Defining cosQ, the Site Required for Termination of Bacteriophage DNA Packaging

Corresponding author: Douglas J. Wieczorek, Department of Microbiology, University of Iowa, 3-315 BSB, Iowa City, IA 52242., wieczorekd{at}mail.medicine.uiowa.edu (E-mail)

Communicating editor: R. MAURER

	ABSTRACT

TOP ABSTRACT MATERIAL AND METHODS RESULTS DISCUSSION LITERATURE CITED

Bacteriophage is a double-stranded DNA virus that processes concatemeric DNA into virion chromosomes by cutting at specific recognition sites termed cos. A cos is composed of three subsites: cosN, the nicking site; cosB, required for packaging initiation; and cosQ, required for termination of chromosome packaging. During packaging termination, nicking of the bottom strand of cosN depends on cosQ, suggesting that cosQ is needed to deliver terminase to the bottom strand of cosN to carry out nicking. In the present work, saturation mutagenesis showed that a 7-bp segment comprises cosQ. A proposal that cosQ function requires an optimal sequence match between cosQ and cosNR, the right cosN half-site, was tested by constructing double cosQ mutants; the behavior of the double mutants was inconsistent with the proposal. Substitutions in the 17-bp region between cosQ and cosN resulted in no major defects in chromosome packaging. Insertional mutagenesis indicated that proper spacing between cosQ and cosN is required. The lethality of integral helical insertions eliminated a model in which DNA looping enables cosQ to deliver a gpA protomer for nicking at cosN. The 7 bp of cosQ coincide exactly with the recognition sequence for the Escherichia coli restriction endonuclease, EcoO109I.

FOLLOWING DNA replication, many large, double-stranded DNA viruses, such as the herpes and pox viruses and bacteriophage , produce concatemeric DNA, end-to-end multimers of individual genomes. Virion chromosomes are generated by cutting the concatemeric substrate prior to and during DNA packaging. Specific ends are generated by viral-encoded proteins that recognize and cleave specific DNA sites on the viral chromosome. The mature -chromosome is a linear chromosome, 48.5 kb in length, with 12-bp cohesive ends at the 5' ends of the strands (SANGER et al. 1982 ). Upon injection into the cell, the cohesive ends anneal and are ligated, forming the cohesive end site (cos) and circularizing the chromosome. Early during infection, the -chromosome is replicated bidirectionally to produce a number of progeny rings. Later, rolling circle replication gives rise to concatemers (FURTH and WICKNER 1983 ).

The phage-encoded terminase enzyme is a key player in the DNA packaging process. Terminase is a heteromultimer composed of two subunits, gpNu1 and gpA (Fig 1). The genes that encode these proteins are located immediately downstream of the cos region at the left end of the mature -chromosome. The packaging strategy used by bacteriophage is as follows (FEISS 1986 ; BECKER and MURIALDO 1990 ; CATALANO et al. 1995 ): Terminase encounters a randomly chosen cos of a concatemer to initiate packaging. The cos region of bacteriophage is composed of three subsites: cosN, cosB, and cosQ (Fig 1). cosN is the site where staggered nicks are introduced by terminase to generate the cohesive ends of virion DNA and represents the junctions of individual chromosomes in the concatemer. Terminase binds the initial cos site of the concatemer through specific interactions of gpNu1 with cosB and gpA with cosN, forming the initial complex (BEAR et al. 1983 ; SHINDER and GOLD 1988 ; HIGGINS and BECKER 1995 ). Terminase's gpA subunits then introduce staggered nicks at cosN resulting in the nicked complex (DAVIDSON and GOLD 1992 ; RUBINCHIK et al. 1994A ; HWANG and FEISS 1996 ; SIPPY ARENS et al. 1999 ). In an ATP-dependent reaction, terminase separates the newly created cohesive ends and remains bound to the left, cosB-containing end, forming a stable complex, complex I (BECKER et al. 1977 ; HIGGINS et al. 1988 ; SIPPY and FEISS 1992 ; RUBINCHIK et al. 1994B ). Complex I binds the empty shell precursor, the prohead, forming the ternary complex, complex II, and ATP-dependent translocation of DNA into the prohead follows (BECKER et al. 1977 ; FRACKMAN et al. 1984 ; DAVIDSON and GOLD 1987 ; WU et al. 1988 ; SIPPY and FEISS 1992 ; YEO and FEISS 1995A , YEO and FEISS 1995B ). The terminase-containing translocation complex remains bound to the prohead until the downstream cos site on the concatemer is encountered. Terminase introduces nicks at the downstream cosN and packaging is thus terminated, a process requiring cosQ (CUE and FEISS 1993 , CUE and FEISS 1998 ). Terminase undocks from the DNA-filled head and binds to the left end of the next chromosome in the concatemer to be packaged, resulting in an assembly analogous to complex I. An average of two to three chromosomes are packaged processively by terminase (EMMONS 1974 ; FEISS et al. 1985 ). Phage construction is completed with the addition of head completion proteins and a tail to the DNA-filled head.

