Assembly and Subunit Stoichiometry of the Functional Helicase-Primase (Primosome) Complex of Bacteriophage T4

Assembly and subunit stoichiometry of the functional helicase-primase (primosome) complex of bacteriophage T4

Davis Jose, Steven E. Weitzel, Debra Jing, and Peter H. von Hippel1

Institute of Molecular Biology and Department of Chemistry, University of Oregon, Eugene, OR 97403

Contributed by Peter H. von Hippel, July 9, 2012 (sent for review May 13, 2011)

Physical biochemical techniques are used to establish the structure, rings of identical subunits and interact with one or more primase subunit stoichiometry, and assembly pathway of the primosome subunits to form functional primosomes (5–7). Thus, the T4 DNA complex of the bacteriophage T4 DNA replication system. Analyti- replication system and its constituent subassemblies comprise the cal ultracentrifugation and fluorescence anisotropy methods show simplest model system for studying many aspects of DNA repli- that the functional T4 primosome consists of six gp41 helicase sub- cation in higher organisms (8). Furthermore, each of the three units that assemble into a hexagon, driven by the binding of six functional protein subassemblies can be loaded onto an appropri- NTPs (or six nonhydrolyzable GTPγS analogues) that are located ate DNA construct in vitro, permitting their mechanisms to be at and stabilize the intersubunit interfaces, together with a single studied separately. DNA helicase activity is required to unwind tightly bound gp61 primase subunit. Assembling the components the genomic DNA at the replication fork and to provide access of the primosome onto a model DNA replication fork is a multistep for the DNA polymerases to their respective templates. This heli- process, but equilibrium cannot be reached along all mixing path- case function must be integrated with that of the primase, which ways. Producing a functional complex requires that the helicase serves as a low processivity RNA polymerase that carries out tem- hexamer be assembled in the presence of the DNA replication fork plate-directed synthesis of RNA (5-mer) primers at approxi- construct prior to the addition of the primase to avoid the forma- mately 1-kb intervals along the lagging strand DNA template. tion of metastable DNA-protein aggregates. The gp41 helicase hex- This functional integration of the helicase and the primase com- amer binds weakly to fork DNA in the absence of primase, but ponents of the T4 system alters both activities within the replica- forms a much more stable primosome complex that expresses full tion complex, and these interactions, together with the assembly and functional helicase (and primase) activities when bound to a pathway and subunit composition of the T4 primosome, form the gp61 primase subunit at a helicase:primase subunit ratio of 6∶1. main focus of this paper. The presence of additional primase subunits does not change the Biophysical studies of T4 primosome complexes (6, 7, 9, 10) molecular mass or helicase activity of the primosome, but signifi- have provided some information about their subunit stoichi- cantly inhibits its primase activity. We develop both an assembly ometry, as well as insight into how these components work in pathway and a minimal mechanistic model for the structure and conjunction with one another. However, two different sets of function of the T4 primosome that are likely to be relevant to observations have come to different conclusions about the stoi- the assembly and function of the replication primosome subassem- chiometry (and thus the mechanism) of the T4 primosome. In blies of higher organisms as well. early work gel shift and protein–protein chemical cross-linking methodologies had shown that the gp61 primase subunit binds DNA–protein complexes ∣ macromolecular machines ∣ to the gp41 helicase hexamer at an approximately 1∶6 subunit duplex DNA unwinding ∣ replication complex assembly stoichiometry (5, 7). In contrast, Benkovic and coworkers have interpreted more recent and indirect fluorescence anisotropy he DNA replication system of bacteriophage T4 contains and scanning calorimetry experiments (10–12) to argue that the Teight different types of protein subunits, several present in primase subunits may interact with gp41 in an approximately 6∶6 multiple copies. Subsets of these components form three stable subunit stoichiometry. This difference is important because in a and functional protein complexes that can be assembled onto primosome containing one primase and six helicase subunits the a model DNA replication fork in vitro to form an integrated primase must be shared, which makes very different functional T4 DNA replication complex that is capable of unwinding the and mechanistic demands on both helicase and primase mechan- parental DNA duplex and synthesizing new viral DNA with isms than does a 6∶6 ratio, as is found in the hexameric helicase of essentially in vivo rates and fidelity (1, 2). These replication sub- bacteriophage T7, where every helicase subunit carries its own assemblies are: (i) the leading- and lagging-strand replication primase domain (13). The replication complex of T7 differs in polymerases that catalyze the template-directed copying of the many important aspects from that of T4 or the replication com- two parental-strands of the DNA genome at each cell division; plexes of higher organisms. (ii) the clamp-clamp loader complex that controls the processivity Given the importance of the bacteriophage T4 replication of the replication process by linking the polymerases to their complex as a model system, we have undertaken additional and respective template strands and also regulates the release and re- careful studies by several methodologies to further examine the cycling of the lagging strand polymerase following the completion issues of subunit stoichiometry and assembly of the T4 primo- of each Okazaki fragment; and (iii) the helicase-primase (primosome. The results increase our understanding of the molecular some) complex that unwinds the double-stranded genome ahead of the replication fork in its capacity as a helicase, while also per- Author contributions: D. Jose and P.H.v.H. designed research; D. Jose, S.E.W. and D. Jing forming template-directed synthesis of the RNA primers that re- performed research; D. Jose, S.E.W., D. Jing and P.H.v.H. analyzed data; and D. Jose and initiate discontinuous lagging-strand DNA synthesis after each P.H.v.H. wrote the paper. Okazaki fragment has been completed. The authors declare no conflict of interest. The subunit components and stoichiometries of the replication 1To whom correspondence should be addressed. E-mail: [email protected]. complexes of higher organisms closely resemble those of T4 This article contains supporting information online at www.pnas.org/lookup/suppl/ (3, 4). Their helicases, in the presence of NTP, all form hexameric doi:10.1073/pnas.1210040109/-/DCSupplemental.

