Compression of the DNA substrate by a viral packaging motor is supported by removal of intercalating dye during translocation

Abstract

Viral genome packaging into capsids is powered by high-force-generating motor proteins. In the presence of all packaging components, ATP-powered translocation in vitro expels all detectable tightly bound YOYO-1 dye from packaged short dsDNA substrates and removes all aminoacridine dye from packaged genomic DNA in vivo. In contrast, in the absence of packaging, the purified T4 packaging ATPase alone can only remove up to ∼1/3 of DNA-bound intercalating YOYO-1 dye molecules in the presence of ATP or ATP-γ-S. In sufficient concentration, intercalating dyes arrest packaging, but rare terminase mutations confer resistance. These distant mutations are highly interdependent in acquiring function and resistance and likely mark motor contact points with the translocating DNA. In stalled Y-DNAs, FRET has shown a decrease in distance from the phage T4 terminase C terminus to portal consistent with a linear motor, and in the Y-stem DNA compression between closely positioned dye pairs. Taken together with prior FRET studies of conformational changes in stalled Y-DNAs, removal of intercalating compounds by the packaging motor demonstrates conformational change in DNA during normal translocation at low packaging resistance and supports a proposed linear “DNA crunching” or torsional compression motor mechanism involving a transient grip-and-release structural change in B form DNA.

Nucleic acid translocation into an empty procapsid is a conserved capsid assembly mechanism found among diverse DNA and RNA viruses (1). High-resolution structures of all of the conserved motor components found among tailed dsDNA bacteriophages have been determined (2–5). The small bacteriophage T4 terminase protein gp16 is required for cutting and packaging the replicative DNA concatemer in vivo but is nonessential and inhibitory for linear DNA packaging in vitro (6). Thus, T4 DNA translocation in vitro can be driven by a two-component motor consisting of a prohead portal ring channel dodecamer situated at a single packaging vertex that is docked during packaging with gp17 terminase-ATPase. A linear packaging motor mechanism proposed that a terminase to portal DNA grip-and-release driven by a motor protein conformational change drives DNA into the prohead by a DNA compression motor stroke (7–10).

It has long been known that acridine dyes inhibit bacteriophage development at concentrations below those that inhibit growth of the bacterial host. Phage mutations that confer resistance to such dyes arise through different mechanisms. Among these are mutations in the phage T4 terminase gene 17, called ac and q, that confer acridine and quinacrine resistance (11). All 9-aminoacridine (9AA) treatments have been shown by electron microscopy to arrest DNA packaging in vivo; however, whether the site of action of 9AA is the DNA, the active site of the packaging enzyme, or the enzyme DNA complex—in fact, acridines are known to inactivate protease active sites or virion proteins associated with DNA (reviewed by Black et al., in ref. 12)—is unknown.

Insertion of intercalating dyes (IDs) between dsDNA base pairs has been intensively studied and is known to cause stretching and unwinding of DNA. Classic IDs are plane aromatic molecules such as 9AA and ethidium bromide (EthBr). Dimeric ID YOYO-1 is widely used because of its over 1,000-fold higher affinity to dsDNA than EthBr. It binds virtually irreversibly to the DNA, and the fluorescence quantum yield of the bound dye is more than 1,000-fold higher than when free in solution. As a result, the background fluorescence from free dye is extremely low, which makes these dyes excellent probes for high-sensitivity quantification of DNA and for imaging of individual DNA molecules (13). At low dye–base pair ratios, the binding mode appears to consist primarily of bisintercalation. Each monomer unit intercalates between adjacent bases, with the benzazolium ring system sandwiched between the pyrimidines and the quinolinium ring between the purine rings, causing the helix to unwind as studied by NMR. Important structural parameters such as the binding site size, the elongation, as well as the untwisting angle per bound YOYO-1 molecule have been reported (14). In a recent study, YOYO-1 dye was used to monitor the helicase activity of the Bacteriodes fragilis AddAB enzyme through its displacement from the unwound dsDNA (15).

