- Research article
- Open Access
A phylogeny and molecular barcodes for Caenorhabditis, with numerous new species from rotting fruits
- Karin C Kiontke†1Email author,
- Marie-Anne Félix†2,
- Michael Ailion3, 6,
- Matthew V Rockman1, 4,
- Christian Braendle5,
- Jean-Baptiste Pénigault2 and
- David HA Fitch1
© Kiontke et al; licensee BioMed Central Ltd. 2011
Received: 10 March 2011
Accepted: 21 November 2011
Published: 21 November 2011
The nematode Caenorhabditis elegans is a major laboratory model in biology. Only ten Caenorhabditis species were available in culture at the onset of this study. Many of them, like C. elegans, were mostly isolated from artificial compost heaps, and their more natural habitat was unknown.
Caenorhabditis nematodes were found to be proliferating in rotten fruits, flowers and stems. By collecting a large worldwide set of such samples, 16 new Caenorhabditis species were discovered. We performed mating tests to establish biological species status and found some instances of semi-fertile or sterile hybrid progeny. We established barcodes for all species using ITS2 rDNA sequences. By obtaining sequence data for two rRNA and nine protein-coding genes, we determined the likely phylogenetic relationships among the 26 species in culture. The new species are part of two well-resolved sister clades that we call the Elegans super-group and the Drosophilae super-group. We further scored phenotypic characters such as reproductive mode, mating behavior and male tail morphology, and discuss their congruence with the phylogeny. A small space between rays 2 and 3 evolved once in the stem species of the Elegans super-group; a narrow fan and spiral copulation evolved once in the stem species of C. angaria, C. sp. 8 and C. sp. 12. Several other character changes occurred convergently. For example, hermaphroditism evolved three times independently in C. elegans, C. briggsae and C. sp. 11. Several species can co-occur in the same location or even the same fruit. At the global level, some species have a cosmopolitan distribution: C. briggsae is particularly widespread, while C. elegans and C. remanei are found mostly or exclusively in temperate regions, and C. brenneri and C. sp. 11 exclusively in tropical zones. Other species have limited distributions, for example C. sp. 5 appears to be restricted to China, C. sp. 7 to West Africa and C. sp. 8 to the Eastern United States.
Caenorhabditis are "fruit worms", not soil nematodes. The 16 new species provide a resource and their phylogeny offers a framework for further studies into the evolution of genomic and phenotypic characters.
The nematode worm Caenorhabditis elegans is a key laboratory model system which has provided key insights into molecular biology (e.g. RNA interference and small RNAs), cell biology (cell polarity, apoptosis), developmental biology (signal transduction pathways, developmental timing) and neurobiology (axon guidance, synaptic function). The use of C. elegans and related species for evolutionary biology has recently increased [see [1–9]]. Several characteristics make this roundworm an interesting species for evolutionary studies, among them the accumulated knowledge on its biology, its simplicity of use (including the ability to cryogenically preserve living strains) and its selfing mode of reproduction with facultative outcrossing.
However, C. elegans lacks an extensive evolutionary framework of closely related species, especially when compared to Drosophila melanogaster. Ten Caenorhabditis species were available in culture prior to this work, compared with over 2000 described Drosophila species . More Caenorhabditis species are known from the literature [11, 12] but they have not been re-isolated and are thus not available for further studies. In addition, molecular divergence among Caenorhabditis species is greater than among Drosophila species . For example C. elegans and C. briggsae, regarded as close relatives within Caenorhabditis, are probably as distant as D. melanogaster and D. ananassae, regarded as fairly distant relatives within Drosophila [13, 14]. The tiny size of these animals and the small number of taxonomists focusing on this taxonomic group may explain in great part the paucity of described species.
A key additional reason for the lack of known diversity in the Caenorhabditis genus is that these so-called "soil nematodes" are rarely found in soil samples. Soil samples yield a variety of nematode species, including a few species of the family Rhabditidae (to which Caenorhabditis belongs). For example, some Oscheius species are readily found in soil samples . However, despite extensive sampling for many years, we failed to isolate Caenorhabditis from soil . Rare positive instances correspond to soil of orchards (e.g. C. elegans strain JU258, Madeira 2001), or soil below trees with rotting fruits (C. sp. 5 JU727). Instead, C. elegans, C. briggsae and C. remanei were found in compost heaps containing decaying vegetal material [11, 17].
We screened rotting vegetal material for the presence of Caenorhabditis and found that Caenorhabditis species are most readily isolated from rotting fruits and flowers, and occasionally from other rotten plant parts (e.g. banana pseudo-stems, but rarely leaves). Focusing on rotten fruit samples, we found a number of new isolates of C. elegans, C. briggsae, C. remanei and C. brenneri. In addition, we collected sixteen new Caenorhabditis species, which dramatically increased the number of Caenorhabditis species presently in culture to 26. We obtained sequences of SSU and LSU rRNA genes and of nine protein-coding genes in these new isolates and built an expanded molecular phylogeny of the genus Caenorhabditis. We tested the new species for reproductive isolation by pairwise mating tests. Finally, we explored the ITS regions of the rRNA gene array for their suitability as a barcode and found that a specimen can reliably be assigned to one of the Caenorhabditis species in culture using the ITS2 sequence. Finally, we report on the evolution of a number of phenotypic characters in the genus, including the mode of reproduction.
Sampling and isolation
Rotting vegetal material was sampled and stored in tubes or plastic bags. Care was taken during storage and transport to provide some oxygen to the samples, and excessive heat was avoided. Caenorhabditis individuals could be retrieved after up to three weeks of sample storage.
Collected samples were placed in the laboratory on standard C. elegans culture plates seeded in the center with the E. coli OP50 strain. The samples were deposited outside the bacterial lawn and humidified by addition of one to three milliliters of water or M9 buffer . Caenorhabditis nematodes are attracted by the E. coli lawn and tend to remain on the lawn, often along the thicker lawn edge. They can often be recognized with a dissecting microscope equipped with transillumination (40-50×) by a combination of morphological criteria: the color of their intestine (light brown), the large intestinal cell nuclei visible as white disks on the brown intestinal cytoplasm background, the long and fine tail of the adult hermaphrodite, the vulva position in the center of the animal, and the short and round tail tip in the adult male . Using a compound microscope equipped with differential interference contrast, further scorable characters are the presence of a round median pharyngeal bulb, the characteristic vulva cell division pattern , and the shape and arrangement of rays in the male tail, including a ray 6 that lacks a tip opening to the outside and which tapers from a wide base .
