The T4 gp43 DNA polymerase (gi:118854, NP_049662), an 898 amino acid residue protein related to the Pol B family, is used in both leading and lagging strand DNA synthesis. The Pol B family includes eukaryotic pol α, δ, and ε. The full length T4 enzyme and the exo^- mutant (D219A) have been cloned, expressed and purified 34 35 . While the structure of the T4 gp43 DNA polymerase has yet to be solved, the enzyme from the RB69 bacteriophage has been solved individually (PDB 1waj) and in complex with a primer template DNA duplex (PDB 1ig9, Figure 3A) 32 36 . The primary sequence alignment reveals that the T4 gp43 DNA polymerase is 62% identical and 74% similar to RB69 gp43 DNA polymerase, a 903 residue protein 37 38 .

E. coli Pol I, the first DNA polymerase discovered by Kornberg, has three domains, an N-terminal 5' to 3' exonuclease (cleaved to create the Klenow fragment), a 3' to 5' editing exonuclease domain, and a C-terminal polymerase domain 5 . The structure of the E. coli Pol I Klenow fragment was described through anthropomorphic terminology of fingers, palm, and thumb domains 39 40 . The RB69 gp43 DNA polymerase has two active sites, the 3' to 5' exonuclease (residues 103 - 339) and the polymerase domain (residues 381 - 903), comparable to Klenow fragment domains 41 . The gp43 DNA polymerase also has an N-terminal domain (residues 1 - 102 and 340 - 380) and a C-terminal tail containing a PCNA interacting peptide (PIP box) motif (residues 883 - 903) that interacts with the 45 sliding clamp protein. The polymerase domain contains a fingers subunit (residues 472 - 571) involved in template display (Ser 565, Lys 560, amd Leu 561) and NTP binding (Asn 564) and a palm domain (residues 381 - 471 and 572 - 699) which contains the active site, a cluster of aspartate residues (Asp 411, 621, 622, 684, and 686) that coordinates the two divalent active site metals (Figure 3B). The T4 gp43 DNA polymerase appears to be active in a monomeric form, however it has been suggested that polymerase dimerization is necessary to coordinate leading and lagging strand synthesis 6 20 .

The gene 45 protein (gi:5354263, NP_049666), a 228 residue protein, is the polymerase-associated processivity clamp, and is a functional analog to the β subunit of E. coli Pol III holoenzyme and the eukaryotic proliferating cell nuclear antigen (PCNA) 8 . All proteins in this family, both dimeric (E. coli β) and trimeric (gp45, PCNA), form a closed ring represented here by the structure of the T4 gp45 (PDB 1czd, Figure 4A) 29 . The diameter of the central opening of all known clamp rings is slightly larger than duplex B-form DNA. When these clamps encircle DNA, basic residues lining the rings (T4 gp45 residues Lys 5 and 12, Arg 124, 128, and 131) interact with backbone phosphates. The clamps have an α/β structure with α-helices creating the inner wall of the ring. The anti-parallel β-sandwich fold forms the outer scaffolding. While most organisms utilize a polymerase clamp, some exceptions are known. For example, bacteriophage T7 gene 5 polymerase sequesters E. coli thioredoxin for use as a processivity factor 42 .

The gp45 related PCNA clamp proteins participate in many protein/DNA interactions including DNA replication, repair, and repair signaling proteins. A multitude of different proteins have been identified that contain a PCNA interaction protein box (PIP box) motif Qxxhxxaa where x is any residue, h is L, I or M, and a is aromatic 43 . In T4, PIP box sequences have been identified in the C-terminal domain of gp43 DNA polymerase, mentioned above, and in the N-terminal domain of RNase H, discussed below. The C-terminal PIP box peptide from RB69 gp43 DNA polymerase has been co-crystallized with RB69 gp45 clamp protein (PDB 1b8h, Figures 3A and 3C) and allows modeling of the gp45 clamp and gp43 DNA polymerase complex (Figure 3A) 31 . The gp45 clamp trails behind the 43 DNA polymerase, coupled through the gp43 C-terminal PIP box bound to a pocket on the outer surface of the gp45 clamp protein. Within RB69 gp45 clamp protein, the binding pocket is primarily hydrophobic (residues Tyr 39, Ile 107, Phe 109, Trp 199, and Val 217) with two basic residues (Arg 32 and Lys 204) interacting with the acidic groups in the PIP box motif. The rate of DNA synthesis, in the presence and absence of gp45 clamp protein, is approximately 400 nucleotides per second, indicating that the accessory gp45 clamp protein does not affect the enzymatic activity of the gp43 DNA polymerase 6 . More discussion about the interactions between T4 gp43 polymerase and T4 gp45 clamp can be found in Geiduschek and Kassavetis, this series. While the gp45 clamp is considered to be a processivity factor, this function may be most prevalent when misincorporation occurs. When a mismatch is introduced, the template strand releases, activating the 3' to 5' exonuclease activity of the gp43 DNA polymerase. During the switch, gp45 clamp maintains the interaction between the replicase and DNA.