We are able to measure DNA packaging at high efficiency (∼20–100%) in vitro with short DNAs ranging from 70 bp to 5 kb, as well as with full-length ∼170 kb genome size DNAs that can produce infectious virions. We quantify packaging in vitro by independent nuclease, Typhoon imager, and fluorescence correlation spectroscopy (FCS) assays. We report here that the purified bacteriophage T4 packaging ATPase reduces ID binding to DNA in the absence of ATP hydrolysis and proheads. Unexpectedly, we find that DNA translocation expels all detectable YOYO-1 from the packaged DNA substrate. We observe similar inhibitory effects in vitro of several IDs on packaging of DNAs of different low molecular weights as well as full-length genomic DNAs in vivo. Thus, we expect that ID inhibition should reflect effects on local DNA structure. Moreover, we find that terminase mutations confer comparable resistance to the different IDs 9AA and EthBr in vivo and 9AA, EthBr, and YOYO-1 in vitro. By mutagenic treatment and site-directed mutagenesis (SDM) we show complex interdependence of three rare mutational sites within the terminase to confer resistance to IDs and to maintain terminase function in the absence of IDs. These results are consistent with the structure and proposed mechanism of the DNA packaging motor.

Results

T4 Terminase Acridine-Resistant Mutants Resist ID Inhibition of Growth and of DNA Packaging.

Previously IDs like 9AA and EthBr have been shown to inhibit bacteriophage development at concentrations below those that inhibit the bacterial host. Two 9AA resistant mutants, ac and acq, the latter a double mutation conferring additional resistance to the intercalating compound quinacrine, were isolated after intensive selection and screening and were found to be located in the terminase gene 17 by genetic mapping (11). The locations of the two terminase mutations in the 17 gene were determined to be ac-A96D and q-F249V. The mutations enhanced T4 phage growth by a factor of 10⁷ (Fig. 1A). When examined for growth inhibition by different IDs, 9AA and EthBr, acq showed greater resistance in comparison with the ac mutant (Fig. 1B). When DNA packaging was determined by growth in the presence of 9AA (3 µg/mL), packaging was inhibited in wild-type (wt) T4 phage, where consistently we saw only partial (∼10 kb) packaging of the full ∼170 kb genomic DNA. The ac mutant produced mostly full genomic T4 DNA along with a lesser amount of the 10 kb DNA, whereas the acq mutant was found to be resistant to the inhibition of DNA packaging and no 10 kb DNA was seen (Fig. 1B). Interestingly, glycerol gradient purified partially or completely filled heads from these infections in the presence of 9AA did not show any 9AA fluorescence by Typhoon imager analysis of the intact and partially filled heads on an agarose gel (Fig. 1C).

Fig. 1.

T4 terminase acridine-resistant mutants are more resistant to inhibition of growth as well as packaging of DNA by IDs. (A) Efficiency of plating of wt and mutant T4 phages with and without IDs (9AA and EthBr). Pulse field gel electrophoresis (PFGE) (B) and agarose (C) gels showing different-length DNAs packaged in the presence of 9AA in T4 wt and acridine-resistant ac and acq mutants. Typhoon image of the agarose gel before (C, ii) and after EthBr staining (C, i) showing no DNA-bound 9AA in the glycerol gradient-purified proheads.

IDs Inhibit in Vitro DNA Packaging and Acridine-Resistant Mutant Terminases Are More Resistant to in Vitro DNA Packaging Inhibition Compared with the wt Terminase.

In this study we have used YOYO-1 dye, which has higher DNA affinity and higher fluorescence quantum yield than EthBr homodimer, 9AA, DAPI, and Hoechst dyes and only fluoresces after binding to DNA. All three IDs (9AA, EthBr, and YOYO-1) were inhibitory for in vitro DNA packaging at a concentration of ∼1 µM (Fig. 2A). At 2.5 μM concentration of 9AA or YOYO-1 wt T4 packaging was 100% inhibited, whereas the acq mutant was only about 50% inhibited (Fig. 2D). Nuclease assays in the presence (+) and absence (–) of YOYO-1 with equal amounts of purified wt, ac, and acq terminases (Fig. 2B) showed that the acq terminase is more resistant to the inhibition than the wt and the ac mutant (Fig. 2 C and D). YOYO-1 and 9AA dyes inhibited packaging in the presence of wt and acq mutant terminases in a similar manner as shown in Fig. 2D.

Fig. 2.

