Aminoacyl-tRNA synthetases (aaRSs) specific for 17 of the 20 amino acids are universally present in cellular life forms. The three exceptions are GlnRS, AsnRS and CysRS. GInRS and AsnRS are missing in many bacteria and archaea because glutamine and asparagine are incorporated into proteins through transamidation of glutamate and aspartate, respectively. CysRS is missing in two archaeal methanogens whose genomes have been sequenced - Methanobacterium thermoautotrophicum and Methanococcus jannaschii [12]. No alternative mechanism for cysteine incorporation into proteins is known; hence the absence of CysRS in these organisms was an enigma.

Two solutions to this puzzle, both unusual, have recently been proposed and experimentally validated. One involves non-orthologous gene displacement, a situation in which the same essential function is carried out by distantly related or even unrelated proteins in different organisms [13,14]. It has been shown that M. jannaschii ProRS, a class II synthetase that is unrelated to the class I CysRS, substituted for the missing CysRS activity [15,16,17]. The other solution involved a new candidate for the role of CysRS, the MJ1477 protein from M. jannaschii. This protein and its orthologs (direct evolutionary counterparts related by vertical descent from a common ancestor) from the bacteria Thermotoga maritima and Deinococcus radiodurans were identified as 'distant orthologs' of the Bacillus subtilis CysRS by using a computational method specifically designed to detect distantly related orthologs [18]. The method is based on application of discriminant analysis to alignment scores, in order to separate the scores for pairs of functionally identical proteins from different genomes from the scores for proteins with different functions. This prediction was then validated experimentally by showing that MJ1477 had CysRS activity in vitro and that an ortholog of MJ1477 from D. radiodurans, DR0705, complemented a CysRS deficient, temperature-sensitive, lethal E. coli mutant strain [18]. An important corollary of these surprising findings is a rapid divergence of the MJ1477 family from CysRS, such that all the catalytic and otherwise functionally important residues characteristic of this enzyme, and also present in other class I aaRSs, have changed. Furthermore, MJ1477 and its orthologs do not have the accessory domains found in all known CysRS, namely the DALR domain (named after a distinct amino-acid signature), which is shared by aaRSs of several specificities, and another domain specific to CysRS [19].

We examined the protein sequences of MJ1477 and its homologs using more traditional computational techniques. Almost all these proteins contain amino-terminal signal pep-tides readily identifiable by using the SignalP program [20], but do not contain any predicted transmembrane segments, and, accordingly, are predicted to be secreted from the cells (Figure 1). Furthermore, iterative database searches using the PSI-BLAST program [9] showed statistically significant sequence similarity between these proteins and an experimentally characterized endo α-1,4-polygalactosaminidase from Pseudomonas species [21]. For example, in a search initiated with the sequence of MJ1477 and a profile inclusion cut-off of 0.01, the polygalactosaminidase sequence was retrieved from the database in the second iteration, followed by other bacterial proteins predicted to possess the same activity. This protein family has several conserved motifs, including a characteristic Dxhp signature (h, hydrophobic residue; p, polar residue), in which the conserved aspartate is likely to directly participate in catalysis (Figure 1). The hybrid-fold-recognition method, which combines sequence-profile analysis with alignment-based secondary-structure prediction [22] and the 3D-PSSM method [23] both suggested a likely α-amylase-like triosephosphate isomerase (TIM) barrel structure for this protein family. Thus, although the identification of MJ1477 as a secreted polygalactosaminidase or a related polysaccharide hydrolase with a different specificity awaits experimental verification, it shows all the signs of a correct computational prediction: statistically significant similarity between the analyzed protein and an experimentally characterized enzyme; conservation of distinct motifs implicated in catalysis; potential presence of a structural fold compatible with the experimentally demonstrated enzymatic activity; and confident prediction of the extracellular localization that is, again, compatible with a polysaccharide hydrolase activity involved in environmental carbohydrate utilization or capsular metabolism. None of this evidence is offered by the analysis that led to the CysRS prediction for MJ1477.

