sign in sign up
<<
Home , Bacillus subtilis, Deinococcus, Deinococcus geothermalis, Deinococcus radiodurans, Methanococcus, Methanogen

Patterns of Temperature Adaptation in from the radiodurans and thermophilus

John H. McDonald Department of Biological Sciences, University of Delaware

Asymmetrical patterns of amino substitution in proteins of living at moderate and high temperatures ( and , respectively) are generally taken to indicate selection favoring different amino at different temperatures due to their biochemical properties. If that were the case, comparisons of different pairs of mesophilic and thermophilic taxa would exhibit similar patterns of substitutional asymmetry. A previous com- parison of mesophilic versus thermophilic with mesophilic versus thermophilic revealed several pairs of amino acids for which one was favored in thermophilic Bacillus and the other was favored in thermophilic Methanococcus. Most of this could be explained by the higher GϩC content of the DNA

of thermophilic Bacillus, a phenomenon not seen in the Methanococcus comparison. Here, I compared the meso- Downloaded from https://academic.oup.com/mbe/article/18/5/741/1018660 by guest on 28 September 2021 philic bacterium and its thermophilic relative , which are similar in GϩC content. Of the 190 pairs of amino acids, 83 exhibited signi®cant substitutional asymmetry, consistent with the pervasive effects of selection. Most of these signi®cantly asymmetrical pairs of amino acids were asymmetrical in the direction predicted from the Methanococcus data, consistent with thermal adaptation resulting from universal biochemical properties of the amino acids. However, 12 pairs of amino acids exhibited asymmetry signi®cantly different from and in the opposite direction of that found in the Methanococcus comparison, and 21 pairs of amino acids exhibited asymmetry that was signi®cantly different from that found in the Bacillus comparison and could not be explained by the greater GϩC content in thermophilic Bacillus. This suggests that selection due to universal biochemical properties of the amino acids and differences in GϩC content are not the only causes of substitutional asymmetry between mesophiles and thermophiles. Instead, selection on taxon-speci®c properties of amino acids, such as their metabolic cost, may play a role in causing asymmetrical patterns of substitution.

Introduction Organisms live at a wide range of temperatures, Most of the sequence comparisons of me- from Ͻ0ЊCtoϾ100ЊC, and they do so despite the dra- sophiles with thermophiles have examined a small num- matic effects of temperature on the function and stability ber of proteins from a broad range of organisms (Argos of their proteins. Understanding how proteins have et al. 1979; MeneÂndez-Arias and Argos 1989; Vogt, adapted to different temperatures therefore has long Woell, and Argos 1997). Some consistent patterns have been the subject of considerable research. One approach been evident, such as arginine being preferred over ly- has been to search for asymmetrical substitution patterns sine at higher temperatures, but the broad taxonomic in protein sequences from organisms living at different samples have made it dif®cult to know whether any gen- temperatures, usually living at moderate eral patterns of thermal adaptation are being obscured temperatures (mesophiles) compared with those living or exaggerated by taxon-speci®c asymmetries that may at high temperatures (thermophiles) (Argos et al. 1979; or may not be thermally adaptive. Recent whole- MeneÂndez-Arias and Argos 1989; Vogt, Woell, and Ar- sequence projects have made it possible to examine a gos 1997; Haney et al. 1999; McDonald, Grasso, and large number of protein sequences from a single pair of Rejto 1999). If protein evolution is largely due to neutral mesophilic versus thermophilic taxa. Haney et al. (1999) processes, or due to adaptation that is speci®c to each compared 115 protein sequences from the mesophilic protein and each site, then the substitutions are expected Methanococcus maripaludis, Methanococcus to be symmetrical: there will be as many aligned sites vannielii, and Methanococcus voltae with the thermo- with amino acid A in the and amino acid B philic Methanococcus jannaschii. They found 26 of the in the as there are sites with the opposite 190 pairs of amino acids to show signi®cant (P Ͻ 0.01) pattern. Signi®cant asymmetry, such as a greater number asymmetry, suggesting that adaptation to temperature of sites with A in the mesophile and B in the thermo- had an effect on a substantial proportion of amino acid phile, is usually interpreted as evidence that selection substitutions. McDonald, Grasso, and Rejto (1999) ex- favors different amino acids at different temperatures, amined a similar data set from the same of and much effort has been put into trying to identify bio- Methanococcus and compared the results with the pat- chemical properties of the amino acids that would ex- terns of asymmetry between the mesophilic bacterium plain this adaptation. and the thermophilic Bacillus stearo- thermophilus. The Bacillus comparison displayed sig- ni®cant (P Ͻ 0.05) substitutional asymmetry at 54 pairs Key words: protein adaptation, thermophile, Deinococcus, of amino acids, and most of the asymmetry was consis- Thermus. tent in direction with that seen in the Methanococcus comparison. However, several pairs of amino acids Address for correspondence and reprints: John H. McDonald, De- partment of Biological Sciences, University of Delaware, Newark, Del- showed patterns of asymmetry that were signi®cantly aware 19716. E-mail: [email protected]. different and opposite in direction in the Methanococcus Mol. Biol. Evol. 18(5):741±749. 2001 and Bacillus data sets, suggesting that taxon-speci®c ᭧ 2001 by the Society for and Evolution. ISSN: 0737-4038 processes were indeed important. Most of these differ-

741 742 McDonald ences consisted of an amino acid with a more GϩC-rich onym of (Degryse, Glansdorff, and codon being favored in thermophilic Bacillus but less Pierard 1978), T. thermophilus and T. aquaticus have common in thermophilic Methanococcus, consistent low similarity in genomic DNA : DNA hybridization with the higher genomic GϩC content of B. stearo- (Manaia et al. 1994; Williams et al. 1995) and 16s se- thermophilus compared with B. subtilis. Because me- quences (Saul et al. 1993), and T. thermophilus can sophilic and thermophilic Methanococcus differ little in grow in media containing 3% NaCl and has a higher genomic GϩC content, McDonald, Grasso, and Rejto maximum growth temperature than T. aquaticus (Man- (1999) suggested that the asymmetrical substitution pat- aia and da Costa 1991). Here, I compared only T. ther- terns seen there gave a better indication of which amino mophilus with D. radiodurans, because T. thermophilus acid substitutions were adaptive at different has a slightly higher optimum growth temperature and temperatures. has more sequences publicly available than does T. Even when mesophiles and thermophiles have the aquaticus.

