Novel and Conserved Protein Macoilin Is Required for Diverse Neuronal Functions in Caenorhabditis elegans

Neural signals are processed in nervous systems of animals responding to variable environmental stimuli. This study shows that a novel and highly conserved protein, macoilin (MACO-1), plays an essential role in diverse neural functions in Caenorhabditis elegans. maco-1 mutants showed abnormal behaviors, including defective locomotion, thermotaxis, and chemotaxis. Expression of human macoilin in the C. elegans nervous system weakly rescued the abnormal thermotactic phenotype of the maco-1 mutants, suggesting that macoilin is functionally conserved across species. Abnormal thermotaxis may have been caused by impaired locomotion of maco-1 mutants. However, calcium imaging of AFD thermosensory neurons and AIY postsynaptic interneurons of maco-1 mutants suggest that macoilin is required for appropriate responses of AFD and AIY neurons to thermal stimuli. Studies on localization of MACO-1 showed that C. elegans and human macoilins are localized mainly to the rough endoplasmic reticulum. Our results suggest that macoilin is required for various neural events, such as the regulation of neuronal activity.


Introduction
Genome projects have determined DNA sequences of various organisms and identified locations of predicted genes in genomes. Surprisingly, numbers of genes are relatively similar across evolutionally divergent animal species. For example, estimated gene numbers are 20000-25000 in human [1], 14000 in Drosophila melanogaster [2], and 19000 in Caenorhabditis elegans [3]. Of the estimated genes in C. elegans, more than 40 percent of the predicted protein products show significant matches to those from other animals [3]. Despite this species homology, functions of these genes are poorly understood. Therefore, further characterization of these proteins could lead to novel insights about important cellular processes, such as gene regulation, protein maturation, signal transduction, and intracellular transport.
Animals use the nervous system to detect stimuli in the surrounding environment. These stimuli are then assessed based on past experiences, and converted into adaptive behaviors. Many neurons are required to act in coordination to execute these processes. However, it is difficult to address individual neuronal function in relation to other neurons in neural circuits because animal brains generally consist of enormous numbers of neurons. The anatomically characterized nervous system of C. elegans is composed of only 302 neurons [4], making it an effective model for studying neuronal function in circuits. Despite their simple nervous systems, C. elegans can sense an array of environmental stimuli and produce appropriate behavioral outputs. For example, worms remember cultivation temperatures in association with food conditions [5][6][7]. This thermotaxis behavior can be divided into two behavioral aspects. First, after cultivation C. elegans with food at a certain temperature for several hours, they will migrate to this cultivation temperature along a temperature gradient [5][6][7]. Previous studies have suggested that this migratory behavior could be explained by the following two opposing strategies: migrations up thermal gradients when located at temperatures below the cultivation temperature (i.e., thermophilic drive); and migrations down thermal gradients when at a higher temperature (i.e., cryophilic drive) [5][6][7]. Work by several groups has been unable to confirm thermophilic drive in C. elegans [8][9][10][11]. However, recent theoretical and experimental studies have shown that the steepness of the thermal gradient drastically changes migration behavior [12,13]. The second behavioral aspect of thermotaxis involves worms leaving isothermal tracks (IT) on a temperature gradient as they migrate to the cultivation temperature associated with food ( Figure 1B and [5,6]). Previous reports have suggested that the neuronspecific calcium sensor-1 (NCS-1) plays an important role in optimizing IT [14]. Specifically, reports showed that a knockout phenotype of ncs-1 animals could be rescued by introducing a wild-type NCS-1 into the AIY interneuron, which is a component of the thermotaxis neural circuit, and that NCS-1 overexpression enhanced the frequency of IT behavior [14].