View larger version (19K): In this window In a new window Download PPT slide   Figure 1. Structure of cos and the terminase genes. The Nu1 and A genes encode the small (gpNu1) and large (gpA) subunits of terminase, and B encodes the portal protein. cos is a tripartite structure consisting of cosQ, required for packaging termination; cosN, the nicking site; and cosB, required for packaging initiation. The three R sequences of cosB are gpNu1 binding sites. I1 is a binding site for E. coli integration host factor. The previously identified sequences of cosQ and cosN are shown at the bottom of the figure, denoted by the boxes. The newly identified sequence of cosQ is denoted by the solid line. cosN exhibits partial twofold rotational symmetry with the center of symmetry indicated by the dot. Numbering of the -sequence and the positions of terminase nicking at cosN, N1 and N2 are also shown.

cosN contains a 22-bp element in which 10 of the base pairs show twofold rotational symmetry (FEISS and BECKER 1983 ). It is proposed that nicking is carried out by symmetrically disposed gpA subunits based on the symmetry found at cosN (BECKER and MURIALDO 1990 ). CUE and FEISS 1998 determined that the nicking of the bottom strand of DNA at cosNR is dependent on the presence of cosQ. In addition, Cue and Feiss noted sequence similarity between cosQ and cosNR and proposed that cosQ is recognized by gpA, as follows. During packaging, the terminase molecules in the packaging complex may be polarized with respect to the prohead and the DNA. The carboxy terminus of gpA is the specificity domain for binding to the portal protein of the prohead (FRACKMAN et al. 1984 ; WU et al. 1988 ; YEO and FEISS 1995A , YEO and FEISS 1995B ), suggesting terminase protomers may have their C-termini bound to the portal vertex with the rest of the protomers extending away from the portal vertex. This polarized arrangement does not match the twofold rotational symmetry found at cosN, which likely requires symmetrically disposed, or depolarized, gpA subunits. Thus, cosQ may function to depolarize a terminase protomer, enabling the protomer to interact with cosNR, since the orientation of cosNR is opposite that of cosQ. In this article, we define the genetic structure of cosQ and test several aspects of the depolarization model.

	MATERIAL AND METHODS

TOP ABSTRACT MATERIAL AND METHODS RESULTS DISCUSSION LITERATURE CITED

Media:
Luria broth (LB), Luria agar (LA), and SOB were prepared as described in SAMBROOK et al. 1989 . Tryptone broth (TB), tryptone agar (TA), and tryptone broth soft agar (TBSA) were prepared as described by ARBER et al. 1983 except that MgSO₄ was added to a final concentration of 10 mM. When required, kanamycin was added to the media at a final concentration of 50 µg/ml, while ampicillin was added at a final concentration of 100 µg/ml for pIBI31-based vectors and 200 µg/ml for pUC19-based vectors.

