Extremely halophilic Archaea (haloarchaea) have adapted to thrive in environments of high salinity, desiccation, and intense solar radiation. These microorganisms require at least 1.5 - 2.5 M NaCl for viability and typically display optimal growth in NaCl concentrations at or above 3.5 M. Haloarchaea commonly inhabit hypersaline environments, e.g. salt lakes, salterns, and heavily salted hides, meats, fish, and sauces [1–3]. Additionally, haloarchaea have been shown to survive space conditions [4] and viable cells have been reported from ancient deep underground salt deposits [5, 6]. Unlike most other extremophilic and archaeal microorganisms, haloarchaea form a monophyletic and coherent taxonomic group, the family Haloarchaeaceae [7].

The Halobacterium sp. NRC-1 genome sequence gave researchers the first opportunity, at the genome level, to probe the mechanisms of adaptation to hypersaline brine [8, 9]. Characterization of the 2 Mbp chromosome and two large megaplasmids showed that the overwhelming majority of predicted proteins were highly acidic, with a pI mode of 4.2, and very few neutral or basic proteins [10, 11]. In contrast, predicted proteins from most other non-haloarchaeal and bacterial organisms had equal fractions of acidic and basic components. The negatively charged residues in haloarchaeal proteins were predominantly found at the protein surface and predicted to function in enhancing their solubility and stability in high salt concentrations. A few individual haloarchaeal proteins have been crystallized, e.g. malate dehydrogenase, dihydrofolate reductase, and DNA sliding clamp (PCNA), and they all display markedly more acidic residues than non-haloarchaeal homologs. They also possess clusters of negative charges on the surface [12–14]. The high prevalence of negatively charged surface residues produces tightly bound hydration shells with salt ions bound at the protein surface [16, 17].

Several previous studies have examined the gene content in haloarchaea, including one aimed at identifying information transfer genes and another concerning metabolic genes [18, 19]. While a significant degree of conservation was found among the essential components of DNA replication, repair, and recombination, transcription, and translation, the study of metabolic genes showed substantially more diversity. Indeed, this diversity was illustrated by the recent identification of genes for a new pathway in central carbon metabolism, the methylaspartate cycle, in several haloarchaea [20]. An additional characteristic observed in most haloarchaeal genomes is the presence of large megaplasmids or minichromosomes which often harbor important or essential genes [21]. Gene content in these large extrachromosomal elements was compared and resulted in the finding of expanded gene families for replication and transcription initiation, e.g. orc and tfb [18], as well as the presence of a variety of genes needed for cell survival, e.g. an amino-acyl tRNA synthetase [9], resistance to arsenic [22], and production of buoyant gas vesicles [9].

In the current study, we present a comprehensive analysis of haloarchaeal genomes aimed at identifying the core haloarchaeal proteins and uniquely haloarchaeal groups. Halophilic Archaea representing thirteen different genera were included, all within the Haloarchaeaceae family. These microorganisms represent both geographic and phylogenetic diversity, including isolates from all 7 continents (Figure 1) and almost half of the genera in this tight clade of the Euryarchaea [2]. The genome-wide analysis produced nearly 800 protein clusters that are completely conserved among sequenced haloarchaea and a subset of 55 protein families that are unique to this family of extremophilic microbes.

Our current study has established core and unique haloarchaeal proteins and assigned likely functions to these conserved haloarchaeal proteins among sequenced haloarchaea. The core haloarchaeal orthologous groups (cHOGs) contained nearly 800 protein clusters that accounted for 21 - 33% of each predicted haloarchaeal proteome. The majority (89%) of the core proteins could be assigned specific or general functions based on association with NCBI KOGs and/or COGs, while the remainder (11%) were novel and could not be correlated to any previously known protein clusters. Based on further analysis of four recently sequenced haloarchaeal genomes and statistical analysis of alignments with non-haloarchaeal homologs, 55 protein clusters (named tucHOGs) were identified as haloarchaeal signature proteins.