13596–13601 ∣ PNAS ∣ August 21, 2012 ∣ vol. 109 ∣ no. 34 www.pnas.org/cgi/doi/10.1073/pnas.1210040109 Downloaded by guest on September 24, 2021 interactions that regulate primosome assembly and its function as a helicase and a primase, as well as providing insight into interactions between the primosome and the other subcomplexes of the T4 DNA replication system and, by extension, our understanding of helicase-primase complexes in higher organisms. Results Assembly of the T4 helicase hexamer depends on the concentrations of gp41 subunits and NTP ligands. Analytical ultracentrifugation was used to characterize the equilibria involved in the assembly of the T4 gp41 helicase hexamer. Sedimentation velocity profiles of gp41 at a subunit concentration of 1.5 μM in the absence of NTP ligands showed only one significant peak with an s20;w of approximately 4.5 S (Fig. 1A; see also Fig. S1). Assuming average hydration, this is close to the sedimentation coefficient expected for these relatively spherical gp41 subunits (14), and confirms earlier findings that gp41 protein exists in solution pri- marily as a monomer at these protein concentrations in the absence of NTP (6). We confirmed also that the addition of GTPγS Fig. 1. Oligomerization of T4 helicase (gp41) subunits as a function of gp41 (a nonhydrolyzable analogue of GTP) to a solution of gp41 mono- and GTPγS concentrations. Sedimentation velocity profiles [(c)s versus s20;w μ γ mers at this concentration drives this replication helicase into a distribution plots] at: (A)1.5 M gp41 (monomers) with no GTP S; (B) 3 μM gp41 and 5 μM GTPγS; (C)3μM gp41 and 30 μM GTPγS; and (D) stable hexameric state. We note that two additional components μ μ γ s γ 3 M gp41 and 60 M GTP S. The approximately 4.5 S peak corresponds appeared in the c(s) versus 20;w distribution plots as the GTP S to gp41 monomers and the peak at approximately 10.5 S to gp41 hexamers. B C concentration was increased (Fig. 1 and ). These components The peaks at intervening s20;w values represent intermediate oligomers of fall at s20;w values between that of the monomer (approximately gp41 subunits and show a progressive shift into hexamers as the concentra- 4.5 S) and the approximately 10.5 S peak that corresponds to the tion of GTPγS is increased. (Additional plots showing complete sedimentation s sedimentation of the hexamer form that dominates at high profiles and residual values for the (c)s versus 20;w distribution plots of B and GTPγS concentrations (Fig. 1 C and D and Fig. S1F).* We sug- D are presented in Fig. S1.) gest that these intermediate peaks represent hexamer assembly intermediates, with the approximately 6 S peak corresponding cence anisotropy experiments in which increasing concentrations of gp61 were added to ssDNA confirmed the appearance of un- to gp41 dimers at approximately 8.2 S to gp41 tetramers. These D assignments are consistent with hydrodynamic modeling calcula- defined primase-DNA aggregates (Fig. S3 ). To assemble the tions (see discussion of hydrodynamic calculations in SI Text). components of the T4 replication system and determine the ratios These intermediate peaks decreased in size as the concentration of individual subunits in the final complex requires that we estab- of GTPγS was increased over the experimentally accessible con- lish an assembly pathway that produces a stable complex and centration range, as expected for an equilibrium system.† At a 20- avoids the formation of metastable aggregates. To this end, a ser- fold molar excess of GTPγS over protein subunits most of the ies of experiments were performed in which various order-of-ad- complexes formed physiologically relevant hexamers (Fig. 1D dition protocols were tested and the concentrations of individual and Fig. S1 D and F) stabilized by NTP ligands binding at the components were varied to define optimal (and equilibrium) mix- gp41 subunit-subunit interfaces. At 3 μM concentrations of gp41 ing conditions for assembling stable and active primosomes on subunits an approximately 60-μM concentration of GTPγS was DNA fork constructs. The details of some of these assembly path- SI Text required to drive the assembly of gp41 hexamers essentially to way experiments are presented in . We tested a possible assembly route involving initial mixing completion under the solution conditions used here and in the γ absence of gp61 and DNA. of helicase and primase subunits in the presence of GTP S, followed by the addition of DNA. This procedure did not result in The Order of Addition of Components is Important in Assembling stable primosome-DNA complexes (Fig. S4). The addition of γ Stable T4 Primosomes. Analytical ultracentrifugation and steady gp41 to mixtures of DNA, gp61, and GTP S also resulted in ag- state fluorescence experiments showed that gp41 hexamers do gregation and precipitation (Fig. S5), showing that this order-of- not form stable complexes with ssDNA or DNA fork constructs addition pathway leads to the formation of metastable aggregates at the concentrations tested (see SI Text and Fig. S2 A and B). rather than stable and equilibrated DNA-bound primosome complexes. Finally, we tried adding gp61 to an initial mixture of DNA Nevertheless, the helicase hexamer does exist as a stable compo- BIOPHYSICS AND and hexameric helicase molecules. Using this assembly protocol

