A central issue in bacterial phylogeny is to understand how different bacterial phyla are related to each other and their branching order from a common ancestor [19] . Phylogenetic trees based on 16S rRNA or other genes/proteins trees have not been able to resolve these critical aspects [2-7,10,18,23,26,30] , with important bearing on numerous other questions (viz. origin of photosynthesis, origin of the eukaryotic cell) [10,15,27] . Important insights in these regards have been provided by the discovery and analyses of conserved indels in many universally distributed proteins [10,12,13,15-17,19] . These conserved indels (referred to as main line signatures) are commonly shared by species from several bacterial bacteria, but they are absent from the others. The genetic events leading to them are postulated to have occurred at important evolutionary branch points and their species distribution patterns provide valuable information regarding the branching order and interrelationships among different bacterial phyla [10,19] .

 

Click to enlarge
One example of such an indel that is found in the RNA polymerase β-subunit (RpoB) is shown in Figure 1. RpoB is a core component of RNA polymerase and it is universally conserved in all species [25] . The RpoB homologs from different proteobacteria, Aquifex, Chlamydiales, Verrucomicrobia, Planctomycetes and Bacteroidetes-Chlorobi-Fibrobacteres contain a large insert (boxed) in a conserved region that is commonly shared by all species from the above groups of bacteria, but which is not found in any species from all other bacterial phyla (e.g. Spirochetes, Deinococcus-Thermus, Chloroflexi, Fusobacteria, Actinobacteria, Thermotoga, Firmicutes) [9] . The absence of this indel in RpoB homologs from various Archaea provides evidence that the bacterial groups lacking this indel are ancestral. Based on its species distribution pattern, this insert was likely introduced in a common ancestor of different proteobacteria, Aquifex and Chlamydiales, Verrucomicrobia, Planctomycetes, Bacteroidetes-Chlorobi-Fibrobacteres after the branching of other group as indicated in the interpretive diagram.

 

Click to enlarge
Another prominent main-line indel has been identified in the DNA Gyrase B (B subunit). In this case a conserved insert of >150 aa is present in the Gyrase B from different Proteobacteria, Aquificales, Chlamydiales, Verrucomicrobia and Planctomycetes, but it is not found in species from any of the other bacterial phyla including Bacteroidetes-Chlorobi-Fibrobacteres ( See Figure 2) [9]. This insert is also not found in archaeal homologs. A comparison of the species distribution pattern of this indel with that found in the RpoB indicates that the insert in Gyrase B was introduced at a later stage in evolution after the branching of Fibrobacters-Chlorobi-Bacteroidetes species. The species distribution patterns of these two indels provide evidence that the Fibrobacters-Chlorobi-Bacteroidetes groups have diverged after the branching of Spirochetes, Deinococcus-Thermus, Chloroflexi, Actinobacteria, Thermotoga and Firmicutes phyla but before the divergence of Proteobacteria, Aquificales, Chlamydiales, Verrucomicrobia and Planctomycetes groups of species (see interpretive diagram in Fig. 2).

 

Similar to the above two indels, many other main line signatures that are helpful in understanding the branching order of bacteria phyla have been discovered in other widely distributed proteins ( see Fig. 3 and various other signatures listed below). Based upon the species distribution patterns of these signatures, they have been introduced at various stages in bacterial evolution as depicted in Fig. 3. Based upon the presence or absence of these signatures, nearly all of the main groups within Bacteria can be clearly distinguished and it is possible to logically deduce their branching order from a common ancestor, as shown in Fig. 3 [9,10,12,13,15-17,19].

 

 

Reliability and Predictive Power of the Indel Model

 

Most of the main-line indels on which the branching pattern shown in Fig. 3 is based were discovered prior to 1997, when the sequence data was quite limited and the number of sequenced genome was <10. However, each of these indels made (or make) very specific predictions about their presence or absence in species from different bacterial phyla. If this branching pattern is reliable and if these indels (i.e. RGCs responsible for them) were introduced in ancestral lineages at the indicated evolutionary stages, then all species from bacterial phyla that lie above the postulated insertion point should contain a given indel, whereas all species from groups that have branched off prior to the insertion point (i.e. lying below the insertion point) should be lacking the indel [10,13-15,17,19] . If the observed distribution of these indels in different genomes follows closely that predicted by the model, it should provide strong evidence that the observed branching pattern is reliable. If these indels were arising independently or if genes containing them were frequently laterally transferred, then their presence or absence in different species will not follow the predicted pattern. The predictions of the model have been tested by comparing the presence or absence of various indels in different completed genomes and they follow{Gupta, 2005 2624 /id;Gupta, 2003 1924 /id;Gupta, 2002 1863 /id} and they follow close. Results of these analyses for 8 of the important main-line signatures (examined at the end of August 2006) are presented below.

