Andersen, Erik et al. (2021) International Worm Meeting "The Caenorhabditis Natural Diversity Resource: expanded and enhanced"

Andersen, Erik, Tanny, Robyn, Evans, Kathryn

[International Worm Meeting, 2021]

Natural diversity offers a treasure trove of variants (or mutations) that can be used to understand quantitative traits and evolutionary processes. The Caenorhabditis elegans Natural Diversity Resource (CeNDR) has archived and disseminated wild strains, whole-genome sequenced those strains, and released variant data and effect predictions to the community. Also, CeNDR facilitates genome-wide association (GWA) mappings to connect natural differences in quantitative traits to variants and molecular mechanisms. Since 2016, CeNDR has shipped more than 4,000 strains and enabled over 3,000 GWA mappings for the community. We continue to archive and sequence new wild strains identified by collaborators and citizen scientists, offering more than 1,300 genetically unique strains. In addition to these new strains, we updated CeNDR to make it more user-friendly, created a broad-sense heritability calculator, modernized the variant browser and variant effect predictions, and created a powerful new GWA mapping portal. These additions and updates make CeNDR more accessible and able to facilitate natural variation studies in laboratories often focused on the laboratory-adapted N2 strain. We present the evolution of CeNDR into the Caenorhabditis Natural Diversity Resource (CaeNDR) where we have incorporated over 1,600 C. briggsae and 700 C. tropicalis wild strains and genome sequences. The same browsers and tools created for C. elegans are available for these species. CaeNDR will expand comparative genomics across the three selfing Caenorhabditis species and enable new approaches to quantitative genetics and population genomics.

Andersen, Erik C. et al. (2013) International Worm Meeting "Using natural variation to decipher the complex genetic causes of C. elegans drug sensitivities."

Andersen, Erik C., Shimko, Tyler

[International Worm Meeting, 2013]

C. elegans is an important model for a variety of biomedically relevant traits, including sensitivities to diverse compounds and pharmaceuticals. Sensitivities to these compounds vary in the human population, and we know little about the genetic causes of this variation. Using C. elegans, we hope to identify and characterize these variant genes exploiting the existing natural genetic and phenotypic variation of the species. Quantitative geneticists use the correlations between genetic and phenotypic differences to identify the variants that control complex traits. Species-wide variation in genome-wide association studies or pairwise strain variation in linkage mapping studies is used to identify variant genes. Using these two approaches, one needs to score the phenotypes of a large number of independent strains. In C. elegans, we must phenotype over one hundred wild strains or nearly six hundred N2xCB4856 recombinant inbred lines in drug susceptibility assays. In order to score these large collections of independent strains for drug susceptibility traits, we optimized highly reproducible and high-throughput growth, fecundity, and pharyngeal pumping assays to score 384 strains per day per researcher in the presence of any aqueous compound. These assays allow us to probe both the acute and chronic phenotypic effects of important drugs, like chemotherapeutics, pesticides, and anthelmintics. To determine the efficacy of these approaches, we scored susceptibility to the most widely applied herbicide paraquat. In one week, we phenotyped growth and fecundity of 97 wild strains and 574 recombinant inbred lines between N2 and CB4856. These data implicated genomic intervals as small as 274 kb with 60 genes that control paraquat sensitivity. To verify these loci, we constructed nearly isogenic lines, in which a genomic region from one strain is isolated in another strain background. These results recapitulate the predicted effects, and we will discuss methods to characterize small phenotypic effects. Additionally, we will discuss recent progress towards scaling up these techniques to a larger set of compounds and pharmaceuticals.

Andersen, Erik C. et al. (2011) International Worm Meeting "Recent chromosome-scale selective sweeps reshaped C. elegans genomic diversity."

Gerke, Justin P., Crissman, Jonathan R., Andersen, Erik C., Ghosh, Raj, Felix, Marie-Anne, Kruglyak, Leonid, Bloom, Joshua S., Shapiro, Joshua A.

