Phylogenomic analysis of proteins that are distinctive of Archaea and its main subgroups and the origin of methanogenesis

Proteins specific for methanogenic archaea and their various subgroups

Currently, the methanogens form the largest group within the Euryarchaeota. They are distinguished from all other prokaryotes by their ability to obtain all or most of their energy via the reduction of CO₂ to methane or by the process of methanogenesis. In the Bergey's manual [75], the methanogenes are divided into 5 distinct orders (viz. Methanobacteriales, Methanococcales, Methanomicrobiales, Methanosarcinales and Methanopyrales). Some studies have suggested that these organisms possess a set of unique enzymes which are responsible for methanogenesis, such as coenzyme M, Factor 420 and methanopterin [76]. However, no systematic study has been carried out thus far to identify proteins that are uniquely present in different methanogens. Our blast searches of proteins from different methanogens have led to identification of 31 proteins, which are uniquely found in various methanogenic archaea. Twenty of these 31 proteins are present in all sequenced methanogens, while 11 proteins are missing only in M. stadtmanae, which is a human intestinal inhabitant (see notes in Table 5). This archaeon generate methane by reduction of methanol with H₂ and lacks many proteins present in the genomes of other methanogens [77, 78]. Thus, it is highly likely that the 11 proteins missing in M. stadtmanae were selectively lost from this species. Therefore, it is very likely that the genes for these 31 proteins that are commonly shared by virtually all methanogens (Table 5(a)) evolved in a common ancestor of all methanogens.

These analyses have also identified 10 proteins that are uniquely shared by various methanogens as well as A. fulgidus (see Table 5(b)). The genes for these proteins likely evolved in a common ancestor of A. fulgidus and various methanogenic archaea and they point to a close relationship between these two groups of organisms (Fig. 3). Ten additional proteins are present in A. fulgidus as well as various Methanosarcinales and M. hungatei (Methanomicrobiales) (Table 5(c)). It is likely that the genes for these proteins also evolved in a common ancestor of A. fulgidus and various methanogenic archaea, but they were selectively lost in other methanogens. Of the proteins that are commonly shared by A. fulgidus and various methanogenic archaea, MMP0607 is reported to be a novel repressor of nif and glnA genes, which are involved in nitrogen assimilation [79]. Interestingly, 2 homologs of this protein are also found in 3 Dehalococcoides species, but nowhere else, which are very likely due to LGT. Protein MMP0984 is the ε-subunit of carbon-monoxide dehydrogenase complex, which is made up of five subunits in different methanogens [80]. The epsilon subunits are required for the reversible oxidation of CO to CO₂ [81]. All of the other components could be found in a few bacterial species, while the ε-subunit is restricted to methanogenic archaea and A. fulgidus [82, 83]. Protein MMP1499 is identified as a transcriptional regulator with a Helix-turn-helix (HTH) motif, but its exact role has not been reported.

Among the genes that are uniquely shared by various methanogenic archaea (or these archaea plus A. fulgidus), two large gene clusters responsible for methanogenesis are found. The proteins MMP1346, MMP1560–MMP1564 and MMP1566–MMP1567 (Table 5) are parts of an eight-component complex, coenzyme M methyltransferase (Mtr), which catalyzes an energy-conserving, sodium-ion-translocating step in methanogenesis from H₂ and CO₂ [84]. M. maripaludis contains all of the known Mtr subunits, but the gene coding for MtrF is fused into the N-terminal region of MtrA [53]. All other methanogenic archaeal genomes contain complete set of mtr genes. It is of interest to note that for the protein MMP1567 (MtrH), homologues with low E-values are also found in two Desulfitobacterium hafniense strains as well as in three Rhizobiales species (Aminobacter lissarensis, Methylobacterium chloromethanicum, and Hyphomicrobium chloromethanicum; α-proteobacteria) (see note in Table 5). These three rhizobiae species can use methyl halides as a sole source of carbon and energy, and all of them possess a set of cmu genes which are essential for methyl chloride degradation [85]. In particular, the CmuB protein which is homologous to MMP1567 transfers a methyl group to methylcobalamin:H₄ folate (H₄F), which is analogous to the reverse of the reaction catalyzed by MtrH in archaea [86]. In view of the sequence and functional similarity between MtrH and CmuB proteins, it is likely that the mtrH gene was laterally transferred from a methanogenic archaeon to the common ancestor of the above three rhizobiae species to serve the new functional role. The function of the laterally transferred mtrH related gene in D. hafniense is not known at present.

