a single subunit directs the assembly of the escherichia coli dna sliding clamp loader

Structure

Article

A Single Subunit Directs the Assembly of theEscherichia coli DNA Sliding Clamp LoaderAh Young Park,1,2 Slobodan Jergic,3 Argyris Politis,1,2 Brandon T. Ruotolo,2,4 Daniel Hirshberg,2 Linda L. Jessop,3

Jennifer L. Beck,3 Daniel Barsky,2,5 Mike O’Donnell,6 Nicholas E. Dixon,3 and Carol V. Robinson1,2,*1Department of Chemistry, University of Oxford, South Parks Road, Oxford OX1 3QZ, UK2Department of Chemistry, University of Cambridge, Lensfield Road, Cambridge CB2 1EW, UK3School of Chemistry, University of Wollongong, Wollongong NSW 2522, Australia4Department of Chemistry, University of Michigan, 930 North University Drive, Ann Arbor, MI 48109, USA5Physical and Life Sciences Directorate, Lawrence Livermore National Laboratory, L-372, Livermore, CA 94550, USA6Rockefeller University, Howard Hughes Medical Institute, 1230 York Avenue, New York, NY 10065, USA*Correspondence: [email protected]

DOI 10.1016/j.str.2010.01.009

SUMMARY

Multi-protein clamp loader complexes are requiredto load sliding clamps onto DNA. In Escherichia colithe clamp loader contains three DnaX (t/g) proteins,d, and d0, which together form an asymmetric pen-tameric ring that also interacts with cc. Here weused mass spectrometry to examine the assemblyand dynamics of the clamp loader complex. We findthat g exists exclusively as a stable homotetramer,while t is in a monomer-dimer-trimer-tetramer equi-librium. d0 plays a direct role in the assembly as at/g oligomer breaker, thereby facilitating incorpora-tion of lower oligomers. With d0, both d and cc stabi-lize the trimeric form of DnaX, thus completing theassembly. When t and g are present simultaneously,mimicking the situation in vivo, subunit exchangebetween t and g tetramers occurs rapidly to formheterocomplexes but is retarded when dd0 is present.The implications for intracellular assembly of theDNA polymerase III holoenzyme are discussed.

INTRODUCTION

Chromosomal replication in Escherichia coli is carried out by the

DNA polymerase III (Pol III) holoenzyme, a complex containing

ten different subunits. Three subcomplexes, the replicase core

(the a polymerase subunit, the 3 proofreader, and the accessory

factor q), the sliding clamp (dimeric b), and the seven-subunit

clamp loader complex (comprised of subunits t2gdd0cc or

t3dd0cc) act together in the holoenzyme to deliver highly proces-

sive, accurate, and rapid DNA replication (Bloom, 2009; Johnson

and O’Donnell, 2005; Kornberg and Baker, 1992; Schaeffer et al.,

2005). The clamp loader complex is responsible for loading/

unloading of the b2 sliding clamp (Naktinis et al., 1995), which

encircles the primed template DNA strand and maintains proc-

essivity by tethering the replicase core to the template (Jeruzalmi

et al., 2001, 2002). The clamp loader comprises a central t2gdd0

or t3dd0 pentameric ring that provides the site of ATPase activity,

Structure 18, 2

attached to a cc heterodimer through interaction of c with the

three t or g subunits (Dallmann and McHenry, 1995; Gulbis

et al., 2004; Xiao et al., 1993a, 1993b). The c subunit mediates

a further interaction between the clamp loader and the single-

stranded DNA-binding protein SSB (Yuzhakov et al., 1999).

Unusually, t and g are produced from the same gene, dnaX; g

is a truncated form of t, produced by a programmed translational

frameshift (Tsuchihashi and Kornberg, 1990) such that the two

are identical in their first 431 residues (Domains I–III). All combi-

nations of (t/g)3 in the clamp loader complex have been shown

to form a functioning clamp loader, i.e., t3dd0cc, t2gdd0cc,

tg2dd0cc, and g3dd0cc (Blinkowa and Walker, 1990; Flower and

McHenry, 1990; McHenry, 2003). Although the minimal unit func-

tional in clamp loading is the g complex g3dd0 (O’Donnell and

Kuriyan, 2006), coupled leading and lagging strand DNA replica-

tion requires the presence of at least two t subunits that dimerize

the replicase core through the unique C-terminal domains of t.

These domains, not present in g, interact with both the a subunit

of the core and the hexameric DnaB helicase (Gao and McHenry,

2001a, 2001b; Jergic et al., 2007; Lopez de Saro et al., 2003).

Clamp loaders from bacteria, archaea, and eukaryotes all

contain a heteropentameric central ring of proteins in the AAA+

superfamily of ring-shaped ATPases (Cann and Ishino, 1999;

Cullmann et al., 1995; Neuwald et al., 1999; O’Donnell et al.,

1993). In eukaryotes, five different proteins build the central

ring, whereas in bacteriophage T4, a monomer of the protein

gp62 combines with a tetramer of gp44. Similar to T4, the

archaeal clamp loader consists of two different subunits with

stoichiometry that is still unclear. In E. coli, the composition is

particularly intriguing since a previous study indicated that while

DnaX (t/g) is in a monomer-tetramer equilibrium in solution, dd0

somehow forces DnaX into a trimer in creating the central ring

(Pritchard et al., 2000). The mechanism by which this occurs,

however, has not been established.