nent within the T4 replication system at the DNA replication fork. COMPUTATIONAL BIOLOGY γ We had previously suggested that T4 gp61 primase might play this we were able to show that adding GTP S to a solution containing stabilizing role (7). However, undefined primase-DNA aggregates the DNA construct, followed by the addition of gp41 subunits, are formed when T4 primase, which sediments as a monomer with resulted in the formation of unstable DNA-gp41 helicase hexam- an s20;w of approximately 2.6 S in isolation (Fig. S3A), is mixed er complexes that could then be stabilized by adding primase. with ssDNA or DNA fork constructs (Fig. S3 B and C). Fluores- This order-of-addition pathway led to clear and reproducible sedimentation velocity boundaries (Fig. 2 A–D) and to well- defined stoichiometric complexes in fluorescence anisotropy *We note that these c(s) versus S distribution plots show a nonzero point (and a peak) titrations (Fig. 2E). at each sedimentation coefficient value at which a resolvable boundary is centered in a sedimentation velocity run. Points are taken at 0.1 S intervals. The height of each ∶ peak represents the concentration of the corresponding component, but we note A Stable Primosome Complex is Formed with DNA at a 6 1 Helicase- that spreading due to homogenous (Gaussian) diffusion has been removed from these Primase Subunit Stoichiometry. To achieve maximal processivity representations and therefore the apparent width of each peak (when it contains more and activity, a functional helicase must bind tightly to the DNA than one point) is due only to the presence of minor apparently independently sediment- ing components detected by the program. For further details of the steps involved in the fork construct that serves as an unwinding substrate. Sedimenta- analysis of such sedimentation velocity experiments see SI Text, Fig. S1. tion velocity experiments showed that the addition of even small †The GTPγS concentrations used were ≤60 μM; at higher concentrations the optical density amounts of gp61 increased the stability of complexes between of this ligand at 280 nm overwhelmed the optical density of the protein components. gp41 hexamers and DNA (Fig. S6). Titration of gp61 into pre-

Jose et al. PNAS ∣ August 21, 2012 ∣ vol. 109 ∣ no. 34 ∣ 13597 Downloaded by guest on September 24, 2021 unit ratio for the functional T4 primosome subassembly. As stated above, our successful approach involved forming weakly associating DNA-gp41-GTPγS complexes and then adding gp61. This procedure permits the system to equilibrate without forming metastable primase-DNA aggregates. Experiments were performed in which the concentration of input gp61 subunits was varied relative to the concentration of the DNA-gp41-GTPγS components present (Fig. 2). Components mixed at an input (subunit or molecule) ratio of 6∶1∶1 (gp41∶gp61∶DNA) showed a major sedimentation peak at s20;w ¼ 12.1 S, and smaller peaks at 2.6 S and 4.5 S (Fig. 2A). These smaller peaks sediment with the s20;w values characteristic of gp41 and gp61 monomers (Fig. 1A and Fig. S3A). We suggest that the 12.1 S peak corresponds to an equilibrium ternary (DNA-gp41-gp61) primosome complex that contains one primase subunit per gp41 hexamer. The s20;w of this component is substantially greater than the 10.5 S value measured for the GTPγS-stabilized gp41 hexamer in the absence of primase (Fig. 1D).‡