IDs inhibit in vitro DNA packaging. (A) Nuclease assays showing inhibition of packaging in the presence of YOYO-1, 9AA, and EthBr dyes (1 μM). (B) SDS/PAGE gel image showing purified ac, acq, and wt terminases. (C) The acq terminase mutant more resistant to packaging inhibition by 9AA (500 nM) compared with ac and wt. (D) YOYO1 and 9AA dyes inhibit DNA packaging in a similar manner, and acq is more resistant to inhibition than the wt terminase.

Terminase Releases IDs from DNA in the Absence of Translocation.

Upon electrophoresis, YOYO-1–bound DNAs ran higher on an agarose gel compared with the unbound DNA, suggesting a change in the conformation and/or molecular weight of the DNA–dye complex (Fig. 3A). When the YOYO-1–bound DNAs (70 bp, 280 bp, and 5 kb) were incubated with wt terminase at 37 °C for 30 min, an increase in DNA mobility was observed by agarose gel electrophoresis analyzed on Typhoon imager and EthBr staining (Fig. 3). Similar results were observed in FCS experiments, when the three different lengths of YOYO-1–bound DNAs (70 bp, 280 bp, and 5 kb) were incubated with terminase. The autocorrelation of the YOYO-1–bound 70 bp, 280 bp, and 5 kb DNAs with and without terminase could be fitted by a single-species diffusion model, consistent with a single fluorescent species. The diffusion coefficients of the YOYO-1–bound 70 bp, 280 bp, and 5 kb DNAs without terminase were around 70 μm²/s, 20 μm²/s, and 2 μm²/s and with terminase were 85 μm²/s, 30 μm²/s, and 3 μm²/s, respectively (Table 1). A fluorescence intensity decrease of ∼5 counts/ms was observed in the presence of terminase for YOYO-1–bound 70 bp DNA (Fig. 3B), and similar fluorescence intensity decrease was also observed for YOYO-1–bound 280 bp and 5 kb DNAs (Fig. S1). A decrease in fluorescence intensity was also seen with YOYO-1–bound 5 kb DNA in the presence of terminase by agarose gel Typhoon image analysis of fluorescence (Fig. 3C). The dye removal by the terminase alone is in the range of 18–35% as calculated from decrease in the photon counts for our FCS experiments (Table S1). This analysis shows that there is an increase in the diffusion coefficients and decrease in fluorescence intensities of the YOYO-1–bound DNAs in the presence of terminase, clearly suggesting that terminase can release some amount of ID YOYO-1 from DNA. Fig. 3C shows partial removal of YOYO-1 by terminase alone, which is enhanced in the presence of ATP or ATP-γ-S. Analysis of gp17/DNA–dye binding in the presence or absence of ATP (Fig. S2) shows that when gp17 binding to the DNA reaches saturation, only partial dye release is observed, although dye release is enhanced by ATP or ATP-γ-S even in the absence of translocation into the proheads.

Fig. 3.

Terminase releases IDs from DNA. (A) Gel mobility shift assays showing the shift in YOYO-1–bound (500 nM) 70 bp, 280 bp, and 5 kb DNAs in the presence of terminase. (B) Decrease in fluorescence intensity of the 70 bp DNA–YOYO-1 complex in the presence of terminase observed in the FCS measurements. (C) Mobility shift and reduction in fluorescence intensity of YOYO-1–bound 5 kb DNA by terminase in the presence or absence of ATP or γ-S-ATP in the reaction mixture (C, i—Typhoon image) but no change in DNA concentration (C, ii—EthBr stained gel).

Table 1.

FCS diffusion coefficients of the DNA–dye complexes

Sample with 500 nM YOYO-1	Diffusion coefficients (D) (μm2/s)
70 bp DNA	70 ± 6
70 bp DNA + terminase alone or with full packaging mix	85 ± 8
280 bp DNA	20 ± 1
280 bp DNA + terminase alone or with full packaging mix	30 ± 2
5 kb DNA	2 ± 0.1
5 kb DNA + terminase alone	3 ± 0.2

Values of the DNA–dye complexes alone, with terminase, or following packaging in the full packaging mix are given as average ± SE as previously determined (refs. 8 and 10).

DNA Translocation Removes IDs from the DNA–Dye Substrate Complex.

