Phylogenetic relationships of symbiotic spirochetes in the gut of diverse termites

Abstract

Phylogenetic relationships of symbiotic spirochetes in the gut of diverse termites were analyzed without cultivation of these microorganisms. A portion of the 16S rDNA (ca. 850 bp) was amplified directly from DNA of the mixed population in the gut by PCR and cloned. A total of 30 spirochetal phylotypes affiliated with the treponemes were identified from four termite species and they were compared with those already reported from other termites. They represented separate lines of descent from any known species of Treponema, and they were divided into two discrete clusters; one was related to Spirochaeta stenostrepta and S. caldaria, and the other was grouped together with members of the Treponema bryantii subgroup. Although some sequences from evolutionarily related termites showed close similarity, most of the sequences of spirochetes were dissimilar among different termite species, and spirochetal sequences from a single termite species occurred in several distinct phylogenetic positions. These findings suggest that termites constitute a rich reservoir of novel spirochetal diversity and that evolution of the symbiosis is not simple.

Termite, Spirochete, Treponema, Phylogeny, Evolution, Symbiosis

1 Introduction

One of the most fascinating examples of symbiosis is that displayed between termites and the microorganisms in their hindgut consisting of both pro- and eukaryotes [1,2]. Spirochetes are one of the most abundant, consistently present, and morphologically distinct groups of bacteria present in termite hindguts [3,4]. Spirochetes rarely occur in nature in a density and morphological diversity as great as that in the gut of termites. Based on comparative morphometric analysis, genera of large spirochetes have been re-verified, revised or created [5]. Also, many undescribed spirochetal morphotypes, especially smaller ones, are present in the termite hindgut, often in impressive numbers. They exist free in the gut fluid of all termites. In so-called lower termites, they are also found attached to the surface of protists in the hindgut or entirely engulfed within the protists [3,4]. The impressive consequence of the surface attachment is the propulsion of certain protists as a result of the coordinated and synchronous undulations of thousands of attached spirochetes, known as motility symbiosis [6]. Recently, pure cultures of the gut spirochetes have been obtained from the lower termite Zootermopsis angusticollis (family Termopsidae) and have been shown to catalyze the synthesis of acetate from H₂ plus CO₂[7]. The findings imply an important role for spirochetes in termite nutrition, and help to reconcile the dominance of acetogenesis over methanogenesis as an H₂ sink in termite hindguts [8].

Termites (order Isoptera) are a complex assemblage of species showing considerable variation in terms of biology, behavior and nutritional ecology, and are divided into two groups, so-called lower and higher termites [9]. The lower termites, a group comprised of six evolutionarily distinct termite families, harbor dense and diverse populations of both bacteria and cellulolytic flagellated protists in their hindgut. The higher termites, a group comprised of only one family, but approximately 75% of all termite species, also harbor a dense and diverse population of gut bacteria, but they typically lack eukaryotic protists. The higher termites are the most evolved and divergent group within the order Isoptera. Given the existence of more than 2000 described species of termites and considering the host specificity of some symbionts, a great diversity of spirochetes in the termite gut seems likely.

The phylogenetic positions of the termite gut spirochetes have been reported as determined through analysis of 16S rDNA amplified directly from DNA obtained from the microflora in the Australian lower termite, Mastotermes darwiniensis (family Mastotermitidae) [10,11] and the African higher termite, Nasutitermes lujae (family Termitidae) [12] without in vitro cultivation of these microorganisms. Analysis of 16S rDNA clones obtained by PCR amplification from termite gut contents has also revealed considerable phylogenetic diversity among spirochete and non-spirochete members of the community, including many novel phylotypes not yet representing in culture, for the Japanese lower termites Reticulitermes speratus (family Rhinotermitidae) [13] and Cryptotermes domesticus (family Kalotermitidae) [14]. These analyses as well as the analysis of the strains in pure cultures [7] reveal that spirochetes in the gut of termites are affiliated with the treponemes, but none were closely related to any known species of Treponema. In this study, we investigated the phylogeny of the spirochetes in diverse termite species based on 16S rDNA sequences. Portions of 16S rDNA of spirochetes were amplified and cloned from four additional termite species, and were compared to those previously reported, especially in order to consider the evolutionary relationships between the termites and their symbionts.

2 Materials and methods

2.1 Termites

Two higher termites, the soil-feeding termite Pericapritermes nitobei and the wood-feeding termite Nasutitermes takasagoensis (both belonging to the family Termitidae), and two lower termites, Glyptotermes fuscus (family Kalotermitidae) and Hodotermopsis sjoestedti (family Termopsidae), were used. The former two were collected on Iriomote Island, Japan, and the latter two were collected on Okinawa Island and Yakushima Island, Japan, respectively.

