Acetogenesis from H2 Plus CO2 by Spirochetes from Termite Guts

See allHide authors and affiliations

Science  29 Jan 1999:
Vol. 283, Issue 5402, pp. 686-689
DOI: 10.1126/science.283.5402.686


Pure cultures of termite gut spirochetes were obtained and were shown to catalyze the synthesis of acetate from H2 plus CO2. The 16S ribosomal DNA sequences of two strains were 98 percent similar and were affiliated with those of the genus Treponema. However, neither was closely related to any known treponeme. These findings imply an important role for spirochetes in termite nutrition, help to reconcile the dominance of acetogenesis over methanogenesis as an H2 sink in termite hindguts, suggest that the motility of termite gut protozoa by means of attached spirochetes may be based on interspecies H2 transfer, and underscore the importance of termites as a rich reservoir of novel microbial diversity.

There are few, if any, habitats on Earth in which spirochetes are such major members of the microbial community as in the gut of termites (1). As many as half of the prokaryotes in termite guts are spirochetes (2), which range in size from small cells (0.1 to 0.2 μm by 3 μm) to much larger ones (1 by 100 μm). However, since they were first observed by Leidy over a century ago (3), none had ever been obtained in pure culture. Recent analyses of spirochetal 16S ribosomal RNA (rRNA)–encoding genes (16S rDNA) amplified by polymerase chain reaction (PCR) from termite guts revealed that they were affiliated with the treponemes, but none were closely related to any known species ofTreponema (2, 4, 5).

We established enrichment cultures of spirochetes from hindgut contents of Zootermopsis angusticollis (Hagen) (Isoptera; Termopsidae) by using an anoxic medium under H2 plus CO2 (6). The medium contained rifamycin and phosphomycin (two drugs to which many spirochetes are resistant), as well as bromoethanesulfonate to inhibit the growth of H2-consuming methanogens (7). During 10 to 12 weeks of incubation at 23°C, growth of a mixture of spirochetes (each 0.2 to 0.3 μm by 5 to 15 μm in size) was accompanied by consumption of H2 and CO2 and by formation of up to 30 mM acetate (8). Little or no spirochetal growth or acetate production occurred if the H2 in the headspace was replaced by N2. Two spirochete strains were isolated from an enrichment in which spirochetes outnumbered nonspirochetal bacteria by about 50 to 1.

Strains ZAS-1 and ZAS-2 were similar in morphology and size (0.2 μm by 3 to 7 μm) (Fig. 1A). Both had two periplasmic flagella (each inserted at opposite ends and overlapping for most of the length of the cells) interposed between the protoplasmic cylinder and the outer sheath (Fig. 1, B and C). The nucleotide sequences of the 16S rDNAs of ZAS-1 and ZAS-2 were 98% similar and were affiliated with those of the genusTreponema (Fig. 1D). Consistent with this assignment were the presence of phylum- and genus-level “signature” nucleotides in the inferred 16S rRNA sequences (9). However, neither strain was closely related [that is, bore >97% sequence similarity (10)] to any known species ofTreponema. Phylogenetically, they grouped within a cluster of 16S rDNA clones from not-yet-cultured termite gut treponemes that ranged from 89% similar (clone NL1) to 97% similar (clones RFS3 and RFS25) and that included a clone (ZAS89; 95% similarity) from Z. angusticollis (11). The most similar sequences from cultivated relatives were from Spirochaeta caldaria and S. stenostrepta (92 to 93% similarity), two anaerobic spirochetes that are currently assigned to the genusSpirochaeta because they are free-living but that group within the treponemes on the basis of 16S rRNA sequence (2, 4, 5, 8). These results implied that ZAS-1 and ZAS-2 represented at least one new species of Treponema. However, we are postponing assignment of a species epithet or epithets until more is known about them.

