Molecular diversity and relationships of fig associated nematodes from South Africa

Nematodes of figs and fig wasps have received limited attention in Africa since their discovery in 1973. Sixteen of the 25 species of native South African figs were sampled for nematode associates using molecular barcoding with three loci (SSU, LSU D2-D3 and mtCOI) and fourteen (93%) were positive for at least one nematode species. Thirty-three putative species of nematodes were identified and classified according to the loci that were amplified and successfully sequenced. Fourteen putative nematode species were classified as Aphelenchoididae, of which nine were identified as Ficophagus from four species of Ficus from the section Galoglychia (i.e., five ex F. burkei including one shared with F. natalensis, one ex F. glumosa, one ex F. lutea, and one ex F. stuhlmannii) and one species ex F. sur from the section Sycomorus. In addition, there were four nematode species classified as Schistonchus s.s. from section Galoglychia figs (i.e., one ex F. burkei, two ex F. trichopoda, and one ex F. glumosa). There was also one species of Bursaphelenchus nematode recovered from F. sur from the section Sycomorus. Sixteen putative nematode species were classified as Diplogastridae, of which eight occurred in two clades of what is currently called Parasitodiplogaster with one (P. salicifoliae) being recovered from two Ficus species in the section Urostigma (F. salicifolia and F. ingens) and seven diplogastrids being associated with six species of Ficus from the section Galoglychia (i.e., two ex F. burkei including P. sycophilon, one ex F. stuhlmannii, one ex F. burtt-davyi, one ex F. trichopoda, one ex F. abutilifolia and one ex F. sansibarica). Three Acrostichus spp., a Teratodiplogaster and a Pristionchus species were recovered from F. sur and two Teratodiplogaster spp. and Pristionchus sycomori were recovered from F. sycomorus with both Ficus species belonging to the subgenus and section Sycomorus. The identities of the previously described T. martini and Parasitodiplogaster doliostoma (= Pristionchus sp. 35) are discussed. Lastly, there was a panagrolaimid identified from F. petersii.


Introduction
Martin et al. [1] first reported about the amazing morphological diversity of nematodes inside figs from six native species of Ficus from Zimbabwe in 1973 and invited participation from the places open to the public. Endangered or protected species were not involved with the present study.

Nematode collection
Fig syconia in various phases (phases B-E) were collected from 25 Ficus species from many locations across South Africa (Fig 1, Table 1, S1 Table in [20]. Because nematodes are vectored into the developing fig only after the syconia becomes receptive at phase B (female), phase A (pre-receptive or pre-female) figs were not used. In most cases, the ideal phases for nematode collection were early to late interfloral (phase C) or male (phase D) figs which were always used, if available. However, because fruit availability can be unpredictable, we collected phase B and E if it was the only fruit available. An average of 50 syconia were collected per crop and never less than 30. Collected figs were cut into ca 1-2 mm thick pieces and soaked in water for 10-15 min. Nematodes recovered from the sliced figs were hand-picked, identified to their taxonomic group at genus or family level at low magnification under a dissecting microscope according to body and stomatal/stylet shapes, and 3-5 individuals for each morphotypes were placed into either 100% ethanol or nematode digestion buffer (NDB) [21,22], but less than that if there were insufficient specimens recovered. The rest of the individuals were killed by heat and fixed in formalin and processed into glycerin [23] as morphological vouchers. In addition, some individuals isolated from F. sur and F. sycomorus were transferred to DESS [24] to eventually compare their typological characters and molecular phylogenetic status for single nematodes.

