INVESTIGATION

Systematic Analyses of rpm-1 Suppressors RevealRoles for ESS-2 in mRNA Splicing in

Caenorhabditis elegansKentaro Noma,*,† Alexandr Goncharov,*,† and Yishi Jin*,†,1

*Neurobiology Section, Division of Biological Sciences, and †Howard Hughes Medical Institute, University of California, San Diego,La Jolla, California 92093

ABSTRACT The PHR (Pam/Highwire/RPM-1) family of ubiquitin E3 ligases plays conserved roles in axon patterning and synaptic de-velopment. Genetic modifier analysis has greatly aided the discovery of the signal transduction cascades regulated by these proteins. InCaenorhabditis elegans, loss of function in rpm-1 causes axon overgrowth and aberrant presynaptic morphology, yet the mutant animalsexhibit little behavioral deficits. Strikingly, rpm-1mutations strongly synergize with loss of function in the presynaptic active zone assemblyfactors, syd-1 and syd-2, resulting in severe locomotor deficits. Here, we provide ultrastructural evidence that double mutants, betweenrpm-1 and syd-1 or syd-2, dramatically impair synapse formation. Taking advantage of the synthetic locomotor defects to select forgenetic suppressors, previous studies have identified the DLK-1 MAP kinase cascade negatively regulated by RPM-1. We now reporta comprehensive analysis of a large number of suppressor mutations of this screen. Our results highlight the functional specificity of theDLK-1 cascade in synaptogenesis. We also identified two previously uncharacterized genes. One encodes a novel protein, SUPR-1, thatacts cell autonomously to antagonize RPM-1. The other affects a conserved protein ESS-2, the homolog of human ES2 or DGCR14. Lossof function in ess-2 suppresses rpm-1 only in the presence of a dlk-1 splice acceptor mutation. We show that ESS-2 acts to promoteaccurate mRNA splicing when the splice site is compromised. The human DGCR14/ES2 resides in a deleted chromosomal regionimplicated in DiGeorge syndrome, and its mutation has shown high probability as a risk factor for schizophrenia. Our findings providethe first functional evidence that this family of proteins regulate mRNA splicing in a context-specific manner.

PROPER synapse development ensures precise wiring andefficient transmission of information in the nervous sys-

tem. In the presynaptic terminals, dense and organized accu-mulation of synaptic vesicles around the docking site calledactive zone is essential for rapid release of neurotransmitters(Sudhof 2012). Forward genetic screens in Caenorhabditiselegans, aided by visualization of synapse morphology usingfluorescent reporters, have made important contributions touncover conserved genes and pathways regulating presynap-tic assembly (Jin and Garner 2008; Ou and Shen 2010).Many such screens have identified a common set of genesthat regulate specific aspects of presynaptic differentiationin many neurons (Ackley and Jin 2004; Jin 2005; Maeder

and Shen 2011). Among them are SYD-2/a-Liprin (LAR in-teracting protein), which has a central role in presynapticactive zone formation (Zhen and Jin 1999; Dai et al. 2006);SYD-1, a PDZ-RhoGAP protein that acts upstream of SYD-2(Patel and Shen 2009; Hallam et al. 2002); and RPM-1, an E3ubiquitin ligase that regulates synaptic organization (Schaeferet al. 2000; Zhen et al. 2000).

Despite their strong defects in synaptic morphology, singleloss-of-function (lf) mutants of rpm-1, syd-1, or syd-2 generallyexhibit subtle to mild impairment in most behavioral assays.However, double mutants of rpm-1(lf) with syd-1(lf) or syd-2(lf) show more severe uncoordinated movement thanexpected from simple additivity of the single mutants (Liaoet al. 2004; Nakata et al. 2005), indicating that rpm-1 acts inparallel to syd-1 and syd-2. This synthetic behavioral deficit ofdouble mutants has allowed powerful selections for geneticsuppressor mutations. Previous characterization of multiplerpm-1(lf) suppressors uncovered the DLK-1 MAP kinase path-way, consisting of dlk-1/MAPKKK, mkk-4/MAPKK, pmk-3/MAPK, and mak-2/MAPKAP2, uev-3/E2 ubiquitin-conjugating

Copyright © 2014 by the Genetics Society of Americadoi: 10.1534/genetics.114.167841Manuscript received July 8, 2014; accepted for publication August 28, 2014; publishedEarly Online September 5, 2014.Supporting information is available online at http://www.genetics.org/lookup/suppl/doi:10.1534/genetics.114.167841/-/DC1.1Corresponding author: 9500 Gilman Drive, La Jolla, CA 92093-0368.E-mail: yijin@ucsd.edu

Genetics, Vol. 198, 1101–1115 November 2014 1101

enzyme variant, and cebp-1, a bZip-domain transcription fac-tor acting downstream of the DLK-1 kinase cascade (Nakataet al. 2005; Yan et al. 2009; Trujillo et al. 2010). Furtheranalysis of specific DLK-1 missense alterations also providedkey insights into the mechanism of DLK-1 activation (Yan andJin 2012). In addition, the identification of a gain-of-functionallele of syd-2 as a suppressor of syd-1 revealed the order ofactivity in presynaptic active zone assembly (Dai et al. 2006).

To gain a more complete understanding of the geneticinteraction landscape on the behavioral suppression of rpm-1; syd-1 or rpm-1; syd-2, here, we have characterized 95suppressors. We uncovered a new allele of the sup-5 ambersuppressor gene as the sole mutation suppressing syd-2. Ofthe 94 rpm-1 suppressors, 92 affected previously knowncomponents of the DLK-1 cascade. Interestingly, althoughthe MLK-1 kinase cascade acts in parallel to the DLK-1 path-way in axon regeneration (Nix et al. 2011), we did notidentify any mutations in the MLK-1 cascade in our screen.We showed that the loss-of-function mutations in this path-way do not suppress rpm-1 developmental defects. Of thetwo remaining rpm-1 suppressors, one defined a novel genesupr-1. The other suppressor resulted in a loss of function inthe homolog of the human DGCR14 gene, ess-2. Our analy-sis of a synthetic interaction between ess-2 and a splicingmutation of dlk-1 provides the first evidence that ess-2 isfunctionally important in mRNA splicing.

Materials and Methods

C. elegans genetics and strains

We maintained C. elegans strains on NGM plates as de-scribed (Brenner 1974). We used juIs1[Punc-25-SNB-1::GFP] (Hallam and Jin 1998) to visualize presynaptic termi-nals of GABAergic motor neurons, muIs32[Pmec-7-GFP](Ch’ng et al. 2003) for axonal morphology of mechanosen-sory neurons. All strains were maintained at 22.5�. Muta-tions and integrated lines were generally outcrossed to N2more than four times before analyses. Strains and their gen-otypes are summarized in Supporting Information, Table S2.

Suppressor screen, genetic mapping, and whole-genome sequencing

Screens for suppressors of syd-1(lf); juIs1; rpm-1(lf) or juIs1;rpm-1(lf); syd-2(lf) were carried out using ethyl methane-sulfonate (EMS) mutagenesis following standard proce-dures as described (Brenner 1974; Nakata et al. 2005).These double-mutant animals had minimal voluntary move-ments and could not move continuously even upon physicalstimulation such as tapping. Suppressed worms showed con-tinuous sinusoidal movements although they were slowerthan wild type (Figure S2). We first sorted new suppressorsinto complementation groups by testing known loci, follow-ing the scheme shown in Figure S1. We then determined themolecular lesions in these mutations by Sanger sequencing.The remaining candidates were either sequenced for thecoding sequences of mak-2, uev-3, and cebp-1 or subjected

to whole-genome sequencing (WGS). In brief, genomic DNAwas prepared from three 15-cm plates for each strain usingPuregene Cell and Tissue kit (Qiagen, Valencia, CA) accord-ing to the manufacturer’s instructions. Purified genomic DNAswere analyzed at Beijing Genomics Institute (BGI Americas).The obtained raw sequences were compared to the refer-ence sequences in the WormBase using MAQGene software(Bigelow et al. 2009). By subtracting common variantsamong the cohort of suppressor strains, we determinedunique homozygous single nucleotide polymorphism(SNP) variants for each strain. We then outcrossed eachsuppressor strain to N2 and performed further linkage anal-ysis on the isolated recombinants based on the phenotypicsuppression using the unique SNPs. ju1041 was mapped tothe middle of chromosome III, in a region containing themutation in sup-5 gene. ju1118 was mapped to the middleof chromosome I, and only supr-1 had a premature stopcodon in this region. The suppression in CZ5301 dlk-1(ju600)I; syd-1(ju82)II; ess-2(ju1117)III; juIs1IV; rpm-1(ju44)V showed linkage to two regions, the right arm ofchromosome I and the middle of chromosome III. We founda splice mutation in dlk-1 on chromosome I. Complementa-tion test of dlk-1(ju600) to dlk-1(0) showed that dlk-1(ju600) per se has no noticeable suppression on behavioralphenotypes of syd-1(lf); rpm-1(lf). We then found a prema-ture stop codon mutation in the ess-2 gene on chromosome IIIto be necessary for the suppression of rpm-1(lf).

Behavioral analysis using a multiworm tracker

Locomotion speed was recorded using an automated multi-worm tracking system as previously described (Pokala et al.2014). Briefly, to restrict animal movement to the trackingarea, filter paper with a 1-square-inch hole was soaked with100 mM CuCl2 solution and placed onto a food-free agarplate. Ten young adult worms were transferred to a food-freeassay plate after briefly washing in M9 solution, and allowedto crawl for 20 min before recording. Video recordings wereacquired at 3 Hz using Streampix software (Norpix, Quebec,Canada) and a PL B-741F camera (PixeLINK, Ontario, Can-ada) with a Zoom 7000 lens (Navitar, Rochester, NY), andwere analyzed with custom software written in MATLAB(Mathworks). The average velocity of forward locomotionwas obtained from 5–10 worms for each session, and threesessions were averaged to obtain the average speed for eachgenotype.

DNA constructs

cDNAs were amplified from an N2 cDNA library using gene-specific primers shown below. The cDNAs were cloned intopCR8 or pDONR221 vector to make entry clones compatiblewith the Gateway system (Invitrogen, Carlsbad, CA), andthe DNA sequences were confirmed by Sanger sequencing.Wild-type (wt) ess-2 cDNA (F42H10.7b in WormBase, 1596bp including ATG to the stop codon; see Figure 5A) wascloned using the primers YJ9222: 59-ATGtcgtcgttcgacaaaaat-39and YJ9172: 59-TTAgaaaaagtctccagcatt-39 (pCZGY2377).

1102 K. Noma, A. Goncharov, and Y. Jin

ess-2(ju1117) and ess-2(ok3569) cDNAs (pCZGY2552 andpCZGY2553) were cloned from the cDNA libraries obtainedfrom each mutant strain. These entry vectors were recombinedwith Pmec-4-GTW (pCZGY553) to make Pmec-4–ESS-2 (wt,pCZGY2379; ju1117, pCZGY2554; ok3569, pCZGY2555) forthe rescue experiments. pCZGY2377 was recombined withPmec-4–GFP–GTW (pCZGY603) to generate Pmec-4–GFP::ESS-2 (pCZGY2381). To construct GFP-tagged ESS-2, ess-2cDNA (1587 bp without stop codon) was cloned using theprimers (YJ9222 and YJ10311: 59-gaaaaagtctccagcatttgc-39).This entry clone (pCZGY2378) was recombined with Pmec-4–GTW–GFP (pCZGY858) to make Pmec-4–ESS-2::GFP(pCZGY2380). supr-1 cDNA (Y71F9AR.3 in WormBase,2241 bp including ATG to the stop codon) was cloned usingthe primers YJ10312: 59-ATGaagctagaggattcttgtgcc-39 andYJ10313: 59-TCAatgttctaaatttttcgaaacaat-39. This entryclone (pCZGY2382) was recombined with pCZGY553 orpCZGY603 to generate Pmec-4–SUPR-1 (pCZGY2383) orPmec-4–GFP::SUPR-1 (pCZGY2384). A short isoform of supr-1lacking exon 4 (see Figure 3A) was identified, but not used inthis study. rpm-1 cDNA (C01B7.6 in WormBase) was clonedinto pDONR221 vector as three restriction enzyme-digestedfragments ligated into one piece (pCZGY2005). GFP wasinserted into the PstI site (pCZGY2017). This entry vectorwas recombined with Prgef-1–GTW (pCZGY66) to makePrgef-1–RPM-1::GFP (pCZGY2551).

Generation of transgenic C. elegans

Multicopy transgenic animals were generated as described(Mello et al. 1991). Pttx-3–RFP was used as a coinjectionmarker to adjust the total amount of DNA to 100 ng/ml. Torescue ess-2, we amplified the genomic fragment from N2 lysisusing YJ10528: 59-ccggccaaaggaacatccgaac-39 and YJ10529:59-aagtggataaaggtccggcaaga-39, and the DNA products werepurified and injected into CZ17517 dlk-1(ju600); muIs32; ess-2(tm4246); rpm-1(ju44) at 1 ng/ml. pCZGY2379, pCZGY2554,or pCZGY2555 at 10 ng/ml was injected into CZ17517.pCZGY2383 at 10 ng/ml was injected into CZ17724 supr-1(ju1118); muIs32; rpm-1(ju44) animals. For GFP-taggedESS-2 and SURP-1 the following constructs were injected totest their rescue activities: pCZGY2380 or pCZGY2381 at 90ng/ml was injected into CZ17517. pCZGY2384 at 10 ng/mlwas injected into CZ17724. These strains were outcrossed toN2 to get rid of all the background mutations to observe GFPsignals. In general, more than three lines were analyzed foreach construct, and quantification data of one representativeline was shown. For RPM-1::GFP transgene (juEx3996),pCZGY2383 at 50 ng/ml was injected into CZ1234 rpm-1(ju23) animals. This extrachromosomal array (juEx3996)was integrated into a chromosome to generate juIs390 follow-ing ultraviolet trimethylpsoralen (UV/TMP) mutagenesis.