The thermotaxis phenotype of C. elegans, which was achieved by laser-ablating specific neurons, has been examined previously, and results led to the hypothesis of a neural circuit for thermotaxis [6]. This neural circuit is presently one of the simplest circuits to regulate animal behavior. By capitalizing on this simple structure, we expect that the study of thermotaxis behavior will reveal molecular mechanisms of thermosensation and concepts of neural plasticity and neuronal function. Several molecular components related to thermotaxis behavior have been identified. The three guanylyl cyclases, GCY-8, GCY-18 and GCY-23, which are expressed in AFD thermosensory neurons, appear to redundantly produce cyclic guanosine monophosphate (cGMP) as a second messenger. The cGMP-dependent cation channel, which is composed of TAX-2 and TAX-4, is crucial in activating AFD neurons [15][16][17]. TAX-6 (calcineurin) and TTX-4 (PKC-1; nPKC-epsilon/eta), both of which act as negative regulators of AFD thermosensory signaling, have also been reported [18,19]. Calcium imaging using the genetically encoded calcium indicator cameleon can display fluorescence changes from variable intracellular calcium concentrations [20]. This imaging has shown that AFD neurons respond to thermal stimuli in a cultivation temperature-dependent manner [21]. This response occurred even after disconnecting the AFD dendrites from their cell bodies [22], indicating that temperature memories are stored in AFD neurons. Consistent with findings in AFD neurons, recent calcium imaging analyses also suggest that AWC thermosensory neurons can also store temperature memories [23,24].
In this study, we found that the novel protein MACO-1 is required for diverse neuronal functions in C. elegans. maco-1/ttx-8 mutants exhibited abnormalities in thermotaxis and chemotaxis to odorants and sodium chloride (NaCl). The maco-1 gene encodes a conserved novel protein that is expressed in many neurons. A previous study reported that the mouse ortholog of MACO-1, macoilin (C61), showed neuron-specific expression [25], although its function remained largely unknown. Rescue experiments revealed that simultaneous expression of MACO-1 in AFD, AIY and AIZ neurons of maco-1 mutants is required for thermotaxis behavior. We also found that maco-1 mutants showed abnormal locomotion resulting from frequent turns. Although defective locomotion of maco-1 mutants might have caused abnormal thermotaxis, calcium imaging indicated that MACO-1 is required for activation of AFD and AIY neurons. In vivo co-localization studies have shown that both MACO-1 and human macoilin are localized to the rough endoplasmic reticulum (rER). All together, our results suggest that macoilin is required for a variety of intracellular events that influence basic neuronal function across species.
Results maco-1/ttx-8 mutants showed prominent thermotaxis defects and slight chemotaxis defects To elucidate molecular mechanisms of thermotaxis, we performed a genetic screen and isolated several mutants that were thermotaxis defective [19]. Two of these mutants, maco-1(nj21) and maco-1(nj34), were isolated as novel mutants. Two other mutant alleles, maco-1(tm2917) and maco-1(ok3165), were isolated by the National Bioresource Project (Japan) and the C. elegans Gene Knockout Consortium, respectively. All four alleles were genetically recessive (data not shown). Despite obvious defects in thermotaxis after cultivation at 20uC ( Figure 1B and 1C), maco-1(nj21) and maco-1(nj34) mutants were slightly defective in chemotaxis to odorants ( Figure 1D and Figure S1) or to NaCl ( Figure 1E). These results suggest that MACO-1 is necessary for animals to respond normally to various stimuli, including temperature, odorants and NaCl. Strong thermotaxis defects in maco-1 mutants after cultivation at 20uC suggest that MACO-1 may play an important role in thermotaxis, especially after cultivation at 20uC. Alternatively, the thermotaxis defects in maco-1 mutants could have been more easily detected in thermotaxis assays for 20uC-grown animals (20uC TTX assay) than 17uC or 23uC TTX assays. Specifically, during 20uC TTX assays, animals must migrate and remain in the 20uC region, which is shown as the light grey zone in the left schematic of Figure 1A. By contrast, during 17uC and 23uC TTX assays, animals can achieve thermotactic behaviors by migrating to 17uC and 23uC regions at the center and the periphery of a 9 cm plate, respectively. These areas are places where animals tend to remain due to the nature of assay plates and non-linear temperature gradients.
