Biological properties of LBCA
The list of 25 oldest bacteria indicated by their VARS-IARS bitscores were headed by Mau and Tpo, and dominated by Clostridia (Table 1). Among non-Firmicute bacteria, only Thermotogae appeared on the list. In comparison, Fusobacteria only attained a bitscore of 312 for F. ulcerans; and Cyanobacteria attained a bitscore of 269 for Synechococcus. The low bitscore of 148 for Aquifex aeolicus was not unexpected on account of its evolution of a substantial oxic network for aerobic metabolism13. While the primitive character of clostridial metabolism among bacteria was clearly regonized14, not all Clostridia exhibited high VARS-IARS bitscores: the pathogenic Clostridia were characterized by modest bitscores likely because of the tortuous evolution of their VARS and/or IARS in the course of adapting possibly to a variety of host organisms, which was in agreement with the phylogenetic tree of Firmicutes based on phosphoglycerate kinase where C. thermocellum clustered with Thermoanaerobacter tencongensis at a lower branching position than C. tetani and C. perfringens15. Since the G + C content, when estimated from genome sequence, varied by no more than 1% within species, it represented a useful taxonomic descriptor16; and the higher G + C contents of 55.5% for Mau and 45.9% for Tpo compared to the 28.6% for C. tetani, 28.6% for C. botulinum and 27.7–28.7% for C. perfringens were indicative of the latter species having to adapt to different hosts. Notably, among the top-bitscoring bacteria Halobacteroides halobius17, Halothermothrix orenii18, Carboxydocella thermautotrophica19, Caldicellulosiruptor, Themoanaerobacter and the non-Firmicute Thermotogae were hydrogen-producers20. While growth with CO yielded acetate in most Moorella thermoacetica strains, such growth produced mainly hydrogen in its atypical AMP strain21; and the same-genus Moorella stamsii displayed an ability to ferment a range of sugars to form both hydrogen and acetate22. The major end products of pyruvate fermentation by Mau included H2, CO2 and acetate23. Tpo contributed to a hydrogen producing biocathode24; and Thermincola carboxydiphila, a close relative sharing 99% SSU rRNA identity with Tpo, could oxidize CO with H2O to form CO2 and H225.
Table 1
Bacteria with top VARS-IARS bitscores.
Species name | Phylum | Class | Bitscore |
Mahella australiensis 50 − 1 BON | Firmicutes | Clostridia | 378 |
Thermincola potens JR | Firmicutes | Clostridia | 377 |
Halobacteroides halobius DSM 5150 | Firmicutes | Clostridia | 375 |
Halothermothrix orenii H 168 | Firmicutes | Clostridia | 373 |
Desulfosporosinus orientis DSM 765 | Firmicutes | Clostridia | 372 |
Caldicellulosiruptor lactoaceticus 6A | Firmicutes | Clostridia | 368 |
Carboxydocella thermautotrophica | Firmicutes | Clostridia | 366 |
Thermotoga sp. RQ7 | Thermotogae | Thermotogae | 363 |
Moorella thermoacetica ATCC 39073 | Firmicutes | Clostridia | 361 |
Marinitoga sp. 1137 | Thermotogae | Thermotogae | 355 |
Caldanaerobacter subterraneus subsp. tengcongensis MB4 | Firmicutes | Clostridia | 353 |
Carboxydothermus hydrogenoformans Z-2901 | Firmicutes | Clostridia | 353 |
Pelotomaculum thermopropionicum SI | Firmicutes | Clostridia | 353 |
Thermosipho africanus TCF52B | Thermotogae | Thermotogae | 352 |
Fervidobacterium pennivorans DSM 9078 | Thermotogae | Thermotogae | 350 |
Thermoanaerobacter mathranii subsp. mathranii str. A3 | Firmicutes | Clostridia | 349 |
Desulfitobacterium dehalogenans ATCC 51507 | Firmicutes | Clostridia | 347 |
Acidaminococcus fermentans DSM 20731 | Firmicutes | Negativicutes | 347 |
Thermoclostridium stercorarium | Firmicutes | Clostridia | 345 |
Pseudothermotoga lettingae TMO | Thermotogae | Thermotogae | 343 |
Desulfotomaculum ferrireducens | Firmicutes | Clostridia | 343 |
Hungateiclostridium thermocellum ATCC 27405 | Firmicutes | Clostridia | 342 |