When 5 kb DNA equilibrated with different concentrations of YOYO-1 (250 nM, 500 nM, and 1 μM) was packaged in vitro, as expected, less packaging of DNA was observed with increasing concentrations of dye, but unexpectedly, no fluorescence was observed in the packaged DNA. This was determined by Typhoon images of intact proheads in an agarose gel before and after EthBr staining (Fig. 4A), suggesting that the DNA packaged into the prohead does not have any bound YOYO-1 dye. Some background fluorescence was seen by Typhoon image analysis when proheads alone were incubated with 1 µM YOYO-1, even though there is no DNA in the proheads as seen in the nuclease assay gels (Fig. 4A). To verify this result, packaging of smaller DNAs (70 bp and 280 bp) was carried out and checked by nuclease assays as well as by FCS measurements (Fig. 4 B and C). The FCS autocorrelation curves of the YOYO-1–bound 70 bp and 280 bp DNA substrates in a negative reaction control lacking proheads could be fitted by a single-species diffusion model, consistent with single fluorescent species with diffusion coefficients of about 85 μm²/s and 30 μm²/s, respectively (Fig. 3B and Table 1). The FCS correlations obtained in the full packaging reactions were apparently identical and showed no prohead-like diffusion, as reported earlier with a slow species of diffusion coefficient of ∼1–3 μm²/s (10), and could be fitted to the same fast diffusing species as shown in our previous studies of covalently dye-labeled dsDNAs of this size range (Fig. 4C). The analysis showed further that even though efficient packaging of the DNAs was seen by nuclease assay gels (Fig. 4B), no prohead-like diffusion of the packaged DNA was observed in the FCS assays (Fig. 4C), thus confirming the Typhoon results showing that the terminase in the presence of the complete packaging mix completely removes the YOYO-1 intercalated dye from the complex and the DNA without dye is packaged into the proheads. Quenching of YOYO-1 in the DNA packaged proheads is inconsistent with the observation that YOYO-1 can slowly diffuse into the proheads and bind to the packaged DNA judged by the Typhoon gel analysis (Fig. S3). The wt and acq terminases alone were comparable in releasing dye from the YOYO-1 DNA complex (Fig. 4 D, i), whereas in the full packaging mix acq terminase released significantly more dye (Fig. 4 D, ii). The mutations in the ac and acq terminases may result in tighter binding to the translocating dsDNA, which in some way releases more dye from the dye–DNA complex compared with the wt terminase during packaging.

Fig. 4.

DNA is packaged after removal of ID from the DNA–dye complex. (A) Packaging of 5 kb DNA in the presence of different concentrations of YOYO-1. (A, i) Typhoon image, (A, ii) same gel stained with EthBr, and (A, iii) nuclease assay gel. (B) Nuclease assay gels showing packaging of the 70 bp and 280 bp DNAs in the presence of YOYO-1 (500 nM). (C) FCS data showing fitting to a single fast diffusing component (DNA–dye complex) and no prohead-like diffusion of the YOYO-1–bound 70 bp and 280 bp DNA in the full packaging mixture (Pmix). (D, i) Typhoon and EthBr-stained gel images showing dye release by wt and acq terminases alone. (D, ii) EthBr-stained agarose gel showing packaging of 5 kb DNA in the absence and presence of 1 μM YOYO-1 by wt and acq terminases.

ac, q, and N Sites Coordinate with Each Other and May Be Near Terminase DNA Binding Sites.

We determined that ac and q each confer 9AA resistance as single mutations, but more effectively when combined, and by mutagenesis that such mutational sites in the terminase are rare. In fact, we were unable to isolate a single additional 9AA resistance mutation comparable to the ac and q mutations (ac and q are transversion mutations) at other sites in gene 17 following heavy hydroxylamine mutagenesis of plasmid gene 17 followed by strong selection for recombinant mutant genomes. Also by screening for spontaneous mutations, we found only identical substitutions at the same sites. However, mutagenesis identified a third interacting site, N267, that modified ID resistance at the ac and q sites. Introduction of single and multiple amber codons at the ac, q, and N sites by SDM allowed us to determine the viability and 9AA resistance of 13 different single and double amino acid replacements at these three sites. This analysis shows that the three sites are highly cooperative and interdependent both in allowing terminase function itself in the absence of IDs and also in producing ID resistance; for example, we were able to introduce an amber codon at N only following a mutation at q that changed F to L as shown by the arrow in Fig. 5A. Although this substitution allowed an N amber codon to be introduced, it could be grown only with Y or F substitution for W at this site, and these changes rendered the terminase ID sensitive, whereas the single q site L was resistant. Additionally, the combination L249, L96 was dye resistant, whereas the single L96 substitution was sensitive; similarly, 96Y conferred viability on the otherwise lethal 249Y substitution. Some of the interdependencies of the ac, q, and N sites for ID resistance and terminase function are summarized in Fig. 5A. Although complex, these results show that these three sites coordinately participate in the same terminase function. The location of the three sites in the gp17 crystal structure (4) is shown in Fig. 5A.

