Endonuclease and Helicase Activities of Bacteriophage λ Terminase: Changing Nearby Residue 515 Restores Activity to the gpA K497D Mutant Enzyme

Virology 277, 204–214 (2000) doi:10.1006/viro.2000.0591, available online at http://www.idealibrary.com on

Endonuclease and Helicase Activities of Bacteriophage ␭ Terminase: Changing Nearby Residue 515 Restores Activity to the gpA K497D Mutant Enzyme Young Hwang, 1 Julie Qi Hang, 2 Jayson Neagle, Carol Duffy, and Michael Feiss 3 Department of Microbiology, College of Medicine, University of Iowa, Iowa City, Iowa 52242 Received July 17, 2000; accepted August 14, 2000 Terminase, the DNA packaging enzyme of bacteriophage ␭, is a heteromultimer of gpNu1 and gpA subunits. In an earlier investigation, a lethal mutation changing gpA residue 497 from lysine to aspartic acid (K497D) was found to cause a mild change in the high-affinity ATPase that resides in gpA and a severe defect in the endonuclease activity of terminase. The K497D terminase efficiently sponsored packaging of mature ␭ DNA into proheads. In the present work, K497D terminase was found to have a severe defect in the cohesive end separation, or helicase, activity. Plaque-forming pseudorevertants of ␭ A K497D were found to carry mutations in A that suppressed the lethality of the A K497D mutation. The two suppressor mutations identified, A E515G and A E515K, affected residue 515, which is located near the putative P-loop of gpA. A codon substitution study of codon 515 showed that hydrophobic and basic residues suppress the K497D defect, but hydrophilic and acidic residues do not. The E515G change was demonstrated to reverse the endonuclease and helicase defects caused by the K497D change. Moreover, the gpA K497D E515G enzyme was found to have kinetic constants for the high-affinity ATPase center similar to those of the wild type enzyme, and the endonuclease activity of the K497D E515G enzyme was stimulated by ATP to an extent similar to the ATP stimulation of the endonuclease activity of the wild type enzyme. © 2000 Academic Press Key Words: virus assembly; genome encapsidation; virus genome packaging; DNA packaging.

INTRODUCTION

required for initiation of DNA packaging. Terminase is also proposed to be involved in ATP hydrolysis-dependent translocation of ␭ DNA into the prohead. The DNA site that is recognized and cut by terminase is termed cos. ␭ cos consists of three subsites, cosQ, cosN, and cosB (Fig. 1). cosB consists of three binding sites, R3, R2, and R1, for the small subunit (gpNu1) of terminase (Shinder and Gold, 1988). Between R3 and R2 is I1, a binding site for integration host factor (IHF), the Escherichia coli DNA binding and bending protein (Kosturko et al., 1989; Xin et al., 1993; Xin and Feiss, 1988, 1993). cosN is the nicking site at which terminase introduces nicks, staggered by 12 bp, to generate the cohesive ends of mature ␭ chromosomes. cosQ is required for termination of DNA packaging (Cue and Feiss, 1993). Genetics studies indicate that the terminase subunits consist of a series of functional domains. The first 91 amino acids of gpNu1 contain a domain for DNA binding (Frackman et al., 1985); this segment contains a putative helix-turn-helix DNA-binding motif involving residues 5 to 24 (A. Becker, cited in Feiss, 1986; Kypr and Mrazek, 1986), followed by a putative ATPase center starting at about residue 29 (Fig. 1) (Becker and Murialdo, 1990). The carboxyl-terminus of gpNu1 is a domain for binding gpA (Frackman et al., 1985). The amino-terminal 48 residues of gpA contain a domain for binding gpNu1 (Wu et al., 1988) Other domains of gpA include a putative ATPase center starting near residue 491 (Guo et al., 1987) and a putative basic leucine zipper including residues

Terminase is the virus-encoded enzyme that mediates packaging of ␭ DNA into the prohead, the protein shell that is the capsid precursor. Terminase is a heteromultimer composed of gpNu1 4 (181 aa), the product of gene Nu1, and gpA (641 aa), the product of gene A (Becker and Murialdo, 1990; Catalano et al., 1995; Cue and Feiss, 1993; Feiss, 1986). The activities of terminase include (i) binding to ␭ DNA, (ii) ATP-stimulated endonucleolytic cleavage of concatemeric ␭ DNA to generate the cohesive ends of virion DNA, (iii) ATP hydrolysis-dependent separation of the cohesive ends, and (iv) recognition of ␭ proheads to form a DNA-terminase-prohead complex

1 Present address: Stratagene Inc., 11011 N Torrey Pines Road, La Jolla, CA 92037. 2 Present address: Eli Lilly, Inc., Indianapolis, IN 46285-0438. 3 To whom correspondence and reprint requests should be addressed at University of Iowa College of Medicine, Dept. of Microbiology, 3403 Bowen Science Bldg., Iowa City, IA 52242-1109. Fax: (319) 335-9006. E-mail: [email protected]. 4 Abbreviations used: SDS–PAGE, sodium dodecyl sulfate–polyacrylamide gel electrophoresis; gpNu1, the product of gene Nu1, etc.; bZip, basic leucine zipper; PFU, plaque-forming units. Mutations are indicated by the resulting amino acid change, so that the A K497D mutation is a change in the 497th codon of the A gene, changing the normal wild type lysine codon to an aspartic acid codon. Mutant terminases are indicated by noting the change in the relevant terminase, so that gpA K497D terminase is comprised of normal gpNu1 and gpA with the K497D change.

0042-6822/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.

204

TERMINASE AND DNA PACKAGING

205

FIG. 1. Genes and proteins involved in DNA packaging. The top of the diagram shows the cos-terminase segment of the ␭ chromosome. cos contains all the sites required for DNA packaging; and genes Nu1 and A encode the small (gpNu1) and large (gpA) subunits of terminase, respectively. Below the gene map are the domain structures of gpNu1 and gpA. For gpNu1, HTH represents the putative helix-turn-helix domain for binding cosB, ATP represents a putative ATP binding center, and gpA represents a specificity domain for binding gpA. For gpA, ATP represents a putative ATP binding center, gpNu1 represents a specificity domain for binding gpNu1, bZIP is a putative basic leucine zipper domain, and Pro indicates a domain for interaction with the prohead. Asterisks indicate the locations of mutations that specifically inactivate the endonuclease activity of terminase (Davidson and Gold, 1992). At the bottom is a diagram of the cos subsites and the gpA sequence from bp 491 to bp 515. cosN is the site at which terminase introduces staggered nicks to generate the cohesive ends of mature ␭ molecules, and the staggered vertical lines in cosN indicate the nick positions. cosB and cosQ are required for packaging initiation and termination, respectively. The coordinates of the gene endpoints are: Nu1, bp 191 to bp 737; A, bp 711 to bp 2633; and W, bp 2633 to bp 2836, where the bp bounding the cosN nick site on the top strand are designated bp 48502 for the bp to the left and bp 1 for the bp to the right (Daniels, 1983). The indicated locations of restriction endonuclease sites used for subcloning are: R, EcoRI, bp 194; M, MluI, bp 458; S, SphI, bp 2212; and O, EcoO109I, bp 48495 and bp 2815.

