Genome Signature Difference between Deinococcus radiodurans and Thermus thermophilus.

The extremely radioresistant bacteria of the genus Deinococcus and the extremely thermophilic bacteria of the genus Thermus belong to a common taxonomic group. Considering the distinct living environments of Deinococcus and Thermus, different genes would have been acquired through horizontal gene transfer after their divergence from a common ancestor. Their guanine-cytosine (GC) contents are similar; however, we hypothesized that their genomic signatures would be different. Our findings indicated that the genomes of Deinococcus radiodurans and Thermus thermophilus have different tetranucleotide frequencies. This analysis showed that the genome signature of D. radiodurans is most similar to that of Pseudomonas aeruginosa, whereas the genome signature of T. thermophilus is most similar to that of Thermanaerovibrio acidaminovorans. This difference in genome signatures may be related to the different evolutionary backgrounds of the 2 genera after their divergence from a common ancestor.

1. Introduction

In the present bacterial taxonomic system, the extremely radioresistant bacteria of the genus Deinococcus and the extremely thermophilic bacteria of the genus Thermus belong to a common lineage with remarkably different characteristics [1, 2]. Comparative genomic analyses have shown that after their divergence from a common ancestor, Deinococcus species seem to have acquired numerous genes from various other bacteria to survive different kinds of environmental stresses, whereas Thermus species have acquired genes from thermophilic archaea and bacteria to adapt to high-temperature environments [3]. For example, the aspartate kinase gene of Deinococcus radiodurans has a different evolutionary history from that of Thermus thermophilus [4]. In addition, D. radiodurans has several unique protein families [5] and genomic characters [6], and there is no genome-wide synteny between D. radiodurans and T. thermophilus [7]. However, phylogenetic analyses based on both orthologous protein sequence comparison and gene content comparison have shown that the genomes of Deinococcus and Thermus are most closely related with each other [3, 8]. The trinucleotide usage correlations have been used to predict the functional similarity between two RecA orthologs of bacteria including D. radiodurans and T. thermophilus [9].

If the genes acquired through horizontal gene transfers are different between Deinococcus and Thermus, then the genomic base composition (GC content) and/or genome signature can be hypothesized to also be different between these 2 genera. However, the GC content of D. radiodurans (67%) is similar to that of T. thermophilus (69.4%). The genome signature, on the other hand, is a powerful basis for comparing different bacterial genomes [11-19].

Phylogenetic analyses based on genome signature comparison have been developed, and these analyses are useful for metagenomics studies [20]. It was reported that comparative study using the frequency of tetranucleotides is a powerful tool for the bacterial genome comparison [21]. In this study, we compared the relative frequencies of tetranucleotides in 89 bacterial genome sequences and determined the phylogenetic positions of D. radiodurans and T. thermophilus.

2. Methods

2.1. Construction of Phylogenetic Relationships Based on the Relative Frequencies of Tetranucleotides in 89 Genome Sequences

We compared the relative frequencies of tetranucleotides in the genome sequences. The frequencies of the 89 bacteria were obtained from OligoWeb (oligonucleotide frequency search, http://insilico.ehu.es/oligoweb/). The 89 bacterial species are part of a list that which covers a wide range of bacterial species published in a previous report [8]. Each frequency vector consisted of 256 ( = 44) elements. The Euclidean distance between 2 vectors was calculated using the software package R (language and environment for statistical computing, http://www.R-project.org). On the basis of the distance matrix, a neighbor-joining tree was constructed using the MEGA software [10].

2.2. Ranking Based on Similarities between the Relative Frequencies of Tetranucleotides according to Correspondence Analysis

Correspondence analysis [22], which is a multivariate analysis method for profile data, was performed against the relative frequencies of tetranucleotides in 89 genomes. Correspondence analysis summarizes an originally high-dimensional data matrix (rows (tetranucleotides) and columns (genomes)) into a low-dimensional projection (space) [23, 24]. Scores (coordinates) in the low-dimensional space are given to each genome. The distance between plots (genomes) in a low-dimensional space theoretically depends on the degree of similarity in the relative frequencies of tetranucleotides: a short distance means similar relative frequencies of tetranucleotides between genomes, whereas a long distance means different relative frequencies. Thus, distance can be used as an index for similarity among genomes in the relative frequencies of tetranucleotides. Distances between all genome pairs were calculated, and then a ranking for distances was obtained.

