Altered Function of the DnaJ Family Cochaperone DNJ-17 Modulates Locomotor Circuit Activity in a Caenorhabditis elegans Seizure Model

The highly conserved cochaperone DnaJ/Hsp40 family proteins are known to interact with molecular chaperone Hsp70, and can regulate many cellular processes including protein folding, translocation, and degradation. In studies of Caenorhabditis elegans locomotion mutants, we identified a gain-of-function (gf) mutation in dnj-17 closely linked to the widely used e156 null allele of C. elegans GAD (glutamic acid decarboxylase) unc-25. dnj-17 encodes a DnaJ protein orthologous to human DNAJA5. In C. elegans DNJ-17 is a cytosolic protein and is broadly expressed in many tissues. dnj-17(gf) causes a single amino acid substitution in a conserved domain, and behaves as a hypermorphic mutation. The effect of this dnj-17(gf) is most prominent in mutants lacking GABA synaptic transmission. In a seizure model caused by a mutation in the ionotropic acetylcholine receptor acr-2(gf), dnj-17(gf) exacerbates the convulsion phenotype in conjunction with absence of GABA. Null mutants of dnj-17 show mild resistance to aldicarb, while dnj-17(gf) is hypersensitive. These results highlight the importance of DnaJ proteins in regulation of C. elegans locomotor circuit, and provide insights into the in vivo roles of DnaJ proteins in humans.

GABA glutamic acid decarboxylase acetylcholine receptor excitation inhibition balance Cells have molecular mechanisms to protect themselves from the stress caused by misfolded or aggregated proteins. DnaJ/Hsp40 family proteins are highly conserved through evolution and act as cochaperones by interacting with and activating the ATPase activity of Hsp70 chaperone proteins (Ohtsuka and Suzuki 2000;Qiu et al. 2006). Together, Hsp40 and Hsp70 help folding of nascent proteins and refolding and degradation of misfolded proteins. Accumulation of protein aggregation underlies various human diseases including neurodegenerative diseases such as Parkinson's disease and Huntington's disease (Muchowski and Wacker 2005;Sherman and Goldberg 2001). Mutations in DnaJ/Hsp40 proteins have been associated with such diseases, suggesting the importance of cochaperones in cellular protein homeostasis (Blumen et al. 2012;Borrell-Pagès et al. 2006;Trinh and Farrer 2013). In C. elegans, overexpression of polyglutamine repeats in muscles or neurons causes formation of protein aggregation in an agedependent manner (Brignull et al. 2006a,b), similar to that observed in human polyglutamine diseases (Sakahira et al. 2002). In addition, excess excitatory neuronal signaling at the neuromuscular junction causes locomotion defects and increased protein aggregation in muscles in a C. elegans polyglutamine disease model, suggesting that neuronal activity can affect protein homeostasis in other tissues (Garcia et al. 2007).
Functions of neural circuits depend critically on balanced activity between excitatory and inhibitory transmission. In C. elegans, locomotion is controlled by the coordinated activities of excitatory cholinergic and inhibitory GABAergic motor neurons (Von Stetina et al. 2006). GABA plays crucial roles in the nervous system of both vertebrates and invertebrates. In C. elegans, mutants affecting GABA transmission were isolated from forward genetic screens for locomotor defects (Brenner 1974;Jin et al. 1999;McIntire et al. 1993). The C. elegans genes required for GABA neurotransmission including unc-25/GAD (Jin et al. 1999), unc-47/VGAT (McIntire et al. 1997, and unc-49/GABA A R (Bamber et al. 1999;Richmond and Jorgensen 1999) are highly conserved among animals. Analysis of unc-25/GAD mutants has revealed that the canonical reference allele unc-25(e156) causes a premature termination codon (Trp383amber) in the enzymatic domain; e156 mutants completely lack GABA immunoreactivity and have been widely used as representative of complete loss of GABA function.
The nicotinic acetylcholine receptor subunit acr-2 is expressed in the cholinergic motor neurons. A gain-of-function mutation of acr-2 causes increased cholinergic motor neuron activity accompanied by decreased GABAergic motor neuron activity, generating excitationinhibition (E/I) imbalance in locomotor circuit. acr-2(gf) animals exhibit a characteristic repetitive convulsion behavior, the frequency of which provides a quantitative measure of E/I imbalance (Jospin et al. 2009;Stawicki et al. 2013).
Through studying the effects of defective GABAergic transmission on acr-2(gf) animals, we unexpectedly found a gain-of-function mutation in a cochaperone protein dnj-17 to be present in the widely used strain . We show that the DNJ-17 gain-of-function mutation behaves in a hypermorphic manner, and exacerbates excitationinhibition imbalance in acr-2(gf). Null mutations of dnj-17 exhibit mild resistance to aldicarb, suggesting a role in modulating neurotransmission. Homologs of DNJ-17 include human DNAJA5, which is expressed in the brain and other tissues. Our findings provide insights into the in vivo function of these cochaperone proteins.

