• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of embojLink to Publisher's site
EMBO J. Mar 9, 2005; 24(5): 885–894.
Published online Feb 3, 2005. doi:  10.1038/sj.emboj.7600568
PMCID: PMC554121

Structural insights into the first incision reaction during nucleotide excision repair


Nucleotide excision repair is a highly conserved DNA repair mechanism present in all kingdoms of life. The incision reaction is a critical step for damage removal and is accomplished by the UvrC protein in eubacteria. No structural information is so far available for the 3′ incision reaction. Here we report the crystal structure of the N-terminal catalytic domain of UvrC at 1.5 Å resolution, which catalyzes the 3′ incision reaction and shares homology with the catalytic domain of the GIY-YIG family of intron-encoded homing endonucleases. The structure reveals a patch of highly conserved residues surrounding a catalytic magnesium-water cluster, suggesting that the metal binding site is an essential feature of UvrC and all GIY-YIG endonuclease domains. Structural and biochemical data strongly suggest that the N-terminal endonuclease domain of UvrC utilizes a novel one-metal mechanism to cleave the phosphodiester bond.

Keywords: crystallography, DNA damage, DNA repair, nucleotide excision repair, UvrC


Nucleotide excision repair (NER) stands apart from other DNA repair mechanisms available to the cell in its ability to recognize a broad range of structurally unrelated DNA damages (Van Houten, 1990; Friedberg et al, 1995; Lloyd and Van Houten, 1995; Sancar, 1996; Goosen and Moolenaar, 2001) including carcinogenic cyclobutane pyrimidine dimers induced by UV radiation, benzo[a]pyrene-guanine adducts caused by smoking and burning of fossil fuels, and guanine-cisplatinum adducts formed during cancer chemotherapy (Sancar, 1994). The strategy employed by NER is the same in all three kingdoms of life. NER in prokaryotes was one of the first repair mechanisms discovered (Boyce and Howard-Flanders, 1964; Setlow and Carrier, 1964) and is mediated by the UvrA, UvrB and UvrC proteins. These three proteins recognize and cleave damaged DNA in an ATP-dependent multistep reaction. UvrA is involved in damage recognition and either forms a heterotrimeric (UvrA2UvrB) (reviewed in Theis et al, 2000) or heterotetrameric (UvrA2UvrB2) (Verhoeven et al, 2002) complex with UvrB. This complex is believed to scan the DNA helix for conformational perturbations induced by DNA lesions (Theis et al, 2000). After the damage has been identified, UvrA dissociates from the protein–DNA complex, leaving UvrB bound to the DNA (Orren and Sancar, 1990) forming a stable preincision complex (Theis et al, 2000; Skorvaga et al, 2002). UvrC binds to this complex and mediates the incision three or four nucleotides 3′ to the damaged site, followed by a second incision seven nucleotides 5′ to the damaged site (Sancar and Rupp, 1983; Lin and Sancar, 1992a, 1992b; Verhoeven et al, 2000). UvrD (helicase II) and DNA polymerase I (polI) are required for turnover of the UvrABC proteins (Caron et al, 1985; Husain et al, 1985). UvrD removes both UvrC and the oligonucleotide containing the lesion, while UvrB remains bound to the gapped DNA until it is displaced by DNA polI (Orren et al, 1992). The reaction is completed by DNA ligase, which closes the nicked DNA. This multistep process of DNA recognition and repair ensures a high degree of discrimination between the damaged and nondamaged strand.

Site-directed mutagenesis and sequence alignments have shown that UvrC catalyzes both the 3′ and 5′ incisions and each of these incisions is performed by a distinct catalytic site that can be inactivated independently (Lin and Sancar, 1992b; Verhoeven et al, 2000). The domain responsible for 3′ incision is located in the N-terminal half of the molecule and consists of approximately the first hundred residues. This domain shares limited homology with a small module found in members of the GIY-YIG endonuclease family (Aravind et al, 1999). Also included in the N-terminal half is a region that interacts with the C-terminal domain of UvrB (Aravind et al, 1999). The 5′ catalytic domain, which is distantly related to Escherichia coli endonuclease V, is located in the C-terminal half of the protein along with two helix-hairpin-helix motifs employed in DNA binding (Aravind et al, 1999). After recruitment to the UvrB:DNA preincision complex, UvrC first catalyzes cleavage of the DNA on the 3′ side of the lesion (Verhoeven et al, 2000). This incision requires the interaction between the C-terminal domain of UvrB and the homologous UvrB binding domain of UvrC (Moolenaar et al, 1995, 1998a), which is not required for 5′ incision (Moolenaar et al, 1995).

In order to obtain a better understanding of the 3′ incision event, we have solved the crystal structure of the N-terminal endonuclease domain of UvrC from two different thermophilic organisms, Bacillus caldotenax and Thermotoga maritima, at 2.0 and 1.5 Å resolution, respectively. This domain shares structural and sequence similarity to the catalytic domain found in I-TevI, a GIY-YIG homing endonuclease (Van Roey et al, 2002). I-TevI is one of at least 60 known GIY-YIG endonuclease family members that are present in bacteriophage T4, bacteria, archaea, algal chloroplasts and mitochondria, and fungal mitochondria. Members of this family are characterized by a 70–100 residues long module containing a conserved GIY-(X9–11)-YIG motif.

We have identified a patch of highly conserved residues on the surface of the N-terminal domain of UvrC to which a single divalent cation is bound. The residues that form the metal binding pocket are conserved throughout all GIY-YIG endonucleases, suggesting the site to be a common feature of all family members. We mutated seven amino acids within the conserved patch and analyzed whether full-length UvrC was still able to incise damaged DNA, in the complete UvrABC reaction. Combined with the structural data, the results indicate that the conserved patch is the active site and the bound divalent cation is the catalytic metal. Based on our data, we propose that UvrC uses a novel one-metal mechanism to catalyze cleavage of the fourth or fifth phosphodiester bond 3′ to the DNA lesion.


