Minichromosome maintenance helicase activityis controlled by N- and C-terminal motifsand requires the ATPase domain helix-2 insertElizabeth R. Jenkinson and James P. J. Chong*

Department of Biology, University of York, P.O. Box 373, York YO10 5YW, United Kingdom

Edited by Bruce W. Stillman, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, and approved March 23, 2006 (received for review October 25, 2005)

The minichromosome maintenance (MCM) proteins are essentialconserved proteins required for DNA replication in archaea andeukaryotes. MCM proteins are believed to provide the replicativehelicase activity that unwinds template DNA ahead of the replica-tion fork. Consistent with this hypothesis, MCM proteins can formhexameric complexes that possess ATP-dependent DNA unwindingactivity. The molecular mechanism by which the energy of ATPhydrolysis is harnessed to DNA unwinding is unknown, althoughthe ATPase activity has been attributed to a highly conservedAAA� family ATPase domain. Here we show that changes to N-and C-terminal motifs in the single MCM protein from the archaeonMethanothermobacter thermautotrophicus (MthMCM) can modu-late ATP hydrolysis, DNA binding, and duplex unwinding. Further-more, these motifs appear to influence the movement of the �-�-�insert in helix-2 of the MCM ATPase domain. Removal of this motiffrom MthMCM increased dsDNA-stimulated ATP hydrolysis andincreased the affinity of the mutant complex for ssDNA and dsDNA.Deletion of the helix-2 insert additionally resulted in the abroga-tion of DNA unwinding. Our results provide significant insight intothe molecular mechanisms used by the MCM helicase to bothregulate and execute DNA unwinding.

archaea � DNA helicase

The minichromosome maintenance (MCM) proteins are con-served throughout eukaryotes and are essential for DNA

replication initiation (1, 2) and progression (2). All eukaryoticorganisms possess six highly related MCM proteins (MCM2–7) (3)that form high-molecular-weight complexes consistent with hetero-hexamers. Although homohexameric DNA helicases are relativelycommon and include the well studied SV40 large T antigen,papillomavirus E1, and phage T4 gene 4 protein, the MCMs areunique in forming a heterohexameric helicase. The evolutionaryadvantages presented by the additional complexity of a heterohex-americ complex are not clear but may include specialization ofindividual components to allow interaction with a range of proteinpartners, allowing the replicative helicase to respond to a variety ofcellular conditions. For example, recent evidence suggests that, inaddition to DNA replication, MCMs play a role in transcriptionalactivation (4) and DNA damage checkpoint progression (5, 6).

The primary sequence of MCM proteins can be divided intothree domains (7, 8): an N-terminal domain possessing a zinc-fingermotif, a central AAA� ATPase domain, and a C-terminal domainof unknown function (Fig. 5a). Consistent with this, eukaryoticMCM complexes have been shown to possess ATPase activity(9–12). In addition to the Walker A and B motifs of the ATPasedomain, an arginine finger is required for ATP hydrolysis (9).

Although much evidence suggests that heterohexameric com-plexes of all six MCM proteins form (9, 11, 13, 14) and all six MCMproteins are required for DNA replication progression (2), to datethere is no biochemical evidence of such a complex possessingprocessive helicase activity. An MCM 4,6,7 subcomplex possessesweak helicase activity in vitro (10, 11) that is improved in thepresence of forked DNA substrates and ssDNA-binding protein(15). It is possible that in vivo MCM complex processivity is further

improved by posttranslational modifications (e.g., ref. 16) or thepresence of additional processivity factors.

The complexity of the eukaryotic MCM system means thataddressing the molecular mechanisms required for unwinding hasbeen difficult. The archaea also possess MCM proteins but in asimplified form. Almost all archaeal species possess a single MCMthat forms homohexameric complexes. Archaeal MCMs are usefulmodels for understanding some of the molecular mechanismsunderlying the eukaryotic heterohexameric MCM complex (17).Biochemical studies of the Methanothermobacter thermautotrophi-cus MCM (MthMCM) demonstrated that this protein formsdouble-hexamer complexes that bind ssDNA and dsDNA. Mth-MCM also possesses dsDNA-stimulated ATPase and ATP-dependent DNA unwinding activities (18–20). A crystal structureof the N-terminal domain of MthMCM showed the subunits forma dodecameric ring-shaped complex with a central channel largeenough to accommodate dsDNA (21). Hexameric, heptameric, anddouble hexameric complexes have been observed by 3D recon-struction of the full-length protein from electron microscope im-ages (8, 22, 23).