Fig. 5.

T4 terminase ID-resistant mutations and terminase-portal interacting translocation functions. (A) ac96, q249, and N267 sites linearly align with each other and with portal clip interacting residue 473 (circled in the gp17 monomer structure) and show complex interdependencies in ID resistance and terminase function. Various interdependent (shown by arrows) amino acid substitutions at residues 96(ac), 249(q), and 267(N) in the terminase leading to either ID resistance (above line), sensitivity (below line), or loss of terminase function (–) are shown together with the 3D structure. (B) Packaging translocation ts and cs mutations and intra- and intergenic suppressors of the packaging motor proteins. Terminase ts mutations are marked in black, and portal cs mutations are marked in red. Intragenic (18) suppressors are shown with blue arrows, and intergenic (18–21) suppressors are shown by red arrows. Only the portal clip region and translocation channel are shown.

Various temperature-sensitive (ts) mutations affecting DNA translocation were previously located at the C-terminal end of gp17 (reviewed in ref. 16 or newly isolated by us, in Materials and Methods). Two ts mutants showed changes at residue Y562. One is a deletion mutant, and in the other, Y is substituted by A. The ts mutation at residue Y562 of gp17 (Y deletion) is suppressed by E573K mutation in the same gene. One ts mutation, S546G, was suppressed by a distant D584N mutation. Another ts mutation, D473N, was suppressed by portal cold-sensitive (cs) mutation csM308I located in the clip interaction region (Fig. 5B). All these mutations in the C-domain of the gp17 gene affecting the DNA translocation suggest its active participation in DNA translocation in conjunction with the portal clip region (3, 10).

Discussion

The principal finding of our study is that DNA packaging expels the IDs 9AA, EthBr, and YOYO-1 from DNA in vivo and in vitro. DNA saturating levels of T4 terminase remove ∼18–35% bound YOYO-1 from the equilibrated dye–DNA complex in the absence of proheads and ATP-powered translocation. The terminase is qualitatively different with all packaging components present during translocation in eliminating all detectable YOYO-1 bound to even very short DNAs (70 bp or 280 bp) that are packaged. Resistance to packaging of short DNAs from DNA pressure within the prohead is expected to be low (17, 18). It is thus unlikely that it is the packaged conformation of the DNA that quantitatively removes the dye but rather the translocation itself; indeed, the packaged DNA in the intact prohead can be readily stained with EthBr and also YOYO-1 (Figs. 1 and 4 and Fig. S3). Removal of the IDs during translocation is at first sight surprising, as the T4 motor can efficiently package DNA labeled with dye covalently attached to the DNA ends or to the bases (19). The Φ29 packaging ATPase is also packaging tolerant of heavily modified duplex DNA (20). And although ID removal is expected and in fact shown for a helicase that separates the DNA strands, this is not expected for a packaging translocase that moves intact B form DNA from outside to inside the prohead. However, this surprising observation is consistent with previous observations including the effect of nicking short DNA substrates, which was demonstrated to disengage the DNA from the motor (7). This finding and others led to a proposal that a linear DNA grip-and-release motor mechanism transiently compresses B form DNA during translocation. In fact, previous FRET studies of stalled Y-DNAs directly supported torsional compression of the Y-stem duplex by 20–25% (8, 10). These studies supported the proposed compression motor mechanism of DNA translocation and also showed that conformational changes in both the motor proteins and the DNA substrate itself are associated with the power stroke of the packaging motor. Our present studies support such a conformational change in the DNA that is apparently found not only in stalled Y-DNA but also accompanies active translocation of linear DNA into the prohead. Compression of the B form duplex would be predicted to expel intercalating compounds from between the base pairs, as it is established that insertion of IDs between bases unwinds, stretches, and stiffens the DNA without affecting the bending rigidity (14). Indeed, it has been shown that YOYO-1 binding is very stable and is only affected by stretching forces exceeding 10 pN (14); thus, the complete removal of IDs from DNA without changing the integrity of the DNA appears most probably caused by an opposing compression of the duplex by the packaging motor that can generate forces as high as ∼60 pN. Also, structural studies of the DNA–YOYO-1 and DNA–TOTO complexes shows that this structure is more consistent with B than A form DNA (A form is 33% compressed compared with B), thus possibly accounting for comparable packaging inhibitory effects of 9AA and the more strongly binding YOYO-1 in our experiments (21, 22). Overall, our direct FRET measurements showing likely DNA compression in the stalled Y-DNA stem and these results showing ID removal during translocation by the powerful motor appear to be consistent and confirmatory (8, 10).