Crystal structure of the N-terminal endonuclease domain of UvrC

The N-terminal catalytic domain of UvrC was initially cloned from B. caldotenax (UvrCN-Bca; residues 1–98) and its structure was solved by multiwavelength anomalous diffraction (MAD) (Table I). The protein crystallized in space group C2 and contained four molecules in the asymmetric unit arranged as a tetramer with C4 symmetry. Each of the four subunits has nearly identical conformations, with an average root mean square (r.m.s.) deviation of 0.62 Å for the Cα atoms of residues 8–94. The structure was refined at 2.0 Å resolution to an R-factor of 0.203 and Rfree of 0.252 (Table I). Residues 95–98 of subunit A, 1–8 and 95–98 of subunit B, 95–98 of subunit C, as well as 1 and 96–98 of subunit D are disordered.

Table 1
Crystallographic statistics

The corresponding domain was also cloned from T. maritima (UvrCN-Tma; residues 1–97) and crystallized in space group P43212 containing one molecule in the asymmetric unit. Crystals were soaked with manganese or magnesium chloride. The structure of UvrCN-Tma bound to manganese was solved by molecular replacement using the UvrCN-Bca monomer as a search model (see Materials and methods). This structure was refined at 1.5 Å resolution to an R-factor of 0.167 and Rfree of 0.185 (Table I), and consists of residues 1–89. The structure of magnesium-bound UvrCN-Tma was solved using difference Fourier methods (Table I). For the remainder of the discussion, residue numbering will correspond to UvrC from T. maritima, unless specified otherwise.

The B. caldotenax and T. maritima structures have similar conformations with an r.m.s. deviation of 1.55 Å for 88 Cα atoms. Both have an αββααβα(α) topology with the C-terminal α-helix (α5) not present in UvrCN-Tma. The core of the molecule is a three-stranded β-sheet, which is flanked by the first three helices on one side and helix 4 on the other (Figure 1). Helices α1, α2 and α4 run approximately parallel to the β-sheet, while helix α3 is positioned almost perpendicular to these secondary structural elements, making contact only with the bottom edge of the sheet. In UvrCN-Bca helix α3 also interacts with the C-terminus, connecting the two helical halves of the protein across the β-sheet.

Figure 1
Stereo view of the N-terminal endonuclease domain of UvrCN-Tma. The central β-sheet (β1–β3) is shown in yellow and the surrounding helices (α1–α4) in green. Conserved residues are shown in ball-and-stick ...

Comparison to the catalytic domain of I-TevI

A search using DALI (Holm and Sander, 1995), a network service for comparing three-dimensional protein structures (http://www.ebi.ac.uk/dali), identified the catalytic domain of I-TevI (Van Roey et al, 2002) (PDB codes 1LN0 and 1MK0) as the only structure with a fold similar to the N-terminal domain of UvrC (Z-score of 7.8). I-TevI is a member of the GIY-YIG family of homing endonucleases, which in turn belong to the larger GIY-YIG superfamily that includes UvrC. The only entity in common among all superfamily members is the domain presented here. It is a small, 70–100 residues module, containing a conserved GIY-(X9–11)-YIG motif (Van Roey et al, 2002) (Gly 17, Val 18, Tyr 19–Tyr 29, Ile 30, Gly 31), four invariant residues (Gly 31, Arg 39, Glu 76, Asn 88) and two highly conserved residues (Tyr 19, Tyr 29) (Figure 1). These amino acids are all located in proximity to each other and form, in part, a highly conserved surface.

The catalytic domains of UvrC and I-TevI have likely diverged long ago as reflected by their low sequence identity of only 15% (Figure 2B). UvrCN-Tma superimposes onto the catalytic domain of I-TevI with an r.m.s. deviation of 2.2 Å for 60 out of 89 possible Cα atoms. However, there are notable differences in secondary and tertiary structure (Figure 2A). UvrCN-Tma/Bca contains an additional helix, α1, as compared to I-TevI. The function of α1 in UvrC is undoubtedly structural, as the residues that form this helix are not conserved among UvrCs, and sequence alignments show that like I-TevI, certain UvrC proteins lack this N-terminal helix (Figure 3B). Secondly, the fragment in UvrCN-Tma/Bca spanning α2 and β3, which includes α3, is not structurally conserved compared to I-TevI. Regardless of this dissimilarity, the position of Ile 54 from UvrCN-Tma (Leu 56 in UvrCBca) is conserved and superimposes well with Leu 45 in I-TevI (Figure 2A and B). This residue is structurally important and stabilizes the hydrophobic core of the domain. Lastly, UvrCN-Bca contains an additional helix, α5, at its C-terminus that is replaced by a loop region in both I-TevI and UvrCN-Tma (Figure 2A and B).

Figure 2
Structural comparison of the two 3′ endonuclease domains from UvrC and the analogous domain in I-TevI. (A) Following their superposition, the three proteins were separated and displayed side by side: UvrCN-Tma (left), UvrCN-Bca (middle) and I- ...
Figure 3
Sequence conservation of the 3′ endonuclease domain of UvrC. (A) Two different views of the N-terminal domain in surface representation. The right view is rotated 90° relative to the left view around the vertical axis. Only the most conserved ...

Both UvrCN-Tma/Bca and I-TevI contain a strictly conserved glutamate, arginine and asparagine, and a pair of highly conserved tyrosine residues (Tyr 19 and Tyr 29 in UvrCTma; Tyr 6 and Tyr 17 in I-TevI) (Figure 2A). The third tyrosine, Tyr 43 in UvrCN-Tma, corresponds to His 31 in I-TevI (Figure 2A and B). Sequence alignments reveal that this residue is either a tyrosine or a histidine in all UvrCs (Figure 3B) and all other GIY-YIG family members. Both Tyr 43 of UvrCN-Tma and His 31 of I-TevI form a hydrogen bond to Tyr 19 and Tyr 6, respectively (Figure 2A). His 31 of I-TevI in turn forms a hydrogen bond to His 40, which occupies the position of Val 51 in UvrC. In place of a hydrogen bond, Val 51 forms van der Waals interactions with Tyr 43.