Kosmotoga olearia TBF 19.5.1 | Thermotogae | Thermotogae | 340 |
Halanaerobium hydrogeniformans | Firmicutes | Clostridia | 340 |
Anoxybacter fermentans | Firmicutes | Clostridia | 339 |
The attributes of Archaeal Progenitor
Since the primitive Clostridia were anaerobic chemoorganotrophs in possession of a glycolytic sequence and an inclination to produce hydrogen, the Archaeal Progenitor of bacteria might share some of these cellular traits. Microbial formation of hydrogen can proceed through biophotolysis as in cyanobacteria, photofermentation as in purple non-sulfur bacteria, electrohydrogenesis involving bacterial consortia and Proteobacteria with outer membrane c-type cytochromes, and dark fermentation as in Clostridia26. Numerous archaeons are hydrogenotrophic methanogens27, with the exception of hydrogen-cycling Methanosarcina where intracellular production of hydrogen is coupled to its extracellular participation in generating a transmembrane electrochemical gradient28. The leading archaea in the development of glycolysis are Thermococcales and Desulfurococcales, which can degrade 100% of glucose via a modified Embden-Meyerhof-Parnas (EMP) pathway29. Since Desulfurococcales employ hydrogen to reduce sulfur, whereas Thermococcales such as Pyrococcus furiosis and Thermococcus kodakarensis generate hydrogen through ferredoxin-NADH based on dark fermentation, Thermococcales would be more probable than Desulfurococcales as the origin of bacterial glycolysis. Thermococcus would be more probable than Pyrococcus as well due to the replacement of glyceraldehyde 3-phosphate (GAP) dehydrogenase by GAP:ferredoxin oxidoreductase in the latter’s glycolytic mechanism20. Although Clostridia utilized an EMP pathway for glycolysis with ATP-dependent kinases, while Thermococci preferred to use ADP-dependent kinases, the ADP-dependent phosphofructokinase of Thermococcus kodakarensis retained 20% activity when ADP was replaced by ATP30, suggesting that it might not be very difficult for Thermococci to switch to ATP-dependent kinases. Therefore Thermococci were faced with relatively few competing candidate Archaeal Progenitors capable of bequeathing to the bacteria both a modified EMP pathway and a dark fermentation mechanism for producing hydrogen. The comparable G + C contents of different Thermococcus clades at 45.5%-59.6%31, in the vicinity of the G + C contents of Mau and Tpo at 55.5% and 45.9% respectively, would remove any notion that Firmicutes are necessarily low in G + C, and facilitate a potential transformation of Thermococci as Archaeal Progenitor to form Clostridia.
Importance of high biodiversity sites
Mau was isolated from an oil reservoir23, Clostridium kogasensis from the soil under a corroded gas pipeline32, and Thermococcus sibiricus from a high temperature oil reservoir33. Thermococcus and Firmicutes were among the most plentiful archaea and bacteria respectively in reservoirs formed by the pressurized water employed to displace oil from the oil-bearing strata34, and Thermincola were abundant in reservoirs above 90oC35. Thus geobiological sites favorable for a transition from Thermococci to Clostridia would include petroleum-bearing niches. Notably, a survey of fifteen Thermococcus species revealed that none of the species originating outside of hydrothermal vents metabolized maltose as energy substrate, while a majority of the species originating from hydrothermal vents, including T. aggregans, T. guaymasensis, T. fumicolans, T. hydrothermalis and T. profundus all metabolized maltose36, suggesting that high biodiversity sites such as the vents would enhance divergence of the chemoorganotrophic Thermococcus from their cellular norm to develop breakthrough properties including bacterial-type sugar glycolysis based on a modified EMP pathway.