same genomic GϩC content, it would be hasty to inter- Downloaded from https://academic.oup.com/mbe/article/18/5/741/1018660 by guest on 28 September 2021 pret all asymmetrical substitution patterns between them Materials and Methods as evidence for thermal adaptation, because there are other processes besides changes in GϩC content that All available protein sequences (including frag- could cause taxon-speci®c patterns of substitutional ments) for T. thermophilus were downloaded from the asymmetry. Amino acids vary in bioenergetic cost, and SwissProt, TrEMBL, and TrEMBLNew databases in Oc- amino acids with lower costs presumably will be fa- tober 2000. Sequences less than 20 amino acids long vored over functionally equivalent amino acids (Craig were discarded. For each sequence, the most similar pro- and Weber 1998; Craig et al. 1999). The relative bio- tein sequence from the complete D. radiodurans ge- energetic costs of different amino acids may vary among nome (White et al. 1999) was identi®ed using BLAST species, depending on the availability for uptake of each (Altschul et al. 1997) servers at the Institute for Geno- amino acid in the environment, the biosynthetic path- mic Research (http://www.tigr.org/tdb/CMR/gdr/htmls/ ways used to synthesize each amino acid, the abundance SeqSearch.html) and the National Center for Biotech- of raw materials for biosynthesis, and the effect of tem- nology Information (http://www.ncbi.nlm.nih.gov/blast/ perature and other environmental variables on the bio- blast.cgi). Where more than one sequence from T. ther- synthetic pathways. Environmental variables other than mophilus matched a single sequence in D. radiodurans, temperature, such as salinity, pH, and hydrostatic pres- presumably re¯ecting a duplication in T. thermo- sure, might also cause adaptive substitutional asymme- philus or loss of a duplicate in D. radiodurans, only the try that is unrelated to temperature. one sequence with the greatest identity to the D. ra- Only patterns of substitutional asymmetry that are diodurans sequence was used. Pairs of amino acid se- repeatedly observed in comparisons of mesophiles quences with less than 35% sequence identity were dis- paired with related thermophiles will be robust evidence carded. The resulting data set consisted of 186 sequenc- for thermal adaptation, while inconsistent patterns of es from T. thermophilus with matches from D. asymmetry could have a variety of possible explana- radiodurans. tions. Here, I compared sequences from the mesophilic Matching sequences were aligned using CLUSTAL bacterium Deinococcus radiodurans, whose genome has W (Thompson, Higgins, and Gibson 1994). Ambigu- been completely sequenced (White et al. 1999), with ously aligned sites adjacent to gaps were omitted, with Thermus thermophilus, a thermophilic bacterium that is the omitted sites extending from the gap to the nearest related to Deinococcus (Hensel et al. 1986; Weisburg, pair of adjacent sites that were both identical in the two Giovannoni, and Woese 1989). The resulting patterns sequences. The total data set consisted of 49,337 aligned were then compared with those observed earlier in com- amino acid sites, of which 18,041 were different be- parisons of mesophilic and thermophilic Methanococcus tween the species. The number of aligned sites exhib- and Bacillus to determine which asymmetries were con- iting each of the 190 possible pairwise patterns of dif- sistent and which differed among pairs of species. ference was then counted. For each pair of amino acids, Little is known about the natural history of D. ra- the signi®cance of the deviation from the expected 50: diodurans (Murray 1992). It can survive remarkable 50 ratio was tested using the log likelihood ratio test (G- amounts of gamma radiation, which may be a byproduct test) with the Williams correction for continuity (Sokal of adaptation to resistance (Mattimore and and Rohlf 1981); if the total number of sites was less Battista 1996), and it can also withstand intense ultra- than 50, Fisher's exact test was used. The ratio for D. violet radiation (Minton 1994) and desiccation (Sanders radiodurans versus T. thermophilus was compared with and Maxcy 1979). Deinococcus radiodurans has a ge- the ratio for previously published Bacillus and Meth- nomic GϩC content of 66.6% (White et al. 1999) and anococcus comparisons (McDonald, Grasso, and Rejto an optimal growth temperature of 25±30ЊC (Murray 1999) using a 2 ϫ 2 contingency table test with the 1992). Thermus thermophilus in hot springs and Williams correction (Sokal and Rohlf 1981); if the total arti®cial hot water environments. The type HB-8, number of sites was less than 50 in either comparison, which is used for most sequences, has an optimal growth Fisher's exact test was used. temperature of 73ЊC (Williams and da Costa 1992) and The information from the 190 pairwise compari- aGϩC content of 64.7% (Manaia et al. 1994). While sons was summarized into a single ranking of the amino T. thermophilus is sometimes considered a junior syn- acids from least preferred to most preferred at higher Temperature Adaptation in Proteins 743

temperatures by assigning a thermal asymmetry index Table 1 (TAI) re¯ecting the direction and magnitude of the Asymmetrical Patterns in Deinococcus radiodurans Versus asymmetries involving that amino acid (McDonald, Thermus thermophilus Grasso, and Rejto 1999). TAI values were assigned to PROPORTION minimize the difference between the predicted asym- DEINOCOCCUS/THERMUS AS SHOWNa metry for each pair of amino acids (a function of the m t Pref. Opp. P Deino Bac Meth difference in TAI values) and the observed asymmetry. S A 329 162 *** 0.670 0.715 0.728 Because only the difference in TAI values between ami- T V 194 78 *** 0.713 0.644 0.659 no acids was relevant, the TAI values were standardized Q R 183 73 *** 0.715 0.684 0.895 so that the average value was 1. A R 211 94 *** 0.692 0.683 0.650 To estimate the amount of divergence between N R 57 11 *** 0.838 0.762 0.800 A L 226 125 *** 0.644 0.578 0.615 pairs of taxa, 17 proteins were identi®ed that were pre- T L 99 38 *** 0.723 0.602 0.647