Cultures were established by isolation of a single animal with a female soma. For gonochoristic species, we picked either a female with a copulatory plug, or one female and one male. The mode of reproduction was determined by isolating virgin L4 females/hermaphrodites and scoring for the presence of progeny. For selfing isolates, isogenic strains were produced by isolating a single larva for a few (3-6) generations. For male-female isolates, strains were established as isofemale lines or cryopreserved as large, multi-founder populations to maintain the sample's genetic diversity. Most strains are listed in WormBase (http://www.wormbase.org) and in http://www.justbio.com/worms/index.php and will soon be listed with distribution maps in RhabditinaDB (http://wormtails.bio.nyu.edu).
Species identification and mating tests
Males and females from new Caenorhabditis isolates were compared at the morphological level with known species either by studying them alongside individuals from cultured strains or by consulting published species descriptions. If the morphology of a new isolate was not unique, mating tests were performed with individuals of morphologically similar strains.
Isolates with a male-female mode of reproduction
3-6 fourth larval stage (L4) females and 3-6 L4 or adult males were placed on a 55-mm Petri dish seeded with E. coli OP50 to mate. The cross was scored regularly for the presence of laid embryos, hatched larvae, and fertile adults. Most crosses were performed at least in duplicates. Many crosses between different Caenorhabditis species were successful at the mating behavior level, as evidenced by the presence of a mating plug deposited by the male on the female vulva, yet no embryos were laid. Other crosses produced dead embryos, and in some cases sterile female larval and adult progeny (Additional File 1). F1 sterility was assessed by placing F1 females either with sibling males from the same cross (if any) or males from either parental genotype.
Isolates with a selfing mode of reproduction
Hermaphrodites mostly produce hermaphrodite progeny upon selfing, and rare males by non-disjunction of the × chromosome. The cross-progeny of (cross-fertile) hermaphrodites and males consists of about 50% males. To test for cross-fertility, 3-6 hermaphrodites of one isolate were placed with 3-6 males of the other isolate on a 55-mm Petri dish seeded with E. coli OP50. The presence of numerous males (over 20%, much more than on control plates seeded with only hermaphrodites) on a plate indicates a successful cross and provides a test for biological species .
PCR, sequencing and sequence alignment
We attempted to obtain partial sequences of 11 genes for all Caenorhabditis species: genes for SSU and LSU rRNA, the largest subunit of RNA polymerase II (RNAP2, ama-1 in C. elegans) as in  and lin-44 (encoding a Wnt signaling factor), par-6 (encoding a PDZ-domain-containing protein), pkc-3 (encoding an atypical protein kinase C) and the orthologs of C. elegans genes ZK686.3 (G43, orthologous to the putative tumor suppressor N33), W02B12.9 (G140, orthologous to the mitochondrial carrier protein MRS3/4), ZK795.3 (G3857, orthologous to a U3 small nucleolar ribonucleoprotein component), Y97E10AL.2 (OMCL4763, a predicted alpha/beta hydrolase) and Y45G12B.2a (OMCL4988, a predicted E3 ubiquitin ligase). Most of these genes were chosen for their expected information content for phylogenetic analysis as derived from genome sequences of six Caenorhabditis species (C. elegans, C. brenneri, C. briggsae, C. remanei, C. japonica and C. angaria) and some EST sequences for C. sp. 5 . That is, from candidate genes with unambiguous alignments among 1:1 orthologs, we chose those which provided some resolution in a phylogenetic analysis with the above six species (data not shown). Such genes promised to have sufficient nucleotide variation to resolve relationships between closely related Caenorhabditis species. Degenerate primers were designed to regions conserved in these Caenorhabditis species and in Pristionchus pacificus (see Additional File 2). SSU and LSU rDNA was amplified from worm lysates as described previously . Sequences of protein coding genes were amplified from cDNA as follows. Total RNA was isolated from mixed stage worms with the Qiagen RNeasy Mini kit following the tissue protocol or with Trizol. RT PCR was performed either with the Qiagen OneStep RT PCR kit using specific primers, or by first-strand cDNA synthesis (with the Transcriptor High Fidelity cDNA Synthesis kit (Roche) using anchored oligo-dT primers, or with the Protoscript kit by New England Biolabs, or with the Invitrogen Superscript III kit using random primers), followed by PCR with gene-specific primers. PCR products were purified and sequenced through Agencourt, or in-house with the Wizard VC Gel and PCR clean-up System (Promega) and the ABI BigDye Terminator v3.1 cycle sequencing kit (Applied Biosystems). Electrophoresis was performed on an ABI 3730 DNA Analyzer. Some PCR products were cleaned using the Zymoclean Gel DNA Recovery kit and sequenced at the University of Utah core facility. Sequences were assembled with Sequencher (Gene Codes) and deposited at GenBank under the accession numbers listed in Additional File 3. Alignments for protein coding genes were generated with ClustalX  and were in some cases manually improved by aligning the amino acid sequence. The rRNA sequences were also aligned with ClustalX. These alignments were largely unambiguous.
As barcodes for quick and easy identification of Caenorhabditis species, we explored the ITS regions between SSU rRNA and LSU rRNA. It proved easiest to amplify ITS1 and ITS2 separately with one primer in the highly conserved 5.8S region in each case (Additional File 2). Sequences were generated as described above. The sequences, which contained highly conserved anchor regions in the rRNA genes, were aligned using ClustalX or Muscle, which is optimized for aligning sequences with highly diverged segments, such as introns and intergenic sequences .
The data file used for phylogenetic analyses was a concatenated alignment of the eleven gene segments listed in the previous section (excluding ITS sequences, Additional File 4). Although some of these sequences were missing for some taxa, the full dataset was used for all phylogenetic analyses discussed here. There has been some controversy regarding the treatment of datasets with missing data in phylogenetic analyses [23–25]. Many simulations and tests with empirical data have demonstrated that using datasets in which some taxa are missing even large amounts of data do not generally suffer ill effects. Instead, better accuracy and resolution are generally obtained if characters with missing data are included than if they are excluded . Consistent with these results, we also found that excluding characters with missing data resulted in poorer resolution (but not significantly different topologies) than if all characters were included (data not presented here). We thus only report our analyses with the full dataset.