The conserved surface

Mapping the sequence conservation of UvrC proteins from different organisms onto the surface reveals a conserved patch of amino acids (Figure 3A and B). Six strictly conserved residues are located in the center on one side of the surface: Tyr 19, Tyr 29, Lys 32, Arg 39, Glu 76 and Asn 88. These strictly conserved amino acids are surrounded by a number of highly conserved residues: Tyr 43, Phe 73 and Ile 80. The conserved residues form a shallow, concave surface with dimensions of 16 Å × 15 Å and 5 Å deep, which could easily accommodate double-stranded DNA. This is in agreement with biochemical data, which have shown that the 3′ endonuclease site only recognizes double-stranded DNA as a substrate and DNA containing an unpaired region of more than eight nucleotides overlapping the 3′ incision site is not incised (Zou and Van Houten, 1999). The cleft is sufficient for nuclease activity; however, the isolated domain does not bind to DNA (data not shown) and requires either the UvrB interacting domain and/or the C-terminal helix-hairpin-helix DNA binding domain of UvrC for catalysis to occur.

The divalent cation

Phosphodiester bonds, although thermodynamically labile, require large activation energies for cleavage at physiological pH (Galburt and Stoddard, 2002). This is mainly due to the negative charge of the phosphoryl group at this pH, which repels potential attacking nucleophiles. Three chemical entities are required to catalyze efficiently the cleavage of the phosphodiester bond: a general base to position and activate the nucleophile (usually a water molecule) for inline attack of the 5′ phosphate, a general acid to protonate the 3′ leaving group and a Lewis acid to stabilize the pentacovalent phosphoanion transition state. A number of different strategies to satisfy these requirements have evolved. A common feature of most nucleases is the use of one, two or even three divalent cations in the active site to lower the free energy of the transition state (Galburt and Stoddard, 2002). Metal ions can decrease the pKa of coordinating water molecules, resulting in a bound hydroxide, which could take the role of either a nucleophile or a general base. Alternatively, a metal-coordinated water molecule can be acidic and has the potential to serve as the general acid necessary to protonate the 3′ OH leaving group. However, the most substantial reduction in free energy derives from the metal's ability to stabilize the negative charge of the phosphoanion transition state (Galburt and Stoddard, 2002).

No divalent cation was observed in the B. caldotenax structure after soaking and cocrystallization attempts. This is presumably due to the crystallization conditions, which contained high salt concentrations at low pH (5.4). In addition, there were protein–protein interactions in proximity to the only strictly conserved negatively charged amino acid, Glu 76, where metal binding was predicted to occur. Fortunately, crystals of UvrCN-Tma grew in PEG 8000 at high pH (8.5), a condition more suitable for soaking experiments. Additionally, Glu 76 was completely solvent accessible with no nearby protein–protein interactions. These crystals were briefly soaked with either MnCl2 or MgCl2 and the resulting structures provided similar results: a single divalent cation coordinated by Glu 76 and five well-ordered water molecules in an octahedral arrangement (Figure 4). These structures provide the first view of the metal and its exact coordination geometry for any member of the GIY-YIG superfamily and thus structural insight into the catalytic mechanism. The bound manganese was clearly identified in anomalous density maps, while omit maps revealed the octahedral arrangement of the chelating water molecules (Figure 4). The refined density for the water molecules in both structures is unambiguous with B-factors ranging from 17.1 to 23.6 Å2 for the manganese structure and 21.1 to 28.1 Å2 for the magnesium structure. The B-factors for the manganese and magnesium ion are 18.8 and 25.6 Å2, respectively. The manganese structure superimposes onto the magnesium structure with an r.m.s. deviation of 0.1 Å, and the two metal-water clusters superimpose with an r.m.s. deviation of 0.1 Å as well.

Figure 4
Stereo view of the active site of the 3′ endonuclease domain. The metal ion is shown as a magenta sphere and the five surrounding water molecules as red spheres. Hydrogen bonds are shown as dotted lines. A simulated annealing omit map omitting ...

Although Glu 76 is the only protein residue bound directly to the metal, the water molecules coordinating to the metal form additional contacts to the protein (Figure 4). One of the waters forms hydrogen bonds to both the hydroxyl group of Tyr 29 and the main-chain carbonyl of Ile 30. A second water forms a hydrogen bond to the carboxylate of Glu 76, while a third water forms a hydrogen bond to the backbone amide of Lys 32. Residues forming the binding pocket for the metal-water cluster are highly conserved, suggesting the metal binding site to be a conserved feature of all GIY-YIG family members.

Mutational analysis of the conserved surface

Based on our structural data, we generated mutations of the highly conserved surface residues Y29A, Y29F, K32A, R39A, F73A, F73E, E76A, I80A, I80E and N88A in full-length UvrCTma and tested these mutant proteins in an incision assay using a defined substrate and the two other NER proteins UvrABca and UvrBBca (Figure 5). In addition, we verified that all of the mutants were able to bind and crosslink to double-stranded DNA. Electrophoretic mobility shift assays were performed using wild-type and mutant UvrC proteins (in the absence of UvrA or UvrB) with a 40 bp duplex containing a centrally located (+)-trans-BPDE adduct. Crosslinking was performed using wild-type and mutant UvrC proteins (in the absence of UvrA or UvrB) with 50 bp duplex containing a site-specific arylazido-modified photoaffinity reagent (as described in DellaVecchia et al, 2004) (data not shown).

Figure 5
Incision activity of T. maritima UvrC mutants. The 5′ (A) or 3′ (B) end-labeled 50-mer double-stranded DNA substrates containing a centrally located fluorescein (FldT) were incubated with 20 nM UvrABca, 100 nM UvrBBca and 12.5 nM of the ...

Mutation of the sole metal ligand, Glu 76, to alanine renders UvrC unable to mediate either the 3′ or 5′ incision (Figure 5). UvrC from Tma, like E. coli UvrC, cannot achieve 5′ incision without prior 3′ incision, and thus inactivation of the 3′ nuclease active site inhibits the 5′ nuclease activity. Other inactive mutants were N88A, Y29A, Y29F, R39A and I80E. Mutant I80A showed an approximately 50% reduction in activity. Mutation of Lys 32 to alanine resulted in a protein that was 25–30% less active than wild-type UvrC and mutation of Phe 73 to either alanine or glutamate resulted in an enzyme with wild-type UvrC activity (Figure 5). The Y19F and Y43F mutants did not overexpress and could not be studied. In addition to the full-length mutants, we generated mutations in the isolated N-terminal domain (Y19F, Y29F, Y43F and N88A) and determined the structures of these mutants to inspect whether structural changes within the mutants lead to the inactivation of the enzyme (see Supplementary Table I).


