DNA MMR (mismatch repair) is an excision repair system that removes mismatched bases generated primarily by failure of the 3′–5′ proofreading activity associated with replicative DNA polymerases. MutL proteins homologous to human PMS2 are the endonucleases that introduce the entry point of the excision reaction. Deficiency in PMS2 function is one of the major etiologies of hereditary non-polyposis colorectal cancers in humans. Although recent studies revealed that the CTD (C-terminal domain) of MutL harbours weak endonuclease activity, the regulatory mechanism of this activity remains unknown. In this paper, we characterize in detail the CTD and NTD (N-terminal domain) of aqMutL (Aquifex aeolicus MutL). On the one hand, CTD existed as a dimer in solution and showed weak DNA-binding and Mn2+-dependent endonuclease activities. On the other hand, NTD was monomeric and exhibited a relatively strong DNA-binding activity. It was also clarified that NTD promotes the endonuclease activity of CTD. NTD-mediated activation of CTD was abolished by depletion of the zinc-ion from the reaction mixture or by the substitution of the zinc-binding cysteine residue in CTD with an alanine. On the basis of these results, we propose a model for the intramolecular regulatory mechanism of MutL endonuclease activity.
DNA MMR (mismatch repair) is a well-conserved DNA repair system and corrects mismatched base pairs produced through replication errors, genetic recombinations and other processes [1,2]. Deficiencies in MMR have been implicated in more than 90% of human hereditary non-polyposis colorectal cancers . The mechanism of MMR has been well characterized in Escherichia coli. The MMR system of this bacterium has been reconstituted using recombinant proteins . In the E. coli MMR system, a mismatched base is recognized by a MutS homodimer [5,6]. A MutL homodimer interacts with the MutS–DNA complex, and then the MutH restriction endonuclease is activated by MutL . MutH nicks the unmethylated strand at the hemi-methylated GATC site to introduce an entry point for the excision reaction and to direct the repair to the error-containing DNA strand. The error-containing strand is removed by helicases and exonucleases, and a new strand is synthesized by DNA polymerase III and ligase. Although homologues of E. coli MutS and MutL exist in almost all organisms, no homologue of E. coli MutH has been identified in the majority of organisms including eukaryotes and most bacteria .
In humans and yeasts, it has been verified that a strand discontinuity serves as the signal that directs the repair to the error-containing strand of a mismatched heteroduplex [2,4]. In newly synthesized strands, the discontinuities can exist as 3′ ends or termini of Okazaki fragments. For biochemical characterization of MMR, a nicked plasmid DNA has been used as a model substrate containing a strand discontinuity. For this assay system, it has been reported that the shorter path from a nick to the mismatch is removed by the excision reaction , indicating that 5′- and 3′-directed MMR are distinct. Intriguingly, the 5′ to 3′ exonuclease activity of EXO1 is required for both 5′- and 3′-directed strand removal [10,11]. The requirement of 5′ to 3′ exo-nuclease activity for the 3′-directed excision reaction was explained by the discovery that human and yeast MutL homologues, MutLα (MLH1-PMS2 heterodimer) proteins, harbour endonuclease activity that can incise on both sides of a mismatch, and the 5′ to 3′exonuclease activity of EXO1 removes the DNA segment spanning the mismatch [12–14].
The DQHA(X)2E(X)4 motif in the CTD (C-terminal domain) of the PMS2 subunit of MutLα has also been shown to comprise the metal-binding site, which is essential for the endonuclease activity . Unlike E. coli and closely related bacteria, most MutH-lacking bacteria possess MutL homologues that contain the metal-binding motif [12,15–17]. Therefore, the molecular mechanism established based on the results obtained for eukaryotic systems would be universal for organisms lacking MutH. In agreement with this prediction, recent studies showed that MutL homologues from Thermus thermophilus, Neisseria gonorrhoeae and Aquifex aeolicus possess endonuclease activity [15–18]. The overall structure of the MutL CTD has also been predicted to be common to all organisms, in the absence or presence of the metal-binding motif [19,20]. In fact, the crystal structure of the CTD of Bacillus subtilis MutL endonuclease  is highly homologous to that of E. coli MutL [22,23], which is composed of two subdomains, the internal and external subdomains. The crystal structure of B. subtilis MutL CTD also revealed the existence of characteristic two zinc-binding sites near the nuclease catalytic site , which had been indicated by Kosinski et al. . The zinc-binding sites consist of several conserved motifs including the CPHGRP motif (Figure 1A), which is not found in E. coli MutL CTD. The binding of zinc ions causes a rearrangement of the local structure near the catalytic site of B. subtilis MutL CTD . Although the significance of this structural rearrangement remains unknown, perturbation of the zinc-binding ability leads to the loss of in vivo DNA repair activity .