In this study, we use electrospray ionization (ESI) mass spec-

trometry (MS) to examine the assembly of the E. coli clamp

loader in solution. MS of intact complexes is capable of identi-

fying noncovalent protein interactions unambiguously under

dynamic equilibrium conditions where multiple intermediates

can be populated, revealing their stoichiometry and composi-

tions in real time (reviewed in Hernandez and Robinson, 2001;

Loo, 1997; Sharon and Robinson, 2007; van den Heuvel and

85–292, March 10, 2010 ª2010 Elsevier Ltd All rights reserved 285

mailto:[email protected]

Figure 1. Mass Spectra of the t and g Subunits in Isolation

(A) Mass spectrum of t in 0.1 M NH4OAc at pH 7.6 and 1 mM dithiothreitol. The

major charge state series corresponds to a tetrameric form of t. Series of t in

monomeric, dimeric, and trimeric forms are also observed as the minor charge

state series.

(B) Mass spectrum of g in 0.1 M NH4OAc at pH 6.9. Well resolved charge states

indicate that g forms a well defined tetramer.

(C) Mass spectrum of all four oligomeric species of g in 1 M NH4OAc at pH 6.9.

See also Table S1.

Structure

Assembly Pathway of the Sliding Clamp Loader

Heck, 2004). We set out to assemble the complete clamp loader

complex from its individual components and to study the subunit

exchange dynamics of t and g in solution. Our results reveal that

g4 is more stable than its t4 counterpart and that the principal role

of d0 in the initiation of assembly is to break the tetramers into

smaller oligomeric species. In the presence of d0, d and the cc

dimer further direct the assembly of the clamp loader by stabi-

lizing the trimeric form of (t/g)3 to form the full clamp loader

286 Structure 18, 285–292, March 10, 2010 ª2010 Elsevier Ltd All rig

complex. Our results further show that incubation of tetrameric

g4 with t4 results in rapid exchange of subunits to yield a statis-

tical distribution of heterotetramers. The presence of dd0, in

contrast, prevents t/g exchange such that subunit interactions

in the clamp loader must be preserved on the time scale of

chromosomal replication.

RESULTS

t and g Are Predominantly Tetramers in IsolationTo identify the preferred oligomeric states of the component

subunits that form the clamp loader (t, g, d, d0, and c), mass

spectra were recorded for all proteins in isolation. The proteins

d, d0, and c were found to exist primarily as monomers (see Table

S1 available online). For t, charge state series were observed

corresponding to four distinct species (Figure 1A). The measured

mass of 285,422 ± 44 Da for the major charge state series of t is

in close agreement with the calculated mass of the t tetramer,

284,548 Da (Table S1). The mass increase observed experimen-

tally is attributed to the incomplete removal of solvent and/or

buffer ions under MS instrument conditions used here (McKay

et al., 2006). Additional series of peaks observed in this spectrum

are assigned to populations of monomers, dimers, and trimers

of t, indicating that all four oligomeric species are present in

equilibrium. In contrast, only one charge state series was

observed for g, centered on a 32+ ion, indicating a homogeneous

tetrameric state (Figure 1B). Only by using higher ionic strength

buffer (>1 M NH4OAc) were we able to observe all four oligomeric

species of g (Figure 1C). This shows that g4 is less stable at high

ionic strength, dissociating to intermediate species similar to

those formed by t at lower ionic strength (0.1 M NH4OAc).

Overall, therefore, g4 appears to be more stable to dissociation

than t4.

d0 Is an Oligomer Breaker Creating Smaller Oligomersof t and g

To investigate how the tetrameric states of t/g adopt their estab-

lished trimeric forms in the clamp loader, we used MS to measure

changes in their oligomeric states as either d0 or d was added to

solutions containing the tetramers. The mass spectrum of t and

d0, at 1:2 mole ratio in solution, showed three distinct charge state

series corresponding to t3d0, t2d0, and td0 (Figure 2A; Table S1).

Furthermore, comparing this result to t alone, we observe

a dramatic increase in the population of smaller oligomers of t

in the presence of d0 (Figure 2A compared with Figure 1A). This

indicates that d0 alone is capable of trapping the smaller oligmeric

forms of t. For solutions of g with d0, again smaller oligomers ap-

peared; the subcomplexes g2d0 and g3d0 were observed

(Figure S1A). For neither t nor g were t4d0 or g4d0 subcomplexes

detected. Together these data indicate that the d0 subunit plays

the role of a t4/g4 ‘‘oligomer breaker’’ by selectively interacting

with and trapping trimers and smaller oligomers in solution.

In contrast, the analogous reaction with d in place of d0 showed

no change in the distributions of t/g oligomers and no new inter-

actions between the two components (Figure 2B; Figure S1B).