573 to 616 (Davidson and Gold, 1992). Mutations in the basic leucine zipper (residues 586 and 600), the putative ATPase center (residue 497), and in a third segment (residues 401 and 403) specifically inactivate the endonuclease activity of terminase (Davidson and Gold, 1992). The carboxyl-terminal 32 amino acids of gpA are a domain for prohead binding (Frackman et al., 1984; Wu et al., 1988; Yeo and Feiss, 1995a,b). ATP plays several important roles in DNA packaging. First, ATP stimulates the rate and accuracy of cosN nicking; ATP hydrolysis is not required for this stimulation. Second, ATP hydrolysis is required by a strandseparation activity of terminase that separates the cohesive ends (Fig. 1) (Higgins et al., 1988). Finally, ATP hydrolysis is required for translocating ␭ DNA into the prohead (Becker and Murialdo, 1990; Tomka and Catalano, 1993b). Terminase holoenzyme has two ATP reactive centers, based on the deduced primary amino acid sequences of both gpNu1 and gpA subunits (Becker and Gold, 1988; Guo et al., 1987) and on ATPase assays showing that terminase holoenzyme possesses two kinetically distinct ATPase reaction centers (Gold and Becker, 1983; Tomka and Catalano, 1993a). Additional work shows that gpA has a DNA-independent, high-affinity ATPase center and gpNu1 has a low-affinity ATPase center (Hwang et al.,

1996). For the gpNu1 ATPase, DNA decreases the K m threefold and increases the k cat twofold (Tomka and Catalano, 1993a). In a previous study of the putative ATP binding site of gpA, a mutation changing the conserved lysine residue 497 to aspartic acid was found to affect the high-affinity ATPase activity and the endonuclease activity of holoterminase; the K m was increased 14-fold (Hwang et al., 1996) and the rate of cos cleavage was decreased by a factor of 2000 (Hwang and Feiss, 1996). In the present work, mutations suppressing K497D’s lethality were isolated, and the K497D mutation was shown to inactivate the helicase activity, in addition to inactivating the endonuclease activity. To understand how the suppressor mutations act at the molecular level, we compared the suppressor terminases with the wild type and gpA K497D terminases. RESULTS Isolation and characterization of mutations able to suppress the K497D change in gpA To isolate suppressors of the lethal A K497D mutation, lysates of ␭-P1 A K497D were plated for plaque-forming variants. When no plaque-forming variants were obtained in unmutagenized lysates, ␭-P1 A K497D was

206

HWANG ET AL. TABLE 1 Yields of ␭-P1 Phages With Various A Gene Alleles

␭ Phage ␭ ␭ ␭ ␭

A⫹ A K497D A K497D E515G A K497D E515K

Yield (PFU/induced lysogen)

Relative yield a

Yield b

120 ⫾ 10 ⬍10 ⫺7 34 ⫾ 4 87 ⫾ 4

1.00 ⬍10 ⫺9 0.28 0.73

25 ⫾ 4 ⬍10 ⫺7 Nd c Nd

a

The number of PFU released/induced lysogen was given a value of 1.00 for ␭-P1. The yields of PFU for the other phages are expressed as a fraction of this number. b The number of Kn-transducing particles per induced lysogen was determined as described under Materials and Methods. c Nd, not determined.

introduced into the mutD ⫺ strain MF2449, induced, and plated on the standard host, MF1427. Plaque-forming variants of ␭-P1 A K497D were found at a frequency of approximately 10 ⫺7/induced cell. Seven plaque-forming variants were chosen for further study; five formed medium-size plaques on MF1427 and two formed small-size plaques. For each variant, the DNA fragment extending from ␭ bp 1 to bp 2815 was cloned into the plasmid vector pIBI31. The bp 1–2815 DNA fragment carries cosB, the Nu1 and A genes, and part of the W gene (Fig. 1). The 1–2815 segments were subdivided by subcloning, and the subsegments were crossed with ␭-P1 A K497D in marker rescue experiments to map the location of each mutation. For the seven variants, the segment extending from bp 2212 to bp 2815 from each was found to rescue the A K497D mutant. The segment from 2212 to 2815 was sequenced from each variant; each was found to retain the mutation causing the gpA K497D change, plus an additional mutation affecting codon 515 of the A gene. Five variants forming medium-size plaques were all found to carry the change G 2253A, changing codon 515 of gene A from GAA (glutamic acid) to AAA (lysine), that is, gpA E515K. Two variants forming small-size plaques were found to carry the change A 2254G, changing codon 515 from GAA (glutamic acid) to GGA (glycine). The E515G mutation is identical to a mutation isolated as a suppressor of the cosN G 2C C 11G mutations (Arens et al., 1999). Residue 515 is located near a putative P-loop of gpA, which is proposed to be in the segment extending from residue 491 to 497 (Guo et al., 1987) (Fig. 1). To quantify the effects of the E515K and E515G changes on growth of ␭-P1 A K497D, the phage yields were determined for ␭-P1 A ⫹, ␭-P1 A K497D, ␭-P1 A K497D E515G, and ␭-P1 A K497D E515K. The yield of ␭-P1 A K497D was ⬍10 ⫺9 that of the wild type phage (Hwang and Feiss, 1996), whereas the yields of the phages with K497D E515G and K497D E515K changes were 28 and 73% that of the wild type phage, respectively (Table 1). Since ␭-P1 A K497D is unable to form plaques,

the yield was determined by assaying Kn-transducing virions. As described above, the E515G change was previously characterized as being able to suppress the cosN G 2C C 11G mutations. Furthermore, mutations causing the E515A, E515I, E515V, E515F, E515S, E515Q, and E515Y changes were also found to be suppressors of the cosN G 2C C 11G mutations (Arens et al., 1999). Therefore, it was of interest to determine which of the substitutions at position 515 could suppress the K497D change. Effect of changing gpA residue 515 on growth of ␭-P1 A K497D To study the effects of substitutions at position 515 on growth of ␭-P1 A K497D, marker rescue experiments were performed using MF1427 cells lysogenized by ␭-P1 A K497D. Plasmids with substitutions at position 515 were used in crosses with ␭-P1 A K497D. Rescues with plasmids specifying residues A, F, G, V, I, K, and R produced viable recombinants, whereas plasmids specifying residues D, S, Q, and Y did not (Table 2). All of the residues tested were functional in an otherwise wild type background (Table 2). To directly observe the effect of the E515G change on terminase action, the mutations generating the E515G change were introduced into pCM101, the ␭ terminase expression vector (Chow et al., 1987), and gpA E515G and gpA K497D E515G mutant terminases were purified to homogeneity. ATPase, in vitro and in vivo cos cleavage and in vitro DNA packaging activities were compared for the wild type and mutant terminases. TABLE 2 Rescue of ␭-P1 A K497D by Various Substitutions at Residue 515 of gpA