sent in the Bacillus, Methanococcus, and Deinococcus/ S P 77 28 *** 0.733 0.630 0.690 Downloaded from https://academic.oup.com/mbe/article/18/5/741/1018660 by guest on 28 September 2021 Thermus data sets. The six sequences for each protein G A 292 191 *** 0.605 0.579 0.704 were aligned using CLUSTAL W (Thompson, Higgins, S R 76 30 *** 0.717 0.786 0.938 S L 48 14 *** 0.774 0.660 0.810 and Gibson 1994), and the sites containing gaps were T P 41 14 *** 0.745 0.688 0.600 eliminated. In those sites present in all six species, the N H 33 9 *** 0.786 0.696 0.700 proportion of identical sites was calculated for each me- S H 31 8 *** 0.795 0.655 0.818 sophile-versus-thermophile pair. S V 45 17 *** 0.726 0.641 0.697 Q P 41 16 *** 0.719 0.708 0.545 H R 64 32 ** 0.667 0.580 0.714 Results N L 28 8 ** 0.778 0.655 0.929 D Q 58 29 ** 0.667 0.604 0.500 For each of the 190 pairs of amino acids, the neu- G P 50 26 ** 0.658 0.500 0.500 tral model predicted an equal number of aligned sites G R 84 52 ** 0.618 0.679 0.773 Y W 38 18 ** 0.679 0.500 0.667 with each direction of difference. There were 19 pairs Q V 42 21 ** 0.667 0.730 0.500 of amino acids in the Deinococcus/Thermus data set V R 63 37 ** 0.630 0.641 0.750 with fewer than six aligned sites, so they could not be S Q 45 24 * 0.652 0.694 0.706 signi®cantly asymmetrical at the P Ͻ 0.05 level. Of the H P 16 4 * 0.800 0.700 0.667 171 remaining pairs of amino acids, 83 exhibited sig- D T 46 25 * 0.648 0.485 0.478 D G 83 55 * 0.601 0.571 0.451 ni®cant asymmetry (tables 1±4). With this many statis- H E 47 27 * 0.635 0.500 0.600 tical tests, several were expected to be signi®cant at the N S 56 34 * 0.622 0.519 0.494 P Ͻ 0.05 level by chance; after a correction for multiple S T 180 139 * 0.564 0.557 0.652 comparisons (Benjamini and Hochberg 1995), 64 of I R 30 14 * 0.682 0.536 0.538 G L 45 26 * 0.634 0.677 0.500 these remained signi®cant. H Y 50 30 * 0.625 0.523 0.769 The patterns of asymmetry could be summarized D H 25 11 * 0.694 0.537 0.600 by assigning each amino acid a TAI (McDonald, Grasso, M I 53 33 * 0.616 0.506 0.635 and Rejto 1999). These indices were ®t to the data such Q W 6 0 * 1.000 1.000 ND that for each pair of amino acids, the amino acid having M P 12 3 * 0.800 0.875 0.500 I P 12 3 * 0.800 0.600 0.385 a higher index was predicted to be preferred in the ther- M Y 15 5 * 0.750 0.600 0.667 mophile, and the predicted magnitude of the asymmetry A H 39 23 * 0.629 0.535 0.500 was proportional to the difference in the indices. The G F 18 7 * 0.720 0.467 1.000 ordering of amino acids from smallest to largest TAI NOTE.ÐShown are those pairs of amino acids with signi®cantly asymmet- (®g. 1) was consistent with almost all of the signi®cantly rical patterns of substitution in the D. radiodurans versus T. thermophilus com- asymmetrical pairs; only the preferences in T. thermus parison and asymmetry that is not signi®cantly different from that in the Meth- of over tyrosine and of phenylalanine over anococcus or Bacillus data sets of McDonald, Grasso, and Rejto (1999). The tryptophan were not consistent with the ordering based ®rst column (m) shows the amino acid preferred in the mesophile D. radiod- on TAI. urans, and the second column (t) shows that amino acid preferred in the ther- mophile T. thermophilus. Pref. ϭ number of sites with the direction of substi- If the substitutional asymmetry were due to adap- tution shown in the ®rst two columns; Opp. ϭ number of sites with the opposite tation to different temperatures based solely on the bio- direction of substitution; P ϭ probability of obtaining the observed deviation chemical properties of the amino acids, similar patterns from a 1:1 ratio by chance: * P Ͻ 0.05, ** P Ͻ 0.01; *** P Ͻ 0.001. Pairs are of asymmetry would be expected in pairs of taxa with in order of increasing P values. Amino acid abbreviations: A, alanine; C, cys- teine, D, aspartic acid; E, glutamic acid; F, phenylalanine; G, glycine; H, histi- similar differences in growth temperature. There were dine; I, isoleucine; K, ; L, leucine; M, methionine; N, asparagine; P, pro- 42 pairs of amino acids exhibiting patterns of asym- line; Q, glutamine; R, arginine; S, serine; T, threonine; V, valine; W, tryptophan; metry that were signi®cantly different between the Dein- Y, tyrosine. ococcus/Thermus data set and the Bacillus data set of a Proportion of substitutions in each of the three comparisons with the di- McDonald, Grasso, and Rejto (1999) (tables 2 and 3). rection of substitution shown in the ®rst two columns. Deino ϭ D. radiodurans versus T. thermophilus; Bac ϭ Bacillus subtilis versus Bacillus stearothermo- Bacillus stearothermophilus had a higher GϩC genomic philus; Meth ϭ mesophilic Methanococcus versus Methanococcus jannaschii. content than B. subtilis, while T. thermophilus had a Proportion in bold are signi®cantly different from a 1:1 ratio (P Ͻ 0.05). ND ϭ slightly lower GϩC content than D. radiodurans. One no substitutions in either direction. might therefore expect stronger asymmetries favoring amino acids with higher GϩC content codons in B. 744 McDonald

Table 2 (Patterns in Deinococcus/Thermus Different from Those in Bacillus (change in G؉C

DEINOCOCCUS/THERMUS PROPORTION AS SHOWNa DEINO/BAC GϩC m t Pref. Opp. P Deino Bac Meth Pb CHANGEc I L 467 216 *** 0.684 0.521 0.414 *** 1 → 2 A E 357 164 *** 0.685 0.385 0.726 *** 3 → 2 G E 150 37 *** 0.802 0.450 0.500 *** 3 → 2 M L 263 130 *** 0.669 0.581 0.662 * 1 → 2 T R 99 26 *** 0.792 0.663 0.696 * 2 → 3 T K 82 18 *** 0.820 0.549 0.674 *** 2 → 1 N E 71 13 *** 0.845 0.526 0.650 *** 1 → 2 Q K 119 48 *** 0.713 0.425 0.679 *** 2 → 1 D K 48 11 *** 0.814 0.533 0.512 *** 2 → 1 N G 62 24 *** 0.721 0.522 0.508 ** 1 → 3 Downloaded from https://academic.oup.com/mbe/article/18/5/741/1018660 by guest on 28 September 2021 M V 57 22 *** 0.722 0.569 0.639 * 1 → 2 A K 119 67 *** 0.640 0.421 0.609 *** 3 → 1 K R 411 311 *** 0.569 0.724 0.626 *** 1 → 3 L Y 81 41 *** 0.664 0.446 0.459 ** 2 → 1 S F 19 1 *** 0.950 0.667 0.833 * 2 → 1 T Y 22 4 *** 0.846 0.545 0.500 * 2 → 1 I V 489 396 ** 0.553 0.489 0.446 ** 1 → 2 A F 42 20 ** 0.677 0.300 0.750 *** 3 → 1 A Y 36 17 ** 0.679 0.324 0.769 ** 3 → 1 Q M 29 13 * 0.690 0.423 0.571 * 2 → 1 H K 27 13 * 0.675 0.347 0.563 ** 2 → 1 W F 48 30 * 0.615 0.257 0.667 *** 2 → 1 R Y 25 12 * 0.676 0.394 1.000 * 3 → 1 V A 266 224 0.543 0.411 0.462 *** 2 → 3 C K 2 0 1.000 0.000 0.750 * 2 → 1 K P 42 30 0.583 0.802 0.450 ** 1 → 3 E R 149 128 0.538 0.672 0.444 ** 2 → 3 C F 3 0 1.000 0.231 0.750 * 2 → 1 R L 78 71 0.523 0.352 0.333 * 3 → 2