The maco-1 gene encodes a novel protein The maco-1 gene was identified by mapping it with the snipsingle nucleotide polymorphisms (SNPs) method [26] and conducting subsequent rescue experiments with genomic or polymerase chain reaction (PCR) fragments. Defective thermotaxis of maco-1 mutants was rescued by introduction of the cosmid C09C9. However, thermotaxis was not rescued by the cosmid C15E12 or C03F5 (Figure 2A), suggesting that maco-1 is located in the region between cosmids C15E12 and C03F5 within the cosmid C09C9. In addition, a PCR fragment containing D2092.4 gene (PCR1) did not rescue abnormal thermotaxis behavior of maco-1 mutants. However, another PCR fragment containing D2092.5 gene (PCR2) could rescue this behavior (Figure 2A), suggesting that D2092.5 is identical to the maco-1 gene. D2092.5/ maco-1 gene encodes a novel protein that was predicted to have five transmembrane domains in the N-terminus and four coiled coil domains in the C-terminus as determined by the TMHMM Server v. 2.0 [27] and by COILS server [28], respectively ( Figure 2B). Blast searches of the U.S. National Center for Biotechnology Inventory (NCBI) inventory identified maco-1 orthologs in several species, including the invertebrates Drosophila melanogaster, and vertebrates Xenopus laevis, Mus musculus and Homo sapiens ( Figure 2B, 2C and Table S1). The nj21 mutant has a G-to-A substitution at

Author Summary
Any animals, including humans, have to be capable of properly sensing and responding to various environmental stimuli for survival and reproduction. Environmental stimuli are evaluated and, based on past experiences, converted to produce appropriate adaptive behaviors. Even the small, free-living soil nematode Caenorhabditis elegans can sense diverse environment stimuli using a nervous system that consists of only 302 neurons. C. elegans exhibit thermotaxis that allows them to remember ambient temperatures and use this information as a cue to seek and remain near food sources. We show here that a novel and conserved protein, MACO-1, is important for proper execution of thermotaxis by C. elegans. The maco-1 gene was originally identified in a novel thermotaxisdefective mutant. Abnormal thermotaxis of maco-1 mutants was weakly but apparently rescued by expressing human MACO-1, suggesting that MACO-1 is functionally conserved among species. MACO-1 protein is required for appropriate regulation of neuronal activity, and the activity of neurons required for thermotaxis is impaired in maco-1 mutants. This analysis provides an important basis for the function of the macoilin family. The center of the assay plate (a 9 cmdiameter petri dish) is approximately 17uC and the periphery is approximately 25uC. Tracks of animals were categorized into four groups after onehour of free movement. Animals that moved to the 17uC region (center) were classified as '17'. Animals that moved to the 20uC region or 25uC region (periphery) were classified as '20' or '25', respectively. Animals that moved randomly around both 17uC and 25uC regions were classified as '17/25'. (B) Tracks of wild-type and maco-1 mutant animals showing thermotaxis behaviors on a radial temperature gradient ranging from 17uC (center) to 25uC (periphery). Animals were grown at 20uC. (C) Thermotaxis of wild-type and maco-1 mutant animals cultivated at 17uC (upper graph), 20uC (middle the splice acceptor site of the 9th intron. The nj34 mutant has two nonsense mutations (W231stop and Q761stop) and a missense mutation (A805T) ( Figure 2D). The tm2917 mutant has a 780-base pair (bp) deletion and 1-bp insertion just before the first coiled coil domain ( Figure 2D). The ok3165 mutant has a 972-bp deletion and 1-bp insertion in the coiled coil region ( Figure 2D).