RNA preparation, RT-PCR, and qRT-PCR

Young adult worms were collected from three 60-mm NGMplates. Total RNA was extracted from whole animals usingTRIzol reagent (Invitrogen) and treated with DNase I (Roche,

Indianapolis, IN) for 20 min at 37�. Two micrograms of totalRNAwere reverse transcribed into cDNA using random primersand the SuperScript III First-Strand Synthesis system (Invitro-gen) following the manufacturer’s instructions. For semiquan-titative RT-PCR, cDNA fragments were amplified using theprimers: dlk-1 pair flanking ju600 mutation site (YJ10323:59-gtcgaaacccagcctgaac-39 and YJ10324: 59-ctcggtctccttcagttg-39; see Figure 6A), dpy-1 pair flanking e128mutation site (D1:59-cacattttcaagtgaagacttgag-39 and D2: 59-ggtcctggaatgcaa-catcct-39) (Aroian et al. 1993), and ama-1 pair (YJ2492: 59-gcattgtctcacgcgttcag-39 and YJ2493: 59-ttcttccttctccgctgctc-39)as a control and analyzed by agarose gel electrophoresis. Forquantitative RT-PCR Power SYBR Green PCR Master Mix (Ap-plied Biosystems, Foster City, CA) the kit was used for thereactions and the products were detected in real time usingthe ABI Prism 7000 Sequence Detection system. cDNAs wereamplified using the following primer pairs in which forward orreverse primers were designed to anneal the exon–exon junc-tion to avoid amplification from genomic DNA (see Figure 6A):dlk-1 exon 3–4 pairs (YJ10317: 59-gggctcttttcgtgttttcagc-39 andYJ10318: 59-cctcggatttttcaatttcaactg-39), dlk-1 exon 5–6 pairs(YJ10319: 59-tgttggaacaaacattctgagcc-39 and YJ10320: 59-ggagaatgagggacgattgcg-39), dlk-1 exon 8–9 pairs (YJ10321:59-gcgaatccgccaaaaatcctac-39 and YJ10322: 59-cggaatccctggccggtgat-39), and ribosomal subunit S25 (rps-25) pairs asan internal control (rps-25 F: 59-ctctacaaggaggtcatcacc-39and rps-25 R: 59-gacctgtccgtgatgatgaacg-39). The dlk-1mRNA levels were normalized to rps-25 for each productand then normalized to the control strain.

Bright-field and fluorescence microscopy

Bright-field images were photographed using a Leica MZ95stereomicroscope (Leica Microsystems, Buffalo Grove, IL).For quantification of the morphology of mechanosensoryneurons using muIs32[Pmec-7-GFP] reporter, young adultswere immobilized in 1% phenoxypropanol (TCI America,Portland, OR) in M9 buffer and scored under a Zeiss Axioplan2 microscope equipped with Chroma HQ filters and a 363(NA = 1.4) objective lens. Images of the fluorescent reporterswere collected from young adults immobilized in M9 with 1%phenoxypropanol using a Zeiss LSM710 confocal microscopeequipped with 363 (NA = 1.4) or 3100 (NA = 1.46) objec-tive lens. Images shown are maximum-intensity projectionsobtained from several z-sections (0.5–1.0 mm/section).

Electron microscopy

CZ333 juIs1, CZ840 juIs1; rpm-1(ju41); syd-2(ju37), andCZ1797 syd-1(ju82); juIs1; rpm-1(ju44) young adult animalswere fixed using glutaraldehyde and osmium as described(Hallam et al. 2002). Midanterior body was serially sectionedat 40–50 nm thickness. The ventral and dorsal nerve cordswere photographed using a JEOL 1200EX transmission elec-tron microscope (JEOL, Tokyo, Japan) at 80 kV. Presynapticterminals of neuromuscular junctions were defined by theexistence of clear synaptic vesicles and electron-dense activezones. The types of the motor neurons, cholinergic or

ESS-2 Regulates mRNA Splicing 1103

GABAergic, were defined by the synaptic connections (Whiteet al. 1986). Synaptic vesicles were morphologically deter-mined as 35–50 nm clear vesicles, and their numbers weremanually counted in the sections containing the active zone.

Statistical analysis

We analyzed the rpm-1(lf) phenotypes of the mechanosen-sory neurons using Fisher’s exact test in GraphPad Prism 5.0(GraphPad Software, La Jolla, CA). Locomotion speed andsynaptic vesicle numbers were analyzed using one-wayANOVA and Bonferroni’s multiple comparison test in Graph-Pad Prism 5.0.

Results

Presynaptic development is severely disrupted indouble mutants of rpm-1 with syd-1 or syd-2

We previously reported that null (0) mutations of rpm-1 resultin disorganized synapses (Zhen et al. 2000), and null muta-tions of syd-1 or syd-2 alter presynaptic active zone assembly(Zhen and Jin 1999; Hallam et al. 2002). By gross movement,rpm-1(0) is indistinguishable from wild type, and syd-1(0) andsyd-2(0) show slow movement while maintaining sinusoidalpattern (Figure 1, A and B). Strikingly, rpm-1; syd-1 and rpm-1; syd-2 double mutants show severely uncoordinated behav-ior from young larvae throughout mature adults (Figure 1B).Their body size and brood size are also smaller than wild-typeanimals or each single mutant (Figure 1A, and data notshown). We quantified locomotion behavior using a multi-worm tracker. Wild-type worms showed a coordinated sinusoi-dal movement, with an average forward locomotion speed of167.1 6 15.6 mm/sec. While movement speed for each singlemutant was similar to that of wild type, the rpm-1; syd-2 dou-ble mutants greatly reduced the speed to 18.2 6 0.7 mm/sec(mean 6 SEM, n = 3 experiments, 5–10 worms/experiment)(Figure 1, A and B). Panneuronal expression of rpm-1(+)using a transgene juIs390[Prgef-1-RPM-1::GFP] rescued boththe uncoordinated behavior and small body size of rpm-1;syd-2 double mutants (Figure 1, A and B), suggesting thatthe behavioral and body size defects arise from the loss ofrpm-1 function in neurons and that the body size may bea secondary effect associated with severe paralysis.

The behavioral deficits of these double mutants arestrongly correlated with severe disruption of synapses inmany neurons. The motor neurons form en passant synapsesonto ventral and dorsal body wall muscles (White et al.1986). A synaptic vesicle marker such as juIs1[Punc-25-SNB-1::GFP] readily enables visualization of the presynapticterminals of GABAergic motor neurons (Figure 1C) (Hallamand Jin 1998). rpm-1, syd-1, and syd-2 single mutant ani-mals each display distinctive alterations in presynaptic mor-phology, as previously described (Figure 1C) (Zhen and Jin1999; Zhen et al. 2000; Hallam et al. 2002). In rpm-1; syd-1or rpm-1; syd-2 double mutants, SNB-1::GFP puncta werebarely detectable in the presumed presynaptic regions alongthe nerve processes (Figure 1C). Examination of several

synaptic markers in other types of neurons revealed similarreduction in synapse number and abnormal synapse morphol-ogy (data not shown). The morphological defects of presyn-aptic terminals were observed from young larvae throughadults, consistent with the behavioral deficit throughoutdevelopment.

We further examined the presynaptic architecture ofmotor neurons using electron microscopy on serial ultrathinsections. Presynaptic terminals of these neurons are identi-fied as axonal swellings containing clusters of synapticvesicles and electron dense active zones (Figure 1D) (Whiteet al. 1986). We analyzed segments of ventral and dorsalnerve cords (�250 sections, corresponding to �11.25 mm)in two mutant strains (syd-1; juIs1; rpm-1 and juIs1; rpm-1;syd-2) and juIs1 as a control. Both double mutants displayeddramatically reduced numbers of synapses in cholinergicand GABAergic motor neurons (Table 1 and Figure 1, Eand F). Furthermore, the few remaining synapses containedvery few synaptic vesicles that were scattered near the pre-synaptic dense projections (Figure 1, D–F), indicating thatpresynaptic terminals were poorly developed and/or main-tained in the double mutants. These analyses show that therpm-1 pathway and the presynaptic active zone assemblypathway via syd-1/syd-2 function in parallel to promote pre-synaptic development.

Most suppressors of the developmental defects of rpm-1 affect the DLK-1 kinase cascade

We exploited the synthetic locomotor deficits of rpm-1; syd-1and rpm-1; syd-2 double mutants to select for mutations thatimproved movement (see Materials and Methods). Subse-quent analyses of the GABAergic presynaptic morphology us-ing the Punc-25-SNB-1::GFP marker allowed us to sort thesuppressor mutations into groups that displayed specific sup-pression on each gene. Although the screen should have anequal chance to isolate mutations suppressing either gene, ofthe 95 suppressors analyzed, we found only one allele,ju1041, that specifically suppressed syd-2(ju37), but notrpm-1(lf) (data not shown). syd-2(ju37) results in an amberstop codon at Glu397 (Zhen and Jin 1999). We mappedju1041 to the middle of chromosome III (Figure 2C) andidentified a single nucleotide change from C to T in sup-5,a tryptophan tRNA (seeMaterials and Methods). The mutatedtRNA in ju1041 can encode tryptophan for the amber stopcodon as previously described (Waterston and Brenner 1978;Wills et al. 1983). The isolation of sup-5(ju1041) as the onlysuppressor of syd-2(ju37) suggests that few genes can bemutated to bypass the requirement for syd-2(lf) in synapsedevelopment and function.

The remaining 94 suppressors we characterized werespecific for rpm-1. We previously reported multiple rpm-1suppressors that cause loss of function in components of theDLK-1/MKK-4/PMK-3 MAP kinase pathway (Figure 2A andFigure S2) (Nakata et al. 2005; Yan et al. 2009; Trujillo et al.2010). We therefore devised a complementation workflowto test whether new suppressors affected these known genes

1104 K. Noma, A. Goncharov, and Y. Jin

(Figure S1 and Materials and Methods). Among 94 suppres-sors, 41 were determined to be alleles of dlk-1 and 26 werealleles of pmk-3 (Figure 2B and Table S1). Among the newsuppressors that were not dlk-1 or pmk-3, some were linked

to syd-2 on the X chromosome, where mkk-4 and cebp-1reside (Figure 2C and Figure S2). Subsequent DNA sequenc-ing analysis of candidate genes, in conjunction with whole-genome sequencing of selected suppressors, revealed single

Figure 1 rpm-1(lf) synergizes with syd-1(lf) and syd-2(lf) to impair locomotion and presynaptic morphologies. (A) Bright-field images showing the bodyshapes of adult animals for the indicated genotypes. Bar, 500 mm. (B) Locomotion speed of worms for the indicated genotypes was analyzed usinga multiworm tracker. Average velocity of the forward movement was calculated for 15-min sessions. Error bar indicates SEM. Statistics: one-wayANOVA, *P , 0.05. (C) Presynaptic terminals in the dorsal cord visualized by juIs1[Punc-25-SNB-1::GFP], genotype of animals indicated. In rpm-1(lf)mutant, the numbers of presynaptic terminals are decreased and remaining synapses either overdeveloped or immature (Zhen et al. 2000). syd-2(lf)mutants have comparable number of synapses but the synaptic vesicles show diffused pattern (Zhen and Jin 1999). Hardly any SNB-1::GFP was observedin rpm-1(lf); syd-2(lf). Bar, 20 mm. (D) Electron micrographs of presynaptic terminals. Arrowheads and arrows in a–d indicate synaptic vesicles and activezones, respectively. In a9–d9 presynaptic varicosities of cholinergic and GABAergic neurons are shown in magenta and blue, respectively. Muscles areshown in green. AZ, active zone. Bar, 500 nm. (E and F) Average number of synaptic vesicles in sections containing active zones was analyzed forcholinergic neurons (E) and GABAergic neurons (F). n = no. of sections. Error bar indicates SEM. Statistics: one-way ANOVA, compared with control.***P , 0.001. Alleles and transgene used: rpm-1(ju23), rpm-1(ju44), rpm-1(ju41), syd-1(ju82), syd-2(ju37), and juIs390[Prgef-1-RPM-1::GFP].

ESS-2 Regulates mRNA Splicing 1105

nucleotide changes in all six genes of the DLK-1 pathway, assummarized in Table S1 and Figure 2B. Most of the nucleo-tide changes in dlk-1, mkk-4, or pmk-3 are predicted to causemissense alterations in the kinase domains (Yan and Jin2012) (Figure 2D, Figure S3, and Figure S4). Seven dlk-1alleles, two pmk-3 alleles, and one mak-2 allele were inde-pendent isolates of the same nucleotide changes from differ-ent mutagenized parents, suggesting that the mutagenesishits are likely saturated in the context of this screen for thebehavioral suppression (Table S1). Among the collection ofnew alleles, we found several mutations that revealed poten-tial roles of new functional domains. For example, two newalleles of pmk-3, ju1010 and ju592, are nonsense mutationsafter the kinase domain, suggesting that the C terminus ofPMK-3 may regulate the protein activity or stability. cebp-1(ju634) altered arginine 63 to proline in the N terminus ofthe protein, whereas the majority of the previously reportedcebp-1 mutations affect the bZip domain (Yan et al. 2009;Bounoutas et al. 2011) (Figure 2D and Table S1). In sum-mary, 98% of the suppressor mutations identified in ourscreens affected the six genes of the DLK-1 kinase cascade(Figure 2A and Table S1), and the number of the identifiedalleles roughly correlated with the size of the coding regionsof the genes (Figure 2B).