Varying the Input Ratios of gp41 and gp61 Subunits. This 6∶1 helicase∶primase subunit assignment was further tested in sedimentation experiments by varying the subunit input ratios. In one set of experiments the input gp61 concentration was increased either sixfold or ninefold, resulting in a 6∶6 (Fig. 2B)or6∶9 (Fig. 2C) helicase∶primase input subunit ratios. The formation of stable primosome complexes containing a gp41 helicase hexamer and more than one primase subunit should increase the observed s20;w values for the ternary complexes at the higher gp61 input Fig. 2. Association states of ssDNA-gp41-gp61 complexes in the presence of levels. We found that a s20;w of 12.1 0.5 S was maintained at GTP or GTPγS show that the helicase∶primase subunit ratio in the T4 primo- all three input ratios, suggesting that larger stable primosome 6∶1 s some is . Sedimentation velocity c(s) versus 20;w distribution plots for: (A) complexes containing more primase subunits per helicase hexam- μ μ γ 41∶ 61 500 nM dT45,3 M gp41, 30 M GTP S and 500 nM gp61 (ratio of gp gp er do not form (Fig. 2D). These experiments also showed that the 6∶1 μ μ μ γ μ subunits is ); (B)3 MdT45,3 M gp41, 60 M GTP S and 3 M gp61 (ratio gp61 subunits present beyond those needed to form the 6∶1 com- of gp41∶gp61 subunits is 1∶1); and (C)3μMdT45,3μM gp41, 60 μM GTPγS plex sedimented at approximately 2.2 S, close to the s20;w for free and 4.3 μM gp61 (ratio of gp41∶gp61 subunits is 2∶3). All plots show three gp61 monomers, while the primosome peak remained at approxi- peaks with s20;w values of approximately 4.5 S (gp41 monomers), approximately 2.6 S (gp61 monomers) and approximately 12.1 S (DNA-primosome mately 12.1 S. Similar experiments were performed with solutions complex). (D) Primosome-DNA s20;w values as a function of increasing concen- containing input gp41 concentrations in excess of the 6∶1 trations of gp61 subunits per gp41 hexamer. (E) Steady state fluorescence helicase∶primase subunit ratio (see SI Text and Fig. S6). Here, the anisotropy monitoring the titration of gp61 into a solution containing excess gp41 appeared as gp41 oligomer peaks that sedimented at 3 μM5′-Oregon green-labeled ssDNA, 3 μM gp41 (monomers), and 60 μM approximately 10.5 S or less, while the largest (presumably pri- γ GTP S. The anisotropy data show a stoichiometric break point at approxi- mosome) peak continued to sediment at 12.1 S. mately 500 nM gp61, corresponding to a T4 primosome complex containing Fluorescence anisotropy experiments in which gp61 was gp41 and gp61 subunits (and DNA constructs) in a 6∶1∶1 molar ratio. titrated into a solution containing premixed DNA-gp41-GTPγS μ E formed gp41-GTPγS-DNA complexes, monitored by fluores- solutions at a total gp41 subunit concentration of 3 M (Fig. 2 ) cence anisotropy (Fig. 2E), confirmed this result. Given that gp61 were also performed. The addition of gp61 resulted in a steep increase in anisotropy and a clear breakpoint at approximately can interact with both gp41 and DNA, it was necessary to deter- 500 nM, again confirming a 6∶1 association stoichiometry for mine the subunit stoichiometry at which these components inter- the T4 DNA-primosome complex. In contrast, the addition of act to form a stable and active T4 helicase-primase (primosome) gp41 to a preformed complex of gp61 and DNA under the same complex. solution conditions showed an initial drop in anisotropy and then As summarized above, two sets of experiments in the T4 DNA an increase, again suggesting that undefined and concentration- replication system literature have previously dealt with the sub- dependent aggregation occurs under these experimental condi- unit stoichiometry of these complexes and came to different con- tions (Fig. S4B). We conclude that the assembly pathway clusions. Using physical methods, Dong, et al. (7) showed that the described here and a 6∶1 gp41∶gp61 subunit ratio lead to a phy- gp41∶gp61 subunit ratio in these complexes was approximately 6∶1 5∶1 sically stable and discrete T4 primosome-DNA complex. , consistent with an earlier approximately subunit ratio To further test this conclusion we also performed hydrody- measured in functional studies by Richardson and Nossal (5). namic modeling calculations (15, 16) in which the complexes were – In contrast, Benkovic and coworkers (10 12) interpreted their represented by constellations of appropriately sized spherical scanning calorimetry, fluorescence anisotropy, and electron beads in varying geometries, and values of s20;w were calculated microscopy studies on T4 helicase-primase complexes to suggest and compared with our experimental measurements of s20;w. that the relevant subunit ratio might be 6∶6. We note that these The calculations are summarized in Fig. S7 and show that, while workers formed their primase-DNA complexes by direct mixing these hydrodynamic calculations cannot help us to discriminate and that their fluorescence anisotropy titrations (figure 1 in ref. between reasonable primosome structures containing from one 11) did not reach stable plateaus. Rather, their published titration curves closely resemble those of Fig. S3D, which we have shown ‡ to reflect the formation of undefined DNA-primase aggregates. The estimated standard error of the determination of these sedimentation coefficients between different runs is approximately 0.6 S, and the standard error for the position Further assembly experiments were performed to test these of a peak on an individual c(s) versus s20;w plot is (by definition) 0.1 S, because that is the contrasting stoichiometry conclusions and define the correct sub- spacing at which we took points in the experiments.