Strains:
The standard -strain used was -P1:5R cI857 Kn^r nin5; it is designated simply as -P1 in the text. This strain carries a 10-kb segment of phage P1 DNA encoding functions for plasmid replication and partitioning (STERNBERG and AUSTIN 1983 ). The -P1:5R cI857 Kn^r nin5 prophage replicates as a single copy plasmid using the P1 replication machinery. The cI857 mutation renders the repressor heat labile, allowing prophage induction at 42°. The -P1:5R cI857 Kn^r nin5 also carries a 1.3-kb kanamycin-resistance cassette and has a genome size of 46.2 kb (PAL and CHATTORAJ 1988 ). The -P1 cosQ derivative of -P1:5R cI857 Kn^r nin5 contains a 14-bp deletion of cosQ from 48,470 to 48,483 (CUE and FEISS 1993 ). The standard bacterial hosts used were MF1427, a galK derivative of the E. coli C strain C1a (SIX and KLUG 1973 ), and MF2049, a ⁺ derivative of MF1427.

General recombinant DNA techniques:
Enzymes for recombinant DNA manipulations were purchased from New England Biolabs (Beverly, MA) and Boehringer Mannheim (Indianapolis) and used according to the manufacturer's recommendations. For cloning segments of DNA, the commercial vectors pBluescript II SK(-), pIBI31, and pUC19 were used. Plasmid DNA, PCR purifications, and DNA fragments purified from agarose gels were prepared with reagents purchased from QIAGEN (Valencia, CA). Preparation of competent cells and transformations were performed as described by HANAHAN 1983 . DNA sequencing reactions were performed using dye terminator cycle sequencing chemistry with AmpliTaq DNA polymerase, FS enzyme (PE Biosystems, Foster City, CA). The reactions were analyzed with an Applied Biosystems model 373 stretch fluorescent automated sequencer at the University of Iowa DNA facility.

Sequence designations:
All references to -sequence are based on the numbering convention described by DANIELS et al. 1983 . Numbering of the -sequence begins with the first base of the left cohesive end and continues along 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.

Mutagenesis of cosQ and cosQ–cosN spacing region:
For all cosQ point mutations from bp 48,478 to 48,483, cosQ mutants were constructed by the generation of oligonucleotides bearing the given mutations. Top and bottom strand oligonucleotides were generated bearing the given mutations as well as EcoO109I and BamHI sticky ends. Oligos were annealed and ligated into EcoO109I ( bp 48,473) and BamHI digested pBUC1. pBUC1 is a pBluescript SK(-) based vector with a DNA insert from bp 48,446 to 458 as well as a BamHI restriction site engineered at bp 48,486. Upon ligation with the oligos, the wild-type cosQ–cosN spacer sequence is regained. For all cosQ point mutations from bp 48,468 to 48,473, cosQ mutants were constructed by the generation of top and bottom strand oligonucleotides bearing the given mutations as well as EcoO109I and XbaI sticky ends. Oligos were annealed and ligated into EcoO109I and XbaI ( bp 48,446) digested pBUC8. pBUC8 is a version of pBUC1 with a wild-type DNA insert from bp 47,712 to 458. cosQ3 (G_48,475C) was constructed by oligo-directed mutagenesis. cosQ G_48,475A and cosQ G_48,475T were isolated as plaque-forming revertants of cosQ3. For all cosQ point mutations from bp 48,476 and 48,477 (with the exception of cosQ1), cosQ mutants were constructed by the generation of top and bottom strand oligonucleotides bearing the given mutations as well as XbaI and AgeI sticky ends. Oligos were annealed and ligated into XbaI and AgeI digested pBUC9. pBUC9 is a version of pBUC8 containing cosQ T_48,481A resulting in an AgeI site at bp 48,481. Upon ligation, the wild-type cosQ sequence is regained. cosQ1 (C_48,477T) was constructed as described by CUE and FEISS 1992 .