NOTE.ÐShown are those pairs of amino acids with a difference in maximum G ϩ C content of their condons and patterns of asymmetry in the Deinococcus/ Thermus and Bacillus data sets that are signi®cantly different from each other. The ®rst column (m) shows the amino acid preferred in the mesophile Deinococcus radiodurans, and the second column (t) shows the amino acid preferred in the thermophile Thermus thermophilus. Pref. ϭ number of sites with the direction of substitution shown in the ®rst two columns; Opp. ϭ number of sites with the opposite direction of substitution; P ϭ probability of obtaining the observed deviation from a 1:1 ratio by chance: * P Ͻ 0.05; ** P Ͻ 0.01; *** P Ͻ 0.001. Amino acid abbreviations: A, alanine; C, cysteine; D, aspartic acid; E, glutamic acid; F, phenylalanine; G, glycine; H, histidine; I, isoleucine; K, lysine; L, leucine; M, methionine; N, asparagine; P, proline; Q, glutamine; R, arginine; S, serine; T, threonine; V, valine; W, tryptophan; Y, tyrosine. a Proportion of substitutions in each of the three comparisons with the direction of substitution shown in the ®rst two columns. Deino ϭ D. radiodurans versus T. thermophilus; Bac ϭ Bacillus subtilis versus Bacillus stearothermophilus; Meth ϭ mesophilic Methanococcus versus Methanococcus jannaschii. Proportions in bold are signi®cantly different from a 1:1 ratio (P Ͻ 0.05). b Probability of obtaining the observed difference in ratios between the Deinococcus/Thermus and Bacillus comparisons by chance. c Change in the maximum number of GϩC in the codons for the amino acids. Bold entries in this column indicate that asymmetry that increases the maximum GϩC content in the thermophile was stronger in the Bacillus comparison. stearothermophilus, as seen when the asymmetries in anything, one might expect greater magnitudes of asym- the Bacillus and Methanococcus data sets were com- metry in the Methanococcus comparison, since meso- pared (McDonald, Grasso, and Rejto 1999). There were philic and thermophilic Methanococcus have a greater 21 pairs for which the amino acid with greater GϩC difference in optimal growth temperatures than do D. content was more strongly favored in B. stearothermo- radiodurans and T. thermophilus. However, of the 22 philus, and there were only 8 with the opposite pattern pairs of amino acids with signi®cantly different patterns (table 2), suggesting that whatever determines genomic of asymmetry in the Methanococcus and Thermus/Dein- GϩC content is affecting the substitutional asymmetries. ococcus data sets (table 4), 12 were asymmetrical in However, in addition to the eight asymmetries where opposite directions in the two pairs of taxa, and 8 were lower GϩC content is favored in B. stearothermophilus, asymmetrical in the same direction but with a greater there were 13 asymmetries that did not change the GϩC magnitude of asymmetry in the Deinococcus/Thermus content of the codons and yet showed signi®cantly dif- data set. Only the preferences of lysine over serine and ferent asymmetries in the Bacillus and Deinococcus/ of isoleucine over threonine were signi®cantly stronger Thermus data sets (table 3). in the Methanococcus comparison. Mesophilic and thermophilic Methanococcus have Another way of comparing the patterns of asym- similar GϩC contents, so if universal biochemical prop- metry between the Methanococcus and Deinococcus/ erties and GϩC content were the only factors affecting Thermus data sets was to examine the correlation of the substitutional asymmetry, the patterns would be similar thermal asymmetry indices in the two pairs of taxa (®g. in the Methanococcus data set of McDonald, Rejto, and 2). They were signi®cantly correlated (r2 ϭ 0.60, P Ͻ Grasso (1999) and the Deinococcus/Thermus data set. If 0.0001), which suggests that the substitutional asym- Temperature Adaptation in Proteins 745

metry has similar causes in the two pairs of taxa. How- Table 3 ever, there were four amino acidsÐisoleucine, cysteine, Pattern in Deinococcus/Thermus Different from Those in (asparagine, and aspartic acidÐthat had noticeably lower Bacillus (no change in G؉C TAI values in the Deinococcus/Thermus comparison, in- DEINOCOCCUS/THERMUS PROPORTION AS SHOWNa DEINO/ dicating that they were selected against more strongly BAC b in T. thermus than in M. jannaschii. Of the 22 pairs of m t Pref. Opp. P Deino Bac Meth P amino acids that differed signi®cantly in amount of D E 509 139 *** 0.785 0.525 0.580 *** asymmetry between the two pairs of taxa, 14 involved A P 194 50 *** 0.795 0.639 0.765 *** Q E 228 99 *** 0.697 0.497 0.610 *** one of these four amino acids being more strongly se- T E 122 41 *** 0.748 0.542 0.754 *** lected against in T. thermus than in M. jannaschii (table Q L 82 22 *** 0.788 0.545 0.824 ** 4). S E 101 40 *** 0.716 0.572 0.723 ** Because the amount of divergence between a pair N K 44 10 *** 0.815 0.455 0.637 *** V L 442 324 *** 0.577 0.486 0.495 **

of species could affect their patterns of substitutional Downloaded from https://academic.oup.com/mbe/article/18/5/741/1018660 by guest on 28 September 2021 D L 26 4 *** 0.867 0.500 0.600 * asymmetry, the divergences between the pairs of me- Q H 39 16 ** 0.709 0.541 0.800 * sophilic and thermophilic taxa were estimated. The 17 C L 15 4 * 0.789 0.381 0.200 * proteins that were present in all of the Methanococcus, M K 13 6 0.684 0.375 0.409 * Bacillus, and Deinococcus/Thermus data sets contained W L 24 20 0.545 0.250 0.333 *