Molecular profiles and phylogeny
The nematodes transferred to 100% ethanol or NDB were used for molecular samples to survey the diversity of nematode genotypes associated with different fig species from different locations, i.e., DNA was extracted from ethanol-fixed material using a QIAamp 1 DNA Micro Kit (Qiagen, Germany) and NDB-treated samples were digested under 55˚C for 20 min. Whereas the DESS-fixed materials were re-hydrated in a drop of sterile water, photo-documented (see below for methods), transferred to NDB and digested for use as template DNA. We attempted to amplify and sequence three genetic loci, i.e., nearly full-length small subunit (SSU) and D2-D3 expansion segments of the large subunit (LSU) of ribosomal RNA genes and a partial region of the mitochondrial cytochrome oxidase subunit I (mtCOI) for all DNA template samples. Methodologies of Ye et al. [25] and Kanzaki and Futai [26] were employed for sequencing ribosomal RNAs and mtCOI. Briefly, amplification of SSU, LSU and mtCOI were attempted with primer sets, SSUF07 (5'-AAA GAT TAA GCC ATG CAT G-3') and SSUnR (5'-TTA CGACTT TTG CCC GGT TC-3'), D2a (5'-ACA AGT ACC GTG AGG GAA AGT TG-3') and D3b (5'-TCG GAA GGA ACC AGC TAC TA-3'), and COIF1 (5'-CCT ACT ATG ATT GGT GGT TTT GGT AAT-3') and COIR2 (5'-GTA GCA GTA AAA TAA GCA CG-3'), respectively. Upon successful amplification, sequences were determined with Sanger sequencing using the purified amplicons. Sequences with ambiguous or poor quality electropherogram/chromatograms were omitted. The sequences were compared with those deposited in the GenBank database using BLASTN program (https://blast.ncbi.nlm.nih.gov/Blast.cgi?PROGRAM=blastn&PAGE_TYPE= BlastSearch&LINK_LOC=blasthome) for initial identification, and the generic status of all compared sequences were confirmed by their phylogenetic status. High quality sequences are summarized in Table 1 and S1 Table in S2 File. Because two groups (= families), Aphelenchoididae and Diplogastridae were recognized, these families were phylogenetically analyzed separately for each locus using Bayesian Inference (BI) to determine their phylogenetic status within family (SSU and D2-D3) or other figassociates (mtCOI). Compared sequences and their GenBank accession numbers are shown in Figs 2-7. The compared sequences were selected according to previous family-wide analyses, i.e., Kanzaki et al. [27] and Davies et al. [3] for Aphelenchoididae and Wöhr et al. [17], Zeng et al. [28] and Gonzalez et al. [29] for Diplogastridae.
Compared sequences were aligned using the MAFFT program [30,31] (https://mafft.cbrc. jp/alignment/server/index.html) with the default settings. The base-substitution models for each gene were determined using the MEGA 7 software [32] and the Akaike information criterion. Combined Bayesian analysis was performed using the MrBayes 3.2 software [33,34]; four chains were run for 4 × 10 6 generations, and Markov chains were sampled at intervals of 100 generations [35]. Two independent runs were performed, and after confirming convergence and discarding the first 2 × 10 6 generations as 'burn in', the remaining topologies were used to generate a 50% majority-rule consensus tree.

Morphological observation
Several DESS-fixed materials were rehydrated and observed for their typological characters, particularly stomatal/stylet and pharyngeal morphologies and male and female tail characters which are usually used for diagnostic characters, prior to digestion. The materials were temporarily mounted on an agar pad, observed under a light microscope (Eclipse 80i, Nikon) with differential interference contrast (DIC) optics and photomicrographed using a microscope digital camera system (MC170 HD, Leica) connected to the microscope. The micrographs were edited using Photoshop Elements 3.0 (Adobe) for constructing micrographic figures (S1-S7 Figs in S1 File). Some formalin-fixed materials were also observed for identification of nematodes to the genus or family-level, and some typological characters were noted and illustrated using a camera-lucida drawing system (S8 Fig in S1 File).  Table 1 and S1 and S2 Tables in S2 File.