In this study, we have crystallized and solved the structures of the N-terminal endonuclease domain of UvrC from two different thermophilic bacteria, T. maritima and B. caldotenax. Multiple sequence alignments have revealed that this domain is the only domain shared among all GIY-YIG family members (Kowalski et al, 1999), suggesting it to contain all the information necessary for catalysis. We have identified a patch of highly conserved residues on the surface of the N-terminal domain of UvrC containing a single divalent cation, which is coordinated by Glu 76 and five water molecules in an octahedral arrangement. We have mutated residues within the conserved region in full-length UvrC and analyzed these mutants for incision activity in the complete UvrABC reaction. These results identify residues important in catalysis and clearly indicate that the conserved region is the active site and the bound divalent cation is the catalytic metal.

The active site

A single divalent cation is bound to the conserved surface of the N-terminal domain of UvrC and coordinated exclusively by Glu 76, which is invariant in all known GIY-YIG family members (Figure 4 and and2B).2B). A similar interaction between the equivalent glutamate in I-TevI (Glu 75) and a manganese ion has been reported (Van Roey et al, 2002). Mutation of this glutamate to alanine renders UvrC inactive (Figure 5). The analogous mutation in I-TevI gives similar results (Derbyshire et al, 1997; Kowalski et al, 1999). Likewise, mutation of Ile 80, which forms part of the metal binding pocket, significantly reduces UvrC's activity in the case of I80A and essentially inactivates the protein when mutated to glutamate (Figure 5). In addition, the invariant residue Gly 31 is positioned just behind the bound metal and a side chain at this position would lead to steric interference (Figure 1). Mutation of the analogous glycine (Gly 19) to alanine in I-TevI yields a protein with no detectable activity (Kowalski et al, 1999). Furthermore, all residues that hydrogen bond to the metal-coordinated water molecules are conserved in UvrC. A similar octahedral arrangement is seen in the crystal structure of Serratia endonuclease (Miller et al, 1999), which also coordinates a single, catalytic divalent cation using a lone protein residue, in this case an asparagine. These observations strongly suggest that the bound divalent cation in UvrC is catalytically important and fulfills the role of the Lewis acid in catalysis as suggested for most nucleases (Galburt and Stoddard, 2002).

Since there is no obvious candidate to act as the general acid in the active site, we hypothesize that one of the water molecules coordinated to the metal performs this function. A similar one-metal mechanism is observed in the unrelated His-Cys box homing endonuclease I-PpoI, which, like UvrC, coordinates the divalent cation with only one residue, Asn 119 (Galburt, 1999). The active site of I-PpoI is similar to that of Serratia endonuclease mentioned above. In addition, there is no obvious general base in the N-terminal domain of UvrC, which is required to activate a nucleophilic water. In I-PpoI, a conserved histidine in the active site (His 98) fulfills this role by positioning and deprotonating the nucleophilic water (Galburt, 1999). In UvrC, one of the highly conserved tyrosines (Tyr 19, Tyr 29 or Tyr 43) may fulfill this function as there is no histidine present. Since the two full-length UvrC mutants Y19F and Y43F did not overexpress, we were only able to analyze the activity of the Y29F mutant showing that it has no detectable activity (Figure 5). Furthermore, the side-chain hydroxyl of Tyr 29 protrudes from the domain's surface and is in close proximity to the divalent cation (4.3 Å), which places it in a region of high positive charge (Figure 6). This arrangement could be required to lower its pKa so that it is able to serve as a general base. The hydroxyl of Tyr 29 also forms a hydrogen bond to one of the metal-coordinated water molecules, which in turn hydrogen bonds to the backbone carbonyl of Ile 30 (Figure 4). Assuming that this water molecule is a hydroxide, Tyr 29 could accept a proton from a nucleophilic water while transferring its proton to the metal-bound hydroxide. This suggests that Tyr 29 acts as a proton shuttle with the final acceptor being the metal-bound hydroxide (Figure 7). To analyze the importance of this residue, we solved the crystal structure of the isolated N-terminal domain Y29F mutant bound to manganese (see Supplementary Table I), confirming that it is identical to the wild-type protein with an r.m.s. deviation of 0.07 Å for all 89 Cα atoms. The analogous mutation in I-TevI (Y17A) retains 1% activity compared to the wild-type I-TevI although it is structurally compromised (Kowalski et al, 1999). This does not appear to be the case for the Y29F UvrC mutant since overexpression of the full-length mutant protein is comparable to wild type and the structures of the Y29F and wild type isolated domains are essentially identical. Perhaps the instability of I-TevI is due to the alanine mutation as opposed to phenylalanine in UvrC. However, we observe no change in expression level or solubility of full-length UvrC Y29A, suggesting that the GIY-YIG domain of UvrC could be more stable in general.

Figure 6
Electrostatic surface representation of the 3′ endonuclease domain calculated in the absence (left) and presence (right) of the bound metal ion. Surface charge was calculated using SPOCK at an ionic strength of 100 mM and is contoured at ±10 ...
Figure 7
Proposed reaction mechanism for 3′ phosphodiester bond cleavage by UvrC. The metal ion fulfills the role of the Lewis acid and one of the water molecules coordinated to the metal acts as a general acid. Tyr 29 acts as the general base and hydrogen ...

We do recognize that there are two other conserved tyrosines in the active site, Tyr 19 and Tyr 43, and we solved the structures of the N-terminal domain mutants Y19F and Y43F, which, in contrast to the full-length enzyme, overexpressed well. Similar to Y29F, both Y19F and Y43F bound the divalent cation and were nearly identical to the wild-type N-terminal domain with an r.m.s. deviation of 0.24 and 0.26 Å for all 89 Cα atoms, respectively. These two tyrosines were initially thought to be structurally important since they form hydrogen bonds to each other (Figure 4), but no significant structural perturbations can be observed upon mutation to phenylalanine. A mutation similar to Y19F was made in I-TevI (Y6A), which inactivated the protein and rendered it insoluble and/or unstable (Kowalski et al, 1999). The role of Tyr 19 and Tyr 43 in catalysis awaits further analysis. However, a close look at the position of the three tyrosines favors Tyr 29 as the general base, since Tyr 19 and Tyr 43 do not extend from the surface and are more distant from the divalent cation, resulting in an environment that is less positively charged and therefore less likely to lower the pKa (Figure 6).