MACO-1 is mainly localized to the rER
To study subcellular localization of MACO-1, we introduced the almost full-length genomic maco-1::GFP, which can rescue abnormal thermotaxis of maco-1 mutants ( Figure S2A), into wildtype animals and observed green fluorescent protein (GFP) fluorescence. Expression of this fusion protein was observed in the cell body but not in the neurite ( Figure 3A and 3B). Because MACO-1 was found in the cell body and is predicted to be a membrane protein, we investigated whether the protein is localized to organelles, such as the ER or the Golgi bodies. We made several transgenic lines co-expressing cyan fluorescent protein (CFP)-or yellow fluorescent protein (YFP)-fused MACO-1 and one of the following organelle markers: Translocating-chain associating membrane protein (CFP::TRAM) as a rER marker; Mannosidase (MANS::CFP) as a Golgi marker and membrane anchored YFP (YFP::GPI) as a plasma membrane marker [29]. We found that functional YFP::MACO-1 fusion protein was colocalized with the rER marker, CFP::TRAM in AFD neurons ( Figure 3C-3E, Figure S2C and Text S1). Similarly, YFP fusion protein with human macoilin, YFP::FLJ10747, was also colocalized with CFP::TRAM in AFD neurons ( Figure

MACO-1 functions in neurons and is conserved across species
To investigate the expression pattern of maco-1, we constructed the maco-1 promoter::GFP reporter that used an approximately 5kilobase (kb) upstream region from the ATG initiation codon as a maco-1 promoter. This construct was then introduced into wildtype animals. We found that strong expression was observed in many neurons ( Figure S3A and Text S1), while very weak expression was also detected in tissues. Thermotaxis defects were consistently rescued when maco-1 cDNA was expressed in almost all neurons of maco-1 mutants using the unc-14 promoter, indicating that MACO-1 acts in neurons (lower left panel of Figure 4A and Figure 4D).
To examine whether MACO-1 is conserved functionally among species, we introduced FLJ10747, a predicted human ortholog of MACO-1, into almost all neurons of maco-1 mutants using the unc-14 promoter. We then evaluated the thermotaxis phenotype of the transgenic strain (lower right panel of Figure 4A, Figure 4D, Figure S2B and Text S1). FLJ10747 rescued, albeit weakly, abnormal thermotaxis phenotype of maco-1 mutants, suggesting that MACO-1 is functionally conserved across species.
Both transmembrane domains and coiled coil domains are required for appropriate function of MACO-1 MACO-1 was predicted to have transmembrane domains in the N-terminus and coiled coil domains in the C-terminus ( Figure 2B). To examine whether these regions are necessary for MACO-1 function, we constructed two truncated proteins, MACO-1 TM (lacks coiled coil region) and MACO-1 CC (lacks transmembrane region), and conducted rescue experiments using each protein ( Figure 4C and 4D). Expression of MACO-1 TM did not rescue in nearly all neurons of maco-1 mutants using the unc-14 promoter, while MACO-1 CC slightly rescued the abnormal thermotaxis of maco-1 mutants ( Figure 4D). We also conducted rescue experiments using YFP::MACO-1 TM or YFP::MACO-1 CC fusion proteins to address whether degradation of these truncated proteins or ER quality influenced rescue experiment results. We verified that YFP fluorescence was observed in both truncated fusion proteins ( Figure S4J-S4M and Text S1). YFP::MACO-1 TM did not rescue and YFP::MACO-1 CC slightly rescued defective thermotaxis of maco-1 mutants ( Figure S2C and Text S1). We also found that YFP::MACO-1 TM appeared to localize to the rER ( Figure S4D-S4F and Text S1). However, YFP::MACO-1 CC localization to the rER was minimal, but rather was localized to cell bodies and neurites ( Figure S4G-S4I, S4L, S4M and Text S1). These results suggest that MACO-1 TM regions are necessary for appropriate MACO-1 localization, and MACO-1 CC regions are important for proper function of MACO-1.