MCMs possess two �-hairpins, one located in the N-terminaldomain, which has a role in DNA binding (21, 24), and thepresensor 1 �-hairpin in the ATPase domain (PS1BH), which isrequired for DNA duplex unwinding (24). The PS1BH is charac-teristic of superfamily 3 helicases and has been observed to moveup to 17 Å during ATP binding and hydrolysis (25). This ATPhydrolysis-coupled movement is believed to cause translocation ofDNA relative to the helicase complex; a point mutation of thePS1BH that would prevent DNA binding caused the abrogation ofhelicase activity in the Sulfolobus MCM (24).

Here we investigate the role of N- and C-terminal residues onMCM activity. We show that amino acids at both ends of the proteinplay roles in the modulation of ATP hydrolysis, DNA binding, andprocessivity of DNA unwinding. Changes to N- and C-terminalmotifs can be related to hydrophobicity changes consistent withconformational changes in the ATPase domain of the protein.Finally, we demonstrate the key role of an ATPase domain �-�-�insert in transducing the energy of ATP hydrolysis into DNAunwinding. Among the superfamily 3 helicases, this motif is founduniquely in the MCMs. Together, our results suggest that ATPhydrolysis is optimized for DNA unwinding in the MCMs, and thathelicase activity is carefully controlled.

ResultsGeneration of N- and C-Terminal Mutations. A series of N-terminalMthMCM deletion mutants were generated by site-directed mu-

Conflict of interest statement: No conflicts declared.

This paper was submitted directly (Track II) to the PNAS office.

Abbreviations: MCM, minichromosome maintenance; MthMCM, Methanothermobacterthermautotrophicus MCM; h-2i, helix-2 insert; PS1BH, presensor 1 �-hairpin in the ATPasedomain.

*To whom correspondence should be addressed. E-mail: jpjc1@york.ac.uk.

© 2006 by The National Academy of Sciences of the USA

www.pnas.org�cgi�doi�10.1073�pnas.0509297103 PNAS � May 16, 2006 � vol. 103 � no. 20 � 7613–7618

BIO

CHEM

ISTR

Y

Dow

nloa

ded

by g

uest

on

Janu

ary

10, 2

021

tagenesis, purified, and tested for dsDNA-stimulated ATPase ac-tivity (not shown). We observed that the MthMCM complex lostATPase activity between residues 89 and 105. An alignment ofMCM sequences from a range of species identified a hydrophobicpatch containing a single conserved arginine (R98) that we mutatedto alanine to produce MthMCM R98A. Limited proteolysis of thefull-length protein allowed us to identify a cleavable C-terminaldomain. We recreated this truncated protein (�597) by the intro-duction of a stop codon into the expression vector. The R98A and�597 mutations were combined to produce a third mutant con-struct designated RA�C. WT MthMCM, R98A, �597, and RA�Cproteins were expressed and purified. The mutant proteins allformed stable high-molecular-weight complexes consistent withdouble hexamers, suggesting that neither the R98 nor the C-terminal 72 amino acids play any significant role in the assembly ormaintenance of multimeric MthMCM complexes.

ATPase Activity of MthMCM Mutants. The WT and mutant proteinswere subjected to ATPase assays in the presence of single- ordouble-stranded closed circular DNA (Fig. 1). As previously re-ported, the WT protein showed dsDNA-stimulated ATPase activ-ity. Compared with WT, in the presence of dsDNA, the R98Amutant showed reduced levels of ATPase activity, whereas the�597 mutant showed an �3-fold increase in ATP hydrolysis. In thepresence of dsDNA, RA�C showed low levels of ATP hydrolysisthat were above background. ATPase activity of RA�C in thepresence of ssDNA was identical to background in the absence ofDNA (Fig. 1d Inset). We conclude that residues in both the N- andC-terminal domains of MthMCM can influence dsDNA-stimulatedATPase activity.