The MLK-1 kinase cascade has a minor role in thedevelopmental phenotypes of rpm-1 mutants

Among other MAP kinase cascades in C. elegans (Sakaguchiet al. 2004), the MLK-1/MEK-1/KGB-1 pathway, has recentlybeen shown to act in parallel to the DLK-1/MKK-4/PMK-3pathway in axon regeneration (Nix et al. 2011). MLK-1::GFP expression levels are increased in rpm-1 mutants, sug-gesting MLK-1, like DLK-1, may be a substrate for degrada-tion by RPM-1. The mlk-1, mek-1, and kgb-1 genes arecomparable in size to the genes of the dlk-1 pathway (Figure2B), and null mutants in these genes are viable and behav-iorally normal. However, we found none of the suppressormutations isolated in our screen affected the genes in the

MLK-1 pathway (Table S1). We therefore constructed a strainof kgb-1(0); rpm-1(lf); syd-2(lf), and found no improvementon locomotor phenotype (Figure 3A). Previous studiesshowed that mlk-1(0) and mek-1(0) did not suppress the pre-synaptic defects in GABAergic neurons of rpm-1(lf) (Nakataet al. 2005). Here, to further assess the activity of the mlk-1/mek-1/kgb-1 in neuronal development, we analyzed the axo-nal development of mechanosensory neurons. In rpm-1(lf)mutants the Posterior Lateral Microtubule (PLM) neurons dis-play highly penetrant axon overshooting and frequent ab-sence of synaptic branches (Schaefer et al. 2000) (Figure 3,B–D). These PLM axon developmental defects of rpm-1(lf) arecompletely suppressed by loss-of-function mutations in theDLK-1 cascade (Yan et al. 2009; Trujillo et al. 2010) (Figure3, C and D). In contrast, null mutations in mlk-1, mek-1, orkgb-1 showed rather mild suppression (Figure 3, C and D).kgb-2 is a closely related homolog of kgb-1, sharing 84%identity. We found that the kgb-1(0) kgb-2(0) double mutantalso did not strongly suppress rpm-1(lf) PLM axonal defects(Figure 3D). This additional analysis demonstrates the highlyspecific regulation of the DLK-1/MKK-4/PMK-3 cascade byRPM-1 in synapse and axon development.

Loss of function in a novel protein SUPR-1suppresses rpm-1(lf)

We mapped the rpm-1 suppressor mutation ju1118 to themiddle of chromosome I, following whole-genome sequenc-ing analysis (Figure 2C and Materials and Methods). Wefound a single nucleotide change, G875A, resulting inTrp292 to a stop codon mutation in the predicted geneY71-F9AR.3, hereby named supr-1 (suppressor of rpm-1) (Figure4A and Figure S5). supr-1(ju1118) was a weak suppressorby the criterion of locomotion behavior (data not shown).Quantification of the suppression of rpm-1(lf) on the PLMaxon defects also showed that supr-1(ju1118) exhibitedweaker suppression activity than those due to complete lossof function in the dlk-1 cascade (Figure 4B). We verified thesupr-1 gene structure by RT-PCR and cDNA analysis. We

Table 1 Severe reduction of synapse number and synaptic vesicles in rpm-1; syd-1 or rpm-1; syd-2 mutants

Genotype Nerve cord Type

No. of synapses/10 mm

Average length of synapses, mmworm1 worm2

juIs1 VC ACh 8.7 4.4 0.58 6 0.03 (n = 49)GABA 4.2 1.8 0.86 6 0.05 (n = 23)

DC ACh 5.1 4.9 0.66 6 0.05 (n = 34)GABA 2.4 1.8 1.01 6 0.07 (n = 15)

juIs1; rpm-1( ju41); syd-2( ju37) VC ACh 1.8 3.6 0.47 6 0.06 (n = 6)GABA 0.9 0.9 0.50 6 0.09 (n = 2)

DC ACh 0.9 0.9 0.47 6 0.11 (n = 2)GABA 0.0 0.9 0.36 (n = 1)

syd-1( ju82); juIs1; rpm-1( ju44) VC ACh 0.0 4.4 0.68 6 0.10 (n = 5)GABA 0.0 0.9 0.54 (n = 1)

DC ACh 0.0 6.2 0.43 6 0.06 (n = 7)GABA 0.9 1.8 0.56 6 0.11 (n = 3)

Young adult worms were chemically fixed, processed, and serially sectioned. The ventral and dorsal cords of two worms per genotype were examined by electronmicroscope. Numbers of sections analyzed were �1000 for juIs1 worm 1 (45 mm), �500 for juIs1 worm 2 (22.5 mm), and �250 for the others (11.25 mm). The synapses weredefined as a varicosity containing synaptic vesicles and one or a few active zones (White et al. 1986). Synaptic vesicles were counted in the sections containing an active zoneand averaged over sections. The length of a synapse was calculated by multiplying the number of sections containing synaptic vesicles by 50 nm. n = no. of synapses.

1106 K. Noma, A. Goncharov, and Y. Jin

identified two mRNA transcripts that differed in alternativeexon 4, corresponding to two protein isoforms of 746 aminoacids (long), and of 688 amino acids (short) (Figure 4A andFigure S5). To verify that supr-1 is the causative gene forsuppressing rpm-1(lf) phenotypes, we expressed the SUPR-1long isoform cDNA under the mechanosensory neuron-specificpromoter (Pmec-4). This transgene fully rescued the suppres-sion phenotypes of supr-1(ju1118) (Figure 4C), confirmingthat supr-1(ju1118) is a loss of function mutation, and further,indicating that SUPR-1 functions cell autonomously.

The predicted SUPR-1 protein lacks known functionaldomains and has closely related proteins in other Caenorhab-ditis species (Figure S5). To gain insights into SUPR-1 func-tion, we examined its localization in mechanosensoryneurons using GFP-tagged SUPR-1 (long). This GFP-taggedconstruct fully rescued the supr-1(ju1118) effects on rpm-1(lf) (Figure 4C). GFP::SUPR-1 was primarily localized inthe nucleus and weakly detectable in the somatic cytoplasmand axon (Figure 4D). We also analyzed two additional muta-tions of supr-1, tm3467, and gk106345. tm3467 is a 327-bp

Figure 2 Most of the rpm-1(lf) suppressors are mutations in the core components of the DLK-1 cascade. (A) Schematic diagram of the knowncomponents of the DLK-1 pathway. E3 ubiquitin ligase RPM-1 down-regulates DLK-1 MAPKK kinase by degradation. (B) The total number of allelesfor each gene in the DLK-1 cascade independently isolated in the suppressor screen was plotted against the length of the coding sequence. The methodof “least squares” was used to obtain the linear approximation. The line represents y = 22.1x 2 16.6. R2 = 0.81112. The sizes of mlk-1, mek-1, andkgb-1 are shown along the x-axis; no mutations were identified for these genes in our screen. (C) Genetic positions of the rpm-1, syd-1, syd-2, and theirsuppressor genes on the C. elegans chromosomes. Unit represents centimorgans (cM). (D) The positions of nonsense (red) and missense (blue) mutationsisolated in the screen are plotted along the protein sequences. Functionally defined domains are shown as light green boxes. Positions of select allelesare indicated under the protein sequences (see Results).

ESS-2 Regulates mRNA Splicing 1107

deletion removing the splice acceptor site of intron 7 anda part of exon 8. gk106345 is a splice donor mutation inthe exon 10/intron 10 junction (Figure 4A and Figure S5).mRNA may be produced from each allele using cryptic splicesites, but the encoded proteins would likely have an alteredC-terminal region from Gly505 and Asn590 for tm3467 andgk106345, respectively. We found that tm3467 did not sup-press the PLM morphological defects of rpm-1(lf), whilegk106345 showed partial suppression, suggesting theSUPR-1 C terminus may likely be dispensible for its function(Figure 4B). Taken together, we identified a novel proteinSUPR-1 that functions to antagonize rpm-1.

Loss of function in ess-2 synergizes with a spliceacceptor mutation in dlk-1 to suppress rpm-1(lf)

In the course of analyzing a suppressor strain isolated fromscreening in the rpm-1(ju44); syd-1(ju82) background, weidentified the ju600 allele as containing a single nucleotide

change at the 39 splice acceptor site (CAG to CAA) in the thirdintron of dlk-1. This change is predicted to result in the re-tention of intron 3 or skipping of exon 4, which would poten-tially lead to a premature stop codon before the kinasedomain. However, we found that dlk-1(ju600) alone did notshow detectable suppression on rpm-1(lf) (Figure 5D). Byinspecting animals from outcrossing of the initial suppressorisolate, we detected a background-dependent suppression ac-tivity of dlk-1(ju600) that showed linkage to chromosome III.Through further genetic mapping using unique SNPs identi-fied by whole-genome sequencing of the original suppressorstrain, we identified a nonsense mutation in the ess-2 (ES2similar) gene (Dimitriadi et al. 2010) as required for dlk-1(ju600) to suppress rpm-1(lf) (Figure 5D). ESS-2 consistsof 522 amino acids, and ess-2(ju1117) alters leucine 428 toa stop codon (Figure 5, A and B). The ess-2(ju1117) singlemutation did not suppress rpm-1(lf), but the dlk-1(ju600); ess-2(ju1117) double mutant suppressed rpm-1(lf), comparable

Figure 3 Loss of function in the core components of the MLK-1 cascade has modest effects on rpm-1(lf) developmental phenotypes. (A) The averagevelocities of the forward movement of the worms were analyzed as in Figure 1B. n = 3 sessions, 5–10 worms in each session. Error bar indicates SEM.Statistics: one-way ANOVA; ns, not significant (P . 0.05). (B) Schematic diagram of Anterior Lateral Microtubule (ALM) and PLM mechanosensoryneurons shown in light green. In rpm-1(lf) mutants, ALM and PLM overshoot and hook back, and PLM shows no branch in �50% of the animals(Schaefer et al. 2000). The micrographs of boxed region are shown in C. (C) The ALM and PLM were visualized bymuIs32[Pmec-7-GFP] for the indicatedgenotypes. Arrows and arrowhead represent the axonal branch and tip of PLM neurons, respectively. Bar, 20 mm. (D) Fraction of worms showing PLMdefects. Open bars and solid bars indicate branching and hook defects, respectively. N = no. of animals. dpy-11(e224) was used as a control for dpy-11(e224) mlk-1(km19). Statistics: Fisher’s exact test, compared with rpm-1(ju44) to show suppression. ***P , 0.001, **P , 0.01, *P , 0.05.

1108 K. Noma, A. Goncharov, and Y. Jin

to the suppression by a dlk-1 null allele (Figure 5D). Wefurther analyzed two deletion alleles of ess-2, tm4246, andok3569 (Figure 5A and Figure S6). tm4246 deletes a spliceacceptor site of intron 1 through part of intron 4, and likelycauses an early stop codon. ok3569 removes parts of exon 3and exon 4. We performed RT-PCR on ok3569 animals andfound that the remaining transcripts would result in an in-frame protein that lacks amino acids from Glu166 to Asp314including a conserved coiled-coil domain (Figure 5B and Fig-ure S6). Both ess-2 alleles showed no suppression of rpm-1(lf)on their own and acted as synthetic suppressors with dlk-1(ju600) (Figure 5D). Furthermore, we generated transgenesexpressing full-length ess-2 genomic DNA, including 916 bp ofthe ess-2 promoter region, and observed rescue of the syn-thetic suppression effects of dlk-1(ju600); ess-2(tm4246) (Fig-ure 5E). These data demonstrate that ess-2 is the causativegene for this synthetic suppression with dlk-1(ju600). A tran-

scriptional reporter driven by 916-bp upstream sequences ofess-2 showed broad tissue expression, including neurons,pharynx, body wall muscles, and seam cells (data not shown).We thus generated transgenes expressing the ess-2 cDNA un-der the control of mechanosensory neuron-specific promoterand found that this expression fully rescued the syntheticsuppression effects on PLM in a cell-autonomous manner(Figure 5E). We infer that the dlk-1(ju600) splice site muta-tion is a cryptic dlk-1 allele. Loss of function in ess-2 enhancesju600 to a complete loss of dlk-1 function, resulting in theobserved synthetic suppression of rpm-1(lf).

ESS-2 is required for mRNA splicing of mutant dlk-1(ju600) transcripts

C. elegans ESS-2 is named after the human protein ES2, alsoknown as DiGeorge syndrome critical region 14 (DGCR14)or DiGeorge syndrome protein I (DGSI) (Lindsay et al. 1996;

Figure 4 supr-1 is a novel gene antagonizing rpm-1. (A) Schematic diagram of the genomic locus of supr-1 (Y71F9AR.3 in WormBase). There are twoisoforms produced from the locus as described inMaterial and Methods, but we focused on the long isoform in this study. (B and C) The fraction of wormsshowing defects of PLM axonal morphologies, as described in Figure 3, A and B. N = no. of animals. Statistics: Fisher’s exact test, compared with rpm-1(ju44)to show suppression in B and with supr-1(ju1118); rpm-1(ju44) to show rescue activity in C. ***P, 0.001; ns, not significant (P. 0.05). (D) Localizations ofSUPR-1 in the mechanosensory neurons were visualized by a transgene juEx5517[Pmec-4-GFP::SUPR-1]. Arrows indicate nucleus. Bar, 20 mm.