[International Worm Meeting, 2011]

Even though C. elegans is central to research in molecular, cell, and developmental biology, nearly all of the research on this organism has been conducted on a single strain - N2. Comparatively little is known about the genomic and phenotypic variation of the species or its evolutionary history. I will report the characterization of C. elegans genetic variation using high-throughput selective sequencing of eight megabases from a worldwide collection of 200 wild strains representing nearly every known collection location. The identified set of 41,188 single nucleotide polymorphisms (SNPs) was used to define the most genotypically diverse strains for future quantitative genetic studies. We used these SNPs and strains to map loci controlling novel quantitative traits to haplotypes of less than 50 kb by genome-wide association studies. Unexpectedly, we found that the genome is dominated by a set of commonly shared haplotypes that each spans many megabases, suggesting recent shared ancestry. Furthermore, the distribution of polymorphism is strikingly non-uniform and correlates with variation in recombination rate found between the N2 and CB4856 strains. Simulations show that these patterns were generated by selective sweeps that have dramatically reduced variation worldwide and suggest that at least one of these sweeps occurred in the past few hundred years. Likely the result of human activity, these sweeps dramatically reshaped the global C. elegans population.

Andersen, Erik C. et al. (2011) International Worm Meeting "N2-specific differences in behavior, growth, and physiology are explained by a laboratory-derived allele of npr-1."

Andersen, Erik C., Kruglyak, Leonid, Bloom, Josh S., Gerke, Justin P., Ghosh, Raj

[International Worm Meeting, 2011]

Over the past forty years, an amazing amount of research has been performed on a single C. elegans strain: N2. This strain may have gone through as many as 1000 generations (over 16 years) before long-term storage was discovered (1). During that time, N2 likely accumulated many mutations through selection in the laboratory. We know of at least two genes that are affected by laboratory-derived mutations, npr-1 and glb-5, which regulate an oxygen avoidance behavior in the presence of food (1). Not only is the N2 strain easier to pick and propagate, as clumping, bordering, and burrowing are highly reduced, but we will present evidence that it grows larger, has more offspring, crawls faster, and is more resistant to pathogen infection than wild C. elegans strains because of the lab-derived allele of npr-1. We mapped quantitative trait loci (QTL) for each of these traits using an advanced intercross recombinant inbred line (RIAIL) collection between N2 and the wild isolate from Hawaii CB4856. Additionally, we mapped thousands of gene expression differences between these two strains to QTL overlapping npr-1. We will present data that suggest npr-1 is the causal gene for many of these traits, using nearly isogenic lines (NILs) and loss-of-function alleles, and the growth and physiological differences are caused by the npr-1-mediated oxygen avoidance behaviors. We are interested in identifying natural variants between C. elegans strains for a variety of phenotypic traits. Nevertheless, npr-1 makes complex trait mapping more difficult, as it explains the majority of the genetic trait variance for crosses involving N2. To circumvent the laboratory-derived npr-1 allele, we constructed a collection of 359 RIAILs using CB4856 and a nearly isogenic line that has the normal form of npr-1 in an otherwise N2 background. However, this NIL with corrected npr-1 has other N2-specific laboratory-derived mutations, so using high-throughput selective sequencing, we have identified the most divergent C. elegans strains for use in future quantitative genetic studies. Because assays on agar plates lead to differences in food exposure, oxygen concentration, and humidity, we have optimized high-throughput liquid assays of growth and fecundity using robotic liquid handling, worm sorting, and image analysis. (1) McGrath et al. Neuron 2009, (61)5: 692-699.

Andersen, Erik C. et al. (2009) International Worm Meeting "Quantitative genetic analyses of C. elegans intraspecies variation in responses to Pseudomonas aeruginosa and Staphylococcus aureus."

Reddy, Kirthi, Okhah, Zachary, Kruglyak, Leonid, Andersen, Erik C., Kim, Dennis H.