Molecular biology and transgenes
Molecular biology was performed following standard methods. Gateway recombination technology (Invitrogen, CA) was used for expression vectors. Table S2 describes the details of constructs generated in this study. We amplified 3.5 kb genomic sequences of dnj-17 with 0.9 kb 59 upstream sequences to 0.1 kb 39 downstream region using the following primers: YJ11121 AAACTCCATCAACCTGACTTCCCTG and YJ11122 TTGCCCATTATTCTTCCCGAAAC. To determine the gene structure of dnj-17, we isolated mRNAs from mixed-stage animals of N2 wild type and CB156 unc-25(e156) dnj-17(ju1162) using Trizol (Thermo-Fisher Scientific). Complementary DNA (cDNA) synthesis was performed using SuperScript III (ThermoFisher Scientific), with random primers according to the manufacturers' instructions. We performed RT-PCR using SL1 primer GTTTAATTACCCAAGTTTGAG and a reverse primer p3 GCGACCAGATTCCTAATTTGCTCGTTC designed on the junction of exon 3 and exon 4 to determine the first exon of dnj-17 mRNA, and p2 ATGAAATGCCATTACGAAGTGCTC and p4 AATGTTTCACCAATCCTCATCATCC primers designed on the first and sixth exon to verify the coding sequence. Sequences of all clones were verified by Sanger sequencing. Protein domain analysis was performed using NCBI domain database (Marchler-Bauer et al. 2015) and Treefam (Li et al. 2006).
Generation of deletion alleles of dnj-17 by CRISPR-Cas9 editing dnj-17(ju1239) and dnj-17(ju1276) deletion alleles were generated by CRISPR-Cas9 editing in the germline, using modifications of previously described methods (Dickinson et al. 2013) (Z. Wang and Y.J., unpublished data). Briefly, adult animals were injected with the Cas9-sgRNA expression constructs (pCZGY2647 and pCZGY2646, made from pDD162 with sgRNA) and Pmyo-2-mCherry as a co-injection marker. F1 animals expressing mCherry in pharynx were isolated, allowed to lay eggs, and then genotyped for dnj-17 to detect deletions. The F2 progeny of F1 animals with deletions were isolated to establish strains containing dnj-17 deletion. sgRNA sequences used to target dnj-17 are the following: ACAGAAAACTAGCGCTCAAA and GAGTTTGGCGACAAGGATAC. ju1239 was generated following microinjection into N2 animals. ju1276 was generated following microinjection into CB156 unc-25(e156) animals.
To analyze the temperature effects on dnj-17(ju1239), we examined the growth and locomotion of N2 and CZ21429 dnj-17(ju1239) under different temperature conditions. Briefly, 10 gravid adults of each strain were allowed to lay eggs for 6 hr at 22.5°. Then, adult animals were removed, and the plates with embryos were kept under 15°, 22.5°, or 25°. Hatched progeny were kept under the same temperature, and their growth and general locomotion were visually scored once within 16224 hr. When the progeny reached L4 stage, animals from each condition were placed onto individual plates. Number of eggs laid by each animal was scored to compare the brood size. The experiment was repeated twice.