The mechanism for loading of the ring shaped PCNA clamps onto duplex DNA is a conundrum; imagine a magician's linking rings taken apart and reassembled without an obvious point for opening. The clamp loaders, the magicians opening the PCNA rings, belong to the AAA + ATPase family which include the E. coli gamma (γ) complex and eukaryotic replication factor C (RF-C) 44 45 . The clamp loaders bind to the sliding clamps, open the rings through ATP hydrolysis, and then close the sliding clamps around DNA, delivering these ring proteins to initiating replisomes or to sites of DNA repair. The gp44 clamp loader protein (gi:5354262, NP_049665) is a 319 residue, two-domain, homotetrameric protein. The N-domain of gp44 clamp loader protein has a Walker A p-loop motif (residues 45-52, GTRGVGKT) 38 . The gp62 clamp loader protein (gi:5354306, NP_049664) at 187 residues, is half the size of gp44 clamp loader protein and must be co-expressed with gp44 protein to form an active recombinant complex 46 .

The T4 gp44/62 clamp loader complex is analogous to the E. coli heteropentameric γ complex (γ₃δ'δ) and yeast RF-C despite an almost complete lack of sequence homology with these clamp loaders 46 . The yeast p36, p37, and p40 subunits of RF-C are equivalent to the E. coli γ, yeast p38 subunit is equivalent to δ', and yeast p140 subunit is equivalent to δ 47 . The T4 homotetrameric gp44 clamp loader protein is equivalent to the E. coli γ₃δ' and T4 gp62 clamp loader is equivalent to the E. coli δ. The first architectural view of clamp loaders came from the collaborative efforts of John Kuriyan and Mike O'Donnell who have completed crystal structures of several components of the E. coli Pol III holoenzyme including the ψ-χ complex (PDB 1em8), the β-δ complex (PDB 1jqj) and the full γ complex γ₃δ'δ (PDB 1jr3, Figure 4B) 48 49 50 . More recently, the yeast RF-C complex has been solved (PDB 1sxj) 47 . Mechanisms of all clamp loaders are likely very similar, therefore comparison of T4 gp44/62 clamp loader protein with the E. coli model system is most appropriate. The E. coli γ₃δ', referred to as the motor/stator (equivalent to T4 gp44 clamp loader protein), binds and hydrolyzes ATP, while the δ subunit, known as the wrench (equivalent to T4 gp62 clamp loader protein), binds to the β clamp (T4 gp45 clamp protein). The E. coli γ complex is comparable in size to the E. coli β clamp and the two proteins interact face to face, with one side of the β clamp dimer interface bound to the δ (wrench) subunit, and the other positioned against the δ' (stator). Upon hydrolysis of ATP, the γ (motor) domains rotate, the δ subunit pulls on one side of a β clamp interface as the δ' subunit pushes against the other side of the β clamp, resulting in ring opening. For the T4 system, interaction with DNA and the presence of the gp43 DNA polymerase releases the gp45 clamp from the gp44/62 clamp loader. In the absence of gp43 DNA polymerase, the gp44/62 clamp loader complex becomes a clamp unloader 6 . Current models of the E. coli Pol III holoenzyme have leading and lagging strand synthesis coordinated with a single clamp loader coupled to two DNA polymerases through the τ subunit and to single-stranded DNA binding protein through the χ subunit 51 . There are no T4 encoded proteins that are comparable to E. coli τ or χ.