13598 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1210040109 Jose et al. Downloaded by guest on September 24, 2021 to three primase subunits per gp41 hexamer, complexes contain- Discussion ing six primase subunits per helicase hexamer are clearly ruled Overview. The following results relevant to the structure, function, out by this approach as well. subunit stoichiometry and assembly pathway of the T4 primosome complex have been established or confirmed. (i) The T4 Functional Unwinding and Priming Assays Confirm an Approximately primosome comprises six helicase (gp41) and one primase (gp61) 6∶1 gp41∶gp61 Subunit Stoichiometry for the T4 Primosome. Finally, subunits, additional primase subunits are not bound, and a pri- we ask whether complexes assembled according to the above pro- mosome complex with this 6∶1 subunit stoichiometry is stable tocols and subunit stoichiometries are also optimal in functional and fully active as both a helicase and a primase. (ii) The helicase assays. Previous experiments in which Richardson and Nossal (5) hexamer of the primosome assembles from free gp41 monomers examined the effects of primase on the helicase activity demon- via dimer and tetramer intermediates in the presence of hydro- strated that the T4 primosome unwound a helicase assay sub- lyzable (GTP) or non-hydrolyzable (GTPγS) ribonucleotide tri- iii strate construct (formed from M13mp2 ssDNA bound to a 51-bp phosphates. ( ) Primosome assembly is an equilibrium process complementary strand with a 15-nt long noncomplementary and intermediates can be detected by ultracentrifugation in the 3′ tail) much more efficiently than did gp41 helicase hexamer presence of limited concentrations of helicase or primase subunits or GTPγS. (iv) An NTP-saturated primosome complex with a alone. Furthermore, these experiments showed that helicase 6∶1 helicase∶primase subunit composition binds stably to ssDNA unwinding activity plotted as a function of input gp61 reached ∶ or to a DNA replication fork, while an NTP-saturated gp41 hex- a plateau at a helicase primase subunit ratio of approximately americ helicase alone binds much more weakly. 6∶1 (5). These results are included in Fig. 3 to compare with We have defined a primosome assembly (order-of-addition) new experiments that we carried out to examine the physiological pathway that leads to a stable equilibrium complex with model priming activity of the T4 primosome complex, using the RNA DNA replication forks. This pathway is shown in Fig. 4. We have priming assay described in Materials and Methods and SI Text. also shown that the use of other assembly protocols, such as As previously shown (17), the addition of gp61 to gp41, adding gp61 to a forked DNA construct in the absence of stable together with DNA substrates containing suitable priming tem- gp41 hexamers, leads to aggregation and metastable complexes. plate sequences, results in the synthesis of pentameric RNA pri- We suggest that this may explain some earlier fluorescence ani- mers on the template DNA. Here (Fig. 3), we demonstrate that sotropy and titration calorimetry data that seemed to suggest heli- the titration of gp61 subunits into a primase assay system contain- case-primase subunit stoichiometries approaching 6∶6 (10–12). ing a fixed concentration of gp41 helicase hexamers, ssDNA and Clearly even in equilibrium situations one cannot assume that a excess concentrations of all four canonical NTPs resulted in a stable and functional macromolecular complex has been formed progressive increase in the rate of formation of pentameric just because the components are mixed in correct proportions at RNA primers, and that this primase activity reached a maximum concentrations above the relevant dissociation constants; rather it must be demonstrated directly that the mixing protocol does not value at an input primase to helicase subunit ratio of 1 to 1.5 gp61 result in the formation of undefined complexes that are trapped subunits per helicase hexamer (grey zone in Fig. 3). The addition in metastable states. In vivo this pathway problem is probably of more primase subunits resulted in a progressive decrease in the avoided by specific or nonspecific chaperones that help macro- rate of primer formation, showing that the addition of primase molecular complexes attain conformational and compositional beyond an approximately 6∶1 helicase∶primase subunit ratio in- equilibrium. hibits RNA primer synthesis. At a 6∶6 molar ratio of gp41∶gp61 subunits this inhibition was essentially complete (Fig. 3). We Significance of a 6∶1 Helicase∶Primase Subunit Stoichiometry. The suggest that this inhibition may reflect the occlusion of priming operation of the primosome with a single primase subunit puts sequences as the excess gp61 subunits form metastable ssDNA- important constraints on mechanistic models for helicase and primase aggregates with the ssDNA template. primase functions, both for T4 and for the replication complexes of higher organisms that seem also to function with significantly fewer primase than helicase subunits. The primase monomer is smaller than a helicase subunit, making it unlikely that more than one or two helicase subunits can be in contact with the primase subunit at any one time. Our demonstration here that the primase subunit plays a central role in binding the primosome to the replication fork suggests, for any model in which all the subunits of the gp41 hexamer play active roles, that the primase is likely to be

repositioned on the helicase hexamer at every iteration of the BIOPHYSICS AND helicase cycle that moves the primosome complex progressively into the replication fork. COMPUTATIONAL BIOLOGY

