Comparative sequence analysis and predicted phylogeny of the DNA-dependent RNA polymerase beta subunits of Staphylococcus aureus and other eubacteria.

Despite a composition varying from 5 to > 10 subunits, the multimeric DNA-dependent RNA polymerases of eubacteria, chloroplasts, archaebacteria, and the eukaryotic nucleus form a homologous protein family. This is particularly evident in sequence conservation of the largest (typically > 100 kDa) subunits, equivalent to the p and p ' subunits of the eubacterial enzyme [reviewed in I]. The rpoRCgenomic region, encoding the p and p ' subunits of RNA polymerase, was cloned from the Grampositive eubacteriun Staphylococcus aureus [2] and the nucleotide sequence of the rpoB gene determined. The rpol3 coding sequence was preceded by an extended ribosome binding site and possessed a TTG initiation codon, both expected features for low(GtC) Grampositive bacteria [ 3 ] . Codon usage was similar to that of other S. aureus genes [ 3 ] . The predicted S. aureus p subunit consisted of 1182 amino acids with a size of 133 kDa, in reasonable agreement with an estimate of 150 kDa from SDS PAGE [ 4 ] . The S. aureus p subunit sequence was aligned with other p-homologues using Clustal V software [ 5 ] . AS expected, the most extensive sequence similarity was to other eubacterial $ subunits, although some 13 colinear 'domains' were recognizably conserved in all sequences. These included regions implicated in the polymerization reaction, substrate binding, and interaction with the transcription inhibitors rifampicin and streptolydigin [ 6 ] . Amino acid substitution frequencies [7] were calculated from the aligned conserved regions (592 residues per sequence) and a predicted phylogenetic tree was generated using the Neighbor-Joining (NJ) distance-matrix algorithm [8] with 'bootstrap' resampling [9]. This tree (Fig. 1) was consistent with the expected evolutionary relationships between monocot, dicot, bryophyte, and algal chloroplasts, and between enterobacteria and fluorescent pseudomonads, indicating that the method was reliable, at least over these regions of the tree. The relatively-close evolutionary distance between the eubacterial and chloroplast sequences was consistent with the 'endosymbiont' origin of chloroplasts. A specific relationship of S. aureus (Gram-positive) with the proteobacteria (Gramnegative) was indicated ('bootstrap' frequency, f = 1.0). rather than the Gram-positive-cyanobacterial/ chloroplast grouping suggested from analysis of rRNA sequences [lo]. However, this may be a consequence of a more rapid rate of sequence divergence resulting in artifactual clustering of the cyanobacterialchloroplast lineage with the distantly related archaebacterial-eukaryotic group. In common with most other studies [ lo-13 , cf 141. a monophyletic origin of the archaebacteria was favoured, albeit only weakly (f = 0.54). Within the archaebacterial cluster, Sulfolobus formed an outgroup to the Xalobacterium-Methanobacterium pair (f = 0.91) [10-12, cf 131. This is consistent with the unique 'halophile-methanogen' characteristic that RpoBrelated sequences are divided between 8' and 8'' subunits [ I l l . In contrast with the proposed 'mozaic' origin of the eukaryotic nucleus [ I l l , but in agreement with other studies using RpoC-related sequences [ 1 2 , 1 3 ] , this analysis favoured a monophyletic origin for the GEOFFREY C. ROWLAND, MOHAMED ABOSBKIWA I r i D. mlamgaster pol I1