves. The array generated

      50% rescue of the life-span. The experiment was repeated twice with similar results.

      Life-span reduction

      The nfi-1 mutant worms have a median life span of 10.00 ± 0.69 days as compared to 14.00 ± 0.76 days for wild type worms (p < 0.001) (Fig. ​ (Fig.6B). 6B ). Mutant worms become progressively more sluggish and flaccid as they age. We are currently investigating whether this lifespan reduction is due to the apparent progressive muscle weakness and partial paralysis seen or to a direct effect on known ageing pathways [43-45].

      Pharyngeal pumping rate defect

      Since the Unc and Egl phenotypes of nfi-1 mutant worms could reflect aberrant muscle function, we examined another process that can influenced by muscle defects, pharyngeal pumping rate. Pharyngeal pumping rates are reduced in nfi-1 mutant worms (Table ​ (Table1, 1 , Days 1𠄴), with more severe reductions in older vs. younger adult animals (Table ​ (Table1, 1 , Days 3ɤ vs. Day 1).

      Table 1

      Rescue of pharyngeal pumping defect by nfi-1 transgene

      Day a N2nfi-1N2 qaEx507 b nfi-1 qaEx507 b
      1 (n = 10)241 ± 5228 ± 9 c 243 ± 5241 ± 8 e
      2 (n = 12)242 ± 10187 ± 64 d 237 ± 8195 ± 78 d
      3 (n = 12)214 ± 18125 ± 79 c 183 ± 78200 ± 72
      4 (n = 12)160 ± 6292.5 ± 86 d 183 ± 79133 ± 90

      a Pharyngeal pumping was counted for one minute starting on day one, when animals first reached adulthood, for four consecutive days.

      b N2 and nfi-1 mutant worms were transformed with a transgene containing the nfi-1 gene and upstream promoter and rol-6 (qaEx507).

      c This value is statistically significantly different from N2 (P < 0.01)

      d This value is statistically significantly different from N2 (P < 0.05)

      e This value is statistically significantly different from nfi-1 (P < 0.01)

      Rescue of nfi-1 mutant with nfi-1 transgene

      Transgenic rescue was performed to test whether loss of the nfi-1 gene was responsible for the observed phenotypes. Transgenic strain XA512 qaEx507 was made by injecting a plasmid containing a 10 kb region of genomic DNA including the nfi-1 coding region and

      4 kb of upstream promoter region together with a rol-6(gf) expressing plasmid into the gonads of N2 worms. We crossed the resulting transgenic array into nfi-1 mutant worms to produce strain XA550 nfi-1(qa524) qaEx507. Egg-laying was completely rescued in nfi-1 qaEx507 worms when compared to the nfi-1 mutant worms (Fig. ​ (Fig.6A). 6A ). In addition, the median life span in nfi-1 qaEx507 worms of 13.0 ± 0.7 days was significantly longer than the 10.0 ± 0.7 day life span of nfi-1 mutant worms (p < 0.05), but slightly less than the 14.0 ± 0.8 day life span of N2 worms (Fig. ​ (Fig.6B). 6B ). The N2 qaEx507 stain used as a control has a 14.00 ± 0.49 day median life span, identical to that in non-transgenic N2 worms. The pharyngeal pumping defect was also partially rescued by transgenic expression of nfi-1 (Table ​ (Table1, 1 , nfi-1 qaEx507 vs. nfi-1). The nfi-1 transgene had little or no effect on pumping rates in N2 worms (N2 qaEx507 vs. N2). These data provide a well-defined developmental system affected by loss of nfi-1 that can be examined for cell-autonomous or inductive roles of nfi-1. The presence of the rol-6 marker gene prevented scoring of rescue of the locomotion phenotype in nfi-1 qaEx507 worms.

      Our in situ hybridization data indicate that nfi-1 transcripts are provided maternally. To test whether maternal nfi-1 transcripts could rescue the nfi-1 Egl phenotype, nfi-1 qa524/qa524, nfi-1 qa524/+ and +/+ progeny of heterozygous nfi-1 qa524/+ parents were scored for the egg-laying defect (Fig. ​ (Fig.7). 7 ). Bagging was seen in 41% of the resulting nfi-1 qa524/ qa524 worms, in 20.7% of nfi-1 qa524/+ worms but in σ% of +/+ animals. These data indicate the absence of maternal transcript rescue and possible haploinsufficiency at the nfi-1 locus. Thus, transgenic replacement of nfi-1 yields either partial (pumping rate and lifespan) or complete (egg-laying) rescue of the phenotypes seen in the nfi-1 mutant worms whereas maternal nfi-1 transcripts are insufficient to rescue the egg-laying defect.

      Haploinsufficiency of nfi-1 locus. Egg-laying defect in progeny of nfi-1 heterozygous mutant animals are shown. Bagging was scored in all progeny of two nfi-1(qa524/+) heterozygous worms (n = 146) derived from eggs laid over 7 hours. All worms were genotyped by single-worm PCR. Bars represent % of bagging in wild type (+/+), nfi-1 heterozygous (qa524/+) and nfi-1 homozygous worms (qa524/qa524). The number of worms scored of each genotype are shown above the bars.

      Ce titin expression is reduced in nfi-1 mutant worms

      We used cDNA microarrays to identify genes whose expression is affected by loss of nfi-1. Such genes could be either direct or indirect targets of nfi-1. Poly A+ RNA was purified from wild type and nfi-1 mutant synchronized gravid adults, labeled, and used to probe DNA microarrays containing

      17,000 C. elegans genes (Stanford Microarray Database). We analyzed RNA from gravid adults because the phenotype differences between nfi-1 mutant and wild type animals are clearer in adults than at earlier stages. The nfi-1 gene was scored as the most down-regulated gene in nfi-1 mutant worms in all experiments (data not shown). Several dozen genes showed small apparent reductions or increases in levels (2𠄳 fold) in mutant adults (data not shown) but only one gene, C. elegans titin (Ce titin) showed larger changes.

      Ce titin (also known as tag-58, t emporarily a ssigned g ene 58) was predicted to be 5.7-fold lower in mutant worms by microarray analysis. Quantitative PCR confirmed that Ce titin is reduced 8� fold in adult nfi-1mutant worms (Fig. ​ (Fig.8). 8 ). To date this is the gene that shows the largest decrease in expression in nfi-1 mutant worms. A search of the Ce titin gene reveals no overabundance of NFI binding sites (data not shown). Also, a CeTPro transgene expressing a translational fusion of GFP to the 5'-end of the Ce titin gene [46] appears to be expressed at similar levels in WT and nfi-1 mutant worms (data not shown). It will be important in future studies to determine whether Ce titin is a direct or indirect target of nfi-1 and the possible role of Ce titin in the phenotypes observed.

      Down regulation of Ce titin expression in nfi-1 mutants assessed by QPCR. Bars represent fold changes in Ce titin transcript level in wild type N2 vs. nfi-1 mutant worms. RNA samples were obtained from 3 independent synchronized adult worm populations for each genotype.

      Expression of nfi-1::GFP reporter transgenes

      One limitation of in situ hybridization in C. elegans is that in older embryos and postembryonic stages it is sometimes not sensitive enough to unambiguously identify individual cells. Since we saw little or no nfi-1 expression in muscle by in situ hybridization, in an effort to develop a more sensitive assay for nfi-1 expression we constructed two GFP-reporter transgenes (Fig. ​ (Fig.9A). 9A ). In Pro1CeNFI::GFP, 4 kb of genomic DNA sequence upstream of the nfi-1 open reading frame and the sequence encoding the first four residues of the CeNFI protein was fused in frame to GFP. In Pro2CeNFI::GFP, the 4 kb promoter region and the sequence encoding the first 94 residues of CeNFI has been fused to GFP. GFP expression for both transgenes was detected in embryos (Fig. ​ (Fig.9B). 9B ). Faint GFP expression is first detected at the late gastrulation stage of embryogenesis (𾌀 cells) as a diffuse green glow throughout the embryo. By the comma stage expression is detected in many cells along the outer edge of the embryo and expression continues through embryogenesis and is detected in L1-L4 larvae in many of the same cells as in adults. Adult transgenic animals show GFP expression in muscles, neurons and intestinal cells (Fig. 9B–I ). Among the muscles, fluorescence was strongest in the pharynx and head muscles, was observed with less frequency in other body wall muscles and was seen occasionally in vulva muscles. Expression was also seen in two pairs of neurons located near the posterior bulb of the pharynx, and in several as yet unidentified tail neurons. Expression patterns in multiple transgenic lines from each reporter were similar with the exception that Pro2CeNFI::GFP expression was detected more consistently in head neurons and Pro1CeNFI::GFP was seen with higher frequency in body-wall muscles. However expression of both transgenes was mosaic, showing expression in only subsets of cells and animals in each population. Since mosaic expression of GFP was seen in transgenic strains from both arrays, transgenic array Pro1CeNFI::GFP was integrated by γ-irradiation. However similar mosaic expression was seen with the integrated array (data not shown). These data may indicate that additional elements are needed for stable regulation of nfi-1 expression and that such elements may be located further downstream in the nfi-1 genomic sequence.

      Expression pattern of the nfi-1::GFP reporter transgenes. The nfi-1 locus and structure of the nfi-1-GFP fusion constructs are shown (A). Nfi-1 coding regions are shown in black, gray boxes indicate alternatively spliced exons, untranslated regions are in white. GFP is shown as a hatched box. Expression of nfi-1-GFP reporter constructs was observed in embryos (B), intestinal cells (C), body wall muscles (D, E), pharynx (F), egg-laying muscles (G), several head (H) and tail neurons (I). Expression was assessed using a FXA Nikon microscope (B-D, F-I) and a Bio-Rad confocal microscope (E).


      Neural Circuits of Sexual Behavior in Caenorhabditis elegans

      The recently determined connectome of the Caenorhabditis elegans adult male, together with the known connectome of the hermaphrodite, opens up the possibility for a comprehensive description of sexual dimorphism in this species and the identification and study of the neural circuits underlying sexual behaviors. The C. elegans nervous system consists of 294 neurons shared by both sexes plus neurons unique to each sex, 8 in the hermaphrodite and 91 in the male. The sex-specific neurons are well integrated within the remainder of the nervous system in the male, 16% of the input to the shared component comes from male-specific neurons. Although sex-specific neurons are involved primarily, but not exclusively, in controlling sex-unique behavior—egg-laying in the hermaphrodite and copulation in the male—these neurons act together with shared neurons to make navigational choices that optimize reproductive success. Sex differences in general behaviors are underlain by considerable dimorphism within the shared component of the nervous system itself, including dimorphism in synaptic connectivity.


      Materials and methods

      C. elegans strains were grown and maintained under standard conditions (Brenner, 1974). A complete listing of all strains used in this study and their genotypes is located in Supplementary file 1.

      Molecular biology and transgenesis

      cDNA corresponding to the entire coding sequences of unc-31 (isoform a), daf-2 (isoform a), age-1 (isoform a), tom-1 (isoform a), and the ins-1 genomic region were amplified by PCR and expressed under cell-selective promoters. unc-17 cDNA was synthesized (GenScript) and expressed under a cell-selective promoter. For cha-1 and cho-1 knockdown experiments, 1 kb fragments corresponding to exons 3–7 and the 3′ end of the gene, respectively, in the sense and antisense orientation were synthesized (GenScript). Neuron-selective RNAi transgenes were created as previously described by co-injection of equal concentrations of sense and antisense oriented gene fragments driven by cell-specific promoters (Esposito et al., 2007). Cell-specific expression was achieved using the following promoters: ceh-36deletion or odr-3 for both AWC, str-2 for AWC ON , srsx-3 for AWC OFF , gpa-4 for AWA and ASI, gpa-4deletion for AWA, gcy-7 for ASEL, gcy-5 for ASER, str-1 for AWB, sre-1 for ADL, srh-142 for ADF, gcy-8 for AFD, ops-1 for ASG, sra-6 for ASH, trx-1 for ASJ and sra-9 for ASK. For all experiments, a splice leader (SL2) fused to a mCherry or gfp transgene was used to confirm cell-specific expression of the gene of interest.

      Germline transformations were performed by microinjection of plasmids (Mello and Fire, 1995) at concentrations between 25 and 200 ng/μl with 10 ng/μl of unc-122::rfp, unc-122::gfp or elt-2::gfp as co-injection markers. For rescue and OE experiments, DNA was injected into mutant or wild-type C. elegans carrying GCaMP arrays.

      Calcium imaging

      Transgenic worms expressing GCaMP calcium indicators under a cell-selective promoter were grown to day 1 or day 5 of adulthood and trapped in a custom designed PDMS microfluidic device and exposed to odor stimuli (Chalasani et al., 2007 Chronis et al., 2007). For aging experiments, a new PDMS device with larger channels was designed to trap and stimulate day 5 adult worms (Chokshi et al., 2010). Older, day 6 adult worms exhibit much larger variation in whole animal size than day 5 adults (see Figure 5—figure supplement 1A) and could not be trapped consistently without introducing bias into the experiment. For aging experiments, animals were transferred to new OP50 bacteria plates every other day to track the aging animals and to avoid contamination by their progeny. Additionally, for whole animal RNAi experiments to knockdown rab-10 and hsf-1, animals were fed either control (Ctrl) empty pL4440, rab-10 RNAi or hsf-1 RNAi expressing bacteria beginning at day 1 of adulthood as previously described (Hansen et al., 2005).

      Fluorescence from the neuronal cell body was captured using a Zeiss inverted compound microscope for 3 min. We first captured 10 s of baseline activity (t = 0–10 s) in chemotaxis assay buffer (5 mM K3PO4 (pH 6), 1 mM CaCl2, 1 mM MgSO4, and 50 mM NaCl), then 2 min (t = 10–130 s) of exposure to an odor (or salt) stimulus dissolved in chemotaxis buffer, and lastly 50 s (t = 130–180 s) of buffer only. BZ refers to a 0.005% vol/vol dilution in chemotaxis assay buffer, except where low BZ (0.0001% vol/vol) or high BZ (0.1% vol/vol) is specifically mentioned. Additionally, a 0.1% vol/vol dilution of 2-nonanone and 50 mM sodium chloride stimulus were used as indicated. For arecoline experiments, worms were pre-treated with 0.15 mM arecoline in chemotaxis buffer for approximately 20 min and immediately imaged in the presence of the drug. Laser ablations of the paired AWC, AWA, AWB or ASE sensory neurons, along with mock ablations, were performed as previously described (Bargmann and Avery, 1995) in transgenic animals expressing GCaMP. In all experiments, a single neuron was imaged in each animal, and each animal was imaged only once. Wild-type Ctrls, mutants, and transgenic or drug treated strains for each figure were imaged in alternation, in the same session.

      We used Metamorph and an EMCCD camera (Photometrics) to capture images at a rate of 10 frames per second. A MATLAB script was used to analyze the average fluorescence for the cell body region of interest and to plot the percent change in fluorescence for the region of interest relative to F0, as previously described (Chalasani et al., 2007). Specifically, data was plotted and statistical analysis was performed as follows: (1) for line graphs of ΔF/F over time (Figures 1–4 and corresponding figure supplements), the average fluorescence in a 8 s window (t = 1–9 s) was set as F0. Average and standard error at each time point were generated and plotted using MATLAB scripts, as previously described (Leinwand and Chalasani, 2013). (2) For heat maps (Figures 5–7 and corresponding figure supplements), the average fluorescence in a 8 s window (t = 1–9 s) was set as F0.