[International Worm Meeting, 2009]

The nematode Caenorhabditis elegans responds to pathogenic bacteria with conserved innate immune responses and pathogen avoidance behaviors. We investigated natural variation in C. elegans responses to different pathogen infections, using a collection of advanced intercross recombinant inbred lines (RILs) between the laboratory wild-type strain N2 and a wild isolate from Hawaii CB4856. Initially, we chose to focus on the opportunistic human pathogen Pseudomonas aeruginosa, a gram-negative, aerobic bacterium. Quantitative trait mapping of the susceptibilities of 126 RILs led to the identification of a polymorphism in the npr-1 gene as causal in the difference in susceptibility to P. aeruginosa between these two strains. We showed that the mechanism of NPR-1-mediated pathogen resistance is through oxygen-dependent behavioral avoidance rather than through direct regulation of innate immunity. Because the npr-1 allele present in the N2 strain is a lab-derived gain-of-function allele (1), we sought natural variants involved in pathogen resistance by scoring susceptibility to P. aeruginosa in only those RILs with the ancestral npr-1 allele (found in CB4856). Through quantitative genetic analysis of 89 additional strains, we identified a second quantitative trait locus (QTL) on LGV. We will discuss our progress in determining the gene underlying this novel QTL. Subsequently, we investigated differences in the susceptibilities of N2 and CB4856 to Bacillus thuringiensis, Candida albicans, Cryptococcus neoformans, Enterococcus faecalis, Serratia marcescens, and Staphylococcus aureus. Assays using the gram-positive bacterium S. aureus had the most robust survival difference between the two parental strains. Quantitative trait mapping of the susceptibilities of 73 RILs led to a causal polymorphism in the npr-1 gene, indicating that behavior determines the susceptibility to this pathogen as well. We will present our progress on the identification of additional QTL controlling susceptibility to S. aureus. (1) McGrath PT et al. Neuron 2009; 61(5):692-699.

Erik Andersen et al. (2005) International Worm Meeting "Cloning and Characterization of Suppressors of the Synthetic Multivulva Phenotype"

Bob Horvitz, Xiaowei Lu, Scott Clark, Erik Andersen

[International Worm Meeting, 2005]

The synthetic Multivulva (synMuv) genes are grouped into at least three functionally redundant classes, A, B, and C, that negatively regulate the specification of vulval cell fates. Animals mutant for one or more genes within the same class are non-Muv. Animals mutant for genes within any two classes are Muv. Some of the identified class B gene products are homologs of members of at least two transcriptional repression complexes.To identify loci that interact genetically with the synMuv genes, we performed two screens for synMuv suppressors. From these screens, we isolated 166 suppressors of the synMuv phenotype of lin-15AB(n765ts) animals and 43 suppressors of the synMuv phenotype of lin-53(n833); lin-15A(n767) animals. The synMuv suppressor isw-1 may act with the genes encoding homologs of the Drosophila NURF complex to antagonize the synMuv phenotype, because loss of the NURF301 and NURF38 homologs suppresses the synMuv phenotype of lin-15AB(n765ts) animals. In our characterization of histone methyltransferase genes (see abstract by Andersen and Horvitz), we found that mes-2 and mes-4 are synMuv suppressors. Loss-of-function mutations in mes-3 and mes-6 but not mes-1 suppress the synMuv phenotype. MES-2 and MES-6 are homologs of the Polycomb proteins E(z) and ESC, respectively. The Strome laboratory has shown that MES-2, MES-3, and MES-6 form a complex in embryos, methylate histone H3 lysine 27, and may play a role in the localization of MES-4. We are currently investigating how histone methylation can influence the synMuv phenotype and whether the predicted NURF complex homologs and the mes genes interact. We identified lst-3(n2070) as a dominant suppressor of the synMuv phenotype of lin-15AB(n765ts) mutants. The Greenwald laboratory has implicated lst-3 through bioinformatics as a putative target gene of LAG-1, a lin-12 and glp-1 Notch effector, and postulated that it may mediate the interplay between the RTK/Ras and Notch signal transduction pathways. The dominant phenotype conferred by lst-3(n2070) affects the putative DNA-binding domain and may increase wild-type function. We isolated several nonsense cis-dominant suppressor alleles and a deletion allele to study loss of lst-3 function. Putative null mutations of lst-3 confer a synMuv phenotype in combination with mutations in class A but not class B synMuv genes. These gain-of-function and loss-of-function alleles of lst-3 may allow for a genetic analysis of the interplay between the RTK/Ras, Notch, and synMuv pathways.