Single-stranded DNA binding proteins have an oligonucleotide-oligosaccharide binding fold (OB fold), an open curved antiparallel β-sheet 52 53 . The aromatic residues within the OB fold stack with bases, thereby reducing the rate of spontaneous deamination of single-stranded DNA 54 . The OB fold is typically lined with basic residues for interaction with the phosphate backbone to increase stability of the interaction. Cooperative binding of ssb proteins assists in unwinding the DNA duplex at replication forks, recombination intermediates, and origins of replication. The T4 gp32 single-stranded DNA binding protein (gi:5354247, NP_049854) is a 301 residue protein consisting of three domain. The N-terminal basic B-domain (residues 1 - 21) is involved in cooperative interactions, likely through two conformations 55 . In the absence of DNA, the unstructured N-terminal domain interferes with the protein multimerization. In the presence of DNA, the lysine residues within the N-terminal peptide presumably interact with the phosphate backbone of DNA. Organization of the gp32 N-terminus by DNA creates the cooperative binding site for assembly of gp32 ssb filaments 56 .

The crystal structure of the core domain of T4 gp32 ssb protein (residues 22 - 239) containing the single OB fold has been solved (Figure 5A) 33 . Two extended and two short antiparallel β-strands form the open cavity of the OB fold for nucleotide interaction. Two helical regions stabilize the β-strands, the smaller of which, located at the N-terminus of the core, has a structural zinc finger motif (residues His 64, and Cys 77, 87, and 90). The C-terminal acidic domain A-domain (residues 240 - 301) is involved in protein assembly, interacting with other T4 proteins, including gp61 primase, gp59 helicase assembly protein, and RNase H 57 . We have successfully crystallized the gp32(-B) construct (residues 21 - 301), but have found the A-domain disordered in the crystals with only the gp32 ssb core visible in the electron density maps (Hinerman, unpublished data). The analogous protein in eukaryotes is the heterotrimeric replication protein A (RPA) 58 . Several structures of Archaeal and Eukaryotic RPAs have been reported including the crystal structure of a core fragment of human RPA70 59 60 . The RPA70 protein is the largest of the three proteins in the RPA complex and has two OB fold motifs with 9 bases of single-stranded DNA bound (PDB 1jmc). The E. coli ssb contains four OB fold motifs and functions as a homotetramer. A structure of the full length version of E. coli ssb (PDB 1sru) presents evidence that the C terminus (equivalent to the T4 32 A domain) is also disordered 61 .

The replicative helicase family of enzymes, which includes bacteriophages T4 gp41 helicase and T7 gene 4 helicase, E. coli DnaB, and the eukaryotic MCM proteins, are responsible for the unwinding of duplex DNA in front of the leading strand replisome 62 . The T4 gp41 protein (gi:9632635, NP_049654) is the 475 residue helicase subunit of the primase(gp61)-helicase(gp41) complex and a member of the p-loop NTPase family of proteins 63 . Similar to other replicative helicases, the gp41 helicase assembles by surrounding the lagging strand and excluding the leading strand of DNA. ATP hydrolysis translocates the enzyme 5' to 3' along the lagging DNA strand, thereby unwinding the DNA duplex approximately one base pair per hydrolyzed ATP molecule. Efforts to crystallize full length or truncated gp41 helicase individually, in complex with nucleotide analogs, or in complex with other T4 replication proteins have not been successful in part due to the limited solubility of this protein. In addition, the protein is a heterogeneous mixture of dimers, trimers and hexamers, according to dynamic light scattering measurements. The solubility of T4 41 helicase can be improved to greater than 40 mg/ml of homogenous hexamers by eliminating salt and using buffer alone (10 mM TAPS pH 8.5) 64 . However, the low ionic strength crystal screen does not produce crystals 65 . To understand the T4 gp41 helicase, we must therefore look to related model systems.

Like T4 41 helicase, efforts to crystallize E. coli DnaB have met with minimal success. Thus far only a fragment of the non-hexameric N-terminal domain (PDB 1b79) has been crystallized successfully for structural determinations 66 . More recently, thermal stable eubacteria (Bacillus and Geobacillus stearothermophilis) have been utilized by the Steitz lab to yield more complete structures of the helicase-primase complex (PDB 2r6c and 2r6a, respectively) 67 . A large central opening in the hexamer appears to be the appropriate size for enveloping single-stranded DNA, as it is too small for duplex DNA. Collaborative efforts between the Wigley and Ellenberger groups revealed the hexameric structure of T7 gene 4 helicase domain alone (residues 261 - 549, PDB 1eOk) and in complex with a non-hydrolyzable ATP analog (PDB 1e0h) 68 . Interestingly, the central opening in the T7 gene 4 helicase hexamer is smaller than other comparable helicase, suggesting that a fairly large rearrangement is necessary to accomplish DNA binding. A more complete structure from the Ellenberger lab of T7 gene 4 helicase that includes a large segment of the N-terminal primase domain (residues 64 - 566) reveals a heptameric complex with a larger central opening (Figure 5B) 69 . Both the eubacterial and bacteriophage helicase have similar α/β folds. The C-terminal Rec A like domain follows 6-fold symmetry and has nucleotide binding sites at each interface. In the eubacterial structures, the helical N-domains alternate orientation and follow a three-fold symmetry with domain swapping. The T4 gp41 helicase is a hexameric two-domain protein with Walker A p-loop motif (residues 197 - 204, GVNVGKS) located at the beginning of the conserved NTPase domain (residues 170 - 380), likely near the protein:protein interfaces, similar to the T7 helicase structure.

