Genetic Evidence That Recognition of cos@ the Signal for

Copyright 0 1997 by the Genetics Society of America

Genetic Evidence That Recognition of cos@ the Signal for Termination of Phage A DNA Packaging, Depends on the Extent of Head Filling

David Cue’ and Michael Feiss

Department of Microbiology, University of Iowa, Iowa City, Iowa 52242 Manuscript received March 11, 1997

Accepted for publication May 27, 1997

ABSTRACT Packaging a phage A chromosome involves cutting the chromosome from a concatemer and translocat-

ing the DNA into a prohead. The cutting site, cos, consists of three subsites: COSN, the nicking site; c o d , a site required for packaging initiation; and cosQ a site required for termination of packaging. COSB contains three binding sites (R sequences) for gpNul, the small subunit of terminase. Because cosQ has sequence identity to the R sequences, it has been proposed that cosQ is also recognized by gpNul. Suppressors of c o d mutations were unable to suppress a COSQ point mutation. Suppressors of a cosQ mutation (cos@) were isolated and found to be of three sorts, the first affecting a base pair in COSQ.

The second type of cos4 suppression involved increasing the length of the phage chromosome to a length near to the maximum capacity of the head shell. A third class of suppressors were missense mutations in gene B, which encodes the portal protein of the virion. It is speculated that increasing DNA length and altering the portal protein may reduce the rate of translocation, thereby increasing the efficiency of recognition of the mutant COSQ None of the cosQsuppressors was able to suppress COSB mutations. Because COSQ and COSB mutations are suppressed by very different types of suppressors, it is concluded that cosQand the R sequences of c o d are recognized by different DNA-binding determinants.

M ANY large dsDNA viruses, such as the tailed bacteriophages and the herpes viruses, produce

multichromosomal lengths DNA as a result of replication and recombination. During virion assembly, the multimeric DNA must be processed to generate unit- length virion DNA. For viruses such as phage A that have unique (non-permuted) chromosomes, specific chromosome ends must be generated by recognition and cleavage of specific DNA sites by viral packaging proteins. DNA from A virions is a linear duplex, 48,502 bp long, with complementary, 12-base-long extensions at the 5’ ends of the strands. Intracellular A DNA is in the form of concatemers, end-teend multimers produced by rolling circle replication and recombination. Virion DNA is generated by the introduction of nicks, staggered by 12 bp, into the concatemeric DNA. The segment of DNA required for efficient packaging of a A chromosome is called cos; nicks are introduced at a subsite of cos, cosN. The nicks are introduced by a phage- encoded, multifunctional DNA packaging enzyme, terminase. Terminase is a heteromultimer of two subunits, gpNul (21 kD) and gpA (74 kD), the products of the A Nul and A genes, respectively. The endonuclease activity resides in gpA (DAVIDSON et al. 1992; RUBINCHIK et al. 1994). Many of the base pairs at the site of nicking are rotationally symmetric, suggesting that symmetri-

Cmesponding authm: Michael Feiss, Department of Microbiology, University of Iowa, Iowa City, IA 52242. E-mail: [email protected]

Presmt address: Department of Microbiology, University of Minne- sota, 420 Delaware St., Minneapolis, MN 55455.

Genetics 147: 7-17 (September, 1997)

cally disposed terminases are involved in nicking COSN. Early studies indicated that the sequences needed for packaging a A chromosome covered -200 bp, an extent much greater than COSN (FEISS et al. 1983; HOHN 1983; MIWA and MATUSBARA 1983). In 1984, BEAR et al. described a A mutant, A cos154, that was unable to form plaques on a host lacking integration host factor (IHF), the Escherichia coli sitespecific DNA binding protein that sharply bends DNA when bound (reviewed in FRIEDMAN 1988; RICE et al. 1996). cos154 was a Gto-T transition mutation affecting bp 160 of A DNA. The cos154 mutation was located in a sequence, called R1, that was re- peated nearby (Figure 1; BEAR et al. 1984). BEAR et al. identified four repeats of the sequence, called Rl-R4. R4 was found near the right end of A virion DNA, and R3, R2 and R1 were between COSN and the start of the Nul gene at the left end of virion DNA. Additionally, a match to the consensus IHF binding site was found between R3 and R2. It was proposed that the R sequences were terminase binding sites, and that the cos154 mutation weakened binding of terminase to R1. BEAR et al. (1984) further proposed that IHF assisted in the binding of terminase to the R sequences, perhaps by facilitating cooperative interactions between terminases bound at the R sequences. In this view, the cos154 mutation weakened terminase binding to cos so that IHF was required for stable binding to the mutant cos. More recent work has supported the speculations of BEAR et al. The small subunit of terminase, gpNul, has been shown to bind to R3, R2 and R1 (though not R4), and the cos154 mutation has been shown to weaken

8 D. Cue and M. Feiss

cos Nul

binding @Nul ATP bZip Pro, endonuclease * * *

I I

N2 cosm

A C A G G T T j $ k k ? m i G G 1 -St rand TGTCCAA C&/c&& +G &CC r-Strand

a COSNR N1

FIGURE 1.-cos and the terminase genes. The Nul and A genes encode the small (gpNu1) and large (gpA) subunits of terminase, gene Wencodes a protein involved in head stabilization and tail attachment, and gene B encodes the portal protein of the prohead. Relow the Nul and A genes are diagrams of the gpNul and gpA polypeptides and their domain structures (reviewed in FF.ISS, 1986; MURIALDO, 1989; BECKER and MURIALDO 1990; CATAIAZNO rt al. 109.3). For gpNu1, HTH denotes the location of a putative helix-turn-helix DNA binding motif, ATP indicates a match to known ATPases (RECKER and Germ 1988) and the likely site of the gpNul ATPase activity, and gpA indicates the domain for binding gpA. For gpA, gpNu1 and Pro indicate the sites of interaction with gpNu1 and the portal protein of the prohead, respectively; ATP indicates a putative ATPase center and the likely site of the gpA ATPase activity, and bZip indicates a putative basic leucine zipper (DAVIIWOX and GOLD 1992). Asterisks indicate sites where mutational changes specifically abolish the endonuclease activity of terminase. cos is tripartite, consisting of 60.~4, the termination signal; cosN, the nicking site; and cnsB, the terminase binding site. The three R sequences of cnsB are sites bound by gpNul, presumably through the HTH, and I1 is a binding site for integration host factor, IHF, the sitespecific DNA hinding/bending protein of E coli. The cosNsequence and the terminase nicking sites are shown at the bottom of the fimre. The N2 nick site is between bp 48,502 and bp 1; the N1 nick site is between bp 12 and bp 13. Rase pairs of cnsN showing"twofo1d rotational symmetly are boxed.

gpNu1 binding to R1 (SHINDER and GOLD 1988). Muta- tional work has shown that the Cto-T transition of cos154, when placed in the analogous bp positions of R2 or R3, causes IHF dependence (CUE and FEES 1992a,b; 1993a). A strong IHF binding site, 11, has been demonstrated to be located between R3 and R2 (XIN and FEISS 1988, 1993; KOSTURKO et nl. 1989; MENDELSON et nl. 1991; XIN et nl. 1993). The DNA segment from R3 through R1 has been named cosR, the terminase binding site. cosB has been shown to stimulate the efficiency and accuracy of the cos cleavage reaction, and to be essential for a postcleavage step required for initiation of DNA packaging (HIGGINS et nl. 1988; CUE and FEISS 1993a; HICCINS and BECKER 1994a,b; CAI f t nl. 1997). A triple mutant A c o d R3-R2-R1-, was found to be unable to form plaques, and viable pseudorevertant.. were found to have suppressing mutations in Nul (CUE and FEES 1992a,b). Many studies have led to the following model of A DNA packaging (see BECKER and Mu-

