Eukaryotic DNA ligases: structural and functional insights

Abstract

DNA ligases are required for DNA replication, repair, and recombination. In eukaryotes, there are three families of ATP-dependent DNA ligases. Members of the DNA ligase I and IV families are found in all eukaryotes, whereas DNA ligase III family members are restricted to vertebrates. These enzymes share a common catalytic region comprising a DNA-binding domain, a nucleotidyltransferase (NTase) domain, and an oligonucleotide/oligosaccharide binding (OB)-fold domain. The catalytic region encircles nicked DNA with each of the domains contacting the DNA duplex. The unique segments adjacent to the catalytic region of eukaryotic DNA ligases are involved in specific protein-protein interactions with a growing number of DNA replication and repair proteins. These interactions determine the specific cellular functions of the DNA ligase isozymes. In mammals, defects in DNA ligation have been linked with an increased incidence of cancer and neurodegeneration.

Figure 1

Enzymatic ligation of DNA. The three-step reaction catalyzed by DNA ligases (Ligs) results in the serial transfer of AMP (adenosine 5′-monophosphate) to an active-site lysine (step 1) then to the 5′-PO₄ end of DNA (step 2). During step 3, the 3′-OH end of a second DNA strand attacks the 5′-PO₄ to release AMP and generate the ligated DNA product. Eukaryotic and archaeal DNA ligases use an ATP cofactor, whereas most bacterial ligases use NAD⁺.

Figure 2

Comparative anatomy of DNA ligases. Examples of viral, bacterial, and mammalian DNA ligases are shown bound to DNA. The DNA ligases from Chlorella virus PCBV-1, Protein Data Bank (pdb) 2q2t, Escherichia coli (LigA protein; pdb 2owo), and human ligase I (pdb 1×9n) have a similar ring-shaped structures in complex with a nicked DNA substrate. A conserved catalytic core, consisting of the NTase domain (green) and OB-fold domain (yellow), contacts the 3′-OH and 5′-PO₄ ends of DNA during steps 2 and 3 of the end-joining reaction (cf. Figure 1). An additional domain (red) in each enzyme interacts with DNA, completing the ring-shaped structure. The Chlorella ligase has an insertion in the OB-fold domain, termed the latch, which is disordered in the absence of DNA and bridges between the two domains of the catalytic core (16). E. coli LigA has a C-terminal helix-hairpin-helix (HhH) (25, 35) domain engaging the DNA that is functionally analogous to the N-terminal DNA-binding domain (DBD) (26) of human DNA ligase I. A small N-terminal motif in the LigA protein is important for sensing the NAD⁺ cofactor (25). Abbreviations: OB-fold, oligonucleotide/oligosaccharide binding fold; NTase, nucleotidyltransferase.

Figure 3

The “jackknife model” of nick sensing by ligase III. (a) DNA ligase III has an additional zinc finger (ZnF) that binds to nicks and gaps in DNA (b), stimulating DNA nick-joining by ligase III (c). The ZnF domain added in trans inhibits ligation by the catalytic core of ligase III, consistent with a competitive binding interaction. The ZnF and adjacent DNA-binding domain (DBD) constitute one DNA-binding surface of ligase III, and the catalytic core consisting of the nucleotidyltransferase (NTase) and OB-fold domains constitutes a second DNA-binding moiety. The jackknife model posits a handoff of the nicked DNA substrate from the ZnF-DBD to the catalytic core through protein motions resembling the opening and closing of blades of a jackknife. In this model, the ZnF-DBD initially recruits ligase III to an interruption in the DNA backbone, and then the DNA ends are engaged by the NTase-OB-fold catalytic core for the ligation reaction. (d) The ZnF also contributes strongly to the intermolecular ligation of linear DNA molecules. In this case, the two independent DNA-binding surfaces of ligase III (ZnF-DBD versus NTase-OB-fold) could align two DNAs for ligation (157). Abbreviations: AdD, adenylation domain; OB-fold, oligonucleotide/oligosaccharide binding fold.

Figure 4

DNA ligase I family. Alignment of the DNA ligases encoded by the Saccharomyces cerevisiae CDC9 and human LIG1 genes. The DNA-binding domain (DBD, red) and the catalytic core composed of the adenylation and OB-fold domains (gray) and the nuclear localization signals (NLS) (yellow) are indicated. The position of the active-site lysine residues (K419, K442, K568), within the adenylation domain that form the covalent bond with AMP are shown. Two polypeptides are generated from the CDC9 gene by alternative translation initiation. The longer polypeptide has an N-terminal mitochondrial (mito) leader sequence (MLS, turquoise). The N-terminal region of human DNA ligase I (hLigI) contains a replication factory-targeting sequence (RFTS, purple) and four serine (Ser) residues, which are phosphorylated, in addition to the NLS (–44). Abbreviations: aa, amino acids.