The precise functions of the signature proteins are not clear because of their unique nature and the dearth of experimental studies focused on these genes. Only a single example among the truly unique haloarchaeal orthologous groups, Ral (tucHOG0456), was examined in any previous experimental work and was suggested to function in double-stranded DNA break repair and desiccation/radiation tolerance in the model haloarchaeon, Halobacterium sp. NRC-1 [27]. Transcriptome analysis of both UV irradiated Halobacterium sp. NRC-1 and its highly ionizing radiation resistant mutants showed an up-regulation of the rfa3-rfa8-ral operon, consistent with their involvement in DNA repair and protection [27]. Due to the transcriptional linkage of the three genes, and the presence of oligonucleotide binding (OB) folds in rfa3 and rfa8, the ral gene was also hypothesized to function as part of the eukaryotic-type single-stranded DNA binding RPA complex. However, analysis of the amino acid sequence of Ral did not reveal an OB fold domain, and it is not clear whether it serves as the third subunit of the RPA complex, replacing the RPA14 subunit found in eukaryotic organisms. While additional experimental studies are still required to determine the precise function of Ral, the possibility that it, as well as those of the other uniquely conserved haloarchaeal proteins, functions in adaptation of these organisms to their naturally extreme environments is an attractive hypothesis.

A somewhat larger (83) group of protein clusters, unique core haloarchaeal proteins (ucHOGs), includes 28 members which are nearly unique to haloarchaea (nucHOGs) and 55 which are truly unique to haloarchaea (tucHOGs). Our bioinformatic analysis of the ucHOGs suggested that they are quite typical of haloarchaeal proteins in pI, molecular weight, and GC-composition of their genes. The average pI of the ucHOGs is 4.7, similar to other haloarchaeal proteins (see Additional file 2). Similarly, the average G + C content of the ucHOGs are typical for each haloarchaeal chromosome (ranging from 68.5% for Halobacterium sp NRC-1 to 48.0% for H. walsbyi) (see Additional file 3). Their average molecular weight, 19.8 kDa, is somewhat smaller than predicted haloarchaeal proteins in general, 31 kDa (see Additional file 4). Their smaller size is consistent with their role as accessories to protein complexes, as suggested for the Ral protein in single-stranded DNA binding and DNA repair and protection. For example, as a group, ucHOGs may improve the activity or function of complexes in the cytoplasm with essentially saturating concentrations of KCl [10]. The great majority of ucHOGs appear to be soluble proteins (unpublished data).

The genomic distribution of ucHOG protein genes was examined and they were found to map overwhelmingly on the chromosomes in all of the haloarchaeal microorganisms (see Additional file 1). In the case of Halobacterium sp. NRC-1, all of the tucHOGs and all but one of the nucHOGs were located on the chromosome (Figure 4). The haloarchaea do not contain more than one or at most two of these proteins on megaplasmids. These findings suggest that the ucHOG proteins serve integral functions in these microorganisms and are likely important and possibly critical for survival. In addition to their likely important function, the ucHOGs, and especially the signature proteins (tucHOGs) and their genes, will also be useful as markers for the presence of members of the Haloarchaeaceae family in the environment.

Of the 83 ucHOGs, 28 were not completely unique to haloarchaea, with one or a few homologs present in non-haloarchaea (see Additional file 1). A large fraction (46%) of the hits were to methanogenic Archaea belonging to the Methanosarcinaceae, Methanosaetaceae, and Methanocellaceae families, which are relatively close to haloarchaea based on phylogenetic analysis of 16S sequences and include some moderate halophiles [1]. There were also a number of hits to halophilic bacteria, e.g., Salinibacter ruber, which may be the result of lateral gene transfer between species in a common environment [10]. Of the clusters determined to not be uniquely haloarchaeal, 14 were associated with archaeal COGs (arCOGs) containing non-haloarchaeal homologs, consistent with their presence in more than a single family of Archaea [28] (see Additional file 1). This may reflect the distinct and common ancestry of the Archaea.