Nematode identification
Nematodes were identified based on their typological characters and phylogenetic status.
These are all considered to be new nematode species and were isolated from various sections/subsections of figs in South Africa, namely subgenus Urostigma and section Galoglychia Parasitodiplogaster spp. were recovered from various figs in the subgenus Urostigma and phylogenetic groupings were in accordance with specific sections and subsections of host figs in South Africa, i.e., P. salicifoliae was isolated only from section Urostigma and subsection  (Figs 2-4). This contrasts with Pristionchus, Teratodiplogaster and Acrostichus which were found exclusively from figs in the subgenus and section Sycomorus and each was inferred as being monophyletic within each genus (Figs 2-4). Because we observed only one species of Parasitodiplogaster (n. sp. 2) associated with sycones of F. burtt-davyi from South Africa it is quite possible that it is conspecific with the species studied for its fig and fig  wasp host dynamics by Jauharlina et al. [16].
The typological characters observed in DESS-and formalin-fixed materials are mostly in accordance with the phylogenetic status, i.e., the examined characters were basically identical to the generic characters or specific characters of nominal species, although some parts, e.g., male genital papillae, were sometimes not clear (S1-S8 Figs in S1 File).
In addition, a short fragment of LSU with ca 400 bp of an unidentified panagrolaimid species was recognized from F. petersii Warburg. However, further identification or phylogenetic analysis were not conducted for the nematode because the sequences closest to it were Macrolaimus spp. (Panagrolaimomorpha) with only 80% identity match. First, there was the issue of extremes in polyphenism of up to five morphotypes per genotype for a clade of, at that point, unknown subgenus Sycomorus fig associated Pristionchus species [8]. Within two Pristionchus spp. found in the present study, Pristionchus sp. 35 is considered conspecific to "Parasitodiplogaster" doliostoma, because although Pristionchus sp. 35 has not been formally described, it is phylogenetically close to P. borbonicus and P. sycomori, and the stomatal morphology of "P. doliostoma" is clearly similar to that of the type IV morph of these two Pristionchus spp. [8,15], i.e., "P. doliostoma" is hypothesized as one morphotype of Pristionchus sp. 35. The species will need to be transferred to Pristionchus, but a formal taxonomic revision will be presented elsewhere. This polymorphism helps explain the misclassification of "Parasitodiplogaster" doliostoma by Kanzaki et al. [15] from original preserved specimens from Martin et. al [1]. Data presented herein and in Susoy et al. [8] for Pristionchus sp. 35 (= Parasitodiplogaster cf. doliostoma) and P. sycomori demonstrates that about half of the nematode species diversity Martin et al. [1] observed in each Sycomorus fig (i.e., F. sur and F. sycomorus) was probably due to a single species of Pristionchus that can manifest in morphotypes as divergent as five different diplogastrid genera, presumably to fill niches during the phenology of the developing syconium [8]. There is also the newly documented situation  (Figs 5 and 6) it is possible that some odd manifestation of trophic morphology in this genus of Sycomorus associated nematodes is yet to be observed. Parasitodiplogaster was reported from F. sycomorus in Susoy et al. [8] but was not recovered from Sycomorus figs in this study. Teratodiplogaster (i.e., T. cf. martini and Teratodiplogaster n. spp. 1 and 2) was delimited to Sycomorus figs and interestingly this clade was sister to the P. salicifoliae clade of nematodes from Urostigma/Urostigma figs (Fig 6).
Second, and in contrast to the intraspecific morphological diversity in Pristionchus spp., fig  and fig wasp associated aphelenchoidids show highly conserved (convergent) morphology which resulted for many years in the lumping of three disparate genera into Schistonchus sensu latu which have now been separated into Schistonchus s.s., Martininema and Ficophagus [3]. In  [3,27,29]. For example, only one aphelenchoidid, S. africanus has been described from the studied region from F. burkei [13], and thus, one of the six aphelenchoidids (five Ficophagus and one Schistonchus spp.) recovered from F. burkei in this study could be conspecific to S. africanus. However, because of the typological similarity among fig-associated aphelenchoidids and the lack of detailed morphological and genetic information in the original description [9,13], we could not resolve or further characterize S. africanus using current taxonomic standards. In addition, the number and quality of materials collected in the present study were not sufficient to characterize each genotype typologically. Therefore, the conspecificity and the phylogenetic status of S. africanus with any of the F. burkei isolates in this study remains unclear and will require a reverse taxonomy approach. In addition, a species of Bursaphelenchus which is closely related to B. sycophilus isolated from F. variegata in Japan (Figs 2 and 3) [27] was isolated from F. sur from Pretoria corroborating a subgeneric Sycomorus fig co-association for this fig-derived Bursaphelenchus clade. Susoy et al. [8] reported discovery of a Bursaphelenchus from F. sycomorus from Pretoria that will require further work for phylogenetic placement.
Lastly, there is the issue of mixed lineages and incomplete lineage sorting from movement of species of nematodes among the different pollinators or inquilines that might visit different Ficus species [37,38]. Clearly, Martin et al. [1] made an extraordinary discovery about fig-associated nematodes in Africa, especially when it comes to Sycomorus clade figs, which is corroborated and further elucidated here with the aid of molecular inferences (Figs 2-7).
In addition to the two typical fig-associated families (Aphelenchoididae and Diplogastridae), an unidentified panagrolaimid was recognized in the present study. The precise phylogenetic status of the panagrolaimid is still unknown. This finding suggests that nematode colonization in fig syconia has occurred many times, more than we expected, and further diversity of fig associated nematodes will be discovered by more extensive collecting. Poinar [12] described P. sycophilon as the first member of this genus from F. burkei (subsection Chlamydodorae) from Zimbabwe without molecular data. We can now place P. sycophilon with six new Parasitodiplogaster species that are all associated with Urostigma subgenus, section Galoglychia figs from three subsections (Caulocarpae, Chlamydodorae and Platyphyllae) which form an African delimited Parasitodiplogaster clade that is associated with an African  [42][43][44].

Host fig association
Multiple colonization events and the duplication of lineages are also well documented in wood and bark beetles and their associated nematodes. A species of bark beetle often harbors several different lineages of aphelenchoidids and diplogastrids [45,46], and two closely related Bursaphelenchus spp. (B. luxuriosae Kanzaki & Futai and B. acaloleptae Kanzaki, Ekino, Maehara, Aikawa & Giblin-Davis) share the same host (Araliaceae trees) and carrier beetle (Acalolepta luxuriosa (Bates)) [47].
The wasp associations of the nematodes, e.g., host range, were not examined in the present study.  [37,38,48,49]. In addition, the host switching, duplication of the lineages, and extinction and fill-in of the wasp lineages can drive the diversification of figs and their pollinating wasps [49,50]. Because the nematodes are tightly associated with figs and fig wasp as the habitat/host and carrier/host, their diversification and host switching are likely to be affected by those of the carrier host. Further collections of figs, wasps and nematodes and specifying their host/carrier ranges, i.e., carrier wasp range of nematodes and host range of wasps, are necessary to understand their tripartite associations.
In the present study, we focused on the diversity of nematodes in relation to their host/habi- survey. Thus, there are potentially many more nematode species and interesting life history discoveries to be made to add upon what is reported here. There is also much reverse taxonomy needed to finish tying the genotypes reported in this study with well-defined morphotypes and biological/life history traits for South African species of fig associated nematodes as recently initiated by Wöhr et al. [17,18] and Susoy et al. [8].