Amino acid

Rescue of A K497D a

Viability in a wild type background b

E D G A I V F K R S Q Y

⫺ ⫺ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫺ ⫺ ⫺

⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹

a Crosses of ␭-P1 A K497D versus derivatives of plasmid pHW436 carrying various alleles at codon 515 of the A gene. The ␭ DNA insert in plasmid pHW436 lacks codon 497 of the A gene. Plaque-forming recombinants were scored by plating on MF1427. b Crosses of ␭-P1 Aam55 versus plasmids carrying various codons at codon 515 of gene A were carried out by induction of an MF1427 (␭-P1 cI857 Aam515) lysogen transformed by derivatives of pHW436. Plaque-forming recombinants were scored by plating on MF1427.

TERMINASE AND DNA PACKAGING TABLE 3 Kinetic Parameters of the ATPase Activities of Terminases Low-affinity ATP binding site

High-affinity ATP binding site

Terminase

k cat (min ⫺1) a

K m (␮M)

k cat (min ⫺1)

K m (␮M)

Wild type b gpA K497D gpA K497D E515G gpA E515G

84 128 80 79

469 1999 480 532

38 38 42 14

4.6 64 7.1 10.2

k cat ⫽ k obs/[enzyme]. Assays were done in the presence of 1.5 nM ␭ virion DNA. b Data for wild type enzyme are taken from Hwang et al. (1996). a

Effect of the E515G change on ATPase activity

207

terminase was indistinguishable from that of the wild type enzyme. In the absence of ATP, the cos cleavage activities of the gpA K497D E515G and gpA E515G terminases were indistinguishable from that of the wild type enzyme (Fig. 2B). The rates of cos cleavage of wild type, gpA K497D E515G, and gpA E515G terminases in the absence of ATP were 1/50–1/100 those in the presence of ATP (Fig. 2B), showing that ATP stimulates the cos cleavage activity of each of these terminases. These results show that the E515G change restores the endonuclease activity of the gpA K497D enzyme to wild type levels and returns the ATP stimulation of the endonuclease to that of the wild type enzyme. Note also that in the absence of ATP, the gpA K497D enzyme has a cos cleavage activity about 10-fold below the activities of the other enzymes, indicating that the gpA K497D enzyme has a defect in the basal

Wild type terminase has two ATPase activities: a lowaffinity (K m ⬃ 500 ␮M) center in gpNu1 and a high-affinity (K m ⬃ 5 ␮M) center in gpA (Hwang et al., 1996; Tomka and Catalano, 1993a). The high-affinity center in gpA K497D terminase has a reduced ATP affinity, since the K m was increased 14-fold. There was also a mild (fourfold) decrease in the ATP affinity of the low-affinity center (Hwang et al., 1996). ATPase assays showed that kinetic constants for ATP hydrolysis for the gpA K497D E515G terminase were within experimental error of those for the wild type terminase, while the kinetic constants for ATP hydrolysis for the gpA E515G terminase differed modestly from those of the wild type enzyme (Table 3). In sum, the E515G change corrects the mild ATP-binding defect in the high-affinity ATPase center caused by the K497D change. Effect of the E515G change on the endonuclease activity of terminase, in vitro and in vivo To determine the effect of the mutation causing the E515G change on the cos cleavage activity, the purified wild type, gpA K497D, gpA K497D E515G, and gpA E515G terminases were studied over a range of DNA concentrations (Fig. 2A). The cleavage rate of gpA K497D terminase was 1/2000 that of the wild type terminase at 1.2 mM ATP. Note that the level of ATP used was saturating for the high-affinity site and near the K m for the lowaffinity site. Although the level of ATP was not saturating for the low-affinity ATPase center in gpNu1 of the gpA K497D terminase, the level was sufficient to saturate roughly half of that site. Since both sites of half of the gpA K497D enzyme were filled, we conclude that the endonuclease defect is not corrected by ATP binding. The reduced cos cleavage activities remained low, even when the substrate DNA was raised to 130 nM (Hwang and Feiss, 1996). The rate of cos cleavage of gpA K497D E515G terminase was 50% that of the wild type enzyme at saturating DNA. The rate of cos cleavage of gpA E515G

FIG. 2. In vitro cos cleavage. (A) Effect of substrate DNA concentration on in vitro cos cleavage by wild type, gpA K497D, gpA K497D E515G, and gpA E515G mutant terminases. The in vitro cos cleavage assay was performed as described under Materials and Methods using ScaI-linearized pSX1 as a substrate DNA. (B) ATP dependence of in vitro cos cleavage by wild type, gpA K497D, gpA K497D E515G, and gpA E515G mutant terminases. ATP (1.2 mM) and terminase (50 nM) were used.

208

HWANG ET AL.

endonuclease activity, i.e., the activity unstimulated by ATP. An in vivo cos cleavage assay was performed to determine whether gpA K497D, gpA K497D E515G, and gpA E515G terminases were able to cleave cos efficiently in vivo. For examination of cos cleavage in vivo, DNA from induced lysogens was isolated, cut with ClaI restriction enzyme, and probed. The ClaI restriction enzyme fragment containing uncut cos is 6.3 kb in length and is called the joint fragment (J). Cleavage of the joint fragment by terminase generates a 2.1-kb piece (called R end) from the right (cosQ) end of the ␭ chromosome and a 4.2-kb piece (called L end) from the left (cosB) end (Fig. 3A). No cos cleavage was observed for ␭-P1 A K497D (⬍2%; Fig. 3B, lane 2). In contrast, substantial levels of mature left (L) and right (R) chromosomal end fragments were found for the positive control phage, wild type ␭-P1 (65%; Fig. 3B, lane 1), ␭-P1 A K497D E515G (32%; Fig. 3B, lane 3), and ␭-P1 A E515G (66%; Fig. 3B, lane 4), confirming that a primary effect of the E515G change is on the ability to cut cos. Effects of the K497D and E515G changes on the helicase activity of terminase Recent work has demonstrated that terminase has an ATP-dependent helicase activity that separates the cohesive ends after they have been created by the endonuclease activity (Higgins et al., 1988; Rubinchik et al., 1994; Yang and Catalano, 1997). To investigate a possible functional relationship between the endonuclease and helicase activities, we studied the effects of the K497D and E515G changes on the helicase activity of terminase. The helicase assay used contained as substrate AccI-digested mature ␭ DNA that had been heated to anneal the cohesive ends. Separation of the cohesive ends converts the 7.8-kb joint piece (J) into 5.6-kb (R) and 2.2-kb (L) product pieces that were resolved by agarose gel electrophoresis (Fig. 4). The K497D enzyme lacked helicase activity and the E515G and K497D E515G enzymes had helicase activities similar to that of the wild type enzyme (Table 4). The results indicated a close functional relationship between the endonuclease and helicase activities of terminase, and as one might expect for a suppressor restoring viability to ␭ A K497D, the E515G suppressor restored both activities. Effect of the E515G change on in vitro DNA packaging To measure packaging activity, the wild type, gpA K497D, gpA K497D E515G, and gpA E515G mutant terminases were added to a crude in vitro packaging system containing virion DNA as the packaging substrate. Since the virion DNA contains cohesive ends, the endonuclease activity of terminase is not required in this packaging

