New insights into the structures and interactions of bacterial Y-family DNA polymerases

Abstract

Bacterial Y-family DNA polymerases are usually classified into DinB (Pol IV), UmuC (the catalytic subunit of Pol V) and ImuB, a catalytically dead essential component of the ImuA-ImuB-DnaE2 mutasome. However, the true diversity of Y-family polymerases is unknown. Furthermore, for most of them the structures are unavailable and interactions are poorly characterized. To gain a better understanding of bacterial Y-family DNA polymerases, we performed a detailed computational study. It revealed substantial diversity, far exceeding traditional classification. We found that a large number of subfamilies feature a C-terminal extension next to the common Y-family region. Unexpectedly, in most C-terminal extensions we identified a region homologous to the N-terminal oligomerization motif of RecA. This finding implies a universal mode of interaction between Y-family members and RecA (or ImuA), in the case of Pol V strongly supported by experimental data. In gram-positive bacteria, we identified a putative Pol V counterpart composed of a Y-family polymerase, a YolD homolog and RecA. We also found ImuA-ImuB-DnaE2 variants lacking ImuA, but retaining active or inactive Y-family polymerase, a standalone ImuB C-terminal domain and/or DnaE2. In summary, our analyses revealed that, despite considerable diversity, bacterial Y-family polymerases share previously unanticipated similarities in their structural domains/motifs and interactions.

Figure 1.

Clustering series of full-length sequences of Y-family polymerases based on their pairwise similarities. Each dot represents a single sequence and connections between them represent pairwise similarities. Results (A–C) are shown at three different similarity cutoffs defined by P-values. The number of sequences is the same in all three cases. Discerned polymerase groups are denoted by different colors and some are labeled. ImuB1 is represented by Pseudomonas aeruginosa ImuB (accession: AAG04059.1); ImuB2, Mycobacterium tuberculosis ImuB (CCP46215.1); UmuC, Escherichia coli UmuC (AAC74268.1); ‘ϵ-prot PolIV-like’, Arcobacter nitrofigilis PolIV (ADG92494.1); ‘acPolY2-like’, Acidobacterium capsulatum PolY2 (ACO33562.1); YqjH-like, Mycoplasma hominis MucB (CAX37259.1); YqjH, YqjW and UvrX, Bacillus subtilis PolY1 (CAB14319.2), PolY2 (CAB14303.1) and UvrX (CAB14068.2), respectively; UvrX-like, Bifidobacterium bifidum PolY (ADP36361.1); DinX and DinP, M. tuberculosis DinX (CCP44301.1) and DinP (CCP45865.1) respectively; ‘msDinB3-like’, Mycobacterium smegmatis DinB3 (ABK74774.1); ‘Bacteroid PolY2’, Beliella baltica DinB (AFL82914.1); ‘scDinB2-like’, Streptomyces coelicolor DinB2 (CAB50953.1); ‘gbDinB2-like’, Geobacter bemidjiensis DinB2 (ACH37529.1); PolY-core, E. coli DinB (AAC73335.1).

Figure 2.

Distribution of DUF4113-containing C-terminal extensions among Y-family polymerases. Clustering snapshot is the same as in Figure 1C. Group labeling corresponds to Figure 1. Sequences are colored based on DUF4113 detection results. Red and orange colors indicate highly confident and less confident match to DUF4113, respectively; blue color indicates sequences having C-terminal regions without DUF4113; gray color indicates sequences that do not have additional C-terminal extensions.

Figure 3.

Schematic representation of domain architectures of representative Y-family DNA polymerases (Eco, E. coli; Bsu, B. subtilis; Mtu, M. tuberculosis). Bsu UvrX has identical domain composition as YqjW, their conserved C-terminal motif is denoted as a dark green shape.

Figure 4.

Sequence logos for the RecA-NT motif and corresponding motifs of UmuC (Escherichia coli) and groups with representatives in Bacillus subtilis. Highly similar motifs of YqjW, UvrX and UvrX-like polymerases were merged into a single logo. Secondary structure (α-β) for RecA-NT is depicted above.

Figure 5.

Comparison of Escherichia coli RecA-NT and UmuC-CT complexes with RecA. (A) Schematic representation of RecA filament on ssDNA. Arrow indicates the only free interaction site on the RecA surface at the 3′-end of RecA filament. (B) Schematic and structural comparison of the two complexes. (C) Interface energetic scores depicted as colored surfaces for RecA bound to RecA-NT (left), not bound to any motif (middle) and bound to UmuC-CT (right). Blue and red colors correspond to estimated favorable-unfavorable surface energy. (D) Zoomed-in interaction interface between E. coli UmuC-CT and RecA. The corresponding part of the structure is indicated with a dashed line in part (B). UmuC-CT residues S370, R367 and Q372 that have been shown to crosslink to the RecA position 113 are highlighted. The RecA S117 residue located at the interface is also highlighted.

Figure 6.

Phylogenetic tree of YqjH(-like), YqjW and UvrX(-like) sequences. Typical operons of selected representatives are shown with corresponding accession numbers.

Figure 7.

Comparison of Escherichia coli UmuD′ (left, PDB ID: 1AY9), Bacillus subtilis YolD model (middle) and cyanobacterial (Nostoc sp.) Hfq protein (right, PDB ID: 3HFN). Common structural parts are colored green.

Figure 8.

Typical PolY operons with DnaE2 and/or ImuB-C in representative organisms (corresponding PolY accession ID indicated in parentheses): Pae, Pseudomonas aeruginosa (AAG04059.1); Mtu, Mycobacterium tuberculosis (CCP46215.1); Ktu, Kyrpidia tusciae (ADG05750.1); Gsu, Geobacter sulfurreducens (AAR33378.1); Sco, Streptomyces coelicolor (CAB50953.1); Cal, Cellulophaga algicola (ADV50500.1). RecA-NT-like motifs of PolYs are indicated with orange rectangles. Inactive polymerases are marked with pink ‘X’.

Figure 9.

Comparison of Tudor domains from three proteins: (A) Modeled Tudor-like domain of Mycobacterium tuberculosis DinX; (B) Tudor domain of Geobacillus stearothermophilus PcrA/UvrD helicase (PDB ID: 5DMA); (C) Tudor domain of CarD bound to RNAP β subunit (light blue) (PDB ID: 4KBM). Common parts are highlighted in green. Conserved residues determined to be important for PcrA/UvrD binding to RNAP are labeled and their side chains are shown. Side chains of residues in corresponding positions of other structures are also shown.

Figure 10.

Schematic comparison of known and proposed complexes and interactions of three mutasomes: (A) Escherichia coli PolV, (B) Bacillus subtilis YqjW (also represents UvrX) and (C) Mycobacterium tuberculosis ImuA-ImuB-DnaE2. Components with similar structures are shown in the same color. Wide black arrows indicate either known interactions or interactions predicted with high confidence and supported by structural models, gray-contoured arrows indicate putative interactions without support of structural models, and narrow black arrows denote transient interactions with the β-clamp. Green star and red X indicate correspondingly the presence and the absence of the polymerase active site.