Figure 1

(A) Phase contrast and (B and C) electron micrographs of transverse-sectioned (B) and intact (C) cells of ZAS-1. The two periplasmic flagella [arrows in (B)] exhibit a subterminal insertion into the protoplasmic cylinder [arrow in (C)]. Scale bars, 10 μm (A) and 0.1 μm [(B) and (C)]. (D) Phylogenetic tree inferred from 16S rDNA sequences of ZAS-1, ZAS-2, representative known spirochetes, and spirochetal 16S rDNA clones generated directly from termite gut contents (11,28). A maximum likelihood technique (fastDNAml) was used to generate the tree, which represents the topology consistentlyretrieved after jumbling the order of sequence addition and permitting global rearrangements. Numbers adjacent to nodes indicate bootstrap values >50% derived from 100 replicate trees generated with either fastDNAml (values above the branches) or PAUP (values below the branches). The asterisk indicates a value of 52. The homologous sequence from E. coli was used as an outgroup (not shown). The scale bar represents units of evolutionary distance and is based on sequence divergence. For Serpulina hyodysenteriae,† see (29).

ZAS-1 and ZAS-2 grew poorly in the medium used for enrichments (6). At 23°C their doubling time was ≥10 days and cell yields were <108 cells/ml. Growth was markedly improved in a medium containing yeast autolysate (YA) and a cofactor solution (12) and by increasing the incubation temperature. In 4YACo medium at 30°C, ZAS-1 grew with a doubling time of 23 to 24 hours to densities of 1.4 × 109 cells/ml, and ZAS-2 grew with a doubling time of 48 hours to 2.8 × 108 cells/ml. Little or no growth of either strain occurred if the cofactor solution or YA was omitted, and for yet-unknown reasons, YA could not be replaced by commercial yeast extracts (13).

The growth of ZAS-2 was dependent on the presence of H2, whose consumption was largely accounted for by the following equations: 4 H2 + 2 CO2 → CH3COOH + 2 H2O (plus H2 + CO2 → HCOOH) (Fig. 2A). Under H2plus 14CO2, the major product was14C-acetate, 80% of which was derived from14CO2 and the 14C label of which was distributed equally between both C atoms (Tables 1 and2). These results implied that acetogenesis from H2 plus14CO2 supported most of the growth of ZAS-2, although a small amount of acetate was also formed from other medium components.

Figure 2

Growth of (A) ZAS-2 in 4YACo medium under 80% H2 and (B and C) ZAS-1 in 2YACo medium under 80% N2 (B) or 80% H2 (C) (the balance was CO2) at 101 kPa. Negative pressure created by consumption of H2 plus CO2 was periodically balanced by the addition of 100% N2. Final concentrations of formate (not plotted) were (A) 3.0, (B) 8.1, and (C) <1.0 μmol/ml. Recovery of H2-derived electrons as acetate plus formate in experiments with ZAS-2 (which did not grow without H2) ranged from 88 to 98%. Representative results from three independent experiments are shown. OD600 nm, optical density at 600 nm.

Table 1

Products formed from 14CO2 by termite gut spirochetes. 14CO2 is used here to mean the total 14CO2 ↔ H14CO3  equilibrium mixture existing in the culture medium, which was under a headspace of 20% CO2(30). dpm, disintegrations per minute.

View this table:
Table 2

Distribution of 14C in acetate produced from 14CO2 by termite gut spirochetes. Acetate from the cultures shown in Table 1 (0.5 to 1.0 μmol each) was degraded as previously described (17). Recovery of the 14C label from methyl- and carboxyl-labeled acetate standards (American Radiolabelled Chemicals) was 92.4 and 98.5%, respectively, with <1% cross contamination.

View this table:

In contrast, ZAS-1 grew equally well under N2 or H2, in each case producing acetate as a major product (Fig. 2, B and C, respectively). However H2, when present, decreased from 80% of the gas phase to 31% (from 88.5 to 39.0 μmol per milliliter of culture). Under N2 plus14CO2, 13% of the acetate and 63% of the formate were derived from 14CO2 (Table 1). Under these conditions, reduction of 14CO2to acetate and formate must have involved the oxidation of a component or components of YA that may have also been a source of C for acetate production, as the specific activity of acetate formed was far less than twice that of the initial 14CO2. Under H2 plus 14CO2, H2oxidation appeared to supply electrons for CO2 reduction, sparing to some extent the oxidation of YA components, as the proportion of acetate and formate derived from14CO2 rose to 30 and 83%, respectively. Under either gas phase, however, both C atoms of acetate became labeled, with a slightly greater proportion of 14C (55 to 60%) being present in the carboxyl group (Table 2). This may reflect a dilution of 14C label entering the methyl group by an unlabeled methyl donor or donors present in YA, or it may reflect an exchange occurring between 14CO2 and the carboxyl group of unlabeled acetate derived from YA components. Such preferential exchange between CO2 and the carboxyl group of acetate has been reported for Acetobacterium woodii(14). H2-grown cells of ZAS-1 and ZAS-2 exhibited CO dehydrogenase, hydrogenase, and formate dehydrogenase activities, which suggests that acetogenesis may occur via the Wood-Ljungdahl (acetyl-CoA) pathway (15). The respective activities (expressed as units per milligram of protein) were 1.28, 0.47, and 0.13 for ZAS-1 and 0.64, 0.45, and 0.20 for ZAS-2.

The results described here reveal an activity previously unknown in the spirochete phylum of bacteria: acetogenesis from H2plus CO2. Hydrogenotrophy may thus be an important property to consider in efforts to isolate other not-yet-cultured spirochetes. It would also be worthwhile to reexamine already-cultured spirochetes for their ability to conserve energy by H2 oxidation; for example, the oral treponeme Treponema denticola, which was found to achieve a higher cell density if H2 was included in the gas phase (16).

Acetate formed by hindgut microbes is a major carbon and energy source for termites. Its oxidation supports as much as 100% of the insect's respiratory requirements, and up to one-third of this acetate can arise from H2 plus CO2(17). Given the abundance of spirochetes in termite guts and the likelihood that ZAS-1 and ZAS-2 are not the only acetate formers among them, it seems safe to conclude that spirochetes make a substantial contribution to termite nutrition. Hence it is not surprising that their elimination from guts results in decreased survival of termites (18). The ability of spirochetes to use H2 as a reductant for acetogenesis also helps reconcile the enigmatic dominance of acetogenesis over methanogenesis as an H2 sink in many termites (19). Ebert and Brune (20) showed that the highest concentration of H2in hindguts of Reticulitermes flavipes was in the central region, being produced by the large biomass of protozoa that occurs there. However, two zones of H2 consumption exist that deplete most of the H2 as it diffuses radially outward: a major zone in the central region itself and a minor one near the gut epithelium. The latter can be attributed to the dense population of methanogens, which, for yet-unknown reasons, preferentially colonize the microoxic region near the gut wall and hence are furthest downstream in the outwardly diffusing H2 gradient (21). The major zone of H2 consumption is now very likely attributable to spirochetes, which course among (and are often attached to) (1) protozoa in the anoxic central region and rarely, if ever, colonize the hindgut wall. In the lumen they can enjoy H2 concentrations as high as 50 mbar [about 50,000 parts per million by volume (ppmv) (20)], which is well above the H2 thresholds that are typical of most H2-utilizing acetogens (362 to 4660 ppmv), which in turn are 10- to 100-fold higher than those of H2-consuming methanogens (19). It may well be that the attachment of spirochetes to hindgut protozoa, which in at least one case results in a spectacular motility symbiosis (22), is based on interspecies H2 transfer from protozoa to spirochetes. Thus, our results support the notion that the spatial distribution of acetogens and methanogens in situ is a major factor affecting their coexistence in termite guts and their relative success as H2 consumers (20).

Nonspirochetal acetogens previously isolated from termite guts have proven to be new species [Sporomusa termitida, Acetonema longum, and Clostridium mayombei (23)], as have the methanogens Methanobrevibacter cuticularis, M. curvatus, and M. filiformis (21). Analysis of 16S rDNA clones obtained by PCR amplification from termite gut contents has revealed considerable phylogenetic diversity among spirochete and nonspirochete members of the community, including many novel phylotypes not yet represented in culture (2, 4, 5,23). Our present results underscore the growing recognition of termites as a rich reservoir of novel microbial diversity (25).

  • To whom correspondence should be addressed. E-mail: breznak{at}


View Abstract

Navigate This Article