RIALDO 1990; CATAIANO ff nl. 1995). Terminase binds cosN and COSR to form an initial complex, followed by nicking of the DNA to generate the nicked complex. Separation of the cohesive ends, by terrninase, leads to formation of complex I in which terminase is tightly bound to the left chromosome end (SIPW and FEISS 1992; K ~ Z M I N O V et nl. 1994). Terminase then binds a prohead, the empty protein shell into which DNA is to be packaged, and DNA packaging ensues. Transloca- tion of DNA requires extensive ATP hydrolysis, perhaps by terminase or some other component of the translocation complex. When the translocation complex en- counters the downstream cos along the concatemer, cosN nicking and cohesive and separation complete packaging of a A chromosome. Unlike R3, R2, and R1, R4 is not required for initiation of DNA packaging. The Cto-T transition mutation (C,,,,,,T), when placed in R4, was found not to affect initiation of DNA packaging, but rather to cause a partial defect in cleavage of the

Termination of A DNA Packaging 9

downstream cos site, i e . , to affect termination of DNA packaging (CUE and FEISS 1993b). Furthermore, a 1 4 bp deletion of R4 was found to not affect initiation of DNA packaging and to cause a lethal defect in terminal cos cleavage (CUE and FEISS 199313). These results, to- gether with the inability of gpNul to bind R4, suggested a different function for R4, and the site was renamed cosQ. The R4 point mutation, C48,477T, was renamed cos@, and the 14bp deletion was named the AcosQ mutation (CUE and FEISS 199313). Further work shows that the cos cleavage defect of cosQmutants is the failure to nick the bottom strand of COSN during termination (D. CUE and M. FEISS, unpublished results). In the present work, we describe additional experiments indicating that cosQfunction is distinct from that of cosB, and describe the nature of suppressors of the cos@ mutation.

MATERIALS AND METHODS

Media: Luna Broth (LB), Luna agar (LA), 2 X YT broth and SOB were prepared as described in SAMBROOK et al. (1989). When required, kanamycin and ampicillin were used at final concentrations of 50 and 100 pg/ml, respectively. X- gal plates were LA plus isopropyl-@-Bthiogalactopyranoside (IPTG) and X-gal at final concentrations of 0.1 mM and 0.02% (w/v) , respectively.

S t r h The standard A strain used was A-P1:5R cB57 Kn' nin5; this phage will be referred to as A or A cos+ in the text. A-P1:5R cB57 nin5 carries a 10-kb segment of phage P1 DNA encoding functions for plasmid replication and partitioning (STERNBERG and AUSTIN 1983). The A-P1:5R prophage repli- cates as a single copy plasmid using the P1 replication machinery. Upon inactivation of the cB57 repressor at 42", the prophage is induced to carry out the A lytic cycle. A derivative of A-P1:5R ~1857 nin5 carrying a 1.3-kb kanamycin-resistance cassette was employed in our studies; this phage, A-P1:5R cB57 nin5 Kn' has a genome size of -46.2 kb (PAL and CHATTORAJ 1988). Mutant derivatives of A-P1:5R cR357 nin5 Kn' were described earlier (CUE and FEISS 1992a,b, 1993a,b) or constructed for this work using standard genetic methods (ARBER et al. 1983). The standard bacterial hosts were MF1427, a galK derivative of the Escherichia coli C strain Cla (SIX and KLUC 1973). MF1427 is IHF+, i.e., produces integration host factor. IHF- strains MF1493 and MF1972 are h i d - and himA- hip- derivatives of MF1427, respectively (CUE .and FEISS 1992a). MF844 is an hdR2 K-12 strain.

Sequence designations and nomenclature: All references to A sequences are based on the numbering convention described in DANIELS et al. (1983). Numbering of the A sequence begins with the first base of the left cohesive end and contin- ues along the 1 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. The single letter des- ignation for amino acids is used.

General recombinant techniques: Enzymes for recombinant DNA manipulations were purchased and used according to the manufacturer's recommendations. DNA sequencing (SANGER et al. 1977) was performed with reagents purchased from Promega and Pharmacia. Plasmid DNA and M13 replica- tive-form DNA were prepared as described by BIRNBOIM and DOLY (1979). DNA fragments were purified from agarose gels as described by VOGELSTEIN and GILLESP~E (1979). Prepara- tion of competent cells and transformation were performed as described by HANAHAN (1983). For cloning segments of A

DNA, the commercial vectors pIBI3O and pIBI3l were used (IBI, Inc.). Single-stranded DNA from pIBI plasmids was prepared as described by VIEIRA and MESSING (1987).

Phage crosses: Plasmids bearing DNA fragments cloned from A-P1:5R cB57 nin5 Kn' or one of the phage variants were introduced, by transformation, into MF1427 lysogenic for a mutant prophage. Transformed lysogens were selected by plating at 31" on LA containing kanamycin and ampicillin. The transformed lysogens were grown overnight in LB plus antibiotics at 31"; the overnight cultures were diluted 1 : lOO into LB and grown to -2-4 X lo' cells/ml at 31". Prophages were induced by incubation at 42" for 20 min, followed by incubation at 37" for 60 min. The lysates were treated with CHCls and titered on MF1427 and/or MF1972 as indicated. TB agar and TBSA were used for phage platings; incubations were at 37".

Phage yield determinations: MF1427 (IHF+) or MF1972 (IHF-), lysogenizedwith A-P1:5R cB57 nin5 Kn' or a derivative was grown overnight with aeration in LB plus kanamycin at 31". The cultures were diluted into LB (1:150 dilution) and grown to -2 x lo7 cell/ml; portions of each culture were removed at this time, diluted and plated on LB plus kanamycin; the plates were incubated overnight at 31" to determine the viable lysogen titer in each culture. Prophages were induced and lysates titered as described above.

RESULTS

Test of suppression of the msQl mutation by Nul missense mutations: Pseudorevertants of mutant phages contain suppressor mutations that provide information about the nature of the original defect. In previous studies, Nul missense mutations that were second-site suppressors of cosB mutations were described (GRANS TON et al. 1988; CUE and FEISS 199213). The Nulmsl and Nulms2 mutations were isolated as suppressors of the lethal COSB R3-R2-R1- mutation; these Nul mutations increased the burst size of the R33WR1- phage to a level that closely approximates the burst of A+ in an IHF' host. Plaque formation by the A cosB R3-R2-R1- Nulms phages was dependent upon IHF, however. The Nulmsl and Nulms2 mutations cause the changes L40F and LNI in gpNul, respectively (CUE and FEISS 1992b). The effects of the Nulms suppressors on the cos@ mutation were examined to see if the cos@ mutation is suppressed by the Nulms suppressors in a manner analogous to the R3, R2 and R1 mutations. To quantitate the effects of the Nulmsl and Nulms2 mutations on the burst of A cos@, we compared the bursts of cos@ and cos@ phages that were Nul+, Nulmsl and Nulms2 (Table 1). Our standard A phage, A-P1:5R cB57 nin5 Kn', hereafter simply called A, is a kanamycin-transducing phage that carries the plasmid replication and parti- tion functions of phage P1 (STERNBERG and AUSTIN, 1983; PAL and CHATTORAJ, 1988).

For A cos+ in the IHF- host, the Nulmsl and Nulms2 mutations increase the yields of progeny phage 3.5-fold and greater than fourfold, respectively. The bursts of the Nulmsl and Nulms2 phages, on the IHF- host, are approximately equal to the burst of the wild-type phage on the IHF+ host. In contrast, the burst of A COS+ in an


TABLE 1

Burst sizes of Nul' and Nulms derivatives of A cos" and A coal

Phage Yield on IHF' host Relative yield Yield on IHF- host Relative yield

A cos+ 104.8 (1 1.5) =1.0 27.3 (6.1) =1.0 A cos+ Nulmsl 46.5 (13.9) 0.44 95.7 (0.5) 3.5 A cos+ Nulms2 54.9 (16.7) 0.52 118.0 (4.0) 4.3 A cos@ 6.3 (0.06) 0.06 0.8 (0.2) 0.03 A cos@ Nulmsl 3.6 (0.3) 0.03 3.0 (0.8) 0.11 A cos@ Nulms2 3.5 (0.7) 0.03 4.8 (0.18) 0.18

Numbers in parentheses are standard deviations.