FIG. 3. DNA processing in vivo. (A) Rationale of the in vivo cos cleavage assay. Top line: A segment of concatemeric ␭ DNA containing cos. Cleavage of concatemeric DNA that is uncut by terminase generates a joint, cos-containing DNA segment, labeled J, that is 6.3 kb in length. Cleavage of an initial cos site by terminase generates one mature right and one mature left chromosomal end, which proceeds to the right of the cleaved cos site. The right (R) and left (L) chromosomal end pieces created by terminase cleavage of the 6.3-kb joint piece are 2.1 and 4.2 kb in length, respectively, as shown on the lower line. The distribution of joint, left end, and right end pieces is determined by separation of the ClaI digest of intracellular DNA by agarose gel electrophoresis and detected by Southern blotting. (B) In vivo cos cleavage assays. Total nucleic acids were isolated from ␭-infected MF1427 cells, digested with ClaI, separated by electrophoresis on a 0.8% agarose gel, transferred to a nitrocellulose membrane, and hybridized with 32 P-labeled pHW14. As indicated, DNAs isolated from induced lysogens of wild type ␭-P1 A K497D, ␭-P1 A K497D E515G, and ␭-P1 A E515G were studied. The positions of the bands that correspond to the uncut immature chromosome (J), the mature left chromosomal ends (L), and the mature right chromosome ends (R) are indicated.

assay. Packaging was followed by measuring the assembly of plaque-forming units (PFU). The gpA K497D terminase had a reproducible mild defect in packaging at limiting concentrations. Reactions with excess mutant terminase were as effective as reactions with the wild type enzyme (Hwang and Feiss, 1996). The E515G change corrects the mild packaging defect of K497D at limiting concentrations (Fig. 5). The packaging activities of gpA K497D, gpA K497D E515G, and gpA E515G mutant terminases were indistinguishable from that of wild type terminase (Fig. 5) at levels of terminase equivalent to the in vivo level of approximately 140 nM.

TERMINASE AND DNA PACKAGING

FIG. 4. Effects of the K497D and E515G changes on the helicase activity of terminase. The substrate was AccI-digested ␭ DNA, heated to anneal the cohesive ends of the 2.2-kb left (L) and 5.6-kb right (R) genomic end pieces to generate the 7.8-kb joint (J) piece. Helicase separation of the annealed cohesive ends of the J piece to produce the L and R product pieces was followed by agarose gel electrophoresis, blotting, and probing with 32P-labeled ␭ DNA. The “Heated 70°C” lane contained substrate DNA heated to artificially separate the annealed cohesive ends; the “No Terminase” lane indicates a reaction with terminase omitted; the “BSA” lane indicates a reaction in which bovine serum albumin was used in place of terminase; and the remaining lanes show results for reactions containing wild type, gpA K497D, gpA E515G, and gpA K497D E515 terminases as indicated.

DISCUSSION The A E515K and A E515G mutations were recovered as suppressors of the lethal A K497D mutation. Further mutational work showed that a variety of changes affecting residue 515 correct the lethality of the A K497D mutation. The E515G change has been shown to correct the defects in the endonuclease, helicase, and highaffinity ATPase activities of gpA K497D terminase. We next discuss the significance of these results in light of what is known about the nature of the gpA K497D defect. The A K497D mutation In a previous study of gpA’s high-affinity ATPase, the conserved lysine of the putative P-loop, K497, was changed (Hwang et al., 1996; Hwang and Feiss, 1996). Comparing the ATPase activities of wild type and mutant terminases showed that the K497D change altered the TABLE 4 Helicase Activities of Terminases

Terminase

Helicase activity (nM of product/min) a

Wild type gpA K497D gpA E515G gpA K497D E515G

0.47 ⬍0.02 0.44 0.35

a Helicase assay carried out as described under Materials and Methods.

209

FIG. 5. Effect of the K497D and E515G changes on in vitro DNA packaging in a crude extract system. The in vitro packaging assay was performed using mature ␭ DNA as described under Materials and Methods. The packaging activity is defined as PFU/␮g ␭ DNA.

high-affinity ATPase site of terminase; the K m for ATP was increased 14-fold and the k cat was unchanged (Table 3) (Hwang et al., 1996). Furthermore, the rate of cos cleavage by K497D terminase decreased 2000-fold (Hwang and Feiss, 1996). The endonuclease of gpA K497D enzyme was defective even when ATP was present at a level sufficient to saturate the altered site, showing that the gpA K497D endonuclease defect was not simply a consequence of the lowered affinity for ATP. Based on this data, the K497D mutation was proposed to disrupt communication between the ATP binding and endonuclease domains. In the present work we have shown that the K497D change additionally inactivates terminase’s helicase. This is the first description of a helicase mutant of terminase. Recent work indicates that the interactions of terminase with ATP may be more complex than previously thought. Photocrosslinking with analogues of ATP carrying azido groups has shown that ATP-interactive amino acids are located in the amino terminus of gpA, residues far from the putative P-loop in residues 491–497 (Babbar and Gold, 1998; J. Q. Hang and M. Feiss, unpublished observations). In the work from this laboratory, 8-azidoATP was found to photocrosslink to residues Y46 and K84 of gpA. Mutational alteration of A gene codons 46 and 84 resulted in mutant enzymes with severe defects in gpA’s high-affinity ATPase, defects much more severe than the K m change observed for gpA K497D terminase. The residue 46 and 84 mutant enzymes had normal helicase and endonuclease activities but failed to sponsor virion assembly. Thus the residue 46 and 84 mutants have activities and defects that are reciprocal to those of the K497D mutant enzyme. [It is important to note that the high-affinity ATPase activity of gpA is a basal activity. Terminase engaged in the helicase and translocation steps of DNA packaging is expected to exhibit ATPase activities that are specifically activated during the helicase and translocation steps (Guo et al., 1987; Morita et

210

HWANG ET AL.