A Simple Mechanistic Model for the T4 Primosome Helicase. A con- Fig. 3. The primase and helicase activities of the T4 primosome are optimal at an approximately 6∶1 gp41∶gp61 subunit ratio. RNA primer synthesis ac- ceptually simple (and speculative) model for primosome function tivity (filled circles) measures the amount of ½α32PrCTP incorporated into RNA at the replication fork, which is based on the literature and the primer pentamers after a reaction time of 20 min (see Materials and Meth- issues of subunit stoichiometry and interaction that we have ods). Increasing concentrations of gp61 subunits were titrated into a reaction defined here and is also compatible with what we know about mix containing a fixed concentration of gp41 helicase hexamers. The priming hexameric replication helicases that react with limited numbers activity reached a maximum at a gp61 concentration ratio somewhat in ex- of primase subunits in higher organisms, is outlined in Fig. S8. cess of one primase subunit per gp41 helicase hexamer and then decreased It has the following features. (i) The primosome is positioned progressively as more gp61 was added (see text). The helicase data were at the replication fork with the lagging DNA template strand pas- taken from Fig. 4 of Richardson and Nossal (5), and show that the unwinding sing through the center of the helicase hexamer and the leading activity of either 50 pmols (open squares) or 100 pmols (open circles) of T4 gp41 hexamer on a helicase substrate consisting of a 3-tailed ssDNA M13 template strand bound to the outside of approximately two gp41 template complexed with a 51 bp duplex construct reached a plateau value subunits. This is compatible with a ssDNA binding site size for the at a gp61 subunit concentration ratio of approximately one gp61 subunit per T4 helicase of approximately 20 nts (18) and is also consistent gp41 hexamer. with the binding site size that has been demonstrated for the

Jose et al. PNAS ∣ August 21, 2012 ∣ vol. 109 ∣ no. 34 ∣ 13599 Downloaded by guest on September 24, 2021 nd tra g s gin Lag 3’ Premixed 5’ L helicase-primase Forked DNA ead ings on complex tr ti and a reg Ag Primase gp61 g g g reg A ati on

GTP or ATP Helicase gp41

d tran g s nd in tra gg g s La gin Lag 3’ 3’

5’ 5’ Le L adin ead + gs ings tran tr d and DNA.(gp41)6

Primase gp61 d ran st ing gg La 3’ 5’ Lea dings tr and

DNA.(gp41)6.(gp61)

Fig. 4. Assembly pathway for the T4 DNA replication primosome. The primosome components are: gp41 subunits (blue ovals), gp61 subunit (red ellipse), GTP or GTPγS (yellow rectangles), and GDP (violet rectangle). The gp61 primase subunits are shown as elongated ellipsoids to reflect their higher frictional ratios compared with the relatively globular helicase subunits. The potential assembly pathways on the Right or Left of the figure lead to the formation of metastable aggregates, while the central pathway leads to a well-defined equilibrium complex with 6∶1 helicase to primase subunit stoichiometry, as shown by the included titration plots. For further details see text.

closely related hexameric dnaB helicase of E. coli (19). (ii) The by one gp41 subunit position. The primase subunit could achieve primase subunit binds to two ssDNA-bound gp41 subunits at the this reorientation by dissociating and rebinding, or it could re- replication fork (20), bringing it into effective contact with the main bound to one of the ssDNA strands and/or gp41 subunits NTP ligand that lies between these subunits. This primase binding and “pivot” into its new position without dissociation. proximity could then specifically “activate” the hydrolysis of This model is consistent with a step-wise and sequential hydro- the NTP located at this position, while not perturbing the NTP lysis of the primosome-bound NTPs around the helicase ring. We ligands that lie between and stabilize other gp41 subunit inter- emphasize that this mechanism requires a “mobile” primase and faces and are not directly adjacent to the primase subunit.§ The is thus fundamentally different from one that might apply to a T7 primase subunit likely binds leading strand ssDNA at the fork and DNA helicase-primase complex in which each subunit of the T7 functional arguments show that in the complete replication com- helicase-primase complex carries both a helicase and a primase plex the primase must also lie sufficiently close to the lagging domain. In a recent single molecule study Sun et al. (22) have also strand to permit it to use this ssDNA as templating sequences proposed a mechanism for the T7 helicase in which the NTPs for RNA synthesis. This primer synthesis could take place as bound at the interface fire sequentially and “cooperatively” as the the separated ssDNA of the lagging strand enters the central hole helicase moves into the replication fork. We note that the model in the helicase hexamer. (iii) The activated NTP of the primo- outlined above is basically a Brownian ratchet, with the primase playing the role of the ratchet. Other proposed mechanisms for some helicase then hydrolyzes to NDP and Pi, transiently releas- ing the primase subunit from the lagging strand ssDNA and also hexameric helicases, developed largely on the basis of crystal specifically weakening the interaction of the approximately 2 structures, are conceptually closer to power-stroke models (23, gp41 subunits with the fork junction. This could permit the heli- 24). Elsewhere (DJ, SEW and PHvH; manuscript under review) case hexamer to rotate by one subunit relative to the fork and also we have further developed and refined the molecular details of to “move into” the dsDNA of the fork by one nucleotide position this mechanistic model on the basis of near UV spectral measure- as a consequence of local fork “breathing” (21), followed by ments with site-specifically placed DNA analogue base probes the rebinding of the next pair of gp41 subunits at the replication that can directly monitor dsDNA base pair opening and closing fork and the concomitant rebinding of the primase to the fork events and the stacking and unstacking of vicinal bases within the DNA, but now translocated around the hexamer helicase ring replication fork as the T4 primosome helicase advances. Materials and Methods §We note that a similar NTP-activation mechanism has recently been described for the Materials. Unless stated otherwise, all experiments were performed at clamp-loading complex of the T4 replication system (25, 26). 20 °C in buffer containing 20 mM HEPES (pH 7.5), 150 mM K(OAc), 10 mM