Erik Andersen et al. (2006) Development & Evolution Meeting "lst-3 Negatively Regulates the Vulval Cell-fate Decision"

Bob Horvitz, Melanie Worley, Erik Andersen, Scott Clark

[Development & Evolution Meeting, 2006]

The C. elegans vulva is formed by the descendants of three of six equipotent hypodermal blast cells. Two distinct vulval cell fates are specified by the actions of at least two conserved antagonistic signaling pathways, Ras and Notch. Mutations affecting these pathways cause the generation of ectopic vulval cells and a multivulva (Muv) phenotype. To understand more about factors controlling the vulval cell-fate decision, we performed a screen for suppressors of the synthetic multivulva (synMuv) phenotype. The synMuv genes are grouped into at least three functionally redundant classes, A, B and C, that negatively regulate the specification of vulval cell fates. Animals mutant for one or more genes within the same class are non-Muv, whereas animals mutant for genes within any two classes are Muv. We identified n2070 as a dominant suppressor of the synMuv phenotype of lin-15AB(n765) mutants. Our gene dosage experiments suggest that the dominant phenotype conferred by n2070 is caused by an increase in wild-type function. We mapped n2070 to a 120-gene interval on LGIV and then used RNAi to inactivate the genes in the interval to eliminate the synMuv suppression phenotype of n2070. RNAi of only lst-3 (lst, lateral signaling target) reverted the synMuv suppression phenotype caused by n2070. In a screen for cis-dominant suppressors of n2070, we identified three independent isolates with nonsense mutations in lst-3. n2070 causes a proline-to-leucine substitution in a putative DNA-binding domain of LST-3. The Greenwald laboratory identified lst-3 as a possible transcriptional target of LAG-1, a lin-12 Notch effector, and suggested that lst-3 mediates the interplay between the Ras and Notch pathways. lst-3 encodes a transcription factor similar to mammalian CARP-1, a cell cycle and apoptosis regulator. We isolated a deletion allele of lst-3. This allele confers a synMuv phenotype in combination with mutations in class A or C but not with mutations in class B synMuv genes. Through gene expression analysis of gain-of-function and loss-of-function lst-3 mutants, we have identified candidate LST-3 target genes. We will discuss how the modulation of lst-3 function might affect the Ras and Notch pathways.

Erik Andersen et al. (2004) East Coast Worm Meeting "met-1 and met-2, Two Putative Histone Methyltransferases, May Act as Synthetic Multivulva Genes"

Erik Andersen, Bob Horvitz

[East Coast Worm Meeting, 2004]