Thus, even though t3 was present in solution, d did not bind to

it to form the t3d subcomplex. This suggests two possibilities.

One is that the specific conformation of t/g produced by prior

binding of d0 is critical for association of d. The other is that

hts reserved

Figure 2. Mass Spectra of t in the Presence of Various Subunits

(A) In the presence of d0, t forms three subcomplexes, t3d0, t2d0, and td0 but

not t4d0.

(B) None of the oligomeric species of t associates with d.

(C) The cc complex binds to t, forming t4cc and t3cc.

(D) d does not associate with t in the presence of cc. See also Figure S1 for the

g subunit interaction and Table S1.

Structure


Structure 18, 2

d requires d0 to be present for binding t/g more tightly through

cooperative interactions with t/g. In either case, only d0 is

capable of initiating the assembly of the clamp loader by

breaking tetrameric (t/g)4 oligomers and making possible the

binding of d. As expected, when both d and d0 are added to t/g,

one predominant series of peaks was observed at high m/z and

assigned to the complete (t/g)3dd0 pentamer (Figure 3A; Fig-

ure S2A compared with Figure 2A and Figure S1A, where

multiple assemblies are formed in the presence of d0 alone).

This implies that after d0 breaks the tetramer of t or g, binding

of d stabilizes the trimeric state to drive the correct stoichiometry

of t/g in the clamp loader complex.

Interaction between the Pentameric Ring and cc

Assembly of the complete clamp loader complex requires

binding of the cc heterodimer to the central (t/g)3dd0 ring to

form (t/g)3dd0cc. We predicted that cc would stabilize the trimer

of t/g, since the recent crystal structure of g3dd0 bound to the

N-terminal interacting peptide of c shows that the peptide

interacts with all three g subunits at the collar domain but neither

d nor d0 (Simonetta et al., 2009). To investigate this, first we

examined the interaction of cc with other clamp loader subunits

(g, t, d, and d0) individually. We did not see any interaction

between cc and d or d0 as would be expected from the crystal

structure (Simonetta et al., 2009) and other biochemical studies

(Gao and McHenry, 2001c; Xiao et al., 1993b). In the mass spec-

trum of cc with t (Figure 2C), two distinct series of peaks were

observed, corresponding to homogenous populations of t4cc

and t3cc (Table S1). Since neither t2cc nor tcc was observed,

it can be inferred that cc binds more stably to t4/t3 and directs

assembly to (or traps) higher oligomeric species of t. Similarly,

the cc complex associates with g4, resulting in g4cc and g3cc

complexes (Figure S1C and Table S1). The effect of cc in stabi-

lizing higher oligmeric species is in contrast to d0, which, as we

have seen, specifically disrupts t and g tetramers. Interestingly,

for both t and g, t4cc and g4cc are major species, not t3cc and

g3cc. This indicates that cc preferentially associates with tetra-

mers of t/g rather than trimers. This may be simply because c

binds to only three subunits of g in a tetramer, consequently

there could be two different overlapping binding sites for c.

Given that the only oligomer breaker among the canonical

clamp loader subunits is d0, we investigated the order of binding

of d and cc to t3d0/t2d0/td0. When d was added to solutions

containing both t and d0, the central ring of the clamp loader,

t3dd0, formed rapidly, without incubation (Figure 3A). No detect-

able intermediate states were detected (e.g., t2dd0), equivalent to

when d and d0 were added together to t. As might be expected

from the g3dd0-c peptide structure (Simonetta et al., 2009), addi-

tion of cc to t3d0/t2d0/td0 produced the t3d0cc complex as the

major species (Figure 3B). Our insight into the assembly grows

further by noting that little t2d0cc was detected, implying that


Figure 3. Assembly Pathway of the Clamp Loader and Mass Spectra of Complexes Formed Along the Pathway

First, d0 initiates the assembly by dissociating tetramers of t into trimers, the functional unit of the clamp loader. Both d or cc interact with the initial t3d0 subcom-

plex, resulting in either t3dd0 or t3d0cc subcomplexes. Further addition of cc or d leads to the final clamp loader, t3dd0cc. Mass spectra are recorded for t3dd0 (A),

t3d0cc (B), and t3dd0cc (C and D), when cc was added to (A) and d to (B). For g, analogous results were observed in Figure S2, suggesting identical assembly

pathways. See also Table S1. Note that although all four species of oligomers of t are present in solution, only tetramers of t are shown initially.

Structure


in the presence of d0, cc forms the most stable complex with the

t trimer. The role of cc in stabilizing the canonical clamp loader

assembly is more clearly observed when the complex is con-

structed using g rather than t, since no g2d0cc was observed

in solution (Figure S2B). In addition, similar results were

observed when d0 was added to the (t/g)4cc and (t/g)3cc

(Figure 3B; Figure S2B). Therefore, cc stabilizes the trimer of

t/g in the presence of d0, guiding the assembly of the functional

stoichiometry of t/g in the clamp loader. This role is analogous

to that of d in the formation of the heteropentamer.

It is possible, however, that alternative assembly routes exist.