C-terminal domains of MutL homologues
MutL homologues belong to the GHL ATPase family [24–26], which contain the Bergerat ATP-binding fold in their NTD (N-terminal domain) and undergoes large conformational changes upon ATP-binding and/or -hydrolysis [27,28]. It is believed that ATPase cycle-dependent conformational changes of a MutL homologue are necessary to perform the MMR reaction. Recently, ATP-induced large conformational changes of Saccharomyces cerevisiae MutLα and Thermotoga maritima MutL were observed with atomic force microscopy  and small-angle X-ray scattering analysis , respectively. Since the endonuclease activity of MutL CTD is extremely weak and non-specific [16,17,21], it is expected that there is a regulatory mechanism to ensure the strong activity specific for mismatched DNA. The observed ATP-dependent conformational change of MutL may be involved in such a regulatory mechanism.
In this study, we prepared Aquifex aeolicus MutL (aqMutL) CTD and NTD, and investigated their biochemical characteristics. aqMutL CTD showed extremely weak DNA-binding and endonuclease activity. In contrast, aqMutL NTD exhibited a relatively strong DNA-binding activity and an ATP-dependent conformational change. Finally, it was clarified that aqMutL NTD stimulates the endonuclease activity of CTD in a zinc-ion-dependent manner.
MATERIALS AND METHODS
Construction of expression plasmids
DNA fragments expressing aqMutL CTD and NTD were generated by PCR using pET-11a/aqMutL  as a template. The forward and reverse primers used for the amplification were 5′-ATATCATATGCCCCTCTCTCAACCTGTGAAAACTTAC-3′ and 5′-ATATAGATCTTTATCAGTAATTCCTGCCTACTTTTTCGTA-3′ (BEX), and 5′-TATACATATGTTTGTAAAGTTACTTCCTCC-3′ and 5′-TATAAGATCTACTAAATATCTACGATAGGTTTCTCTTTTC-3′ for aqMutL CTD and NTD genes, respectively. The forward and reverse primers contained Nde I and Bgl II sites, respectively (underlined). The amplified aqMutL CTD and NTD fragments were ligated into the Nde I and Bam HI sites of pET-11a (Novagen) to yield pET-11a/aqMutL CTD and pET-11a/aqMutL NTD plasmids. Sequence analysis revealed that the constructions were error free.
Expression and purification of proteins
E. coli BL21(DE3) (Novagen) was transformed with pET-11a/aqMutL CTD and then cultivated at 37°C in 1.5 litres of YT medium [0.8% (w/v) tryptone (Difco), 0.5% (w/v) yeast extract (Oriental Yeast) and 0.5% (w/v) NaCl]  containing 50 μg/ml ampicillin. When the density of cultures reached 4 × 108 cells/ml, isopropyl β-D-thiogalactopyranoside (Wako) was added to 0.1 mM. The cells were grown at 37°C for 4 h after induction and then harvested by centrifugation. The cells were lysed by sonication in buffer I (20 mM Tris/HCl, pH 8.0, and 50 mM NaCl) and heated at 70°C for 10 min. After centrifugation at 48000 g for 60 min, the resulting supernatant was loaded on to a SP-Sepharose column (40 ml) (GE Healthcare Biosciences) pre-equilibrated with buffer I. Although aqMutL CTD is a highly acidic protein, it binds to SP-Sepharose at pH 8.0. The column was washed with 100 ml of buffer I and then eluted with a 300 ml gradient of 0.05–1 M NaCl in buffer I. The fractions containing aqMutL CTD were detected by SDS/PAGE and collected, and ammonium sulfate was added to the fraction to yield a final concentration of 1 M. The solution was loaded onto a Toyopearl-Phenyl column (40 ml) (TOSOH) equilibrated with buffer I containing 1 M ammonium sulfate. The column was washed with 100 ml of buffer I containing 1 M ammonium sulfate, and then eluted with a 300 ml gradient of 1–0 M ammonium sulfate in buffer I. The fractions containing aqMutL CTD were detected by SDS/PAGE and concentrated with a Vivaspin concentrator (Vivascience). The concentrated solution was applied to a Superdex 75 HR column (24 ml; GE Healthcare Biosciences) pre-equilibrated with buffer I using an ÄKTA system (GE Healthcare Biosciences). The full-length aqMutL was over expressed and purified by the same procedure as that for aqMutL CTD using pET-11a/aqMutL plasmid. aqMutL NTD was also over expressed and purified by the same procedure as that for aqMutL CTD using a Rosetta 2(DE3) (Novagen) competent cell and pET-11a/aqMutL NTD.
The molar absorption coefficients (ϵ) of aqMutL CTD, the full-length aqMutL and aqMutL NTD were calculated to be 19 005, 30 765 and 11 760 M−1·cm−1, respectively, at an absorption maximum of ~278 nm, using the procedure described previously .