 

Distribution of Mainline Indels in Completed Bacterial Genomes

Protein	Signature	No. of genomes with indels Expected/found	No. of genomes lacking indel Expected/found	Exceptions observed
Hsp70	21-23 aa insert	410/410	167/167	0
Hsp60	1 aa insert	402/404	165/163	2
Ala RS	4 aa insert	369/370	207/206	1
RpoB	~190 aa insert	370/370	207/207	0
RpoC	>120 aa insert b	402/402	165/165	0
IP	2 aa insert	293/288	124/129	5
CTP synthase	10 aa insert	328/305	241/264	23 (most delta proteo) a
GyrB	~160aa	380/380	254/254	0 c

a CTP synthase originally examined when few delta proteobacteria genomes were available. The branching order of the delta/epsilon subgroups has not been resolved as yet. Recent examination suggests the CTP insert originated after the delta proteo before the epsilon, alpha, beta and gamma subdivisions, hence the absence of the 10 aa in most deltas.

b Insert larger in all cyanobacteria (>600aa)

c Solibacter and Acidobacteria (both Acidobacteria group) also contain indel, don’t know if this is an exception as we don’t know where Acidobacteria are placed

 

This table indicates the numbers of bacterial species in which these proteins have been found and compare the expected absence or presence of these indels in various species (as predicted by the model) with that actually observed. As seen, the model predicted with very high degree of accuracy the presence or absence of these indels in various species/genomes. For a number of proteins that are present in virtually all of the sequenced genomes (viz. Hsp70, Hsp60, RpoB, RpoC, GyrB, AlaRS), <10 exceptions to the predicted patterns are observed in >4000 observations. The ability of the indel model (Fig. 3) to predict with remarkable accuracy (>99%) the presence or absence of these indels in various genomes/species provides compelling evidence that the branching pattern suggested by these indels is reliable [13-15,17,19] . In case of CTP synthase, the insert was originally indicated to be specific for all proteobacteria {Gupta, 2000 1308 /id}. However, at that time sequence information for no delta proteobacteria was available. However, the absence of CTP synthase insert in most delta proteobacteria indicates that this insert was introduced in this gene after the divergence of delta proteobacteria and it is a shared characteristics of species from the epsilon, alpha, beta and gamma proteobacterial subdivisions.

The branching order of the bacterial phyla as deduced based on conserved indels, with a few exceptions, is generally in accordance with phylogenetic trees based on 16S rRNA as well as many other genes and proteins [2,4-6,10,18,26,30] . In most published trees, groups such as Thermotoga, Deinococcus-Thermus, Cyanobacteria and Green nonsulfur bacteria show deep branching, whereas other groups such as Proteobacteria and Chlamydiae and Bacteriodetes-Chlorobi are late branching lineages. A recent phylogenetic study based on concatenated sequences for 31 universally conserved proteins also strongly indicates that the Firmicutes constitute the deepest branching lineage with the bacteria [4] . The deep branching of Firmicutes, Actinobacteria as well Thermotoga, Deinococcus-Thermus is also supported by comparison of gene order arrangements in bacterial genomes [21,28] . Studies by Lake and coworkers also provide strong evidence that the root of the bacterial tree does not lay within the Gram-negative (or diderm) bacteria [10,22,29] . One notable difference between the present branching pattern and that seen in 16S rRNA trees concerns the deep branching of Aquifex in the latter studies [8,26] . However, this anomaly has now been shown to be due to the very high G+C content of Aquifex rRNA, which is a common characteristic of various hyperthermophilic organisms that results in their clustering with the Archaea and deep branching in the rRNA trees [24] .

 

 

The branching order of different groups as deduced by indel analysis is also consistent with the major structural differences seen within Bacteria (or prokaryotes) (Fig. 4). Bacteriacan be divided into two distinct groups, depending upon whether they are bounded by one membrane (monoderms) or two different membranes separated by a periplasmic compartment (diderms) [3,10,11,14,19,20,28] .These two groups roughly correspond to the Gram-positive and Gram-negative bacteria. The model derived based on conserved indels support this important structural distinction and indicate that of these two groups the monoderms or Gram-positive bacteria are ancestral (Fig. 3). The deduced branching order places Deinococcus-Thermus group in an intermediate position between the Gram-positive and Gram-negative bacteria (Fig. 4). This placement is in accordance with the unique structural characteristics of Deinococcus, which contains a thick peptidoglycan layer and show positive gram-staining, but they are surrounded by both inner and outer cell membranes, similar to various gram-negative bacteria [1] . These observations indicate that Deinococcus represents an evolutionary intermediate in the transition from Gram-positive bacteria to Gram-negative bacteria. Thus, the picture of bacterial phylogeny based on conserved RGCs is supported by many different lines of evidence including a good correspondence between genotype-phenotype, which suggests that it is reliable.

 

List of Mainline Signatures

Main Line Signature in Alanyl t-RNA Synthetase

Main Line Signature in FtsZ Protein

Main Line Signature in DNA Gyrase B

Main Line Signature in Hsp60 Protein

Main Line Monoderm-Diderm Signature in the Hsp70 Protein

Conserved Insert in the Hsp70 Protein Specific for βγ-Proteobacteria

Main Line Signature in Inorganic Pyrophosphatase

Conserved Insert in Valyl-tRNA Synthetase for βγ-Proteobacteria

Main line Signature in Hsp90 Protein Distinguishing Gram-positive and Gram-negative Bacteria

Sec A Signature Specific for αβγ- Proteobacteria

Main Line Signature in RNA Polymerase b Subunit (RpoB)

A 1 aa Deletion in Lon Protease that is Specific for αβγ- Proteobacteria

Conserved Main-line Signature in the Rho Protein

Main Line Signature in RNA Polymerase b ’ Subunit (RpoC)

[Go to Top]