To investigate this we examined whether cc can support d in

binding of t/g. Mass spectra recorded after addition of d to

solutions of (t/g)4cc and (t/g)3cc showed no evidence of

(t/g)3dcc formation (Figure 2D; Figure S1D). Unlike d0, therefore,

the cc complex does not assist d in binding. Thus, we have

shown conclusively that only d0 can initiate the assembly of

the full clamp loader and subsequently enable binding of cc

and/or d (Figure 3).

Subunit Exchange Dynamics of t and g

So far we have studied the clamp loader assembly with either t or

g, but since all combinations of (t/g)3 form functional clamp

loaders, it is important to consider how the incorporation of t

versus g is established. First, we examined subunit exchange

between t and g in the absence of other clamp loader proteins.

Under the same buffer conditions and concentrations, solutions

of t and g were combined and mass spectra were recorded at

fixed time points to monitor the exchange reaction. All five


possible tetrameric species of t and g (t4, t3g, t2g2, tg3, and g4)

were formed immediately in a statistical distribution (Figure 4A)

that showed no further changes with time (Figure S3). This

indicates that subunit exchange within homotetramers of t and

g occurs rapidly and reaches equilibrium within seconds in vitro.

To see whether the rapid exchange dynamics observed for t

and g also occur in the presence of dd0, we mixed solutions

containing equimolar t3dd0 and g4 for specific time intervals.

Soon after mixing, we observed peaks corresponding to t3dd0,

g4, and, interestingly, g3dd0 (Figure 4B, bottom). Given that

t3dd0 is in equilibrium with many subspecies such as t3 and dd0

(Figure 3A), it is likely that the appearance of g3dd0 is not due to

g3 replacing t3 in t3dd0, but to the availability of free dd0 that could

convert g4 into the g3dd0 complex as described above. After

incubation for 1 hr, small peaks indicate minimal exchange of

t/g subunits (Figure 4B, middle). Nevertheless, all four forms of

(t/g)3dd0 were readily identified after 24 hr (Figure 4B, top). These

experiments demonstrate that in the presence of dd0, subunit

exchange between t and g takes place on a much longer time

scale relative to the situation where t and g alone are incubated.

Since this longer time scale is beyond the time required for chro-

mosome replication (�40 min), we surmise that dd0 prevents

significant subunit exchange in vivo.

DISCUSSION

Unique Roles for d and d0 in the Clamp Loader AssemblyIn this study we used MS to follow an assembly pathway for the

seven-subunit E. coli DNA sliding clamp loader, an important

hts reserved

Figure 4. Subunit Exchange Dynamics

between t and g

(A) t and g undergo rapid subunit exchange at 0�C

and form all five possible tetrameric species, t4,

t3g, t2g2, tg3, and g4, with the dead time of the

experiment <15 s. See also Figure S3.

(B) Mass spectra of t3dd0 in the presence of g4. At

t = 0, free dd0 interacts with g4 directly and forms

the g complex (bottom spectrum). At t = 1 hr, low

intensity peaks are observed (not labeled, middle

spectrum), which increase in intensity after 24 hr

and are assigned to all four possible complexes,

g3dd0, gt2dd0, g2tdd0, and t3dd0 (top spectrum).

See also Table S1.

Structure


example of an AAA+ motor class of multi-protein complex.

Although this clamp loader in E. coli contains trimeric t/g proteins,

previous studies using ultracentrifugation have suggested that

t/g in isolation exists in both monomeric and tetrameric forms

Structure 18, 285–292, March 10, 2010

(Dallmann and McHenry, 1995). As a

consequence, it was not clear how such

an oligomeric population is converted

into trimers in the presence of dd0 to form

the functional clamp loader. Here we

confirm that both t and g adopt a predom-

inantly tetrameric state but, in contrast to

the earlier work, dimers and trimers were

observed in addition to monomers.

Furthermore, we found that g4 is more

stable than t4 since, unlike t, the

tetramers of g do not dissociate under

similar experimental conditions.

By determining interactions among

components of the clamp loader, we

identified the crucial role of d0 in breaking

the t/g tetramer, to displace the equilib-

rium among t/g oligomers toward the

smaller oligomeric species. It is not clear

how d0 displaces t/g within the t/g oligo-

mers; however, we predict that the rigidity

within d0 offers constant binding sites for

t/g, which subsequently secures a more

stable conformation. It is also possible

that in the presence of d0, the fourth t/g

may be prevented sterically from forming

the circular architecture. In contrast, we

found that the d subunit alone could not

interact with or disrupt any of the t/g olig-

omers, in accord with previous studies

(Onrust et al., 1995a). In particular, al-

though even in the absence of d0 some

t3 exists in equilibrium, we observed

that d did not interact even with t3. There-

fore, the data led us to speculate that

d0 assists d in binding to (t/g)3 either by

facilitating changes in the conformation

within (t/g)3, revealing interaction sur-

faces on t/g, or simply by placing d next

to g as the result of binding d0. Thus, we

have identified two important roles of d0 in the assembly of the

clamp loader: first as a breaker of the t/g oligomer and second

as a promoter for binding d and consequently completing the

pentameric central ring. Moreover, since we observed no smaller

ª2010 Elsevier Ltd All rights reserved 289

Structure


oligomeric species of t/g in the presence of both d and d0, it

appears that d not only completes the central ring but that it

also stabilizes the trimeric t/g in the presence of d0.