Therefore we are forced to conclude that MJ1477 and its homologs are not related to CysRS and there is nothing in the computational analysis of these proteins that would point to an aaRS activity. In contrast, we predict these proteins to be extracellular polygalactosaminidases or similar polysaccharide hydrolases. The polysaccharide hydrolase and aaRS functions seem to be essentially incompatible. First, a secreted enzyme is unlikely to function as an aaRS whose site of action is, by definition, intracellular. Second, even if an entirely new class of aaRSs is postulated, the reaction catalyzed by this new aaRS does not resemble polysaccharide hydrolysis or its reversal. Aminoacyl-tRNA synthetases catalyze a succession of reactions, which involve: hydrolysis of the α-β phosphate bond in ATP; condensation of AMP with the cognate amino acid, resulting in the formation of an aminoacyl-adenylate; displacement of the AMP moiety of the aminoacyl-adenylate with the cognate tRNA, producing aminoacyl-tRNA. Even if the two condensation reactions, in very general terms, could be considered a reversal of the polysaccharide hydrolysis reaction, there is no indication that polysaccharide hydrolases could bind and hydrolyze ATP, and the multiple alignment of the MJ1477 family did not include any conserved signatures typical of potential phosphate-binding loops (Figure 1). Neither does this family contain any recognizable RNA-binding domains. Finally, M. thermoautotrophicum does not encode any homologs of MJ1477, ruling out the possibility that this family encompasses CysRS of both archaeal methanogens. Taken together, these observations appear to effectively refute the prediction of a CysRS activity, thus pitting computational results against experimental data.

Dihydropteroate synthase (DHPS) catalyzes the condensation of p-aminobenzoic acid with 7,8-dihydro-6-hydroxymethylpterin pyrophosphate to give 7,8-dihydropteroate, an intermediate in folate metabolism. The protein from Staphylococcus, a Gram-positive bacterium, has been crystallized and shown to adopt a TIM-barrel structure [24]. Although it has been indicated that no DHPS could be detected in archaeal genomes [25], orthologs of bacterial DHPS are readily identifiable in all archaea; this enzyme is missing only in animals and in several intracellular bacterial pathogens, such as Rickettsia prowazekii, spirochetes and mycoplasmas (COG0294 in the database of Clusters of Orthologous Groups of proteins (COGs)) [26]. Most archaea have a distinct version of DHPS that shows relatively low sequence similarity to the bacterial orthologs and contains an additional uncharacterized carboxy-terminal domain. This previously undetected domain is also present in some other enzymes of pterin biosynthesis, such as tetrahydromethanopterin-S-methyltransferase from Streptomyces (L.M.I., LA. and E.V.K., unpublished observation). Some archaeal species, including Thermoplasma and Halobacterium, have the bacterial-type DHPS, which was probably acquired by horizontal gene transfer and displaced the original archaeal version. Despite the relatively low sequence similarity to bacterial DHPS, all archaeal orthologs have the conserved catalytic residues identified in DHPS (Figure 2) and are confidently predicted, by the hybrid-fold-recognition method, to assume the same fold as DHPS from Pneumocystis carinii and Staphylococcus aureus whose crystal structures have been determined.

An analysis using ORF, a program developed to recognize folds by comparing predicted secondary structures of proteins ([27]; we are unaware of a published detailed description of this method), identified MJ0301 as a homolog of DHPS, although, given the low sequence similarity, a convergent origin of the relationship between MJ0301 and DHPS was deemed likely (there seems to be a terminological confusion involved here, but we are quoting the results of the original computational analysis of this protein as they have been presented). It was acknowledged that MJ0107 (a member of COG0294) could be identified as a possible homolog of DHPS by sequence-based methods, and this protein was assayed for dihydropteroate synthase activity, but none was detected [25]. In contrast, DHPS activity (albeit relatively low) was shown in vitro for the partially purified MJ0301 protein [25]. However, MJ0301 has been shown to belong to the metallo-β-lactamase superfamily of enzymes and, in the evolutionary classification of metallo-β-lactamases, belongs to an archaea-specific family (Figure 2; COG1237) [28]. Metallo-β-lactamases encompass a wide range of metal-dependent hydrolytic and oxidoreductase activities with a variety of substrates and are particularly abundant in archaea where some of them are involved in RNA processing [28]. None of these enzymes catalyzes a reaction resembling the condensation reaction catalyzed by DHPS. The characteristic motifs of metallo-β-lactamases, which mostly include metal-binding histidines, are highly conserved in MJ0301 and its orthologs (Figure 3). In contrast, most of the MJ0301 residues described as equivalent to the functionally important residues of Escherichia coli dihydropteroate synthase are not conserved, even among the archaeal orthologs of this protein. Finally, the β-lactamase fold consists of two subdomains of the β4-α-β-α topology whose β sheets are sandwiched against each other; in structural terms, these domains are completely different from the TIM-barrel, with which the ORF program matched the MJ0301 structural prediction. Taken together, these observations are sufficient to reject the proposed relationship between MJ0301 and dihydropteroate synthases.