I-PpoI contains a strictly conserved arginine, Arg 61, as do all GIY-YIG family members. The function of the arginine in I-PpoI is to stabilize the negative charge of the free 5′-phosphate after DNA cleavage (Galburt, 1999). A similar role can be imagined for Arg 39 in UvrC (Figure 7). The side chain of Arg 39 is completely solvent exposed and highly flexible as observed by incomplete side-chain density. The critical role of this conserved arginine for the activity of the N-terminal domain of UvrC and I-TevI has been previously demonstrated by site-directed mutagenesis (Derbyshire et al, 1997; Verhoeven et al, 2000). Here we confirm that a mutation of this conserved Arg 39 to alanine renders UvrC catalytically inactive (Figure 5). A second strictly conserved residue in UvrC, which may also be involved in stabilizing the product, is Lys 32. Lysine 32 is completely solvent accessible and is positioned in close proximity to the bound metal and Arg 39 (Figure 4). However, this residue is not strictly conserved within other GIY-YIG family members and is replaced by a serine in I-TevI (Figure 2A). Mutation of Lys 32 to alanine reduces its activity by 25–30% compared to wild-type UvrC (Figure 5), suggesting that it may play a role in stabilizing the negative charge of the free 5′-phosphate, but is clearly less important than Arg 39.

The role of Asn 88

The last of the four invariant residues is Asn 88. We originally predicted this residue to be structurally important since it forms two hydrogen bonds to Ile 30: one to the backbone amide and one to the backbone carbonyl (Figure 4). The backbone carbonyl of Ile 30, in turn, hydrogen bonds to the same manganese-coordinated water molecule as Tyr 29 (Figure 4). We solved the structure of the N88A mutant since it is in close proximity to the bound metal (Figure 4) and also renders the protein inactive (Figure 5). In I-TevI, the analogous mutation (N90A) leads to an enzyme with only 3% activity compared to the wild type protein (Kowalski et al, 1999). Like the three tyrosine mutant structures, N88A can still bind the divalent cation and the structure is nearly identical to wild-type UvrC with an r.m.s. deviation of 0.12 Å over all 89 Cα atoms. Therefore, N88A does not perturb the active site architecture as a water molecule substitutes for the side chain of Asn 88 in the structure. Due to its position at the C-terminus of the 3′ endonuclease domain, it can be speculated that this residue is important to position the catalytic domain correctly with respect to the other domains of UvrC. This is a critical requirement since it has been shown that the interaction between UvrC and the UvrB–DNA complex is located in a region C-terminal to the 3′ endonuclease domain (Moolenaar et al, 1997). Upon interaction of UvrC with the UvrB–DNA complex, the N-terminal domain would then be located in close proximity to the DNA, thus allowing the incision reaction.

Protein–DNA interactions

The GIY-YIG endonuclease domain, when isolated from either UvrC (data not shown) or I-TevI (Van Roey et al, 2002), is not able to bind or incise DNA. Apparently, a separate DNA binding and/or protein interacting domain is required to position the GIY-YIG domain properly on the DNA. For UvrC, this would be the UvrB binding domain and/or the C-terminal helix-hairpin-helix DNA binding motif (Moolenaar et al, 1997, 1998b; Sohi et al, 2000). Van Roey et al (2002) proposed that the GIY-YIG domain might interact with DNA in a similar way as I-PpoI (Galburt, 1999) based on similarities in the active sites of the two proteins. The amino acids in I-PpoI and UvrC that coordinate the divalent metal ion are both located in an α-helix (Asn 119 in α2 of I-PpoI and Glu 76 in α4 of UvrCN-Tma). I-PpoI inserts this helix into the minor groove (Flick et al, 1998) of the DNA (Figure 8), which leads to several protein–DNA interactions. We superimposed helix α4 of UvrC with the corresponding helix in I-PpoI while maintaining the position of Glu 76 (UvrC) with respect to Asn 119 (I-PpoI). This results in the overlap of Tyr 29 (UvrC) onto His 98 (I-PpoI) (Figure 8), which was initially encouraging since we hypothesize that Tyr 29 is the general base similar to His 98. Similar results were observed when I-TevI was superimposed with I-PpoI (Van Roey et al, 2002). However, in addition to Asn 119, another important residue is located in helix α2 of I-PpoI, Leu 116, which points into the minor groove and forms van der Waals contacts with an adenine. Mutation of Leu 116 has been reported to reduce the catalytic efficiency of I-PpoI (Galburt, 1999). After superposition with UvrC, Leu 116 aligned structurally with Phe 73 (Figure 8), a type conserved hydrophobic residue (Figure 3A and B). We mutated Phe 73 to examine its importance in comparison to I-PpoI. A mutation to either alanine or glutamate resulted in an enzyme that was as active as the wild-type protein (Figure 5), indicating that Phe 73 is most likely not involved in protein–DNA interactions or in a different way compared to I-PpoI. In addition to the mutational data, the superposition of UvrC and I-PpoI revealed that the catalytic metals of both enzymes are 3.6 Å apart (Figure 8). These results strongly suggest that the interaction of the GIY-YIG domain with DNA is different from that observed for I-PpoI despite the similarities in the positions of active site residues.

Figure 8
Superposition of helix α4 of the 3′ endonuclease domain of UvrC onto helix α2 of the I-PpoI–DNA complex. I-PpoI and UvrCN-Tma are illustrated as gray and green ribbon diagrams, respectively. Selected side chains of I-Ppo ...


We have solved the crystal structure of the N-terminal GIY-YIG endonuclease domain from two thermophilic organisms, B. caldotenax and T. maritima. This domain is shared by a number of very different proteins all belonging to the GIY-YIG family of endonucleases. In this study, we have characterized the active site of the N-terminal GIY-YIG endonuclease domain of UvrC including the catalytic divalent cation and its coordination geometry. Our results indicate that the N-terminal endonuclease domain of UvrC utilizes a novel one-metal mechanism to cleave the phosphodiester bond. We propose a mechanism where the Lewis acid is the divalent cation, the general acid is a metal-coordinated water molecule and the general base is Tyr 29.