Proper function of MACO-1 is necessary for normal thermotaxis in many neurons
To identify cells that require MACO-1 activity for sensory behavior, maco-1 cDNA was expressed by using various cellspecific promoters. Expression of MACO-1 in AFD, AWC, AIY, AIZ and RIA neurons, which are essential for thermotaxis behavior ( Figure 4B and [6,24]), partially rescued the thermotaxis defect of maco-1 mutants ( Figure 4D). By contrast, specific expression of maco-1 cDNA in AFD, AWC, AIY AIZ or RIA neurons individually did not rescue or only slightly rescued the abnormal thermotaxis phenotype of maco-1 mutants ( Figure 4D). Thus, we propose that simultaneous expression of MACO-1 in multiple neurons of the thermotaxis neural circuit is necessary to rescue abnormal thermotaxis among maco-1 mutants. We attempted but were unable to express MACO-1 in several combinations of neurons in maco-1 mutants to find the minimum set of neurons required to ensure full rescue of thermotaxis behavior. However, the expression of MACO-1 in AFD, AIY and AIZ neurons rescued maco-1 mutant defects similar to the expression of MACO-1 in AFD, AWC, AIY, AIZ and RIA neurons ( Figure 4D). Notably, expression of MACO-1 in any other combinations of AFD, AWC, AIY, AIZ and RIA neurons only graph) and 23uC (lower graph). n = 60-380. Single and double asterisks indicate p,0.05 and p,0.01 by ANOVA with a Dunnett's post hoc test, respectively. (D) Chemotaxis of wild-type and maco-1 mutant animals for odorants. Odorant concentrations are described in the Materials and Methods section. For each assay, 50-300 animals were analyzed. A single asterisk indicates p,0.05 compared to wild-type by ANOVA with a Dunnett's post hoc test. Bars represent the mean of three independent assays with error bars showing the standard error of the mean (SEM). (E) Chemotaxis to sodium chloride (NaCl). For each genotype, 40-120 animals were individually assayed. The phenotypic categories are described in the Materials and Methods section. Mean fractions categorized as 'Normal' were 0.8660.05 among wild-type animals, 0.4060.04 in maco-1(nj21) mutants and 0.2360.18 in maco-1(nj34) mutants. Statistically significant differences were observed between the fraction of 'Normal' of wild-type and maco-1 mutants (p,0.01 by ANOVA with a Dunnett's post hoc test). doi:10.1371/journal.pgen.1001384.g001 weakly rescued the defect ( Figure 4D). All together, these results suggest that MACO-1 activity in AFD, AIY and AIZ neurons is required for thermotaxis behavior, and MACO-1 activity is needed in many neurons to achieve full rescue.
Abnormal thermotaxis behavior is not attributable to indirect effects of defective locomotion in maco-1 mutants Cell-specific rescue experiments have implicated that a functional defect of MACO-1 in AFD, AIY and AIZ neurons caused defects in thermotaxis of maco-1 mutants. However, we also observed that maco-1 mutants changed their direction of movement more frequently than wild-type animals based on their tracks left during TTX assays ( Figure 1B). This observation raised another possibility that locomotory defects secondarily causes the abnormal thermotaxis behavior of maco-1 mutants.
Spontaneous locomotion of C. elegans is roughly classified into two categories. The first category, runs, includes long series of sinusoidal-swimming movement. The second category, turns, is divided into two types. Omega turns involve the animal curling its head toward the tail and continues to propel forward. Reversals involve the animal moving backwards for several seconds and then moving forwards in a new direction [30,31]. To quantify the potential locomotory defect, we constructed curvature maps of animal bodies along the head-to-tail axis using a tracking system and homemade MATLAB programs. Locomotion was represented by patterns in the curvature map ( Figure 5A and 5C-5E). Wildtype animals spent most of their time moving forward ( Figure 5C and left graph of 5F). Conversely, forward movement was interrupted by omega turns at short intervals among maco-1 mutant animals ( Figure 5D and middle graph of 5F). A similar tendency was observed in transgenic maco-1 mutants ( Figure 5E and left graph of 5F), whose abnormal thermotaxis was partially rescued by expression of maco-1 cDNA in AFD, AWC, AIY, AIZ and RIA neurons ( Figure 4D). This finding supports that defective locomotion does not cause abnormal thermotaxis in maco-1 mutants.