Thymidylate synthase is a central enzyme of pyrimidine metabolism that catalyzes the formation of deoxythymidine monophosphate (dTMP) from deoxyuridine monophosphate (dUMP) by transfer of a methyl group to its pyrimidine ring. This reaction is catalyzed by at least two unrelated enzymes. The canonical thymidylate synthase (TS), such as the E. coli ThyA, is a protein with a distinct α/β-fold that transfers a methyl group to dUMP from 5,10-methylenetetrahydrofolate [29]. This classic TS is readily identifiable in many (but not all) bacteria, eukaryotes and three archaeal species, Archaeoglobus fulgidus, M. jannaschii, and M. thermoautotrophicum (COG0207). The archaeal members of the TS family share with their bacterial orthologs all the conserved residues involved in catalysis (Figure 4).

An alternative TS or its subunit is predicted to be encoded by a gene from Dictyostelium that rescues a slime mold mutant auxotrophic for thymidylate [30]. This protein is not homologous to the canonical TS, but its orthologs in bacteria and archaea show an almost perfect complementary phyletic distribution (COG1351).

In a screen for the TS in M. jannaschii, the ORF method picked the MJ0757 protein as the most likely homolog of the canonical TS family [27]. In the validation experiment, MJ0757 overexpressed in E. coli was shown to possess TS activity [25]. Sequence searches show that MJ0757 belongs to a small family of euryarchaea-specific proteins of uncharacterized function (COG1810). Of the 17 residues reported to be conserved between MJ0757 and the TS family, only seven were conserved thoughout the MJ0757 family (Figure 5). Moreover, a comparison of the secondary structure elements derived from the reported three-dimensional model of MJ0757 [27] and those derived from a prediction generated using a multiple alignment query with the structure-prediction program PHD (such predictions typically exceed 70% accuracy), showed an overlap of just two of the 16 or so secondary structural elements (Figure 5). Conversely, several sequence motifs that are characteristic of the MJ0757 family did not overlap with the conserved regions in the MJ0757-TS alignment (Figure 5). Furthermore, some, but not all, members of the MJ0757 family contain an amino-terminal insertion of a small, metal-chelating module (Figure 5), which was used to improve the alignment with the E. coli TS [25], although this region was variable even within the MJ0757 family itself. On the basis of these observations, a relationship between MJ0757 and the canonical TS has to be rejected. The actual fold and function of MJ0757 and its homologs cannot be predicted at present. However, these proteins have several features that suggest that they might be metal-dependent enzymes potentially involved in redox reactions. These suggestive features include the fusion with a ferredoxin domain seen in the M. thermoautotrophicum member MTH601, the insertion of the metal-binding module in certain members, including MJ0757 (see above), and the presence of three cysteines that are conserved throughout this family.

Viral movement proteins (MPs) are encoded by diverse, unrelated families of plant viruses, such as positive-strand RNA, negative-strand RNA, single-stranded DNA and double-stranded DNA viruses, and are essential for cell-to-cell movement of all these viruses [31,32]. To isolate potential host homologs of the red clover necrotic mosaic virus (RCNMV) MP, antibodies to this protein were used to screen phloem extracts of Cucurbita maxima, resulting in the detection of a protein designated Cmpp16. This protein was identified as a 'paralog' (generally, this term refers to homologous genes related by duplication within the same genome) of the viral MPs on the basis of sequence similarity detected using the Megalign program [33]. Subsequently, Cmpp16 was shown to bind RNA, which is a common property of viral MPs, and to induce an increase of the size-exclusion limit of plasmodesmata, also a mechanism associated with the MPs [33].

However, computational analysis of the Cmpp16 sequence reveals a picture that is incompatible with a homologous relationship with MPs. Cmpp16 consists mostly of a C2 domain that is readily detected by PSI-BLAST or by profile-searching engines such as the CD-search [34]. The Cmpp16 sequence contains all critical residues of the C2 domain (Figure 6). C2 domains bind a variety of substrates, such as Ca²⁺, phospholipids, inositol polyphosphates and other proteins, but apparently not RNA [35]. There is no detectable similarity between C2 domains and the MPs, and conserved motifs in the published alignment of Cmpp16 and the RCNMV MP do not correspond to those in C2 domains; moreover, many of the residues described as conserved in Cmpp16 and MP are not conserved within the viral movement protein family itself. Thus, we conclude that viral MPs and Cmpp16, a C2-domain protein, are not homologs. Subsequently, a similar methodology has been employed to detect a relationship between Cmpp36 (a cytochrome B5 reductase), Cmpp16 and the RCNMV movement protein [36]. As in the above case of Cmpp16, this relationship of a cytochrome B5 reductase with the viral movement proteins appears to be spurious (data not shown).