      To quantify calcium responses, F0 was consistently set to the average fluorescence signal from 1 s to 9 s prior to the relevant change (addition or removal) of stimulus. For statistical analysis, the average fluorescence and standard error were calculated for each animal over a short period corresponding to the duration of a response. Specifically, to analyze on responses to the addition of stimulus, the average fluorescence and standard error were calculated in the 10 s period following the addition of odor or salt (t = 10–20 s). For AWA neurons, the response duration was very brief therefore, a 4 s time period was used instead (t = 10–14 s) so that small, fast responses could be appropriately quantified. To analyze off responses to the removal of stimulus, the average fluorescence and standard error were calculated in the period following the removal of odor (t = 130–140 s for all cells except ASE, and t = 130–145 for the slower, longer duration ASE responses). Traces in which an averaged ΔF/F of greater than 600% was recorded were excluded as they are likely to be artifacts of the neurons moving out of the focal plane and these usually account for less than 1% of the traces collected. To determine whether there was an odor-evoked increase or suppression of the calcium signal (see Figure 1C), the average fluorescence in these time windows in buffer only trials was compared (by a two-tailed unpaired t-test) to the average fluorescence in odor stimulation trials, for each neuron. The maximum ΔF/F in these time periods following odor addition or removal and the time to reach this maximum ΔF/F (from the stimulus change, in seconds) were also quantified (see Figure 5G, Figure 1—figure supplement 1A,B and Figure 5—figure supplement 1B). More specifically:

      (1) For bar graphs of averaged ΔF/F after odor addition or removal (Figures 2–4, Figure 3—figure supplement 1 and Figure 4—figure supplement 1): (a) F0 was set to the average fluorescence from 1–9 s for quantification of AWA neuron responses to the addition of BZ stimulus and (b) F0 was set to the average fluorescence from 121–129 s for quantification of AWC, ASE and AWB responses to the removal of BZ or 2-nonanone. Two-tailed unpaired t-tests were used to compare the responses of different genotypes or cell ablation conditions, and the Bonferroni correction was used to adjust for multiple comparisons.

      (2) For scatter plots of maximum ΔF/F (Figure 1—figure supplement 1A and Figure 5G) and scatter plots of averaged ΔF/F after stimulus change (Figure 5H, Figure 6—figure supplement 1B and Figure 7—figure supplement 1B): (a) for AWA neurons' response to the addition of odor stimulus F0 was set to the average fluorescence from 1–9 s and (b) for AWC ON , ASEL and AWB responses to odor stimulus removal F0 was set to the average fluorescence from 121–129 s. For the subset of odor-responsive neurons (exceeding the 10% ΔF/F cut-off), the averaged ΔF/F after the stimulus change and the time to the maximum ΔF/F were also analyzed using two-tailed unpaired t-tests to compare different ages or genotypes (Figure 5H, Figure 5—figure supplement 1B, Figure 6—figure supplement 1B and Figure 7—figure supplement 1B). Furthermore, considering only the odor responsive neurons, no significant differences were observed in the magnitude of the odor-evoked suppression of young and aged animals (comparing the average fluorescence in ten second windows tiling the period of odor stimulation, by two-tailed t-test), indicating that our subsequent analyzes of the odor removal time period are not biased by the choice of the F0.

      (3) For bar graph quantifications of the % odor or salt responsive neurons in the aging experiments (Figure 5I,K,L, 6F, and 7E and the corresponding figure supplements): (a) F0 was set to the average fluorescence from 1–9 s for quantification of the percent of AWA and ASH neurons responsive to the addition of BZ stimulus and for the percent of ASEL and AWC neurons responsive to the addition of NaCl salt stimulus. (b) F0 was set to the average fluorescence from 121–129 s for AWC ON , ASEL and AWB responses to BZ or 2-nonanone odor stimulus removal. The percent of odor responsive neurons was calculated by determining the proportion of cells displaying an average fluorescence (ΔF/F) greater than 10% after odor addition (for AWA and ASH) or odor removal (all other neurons). 10% ΔF/F was used as the cut-off for odor responsiveness because, for all neurons imaged, changing buffer around the nose of the animal elicited a response smaller than this cut-off. Similarly, neurons displaying an average fluorescence (ΔF/F) greater than 10% after salt addition were considered to be salt responsive. A two-tailed Chi–Square test was used to compare the percent of odor or salt responsive neurons in different conditions.

      Chemotaxis assays

      Odor chemotaxis assays were performed as previously described (Ward, 1973). For aging assays, worms were synchronized by hatch offs in which 8 young adult worms were given 150 min to lay eggs on a large plate before being picked off. These eggs were grown at 20° until the appropriate day of adulthood, except for glp-1 mutants, which were raised at the restrictive temperature, 25°. Aging animals were transferred to new bacteria plates every other day to track the aging animals and to avoid contamination by their progeny. Chemotaxis assays were performed on 2% agar plates (10 cm diameter) containing 5 mM potassium phosphate (pH 6), 1 mM CaCl2 and 1 mM MgSO4. Animals were washed once in M9 and three times in chemotaxis buffer (5 mM K3PO4 (pH 6), 1 mM CaCl2 and 1 mM MgSO4). For arecoline chemotaxis experiments, 0.15 mM arecoline was added to the M9 and chemotaxis buffer washes, yielding a 16–20 min drug treatment immediately prior to the behavioral experiment. Odor concentration gradients were established by spotting diluted BZ (0.2% vol/vol, in ethanol) near the edge of the plate, with a Ctrl 1 μl of ethanol spotted at the opposite end of the plate. Where noted, 1 μl of neat BZ was used for high concentration point source assays. For 2-nonanone experiments, a 50% vol/vol dilution of 2-nonanone in ethanol was used. For salt chemotaxis experiments, salt gradients were established by placing a Ctrl or a high salt (500 mM NaCl) agar plug on the assay plate and allowing 16–20 hr for the salt to diffuse and form a gradient (Leinwand and Chalasani, 2013). 1 μl of sodium azide was added to the odor (or salt) and the Ctrl spots to anesthetize animals reaching the end points. Washed worms were placed on the plate and allowed to move freely for one hour. The chemotaxis index was computed as the number of worms in the region near the odor (or salt) minus the worms in the region near the Ctrl divided by the total number of worms that moved beyond the origin. Nine or more assays were performed, over at least three different days. Two-tailed unpaired t-tests were used to compare the responses of different genotypes or ages, and the Bonferroni correction was used to adjust for multiple comparisons.

      Correlated chemotaxis and imaging experiments

      Transgenic worms bearing GCaMP arrays, synchronized by a hatch off as described above, were grown until day 5 of adulthood at 20°. Aging animals were transferred to new bacteria plates every other day to track the aging animals and to avoid contamination by their progeny. Animals were tested in a (0.2% vol/vol) BZ odor chemotaxis assay as above, with two modifications. First, no sodium azide was used to paralyze the animals. Second, animals were given only 30 min to move freely on the chemotaxis plate. The chemotaxis assay plate was then cut into three regions corresponding to the BZ odor side, the middle, and the ethanol Ctrl region immediately after 30 min and worms were washed off each section separately and allowed to recover on OP50 bacteria plates for at least 90 min. Worms from the odor and the Ctrl sections of the chemotaxis assay were imaged in alternation as described above.

      Lifespan assays

      Worms, synchronized by a hatch off as described above, were grown until day 1 or 5 of adulthood at 20°. To sort animals on the basis of their chemotaxis performance, wild-type animals were tested in a (0.2% vol/vol) BZ odor or (500 mM NaCl) salt chemotaxis assay as above, but without sodium azide and with only 30 min for the animals to move freely on the chemotaxis plate. The chemotaxis assay plate was then cut into a BZ odor (or salt), middle, and Ctrl region and worms were washed off each section separately. 100 adults from the odor (or salt) region or the Ctrl region were transferred onto 10 small OP50 plates (10 adults per plate) and grown at 20°. For experiments with transgenic animals, day 1 animals bearing the appropriate transgene were picked from the hatch off plate directly onto 10 small OP50 plates (10 adults per plate) and grown at 20°. In all experiments, aging animals were transferred to new bacteria plates every other day to track the aging animals and to avoid contamination by their progeny. Survival was analyzed every other day and worms were scored alive or dead based on their response to a gentle head touch (or lack thereof) as previously described (Kenyon et al., 1993). Worms were censored if they bagged, exploded or desiccated on the side of the plate. The chemotaxis assay followed by lifespan analysis or lifespan assays with transgenic animals were repeated two or three times per condition as indicated, beginning on separate days. The percent change in mean survival was calculated as the mean survival of animals from the odor side minus the mean survival of animals from the Ctrl side divided by the mean odor side survival or the mean transgenic animal survival minus the mean wild-type survival divided by the mean wild-type survival. Statistical analysis of lifespan was performed by the Mantel–Cox Log–Rank test, using GraphPad Prism and OASIS (Yang et al., 2011).

      Speed analysis

      Chemotaxis assays to BZ were set up as described above, but with modifications to enable automated analysis of animal speed. 200 mM Cu(II)SO4-soaked filter paper was placed on a standard chemotaxis assay plate to contain the worms in a reduced chemotaxis arena (1.25 by 1.25 inch square). 1 μl of BZ (0.2% vol/vol dilution in ethanol) and a Ctrl 1 μl of ethanol were spotted at opposite corners of the square arena, without any paralytic. After washing, only 5 worms were placed on the chemotaxis plate this number minimized collisions and enabled more accurate tracking. The movement of the animals was tracked over 60 min using a Pixelink camera and speed was analyzed using previously published MATLAB scripts to track the centroid of the animal (Ramot et al., 2008). The results from eleven chemotaxis plates were averaged for each age. NS indicates p > 0.05, two-tailed t-test.

      Aged worm measurements

      Day 5 and day 6 adult worms from hatch offs performed on three separate days were immobilized with tetramisole and imaged on 2% agarose pads. Images were captured on a Zeiss Observer D1 microscope using a 10× objective with DIC. The perimeters of 55 worms were measured using MetaMorph software.


      A Caenorhabditis elegans Model System for Amylopathy Study

      We describe methods to study aspects of amylopathies in the worm C. elegans. We show how to construct worms expressing human A&beta42 in neurons and how to test their function in behavioral assays. We further show how to obtain primary neuronal cultures that can be used for pharmacological testing.

      Abstract

      Amylopathy is a term that describes abnormal synthesis and accumulation of amyloid beta (A&beta) in tissues with time. A&beta is a hallmark of Alzheimer's disease (AD) and is found in Lewy body dementia, inclusion body myositis and cerebral amyloid angiopathy 1-4 . Amylopathies progressively develop with time. For this reason simple organisms with short lifespans may help to elucidate molecular aspects of these conditions. Here, we describe experimental protocols to study A&beta-mediated neurodegeneration using the worm Caenorhabditis elegans. Thus, we construct transgenic worms by injecting DNA encoding human A&beta42 into the syncytial gonads of adult hermaphrodites. Transformant lines are stabilized by a mutagenesis-induced integration. Nematodes are age synchronized by collecting and seeding their eggs. The function of neurons expressing A&beta42 is tested in opportune behavioral assays (chemotaxis assays). Primary neuronal cultures obtained from embryos are used to complement behavioral data and to test the neuroprotective effects of anti-apoptotic compounds.

      Introduction

      Amyloid beta (A&beta) is a peptide of 36-43 amino acids that is formed after sequential cleavage of the amyloid precursor protein (APP) by &beta and &gamma secretases 1 . The &gamma secretase processes the C-terminal end of the A&beta peptide and is responsible for its variable lengths 5 . The most common forms of A&beta are A&beta40 and A&beta42, the latter being more commonly associated to pathologic conditions such as AD 5 . At high concentrations A&beta form &beta-sheets that aggregate to form amyloid fibrils 6 . Fibrils deposits are the main component of senile plaques surrounding neurons. Both plaques and diffusible, non-plaque A&beta oligomers, are thought to constitute the underlying pathogenic forms of A&beta.

      Laboratory study of neuronal amylopathies is complicated by the fact that these conditions progress with time. Therefore, it is important to develop genetically tractable animal models-complementary to mice-with short life span. These models can be used to elucidate specific aspects of amylopathies-typically cellular and molecular-and by virtue of their simplicity, help to capture the essence of the problem. The worm Caenorhabditis elegans falls is this category. It has a short life span,

      20 days and in addition basic cellular processes including regulation of gene expression, protein trafficking, neuronal connectivity, synaptogenesis, cell signaling, and death are similar to mammalian 7 . Unique features of the worm include powerful genetics and lack of a vessel system, which enables to study neuronal damage independently of vascular damage. On the other hand, the lack of a brain limits the use of C. elegans to studying many aspects of neurodegeneration. In addition, the reproduction and identification of anatomical distributions of lesions cannot be performed in this organism. Other limitations include the difficulty to assess both differences in gene expression profiles and impairment of complex behavior and memory function. Here we describe methods to generate C. elegans models of amylopathies.

      Subscription Required. Please recommend JoVE to your librarian.

      Protocol

      1. Construction of Transgenic Worms

      1. Transformation.
        1. Prepare injection pads. Place a drop of hot, 2% agarose dissolved in water, onto a glass coverslip. Quickly place a second coverslip on the drop and lightly tap it. After the agarose is solidified, slide coverslips apart, and bake the coverslip-pad in a vacuum oven at 80 °C O/N.
        2. Pull pipettes. We use a Sutter P-97 puller to pull 1/0.5 mm O.D./I.D. borosilicate capillaries with filament. Pipettes are forged with closed tip which is broken open at a later stage.
        3. Prepare injection mix. Make the injection mixture containing the DNA of interest (20 ng/&mul per construct) plus empty plasmid DNA to a final concentration of 150 ng/&mul. Centrifuge the injection mixture to remove any contaminant. This step is crucial because contaminants reduce transformation efficiency. Transfer the top 5 &mul (debris and other contaminants collect in the bottom) to a fresh tube to be used in injection.
        4. Load injection mix by capillarity. The bottom of the pipette is immersed in the injection mix. Typically 0.5 &mul are sufficient to inject 100 worms.
        5. Load injection pipette. Insert the pipette onto the holder of the micromanipulator and break it open by rubbing its tip against the edge of a cover slip mounted on a glass slide or alternatively against debris on the agarose pad. This is a crucial step as too large tips damage the worm and too small tips are easily clogged. The quality of the broken tip can be judged by the shape and most importantly by the flow rate. The flow rate can be assessed by the size of the bubbles flowing from the open tip immersed in a drop of halocarbon 700 oil in response to an injection pressure.
        6. Transfer worms to injection pad. Place a drop of 700 halocarbon oil on an injection pad and transfer several (1 or 2 for beginners) worms to the oil. Use a worm pick to push the worms down onto the pad until they adhere to the agarose. Worms should be oriented in rows with ventral sides facing the same direction. Avoid touching worm's head. If after several attempts the worms fail to adhere to the pad, replace with a fresh pad or increase agarose thickness and/or concentration.
        7. Insert the pipette into the worm. By moving the stage, position the worm under the pipette. Position the pipette in the central core of the gonad because its cytoplasm is shared by many germ cell nuclei. This increases the likelihood to deliver injected DNA to many progeny. The pipette should lie almost parallel to the worm (

        1. Place 40 healthy, well-fed, L4 worms into a fresh plate
        2. Irradiate with &gamma-ray with 4,000 rads for 40 min. Transfer irradiated worms (P0) in fresh, OP50 seeded plates (4 worms/plate). Worms can alternatively be irradiated with a dose of 300 J/m 2 UVs.
        3. Transfer 10-20 F1 transformants from each plate to individual plates, label the plates with the corresponding P0 origin.
        4. Single out 2-4 F2 transformants for each F1 to separate plates. Check the F3 progeny for 100% transmission of the transformation marker. Typically, 1-3% of progeny from an irradiated worm will have an integration event.
        5. Make stocks of three or more independent transformant lines.

        1. Age-synchronization.
          1. Grow worms in standard 10 cm NGM plates + OP50 E. coli until a large population of gravid adults is reached (3-5 days).
          2. Collect the worms in 50 ml Falcon tubes by suspending them in 1 ml M9 buffer.
          3. Add 5 ml M9 buffer and centrifuge at 450 x g for 3 min. Discard supernatant. Repeat 3-4x.
          4. At the end of the last centrifugation, remove the supernatant and add 10 volumes of basic hypochlorite solution (0.5 M NaOH, 1% hypochlorite freshly mixed) to the pelleted worms. Incubate at RT for

          1. Prepare several pieces of agar roughly 0.5x0.5x0.5 cm. We deposit the agar in a 10-cm plate and we cut the agar chunks from there.
          2. Soak the agar chunk in a solution containing the desired attractant (at saturating, near-saturating concentrations) for 2 hr. Typical attractants are lysine (0.5 M), biotin (0.2 M).
          3. Deposit an agar chunk in a 10 cm test plate in which the location of a test spot and a control spot have been marked (Figure 1A). Allow equilibration and formation of a gradient O/N (Figure 1B). Prepare 5 plates for a single experiment.
          4. Prior the experiment add 10 &mul of 20 mM NaN3 (anesthetic) to each spot.
          5. Place 20 age-synchronized worms in the center of the plate. Place the plate in the incubator at 20 °C.
          6. After 1 hr, count the animals on the test/control spots and calculate the chemotaxis index, (C.I.) as follow: where N, Ntest and NCnt., indicate the total number of animals, the number of animals in the test spot and the number of animals in the control spot.