3. Results and Discussion

In the neighbor-joining tree (Figure 1), D. radiodurans is located in the high-GC-content cluster, whereas T. thermophilus is grouped with Thermanaerovibrio acidaminovorans and their group is located away from the high-GC-content cluster. The neighbor-joining tree (Figure 1) was greatly influenced by the genomic GC content bias; most of the well-defined major taxonomic groups did not form a monophyletic lineage. This result indicates that each constituent of the well-defined major group has diversified by changing its genome signature during evolution. It is consistent with a previous paper indicating that microorganisms with a similar GC content have similar genome signature patterns [25].

Figure 1

Neighbor-joining tree based on tetranucleotide sequence frequencies in 89 genomes. The frequencies for 89 bacteria were obtained from OligoWeb (oligonucleotide frequency search, http://insilico.ehu.es/oligoweb/). Each frequency vector consisted of 256 elements. The Euclidean distance between 2 vectors was calculated using the software package R (language and environment for statistical computing, http://www.R-project.org). On the basis of the distance matrix, a neighbor-joining tree was constructed using the MEGA software [10]. Numbers in parentheses indicate the GC content (percentage) of each genome sequence. Arrows indicate the positions of Thermus thermophilus and Deinococcus radiodurans.

Phylogenetic analysis according to genome signature comparison is not based on multiple alignment data. Thus, bootstrap analysis cannot be performed. In this paper, we estimated the similarity between 2 different tetranucleotide frequencies by using correspondence analysis. The correspondence analysis showed that the genome signature of D. radiodurans is most similar to that of Pseudomonas aeruginosa (Table 1), whereas the genome signature of T. thermophilus is most similar to that of Th. acidaminovorans (Table 2). Although the D. radiodurans genome signature has similarity to 18 bacterial species within the distance 0.5, the T. thermophilus genome signature has similarity only to Th. acidaminovorans within the same distance (Table 2). These results indicate that T. thermophilus has a different genome signature from those of bacteria included in the high-GC-content cluster (Figure 1).

Table 1

Distance between Deinococcus radiodurans and each bacterium using correspondence analysis.