To test the data for robustness to method of phylogenetic inference, we compared the results from analyses with weighted maximum parsimony (wMP), maximum likelihood (ML) and Bayesian inference (BI). Robustness of the data to character representation was tested using bootstrap and jackknife analyses.
The wMP analysis was performed with PAUP* ver4b10 . A transversion was weighted twice a transition as in previous analyses of this taxon . A jackknife analysis was performed with 1000 replicates and two addition sequence replicates in each round.
The ML analysis (a 100-replicate bootstrap and a thorough heuristic search) was run with RAxML ver. 7.2.8 ("BlackBox" version) via the CIPRES Science Gateway on the TeraGrid of NSF [27–30]. A six-parameter substitution model was used with a gamma correction for rate differences across sites (using 25 discrete categories of sites) and a correction for unvarying sites (GTR+Γ+I). Parameters were estimated from the data. The shape parameter for the gamma distribution of rates was α = 0.44081. Estimated proportions of nucleotides were: π(A) = 0.264, π(C) = 0.217, π(G) = 0.260, π(T) = 0.259. Estimated rates for the GTR model were: f(AC) = 1.390, f(AG) = 3.120, f(AT) = 1.182, f(CG) = 1.115, f(CT) = 5.790, relative to f(GT) = 1.000.
Another analysis of the same dataset was performed using Bayesian Inference (BI) as implemented in MrBayes ver. 3.1.2 [31, 32] via the CIPRES portal [27, 28, 33]. A six-parameter substitution model was used with a gamma correction for rate differences across sites and an estimate for the proportion of invariant sites (GTR+Γ+I). The analysis was stopped automatically by MrBayes at 4,055,000 generations (due to convergence of all parameters). Trees and parameters were sampled every 1000 generations for a total of 1,556 samples. Burnin was set to 50% of the samples to calculate the clade credibility values (posterior probabilities) and to estimate the model parameters, which were: π(A) = 0.265, π(C) = 0.217, π(G) = 0.256, π(T) = 0.262. Estimated rates for the GTR model were: f(AC) = 1.296, f(AG) = 2.284, f(AT) = 1.108, f(CG) = 1.147, f(CT) = 3.678, relative to f(GT) = 1.000. In the final tree, only one branch had a clade credibility value less than 100 (i.e., a branch that placed C. sp. 20 with C. angaria, C. drosophilae, C, sp. 2, 8 and 12, exclusive of C. sp. 6 and 13).
To estimate the genetic divergence within Caenorhabditis (and for comparisons with other taxa), we calculated the amount of nucleotide change along the branches of the phylogeny using maximum likelihood implemented in PAUP* and the sequences of RNAP2. This gene was used because previous results showed that the rRNA genes--but not RNAP2--display significant heterotachy , and RNAP2 was the only protein-coding gene that we could sequence for all species. All parameters for a general time-reversible model were estimated from the data. For comparison, we also calculated branch lengths for an RNAP2 dataset from 12 Drosophila species with the topology from . The data matrix is found in Additional File 5.
To test the utility of ITS2 sequences (i.e. the intergenic region between 5.8S and LSU rRNA genes) for distinguishing which isolates belong to which Caenorhabditis species, we sequenced this region for several strains of the species that were isolated more than once. The sequences, which contained the highly conserved 5.8S rDNA at the 5' end and part of the LSU rDNA at the 3' end, were aligned with ClustalX and then trimmed down to the ITS2 sequence only, following the annotation of the rRNA gene structure of C. elegans ). This data matrix is presented in Additional File 6. This alignment was used to determine the pairwise differences between species and strains. To represent these differences graphically, we calculated the branch length of a tree for all strains. This tree was reconstructed based on the ITS2 sequences with MP, using the species phylogeny as a constraint. A heuristic search yielded eight most parsimonious trees, one of which was chosen for further analysis. Branch lengths of this tree were determined by parsimony and include indels (one change per gap, regardless of the length of the gap). Changes with ambiguous branch assignment were optimized with ACCTRAN. ACCTRAN was used to offset somewhat the underestimation by parsimony of changes occurring in deeper branches. Correcting for superimposed changes demonstrates even greater discernibility between intra- and interspecific differences, suggesting that the parsimony approach is conservative. Pairwise intraspecific differences were manually tabulated as transitions, transversions, and indels.
Results and discussion
Sampling of rotting plant parts yielded many Caenorhabditisisolates including 16 new species
By sampling rotting fruits, flowers and stems in various temperate and tropical regions of the world, several hundred cultures of different Caenorhabditis species were established. Crosses with established cultures of known species revealed that many of the new isolates belonged to four already described species, namely C. elegans, C. briggsae, C. remanei and C. brenneri (Additional File 7).
New Caenorhabditis species, region where they were isolated and reproductive mode
Mode of reproduction
Some of the new species were sampled many times, whereas some others are currently represented by single isolates (Additional File 8). For the species sampled many times, we could not find any clear substrate preference. For example, C. sp. 11 was found in rotting flowers, fruits and banana pseudo-stems, like C. briggsae. All of these habitats, however, are rich in nutrients, bacteria and likely yeasts, and may provide similar conditions as habitats for the species.
The phylogeny supports the following grouping: C. sp. 1 branches off first and the second branch is C. plicata. This branching pattern is the same as in a previous analysis with only 11 Caenorhabditis species but with 54 species outside of Caenorhabditis , suggesting that the choice of P. pacificus as the outgroup representative had no effect on the overall tree topology. The remaining Caenorhabditis species fall into two monophyletic groups, the Elegans super-group and the Drosophilae super-group (Figure 2). Within the Elegans super-group, we find two subclades, which we call the Japonica group and the Elegans group. The Japonica group consists of C. spp. 7, 14, 17-19 and C. japonica. The monophyly of this group is well supported by likelihood analyses and less so by wMP. C. sp. 15 appears to be the sister species of the Japonica group. The Elegans group comprises the remaining Elegans super-group species. Their relationships are highly supported in all analyses. C. elegans forms the first branch of this group. The other Elegans group species fall in two clades, one comprising C. briggsae, C. remanei and C. spp. 5 and 9, the other one comprising C. brenneri and C. spp. 10, 11 and 16. None of the 26 Caenorhabditis species in this analysis is the sister species of C. elegans. Within the Drosophilae super-group, C. drosophilae and C. sp. 2 form the Drosophilae group which is the sister taxon to the highly supported Angaria group composed of C. angaria plus C. sp. 12 and C. sp. 8. Of the other three species in the Drosophilae super-group, C. sp. 6 and 13 are sister species, and C. sp. 20 possibly forms the first branch.