al., 1993).] The contrast between the severe ATPase defects caused by changing residues 46 and 84 and the mild defect caused by changing residue 497 raises the possibility that the mild effect of the K497D change on the high-affinity ATPase may be an indirect effect on the N-terminal ATPase center of which residues 46 and 84 are a part. It is also the case that changing residues 46 and 84 had no effect on the ATP-stimulated endonuclease and ATP-dependent helicase activities, suggesting that, in addition to the amino terminal ATPase involved in post-cos cleavage stages of packaging, gpA has a carboxyl-terminal ATPase associated with the endonuclease and helicase centers. The proposed carboxyl-terminal ATPase center associated with the endonuclease and helicase activities appears not to be detectable under the assay conditions used here. Thus it is not clear that the K497D change affects an ATPase center involved in the endonuclease and helicase activities. Suppression of the A K497D defect The helicase and endonuclease activities of K497D E515G terminase are near to those of the wild type enzyme (Table 4; Fig. 2A). Also, the E515G change alone had no significant effects on the endonuclease and helicase activities of terminase. The endonuclease activity of the gpA K497D E515G enzyme showed an ATP dependence similar to that of the wild type enzyme (Fig. 2B). The basal endonuclease activities of the wild type, gpA K497D E515G, and gpA E515G enzymes were also similar, indicating that the E515G change also suppresses the defect in the basal endonuclease activity of the gpA K497D enzyme (Fig. 2B). Thus E515G must restore the endonuclease of the gpA K497D enzyme to a configuration near to that of the wild type enzyme, both in the presence and absence of ATP. Similarly, the gpA K497D E515G high-affinity ATPase center binds ATP with an affinity near to that of the wild type enzyme, whereas the single E515G change caused only mild differences from the wild type enzyme in ATP affinity and hydrolysis (Table 3). Packaging of mature DNA in the crude system depends sharply on the terminase concentration, suggesting that assembly of multiple terminase protomers is required in the packaging process (Becker, 1977; Hwang and Feiss, 1995). In the crude system, there is a reproducible mild defect in DNA packaging by the K497D enzyme, at limiting terminase concentrations (Hwang and Feiss, 1996) (Fig. 5). The defect does not affect the maximal level of packaging, indicating that high concentrations of K497D terminase compensate for the defect. The compensation by high terminase concentrations suggests that the K497D change weakens a protein– protein interaction and/or a protein–DNA interaction involved in assembly. The interaction weakened could be an interaction between any of the components of the

packaging machine: terminase subunits, cos subsites, and the prohead’s portal protein. The E515G change corrects this mild DNA packaging defect caused by K497D. In the wild type enzyme, it is likely that any role of residue 515 in cos cleavage is indirect, because all of the 12 amino acid residues tested allowed plaque formation, including acidic (D), basic (K and R), hydrophobic (G, A, I, V, and F), and hydrophilic (S, Q, and Y) residues (Table 2). We have not carried out burst size determinations for phages with substitutions at residue 515. Since the test for ability to form plaques is only a qualitative test, possible subtle differences between phages with residue 515 substitutions have not been studied. The broad spectrum of residues which suppress the K497D change, including basic (K and R) and hydrophobic ones (G, A, I, V, and F), also suggests that residue 515 only indirectly influences the interaction of the endonuclease center with cosN. Again, it seems likely that there will be subtle differences between the endonuclease and helicase centers of the wild type and gpA K497D E515G enzymes. Information about the structure of terminase is required for a deeper understanding of these interactions. Residue 515 and suppression of cosN defects Mutations in cosN reduce both the affinity of terminase for cosN and a post-DNA binding step, such that the rate of cosN cleavage is diminished (Arens et al., 1999; Xu and Feiss, 1991). Mutations affecting gpA have been recovered as suppressors that compensate for the cos cleavage defects of cosN mutations (Arens et al., 1999). The suppressor mutations are nonspecific, in that the altered terminases are compatible with wild type and mutant cosNs. The gpA changes caused by these suppressors clearly alter the ability of terminase to interact with cosN. The cosN suppressor mutations are missense mutations that alter residues near the gpA P-loop: the residues affected are 504, 509, and, interestingly, 515. Residues present at position 515 of gpA that suppress cosN defects include hydrophobic and uncharged polar residues, although basic and acidic residues do not suppress cosN defects. The spectrum of residues that suppress cosN defects partially overlaps that of the suppressors of the K497D change. Again, the variety of residues that can act as cosN suppressors suggests that the changes at 515 indirectly affect the configuration of the endonuclease, such that the nature of its interaction with cosN is subtly changed. MATERIALS AND METHODS Media Tryptone broth, tryptone agar, and tryptone soft agar were prepared as described by Arber et al. (1983), except that MgSO 4 was added to a final concentration of 10 mM.

TERMINASE AND DNA PACKAGING

Luria Broth (LB), Luria agar (LA), and SOB were prepared as described (Sambrook et al., 1989). When required, kanamycin and ampicillin were used at final concentrations of 50 and 100 ␮g/ml, respectively. Final concentrations of isopropyl-␤-D-thiogalactopyranoside (IPTG) and X-gal used in LA were 0.1 mM and 0.02% (w/v), respectively (Vieira, 1987). Bacteria, phages, and plasmids The standard ␭ strain used was ␭-P1:5R cI857 Kn R nin5, which carries a 10-kb segment of phage P1 DNA encoding the plasmid replication and partitioning functions (Sternberg and Austin, 1983). The ␭-P1:5R prophage replicates as a single-copy plasmid using the P1 replication machinery. Upon inactivation of the cI857 repressor at 42°C, the prophage is induced to carry out the ␭ lytic cycle. A derivative of ␭-P1:5R cI857 nin5 carrying a 1.3-kb kanamycin cassette was employed in our studies; this phage, ␭-P1:5R cI857 Kn R nin5, or simply ␭-P1, has a genome size of approximately 46.2 kb (Pal and Chattoraj, 1988). ␭-P1 A K497D was described elsewhere (Hwang and Feiss, 1996). Phage M13KO7 was obtained from Stratagene Cloning Systems (La Jolla, CA). ␭-P1 Aam515, with an amber mutation at codon 515 of gene A, was constructed by Sara Max (Arens et al., 1999). The nonsuppressing host MF1427 is a galK mutant of C1a (Six and Klug, 1973), and the suppressing host was the supF strain C-4518 (King et al., 1992). The standard terminase expression was vector pCM101 (Chow et al., 1987) and the host was OR1265 (Reyes et al., 1979). A derivative of pCM101 carrying the A K497D change was described earlier (Hwang and Feiss, 1996); additional derivatives of pCM101 carrying A E515G or A K497D E515G were constructed using standard cloning techniques. Cosmid pSX1 (Xu and Feiss, 1991) is a derivative of pUC19 carrying the ␭ DNA segment extending from the BclI site at 47492 to an EcoRI site at 194 (Williams and Blattner, 1979). pSF1 is a derivative of pBR322 with an insert of ␭ DNA extending from the HindIII site at bp 44141 through cos to the BamHI site at bp 5505 (Frackman et al., 1984). pHW14 is a derivative of pBR322 and carries the ␭ DNA segment extending from the ClaI site at bp 46438 to the ClaI site at bp 4198 (Hwang and Feiss, 1996). pHW436, constructed using standard cloning techniques, is a derivative of the commercial vector pIBI (International Biotechnologies, New Haven, CT) containing the ␭ DNA segment extending from the SphI site at bp 2212 to the EcoO109 site at bp 2815. MF1975 is a ␭ ⫹ lysogen of MF1427. Packaging extracts were obtained from induced cultures of MF1427 [␭ cI857 Sam7 ⌬(cos-Nu1-A)::Kn R] (Hwang and Feiss, 1995). Strain XL1Blue (Stratagene Cloning Systems) was used as a transformation recipient. XL1-Blue and BW313 were used as the dut ⫹ ung ⫹ and dut ⫺ ung ⫺ strains, respectively, for site-specific mutagenesis (Kunkel, 1985). MF2449, a