CD measurements were carried out with a Jasco spectropolarimeter, model J-720W (Jasco). All measurements were performed using a 0.1cm cell. The residue molar ellipticity [θ] was defined as 100 θobs /(lc), where θobs was the observed ellipticity, l the length of the light path in centimetres and c the residue molar concentration of each protein. The measurements were performed at 25°C in a solution comprising 50 mM potassium phosphate (pH 7.0), 100 mM KCl and 5 μM aqMutL CTD.
The nuclease assay was performed by modifying the procedure reported previously . Supercoiled plasmid DNA (pT7Blue; Novagen) was prepared by using a minipreparation kit, Wizard plus SV Minipreps (Promega). Since contaminations of EDTA, salts, denaturant and other agents significantly perturb the results in this experiment, the washing step was carefully performed. The plasmid DNA was dissolved in H2O, stored at −20°C, and subjected to the nuclease assay in a few days. The 50 ng/μl supercoiled pT7Blue was incubated with various concentrations of freshly prepared aqMutL CTD in 100 mM Hepes (pH 7.5), 100 mM KCl, 5 mM MnCl2 and 1 mM dithiothreitol at 55°C for 30 min, unless otherwise noted in the Figure legends. In the metal dependence experiment, 5 mM MgCl2, NiCl2, CoCl2 or CaCl2 was added in place of MnCl2. In the examination of the effect of aqMutL NTD on the endonuclease activity of CTD, ZnCl2 was added to yield a final concentration of 0.5 or 1 mM, and NTD was pre-incubated with CTD prior to the addition of substrate DNA (details are shown in the Figure legends). The reactions were stopped by adding 5×loading buffer (5 mM EDTA, 1% SDS, 50% glycerol and 0.05% Bromophenol Blue). The reaction solutions were loaded onto a 1.0% agarose S (Takara) gel containing 0.5× TBE buffer (89 mM Tris-borate and 2 mM EDTA) and electrophoresed in the same buffer. The gel was stained with ethidium bromide and the DNA fragments were detected under UV light at 254 nm. The amounts of digested and undigested DNA were determined using the ImageJ software, which is a freely available image processing and analysis program developed at the National Institutes of Health.
Electrophoretic mobility-shift assay
Synthesized 30-mer single-stranded DNA, 5′-CGGTATCTTGACTATGACCGCTCTACGAGC-3′ (BEX), was annealed to complementary single-stranded DNA (BEX) to obtain 30-bp double-stranded DNA. The 200 nM 30-bp double-stranded DNA was incubated with various concentrations of aqMutL CTD or NTD in 80 mM Hepes (pH 7.5), 100 mM KCl, 1 mM dithiothreitol and 0.2 mg/ml BSA for 30 min at 25°C. The concentrations of proteins used are indicated at the top of the Figure panels. The reaction mixture (10 μl) was loaded on to an SDS/9% polyacrylamide gel and then electrophoresed in 1× TBE buffer (89 mM Tris-borate and 2 mM EDTA). The gel was stained with SYBR Gold (Invitrogen) and detected under UV light at 254 nm. The amounts of free and shifted DNAs were determined using the ImageJ software.
Surface plasmon resonance assay
The surface plasmon resonance assay was performed using a BIAcore 3000 instrument (BIAcore AB). The 5′-biotinylated 41-mer single-stranded DNA 5′-CGCCGAATTGCTAGCAAGCTATCGAGTCTAAAAATTCGGCT-3′ (BEX Co.) was synthesized and hybridized with a complementary strand (BEX) to yield 41-bp double-stranded DNA. The 10 μM biotinylated 41-bp double-stranded DNA was immobilized on the streptoavidin-coated sensor chip to obtain the proper resonance units. Experiments were performed at 25°C in a buffer containing 10 mM Hepes (pH 7.4), 3 mM EDTA, 0.005% Tween 20 and 54 mM NaCl. A 50 μl volume of 10 μM aqMutL CTD was injected at a flow rate of 10 μl/min. Bound species were dissociated by injection of 2 M NaCl containing 0.2% SDS. The control sensorgram for the sensor chip containing no DNA was subtracted from measured sensorgrams.
Size-exclusion chromatography was performed at 25°C on a Superdex 75 HR column (1×30 cm; GE Healthcare Biosciences) using an ÄKTA system (GE Healthcare Biosciences). A 100 μl volume of aqMutL CTD (4, 8, 12 or 16 μM) or aqMutL NTD (10 μM) was incubated in 50 mM Tris/HCl (pH 7.5) and 300 mM KCl at 25°C for 16 h. Then, the resulting protein solution was loaded onto the Superdex 75 HR column and eluted at the flow rate of 0.5 ml/min in the same buffer. The elution profile was monitored by recording the absorbance at 280 and 260 nm. The Superdex 75 HR column was calibrated using BSA (66 kDa), ovalbumin (45 kDa), carbonic anhydrase (29 kDa) and cytochrome c (12.4 kDa).