A similar stabilizing effect of cc on (t/g)3 was observed, further

ensuring the correct stoichiometry of t/g in the heteropentameric

clamp loader complex. However, we found that this locking of

t/g into a trimeric form by cc is only established in the presence

of d0. Reconciling our results with the g3dd0-c peptide structure

(Simonetta et al., 2009), we conclude that cc does not break

the t/g tetramer; it may nonetheless stabilize a trimer within the

tetramer. It may be that cc stabilizes the t/g trimer, which in

turn assists in entry of d0, providing an alternative assembly

pathway of the clamp loader complex. Assembly dynamics of

the clamp loader complex may also differ in the presence of

nucleotides, although the clamp loader assembles indepen-

dently in vitro without nucleotides. Nonetheless, it is clear that

d0 plays a crucial role as an oligomer breaker in directing the

correct assembly of the clamp loader.

Final Stoichiometry of t versus g

in the Clamp Loader ComplexIn vivo, DNA replication proceeds on both the leading and

lagging template strands of DNA within a single DNA replisome.

Each strand requires a replicative polymerase (the a3q core),

tethered to the clamp loader complex through a via the C

terminus of a t subunit within the assembly. Therefore, for

‘‘coupled’’ replication to take place, at least two t subunits per

clamp loader are required [i.e., (a3q-t)2gdd0cc]. However,

whether there are two or three t subunits within a functioning

replisome in vivo is yet to be determined.

Prior work on assembly of the holoenzyme and clamp loader

complexes from the component subunits in vitro (Onrust et al.,

1995a; Pritchard and McHenry, 2001) is consistent with the prop-

osition that exchange of the t and g subunits can only occur

efficiently prior to their assembly with d and d0 into the active

clamp loader complexes. Nevertheless, the picture has been

complicated by belief that (a) simply mixing all ten subunits

produces only the t3 holoenzyme (McInerney et al., 2007; Pritch-

ard et al., 2000; Pritchard and McHenry, 2001), (b) that Pol III*

(holoenzyme lacking b) isolated directly from E. coli contains

both the t3 holoenzyme and the g3 clamp loader complex, lacking

the heterotrimeric t1g2 and t2g1 species (McInerney et al., 2007),

and (c) that Pol III’, a species with the composition (a3q-t)2, is an

intermediate in the holoenzyme assembly (Pritchard and

McHenry, 2001). It has been shown that in vitro assembly of

t1g2 and t2g1 clamp loader complexes requires lengthy preincu-

bation of separately purified t and g to allow exchange before

addition of the other subunits (Onrust et al., 1995a, 1995b) and

that t/g exchange is prevented when the other clamp loader

subunits are present (Onrust et al., 1995b; Pritchard and

McHenry, 2001). Nevertheless, overexpression of t and g from

wild-type dnaX in a single operon with genes encoding the other

four subunits produced g3, t1g2, and t2g1 clamp loader com-

plexes, indicating fast exchange of newly translated t and g

in vivo, at least when at elevated concentrations (Pritchard

et al., 2000).

These previous studies used chromatographic methods to

study clamp loader assembly. In this study, we exploited the

ability of ESI MS to rapidly determine compositions of heteroge-


neous complexes to show that all combinations of mixed

tetramers of t and g were assembled immediately upon incuba-

tion of the two proteins under our buffer conditions (Figure 4A).

This is expected given that the two proteins exist in solution as

equilibrium mixtures of oligomers. That the distribution of mixed

tetramers appears to be statistical (Figure 4B) indicates that

there is no preferential association between t subunits; i.e., all

interactions that determine assembly among subunits are

restricted to the N-terminal region common to the two proteins.

Because subunit exchange occurs rapidly, it is reasonable

to think that all combinations of t and g would assemble into

mixed tetramers following translation in vivo, and thus it

seems very likely that E. coli makes all Pol III*/clamp loader

assemblies, including g3dd0cc, (a3q-t)g2dd0cc, (a3q-t)2gdd0cc,

and (a3q-t)3dd0cc, although only the last two complexes are

capable of coupled leading and lagging strand replication (McI-

nerney et al., 2007). Interestingly, g is not essential for cell

viability (Blinkova et al., 1993), which implies that (a3q-t)3dd0cc

not only accomplishes coupled DNA synthesis but must also

be capable of clamp loading both in replisome assembly and

elsewhere. It has been suggested that g3dd0cc may function in

clamp loading for other DNA maintenance functions (Bloom,

2009). We have also confirmed that dd0 prevents the subunit

exchange between t and g. Thus, the compositions of individual

clamp loader/holoenzyme assemblies are likely to be preserved

as they carry out their specific functions, (a3q-t)3dd0cc and

(a3q-t)2gdd0cc in replication and g3dd0cc in clamp loading (or

unloading) for other activities. While there is no clear role for

(a3q-t)g2dd0cc, which would be incapable of lagging strand

synthesis at a fork, its comparative lack of necessary protein-

protein and DNA interactions would probably place it at

a competitive disadvantage in functional replisome assembly.