ESS-2 Regulates mRNA Splicing 1109

Figure 5 dlk-1(ju600) splice acceptor mutation suppresses rpm-1(lf) together with ess-2(lf). (A) Schematic diagram of the genomic locus of ess-2 andthe DNA constructs for transgenes. The ess-2 locus produces two isoforms, a and b (F42H10.7a and b in WormBase) that differ in six extra nucleotides atthe beginning of exon 5 (asterisk). (B) Schematic diagrams of the domain structures of ESS-2 and its orthologs. Light green boxes indicate putativecoiled-coil domains. (C) Dendrogram of the ESS-2 orthologs. The dendrogram was made using Clustal Omega. The names of ESS-2 homologs andaccession nos. in UniProt are S. pombe: Bip1, O59793; C. elegans: ESS-2, P34420; C. briggsae: ESS-2, A8XT26; C. remanei: ESS-2, E3MI09; Drosophilamelanogaster: ES2/DGCR14, O44424; Xenopus tropicalis, DGCR14, B1H164; Danio rerio: DGCR14, F1QAX0; Mus musculus: ES2/DGCR14/DGSI,O70279; and Homo sapiens: ES2/DGCR14/DGSI, Q96DF8. (D and E) The fraction of worms showing the defects of PLM axonal morphologies, asdescribed in Figure 3, A and B. N = no. of animals. Statistics: Fisher’s exact test, compared with rpm-1(ju44) to show suppression in D and with dlk-1(ju600); ess-2(tm4246); rpm-1(ju44) to show rescue activity in E. ***P, 0.001; **P, 0.01; ns, not significant (P. 0.05). (F) Localization of ESS-2 in thePLM mechanosensory neuron was visualized by a transgene juEx5859[Pmec-4-ESS-2::GFP]. Arrowhead and arrow indicate nucleus and punctum,respectively. Bar, 20 mm.

1110 K. Noma, A. Goncharov, and Y. Jin

Rizzu et al. 1996; Gong et al. 1997). This protein family isconserved from fission yeast Schizosaccharomyces pombe tohuman and contains two conserved coiled-coil regions in itsN-terminal based on the MARCOIL program (Delorenzi andSpeed 2002) (Figure 5B and Figure S6). Both yeast homologBis1 and mouse ES2 are localized in the nucleus (Lindsayet al. 1998; Taricani et al. 2002). A recent large-scale proteininteractome study identified human ES2/DGCR14 as a non-core component of the spliceosome C complex (Hegele et al.2012), suggesting that the ESS-2 protein family might func-tion in mRNA splicing.

To investigate how ESS-2 functions, we next examinedESS-2 protein localization. We tagged ESS-2 with GFP ateither the N or C terminus. Both constructs are fully func-tional based on transgenic rescue activity (Figure 5E). GFP-tagged ESS-2 was primarily localized to the nucleus, withoccasional detection of one or a few bright nuclear puncta(Figure 5F). Similar nuclear puncta have also been reportedfor the fission yeast homolog, Bis1 (Taricani et al. 2002).Expression of cDNA isolated from ess-2(ok3569) did not res-cue ess-2(tm4246). Overexpression of cDNA containing ess-2(ju1117) mutation partially rescued ess-2(tm4246) (Figure5E). This analysis suggests that the conserved coiled-coil do-main is critical for ESS-2 function, and that C-terminally trun-cated ESS-2 protein produced in ess-2(ju1117) retains partialactivity.

The observation that dlk-1(ju600) single mutants did noteliminate dlk-1 activity suggests that this mutant likely pro-duces sufficient functional DLK-1, despite the mutation alter-ing the conserved splice acceptor consensus sequence. Earlystudies have indicated that the CAG in the splice acceptor sitecan be dispensable for mRNA splicing in C. elegans (Aroianet al. 1993). We thus examined the splicing of intron 3 ofdlk-1 by RT-PCR. Analysis of semiquantitative RT-PCR prod-ucts showed that transcripts corresponding to the fully splicedtranscripts were produced from dlk-1(ju600) animals to a com-parable level of those from wild-type animals (Figure 6B).Sequence analysis also confirmed that mRNA splicing in dlk-1(ju600) used the same site as in dlk-1(wt) despite the nucle-otide change at the splice acceptor site. On the other hand,transcripts corresponding to the unspliced transcript were onlyobserved in dlk-1(ju600) (Figure 6B). In ess-2(tm4246) singlemutants, we detected only fully spliced dlk-1 transcripts, sim-ilar to wild-type animals (Figure 6B). However, in dlk-1(ju600); ess-2(tm4246) double mutants the amount of fullyspliced dlk-1 transcripts was significantly decreased, comparedto dlk-1(ju600) and wild type (Figure 6B). It is possible thatRT-PCR may not detect transcript intermediates, such as thosestalled in a Lariat structure (see Discussion). To more ac-curately assess the effects on mRNA splicing, we performedquantitative RT-PCR (qRT-PCR), using primers to specificallydetect the transcripts in which intron 3 appeared to be splicedout (Figure 6A, magenta). Consistent with the semiquantita-tive RT-PCR data, we observed no detectable changes in ei-ther dlk-1(ju600) or ess-2(ju1117) single mutant, compared towild type. However, spliced transcripts without intron 3 were

dramatically decreased in dlk-1(ju600); ess-2(ju1117) doublemutant (Figure 6C, magenta). As a control for total mRNAlevel of dlk-1, we designed primers to detect exon 5–6 andexon 8–9 junctions (Figure 6A, gray and blue). qRT-PCR anal-ysis showed that the amount of the dlk-1 transcripts was notaltered in dlk-1(ju600); ess-2(ju1117) (Figure 6C, gray andblue), indicating that loss of function in ess-2 specificallyaffects intron 3 splicing of dlk-1.

We next asked if ESS-2 has a similar role in splicing ofother genes. The dpy-10(e128) mutation alters the spliceacceptor from CAG to CAA in intron 1 of isoform “a” (Figure6D). We performed RT-PCR for this exon/intron junctionand detected both spliced and unspliced transcripts in dpy-10(e128), consistent with a previous report (Aroian et al.1993) (Figure 6E). In the dpy-10(e128); ess-2(tm4246) dou-ble mutant, however, both spliced and unspliced bands weresignificantly reduced, suggesting that ESS-2 affects splicingefficiency of dpy-10 transcripts when they carry the muta-tion in a splice acceptor site (Figure 6E). The effect of ess-2(lf) on the “a” isoform did not lead to further enhancementof dumpy body shape in ess-2(lf); dpy-10(e128) compared todpy-10(e128), possibly because of residual activities of the“b” isoform (Figure 6D). Taken together, these data supporta conclusion that ESS-2 acts to facilitate RNA splicing, par-ticularly when pre-mRNA is compromised or contains non-canonical sequences at the splice acceptor site.

Discussion

Large-scale analysis of genetic suppressors of rpm-1supports the functional specificity of the DLK-1 kinasecascade in synapse formation

Synapse formation is a terminal step in neuronal differenti-ation and involves coordinated action of multiple pathways,including recruitment and retention of synaptic vesicles,formation, and remodeling of synaptic cytoskeleton, assemblyand alignment of junctional structures at pre- and postsyn-aptic sites. Genetic studies in C. elegans over the past decadehave made important contributions to the delineation of keymolecules and pathways in each specific aspect of synapseformation (Ackley and Jin 2004; Yan et al. 2011). Here, ouranalysis of rpm-1; syd-1 and rpm-1; syd-2 reinforces the no-tion that synaptogenesis is regulated by multiple parallelpathways. While removing each gene individually is not se-verely detrimental to animal behavior, these double mutantsexhibit severely impaired movement, correlating to the dra-matic reduction of synapse number and poorly developedremaining synapses. It is interesting to note that the locomo-tion defect of these double mutants is less severe than that ofthe kinesin unc-104(lf), in which synaptic vesicles are nearlydepleted while the localization of synaptic active zone pro-teins appears to be unperturbed (Hall and Hedgecock 1991;Zhen and Jin 1999). The behavioral outcome of rpm-1(lf);syd-2(lf) might indicate that some residual neuronal activityremains in these animals due to compensation or use of dif-ferent means of synaptic transmission.

ESS-2 Regulates mRNA Splicing 1111

We selected genetic suppressor mutations based onbehavioral improvement of the uncoordinated locomotion ofrpm-1 double mutants with syd-1 or syd-2. We can draw thefollowing conclusions from the analyses of a large number ofgenetic suppressor mutations. First, among 94 rpm-1(lf) sup-pressor alleles, we found nine alleles that were independent

hits of the same nucleotide. Although the small number ofgenes and high hit rate per gene preclude the classical appli-cation of the Poisson distribution (Pollock and Larkin 2004),we reckon that this screen is likely saturated for the majorcomponents of the DLK-1 pathway. The design of our suppres-sor screen biases against genes whose loss of function results

Figure 6 ESS-2 regulates mRNA splicing of cryptic transcripts. (A and D) Schematic diagrams of genomic loci of dlk-1 and dpy-10. The dlk-1 locusproduces at least two isoforms, long (F33E2.2a) and short (F33E2.2c). ju600 is a G-to-A substitution mutation located in the splice acceptor site of intron3 as indicated by the arrowhead. Primers used for qRT-PCR in C are shown as bars and dashed lines between long and short isoforms. (B and E)Amounts of PCR products were analyzed by semiquantitative RT-PCR using dlk-1 or dpy-1 primers (red arrows). The positions of splice acceptormutations in dlk-1(ju600) and dpy-10(e128) are shown by asterisks. The predicted sizes of spliced and unspliced fragments are indicated on the rightside. ama-1 encoding a large subunit of RNA polymerase II was used as a control. (C) Amounts of PCR products were analyzed by quantitative RT-PCR.The colors of bars match the amplified fragments indicated in A. All the strains in this experiment have muIs32[Pmec-7-GFP] and rpm-1(ju44) in thebackground. Error bar represents SEM; n = 2 biological replicates.

1112 K. Noma, A. Goncharov, and Y. Jin

in lethality or severe movement defects. Yet, we might expectto identify partial loss-of-function mutations in known essen-tial genes as some suppressor mutations in the dlk-1 cascadeshow rather weak activity. The absence of such mutationsleads us to speculate that other genes may only have minorroles or act in parallel to modulate the six core components ofthe dlk-1 cascade. Second, null mutants in the MLK-1/MEK-1/KGB-1 cascade are also homozygous viable; yet, our analysisindicates that this pathway does not have a major role in thecontext of synapse formation. The interaction between RPM-1and this kinase pathway may be specific to other cellularcontexts, such as axon regeneration (Nix et al. 2011). Third,most missense mutations in DLK-1/MKK-4/PMK-3 are in thekinase domains, highlighting the importance of the phosphor-relay in their action. The collection of functionally importantmutations would provide additional information about howthese proteins work when the three-dimensional structures ofthese proteins are available. Fourth, newly identified muta-tions in this study provide insights into how the componentsof the DLK-1 pathway function. For instance, the nonsensemutations in the C terminus of PMK-3 reside outside of thecanonical kinase domain, implying its possible regulatory role.Furthermore, the missense mutation of CEBP-1 suggests animportant role of the N terminus in its activity. As shown forother CEBP-1 orthologs, the N terminus of CEBP-1 may act asa transactivation domain (Pei and Shih 1991; Williams et al.1995). Fifth, besides the core components of the DLK-1 kinasecascade, we identified a novel gene supr-1 that antagonizesrpm-1. The lack of known motifs precludes us from drawingstrong conclusions regarding the function of supr-1. Nonethe-less, the observation that it is predominantly localized to thenucleus of neurons suggests that it may act through transcrip-tional or post-transcriptional modulation. Future studies, suchas identifying its binding partners and/or targets will addresssuch possibilities. Lastly, although our suppressor screen hasan equal chance to identify mutations suppressing the behav-ioral defects of syd-1 or syd-2, we found only a sup-5mutationas a suppressor of syd-2(Q397X) (this study) and a gain-of-function mutation in syd-2 as a suppressor mutation of syd-1(Dai et al. 2006). This outcome contrasts to the large numberof suppressors of rpm-1, and clearly reflects the nature ofsignaling contexts of these two pathways. RPM-1 negativelyregulates a signal transduction cascade that largely acts ina linear manner, while SYD-1 and SYD-2 function in a con-certed manner to promote the proper assembly or organizationof an ensemble of proteins in presynaptic dense projections.