DNA-Binding Activity of MthMCM Mutants. We tested the DNA-binding activity of the MCM mutants using an EMSA. Both single-and double-stranded short linear DNA templates were tested in thepresence and absence of ATP. Table 1 lists dissociation constants(Kd) for substrate saturation by WT and mutant proteins deter-mined from the data in Fig. 7, which is published as supportinginformation on the PNAS web site. The MthMCM complex shows

a preference for binding ssDNA rather than dsDNA. Consistentwith previous results (18), DNA binding by the WT MthMCM wasreduced in the presence of hydrolyzable ATP. All of the mutantsshowed approximately WT levels of dsDNA binding in the presenceof ATP (Fig. 7). The R98A protein showed reduced affinity forssDNA, whereas the �597 protein showed a higher than WTaffinity for ssDNA and dsDNA under these conditions. Our resultsclearly indicate that the presence of hydrolyzable ATP reduces theaffinity of MthMCM for DNA substrates. Three obvious mecha-nisms could explain these observations: (i) the presence of ATPcould cause translocation of the DNA through the MCM complexfollowed by substrate release, (ii) ATP and DNA could compete fora particular binding site, or (iii) ATP may cause a conformationalchange that reduces binding of DNA.

Helicase Activity of MthMCM Mutants. DNA unwinding was mea-sured by strand displacement by using a 3� tailed DNA substrate(Fig. 2). We observed that, in all cases, very high MCM to DNAratios inhibited strand displacement. Surprisingly, despite theobserved differences in ATPase activity and ssDNA affinity, theR98A and �597 proteins (Fig. 2 b and c) showed WT levels ofhelicase activity. Even more strikingly, despite having very lowATPase activity compared with WT, the combined RA�CMthMCM mutant (Fig. 2d) was able to unwind a larger proportionof the helicase substrate at lower protein concentrations than either

Fig. 1. ATPase activity is enhanced in �597 but reduced in other MthMCMmutants. Quantification of ATP hydrolysis (calibrated from a standard curveproduced by substrate dilution) for WT (a), R98A (b), �597 (c), and RA�C (d) inthe presence of dsDNA (closed circles, solid lines) or ssDNA (open circles,dotted lines). ATP hydrolysis in the absence of DNA is shown for RA�C (d Inset,open squares).

Table 1. Kd values for short DNA substrate saturation byMthMCM proteins (nM 12-mer)

ssDNA dsDNA

�ATP �ATP �ATP �ATP

WT 60 � 21 65 � 8 ND* NDR98A �100 66 � 30 428 � 148 ND�597 8 � 2 �225 34 � 8 NDRA�C �100 �225 58 � 16 ND

*ND Kd in all cases is �1 �M 12-mer.

Fig. 2. DNA unwinding activity is enhanced only in the RA�C MthMCMmutant. Quantification of the amount of radiolabeled oligonucleotide dis-placed from a duplex possessing a 3� tail by WT (a), R98A (b), �597 (c), andRA�C (d) MthMCM proteins. Protein amounts are fmol of 12-mer complex ina 20-�l reaction.

7614 � www.pnas.org�cgi�doi�10.1073�pnas.0509297103 Jenkinson and Chong

Dow

nloa

ded

by g

uest

on

Janu

ary

10, 2

021

WT or the individual mutant proteins. To further address this point,we performed a time course experiment using a long substrate toexamine the rate of duplex unwinding and the processivity of thedifferent mutants. Fig. 3 shows that R98A is slightly more processivethan WT over 30 min (compare lanes 20 and 8, respectively),whereas �597 is significantly more processive than WT. RA�Cdisplays the best processivity and has a faster rate of unwinding than�597 (compare 5-min time point, Fig. 3, lanes 10 and 23, respec-tively). In the context of our other results, this suggests that DNAunwinding by the MCM complex does not require high levels ofATP hydrolysis to be processive. Thus the N- and C-terminal motifsmay act to modulate the efficiency with which ATP hydrolysis isutilized in DNA unwinding.