The proteins MMP1555–MMP1559 in Table 5 form another gene cluster, encoding the subunits of Methyl-coenzyme M reductase (MCR). This complex catalyzes the final reaction of the energy conserving pathway in which methylcoenzyme M and coenzyme B are converted to methane and the heterodisulfide CoM-S-S-CoB [87, 88]. Except for these proteins, the other proteins listed in Table 5 are of putative or unknown functions. It is likely that these proteins are involved in some aspects of methanogenesis or other unknown pathways unique to methanogenic archaea. These proteins provide molecular markers for methanogens, which can be used for identification of new archaeal species capable of methane production.

The blast searches of the M. maripaludis [53] and M. kandleri [56] genomes have identified 10 proteins that are uniquely shared by all of the following species belonging to the orders Methanobacteriales (M. thermoautotrophicus), Methanococcales (M. jannaschii, M. maripaludis) and Methanopyrales (M. kandleri) (Table 6(b)). Of these, only 2 proteins are present in M. stadtmanae, which is also a Methanobacteriales that has lost most of its genes due to its adaptation to the human intestine [78]. The genes for these 10 proteins likely evolved in a common ancestor of the above groups of methanogens (Fig. 3), which corresponds to the cluster of methanogenic archaea referred to as "Class I methanogens" [13]. Interestingly, these studies have also identified 10 proteins that are uniquely shared by these methanogenic orders and M. hungatei (see Table 6(a)), which branches distantly in phylogenetic trees [13]. The unique presence of these proteins in these methanogens suggests that species from these groups shared a common ancestor exclusive of other methanogenic archaea (Fig. 3).

Fifteen additional proteins discovered in this work (Table 6(c)) are uniquely present in M. kandleri and various Methanobacteriales indicating that these two groups are more closely related to each other than the Methanococcales (Fig. 3). We have also come across 7 proteins that are uniquely shared by Methanococcales and Methanobacteriales (Table 6(d)), and 4 proteins that are only present in Methanococcales and Methanopyrales (Table 6(e)). The most likely explanation to account for the species distributions of these latter proteins is that their genes also originated in a common ancestor of the above three groups of methanogens, but were selectively lost in either the Methanobacteriales or Methanopyrales lineages. These analyses have also identified 14 additional proteins that are uniquely present in all 5 Methanosarcinales species (Table 6(f)), as well as 7 proteins that are only found in various Methanosarcinales and M. hungatei (Table 6(g)). Lastly, these studies have also identified 55 proteins that are uniquely present in M. maripaludis and M. jannaschii (Methanococcales, see Additional file 3(a)) and 68 proteins that are only present in M. burtonii and 3 Methanosarcina species, all belonging to the Methanosarcinaceae family (see Additional file 3(b)) (Fig. 3) indicating that they are likely distinctive characteristics of species from these groups.

Of the proteins that are uniquely found in Methanococcales, Methanobacteriales, Methanopyrales and Methanomicrobiales, 12 proteins viz. MMP1448–MMP1454, MMP1456, MMP1458–MMP1460 and MMP1467 are from a big gene cluster eha, which encodes the multisubunit membrane-bound [Ni-Fe] hydrogenase [89]. Two of these proteins, MMP1456 and MMP1458, are only found in Methanococcales (Table 6(e)). The whole eha operon is composed of 20 ORFs in the genome of M. thermoautotrophicus and of these only these 12 proteins are restricted to these methanogens while the other subunits have counterparts in bacteria. The precise roles of these 12 proteins, which are predicted to be integral membrane proteins in the hydrogenase complex, have not been determined [89]. Among the other proteins that are specific for these groups of methanogens, MMP0127 and MMP1716 are Hmd homologs, which catalyze the reversible dehydrogenation of N⁵, N¹⁰-methylenetetrahydromethanopterin [90]. In the proteins that are specific for the Methanococcales (see Additional file 3(a)), one large gene cluster (MMP0233–MMP0240) is found, but no information is available concerning its possible function. Except for these proteins, all other proteins that are specific for these methanogenic archaea are of unknown or putative function.