MACO-1 is required for proper activation of AFD and AIY neurons
Cell-specific rescue experiments indicated that MACO-1 activity, at least in AFD, AIY and AIZ neurons, is necessary for normal thermotaxis behavior ( Figure 4D). To examine whether the MACO-1 functional defect influenced activities of these neurons, we conducted calcium imaging experiments and monitored responses of AFD and AIY neurons from maco-1(nj34) mutants to thermal stimuli. We found that the ratio change in AFD neurons of maco-1(nj34) mutants was significantly less than that of wild-type animals ( Figure 6A). Previous work has shown that the ratio change in AFD neurons of tax-4(p678) mutants was hardly detected [21]. AFD neurons of the tax-4(p678) mutant do not respond properly to temperature because of defective cGMPdependent cation channels composed of TAX-2 and TAX-4. The average of maximum ratio change from baseline 6 SEM in maco-1(nj34) mutants was slightly larger (4.360.8) relative to tax-4(p678) mutants (3.360.6) [21] (Figure 6A), suggesting that AFD neuronal response to thermal stimuli among maco-1(nj34) mutants was not completely lost but partially reduced. Similarly, the ratio change in AIY neurons of maco-1 mutants was also lower than that of wildtype animals ( Figure 6B). Taken together, our data suggest that activities of both AFD and AIY neurons in maco-1 mutants are partially decreased in response to thermal change, further indicating that MACO-1 is required for activation of AFD and AIY neurons. We also examined morphologies of these neurons in maco-1 mutants by expressing GFP in AFD and AIY neurons to determine whether the decreased activity of AFD and AIY neurons was caused by developmental or functional defects. The AFD and AIY neurons in maco-1 mutants appeared normal ( Figure  S3B, S3C and Text S1), suggesting that MACO-1 is not required for development but rather for proper function of AFD and AIY neurons.

MACO-1 is required for proper activation of neurons
We found that proper function of MACO-1 in AFD, AIY and AIZ neurons is necessary for normal thermotaxis behavior. Consistent with this result, calcium imaging suggested that MACO-1 is required for accurate activation of AFD and AIY neurons. Rescue experiments also indicate that appropriate MACO-1 function in many other interneurons and motor neurons is probably required for full rescue of abnormal thermotaxis behavior. These results suggest that MACO-1 is required broadly for proper neural activation. Consistent with this postulate, maco-1 mutants showed additional behavioral defects in chemotaxis and locomotion.

MACO-1 likely plays a role in assisting fundamental neuronal function
Co-localization studies revealed that MACO-1 was mainly localized to the rER. The rER is generally thought to be required for protein synthesis, folding and translocation of secretory and transmembrane proteins. Thus, functional defects of MACO-1 in the rER might alter these events, resulting in abnormal phenotypes in maco-1 mutants.
Previous studies have reported that Resistant to Inhibitors of Cholinesterase 3 (RIC-3) is localized to the ER and required for the maturation (the folding, assembly or transport) of nicotinic acetylcholine receptors [32][33][34][35][36][37][38][39]. RIC-3 is predicted to have two transmembrane domains followed by coiled-coil domains. Similar structural motifs and intracellular localization patterns of RIC-3 and MACO-1 suggest that MACO-1 might be involved in the maturation of membrane proteins like RIC-3. To explore this . Results of rescue experiments among thermotaxis-defective maco-1 mutants are indicated as + (rescued) or 2 (not rescued) on the right side. Numbers in parentheses indicate the fraction of rescued lines. For each genotype, 120-180 animals were individually assayed for thermotaxis. The maco-1 gene was mapped to a 100-kb region between two SNPs (1.1 and 1.18) on chromosome I. We conducted rescue experiments using overlapping cosmids to encompass this 100-kb region. (B) Predicted structure of maco-1 gene family. Black and gray boxes indicate transmembrane domains and coiled coil domains, respectively. Numbers of amino acids are shown on the right. (C) Unrooted dendrogram of MACO-1. The dendrogram was generated using the freely available program, Clustal W [52] (http://clustalw.genome.jp/) and NJplot [53]. (D) Comparison of sequences of MACO-1 and orthologs of MACO-1 in other species and mutation sites of maco-1. Black and gray boxes highlight identical and similar residues, respectively. Black bars and