IHF+ host is reduced about twofold by the Nulmsl and ms2 mutations. The Nul missense mutations affect the burst of the cos@ phage in a manner that closely parallels the effects the Nul mutations have on the cos+ phage; under IHF- conditions, the Nul mutations increase the burst of the cos@ phage to an extent sufficient for plaque formation. Under IHF' conditions, as with the cos+ phage, the Nul mutations lead to a small decrease in burst size.

The cos@ phage responds to the Nul missense mutations in the same manner and to nearly the same extent as the cos+ phage. This is the result that might be antici- pated if the Nul mutations affect the terminase-cosB interaction without affecting the terminase-cosQinterac- tion; both A' and A cos@ have wild-type cosB regions. These experiments indicate that the Nul missense mutations, which were isolated as suppressors of cosB mutations, do not directly suppress the cos@ mutation. We conclude that either terminase binding to c o s 4 occurs via a different DNA recognition determinant than does terminase binding to R1, R2 and R3 and/or that the cosQsequence is required at a different step in the DNA packaging process than is COSB.

Suppression of the cosQ1 mutation by a second cos4 mutation: To see what sorts of mutations would allow suppression of the cos@ mutation, IHF-independent variants of the cos@ mutant were selected for by plating A COS@ on the IHF- host, MF1493. Variants of the cos@ mutant were found at a frequency of - lop7. Twenty-five of these variants were isolated and replated on MF1427 (IHF') and MF1972 (IHF-); the plating characteristics of this group of variants were very uniform; all formed plaques that were approximately wild type in size on the IHF+ host and moderate-sized on the IHF- host.

Two of these variants, RevlO and Rev24 of X cos@, were further characterized. Small plasmid libraries were constructed and marker rescue experiments were performed to map the suppressors. The results indicated that DNA segments carrying the EOSQ site were able to rescue the cos@ mutation (Table 2).

The sequences of the RevlO and Rev24 DNAs were determined from bp 48,299 to bp 194; RevlO and Rev24 were found to have retained the cos@ mutation (C48,477T) and to have an additional Gto-A transition

at bp 48,473 (also within the cosQsequence). We desig- nate the G48,47& change as the cos@ mutation. The net effect of the two mutations is to alter the wild-type COSQ

sequence from 5'-CGGGTCCTTTCC-3' (top strand) to

The bur;& of the RevlO and Rev24 ( X cos41 cos@) phages were found to be -67 pfu/induced lysogen under IHF' conditions and 14 pfu/induced lysogen in the absence of IHF, as compared with bursts of 6.3 pfu/induced IHF+ lysogen and 0.8 pfu/induced IHF- lysogen for A cos@.

Suppression of the cosQ1 mutation by an increase in chromosome length: During the course of performing the marker rescue experiments described above, it ap- peared that A DNA fragments that were cloned from the cosQl RevlO and cos@ Rev24 phages carried mutations that were responsible for the suppression of the cos@ mutation. These DNA fragments, corresponding to the wild-type A bp 17,791-19,996, carried the X lorn gene and portions of the Jand of401 coding sequences. When these plasmids were crossed with the cos@ phage, recombinant phages that formed large plaques on MF1427 were readily obtained; these recombinant

TABLE 2

Mapping the suppressor in Rev24 of A cos@: crosses between A coal and cloned DNA segments of

A cos+ or Rev24 of A c o a l

5"CAGGTTCTTTCC-3'.

Recombinants (pfu/ml)

Cloned DNA Phage yield On IHF' On IHF- segment (pfu/ml) cells cells

A cos+ 47,942-194 4.6 X IO7 5.0 X lo5 1.4 X 10" 1-2819 1.9 x lox < lo4 < 103

Rev24 of A cos@ 44,141-194 1.0 x 10' 2.7 x 10" 1.3 x lo6 48,299-194 3.1 X lo8 1.4 X 10" 1.0 X lo5 1-2819 1.5 X lo8 < 104 < lo1

Restriction enzyme sites used to generate A DNA inserts were Hind111 (44, 141), BclI ( A bp 47,942), Him11 (48,296), EcoRI (194), and EcOo109 (2815). The end point at bp 1 was produced by filling in the left cohesive end of virion DNA with DNA polymerase.


phages, however, did not form plaques on an IHF- host (data not shown). As a control for these experiments, the correspond-

ing DNA fragment was cloned from the wild-type phage. Crosses performed between the plasmid carrying the wild-type DNA and the cos41 phage also yielded large- plaque forming recombinants. The results of these crosses indicated that while the plasmids carrying the cloned phage DNA somehow suppressed the cos@ mutation, this suppression was not due to the presence of a mutation within the DNA fragments that had been cloned from the cos@ variants.

A likely explanation for these results was that, because a portion of the A DNA segment carried by these plasmids is from the "nonessential" b region of the A genome (COURT and OPPENHEIM 1983), the plasmids would be free to integrate within the cos@ phage chromosome, increasing its length from 46.2 to 51.3 kb. Suppression of the cos@ mutation could be the result of either an increase in the length of the phage chromosome or the duplication of a A gene.

If the idea that suppression of the cos@ mutation resulted from plasmid integration was correct, then the phages forming large plaques that resulted from the crosses would be predicted to transduce ApR (the ApR determinant being carried by the plasmid vector) with high efficiency. Several of the phages that formed large plaques were isolated and the abilities of the phages to transduce ApR were tested. All of the recombinant phages that were tested were found to be capable of efficiently transducing ApR (data not shown). The ability of the recombinant phages to transduce ApR was however readily lost; when this loss occurred, the resulting phages invariably formed small plaques, indistin- guishable from plaques of the cosQl parent phage.

The results of these experiments verified that it is the insertion of the b region-bearing plasmids into the phage chromosome that results in suppression of the cosQl mutation. It seemed that the most likely effect these insertions would have is to result in an increase in the length of the cosQl phage chromosome and that suppression of the cos@ mutation was a direct result of the increase in length. The possibility remained, however, that other effects, such as the duplication of the A lorn gene or the activation of a cryptic gene by the insertions, could be responsible for suppression of the cos@ mutation.

To test whether the suppression effect was directly due to an increase in chromosome length, plasmids were constructed that carried various subfragments of the 17,791-19,996 A DNA fragment and the abilities of these plasmids to suppress the cos@ mutation were tested. The inserts carried by these plasmids, the sizes of the plasmids and the results of the crosses with the cos@ phage are listed in Table 3.

Plasmid insert A, in Table 3, is the original 17,791- 19,996 cloned fragment; inserts B-E represent various

TABLE 3

Suppression of the msQl mutation by plasmid integration

Plasmid Total Recombinants Plasmid insert" size (pfu/ml) (pfu/ml)

1 7 I Zom I4OI'l 5,131 8.76 X 10' 1.88 X lo6

pJ 3,477 2.46 X 107 < lo3

4,319 1.36 X lo7 5.25 X lo5

3,484 1.07 X 10' 3.6 X lo5

[3110'1 3,938 1.51 X lo9 2.78 X lo6

No insert 2,923 1.00 X 10' < lo3

"Diagrams indicate extents of A genes cloned into pIBI31. Apostrophes indicate gene truncations. The coordinates of the inserts are as follows: A, 17,791 (MZuI) to 19,996 (MZuI); B, 18,556 (KpnI) to 19,996 (MZuI); C, 17,791 (MluI) to 18,385 (EcoRV); D, 19,397 (SmaI) to 19,996 (MluI); E, 18,385 (EcoRV) to 19,397 (SmaI).

subcloned segments of insert A. Crosses performed with plasmids carrying inserts B, D and E also resulted in the recovery of recombinant phages that formed large plaques on IHF+ E. coli. Plasmid inserts D and E carry the 19,400-19,996 and 18,387-19,399 A DNA segments, respectively; these two plasmids do not carry any common A DNA sequences, nor does either plasmid code for any known A protein, yet both plasmids are capable of suppressing the cos@ mutation. Therefore, the likeliest explanation for the suppression of the cos@ mutation is that the suppression is due to the integration of the plasmid into the phage chromosome; the resulting increase in chromosome length is directly responsible for suppression of the cosQ1 mutation. The plasmid carrying insert C (Table 3) would not be ex- pected to lead to the isolation of recombinant phages, since the integration of the plasmid carrying this insert would result in the disruption of the essential A Jgene. Additional studies showed that the cos@ mutation was suppressed by insertions elsewhere in the chromosome, i e . , at the extreme left and right ends (D. CUE, unpublished observations).