Mutations can lead to a terminase with enhanced ID resistance to packaging inhibition by several IDs. Our experiments suggest that acq terminase is more able to release dyes from the dye–DNA complex during packaging and hence packages more DNA compared with the wt terminase, which could be attributed to the higher binding affinity of this mutant terminase to the dsDNA. It has been found that in bacteriophage Φ29, the contact of packaging ATPase to ATP or γ-S-ATP resulted in conformational change to a higher binding affinity toward dsDNA (23). Although ATP hydrolysis is not required to release some dye from the dsDNAs, the presence of ATP or ATP-γ-S does enhance the efficiency of dye removal by the terminase, suggesting that the acq mutations might be causing a similar, possibly active translocation-like conformational change in the terminases, making them more ID resistant.

We have shown by SDM that rare dye-resistant terminase mutations affecting DNA translocation are located in the N-terminal ATPase domain. In bacteriophage lambda, two translocation defective mutants (K84A and Y46F) in the ATPase domain were also characterized (24). The T4 ID-resistant sites have been shown by SDM to have a high degree of functional coordination despite being separated by substantial 3D distances in the gp17 3D crystal structure (Fig. 5A). In view of their interdependence and the association of these three ID-resistance sites with enhanced removal of IDs from the translocated DNA, the simplest explanation is that these mutant sites are located near to points of contact made between the terminase and the translocating DNA. Likely coordinated binding of DNA near to these sites requires amino acids at these sites to be compatible with coupled association with the translocating DNA. Interestingly, although the ac, q, and N sites are not located in known terminase helicase-like and ATPase motifs, they are linearly aligned in the 3D structure, with ac and N separated by about 35 Å or one turn of the helix. Although less direct explanations of the behaviors of mutations in these sites in the terminase are possible, there is no other direct or conflicting structural evidence for terminase–DNA translocation contact points made in either the T4 or other terminases. One structure of a terminase docked portal has been proposed (4) that does not apparently fit to this alignment, but it is quite possible that other docked structures are compatible with the high-resolution gp17 crystal structure; for example, one is shown aligning the ac, q, N, and exposed clip portal channel that apparently interacts with terminase residue 473 (Fig. 5 A and B). There is ample evidence including direct FRET measurements (10) that suggest that this terminase-portal structure is highly mobile and may be different in the initiation and stalled Y-DNA complex conformations. It is tempting to speculate that the DNA duplex is gripped differently at two positions during translocation: at one point the phosphodiester backbone could be gripped by the highly basic peptide (7 of 21 residues are basic surrounding residue 96) and, at the second, by insertion or other hydrophobic interaction with the stacked bases in a region of about one helical turn away near to residues F249 and W267 (Fig. 5A). The comparable partial removal of dyes by the wt and acq mutant terminases alone from the DNA dye complex is likely by the ATP hydrolysis independent gripping and release of the DNA by these terminases. More efficient removal of dye by the acq mutant during packaging suggests tighter gripping of the dsDNA between these two regions of the terminase, which might help to explain how high force can be generated during ATP-generated terminase conformational change and how interactions at these two sites might be highly coordinated.