DLS (dynamic light scattering) experiment
aqMutL CTD (1.2 mg/ml) was prepared in 20 mM Tris/HCl (pH 8.0) containing 150 mM NaCl, and was passed through a 0.02 μm, VectaSpin Micro-centrifuge filter (Whatman). A 25 μl volume of the protein solution was loaded into a quartz cuvette, and then analysed using a DLS instrument, DynaPro MSXTC/12/F with a gallium-arsenite diode laser, DynaPro-99-E-50 (Protein Solutions Inc., Charlottesville, U.S.A) at 25°C. The data were analysed using the Dynamics version 5.26.60 (Protein Solutions). Two independently prepared samples were analysed and 50 measurements were made for each sample. The hydrodynamic radius (Rh) value was calculated with the Stokes–Einstein equation (eqn 1) using the obtained translational diffusion coefficient (DT) :
where kB is the Boltzmann constant, T the absolute temperature, η the solvent viscosity and Rh the hydrodynamic radius. Molecular mass of the protein in the solution was estimated from Rh using an empirical curve of known proteins (eqn 2).
aqMutL CTD (5 or 10 μM) was incubated with 0.1, 0.3, 0.6, 1 or 2 μM DIDS (4,4′-di-isothiocyanostilbene-2,2′-disulfonate) (Dojindo) in 20 mM Hepes (pH 7.5), 100 mM KCl and 1 mM dithiothreitol at 37°C for 30 min. Reaction products were separated by SDS/12.5% PAGE and stained with Coomassie Brilliant Blue R-250.
Materials for MS
A nano-scale analytical column, NS-AC-10–C18 BioSphere C18 (5 μm particle size, 120 Å pore size, 75 μm inner diameter, 100 mm length) and a guard column, NS-MP-10 BioSphere C18 (5 μm particle size, 120 Å pore size, 100 μm inner diameter and 20 mm length), were purchased from NanoSeparations. HPLC grade Solution A (0.1% formic acid in double distilled water) and Solution B (0.1% formic acid in acetonitrile) were purchased from Wako. Modified trypsin (sequencing grade) was purchased from Promega.
In-gel trypsin digestion
The protein spot was excised from the SDS/PAGE after washing the gel twice with distilled water. The excised protein spot was further washed a couple of times with 100 mM ammonium bicarbonate buffer (pH 8.0) and 100% acetonitrile alternately and then completely dried with a speed vacuum drier, Sakuma EC-57CS for 1 h. A1 μl volume of trypsin solution (20 ng/μl, w/v) was directly applied to the gel followed by rehydration for 30 min on ice and then 9 μl of trypsin digestion buffer (40 mM ammonium bicarbonate, pH 8.0, 10% acetonitrile) was added. Tryptic digestion was carried out at 37°C overnight.
nLC (nano-scale liquid chromatography)-MS/MS (tandem MS) and data analyses
The tryptic-digested peptides were separated with an EASY-nLC (Proxeon), and their mass spectra were analysed with a micrOTOF-QII (Bruker Daltonics). Approx. 100 ng of tryptic-digested peptides was injected into the pre-column at a flow rate of 10 μl/min using the autosampler of the EASY-nLC, and the pre-column was washed with Solution A at a flow rate of 10 μl/min for 10 min. The desalted peptides were subsequently separated on an analytical column at a flow rate of 200 nl/min. The solvent gradient was started at 5% Solution B and linearly increased to 50% Solution B in 60 min. The peptides separated on a C18 column were introduced into the micrOTOF-QII mass spectrometer with nanoelectrospray ionization in the positive mode. The capillary voltage was −4.5 kV and drying gas temperature was set at 200°C. The collision energy of the quadrupole for MS/MS fragmentation of each peptide using Ar gas was set at 20–40 eV depending on its charged state and size. MS/MS fragmentation of precursor ions was carried out automatically in the scan mode with the three most abundant precursor ions in each precursor ion scan selected.
The mass spectra acquired with the micrOTOF-QII were processed with DataAnalysis 3.4 (Bruker Daltonics) and Biotools 3.2 (Bruker Daltonics). The mass tolerance for both the precursor and MS/MS fragmented ions was set at ±0.1 Da. The amino acid sequences of peptides that were not coupled with DIDS were automatically identified by Biotools 3.2 with Sequence Editor 3.2 (Bruker Daltonics), whereas the sequence of the DIDS cross-linked peptide was manually determined with 0.05 Da MS/MS tolerance.
Limited proteolysis was performed according to the procedure reported previously . In short, an aliquot of full-length (4 μM), CTD (10 μM) or NTD (5 μM) of aqMutL was reacted with various concentrations of trypsin. The concentrations of trypsin are indicated at the top of the Figure panels. The reaction was performed in a buffer consisting of 50 mM Tris/HCl (pH 7.5), 100 mM KCl, 1 mM dithiothreitol, 5 mM MgCl2 and 0 or 0.3 mM ATP at 37°C for 10 min, and stopped by the addition of an equal volume of SDS-containing dye (125 mM Tris/HCl, pH 7.5, 10% 2-mercaptoethanol, 4% SDS, 10% sucrose and 0.01% Bromophenol Blue). The digests were separated by SDS/7.5% PAGE and stained with Coomassie Brilliant Blue.