The temperature preferences of the species correlate with the latitude of their geographic distribution. Indeed, different from C. elegans, C. briggsae and C. remanei, the tropical species C. spp. 7, 9, 10, 11 do not grow at 15°C. In contrast, C. spp. 9, 10 and 11 can grow at 30°C (but not 33°C), a character they share with C. briggsae but not with C. elegans [; M. Ailion, unpublished observations].
Several species often co-occur in the same location and sometimes in the same individual fruit. For example, C. briggsae, C. brenneri and C. sp. 10 were all found in the same garden in Kanjirapally, Kerala, India; C. elegans, C. briggsae and C. sp. 13 co-occurred in rotting apples in an orchard in Orsay, France; C. briggsae, C. brenneri and C. sp. 11 were found together in one garden in La Réunion; C. briggsae and C. sp. 14 were isolated from the same chestnut in Moorea; C. elegans and C. sp. 6 were found in rotting apples from the same tree in Amares, Portugal; C. briggsae and C. sp. 8 were found in the same rotting persimmon in New York; and C. briggsae and C. sp. 15 were found in the same small sample of rotting flowers in Kauai, Hawaii.
To determine whether any evolutionary pattern for phenotypic characters can be discerned, we mapped several such characters onto our phylogeny (Additional File 10). Most informative morphological characters in rhabditids are associated with the male reproductive organs [12, 38]. We therefore analyzed the 26 species for the shapes of their spicules and of their male tail with its fan and sensory rays. The evolution of reproductive modes is also of particular interest in Caenorhabditis.
Caenorhabditis species have one of two modes of reproduction. They can be gonochoristic (male-female), like C. remanei and C. brenneri, or they can be androdioecious with selfing hermaphrodites and facultative males, like C. elegans and C. briggsae. Of the new species, only C. sp. 11 presents the selfing mode of reproduction (Table 1). Tracing this character along the branches of the phylogenetic tree of Caenorhabditis reveals that hermaphroditism likely evolved independently in each of the three lineages (Figure 2). Assuming that evolution of hermaphroditism and gonochorism are equally likely, this scenario requires three evolutionary steps, whereas the alternative hypothesis that hermaphroditism evolved once requires six evolutionary steps (one gain of hermaphroditism and five reversals to gonochorism). Recent studies [39–42] have discovered multiple differences in the genetic underpinnings of sex determination in C. elegans and C. briggsae, supporting the hypothesis that hermaphroditism evolved convergently in these two species.
Morphology and other phenotypic characters
Another conflict between the molecular data (supporting the phylogeny presented here) and a morphological character concerns the position of one of the three pairs of rays which are attached to the dorsal surface of the fan. In all species of the Drosophilae super-group, the three dorsal rays are in positions 1, 4 and 7, counted from anterior, whereas in all species of the Elegans super-group the dorsal rays are in position 1, 5 and 7 (Figure 6). Thus, the middle dorsal ray is in a different position. A middle dorsal ray in position 5 is also found in C. sp. 1 and in the Caenorhabditis sister group (the Protorhabditis group, see  for characters of the stem species), but not in C. plicata. Since the position of C. sp. 1, C. plicata, the Drosophilae super-group and the Elegans super-group are well supported, the distribution of this character requires at least two evolutionary steps, either two shifts of the middle dorsal ray from position 5 to 4 in the Drosophilae super-group and in C. plicata, or one shift from position 5 to 4 after the branch to C. sp. 1 and a reversal to the ancestral situation (middle dorsal ray in pos. 5) in the Elegans super-group. Likewise, the occurrence of a particularly short ray 4 is homoplasious, as it is found in C. japonica, C. sp. 14, 17 and 19. The distribution of this character requires at least three evolutionary steps.
The ability to start an RNA interference response upon external administration of double-stranded actin RNAs was tested on all new species (Nuez and Félix, submitted). This character has a complex distribution. Competence to external RNAi is absent in all species of the sister clade of C. elegans, in the clade comprising C. spp. 17, 18 and 19, in C. plicata and in all species of the Angaria and Drosophilae group but not in C. sp. 6, C. sp. 13 and C. sp. 20 (Nuez and Félix, submitted). Mapping of this character onto the phylogeny suggests that competence to respond to external dsRNA may have been present in the Caenorhabditis stem species and was lost four times independently.
There are also characters which are mostly unambiguously distributed and thus constitute reliable apomorphic characters for certain monophyletic groups that are supported by molecular data: In most species of the Elegans super-group, the distal end of the fan is notched. This terminal notch is particularly large in the monophyletic group consisting of C. sp. 17, 18 and 19. The Elegans super-group is characterized by a small space between the second and third ray (smaller than the space between the third and fourth ray or of similar size), when compared to the other Caenorhabditis species in which this space is always bigger (Figure 6). A particularly narrow fan is found in C. angaria, C. sp. 8 and C. sp. 12. Also, these species share papilliform phasmids, a very similar spicule shape, and a spiral mating position in which the male is coiled around the female (Figures 5 and 6). In addition, C. angaria, C. sp. 8 and C. sp. 12 all have a short stoma with a bifurcated projection at each sector of the metastegostom (see illustrations in  for C. angaria). Thus, the monophyly of the Angaria group is exceptionally well supported by morphological and molecular characters. A number of previously mentioned phenotypic characters support the sister group relationships of the Angaria and Drosophilae groups even though their distribution across all Caenorhabditis species is homoplastic: In all species of both clades the spicule is short, the fan is open and lacks a serrated edge, and all species are insensitive to externally administered double-stranded RNA (against actin). A sister group relationship of these two clades has been proposed earlier , based on molecular data and on the presence of one semicircular flap on each lip seen on SEM images of C. angaria and C. drosophilae. The presence of this flap needs to be confirmed for the other species of these groups, but could constitute a further apomorphy. The geographic distribution of their members suggest that the Angaria and Drosophilae groups originated in the New World.