The fact that the cos@ mutation is partially s u p pressed by the insertion of DNA within the central region of the A chromosome, suggested that the cos@ mutation has an effect on the terminal step of DNA packaging (see DISCUSSION). The question arises then as to how a COSB mutant would respond to an increase in chromosome length.

To test the response of a cosB mutant phage to an increase in chromosome length, a cross was performed between the b region-bearing plasmid and A cosB R3-R2-R1- Nulms3. This phage was chosen because it


TABLE 4

Mapping the suppressor in Rev29 of A cosQ1: mosses between A coal and cloned DNA segments of A

cos+ or Rev29 of A cosQ1

Cloned DNA Phage yield Recombinants segment (pfu/ml) (pfu/ml)

X cos+ DNA 458-5548 4.4 X lo7 < lo4 47,942-194 9.2 X lo7 1.9 x lo5

Rev29 of X cosQl 458-5548 3.0 X 10' 2.4 X lo6 458-4470 1.0 x lo* 1.0 x lo4 2815-5548 2.6 X 10' 9.5 X lo4 2815-4198" 3.3 x 10' < lo4 3800-5548 3.5 x 108 4.0 X lo5 2815-4470 4.6 X lo* 3.0 x lo4 458-2560 8.45 X lo8 < 104 458-2815 7.0 x lo8 < lo4 4198-5548 3.3 x 108 < lo4

Restriction enzyme sites used to generate A DNA inserts, in addition to those given in Table 2 were as follows: MZuI (458), Psi1 (2560), RsdI (3800), CluI (4198), AvuI (4470), and MZuI (5548).

a Although the segment from 2815 to 4198 carries the Bms6 mutation, no recombinant phage was isolated. This likely resulted from the fact that the mutation at bp 4184 is only 17 bp from the end of the cloned DNA fragment.

has a phenotype that is very similar to that of the cos41 mutant; it forms small plaques on an IHF+ host and does not make plaques on IHF- cells (CUE and FEW 1992b). The Nulms3 mutation causes a Qq7K change in gpNul (CUE and FEISS 1992b). The cross resulted in the isolation of a significant number of ApR transducing particles, but no large-plaque forming phages were ob- served (data not shown). This result argues that increasing the chromosome length of a cosB mutant phage does not result ih any attenuation of the effect of the mutation. It thus appears that this type of suppression is specific for the cosy1 mutation.

Suppression of the cosQ1 mutation by a portal protein mutation: From a search for additional revertants of X cosQ1, a novel variant, Rev29, was found that formed small plaques on IHF+ cells (MF1427), in contrast to the tiny plaques formed by X cosQl. A library of the Rev29 DNA was prepared and used in crosses with X cos@ to determine whether there was a suppressor and, if so, to locate the suppressor (Table 4). Initial crosses demonstrated the presence of a suppressor in the MZuI segment extending from 458 to 5548, and further studies located the suppressor to a segment extending from 3800 to 4470. This segment was sequenced and a single mutation, Bms6, was identified, a C-to-T transition at bp 4184 that changes codon 450 of the B head assembly gene from serine to phenylalanine. Gene B encodes the portal protein, which makes up the portal vertex of the prohead. The portal vertex is thought to be the vertex at which DNA enters and exits the head shell, and is

TABLE 5

Suppression of the cosQ1 mutation by mutations affecting the portal protein

Yield Yield (phageshnduced (phages/induced

Phage IHF' cell) IHF- cell)

X cosQ+ B' 83 (10.5) 66.8 (11) X cosQ' Bms6 51.1 (8.0) 66.5 (18) X cosQ+ Bms8 47.2 (12.8) 56.7 (12.5) X cosQl B+ 5.4 (1.4) 2.2 (1.6) X cos@ Bms6 17.3 (1.0) 7.7 (4.0) X cosQ1 Bms8 9.7 (3.9) 4.9 (2.8)

Numbers in parentheses are standard deviations.

the site of tail attachment (BAZINET and KING 1985). Recently another missense mutation was identified in B as a suppressor of a mutation altering the prohead binding domain of terminase (YEO and FEES 1995). This mutation, Bms8, is a C-to-T transition mutation at bp 3826 that changes codon 331 of the B gene from proline to serine. In crosses analogous to those used to map the Rev29 mutation, the PSs1S change in gpB was also found to be a suppressor of the cos@ mutation. [In the reciprocal experiment, the Bms6 mutation was found not to suppress the terminase mutation affecting the prohead binding domain of gpA (A. YEO, personal communication).] The effects of the B m s suppressors are mild; the burst size data of Table 5 show that Bms6 and Bms8 increase the yield about three- and twofold, respectively, both in the presence and absence of IHF. The weakness of the Bms8 suppression may be due in part to the fact that the B m s 8 mutation has been shown to cause a threefold decrease in the efficiency of interaction of B m s 8 proheads with wild-type terminase (YEO

and FEW 1995). Additional crosses showed that the Bms6 suppressor was unable to suppress the IHF-dependence of X cosB R1-, X cosB R3-, and X cosB Rl-R2-R3- Nulms3; nor did the Bms6 suppressor allow A cosB R3-RTRl- or X AcosQ to form plaques on IHF+ or IHF- cells.

Suppression of the defects of h cosQ1 cosB R3-R2-R1- by the Nulmsl and Bms6 suppressors: Finding the cos@ mutation is suppressed by suppressors distinctly different from suppressors of COSB mutations makes the prediction that a phage with both cosy and cosB defects would require two suppressors, one for the cosQ defect and one for the cosB defect. This prediction was confirmed by examining the suppression requirements of X cos@ cosB R3-R2-R1-. X COS@ COSB R3-R2-RlP was crossed with a series of plasmids containing inserts and the resulting lysates examined for viable recombinants. Crossing X cos@ cosB R3-R2-R1- with a plasmid carrying an insert extending from 194 to 5505 and containing the Nulmsl and Bms6 mutations resulted in viable recombinants that formed tiny plaques, whereas crosses us. plasmid inserts lacking one

Termination of A DNA Packaging

TABLE 6

Suppression of the cosQ and COSR mutations by the h s 6 and Nulmsl suppressors

A insert Titer on MF844 Phage Plasmid A insert markers (pfu/ml)

A cos@ cosB R1- R2- R3- pSXl 47,942-194 cosQ+ cosB+ 1.4 X lo5 A cos@ cosB R1- R2- R3- pSXlO2 194-5505 Nul' - B+ <10 A cos@ cosB R1- R2- R3- pRVl22 458-5548 Bms6 <10 A COS@ cosB R1- R2- R3- pRV5 194-2819 Nulmsl <10 A cos@ cosB R1- R2- R3- pRV181 194-5505 Nulmsl Bms6 1.6 X lo3"

"These A cosQ1 COSB R1- R2- R3- Nulmsl Bms6 recombinants formed tiny plaques on MF844 and were

13

unable to form plaques on MF1427.

or both of the suppressor mutations failed to generate recombinants (Table 6). An additional study showed that the defects of h cosQl cosB R3-RTRl- could be suppressed with a combination of the Nulmsl suppressor and a chromosomal insertion (D. CUE, unpublished observations).