Several terminase ts mutations that block initiation of DNA packaging when grown at high temperature are found to be located in its N-terminal portion—here, a mutation in the hinge region of the terminase is able to suppress a portal cs (csD281E) mutation that also blocks packaging initiation (Fig. 5B) (reviewed in ref. 16). A different phenotype is associated with several terminase ts mutations near its C terminus: accumulation of partially DNA filled proheads (16). Such mutants are thus DNA translocation defective, similar to the ID effects seen on translocation. In addition to local intragenic terminase suppressors of these C-terminal ts mutations that we have recently isolated—for example, D584N suppresses tsS546G, and E573K suppresses ts deletion 562Y—a portal cs mutation csM308I located in the clip interaction region suppresses 17 tsD473N (Fig. 5 A and B). It has been shown by FRET that the terminase C terminus is located ∼57 Å from the portal and approaches the portal by about 6 Å in the stalled Y-DNA substrate (10). Taken together, these observations suggest that the portal plays a significant role in DNA translocation that is apparently coupled to interaction with the terminase C-terminal nuclease domain. The apparent closer approach of the terminase to the portal during packaging that is seen in the stalled Y-DNA, coupled with a grip-and-release of the DNA by the portal, may account for a transient DNA compression that is connected to the power stroke of the motor complex and to removal of IDs from dsDNA.

Materials and Methods

Bacteriophages, Bacterial Strains, and Growth of Phage Infected Bacteria.

T4D phage empty large proheads were prepared for packaging in vitro by centrifugation and column chromatography as previously described (6). Escherichia coli DH10B was used for transformation of plasmids in cloning as well as for SDM and hydroxylamine mutagenesis experiments. BL21 (DE3) pLys was used for expression of recombinant proteins. Thirteen different E. coli suppressor strains that insert known different amino acids at amber codons were used to characterize terminase SDM amber mutant sites (25). Luria–Bertani (LB) medium was used for growth of cells, except for phage infection experiments in which M9S/20% LB medium (12) was used. Phage infection in the presence of 9AA was carried out at 30 °C, where the bacteria were grown at 37 °C and shifted down to 30 °C immediately before infection. The 9AA (3 µg/mL) was added to the culture 30 s after the first infection. After 9AA treatment, heads were prepared by glycerol gradient centrifugation similar to the purification of proheads described in ref. 6, but no column purification was done. Plating was carried out on plates containing 30 mL of Bottom agar (with or without 9AA and EthBr at 1 µg/mL) to determine acridine resistance and sensitivity (12).

IDs.

EthBr and 9AA powder were from Sigma and YOYO-1 from Molecular Probes (Invitrogen).

Mutagensis.

Hydroxylamine mutagenesis (26), spontaneous mutagenesis, and SDM of the gp17 gene was carried out as described in SI Mutagenesis Procedure.

YOYO-1 Staining of the dsDNAs.

The double-stranded linear DNAs used in this study are 5-kb-length Pl16 DNA linearized by digesting it at the unique PstI site (6), a 280 bp gene20 PCR product, and a 70 bp DNA prepared using oligonucleotides synthesized from Sigma. In all of the experiments, the estimated staining ratio was roughly 1/20–1/5 dye molecules per DNA base pair. Terminase–DNA ratio was constant, with ∼3 gp17/100 bp of DNA in most of the experiments, except for concentration-dependent experiments where different ratios of terminase–DNA were used, as shown in Fig. 3C and Fig. S2. The concentrations of DNA and YOYO-1 were determined spectrophotometrically using the extinction coefficient, є = 6,600 M^–l·cm^–1 for DNA bases and є = 96,100 M^–l·cm^–1 for YOYO-1. All samples had a final total volume of 16 µL in 2× packaging buffer (10) with 10 mM ATP/ ATP-γ-S or no ATP and were prepared by adding 1 µL of DNA stock to the dye diluted in the buffer, and subsequently incubating at 50 °C for 2 h to achieve homogeneous staining of all molecules (13). To make sure the sample was equilibrated—that is, that it only gave one band in electrophoresis—the samples were always run on 1.5% (wt/vol) agarose and at a 100 V constant field electrophoresis. Western blotting was carried out using a gp17 antiserum as previously described (27) after running the samples on 5% native PAGE gel under cold conditions.

Cloning, Expression, and Purification of ac and acq Terminases.

The ac and acq gp 17 genes were cloned as the wt gp 17 gene in pTYB2 plasmids and were expressed in BL21 (DE3) pLysS, successfully making the gp17-intein/chitin-binding domain fusion (27). Full-length active ac and acq gp17 was purified based on the Impact system of cloning (New England Biolabs, Inc.).

DNA Packaging Assays and Analyses.

DNA packaging, analysis of gels on Typhoon imager, and FCS measurements on a Picoquant microtime system were carried out as previously described (8, 10).