Bacterial species	Distance
Pseudomonas aeruginosa PO1	0.297932379
Myxococcus xanthus	0.305390764
Azorhizobium caulinodans	0.308895493
Ralstonia solanacearum	0.309212661
Gloeobacter violaceus	0.317496648
Symbiobacterium thermophilum	0.324422553
Thermomonospora curvata	0.347077134
Opitutus terrae	0.376683191
Acidobacterium capsulatum	0.378916616
Gemmatimonas aurantiaca	0.383939504
Rhodobacter sphaeroides 2.4.1	0.386383492
Rhodospirillum rubrum	0.392789705
Streptomyces griseus	0.415746597
Geobacter sulfurreducens	0.425877427
Agrobacterium tumefaciens	0.457788385
Thermomicrobium roseum	0.460897799
Syntrophobacter fumaroxidans	0.470005872
Sphingomonas wittichii	0.478630032
Desulfohalobium retbaense	0.50752939
Heliobacterium modesticaldum	0.512911658
Chloroflexus aurantiacus	0.53688488
Pirellula staleyi	0.540489386
Desulfatibacillum alkenivorans	0.618176651
Pelotomaculum thermopropionicum	0.636637282
Moorella thermoacetica	0.637983756
Xylella fastidiosa 9a5c	0.655118109
Escherichia coli K-12 MG1655	0.671407958
Neisseria meningitidis MC58	0.679417806
Thermanaerovibrio acidaminovorans	0.707497366
Nitrosomonas europaea ATCC 19718	0.718956013
Fibrobacter succinogenes	0.773393097
Dehalococcoides ethenogenes	0.793600646
Vibrio cholerae N16961	0.794460696
Desulfitobacterium hafniense DCB-2	0.823845007
Thermus thermophilus	0.831109438
Shewanella oneidensis	0.84848937
Alteromonas macleodii	0.881823764
Aminobacterium colombiense	0.889238858
Thermobaculum terrenum	0.899243716
Syntrophomonas wolfei	0.90576094
Bacillus subtilis	0.913613719
Coprothermobacter proteolyticus	0.923779953
Chlorobium chlorochromatii	0.926043337
Coxiella burnetii RSA 493	0.929681834
Thermotoga maritima	0.952651677
Bacteroides thetaiotaomicron	0.958944885
Denitrovibrio acetiphilus	0.966489936
Kosmotoga olearia	0.998958025
Carboxydothermus hydrogenoformans	1.012583789
Nostoc sp. PCC 7120	1.014447775
Aquifex aeolicus	1.03027576
Chlamydia trachomatis D/UW-3/CX	1.041383827
Elusimicrobium minutum	1.06077929
Haemophilus influenzae Rd KW20	1.084974973
Veillonella parvula	1.10092918
Helicobacter pylori 26695	1.124775019
Cyanothece sp. ATCC 51142	1.126779861
Thermoanaerobacter tengcongensis	1.139238445
Halothermothrix orenii	1.149150516
Eubacterium eligens	1.164829099
Natranaerobius thermophilus	1.167863816
Prochlorococcus marinus CCMP1375	1.174664974
Fervidobacterium nodosum	1.195233916
Caldicellulosiruptor saccharolyticus	1.19880562
Caldicellulosiruptor bescii	1.199097055
Persephonella marina	1.209862481
Leptospira interrogans serovar lai 56601	1.221066506
Anaerococcus prevotii	1.224535688
Petrotoga mobilis	1.231307366
Thermodesulfovibrio yellowstonii	1.242134666
Trichodesmium erythraeum	1.246593564
Sebaldella termitidis	1.270114395
Dictyoglomus turgidum	1.29240584
Dictyoglomus thermophilum	1.297069077
Thermosipho melanesiensis	1.324630145
Deferribacter desulfuricans	1.331638037
Clostridium acetobutylicum	1.357082068
Mycoplasma genitalium	1.360597739
Campylobacter jejuni NCTC 11168	1.374681774
Leptotrichia buccalis	1.383345312
Rickettsia prowazekii	1.426681449
Borrelia burgdorferi B31	1.431569209
Candidatus Phytoplasma asteris	1.471567529
Mesoplasma florum	1.477622916
Fusobacterium nucleatum	1.487576702
Brachyspira hyodysenteriae	1.517447262
Streptobacillus moniliformis	1.535004291
Ureaplasma parvum ATCC 700970	1.559892696

Table 2

Distance between Thermus thermophilus and each bacterium using correspondence analysis.