The progression of the DNA replisome is restricted in the absence of either gp32 ssb protein or the gp41 helicase 6 . In the presence of gp32 ssb protein, loading of the gp41 helicase is inhibited. In the absence of gp32 ssb protein, the addition of gp41 helicase improves the rate of DNA synthesis but displays a significant lag prior to reaching maximal DNA synthesis 13 . The gp59 helicase loading protein (gi:5354296, NP_049856) is a 217 residue protein that alleviates the lag phase of gp41 helicase 13 14 . In the presence gp32 ssb protein, the loading of gp41 helicase requires gp59 helicase loading protein. This activity is similar to the E. coli DnaC loading of DnaB helicase 70 71 . Initially, 59 helicase loading protein was thought to be a single-stranded DNA binding protein that competes with 32 ssb protein on the lagging strand 13 72 . In that model, the presence of gp59 protein within the gp32 filament presumably created a docking site for gp41 helicase. However, the gp59 helicase loading protein is currently known to have more specific binding affinity for branched and Holliday junctions 16 17 . This activity is comparable to the E. coli replication rescue protein, PriA, which was first described as the PAS recognition protein (n' protein) in φX174 phage replication 73 . Using short pseudo-Y junction DNA substrates, gp59 helicase loading protein has been shown to recruit gp32 ssb protein to the 5' (lagging strand) arm, a scenario relevant to replication fork assembly 74 .

The high-resolution crystal structure of 59 helicase loading protein reveals a two-domain, α-helical structure that has no obvious cleft for DNA binding 17 . The E. coli helicase loader, DnaC, is also a two domain protein. However, the C-terminal domain of DnaC is an AAA + ATPase related to DnaA, as revealed by the structure of a truncated DnaC from Aquifex aeolicus (pdb 3ec2) 75 . The DnaC N-domain interacts with the hexameric DnaB in a one-to-one ratio forming a second hexameric ring. Sequence alignments of gp59 helicase loading protein reveal an "ORFaned" (orphaned open reading frame) protein; a protein that is unique to the T-even and other related bacteriophages 4 17 . Interestingly, searches for structural alignments of the gp59 protein, using both Dali 76 and combinatorial extension 77 , have revealed partial homology with the eukaryotic high mobility group 1A (HMG1A) protein, a nuclear protein involved in chromatin remodeling 78 . Using the HMG1A:DNA structure as a guide, we have successfully modeled gp59 helicase assembly protein bound to a branched DNA substrate which suggests a possible mode of cooperative interaction with 32 ssb protein (Figure 5C) 17 . Attempts to co-crystallize gp59 protein with DNA, or with gp41 helicase, or with gp32 ssb constructs have all been unsuccessful. The 59 helicase assembly protein combined with 32(-B) ssb protein yields a homogenous solution of heterodimers, amenable for small angle X-ray scattering analysis (Hinerman, unpublished data).