          3. Primary Embryonic Cell Culture

          1. Lyse worms as described in age-synchronization.
          2. Stop the lysis reaction by adding the same volume of sterile egg buffer and centrifuge at 450 x g for 5 min. Gently discard supernatant being careful to not lose pelleted eggs. Repeat 2-3x or until supernatant is clear.
          3. Resuspend pelleted eggs (and carcasses) in 2 ml sterile egg buffer and add 2 ml of sterile 60% sucrose in egg buffer. Mix this solution until eggs are completely resuspended (as they tend to form clumps under centrifugation) by hand or by vortexing.
          4. Centrifuge at 450 x g for 15 min.
          5. Carefully transfer supernatant (containing the eggs) to a sterile tube. Discard pellet which contains carcasses and other by-products of lysis.
          6. Remove residual sucrose by resuspending the eggs in egg buffer and centrifuging at 450 x g for 5 min. Gently collect and discard the supernatant. Repeat 3x.
          7. Under a laminar hood resuspend pelleted eggs in sterile egg buffer containing 1 U/ml chitinase at RT to digest the eggshells. After 30 min start to monitor the reaction (under an inverted cell culture microscope). Each batch of chitinase has a slightly different activity. Typically digestion is completed in 1 hr.
          8. When roughly 70-80% eggshells are digested by chitinase add CM-15 (L-15 cell culture medium containing 10% fetal bovine serum, 50 U/ml penicillin, and 50 &mug/ml streptomycin). Dissociate cells using a syringe with a 27 gauge. Filter the cell suspension with a 5.0 &mum filter to remove intact embryos, clumps of cells and larvae.
          9. Pellet the dissociated cell suspension by centrifugation at 450 x g for 15 min. Remove the supernatant and resuspend the pellet in CM-15 cell culture medium.
          10. Plate dissociated cells on glass cover slips previously coated with peanut lectin (0.1 mg/ml) dissolved in water. Note: cells must adhere to the substrate in order to differentiate.
          11. Cells can be maintained at RT (16-20 °C) in air for more than 2 weeks.

          Subscription Required. Please recommend JoVE to your librarian.

          Representative Results

          With our protocols we study the effects of human A&beta42 oligomer on neuronal function 8 . A fragment encoding human A&beta42 and the artificial signal peptide coding sequence of Fire vector pPD50.52 was amplified from construct PCL12 9 using primers that introduced a Sma 1 restriction endonuclease site at the ends. The fragment was then inserted into a construct containing a 2,481-bp flp-6 promoter sequence in the pPD95.75 Fire vector between the unique Sma 1 site 10 . Using the transformation techniques described in protocol 1 we constructed a transgenic worm expressing A&beta42 in the ASE neurons (FDX(ses25) strain) 8 . To mark positive transformants we used the Pgcy-5::GFP reporter which specifically drives GFP expression in the ASE right (ASER) neuron 11 . Worms stick to the pad because the dry agarose absorbs their water. Therefore it is crucial that the animals are placed onto the injection pad and injected relatively quickly, because otherwise they will desiccate and die. The percentage of F1 progeny that carries a transmissible extrachromosomal array may vary. Typical values are in the 3-7% range. It is important that the injection mix (1.1.3) contains non-encoding DNA sharing sequence homology with the transgenic DNA (usually empty vector) because the DNAs undergo homologous recombination with each other. However, if overexpression of the transgene is a problem the injection mix should be supplemented with 50-100 ng/&mul of genomic DNA digested with Sca 1. ASE neurons detect water soluble attractants such as biotin and therefore their function can be assessed in behavioral assays (chemotaxis assay, protocol 2 and Figures 1A and 1B) 12 . In a typical experiment, we tested seven day old worms expressing Pgcy-5::GFP alone (DA1262 strain) or with A&beta42 (FDX(ses25)) for chemotaxis to biotin. In young worms (3-4 day old) the effects of A&beta42 expression are modest but already detectable (

          10% decrease in chemotaxis index, see ref. 8 ). Representative results of this experiment are shown in Figures 1C and 1D. Most DA1262 worms were found in, or nearby, the attractant spot (Figure 1C). By contrast only a few worms expressing A&beta42 could find the attractant spot (Figure 1D). We tested 100 animals/genotype distributed in 5 test plates/genotype obtaining a chemotaxis index for biotin 0.68±0.09 and 0.12±0.04 for DA1262 and FDX(ses25), respectively. Active worms were individually picked and transferred to the center of the test plate. It is important to not only quickly, but also gently transfer the worms because otherwise they may remain inactive for several minutes and fail to track the attractant. At the end of the experiment we suggest monitor worms activity by looking at their paths. If only a few tracks are visible we usually discard the plate.

          GFP fluorescence in the ASER neurons of FDX(ses25) worms disappears within the first eleven days of life (not shown). This suggests that these cells undergo apoptosis due to the presence of A&beta42. Therefore we determined whether a broad-spectrum inhibitor of apoptosis such as caspase inhibitor N-(2-Quinolyl)valyl-aspartyl-(2,6-difluorophenoxy)methyl Ketone (Q-VD-OPh) 13 could stop the loss of ASER cells. To this end we employed cultured primary ASER neurons from embryos 14 , which were prepared as described in protocol 3. Representative images of an ASER neuron in a young (4 day old) FDX(ses25) worm along with images of neurons in culture are shown in Figures 2A-C. Cultured ASER cells were short-lived (Figure 2D). As expected incubation with 0.3 &mug/ml Q-VD-OPh freshly supplemented daily, completely stopped the loss of fluorescent ASER neurons. When working with primary neurons in culture it is important to maintain an opportune cell density. This parameter mainly depends on the number of worms used to extract the eggs. We measure cell density with a standard hemocytometer and take particular care to maintain cell density constant from cell culture to cell culture. For pharmacological experiments such as those described here we worked with

          200,000 cells/cm 2 which was obtained by harvesting four confluent 10 cm plates. Cells were plated in 12 wells of a 24-well plate. For optimal results it is also important to dissociate the embryonic cells before seeding as they tend to form clumps. We use a syringe with a 27 gauge needle and gentry aspirate the suspension back and forth a couple of times. This is generally sufficient to dissociate most of the cells (especially in the beginning we suggest to check the suspension under a microscope).


          Figure 1. Chemotaxis assay. A. Representative chemotaxis plate. Attractant (biotin) and control spots are marked with an hollow and a filled circle. B. A chemotaxis plate loaded with a 0.5-cm diameter piece of agar used to establish a gradient of biotin. The piece of agar was cut from a 10-cm plate using the top side of a glass Pasteur pipette. C. Representative distribution of worms around the attractant spot. In this example the majority of DA1262 worms were able to localize the source of attractant (biotin). D. As in C. for FDX(ses25) worms. These biotin-insensitive worms exhibited a scattered distribution around the plate and only few were found near the attractant spot.


          Figure 2. Culture of C. elegans embryonic cells. A. Fluorescence microscopy image (left picture) and bright light (right picture) taken from a FDX(ses25) transgenic worm head. This worm expresses GFP in the ASER neuron driven by the gcy-5 promoter. B. Bright light image of a culture of FDX(ses25) embryonic cells. Scale bar is 5 &mum. C. Fluorescence microscopy image of a cultured FDX(ses25) ASER neuron. Images were taken with an Olympus BX61 microscope equipped with a digital camera. D. Representative experiment testing the viability of cultured, age-synchronized, ASER neurons. Cells were obtained from DA1262 embryos (hollow circles) or FDX(ses25) embryos maintained in the absence/presence of 300 ng/ml Q-VD-Oph (hollow and filled squares, respectively). The disappearance of GFP fluorescence was used as a measure of a neuron's viability. The experiment started with

          300 fluorescent ASER neurons. Viability was calculated as 100*(number of fluorescent cells at day X divided by number of fluorescent cells at day 1). Click here to view larger figure.

          Subscription Required. Please recommend JoVE to your librarian.

          Discussion

          Here we describe a combined approach, to study cellular and molecular aspects of amylopathies using C. elegans. The advantages of this approach include: 1) low cost. C.elegans is maintained in normal Petri dish seeded with bacteria, at room temperature. 2) Powerful genetics. Transgenic animals can be obtained in few months and a wide array of promoter sequences is available to drive expression of the desired gene in specific neurons. 3) Simple, well-characterized, nervous system. C. elegans possesses a remarkably simple nervous system (302 neurons). This simplicity has afforded extensive characterization of the worm's nervous system including cell lineage, specific function/role of each neuron and its synaptic connections. The limitations of C. elegans include the small size of the cells, which hinders the application of standard biochemical techniques such as immunohistochemistry and a thick skin (cuticle) which insulates neurons form the external environment. Therefore, pharmacological approaches may be of limited efficacy in C. elegans. Cultures of primary cells represent a valid strategy to partially ameliorate this problem.

          The critical steps in these experimental protocols include starting with large quantities of worms and monitoring the preparations in order to not lose eggs, embryos etc. It is also crucial to learn how to handle the animals during injection. Keeping worms in the agarose pad too long, punching them with a large pipette or injecting too much DNA mix can irreversibly damage them. Transformation efficiency, reproducibility and transgene expression may vary. Efficiency is related to the purity of the injection mix and its composition (presence of non-coding DNA). Extrachromosomal arrays vary from animal to animal therefore it is critical to establish at least 2-3, independent lines per transgene. On the other hand, the addition of digested genomic DNA to the mix represents a valid strategy to reduce transgenic over-expression. Chemotaxis assays are relatively trouble-free. However it is crucial to maintain consistency in the concentration gradient in order to avoid false results. Use pieces of agar of the same size-for example cut them with a Pasteur pipette-and maintain the equilibration time constant. If any, remove excess solution with a cotton swab.

          In conclusion, here we provide an example of how a simple, genetically tractable organism can be exploited to investigate molecular aspects of amylopathies. The same experimental techniques could be applied to the study of other neuronal proteins and to the generation of new animal models of disease.

          Subscription Required. Please recommend JoVE to your librarian.

          Disclosures

          The authors declare that they have no competing financial interests.

          Acknowledgments

          We thank Dr. Shuang Liu for critical reading of the manuscript. The PCL12 construct was a gift form Dr. Christopher D. Link. This work was supported by two National Science Foundation grants (0842708 and 1026958) and an AHA grant (09GRNT2250529) to FS.


          Neural Circuits of Sexual Behavior in Caenorhabditis elegans

          The recently determined connectome of the Caenorhabditis elegans adult male, together with the known connectome of the hermaphrodite, opens up the possibility for a comprehensive description of sexual dimorphism in this species and the identification and study of the neural circuits underlying sexual behaviors. The C. elegans nervous system consists of 294 neurons shared by both sexes plus neurons unique to each sex, 8 in the hermaphrodite and 91 in the male. The sex-specific neurons are well integrated within the remainder of the nervous system in the male, 16% of the input to the shared component comes from male-specific neurons. Although sex-specific neurons are involved primarily, but not exclusively, in controlling sex-unique behavior—egg-laying in the hermaphrodite and copulation in the male—these neurons act together with shared neurons to make navigational choices that optimize reproductive success. Sex differences in general behaviors are underlain by considerable dimorphism within the shared component of the nervous system itself, including dimorphism in synaptic connectivity.


          Results

          Alternative splicing and promoter of the nfi-1 gene

          The C. elegans nuclear factor I gene (nfi-1) was first identified by the C. elegans sequencing consortium by its homology to mammalian NFI genes [14]. Only a single NFI gene is present in the C. elegans genome (Fig. ​ (Fig.1A) 1A ) while vertebrates possess 4 NFI genes. To confirm the structure of the nfi-1 gene, we obtained cDNAs from the Kohara and TIGR libraries and made primers within predicted exons for production of cDNAs from RNA of adult worms and purified eggs. These cDNAs confirmed the presence of 14 exons and showed alternative splicing of exons 2 and 13 (Fig. ​ (Fig.1A). 1A ). The differential splicing pattern of nfi-1 indicates that different isoforms may be expressed in different cell types. Primary transcripts from most C. elegans genes are trans-spliced to SL1 leader sequences and detection of SL1-linked transcripts is frequently used to assess the initially transcribed exons of genes [37]. We confirmed the start site of the gene using PCR with an SL1 primer for SL1-containing transcripts and a primer in exon 2.

          A) Alternatively spliced products of the nfi-1 gene. The nfi-1 gene is shown as a line with exons as numbered solid boxes and alternatively spliced exons as gray boxes. Arrows show the direction of transcription. Below the line are cDNAs from the Kohara and TIGR libraries the vertical bar indicates the 5' end of the cDNA. In yK213C10 the letters a and b above exon 2 denote that it is alternatively spliced generating 2a and 2b and the arrow on the right denote undetermined sequence in the cDNA. The asterisk (*) on yk42f10 denotes an alternative 3'splice acceptor site in exon 13 used in yk42f10 but not in CEESQ09. Below are depictions of cDNAs obtained by RT-PCR from total and polyA+ RNA of whole worms or isolated eggs using nfi-1 exon-specific primers (RT-PCR nfi-1 primers). Lastly we show cDNA obtained by PCR using SL1 primer and nfi-1 exon-specific primers (RT-PCR SL1-linked products). Some of the alternatively spliced cDNAs have been described previously [14]. B) Comparison of CeNFI-DBD with the consensus mouse NFI DBD.The sequence of the CeNFI-DBD (top line) is aligned to a consensus sequence from the 4 mouse NFI-DBDs (bottom line). Gaps in the mouse consensus indicate residues that are not identical between the 4 mouse NFI-DBDs. The dash in the 6 th aligned row indicates a single insertion in the CeNFI-DBD sequence needed to align it with the mouse consensus. Dark gray boxes show identical residues in CeNFI-DBD and mouse NFI-DBDs, light gray boxes show residues not identical but similar between CeNFI and mouse NFI-DBDs, unboxed residues with black letters show residues that are not similar between the CeNFI and mouse DBDs, and gray letters in the CeNFI-DBD above gaps in the mouse consensus indicate positions where the 4 mouse genes are not identical. 151 of 190 residues are identical in all 4 mouse NFI-DBDs while 60 of these 151 residues are different in the CeNFI-DBD and 27 of these 60 differences are non-conservative substitutions. The alignment was done in Macvector 6.5.3 using the ClustalW similarity matrix.

          DNA binding by CeNFI is indistinguishable from that of human NFI-C

          The predicted DNA-binding domain of nfi-1 (CeNFI-DBD) shares homology with the vertebrate NFI proteins, however 60 of 151 residues that are completely conserved among the 4 mouse NFI proteins are changed in CeNFI (Fig. ​ (Fig.1B). 1B ). To assess the DNA-binding activity and specificity of CeNFI, the DNA-binding domain (DBD) of CeNFI was cloned in frame with a 6 histidine-tag, expressed in E. coli, partially purified by nickel-affinity chromatography, and was used for in vitro DNA binding assays. The DNA-binding activity of CeNFI-H6 was indistinguishable from that of the human hNFI-C220H6, with both proteins binding the NFI-site oligo 2.6 (Fig. ​ (Fig.2A) 2A ) but not the C2 oligo containing a point mutation that prevents vertebrate NFI binding [19,38]. To ask if wild-type C.elegans contains a protein with similar DNA binding properties as CeNFI-H6, extracts were prepared from a mixed population of worms and tested for binding to the 2.6 and C2 sites. As expected, proteins were detected that bound to the 2.6 but not the C2 site, confirming the binding specificity of CeNFI (Fig. ​ (Fig.2B). 2B ). The specificity of the native and recombinant CeNFI proteins was also measured by competition of binding of the 2.6 oligo with various unlabeled oligonucleotides and was indistinguishable from the specificity of hNFI-C220-H6 (data not shown). Thus despite the differences between CeNFI and vertebrate NFIs in conserved residues of their DNA-binding domains, their DNA-binding activities are indistinguishable.