DISCUSSION

Evidence that cosQand cosBperfom distinctly different roles in X DNA packaging: An abundance of genetic and molecular evidence indicates that cosQ and cosB play distinctly different roles in X DNA packaging reactions. COSB stimulates the efficiency and accuracy of nicking of cosN and is essential for initiation of DNA packaging (CUE and FEISS 1992a,b, 1993a; HIGGINS and BECKER 1988, 1994a,b). Mutations of cosB that are quite defective for packaging initiation do not affect termination of packaging. cos4 is not required for initiation of DNA packaging, but is required for termination of packaging (CUE and FEISS 1993b). GpNul binds the three R sequences of cosB, but not cos4 (SHINDER and GOLD 1988). Here we present further genetic evidence, derived from suppressors, that cosQand cosB play different roles in DNA packaging.

The Nulmsl and Nulms2 mutations were isolated as suppressors of cosB mutations (CUE and FEISS 1992b). The Nulmsl and Nulms2 mutations are strong suppressors of all of the cosB mutations tested; both mutations permit plaque formation by a phage, X cosB R3-R2-RlP, carrying a transition mutation in each of the R sequences of cosB and promote IHF-independent growth by all IHF-dependent COSB mutants that have been tested.

Despite the fact that the Nulmsl and ms2 mutations function as efficient suppressors of cosB mutations, neither of these Nul mutations has a large effect on the plating characteristics of the cos@ phage. The cosQ1 mutant, like the COS+ phage, has a wild-type c o d region and responds to the terminase mutations in a manner that closely parallels the response of the cos+ phage; the Nulmsl and Nulms2 mutations both decrease the burst of the cos@ phage about twofold on an IHF' host and

increase the phage burst several fold on an IHF- host. The increase in burst size is sufficient to allow the cos@ Nulmsl and Nulms2 phages to form plaques in the absence of IHF, but the ability of the Nul missense mutations to allow IHF-independent plaque formation by the cos@ mutant cannot be regarded as direct suppression, but rather reflects the response of a cosB+ phage to the Nul mutations. W h y the Nulmsl and Nulms2 suppressors decrease the burst size of both cos@ and cos+ phages is not clear; it has been speculated that the interaction of the suppressor gpNul with cosB may be strengthened such that a necessary disas- sembly step is less efficient (CUE and FEISS 1992b).

In contrast, the Nulmsl and Nulms2 mutations increase the burst of X COSB R3-R2-R1- phage nearly 1000-fold on an IHF' host. These same Nul mutations also increase the burst of the R3-R2-RlP phage on an IHF- host by -100-fold, an increase insufficient to allow plaque formation (CUE and FEISS 1992b). In sum, the Nul missense mutations can be regarded as specific suppressors of COSB mutations.

The fact that the Nul mutations can lead to suppression of point mutations in R3, R2 and R1 (CUE and FEISS 1992b), but do not suppress the cos@ mutation suggests that cosQ unlike R1, R2 and R3, is not a gpNul binding site. SHINDER and GOLD (1988) demonstrated that the small terminase subunit does bind to R1, R2 and R3; these same researchers (SHINDER and GOLD 1989) later reported that gpNul did not bind to COSQ

but that, somehow, COSQ may aid in gpNul binding to the other three R sites. Our results suggest a different role for cos4 in DNA packaging.

Although cos4 was originally identified by its sequence homology to the other R sequences (BEAR et al. 1984), the COSQ sequence shares homology with COSN half-sites as well as with the R sequences (Fig. 2), suggesting that terminase may recognize cosQ by the same specificity determinant that is used to recognize COSN. This suggestion is supported by our finding that the related lambdoid phage, 21, has a cosQ sequence that is identical to the X cosQ sequence (SMITH and FEISS 1993). X and 21 have terminase genes that are -60% homologous (reviewed in FEISS 1986). Although X and

14

a .

R m u t a t i o n s :

R1: 151

R2 : 124

R3 : 49

cosQ: 48468

cosQl:

b.

cosNL: 48501

cosQ: 48468

cosNR : 14

D. Cue and M. Feiss

T

CTGTCGTTTCCTTTCT 166

CTGTCGTTTCCTTTCT 109

AAGGCGTTTCCGTTCT 64

................ ................ . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . TTTACGGGTCCTTTCC 48483

i

1 CGGGGCGG 6

TTTACGGGTCCTTTCC 48483

CGAGGTCG 7

. . . . . . . . . .

. . . . . . . .

1 FIGURE 2,”Sequence relationships of COSQ cosNand the R

sites of COSB. (a) Alignment cosQwith the R sequences of COSB. Sequences are numbered according to the standard h system; note that the R2 sequence is that of the bottom strand. The site of the cos154 mutation of R1 (Rl-; BEAR et al. 1984) is in bold; this bp, conserved in R2, R3, and COSQ is the bp changed to generate the R2-, R3-, and cos@ mutations are also in bold (CUE and FEISS 1992a). Deletion and point mutation studies confine cos4 to the segment from 48,471 through 48,485 (HOHN 1983; D. WIECZOREK and M. FEISS, unpublished observations). (b) Alignment of the cos4 sequence and the cosN half-sites. The cosQ sequence is aligned with cosNL (above), the left half-site of COSN, and cosNR (below), the right halfsite of cosN. Note that the cosNR sequence is the bottom strand sequence. Arrows indicate the N2 nick site in cosNL and the N1 nick site in cosNR. The cosN half-site sequences shown are restricted to the segment of twofold rotational symmetry (bp 48501 to 14).

21 have cosB regions that are similar in size and organi- zation (MILLER and FEISS 1985), the COSB sequences have diverged sufficiently so that neither A nor 21 terminase efficiently recognizes the cosB of the other phage for initiation of DNA packaging (FEISS et al. 1981; FEISS and WIDNER 1982). These divergent packaging specificities are reflected in the divergence of the R1, R2 and R3 sites between the two phages and between the putative helix-turn-helix DNA-binding motifs of A gpNul and 21 gpl (MILLER and FEES 1985; FEISS 1986; BECKER and MURIALDO 1990). The fact that the cosN and cosQ sequences of A and 21 are identical, whereas the cosB sequences have diverged, also argues that cosQmay not be a gpNul binding site. We suggest that COSQ, like COSN is bound by gpA, the terminase subunit with the endonuclease activity.

Suppression of the cosQ1 mutation by a second cosQ

mutation: The cosQ1 mutation (C48,477T) is suppressed by a second mutation within COSQ i.e., by the cos@ mutation (G48,473A). While the cos@ mutation is proposed to restore cosQfunction, we have no information on how the cos@ or cos@ mutations affect cosQfunc- tion. The speculative proposal that cosQ and cosN half- sites are recognized similarly leads to an examination of the effects of the cosQmutations on sequence identity between COSQ and the COSN half-sites (Figure 2b). cos4 has 5-bp matches when aligned with the 8-bp of cosNL, and the cos@ and cos@ mutations each eliminate a bp match, SO that cosQwith the cosQ1 and cos@ mutations has only a 3-bp match with cosNL. cosQhas 4bp matches when aligned with the 8-bp of cosNR. The cos@ mutation increases the cosQJcosNR match to 5 bp, and the cosQ site with both the cos@ and cos@ mutations has a match of 4 bp. Thus, in suppressing cosQl, the cos@ mutation restores the number of matching bp to the wild-type value of 4. It is possible that the bp match between wild-type cosQ and cosNR is optimal for cos4 function, and that optimal matching is restored in the cos@ cos@ doubly mutant site. It is hoped these highly speculative suggestions will encourage further study into cosQ recognition. It has recently been found that a lethal cosQ mutation blocks nicking of cosNR, while having little effect on cosNL nicking, so there is evidence of a functional link between cosQand cosNR (D. CUE and M. FEISS, unpublished results).