Bacterial species	Distance
Thermanaerovibrio acidaminovorans	0.468763255
Symbiobacterium thermophilum	0.686400076
Geobacter sulfurreducens	0.756754453
Myxococcus xanthus	0.772836176
Streptomyces griseus	0.786527308
Thermomonospora curvata	0.791039191
Moorella thermoacetica	0.806329416
Syntrophobacter fumaroxidans	0.825184063
Deinococcus radiodurans	0.831109438
Desulfohalobium retbaense	0.835469081
Rhodospirillum rubrum	0.836862939
Azorhizobium caulinodans	0.837497899
Gloeobacter violaceus	0.847382695
Rhodobacter sphaeroides 2.4.1	0.857474011
Desulfatibacillum alkenivorans	0.876877944
Heliobacterium modesticaldum	0.886943785
Pseudomonas aeruginosa PO1	0.902403886
Pelotomaculum thermopropionicum	0.910464775
Acidobacterium capsulatum	0.940977424
Thermomicrobium roseum	0.958396462
Agrobacterium tumefaciens	0.993864461
Gemmatimonas aurantiaca	0.993867563
Ralstonia solanacearum	0.99540692
Opitutus terrae	1.014357577
Sphingomonas wittichii	1.018425039
Chloroflexus aurantiacus	1.027585883
Pirellula staleyi	1.047176443
Desulfitobacterium hafniense DCB-2	1.051272244
Dehalococcoides ethenogenes	1.071801398
Xylella fastidiosa 9a5c	1.080146527
Thermobaculum terrenum	1.103102039
Aminobacterium colombiense	1.103447745
Syntrophomonas wolfei	1.119525557
Nitrosomonas europaea ATCC 19718	1.125942985
Escherichia coli K-12 MG1655	1.136087269
Neisseria meningitidis MC58	1.137392967
Fibrobacter succinogenes	1.147727362
Aquifex aeolicus	1.154770307
Thermotoga maritima	1.163190235
Coprothermobacter proteolyticus	1.187035315
Vibrio cholerae N16961	1.194131544
Carboxydothermus hydrogenoformans	1.202997317
Shewanella oneidensis	1.207081448
Bacillus subtilis	1.236980427
Coxiella burnetii RSA 493	1.237627206
Kosmotoga olearia	1.240198963
Alteromonas macleodii	1.241401986
Bacteroides thetaiotaomicron	1.250498401
Chlamydia trachomatis D/UW-3/CX	1.259097769
Chlorobium chlorochromatii	1.264256111
Denitrovibrio acetiphilus	1.264320363
Nostoc sp. PCC 7120	1.283892849
Halothermothrix orenii	1.307140057
Thermoanaerobacter tengcongensis	1.321852789
Elusimicrobium minutum	1.327006319
Cyanothece sp. ATCC 51142	1.338924672
Helicobacter pylori 26695	1.353623157
Veillonella parvula	1.366604516
Natranaerobius thermophilus	1.374016605
Persephonella marina	1.384851067
Prochlorococcus marinus CCMP1375	1.392425502
Haemophilus influenzae Rd KW20	1.392980033
Anaerococcus prevotii	1.394012634
Eubacterium eligens	1.420199298
Dictyoglomus turgidum	1.42068199
Caldicellulosiruptor saccharolyticus	1.428805275
Caldicellulosiruptor bescii	1.430940559
Dictyoglomus thermophilum	1.432160811
Petrotoga mobilis	1.43247619
Fervidobacterium nodosum	1.436232766
Leptospira interrogans serovar lai 56601	1.445508054
Thermodesulfovibrio yellowstonii	1.445636432
Trichodesmium erythraeum	1.459523665
Sebaldella termitidis	1.491593819
Thermosipho melanesiensis	1.522817305
Deferribacter desulfuricans	1.541728701
Clostridium acetobutylicum	1.553667164
Mycoplasma genitalium	1.586376378
Campylobacter jejuni NCTC 11168	1.590027263
Leptotrichia buccalis	1.598390503
Borrelia burgdorferi B31	1.626448618
Rickettsia prowazekii	1.653875547
Candidatus Phytoplasma asteris	1.673704846
Fusobacterium nucleatum	1.674099107
Mesoplasma florum	1.701326765
Streptobacillus moniliformis	1.715886446
Brachyspira hyodysenteriae	1.717967185
Ureaplasma parvum ATCC 700970	1.784252531

Although Pearson's correlation coefficient between the tetranucleotide frequencies of genomes of D. radiodurans and T. thermophilus is 0.630 (Figure 2), that between the tetranucleotide frequencies of genomes of D. radiodurans and Pseudomonas aeruginosa is 0.935 (Figure 3) and that between the tetranucleotide frequencies of genomes of Th. acidaminovorans and T. thermophilus is 0.914 (Figure 4). These results support the results of the neighbor-joining and correspondence analyses.

Figure 2

Scatter plot between the tetranucleotide frequencies of the genomes of Deinococcus radiodurans and Thermus thermophilus.

Figure 3

Scatter plot between the tetranucleotide frequencies of the genomes of Deinococcus radiodurans and Pseudomonas aeruginosa.

Figure 4

Scatter plot between the tetranucleotide frequencies of the genomes of Thermanaerovibrio acidaminovorans and Thermus thermophilus.

The frequency of horizontal gene transfer between different bacteria may be associated with genome signature similarity. However, the tree topology based on genome signature (Figure 1) is different from that based on gene content [8]. This is caused by, among others, an amelioration of the horizontally transferred genes [26]. Our findings strongly support the previous report that Deinococcus has acquired genes from various other bacteria to survive different kinds of environmental stresses, whereas Thermus has acquired genes from thermophilic bacteria to adapt to high-temperature environments [3].