Materials and methods

Protein expression and purification

N-terminal domains of UvrC from B. caldotenax (UvrCN-Bca; residues 1–98) and T. maritima (UvrCN-Tma; residues 1–97) were cloned into the pTXB1 vector (New England Biolabs) under control of the T7 promoter. The proteins were purified using the T7 IMPACTtm system (NEB). The proteins were expressed in BL21-CodonPlus®(DE3)-RIL cells and purified by chitin-affinity chromatography and size-exclusion chromatography. The concentration of UvrCN-Tma and UvrCN-Bca was determined by its absorption at 280 nm using calculated extinction coefficients of 7680 and 6400 M−1 cm−1, respectively.

Selenomethionine-substituted UvrCN-Bca was expressed in BL21-(DE3)-RIL cells by methionine biosynthesis inhibition. Cells were grown in M9 minimal medium with 0.2% glucose and 5% (v/v) glycerol as carbon sources. Purification was as described for the native protein with the exception of adding 5 mM DTT to the size-exclusion chromatography buffer.

Crystallization and data collection

Crystals of native and selenomethionine-derivatized UvrCN-Bca were grown by vapor diffusion, equilibrating equal volumes of protein solution (16.0 mg/ml) and precipitant solution containing 2 M (NH4)2SO4, 5 mM DTT and 100 mM sodium acetate pH 5.4, against a reservoir solution containing 2 M (NH4)2SO4, 5 mM DTT, 100 mM sodium acetate pH 5.4 and 250 mM NaCl. Crystals of UvrCN-Tma were grown by vapor diffusion, equilibrating equal volumes of protein solution (16.0 mg/ml) and precipitant solution containing 26% PEG 8000 and 100 mM Tris pH 8.5, against a reservoir solution containing 26% PEG 8000, 100 mM Tris pH 8.5 and 250 mM NaCl. Diffraction data of UvrCN-Bca and UvrCN-Tma crystals, cryocooled in liquid nitrogen, were collected at beam lines X26C and X12B, and X26C and X25, respectively, at the National Synchrotron Light Source at Brookhaven National Laboratory. Diffraction data were indexed, integrated and scaled using the HKL software (Otwinowski and Minor, 1997). Crystals of UvrCN-Bca belong to space group C2 with a=86.5 Å, b=86.7, c=67.9 and β=120.2 Å, and contain four molecules per asymmetric unit. Crystals of UvrCN-Tma belong to space group P43212 with a=55.0 Å and c=109.0 Å, and contain one molecule per asymmetric unit.

UvrCN-Tma crystals were transferred into 19% PEG 8000, 100 mM Tris pH 8.5, 25% glycerol and either 200 mM MnCl2 or 300 mM MgCl2. Crystals were soaked stepwise in solutions with increasing divalent cation and glycerol concentrations while simultaneously removing NaCl from the mother liquor. The total exposure time to the divalent cation was 1 h.

Structure solution and refinement

The structure of UvrCN-Bca was determined by MAD phasing using SOLVE (Terwilliger and Berendzen, 1999), which located 10 out of the 16 selenium sites. Phase refinement was performed with RESOLVE (Terwilliger, 2000), gradually extending the phases from 3.0 to 2.0 Å resulting in an ~50% complete model. The remainder of the model was built with O (Jones et al, 1991) and refinement was continued with REFMAC (Murshudov et al, 1997). Four-fold noncrystallographic symmetry (NCS) restraints were maintained initially and removed in the final refinement cycles. Water molecules were added with ARP (Perrakis et al, 1999). TLS refinement (one TLS group per monomer) was used in the final stages to account for overall anisotropic motion of the molecules. Additional calculations were performed using the CCP4 suite of programs (Bailey, 1994).

The structure of UvrCN-Tma bound to manganese was determined by molecular replacement using AMORE (Navaza, 1994). The search model was a modified monomer from the UvrCN-Bca structure where 18 dissimilar residues were changed to alanine. The resulting structure was initially refined at 1.5 Å by simulated annealing using CNS (Brunger et al, 1998). Side chains from UvrCN-Bca were replaced with those from UvrCN-Tma and refinement was continued with REFMAC and O.

UvrCN-Tma mutants

Single amino-acid residue substitution mutants of UvrCN-Tma were generated with the QuickChange Site-Directed Mutagenesis Kit (Stratagene) using pTXB1-uvrCN-Tma as template, and sense and antisense oligonucleotides specific for each mutant as PCR primers.

DNA substrates

All DNA substrates were synthesized by Sigma-Genosys (Woodlands, TX). The DNA sequence of the 50-mer double-stranded substrate containing a single internal fluorescein (FldT) adduct was F26 (5′-GACTACGTACTGTTACGGCTCCAT C[FldT]CTACCGCAATCAGGCCAGATCTGC-3′) while the complementary strand was NDB (5′-GCAGATCTGGCCTGATTGCGGTAGAGATGGAGCCG TAACAGTACGTAGTC-3′). The F26 strand was 5′ end-labeled using OptiKinase (USB Corporation) and [γ-32P]ATP (3000 Ci/mmol; Amersham Biosciences) according to the manufacturer's instructions. The reaction was terminated by the addition of EDTA and the enzyme was heat denatured by incubation for 10 min at 65°C. Alternatively, the F26 oligo was 3′ labeled using terminal transferase (Roche) and [α-32P]dideoxyATP (3000 Ci/mmol; Amersham Biosciences) according to the manufacturer's instructions. The reaction was terminated as described above. After either labeling reaction, free nucleotides were removed by gel filtration chromatography (Micro Biospin-6, Bio-Rad). The labeled oligonucleotide was annealed to the complementary oligonucleotide (NDB) using a 20% molar excess of NDB. The double-stranded character of the 50-mer duplex was confirmed by native polyacrylamide gel electrophoresis.