2,576 aligned amino acids after sites aligned with gaps NOTE.ÐShown are those pairs of amino acids with no difference in maxi- were eliminated. At these sites, there was 64.6% diver- mum GϩC content of their codons and patterns of asymmetry in the Deinococ- gence between D. radiodurans and T. thermophilus, cus/Thermus and Bacillus comparison that are signi®cantly different from each 65.7% divergence between the mesophilic Methanococ- other. The ®rst column (m) shows the amino acid preferred in the mesophile cus and M. jannaschii, and 79.6% divergence between Deinococcus radiodurans, and the second column (t) shows the amino acid preferred in the thermophile Thermus thermophilus. Pref. ϭ number of sites with B. subtilis and B. stearothermophilus. the direction of substitution shown in the ®rst two columns; Opp. ϭ number of sites with the opposite direction of substitution; P ϭ probability of obtaining Discussion the observed deviation from a 1:1 ratio by chance: * P Ͻ 0.05; ** P Ͻ 0.01; *** P Ͻ 0.001. Amino acid abbreviations: A, alanine; C, cysteine; D, aspartic Asymmetrical substitution patterns between meso- acid; E, glutamic acid; F, phenylalanine; G, glycine; H, histidine; I, isoleucine; philes and thermophiles, such as those observed here K, lysine; L, leucine; M, methionine; N, asparagine; P, proline; Q, glutamine; between D. radiodurans and T. thermophilus and earlier R, arginine; S, serine; T, threonine; V, valine; W, tryptophan; Y, tyrosine. a Proportion of substitutions in each of the three comparison with the direc- in Methanococcus and Bacillus (Haney et al. 1999; tion of substitution shown in the ®rst two columns. Deino ϭ D. radiodurans McDonald, Grasso, and Rejto 1999), are usually inter- versus T. thermophilus; Bac ϭ Bacillus subtilis versus Bacillus stearothermo- preted as evidence for adaptation to different tempera- philus; Meth ϭ mesophilic Methanococcus versus Methanococcus jannaschii. tures due to biochemical properties of the amino acids Proportions in bold are signi®cantly different from a 1:1 ratio (P Ͻ 0.05). b Probability of obtaining the observed difference in ratios between the Dein- (Argos et al. 1979; MeneÂndez-Arias and Argos 1989; ococcus/Thermus and Bacillus comparisons by chance. Vogt, Woell, and Argos 1997; Haney et al. 1999). If this were the sole cause of asymmetrical substitution, how- er abundances of those amino acids with GϩC-rich co- ever, the patterns would be similar in all comparisons dons (Lobry 1997 and references therein). In the com- of mesophiles with thermophiles. There were many pairs parison of the mesophile B. subtilis (GϩC content of amino acids with consistent patterns of asymmetry in 43.5%) with the thermophile B. stearothermophilus the Bacillus, Methanococcus, and Deinococcus/Thermus (GϩC content 52%), the majority of signi®cantly asym- comparisons, but there were as many pairs of amino metrical pairs of amino acids favor amino acids with acids with signi®cantly different patterns of asymmetry codons with higher GϩC content in B. stearothermo- across the three taxa. This indicates that processes other philus, while a very small number decrease the GϩC than selection due to biochemical properties of the ami- content (McDonald, Grasso, and Rejto 1999). Of the 15 no acids affect the patterns of amino substitution be- pairs of amino acids that differ signi®cantly in substi- tween mesophiles and thermophiles. Here, I review tution ratio between Methanococcus and Bacillus and some possible reasons for different patterns of substi- are in the opposite direction, all but two have an amino tutional asymmetry in different pairs of mesophiles and acid with higher GϩC content being favored in ther- thermophiles. mophilic Bacillus and less common in thermophilic Methanococcus. This suggests that some of the asym- Difference in G C Content ϩ metrical substitution patterns observed in Bacillus are While there is no overall correlation of higher due to the differing GϩC content of the amino acids' GϩC content with higher habitat temperature across all codons, not their biochemical properties. For the Meth- prokaryotes (Galtier and Lobry 1997), thermophiles ex- anococcus and Deinococcus/Thermus comparisons, hibit higher GϩC content than mesophiles within some where the mesophiles and the thermophiles are similar taxa, such as Bacillus (Claus and Berkeley 1986) and in GϩC content, this should not play a major role in (Whitman, Bowen, and Boone producing the asymmetry. 1992). It is not known whether higher GϩC content A more subtle way in which GϩC content might re¯ects selection for greater DNA stability or a change affect patterns of adaptive asymmetry could result from in the mutation process (Mooers and Holmes 2000). Or- the differences in the GϩC content of the mesophile/ ganisms with higher GϩC content generally have great- thermophile pairs. Imagine a site with amino acid A in 746 McDonald

Table 4 Patterns in Deinococcus/Thermus Different from Those in Methanococcus

DEINOCOCCUS/THERMUS PROPORTION AS SHOWNa DEINO/ METH m t Pref. Opp. P Deino Bac Meth Pb D E 509 139 *** 0.785 0.525 0.580 *** I L 467 216 *** 0.684 0.521 0.414 *** G E 150 37 *** 0.802 0.450 0.500 *** T K 82 18 *** 0.820 0.549 0.674 * N E 71 13 *** 0.845 0.526 0.650 ** D R 77 21 *** 0.786 0.680 0.455 * D P 59 12 *** 0.831 0.733 0.545 * D K 48 11 *** 0.814 0.533 0.512 *** N K 44 10 *** 0.815 0.455 0.637 * Downloaded from https://academic.oup.com/mbe/article/18/5/741/1018660 by guest on 28 September 2021 D A 108 52 *** 0.675 0.607 0.313 *** T A 182 108 *** 0.628 0.568 0.490 * V L 442 324 *** 0.577 0.486 0.495 * N G 62 24 *** 0.721 0.522 0.508 ** FIG. 1.ÐThermal asymmetry index of amino acids, standardized C A 44 14 *** 0.759 0.605 0.450 * so that the average is 1.0. A, Deinococcus radiodurans versus Thermus S K 48 18 *** 0.727 0.646 0.864 * thermophilus. B, Bacillus subtilis versus Bacillus stearothermophilus. L Y 81 41 *** 0.664 0.446 0.459 * C, Mesophilic Methanococcus versus Methanococcus jannaschii. Ami- I V 489 396 ** 0.553 0.489 0.446 *** no acids with a greater asymmetry index are preferred in thermophiles. I F 43 21 ** 0.672 0.523 0.452 * Abbreviations of amino acids are as in table 1. C L 15 4 * 0.789 0.381 0.200 * T I 39 30 0.565 0.587 0.724 * Y Q 6 3 0.667 0.385 0.000 * creased in frequency in the lineage where they were K L 47 38 0.553 0.556 0.333 * newly neutral would be dif®cult to distinguish from those that increased due to positive selection. Because NOTE.ÐShown are those pairs of amino acids with patterns of asymmetry in the Deinococcus/ Thermus and Bacillus comparison that are signi®cantly dif- some Deinococcus species (Ferreira et al. 1997) and all ferent from each other. The ®rst column (m) shows the amino acid preferred in known Thermus species are thermophilic, the common the mesophile Deinococcus radiodurans, and the second column (t) shows the ancestor of the Deinococcus/Thermus group was prob- amino acid preferred in the thermophile Thermus thermophilus. Pref. ϭ number ably a thermophile, as were the common ancestors of of sites with the direction of substitution shown in the ®rst two columns; Opp. ϭ number of sites with the opposite direction of substitution; P ϭ probability the Methanococcus spp. (Keswani et al. 1996) and B. of obtaining the observed deviation from a 1:1 ratio by chance: * P Ͻ 0.05; subtilis and B. stearothermophilus (Ochi 1994). If there ** P Ͻ 0.01; *** P Ͻ 0.001. Amino acid abbreviations: A, alanine; C, cysteine; are many sites with less selective constraint at moderate D, aspartic acid; E, glutamic acid; F, phenylalanine; G, glycine; H, histidine; I, temperatures than at high temperatures, similar patterns isoleucine; K, lysine; L, leucine; M, methionine; N, asparagine; P, proline; Q, of asymmetry could have arisen in all three comparisons glutamine; R, arginine; S, serine; T, threonine; V, valine; W, tryptophan; Y, ty- rosine. Note that several pairs in table 4 also appear in table 2 or 3. a Proportion of substitutions in each of the three comparisons with the di- rection of substitution shown in the ®rst two columns. Deino ϭ D. radiodurans versus T. thermophilus; Bac ϭ Bacillus subtilis versus Bacillus stearothermo- philus; Meth ϭ mesophilic Methanococcus versus Methanococcus jannaschii. Proportions in bold are signi®cantly different from a 1:1 ratio (P Ͻ 0.05). b Probability of obtaining the observed difference in ratios between the Dein- ococcus/Thermus and Methanococcus comparisons by chance. a mesophile where either amino acid B or amino acid C would have equally adaptive biochemical properties in a thermophile. If the codons for B have higher GϩC content than the codons for C, the A-to-B substitution might be more common in a mesophile/thermophile pair with high GϩC content (such as Deinococcus/Thermus), while the A-to-C substitution might be more common in a mesophile/thermophile pair with low GϩC content (such as Methanococcus). It will be necessary to com- pare multiple mesophile/thermophile pairs with similar GϩC contents to evaluate the importance of this effect.