The gp61 DNA dependent RNA polymerase (gi:5354295, NP_049648) is a 348 residue enzyme that is responsible for the synthesis of short RNA primers used to initiate lagging strand DNA synthesis. In the absence of gp41 helicase and gp32 ssb proteins, the gp61 primase synthesizes ppp(Pu)pC dimers that are not recognized by DNA polymerase 79 80 . A monomer of gp61 primase and a hexamer of gp41 helicase are essential components of the initiating primosome 63 81 . Each subunit of the hexameric gp41 helicase has the ability to bind a gp61 primase. Higher occupancies of association have been reported but physiological relevance is unclear 82 83 . When associated with gp41 helicase, the gp61 primase synthesizes pentaprimers that begin with 5'-pppApC onto template 3'-TG; a very short primer that does not remain annealed in the absence of protein 79 . An interaction between gp32 ssb protein and gp61 primase likely coordinates the handoff of the RNA primer to the gp43 DNA polymerase, establishing a synergy between leading strand progression and lagging strand synthesis 84 . The gp32 ssb protein will bind to single-stranded DNA unwound by gp41 helicase. This activity inhibits the majority of 3'-TG template sites for gp61 primase and therefore increases the size of Okazaki fragments 6 . Activity of gp61 primase is obligate to the activity of the gp41 helicase. The polymerase accessory proteins, gp45 clamp and gp44/62 clamp loader, are essential for primer synthesis when DNA is covered by gp32 ssb protein 85 . Truncation of 20 amino acids from the C-terminus of gp41 helicase protein retains interaction with gp61 primase but eliminates gp45 clamp and gp44/62 clamp loader stimulation of primase activity 86 .

The gp61 primase contains an N-terminal zinc finger DNA binding domain (residues cys 37, 40, 65, and 68) and a central toprim catalytic core domain (residues 179 - 208) 87 88 . Crystallization trials of full length gp61 primase and complexes with gp41 helicase have been unsuccessful. Publication of a preliminary crystallization report of gp61 primase C-terminal domain (residues 192 - 342) was limited in resolution, and a crystal structure has not yet been published 89 . A structure of the toprim core fragment of E. coli DnaG primase (residues 110 to 433 of 582) has been solved, concurrently by the Berger and Kuriyan labs (PDB 1dd9, 90 ) (PDB 1eqn, 91 ). To accomplish this, the N-terminal Zn finger and the C-terminal DnaB interacting domain were removed. More recently, this same DnaG fragment has been resolved in complex with single-stranded DNA revealing a binding track adjacent to the toprim domain (PDB 3b39, 92 ). Other known primase structures include the Stearothermophilis enzymes solved in complex with helicase (discussed above) and the primase domain of T7 gene 4 primase (PDB 1nui) (Figure 5D) 69 . The primase domain of T7 gene 4 is comprised of the N-terminal Zn finger (residues 1 - 62) and toprim domain (residues 63 - 255). This structure is actually a primase-helicase fusion protein.

RNase H activity of the bacteriophage T4 rnh gene product (gi:5354347, NP_049859) was first reported by Hollingsworth and Nossal 24 . The structure of the 305 residue enzyme with two metals bound in the active site was completed in collaboration with the Nossal laboratory (PDB 1tfr) (Figure 6A) 28 . Mutations of highly conserved residues which abrogate activity are associated with the two hydrated magnesium ions 93 . The site I metal is coordinated by four highly conserved aspartate residues (D19, D71, D132, and D155) and mutation of any one to asparagines eliminates nuclease activity. The site II metal is fully hydrated and hydrogen bonded to three aspartates (D132, D157, and D200) and to the imino nitrogen of an arginine, R79. T4 RNase H has 5' to 3' exonuclease activity on RNA/DNA, DNA/DNA 3'overhang, and nicked substrate, with 5' to 3' endonuclease activity on 5' fork and flap DNA substrates. The crystal structure of T4 RNase H in complex with a pseudo Y junction DNA substrate has been solved (PDB 2ihn, Figure 6B) 94 . To obtain this structure, it was necessary to use an active site mutant (D132N); Asp132 is the only residue in RNase H that is inner sphere coordinated to the active site metals 28 .

The processivity of RNase H exonuclease activity is enhanced by the gp32 ssb protein. Protein interactions can be abrogated by mutations in the C-terminal domain of RNase H 22 and within the core domain of gp32 ssb protein (Mueser, unpublished data). Full length gp32 ssb protein and RNase H do not interact in the absence of DNA substrate. Removal of the N-terminal peptide of gp32 ssb protein (gp32(-B)), responsible for gp32 ssb cooperativity, yields a protein that has high affinity for RNase H. It is likely that the reorganization of gp32 B-domain when bound to DNA reveals a binding site for RNase H and therefore helps to coordinate 5'-3' primer removal after extension by the DNA polymerase. This is compatible with the model proposed for the cooperative self assembly of gp32 protein. The structure of RNase H in complex with gp32(-B) has been solved using X-ray crystallography and small angle X-ray scattering (Mueser, unpublished data) (Figure 6C). The gp45 clamp protein enhances the processivity of RNase H on nicked and flap DNA substrates 23 . Removal of the N-terminal peptide of RNase H eliminates the interaction between RNase H and gp45 clamp protein and decreases processivity of RNase H. The structure of the N-terminal peptide of RNase H in complex with gp45 clamp protein reveals that binding occurs within the gp45 clamp PIP-box motif of RNase H (Devos, unpublished data).