gray bars above the sequences indicate predicted transmembrane domains (TM1-TM5) and coiled coil domains (CC1-CC4) of MACO-1 in C. elegans, respectively. Boxes below the amino-acid alignment correspond to mutation sites of tm2917 and ok3165. doi:10.1371/journal.pgen.1001384.g002  The proposed neural circuit model for thermotaxis [6]. Temperature is sensed by AFD and AWC sensory neurons. The AIY-RIA circuit is involved in thermophilic behavior (i.e., movements to temperatures higher than the cultivation temperature indicated as 'T'). The AIZ-RIA circuit is involved in cryophilic behavior (i.e., movements to temperatures lower than the cultivation temperature indicated as 'C'). RIA neurons are proposed to integrate these signals.  Table S2. Bars show the percentage of animals that moved to the cultivation temperature (20uC). Error bar indicates the standard error of the mean (SEM). Asterisks indicate statistically significant differences between fractions of 20 of 20uC-cultivated transgenic strains relative to fractions of 20 of 20uC-cultivated maco-1 mutants. One asterisk denotes statistical significance at p,0.05 and two asterisks denote statistical significance at p,0.01 (ANOVA with a Dunnett's post hoc test; n = 120-680). doi:10.1371/journal.pgen.1001384.g004 possibility, we observed the localization of several membrane proteins that could influence abnormal behaviors of maco-1 mutants. We found slightly increased spacing between adjacent SNB-1::GFP puncta (presynaptic marker) in DD motor neurons of maco-1 mutants compared with wild-type animals ( Figure S5, Text S1 and [40,41]). This change could be attributable to MACO-1 being directly associated with the maturation of presynaptic proteins. However, we cannot exclude the possibility of reduced excitability or altered Ca 2+ signaling in maco-1 mutant neurons resulting in weak mislocalization of presynaptic proteins. Further studies on MACO-1 targets may reveal its molecular function and involvement in several intracellular events, such as the maturation of presynaptic proteins and proper neuronal activity.

Behavioral analysis
Animals were grown at 20uC for all behavioral assays in this study, unless otherwise described. The procedure for the thermotaxis assay using radial temperature gradients was based on previously published research by Mori and Ohshima [6] with some modification. Evaluation of thermotaxis was consistent with methods of Mohri et al. [7]. Thermotaxis of individual animals on a radial temperature gradient was evaluated using the following four phenotypic categories: animals that moved to the cold region (i.e., center of plate) were classified as '17'; animals that moved to the 20uC region were classified as '20'; animals that moved to warm regions (i.e., plate periphery) were classified as '25'; animals that moved to cold and warm regions were classified as '17/25'. Chemotaxis to odorants was assayed according to Bargmann et al. [44], except that the assay plates contained a slightly different   -1(nj34). (E) maco-1(nj34) expressing maco-1 cDNA in AFD, AWC, AIY, AIZ and RIA neurons. This transgenic strain showed a partially rescued phenotype in thermotaxis behavior ( Figure 4D). (F) Duration between omega turns. Initial 1000-second video images of each animal were analyzed by image-processing software, MATLAB. n = 4 (wild-type), 3 (maco-1) and 2 (transgenic strain). doi:10.1371/journal.pgen.1001384.g005 medium (2% agar, 1 mM magnesium sulfate (MgSO 4 ), 1 mM calcium chloride (CaCl 2 ) and 25 mM potassium phosphate (pH 6)). Dilutions of odorants with ethanol were as follows: 1:1000 diacetyl, 1 mg/ml pyrazine, 1:100 isoamyl alcohol, and 1:100 benzaldehyde. The chemotaxis index was calculated according to Bargmann et al. [44]. Chemotaxis to NaCl was essentially as described previously [15]. Briefly, animals were scored as 'normal' when they migrated repeatedly to the concentration peak or remained at the peak during the assay. The animal was scored as 'partially defective' when it migrated toward the peak only once or stayed very briefly at the peak. The animal was scored as 'defective' when it failed to migrate to the peak.
The unc-14p::FLJ10747 shown in Figure 4 was constructed by ligation of the unc-14 promoter into pPD49.26 vector to generate pMYA171. The full-length cDNA of human macoilin, FLJ10747 cDNA, was PCR amplified from a Mammalian Gene Collection (MGC) clone (Invitrogen), which contained full-length FLJ10747 cDNA. This product was inserted into pMYA171 to generate unc-14p::FLJ10747 (pMYA172).