          A) Specific NFI DNA-binding activity of recombinant CeNFI-H6. Partially purified recombinant H6-tagged CeNFI (containing parts of exons 2𠄸) and human NFI-C220 were incubated with a duplex oligonucleotide containing an NFI binding site (2.6, even lanes) or an oligo with a single point mutation that abolishes NFI binding (C2, odd lanes) and analyzed on a 6.5% non-denaturing polyacrylamide gel. Lanes 1 and 2, crude E. coli extract (neg. control) lanes 3 and 4,

          5 ng partially purified CeNFI-H6 lanes 5 and 6,

          40 ng partially purified CeNFI-H6 lanes 7 and 8,

          5 ng purified human NFI-C220H6. See bottom of panel for sequences of oligonucleotides. B) Specific NFI DNA-binding activity in worm extracts. Nuclear extracts of a mixed population of C. elegans were prepared and used in a gel mobility shift assay with an oligonucleotide that contains an NFI-binding site (lanes 1 & 2, 2.6) or the same oligo with a single point mutation that abolishes NFI binding (lane 3, C2). Lane 1, no extract Lanes 2 and 3, C. elegans extract (

          10 μg). See A for sequences of oligonucleotides.

          Expression of nfi-1 in embryo and adults

          In the mouse, the 4 NFI genes are expressed in a complex overlapping pattern during embryogenesis and in adult tissues [13]. To assess the expression pattern of nfi-1 transcripts in C. elegans, a digoxin-labeled antisense probe from the 3' end of the nfi-1 transcript was used for in-situ hybridization to fixed embryos and whole worms [39]. nfi-1 transcripts are present in the one-cell (not shown) and two-cell stages (Fig. ​ (Fig.3a) 3a ) prior to the onset of zygotic transcription [40,41], indicating that the transcript is maternally inherited. Expression continues in most cells of the embryo throughout early and mid-embryogenesis (Fig. 3b𠄽 ) but decreases after gastrulation and no expression is seen in L1 larvae (Fig. 3f–i and data not shown). As expected for a maternally inherited transcript, expression of nfi-1 reappears in the adult gonad (Fig. 3j–m ). Expression of nfi-1 transcripts are also seen in the cytoplasm of gut cells (Fig. 3j–m ). No signal is seen using control sense probes (Fig. 3n–o and data not shown). This in situ expression pattern indicates that nfi-1 could function both early in embryogenesis and in adult worms.

          Expression of endogenous nfi-1 mRNA in embryo and adults. N2 worms were fixed and hybridized with digoxigenin (DIG)-labeled antisense nfi-1 probe (from plasmid yk42f10) and bound probe was detected using alkaline phosphatase-conjugated anti-DIG antibodies. Panels : a, 2 cell embryo b, 4 cell embryo c, 24 cell embryo d, beginning gastrulation e, mid-gastrulation f, late gastrulation g, comma stage h, 1.5-fold stage and i, 2-fold stage. Embryos in a-d and f-i are positive for staining. We are currently investigating the apparent loss of signal at mid-gastrulation (panel e). Panels j-m, antisense probe, adults with exposed internal organs: mature oocytes intestine gonadal germ cells n-o, control sense probe, no specific staining is seen. Right panels k, m, o are two-fold magnifications of those on the left.

          Isolation of a null allele of nfi-1

          To assess the role of nfi-1 in worms, we isolated a nfi-1 mutant using a reverse genetic approach based on the insertion-excision of the transposon Tc1 [42]. Worms carrying an

          2 kb deletion in nfi-1 were isolated by PCR screening and sib-selection (Fig. 4A,B ). Sequence analysis of the nfi-1 transposon excision allele (designated nfi-1(qa524)) showed loss of the genomic region corresponding to nucleotides 14754� of cosmid ZK1290. This eliminates the first 6 exons of nfi-1 including sequences encoding the DNA-binding domain. RT-PCR shows the absence of nfi-1 transcripts in mutant worms (Fig. ​ (Fig.4C). 4C ). A gel mobility shift assay, using nuclear extracts prepared from mix-stage populations of C. elegans wild type and nfi-1 mutant worms confirmed the loss of CeNFI DNA-binding activity in the mutant worms (Fig ​ (Fig4D). 4D ). Thus, this mutation in the nfi-1 gene is a null allele. The survival of worms homozygous for the null allele shows that nfi-1 is not essential for worm survival. The nfi-1 mutant allele was backcrossed 12 times to the wild-type N2 strain to remove unwanted mutations prior to assessment of the phenotype.

          A) Deletion in C. elegans nfi-1 gene. The relative locations of confirmed exons (boxes) are shown. Square brackets indicate the location of the NFI-1 DNA-binding domain and the region deleted in nfi-1 mutant. The arrows indicate the position of the Tc1 insertion in an intron of the nfi-1 gene in strain NL747 pk240 and locations of PCR primers used in the screening for the deletion. B) Single-worm PCR reactions on N2 worm and nfi-1 homozygous mutant isolated by sib-selection. The arrows indicate a 1007 bp and 2969 bp PCR products corresponding to the nfi-1 mutant (qa524) and wild type alleles respectively. Lane 2 is 1 kb DNA ladder. C) RT-PCR on N2 worms and nfi-1 homozygous mutants. The arrows indicate a 480 bp and 320 bp RT-PCR products amplified using total RNA obtained from N2 worms (lane 1). nfi-1 mutants show loss of nfi-1 transcripts (lane 2). Lane (3) is 100 bp DNA ladder. D) Loss of NFI DNA-binding activity in extracts of nfi-1 mutants. Nuclear extracts of a mixed population of N2 worms and nfi-1 mutants were prepared and used in a gel mobility shift assay with an oligonucleotide (2.6) that contains an NFI-binding site (lanes 3, 5) or the same oligo with a single point mutation that abolishes NFI binding (C2) (lane 4, 6). See Fig. 2A for sequences of oligonucleotides. Extract of nfi-1 mutants show loss of NFI DNA-binding activity (lanes 5, 6). Lanes 1 & 2, no extract.

          Phenotype of nfi-1 mutant worms

          Locomotion defect

          Loss of nfi-1 results in a body movement defect (Unc, uncoordinated). nfi-1 mutant animals are fairly active and healthy, but are sluggish and flaccid at rest and slightly longer and thinner than N2 worms (Fig. 5A,D ). While wild type worms usually move in straight long lines, nfi-1 mutant worms often change direction abruptly (Fig. 5B,E ). The nfi-1mutant worms produce less regular tracks on the bacterial lawn with higher amplitude, and sometimes are slightly coiled when compared to wild-type worms (Fig. 5C,F ). This phenotype is more severe in older adults.

          Locomotion in nfi-1 mutants. Single young adults were spotted in the center of fresh plates and left for 10 min. (A, D) Photographs of N2 worms and nfi-1 mutants (B, E) Track patterns of N2 worms and nfi-1 mutants (C, F) Track patterns of N2 worms and nfi-1 mutants with higher magnification. Note less regular tracks in nfi-1 mutant vs. N2 worms.

          Egg-laying defect (Egl)

          17�% of older nfi-1 mutants have a "bag of worms" phenotype, where the mother is unable to lay fertilized eggs and fills with hatched progeny (Fig. ​ (Fig.6A). 6A ). Appearance and severity of this egg-laying phenotype correlates with the progressive locomotion defect, as young adult nfi-1 mutant worms do not bag. In young adults, serotonin stimulated egg-laying in both the wild type and nfi-1 mutant animals to a similar extent (data not shown), indicating that the postsynaptic response to serotonin is normal and the contractile apparatus for egg-laying is intact in nfi-1 mutant worms.

          A) Egg-laying defect in nfi-1 mutants and transgenic rescue. Bagging was measured in wild-type N2, nfi-1, N2 worms carrying the transgenic array qaEx507(N2 qaEx507) and nfi-1 worms carrying this array (nfi-1 qaEx507). Bars represent % of bagging as the mean of 3𠄴 independent experiments and error bars show the standard deviation. 30� worms of each genotype were scored in each independent experiment. N2 and N2 qaEx507showed υ% bagging. The nfi-1 mutant worms showed

          30% bagging while the rescued nfi-1 qaEx507showed υ% bagging. B) Shortened life span in nfi-1 mutants and transgenic rescue. Survival curves for the strains described above N2 (n = 57), nfi-1 (n = 58), N2 qaEx507 (n = 52) and nfi-1 qaEx507 (n = 31) are shown. Kaplan-Meier analysis (SPSS11 software) was use to determine median, percentile and p values (log rank test) and Excel was used to construct survival curves. The array generated

          50% rescue of the life-span. The experiment was repeated twice with similar results.

          Life-span reduction

          The nfi-1 mutant worms have a median life span of 10.00 ± 0.69 days as compared to 14.00 ± 0.76 days for wild type worms (p < 0.001) (Fig. ​ (Fig.6B). 6B ). Mutant worms become progressively more sluggish and flaccid as they age. We are currently investigating whether this lifespan reduction is due to the apparent progressive muscle weakness and partial paralysis seen or to a direct effect on known ageing pathways [43-45].

          Pharyngeal pumping rate defect

          Since the Unc and Egl phenotypes of nfi-1 mutant worms could reflect aberrant muscle function, we examined another process that can influenced by muscle defects, pharyngeal pumping rate. Pharyngeal pumping rates are reduced in nfi-1 mutant worms (Table ​ (Table1, 1 , Days 1𠄴), with more severe reductions in older vs. younger adult animals (Table ​ (Table1, 1 , Days 3ɤ vs. Day 1).

          Table 1

          Rescue of pharyngeal pumping defect by nfi-1 transgene

          Day a N2nfi-1N2 qaEx507 b nfi-1 qaEx507 b
          1 (n = 10)241 ± 5228 ± 9 c 243 ± 5241 ± 8 e
          2 (n = 12)242 ± 10187 ± 64 d 237 ± 8195 ± 78 d
          3 (n = 12)214 ± 18125 ± 79 c 183 ± 78200 ± 72
          4 (n = 12)160 ± 6292.5 ± 86 d 183 ± 79133 ± 90

          a Pharyngeal pumping was counted for one minute starting on day one, when animals first reached adulthood, for four consecutive days.

          b N2 and nfi-1 mutant worms were transformed with a transgene containing the nfi-1 gene and upstream promoter and rol-6 (qaEx507).

          c This value is statistically significantly different from N2 (P < 0.01)

          d This value is statistically significantly different from N2 (P < 0.05)

          e This value is statistically significantly different from nfi-1 (P < 0.01)

          Rescue of nfi-1 mutant with nfi-1 transgene

          Transgenic rescue was performed to test whether loss of the nfi-1 gene was responsible for the observed phenotypes. Transgenic strain XA512 qaEx507 was made by injecting a plasmid containing a 10 kb region of genomic DNA including the nfi-1 coding region and

          4 kb of upstream promoter region together with a rol-6(gf) expressing plasmid into the gonads of N2 worms. We crossed the resulting transgenic array into nfi-1 mutant worms to produce strain XA550 nfi-1(qa524) qaEx507. Egg-laying was completely rescued in nfi-1 qaEx507 worms when compared to the nfi-1 mutant worms (Fig. ​ (Fig.6A). 6A ). In addition, the median life span in nfi-1 qaEx507 worms of 13.0 ± 0.7 days was significantly longer than the 10.0 ± 0.7 day life span of nfi-1 mutant worms (p < 0.05), but slightly less than the 14.0 ± 0.8 day life span of N2 worms (Fig. ​ (Fig.6B). 6B ). The N2 qaEx507 stain used as a control has a 14.00 ± 0.49 day median life span, identical to that in non-transgenic N2 worms. The pharyngeal pumping defect was also partially rescued by transgenic expression of nfi-1 (Table ​ (Table1, 1 , nfi-1 qaEx507 vs. nfi-1). The nfi-1 transgene had little or no effect on pumping rates in N2 worms (N2 qaEx507 vs. N2). These data provide a well-defined developmental system affected by loss of nfi-1 that can be examined for cell-autonomous or inductive roles of nfi-1. The presence of the rol-6 marker gene prevented scoring of rescue of the locomotion phenotype in nfi-1 qaEx507 worms.

          Our in situ hybridization data indicate that nfi-1 transcripts are provided maternally. To test whether maternal nfi-1 transcripts could rescue the nfi-1 Egl phenotype, nfi-1 qa524/qa524, nfi-1 qa524/+ and +/+ progeny of heterozygous nfi-1 qa524/+ parents were scored for the egg-laying defect (Fig. ​ (Fig.7). 7 ). Bagging was seen in 41% of the resulting nfi-1 qa524/ qa524 worms, in 20.7% of nfi-1 qa524/+ worms but in σ% of +/+ animals. These data indicate the absence of maternal transcript rescue and possible haploinsufficiency at the nfi-1 locus. Thus, transgenic replacement of nfi-1 yields either partial (pumping rate and lifespan) or complete (egg-laying) rescue of the phenotypes seen in the nfi-1 mutant worms whereas maternal nfi-1 transcripts are insufficient to rescue the egg-laying defect.

          Haploinsufficiency of nfi-1 locus. Egg-laying defect in progeny of nfi-1 heterozygous mutant animals are shown. Bagging was scored in all progeny of two nfi-1(qa524/+) heterozygous worms (n = 146) derived from eggs laid over 7 hours. All worms were genotyped by single-worm PCR. Bars represent % of bagging in wild type (+/+), nfi-1 heterozygous (qa524/+) and nfi-1 homozygous worms (qa524/qa524). The number of worms scored of each genotype are shown above the bars.

          Ce titin expression is reduced in nfi-1 mutant worms

          We used cDNA microarrays to identify genes whose expression is affected by loss of nfi-1. Such genes could be either direct or indirect targets of nfi-1. Poly A+ RNA was purified from wild type and nfi-1 mutant synchronized gravid adults, labeled, and used to probe DNA microarrays containing

          17,000 C. elegans genes (Stanford Microarray Database). We analyzed RNA from gravid adults because the phenotype differences between nfi-1 mutant and wild type animals are clearer in adults than at earlier stages. The nfi-1 gene was scored as the most down-regulated gene in nfi-1 mutant worms in all experiments (data not shown). Several dozen genes showed small apparent reductions or increases in levels (2𠄳 fold) in mutant adults (data not shown) but only one gene, C. elegans titin (Ce titin) showed larger changes.

          Ce titin (also known as tag-58, t emporarily a ssigned g ene 58) was predicted to be 5.7-fold lower in mutant worms by microarray analysis. Quantitative PCR confirmed that Ce titin is reduced 8� fold in adult nfi-1mutant worms (Fig. ​ (Fig.8). 8 ). To date this is the gene that shows the largest decrease in expression in nfi-1 mutant worms. A search of the Ce titin gene reveals no overabundance of NFI binding sites (data not shown). Also, a CeTPro transgene expressing a translational fusion of GFP to the 5'-end of the Ce titin gene [46] appears to be expressed at similar levels in WT and nfi-1 mutant worms (data not shown). It will be important in future studies to determine whether Ce titin is a direct or indirect target of nfi-1 and the possible role of Ce titin in the phenotypes observed.

          Down regulation of Ce titin expression in nfi-1 mutants assessed by QPCR. Bars represent fold changes in Ce titin transcript level in wild type N2 vs. nfi-1 mutant worms. RNA samples were obtained from 3 independent synchronized adult worm populations for each genotype.

          Expression of nfi-1::GFP reporter transgenes

          One limitation of in situ hybridization in C. elegans is that in older embryos and postembryonic stages it is sometimes not sensitive enough to unambiguously identify individual cells. Since we saw little or no nfi-1 expression in muscle by in situ hybridization, in an effort to develop a more sensitive assay for nfi-1 expression we constructed two GFP-reporter transgenes (Fig. ​ (Fig.9A). 9A ). In Pro1CeNFI::GFP, 4 kb of genomic DNA sequence upstream of the nfi-1 open reading frame and the sequence encoding the first four residues of the CeNFI protein was fused in frame to GFP. In Pro2CeNFI::GFP, the 4 kb promoter region and the sequence encoding the first 94 residues of CeNFI has been fused to GFP. GFP expression for both transgenes was detected in embryos (Fig. ​ (Fig.9B). 9B ). Faint GFP expression is first detected at the late gastrulation stage of embryogenesis (𾌀 cells) as a diffuse green glow throughout the embryo. By the comma stage expression is detected in many cells along the outer edge of the embryo and expression continues through embryogenesis and is detected in L1-L4 larvae in many of the same cells as in adults. Adult transgenic animals show GFP expression in muscles, neurons and intestinal cells (Fig. 9B–I ). Among the muscles, fluorescence was strongest in the pharynx and head muscles, was observed with less frequency in other body wall muscles and was seen occasionally in vulva muscles. Expression was also seen in two pairs of neurons located near the posterior bulb of the pharynx, and in several as yet unidentified tail neurons. Expression patterns in multiple transgenic lines from each reporter were similar with the exception that Pro2CeNFI::GFP expression was detected more consistently in head neurons and Pro1CeNFI::GFP was seen with higher frequency in body-wall muscles. However expression of both transgenes was mosaic, showing expression in only subsets of cells and animals in each population. Since mosaic expression of GFP was seen in transgenic strains from both arrays, transgenic array Pro1CeNFI::GFP was integrated by γ-irradiation. However similar mosaic expression was seen with the integrated array (data not shown). These data may indicate that additional elements are needed for stable regulation of nfi-1 expression and that such elements may be located further downstream in the nfi-1 genomic sequence.