In summary we have addressed two major problems: first,

how the central pentameric ring of the clamp loader in E. coli

is formed from six proteins (a tetramer of t/g, d, and d0) and,

second, how the cell avoids creating a replisome with fewer

than two copies of t. The first problem is solved in a surprising

way, by d0 first breaking the tetramers of t/g and then (and only

then) enabling d to bind and lock in the central pentameric ring.

This solution to the first problem also turns out to be related to

the second. Our subunit exchange studies between t and g

have indicated that all combinations of (t/g)3dd0 are likely

formed in vivo; yet, once any are formed, the presence of

d and d0 prevents the exchange of t and g subunits within the

time scale of DNA synthesis. Once the replisome is formed

with either t2gdd0 or t3dd0, the stability of these central pen-

tameric rings therefore prevents them from exchanging to a

form with fewer than two copies of t. Thus, the second problem

is apparently solved, not by the abolition of tg2dd0 and g3dd0,

but by their exclusion from the assembly of the functional

replisome.

EXPERIMENTAL PROCEDURES

Proteins and Sample Preparation for MS

Each subunit, t, g, d, d0, c, cc, and a, was overexpressed in E. coli and purified

as described (Ozawa et al., 2008; Tanner et al., 2008; Wijffels et al., 2004;

Xiao et al., 1993a). For mass spectra of individual proteins, t (10.2 mg/ml),

g (3.2 mg/ml), d (3.8 mg/ml), d0 (4.0 mg/ml), c (2.4 mg/ml), cc (4.8 mg/ml),

and a (9.2 mg/ml) were buffer exchanged into appropriate NH4OAc buffers

hts reserved

Structure


using Vivaspin 500 with 10–50 kDa molecular weight cutoff, depending on the

size of the protein (Sartorius). Various concentrations and pHs of NH4OAc

buffers were used for each protein: 0.1 M NH4OAc at pH 7.6 containing

1 mM dithiothreitol for t, 0.1 M NH4OAc at pH 6.9 for g, 1 M NH4OAc at

pH 6.9 for d, 0.5 M NH4OAc at pH 6.9 for d0, and 0.1 M NH4OAc at pH 7.6

for cc and c. Concentrations for t, g, d, d0, c, cc, and a were measured spec-

trophotometrically at 280 nm, using 3280 = 46380, 20940, 46830, 60440, 29400,

53680, and 96880 M�1cm�1, respectively. The final concentrations of the

proteins were in the range 5 to 10 mM. For subunit exchange experiments, t4

or t3dd0 and g4 at 5 mM, subsequent to buffer exchange in 0.1 M NH4OAc at

pH 7.6 and 1 mM dithiothreitol, were mixed at 0�C. MS were recorded every

10 min for 1 hr and after 24 hr.

Mass Spectrometry

Nanoflow ESI MS and tandem MS experiments were conducted on a high

mass Q-TOF type instrument (Sobott et al., 2002) adapted for a QSTAR XL

platform (Chernushevich and Thomson, 2004). Typically, 2 ml aliquots of

solution were electrosprayed from gold-coated borosilicate capillaries

prepared in-house as described (Hernandez and Robinson, 2007). MS

experiments were conducted at a capillary voltage up to 1.2 kV, declustering

potential 100–150 V, focusing potential 200–250V, declustering potential two

15–55 V, collision energy up to 180 V, and MCP 2350 V. All spectra were

calibrated externally using a solution of cesium iodide (100 mg/ml). Spectra

are shown here with minimal smoothing and without background subtraction.

SUPPLEMENTAL INFORMATION

Supplemental Information includes three figures and one table and can be

found with this article online at doi:10.1016/j.str.2010.01.009.

ACKNOWLEDGMENTS

We acknowledge funding from the Biotechnology and Biological Sciences

Research Council (to A.Y.P., A.P., and D.B.), the Royal Society and Walter’s

Kundert Trust (to C.V.R.), the Australian Research Council (to J.L.B.) and

Australian Professorial Fellowship (to N.E.D), and Waters (B.T.R.) and sabbat-

ical funding from Lawrence Livermore National Laboratory (to D.B.).

Received: November 20, 2009

Revised: January 19, 2010

Accepted: January 22, 2010

Published: March 9, 2010

REFERENCES

Blinkova, A., Hervas, C., Stukenberg, P.T., Onrust, R., O’Donnell, M.E., and

Walker, J.R. (1993). The Escherichia coli DNA polymerase III holoenzyme

contains both products of the dnaX gene, t and g, but only t is essential. J.

Bacteriol. 175, 6018–6027.

Blinkowa, A.L., and Walker, J.R. (1990). Programmed ribosomal frameshifting

generates the Escherichia coli DNA polymerase III g subunit from within the t

subunit reading frame. Nucleic Acids Res. 18, 1725–1729.

Bloom, L.B. (2009). Loading clamps for DNA replication and repair. DNA

Repair (Amst.) 8, 570–578.