Suppression of the cosQ1 mutation by DNA insertions: The effect of the cos@ mutation has been found to be largely suppressed by the insertion of DNA segments within the chromosome of the cos@ phage; plasmids carrying a portion of the A b region can integrate within the phage genome and promote the formation of near wild-type size plaques on an IHF+ host. The A b region is an 8.6-kb DNA segment in the central region of the A genome, between the A J gene and the A att region (COURT and OPPENHEIM 1983). Although the b region encodes for several proteins, the region contains no genes that are essential for A growth (HENDRIX and DUDA 1992).

Since the b region is nonessential, plasmids that carry segments of this region can integrate within the phage chromosome without disrupting essential genes. The net effect of the integration of the plasmids listed in Table 3 is to increase the length of the A cos@ chromosome from 46.2 kb to -50-51 kb. The experiments summarized above demonstrate that plasmid integration is both necessary and sufficient for suppression of the cos41 mutation. The results presented in Table 3 demonstrate that a variety of plasmids can suppress the cosQl mutation, despite the fact that the plasmids do not carry any common A DNA sequences or any intact A genes; thus the suppression effect is the direct result of an increase in chromosome length. A chromosome insertion does not result in an enhanced plaque-forming ability for the COSB mutant A cosB R3-R2-R1-


Nulms3, a phage that forms plaques that are of approximately the same size as X cos@ plaques. Thus, the ability of DNA insertions to suppress cos mutations appears to be specific for the cos@ mutant. The fact that an insertion in the central portion of the X chromosome can suppress the effects of the cosQl mutation argues that the cosQ1 mutation affects the termination of DNA packaging. The termination of X DNA packaging, or the cutting of the downstream cos site, is influenced by chromosome length and is independent of cosB specificity, i.e., X us. 21 (FEISS et al. 1981; FEISS and WIDNER 1982; FEISS and BECKER 1983). FEISS and BECKER (1983) speculated about the molecular explanation for the effect of chromosome length on terminal cos cleavage. One proposed model states that the velocity of DNA entry into the phage head is dependent upon DNA length and that the rate of DNA packaging slows progressively as increasing amounts of DNA are packaged into the head; the ability of terminase to cut at the terminal cos could be inversely dependent on the packaging rate. There is some in vitro evidence to support the idea that the rate of DNA packaging does decrease as packaging proceeds (SHIBATA et al. 1987). The cos@ mutant has a chromosome that is 46.2 kb in length, a size that would not be predicted to influence the efficiency of DNA packaging (FEISS et al. 1977; FEISS and SIEGELE 1979), yet insertions of 3.5-5.1 kb of DNA result in the partial suppression of the cos41 mutation. A reasonable explanation for the cos@ defect is that the cos@ mutation interferes with the ability of terminase to cut the terminal cos site. Increasing the size of the cos@ chromosome from 46 to 250 kb might suppress the cos@ defect by slowing DNA packaging to a rate such that terminase can recognize the mutant cosQand cut the terminal cos site.

Suppression of the cos@ mutation by DNA insertions is heavily dependent upon IHF. The recombinant cos@ phages with DNA insertions form plaques that are nearly wild type in size on an IHF+ host, but fail to form plaques on an IHF- host. Thus, IHF contributes to suppression of the cos@ mutation. Apparently, neither an increase in chromosome size nor the presence of IHF alone is sufficient to alter the phenotype of A

Earlier work demonstrated that the presence of wild- type cosB is not required for either terminal cos cleavage or processive DNA packaging (FEES et al. 1981, 1985; FEISS and WIDNER 1982). The model proposed (FEISS et al. 1985) as a result of these earlier experiments is as follows. Packaging is initiated at a cosB+ cos site, and DNA translocation delivers terminase to the terminal cos site. The terminal COSN is directly recognized and cut by terminase. Following the packaging of the initial cosB+ chromosome, terminase remains bound to the left end of the next chromosome, binds a prohead and initiates packaging of the next chromosome. Although A terminase requires cosB of A for initiation of packag-

cos@.

ing, termination and processivity are carried out at a downstream COSBX or cosB421, indicating that the sequence requirements for termination and processivity differ from those required for packaging initiation (FEISS and WIDNER 1982; FEISS et al. 1985). More recent results established that cosQis required for the recognition of terminal cos sites (CUE and FEISS 199313); the IHFdependence of suppession of cos@ by DNA insertion suggests that cosB may also play a role in packaging termination.

While there are a number of possible mechanisms whereby IHF could contribute to cos@ suppression, we favor a model where IHF binding within cosB affects the efficiency of packaging termination. We have recently found that a downstream cos consisting of cosQand cosN is inadequate to support efficient packaging termination. Rather, we have found that the sequences within cosB function with cosQ and cosN to promote packaging termination and processivity. Thus it is possible that IHF functions similarly during packaging termination and initiation.

We previously reported (CUE and FEISS 1992b) that IHF contributes little to suppression of the cosQ1 mutation. While these results seem to be at odds with the results reported here, it seems likely that IHF plays only a minor role in termination and processive packaging of cos+ chromosomes. It may be only under specific circumstances (i .e. , packaging of X cosQ1 chromosomes that are >50 kb in length) that the contributory role is apparent.

Suppression of the cosQ1 mutation by a portal protein mutation: The product of the B gene of X is the portal protein, which assembles into a gear-shaped dodecamer that serves as the site for assembly of the shell of the prohead and forms the portal vertex of the prohead (KOCHAN and MURIALDO 1983; KOCHAN et al. 1984; reviewed in GEORGOPOULOS et al. 1983). GpB is predicted to be a protein of 533 amino acids (SANGER et al. 1982). During prohead assembly, a majority of the gpB is processed to gpB* by removal of the N-terminal 20 amino acids (CASJENS and HENDRIX 1974; WALKER et al. 1982). In addition to serving as the site of prohead shell assembly, the portal vertex is the site of tail attachment (TSUI and HENDRIX 1980). The portal vertex interacts with the C-terminus of the large terminase subunit, gpA, during DNA packaging (YEO and FEISS 1995) and is likely the DNA entry and exit site during packaging and injection, respectively (reviewed in VALPUESTA and CARRASCOSA 1994; CATALAN0 et al. 1995). The gpB dodecamer has a central hole sufficiently large to allow DNA passage, and a number of DNA packaging models include packaging of the DNA through this central hole (HENDRIX 1978; EARNSHAW and CASJENS 1980; DUBE et al. 1993), although other models involve wrapping of the DNA around the outside of the structure (TURN- QUIST et al. 1992). There is evidence that the portal vertex is intimately involved in the DNA packaging pro-


cess, as follows. Phages P22 and SPPl are phages that package DNA by a headful mechanism: following an initial, site-specific cut, chromosomes are cut from the concatemeric precursor DNA by nonspecific cuts that appear to be triggered by filling of the head. Mutations affecting the portal proteins of P22 and SPPl have been found that increase and decrease the length of the packaged DNA molecules, respectively (CASJENS et al. 1992; TAVARES et al. 1992). These mutations indicate a role for the portal vertex in sensing the amount of DNA that has been packaged, by sensing (1) the packaging density, (2) the rate of DNA translocation, or (3) the energy required to continue packaging (CASJENS et al. 1992).

For A, there is a length-sensing aspect to termination of DNA packaging, in the cleavage of the downstream cos becomes progressively less efficient as the chromosome is shortened. That is, over decreasing chromosome lengths from 0.90 (43.6 kb) to 0.78 (37.8 kb) that of wild type (48.5 kb), the efficiency of terminal cos cleavage declines from 100% to -75% (FEISS and SIEGELE 1979). It seems likely that the A portal protein, like those of P22 and SPP1, may play a role in sensing how much DNA has been packaged into the head. Here we find that mutations affecting the A portal protein suppress the defect of a leaky cos4 mutation. We con- sider two models for how changes in the portal protein might affect COSQ recognition.