DNA- and Nucleotide-Induced Changes in Fluorescence. The MthMCMprotein contains only a single tryptophan residue (W363),

located between the Walker A and B motifs of the centralATPase domain of the protein (Fig. 5a). To determine whetherthis region of the protein was influenced by the R98A and �597mutations, we performed steady-state tryptophan fluorescencemeasurements. WT and mutant MCM proteins were incubatedwith nucleotide analogues or DNA, and the resulting signalswere compared with those obtained in the absence of substrate(Fig. 4). Fluorescence changes consistent with the tryptophanbeing exposed to a more aqueous environment were observed inthe WT protein on the addition of ATP, ADP, or nonhydrolyz-able adenylyl-imidodiphosphate (AMP-PNP) (Fig. 4a). Thesechanges were also observed in the R98A protein, although thechange in signal between different analogues was less thanobserved for WT. The fluorescence changes were almost com-pletely lost in the �597 protein (Fig. 4a). Combining bothmutations produced an intermediate phenotype. The patterns ofchanges on addition of ssDNA or dsDNA (Fig. 4 b and c) weresimilar to those observed on nucleotide addition, although in allcases, addition of DNA resulted in larger fluorescence changesthan the addition of nucleotides. In all cases, the �597 proteindeviated most from WT and displayed changes consistent with

Fig. 3. Time course of helicase activity shows increased processivity ofunwinding in �597 and RA�C compared with WT. WT (lanes 5–8), R98A (lanes17–20), �597 (lanes 22–25), or RA�C (lanes 9–12) MCM complex (20 nM12-mer) were incubated with 8 nM DNA substrate in the presence of 4 mM ATPfor 1 min (lanes 5, 9, 17, and 22), 5 min (lanes 6, 10, 18, and 23), 10 min (lanes7, 11, 19, and 24), or 30 min (lanes 8, 12, 20, and 25). Controls were labeled24-residue oligonucleotide (lanes 1 and 14), sizing ladder (lanes 2 and 13), DNAsubstrate incubated on ice (lanes 3 and 15), boiled substrate (lanes 4 and 16),and mock-treated substrate (lane 21).

Fig. 4. Movement of W363 in MthMCM. Relative changes in fluorescencemeasured on addition of nucleotide analogues (a), ssDNA (b), or dsDNA (c) to0.3 �M protein are shown with an indication of whether the change isconsistent with the environment of W363 becoming more or less hydrophobic.Bars are the average values from the measurement of at least three indepen-dent samples and are relative to the fluorescence observed in the absence ofsubstrate. PNP, purine nucleoside phosphorylase.

Jenkinson and Chong PNAS � May 16, 2006 � vol. 103 � no. 20 � 7615

BIO

CHEM

ISTR

Y

Dow

nloa

ded

by g

uest

on

Janu

ary

10, 2

021

the least movement in the locale of W363. These effects werelargely overcome in the RA�C double mutant. This suggests thatthe C-terminal domain of MCM exerts a strong conformationalinfluence over the ATPase domain of the protein. Moreover, theenvironment around W363 is additionally modulated by theN-terminal domain of the protein, with the R98 residue alsohaving a significant influence on the conformation of the centralportion of the protein in a nucleotide- and DNA-dependentmanner.

Molecular Modeling of the Helix-2 Insert (h-2i). We precisely locatedW363 in the MCM ATPase domain so we could investigate thespecific role of the N- and C-terminal motifs on activity. Bioinfor-matic analysis of the AAA� ATPase family suggests that MCMscan be differentiated from Tag and other superfamily 3 helicases bya h-2i between the Walker A and B motifs (26). This feature is notconserved at the level of amino acid sequence, but a �-�-�structural motif is predicted that would include W363 in MthMCM(Fig. 5 a and b). The crystal structure of the AAA� domain ofmagnesium chelatase, another h-2i clade protein, revealed that theh-2i forms a �-�-� loop positioned near the PS1BH (27). Wegenerated a model of the N-terminal and ATPase domains of

MCM from the magnesium chelatase ATPase structure and theN-terminal structure of MthMCM using the position of thesedomains proposed by Pape et al. (8) (Fig. 5 c and d). The modelallowed us to visualize the putative location of the h-2i. It suggestedthat the h-2i was likely to interact with the PS1BH and protrude intothe central channel of the MCM complex, potentially causing anobstruction to ‘‘loading’’ of substrate DNA. Thus our fluorescencemeasurements are likely reflecting the movement of the h-2i inresponse to nucleotide and DNA substrates.