UvrABC incision assay

The 5′ or 3′ end-labeled duplex DNA (2 nM) was incised by the UvrABC enzymes (20 nM Bca UvrA, 100 nM Bca UvrB, 12.5 nM Tma UvrC) in 20 μl of UvrABC buffer (50 mM Tris–HCl pH 7.5, 50 mM KCl, 10 mM MgCl2, 1 mM ATP and 5 mM DTT) at 55°C for 30 min. The reaction was terminated by addition of EDTA (20 mM). A 2 μl portion of the reaction was added to 5 μl of formamide and blue dextran, and then heated to 85°C for 10 min. The incision products were resolved on a 10% denaturing polyacrylamide gel and electrophoresis was performed at 325 V for 40 min in Tris–Borate–EDTA buffer (89 mM Tris, 89 mM boric acid and 2 mM EDTA). The gels were dried and exposed to a phosphorimager screen (Molecular Dynamics) overnight. The incision efficiency was calculated using the Molecular Dynamics software ImageQuant.

Supplementary Material

Supplementary Table I


We thank Dr Annie Heroux of X26C, Dr Dieter Schneider of X12B and Dr Michael Becker of X25 for excellent technical assistance at the National Synchrotron Light Source (NSLS). This research was supported by DOE grant (DEFG02-01ER63073), NIH grant (GM070873), and the Pew Scholars Program in the Biomedical Sciences to CK. Beamline X26C at the NSLS in Brookhaven is supported in part by the State University of New York and its Research Foundation.


Accession numbers Coordinates for native, Mn2+-bound, Mg2+-bound, Y19F, Y43F, Y29F and N88A T. maritima UvrC, and native B. caldotenax UvrC have been deposited in the Protein Data Bank with accession codes 1YCZ, 1YD0, 1YD1, 1YD2, 1YD3, 1YD4, 1YD5 and 1YD6, respectively.


