Evolution of DNA polymerases: an inactivated polymerase-exonuclease module in Pol epsilon and a chimeric origin of eukaryotic polymerases from two classes of archaeal ancestors

Abstract

Background: Evolution of DNA polymerases, the key enzymes of DNA replication and repair, is central to any reconstruction of the history of cellular life. However, the details of the evolutionary relationships between DNA polymerases of archaea and eukaryotes remain unresolved.

Results: We performed a comparative analysis of archaeal, eukaryotic, and bacterial B-family DNA polymerases, which are the main replicative polymerases in archaea and eukaryotes, combined with an analysis of domain architectures. Surprisingly, we found that eukaryotic Polymerase epsilon consists of two tandem exonuclease-polymerase modules, the active N-terminal module and a C-terminal module in which both enzymatic domains are inactivated. The two modules are only distantly related to each other, an observation that suggests the possibility that Pol epsilon evolved as a result of insertion and subsequent inactivation of a distinct polymerase, possibly, of bacterial descent, upstream of the C-terminal Zn-fingers, rather than by tandem duplication. The presence of an inactivated exonuclease-polymerase module in Pol epsilon parallels a similar inactivation of both enzymatic domains in a distinct family of archaeal B-family polymerases. The results of phylogenetic analysis indicate that eukaryotic B-family polymerases, most likely, originate from two distantly related archaeal B-family polymerases, one form giving rise to Pol epsilon, and the other one to the common ancestor of Pol alpha, Pol delta, and Pol zeta. The C-terminal Zn-fingers that are present in all eukaryotic B-family polymerases, unexpectedly, are homologous to the Zn-finger of archaeal D-family DNA polymerases that are otherwise unrelated to the B family. The Zn-finger of Polepsilon shows a markedly greater similarity to the counterpart in archaeal PolD than the Zn-fingers of other eukaryotic B-family polymerases.

Conclusion: Evolution of eukaryotic DNA polymerases seems to have involved previously unnoticed complex events. We hypothesize that the archaeal ancestor of eukaryotes encoded three DNA polymerases, namely, two distinct B-family polymerases and a D-family polymerase all of which contributed to the evolution of the eukaryotic replication machinery. The Zn-finger might have been acquired from PolD by the B-family form that gave rise to Pol epsilon prior to or in the course of eukaryogenesis, and subsequently, was captured by the ancestor of the other B-family eukaryotic polymerases. The inactivated polymerase-exonuclease module of Pol epsilon might have evolved by fusion with a distinct polymerase, rather than by duplication of the active module of Pol epsilon, and is likely to play an important role in the assembly of eukaryotic replication and repair complexes.

Figure 1

The conserved motifs of exonuclease and polymerase catalytic domains of active B-family polymerases compared to inactivated C-terminal domains of polymerases ε. The motifs are represented as four sequence LOGOs, from top to bottom: all active Exo domains of B-family polymerases from the alignment in Additional File 1 (archaeal, proteobacterial, and eukaryotic Polδ and N-terminal domain of Polε); inactivated C-terminal domain of Polε; all active Pol domains; inactivated C-terminal domain of Polε. The motifs that contribute to the active centers are denoted Exo I-III and Region I-III for the Exo and Pol domains, respectively, and the catalytic residues are shown by #.

Figure 2

A schematic diagram of conserved blocks and specific inserts in the most conserved part of the alignment of polymerase catalytic domain of different groups of B-family DNA polymerases. For the actual alignment, see Additional File 1.

Figure 3

Multiple alignment of the two-Zn-finger modules of eukaryotic Pol α, ζ, δ, and ε, and the single Zn-finger of archaeal PolD. The sequences are denoted by their GI numbers and species names. The positions of the first and the last residues of the aligned region in the corresponding protein are indicated for each sequence. The numbers within the alignment represent poorly conserved inserts that are not shown. The cysteine residues that are essential for Zn-binding are shown by reverse shading. The coloring is based on the consensus shown underneath the alignment; 'h' indicates hydrophobic residues (ACFILMVWY), 'p' indicates polar residues (EDKRNQHTS). Additional consensus line at the top of archaeal polymerase II alignment indicates additional conservation between polymerase ε and archaeal polymerase II: 's' indicates small residues (ACDGNPSTV). The predicted secondary structure is shown above the alignment and is compared to the NMR structure that is available for human Pol α (pdb: 1N5G) [67] that is shown on top of the Pol α alignment; 'H' indicates α-helix, 'E' indicates extended conformation (β-strand) and 'T' indicates a turn.

Figure 4

Unrooted phylogenetic tree of B-family DNA polymerases. The tree was constructed using the conserved blocks from the Exo-Pol alignment (see Additional File 1). The tree is rendered as a scheme, with only the major groups denoted; for the complete tree, with all species indicated, and trees constructed with alternative methods, see Additional File 2. The tree is overlaid with schematics of domain architectures which are given for representatives of each group (Saccharomyces cerevisiae sequences for polymerases α, ζ, δ, ε and those from Sulfolobus solfataricus, Pyrococcus furiosus (pdb:2JGU), and Escherichia coli (pdb:1Q8I) for archaeal minor, archaeal major and proteobacterial groups, respectively). The domains are shown roughly to scale. Inactivated C-terminal domains of polymerases ε are crossed. Dashed line indicates the portion of the sequences corresponding the adjacent tree branch. Zn-finger 2* denotes the distinct version of this module in Pole that is highly similar to the Zn-finger of archaeal PolD (see text for details). The proposed position of the root is shown by an arrow.

Figure 5

A putative evolutionary scenario for the origin of eukaryotic B-family DNA polymerases from prokaryotic ancestral forms. The scheme is rendered within the framework of the symbiotic scenario of the origin of eukaryotes whereby the symbiosis of an archaeon with an α-proteobacterium gave rise to the mitochondrion and triggered eukaryogenesis. The domains are designated by unique shapes as in Figure 4. PolBM, the "major" form of archaeal B-family DNA polymerase (PolBI [43]); PolBm, "minor" form of archaeal B-family DNA polymerase (PolBII [43]; PolDs, small subunit of archael PolD (active exonuclease). Inactivation of PolDs in the protoeukaryote (the Last Eukaryotic Common Ancestor, LECA) is denoted by crosses. The origin of Pol ε is depicted as insertion of a bacterial B-family polymerase between the catalytically active module derived from the archaeal PolB-M and the Zn-finger derived from the archaeal PolD.