Figure 5

DNA ligase III family. Alignment of the DNA ligases encoded by vaccinia virus and human LIG3. The DNA-binding domain (DBD, red), catalytic core composed of the adenylation and OB-fold domains (gray), and nuclear localization signals (NLS, black) are indicated. The position of the active-site lysine residues (K508, K421, K231) within the adenylation domain that form the covalent bond with AMP are shown. Multiple DNA ligase polypeptides are encoded by the LIG3 gene. The N-terminal region of each of the polypeptides contains a zinc finger (ZnF, orange) and a serine residue (Ser123 or Ser210) that is phosphorylated (79). Two transcripts encoding hLigIIIα and hLigIIIβ are generated by alternative splicing. DNA ligase IIIα polypeptides have a breast and ovarian cancer susceptibility protein 1 C-terminal (BRCT) domain (green), whereas DNA ligase IIIβ polypeptides have a C-terminal NLS (gray) (62). Two polypeptides are generated from each of the transcripts by alternative translation initiation. The longer polypeptides have an N-terminal mitochondrial leader sequence (MLS, turquoise) (60). Abbreviations: aa, amino acids; mito, mitochondria.

Figure 6

DNA ligase IV family. Alignment of the DNA ligases encoded by the S. cerevisiae DNL4 and human LIG4 genes. The DNA-binding domain (DBD, red) and the catalytic core composed of the adenylation and OB-fold domains (gray) are indicated. The position of the active-site lysine residues (K282, K213) within the adenylation domain that form the covalent bond with AMP are shown. The positions of the tandem BRCT domains (green) in the C-terminal regions of yeast Dnl4 and human DNA ligase IV (human LigIV) and a phosphorylated serine (Ser650) residue in human DNA ligase IV are shown (89). Abbreviation: aa, amino acids.

Figure 7

Model for the interaction of human DNA ligase I with DNA-linked homotrimeric proliferating cell nuclear antigen (PCNA). (a) DNA ligase I in an elongated conformation docks onto the PCNA ring via an interaction between the N-terminal PIP box of DNA ligase I (light blue) and the interdomain connector loop of a PCNA monomer. (b) The initial docking by the PIP box that is flexibly linked to the catalytic core of ligase I facilitates an interaction between the DNA-binding domain (DBD) (red) domain and the PCNA ring. At this stage, the catalytic region remains in an extended conformation. (c) Subsequent interactions with nicked DNA orchestrate the transition of the catalytic region of DNA ligase I from the extended to a closed ring conformation with each of the domains, DBD (red), adenylation (green), and OB-fold (yellow) contacting the DNA. This model is based on structural studies of human ligase I bound to DNA (26) and the Sulfolobus DNA ligase complexed with PCNA (112).

Figure 8

Protein partners of the DNA ligase IIIα/XRCC1 complex. Diagram showing the regions of the DNA ligase IIIα (pink) and XRCC1 (blue) that are involved in interactions with other DNA repair proteins (, , , –126). Abbreviations: APE1, apurinic/apyrimidinic site endonuclease 1; APLF, aprataxin- and PNK-like factor; BRCT I and II, breast and ovarian cancer susceptibility protein 1 C-terminal domains I and II; DBD, DNA-binding domain; DNA Polβ, DNA polymerase β; NEIL1/2, nei endonuclease VIII-like 1/2; NLS, nuclear localization signal; NTD, nucleotidyl transferase domain; OGG1, 8-oxoguanosine DNA glycosylase; PARP1, poly(ADP-ribose) polymerase 1; PCNA, proliferating cell nuclear antigen; PNK, polynucleotide kinase; TDP1, tyrosyl-DNA phosphodiesterase 1; XRCC1, X-ray cross-complementing 1; ZnF, zinc finger.

Figure 9

DNA end-bridging and end-processing complexes formed by human NHEJ proteins. (a) A model of the end-bridging complex formed by heterodimeric Ku, DNA PK_cs/Artemis, and DNA ligase IV/XRCC4. (b) Diagram showing the protein-protein interactions among the NHEJ factors involved in the end-processing and ligation steps of NHEJ (–142). Pol X refers to Pol λ, Pol μ, and terminal transferase (138, 139). Abbreviations: DNA PK_cs, catalytic subunit of the DNA-dependent protein kinase; PNK, polynucleotide kinase; X4, X-ray cross-complementing 4 or XRCC4; XLF, XRCC4-like factor. This protein is also known as Cernunnos.

Figure 10

End-bridging and end-processing complexes formed by yeast nonhomologous end-joining (NHEJ) proteins. (a) Diagram showing the protein-protein interaction among the core NHEJ factors, heterodimeric Ku, Dnl4/Lif1, and Mre11/Rad50/Xrs2. (b) Diagram showing the protein-protein interactions among the NHEJ factors involved in the end-processing and ligation steps of NHEJ. The role of Nej1, an ortholog of human XLF/Cernunnos, has yet to be determined (–150). Abbreviations: BRCT, breast and ovarian cancer susceptibility protein 1 C-terminal; DBD, DNA-binding domain; Dnl4, homolog of human DNA ligase IV; Lif1, ligase interacting factor 1, an ortholog of human XRCC4; Mre11, Rad50, Xrs2, subunits of the Mre11/Rad50/Xrs2 protein complex that processes double-strand DNA breaks to enable repair; yKu70 and yKu80, homologs of human Ku70 and Ku80.