  • Aravind L, Walker DR, Koonin EV (1999) Conserved domains in DNA repair proteins and evolution of repair systems. Nucleic Acids Res 27: 1223–1242 [PMC free article] [PubMed]
  • Bailey S (1994) The CCP4 suite—programs for protein crystallography. Acta Crystallogr D 50: 760–763 [PubMed]
  • Boyce RP, Howard-Flanders P (1964) Release of ultra-violet light-induced thymine dimers from DNA in E. coli K-12. Proc Natl Acad Sci USA 51: 293–300 [PMC free article] [PubMed]
  • Brunger AT, Adams PD, Clore GM, DeLano WL, Gros P, Grosse-Kunstleve RW, Jiang JS, Kuszewski J, Nilges N, Pannu NS, Read RJ, Rice LM, Simonson T, Warren GL (1998) Crystallography and NMR systems (CNS): a new software system for macromolecular structure determination. Acta Crystallogr D 54: 905–921 [PubMed]
  • Caron PR, Kushner SR, Grossman L (1985) Involvement of helicase-II (UvrD gene product) and DNA Polymerase-I in excision mediated by the UvrABC protein complex. Proc Natl Acad Sci USA 82: 4925–4929 [PMC free article] [PubMed]
  • Christopher JA, Baldwin TO (1998) SPOCK: real-time collaborative molecular modeling. J Mol Graph 16: 285
  • DellaVecchia MJ, Croteau DL, Skorvaga M, Dezhurov OI, Lavrik B, Van Houten B (2004) Analyzing the handoff of DNA from UvrA to UvrB utilizing DNA–protein photoaffinity labeling. J Biol Chem 279: 45245–45256 [PubMed]
  • Derbyshire V, Kowalski JC, Dansereau JT, Hauer CR, Belfort M (1997) Two-domain structure of the td intron-encoded endonuclease I-TevI correlates with the two-domain configuration of the homing site. J Mol Biol 265: 494–506 [PubMed]
  • Flick KE, Jurica MS, Monnat RJ Jr, Stoddard BL (1998) DNA binding and cleavage by the nuclear intron-encoded homing endonuclease I-PpoI. Nature 394: 96–101 [PubMed]
  • Friedberg E, Walker J, Siede W (1995) DNA Repair and Mutagenesis. Washingtion, DC: American Society of Microbiology Press
  • Galburt EA (1999) A novel endonuclease mechanism directly visualized for I-PpoI. Nat Struct Biol 6: 1096–1099 [PubMed]
  • Galburt EA, Stoddard BL (2002) Catalytic mechanisms of restriction and homing endonucleases. Biochemistry 41: 13851–13860 [PubMed]
  • Goosen N, Moolenaar GF (2001) Role of ATP hydrolysis by UvrA and UvrB during nucleotide excision repair. Res Microbiol 152: 401–409 [PubMed]
  • Holm L, Sander C (1995) Dali: a network tool for protein structure comparison. Trends Biochem Sci 20: 478–480 [PubMed]
  • Husain I, Houten BV, Thomas DC, Abdel-Monem M, Sancar A (1985) Effect of DNA polymerase I and DNA helicase II on the turnover rate of UvrABC excision nuclease. Proc Natl Acad Sci USA 82: 6774–6778 [PMC free article] [PubMed]
  • Jones TA, Zou JY, Cowan SW, Kjeldgaard M (1991) Improved methods for building protein models in electron density maps and the location of errors in these models. Acta Crystallogr A 47: 110–119 [PubMed]
  • Koradi R, Billeter M, Wuthrich K (1996) MOLMOL: a program for display and analysis of macromolecular structures. J Mol Graph 14: 29–32 [PubMed]
  • Kowalski JC, Belfort M, Stapleton MA, Holpert M, Dansereau JT, Pietrokovski S, Baxter SM, Derbyshire V (1999) Configuration of the catalytic GIY-YIG domain of intron endonuclease I-TevI: coincidence of computational and molecular findings. Nucleic Acids Res 27: 2115–2125 [PMC free article] [PubMed]
  • Kraulis PJ (1991) MOLSCRIPT—a program to produce both detailed and schematic plots of protein structures. J Appl Crystallogr 24: 946–950
  • Laskowski RA, MacArthur MW, Moss DS, Thornton JM (1993) PROCHECK: a program to check the stereochemical quality of protein structures. J Appl Crystallogr 26: 283–291
  • Lin JJ, Sancar A (1992a) (A)BC excinuclease: the Escherichia coli nucleotide repair enzyme. Mol Microbiol 6: 2219–2224 [PubMed]
  • Lin JJ, Sancar A (1992b) Evidence for 5′ incision by UvrC through a catalytic site involving Asp399, Asp438, Asp466, and His538 residues. J Biol Chem 267: 17688–17692 [PubMed]
  • Lloyd RS, Van Houten B (1995) DNA damage recognition. In DNA Repair Mechanisms: Impact on Human Diseases and Cancer, Vos J-M (ed) pp 25–66. Austin, TX: RG Landes Company, Biomedical Publishers
  • Merritt EA, Murphy MEP (1994) Raster3D Version 2.0—a program for photorealistic molecular graphics. Acta Crystallogr D 50: 869–873 [PubMed]
  • Miller MD, Cai J, Krause KL (1999) The active site of Serratia endonuclease contains a conserved magnesium-water cluster. J Mol Biol 288: 975–987 [PubMed]
  • Moolenaar GF, Bazuine M, van Knippenberg IC, Visse R, Goosen N (1998a) Characterization of the Escherichia coli damage-independent UvrBC endonuclease activity. J Biol Chem 273: 34896–34903 [PubMed]
  • Moolenaar GF, Franken KLMC, Dijkstra DM, Thomas-Oates JE, Visse R, van de Putte P, Goosen N (1995) The C-terminal region of the UvrB protein of Escherichia coli contains an important determinant for UvrC binding to the preincision complex but not the catalytic site for 3′-incision. J Biol Chem 270: 30508–30515 [PubMed]
  • Moolenaar GF, Franken KLMC, van de Putte P, Goosen N (1997) Function of the homologous regions of the Escherichia coli DNA excision repair proteins UvrB and UvrC in stabilization of the UvrBC–DNA complex and in 3′-incision. Mutat Res 385: 195–203 [PubMed]
  • Moolenaar GF, Uiterkamp RS, Zwijnenburg DA, Goosen N (1998b) The C-terminal region of the Escherichia coli UvrC protein, which is homologous to the C-terminal region of the human ERCC1 protein, is involved in DNA binding and 5′ incision. Nucleic Acids Res 26: 462–468 [PMC free article] [PubMed]
  • Murshudov G, Vagin A, Dodson E (1997) Refinement of macromolecular structures by the maximum likelihood method. Acta Crystallogr D 53: 240–255 [PubMed]
  • Navaza J (1994) AMoRe: an automated package for molecular replacement. Acta Crystallogr A 50: 157–163
  • Orren DK, Sancar A (1990) Formation and enzymatic properties of the UvrB–DNA complex. J Biol Chem 265: 15796–15803 [PubMed]
  • Orren DK, Selby CP, Hearst JE, Sancar A (1992) Post-incision steps of nucleotide excision repair in Escherichia coli. Disassembly of the UvrBC-DNA complex by helicase II and DNA polymerase I. J Biol Chem 267: 780–788 [PubMed]
  • Otwinowski Z, Minor W (1997) Processing of X-ray diffraction data collected in oscillation mode. In Methods in Enzymology, Carter CW, Sweet RM (eds) Vol. 276, pp 307–326. New York: Academic Press
  • Perrakis A, Morris R, Lamzin VS (1999) Automated protein model building combined with iterative structure refinement. Nat Struct Biol 6: 458–463 [PubMed]
  • Sancar A (1994) Mechanisms of DNA excision repair. Science 266: 1954–1956 [PubMed]
  • Sancar A (1996) DNA excision repair. Annu Rev Biochem 65: 43–81 [PubMed]
  • Sancar A, Rupp WD (1983) A novel repair enzyme: UvrABC excision nuclease of Escherichia coli cuts a DNA strand on both sides of the damaged region. Cell 33: 249–260 [PubMed]
  • Setlow RB, Carrier WL (1964) The disappearance of thymine dimers from DNA: an error-correcting mechanism. Proc Natl Acad Sci USA 51: 226–231 [PMC free article] [PubMed]
  • Skorvaga M, Theis K, Mandavilli BS, Kisker C, Van Houten B (2002) The beta-hairpin motif of UvrB is essential for DNA binding, damage processing, and UvrC-mediated incisions. J Biol Chem 277: 1553–1559 [PubMed]
  • Sohi M, Alexandrovich A, Moolenaar G, Visse R, Goosen N, Vernede X, Fontecilla-Camps J, Champness J, Sanderson MR (2000) Crystal structure of Escherichia coli UvrB C-terminal domain, and a model for UvrB–UvrC interaction. FEBS Lett 465: 161–164 [PubMed]
  • Terwilliger TC (2000) Maximum likelihood density modification. Acta Crystallogr D 56: 965–972 [PMC free article] [PubMed]
  • Terwilliger TC, Berendzen J (1999) Automated MAD and MIR structure solution. Acta Crystallogr D 55: 849–861 [PMC free article] [PubMed]
  • Theis K, Skorvaga M, Machius M, Nakagawa N, Van Houten B, Kisker C (2000) The nucleotide excision repair protein UvrB, a helicase-like enzyme with a catch. Mutat Res 460: 277–300 [PubMed]
  • Van Houten B (1990) Nucleotide excision repair in Escherichia coli. Microbiol Rev 54: 18–51 [PMC free article] [PubMed]
  • Van Roey P, Meehan L, Kowalski JC, Belfort M, Derbyshire V (2002) Catalytic domain structure and hypothesis for function of GIY-YIG intron endonuclease I-TevI. Nat Struct Biol 9: 806–811 [PubMed]
  • Verhoeven EE, Wyman C, Moolenaar GF, Goosen N (2002) The presence of two UvrB subunits in the UvrAB complex ensures damage detection in both DNA strands. EMBO J 21: 4196–4205 [PMC free article] [PubMed]
  • Verhoeven EEA, van Kesteren M, Moolenaar GF, Visse R, Goosen N (2000) Catalytic sites for 3′- and 5′ incision of E. coli excision repair are both located in UvrC. J Biol Chem 275: 5120–5123 [PubMed]
  • Zou Y, Van Houten B (1999) Strand opening by the UvrA2B complex allows dynamic recognition of DNA damage. EMBO J 18: 4889–4901 [PMC free article] [PubMed]

Articles from The EMBO Journal are provided here courtesy of The European Molecular Biology Organization
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...