211

zae-13::Tn10 mutD5 derivative of C1a, was constructed by and generously supplied by Dr. Mel Sunshine. C4518, a supF E. coli C strain (obtained from Dr. Mel Sunshine), was used as the host for titering ␭ cI857 Sam7 virions produced in in vitro virus assembly assays. General recombinant DNA techniques Restriction enzymes and and T4 polynucleotide kinase were from New England Biolabs (Beverly, MA). Bacteriophage T4 DNA ligase, the Klenow fragment of DNA polymerase I, and calf intestinal alkaline phosphatase were from Boehringer-Mannheim (Indianapolis, IN). These enzymes were used according to the supplier’s recommendations. Nucleoside triphosphates were from Boehringer-Mannheim. [␣- 32P]dCTP and [␣- 32P]ATP were from Amersham (Piscataway, NJ). ␭ DNA was purchased from New England Biolabs. Standard methods were used for plasmid DNA preparation (Birnboim and Doly, 1979), purification of DNA fragments from agarose gels (Vogelstein and Gillespie, 1979), and preparation of competent cells and transformation (Hanahan, 1983). Sequence designations and nomenclature All references to ␭ sequence positions are based on the standard numbering convention (Daniels, 1983). Numbering of the ␭ sequence begins with the first base of the left cohesive end and continues along the ␭ strand (the top strand) in a 5⬘ to 3⬘ direction. The position of each restriction cut site is given as the first nucleotide of the recognition sequence. Single-letter designations for amino acids are used. In vivo mutagenesis Mutagenesis of ␭-P1 A K497D was carried out by heat induction of a culture of the mutD ⫺ lysogen MF2449 (␭-P1 A K497D). Cells were grown in LB as described (Fowler, 1974). Burst size determinations Burst sizes were determined by the method described previously (Hwang and Feiss, 1996). Cultures of lysogens were grown in LB at 31°C to 5 ⫻ 10 7 cells/ml. The cultures were then induced by heating to 42°C for 15 min, followed by incubation at 37°C for 70 min. Plaqueforming phages were assayed by plating on MF1427 at 37°C. For the assay of Kn R-transducing particles, lysates were diluted in 10 mM MgSO 4. A 100-␮l aliquot of diluted lysate was added to 100 ␮l of stationary phase MF1975. The infected cells were incubated at 31°C for 60 min and then plated on LA ⫹ kanamycin plates at 31°C. In vitro site-specific mutagenesis Nine mutations affecting codon 515 of gene A were constructed. The A mutations caused the following

212

HWANG ET AL.

changes in gpA: E515D, E515R, E515A, E515V, E515I, E515F, E515S, E515Q, and E515Y. Mutagenesis was performed using the site-specific, oligonucleotide-directed method (Kunkel, 1985). Briefly, to construct mutations in A, BW313 was transformed with pHW436 and uridylated single-stranded template DNA corresponding to the plasmid was generated by infection with the helper phage M13KO7 (Vieira, 1987). Plasmid pHW436 contains the mutation E515G that also creates an HinfI site, so the mutagenic primers used to create the new 515 mutations could be identified as having eliminated the HinfI site. After in vitro synthesis of the complementary strand using the Klenow fragment of DNA polymerase I, the double-stranded circular DNAs were used to transform E. coli XL1-Blue (dut ⫹ ung ⫹); the mutagenized plasmids were isolated and the presence of the desired mutations was confirmed by restriction digestion and direct nucleotide sequencing (Sanger et al., 1977). To introduce terminase mutations into the terminase expression vector pCM101, standard recombinant DNA manipulations were used. To determine whether the codon 515 missense mutations were able to suppress the A K497D mutation, derivatives of pHW436 carrying the codon 515 missense mutations were used to transform MF1427 (␭-P1 A K497D); transformants were subsequently heat-induced and the resulting lysates plated for plaque-forming recombinants. Note that the ␭ DNA insert in pHW436 and derivatives extends rightward from the SphI site at bp 2212; accordingly the plasmids lack the wild type allele at codon 497, which is bp 2198–2200. To determine whether the missense mutations at codon 515 were viable in an otherwise wild type ␭-P1 background, the pHW436based plasmids were crossed with ␭-P1 Aam515, and the lysates plated for plaque-forming, am ⫹ recombinants on MF1427. Preparation of purified terminase and packaging extracts E. coli OR1265 (Reyes et al., 1979) carrying the terminase expression vector pCM101, or a derivative with a mutation altering codon 497 and/or 515 of gene A, was grown, induced, and harvested as described (Chow et al., 1987) for preparation of terminase. Terminase purification was carried out using the method described previously (Hwang et al., 1996). Packaging extracts supplying all the protein components required for virion assembly, except terminase, were prepared from induced cells of MF1427 [␭ cI857 Sam7 ⌬(cos-Nu1-A)::Kn R ] as previously described (Hwang and Feiss, 1995). Purified terminases were homogeneous as determined by SDS–PAGE. The protein concentration of terminase was determined by the Bradford method (Bradford, 1976).