Prediction of the tertiary structure of aqMutL CTD
The N-terminus of aqMutL CTD (amino acid residues 316–425 of aqMutL) was determined on the basis of amino acid sequence comparison of MutL homologues from MutH-less organisms (Figure 1A). Although the C-terminal region of aqMutL lacks approx. 90 amino acid residues corresponding to the central region of the CTD of a MutL homologue (Figure 1A), the C-terminal region contains a metal-binding motif, DQHA(X)2E(X)4E, which is essential for the endonuclease activity. We constructed a model structure of aqMutL CTD by using the Rosetta program provided by D. Baker's group [36–38]. The Rosetta program built a model structure by means of the Rosetta fragment insertion method . The model structure suggests that aqMutL CTD lacks the internal subdomain, as expected from the amino acid sequence alignment (Figures 1B and 1C).
Biochemical properties of aqMutL CTD
aqMutL CTD was successfully overexpressed in E. coli and purified to homogeneity, as shown in Figure 2(A) (the calculated molecular mass of aqMutL CTD is 13 kDa). It should be noted that the aqMutL CTD was extremely soluble although the wild-type aqMutL exhibited limited solubility. The far-UV CD spectrum of purified aqMutL CTD was examined to obtain information on the conformation of its polypeptide backbone. The spectrum showed a broad negative signal from 210 to 220 nm (Figure 2B) and suggested the presence of α-helix and β-sheet. The helix content of aqMutL CTD was estimated to be approx. 20% on the basis of the [θ]222 value . The thermostability of aqMutL CTD was also examined based on the mean residue ellipticity at 222 nm. As a result, aqMutL CTD was found to be stable up to 95°C (Figure 2C), indicating the proper folding and the extreme thermostability of aqMutL CTD.
CD measurement of aqMutL CTD
To confirm that the recombinant aqMutL CTD is biochemically active, the endonuclease activity of aqMutL CTD was examined using a CCC (covalently closed circular form) of plasmid DNA as a substrate. The CCC of plasmid DNA is often used in in vitro experiments to test the latent endonuclease activities of DNA repair and recombination enzymes [12,41,42]. As shown in Figure 3(A), in the presence of manganese ions, aqMutL CTD relaxed the CCC of plasmid DNA to generate the nicked OC (open-circular form) of the plasmid DNA, indicating that aqMutL CTD retains the endonuclease activity. This is consistent with the recent report that bacterial MutL CTD retains the endonuclease activity [16,17]. It was revealed that nickel and cobalt ions can substitute for manganese ions (Figure 3B). Such divalent ion-specificity of the nuclease activity of aqMutL CTD is exactly the same as that of full-length aqMutL . Interestingly, the endonuclease activity of aqMutL CTD was not stimulated by calcium ions, unlike that of N. gonorrhoeae MutL CTD . The activity increased with increasing reaction temperatures from 25 to 70°C (Figure 3C). These results strongly show that the C-terminal region of aqMutL is sufficient to form the active site of the extremely thermostable endonuclease domain. Although aqMutL CTD showed endonuclease activity in the presence of manganese, nickel or cobalt ions, which metal ion is required for the endonuclease activity in vivo should be carefully investigated, as pointed out by Kosinski et al. . Although we could not detect the DNA-binding activity of aqMutL CTD by electrophoretic mobility-shift assay (Figure 3D), the surface plasmon resonance experiment revealed its extremely weak DNA-binding activity (Figure 3E). The dissociation constant for 41-bp double-stranded DNA was approx. 40 μM.
Endonuclease activity of aqMutL CTD
Quaternary structure of aqMutL CTD
The oligomeric state of aqMutL CTD was examined by means of a DLS experiment. The measurement gave an Rh value of 2.59 nm, suggesting that the molecular mass of the particle in the sample is 31.3 kDa (Figure 4A). Since the molecular mass of the single aqMutL CTD was calculated to be 13 kDa according to its amino acid sequence, this result implies that aqMutL CTD exists in a dimeric state in solution. To further analyse the self-association ability of aqMutL CTD, size-exclusion chromatography was performed. The elution profile of aqMutL CTD exhibited a single peak at 30 kDa corresponding to a dimeric state (Figure 4B). The dimerization of aqMutL CTD was also confirmed by the cross-linking experiments involving the homo-bifunctional cross-linker, DIDS, which cross-links two amino groups that are in close proximity to each other [33,43]. The covalently cross-linked aqMutL CTD molecules became detectable as the concentration of DIDS increased (Figure 4C). The same amount of the dimeric molecule was detected regardless of the concentration of the protein, suggesting the dimerization of aqMutL CTD. This is in good agreement with the results obtained in DLS and size-exclusion chromatography experiments. As previously reported, the C-terminal region is essential for the dimerization property of bacterial MutL [22,44]. Such dimerization ability should be a common feature of the CTD of MutL homologues regardless of sequence diversity.