Characterization of the MthMCM �h-2i Mutant. To determinewhether the h-2i plays a role in MCM function, we deleted the 15amino acids of the h2-i in MthMCM. The �h2-i protein was solubleand formed high-molecular-weight complexes comparable to thosepreviously described for MthMCM and consistent with a functionaldodecamer (data not shown). ATPase activity of the mutantprotein was compared with WT (Fig. 6a). In the presence of ssDNAand 50 �M ATP, the �h-2i protein was comparable to WT, showingno stimulation of ATPase activity (1.09 � 0.08 pmol ADP per mincompared with 1.09 � 0.18 pmol ADP per min for WT). However,in the presence of 50 �M ATP and dsDNA, �h-2i showed a�12-fold increase in dsDNA-stimulated ATPase activity (41.48 �

Fig. 5. Identification and modeling of the h-2i in MthMCM. (a) Domain structure of MCMs, showing zinc-finger containing N-terminal domain, central AAA�ATPase domain containing Walker A and B motifs, and h-2i. (b) Excerpt from CLUSTALX alignment of the MCM ATPase domain showing the Walker A and B motifs(orange), the conserved presensor 1 �-hairpin lysine (blue), and the �-�-� h-2i (magenta). MCM sequences are Mth, M. thermautotrophicus; Sso, Sulfolobussolfataricus; Sc, Saccharomyces cerevisiae; and Hs, Homo sapiens. Consensus residues are shown in gray and are based on 100% agreement with the Betts andRussell scheme (33): *, identical; �, favorable substitution; �, neutral substitution. Side (c) and top (d) views of a molecular model of the N-terminal and ATPasedomains of MthMCM. Walker motifs (orange) and h-2i (magenta) are highlighted in the magnesium chelatase structure (green). N-terminal crystal structure ofMCM (magenta). Electron density mesh indicates overall complex conformation.

7616 � www.pnas.org�cgi�doi�10.1073�pnas.0509297103 Jenkinson and Chong

Dow

nloa

ded

by g

uest

on

Janu

ary

10, 2

021

15.87 pmol ADP per min vs. 3.52 � 0.57 pmol ADP per min forWT), confirming that the protein was functional and suggesting thatthe h-2i normally inhibits ATP hydrolysis.

DNA binding of �h-2i was characterized by using EMSA. Themutant protein bound to both ssDNA and dsDNA substrates (Fig.6 b and c) with much greater affinity than WT (Tables 1 and 2). Incontrast to WT, where the presence of ATP caused a considerableincrease in Kd, the presence of ATP actually reduced the Kd of �h-2ifor both DNA substrates. Using a strand displacement assay, we

determined whether �h-2i possessed DNA helicase activity. Asreported (18–20), the WT protein displayed magnesium- andATP-dependent DNA helicase activity in a concentration-dependent manner (Fig. 6d). In contrast, we were unable to detectany helicase activity for �h-2i (Fig. 6e) despite its increased ATPaseactivity. Consistent with the EMSA results (Fig. 6 b and c), weobserved a protein-dependent mobility shift of the helicase sub-strate when the reactions were incubated on ice. The proportion ofDNA substrate shifted was greater in the presence of the �h-2imutant than in the presence of WT protein.

Our results strongly suggest that the MCM helix-2 �-�-� insert isessential for coupling the energy of ATP hydrolysis to DNAunwinding. In contrast to the Sulfolobus MCM PS1BH mutant (24),deletion of the h-2i from MthMCM caused a large increase indsDNA-stimulated ATP hydrolysis and increased the affinity of themutant complex for both ssDNA and dsDNA, even in the presenceof ATP. The observed increase in ATP hydrolysis on deletion of theh-2i suggests that the MCM complex is capable of much higherlevels of ATP hydrolysis than are normally observed for the WTprotein in vitro.

DiscussionWe have identified an N-terminal residue, R98, and a 72-aaC-terminal domain in MthMCM that affect both ATP hydrolysisand DNA binding. The mutations we have generated in theseregions could reflect changes brought about in the MCM complexin vivo by protein–protein interactions or posttranslational modifi-cations. Removal of both these influences resulted in an increasedrate and processivity of DNA unwinding activity. Addition ofsubstrates to the protein caused changes in the environment of theh-2i located between the Walker A and B motifs of the ATPasedomain, consistent with conformational changes. The h-2i locale isadditionally modulated by the N- and C-terminal motifs identified;these residues appear to have antagonistic effects on the h-2i in thepresence of DNA or nucleotide. We predict that the h-2i protrudesinto the central channel of the MCM complex alongside thePS1BH. The poor sequence conservation found in the h-2i relativeto the surrounding AAA� motifs suggests that whatever role thismotif plays in MCM activity, it uses a mechanical mechanism thatdoes not require specific amino acid interactions.