ATPase activity of terminase The standard assay was performed as previously described (Hwang et al., 1996). Briefly, hydrolysis of [␣- 32P]ATP to [␣- 32P]ADP was measured under standard assay conditions. The standard assay was performed in a 20-␮l reaction containing 50 mM Tris–HCl (pH 9.0), 10 mM MgCl 2, 6 mM spermidine, 7 mM ␤-mercaptoethanol, 1 mM EDTA, 0.15 nM ␭ DNA, and various concentrations of [␣- 32P]ATP. Reaction products were separated on cellulose PEI-F thin-layer chromatography plates (J. T. Baker, Phillipsburg, NJ), followed by quantification with a radioanalytic imaging system (AMBIS Systems, San Diego, CA). The kinetic constants for ATP hydrolysis were determined by nonlinear regression analysis of the experimental data using the SigmaPlot data analysis program (Jandel Scientific, San Rafael, CA). In vivo cos cleavage The assay for phage DNA maturation in vivo was performed as previously described (Murialdo and Fife, 1987). Briefly, heat-induced cultures were rapidly chilled and centrifuged, and the resuspended cells were subjected to four extractions with phenol/CHCl 3, followed by ethanol precipitation and resuspension of the nucleic acids. The intracellular DNA was cut with ClaI restriction enzyme, and the DNA separated by agarose gel electrophoresis. Following electrophoresis, the DNA was transferred onto a GeneScreen Plus (New England Nuclear, Boston, MA) membrane using a vacuum blotting transfer system (American Bionetics, Hayward, CA). DNA hybridizations were performed using HindIII-digested 32P-labeled pHW14 as a probe. The radiolabeled nucleotides were quantified by radioanalytic imaging. In vitro cos cleavage For the in vitro cos cleavage assay, cos-containing pSX1 DNA was linearized with ScaI. Linearized pSX1 yields 1099- and 2320-bp fragments when cut at cos by terminase. The reactions were carried out as described (Chow et al., 1987). The standard 20-␮l reaction contained 50 mM Tris–HCl (pH 9.0), 10 mM MgCl 2, 6 mM spermidine, 6 mM putrescine, 7 mM ␤-mercaptoethanol, 1 mM EDTA, 1.2 mM ATP, 50 nM terminase, and various concentrations of linearized pSX1 DNA as indicated under Results. Prior to separation of products by 0.8% agarose gel electrophoresis, the samples were heated at 65°C for 5 min to separate any joined cohesive ends. Following electrophoresis, the DNA was transferred onto a GeneScreen Plus (New England Nuclear) membrane. DNA hybridizations were performed using 32P-labeled ScaI-treated pSX1 DNA as a probe. The radiolabeled DNAs were quantified by radioanalytic imaging. The initial velocity of cos cleavage was determined within the linear range of each reaction.

TERMINASE AND DNA PACKAGING

Helicase assay The helicase assay was performed as previously described (Yang and Catalano, 1997). The substrate was prepared by cutting mature ␭ DNA with AccI restriction enzyme. The cut DNA contained 10 fragments, including the 5.58-kb mature genomic right end and the 2.19-kb mature genomic left end. The genomic right and left ends were annealed by a 3-h incubation at 50°C in the presence of 0.01 M MgCl 2 to produce the 7.77-kb joint piece that is the substrate for the helicase activity. The helicase reactions were performed in the presence of a blocking oligonucleotide designed to anneal with the cohesive end of the left chromosomal end. The sequence of the blocking oligonucleotide was: 5⬘-AGGTCGCCGCCCGGGGGGG G P 3⬘-CCCCCCC C Each strand separation reaction had a total volume of 20 ␮l and contained 10 mM Tris–HCl (pH 8.0), 10 mM MgCl 2, 1 mM ATP, 100 ␮g/ml BSA, 2.5 mM blocking oligonucleotide, and 2.5 nM of DNA substrate. The reaction was initiated by the addition of terminase holoenzyme to a final concentration of 50 nM terminase and incubated at 37°C. After 3 min, each reaction was stopped by the addition of 3 ␮l agarose loading buffer (200 mM EDTA, 25% glycerol, 0.025% bromphenol blue). The DNA was fractionated on a 0.8% agarose gel. After electrophoresis at 100 V, the DNA was transferred onto a GeneScreen membrane. DNA hybridization was completed using [␣- 32P]-labeled pSF1 DNA as the probe. Quantification of strand separation was determined by phosphorimaging. In vitro DNA packaging The in vitro packaging assay was carried out as previously described (Hwang and Feiss, 1996). Briefly, a crude packaging mixture containing an extract of induced cells of MF1427 [␭ cI857 Sam7 ⌬(cos-Nu1-A)::Kn R] plus various concentrations of purified terminase and mature ␭ DNA was incubated under standard conditions. After a 30-min incubation period for phage assembly, dilutions were assayed for plaque-forming units by plating on C4518. ACKNOWLEDGMENTS We thank Carlos Catalano for discussions and comments on the manuscript. We thank our coworkers, Zhi-Hao Cai, David Cue, Carol Duffy, Qi Hang, Hillary Johnson, Jenny Meyer, John Randell, Jean Sippy, Michael Smith, Feodor Tereshchenko, Doug Wieczorek, and Ashly Yeo, for advice and interest during the course of this work. This research was supported by NIH research Grant GM-51611.

REFERENCES Arber, W., Enquist, L., Hohn, B., Murray, N. E., and Murray, K. (1983). Experimental methods for use with lambda. In “Lambda II” (R. W.

213

Hendrix, J. W. Roberts, F. W. Stahl, and R. A. Weisberg, Eds.), pp. 433–366. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Arens, J. S., Hang, Q., Hwang, Y., Tuma, B., Max, S., and Feiss, M. (1999). Mutations that extend the specificity of the endonuclease activity of lambda terminase. J. Bacteriol. 181, 218–224. Babbar, B. K., and Gold, M. (1998). ATP-reactive sites in the bacteriophage ␭ packaging protein terminase lie in the N-termini of its subunits, gpA and gpNu1. Virology 247, 251–264. Becker, A., and Gold, M. (1988). Prediction of an ATP reactive center in the small subunit, gpNu1, of the phage lambda terminase enzyme. J. Mol. Biol. 199, 219–222. Becker, A., and Murialdo, H. (1990). Bacteriophage lambda DNA: The beginning of the end. J. Bacteriol. 172, 2819–2824. Becker, A., Murialdo, H., and Gold, M. (1977). Early events in the in vitro packaging of bacteriophage DNA. Virology 78, 291–305. Birnboim, H. C., and Doly, J. (1979). A rapid alkaline extraction procedure for screening recombinant plasmid DNA. Nucleic Acids Res. 7, 1513–1523. Bradford, M. M. (1976). A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein dye binding. Anal. Biochem. 72, 248–254. Catalano, C. E., Cue, D., and Feiss, M. (1995). Virus DNA packaging: The strategy used by phage lambda. Mol. Microbiol. 16, 1075–1086. Chow, S., Daub, E., and Murialdo, H. (1987). The overproduction of DNA terminase of coliphage lambda. Gene 60, 277–289. Cue, D., and Feiss, M. (1993). A site required for termination of packaging of the phage lambda chromosome. Proc. Natl. Acad. Sci. USA 90, 9290–9294. Daniels, D., Schroeder, J., Szybalski, W., Sanger, F., Coulsen, A., Hong, D., Hill, D., Peterson, G., and Blattner, F. (1983). Completed annotated lambda sequence. In “Lambda II” (R. W. Hendrix, J. W. Roberts, F. W. Stahl, and R. A. Weisberg, Eds.), pp. 519–676. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Davidson, A., and Gold, M. (1992). Mutations abolishing the endonuclease activity of bacteriophage ␭ terminase lie in two distinct regions of the A gene, one of which may encode a leucine zipper DNA binding domain. Virology 161, 305–315. Feiss, M. (1986). Terminase and the recognition, cutting and packaging of ␭ chromosomes. Trends Genet. 2, 100–104. Fowler, R. G., Degner, G. E., and Cox, E. C. (1974). Mutational specificity of a conditional Escherichia coli mutator, mutD5. Mol. Gen. Genet. 133, 179–191. Frackman, S., Siegele, D. A., and Feiss, M. (1984). A functional domain of bacteriophage lambda terminase for prohead binding. J. Mol. Biol. 180, 283–300. Frackman, S., Siegele, D. A., and Feiss, M. (1985). The terminase of bacteriophage lambda: Functional domains for cosB binding and multimer assembly. J. Mol. Biol. 183, 225–238. Gold, M., and Becker, A. (1983). The bacteriophage ␭ terminase: Partial purification and preliminary characterization of properties. J. Biol. Chem. 258, 14619–14625. Guo, P., Peterson, C., and Anderson, D. (1987). Prohead and DNA-gp3dependent ATPase activity of the DNA packaging protein gp16 of bacteriophage ␾29. J. Mol. Biol. 197, 229–236. Hanahan, D. (1983). Studies on transformation of Escherichia coli with plasmids. J. Mol. Biol. 166, 557–580. Higgins, R. R., Lucko, H. J., and Becker, A. (1988). Mechanism of cos DNA cleavage by bacteriophage lambda terminase: Multiple roles of ATP. Cell 54, 765–775. Hwang, Y., Catalano, C. E., and Feiss, M. (1996). Kinetic and mutational dissection of the two ATPase activities of terminase, the DNA packaging enzyme of bacteriophage Chi. Biochemistry 35, 2796–2803. Hwang, Y., and Feiss, M. (1995). A defined system for in vitro lambda DNA packaging. Virology 211, 367–376. Hwang, Y., and Feiss, M. (1996). Mutations affecting the high affinity ATPase center of gpA, the large subunit of bacteriophage lambda