          Expression pattern of the nfi-1::GFP reporter transgenes. The nfi-1 locus and structure of the nfi-1-GFP fusion constructs are shown (A). Nfi-1 coding regions are shown in black, gray boxes indicate alternatively spliced exons, untranslated regions are in white. GFP is shown as a hatched box. Expression of nfi-1-GFP reporter constructs was observed in embryos (B), intestinal cells (C), body wall muscles (D, E), pharynx (F), egg-laying muscles (G), several head (H) and tail neurons (I). Expression was assessed using a FXA Nikon microscope (B-D, F-I) and a Bio-Rad confocal microscope (E).


          Robustness of Optimization Results to Small Variations of Parameters

          To determine the robustness of the wire-minimized solution, we explored several aspects of the cost function and assessed their impact on the ability to predict neuronal layout.

          First, we analyze the sensitivity of the wire-minimized layout to the normalization coefficient α and the exponent ζ. As mentioned, our cost formulation accounts for multiple synapses on a given neurite by normalizing connection weights by the average number of synapses per neurite (α = 29.3). We test how the predicted layout changes by varying α between 1 and 45. Because the choice of the quadratic form of the cost function may seem arbitrary, we also varied the power of wire length in the cost function, ζ in Eqs. 2 and 3 between values of 1 and 4. As argued previously, the wiring cost is likely to scale supralinearly (ζ > 1) with distance between neurons (26). If so, the minimization problem is convex and can be efficiently solved numerically. The lowest mean deviation, 9.71%, is achieved by using the cost function with normalization coefficient ≈27 and exponent ≈2 (see General Power-Law Cost Function in Supporting Text and Fig. 6, which are published as supporting information on the PNAS web site). Interestingly, these values are close to those chosen from biological considerations and validate the quadratic cost function.

          Second, we test the importance of synaptic multiplicity between neurons. Instead of a wire dedicated to each synapse between cells (Fig. 1 C Inset), we use a single wire to connect a given pair of neurons regardless of the number of synapses (Fig. 1 D Inset). In other words, we minimize the quadratic cost function with a binary connection matrix (only 0 or 1 elements in the matrix A from Eq. 2 ). Using ζ = 2, the lowest mean deviation between predicted and actual position (9.82%) is higher than the result from a synapse-number weighted cost function and was found at α = 8. In the actual worm, the average number of synaptic partners (as opposed to individual synapses) per neurite is 12.2, close to the optimal value of α obtained from the binary connection matrix.

          To summarize, we find that various reasonable cost functions predict neuronal placement incomparably better than the random one. Although mean deviations vary somewhat between different cost functions, they are not far from the best known solution. Thus the wire length minimization approach is rather robust. Because the quadratic cost function can be solved exactly and is reasonably close to the best-known solution, it may serve as the reference predicted layout. Although the predicted placement is only approximately correct, we recall that the problem was solved in one dimension. Such dimensionality reduction may introduce errors on the order of the inverse aspect ratio of the worm, just under 10%. Because the mean deviations we report approach this range, wiring optimization results are encouraging.


          Discussion

          C. elegans dietary choice behavior

          Here, we describe novel behavioral paradigms in C. elegans, food seeking and food preference. We have identified five worm bacterial foods to establish a range of food quality as measured by the food's ability to support growth (see Fig. S1 in supplementary material). Remarkably, an animal as simple as C. elegans can exhibit dietary choice(Fig. 1). Worms preferred high quality food, i.e. that better supported growth. This choice developed with time, suggesting that animals needed to try the food to make a decision. Using eat mutants, we showed that this choice requires food assessment via feeding. This is similar to mammals, which are also capable of selecting food that better supports growth, and try foods before making a choice(Osborne and Mendel, 1918 Young, 1941).

          AIY neurons function to extend food-seeking periods. Trajectory on mediocre food, E. coli DA837, of (A) a ttx-3 mutant and (B) an animal whose AIY neurons have been killed. Compare to wild-type in Fig. 4C. ttx-3 mutant trajectories did not span the whole lawn and there were far fewer long straight roaming events. Trajectories of AIY worms also had fewer straight long movements than wild-type controls. (C,D) Movement duration distribution of (C) wild type, ttx-3, osm-6, osm-6ttx-3 and (D) AIY-ablated animals, all tested on E. coli DA837 food. N=10 for WT, 10 for ttx-3, 6 for osm-6, 6 for osm-6ttx-3, 10 for AIY ablations and 8 for ttx-3p::GFP controls. (E) ttx-3 was defective in the food preference behavior if bacterial foods were located at a small distance from each other. By 3 h, all ttx-3 worms found food,but there was no preference in the harder arrangement. In contrast to ttx-3, osm-6 animals took longer to discriminate between good and bad food, but they finally managed to make the right choice even if foods were located at a distance. Values are means ± s.e.m. * Different from the wild type (P<0.01) † different from ttx-3 (P<0.01 Student's t-test). (F) Biased food preference for E. coli HB101 over B. megaterium of mutants and animals with laser-ablated neurons. The fraction of animals that reached the central colony of good food, E. coli HB101, was determined. ttx-3 mutants and AIY-ablated animals performed worse than controls. In laser ablation experiments, worms were counted after 20 h. For tests on mutants, the number of assays is 18 for WT, 15-17 for ttx-3 alleles and 6-15 for various mutants tested. For laser ablations, number of worms found in the center and the total number of worms tested is indicated next to the bars. Values are means ± s.e.m. * Different from the wild type (P<0.01 Student's t-test) † Different from the ablation control(P<0.01 χ 2 test of independence).

          AIY neurons function to extend food-seeking periods. Trajectory on mediocre food, E. coli DA837, of (A) a ttx-3 mutant and (B) an animal whose AIY neurons have been killed. Compare to wild-type in Fig. 4C. ttx-3 mutant trajectories did not span the whole lawn and there were far fewer long straight roaming events. Trajectories of AIY worms also had fewer straight long movements than wild-type controls. (C,D) Movement duration distribution of (C) wild type, ttx-3, osm-6, osm-6ttx-3 and (D) AIY-ablated animals, all tested on E. coli DA837 food. N=10 for WT, 10 for ttx-3, 6 for osm-6, 6 for osm-6ttx-3, 10 for AIY ablations and 8 for ttx-3p::GFP controls. (E) ttx-3 was defective in the food preference behavior if bacterial foods were located at a small distance from each other. By 3 h, all ttx-3 worms found food,but there was no preference in the harder arrangement. In contrast to ttx-3, osm-6 animals took longer to discriminate between good and bad food, but they finally managed to make the right choice even if foods were located at a distance. Values are means ± s.e.m. * Different from the wild type (P<0.01) † different from ttx-3 (P<0.01 Student's t-test). (F) Biased food preference for E. coli HB101 over B. megaterium of mutants and animals with laser-ablated neurons. The fraction of animals that reached the central colony of good food, E. coli HB101, was determined. ttx-3 mutants and AIY-ablated animals performed worse than controls. In laser ablation experiments, worms were counted after 20 h. For tests on mutants, the number of assays is 18 for WT, 15-17 for ttx-3 alleles and 6-15 for various mutants tested. For laser ablations, number of worms found in the center and the total number of worms tested is indicated next to the bars. Values are means ± s.e.m. * Different from the wild type (P<0.01 Student's t-test) † Different from the ablation control(P<0.01 χ 2 test of independence).

          Next, we showed that C. elegans left hard-to-eat bacteria(Fig. 2), and, like food preference, this behavior required food quality assessment. Previously, it was generally thought that once C. elegans finds food, it stays there and eats until death or until the source is exhausted, although Lipton et al.(2004) have found recently that adult males leave food in search of hermaphrodites. Leaving experiments showed that even after food is found, the animal could decide to stay in the food or leave, and this decision was based on the assessment of whether the food was good or bad. Leaving behavior is a compromise: on the one hand, the worm risks losing the food that has already been found and ending up in an adverse environment, or, on the other hand, there is a chance of finding even better food.

          Leaving behavior that we describe here is somewhat related to the known phenomenon of adaptation to a volatile or soluble attractant. Upon extended exposure to an odor (typically, 1 h or more) in the absence of food, odortaxis to this odor dwindles (Colbert and Bargmann, 1995 Colbert and Bargmann, 1997). Likewise, attraction to a soluble chemical switches to avoidance after 3-4 h exposure(Saeki et al., 2001) in the absence of food. In view of our results, adaptation is an increase of a food-seeking behavior because of the lack of reinforcement. Consistent with this, Nuttley et al. (2002)have shown that in the presence of food, chemoattraction is suppressed. Also,aerotaxis fades in the presence of food(Gray et al., 2004). And, if animals are conditioned to an odor or taste in the presence of food, no adaptation occurs. If worms are adapted to the stimulus in the absence of food, but then briefly exposed to food, chemoattraction robustly revives(Nuttley et al., 2002 Saeki et al., 2001). These and our data suggest the food feeds back on behavior after it is eaten and acts as a reinforcer in C. elegans.

          Effect of previous dietary experience

          If bacterial food was switched from good to mediocre, C. elegansappetitive behavior was increased compared to animals that had not experienced good food (Fig. 3). Previous experience of good food made worms more risk-loving, more willing to explore.

          In the leaving experiment (Fig. 3C), the time course of the effect of experience could be observed: the enhancement of leaving was not high at the very beginning, but reached a maximum after 0.5-1 h. This indicates that time was needed to assess new conditions, followed by comparison and output. If a worm is taken off food, there is a period of about half an hour of area-restricted search with frequent reversals and turns, followed by active `running', when reversals and turns are suppressed (Gray et al.,2005 Hills et al.,2004). This time, however, is much shorter than that required to deplete fat stores, which is about 6 h(McKay et al., 2003). Probably for worms, which feed continuously throughout life to support a 3-day life cycle, even brief food deprivation or brief decline in food quality is an alert signal that motivates them to explore the environment.

          One might propose a trivial explanation for the results in Fig. 3: well-fed worms are simply healthier and explore the environment more actively than unfed ones,which strive to save energy. We think this is unlikely, because in the continuous presence of good food, both leaving behavior(Fig. 2) and exploratory activity (Fig. 4) were suppressed, while in the continuous presence of poor food, worms were very active. Therefore, it was the switch from good to bad that causes an increase in exploratory behavior.

          At least two other explanations are possible. First, experiencing different quality foods changes worms' satiation (or hunger) status, and that, in turn,affects their food choice and leaving behaviors. More hungry animals tend to accept any first food they encounter, thus their leaving behavior and food choice is less pronounced. Less hungry (more satiated) animals, on the contrary, tend to be more particular about food and their exploratory behavior is more active. Another explanation invokes memory: worms may learn that previous conditions were associated with a different satiation status and compare those with the new conditions. These two mechanisms are not mutually exclusive and may function in parallel.

          In humans, powerful diet preferences can form, especially for nutritious fat- and carbohydrate-rich foods, such as sodas, desserts, pizza, etc. These feeding habits are hard to change. Dieting often results in `food craving',`carbohydrate craving', and `binge eating', which eventually lead to even further increase of food consumption(Capaldi, 1996). These phenotypes are analogous to the increased food-seeking behavior when the food is switched from good to bad in C. elegans. (Of course, the time scale has to be normalized to the life span.) These behaviors are adaptive in the wild, where good food is usually scarce, but in developed countries, where high quality food is always easily available, they contribute to overeating and obesity.

          The C. elegans food-seeking strategy

          C. elegans locomotion, in particular the equilibrium between roaming, rapid straight movement, and dwelling, slow movement with frequent reversals and stops, was affected by the food source(Fig. 4). On poor food,straight rapid movement, called roaming, was drastically increased, while dwelling predominated on high quality food.

          In previous studies, it has been shown that the speed of C. elegans locomotion is increased in the absence of food, and reversals and turns are suppressed after about 30 min in the absence of food(Gray et al., 2005 Hills et al., 2004). Here, we showed that this also happened on food, if the food was hard to eat. This suggests that it is not the mere presence of food that is decisive in regulating C. elegans locomotion, but food quality.

          We identified mutants defective in food preference behavior and two of them, osm-6 and ttx-3 were also defective in roaming. TTX-3 is a transcription factor required for the differentiation of the AIY thermosensory interneuron. The defects of the ttx-3 mutant were partially reproduced by killing AIY. The latter also caused a decrease in duration of roaming periods, suggesting that AIY functions to suppress the roaming-to-dwelling transition and to extend the food-seeking periods. This is consistent with, and extends, the results of other reports, which demonstrated that AIY suppresses reversals and turns(Gray et al., 2005 Tsalik and Hobert, 2003).


          The authors would like to thank Ben King, PhD, and Joel Graber, PhD, for bioinformatics advice and George Sutphin, PhD, for constructive criticism of the manuscript. This work was supported by grants from the National Institute on Aging of the National Institutes of Health (R21AG056743) and by the Ellison Medical Foundation (AG-NS-1087-13), both to AN Rogers. In addition, this project was supported by Institutional Development Awards (IDeA) from the National Institute of General Medical Sciences of the National Institutes of Health (grant numbers P20GM0103423 and P20GM104318, respectively). Some strains were provided by the CGC, which is funded by the NIH Office of Research Infrastructure Programs (P40 OD010440).

          Author Contributions

          JA Rollins: data curation, formal analysis, methodology, writing—original draft, writing—review and editing.

          P Kapahi: conceptualization.

          AN Rogers: conceptualization, funding acquisition, writing—review and editing.


          RESULTS

          Unc-64 Encodes a Syntaxin Homolog

          unc-64 was first described by Brenner (1974) as a mutant with locomotory defects. unc-64 was suspected to encode a component regulating synaptic transmission because it shares a common phenotype with many other synaptic mutants in C. elegans: resistance to the acetylcholinesterase inhibitor aldicarb ( Alfonso et al., 1993 , 1994 Nonet et al., 1993 , 1997 , 1998 Nguyen et al., 1995 Miller et al., 1996 Iwasaki et al., 1997 ). We isolated a C. elegans gene with strong similarity to rat syntaxin 1 that mapped to a region of chromosome III on the physical map, roughly corresponding to the genetic map position of the unc-64 gene (see Figure ​ Figure1 1 and MATERIALS AND METHODS for details). We tested whether the syntaxin-like gene was encoded by unc-64 using germline transformation ( Mello et al., 1991 ). unc-64(e246) mutants harboring a genomic clone of the C. elegans syntaxin gene were phenotypically rescued to wild-type, as determined by both behavioral and aldicarb resistance assays. Additionally, sequencing of the coding region of the candidate syntaxin gene from unc-64 mutant alleles in each case revealed a molecular lesion that could account for the genetic defect (Figure ​ (Figure1A). 1 A). Thus, we conclude that unc-64 encodes a C. elegans syntaxin homolog.