Removal of the h-2i from MthMCM resulted in the abrogationof DNA unwinding despite producing enhanced ATP hydrolysis.Thus the h-2i has a key role in the transduction of energy releasedby ATP hydrolysis into DNA unwinding activity. It should beemphasized that among helicases, the h-2i is unique to the MCMproteins: this motif is not present in other helicases such as Tag, andso Tag and the many other DNA helicases that fall outside the h-2iclade are likely to use a different mechanism for coupling ATPhydrolysis and unwinding. Our fluorescence results are consistentwith movement of the h-2i into a more aqueous environment in thepresence of substrates, suggesting this motif may move into thecentral channel of the MCM complex. In this case, the h-2i wouldbe well positioned to mechanically separate the antiparallel strandsof duplex DNA and act as the recently proposed molecular plough-share (28). A scenario in which the PS1BH binds to the DNAsubstrate, followed by ATP-catalyzed opposing movement of theh-2i and the PS1BH in the central channel, is one mechanism bywhich the efficiency of duplex unwinding could be enhanced.

Fig. 6. Characterization of the MthMCM �h-2i protein. (a) ATPase activity ofWT or �h2-i MthMCM in the absence of DNA or in the presence of 90 fmolclosed circular ssDNA or 2.88 pmol dsDNA in 20-�l reactions. Values arehydrolysis rates for an input concentration of 50 �M ATP. (b) EMSA quantifi-cation (relative to input substrate) of ssDNA or (c) dsDNA-binding activity inthe presence (open circles, dotted lines) or absence (closed circles, solid lines)of 4 mM ATP. Visualization of helicase activity for WT (d) or �h2-i (e) MthMCMprotein using strand displacement. Activity is indicated by the displacement ofa labeled oligonucleotide from the substrate, resulting in a faster-migratingband. Boiled substrate controls indicate the position of the displaced oligo-nucleotide. The supershift observed in the samples incubated on ice is consis-tent with EMSA data. Protein amounts are fmol of a 12-mer complex in a 20-�lreaction.

Table 2. Kd values for short DNA substrate saturation byMthMCM �h-2i protein (nM 12-mer)

ssDNA dsDNA

�ATP �ATP �ATP �ATP

�h-2i 1.6 � 0.2 0.7 � 0.1 8.4 � 3.4 3.7 � 1.0

Jenkinson and Chong PNAS � May 16, 2006 � vol. 103 � no. 20 � 7617

BIO

CHEM

ISTR

Y

Dow

nloa

ded

by g

uest

on

Janu

ary

10, 2

021

An enhanced ability to unwind DNA is potentially very impor-tant for the MCM helicase. The RA�C mutant shows increasedprocessivity of unwinding with decreased ATP hydrolysis, suggest-ing N-terminal components modulate the efficiency of transductionof ATP hydrolysis into duplex unwinding. Consistent with thishypothesis, increased processivity of the Archaeoglobus MCM wasobserved on deletion of 112 amino acids from the N terminus (29).Inhibition of the MthMCM complex by N-terminal interactionswith CDC6 have also been recently described (30). The negativeinfluence of both the C- and N-terminal domains on helicaseactivity may explain why eukaryotic complexes have not yet beenshown to possess efficient helicase activity in vitro, a point thatwould benefit from further investigation in the light of these results.

Materials and MethodsMutant Construction. All mutants were generated by using theStratagene Quik-Change site-directed mutagenesis (SDM) proto-col. Oligonucleotide (Sigma Genosys) sequences are available onrequest. MthMCM amino acid positions were numbered from thesecond methionine in the sequence. pJC025 (18) was used as atemplate to produce R98A (pJC059), �597 (pLJ040), and �h2-i(pLJ042). The double mutant, RA�C (pLJ041), was produced bySDM of pJC059.

Protein Expression and Purification. All constructs were expressed asdescribed (18) and purified over TALON metal affinity resin(Clontech), followed by 50–500 mM NaCl elution from a ResourceQ column (GE Biosciences). Purified proteins were desalted andspin-concentrated (30-kDa molecular weight cutoff, Vivaspin, Sar-torius). Protein concentration was measured, and small aliquotswere snap-frozen before being stored at �80°C.