Self-association of aqMutL CTD
In order to identify the dimerization interface, we identified the cross-linked site in aqMutL CTD by MS/MS. The DIDS cross-linked dimeric aqMutL CTD was in-gel digested with trypsin and extracted from an SDS/PAGE. The digested peptides were separated by nLC and subsequently introduced into the ESI-Qq-TOF mass spectrometer to identify its amino acid sequences by MS/MS. Figure 5(A) presents the total ion chromatography of MS/MS spectra of all detected peptides. The DIDS cross-linked peptide was detected at a retention time of 37.7 min with 995.423 m/z as +2 charged state (Figure 5B), indicating that the monoisotopic neutral mass of the cross-linked product is 1988.830. The result of MS/MS of the 995.423 m/z peak assigned the monoisotopic mass of DIDS, 453.93, between the y7 and y8 ions, and spectral differences between detected y-series ions showed a symmetrical arrangement from DIDS (Figure 5C). The deduced amino acid sequence from MS/MS spectra of the fragmented precursor ion, 995.423 m/z, indicated ‘KVPQSLP-(DIDS)-PLSQPVK’ with the y-series ions, which are dissociated from the C-terminal peptide. The deduced amino acid sequence corresponded to an N-terminal tryptic-digested fragment of aqMutL CTD of which the first methionine residue was processed. In addition, the detected monoisotopic neutral mass of the precursor ion, 1988.830, exactly matched the theoretical mass of DIDS that is cross-linked to the same two N-terminal peptides, 1988.850, within the 10 ppm mass measurement accuracy. These results indicate that the imine of the N-terminal proline residue of one aqMutL CTD was cross-linked to that of the other one with DIDS; in other words, two isothiocyanosulfate groups of DIDS were coupled with each N-terminus of two aqMutL CTDs. The α-imino group of an N-terminal proline can contain a lone pair and attack the nucleophilic centre of DIDS. For example, an N-terminal proline is modified by an isothiocyanate group during Edman degradation . The low efficiency of the cross-linking in this experiment would be due to the physiological pH condition under which the experiments were carried out because the pKa value of the α-imino group of the proline is approx. 10. In addition, assuming that aqMutL CTD dimerizes via the external subdomain, the distance between the two N-terminal proline residues is predicted to be approx. 15 Å which is comparable to the length of DIDS (14 Å) (Figure 5D). No additional DIDS cross-linked peptide was detected in this analysis (results not shown). Based on these results, it was concluded that the two N-terminal regions of dimeric aqMutL CTD are closely located in solution.
Mass spectrometric analysis of the cross-linked peptide
Biochemical properties of aqMutL NTD
We also prepared recombinant aqMutL NTD (amino acid residues 1–315 of aqMutL) (Figure 6A) and investigated its biochemical properties. Unlike other MutL homologues, A. aeolicus MutL has no interdomain-linker region between NTD and CTD . Therefore, the C-terminus of aqMutL NTD used here is the N-terminus of aqMutL CTD. Gel filtration analysis revealed that aqMutL NTD exists as a monomeric form in solution (Figure 6B) like E. coli MutL NTD. The overall structure of aqMutL would resemble that of E. coli MutL and human MutLα (Figure 6C), which are expected to dimerize via its CTD [19,46], although there is a significant difference in the length of their interdomain-linker regions.
Purification and CD measurement of aqMutL NTD
The far-UV CD spectrum of aqMutL NTD showed negative signals from 210 to 220 nm, indicating the presence of α-helices and β-sheets (Figure 6D). The intensity of mean residue ellipticity at 222 nm showed no drastic change even when the solution was treated up to 95°C (Figure 6E), indicating that aqMutL NTD also possesses extreme thermostability like aqMutL CTD. Interestingly, the full-length aqMutL showed reduced thermostability (Figure 6F) compared to NTD and CTD although aqMutL has no interdomain-linker region. It can be speculated that interaction between NTD and CTD influences the biochemical characteristics of each domain.
The DNA-binding activity of aqMutL NTD was examined by electrophoretic mobility-shift assay. As shown in Figure 7, aqMutL NTD showed a DNA-binding activity with 50 times lower affinity than the full-length aqMutL. The dissociation constants of NTD and the full-length aqMutL for a 30-bp double-stranded DNA were 6.5 μM and 150 nM, respectively. Since aqMutL CTD showed only extremely weak DNA-binding activity, which cannot be detected by the electrophoretic method (Figure 3D), the high affinity of the full-length aqMutL would be due to the synergistic effect of aqMutL CTD and NTD.