13600 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1210040109 Jose et al. Downloaded by guest on September 24, 2021 MgðOAcÞ2, 0.1 mM EDTA and 1 mM DTT. Purine nucleoside triphosphates and Analytical Ultracentrifugation. Sedimentation velocity experiments were per- their non-hydrolysable analogues were purchased from Sigma-Aldrich. formed in a Beckman Optima XL-I Analytical Ultracentrifuge. Protein and nu- [γ-32P] GTP and [γ-32P] ATP were obtained from NEN (Boston, MA). Unlabeled cleic acid samples were dialyzed extensively against reference buffer [20 mM and 2-AP labeled DNA oligonucleotides were purchased from Integrated HEPES (pH 7.5), 150 mM K(OAc), 10 mM MgðOAcÞ2] prior to centrifugation DNA Technologies (Coralville, IA). DNA oligonucleotides 5′-labeled with and protein and nucleic acid concentrations were re-measured after dialysis. Oregon Green 488 were from Operon (Huntsville, AL). Oligonucleotide con- Sedimentation runs were performed and analyzed by standard techniques; centrations were determined by absorbance at 260 nm using extinction coef- see SI Text and Fig. S1. ficients furnished by the manufacturer. Forked DNA constructs were annealed by heating equimolar concentrations of DNA primer and template Steady State Fluorescence and Anisotropy Experiments. Fluorescence experi- strands at 90 °C for 5 min, followed by gradual (2 h) cooling to 25 °C. The DNA ments were performed with a Jobin-Yvon Fluorolog or Fluoromax spectro- constructs used in this study are shown in Table S1. The T4-coded DNA repli- fluorimeter. The excitation wavelength for Oregon Green labeled samples cation helicase (gp41) and replication primase (gp61) were prepared and con- was 494 nm and emission was monitored at 520 nm. When 2-AP served as centrations determined as previously described (see 6, 20 and SI Text). the fluorescent probe the excitation wavelength was 315 nm and emission spectra were recorded from 330 to 400 nm. Binding curves were fit to Eq. 1: Primase Assay. Primase assays were performed in a reaction buffer containing μ 33 mM Tris-OAc, 125 mM KOAc, 6 mM MgðOAcÞ2, 0.5 mM DTT and 200 M ΔA ΔA ∕2D E D K each of all four canonical rNTPs, in addition to 3.5 mM ATP and 100 nM ¼ T T fð T þ T þ dÞ 32 ½α- PrCTP. Unless otherwise indicated, the reaction mixtures contained 2 1∕2 − ½ðET þ DT þ KdÞ − 4ET DT g; [1] 800 nM gp41 and 50–900 nM gp61 (concentrations as protein subunits) and 100 nM of ssDNA construct (Table S1) in a 10-μl reaction volume. Reaction mixtures were assembled on ice and preincubated at 37 °C for 2 min. Reac- where ΔA is the change in anisotropy at each point in the titration, ΔAT is the tions were started by the addition of the proteins components, incubated at total anisotropy change, DT is the total DNA concentration, ET is the total 37 °C for 30 min and then quenched by adding 10 μl of 95% deionized for- gp61 concentration and Kd is the dissociation constant. mamide containing 20 mM EDTA, 0.25% xylene cyanol and 0.25% bromo- phenol blue. The quenched reactions were subjected to electrophoresis on ACKNOWLEDGMENTS. We are grateful to our laboratory colleagues (especially 22% polyacrylamide 8 M urea gels (19∶1 acrylamide∶bis-acrylamide) at Walt Baase and Sandra Grieve) for many helpful discussions of this work. 40 watts for 90 min and the dried gels were quantitated using an AMBIS These studies were supported by NIH Grant GM-15792 to P.H.v.H., who is also scanner. an American Cancer Society Research Professor of Chemistry.