The DGCR14 protein ESS-2 is a splicing factorthat ensures accurate splicing in acontext-dependent manner

Our analysis of ESS-2 has provided the first functional evi-dence supporting a role of the conserved family of DGCR14 inmRNA splicing. The splicing machinery and process are highlyconserved among metazoans (Rothman and Singson 2011).In C. elegans the 59 splice site has the canonical metazoanconsensus sequence AG/GURAGU (exon/intron), and most

39 splice sites have a conserved sequence UUUUCAG/R(intron/exon). Despite the sequence conservation, a few stud-ies show that mRNA splicing in C. elegans does not havestringent requirement for the CAG at the 39 splice acceptorsite, especially when the intron is short (Aroian et al. 1993;Zhang and Blumenthal 1996). A study demonstrated that anintron of 48 bp, which is a typical length in C. elegans, can bespliced out in the normal position despite mutations in thesplice acceptor site, while longer introns (171 and 283 bp) aremore dependent on canonical sequences at splice sites (Zhangand Blumenthal 1996). However, the mechanism of how ac-curacy of mRNA splicing is maintained for compromised ornoncanonical splice sites still remains poorly understood. Ouranalysis of the splice acceptor mutation dlk-1(ju600) hasadded to our understanding of the mechanisms that ensurethe fidelity of splice site selection. Intron 3 of dlk-1 and intron1 of dpy-10 are 46 bp and 48 bp, respectively, and containCAG at the 39 splice site, which are mutated to CAA in dlk-1(ju600) and dpy-10(e128). Our RT-PCR and qRT-PCR dataindicate that these mutated introns can be properly spliced inthe mutant animals, consistent with predictions from previousstudies (Zhang and Blumenthal 1996). Although loss of func-tion in ess-2 does not alter this splicing step in wild type, it ledto strongly impaired splicing in dlk-1(ju600) and dpy-10(e128). Moreover, the dependency of dlk-1(ju600) on ess-2(lf) to suppress rpm-1(lf) provides functional evidence forthe altered splicing. Together, these data support the role ofESS-2 in ensuring accurate mRNA splicing when the splicesite is compromised and relies on noncanonical sequences.

mRNA splicing generally consists of two catalytic events,lariat intermediate formation in an intron and ligation of thetwo exons. In each step, splicing factors and snRNA compo-nents are dynamically exchanged and rearranged (Mortonand Blumenthal 2011; Hegele et al. 2012). Among themthe ribonucleoprotein complex that catalyzes the ligation stepis called the C complex in which snRNAs and most proteinsare conserved in C. elegans (Morton and Blumenthal 2011)(data not shown). The mammalian ES2/DGCR14 was foundto coprecipitate with the C complex (Hegele et al. 2012),suggesting that ESS-2 may be functioning in the C complexin C. elegans. We found that both fully spliced and unsplicedtranscripts appeared to be reduced in dlk-1(ju600); ess-2(ju1117) mutants, although the amount of transcripts wasnot changed. ess-2(lf) might cause intron 3 of dlk-1(ju600)transcripts to be stalled as a lariat structure, which might notbe detected by RT-PCR and qRT-PCR. Further studies mayaddress the precise effects of ESS-2 and its associated proteinsin mRNA splicing. We observed similar effects on a dpy-10splice acceptor mutant, suggesting that ESS-2 function insplicing is widespread. We did not find any obvious pheno-type nor splice deficiency in ess-2 single mutant. This suggeststhat ESS-2 may play a redundant role with another compo-nent in C complex, or an accessory role in ensuring accuratesplicing. As ESS-2 is a conserved protein from yeast to human,this mechanism ensuring accurate splicing may be conservedin higher organisms.

ESS-2 Regulates mRNA Splicing 1113

The fission yeast ortholog of ES2, Bis1, was found as aninteracting protein to a stress-response nuclear envelopeprotein Ish1 (induced in stationary phase 1) and plays a rolein viability in the stationary phase (Taricani et al. 2002).Human ES2/DGCR14 is highly expressed in heart, brain,and skeletal muscle and is located in the common chromo-some region on 22q11.2 deleted in patients with DiGeorgesyndrome and velocardiofacial syndrome (Lindsay et al.1996; Rizzu et al. 1996; Gong et al. 1997). Although thesesyndromes have been attributed to disruption of the T-boxtranscription factor, Tbx1, based on knockout studies in mice(Jerome and Papaioannou 2001; Lindsay et al. 2001;Merscher et al. 2001), it remains possible that deletion ofES2/DGCR14 may contribute to pathological progress inhuman patients through impaired splicing. Indeed, an anal-ysis of point mutations associated with multiple human ge-netic diseases has estimated that up to 15% of all thesemutations may result in mRNA splicing defects (Krawczaket al. 1992). Moreover, a recent study has suggested thatpromoter polymorphisms in the ES2/DGCR14 gene are as-sociated with schizophrenia (Wang et al. 2006). Thus, ex-amining in vivo targets of ESS-2 in the future will aid inbetter understanding the physiological roles of the ESS-2family of proteins in higher organisms.

Acknowledgments

We gratefully acknowledge Mei Zhen and Xun Huang for theircontribution in the early studies of synergistic interactionsbetween rpm-1 and syd-1 or syd-2 genes. We thank GlorianaTrujillo, Dong Yan, and Neset Ozel for noting the complexgenetic traits of dlk-1(ju600); past and present members ofour laboratory for discussion and advice; and Andrew Chis-holm for critical comments on the manuscript. We thankShohei Mitani and the National Bioresearch Project (TokyoWomen’s Medical University School of Medicine, Tokyo,Japan) and C. elegans knockout consortium for providingdeletion alleles, Navin Pokala and Cornelia Bargmann forproviding the code for the multiworm tracker prior to pub-lication, and Seika Takayanagi-Kiya for her help in trackinganalysis. Some of the strains were obtained from the Caeno-rhabditis Genetics Center, which is supported by grants fromNational Institutes of Health (NIH)–National Center for Re-search Resources. This work was supported by a grant fromNIH to Y.J. (National Institute of Neurological Disorders andStroke R01-035546). K.N. and A.G. are research associatesand Y.J. is an investigator of the Howard Hughes MedicalInstitute.

Literature Cited

Ackley, B. D., and Y. Jin, 2004 Genetic analysis of synaptic targetrecognition and assembly. Trends Neurosci. 27: 540–547.

Aroian, R. V., A. D. Levy, M. Koga, Y. Ohshima, J. M. Kramer et al.,1993 Splicing in Caenorhabditis elegans does not require anAG at the 39 splice acceptor site. Mol. Cell. Biol. 13: 626–637.

Bigelow, H., M. Doitsidou, S. Sarin, and O. Hobert, 2009 MAQGene:software to facilitate C. elegans mutant genome sequence analy-sis. Nat. Methods 6: 549.

Bounoutas, A., J. Kratz, L. Emtage, C. Ma, K. C. Nguyen et al.,2011 Microtubule depolymerization in Caenorhabditis eleganstouch receptor neurons reduces gene expression through a p38MAPK pathway. Proc. Natl. Acad. Sci. USA 108: 3982–3987.

Brenner, S., 1974 The genetics of Caenorhabditis elegans. Genetics77: 71–94.

Ch’ng, Q., L. Williams, Y. S. Lie, M. Sym, J. Whangbo et al.,2003 Identification of genes that regulate a left-right asym-metric neuronal migration in Caenorhabditis elegans. Genetics164: 1355–1367.

Dai, Y., H. Taru, S. L. Deken, B. Grill, B. Ackley et al., 2006 SYD-2Liprin-[alpha] organizes presynaptic active zone formationthrough ELKS. Nat. Neurosci. 9: 1479–1487.

Delorenzi, M., and T. Speed, 2002 An HMM model for coiled-coildomains and a comparison with PSSM-based predictions. Bio-informatics 18: 617–625.

Dimitriadi, M., J. N. Sleigh, A. Walker, H. C. Chang, A. Sen et al.,2010 Conserved genes act as modifiers of invertebrate SMNloss of function defects. PLoS Genet 6: e1001172.

Gong, W., B. S. Emanuel, N. Galili, D. H. Kim, B. Roe et al.,1997 Structural and mutational analysis of a conserved gene(DGSI) from the minimal DiGeorge syndrome critical region.Hum. Mol. Genet. 6: 267–276.

Hall, D. H., and E. M. Hedgecock, 1991 Kinesin-related gene unc-104 is required for axonal transport of synaptic vesicles inC. elegans. Cell 65: 837–847.

Hallam, S. J., and Y. Jin, 1998 lin-14 regulates the timing of syn-aptic remodelling in Caenorhabditis elegans. Nature 395: 78–82.

Hallam, S. J., A. Goncharov, J. McEwen, R. Baran, and Y. Jin,2002 SYD-1, a presynaptic protein with PDZ, C2 and rho-GAP-like domains, specifies axon identity in C. elegans. Nat.Neurosci. 5: 1137–1146.

Hegele, A., A. Kamburov, A. Grossmann, C. Sourlis, S. Wowro et al.,2012 Dynamic protein-protein interaction wiring of the hu-man spliceosome. Mol. Cell 45: 567–580.

Jerome, L. A., and V. E. Papaioannou, 2001 DiGeorge syndromephenotype in mice mutant for the T-box gene, Tbx1. Nat. Genet.27: 286–291.

Jin, Y., 2005 Synaptogenesis (December 23, 2005), WormBook,ed. The C. elegans Research Community, WormBook, doi/10.1895/wormbook.1.44.1, http://www.wormbook.org.

Jin, Y., and C. C. Garner, 2008 Molecular mechanisms of presyn-aptic differentiation. Annu. Rev. Cell Dev. Biol. 24: 237–262.

Krawczak, M., J. Reiss, and D. N. Cooper, 1992 The mutationalspectrum of single base-pair substitutions in mRNA splice junc-tions of human genes: causes and consequences. Hum. Genet.90: 41–54.

Liao, E. H., W. Hung, B. Abrams, and M. Zhen, 2004 An SCF-likeubiquitin ligase complex that controls presynaptic differentia-tion. Nature 430: 345–350.

Lindsay, E. A., P. Rizzu, R. Antonacci, V. Jurecic, J. Delmas-Mataet al., 1996 A transcription map in the CATCH22 critical re-gion: identification, mapping, and ordering of four novel tran-scripts expressed in heart. Genomics 32: 104–112.

Lindsay, E. A., E. L. Harvey, P. J. Scambler, and A. Baldini, 1998 ES2,a gene deleted in DiGeorge syndrome, encodes a nuclear proteinand is expressed during early mouse development, where it sharesan expression domain with a Goosecoid-like gene. Hum. Mol.Genet. 7: 629–635.

Lindsay, E. A., F. Vitelli, H. Su, M. Morishima, T. Huynh et al.,2001 Tbx1 haploinsufficieny in the DiGeorge syndrome regioncauses aortic arch defects in mice. Nature 410: 97–101.

Maeder, C. I., and K. Shen, 2011 Genetic dissection of synapticspecificity. Curr. Opin. Neurobiol. 21: 93–99.

1114 K. Noma, A. Goncharov, and Y. Jin

Mello, C. C., J. M. Kramer, D. Stinchcomb, and V. Ambros,1991 Efficient gene transfer in C. elegans: extrachromosomalmaintenance and integration of transforming sequences. EMBOJ. 10: 3959–3970.

Merscher, S., B. Funke, J. A. Epstein, J. Heyer, A. Puech et al.,2001 TBX1 is responsible for cardiovascular defects invelo-cardio-facial/DiGeorge syndrome. Cell 104: 619–629.

Morton, J. J., and T. Blumenthal, 2011 RNA processing inC. elegans. Methods Cell Biol. 106: 187–217.

Nakata, K., B. Abrams, B. Grill, A. Goncharov, X. Huang et al.,2005 Regulation of a DLK-1 and p38 MAP kinase pathwayby the ubiquitin ligase RPM-1 is required for presynaptic devel-opment. Cell 120: 407–420.

Nix, P., N. Hisamoto, K. Matsumoto, and M. Bastiani, 2011 Axonregeneration requires coordinate activation of p38 and JNKMAPK pathways. Proc. Natl. Acad. Sci. USA 108: 10738–10743.

Ou, C. Y., and K. Shen, 2010 Setting up presynaptic structures atspecific positions. Curr. Opin. Neurobiol. 20: 489–493.

Patel, M. R., and K. Shen, 2009 RSY-1 is a local inhibitor of pre-synaptic assembly in C. elegans. Science 323: 1500–1503.

Pei, D. Q., and C. H. Shih, 1991 An “attenuator domain” is sand-wiched by two distinct transactivation domains in the transcrip-tion factor C/EBP. Mol. Cell. Biol. 11: 1480–1487.

Pokala, N., Q. Liu, A. Gordus, and C. I. Bargmann, 2014 Inducibleand titratable silencing of Caenorhabditis elegans neuronsin vivo with histamine-gated chloride channels. Proc. Natl.Acad. Sci. USA 111: 2770–2775.

Pollock, D. D., and J. C. Larkin, 2004 Estimating the degree ofsaturation in mutant screens. Genetics 168: 489–502.

Rizzu, P., E. A. Lindsay, C. Taylor, H. O’Donnell, A. Levy et al.,1996 Cloning and comparative mapping of a gene from thecommonly deleted region of DiGeorge and Velocardiofacial syn-dromes conserved in C. elegans. Mamm. Genome 7: 639–643.

Rothman, J. H., and A. Singson, 2011 Caenorhabditis elegans:molecular genetics and development. Methods Cell Biol. 106:xv–xviii.

Sakaguchi, A., K. Matsumoto, and N. Hisamoto, 2004 Roles ofMAP kinase cascades in Caenorhabditis elegans. J. Biochem.136: 7–11.

Schaefer, A. M., G. D. Hadwiger, and M. L. Nonet, 2000 rpm-1,a conserved neuronal gene that regulates targeting and synapto-genesis in C. elegans. Neuron 26: 345–356.

Sudhof, T. C., 2012 The presynaptic active zone. Neuron 75: 11–25.