          Three Distinct Forms of Syntaxin Are Encoded by unc-64

          We isolated cDNA clones derived from the unc-64 syntaxin locus. Several complete cDNA clones representing one transcript from the C. elegans syntaxin locus were isolated from an embryonic cDNA library, and partial cDNAs representing two additional transcripts were isolated from first-strand cDNA using the PCR. Each transcript is predicted to encode a product with 63% or 64% identity to human syntaxin 1A. The three transcripts differ only in the last of eight coding exons and thus encode products with identical cytoplasmic domains but different transmembrane anchor domains (Figure ​ (Figure1, 1 , A and B). We refer to these three products as the syntaxin A, B, and C products of the unc-64 locus. At least three distinct transcripts of approximately 1.1 kb, 2.6 kb, and 3.8 kb are present on Northern blots of mixed-staged RNA (Figure ​ (Figure1C). 1 C). The size of the full-length syntaxin B cDNA isolated from the cDNA library is of the appropriate size to correspond with the smaller transcript. The larger transcripts are of an appropriate length to represent the syntaxin A and C cDNAs if they utilize the same polyadenylation site. The splicing of the transmembrane domain of syntaxin in C. elegans is reminiscent of the splicing pattern observed for the rat syntaxin 2 and the mouse syntaxin 3 genes ( Bennett et al., 1993 Ibaraki et al., 1995 ).

          Unc-64 Syntaxin Is Expressed in the Nervous System and Secretory Tissues

          The C. elegans syntaxin sequences are most similar to the vertebrate neuronal-specific syntaxin 1 molecules and hence are likely to be expressed in the nematode nervous system. To examine the expression pattern, antisera were raised against bacterially expressed C. elegans syntaxin fusion protein. An affinity-purified anti-syntaxin antiserum was incubated with fixed whole adult animals, and the antibodies were detected using FITC-conjugated secondary antisera. As expected, syntaxin was detected in the nervous system (Figure ​ (Figure2). 2 ). Syntaxin immunoreactivity was very strong and distributed uniformly along the major process bundles: the nerve ring (Figure ​ (Figure2A), 2 A), ventral cord (Figure ​ (Figure2B), 2 B), and dorsal cord (Figure ​ (Figure2A). 2 A). Syntaxin immunoreactivity also was detected in the vast majority of neuronal cell bodies as well as in commissural and dendritic processes (Figure ​ (Figure2B). 2 B). By contrast, the vesicle-associated markers synaptobrevin, synaptotagmin, and RAB-3 are restricted to the synaptic-rich regions of the nervous system where they show a punctate, rather than uniform, appearance ( Nonet et al., 1993 , 1997 , 1998 ). Syntaxin was also detected in the pharyngeal nervous system (Figure ​ (Figure2A). 2 A). Furthermore, syntaxin was detected outside the nervous system. Immunoreactivity was strong in the uv1 cells of the vulva, in the intestine (Figure ​ (Figure2C), 2 C), and in the spermatheca (Figure ​ (Figure2D). 2 D). Detectable levels of syntaxin protein were not observed in other tissues. Our protein expression data suggest that unc-64 encodes the homolog of the vertebrate syntaxin 1 gene ( Bennett et al., 1992 , 1993 ).

          Expression of unc-64 syntaxin. (A𠄽) Whole adult wild-type hermaphrodite worms fixed and stained with α-UNC-64 primary antibodies and visualized with FITC-conjugated antibodies. (A) Lateral view of the head region showing immunoreactivity in the nerve ring, dorsal cord, and pharyngeal nervous system. (B) Ventral view of the midbody region showing expression in ventral cord axons, neuronal cell bodies, and in commissural and sublateral processes. (C) A lateral view of the midbody showing UNC-64 immunoreactivity on the basolateral surface of the intestine. (D) A lateral view of the midbody showing immunoreactivity in the hermaphrodite spermatheca. The male vas deferens also stains. (E) A lateral view of the vulva viewed using differential interference contrast microscopy. (F) Same plane of focus as panel E showing fluorescence from the UNC-64 A::GFP product in the uv1 cells of the vulva. (G–H) Whole adult wild-type animals fixed and stained for LacZ gene activity expressed under the control of the unc-64 promoter. LacZ activity assayed using X-gal as a substrate. The lacZ gene contains a nuclear localization signal. (G) Lateral view of the head region showing expression of unc-64 ::lacZ in neuronal and intestinal nuclei. (H) Lateral view of the midbody of an adult hermaphrodite showing unc-64::lacZ expression in intestinal nuclei, neuronal nuclei, and the spermatheca. Scale bar, in all panels, 20 μm.

          As it was not feasible using immunohistochemistry to determine which neurons expressed syntaxin, the expression pattern of the SNARE gene was examined using a translational lacZ fusion. A genomic syntaxin fragment (Figure ​ (Figure1C) 1 C) was inserted into a plasmid containing a nuclear-localized lacZ reporter gene ( Fire et al., 1990 ). As expected, syntaxin was expressed in the vast majority of all neurons (Figure ​ (Figure2, 2 , G and H). In addition, syntaxin expression was high in intestinal nuclei, in the spermatheca, and in the uv1 cells of the vulva (Figure ​ (Figure2H). 2 H). In summary, C. elegans syntaxin is expressed in both secretory and neuronal tissues. In neurons, the protein is not specifically associated or concentrated at synaptic release sites rather, it is ubiquitously distributed.

          The three identified products deriving from the C. elegans syntaxin locus differ only in those sequences that are predicted to lie in the lipid bilayer and act as transmembrane anchors. Because production of antibodies capable of distinguishing these products did not seem feasible, we engineered constructs that tagged each product by appending the GFP-coding region to the last codon of each predicted transmembrane domain. The constructs were introduced into wild-type animals by germline transformation, and the expression pattern of the GFP-tagged syntaxins was examined. Syntaxin A and C were expressed in neurons and in five nonneuronal tissues. By contrast, expression of syntaxin B was limited to the five nonneuronal cell types. In neurons, syntaxin A and C were found ubiquitously on neuronal cell bodies, axons, and dendrites. Also stained were the uv1 secretory cells of the vulva, the excretory gland cells, the distal tip cells, the spermatheca, and the intestinal muscle (Figure ​ (Figure2, 2 , E and F). In the uv1 cells, GFP fluorescence was concentrated subcellularly (Figure ​ (Figure2, 2 , E and F). The expression pattern from the GFP, lacZ, and antibody studies were consistent, except that GFP reporter expression was observed in the intestine only with syntaxin C fusions. This expression was limited to the posterior half of the intestine. It is possible that sequences that direct gut expression are absent from the GFP clones since sequences further upstream of the initiation codon are present in the lacZ constructs. In summary, the alternative UNC-64 syntaxin products are expressed in distinct but overlapping neuronal and nonneuronal secretory cells.

          Isolation of Additional unc-64 Syntaxin Mutants

          Additional alleles of the unc-64 gene were isolated in three independent screens. md130 and md1259 were isolated in a screen for aldicarb- resistant mutants occurring spontaneously in a strain with high levels of Tc1 transposition ( Miller et al., 1996 ). js21 was isolated in a general screen for aldicarb-resistant mutants using EMS as a mutagen (M.L.N., unpublished data). Finally, two lethal alleles, js115 and js116, were isolated in a noncomplementation screen (see MATERIALS AND METHODS). Molecular characterization revealed that the e246 and js21 lesions are missense mutations that result in alanine-to-valine substitutions at codons 241 and 248, respectively. These codons are in the H3 domain of syntaxin, a region thought to assume an amphipathic helical structure capable of forming coiled-coil interactions with other proteins ( Hardwick and Pelham, 1992 Calakos et al., 1994 Fasshauer et al., 1997 ). The point mutations in md130 and md1259 disrupt splicing sites (see MATERIALS AND METHODS for details). The js116 mutation results in premature termination in the transmembrane domain of the syntaxin A product, while the js115 lesion results in premature termination at codon 71. Immunohistochemical examination of whole fixed animals revealed that syntaxin product remained in the nervous system of all viable mutants, although the level of syntaxin was reproducibly lower in the md1259 mutant. UNC-64 staining was not detected in js115. All available evidence suggests that js115 represents the null phenotype (see MATERIALS AND METHODS).

          Behavioral Defects of unc-64 Syntaxin Mutants

          All six unc-64 mutants exhibit locomotory abnormalities. The lethal mutants exhibit the most severe motor defects. js115 animals are virtually completely paralyzed and rarely maintain a sinusoidal posture, although they occasionally will make slow head movements (Figure ​ (Figure3). 3 ). js116 animals exhibit very slow motor movements (Figure ​ (Figure3) 3 ) and tend to adopt a coiled position. Both js115 and js116 animals arrest development just after hatching and eventually die as L1 larvae. Behaviorally, the viable js21 and e246 mutants are very lethargic, although they retain the ability to move briefly when prodded. Locomotory abnormalities were quantified by examining the rate of movements of animals imaged using a charge-coupled device camera (Table ​ (Table1). 1 ). Despite exhibiting severe locomotory defects, other behaviors of the hypomorphic mutants were only weakly affected (Table ​ (Table1). 1 ). The rate of pharyngeal pumping was relatively normal in the hypomorphic mutants e246 and js21, although pumping was completely abolished in the js115 null mutant. The motor defecation program was only slightly abnormal in hypomorphs, but completely abolished in js115. It is interesting to note that, compared with other synaptic mutants with similarly severe locomotory defects (e.g., unc-18 and unc-13), the unc-64(e246) mutant has much milder deficiencies in other behaviors. unc-64(e246) mutants also exhibit a tendency to form enduring dauer larvae ( Iwasaki et al., 1997 ), a developmental decision regulated by sensory input ( Bargmann and Horvitz, 1991 ). Two other mutants, md130 and md1259, also exhibit milder behavioral defects, but minimal analysis was performed on these alleles due to the difficulty in determining, with certainty, the resulting gene products. Finally, at the morphological level, muscle, the intestine, the pharynx, the cuticle, and the organization of the nervous system appear normal in both the hypomorphs and the null mutant. In summary, a variety of genetic lesions in the syntaxin gene result in behavioral defects of varying severity. Presumably, all these phenotypes represent manifestations of an underlying impairment of synaptic transmission.

          Phenotype of lethal unc-64 syntaxin mutants. Bright field images of first larval stage animals of wild-type, unc-64(js115), and unc-64(js116) animals on an E. coli bacterial lawn at the indicated time intervals (wild-type in seconds and mutants in minutes). Scale bar, 100 μm.

          Table 1

          Behavioral defects unc-64 mutants

          StrainLocomotion Pharyngeal pumps/minDefecation
          Speed (μm/sec)Efficiency a Cycle time (s)EMC b
          N2 c 161 ±�0.63250 ±�42.0 ±𠂓.398%
          e24613 ±𠂘0.18209 ±�39.3 ±𠂓.387%
          js2151 ±�0.43234 ±�48.8 ±𠂙.280%
          js115 d 0NA0.2 ±𠂐.6NoneNA
          js116 d 0NA3 ±𠂒NoneNA

          Data shown as mean ± SD. A minimum of 18 animals were analyzed. NA, Not applicable. 

          Late Arrest of unc-64 Syntaxin Mutants Is Not the Result of Maternal Contribution

          In D. melanogaster, mosaic clones that lack syntaxin cannot be isolated, suggesting the molecule is required in all cells ( Schulze and Bellen, 1996 ). Furthermore, syntaxin is also required early in development during cellularization of the Drosophila embryo ( Burgess et al., 1997 ). As our most severe mutant arrests development after completing embryogenesis, we examined whether unc-64 syntaxin mutants completed embryogenesis using maternally derived products. We constructed a strain homozygous for js115, which also harbored an extrachromosomal array bearing a visible marker (rol-6), a muscle-specific unc-54::GFP construct, and the wild-type unc-64 syntaxin gene. These animals segregate both wild-type and mutant progeny as the array is not transmitted to all progeny. We examined the progeny of nine mosaic animals lacking syntaxin product in the germline. These animals were identified on the basis of their failure to produce viable progeny. Nearly 99% (n = 740) of the progeny of these mosaic animals completed embryogenesis and arrested as first larval stage animals with a phenotype indistinguishable from the js115 progeny of heterozygous mothers. The secreted cuticle of these animals appeared normal, containing well organized alae (Figure ​ (Figure4). 4 ). Most of the remaining 1% of progeny arrested late in embryonic development as threefold embryos. None of the progeny of the mosaic animals we examined expressed GFP, confirming that the unc-64(+) array was absent from the germline. We also examined viable mosaic animals that lacked unc-64 expression in muscle. Three mosaics that lacked GFP in muscle anterior of the vulva and 11 mosaics that lacked GFP in both dorsal body wall muscle quadrants all moved normally, suggesting that syntaxin is not required in muscle. Additionally, most of the severely uncoordinated mosaic animals we examined retained GFP expression in muscle, consistent with a neuronal site of action for syntaxin.

          Cuticle of L1 js115 syntaxin mutant animals. Lateral view of the cuticle structures (alae) of the L1 larvae of wild-type and homozygous unc-64(js115) animals segregating from a mosaic animal lacking unc-64(+) in the germline. Arrowheads point to the alae (cuticular specializations) observed in L1 larvae. Scale bar, 20 μm.

          In separate experiments, we injected antisense RNA into the germline of wild-type animals in an attempt to inactivate the unc-64 gene by RNA-mediated interference ( Guo and Kemphues, 1995 Rocheleau et al., 1997 ). This technique phenocopies the null phenotype of most genes that provide maternal contributions to the embryo, but is much less efficacious in phenocopying late-acting zygotic genes. Our injections failed to confer the js115 phenotype or an aldicarb-resistant phenotype. In summary, we find no evidence for either a maternal contribution of unc-64 or for a requirement for zygotic syntaxin in the early embryo.

          Synaptic Transmission Defects in unc-64 Mutants

          To assess neurotransmission in unc-64 mutants, we began by examining the effect of cholinergic pharmacological agents. We quantified the sensitivity of unc-64 mutants to aldicarb, a potentiator of released acetylcholine, and to levamisole, an agonist of nicotinic acetylcholine receptors in the nematode ( Lewis et al., 1980a , b ). All of the viable unc-64 mutants examined were resistant to high levels of aldicarb (Figure ​ (Figure5A), 5 A), as was the acetylcholine receptor mutant unc-29 ( Fleming et al., 1997 ). This strongly suggests that cholinergic transmission is reduced in these mutants. However, unlike unc-29, the unc-64 syntaxin mutants were sensitive to levamisole (Figure ​ (Figure5B), 5 B), suggesting an intact postsynaptic apparatus and thus a presynaptic defect. Finally, examination of the lethal mutations revealed that aldicarb enhanced js116 movement but failed to stimulate movement of js115 animals, providing evidence that cholinergic synaptic transmission may be completely abolished in the null mutant.

          Pharmacological properties of unc-64 mutants. (A) Aldicarb sensitivity of unc-64 mutants and the wild type. Shown is the percentage of adult animals paralyzed after a 4-h exposure to various concentrations of aldicarb on plates seeded with E. coli wild-type (•), js21(□), e246 (Δ), and unc-29 (○). (B) Levamisole sensitivity of unc-64 mutants and the wild type. Shown is the percentage of adult animals paralyzed after exposure to 100 μM levamisole at various times. Wild-type (•), js21(□), e246 (Δ), and unc-29 (○).

          To provide direct evidence of synaptic transmission defects in C. elegans syntaxin mutants, we used an extracellular recording technique developed by Raizen and Avery (1994) . This method permits the detection of postsynaptic potentials in pharyngeal muscle resulting from the action of both excitatory and inhibitory motor neurons. MC is an excitatory motor neuron thought to stimulate pharyngeal muscle contraction ( Raizen et al., 1995 ). Its activation is visualized as a single depolarizing transient preceding many pumps. In unc-64 syntaxin mutants, MC neuronal function appeared defective. Specifically, we observed a series of multiple excitatory postsynaptic potentials before initiation of pharyngeal pumps (arrows, Figure ​ Figure6), 6 ), which may reflect nonsynchronous transmitter release from MC. A similar phenotype has previously been observed in aex-3, rab-3, and snb-1 mutants ( Iwasaki et al., 1997 Nonet et al., 1997 , 1998 ).

          Pharyngeal recordings from wild-type and unc-64 mutant animals. Shown are characteristic electrophysiological recordings from the wild-type strain N2 (A), unc-64(e246) (B), unc-64(js21) (C), unc-64(md130) (D), aex-3(y255) (E), snb-1(js17) (F), and unc-64(js21) snb-1(js17) (G). Arrows indicate MC-induced transients, and filled circles indicate M3-induced transients. All traces are millivolts versus time.