Proteins that are specific for Thermococci

Thermococci are obligately thermophilic, strictly anaerobic cocci, which are able to convert elemental sulfur to hydrogen sulfide. Thus, they are so called "extremely thermophilic sulfur metabolizer", which comprise one of the main functional groups within Euryarchaeota. According to the Bergey's Manual [75], the class Thermococci contains only one family, Thermococcaceae, consisting of 2 genera: Thermococcus and Pyrococcus. Currently, 4 species from this family have been completely sequenced (Pyrococcus abyssi, P. horikoshii, P. furiosus and Thermococcus kodakarensis; see Table 1) [52, 91–93]. The blast searches on each protein from P. abyssi have identified 141 proteins that are shared by all 4 of these species (see Additional file 4(a)). All of these proteins show high degree of conservation within Thermococci and they do not have homologs in any other prokaryotes or eukaryotes except one possible LGT event (PAB1493, see note in Additional file 4). The genes for these proteins have likely evolved in a common ancestor of the Thermococci (Fig. 3). Of these proteins, PAB1510 is annotated as TBP-interacting protein (TIP), which forms complex with TBP (TATA-binding protein) to regulate transcription [94]. It is known that the archaeal transcription machinery is strikingly similar to that in eukaryotes [23], but no TBP-binding component was found in archaeal species until the discovery of the TIP in T. kodakaraensis [95, 96]. Most other Themococci-specific proteins are of unknown function, although in a few cases limited similarity to domains in known protein families have been noted. A number of proteins (viz. PAB0643–PAB0644.1n; PAB1821–PAB1826) are clustered together in the P. abyssi genome, and it is possible that they may form functional units and are involved in related functions.

Cohen et al. [52] have reported a large number of proteins which are restricted to the Pyrococcus genus. However, a number of proteins from their list are also found in T. kodakarensis KOD1 [93], whose genome was not available when their work was published. Some proteins are not specific for either Pyrococcus or Thermococci according to our criteria and some of them are only found in one species – P. abyssi. Our analysis of the P. abyssi GE5 genome has also identified 43 proteins that are unique to the Pyrococcus genus (see Additional file 4(b)). Again, almost all of these proteins are of unknown function except PAB2241, which is annotated as RNase P, but this annotation seems arbitrary as it does not show significant sequence similarity to known RNases. The proteins that are uniquely found in the 3 Pyrococcus genomes likely evolved in a common ancestor of this genus (Fig. 4).

Proteins that are specific for Halobacteria

Extreme halophiles constitute another major class within Euryarchaeota. They require 5–10 times the salinity of seawater (ca. 3–5 M NaCl) for optimal growth [17, 97]. In order to grow in such high salinity environments, they have developed a set of physiological adaptation, such as: high internal concentration of potassium chloride, acidic proteome with low pI value, high GC content with GC bias in the wobble position, unique chloride pumps to maintain osmotic balance, etc. [17, 98, 99]. Among archaea, halobacteria also have the unique ability to use solar energy to generate a proton gradient to synthesize ATP. So far, the Class Halobacteria harbors one family with 15 genera and 4 species have been completely sequenced, including Halobacterium sp. NRC-1, Haloarcula marismortui, H. walsbyi and Natronomonas pharaonis [54, 98, 100, 101]. By performing blast searches on each protein in the Halobacterium sp. NRC-1 genome, we have identified 127 proteins, which are only present in all 4 Halobacteria species with only 3 exceptions (see Additional file 5).

Of the proteins listed in this Table, VNG0016H, VNG1096H, VNG2414H and VNG2563H are annotated as DNA-binding proteins or regulators because of the presence of HTH domain, but their exact functions have not been reported. VNG0667G is an ATP-binding protein of ABC transporter family. Several other proteins, such as VNG2089H and VNG2628H, have also been assigned possible functions based on weak similarity to known conserved domains in the CDD database [102], but their exact functions remain to be determined. Because of their high degree of conservation and uniqueness to halobacteria, the genes for these proteins likely evolved in a common ancestor of Halobacteria (Fig. 4) and they are presumably involved in unique physiological functions related to their adaptation to the hypersaline environment. Because of their specificity for Halobacteria, these proteins provide useful biomarkers for this group of species.

In addition to these proteins that are specific to all sequenced halobacterial species, we have also identified a large number of proteins either uniquely shared by 3 halobacterial species or only found in 2 halobacterial species (see Additional files 6 and 7). Surprisingly, these proteins are present in different combinations of halobacterial species. The four-halobacterial species are from 4 different genera within the Halobacteriales order and their relationships are unclear at present. The largest numbers of these proteins (i.e. 56) are uniquely shared by the Haloarcula, Haloquadratum and Natronomonas species, followed by 49 proteins that are restricted to Haloquadratum and Haloarcula. These results suggest that of these three species, Haloquadratum and Haloarcula are more closely related to each other and that Halobacterium might be the deepest branching of the four available halobacterial species (Fig. 4). However, the genome size of these halobacterial species varies and some of these protein sequences are present on plasmids found in these species, which makes it difficult to reliably infer their relationships solely based on the number of shared proteins. Among the proteins that are specific for halobacteria, only few have been assigned possible functions. Protein VNG2178H is annotated as PhiH1-like repressor and VNG0584H is assigned as a Rieske Fe-S protein. Two additional proteins VNG1720H and VNG2562H have been annotated as iron-binding proteins because of their similarity with FhuD and TroA_a domains, respectively [102]. All of the other proteins are of unknown function.