Relaxed Constraint

Relaxation of selective constraint such that many FIG. 2.ÐValues of the thermal asymmetry index (TAI) of amino sites were constrained to have one amino acid in the acids calculated for Deinococcus radiodurans versus T. thermophilus, plotted against the TAI values for mesophilic Methanococcus versus ancestral species but could have more than one adap- Methanococcus jannaschii. The four amino acids with the greatest dif- tively equivalent amino acid in one descendant lineage ferences between the TAI values for the two pairs of taxa are identi®ed: could also lead to asymmetry. Amino acids that in- N, asparagine; D, aspartic acid; C, cysteine; I, isoleucine. Temperature Adaptation in Proteins 747

Table 5 cently diverged of two mesophile-versus-thermophile Frequencies of Amino Acids at Aligned Sites pairs could exhibit greater asymmetry. At those sites Amino Acid D. rad.a T. therm.b Pc where amino acid A is adaptive at cooler temperatures A ...... 10.788 10.535 and amino acid B is adaptive at hotter temperatures, C ...... 0.551 0.458 * adaptation would occur rapidly after two species split D ...... 5.598 4.260 *** and began living in different temperatures. At those sites E ...... 6.934 8.928 *** where amino acids A and B are functionally equivalent F ...... 3.322 3.573 * at both high and low temperatures, neutral substitutions G ...... 9.137 8.593 ** H ...... 2.057 2.043 in both directions would slowly and symmetrically ac- I ...... 4.964 4.177 *** cumulate. The combination of rapid, asymmetrical sub- K ...... 4.333 4.718 ** stitutions due to adaptation and slow, symmetrical sub- L ...... 10.614 12.286 *** stitutions due to neutral processes would result in higher M ...... 2.189 1.743 ***

N ...... 2.649 1.994 *** asymmetry in more recently diverged species. Downloaded from https://academic.oup.com/mbe/article/18/5/741/1018660 by guest on 28 September 2021 P ...... 4.822 5.624 *** If the slow accumulation of neutral substitutions Q ...... 3.476 2.627 *** were causing different amounts of asymmetry in the dif- R ...... 7.019 8.334 *** ferent pairs of taxa, one would expect to see more asym- S ...... 4.469 3.411 *** metry in the less diverged pairs of species. Mesophilic T ...... 5.304 4.331 *** V ...... 8.137 8.338 Methanococcus and M. jannaschii have about the same W ...... 0.926 0.997 amount of divergence (65.7% identity for the 17 pro- Y ...... 2.712 3.028 ** teins analyzed) as D. radiodurans versus T. thermophi- lus (64.6% identity), so the amount of divergence is un- NOTE.ÐAmino acid abbreviations: A, alanine; C, cysteine; D, aspartic acid; E, glutamic acid; F, phenylalanine; G, glycine; H, histidine; I, isoleucine; K, likely to be the cause of the different patterns of asym- lysine; L, leucine; M, methionine; N, asparagine; P, proline; Q, glutamine; R, metry in these two comparisons. Bacillus subtilis and B. arginine; S, serine; T, threonine; V, valine; W, tryptophan; Y, tyrosine. The total stearothermophilus reveal considerably less divergence number of aligned sites was 49,337. (79.6% identity) than the other mesophile-versus-ther- a Composition (%) of the amino acid at aligned sites in Deinococcus ra- diodurans. mophile comparisons. Of those pairs of amino acids b Composition (%) of the amino acid at aligned sites in Thermus thermo- with signi®cantly different asymmetries between the Ba- philus. cillus and Deinococcus/Thermus data sets that cannot be c Probability of obtaining the observed difference in frequency by chance: explained by differences in GϩC content, most have * P Ͻ 0.05; ** P Ͻ 0.01; *** P Ͻ 0.01. less asymmetry in Bacillus than in Methanococcus or Deinococcus/Thermus (table 3). Thus, there is no evi- of mesophiles with thermophiles. Comparing a pair of dence that reduced asymmetry due to slowly accumu- species in which thermophily is the derived state would lating neutral substitutions is important in explaining the help to test this possibility. differences among the three data sets.