1. Alberts BM (1987) Prokaryotic DNA replication mechanisms. Philos Trans R Soc Lond B 14. Erickson HP (2009) Size and shape of protein molecules at the nanometer level deter- Biol Sci 317:395–420. mined by sedimentation, gel filtration, and electron microscopy. Biol Proced Online 2. Nossal NG (1994) Molecular Biology of Bacteriophage T4, ed JD Karam (American 11:32–51. Society for Microbiology, Washington, DC), editor-in-chief. 15. Bloomfield V, Dalton WO, Van Holde KE (1967) Frictional coefficients of multisubunit 3. Mueser TC, Hinerman JM, Devos JM, Boyer RA, Williams KJ (2010) Structural analysis of structures. I. Theory. Biopolymers 5:135–148. bacteriophage T4 DNA replication: A review in the Virology Journal series on bacter- 16. Carrasco B, Garcia de la Torre J (1999) Hydrodynamic properties of rigid particles: Com- iophage T4 and its relatives. Virol J 7:359. parison of different modeling and computational procedures. Biophys J 76:3044–3057. ’ 4. Johnson A, O Donnell M (2005) Cellular DNA replicases: Components and dynamics at 17. Liu CC, Alberts BM (1980) Pentaribonucleotides of mixed sequence are synthesized – the replication fork. Annu Rev Biochem 74:283 315. and efficiently prime de novo DNA chain starts in the T4 bacteriophage DNA replica- 5. Richardson RW, Nossal NG (1989) Characterization of the bacteriophage T4 gene 41 tion system. Proc Natl Acad Sci USA 77:5698–5702. – DNA helicase. J Biol Chem 264:4725 4731. 18. Young MC, Kuhl SB, von Hippel PH (1994) Kinetic theory of ATP-driven translocases on 6. Dong F, Gogol EP, von Hippel PH (1995) The phage T4-coded DNA replication helicase one-dimensional polymer lattices. J Mol Biol 235:1436–1446. (gp41) forms a hexamer upon activation by nucleoside triphosphate. J Biol Chem 19. Bujalowski W, Jezewska MJ (1995) Interactions of Escherichia coli primary replicative 270:7462–7473. helicase DnaB protein with single-stranded DNA. The nucleic acid does not wrap 7. Dong F, von Hippel PH (1996) The ATP-activated hexameric helicase of bacteriophage around the protein hexamer. Biochemistry 34:8513–8519. T4 (gp41) forms a stable primosome with a single subunit of T4-coded primase (gp61). 20. Jing DH, Dong F, Latham GJ, von Hippel PH (1999) Interactions of bacteriophage T4- J Biol Chem 271:19625–19631. coded primase (gp61) with the T4 replication helicase (gp41) and DNA in primosome 8. Sinha NK, Morris CF, Alberts BM (1980) Efficient in vitro replication of double-stranded formation. J Biol Chem 274:27287–27298. DNA templates by a purified T4 bacteriophage replication system. J Biol Chem 21. Jose D, Datta K, Johnson NP, von Hippel PH (2009) Spectroscopic studies of position- 255:4290–4293. specific DNA ‘breathing’ fluctuations at replication forks and primer-template junc- 9. Dong F, Weitzel SE, von Hippel PH (1996) A coupled complex of T4 DNA replication – helicase (gp41) and polymerase (gp43) can perform rapid and processive DNA strand- tions. Proc Natl Acad Sci USA 106:4231 4236. displacement synthesis. Proc Natl Acad Sci USA 93:14456–14461. 22. Sun B, et al. (2011) ATP-induced helicase slippage reveals highly coordinated subunits. – 10. Valentine AM, Ishmael FT, Shier VK, Benkovic SJ (2001) A zinc ribbon protein in DNA Nature 478:132 135. replication: Primer synthesis and macromolecular interactions by the bacteriophage 23. Enemark EJ, Joshua-Tor L (2006) Mechanism of DNA translocation in a replicative hex- – T4 primase. Biochemistry 40:15074–15085. americ helicase. Nature 442:270 275. 11. Yang J, Xi J, Zhuang Z, Benkovic SJ (2005) The oligomeric T4 primase is the functional 24. Thomsen ND, Berger JM (2009) Running in reverse: The structural basis for transloca- form during replication. J Biol Chem 280:25416–25423. tion polarity in hexameric helicases. Cell 139:523–534. 12. Norcum MT, et al. (2005) Architecture of the bacteriophage T4 primosome: Electron 25. Pietroni P, von Hippel PH (2008) Multiple ATP binding is required to stabilize the microscopy studies of helicase (gp41) and primase (gp61). Proc Natl Acad Sci USA ‘activated’ (clamp open) clamp loader of the T4 DNA replication complex. J Biol Chem 102:3623–3626. 283:28338–28353. BIOPHYSICS AND 13. Tabor S, Richardson CC (1981) Template recognition sequence for RNA primer synthesis 26. Kelch BA, Makino DL, O’Donnell M, Kuriyan J (2011) How a DNA polymerase clamp

by gene 4 protein of bacteriophage T7. Proc Natl Acad Sci USA 78:205–209. loader opens a sliding clamp. Science 334:1675–1680. COMPUTATIONAL BIOLOGY

Jose et al. PNAS ∣ August 21, 2012 ∣ vol. 109 ∣ no. 34 ∣ 13601 Downloaded by guest on September 24, 2021