Technical Comments

Evidence for a Family of Archaeal ATPases

Science  07 Mar 1997:
Vol. 275, Issue 5305, pp. 1489-1490
DOI: 10.1126/science.275.5305.1489

The analysis by Carol J. Bult et al. of the Methanococcus jannaschii genome included families of paralogous proteins that did not seem to have counterparts in the current sequence databases (1). The largest of such families consists of 13 chromosomal and three plasmid-encoded proteins, which were found to be highly similar to one another [figure 6 in (1)], but did not show statistically significant similarity to any proteins, thus escaping functional prediction. Our inspection of the alignment, however, indicates that two of the conserved sequence blocks correspond to well-characterized functional motifs: namely, the phosphate-binding P-loop and the Mg2+-binding site that are conserved in a vast variety of ATPases and GTPases (Fig. 1 and 24). Even though most commonly used methods for database search such as BLASTP (5) showed only marginally significant similarity to several ATPases, a new version of the BLASTP program that constructs local alignments with gaps (6) indicated a probability of matching by chance between 10−4 and 10−6 for some of the proteins in the new archaeal family and bacterial DnaA proteins; the conservation was particularly notable in the two ATPase motifs (Fig. 1). Thus, even though these 16 proteins comprise a novel family that is so far represented only in archaea, they appear to belong to a known broad class of proteins, and we predict that they possess ATPase activity.

Fig. 1.

Alignment of the three conserved motifs in the novel family of putative archaeal ATPases. Alignment was constructed using the MACAW program (11). Consensus shows amino acid residues conserved in all of the 16 aligned sequences; h indicates a bulky hydrophobic residue (I, L, V, M, F, Y, W); $ indicates serine or threonine. Distances from the protein N-termini and distances between the alignment blocks are indicated by numbers. Fragments of the Bacillus subtilis DnaA protein and Escherichia coli Fur protein are shown for comparison. Two ATPase motifs and the conserved histidine and cysteine residues in the predicted metal-binding site are shown by reversed type. ATPase motif consensus is from (4); <hhh> indicates that three out of five residues preceding the first invariant G in the P-loop and the first D in the Mg2+-binding motif are bulky and hydrophobic. In addition to the proteins shown, open reading frames MJ0819, MJ0820, and MJ0821 appear to represent remnants of a disrupted gene coding for a putative ATPase of the same family.

Screening of the nonredundant protein sequence database at the National Center for Biotechnology Information (National Institutes of Health, Bethesda, MD), with a bipartite pattern representing the specific forms of the two ATPase motifs conserved in the M. jannaschii family—namely, hhhhGx4- GK[TS]xnhhhhD[DE] (h indicates a bulky hydrophobic residue), selected 271 proteins, all of which are either known to possess ATPase activity or are highly similar to ATPases. In addition to DnaA, this list includes a number of members of the so-called AAA ATPase family (7); the similarity between these proteins and DnaA has been noted before (4). Many of the AAA family proteins possess chaperone-like activity and, in particular, are involved in ATP-dependent proteolysis; examples include bacterial proteins ClpA, ClpB, ClpX, FtsH, and HslU; proteasome components; and yeast HSP78 (7). Members of the novel archaeal protein family could also perform chaperone-like functions. This is particularly plausible, because M. jannaschii does not encode several molecular chaperones that are ubiquitous and highly conserved in bacteria and eukaryotes—namely, members of the HSP70, HSP90, and HSP40 families. It remains to be seen how typical is this situation in archaea.

Finally, the family of putative ATPases contains a third strikingly conserved motif with two invariant histidines and one invariant cysteine (Fig. 1). Even though this motif did not show statistically significant similarity to any proteins in the database, this may be a specific metal-binding site, and some resemblance of the divalent cation-binding motif in bacterial Fur proteins that are metal-dependent transcription regulators (8) could be detected (Fig. 1). Two observations seem relevant: (i) One of the chaperone ATPases, FtsH, contains a metal-binding motif conserved in its bacterial and eukaryotic homologs and is a Zn-dependent protease (9). (ii) Methanococcus jannaschii encodes at least two other putative ATPases, namely, the predicted proteins MJ0578 and MJ0579 that also contain a metal-binding domain, in these cases a ferredoxin-like domain (10).

Thus, analysis of conserved motifs and application of additional methods for sequence database search yields specific functional predictions for archaeal proteins that initially appeared to comprise a unique family. There is little doubt that further exploration of the M. jannaschii genome sequence will bring more interesting findings.

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.

Related Content

Navigate This Article