ITS2 is a suitable barcode for distinguishing Caenorhabditisspecies
Morphology can be used to assign species to the major groups within Caenorhabditis, but some species within these groups look very similar or entirely alike. In fact, the genus contains a host of morphological sibling species. Therefore, morphology alone is not suitable for identifying new species. In this study, species were initially identified via mating tests (Additional File 1). However, with a growing number of species, mating tests become tedious and time-consuming. Unlike morphology, genomic sequences contain many easily accessible species-specific differences. We thus sought a genetic barcode for Caenorhabditis species.
As suitable targets for barcoding in nematodes, De Ley et al.  proposed the use of SSU and LSU rDNA sequences. However, within Caenorhabditis, SSU rDNA is very highly conserved and can be identical in closely related species (e.g. in C. angaria and C. sp. 12). LSU rDNA, specifically the D2D3 region, is usually more variable, but C. briggsae and C. sp. 9 differ in only five positions over the entire LSU locus (three substitutions and two indels). Such a small number of differences can be easily concealed by sequencing errors. Here, we explore the ITS region instead. Both internal transcribed spacers are variable between species. In more distantly related species, the unambiguously alignable portions of the ITS regions are very short. However, the flanking rRNA gene sequences are highly conserved. Thus, PCR with primers in the flanking sequences reliably amplifies the ITS regions. The flanking sequences can also serve as anchors for alignments.
The ITS1 region was often variable within a species and even within one animal, making direct sequencing of PCR products problematic (data not shown). ITS2 was less polymorphic in the strains tested, although here, too, we found four strains with indel polymorphisms that precluded sequencing of the entire region without cloning. Nevertheless, the parts of ITS2 that could be sequenced directly from PCR products were long enough to enable species identification.
Caenorhabditisare "fruit worms", not soil nematodes
Our study has shown beyond doubt that the best place to find Caenorhabditis species is rotten fruit. Of the 26 Caenorhabditis species currently in culture, all but seven have been isolated from rotten fruit, although most species have also been found in other rotting plant material. We do not yet know whether any Caenorhabditis species is a strict fruit specialist, but it is likely that C. japonica is specific for the fruit of Schoepfia jasminodera . The information which we have gathered to date shows that the habitats of these nematodes are strikingly similar to those of Drosophila species. Many Drosophila species, including D. melanogaster, can be found in various fermenting fruit and fungi, but there are also specialists for particular fruit like figs or those that specialize on breeding in flowers . Furthermore, some Drosophila species are specialists for rotten cactus, as are C. drosophilae and C. sp. 2 , and others breed in fungi, a habitat from which C. sp. 1 and C. auriculariae have been isolated . Further samplings of Caenorhabditis species may discover more parallels between the habitats of these genera. It should be noted that at least two Caenorhabditis species live in a habitat quite different from rotten fruit, namely decomposing animal tissue. C. plicata has only been found in carrion and C. bovis (not included here for the lack of material for molecular analyses) lives in the inflamed ears of cattle . Taking all of our sampling data together, it is clear that Caenorhabditis are not soil nematodes. The only stage which is occasionally found in soil is the dauer larva. In this respect, Caenorhabditis species do not differ much from the majority of rhabditid nematode species which also reproduce in substrates rich in nutrients and bacteria. Nematodes that specialize in such habitats often use other animals for dispersal. Such phoresy is indeed an essential part of the life cycle of C. angaria, C. drosophilae and C. japonica where the dauer larvae attach to weevils, drosophilid flies or burrower bugs, respectively [44, 46, 48]. Other species, including C. elegans and C. briggsae, have been isolated from phoretic carriers  and it is likely that phoretic relationships exist for many or all other Caenorhabditis species.
How many Caenorhabditisspecies are there?
Remarks on the geographic range
Our new records corroborate the observation that of all Caenorhabditis species, only C. briggsae and C. elegans are cosmopolitan (Figure 4 and Additional File 12). However, only C. briggsae seems to be equally common in temperate and tropical regions. C. elegans was recorded from tropical Africa at high altitude sites (2000 m and above) in Limuru, Kenya  and Addis Ababa, Ethiopia (data courtesy of Dee Denver) and on islands such as La Réunion (at altitude 1100 m) or Hawaii (at an unknown altitude). Two other species have been found in temperate as well as in tropical regions: C. plicata was found in Germany and Kenya and C. sp. 5 occurs in temperate and tropical China and in Vietnam. The remaining Caenorhabditis species for which we have more than one record were either isolated from tropical or from temperate regions but not from both. Of the 16 Elegans super-group species in culture, ten were found exclusively in the tropics and only two (C. remanei and C. japonica) are exclusively temperate. In contrast, only three of the ten non-Elegans super-group species are only found in the tropics. This suggests a radiation of Elegans super-group species occurred in tropical regions. However, this picture can easily change in the future, since our knowledge of Caenorhabditis biogeography is still very sketchy. The southern hemisphere has been only sparsely sampled and there is almost no information about the distribution of Caenorhabditis species in southern temperate regions.
Phylogeny and character evolution
In this study, we used the sequences of 11 genes to reconstruct the phylogeny of Caenorhabditis species. Our analyses yielded a tree with some very well-resolved areas, but there are also branches which are weakly supported. More sequence information is needed to resolve all branches with high confidence. However, the current phylogeny is a usable working hypothesis which allows us to draw some sound conclusions: (1) Caenorhabditis comprises two monophyletic sister groups, the Elegans super-group with 16 species currently in culture, and the Drosophilae super-group with eight species. (2) None of the species in the Elegans super-group is more closely related to C. elegans than C. brenneri, C. briggsae or C. remanei; i.e., we do not yet know of an extant sister species of C. elegans. (3) Morphologically, the Elegans super-group species are more uniform than the Drosophilae super-group species, making it particularly difficult to identify any of the individual Elegans super-group species by morphology alone. (4) With the exception of C. briggsae and C. sp. 9, all of these species are fully reproductively isolated and genetically divergent.