Repression of euchromatic gene transcription may occur through the specific covalent modification of histone tails leading to the recruitment of trans -acting repressive factors. In one model, through the actions of the Nucleosome Remodeling and Deacetylase (NuRD) complex, lysine 9 of histone H3 (H3K9) is deacetylated and methylated by a histone methyltransferase (HMTase), which ultimately leads to recruitment of heterochromatin protein 1 (HP1) and a transcriptionally silenced state. Among the proteins encoded by the class B synthetic Multivulva (synMuv) genes are homologs of some NuRD complex members, including LIN-53 p48, HDA-1 HDAC, and LET-418 Mi-2. Within the Pn.p cells fated to adopt a non-vulval cell fate, a putative LIN-53/HDA-1/LET-418 transcriptional repression complex may be recruited to genes to be repressed through interaction with LIN-35 Rb and the sequence-specific heterodimeric transcription factor DPL-1 DP/EFL-1 E2F. Through the actions of the histone deacetylase HDA-1, histones may be deacetylated. After subsequent methylation of histones, the C. elegans HP1 homolog (and synMuv protein) HPL-2 may be recruited, leading to the repression of genes that would normally promote a vulval fate. Given this model, we decided to seek HMTase-encoding genes that interact genetically with the synMuv genes. All known HMTases specific for modifying histone tail residues have an enzymatic SET domain. Through BLAST, Pfam, and SMART database searches we identified and classified the 31 genes predicted to encode SET-domain-containing proteins in C. elegans . The HMTases predicted to modify histone H3 lysines 4, 9, or 36 have conserved domains flanking the SET domain. There are nine genes in C. elegans predicted to encode such HMTases. Using existing mutant alleles or deletion alleles that we have generated, we tested all nine of these HMTases for a synMuv phenotype with mutations in synMuv class A or B genes. Additionally, we injected dsRNA to inactivate each of the 31 genes predicted to encode SET domain containing genes and tested for a synMuv phenotype with mutations in synMuv class A or B genes. Mutations in two genes, met-1 and met-2, cause a synMuv phenotype in combination with mutations in all synMuv class A genes but not with mutations in synMuv class B genes. Based upon sequence similarity, met-1 is predicted to encode a histone H3 lysine 36 (H3K36) HMTase. In S. cerevisiae , methylation of H3K36 may play a role in transcriptional elongation or repression. We are currently testing whether perturbations in transcriptional elongation can cause a synMuv phenotype. met-2 is predicted to encode a histone H3K9 HMTase. We are testing whether met-2 plays a role in the recruitment of HPL-2. Interestingly, strains doubly mutant for met-1 and met-2 are synthetically Muv. We will discuss our model for the redundant actions of these putative histone methyltransferases and their interactions with the products of the synMuv genes.

Erik Andersen et al. (2004) East Coast Worm Meeting "Characterization of the Synthetic Multivulva Suppressor Gene isw-1, a Homolog of the Drosophila Chromatin-Remodeling ATPase ISWI"

Scott Clark, Bob Horvitz, Xiaowei Lu, Erik Andersen

[East Coast Worm Meeting, 2004]

The synthetic Multivulva (synMuv) genes are grouped into at least three functionally redundant classes, A, B, and C, that negatively regulate the induction of vulval cell fates. Animals with a mutation in one or more genes within the same class are non-Muv. Animals with mutations in genes within any two classes are Muv. Among some of the identified class B gene products are counterparts of a transcriptional repression complex, and among the identified class C gene products are homologs of a putative transcriptional activation complex. To identify genes that interact genetically with the synMuv genes, we performed two screens for synMuv suppressors. From these screens, 166 suppressors of the synMuv phenotype of lin-15AB(n765ts) animals and 43 suppressors of the synMuv phenotype of lin-53(n833); lin-15A(n767) animals were isolated. We cloned one lin-53(n833); lin-15A(n767) synMuv suppressor. It encodes a homolog of the chromatin-remodeling ATPase ISWI. We named this gene isw-1 . RNAi or a presumptive null allele of isw-1 can suppress the Muv phenotype of most, if not all, synMuv mutant combinations. By contrast, loss of isw-1 function did not suppress the Muv phenotype of a null mutant of lin-1 , a RTK/Ras pathway effector. The loss-of-function Vulvaless phenotype of mutants in the RTK/Ras pathway is epistatic to the synMuv phenotype. We conclude that isw-1 is likely to act downstream of or in parallel to the synMuv genes and upstream of or in parallel to the RTK/Ras pathway. Immunohistochemistry using antisera specific to ISW-1 suggests that ISW-1 is in the nuclei of most if not all cells throughout development. Levels of ISW-1 are not significantly changed in synMuv mutants as compared to the wild type based on quantitative western blotting. We are currently trying to identify in which cells and during which developmental times isw-1 is required to allow the synMuv phenotype. ISWI has been biochemically purified from Drosophila embryo extracts as a component of multiple complexes, including the Nucleosome Remodeling Factor (NURF), the Chromatin Accessibility Complex (CHRAC), and the ATP-utilizing Chromatin assembly and remodeling Factor (ACF). RNAi of C. elegans homologs of NURF, CHRAC, and ACF complex members in a lin-15AB(n765ts) background was used to test for suppression of the synMuv phenotype. A NURF301 homolog and the NURF38 homolog dhp-2 were identified as synMuv suppressor genes likely to act with isw-1 to regulate vulval development. Additionally, a deletion allele of the NURF301 homolog can suppress the synMuv phenotype of lin-15AB(n765ts) animals. Thus, ISW-1 may be acting as a component of a C. elegans NURF-like complex to antagonize the synMuv phenotype. We are mapping and cloning additional mutationally-defined suppressors of the synMuv phenotype in an effort to identify other factors that may act with a putative C. elegans NURF-like complex or in parallel processes to antagonize the synMuv phenotype.