Cann, I.K., and Ishino, Y. (1999). Archaeal DNA replication: identifying the

pieces to solve a puzzle. Genetics 152, 1249–1267.

Chernushevich, I.V., and Thomson, B.A. (2004). Collisional cooling of large ions

in electrospray mass spectrometry. Anal. Chem. 76, 1754–1760.

Cullmann, G., Fien, K., Kobayashi, R., and Stillman, B. (1995). Characterization

of the five replication factor C genes of Saccharomyces cerevisiae. Mol. Cell.

Biol. 15, 4661–4671.

Dallmann, H.G., and McHenry, C.S. (1995). DnaX complex of Escherichia coli

DNA polymerase III holoenzyme. Physical characterization of the DnaX

subunits and complexes. J. Biol. Chem. 270, 29563–29569.

Structure 18, 2

Flower, A.M., and McHenry, C.S. (1990). The g subunit of DNA polymerase III

holoenzyme of Escherichia coli is produced by ribosomal frameshifting. Proc.

Natl. Acad. Sci. USA 87, 3713–3717.

Gao, D., and McHenry, C.S. (2001a). t binds and organizes Escherichia coli

replication through distinct domains. Partial proteolysis of terminally tagged

t to determine candidate domains and to assign domain V as the a binding

domain. J. Biol. Chem. 276, 4433–4440.

Gao, D., and McHenry, C.S. (2001b). t binds and organizes Escherichia coli

replication proteins through distinct domains. Domain IV, located within the

unique C terminus of t, binds the replication fork, helicase, DnaB. J. Biol.

Chem. 276, 4441–4446.

Gao, D., and McHenry, C.S. (2001c). tau binds and organizes Escherichia coli

replication proteins through distinct domains. Domain III, shared by g and t,

binds d d 0 and c c. J. Biol. Chem. 276, 4447–4453.

Gulbis, J.M., Kazmirski, S.L., Finkelstein, J., Kelman, Z., O’Donnell, M., and

Kuriyan, J. (2004). Crystal structure of the chi:psi sub-assembly of the

Escherichia coli DNA polymerase clamp-loader complex. Eur. J. Biochem.

271, 439–449.

Hernandez, H., and Robinson, C.V. (2001). Dynamic protein complexes:

insights from mass spectrometry. J. Biol. Chem. 276, 46685–46688.

Hernandez, H., and Robinson, C.V. (2007). Determining the stoichiometry

and interactions of macromolecular assemblies from mass spectrometry.

Nat. Protoc. 2, 715–726.

Jergic, S., Ozawa, K., Su, X.-C., Scott, D.D., Williams, N.K., Hamdan, S.M.,

Crowther, J.A., Otting, G., and Dixon, N.E. (2007). The unstructured

C-terminus of the t subunit of Escherichia coli DNA polymerase III holoenzyme

is the site of interaction with the a subunit. Nucleic Acids Res. 35, 2813–2824.

Jeruzalmi, D., O’Donnell, M., and Kuriyan, J. (2001). Crystal structure of

the processivity clamp loader gamma (g) complex of E. coli DNA

polymerase III. Cell 106, 429–441.

Jeruzalmi, D., O’Donnell, M., and Kuriyan, J. (2002). Clamp loaders and sliding

clamps. Curr. Opin. Struct. Biol. 12, 217–224.

Johnson, A., and O’Donnell, M. (2005). Cellular DNA replicases: components

and dynamics at the replication fork. Annu. Rev. Biochem. 74, 283–315.

Kornberg, A., and Baker, T.A. (1992). DNA Replication, Second Edition

(New York: W.H. Freeman).

Loo, J.A. (1997). Studying noncovalent protein complexes by electrospray

ionization mass spectrometry. Mass Spectrom. Rev. 16, 1–23.

Lopez de Saro, F.J., Georgescu, R.E., and O’Donnell, M. (2003). A peptide

switch regulates DNA polymerase processivity. Proc. Natl. Acad. Sci. USA

100, 14689–14694.

McHenry, C.S. (2003). Chromosomal replicases as asymmetric dimers:

studies of subunit arrangement and functional consequences. Mol. Microbiol.

49, 1157–1165.

McInerney, P., Johnson, A., Katz, F., and O’Donnell, M. (2007). Characteriza-

tion of triple DNA polymerase replisome. Mol. Cell. 27, 527–538.

McKay, A.R., Ruotolo, B.T., Ilag, L.L., and Robinson, C.V. (2006). Mass

measurements of increased accuracy resolve heterogeneous populations of

intact ribosomes. J. Am. Chem. Soc. 128, 11433–11442.

Naktinis, V., Onrust, R., Fang, L., and O’Donnell, M. (1995). Assembly of a

chromosomal replication machine: two DNA polymerases, a clamp loader,

and sliding clamps in one holoenzyme particle. II. Intermediate complex

between the clamp loader and its clamp. J. Biol. Chem. 270, 13358–13365.

Neuwald, A.F., Aravind, L., Spouge, J.L., and Koonin, E.V. (1999). AAA+:

a class of chaperone-like ATPases associated with the assembly, operation,

and disassembly of protein complexes. Genome Res. 9, 27–43.