Generation of single-copy inserted strains
Single-copy insertion transgenes of Pdnj-17-dnj-17(+) and Pdnj-17-dnj-17(ju1162) were generated at Chromosome II site ttTi5605 using modified vectors (Z. Wang and Y.J., unpublished data). Briefly, N2 young adult animals were injected with the following constructs: a construct (pCZGY3031 or pCZGY3032) containing dnj-17 sequence with ttTi5605 homology arms and a copy of hygromycin resistance gene, a construct (pDD122) which drives expression of Cas9 and sgRNA targeting ttTi5605 in the germline, and a Pmyo-2-mCherry fluorescent co-injection marker. F2 animals were selected for the resistance to hygromycin. Single-copy insertion lines were confirmed by PCR using primers designed outside of the homology arms. Loss of extrachromosomal array was confirmed by PCR and the loss of co-injection marker.
Each insertion line was outcrossed twice before being used in experiments.

Quantification of convulsion behavior
Scoring of convulsions was performed as previously described (Stawicki et al. 2013). Briefly, L4 larvae were transferred to nematode growth medium (NGM) plates seeded with Escherichia coli OP50. On the following day, young adults were transferred to fresh plates with OP50 and visually scored for convulsion behavior under a dissecting microscope. The observer was blinded to the genotype of the animals tested. A convulsion event was defined as a shortening of the animal's body length. The assay was repeated at least twice per genotype in two different generations. Two independent transgenic lines were used for each construct.

Aldicarb assay
One day before the experiment, L4 animals were transferred to fresh plates seeded with OP50. On the next day, 10 animals were transferred to an NGM plate with 500 mM aldicarb. Animal behavior was scored every 30 min. Animals were scored paralyzed when they did not move for more than 5 sec in response to touch stimulus.
Confocal microscopy L4 animals were imaged using a Zeiss LSM 710 confocal microscope (63· objective). Animals were immobilized by 1 mM levamisole and placed on 4% agar pads. Images are maximum-intensity projections of z stacks obtained at 1 mm intervals. ImageJ was used to process the images obtained. Data availability Strains and constructs are described in Table S1 and Table S2 respectively, and are available upon request.

RESULTS AND DISCUSSION
Identification of a missense mutation in dnj-17 in unc-25(e156) strains acr-2(n2420gf) animals show spontaneous convulsion behavior, due to increased cholinergic excitation and reduced GABAergic inhibition (Jospin et al. 2009). We wanted to further examine the effects of GABAergic transmission on the convulsive behavior of acr-2(gf) animals. We generated double mutants of acr-2(gf) with genes essential for GABA signaling, using canonical or null alleles of unc-25/GAD, unc-47/ VGAT, unc-49/GABAR. We used three null mutations of unc-25: e156, n2324, and n2328, which cause amber stop codons at Trp383, Trp291, and Glu486, respectively, and which are all predicted to encode truncated proteins that lack the cofactor binding site and enzymatic activity site at the C-terminus. All double mutants showed increased convulsion frequency compared to acr-2(gf) single mutants ( Figure 1A). While unc-25(n2324) and unc-25(n2328) enhanced acr-2(gf) behavioral defects to similar degrees as unc-47(gk192) and unc-49(e382), unc-25(e156) increased the convulsion frequency significantly more than these four mutations. Further outcrossing of unc-25(e156); acr-2(gf) (MT6648) did not eliminate this enhancement. We thus hypothesized that the ancestral CB156 strain may contain additional modifier mutation(s) linked to unc-25(e156).
dnj-17 encodes a homolog of human DNAJA5 Gene structure (Wormbase WS251) showed that dnj-17 contains seven exons, generating a mature mRNA predicted to encode a protein of 510 amino acids. To verify the dnj-17 gene structure we performed cDNA analyses using mRNA isolated from N2 and CB156 unc-25(e156) dnj-17(ju1162). RT-PCR analyses using SL1 and gene-specific primers revealed that the 59 end of dnj-17 mRNA contained an SL1 leader, but predicted exon 1 was not present in the mature mRNA. We obtained full-length dnj-17 cDNA and found that DNJ-17 protein consists of 485 amino acids (Figure 1, B and C). From NCBI protein domain analysis, the N-terminus of DNJ-17 has a highly conserved DnaJ domain, known to interact with Hsp70 family proteins, and the C-terminal half contains two C2H2-type zinc finger motifs that have been implicated to be important for polypeptide binding (Banecki et al. 1996;Lu and Cyr 1998;Szabo et al. 1996). Relatives of DNJ-17 are found widely in eukaryotes, with orthologs named as JJJ1 in yeast, DNAJA5/DNAJC21 in human, DNAJC21 in mouse, and CG2790 in Drosophila ( Figure 1D). The J domain of DNJ-17 also has a highly conserved HPD motif that is crucial for interaction with Hsp70 proteins (Tsai and Douglas 1996). Yeast JJJ1 activates ATPase activity of Hsp70, and lack of JJJ1 results in cold sensitivity (Meyer et al. 2007). On the other hand, functions of the DNJ-17 family proteins in animals remain mostly unknown, though human DNAJA5 is expressed in several tissues including the brain (Chen et al. 2004). We confirmed that cDNAs from CB156 unc-25(e156) dnj-17(ju1162) contained a single nucleotide change, which causes Asp77 to Lys amino acid substitution (N77K) in the region immediately adjacent to the DnaJ domain (Figure 1, B and D).