Erik Andersen et al. (2005) International Worm Meeting "MET-1 and MET-2, Two Putative Histone Methyltransferases, Are Redundantly Required for Proper Vulval Cell-Fate Determination in C. elegans"

Bob Horvitz, Erik Andersen

[International Worm Meeting, 2005]

The identification and characterization of genes that when mutated affect the vulval cell fate have helped lead to a fundamental understanding of the RTK/Ras, Wnt, and Notch signal transduction pathways. The synthetic Multivulva (synMuv) mutants may help to define a role in the specification of cell fate for chromatin remodeling and transcriptional modulation. The synMuv genes are grouped into at least three functionally redundant classes, A, B, or C, that negatively regulate the specification of the vulval cell fate. Animals mutant for one or more genes of the same class are non-Muv. Animals mutant for genes within any two classes are Muv. Several synMuv genes encode counterparts of proteins implicated in transcriptional repression and chromatin remodeling. Given the possible involvement of chromatin remodeling and the likely conservation of transcriptional repression mechanisms found in other organisms, we investigated whether histone methyltransferases (HMTs) play a role in vulval cell-fate specification.Using database searches and protein alignments, we identified the 33 genes predicted to encode lysine-specific HMTs that contain the evolutionarily-conserved enzymatic SET domain. We inactivated each of these 33 genes by RNAi and assayed for vulval defects. We also deleted a subset of the HMTs with cysteine-rich domains flanking the SET domain that are predicted to be enzymatic HMTs and assayed for vulval defects. Through these two approaches, we found that met-1 or met-2 loss of function conferred a synMuv phenotype in combination with class A or class C synMuv mutations, mes-2 or mes-4 loss of function suppressed the synMuv phenotype of lin-15AB(n765ts) mutants, and several other predicted HMTs are required for embryonic viability. MET-1 is a homolog of yeast Set2p and predicted to methylate histone H3 lysine 36 (H3K36). MET-2 is the homolog of human SETDB1 and predicted to methylate histone H3 lysine 9 (H3K9). Quantitative western blots of wild-type, met-1, and met-2 extracts showed that met-1 and met-2 are required for both H3K9 and H3K36 trimethylation but not for H3K4 nor H3K27 trimethylation in vivo. Interestingly, met-1; met-2 double mutants are Muv, but met-1 and met-2 do not act redundantly with mutations in class B synMuv genes. Little is known about the function of H3K36 HMTs outside of S. cerevisiae and even less is known about interactions with H3K9 HMTs. These studies may lead to further insight into the roles of H9K9 and H3K36 methylation in specifying a developmental cell fate in a multicellular organism.