A closer look at a small number of morphological characters showed that almost none of them had an unambiguous distribution on our current phylogeny. For almost all morphological characters, conflicts existed between each other and/or the molecular data that were used to reconstruct the phylogeny. This observation matches previous findings for character evolution in rhabditid nematodes which showed that homoplasy (convergent or parallel evolution) is a common theme in this group [2, 50, 51]. Importantly, our current study confirms that hermaphroditism evolved convergently in C. briggsae and C. elegans. Furthermore, we found that hermaphroditism evolved a third time in one of the new Caenorhabditis species, C. sp. 11.
New Caenorhabditisspecies as a resource for future studies
The new Caenorhabditis species provide a phylogenetic framework to study the evolution of a number of genomic and phenotypic traits. At the molecular level, previously analyzed species were very distant from each other , and nucleotide turnover at putatively neutral sites was saturated, preventing the application of several molecular evolution tests. The new species provide several cases of more closely related species pairs, especially the C. briggsae/C. sp. 9 [52, 53], C. drosophilae/C. sp. 2 and C. angaria/C. sp. 12 comparisons (Figure 2). In addition, the level of polymorphism within some of the new gonochoristic species is high, e.g. in C. sp. 5 where it is comparable to that of the ascidian Ciona savignyi . Intraspecific genome comparisons in such cases are likely to reveal which parts of the genome are less constrained than other parts. Active genome sequencing is presently ongoing for the new Caenorhabditis species (see http://www.nematodes.org/nematodegenomes), and data are already available for C. spp. 7, 9 and 11 from the Genome Center at Washington University and GenBank, and C. sp. 5 from Genepool at the University of Edinburgh.
Some species provide interesting phenotypic features. For instance, C. sp. 11 provides a third example--after C. elegans and C. briggsae--of independent evolution of hermaphroditism from a gonochoristic ancestor. C. sp. 9 and C. briggsae are the first species pair in Caenorhabditis with partially fertile progeny, providing a genetic entry into species isolation studies [52, 53]. These two species have different modes of reproduction (gonochoristic for C. sp. 9, hermaphroditic for C. briggsae), thus also allowing for genetic studies of reproductive mode evolution. Other species pairs, such as C. angaria and C. sp. 12, offer sterile hybrids of both sexes, and crosses of C. sp. 5 and C. briggsae yield sterile adult females with abnormal gonads (Additional File 1). Another example is sperm size, which is under selection in these nematodes [55, 56]. Hermaphrodites have smaller sperm than males , males in hermaphroditic species have smaller sperm than males in gonochoristic species , and as a surprising extreme, C. sp. 18 males produce giant sperm (Additional File 13). These new species considerably widen the spectrum of phenotypic evolution that can be studied using Caenorhabditis.
We are very grateful to all sample collectors and strain contributors listed in Additional Files 7 and 10, as well as all those who participated in our travels. We thank Asher Cutter and Erich Schwarz for making unpublished data available to us.
NSF grants DEB0922012 and IOB0643047 were awarded to DHAF. MAF, CB and JBP were supported by the CNRS which included a Nouragues Station grant (Program 2009) to MAF and CB. MA was supported by a Helen Hay Whitney Postdoctoral Fellowship and by NIH grant K99MH082109. MVR was supported by NIH R01 GM089972 and an Ellison Foundation New Scholar Award. KK was partially supported by a supplement to NIH NHGRI modENCODE grant 5U01HG004276-04 to Fabio Piano (New York University).
- Barrière A, Félix M-A: Natural variation and population genetics of C. elegans. The C. elegans Research Community. January 9, 2006, 10.1895/wormbook.1.37.1. [http://www.wormbook.org/]Google Scholar
- Kiontke K, Barrière A, Kolotuev I, Podbilewicz B, Sommer RJ, HFitch DHA, Félix M-A: Trends, stasis and drift in the evolution of nematode vulva development. Curr Biol. 2007, 17: 1925-1937. 10.1016/j.cub.2007.10.061.View ArticlePubMedGoogle Scholar
- Kiontke K, Gavin NP, Raynes Y, Roehrig C, Piano F, Fitch DH: Caenorhabditis phylogeny predicts convergence of hermaphroditism and extensive intron loss. Proc Nat Acad Sci USA. 2004, 101 (24): 9003-9008. 10.1073/pnas.0403094101.View ArticlePubMedPubMed CentralGoogle Scholar
- Carvalho S, Barrière A, Pires da Silva A: The world of a worm: a framework for Caenorhabditis evolution. EMBO Reports. 2006, 7: 981-983. 10.1038/sj.embor.7400798.View ArticlePubMedPubMed CentralGoogle Scholar
- Cutter AD, Dey A, Murray RL: Evolution of the Caenorhabditis elegans genome. Mol Biol Evol. 2009, 26: 1199-1234. 10.1093/molbev/msp048.View ArticlePubMedGoogle Scholar
- Delattre M, Félix M-A: Microevolutionary studies in nematodes: a beginning. Bioessays. 2001, 23: 807-819. 10.1002/bies.1116.View ArticlePubMedGoogle Scholar
- Haag ES, Chamberlin H, Coghlan A, Fitch DH, Peters AD, Schulenburg H: Caenorhabditis evolution: if they all look alike, you aren't looking hard enough. Trends Genet. 2007, 23 (3): 101-104. 10.1016/j.tig.2007.01.002.View ArticlePubMedGoogle Scholar