DNA-binding activity of aqMutL NTD
As previously described, the ATP-binding-induced conformational change of MutL homologues can be detected by limited proteolysis [15,29]. The full-length aqMutL also showed an increased tolerance to tryptic digestion in the presence of ATP (Figure 8A), indicating an ATP-dependent conformational change. While ATP did not have any effect on the result of limited proteolysis of aqMutL CTD (Figure 8B), aqMutL NTD exhibited a drastic increase in trypsin-resistance upon addition of ATP (Figure 8C). ATP was found to be more effective on aqMutL NTD than on the full-length aqMutL. In the subsequent GdnHCl (guanidine hydrochloride)-denaturing experiment, the difference between aqMutL NTD and the full-length aqMutL was also distinct. aqMutL NTD and the full-length aqMutL were incubated with various concentrations of GdnHCl in the presence or absence of ATP, and then the intensity of mean residue ellipticity at 222 nm was measured to estimate the fraction of unfolded protein. As shown in Figure 8(D), ATP enhanced the tolerance of aqMutL NTD against the denaturing reagent but exhibited no effect on that of the full-length aqMutL. These differences in the effect of ATP between aqMutL NTD and the full-length aqMutL may also reflect the interaction between NTD and CTD.
ATP-induced conformational change of NTD and the full-length aqMutL
aqMutL NTD stimulated the endonuclease activity of CTD
In order to investigate the interaction between aqMutL NTD and CTD, we examined whether aqMutL NTD has any effect on the endonuclease activity of CTD. As shown in Figures 9(A)–9(C), aqMutL CTD pre-incubated with NTD exhibited enhanced endonuclease activity. The aqMutL NTD solution was confirmed to have no endonuclease activity (Figure 9A, right edge of the panel); BSA has no enhancing effect on the activity (Figure 9D); and the outer solution of NTD also had no effect (results not shown). These results indicate that aqMutL NTD specifically interacts with CTD and stimulates the endonuclease activity. Pre-incubation of aqMutL CTD with NTD (see Figure legends) was highly effective in revealing the enhancing the effect of NTD. Interestingly, the aqMutL NTD-dependent activation of CTD was not observed when the reaction mixture contained no zinc ion (Figures 9E and 9G). It has been reported that CTD of MutL endonuclease contains two zinc-binding sites: one tightly binds a zinc ion and the other seems to show a relatively low affinity to a zinc ion [20,21]. Although the biological function of zinc ions in MutL CTD is unknown, crystallographic analysis revealed that zinc binding induces a local conformational change in MutL CTD . Inductively coupled plasma atomic emission spectroscopy revealed that only 8% of purified aqMutL CTD contained zinc (results not shown). It is thought that, therefore, the addition of zinc ion exhibited a significant effect on this experiment. It should be mentioned that contamination of agents such as EDTA, salts and denaturants significantly perturbed the results of this experiment. Therefore, substrate plasmid DNA was carefully purified and stored.
The endonuclease activity of aqMutL CTD was stimulated by NTD
Then, we prepared an aqMutL CTD C402A mutant. Cys 402 is a part of the CPHGRP motif in aqMutL (Figure 1A). In B. subtilis MutL, the cysteine residue of the CPHGRP motif is essential to form the stronger one of two zinc-binding sites . The aqMutL CTD C402A mutant exhibited endonuclease activity, with a velocity that is almost same as that of wild-type aqMutL CTD (Figure 9F). However, even in the presence of zinc ion, aqMutL CTD C402A was only slightly stimulated by NTD (Figure 9F), and there was a distinct difference between wild-type and C402A CTD (Figure 9G). These results also strongly suggest that aqMutL NTD stimulates the endonuclease activity of CTD in a zinc-dependent manner.
In the present paper, we have described in detail the biochemical characteristics of aqMutL CTD and NTD. Each domain was extremely thermostable (Figures 2C and 6E), highly soluble, and suitable for physicochemical characterization. The stability of aqMutL CTD may partly come from the compactness of its tertiary structure which lacks the internal (also called as ‘regulatory’) subdomain (Figure 1C). Despite the lack of an internal (regulatory) subdomain, our examination of its quaternary structure and enzymatic properties strongly indicates that aqMutL CTD possesses fundamental features that are common among all MutL CTDs. The overall structure of full-length aqMutL is also expected to be minimized compared to other MutL homologues because aqMutL has no interdomain linker region (Figure 6C). These features may confer advantages in performing structural analyses on the full-length MutL.