dnj-17(ju1162) is a gain-of-function mutation
Several lines of evidence support that ju1162 is a gain-of-function mutation of dnj-17. First, we examined the effect of a dnj-17 deletion allele dnj-17(tm570), which removes the C-terminal half of the protein ( Figure 1B). dnj-17(tm570) homozygous animals showed normal growth rate, wild-type locomotion, and did not affect convulsion frequency of acr-2(gf) (Figure 2A). We also generated unc-47(gk192) dnj-17(tm570); acr-2(n2420) triple mutants and found that they resembled unc-47(gk192); acr-2(n2420) double mutants in their convulsion frequency. As dnj-17(tm570) mutants potentially produce mRNAs encoding a truncated protein with intact DnaJ domain, we next generated a deletion allele targeting the DnaJ domain using CRISPR-Cas9-mediated genome editing technology (Dickinson et al. 2013;Friedland et al. 2013) (Z. Wang and Y.J., unpublished results). dnj-17(ju1239) removes a large portion of the DnaJ domain and is predicted to cause a frameshift and premature stop after amino acid 34 ( Figure 1B). Since the null mutation of a yeast protein with J domain, Jjj1, was previously reported to cause cold sensitivity in yeast (Meyer et al. 2007), we examined the viability and locomotion of dnj-17(ju1239) mutants. The mutant animals had similar brood size as wild type under three temperature conditions ( Figure 2B). Their growth rate, body shape, and movement were also indistinguishable from wild type. Finally, dnj-17(ju1239) did not affect the convulsion frequency of acr-2(n2420) (Figure 2A). These observations show that dnj-17 is a nonessential gene for C. elegans development and behavior, and that dnj-17 loss-of-function does not affect convulsion of acr-2(gf) by itself or when GABA transmission is eliminated in unc-47(null) animals.

DNJ-17 is a cytosolic protein expressed in multiple tissues
We next analyzed the expression pattern of dnj-17. We first generated an extrachromosomal transcriptional green fluorescent protein (GFP) reporter using 0.9 kb promoter region of dnj-17. GFP was seen throughout the body with enrichment in the intestine and several cells around the pharynx, and the expression pattern was similar in both wild type and acr-2(gf) background ( Figure 3A). We then made GFP-fused translational DNJ-17 reporters. GFP signals from Pdnj-17-dnj-17:: gfp localized to the cytosol of head neurons, and in other unidentified cells at lower levels throughout the body ( Figure 3B). A weaker but similar pattern was observed in an integrated fosmid expression line which expresses DNJ-17 tagged with C-terminal TY1::EGFP:3xFLAG (not shown) (Zhong et al. 2010). Moreover, DNJ-17(N77K)::GFP showed similarly diffused expression. Both DNJ-17(+)::GFP and DNJ-17(N77K) expression patterns were similar in wild type and in acr-2(gf) background, suggesting that the presence of acr-2(gf) does not largely affect the localization of DNJ-17.