ATP Hydrolysis Assay. Assays were performed as described (18).ATP was mixed with [�-32P]-ATP [3,000 Ci�mmol (1 Ci � 37 GBq),GE Biosciences] and titrated into 20-�l reactions containing 4.32pmol monomeric protein. Reactions contained no DNA or weresupplemented with either 2.88 pmol closed circular dsDNA or 90fmol closed circular ssDNA. The data were calibrated by using adilution series of the ATP stock performed in parallel with theexperiment. Signals were detected by phosphorimaging and quan-tified by using Quantity One (Bio-Rad).

EMSA. Experiments were performed as described (18) by using 2.5nM 5� radiolabeled 30-residue oligonucleotide as the ssDNA sub-strate. The dsDNA substrate was made by annealing the labeledoligonucleotide to a second oligonucleotide with complementarysequence and added to 2.5 nM. ATP was added to a final concen-tration of 4 mM as appropriate. Dissociation constants (Kd) were

determined where saturation was reached by fitting a single rect-angular 2 parameter hyperbola (Eq. 1) to the data using SIGMAPLOT9 (Systat, Point Richmond, CA). In all cases, R2 � 0.99.

y � [MCM]�{Kd � [MCM]} [1]

Helicase Assays. Strand-displacement assays were performed asdescribed (18) with the following modifications: the substrate wasmade by labeling a 26-residue oligonucleotide with [�-32P]-ATP(GE Biosciences) using T4 polynucleotide kinase (New EnglandBiolabs). The labeled oligonucleotide was desalted and annealed toa 50-residue oligonucleotide with sequence complementary to thelabeled oligonucleotide. This resulted in a substrate of 26 base pairsand a 24-nt 3� tail. Twenty-microliter reactions containing 4 mMATP and 20 fmol substrate were incubated at 50°C for 30 minbefore being stopped and separated on a 15% polyacrylamide(19:1) gel in 1 TBE (90 mM Tris�90 mM boric acid�2 mM EDTA,pH 8.0). Processivity assays were performed by using a substratebased on the work of Tabor and Richardson (18, 31) and aredescribed in detail elsewhere (S. Castella, D. Burgin, and C. M.Sanders, personal communication).

Fluorescence Measurements. Control and sample reactions contain-ing 0.5 mM nucleotide or different concentrations of 30-residueoligonucleotide were incubated with 0.3 �M protein for 5 min atroom temperature before being excited at 295 nm in a FluoroMax-2fluorimeter (Jobin Yvon, Longjumeau, France). Emission spectra(310–400 nm) were collected and the data averaged from at leastfour technical and three biological replicates. Values at 340 nm werecompared. Scans of blank samples showed no inner filter effects onaddition of nucleotide or DNA alone.

Molecular Modeling. The Protein Data Bank (PDB) file for mag-nesium chelatase (PDB ID code 1G8P) was edited to removeregions that were not homologous to MthMCM before being fittedto the electron density map of MthMCM (8). The N terminus ofMthMCM (1LTL) was docked in a similar way, and the resultingmodel was visualized by using MACPYMOL, Ver. 0.99 (DeLanoScientific, South San Francisco, CA). The h-2i was identified byusing CLUSTALX (32) alignments of the amino acid sequences of theappropriate regions of magnesium chelatase and MthMCM (26).

We thank S. Onesti (Imperial College, London) for providing theelectron density map of MthMCM and C. Sanders, S. Castella, A. Leech,and M. Polonska for technical support. We are grateful to S. Bell forcommunicating work before publication and to J. Blow, C. Sanders, andM. van der Woude for comments on the manuscript. This work wassupported by a Biotechnology and Biological Sciences Research CouncilDavid Phillips Research Fellowship and University of York Innovationand Research Priming Funds (to J.P.J.C.).

1. Chong, J. P., Mahbubani, H. M., Khoo, C. Y. & Blow, J. J. (1995) Nature 375, 418–421.2. Labib, K., Tercero, J. A. & Diffley, J. F. (2000) Science 288, 1643–1647.3. Chong, J. P., Thommes, P. & Blow, J. J. (1996) Trends Biochem. Sci. 21, 102–106.4. Snyder, M., He, W. & Zhang, J. J. (2005) Proc. Natl. Acad. Sci. USA 102,

14539–14544.5. Cortez, D., Glick, G. & Elledge, S. J. (2004) Proc. Natl. Acad. Sci. USA 101,

10078–10083.6. Tsvetkov, L. & Stern, D. F. (2005) J. Biol. Chem. 280, 11943–11947.7. Chong, J. P. (2005) Nat. Struct. Mol. Biol. 12, 734–736.8. Pape, T., Meka, H., Chen, S., Vicentini, G., van Heel, M. & Onesti, S. (2003) EMBO

Rep. 4, 1079–1083.9. Davey, M. J., Indiani, C. & O’Donnell, M. (2003) J. Biol. Chem. 278, 4491–4499.