We characterized the various biochemical properties of aqMutL CTD and NTD. aqMutL CTD and NTD seemed to contribute to the characteristics of the full-length aqMutL in a synergistic rather than additive manner. The full-length aqMutL showed reduced thermostability compared with aqMutL CTD and NTD. In addition, although aqMutL CTD hardly showed DNA-binding activity, the activity of the full-length aqMutL was 50 times stronger than that of NTD. These results imply interaction between aqMutL CTD and NTD, which affects the property of each domain. The interaction between aqMutL CTD and NTD was directly supported by our finding that the endonuclease activity of aqMutL CTD is stimulated by NTD. Since aqMutL CTD lacks the region corresponding to the internal (regulatory) subdomain, it is thought that the internal (regulatory) subdomain is inessential for the interaction of MutL CTD with NTD. aqMutL NTD is a basic domain (pI=9.6) while CTD is a highly acidic domain (pI=4.4). The negative charges of aqMutL CTD may be responsible for the suppression of non-specific DNA-binding of CTD, and neutralized by interaction with NTD to stabilize the CTD–DNA complex. As pointed out by Pillon et al., strong negative charges are found on the surface of the internal (regulatory) subdomain of MutL CTD, which has been suspected to prevent the DNA-binding activity of CTD. However, aqMutL CTD also hardly bound to DNA (Figure 3D) despite the absence of the internal (regulatory) subdomain (Figure 1A). Therefore, the factor responsible for the prevention of non-specific DNA-binding of MutL CTD would also be located in the external subdomain.
Interestingly, aqMutL NTD-dependent activation of CTD required the addition of zinc ion in the reaction mixture (Figures 9E and 9G). It has been reported that the CTD of MutL endonuclease contains two zinc-binding sites that are composed of DQHAX2EX4E, CPHGRP and SCK motifs [20,21]. The cysteine residue of CPHGRP motif is responsible for binding of one of the two zinc ions [20,21] and critical for in vivo DNA repair activity [15,21]. We prepared an aqMutL CTD C402A mutant in which an alanine residue substitutes for the cysteine residue in the CPHGRP motif, and examined the enhancing effect of aqMutL NTD on the mutant. aqMutL NTD only slightly enhanced the endonuclease activity of CTD C402A mutant (Figures 9F and 9G), indicating that the proper binding of the two zinc ions is essential for the activation of CTD by NTD. This result is in good agreement with our previous finding that substitution of the cysteine of CPHGRP with an alanine results in the perturbation of the ATP-binding-dependent conformational and functional change of the full-length MutL . It should be examined in future studies as to whether zinc ions are directly involved in the interaction between CTD and NTD or indirectly effect the interaction through rearrangement of the local structure.
MutL endonucleases have been reported to undergo ATP-induced conformational and functional changes [15–17,29,30] that are expected to be essential for the regulation of their endonuclease activity. The direct observation by atomic force microscopy has suggested the possible ATP-binding (not hydrolysis)-induced approach of NTD to CTD . It has also been revealed that the physiological concentration of ATP tightly binds to the relatively low concentrations of MutL without any detectable hydrolysing activity , and that ATP-binding suppresses the endonuclease activity of relatively low concentrations of MutL [15,16]. Although observation by atomic force microscopy suggested the ATP-binding-induced approach of MutL NTD to CTD, such approach might be insufficient to stimulate the endonuclease activity. On the other hand, it was clarified that the endonuclease activity of relatively high concentrations of MutL is enhanced by the addition of ATP , and that the non-hydrolysable analogue of ATP has no promoting effect on the endonuclease activity of MutL . These results strongly indicate the requirement of ATP-hydrolysis for the activation of MutL endonuclease activity. ATP-hydrolysis might promote tight contact of MutL NTD with CTD that is enough to stimulate the endonuclease activity. If this is the case, mismatch-specific incision would be assured by controlling the ATPase activity of MutL. Since our previous study suggested that MutL forms a stable complex with MutS in the presence of non-hydrolysable ATP analogue , interaction with MutS may define the timing for ATP-hydrolysis by MutL. We speculate that the mismatch-specific nicking by MutL is achieved by the following mechanisms: (1) ATP-binding induces the approach of MutL NTD to CTD, and the DNA-binding activity of ATP-bound MutL is reduced to prevent the non-specific DNA-binding; (2) The approach of NTD to CTD stabilizes the interaction of MutL with MutS; (3) the interaction with MutS and other MMR proteins allows MutL to hydrolyse ATP; and (4) ATP-hydrolysis induces tight contact between NTD and CTD, which leads to the stimulation of CTD endonuclease activity. It remains to be investigated whether and how ATP-hydrolysis influences the structure and function of MutL endonuclease.
Ryoji Masui, Seiki Kuramitsu and Kenji Fukui designed the experiments. Hitoshi Iino, Kwang Kim and Kenji Fukui performed the experiments and analysed the data. Atsuhiro Shimada performed the experiments. Kenji Fukui and Kwang Kim wrote the paper. All authors discussed the results and commented on the paper.
We thank Dr Akeo Shinkai for his excellent help in the surface plasmon resonance experiment, Dr Yoshitaka Bessho for his valuable advice on this study, Ms Naoko Aoki for nucleotide sequencing analyses in construction of expression plasmids, and Dr Hisashi Naitow and Tomoyuki Tanaka for their excellent help in the DLS experiment.
This work was partly supported by Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science, and Technology of Japan [grant numbers 20570131 (to R.M.) and 20870042 (to K.F.)].