Taricani, L., M. L. Tejada, and P. G. Young, 2002 The fission yeastES2 homologue, Bis1, interacts with the Ish1 stress-responsivenuclear envelope protein. J. Biol. Chem. 277: 10562–10572.

Trujillo, G., K. Nakata, D. Yan, I. N. Maruyama, and Y. Jin, 2010 Aubiquitin E2 variant protein acts in axon termination and syn-aptogenesis in Caenorhabditis elegans. Genetics 186: 135–145.

Wang, H., S. Duan, J. Du, X. Li, Y. Xu et al., 2006 Transmissiondisequilibrium test provides evidence of association betweenpromoter polymorphisms in 22q11 gene DGCR14 and schizo-phrenia. J. Neural Transm. 113: 1551–1561.

Waterston, R. H., and S. Brenner, 1978 A suppressor mutation inthe nematode acting on specific alleles of many genes. Nature275: 715–719.

White, J. G., E. Southgate, J. N. Thomson, and S. Brenner, 1986 Thestructure of the nervous system of the nematode Caenorhabditiselegans. Philos. Trans. R. Soc. Lond. B Biol. Sci. 314: 1–340.

Williams, S. C., M. Baer, A. J. Dillner, and P. F. Johnson,1995 CRP2 (C/EBP beta) contains a bipartite regulatory do-main that controls transcriptional activation, DNA binding andcell specificity. EMBO J. 14: 3170–3183.

Wills, N., R. F. Gesteland, J. Karn, L. Barnett, S. Bolten et al., 1983 Thegenes sup-7 X and sup-5 III of C. elegans suppress amber nonsensemutations via altered transfer RNA. Cell 33: 575–583.

Yan, D., and Y. Jin, 2012 Regulation of DLK-1 kinase activity bycalcium-mediated dissociation from an inhibitory isoform. Neu-ron 76: 534–548.

Yan, D., Z. Wu, A. D. Chisholm, and Y. Jin, 2009 The DLK-1kinase promotes mRNA stability and local translation in C. ele-gans synapses and axon regeneration. Cell 138: 1005–1018.

Yan, D., K. Noma, and Y. Jin, 2011 Expanding views of presyn-aptic terminals: new findings from Caenorhabditis elegans. Curr.Opin. Neurobiol. 22: 431–437.

Zhang, H., and T. Blumenthal, 1996 Functional analysis of anintron 39 splice site in Caenorhabditis elegans. RNA 2: 380–388.

Zhen, M., and Y. Jin, 1999 The liprin protein SYD-2 regulates thedifferentiation of presynaptic termini in C. elegans. Nature 401:371–375.

Zhen, M., X. Huang, B. Bamber, and Y. Jin, 2000 Regulation ofpresynaptic terminal organization by C. elegans RPM-1, a puta-tive guanine nucleotide exchanger with a RING-H2 fingerdomain. Neuron 26: 331–343.

Communicating editor: M. Sundaram

ESS-2 Regulates mRNA Splicing 1115

GENETICSSupporting Information

http://www.genetics.org/lookup/suppl/doi:10.1534/genetics.114.167841/-/DC1

Systematic Analyses of rpm-1 Suppressors RevealRoles for ESS-2 in mRNA Splicing in

Caenorhabditis elegansKentaro Noma, Alexandr Goncharov, and Yishi Jin

Copyright © 2014 by the Genetics Society of AmericaDOI: 10.1534/genetics.114.167841

Sma Unc

suppression

(GFP)

Sma Unc

(GFP)

Sma Unc

(GFP)

suppression

(GFP)

Sma Unc

suppression

A B

C D

m(=dlk-1)I muIs32 II syd-2(lf)Xdlk-1(0)I + +

rpm-1(lf)V

m(=dlk-1) or dlk-1(0)I muIs32 or + IIm(=dlk-1) or dlk-1(0)I muIs32 or + II

rpm-1(lf)V syd-2(lf)X

muIs32 II pmk-3(0)IV rpm-1(lf)V x m(=dlk-1)I rpm-1(lf)V syd-2(lf)X

m(=dlk-1)I muIs32 II pmk-3(0)IV syd-2(lf)X+ + + +

rpm-1(lf)V

m(=dlk-1) or + I muIs32 or + II pmk-3(0) or + IVm(=dlk-1) or + I muIs32 or + II pmk-3(0) or + IV

rpm-1(lf)V syd-2(lf)X

dlk-1(0)I muIs32 II rpm-1(lf)V x rpm-1(lf)V syd-2(lf) m(=mkk-4)X

dlk-1(0)I muIs32 II syd-2(lf) m(=mkk-4)X+ + +

rpm-1(lf)V

dlk-1(0) or + I muIs32 or + IIdlk-1(0) or + I muIs32 or + II

rpm-1(lf)V syd-2(lf) m(=mkk-4)X

muIs32 II pmk-3(0)IV rpm-1(lf)V x rpm-1(lf)V syd-2(lf) m(=mkk-4)X

muIs32 pmk-3(0)IV syd-2(lf) m(=mkk-4)X+ + +

rpm-1(0)V

muIs32 or + II pmk-3(0) or + IVmuIs32 or + II pmk-3(0) or + IV

rpm-1(lf)V syd-2(lf) m(=mkk-4)X

P0

F1

F2

P0

F1

F2

dlk-1(0)I; muIs32 II; rpm-1(lf)V x m(=dlk-1)I rpm-1(lf)V syd-2(lf)X

Figure S1 Complementation tests to dlk-1 or pmk-3. Complementation tests were carried out to determine the identity of new suppressor mutations of rpm-1 with respect to dlk-1, pmk-3 and mkk-4. rpm-1(ju44); syd-2(ju37); suppressor mutant (shown as m) was crossed to dlk-1(tm4024); muIs32; rpm-1(ju44) (A and C) or muIs32; pmk-3(ju485); rpm-1(ju44) (B and D). muIs32-positive F1s are singled and self-fertilized. We examined whether any small and uncoordinated (Sma Unc) worms were present among the F2 progenies. Presence of Sma Unc worms due to rpm-1(lf); syd-2(lf) double-mutant background is an indication of complementation. mkk-4 mutant appears to complement neither dlk-1 nor pmk-3 in this assay because it is tightly linked to syd-2 (Figure 2C). (A) Complementation test with dlk-1(0) when a mutant is dlk-1(lf). They do not complement each other. (B) Complementation test with dlk-1(0) when a mutant is pmk-3(lf). They complement each other. (C and D) Complementation tests with dlk-1(0) and pmk-3(0) when a mutant is mkk-4(lf) in (C) and (D), respectively. Neither dlk-1 nor pmk-3 appears to complement mkk-4.

2 SI K. Noma, A. Goncharov, and Y. Jin

pmk-3

rpm-1; syd-2

dlk-1 mkk-4

BA

cebp-1

N2sy

d-2rp

m-1

dlk-1(t

m4024

) -

dlk-1(t

m4024

)

dlk-1(j

u590)

dlk-1(j

u1054

)

mkk-4(

ju1034

)

pmk-3(ju

693)

mak-2(

ju546)

cebp-1(

ju634)

cebp-1(

ju1019

)0

50

100

150

200

rpm-1; syd-2

Aver

age

spee

d (u

m/s

ec)

Figure S2 rpm-1(lf); syd-2(lf) body shape and movement phenotypes are suppressed by the mutations in the components of the DLK-1 pathway. (A) Bright-field images showing the body shapes of adult animals for the indicated genotypes. Scale bar: 500 µm. (B) The average velocities of the forward movement of the worms were analyzed as in Figure 1B. n=3 sessions, 5-10 worms in each session. Error bar indicates S.E.M.

3 SI K. Noma, A. Goncharov, and Y. Jin

ju512(P111S)I

ju1049(G113D)I

ju516, ju549(G130E)I

ju693(R153X)I

ju658(R164C)I

ju622(E168K)I

ju655(A205T)I

ju597(H228Y)I

ju1053(A239T)I

ju633(A246T)II

ju1045(I249X)II

ju626(R251C)I

ju584(P255S)

ju621,ju627(W301X)I

ju1052(W313X)I

ju548(P348L)I

Catalytic loopju1051(S314L)I

ju632(I414F)I

ju671(H416Y)I

Activation segment

ju485(E428K)I

ju592(W453X)I

ju1010(W448X)I

ATP

ju673(C345Y)I

Figure S3 Alignment of PMK-3 orthologs and locations of pmk-3(lf) alleles identified as rpm-1(lf) suppressors. The alignment was obtained by the ClustalW multiple-alignment program. MAPK14/p38-α is shown as the PMK-3 ortholog in fly and vertebrates. The positions of nonsense (red) and missense (blue) mutations isolated as rpm-1(lf) suppressors in our study are shown above the sequences. Identical and similar residues are highlighted in black and gray, respectively. Kinase domain, based on UniProt, was shaded in yellow. The conserved ATP binding site is shown as a black arrowhead above the sequences. The conserved catalytic loop and activation segment are underlined in green. Species are Ce: Caenorhabditis elegans; Dm: Drosophila melanogaster; Dr: Danio rerio; Mm: Mus musculus. Accession numbers in UniProt are Ce PMK-3: O44514; Dm MAPK14: O61443; Dr MAPK14: Q9DGE2; Mm MAPK14: P47811; Ce PMK-1: Q17446; Ce PMK-2: Q8MXI4; Ce KGB-1: O44408; Ce KGB-2: H2KZI0.

4 SI K. Noma, A. Goncharov, and Y. Jin

ATP

ju1034(W140X)I

ju686(D194N)I

Catalytic loop Activation segment

ju1040(E221V)I

ju1066(E238X)I

ju1028(P268S) I

ju91(G265R)

ju488(C176Y)I

ju473(G206E)I

ju474(Q270*)I

Activation segment

Figure S4 Alignment of MKK-4 homologs with mkk-4(lf) alleles suppressing rpm-1(lf). The alignment was obtained by the ClustalW multiple-alignment program. The direct ortholog of C. elegans MKK-4 in mammals is MKK4 but mice have three other paralogs, MKK7, MKK3 and MKK6. The positions of nonsense (red) and missense (blue) mutations isolated as rpm-1(lf) suppressors in our study are shown above the sequences. Identical and similar residues are highlighted in black and gray, respectively. Kinase domain, based on UniProt, was shaded in yellow. The conserved ATP binding site is shown as a black arrowhead above the sequences. The conserved catalytic loop and activation segment are underlined in green. Species are Ce: Caenorhabditis elegans; Dm: Drosophila melanogaster; Mm: Mus musculus. Accession numbers in UniProt are Ce MKK-4: Q20347; Dm MKK4: O61444; Mm MKK4: P47809; Mm MKK7: Q8CE90-2; Mm MKK6: P70236; Mm MKK3: O09110.

5 SI K. Noma, A. Goncharov, and Y. Jin

10 20 30 40 50 60. . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . |

C. elegans 1 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - M K L E D S - - C A QR N L V L N L L H R E I GT QYC. briggsae 1 V S R T D L GK L A L L D L E QQY I P I R R S S F I S T H I N T M D V T R P L P K F QK H I V S N L L D R E I GS E NC. remanei 1 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - M E F N E K K P K F QQN V V S N L L D R E I GT E Y

70 80 90 100 110 120. . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . |

C. elegans 26 R K - - - - - - - - - - - - - - - Y GS I R N T Y R H P Y QL I A A - V L P D L S A GH S M D R Y D P - - - T A S P S FC. briggsae 61 QK - - - - - - - - - - - - - - - Y K S V QN S F H H H S QL F A S S S QL T N L QL I D F N K C D P - - S I S D T Q FC. remanei 28 R K RM L L F Y L S E K T I E F S D A T N T N T Y R H P Y QL L T S - T K P D T L V Q I D L D K C D K N P D S D QP P F

130 140 150 160 170 180. . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . |

C. elegans 67 S T Y A K C L G F S P D GK Y M A F H S H A R S A I S V L R Y S GL E N - - F V K D N V D L R T F F K K I D T F QL K RC. briggsae 104 D S D A D C L G F S P D GK Y L A F H S H A H S A I S V L K Y T G I E NM K GA K D QS D P L N C F GG F T K F E L A PC. remanei 87 S L L A D C L G F S P D GK Y L A F H S H A R S V V S I F R Y T GA E GGK C A E G I S N P QN Y L R E A GR L D L S S

190 200 210 220 230 240. . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . |

C. elegans 125 F V R T L N C R L N R V L GS WWT D D S S I L I V QL S L N K I Y D E D S D QN E S D E E QL P QA Y D I V N H D A DC. briggsae 164 I T R E H R Y QN E K I I R S WWT D D S S T L I I QL T I E Q F R E E E QA H D E P D E E I L V V A N D I M T T E S YC. remanei 147 F T R D Y N Y E N E RM F GS WWT D D S K A L I V QL S M E S S I D D E D QV V P E D E D V L QL A R QL V E E N R N

250 260 270 280 290 300. . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . |

C. elegans 185 F F A N - - - - - - - - - - - - - - - E L P S T P S - - - - - - E R E A S P P S S S Q F S N S A A P N - - - - - - - - -C. briggsae 224 QN A A D V P D P V AW F D E I V GS D F Y R P GQ - - - - - - - E QY GR S T I S R R T QN F H D R FM F L N Q F T TC. remanei 207 T Y P E GDM V GN F V N E R R R R L E I L R A P A G E G F R D D D L Y E E V R Y N V Y D N T M R P N P E C Y D A D A E

310 320 330 340 350 360. . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . |

C. elegans 214 - - - - - - - - - - - - - - - - - L P E T K QP - P T T T D QY S E T S R T I P F Y L AM E S A A T I S L D G E R V A GC. briggsae 277 L N D V R E Y H H Q E S S GR Y Y E N E F T C E D P P K A I QH R E T S R T M P Y Y S A V K A A A QV A I D E QR F H RC. remanei 267 I D Y GT L E T L F K QH T N G I M P E N E L T N E I S A Y QH R E T S R T I P F T F AM E S A A RM R V D S F K K R G

370 380 390 400 410 420. . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . |

C. elegans 257 E T S R T M P V Y H S L T A S I I D G E R D GW E V D P S I P S T S NWC P - P R E E E E P R E K R QK I G E D S S F SC. briggsae 337 E C S R T M S V F H A V T S S A H H E S E A Y L D E D D T F A S T S K S V R R L R S P N S S Q E GGT K QK R A K T D SC. remanei 327 QT S R T M P V Y H A L T R S I R K S N Y D GY S S D S S D S S D S T D D L - H QD QL P A S Y A Y M Q E S D A T RM L

430 440 450 460 470 480. . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . |

C. elegans 316 D F D S L F D QL T E N L S H N K S Q E N R - - S P AM R S L I E N L R N N P R I S S R QGS P I T N A S GL GY S A PC. briggsae 397 F V S T A I E E AWN S M D D Y E KWN S E N Y P R S V R H V V H L L QK F P E Y A A D F K T S GP N I N C L GY S R PC. remanei 386 A A E E K L K K L QKM N C R C R I C S K R E V S Y GL N G E K K V I V C T D Y WK KWT R K Y GA N L C GL GY S K P

490 500 510 520 530 540. . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . |

C. elegans 374 GS I L R R S R - - - - - - - - - - - - - - - - A S A I P V L P N F H D D - - S A D E E E E E N S E R S D - - - - - - -C. briggsae 457 GS D T R R A S T V R - - - - - - - - - - - - - QK P F S D R F R EM R S F I A GL D QS D P V D K N E K - - - - - P EC. remanei 446 D A Y I R R K H GL R N E R P T V GA Y R D R L R K T L E D L D Y F H K S Y L T A S A R E E R L A GR E D R V QS L K V

550 560 570 580 590 600. . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . |

C. elegans 409 T V M G F L A F D V K T S T T T P L I L R N L N N N F K C HM R N R R L T V C C N E N L I E T Y QL T K S N GK N EW EC. briggsae 499 A T V V F V A F H V E T K KM I K L P F E K V F T P Y H C H GR D R L V V F S C N N D F I QV Y H L S D V D - - H EW IC. remanei 506 D T M L V L A F E I E E D T V C P L Y Y N N I S S K F L C H V R D R T M T I C C N N D T I E I Y H Y S E T T - - H I W E

610 620 630 640 650 660. . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . |

C. elegans 469 K I GQT R L P N L N S F P I D L H Y R C GT A V GM S S RM E E V E QG - - - - - - - - - - - - - - Q I F T D V F N FC. briggsae 557 S QK H F R L S QL D S E P L QL K Y R T GT M P GM P S L N R D V QT G - - - - - - - - - - - - - - S V I GD L F D IC. remanei 564 QQ E D V R H S QN E S Y L V P L D F R V GT M P GQP S R N E E P EM GY A GS I D S I L L K N F R E I F D D L F E S

670 680 690 700 710 720. . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . |

C. elegans 515 H R F H S P I QQD I L T I L H R D A P E H A S R - - - L F GK S P K L T QC HM I S D D L L A V T M T F H K R R S S PC. briggsae 603 R A F N S P I QQD V A T L L GS NM P E Y F N S F K P F L T E T P N I S S I R C L S D GL I A V S M S Y S - - R L A RC. remanei 624 D Y F Y S P V QQD T L T L L H R DM P E V F D T N D A L L K GR P R V F R C H A L S D D L I A V S L K L P - - W F K P

730 740 750 760 770 780. . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . |

C. elegans 572 L H I GA T I S I L D S S I T V Y P N Y H E Y L K L L F T H H L P D L L P R HWT Y L N P T P A H E QP W F K N S L I NC. briggsae 661 I QM GV I V S L V D S T V E V Y A D Y H K F L K L L F T H H L P D L L P R HWT H N N P Y P V T E QS W F K N E I I DC. remanei 682 M Y M GM I I S P D S C N T T L Y P D Y H E F L K L L F A N H L P E L L P R AWT Y F N P Y P A K D QP W F K K E I I N

790 800 810 820 830 840. . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . |

C. elegans 632 V S N P D E L N N F D R L F E V Y H I L I S N F N N V K S T S N P L T D A F F N A A E A E D FM NM GK S T N H Y P R DC. briggsae 721 V S N P K EM T N F R P L N D I Y Q F F D K I WH N V R S V S N P L QC S L F N K T H A Y QL L F N Y L F N D G - - - RC. remanei 742 V S D P R D L N D Y E L L R E T Y E F L S K HWN N L K S T S N P L N D S L F N S T H A Y Q F F S K F I F T N D - - - N

850 860 870 880 890 900. . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . |

C. elegans 692 T S AWN H F H E - - N E E F GA T L Y Y H P I D P V L F L K V GS FWL V A H R QA D V S D D F I V S K N - L E H - -C. briggsae 778 N I T F R E Y D D N L C R A I RM E QY Y H P S D P F K L I K N V Y QWY L A F QP K D P - D N V I V H S N R F K A - -C. remanei 799 R I T F N A P D L - - N K R K H L D V F C H P K D P F A V V N N T Y NW F F G I R K K D Q - D D V I V GV N P L E S D D

. . . . | . . .

C. elegans 746 - - - - - - - -C. briggsae 834 - - - - - - - -C. remanei 856 D D E KM D T N

ju1118(W292X)I

exon4

exon4

Figure S5 Alignment of SUPR-1 homologs. The alignment was obtained by the ClustalW multiple-alignment program. Identical and similar resides are highlighted in black and gray, respectively. The rpm-1(lf) suppressing allele ju1118 is shown in red above the sequences. Accession number for C. elegans SUPR-1 in UniProt is Q95XT1. Accession numbers for C. briggsae and C. remanei supr-1 in Wormbase are CBG04186 and CRE30057, respectively.

6 SI K. Noma, A. Goncharov, and Y. Jin

F S ju1117(L428X)I

ok3569

ok3569

ok3569

ok3569

10 20 30 40 50 60. . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . |

S.pombe 1 - - - - - - - - - - - - - - - - - - - - - M S L A K K I D D D E K QL V K P V N K E QT Y K K P I P L E E D D Y I E GLC.elegans 1 - - - - - - - - - M S S F D K N D E R A QL M V P K S I N E GK V V K L S QK T L V T K K I E R QV V P E E K Y I A GLC.briggsae 1 - - - - - - - - - M S S F D K I D E K A K L I V P K S T N E GK V V K V S QK T L A S K K H E R QV V P E E K Y I A GLD.melanogaster 1 M S A T T R T P A T P GT P GT P GS L AM E V A R V QN S A L A E F K K P T AM V R H K N K P K I L T E E K Y I E EMD.rerio 1 - - - - - - - - - - - - M A E S K E K A V C L R S D V T T V A V R N P V E E - - - - S K K P K R K I L D E E QY I E S LM .musculus 1 - - - - - - - - - - M GT P GT S A GA L F L S S A S A P S R K R A A G E A G E A GV A R S R QR V L D E E E Y I E GL

70 80 90 100 110 120. . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . |

S.pombe 40 S Y I I QQQY F P D L P K L K A E V V L E S E E V GS F D A QN E S R D E K L K Y L I A K N S E D - - - - - P L R K RC.elegans 52 D K I I E K D Y F P H L K KM QA QK E Y L E A V A N K D I N K I K E L QM K F C S T GS V R T D R S F R T P I T T R SC.briggsae 52 D K I I E K D Y F P QL K KM QA QK E Y L E A V A K K D I T K I K E L QM K Y C S T GS I R T D R T S I N P P T T R SD.melanogaster 61 S K I I QR D F F P D L E R L R A QN D Y L D A E S R R D F V QM A E I R E R Y S L GR I S GT GR - - - S T S R R N ND.rerio 45 E K I I QR D F F P D V S K L QA QK D Y L E A E E N GD L E RM R E I A I K Y GS AM A K Y T P R - - - T Y V P H V TM .musculus 51 QT V I QR D F F P D V E K L QA QK E Y L E A E E N GD L E RM R Q I A I K F GS A L GK I S R E - - - P P P P Y V T

130 140 150 160 170 180. . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . |

S.pombe 95 L P S L A I H E I T K A QL D - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - G E N K P I S V A SC.elegans 112 T T - E A P D V S S F D A D T P GP S S A S T S S A H DWM QS P M P F A N E E GD N E A L N R K R K K K K E E T L T SC.briggsae 112 T T S E A P Y D E Q F D S E T P GP S - - S N S GK H DWM QS P M P F A N E E GD N E A I N K K R R K K N Q E T L T SD.melanogaster 118 AM S P A T F E T P V S QA K C S N T P L P N S R A T D T P F S T D GS E K S D A E G - - - - - - R D T T A K L S L D AD.rerio 102 P S T F E T P D - - - GR S A S P S S S QS K S R A A R E - - - - - - - D GK D G E E - - - - - - A S E K D L P S L D RM .musculus 108 P A T F E T P E V H P GS A V L GN K P R P QGR D L D D - - - - - - - A G E A G E E - - - - - - E E K E P L P S L D V

190 200 210 220 230 240. . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . |

S.pombe 120 Y QN K F T S E D N A S F G E L M E D E S R L R A E QH K R R F GV H S QQP S N - - - - - - S I QT I GY S N S D A IC.elegans 171 Y L N K Y T S E D N A S F E E L A K V M R E R E D A R R P WV Y K A E E E H N K N L V T R QA I A A E A D V QL A L K HC.briggsae 170 Y L N K Y T S E D N A S F E E L A K V M R E R E D A R R P WV Y K A E E E H N K N L V V R QA I A A E A D V QL A L K HD.melanogaster 172 F L QK Y T S E D N QS F Q E I I E T A E A K L R QK Y A V L Y N H E K L S A E - - - - - - - - QL QR A L M L P S I ED.rerio 146 F L S K N T S E D N A S F E Q I M E L A E D K D K R R H AWL Y E A E N E Y K E - - - - - - - - R H E QN L A L P S S EM .musculus 155 F L S QY T S E D N A S F Q E I M E V A K E K S H A R H AWL Y QA E E E F E K - - - - - - - - R QK D N L E L P S A E

250 260 270 280 290 300. . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . |

S.pombe 174 K S I AWK E K D K S I K T WN Y QP K N A L M Y T P E T N H S S S - - - L S Q I K K QS T E I QA D A T GL S QS - FC.elegans 231 A V D A D D N R P L N V D NWA Y K AWN T V L F N P D GA A L T P A E I A D A A R K QQT E I N K R GT R F P D S GKC.briggsae 230 A V D S D D N R P L H L D NWK Y K AWN T V L F N P E GA A L T P A E QA E A A K R QK T E I N R K GT R F P D S GRD.melanogaster 224 T Q F E E P D P L R K I E T WN Y T NM N S I M Y V P D GV E Y T E E E R V QL A E R K - QS I QH N A T R L P D E A KD.rerio 198 K QA L E C - T K A GL E T WQY K A K N S L M Y Y P E G E K D D - - - - - T L F K K P - R E V L Y K N T R F E D D P FM .musculus 207 H QA I E S - S QA GV E T WK Y K A K N S L M Y Y P E GV P D E E - - - - QL F K K P - R Q I V H K N T R F L R D P F

310 320 330 340 350 360. . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . |

S.pombe 230 L E A A N QP S T D P I V P P S - - - - - - - - - - - - - - - V N E T D A S V N GY P L V D V N - - - - - - - - - - - -C.elegans 291 L K P S D E AM T R A A V S H A L A N A G - - - - - - K V D F L GN E V T P A N S F K L L E T P N P N P D DM D S P L MC.briggsae 290 L K P S D E AM T R A A V S H A L A N A G - - - - - - K V D F L GN E V T P A N S F K L L E T P N P N P D DM D S P L MD.melanogaster 283 H R EM D T K K L N D E V P QN GA GG - - - - - - - - - - - - A T A T P K V R G F D L L R S P S P R P G E A F S P I MD.rerio 251 C K A L N K S Q I QQA A A L N A Q F K QGK V GP D GK E L L P H E S P K V N GY GY E GA P S P A P GV A E S P L MM .musculus 261 S QA L S R S QL QQA A A L N A QH K QGK V GP D GK E L I P Q E S P R V GG F G F V A T P S P A P GV N E S P L M

370 380 390 400 410 420. . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . |

S.pombe 262 - F GQA V GS - - - - - - - - - - - - - - - S S K S Y F N I P E R P R R E R L H AM R V R D I R S H S T N T T I T S VC.elegans 345 T WG E I D GT P F R L D A P D V T E H S L P GA A P V F K I P E V P Y R E K I A QS M N D S I A A K Y R D K R K V AMC.briggsae 344 T WG E I D GT P F R L D A P D V T E H T L P D A A P I F K I P EM P Y R E K I A QS M N D S I A A K Y R D K R K V A ID.melanogaster 331 T WG E I D GT P F R L D GGD T P L R P - - T QGP S F R I N E N S R R E N I A I A L A E R V S E RM R N QK QM A LD.rerio 311 T WG E I E S T P F R L E GS D T P L V E R - S H GP S F K I P E P GR R E R L GL KM A N E A A A K N R A K K Q E A LM .musculus 321 T WG E V E N T P L R V E GS E S P Y V D R - T P GP T F K I L E P GR R E R L GL KM A N E A A A K N R A K K Q E A L