O’Donnell, M., and Kuriyan, J. (2006). Clamp loaders and replication initiation.

Curr. Opin. Struct. Biol. 16, 35–41.

O’Donnell, M., Onrust, R., Dean, F.B., Chen, M., and Hurwitz, J. (1993).

Homology in accessory proteins of replicative polymerases—E. coli to

humans. Nucleic Acids Res. 21, 1–3.

Onrust, R., Finkelstein, J., Naktinis, V., Turner, J., Fang, L., and O’Donnell, M.

(1995a). Assembly of a chromosomal replication machine: two DNA


http://dx.doi.org/doi:10.1016/j.str.2010.01.009

Structure


polymerases, a clamp loader, and sliding clamps in one holoenzyme particle. I.

Organization of the clamp loader. J. Biol. Chem. 270, 13348–13357.

Onrust, R., Finkelstein, J., Turner, J., Naktinis, Y., and O’Donnell, M. (1995b).

Assembly of a chromosomal replication machine: two DNA polymerases,

a clamp loader, and sliding clamps in one holoenzyme particle. III. Interface

between two polymerases and the clamp loader. J. Biol. Chem. 270, 13366–

13377.

Ozawa, K., Jergic, S., Crowther, J.A., Thompson, P.R., Wijffels, G., Otting, G.,

and Dixon, N.E. (2005). Cell-free protein synthesis in an autoinduction system

for NMR studies of protein-protein interactions. J. Biomol. NMR 32, 235–241.

Pritchard, A.E., Dallmann, H.G., Glover, B.P., and McHenry, C.S. (2000).

A novel assembly mechanism for the DNA polymerase III holoenzyme DnaX

complex: association of deltadelta’ with DnaX(4) forms DnaX(3)deltadelta’.

EMBO J. 19, 6536–6545.

Pritchard, A.E., and McHenry, C.S. (2001). Assemby of DNA polymerase III

holoenzyme: co-assembly of gamma and tau is inhibited by DnaX complex

accessory proteins but stimulated by DNA polymerase III core. J. Biol.

Chem. 276, 35217–35222.

Schaeffer, P.M., Headlam, M.J., and Dixon, N.E. (2005). Protein–protein

interactions in the eubacterial replisome. IUBMB Life 57, 5–12.

Sharon, M., and Robinson, C.V. (2007). The role of mass spectrometry in

structure elucidation of dynamic protein complexes. Annu. Rev. Biochem.

76, 167–193.

Simonetta, K.R., Kazmirski, S.L., Goedken, E.R., Cantor, A.J., Kelch, B.A.,

McNally, R., Seyedin, S.N., Makino, D.L., O’Donnell, M., and Kuriyan, J.

(2009). The mechanism of ATP-dependent primer-template recognition by

a clamp loader complex. Cell 137, 659–671.


Sobott, F., Hernandez, H., McCammon, M.G., Tito, M.A., and Robinson, C.V.

(2002). A tandem mass spectrometer for improved transmission and analysis

of large macromolecular assemblies. Anal. Chem. 74, 1402–1407.

Tanner, N.A., Hamdan, S.M., Jergic, S., Loscha, K.V., Schaeffer, P.M., Dixon,

N.E., and van Oijen, A.M. (2008). Single-molecule studies of fork dynamics in

Escherichia coli DNA replication. Nat. Struct. Mol. Biol. 15, 170–176.

Tsuchihashi, Z., and Kornberg, A. (1990). Translational frameshifting generates

the g subunit of DNA polymerase III holoenzyme. Proc. Natl. Acad. Sci. USA

87, 2516–2520.

van den Heuvel, R.H., and Heck, A.J. (2004). Native protein mass spectrom-

etry: from intact oligomers to functional machineries. Curr. Opin. Chem. Biol.

8, 519–526.

Wijffels, G., Dalrymple, B.P., Prosselkov, P., Kongsuwan, K., Epa, V.C., Lilley,

P.E., Jergic, S., Buchardt, J., Brown, S.E., Alewood, P.F., et al. (2004).

Inhibition of protein interactions with the beta 2 sliding clamp of Escherichia

coli DNA polymerase III by peptides from beta 2-binding proteins. Biochem-

istry 43, 5661–5671.

Xiao, H., Crombie, R., Dong, Z., Onrust, R., and O’Donnell, M. (1993a). DNA

polymerase III accessory proteins. III. holC and holD encoding chi and psi.

J. Biol. Chem. 268, 11773–11778.

Xiao, H., Dong, Z., and O’Donnell, M. (1993b). DNA polymerase III accessory

proteins. IV. Characterization of chi and psi. J. Biol. Chem. 268, 11779–11784.

Yuzhakov, A., Kelman, Z., and O’Donnell, M. (1999). Trading places on DNA—

a three-point switch underlies primer handoff from primase to the replicative

DNA polymerase. Cell 96, 153–163.

hts reserved

a single subunit directs the assembly of the escherichia coli dna sliding clamp loader

Documents