In this regard, the marine hydrothermal vents at Guaymas Basin are known to release an abundance of CO2, H2 and low molecular-weight peptides, hydrocarbons and carbohydrates when magmatic sills intruded into organic-rich sediments up to 28 to 7 Kyr ago37,38, and these compounds gave rise to elevated biodiversity39,40. In a single study, five novel Thermococcus species were identified from these vents based on SSU rRNA, elongation factors EF-1alpha and EF-2, which was indicative of a rapid rate of Thermococus evolution41. More recently, based on 37 single-copy genes including the VARS and IARS genes, a series of metagenome-assembled genomes (MAGs) of Thermococci and Euryarchaeota from these vents were sequenced42,43. The terrestrial subsurface has been revealed likewise as a hotspot of anaerobic biodiversity, where a single site was known to harbor much of the tree of life44: the cold water CO2-driven Crystal Geyser in Utah, originally an abandoned oil exploration well, allowed the sampling of 104 different phylum-level lineages of archaea and bacteria from different depths of the subsurface in the form of genome-resolved metagenomes and single amplified genomes (SAGs), including a novel Euryarchaeota archaeon genome with a VARS-IARS bitscore of 403 (viz. bit403)45,46.
VARS tree and sequence alignment
Archaeal and bacterial species that produce hydrogen through dark fermentation are relatively limited. To examine whether the cellular resemblance between Thermococci and primitive Clostridia represented within regard convergent evolution or phylogenetic relatedness, a two-domain VARS tree was built using the maximum parsimony method47, with the LUCA-proximal Methanopyrus kandleri as root (Fig. 1). There were a major-Thermococcal division consisting of familiar and novel species of Thermococci and Euyarchaeota such as T. litoralis, T. sibiricus and T. archaeon B45 G15, and a minor-Thermococcal division consisting of the novel species T. archaeon B48 G16 from Guaymas Basin, and Euryarchaeota archaeon bit403 from Crystal Geyser together with the clostridial Mahella and Thermincola. Different archaeal species with 300 or higher VARS-IARS bitscores were examined preferentially in the tree to minimize the impact of horizontal gene transfers (HGTs): since HGTs involving both the VARS and IARS genes would be unlikely, while HGTs involving either VARS or IARS alone would lower the similarity bitscore between VARS and IARS substantially, same-species VARS-IARS pairs that have managed to maintain a high bitscore between them could have been relatively protected against HGT perturbation.
In order to test further whether the allocation of Mau and Tpo to the minor-Thermococcal division on the VARS tree was the result of kinship or HGT, Fig. 2 shows oligopeptide segments I-V excerpted from the 49 aligned VARS sequences from Supplementary Figure S1. Their outstanding features were:
Oligo I. All eight of the amino acid positions of this Oligo in Thermincola were identical to those in Pyrococcus.
Oligo II. All nine of the amino acid positions in Thermincola were identical to those in Pyrococcus.
Oligo III. Eight of the nine amino acid positions in each of the Mahella and Thermincola nonapeptides were identical to those of Thermococcus sibiricus.
Oligo IV. The decapeptides in Mahella and Thermincola were identical to those in Paleococcus pacificus and T. archaeon B89 G9.
Oligo V. The heptapeptide in Mahella was identical to those in T. sibiricus, T. archaeon B45 G15, T. sp. 2319x1, T. litoralis, T. archaeon bit391, T. kodakarensis and P. furiosis.
These amino acid identities between the Oligos I-V of clostridial Mahella, Thermincola and some of the Thermococci were in accord with the association of the two primitive clostridial species only with Thermococci and no other archaeal group on the VARS tree48. Since HGTs occur most frequently between closely related species49, inter-domain HGTs would be rare. Furthermore, because Oligos I-V were scattered over different regions of the VARS sequence, five separate HGT events would be required for their horizontal transfer between the two microbial groups, even without counting the numerous additional oligopeptides in the VARS sequence that also exhibited significant Thermococci-Clostridia identities (as marked by red underline in Supplementary Figure S1). Therefore, the probability of Oligos I-V stemming from inter-domain HGTs was vanishingly small, establishing thereby the phylogenetic proximity between primitive Clostridia and a Thermococci-related LBCA.