430 440 450 460 470 480. . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . |

S.pombe 307 D S A S T A L N S Y S - T P N S V S R K L T N L T P A A R R L V A R S - - - - - - - Y L R S P L H GS S P S A S R - - -C.elegans 405 R A A E GA H R T P G F GS K R V S D K L A QL S P A A QK L A T K K L GL KM L P A H K S P F A S P K I M S NWS R PC.briggsae 404 K A A E GA H R T P G F GS K R V S D K L A QL S P A A QK L A T K K L GL KM I P V H R S P F A S P K I GS AWS R SD.melanogaster 389 D T A R R N I GS P - - L I R T NM E R L A S M S P A A QL L A T GK L G I R GT P R L R H T P S P M S GR K R K V T PD.rerio 370 R K V T E N L A S - - - - L T P K GL S P A A L S P A L QR L V N R S S S K Y T D K A L R A S Y T P S P A H R S L G - -M .musculus 380 R R V T E N L A S - - - - L T P K GL S P - AM S P A L QR L V S R T A S K Y T D R A L R A S Y T P S P A R S S H - - -

490 500 510 520 530 540. . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . | . . . . |

S.pombe 355 - - - - - H T A L R T S I P K F S WT P T P R V K S T A P T P K R V - - - - - - - - - - - - - - - - - - - - - - - - - -C.elegans 465 S S S K R S T T P GS AWS R GS T T P GS S WS QGA QT P GT P G I E S M I R R - - - - K GD N V GP S T S A GA AC.briggsae 464 S D S K R S E T P GS AWS R GS GT P GS S WS QGA R T P GT P G I E S M I R K P R R E D D D GA GP S R P S - V AD.melanogaster 447 GV V R S T N T P I L G E P K P K QQA K I S T P A K N V T I D T G - - - - - - - - - - - - - - S T L T D D L L K I P TD.rerio 423 - - - - - A K T P S GGP T T L T P T P T K - - - T R T P A S QD P - - - - - - - - - - - - - - A S I T D N L L QL P -M .musculus 431 - - - - - L K T P A GGP QT P T S T P A P GS A T R T P L T QD P - - - - - - - - - - - - - - A S I T D N L L QL P -

. . . . | . . . .S.pombe 384 - - - - - - - - -C.elegans 521 D R A N A GD F FC.briggsae 523 GR A QA GD F FD.melanogaster 493 K R R T A A D F FD.rerio 461 K R R K A QD Y FM .musculus 472 A R R K A S D F F

Figure S6 Alignment of ESS-2 homologs. The alignment was obtained by the ClustalW multiple-alignment program. Identical and similar resides are highlighted in black and gray, respectively. Deletion in ok3569 mutant was confirmed by RT-PCR and is under-lined in red. Two coiled-coil domains predicted by the MARCOIL program in C. elegans ESS-2 shaded in yellow (See Results). C. elegans ESS-2 b isoform has two additional residues, Phe and Ser, in the C-terminus compared to a isoform (shown as FS in red). The names of ESS-2 homologs and accession numbers in UniProt are the same as in Figure 5C.

7 SI K. Noma, A. Goncharov, and Y. Jin

Table S1 The list of rpm-1(lf) suppressor alleles

* The same mutations that were independently hit more than once from the different parental worms in our screen.

Gene Allele NT change AA change Domain Gene Allele NT change AA change Domaincebp-1 ju1019 G391T, C403T A58S, Q62X mak-2 ju515* G2049A E53K Kinase domaincebp-1 ju634 G407C R63P mak-2 ju567* G2049A E53K Kinase domaincebp-1 ju602 G938A R240K bZip domain mak-2 ju637 C2316T L142F Kinase domaincebp-1 ju640 C968T S250L bZip domain mak-2 ju546 G2727A Splice acceptor Kinase domaincebp-1 ju659 C970T R251C bZip domain mkk-4 ju1034 G635A W140X Kinase domaincebp-1 ju695 G979A A254T bZip domain mkk-4 ju488 G742A C176Y Kinase domaincebp-1 ju1021 C1057T Q280X bZip domain mkk-4 ju686 G841A D194N Kinase domaindlk-1 ju600 G2123A Splice acceptor Kinase domain mkk-4 ju473 G878A G206E Kinase domaindlk-1 ju1046 G2204A G147E Kinase domain mkk-4 ju1040 A923T E221V Kinase domaindlk-1 ju687 G2327A G188E Kinase domain mkk-4 ju1066 G1225T E238X Kinase domaindlk-1 ju477 G2364A M200I Kinase domain mkk-4 ju91 G1350A G265R Kinase domaindlk-1 ju547 G2442A W226X Kinase domain mkk-4 ju1028 C1359T P268S Kinase domaindlk-1 ju624 G2462A G233E Kinase domain mkk-4 ju474 C1365T Q270X Kinase domaindlk-1 ju583 T2465A M234K Kinase domain mkk-4 ju1026 G1405A Splice doner Kinase domaindlk-1 ju692 C2479T Q239X Kinase domain pmk-3 ju512 C1841T P111Sdlk-1 ju589* G2500A D246N Kinase domain pmk-3 ju1049 G1848A G113Ddlk-1 ju660* G2500A D246N Kinase domain pmk-3 ju516* G1899A G130E kinase domaindlk-1 ju691 G2554A D264N Kinase domain pmk-3 ju549* G1899A G130E kinase domaindlk-1 ju551 C2564T T267M Kinase domain pmk-3 ju693 C2017T R153X kinase domaindlk-1 ju635 G2601A M279I Kinase domain pmk-3 ju570 G2341A Splice acceptor kinase domaindlk-1 ju599 G2720A M288I Kinase domain pmk-3 ju658 C2348T R164C kinase domaindlk-1 ju1031 C2746T Q296X Kinase domain pmk-3 ju622 G2360A E168K kinase domaindlk-1 ju648* C2777T S306L Kinase domain pmk-3 ju655 G2516A A205T kinase domaindlk-1 ju1037* C2777T S306L Kinase domain pmk-3 ju597 C2585T H228Y kinase domaindlk-1 ju588 G2796A W312X Kinase domain pmk-3 ju1053 G2618A A239T kinase domaindlk-1 ju625 G2797A E313K Kinase domain pmk-3 ju633 G2639A A246T kinase domaindlk-1 ju1035 G2812A E318K Kinase domain pmk-3 ju1045 T2822A I249K kinase domaindlk-1 ju630 A2828T N323I pmk-3 ju626 G2828T R251C kinase domaindlk-1 ju619* C3006T P347S pmk-3 ju584 C2839T P255S kinase domaindlk-1 ju1029* C3006T P347S pmk-3 ju621* G2979A W301X kinase domaindlk-1 ju513 C3033T Q356X pmk-3 ju627* G2979A W301X kinase domaindlk-1 ju586 G3055A R363H pmk-3 ju1052 G3014A W313X kinase domaindlk-1 ju585* C3060T R365C pmk-3 ju1051 C3017T S314L kinase domaindlk-1 ju623* C3060T R365C pmk-3 ju673 G3110A C345Y kinase domaindlk-1 ju1032* C3060T R365C pmk-3 ju548 C3119T P348L kinase domaindlk-1 ju590 G3409A Splice acceptor pmk-3 ju632 A3316T I414F kinase domaindlk-1 ju591 C3470T L459P Leucine zipper pmk-3 ju671 G3322T H416Y kinase domaindlk-1 ju1044 C3473T Q460X Leucine zipper pmk-3 ju485 G3358A E428Kdlk-1 ju598 C3662T R523X pmk-3 ju1010 G3420A W448Xdlk-1 ju596* G4228A W581X pmk-3 ju592 G3516A W453Xdlk-1 ju1039* G4228A W581X supr-1 ju1118 G1933A W292Xdlk-1 ju649 G4296A W604X uev-3 ju638 G611A Splice acceptordlk-1 ju694 G4346A Splice donor uev-3 ju593 G1875A Splice acceptor UEV domaindlk-1 ju476 After 4467 5 bp insertion uev-3 ju587 G2058A Splice acceptor UEV domaindlk-1 ju653 C5482T Q700X uev-3 ju639 26 bp deletion UEV domaindlk-1 ju636 C5830T R816Xdlk-1 ju475* C5929T R849Xdlk-1 ju656* C5929T R849Xdlk-1 ju620 G5993A G870E Activation peptide

8 SI K. Noma, A. Goncharov, and Y. Jin

Table S2 The list of strainsStrain Genotype

CZ1252 rpm-1(ju44)

CZ900 syd-2(ju37)

CZ1338 juIs1; rpm-1(ju44); syd-2(ju37)

CZ1339 rpm-1(ju44); syd-2(ju37)

CZ17301 rpm-1(ju23); syd-2(ju37); juIs390[Prgef-1-RPM-1::GFP]

CZ15454 dlk-1(ju1054); juIs1; rpm-1(ju44); syd-2(ju37)

CZ15434 juIs1; rpm-1(ju44); mkk-4(ju1034) syd-2(ju37)

CZ6506 pmk-3(ju693) juIs1; rpm-1(ju44); syd-2(ju37)

CZ15419 juIs1; rpm-1(ju44); cebp-1(ju1019) syd-2(ju37)

CZ1234 rpm-1(ju23)

CZ15956 dlk-1(tm4024)

CZ20842 dlk-1(tm4024); juIs1; rpm-1(ju44); syd-2(ju37)

CZ5180 mak-2(ju546) juIs1; rpm-1(ju44); syd-2(ju37)

CZ5320 juIs1; rpm-1(ju44); cebp-1(ju634) syd-2(ju37)

CZ333 juIs1

CZ4926 juIs1; syd-2(ju37)

CZ455 juIs1; rpm-1(ju44)

CZ8835 muIs32

CZ3073 muIs32; rpm-1(ju44)

CZ14005 dlk-1(tm4024); muIs32; rpm-1(ju44)

CZ17305 uev-3(ju638); muIs32; rpm-1(ju44)

CZ8456 muIs32; mak-2(ok2394); rpm-1(ju44)

CZ17322 muIs32; rpm-1(ju44); cebp-1(tm2807)

CZ7412 muIs32; dpy-11(e224) rpm-1(ju44)

CZ18388 muIs32; ml-1(km19) dpy-11(e224) rpm-1(ju44)

CZ17060 muIs32; kgb-1(um3); rpm-1(ju44)

CZ17072 muIs32; kgb-1(um3) kgb-2(km16); rpm-1(ju44)

CZ17589 kgb-1(um3); rpm-1(ju44); syd-2(ju37)

CZ17724 supr-1(ju1118); muIs32; rpm-1(ju44)

CZ17580 supr-1(gk106345); muIs32; rpm-1(ju44)

CZ17522 supr-1(tm3467); muIs32; rpm-1(ju44)

CZ18027 supr-1(ju1118); muIs32; rpm-1(ju44); juEx5384[Pmec-4-SUPR-1(cDNA)]

CZ18414 supr-1(ju1118); muIs32; rpm-1(ju44); juEx5517[Pmec-4-GFP::SUPR-1(cDNA)]

CZ18420 juEx5517[Pmec-4-GFP::SUPR-1(cDNA)]

CZ19707 dlk-1(ju600); muIs32; rpm-1(ju44)

CZ17520 muIs32; ess-2(ju1117); rpm-1(ju44)

CZ17518 muIs32; ess-2(tm4246); rpm-1(ju44)

CZ17585 muIs32; ess-2(ok3569); rpm-1(ju44)

CZ17519 dlk-1(ju600); muIs32; ess-2(ju1117); rpm-1(ju44)

CZ17517 dlk-1(ju600); muIs32; ess-2(tm4246); rpm-1(ju44)

CZ20928 dlk-1(ju600); muIs32; ess-2(ok3569); rpm-1(ju44)

CZ21083 dlk-1(ju600); muIs32; ess-2(tm4246); rpm-1(ju44); juEx6352[Pess-2-ESS-2(gDNA)]

CZ17593 dlk-1(ju600); muIs32; ess-2(tm4246); rpm-1(ju44); juEx5241[Pmec-4-ESS-2(cDNA)]

CZ19400 dlk-1(ju600); muIs32; ess-2(tm4246); rpm-1(ju44); juEx5859[Pmec-4-ESS-2(cDNA)::GFP]

CZ19402 dlk-1(ju600); muIs32; ess-2(tm4246); rpm-1(ju44); juEx5861Pmec-4-GFP::ESS-2(cDNA)]

CZ21081 dlk-1(ju600); muIs32; ess-2(tm4246); rpm-1(ju44); juEx6350[Pmec-4-ESS-2(ok3569)]

CZ21082 dlk-1(ju600); muIs32; ess-2(tm4246); rpm-1(ju44); juEx6351[Pmec-4-ESS-2(ju1117)]

CZ19415 juEx5859[Pmec-4-ESS-2(cDNA)::GFP]

CZ18402 ess-2(tm4246)

CZ19531 dpy-10(e128)

CZ19532 dpy-10(e128); ess-2(tm4246)

*juIs1[Punc-25-SNB-1::GFP], muIs32[Pmec-7-GFP]

9 SI K. Noma, A. Goncharov, and Y. Jin