In the first model it is proposed that the sensing mechanism measures the rate of DNA translocation, and that Bms6 and &ns8 slow the rate of translocation, thereby increasing the efficiency at which the weakened cos@ site is recognized by the translocation complex. While speculative, this proposal fits well with the obser- vation that increased DNA length also suppresses the cos@ mutation, since increased DNA length might also alter the rate of translocation at the time cos@ is en- countered by the translocation machinery.

In a second model, the B suppressor mutations are proposed to alter the sensing mechanism per se. In this model, again assuming that the sensing mechanism measures the rate of translocation, the B mutations are proposed to alter the threshold of the sensing mechanism, so that the transition to the cleavage mode occurs at a higher translocation rate, such that the cos@ site is recognized more efficiently. The first model predicts that the B suppressors affect the translocation rate, while the second model does not. We plan to test whether the B suppressors alter the rate of DNA translocation.

We thank ANDY BECKER, CARLOS CATALANO, HISAO FUJISAWA and our coworkers Z H I - ~ O CAI, CAROL DUFN, QI HANG, HILLARY JOHN-

SON, JENNY MEYER, JOHN RANDELI., JEAN SIPPY, MICHAEI. SMITH, FEODOR TERESHCHENKO, DOUG WIECZOREK and ASHI.Y YEO for advice and interest during the course of this work. This work was supported by National Institutes of Health research grants AI-12581 and GM- 51611.

LITERATURE CITED

ARBER, W., L. ENQUIST, B. HOHN, N. E. MURRAY and K MURRAY, 1983 Experimental methods for use with lambda, pp. 433-466 in Lambda II, edited by R. W. HENDRIX, J. W. ROBERTS, F. W. STAHL and R. A. WEISBERG. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N Y .

BAZINET, C., and J. KING, 1985 The DNA translocating vertex of DS DNA bacteriophage. Annu. Rev. Microbiol. 3 9 109-129.

BEAR, S., D. COURT and D. I. FRIEDMAN, 1984 An accessory role for integration host factor: characterization of a lambda mutant

Virol. 52: 966-972. dependent upon integration host factor for DNA packaging. J.

BECKER, A., and M. GOLD, 1988 Prediction of an ATP reactive center in the small subunit, gpNul, of the phage lambda terminase enzyme. J. Mol. Biol. 199: 219-222.

BECKER, A., and H. MURIALDO, 1990 Bacteriophage A DNA the beginning of the end. J. Bacteriol. 172: 2819-2824.

BIRNBOIM, H. C., and J. DOLY, 1979 A rapid alkaline extraction pro- cedure for screening recombinant plasmid DNA. Nucleic Acids Res. 7: 1513-1523.

CAI, Z-H.,Y. HWANG, D. CUE, C. CATALANO, and M. FEISS, 1997 Muta- tions in Nul, the gene encoding the small subunit of bacteriophage A terminase, suppress the post-cleavage DNA packaging defect of c o d mutations. J. Bacteriol. 179: 2479-2485.

CASJENS, S., E. WCKOFF, M. HAYDEN, L. SMPSON, K. EPPLER et al., 1992 Bacteriophage P22 portal protein is part of the gauge that regulates packing density of intravirion DNA. J. Mol. Biol. 2 2 4

CASJENS, S. R., and R. W. HENDRIX, 1974 Locations and amounts of the major structural proteins in bacteriophage lambda. J. Mol. Biol. 88: 535-545.

CATALANO, C. E., D. CUE, and M. FEISS, 1995 Virus DNA packaging: The strategy used by phage A. Mol. Microbiol. 16: 1075-1086.

COURT, D. L., and A. B. OPPENHEIM, 1983 Phage lambda’s accessory genes, pp. 251-277 in Lambda II, edited by R. W. HENDRIX, J. W. ROBERTS, F. W. STAHL, and R. A. WEISBERG. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

CUE, D., and M. FEISS, 1992a Genetic analysis cosB, the binding site for terminase, the DNA packaging enzyme of bacteriophage A. J. Mol. Biol. 228: 58-71.

CUE, D., and M. FEISS, 1992b Genetic analysis of mutations affecting terminase, the bacteriophage A DNA packaging enzyme, that suppress mutations in COSB, the terminase binding site. J. Mol. Biol. 228: 72-87.

CUE, D., and M. FEISS, 1993a The role of COSB, the binding site for terminase, the DNA packaging enzyme of bacteriophage A, in the nicking reaction. J. Mol. Biol. 234: 594-609.

CUE, D., and M. FEISS, 1993b A site required for termination of packaging of the phage A chromosome. Proc. Natl. Acad. Sci.

DANIELS, D., J. SCHROEDER, W. SZYBALSKI, F. SANCER, A. COULSEN et al., 1983 Complete annotated lambda sequence, pp 519-676 in Lambda II, edited by R. W. HENDRIX, J. W. ROBERTS, F. W. STAHL and R. A. WEISBERG. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N Y .

DAVIDSON, A., and M. GOLD, 1992 Mutations abolishing the endonuclease activity of bacteriophage A terminase lie in two distinct regions of the A gene, one of which may encode a “leucine zipper” DNA binding domain. Virology 1 8 9 21-30.

DUBE, P., P. TAVARES, R. LURZ and M. VAN HEEL, 1993 The portal protein of bacteriophage SPP1: a DNApump with 13-fold symme-

EARNSHAW, W. C., and S. R. CASJENS, 1980 DNA packaging by the double-stranded DNA bacteriophages. Cell 21: 319-331.

FEISS, M., 1986 Terminase and the recognition, cutting and packaging of g chromosomes. Trends Genet. 2: 100-104.

FEISS, M., and A. BECKER, 1983 DNA packaging and cutting, pp. 305-330 in Lambda II, edited by R. W. HENDRIX, J. W. ROBERTS, F. W. STAHL and R. A. WEISBERG. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

FEW, M., and D. A. SEIGELE, 1979 Packaging of the bacteriophage lambda chromosome: Dependence of cos cleavage on chromosome length. Virology 92: 190-200.

FEISS, M., and W. WIDNER, 1982 Bacteriophage A DNA packaging:

1055-1074.

USA. 90: 9290-9294.

try. EMBO J. 12 1303-1310.


Scanning for the terminal cohesive end site during packaging. Proc. Natl. Acad. Sci. USA 79: 3498-3502.

FEISS, M., R. A. FISHER, M. A. CRAY~ON and C. EGNER, 1977 Packag- ing of the bacteriophage A chromosome: Effect of chromosome length. Virology 77: 281-293.

FEISS, M., R. A. FISHER, D. A. SIECELE and W. WIDNER, 1981 Bacterio- phage lambda and 21 packaging specificities, pp 213-222 in Bacteriophage Assembly, edited by M. DEBOW. Alan Liss, New York.

FEISS, M., W. WIDNER, G. MILLER, G. JOHNSON and S. CHRISTIANSEN, 1983 Structure of the bacteriophage lambda cohesive end site: location of the site of terminase binding (cosB) and nicking (coslv) Gene 2 4 207-218.

FEISS, M., J. SIPPY and G. MILLER, 1985 Processive action of terminase during sequential packaging of bacteriophage A chromosomes. J. Mol. Biol. 186: 759-772.

FRIEDMAN, D. I., 1988 Integration host factor: a protein for all rea- sons. Cell 55: 545-554.

GEORGOPOULOS, C., K. TIL^ and S. CASJENS, 1983 Lambdoid phage head assembly, pp. 279-304 in Lambda IZ, edited by R. W. HEN- DRIX, J. W. ROBERTS, F. W. STAHL and R. A. WEISBERG. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

GRANSTON, A. E., D. M. ALESSI, L. EADES and D. FRIEDMAN, 1988 A point mutation in the N u l gene of bacteriophage A facilitates phage growth on Escherichia coli strains with himA and gyrB mutations. Mol. Gen. Genet. 212: 149-156.