10. Ishimi, Y. (1997) J. Biol. Chem. 272, 24508–24513.11. Lee, J. K. & Hurwitz, J. (2000) J. Biol. Chem. 275, 18871–18878.12. Schwacha, A. & Bell, S. P. (2001) Mol. Cell 8, 1093–1104.13. Crevel, G., Ivetic, A., Ohno, K., Yamaguchi, M. & Cotterill, S. (2001) Nucleic Acids

Res. 29, 4834–4842.14. Prokhorova, T. A. & Blow, J. J. (2000) J. Biol. Chem. 275, 2491–2498.15. Lee, J. K. & Hurwitz, J. (2001) Proc. Natl. Acad. Sci. USA 98, 54–59.16. Weinreich, M. & Stillman, B. (1999) EMBO J. 18, 5334–5346.17. Tye, B. K. (2000) Proc. Natl. Acad. Sci. USA 97, 2399–2401.18. Chong, J. P., Hayashi, M. K., Simon, M. N., Xu, R. M. & Stillman, B. (2000) Proc.

Natl. Acad. Sci. USA 97, 1530–1535.19. Kelman, Z., Lee, J. K. & Hurwitz, J. (1999) Proc. Natl. Acad. Sci. USA 96, 14783–14788.

20. Shechter, D. F., Ying, C. Y. & Gautier, J. (2000) J. Biol. Chem. 275, 15049–15059.21. Fletcher, R. J., Bishop, B. E., Leon, R. P., Sclafani, R. A., Ogata, C. M. & Chen, X. S.

(2003) Nat. Struct. Biol. 10, 160–167.22. Gomez-Llorente, Y., Fletcher, R. J., Chen, X. S., Carazo, J. M. & San Martin, C.

(2005) J. Biol. Chem. 280, 40909–40915.23. Yu, X., VanLoock, M. S., Poplawski, A., Kelman, Z., Xiang, T., Tye, B. K. & Egelman,

E. H. (2002) EMBO Rep. 3, 792–797.24. McGeoch, A. T., Trakselis, M. A., Laskey, R. A. & Bell, S. D. (2005) Nat. Struct. Mol.

Biol. 12, 756–762.25. Gai, D., Zhao, R., Li, D., Finkielstein, C. V. & Chen, X. S. (2004) Cell 119, 47–60.26. Iyer, L. M., Leipe, D. D., Koonin, E. V. & Aravind, L. (2004) J. Struct. Biol. 146, 11–31.27. Fodje, M. N., Hansson, A., Hansson, M., Olsen, J. G., Gough, S., Willows, R. D. &

Al-Karadaghi, S. (2001) J. Mol. Biol. 311, 111–122.28. Takahashi, T. S., Wigley, D. B. & Walter, J. C. (2005) Trends Biochem. Sci. 30,

437–444.29. Grainge, I., Scaife, S. & Wigley, D. B. (2003) Nucleic Acids Res. 31, 4888–4898.30. Kasiviswanathan, R., Shin, J. H. & Kelman, Z. (2005) Nucleic Acids Res. 33,

4940–4950.31. Tabor, S. & Richardson, C. C. (1989) Proc. Natl. Acad. Sci. USA 86, 4076–4080.32. Thompson, J. D., Gibson, T. J., Plewniak, F., Jeanmougin, F. & Higgins, D. G. (1997)

Nucleic Acids Res. 25, 4876–4882.33. Betts, M. J. & Russell, R. B. (2003) in Bioinformatics for Geneticists, eds. Barnes,

M. R. & Grey, I. C. (Wiley, Chichester, U.K.), pp. 289–316.

7618 � www.pnas.org�cgi�doi�10.1073�pnas.0509297103 Jenkinson and Chong

Dow

nloa

ded

by g

uest

on

Janu

ary

10, 2

021