          Our analysis concentrated on the examination of inhibitory postsynaptic potentials (IPSPs) produced by the motor neuron M3, a glutaminergic neuron that regulates the duration of pharyngeal muscle contractions ( Avery, 1993 Dent et al., 1997 ). In wild-type animals, M3 transients were regularly spaced and initiated 28 ± 5 ms after depolarization of the pharynx. In unc-64 (e246) animals, the first detectable IPSP often was greatly delayed (102 ± 43 ms Figure ​ Figure6). 6 ). This is most easily illustrated by examining the distribution of IPSPs as a function of time after the initiation of the pharyngeal pump (Figure ​ (Figure7, 7 , A and B). While IPSPs occurred relatively synchronously in wild-type, they were broadly distributed in e246 mutants. However, normalizing the time of the IPSP relative to the first IPSP within the pump reveals that, although the initial IPSP was variably delayed in the mutant, subsequent IPSPs were temporally synchronized (Figure ​ (Figure7, 7 , C and D). A second difference between wild-type and mutant was that the amplitude of M3 IPSPs was relatively constant in wild-type but increased substantially in e246. While in wild-type, the initial IPSP was slightly larger than the final IPSP, in e246 there was a threefold increase in IPSP amplitude by the end of the pump (Figure ​ (Figure7, 7 , C and D). js21 exhibited a similar but less severe defect (our unpublished data). These particular defects in M3 transmission are specific to the syntaxin helical mutants. The splicing mutants md130 and md1259 do not exhibit these alterations but instead resemble aex-3(y255), rab-3(y250), and snb-1(js17) in that their IPSPs were consistently decreased in amplitude yet initiated without delay after contraction of the pharynx (Figure ​ (Figure6). 6 ). In summary, lesions in the H3 domain of syntaxin affect both the kinetics of transmitter release and the amplitude of postsynaptic potentials.

          Analysis of M3 transmission in wild-type and unc-64(e246) animals. The amplitude of the first ([triof]), second (○), third (▪), fourth (Δ), fifth (•), and six (□) M3-induced IPSPs within each pump (see Figure ​ Figure6) 6 ) is plotted as a function of time of occurrence. M3 amplitudes are normalized relative to the mean R-phase amplitude for each record. (A and B) Time of occurrence is plotted relative to initiation of pharyngeal muscle excitation. (C and D) Time of occurrence is normalized to the first M3 IPSP of the same pump. (A) Each group of M3 IPSPs cluster showing the high temporal synchrony of M3 transmitter release in wild-type (pump duration 148 ± 23 ms [mean ± S.D.], N = 130 pumps). (B) In unc-64(e246) temporal synchrony is lost as the timings of each group of M3 IPSPs are distributed over a larger time span. The first M3 IPSP occurs from 22 to 212 ms after muscle excitation (pump duration 209 ± 44 ms [mean ± S.D.], N = 130 pumps). (C) Aligning the IPSPs by timing of the first M3 IPSP again shows temporal clustering in wild-type. Mean IPSP spacings are ֱst to 2nd] 34 ms, ֲnd to 3rd] 31 ms, ֳrd to 4th] 32 ms, ִth to 5th] 32 ms, and ֵth to 6th] 36 ms. The downward sloping linear regression reveals that subsequent wild-type IPSP amplitudes tend to decrease slightly. (D) Similar alignment to the first M3 IPSP in e246 reveals that subsequent IPSPs are, in fact, regularly spaced at ֱst to 2nd] 41 ms, ֲnd to 3rd] 42 ms, ֳrd to 4th] 48 ms, ִth to 5th] 42 ms, and ֵth to 6th] 43-ms mean intervals. The upward sloping regression line demonstrates that mutant IPSP amplitudes increase up to threefold with successive M3 firings.

          Genetic Interactions between Syntaxin and Synaptobrevin Mutants

          Syntaxin interacts tightly with synaptobrevin and SNAP-25 to form a stable ternary complex. This complex is proposed to represent an intermediate in the synaptic vesicle fusion cycle ( Sollner et al., 1993 ). In each of these proteins, the domains participating in the interaction are predicted to contain α-helices capable of assembling into coiled-coil structures ( Hardwick and Pelham, 1992 Fasshauer et al., 1997 ). Lesions that lie on the hydrophobic face of the presumed coiled helix region exist in both C. elegans syntaxin and synaptobrevin ( Nonet et al., 1998 ). We constructed double mutants between the synaptobrevin mutants snb-1(js17), snb-1(js44), and snb-1(md247) and the two missense helical domain syntaxin mutants js21 and e246 (Figure ​ (Figure8). 8 ). When analyzing the phenotypes of all six double mutants, we observed strong synergistic effects with certain allelic combinations. The size of young adult animals (Figure ​ (Figure9) 9 ) and the rates of pharyngeal pumping (Table ​ (Table2) 2 ) were drastically reduced in certain double-mutant combinations. For example, when the weakest synaptobrevin mutant, js44, was combined with the weakest syntaxin mutant, js21, the result was a more severe phenotype than when js44 was paired with the stronger e246 syntaxin lesion. Conversely, snb-1(js17) interacted strongly with e246, but not with js21. Neither syntaxin allele showed interactions with snb-1(md247), whose lesion is outside the helical domain and alters the transmembrane domain of synaptobrevin ( Nonet et al., 1998 ). EPG analysis provided no additional insight, as even the healthier double mutants displayed more severe EPG phenotypes than the singles and failed to exhibit any significant M3 activity (Figure ​ (Figure6). 6 ). This was not unexpected, as this assay examining M3 activity is very sensitive and is able to detect defects in mutants that exhibit only very mild behavioral changes ( Nonet et al., 1997 ).

          Mutations in snb-1 and unc-64 on the hydrophobic faces of α-helices proposed to mediate interactions between synaptobrevin and syntaxin. Two model amphipathic α-helices composed of a repeating seven-amino acid pattern. The residues are labeled a through g. Residues in a and d positions are usually hydrophobic and are proposed to mediate interactions between binding partners. Below, a portion of the sequences of human, fly, yeast, and worm syntaxin and synaptobrevin ( Archer et al., 1990 DiAntonio et al., 1993 Gerst et al., 1992 ) are aligned. Hydrophobic residues in the a and d positions of the predicted α-helices are labeled. The site of lesions in C. elegans syntaxin and synaptobrevin ( Nonet et al., 1998 ) are also labeled. The two sequences are oriented assuming they will interact in a parallel manner ( Hanson et al., 1997 Lin and Scheller, 1997 ).

          Phenotypes of unc-64 snb-1 double- mutant animals. Bright field photographs of young adult animals singly and double mutant for unc-64 syntaxin and snb-1 synaptobrevin mutations. Animals were isolated as L4 larvae and then incubated 1 d before examination. Specific combinations of alleles cause severe behavioral, growth, and pharyngeal pumping defects.

          Table 2

          Pharyngeal pumping in unc-64 snb-1 double mutants

          Pharyngeal pumping (pumps per min) of animals of genotype: unc-64 (horizontal) snb-1 (vertical). 


          Materials and Methods

          Subjects

          All of the measurements were made on 4-d-old adult C. elegans. A total of 2814 C. elegans were used in these experiments. Of these, 2301 were wild-type N2. To assess the role of chemosensory cues in the effects of isolation, 40 osm-6(p811) worms were used. In this strain, the mutation results in disrupted structure of sensory cilia that causes defects in all types of chemosensory behaviors (Dusenbery and Barr, 1980 Peckol et al., 1999). To assess the role of glr-1 in the effects of isolation 124 of glr-1(n2461) and 157 of the glr-1, rescue strain kp537 was used (in these experiments, n values were high because the rescue strain is not integrated and responses were more variable than in wild-type worms). To assess the role of mec-4(e1611) and egl-4(n477) in the effects of isolation, 37 mec-4 worms and 42 egl-4 worms were reared in either isolation or colonies and tested. For confocal imaging, two genetically engineered strains were used: GLR-1::GFP and pmec-7::SNB-1::GFP. GLR-1::GFP worms (n = 65) contain chimeric receptors made up of GLR-1 (a homolog of a kainate/AMPA-type glutamate receptor expressed on the interneurons of the tap withdrawal circuit) tagged with GFP, which has been used to visualize synapses in C. elegans (Rongo and Kaplan, 1999). In pmec-7::SNB-1::GFP (n = 48), we looked at GFP expression of the gene for C. elegans synaptobrevin (snb-1), a protein associated with synaptic vesicles that plays a role in regulating vesicle fusion at the synaptic terminal. SNB-1::GFP was expressed under the control of the mec-7 promoter, which targets this GFP expression to the six mechanosensory neurons of the tap withdrawal circuit (Nonet, 1999).

          Procedure and apparatus

          All plates were 60 mm in diameter and 15 mm deep, filled with 10 ml of nematode growth medium agar, and streaked with one to two drops of E. coli (strain OP50) to produce a circular lawn ∼10 mm in diameter (Brenner, 1974). For behavioral testing, a platinum wire pick was used to transfer adult worms (96-100 h after eggs were laid) from the plates where they were grown to a test plate without any E. coli.

          The test plate, in a plate holder attached to a micromanipulator, rested on the stage of the stereomicroscope (Wild M3Z Leica, Nussloch, Germany). A video camera (Digital 5100 Panasonic, Secaucus, NJ) was attached to the stereomicroscope and also to a video cassette recorder (Panasonic AG1960) and television monitor (PM-1271A NEC, Tokyo, Japan). The date and time were displayed on the monitor via a time-date generator (Panasonic WJ-810) and videotaped. After 2 unrecorded minutes, the stage lamp was turned on, and a single tap was delivered to the side of the plate via a copper rod driven by a Grass Instruments (Quincy, MA) S88 stimulator.

          Rearing condition

          Colony-raised worms were produced by allowing five adult hermaphrodites to lay between 30 and 40 eggs on a Petri plate with E. coli, then removing the adults so that only the eggs remained. Isolate-raised plates were produced by allowing one adult hermaphrodite to lay one egg on a Petri plate with E. coli, then removing the adult and any extra eggs so that only one egg remained. To protect worms from extraneous stimulation, all plates were then placed in foam- or cotton-lined boxes and placed in a 20 ± 0.2°C incubator. Plates were not touched until testing (96-100 h later).

          Conditioned plates

          In one experiment, isolate and colony worms were raised on plates that had contained colonies of worms for 4 d before the experiment. All of these “conditioning” worms were removed from the plates before the egg-laying for the experiment. This was done to provide the chemical stimuli of groups of worms for the isolate-raised worms.

          Larval stimulation

          In experiments with a stimulated isolate-raised condition, worms were treated as described above for the isolate-raised and colony-raised conditions however, in the third larval stage of development (“L3” ∼36 h of age), they were given a brief period of mechanosensory stimulation. (L3 was chosen as the approximate mid-point in larval development. In ongoing studies, we are investigating whether there is a critical period for rescue of the tap response. However, as mentioned, the isolate effect on behavior is very fragile and is easily disrupted by extraneous stimulation in the incubator at any time during the 4 d period, and we are hypothesizing that there will not be a critical period for the rescue of the behavior.) To deliver the same number of stimulations to all worms at exactly the same age, a cardboard box containing plates of L3 isolate and colony worms was removed from the incubator, and 30 mechanosensory stimuli were delivered to all plates simultaneously by dropping the box onto a table from a height of 5 cm 30 times at a 10 s interstimulus interval (ISI). After this stimulation, the box containing the stimulated isolate-raised and stimulated colony-raised worms was returned to the incubator until testing at 4 d of age.

          Scoring and analysis

          Responses to tap were videotaped and later scored using stop-frame video analysis. For each worm videotaped, a tracing was made of the length and width of the worm. Worm length was calculated by measuring the worm from the tip of the head to the tip of the tail. Worm width was calculated by measuring the width of the worm at the midpoint of its body length. A “reversal” was defined as a worm swimming backward in response to the tap within 1 s after the tap.

          Reversals were traced onto acetate sheets and later scanned (ScanJet 3c Hewlett-Packard, Palo Alto, CA) into a Macintosh (Performa 6200CD Apple Computers, Cupertino, CA) computer using DeskScan II software. Tracings were measured using NIH Image software, and all data analyses were performed in Statview 4.5 (Abacus Concepts, Berkeley, CA). Because differences were found in the length of worms raised under different conditions, all of the response to tap data were reported standardized to a percentage of worm length by dividing the distance each worm reversed by that worm's length and multiplying by 100.

          For spontaneous reversals, 52 colony and 44 isolate worms were each placed alone on agar-filled Petri plates and filmed for 6 min. Backward movements of the body during the 6 min period were counted as spontaneous reversals and scored in the same manner as reversals.

          Because the magnitude of the response to tap is affected by a number of factors including temperature, humidity, and age of the agar plates used, a separate N2 group was run on the same day and under the same conditions as the experimental groups for every experiment. Across experiments, there was some variability in the raw scores of the N2 worms as a result of changes in environmental conditions, but the effect of isolation on the response was consistent (isolates significantly smaller than colony). Because we found that the colony/isolate effect was easily disrupted (if the worms experienced even very low levels of vibration during development the effect was lost), we depended on the N2 worms as an indicator of this disruption. If the N2 worms did not show the colony/isolate effect (resulting from a variety of factors including the incubator being bumped, the plates being contaminated, or the worms being sick), the data from that experiment were thrown out, and the experiment was rerun.

          When doing statistical analysis on mutant strains of worms (see Figs. 1a, 3a-d), comparisons were only made within a strain to directly test our experimental manipulation (i.e., colony vs isolate) and not between strains. We did not make statistical comparisons between strains, because mutations are made on a large number of different genetic backgrounds and may differ from our wild-type worms in a number of ways. Therefore, differences between strains in baseline measures were not relevant for this paper.

          Effects of isolation on tap withdrawal response. a, Mean tap response magnitude (Resp. Magnitude) for wild-type (N2) worms and a chemosensory mutant (osm-6). The isolate-raised worms of both strains showed significantly smaller responses to tap than did the colony-raised worms (n = 20 per group). b, Effects of isolation on reversals to heat. Mean reversal magnitude to tap and to heat probe for isolate-raised and colony-raised worms (n = 20 per group). Isolate-raised worms had significantly smaller responses to tap than colony worms. There were no differences between the isolate-raised and colony-raised worms' responses to the heat probe. c, Effects of stimulation on isolated worms. Mean tap response magnitude for colony-raised, isolate-raised, L3 stimulated colony-raised worms, and L3 stimulated isolate-raised worms (L3 worms were raised in colony or isolation but were given brief stimulation during the third larval stage). Brief mechanosensory stimulation in L3 reversed the effects of isolation on behavior. Stimulated isolate-raised responses to tap were larger than the isolate-raised worms and not different from colony-raised or stimulated-colony raised worms. The asterisks represent statistically significant pairs. Error bars represent SEM. WL, Worm length.

          Heat probe experiment

          A heated probe (a scalpel blade heated in an alcohol flame until it glowed red) was moved manually into the path of a worm during forward locomotion so that the probe was perpendicular to the body axis of the worm and contacted neither the worm nor the agar surface. The reversal response to the probe was filmed and scored in the same way as the reversals to tap.

          Egg counting experiment

          Colony and isolate reared plates were set up as described above. Plates were taken from the incubator at either 66 or 70 h of age, and the number of worms (for colony plates) and the number of eggs per plate were counted and recorded using a manual counter.

          Confocal GFP imaging

          Worms were mounted onto 14 × 14 mm three-square slides (one worm per slide Erie Scientific, Erie, PA) using 12 μl of 2,3-butanedione monoxime for paralysis mixed with Sephadex beads (G-50 medium) to prevent worms from being crushed on the microscope stage during imaging. Images were obtained using a Nikon (Tokyo, Japan) Optiphot-2 microscope with an MRC 600 confocal system (Bio-Rad, Hercules, CA) equipped with a krypton/argon laser. GFP was excited using a 488 nm wavelength laser setting with the emitted light collected by passing through a ∼510-550 nm bandpass filter. The images collected from the MRC 600 were captured in a 768 × 512 pixel field of view with the optical sections collected at 0.5 μm intervals. Worms were imaged using a 60× magnification oil lens.

          GLR-1::GFP was expressed along the ventral cord. Therefore, images were collected along the posterior portion of the ventral cord from the tail to the vulva. Images of GLR-1::GFP expression were composed of 10-15 optical sections for each ventral nerve cord segment. Each resultant stack of images was then compiled into a single projection image. These projection images were used for the analyses. The GFP expression in the ventral nerve cord is uniform in thickness. Therefore, the length of GFP expression was measured (Rose et al., 2003).