HANAHAN, D., 1983 Studies on transformation of Escherichia coliwith plasmids. J. Mol. Biol. 166: 557-580.

HENDRIX, R. W., 1978 SymmetIy mismatch and DNA packaging in the large DNA bacteriophages. Proc. Natl. Acad. Sci. USA 75:

HENDRIX, R. W., and R. L. DUDA, 1992 Bacteriophage APaPa: not the mother of all A phages. Science 258: 1145-1148.

HIGGINS, R. R., and A. BECKER, 1994a The A terminase enzyme measures the point of its endonucleolytic attack 47 ? 2 bp away from its site of specific DNA binding, the R site. EMBO J. 13: 6152- 6161.

HIGGINS, R. R., and A. BECKER, 1994b Chromosome end formation in phage A, catalyzed by terminase, is controlled by two DNA elements of cos, cosN and R3, and by ATP. EMBO J. 13: 6162- 6171.

HIGGINS, R., H. LUCKO and A. BECKER, 1988 Mechanism of cos DNA cleavage by bacteriophage A terminase: multiple roles of ATP. Cell 54: 765-775.

HOHN, B., 1983 DNA sequences necessary for packaging of bacteriophage A DNA. Proc. Natl. Acad. Sci. USA. 80: 7456-7460.

K O C M , J., and H. MURIALDO, 1983 Early intermediates in bacteriophage lambda prohead formation. Virology 131: 100-115.

KOCHAN, J., J. L. CARRASCOSA, and H. MURIALDO, 1984 Bacterio-

4779-4783.

phage lambda preconnectors: purification and structure. J. Mol. Biol. 174: 433-447.

KOSTURKO, L., E. DAUB and H. MURIALDO, 1989 The interaction of E. coli integration host factor and lambda cos DNA multiple complex formation and protein-induced bending. Nucleic Acids Res. 17: 329-334.

KUZMINOV, A., E. SCHABTACH and F. W. STAHL, 1994 x sites in combination with RecA protein increase the survival of linear DNA in Escherichia coli by inactivating exoV activity of RecBCD nuclease. EMBO 1. 13: 2764-2776.

MENDELSON, I . , M. GOTTESMAN and A. OPPENHEIM, 1991 HU and integration host factor function as auxiliary proteins in cleavage of phage lambda cohesive end by termink;. J. Bacteriol. 1%: 1670-1676.

MILLER, G., and M. FEISS, 1985 Sequence of the left end of phage 21 DNA. J. Mol. Biol. 183: 239-246.

MIWA, T., and K. MATSUBARA, 1983 Determination of the DNA region necessary for packaging into lambda phage heads. Gene 24: 199-206.

MURIALDO, H., 1991 Bacteriophage lambda DNA maturation and packaging. Annu. Rev. Biochem. 60: 125-153.

PAL, S. K., and D. K. CHATTORAJ, 1988 P1 plasmid replication: initiator sequestration is inadequate to explain control by initiator- binding sites. J. Bacteriol. 172: 2819-2824.

RICE, P. A., SW. YANG, K. MIZUUCHI and H. A. NASH, 1996 Crystal

structure of an IHF-DNA complex: A protein-induced DNA U- turn. Cell 87: 1295-1306.

RUBINCHIK, S., W. PARIUS and M. GOLD, 1994 The in vitro endonuclease activity of gene product A, the large subunit of the bacteriophage A terminase, and its relationship to the endonuclease activity of the holoenzyme. J. Biol. Chem. 269 13575-13585.

SAMBROOK, J., E. F. FRITSCH and T. MANIATIS, 1989 Molecular Clon- ing; A Laboratoly Manual, Ed. 2, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

SANGER, F., A. R. COULSON, G. F. HONG, D. F. HILL and G. B. PET- ERSEN, 1982 Nucleotide sequence of bacteriophage lambda DNA. J. Mol. Biol. 162: 729-773.

SANCER, F., S. MICKLEN and A. R. COULSEN, 1977 DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA. 7 4 5463-5467.

SHIBATA, M., H. FUJISAWA and T. MINIGAWA, 1987 Early events in a defined in vitro system for packaging of bacteriophage T3 DNA. Virology 159: 250-258.

SHINDER, G., and M. GOLD, 1988 The N u l subunit of bacteriophage lambda terminase binds to specific sites in cos DNA. J. Virol. 62:

SHINDER, G., and M. GOLD, 1989 Integration host factor (IHF) simu- lates binding of the gpNul subunit of A terminase to cos DNA. Nucleic Acids Res. 17: 2005-2022.

SIPFY, J., and M. FEISS, 1992 Analysis of a mutation affecting the specificity domain for prohead binding of the bacteriophage A terminase. J. Bacteriol. 174: 850-856.

SIX, E. W., and C. A. C. KLUG, 1973 Bacteriophage P4: a satellite virus depending on a helper such as prophage P2. Virology 51: 327-344.

387-392.

SMITH, M. P., and M. FEISS, 1993 Sites and gene products involved in lambdoid phage DNApackaging. J. Bacteriol. 175: 2393-2399.

STERNBERG, N., and S. AUSTIN, 1983 Isolation and characterization of PI minireplicons, A-P1:5R and A-P1:5L. J. Bacteriol. 153: 800- 812.

TAVARES, P., M. SNVTOS, R. LURZ, G. MORELLI, H. LENCASTRE, et al., 1992 Identification of a eene in Bacillus subtilis bacteriophage SPPl determining the amGunt of packaged DNA. J. Mol: Biol. 225 81-92.

TURNQUIST, S., M. SIMON, E. ECELMAN and D. ANDERSON, 1992 Su- percoiled DNA wraps around the bacteriophage 429 head-tail connector. Proc. Natl. Acad. Sci. USA 89: 10479-10483.

TSUI, L., and R. HENDRIX, 1980 Head-tail connector of bacteriophage lambda. J. Mol. Biol. 142: 419-438.

VALPUESTA, J. M., and J. L. CARRASCOSA, 1994 Structure of viral con- nectors and their function in bacteriophage assembly and DNA packaging. Quart. Rev. Biophys. 27: 107-155.

VIEIRA, J., and J. MESSING, 1987 Production of single-stranded plasmid DNA. Methods Enzymol. 153: 3-11.

VOGELSTEIN, B., and D. GILLESPIE, 1979 Preparative and analytical purification of DNA from agarose. Proc: Natl. Acad. Sci. USA 76: 615-619.

WALKER, J. E., A. D. AUFFRET, A. WE, A. GURNETT, P. HANSICH, et al., 1982 Solid-phase sequence analysis of polypeptides eluted from polyacrylamide gels: an aid to interpretation of DNA sequences exemplified by the Escherichia coli unc operon and bacteriophage lambda. Eur. J. Biochem. 123: 253-260.

XIN, W., Z-H. CAI and M. FEISS, 1993 Function of IHF in A DNA packaging 11. Effects of mutations altering the IHF binding site and the intrinsic bend in cosB on A development. J. Mol. Biol.

XIN, W., and M. FEISS, 1988 The interaction of Escherichia coli integration host factor with the cohesive ends sites of phages A and 21. Nucleic Acids Res. 16: 2015-2029.

XIN, W., and M. FEISS, 1993 Function of IHF in A DNA packaging. I. Identification of the strong binding site for integration host factor and the locus for intrinsic bending in COSB. J. Mol. Biol.

YANISCH-PERRON, C., J. VIEIRA and J. MESSING, 1985 Improved M13 phage cloning vectors and host strains: nucleotide sequences of the M13mp18 and pUC19 vectors. Gene 33: 103-119.

YEO, A,, and M. FEISS, 1995 Specific interaction of terminase, the DNA packaging enzyme of bacteriophage A, with the portal protein of the prohead. J. Mol. Biol. 245: 141-150.

230: 505-515.

230: 492-504.

Communicating editor: R. MAURER

Documents

Genetic Evidence That Recognition of cos@ the Signal for