2.2 PCR amplification, cloning, and sequencing

Thirty to several hundreds worker or pseudergate termites were collected and the termite guts were withdrawn using forceps. The guts were homogenized, and DNA from the mixed population in the whole gut was extracted as described previously [13]. A portion of 16S rDNA was amplified from the extracted DNA by PCR using ExTaq DNA polymerase (Takara) according to the manufacturer's directions. The PCR primers used were the previously described universal primers 519FB and 1392RH [13] which sample the 3′-proximal two-thirds of the 16S rDNA corresponding to Escherichia coli 16S rRNA sequence positions 534–1391. The reaction conditions were as follows: 35 cycles at 94°C for 30 s, 48°C for 45 s, and 72°C for 2 min. PCR products were cloned into the pGEM-T vector (Promega) as described previously [14]. Partial sequences of the isolated clones were determined using the EUB900R primer [14] and compared to each other and to sequences in the databases (for the latter comparison, the BLAST algorithm [15] was used). In the case of the clones from H. sjoestedti, the DNA inserts were amplified by PCR using universal and reverse primers (Takara) which corresponded to both sides of the cloning site on the vector, and the amplified insert DNA was digested with either HpaII or HhaI (Takara) to analyze the restriction fragment length polymorphism (RFLP). Then, the clones showing the same RFLP pattern were grouped together and a representative of the group was partially sequenced (in representative clones of the 26 RFLP groups their partial sequences were determined and 15 sequences (RFLP groups) comprising 21 clones were affiliated with spirochetes). The complete nucleotide sequence of the clone insert was determined using the sequencing primer set for 16S rDNA described previously [14]. Nucleotide sequencing was performed as described previously [14]. The sequence data determined in this study will appear in the DDBJ, EMBL, and GenBank nucleotide sequence databases under the accession numbers shown in Fig. 1.

1

Phylogenetic tree of the 16S rDNA partial sequences of spirochetes from diverse termite species. The tree was inferred by the neighbor joining method on the basis of 871 unambiguously aligned nucleotide positions which corresponded to the Escherichia coli sequence positions 534–1391. Bootstrap values above 50 from 100 resamplings are shown for each of the nodes (neighbor joining and maximum parsimony methods represented above and below the nodes, respectively). The scale bar indicates 0.05 nucleotide substitution per position. As previously reported [22], three Spirochaeta species, S. zuelzerae, S. stenostrepta, and S. caldaria (free-living species of anaerobic spirochetes), are more closely related to the treponemes than to other members of the genus Spirochaeta on the basis of 16S rDNA sequence similarity. The termite Treponema clusters I and II detected upon comparing the spirochetal sequences from termites are indicated at the right side of the tree. Programs used for the phylogenetic analysis are contained in the PHYLIP package (version 3.5c) [26]. Distance matrix was calculated with the Kimura two-parameter in the program DNADIST and the neighbor joining tree was constructed from the evolutionary distance data using NEIGHBOR. Parsimony analysis was conducted with DNAPARS, with jumble order for the taxa addition (repetition 10 times) and global branch swapping options. The bootstrap with 100 replicates was generated using SEQBOOT to estimate the robustness of branches, and the consensus tree was constructed using CONSENSE.

2.3 Phylogenetic analysis

Sequence data used to infer phylogenetic trees were retrieved from the databases and their accession numbers are shown in Fig. 1. The sequence data were aligned using the CLUSTAL W package [16] and checked manually. Nucleotide positions of ambiguous alignments were omitted from the subsequent phylogenetic analysis. The phylogenetic relationships were inferred as described previously [14]. The program fastDNAml [17] was also used to create the maximum likelihood tree using empirical base frequency and local rearrangement of branches. The quartet puzzling analysis was conducted with the program PUZZLE 3.1 [18] with 1000 puzzling steps using the HKY substitution model with an estimated transition-transversion ratio, and site to site substitution rate variation modeled on a gamma distribution with four categories and the shape parameter estimated from the data.

3 Results and discussion

In order to investigate phylogenetic diversity among members of the microbial community in the termite guts, 16S rDNA clones which were amplified from the DNA extracted from the gut microflora by PCR using a set of universal primers were collected, and were analyzed by their partial sequences (ca. 300 bp). Among 52, 27, 34, and 52 16S rDNA clones obtained from the gut communities of two higher termites, P. nitobei and N. takasagoensis, and of two lower termites, G. fuscus and H. sjoestedti, respectively, 15, 26, 20, and 21 clones were affiliated to 16S rDNA of spirochetes. Isolation of a large number of spirochetal 16S rDNA clones from each termite suggests that spirochetes are one of the major populations in the gut community of these termites. Clones sharing high sequence similarity (more than 98%) were grouped together in each termite to obtain 11, 14, 11, and 10 groups, respectively, and whole nucleotide sequences (ca. 850 bp) of representative clones of the groups were determined.