Proteins that are specific for Thermoplasmata

The Thermoplasmata group is comprised of cell wall-less archaea, which resemble the bacterial Mycoplasma species [63]. Generally, they are thermoacidophilic, aerobic or facultative anaerobic, and are able to reduce sulfur to H₂S under anaerobic conditions [19, 55]. To date, this class include three families-Thermoplasmaceae, Picrophilaceae, and Ferroplasmaceae, each represented by one genus [103, 104]. Three complete genomes from this class (T. acidophilum, T. volcanium and P. torridus) are available at present (see Table 1) [19, 55, 63] and Ferroplasma acidarmanus Fer1 genome is draft assembled and sequence information for this is also available in the NCBI database. Our analyses have uncovered 77 proteins that are uniquely present in all four species belonging to this class (see Additional file 8(a)) (Fig. 4). Most of these proteins are present in all four available genomes, but a few are missing in one or two species, which is probably due to gene loss. Besides, we have also identified 33 proteins, which are shared only by the two Thermoplasma species (see Additional file 8(b)) and 17 proteins unique to P. torridus and F. acidarmanus (see Additional file 8(c)). The latter proteins indicate that species from Picrophilaceae and Ferroplasmaceae families are more closely related to each other (Fig. 4). All of these proteins are of unknown or predicted functions.

Proteins restricted to several archaeal lineages or showing sporadic distribution

In addition to the above proteins that are restricted to specific lineages of archaea, we have also identified 63 proteins, which are shared by several archaeal groups (see Table 7). The distribution pattern of these proteins could provide useful insights concerning the phylogenetic relationship between different groups. However, their distribution patterns could also be explained by gene losses in specific lineages or LGT between particular groups. Table 7 shows many proteins that are uniquely shared by various methanogenic archaea, Archaeoglobus and Thermococci. The first 5 proteins in Table 7(a) (PAB0076, PAB0138, PAB0965, PAB1927 and PAB1994) are present in all of the Thermococci and most of the methanogens. Four of these proteins are also present in A. fulgidus. The next 13 proteins in this Table are also uniquely found in most of the Thermococci as well as a number of methanogens and also in many cases in A. fulgidus. In addition, 6 proteins listed in Table 7(b) are only found in various Thermococci and A. fulgidus. These results suggest a closer relationship between the methanogenic archaea, A. fulgidus and Thermococci within the Euryarchaeota lineage. In conjunction with our earlier inference that A. fulgidus forms an outgroup of the methanogenic archaea, these results suggest that the above three groups are related in the following manner: Thermococci → A. fulgidus → Methanogens.

Although the relationship suggested above is the most likely explanation for the observed results, we have also come across three proteins (VNG1263c, MMP11287 and VNG2408c) that are uniquely present in various Halobacteria, A. fulgidus and different methanogens. To account for their species distribution, one has to postulate that their genes have been selectively lost from the Thermococci. In addition, 9 proteins are only found in various Halobacteria as well as Methanosarcinales and Methanomicrobiales (Table 7(c)). Their distribution requires again either selective gene losses from other lineages or LGT from Halobacteria to these methanogens.

Our analyses have also uncovered 30 proteins that are uniquely shared by species from Thermoplasmata and Sulfolobus (see Table 7(d)). Among these proteins, 7 are present in all Thermoplasmata and Sulfolobus species for which sequence information is available, while the remainder are missing in 1 or more species. It has been reported that there has been much lateral gene transfer between T. acidophilum and S. solfataricus, both of which inhabit the same environment [55]. However, the shared presence of these proteins in these two groups could also result from a unique shared ancestry of these thermo-acidophilic archaea.

Another 43 Archaea-specific proteins are sporadically present in different archaeal species (see Additional file 9). A number of proteins in this group are present in a limited number (between 3 to 6) of archaeal species belonging to different groups. There are 2 possible explanations that can account for their sporadic distribution: First, it is possible that some of these genes are the remnants of sequences that also originated in an ancestral lineage of Archaea but they have been selectively lost in many species because they are not required for growth. Second, the sporadic presence of these genes in a number of archaeal species can also be explained if some of these genes originally evolved in a particular group or species of archaea and then transferred to other archaea by LGT [105]. However, in view of the observed specificity of these genes/proteins for archaea, the LGTs in these cases need to be selective and limited to within archaea.