Histone acetyltransferases (HAT) are key regulators of eukaryotic transcription. GCN5-like HATs, which modulate chromatin-associated transcription, belong to a vast superfamily of amino-group acetyl- and myristoyl-transferases with extremely diverse functions [37]. ATF-2 is a basic leucine zipper (b-ZIP) family transcription factor that binds to cyclic AMP-response elements (CRE) and activates transcription [38]. Vertebrate ATF-2 also has an amino-terminal zinc finger, which is involved in transcription activation [39]. Non-vertebrate orthologs of ATF-2, in Drosophila, Caenorhabditis elegans and yeasts, lack the zinc finger. In experiments designed to isolate ATF-2-associated HAT, ATF-2 alone was shown to be sufficient for the acetyltransferase activity. Examining the region of ATF-2 that showed HAT activity, the authors found some sequence similarity and at least one motif resembling the acetyltransferase superfamily and concluded that ATF-2 contained a GCN5-like acetyltransferase domain [40]. Subsequent site-directed mutagenesis supported the importance of the reported acetyltransferase motifs for the HAT activity of ATF-2.

However, profile-based sequence searches and attempts at fold recognition failed to detect any relationship between ATF-2 and the acetyltransferase superfamily. The region designated as having HAT activity and containing the acetyltransferase domain shows poor conservation between orthologs and closely related paralogs of the ATF-2 family, especially in the sequence identified as the most prominent A motif of the acetyltransferase family (Figure 7). Furthermore, complexity analysis using the SEG program, with the parameters adjusted for decomposition of a protein into globular and non-globular regions [41], predicted that the entire region of the ATF-2 protein between the amino-terminal zinc finger and the carboxy-terminal helical b-ZIP was unstructured. This is consistent with the structural prediction derived using the PHD program that indicated no regular secondary structure in this region. Thus, the relationship between ATF-2 and the GCN5-like acetyltransferase superfamily seems to be invalid, leaving the structural basis for the reported acetyltransferase activity of ATF-2 an open issue.

PAS domains are sensory modules in various signal transduction proteins from all major lineages of cellular life [42]. PAS domains are typically implicated in sensing oxygen, redox potential, light and small ligands [43]. In addition, PAS domains are sites for protein-protein interactions and are responsible for the formation of homo- and hetero-dimers in several signal transduction pathways that involve transcriptional activation. A PAS domain has been reported in the transcription factor PIF3 from Arabidopsis, which interacts with a phytochrome photoreceptor and transduces light signals to photoresponsive plant genes [44]. It has been hypothesized that the purported PAS domain of PIF3 directly interacts with the PAS domains of the phytochrome [44]. This hypothesis was later tested experimentally and evidence was presented that the PAS domain of PIF3 indeed was a major contributor to the interaction between the two proteins [45].

PIF3 belongs to a plant-specific family of basic helix-loop-helix (bHLH)-domain- containing proteins that, in addition to the bHLH domain, have an uncharacterized conserved domain at the amino terminus present in single or duplicate copies (L.M.I., I.Z., L.A. and E.V.K., unpublished observations). The PIF3 family currently consists of about eight paralogous proteins in Arabidopsis and an ortholog from rice. The region predicted to be a PAS domain is poorly conserved in the rice ortholog of PIF3 and the paralogs from Arabidopsis. An alignment with the rice ortholog indicated that the proposed PAS domain was a rapidly diverging, compositionally biased sequence (Figure 8). Complexity analysis using the SEG program showed that the reported PAS domain mapped to a region that was predicted to be entirely non-globular. All attempts to objectively detect a PAS domain in PIF3 using sensitive profile methods based on PSI-BLAST-derived scoring matrices or Hidden Markov Models (HMM) failed. Additionally, secondary-structure prediction for the proposed PAS region using PHD indicated that this region is largely unstructured. These observations appear to be sufficient to reject the presence of a PAS domain in PIF3 although the region thought to be a PAS domain could indeed be involved in the interaction with phytochrome.