Since formation of chimeras is a problem in PCR amplification from DNA of mixed populations, presumptive chimeric artifacts were intensively searched and excluded as follows. First, the sequences were analyzed by the CHECK_CHIMERA program of the Ribosomal RNA Database project [19] using the maximum improvement score (MIS) as a crude measurement of the sequences to be chimeric. The sequences with MIS values above 55 [20] were regarded as presumptive chimeras. Secondly, the sequences were separated into 5′- and 3′-short sequence domains and analyzed independently [21]. Sequences nearly identical in one of the domains but not in the other domain were also regarded as presumptive chimeras. We excluded any sequences suspected to be chimeras from both analyses (5 and 11 sequences, respectively). Then, the predicted rRNA secondary structures of the remaining sequences were inspected to confirm that they contained compensatory nucleotide substitutions in stem regions, which serve to maintain the stability of the secondary structure. A total of 30 sequences, that is 8, 9, 6, and 7 sequences from P. nitobei, N. takasagoensis, G. fuscus, and H. sjoestedti, respectively, were obtained as spirochetal phylotypes showing no obvious evidence of chimeric artifacts. The names of the clones were designated according to the initials of the genus and species names of the host termites, i.e. Pn, Nt, Gf, and Hs, respectively. It seems likely that further sampling of the clones will give more diverse sequences, although the sequences identified here were thought to represent major populations of spirochetes in the gut community. However, we did not aim at a more comprehensive analysis of spirochetal diversity in a single termite, rather we preferred to survey many termite species in order to investigate evolutionary relationships between diverse termites and their symbionts.

In addition to the 30 phylotypes identified in this study, several sequences of spirochetes in the gut communities of M. darwiniensis, N. lujae, R. speratus, C. domesticus, and Z. angusticollis have been reported [7,10–14]. We compared these sequences (a total of 53 sequences from nine termite species belonging to eight genera) and Fig. 1 shows their phylogenetic relationships with representatives of known spirochetal species. All of the sequences from termite spirochetes fell within the treponeme cluster of spirochetes. High bootstrap values, 100% for the neighbor-joining method (NJ) and for the maximum parsimony method (MP), respectively, supported this clustering. The presence of phylum and genus level ‘signature’ nucleotides in the inferred 16S rDNA sequences [22] was consistent with this assignment. However, none of the termite spirochetes was closely related to any known species of Treponema, showing less than 92% sequence similarity, and they represented distinct lines of descent within the treponeme cluster of spirochetes. Although only a portion of 16S rDNA (ca. 850 bp) was used in the phylogenetic analysis, the tree topology for the known spirochetes in Fig. 1 was similar to those previously reported [7,10,12,14,22–24].

There was no identical sequence isolated from different termite species. The 30 phylotypes identified in this study were remarkably diverse and represented novel sequences, although they seemed to be clustered together with any of those reported previously from the termites. Consequently, the termite gut spirochetes fell into two discrete clusters, designated termite Treponema clusters I and II (see Fig. 1). Most of the termite spirochetal sequences (45 out of 53) were grouped together with the sequences of S. stenostrepta and S. caldaria (cluster I), showing 87–92% sequence similarity to them. The bootstrap value of 76% for the NJ method supported this grouping, but this grouping was not supported by the MP bootstrap analysis. Cluster I of termite spirochetes consisted of a heterogeneous collection of sequences from all eight termite genera. All of the spirochetal sequences from higher termites fell in this cluster. The other eight sequences were grouped with those of members of the genus Treponema and S. zuelzerae (cluster II), showing 81–89% sequence similarity to them. The bootstrap values of 88% for the NJ method and 79% for the MP method supported this grouping. Especially, the T. bryantii subgroup of treponemes [22] seemed to form a monophyletic lineage together with the termite spirochetes of cluster II, which was supported by bootstrap analyses (77% for NJ and 71% for MP). Members of cluster II consisted of eight sequences from only three lower termites, H. sjoestedti, R. speratus, and M. darwiniensis. Base signature analysis also supported the monophyly between the T. bryantii subgroup and cluster II of the termite spirochetes (e.g. base positions 543, 661, 733, 744, 851, 1310, and 1327 [22]), while cluster I of the termite spirochetes shared base signatures with the T. phagedenis subgroup [22], in spite of their distant relationship in the phylogenetic tree. Furthermore, S. stenostrepta, S. caldaria, and members of cluster I completely shared these base signatures, supporting their grouping.