Unequal Numbers of Mutations Different Sets of Proteins The neutral model yielding symmetrical protein It is possible that certain categories of proteins substitution assumes that the numbers of mutations (the (membrane proteins, highly expressed proteins, ribo- number of generations times the mutation rate per gen- somal proteins, , etc.) exhibit different patterns eration) are equal on the lineages connecting a meso- of adaptive asymmetry from other categories of pro- phile and a thermophile to their common ancestor teins. A mesophile/thermophile comparison that includ- (McDonald, Grasso, and Rejto 1999). If there are more ed mostly proteins from one category could then exhibit generations or more mutations per generation on one different patterns of asymmetry from a mesophile/ther- lineage than on the other, asymmetrical patterns of sub- mophile comparison including mostly proteins from a stitution may result. There is no evidence for the dra- different category. None of the three data sets compared matic difference in substitution rate between mesophiles here has an obvious preponderance of proteins from one and thermophiles required for this process to yield functional category that was rare in one of the other data asymmetrical substitution patterns. In addition, this pro- sets, so it seems unlikely that this is a major cause of cess would not change the overall frequencies of the the different patterns of asymmetry among the data sets. amino acids in proteins, provided that the frequencies of Detailed examination of the patterns of substitutional amino acids in the ancestral species were at mutation/ asymmetry among functional categories will require drift equilibrium. The signi®cant changes in frequency much larger data sets for any statistical power. of several amino acids in the Deinococcus/Thermus (ta- ble 5), Methanococcus, and Bacillus data sets (Mc- Selection Due to Environmental Factors Other than Donald, Grasso, and Rejto 1999) indicate that unequal Temperature mutation rates are not the sole cause of the asymmetry. The environments of mesophiles and thermophiles often differ in other environmental variables; if any of Amount of Divergence these variables favor some amino acids over others, they If some form of adaptation produces substitutional might also cause substitutional asymmetry. For example, asymmetry, it is possible to imagine that the more re- the mesophilic Methanococcus were isolated from shal- 748 McDonald low and intertidal marine and estuarine sediments BLAST and PSI-BLAST: a new generation of protein da- (Stadtman and Barker 1951; Jones, Paynter, and Gupta tabase search programs. Nucleic Acids Res. 25:3389±3402. 1983), while the thermophilic M. jannaschii was isolat- ARGOS, P., M. G. ROSSMANN,U.M.GRAU,H.ZUBER,G. ed from a depth of 2,600 m (Jones et al. 1983). Salinity, FRANK, and J. D. TRATSCHIN. 1979. Thermal stability and hydrostatic pressure, pH, and the abundance of different protein structure. 25:5698±5703. BENJAMINI, Y., and Y. HOCHBERG. 1995. Controlling the false amino acids and their precursors in the environment are discovery rate: a practical and powerful approach to mul- among the environmental variables that could cause se- tiple testing. J. R. Stat. Soc. B 57:289±300. lection favoring different amino acids in different en- CLAUS, D., and R. C. W. BERKELEY. 1986. Bacillus vironments. To identify those asymmetries caused by Cohn 1872. Pp. 1105±1139 in P. H. A. SNEATH, ed. Ber- temperature differences, one ideally would want to com- gey's manual of systematic bacteriology. Vol. 2. Williams pare two species whose environments are identical ex- and Wilkins, Baltimore, Md. cept for temperature. Realistically, it will be necessary CRAIG,C.L,M.HSU,D.KAPLAN, and N. E. PIERCE. 1999. A comparison of the composition of silk proteins produced by

to compare a large number of mesophile/thermophile Downloaded from https://academic.oup.com/mbe/article/18/5/741/1018660 by guest on 28 September 2021 pairs, so that temperature is the only environmental var- spiders and insects. Int. J. Biol. Macromol. 24:109±118. iable that consistently differs between them and could CRAIG, C. L., and R. S. WEBER. 1998. Selection costs of amino acid substitutions in ColE1 and ColIa gene clusters har- therefore explain any consistent asymmetries. bored by . Mol. Biol. Evol. 15:774±776. DEGRYSE, E., N. GLANSDORFF, and A. PIERARD. 1978. A com- Selection Due to Bioenergetic Costs of the Amino parative analysis of extreme thermophilic bacteria belong- Acids ing to the genus Thermus. Arch. Microbiol. 117:189±196. EKIEL, I., K. F. JARRELL, and G. D. SPROTT. 1985. Amino acid In addition to biochemical properties and the GϩC biosynthesis and sodium-dependent transport in Methano- content of their codons, amino acids differ in their cost coccus voltae, as revealed by 13C NMR. Eur. J. Biochem. of uptake or synthesis, and if these bioenergetic costs 149:437±444. vary among species, substitutional asymmetry could re- FERREIRA, A. C., M. F. NOBRE,F.A.RAINEY,M.T.SILVA,R. WAIT,J.BURGHARDT,A.P.CHUNG, and M. S. DA COSTA. sult (Craig and Weber 1998; Craig et al. 1999). For ex- 1997. sp. nov. and Deinococcus ample, the mesophilic M. voltae lives in marine muds murrayi sp. nov., two extremely radiation-resistant and rich in organic material (Whitman et al. 1986), assimi- slightly thermophilic species from hot springs. Int. J. Syst. lates all amino acids tested (Ekiel, Jarrell, and Sprott Bacteriol. 47:939±947. 1985), and is heterotrophic for leucine and isoleucine GALTIER, N., and J. R. LOBRY. 1997. Relationships between (Whitman, Ankwanda, and Wolfe 1982); presumably, it genomic GϩC content, RNA secondary structures, and op- obtains much of its amino acids from its environment. timal growth temperature in prokaryotes. J. Mol. Evol. 44: The thermophile M. jannaschii is autotrophic and has 632±636. limited ability to assimilate amino acids (Sprott, Ekiel, HANEY, P. J., J. H. BADGER,G.L.BULDAK,C.I.REICH,C.R. and Patel 1993). At sites in protein sequences where two WOESE, and G. J. OLSEN. 1999. Thermal adaptation ana- lyzed by comparison of protein sequences from mesophilic or more amino acids are functionally equivalent, the one and extremely thermophilic Methanococcus species. Proc. that was most abundant in its environment presumably Natl. Acad. Sci. USA 96:3578±3583. would be favored in M. voltae, while the one with the HENSEL, R., W. DEMHARTER,O.KANDLER,R.M.KROPPEN- lowest cost of biosynthesis would be favored in M. jan- STEDT, and E. STACKEBRANDT. 1986. Chemotaxonomic and naschii. This could appear to be temperature-related molecular-genetic studies of the genus Thermus: evidence substitutional asymmetry, although the adaptation might for a phylogenetic relationship of Thermus aquaticus and not be caused by the temperature difference. Thermus ruber to the genus Deinococcus. Int. J. Syst. Bac- To summarize, the dramatic substitutional asym- teriol. 36:444±453. metries observed between proteins from mesophiles and JONES, W. J., J. A. LEIGH,F.MAYER,C.R.WOESE, and R. S. thermophiles are inconsistent with a simple neutral mod- WOLFE. 1983. Methanococcus jannaschii sp. nov., an ex- tremely thermophilic from a submarine hydro- el of protein evolution, but they may not all be the result thermal vent. Arch. Microbiol. 136:254±261. of temperature adaptation due to biochemical properties JONES, W. J., M. J. B. PAYNTER, and R. GUPTA. 1983. Char- of the amino acids. The signi®cant differences among acterization of Methanococcus maripaludis sp. nov., a new taxa in amount and direction of asymmetry suggest that methanogen isolated from salt marsh sediment. Arch. Mi- other processes, such as adaptation to environmental crobiol. 135:91±97. variables other than temperature, selection based on bio- KESWANI, J., S. ORKAND,U.PREMACHANDRAN,L.MANDELCO, energetic costs of amino acids, or (for taxa such as Ba- M. J. FRANKLIN, and W. B. WHITMAN. 1996. Phylogeny and cillus) changes in GϩC content, play an important role. of mesophilic Methanococcus spp. and compar- As more data become available, those patterns of asym- ison of rRNA, DNA hybridization, and phenotypic meth- metry that remain consistent across pairs of mesophiles/ ods. Int. J. Syst. Bacteriol. 46:727±735. LOBRY, J. R. 1997. In¯uence of genomic GϩC content on av- thermophiles will become more de®nitely related to erage amino acid composition of proteins from 59 bacterial temperature adaptation. species. Gene 205:309±316. MCDONALD, J. H., A. M. GRASSO, and L. K. REJTO. 1999. LITERATURE CITED Patterns of temperature adaptation in proteins from Meth- anococcus and Bacillus. Mol. Biol. Evol. 16:1785±1790. ALTSCHUL, S. F., T. L. MADDEN,A.A.SCHAFFER,J.H.ZHANG, MANAIA, C. M., and M. S. DA COSTA. 1991. Characterization Z. ZHANG,W.MILLER, and D. J. LIPMAN. 1997. Gapped of halotolerant Thermus isolates from shallow marine hot Temperature Adaptation in Proteins 749