Sequence alignment of T4 RNase H reveals membership to a highly conserved family of nucleases that includes yeast rad27, rad2, human FEN-1, and xeroderma pigmentosa group G (XPG) proteins. The domain structure of both FEN-1 and XPG proteins is designated N, I, and C 95 . The yeast rad2 and human XPG proteins are much larger than the yeast rad27 and human FEN-1 proteins. This is due to a large insertion in the middle of rad2 and XPG proteins between the N and I domains. The N and I domains are not separable in the T4 RNase H protein as the N-domain forms part of the α/β structure responsible for fork binding and half of the active site. The I domain is connected to the N-domain by a bridge region above the active site which is unstructured in the presence of active site metals and DNA substrate. It is this region that corresponds with the position of the large insertions of rad2 and XPG. Curiously, this bridge region of T4 RNase H becomes a highly ordered a-helical structure in the absence of metals. Arg and Lys residues are interdigitated between the active site Asp groups within the highly ordered structure (Mueser, unpublished data). The I domain encompasses the remainder of the larger α/β subdomain and the α-helical H3TH motif responsible for duplex binding. The C-domain is truncated at the helical cap that interacts with gp32 ssb and the PIP motif is located in the N-terminus of T4 RNase H. In the FEN-1 family of proteins, the C-domain, located opposite the H3TH domain, contains a helical cap and an unstructured C-terminal PIP-box motif for interaction with a PCNA clamp.

The T4 gp30 protein (gi: 5354233, NP_049813) is best known as T4 DNA ligase, a 487 residue ATP-dependent ligase. DNA ligases repair nicks in double-stranded DNA containing 3' OH and 5' phosphate ends. Ligases are activated by the covalent modification of a conserved lysine with AMP donated by NADH or ATP. The conserved lysine and the nucleotide binding site reside in the adenylation domain (NTPase domain) of ligases. Sequence alignment of the DNA ligase family Motif 1 (KXDGXR) within the adenylation domain identifies Lys 159 in T4 DNA ligase (159 KADGAR 164) as the moiety for covalent modification 96 . The bacterial ligases are NADH-dependent, while all eukaryotic enzymes are ATP-dependent 97 . Curiously, T4 phage, whose existence is confined within a prokaryote, encodes an ATP-dependent ligase. During repair, the AMP group from the activated ligase is transferred to the 5' phosphate of the DNA nick. This activates the position for condensation with the 3' OH, releasing AMP in the reaction. The T4 ligase has been cloned, expressed, and purified but attempts to crystallize T4 ligase, with and without cofactor, have not been successful. The structure of the bacteriophage T7 ATP-dependent ligase has been solved (PDB 1a0i, Figure 6C) 98 99 , which has a similar fold to T4 DNA ligase 100 . The minimal two-domain structure of the 359 residue T7 ligase has a large central cleft, with the larger N-terminal adenylation domain containing the cofactor binding site and a C-terminal OB domain. In contrast, the larger 671 residue E. coli DNA ligase has five domains; the N-terminal adenylation and OB fold domains, similar to T7 and T4 ligase, including a Zn finger, HtH and BRCT domains present in the C-terminal half of the protein 97 . Sequence alignment of DNA ligases indicate that the highly conserved ligase signature motifs reside in the central DNA binding cleft, the active site lysine, and the nucleotide binding site 98 . Recently, the structure of NAD-dependent E. coli DNA ligase has been solved in complex with nicked DNA containing an adenylated 5' PO₄ (pdb 2owo) 101 . This flexible, multidomain ligase encompasses the duplex DNA with the adenylation domain binding to the nick; a binding mode also found in the human DNA ligase 1 bound to nicked DNA (pdb 1x9n) 102 . T4 DNA ligase is used routinely in molecular cloning for repairing both sticky and blunt ends. The smaller two-domain structure of T4 DNA ligase has lower affinity for DNA than the multidomain ligases. The lack of additional domains to encompass the duplex DNA likely explains the sensitivity of T4 ligase activity to salt concentration.