Within the two large clusters described above, there were several sequence clusters consisting of a single or two evolutionarily related termites. Four sequences, NL1, Nt25, Nt17, and Nt12, formed a monophyletic lineage supported by high bootstrap values (100% for NJ and 92% for MP). These sequences were derived from termite species of the same genus, Nasutitermes, but one inhabiting Africa and the other inhabiting Japan. The other two sets of four sequences (Nt63, Nt59, Pn16, and Pn58; and Hs23, Hs26, Hs27, and Hs32) formed monophyletic lineages supported statistically (96% and 68%; and 100% and 98%, for NJ and MP respectively). N. takasagoensis and P. nitobei, from which clones Nt63, Nt59, Pn16, and Pn58 were isolated, are evolutionarily related and belong to the family Termitidae. The sequences Gf35 and Cd48 showed 95.7% nucleotide identity and they were derived from the termites of the family Kalotermitidae. Recently, sequences of 16S rDNA from spirochetes in the gut of the lower termite Reticulitermes flavipes and Coptotermes formosanus have appeared in the databases (AF068334 to AF068463 [25]), and phylogenetic analysis including these sequences revealed that these sequences were assigned to either the termite Treponema cluster I or II. Most of the sequences from R. flavipes clustered together with those from R. speratus. Both termites belong to the same genus but one inhabits America and the other Japan. Most of the sequences from different termite species shared only less than 95% nucleotide identity, although most of the termites investigated here and previously (R. speratus and C. domesticus) inhabit the Japan Archipelago. Thus, phylogenetically related species of spirochetes were present not in termites inhabiting geographically near places but in evolutionarily related termite species. This finding may suggest the concept of specific diversification of spirochetes within the gut of termites. The sequences Rs1, Rs21, ZAS-1, and ZAS-2 showed significant sequence similarity (97.2–94.9%), and the sequences Rs2 and Hs33 also shared high sequence similarity (97.1%), in spite of the distant phylogenetic relationship of the termite hosts. Interestingly, in situ hybridization experiment with a specific probe for Rs2 and Hs33 identified origins of the sequences as ectosymbiotic spirochetes of dinenymphid protists (our unpublished results).

Other than the sequences clustering with statistical reliability described above, some sequences from higher termites seemed to be grouped into several clusters, and seven sequences from M. darwiniensis (sp40-7 to mpsp15) also seemed to be clustered. However, these clusters were not reliably supported by the bootstrap analyses. Furthermore, these clusters were not always shown in a maximum likelihood tree, and could not be supported in the quartet puzzling analysis. Especially, the grouping and the branching orders of the members of the cluster I remained unresolved. Although the sequence length of the analysis described in this study (ca. 850 bp) gave similar phylogenetic positions of the cultivated treponemes, a comparison with the longer sequence may give statistical reliability to these tenuous clusters. On the other hand, the sequences from the related termite species did not always appear as clusters. Probably, more extensive sampling of the clones within each termite species is necessary. However, these findings suggest that evolution of symbiotic spirochetes is not simple (see below).

It is noted that spirochetal sequences from one termite species tended to occur in several different phylogenetic positions. This observation was inconsistent with a simple evolutionary scenario that one spirochete species was acquired by an ancestor of termites and evolved within the termite gut to the present diversity. Rather, it may suggest that phylogenetically distinct species of spirochetes were acquired during their evolution, probably independently in the course of time, by diverse termite species, and then evolved within the gut to attain the present diversity. Of course, in order to ascertain the complete evolutionary history of termite spirochetes, more comprehensive analysis of more termite species is necessary. It is also noted that identical or closely similar spirochetal sequences (e.g. more than 97% identity) rarely occurred among the termite genera examined so far. In fact, the more termite species were investigated, the more spirochetal phylotypes were obtained. The results reported in this study are still premature in order to describe the real diversity of spirochetes in the termite gut because approximately 270 genera and more than 2000 species of termites in the world may constitute an abundant reservoir of formerly unrecognized spirochetal biodiversity.

Acknowledgments

We thank Y. Murayama and F. Aoki for assistance and I. Yasuda for advice on the collection of termites. This work was partially supported by grants for the Biodesign Research Program, the Genome Research Program, and the Eco Molecular Science Research Program from RIKEN, and by a grant for the International Cooperative Research Project (Bio-Recycle Project) from Japan Science and Technology Corporation. One of us (T.I.) is the recipient of a Special Postdoctoral Research Fellowship from RIKEN.