Prior to our study, an analysis of conserved proteins in the Archaea was first completed on eight archaeal genomes which did not include any haloarchaeal genomes [29]. In this early study, 351 signature proteins present in at least two of the archaeal genomes were identified. In a subsequent study, 11 archaeal genomes were compared, including two haloarchaeal genomes [30]. The number of signature proteins shared by all 11 genomes decreased to only six and an additional 30 were identified in the majority of archaeal genomes. In an analysis of four haloarchaeal genomes, 127 haloarchaeal-specific proteins were reported [30]. Of these, we classified 51 as signature proteins or tucHOGs, 13 as nucHOGs, while the remaining 63 were either missing in one or more of the 13 haloarchaeal genomes or were associated with a COG (see Additional file 5). In another report, ten haloarchaeal genomes were recently compared and 112 'signature' clusters were reported [19], of which we found that 50 were similar to tucHOGs and 11 are like nucHOGs (see Additional file 6).

Several studies aimed at identifying signature proteins in other taxonomic groups have been conducted for organisms from other domains of life. Among bacteria, an analysis of actinobacterial genomes found 29 signature proteins present in the majority of genomes and an additional 204 that are found in some, but not all of the genomes [31]. In another study [32], five Chlamydial genomes and one Parachlamydial genome were compared, and 59 proteins were conserved in all six genomes, coded by hypothetical genes with no known functions. Two subsequent studies of α-proteobacterial genomes reported signature proteins [33, 34]. Initially three genomes were compared and six signature proteins were identified in the majority of α-proteobacterial genomes and an additional 47 proteins were identified in some but not all subgroups [33]. With the increase to 12 α-proteobacterial genomes, further work showed that only four of the original six signature proteins were present in all of the genomes [34]. Among eukaryotes, 300 conserved signature proteins were identified in sequenced genomes, including the deeply branching Giardia lamblia species [35–37].

The entire set of genes within a given species or group of organisms, in essence, the combination of the core and all dispensable genes, is sometimes referred to as the "pan-genome" [38]. With this approach, as more whole genomes become available, the size of the pan-genome increases due to an increase in the number of accessory genes, while the size of the core-genome asymptotically reaches a minimum. While there are numerous studies of species level pan-genomes, there are only a few published studies at the genus or family level. A study of 26 genomes from the Streptococcus genus found that the core-genome contains 611 orthologous groups, which constituted 26 - 30% of any one genome [39]. Analysis of 11 genomes from the Vibrionaceae family found the core-genome of 1,882 orthologous groups constituted 32 - 50% of these genomes [40]. Analysis of six genomes from the Enterobacteriaceae family identified 2,125 core orthologous groups that accounted for 43 - 88% of these genomes [41].

Our result from this study of the Haloarchaeaceae family showed that 21 - 33% of each genome constituted the core-genome and was similar to the results reported in earlier studies on other groups. Moreover, the great majority of core orthologous groups identified in the first nine haloarchaea were conserved in the subsequent four sequenced species. Our preliminary results with analysis of the pan-genome of haloarchaea show an expanding number of dispensable genes among members of this group (data not shown). The sequencing of additional haloarchaeal genomes and metagenomes and further bioinformatic analysis are likely to yield additional insights into the genetic composition of this interesting group of extremophilic microorganisms [42].

The signature and core genes and proteins are valuable concepts for understanding phylogenetic and phenotypic characteristics of coherent groups of organisms. Our analysis of 13 haloarchaea from different genera has established that the haloarchaeal proteome consists of 4,455 orthologous groups (HOGs), 784 of which form the core proteome (cHOGs), and 55 of which constitute haloarchaeal signature proteins (tucHOGs). The conservation of the cHOG and tucHOG clusters suggests that they may be essential or vital for survival. An attractive hypothesis, similar to what has been suggested for Ral, the only tucHOG with a predicted function, is that these small, chromosomally encoded proteins may act as accessory proteins enhancing macromolecular function in extreme conditions.