- Kammenga JE, Phillips PC, de Bono M, Doroszuk A: Beyond induced mutants: using worms to study natural variation in genetic pathways. Trends Genet. 2008, 24: 178-185. 10.1016/j.tig.2008.01.001.View ArticlePubMedGoogle Scholar
- Phillips PC: One perfect worm. Trends Genet. 2006, 22 (8): 405-407. 10.1016/j.tig.2006.06.001.View ArticlePubMedGoogle Scholar
- Markow TA, O'Grady PM: Drosophila biology in the genomic age. Genetics. 2007, 177 (3): 1269-1276. 10.1534/genetics.107.074112.View ArticlePubMedPubMed CentralGoogle Scholar
- Kiontke K, Sudhaus W: Ecology of Caenorhabditis species. The C. elegans Research Community. January 9, 2006, 10.1895/wormbook.1.115.1. [http://www.wormbook.org]Google Scholar
- Sudhaus W, Kiontke K: Phylogeny of Rhabditis subgenus Caenorhabditis (Rhabditidae, Nematoda). J Zool System Evol Res. 1996, 34: 217-233.View ArticleGoogle Scholar
- Cutter AD: Divergence times in Caenorhabditis and Drosophila, inferred from direct estimates of the neutral mutation rate. Mol Biol Evol. 2008, 25: 778-786. 10.1093/molbev/msn024.View ArticlePubMedGoogle Scholar
- Stark A, Lin MF, Kheradpour P, Pedersen JS, Parts L, Carlson JW, Crosby MA, Rasmussen MD, Roy S, Deoras AN, et al: Discovery of functional elements in 12 Drosophila genomes using evolutionary signatures. Nature. 2007, 450 (7167): 219-232. 10.1038/nature06340.View ArticlePubMedPubMed CentralGoogle Scholar
- Baille D, Barrière A, Félix M-A: Oscheius tipulae, a widespread hermaphroditic soil nematode, displays a higher genetic diversity and geographical structure than Caenorhabditis elegans. Mol Ecol. 2008, 17: 1523-1534. 10.1111/j.1365-294X.2008.03697.x.View ArticlePubMedGoogle Scholar
- Felix MA, Braendle C: The natural history of Caenorhabditis elegans. Curr Biol. 2010, 20 (22): R965-969. 10.1016/j.cub.2010.09.050.View ArticlePubMedGoogle Scholar
- Barrière A, Félix M-A: High local genetic diversity and low outcrossing rate in Caenorhabditis elegans natural populations. Curr Biol. 2005, 15: 1176-1184. 10.1016/j.cub.2005.06.022.View ArticlePubMedGoogle Scholar
- Barrière A, Félix M-A: Isolation of C. elegans and related nematodes. The C. elegans Research Community. Janurary 9, 2006, 10.1895/wormbook.1.115.1. [http://www.wormbook.org/]Google Scholar
- Nigon V, Dougherty EC: Reproductive patterns and attempts at reciprocal crossing of Rhabditis elegans Maupas, 1900 and Rhabditis briggsae Dougherty and Nigon, 1949 (Nematoda: Rhabditida). J Exp Zool. 1949, 112 (3): 485-503. 10.1002/jez.1401120307.View ArticlePubMedGoogle Scholar
- Cutter AD, Wasmuth JD, Washington NL: Patterns of molecular evolution in Caenorhabditis preclude ancient origins of selfing. Genetics. 2008, 178 (4): 2093-2104. 10.1534/genetics.107.085787.View ArticlePubMedPubMed CentralGoogle Scholar
- Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F, Higgins DG: The CLUSTAL_X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res. 1997, 25 (24): 4876-4882. 10.1093/nar/25.24.4876.View ArticlePubMedPubMed CentralGoogle Scholar
- Edgar RC: MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 2004, 32 (5): 1792-1797. 10.1093/nar/gkh340.View ArticlePubMedPubMed CentralGoogle Scholar
- Lemmon AR, Brown JM, Stanger-Hall K, Lemmon EM: The effect of ambiguous data on phylogenetic estimates obtained by maximum likelihood and Bayesian inference. Syst Biol. 2009, 58 (1): 130-145. 10.1093/sysbio/syp017.View ArticlePubMedGoogle Scholar
- Wiens JJ: Missing data and the design of phylogenetic analyses. J Biomed Inform. 2006, 39 (1): 34-42. 10.1016/j.jbi.2005.04.001.View ArticlePubMedGoogle Scholar
- Wiens JJ, Morrill MC: Missing data in phylogenetic analysis: reconciling results from simulations and empirical data. Syst Biol. 2011, 60 (5): 719-731. 10.1093/sysbio/syr025.View ArticlePubMedGoogle Scholar
- Swofford DL: Phylogenetic Analysis Using Parsimony (*and Other Methods). Version 4.0. 2002, Sunderland, Massachusetts: Sinauer AssociatesGoogle Scholar
- Miller MA, Pfeiffer W, Schwartz T: Creating the CIPRES science gateway for inference of large phylogenetic trees. Proceedings of the Gateway Computing Environments Workshop (GCE). 2010, New Orleans, 1-8.View ArticleGoogle Scholar
- Pfeiffer W, Stamatakis A: Hybrid MPI/Pthreads parallelization of the RAxML phylogenetics code. Ninth IEEE International Workshop on High Performance Computational Biology (HiCOMB 2010): 2010; Atlanta. 2010Google Scholar
- Shannon AJ, Tyson T, Dix I, Boyd J, Burnell AM: Systemic RNAi mediated gene silencing in the anhydrobiotic nematode Panagrolaimus superbus. BMC Mol Biol. 2008, 9: 58-10.1186/1471-2199-9-58.View ArticlePubMedPubMed CentralGoogle Scholar
- Stamatakis A: RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics. 2006, 22 (21): 2688-2690. 10.1093/bioinformatics/btl446.View ArticlePubMedGoogle Scholar
- Huelsenbeck JP, Ronquist F: MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics. 2001, 17 (8): 754-755. 10.1093/bioinformatics/17.8.754.View ArticlePubMedGoogle Scholar
- Ronquist F, Huelsenbeck JP: MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics. 2003, 19 (12): 1572-1574. 10.1093/bioinformatics/btg180.View ArticlePubMedGoogle Scholar
- Pratas W, Trancoso P, Stamatakis A, Sousa L: Fine-grain parallelism using Multi-core, Cell/BE, and GPU systems: Accelerating the phylogenetic likelihood function. Proceedings of the ICPP: 2009; Vienna. 2009Google Scholar
- Ellis RE, Sulston JE, Coulson AR: The rDNA of C. elegans: sequence and structure. Nucleic Acids Res. 1986, 14 (5): 2345-2364. 10.1093/nar/14.5.2345.View ArticlePubMedPubMed CentralGoogle Scholar