DNJ-17(N77K) behaves as a hypermorph
We next examined the nature of DNJ-17(N77K) using transgenic overexpression. Overexpression of wild-type dnj-17 by genomic sequence of dnj-17 including 0.9 kb upstream promoter region caused increase of acr-2(gf) convulsion frequency ( Figure 4A). This transgene also enhanced convulsion frequency in unc-47(0) mutant background, suggesting that the increase in convulsion by overexpression of dnj-17 is independent of the effect caused by defects in GABAergic transmission. Interestingly, this enhanced effect was also observed by overexpression of dnj-17(ju1162gf). ACR-2 is expressed specifically in neurons (Jospin et al. 2009). However, overexpression of dnj-17 wild type or ju1162(gf) using the Prgef-1 pan-neuronal promoter did not affect convulsion frequency of acr-2(gf) ( Figure 4B). Also, overexpression in muscles using Pmyo-3 promoter did not affect acr-2(gf) convulsion frequency ( Figure 4B). These results suggest that the effect of dnj-17 on convulsion frequency likely requires its expression in multiple tissues.
dnj-17 activity affects the response to aldicarb To further assess the effect of dnj-17 mutations on neurotransmission at the neuromuscular junction, we examined the sensitivity of the mutant animals to an acetylcholine esterase inhibitor aldicarb. dnj-17(ju1239) null animals showed mild resistance to aldicarb, which was rescued by single-copy insertion of dnj-17(+), implying that dnj-17 affects cholinergic transmission at the neuromuscular junction ( Figure 4D). Expression of dnj-17(ju1162gf) caused mildly increased sensitivity to aldicarb, reaching statistical significance at one time point, consistent with this allele being a hypermorph mutation. These results raise a possibility that the function of dnj-17 is required for folding and/or function of proteins in multiple tissues that are involved in cholinergic transmission. Overexpression of wild-type DNJ-17 may also lead to a high level of cholinergic transmission by contributing to folding of the proteins in the pathway.

Perspectives
Other mechanisms might account for the effects of dnj-17(ju1162). The N77K mutation may make DNJ-17 protein prone to form aggregates. DnaJ/Hsp40 proteins bind to misfolded proteins and bring them to Hsp70 (Cheetham and Caplan 1998). The N77K mutation could alter the kinetics for DNJ-17 to detach from the protein(s) it binds to, and prevent the misfolded protein from being degraded, resulting in accumulation of misfolded proteins that cause cellular stress. Such cellular stress could alter neuronal and muscular functions. Another possibility is that the mutation disrupts certain cellular functions. Recently it was reported that an Asn to Ser mutation in the DnaJ domain of human DNAJC13 was found in a family with Parkinson disease, where the disease was transmitted in an autosomal-dominant manner (Vilariño-Güell et al. 2014). The mutant protein exhibited a toxic gain-of-function activity affecting endosomal transport. The N77K mutation in C. elegans DNJ-17 may affect similar cellular functions such as endocytosis and Figure 4 Single-copy expression of dnj-17(ju1162gf) is sufficient to cause increase in convulsion frequency. (A2C) Quantification of convulsion frequency. (A) Overexpression of dnj-17(+) or dnj-17(ju1162gf) by high-copy extrachromosomal arrays causes increase of acr-2(gf) convulsion frequency. (B) Neuron-or muscle-specific overexpression of dnj-17 does not affect convulsion frequency. (C) Single-copy expression of dnj-17(ju1162gf), but not dnj-17(+), causes increase in convulsion frequency. Statistics, one way ANOVA followed by Bonferroni's post hoc test. Ã P , 0.05. Error bars indicate SEM. Numbers in the column indicate sample sizes. (D) Aldicarb resistance of dnj-17(ju1239) null is suppressed by expression of single-copy wild-type dnj-17. dnj-17(ju1239) animals with single-copy expression of dnj-17(ju1162gf) showed mildly increased sensitivity to aldicarb. Statistics, two way ANOVA. Ã P , 0.05 compared to N2 at the given time point. subcellular trafficking, thus disrupting the coordination of the motor neuron circuit.
E. coli has only one gene coding DnaJ/Hsp40, whereas animals typically express multiple DnaJ family members. The DnaJ protein in E. coli has been well characterized, but functions of individual DnaJ/Hsp40 family proteins in animals remain largely unknown. DNAJA5, the closest human homolog of DNJ-17, shows enhanced expression in the brain (Chen et al. 2004) which suggests neuron-specific roles, but its substrates and functions are yet to be characterized. Studies of DnaJ/Hsp40 in animals may lead to further understanding of the physiological mechanisms of protein homeostasis in neurodegenerative diseases.