          GFP expressed in the pmec-7::SNB-1::GFP strain was captured in a single image stack each composed of ∼12-18 optical sections. Because the GFP in these worms was quite faint, the microscope was set to maximal sensitivity for all worms, and intensity was not measured. Worms with no detectable GFP were discarded (∼13.5% of the worms distributed evenly across groups). Because the SNB-1::GFP appeared as one to three clustered bright spots, area measurements were used to quantify GFP expression (Rose et al., 2003).

          Analysis of confocal images GLR-1::GFP.

          Collected projection images were coded and viewed in NIH Image 1.61. A researcher blind to the condition of the worm measured the number and length of GFP expression in the image. Measurements were entered in Statview 4.5 for statistical analysis.

          pmec-7::SNB-1::GFP. Projection files were coded before analysis so that all measurements were made by a researcher blind to the condition of the worm. Single projection images were opened in NIH Image (as above) and viewed as a binary image with the threshold adjusted to allow for measuring of faint images. Area measurements for each region of GFP expression were calculated by outlining the edge of the region and using the area measure function. Area measurements were entered into Statview 4.5 for statistical analysis.

          Final figures for GFP images were generated using Adobe Photoshop 7.0 (Adobe Systems, San Jose, CA).

          Statistical analysis

          In experiments in which only two groups were run, unpaired t tests were used to determine significance. If multiple groups were run, an ANOVA with Fisher's planned least significant difference (PLSD) planned comparisons were used. For all experiments, α was set to p ≤ 0.05.


          Materials and methods

          C. elegans strains were grown and maintained under standard conditions (Brenner, 1974). A complete listing of all strains used in this study and their genotypes is located in Supplementary file 1.

          Molecular biology and transgenesis

          cDNA corresponding to the entire coding sequences of unc-31 (isoform a), daf-2 (isoform a), age-1 (isoform a), tom-1 (isoform a), and the ins-1 genomic region were amplified by PCR and expressed under cell-selective promoters. unc-17 cDNA was synthesized (GenScript) and expressed under a cell-selective promoter. For cha-1 and cho-1 knockdown experiments, 1 kb fragments corresponding to exons 3–7 and the 3′ end of the gene, respectively, in the sense and antisense orientation were synthesized (GenScript). Neuron-selective RNAi transgenes were created as previously described by co-injection of equal concentrations of sense and antisense oriented gene fragments driven by cell-specific promoters (Esposito et al., 2007). Cell-specific expression was achieved using the following promoters: ceh-36deletion or odr-3 for both AWC, str-2 for AWC ON , srsx-3 for AWC OFF , gpa-4 for AWA and ASI, gpa-4deletion for AWA, gcy-7 for ASEL, gcy-5 for ASER, str-1 for AWB, sre-1 for ADL, srh-142 for ADF, gcy-8 for AFD, ops-1 for ASG, sra-6 for ASH, trx-1 for ASJ and sra-9 for ASK. For all experiments, a splice leader (SL2) fused to a mCherry or gfp transgene was used to confirm cell-specific expression of the gene of interest.

          Germline transformations were performed by microinjection of plasmids (Mello and Fire, 1995) at concentrations between 25 and 200 ng/μl with 10 ng/μl of unc-122::rfp, unc-122::gfp or elt-2::gfp as co-injection markers. For rescue and OE experiments, DNA was injected into mutant or wild-type C. elegans carrying GCaMP arrays.

          Calcium imaging

          Transgenic worms expressing GCaMP calcium indicators under a cell-selective promoter were grown to day 1 or day 5 of adulthood and trapped in a custom designed PDMS microfluidic device and exposed to odor stimuli (Chalasani et al., 2007 Chronis et al., 2007). For aging experiments, a new PDMS device with larger channels was designed to trap and stimulate day 5 adult worms (Chokshi et al., 2010). Older, day 6 adult worms exhibit much larger variation in whole animal size than day 5 adults (see Figure 5—figure supplement 1A) and could not be trapped consistently without introducing bias into the experiment. For aging experiments, animals were transferred to new OP50 bacteria plates every other day to track the aging animals and to avoid contamination by their progeny. Additionally, for whole animal RNAi experiments to knockdown rab-10 and hsf-1, animals were fed either control (Ctrl) empty pL4440, rab-10 RNAi or hsf-1 RNAi expressing bacteria beginning at day 1 of adulthood as previously described (Hansen et al., 2005).

          Fluorescence from the neuronal cell body was captured using a Zeiss inverted compound microscope for 3 min. We first captured 10 s of baseline activity (t = 0–10 s) in chemotaxis assay buffer (5 mM K3PO4 (pH 6), 1 mM CaCl2, 1 mM MgSO4, and 50 mM NaCl), then 2 min (t = 10–130 s) of exposure to an odor (or salt) stimulus dissolved in chemotaxis buffer, and lastly 50 s (t = 130–180 s) of buffer only. BZ refers to a 0.005% vol/vol dilution in chemotaxis assay buffer, except where low BZ (0.0001% vol/vol) or high BZ (0.1% vol/vol) is specifically mentioned. Additionally, a 0.1% vol/vol dilution of 2-nonanone and 50 mM sodium chloride stimulus were used as indicated. For arecoline experiments, worms were pre-treated with 0.15 mM arecoline in chemotaxis buffer for approximately 20 min and immediately imaged in the presence of the drug. Laser ablations of the paired AWC, AWA, AWB or ASE sensory neurons, along with mock ablations, were performed as previously described (Bargmann and Avery, 1995) in transgenic animals expressing GCaMP. In all experiments, a single neuron was imaged in each animal, and each animal was imaged only once. Wild-type Ctrls, mutants, and transgenic or drug treated strains for each figure were imaged in alternation, in the same session.

          We used Metamorph and an EMCCD camera (Photometrics) to capture images at a rate of 10 frames per second. A MATLAB script was used to analyze the average fluorescence for the cell body region of interest and to plot the percent change in fluorescence for the region of interest relative to F0, as previously described (Chalasani et al., 2007). Specifically, data was plotted and statistical analysis was performed as follows: (1) for line graphs of ΔF/F over time (Figures 1–4 and corresponding figure supplements), the average fluorescence in a 8 s window (t = 1–9 s) was set as F0. Average and standard error at each time point were generated and plotted using MATLAB scripts, as previously described (Leinwand and Chalasani, 2013). (2) For heat maps (Figures 5–7 and corresponding figure supplements), the average fluorescence in a 8 s window (t = 1–9 s) was set as F0.

          To quantify calcium responses, F0 was consistently set to the average fluorescence signal from 1 s to 9 s prior to the relevant change (addition or removal) of stimulus. For statistical analysis, the average fluorescence and standard error were calculated for each animal over a short period corresponding to the duration of a response. Specifically, to analyze on responses to the addition of stimulus, the average fluorescence and standard error were calculated in the 10 s period following the addition of odor or salt (t = 10–20 s). For AWA neurons, the response duration was very brief therefore, a 4 s time period was used instead (t = 10–14 s) so that small, fast responses could be appropriately quantified. To analyze off responses to the removal of stimulus, the average fluorescence and standard error were calculated in the period following the removal of odor (t = 130–140 s for all cells except ASE, and t = 130–145 for the slower, longer duration ASE responses). Traces in which an averaged ΔF/F of greater than 600% was recorded were excluded as they are likely to be artifacts of the neurons moving out of the focal plane and these usually account for less than 1% of the traces collected. To determine whether there was an odor-evoked increase or suppression of the calcium signal (see Figure 1C), the average fluorescence in these time windows in buffer only trials was compared (by a two-tailed unpaired t-test) to the average fluorescence in odor stimulation trials, for each neuron. The maximum ΔF/F in these time periods following odor addition or removal and the time to reach this maximum ΔF/F (from the stimulus change, in seconds) were also quantified (see Figure 5G, Figure 1—figure supplement 1A,B and Figure 5—figure supplement 1B). More specifically:

          (1) For bar graphs of averaged ΔF/F after odor addition or removal (Figures 2–4, Figure 3—figure supplement 1 and Figure 4—figure supplement 1): (a) F0 was set to the average fluorescence from 1–9 s for quantification of AWA neuron responses to the addition of BZ stimulus and (b) F0 was set to the average fluorescence from 121–129 s for quantification of AWC, ASE and AWB responses to the removal of BZ or 2-nonanone. Two-tailed unpaired t-tests were used to compare the responses of different genotypes or cell ablation conditions, and the Bonferroni correction was used to adjust for multiple comparisons.

          (2) For scatter plots of maximum ΔF/F (Figure 1—figure supplement 1A and Figure 5G) and scatter plots of averaged ΔF/F after stimulus change (Figure 5H, Figure 6—figure supplement 1B and Figure 7—figure supplement 1B): (a) for AWA neurons' response to the addition of odor stimulus F0 was set to the average fluorescence from 1–9 s and (b) for AWC ON , ASEL and AWB responses to odor stimulus removal F0 was set to the average fluorescence from 121–129 s. For the subset of odor-responsive neurons (exceeding the 10% ΔF/F cut-off), the averaged ΔF/F after the stimulus change and the time to the maximum ΔF/F were also analyzed using two-tailed unpaired t-tests to compare different ages or genotypes (Figure 5H, Figure 5—figure supplement 1B, Figure 6—figure supplement 1B and Figure 7—figure supplement 1B). Furthermore, considering only the odor responsive neurons, no significant differences were observed in the magnitude of the odor-evoked suppression of young and aged animals (comparing the average fluorescence in ten second windows tiling the period of odor stimulation, by two-tailed t-test), indicating that our subsequent analyzes of the odor removal time period are not biased by the choice of the F0.

          (3) For bar graph quantifications of the % odor or salt responsive neurons in the aging experiments (Figure 5I,K,L, 6F, and 7E and the corresponding figure supplements): (a) F0 was set to the average fluorescence from 1–9 s for quantification of the percent of AWA and ASH neurons responsive to the addition of BZ stimulus and for the percent of ASEL and AWC neurons responsive to the addition of NaCl salt stimulus. (b) F0 was set to the average fluorescence from 121–129 s for AWC ON , ASEL and AWB responses to BZ or 2-nonanone odor stimulus removal. The percent of odor responsive neurons was calculated by determining the proportion of cells displaying an average fluorescence (ΔF/F) greater than 10% after odor addition (for AWA and ASH) or odor removal (all other neurons). 10% ΔF/F was used as the cut-off for odor responsiveness because, for all neurons imaged, changing buffer around the nose of the animal elicited a response smaller than this cut-off. Similarly, neurons displaying an average fluorescence (ΔF/F) greater than 10% after salt addition were considered to be salt responsive. A two-tailed Chi–Square test was used to compare the percent of odor or salt responsive neurons in different conditions.

          Chemotaxis assays

          Odor chemotaxis assays were performed as previously described (Ward, 1973). For aging assays, worms were synchronized by hatch offs in which 8 young adult worms were given 150 min to lay eggs on a large plate before being picked off. These eggs were grown at 20° until the appropriate day of adulthood, except for glp-1 mutants, which were raised at the restrictive temperature, 25°. Aging animals were transferred to new bacteria plates every other day to track the aging animals and to avoid contamination by their progeny. Chemotaxis assays were performed on 2% agar plates (10 cm diameter) containing 5 mM potassium phosphate (pH 6), 1 mM CaCl2 and 1 mM MgSO4. Animals were washed once in M9 and three times in chemotaxis buffer (5 mM K3PO4 (pH 6), 1 mM CaCl2 and 1 mM MgSO4). For arecoline chemotaxis experiments, 0.15 mM arecoline was added to the M9 and chemotaxis buffer washes, yielding a 16–20 min drug treatment immediately prior to the behavioral experiment. Odor concentration gradients were established by spotting diluted BZ (0.2% vol/vol, in ethanol) near the edge of the plate, with a Ctrl 1 μl of ethanol spotted at the opposite end of the plate. Where noted, 1 μl of neat BZ was used for high concentration point source assays. For 2-nonanone experiments, a 50% vol/vol dilution of 2-nonanone in ethanol was used. For salt chemotaxis experiments, salt gradients were established by placing a Ctrl or a high salt (500 mM NaCl) agar plug on the assay plate and allowing 16–20 hr for the salt to diffuse and form a gradient (Leinwand and Chalasani, 2013). 1 μl of sodium azide was added to the odor (or salt) and the Ctrl spots to anesthetize animals reaching the end points. Washed worms were placed on the plate and allowed to move freely for one hour. The chemotaxis index was computed as the number of worms in the region near the odor (or salt) minus the worms in the region near the Ctrl divided by the total number of worms that moved beyond the origin. Nine or more assays were performed, over at least three different days. Two-tailed unpaired t-tests were used to compare the responses of different genotypes or ages, and the Bonferroni correction was used to adjust for multiple comparisons.

          Correlated chemotaxis and imaging experiments

          Transgenic worms bearing GCaMP arrays, synchronized by a hatch off as described above, were grown until day 5 of adulthood at 20°. Aging animals were transferred to new bacteria plates every other day to track the aging animals and to avoid contamination by their progeny. Animals were tested in a (0.2% vol/vol) BZ odor chemotaxis assay as above, with two modifications. First, no sodium azide was used to paralyze the animals. Second, animals were given only 30 min to move freely on the chemotaxis plate. The chemotaxis assay plate was then cut into three regions corresponding to the BZ odor side, the middle, and the ethanol Ctrl region immediately after 30 min and worms were washed off each section separately and allowed to recover on OP50 bacteria plates for at least 90 min. Worms from the odor and the Ctrl sections of the chemotaxis assay were imaged in alternation as described above.

          Lifespan assays

          Worms, synchronized by a hatch off as described above, were grown until day 1 or 5 of adulthood at 20°. To sort animals on the basis of their chemotaxis performance, wild-type animals were tested in a (0.2% vol/vol) BZ odor or (500 mM NaCl) salt chemotaxis assay as above, but without sodium azide and with only 30 min for the animals to move freely on the chemotaxis plate. The chemotaxis assay plate was then cut into a BZ odor (or salt), middle, and Ctrl region and worms were washed off each section separately. 100 adults from the odor (or salt) region or the Ctrl region were transferred onto 10 small OP50 plates (10 adults per plate) and grown at 20°. For experiments with transgenic animals, day 1 animals bearing the appropriate transgene were picked from the hatch off plate directly onto 10 small OP50 plates (10 adults per plate) and grown at 20°. In all experiments, aging animals were transferred to new bacteria plates every other day to track the aging animals and to avoid contamination by their progeny. Survival was analyzed every other day and worms were scored alive or dead based on their response to a gentle head touch (or lack thereof) as previously described (Kenyon et al., 1993). Worms were censored if they bagged, exploded or desiccated on the side of the plate. The chemotaxis assay followed by lifespan analysis or lifespan assays with transgenic animals were repeated two or three times per condition as indicated, beginning on separate days. The percent change in mean survival was calculated as the mean survival of animals from the odor side minus the mean survival of animals from the Ctrl side divided by the mean odor side survival or the mean transgenic animal survival minus the mean wild-type survival divided by the mean wild-type survival. Statistical analysis of lifespan was performed by the Mantel–Cox Log–Rank test, using GraphPad Prism and OASIS (Yang et al., 2011).

          Speed analysis

          Chemotaxis assays to BZ were set up as described above, but with modifications to enable automated analysis of animal speed. 200 mM Cu(II)SO4-soaked filter paper was placed on a standard chemotaxis assay plate to contain the worms in a reduced chemotaxis arena (1.25 by 1.25 inch square). 1 μl of BZ (0.2% vol/vol dilution in ethanol) and a Ctrl 1 μl of ethanol were spotted at opposite corners of the square arena, without any paralytic. After washing, only 5 worms were placed on the chemotaxis plate this number minimized collisions and enabled more accurate tracking. The movement of the animals was tracked over 60 min using a Pixelink camera and speed was analyzed using previously published MATLAB scripts to track the centroid of the animal (Ramot et al., 2008). The results from eleven chemotaxis plates were averaged for each age. NS indicates p > 0.05, two-tailed t-test.

          Aged worm measurements

          Day 5 and day 6 adult worms from hatch offs performed on three separate days were immobilized with tetramisole and imaged on 2% agarose pads. Images were captured on a Zeiss Observer D1 microscope using a 10× objective with DIC. The perimeters of 55 worms were measured using MetaMorph software.