- Barriere A, Felix MA: High local genetic diversity and low outcrossing rate in Caenorhabditis elegans natural populations. Curr Biol. 2005, 15 (13): 1176-1184. 10.1016/j.cub.2005.06.022.View ArticlePubMedGoogle Scholar
- Sudhaus W, Kiontke K: Comparison of the cryptic nematode species Caenorhabditis brenneri sp. n. and C. remanei (Nematoda: Rhabditidae) with the stem species pattern of the Caenorhabditis Elegans group. Zootoxa. 2007, 1456: 45-62.Google Scholar
- Prasad A, Croydon-Sugarman MJ, Murray RL, Cutter AD: Temperature-dependent fecundity associates with latitude in Caenorhabditis briggsae. Evolution. 2010, 65 (1): 52-63.View ArticlePubMedGoogle Scholar
- Sudhaus W, Fitch DHA: Comparative studies on the phylogeny and systematics of the Rhabditidae (Nematoda). J Nematol. 2001, 33 (1): 1-70.PubMedPubMed CentralGoogle Scholar
- Guo Y, Lang S, Ellis RE: Independent recruitment of F box genes to regulate hermaphrodite development during nematode evolution. Curr Biol. 2009, 19 (21): 1853-1860. 10.1016/j.cub.2009.09.042.View ArticlePubMedGoogle Scholar
- Hill RC, de Carvalho CE, Salogiannis J, Schlager B, Pilgrim D, Haag ES: Genetic flexibility in the convergent evolution of hermaphroditism in Caenorhabditis nematodes. Dev Cell. 2006, 10 (4): 531-538. 10.1016/j.devcel.2006.02.002.View ArticlePubMedGoogle Scholar
- Hill RC, Haag ES: A sensitized genetic background reveals evolution near the terminus of the Caenorhabditis germline sex determination pathway. Evol Dev. 2009, 11 (4): 333-342. 10.1111/j.1525-142X.2009.00340.x.View ArticlePubMedPubMed CentralGoogle Scholar
- Nayak S, Goree J, Schedl T: fog-2 and the evolution of self-fertile hermaphroditism in Caenorhabditis. PLoS Biology. 2005, 3 (1): e6-10.1371/journal.pbio.0030006.View ArticlePubMedPubMed CentralGoogle Scholar
- Pénigault J-B, Félix MA: Evolution of a system sensitive to stochastic noise: P3.p cell fate in Caenorhabditis. Dev Biol. 2011, 357: 419-427. 10.1016/j.ydbio.2011.05.675.View ArticlePubMedGoogle Scholar
- Sudhaus W, Kiontke K, Giblin-Davis RM: Description of Caenorhabditis angaria n. sp. (Nematoda: Rhabditida), an associate of sugarcane and palm weevils (Coleoptera: Curculionidae). Nematology. 2010Google Scholar
- De Ley P, De Ley IT, Morris K, Abebe E, Mundo-Ocampo M, Yoder M, Heras J, Waumann D, Rocha-Olivares A, Jay Burr AH, et al: An integrated approach to fast and informative morphological vouchering of nematodes for applications in molecular barcoding. Philos T Roy Soc Lon B. 2005, 360 (1462): 1945-1958. 10.1098/rstb.2005.1726.View ArticleGoogle Scholar
- Kiontke K, Hironaka M, Sudhaus W: Description of Caenorhabditis japonica n. sp. (Nematoda: Rhabditida) associated with the burrower bug Parastrachia japonensis (Heteroptera: Cydnidae) in Japan. Nematology. 2002, 4: 933-941. 10.1163/156854102321122557.View ArticleGoogle Scholar
- Markow TA, O'Grady PM: Evolutionary genetics of reproductive behavior in Drosophila: connecting the dots. Ann Rev Genetics. 2005, 39: 263-291. 10.1146/annurev.genet.39.073003.112454.View ArticleGoogle Scholar
- Kiontke K: Description of Rhabditis (Caenorhabditis) drosophilae n. sp. and R. (C.) sonorae n. sp. (Nematoda: Rhabditida) from saguaro cactus rot in Arizona. Fundam appl Nematol. 1997, 20: 305-315.Google Scholar
- Dolgin ES, Felix MA, Cutter AD: Hakuna Nematoda: genetic and phenotypic diversity in African isolates of Caenorhabditis elegans and C. briggsae. Heredity. 2008, 100 (3): 304-315. 10.1038/sj.hdy.6801079.View ArticlePubMedGoogle Scholar
- Brauchle M, Kiontke K, MacMenamin P, Fitch DH, Piano F: Evolution of early embryogenesis in rhabditid nematodes. Dev Biol. 2009, 335 (1): 253-262. 10.1016/j.ydbio.2009.07.033.View ArticlePubMedPubMed CentralGoogle Scholar
- Kiontke K, Fitch DHA: The phylogenetic relationships of Caenorhabditis and other rhabditids. The C. elegans Research Community. January 9, 2006, 10.1895/wormbook.1.11.1. [http://www.wormbook.org]Google Scholar
- Cutter AD, Yan W, Tsvetkov N, Sunil S, Felix MA: Molecular population genetics and phenotypic sensitivity to ethanol for a globally diverse sample of the nematode Caenorhabditis briggsae. Mol Ecol. 2010, 19 (4): 798-809. 10.1111/j.1365-294X.2009.04491.x.View ArticlePubMedGoogle Scholar
- Woodruff GC, Eke O, Baird SE, Felix MA, Haag ES: Insights into species divergence and the evolution of hermaphroditism from fertile interspecies hybrids of Caenorhabditis nematodes. Genetics. 2010, 186 (3): 997-1012. 10.1534/genetics.110.120550.View ArticlePubMedPubMed CentralGoogle Scholar
- Wang GX, Ren S, Ren Y, Ai H, Cutter AD: Extremely high molecular diversity within the East Asian nematode Caenorhabditis sp. 5. Mol Ecol. 2010, 19 (22): 5022-5029. 10.1111/j.1365-294X.2010.04862.x.View ArticlePubMedGoogle Scholar
- LaMunyon CW, Ward S: Evolution of larger sperm in response to experimentally increased sperm competition in Caenorhabditis elegans. Proc Biol Sci. 2002, 269 (1496): 1125-1128. 10.1098/rspb.2002.1996.View ArticlePubMedPubMed CentralGoogle Scholar
- Murray RL, Cutter AD: Experimental Evolution of sperm count in protandrous self-fertilizing hermaphrodites. Journal of Experimental Biology. 2011Google Scholar
- LaMunyon CW, Ward S: Evolution of sperm size in nematodes: sperm competition favours larger sperm. Proceedings of the Royal Society of London - Series B: Biological Sciences. 1999, 266 (1416): 263-267. 10.1098/rspb.1999.0631.View ArticlePubMedPubMed CentralGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.