Video analysis
Animal behavior was recorded using a custom made tracking and live imaging system. Well-fed single animal was placed on a TTX plate for tracking experiments. An ordinal analog video camera (Olympus) captured freely moving individual animal for 30 min with adequate sampling rate (30 frames/sec). A computercontrolled microscope stage was automatically moved to center animals in the visual field using a custom image analysis algorithm with microscope software (MetaMorph: Universal Imaging Corporation). Midlines of recorded animals were extracted from each image, and curvatures at certain positions were calculated by analyzing points along midline with constant intervals. Spatiotemporal curvature maps were constructed by expressing the quantity of curvatures with a color-map (x-axis = time and yaxis = position along midline of animal). A smoothing filter (363 dimensions) was applied to the curvature map to reduce noise effects. Omega turn states (i.e., head of the animal curls to touch its own body; Figure 5B) were identified by an independent algorithm and expressed as white color in the curvature map. All video images were analyzed with MATLAB (Mathworks, Natick, MA).

In vivo calcium imaging and data analysis
In vivo calcium imaging was performed according to previous studies [21,50]. Well-fed animals expressing yellow cameleon 3.60 [51] caused by AFD and AIY promoters, gcy-8p::yc3.60 (pSAS309) and AIYp::yc3.60 (pNR86.79), were glued onto a 2% agar pad on glass, immersed in M9 buffer, and covered by cover glass. The agar pad and M9 buffer were kept at the initial imaging temperature. Sample preparation was completed within one minute. The sample was then placed onto a Peltier-based thermocontroller (Tokai Hit, Japan) on the stage of a Nikon E600 or Olympus BX61WI microscope at the initial imaging temperature for two minutes. Fluorescence was introduced into W-View (Hamamatsu Photonics, Japan) or Dual-View (Molecular devices, USA) optic systems. Cyan fluorescent protein (CFP; F480) and yellow fluorescent protein (YFP; F535) images were simultaneously captured by a CCD camera HiSCA or EM-CCD camera C9000-13 ImagEM (Hamamatsu Photonics).
Images were taken with a 200-ms exposure time with 464 or 262 binning. Temperatures on the agar pad were monitored by the thermometer system, DCM-20 (Tokai Hit and Hamamatsu Photonics). For each imaging experiment, fluorescence intensities of F535 and F480 were measured using AquaCosmos (Hamamatsu Photonics) or MetaMorph (Molecular Device) imaging analysis systems. Relative increases or decreases in intracellular Ca 2+ concentrations were measured as increases or decreases in YFP/CFP fluorescence ratios of cameleon protein (ratio change).

Statistical analysis
Error bars in figures indicate standard errors (SEM). The statistical analysis for behavioral experiments was performed by one-way analysis of variance (ANOVA) for multiple comparisons followed by post hoc Dunnett's multiple comparison. A single asterisk and double asterisk indicate statistical significance at p,0.05 and p,0.01, respectively. The value of p is probability. Asterisks indicate statistically significant differences between the index of maco-1(nj21) or maco-1(nj34) and index of wild-type at each concentration. One asterisk denotes statistical significance at the p,0.05 level and two asterisks denote statistical significance at the p,0.01 level (ANOVA with a Dunnett's post hoc test). (E) Quantification of distance between adjacent puncta. We measured the distance between adjacent puncta and categorized it into sixteen classes. The x-axis and yaxis show each class and fraction of distance categorized into one of sixteen classes, respectively. Error bar indicates the standard error of the mean (SEM). The statistical difference was determined by using a two-tailed Student's test. A single asterisk indicates statistically significant difference at p,0.05. n = 27 (wild-type) and 30 (maco-1). Table S2 Expression patterns driven by each promoter. We used lin-11p as an AIZ promoter, glr-3p as a RIA promoter and odr-1p as an AWC promoter ( Figure 4D). Detailed information about expression patterns of unc-14p, gcy-8p, gcy-18p, ttx-3p, odr-1p and flp-13p are shown in previous studies [17,40,43,[45][46][47]