214

HWANG ET AL.

terminase, inactivate the endonuclease activity of terminase. J. Mol. Biol. 261, 524–535. King, R. A., Anders, D. L., and Christie, G. E. (1992). Site-directed mutagenesis of an amino acid residue in the bacteriophage P2 Ogr protein implicated in interaction with Escherichia coli RNA polymerase. Mol. Microbiol. 6, 3313–3320. Kosturko, L., Daub, E., and Murialdo, H. (1989). The interaction of E. coli integration host factor and lambda cos DNA multicomplex formation and protein-induced bending. Nucleic Acids Res. 17, 329–334. Kunkel, T. A. (1985). Rapid and efficient site-specific mutagenesis without phenotypic selection. Proc. Natl. Acad. Sci. USA 82, 488–492. Murialdo, H., and Fife, W. L. (1987). Synthesis of a trans-acting inhibitor of DNA maturation by prohead mutants of phage ␭. Genetics 115, 3–10. Pal, S. K., and Chattoraj, D. K. (1988). P1 plasmid replication: Initiator sequestration is inadequate to explain control by initiator-binding sites. J. Bacteriol. 172, 2819–2824. Reyes, O., Gottesman, M., and Adhya, S. (1979). Formation of lambda lysogens by IS2 recombination: gal operon-lambda P R promotor fusions. Virology 94, 400–408. Rubinchik, S., Parris, W., and Gold, M. (1994). The in vitro ATPases of bacteriophage lambda terminase and its large subunit, gene product A: The relationship with their DNA helicase and packaging activities. J. Biol. Chem. 269, 13586–13593. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989). “Molecular Cloning: A Laboratory Manual,” 2nd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Sanger, F., Micklen, S., and Coulsen, A. R. (1977). DNA sequencing with chain-termination inhibitors. Proc. Natl. Acad. Sci. USA 74, 5463– 5467. Shinder, G., and Gold, M. (1988). The Nu1 subunit of bacteriophage lambda terminase binds to specific sites in cos DNA. J. Virol. 62, 387–392. Six, E., and Klug, C. A. C. (1973). Bacteriophage P4: A satellite virus depending on a helper such as prophage P2. Virology 51, 327–344. Sternberg, N., and Austin, S. (1983). Isolation and characterization of P1 minireplicons, ␭-P1:5R and ␭-P1:5L. J. Bacteriol. 153, 800–812. Tomka, M. A., and Catalano, C. E. (1993a). Kinetic characterization of

the ATPase activity of the DNA packaging enzyme from bacteriophage lambda. Biochemistry 32, 11992–11997. Tomka, M. A., and Catalano, C. E. (1993b). Physical and kinetic characterization of the DNA packaging enzyme from bacteriophage lambda. J. Biol. Chem. 268, 3056–3065. Vieira, J. M. (1987). Production of single-stranded plasmid DNA. Methods Enzymol. 153, 3–11. Vogelstein, B., and Gillespie, D. (1979). Preparative and analytical purification of DNA from agarose. Proc. Natl. Acad. Sci. USA 76, 615– 619. Williams, B. G., and Blattner, F. R. (1979). Construction and characterization of hybrid bacteriophage lambda Charon vectors for DNA cloning. J. Virol. 29, 555–575. Wu, W. F., Christiansen, S., and Feiss, M. (1988). Domains for proteinprotein interactions at the N and C termini of the large subunit of bacteriophage lambda terminase. Genetics 119, 477–484. Xin, W., Cai, Z.-H., and Feiss, M. (1993). Function of IHF in ␭ DNA packaging. II. Effects of mutations altering the IHF binding site and the intrinsic bend in cosB on ␭ development. J. Mol. Biol. 230, 505–515. Xin, W., and Feiss, M. (1988). The interaction of Escherichia coli integration host factor with the cos sites of phage ␭ and 21. Nucleic Acids Res. 16, 2015–2030. Xin, W., and Feiss, M. (1993). Function of IHF in ␭ DNA packaging. I. Identification of the strong binding site for integration host factor and the locus for intrinsic bending in cosB. J. Mol. Biol. 230, 492–504. Xu, S. Y., and Feiss, M. (1991). Structure of the bacteriophage lambda cohesive end site: Genetic analysis of the site (cosN) at which nicks are introduced by terminase. J. Mol. Biol. 220, 281–292. Yang, Q., and Catalano, C. E. C. (1997). Kinetic characterization of the strand separation (“helicase”) activity of the DNA packaging enzyme from bacteriophage ␭. Biochemistry 36, 10638–10645. Yeo, A., and Feiss, M. (1995a). Mutational analysis of the prohead binding domain of the large subunit of terminase, the bacteriophage lambda DNA packaging enzyme. J. Mol. Biol. 245, 126–140. Yeo, A., and Feiss, M. (1995b). Specific interaction of terminase, the DNA packaging enzyme of bacteriophage lambda, with the portal protein of the prohead. J. Mol. Biol. 245, 141–150.