springs on S. Miguel, Azores. J. Gen. Microbiol. 137:2643± STADTMAN, T. C., and H. A. BARKER. 1951. Studies on the 2648. . X. A new formate-decomposing bac- MANAIA, C. M., B. HOSTE,M.C.GUTIERREZ,M.GILLIS,A. terium, Methanococcus vannielii. J. Bacteriol. 62:269±280. VENTOSA,K.KERSTERS, and M. S. DA COSTA. 1994. Hal- THOMPSON, J. D., D. G. HIGGINS, and T. J. GIBSON. 1994. otolerant Thermus strains from marine and terrestrial hot CLUSTAL W: improving the sensitivity of progressive mul- springs belong to Thermus thermophilus (ex Oshima and tiple sequence alignment through sequence weighting, po- Imahori, 1974) nom. rev. emend. Syst. Appl. Microbiol. 17: sition-speci®c gap penalties and weight matrix choice. Nu- 526±532. cleic Acids Res. 22:4673±4680. MATTIMORE, V., and J. R. BATTISTA. 1996. of VOGT, G., S. WOELL, and P. ARGOS. 1997. Protein thermal Deinococcus radiodurans: functions necessary to survive stability, bonds, and pairs. J. Mol. Biol. 269: are also necessary to survive prolonged 631±643. desiccation. J. Bacteriol. 178:633±637. WEISBURG, W. G., S. J. GIOVANNONI, and C. R. WOESE. 1989. MENEÂ NDEZ-ARIAS, L., and P. ARGOS. 1989. Engineering pro- The Deinococcus-Thermus phylum and the effect of rRNA tein thermal stability: sequence statistics point to residue composition on phylogenetic tree construction. Syst. Appl. Downloaded from https://academic.oup.com/mbe/article/18/5/741/1018660 by guest on 28 September 2021 substitutions in alpha helices. J. Mol. Biol. 206:397±405. Microbiol. 11:128±134. HITE ISEN EIDELBERG MINTON, K. W. 1994. DNA repair in the extremely radioresis- W , O., J. A. E ,J.F.H et al. (29 co-authors). tant bacterium Deinococcus radiodurans. Mol. Microbiol. 1999. Genome sequence of the radioresistant bacterium 13:9±15. Deinococcus radiodurans R1. Science 286:1571±1577. WHITMAN, W. B., E. ANKWANDA, and R. S. WOLFE. 1982. MOOERS, A. é., and E. C. HOLMES. 2000. The evolution of Nutrition and carbon of Methanococcus voltae. base composition and phylogenetic inference. Trends Ecol. J. Bacteriol. 149:852±863. Evol. 15:365±369. WHITMAN, W. B., T. L. BOWEN, and D. R. BOONE. 1992. The MURRAY, R. G. E. 1992. The family Deinococcaceae. Pp. methanogenic bacteria. Pp. 719±767 in A. BALOWS,H.G. 3732±3744 in A. BALOWS,H.G.TRUÈ PER,M.DWORKIN,W. TRUÈ PER,M.DWORKIN,W.HARDER, and K.-H. SCHLEIFER, HARDER, and K.-H. SCHLEIFER, eds. The prokaryotes. eds. The prokaryotes. Springer-Verlag, New York. Springer-Verlag, New York. WHITMAN, W. B., J. SHIEH,S.SOHN,D.S.CARAS, and U. CHI O , K. 1994. Phylogenetic diversity in the genus Bacillus PREMACHANDRAN. 1986. Isolation and characterization of and comparative AT-L30 analyses of the 22 mesophilic . Syst. Appl. Microbiol. 7:235± genus Thermoactinomyces and relatives. Microbiology 140: 240. 2165±2171. WILLIAMS, R. A. D., and M. S. DA COSTA. 1992. The genus SANDERS, S. W., and R. B. MAXCY. 1979. Isolation of radia- Thermus and related . Pp. 3745±3753 in A. tion-resistant bacteria without exposure to irradiation. Appl. BALOWS,H.G.TRUÈ PER,M.DWORKIN,W.HARDER, and K.- Environ. Microbiol. 38:436±439. H. SCHLEIFER, eds. The prokaryotes. Springer-Verlag, New SAUL, D. J., A. G. RODRIGO,R.A.REEVES,L.C.WILLIAMS, York. K. M. BORGES,H.W.MORGAN, and P. L. BERGQUIST. 1993. WILLIAMS, R. A. D., K. E. SMITH,S.G.WELCH,J.MICALLEF, Phylogeny of 20 Thermus isolates constructed from 16s ri- and R. V. SHARP. 1995. DNA relatedness of Thermus bosomal RNA gene sequence data. Int. J. Syst. Bacteriol. strains, description of Thermus brockianus sp. nov., and 43:754±760. proposal to reestablish Thermus thermophilus (Oshima and SOKAL, R. R., and F. J. ROHLF. 1981. Biometry. 2nd edition. Imahori). Int. J. Syst. Bacteriol. 45:495±499. W. H. Freeman, San Francisco. SPROTT, G. D., I. EKIEL, and G. B. PATEL. 1993. Metabolic JULIAN ADAMS, reviewing editor pathways in Methanococcus jannaschii and other methan- ogenic bacteria. Appl. Environ. Microbiol. 59:1092±1098. Accepted January 4, 2001