For cosQsub2, a mutagenic primer containing EcoO109I at bp 48,473 and an AflII substitution at bp 48,486 with the sequence 5' ACGGGTCCTTTCCGGCTTAAGGACAGGTTA 3' was constructed and used for PCR across cos using pCF114 as a template. pCF114 is a pIBI31-based vector containing DNA from bp 47,942 to 650 and contains cos (CUE and FEISS 1998 ). The PCR fragment was digested with EcoO109I and EcoRI (at bp 194) and ligated into EcoO109I and EcoRI digested pCF114 to generate pDW202. For cosQsub3, a mutagenic primer containing the above AflII substitution and BstEII substitution at bp 48,494 with the sequence 5' CGGCTTAAGGAGGTAACCCGGGGCGGCGAC 3' was constructed and used for PCR across cos. The fragment was digested with AflII and EcoRI and ligated into AflII and EcoRI digested pDW202 to generate pDW204.

cosQ–cosN spacing mutants were constructed by the generation of oligonucleotides bearing the given mutations. For additions of +1 to +4, top and bottom strand oligonucleotides were generated bearing the given mutations as well as EcoO109I and AflII sticky ends. Oligos were annealed and ligated into EcoO109I and AflII digested pDW202 to form pDW306 (+1) containing a SmaI site, pDW307 (+2) containing a NcoI site, pDW308 (+3) containing an AvrII site, and pDW309 (+4) containing a PstI site. For spacing mutants +5 and +11, top and bottom strand oligonucleotides were generated bearing the given mutations as well as EcoO109I and AflII sticky ends. Oligos were annealed and ligated into EcoO109I and AflII digested pDW202 to form pDW301 (+5) and pDW302 (+11), which contains a BstEII site. For +15, pDW302 was digested with AflII, the overhangs were filled in with Klenow enzyme, and blunt-end ligated. For +21, top and bottom strand oligonucleotides were generated bearing the given mutations and an SpeI site as well as BstEII sticky ends and annealed. pDW302 was digested with BstEII and the oligo linker was inserted to generate pDW305; +31 was generated through the double insertion of the +21 oligonucleotide in pDW302 to generate pDW304. Plasmids were confirmed by restriction enzyme digestion and DNA sequencing (see Table 3 and Table 4 for sequence information).

View this table: In this window In a new window   Table 2. Virus yields of cosQ single and double mutants

View this table: In this window In a new window   Table 4. Virus yields of cosQ–cosN substitution and spacing mutants

Introduction of cosQ mutations into the -genome:
Plasmids bearing DNA fragments containing the cosQ variants were introduced, by transformation, into MF1427 lysogenic for -P1 cosQ. Transformed lysogens were selected for by plating at 31° on LA containing kanamycin and ampicillin. The transformed lysogens were grown overnight in LB plus antibiotics at 31°. Overnight cultures were diluted 1:125 into LB and grown to 3 x 10⁷ 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 CHCl₃ and titered on MF1427. For recombinant phage unable to form plaques, kanamycin-resistance (Kn^r) transducing particles were titered by diluting lysates in 10 mM MgSO₄, and diluted lysates were used to infect MF1427 in TB plus 0.2% maltose. The infected cells were incubated at 31° for 60 min and then plated on LA containing kanamycin at 31°. PCR amplification followed by restriction enzyme analysis and DNA sequencing were performed to verify the crosses.

Phage yield determinations:
MF1427 lysogenized with -P1:5R cI857 nin5 Kn^r or a derivative were grown overnight with aeration in LB plus kanamycin at 31°. The cultures were diluted into LB (1:125 dilution) and grown to 3 x 10⁷ cells/ml at 31°. Portions of each culture were removed at this time, diluted 1:10,000 in 10 mM MgSO₄, and plated on LA plus kanamycin. Plates were incubated overnight at 31° to determine the number of viable lysogens in each culture. Prophages were induced as described above, and lysates were titered on MF1427. Kanamycin-resistance (Kn^r) transducing particles were titered by diluting lysates in 10 mM MgSO₄, and diluted lysates were used to infect MF2049 (MF1427 +) in TB plus 0.2% maltose. The infected cells were incubated at 31° for 60 min and then plated on LA containing kanamycin at 31°.

In vivo cosmid packaging assay:
MF1427 was lysogenized with -P1 cosQ and then transformed with pIBI31 containing the 5-, 11-, 15-, 21-, and 31-bp additions between cosQ and cosN. The transformed lysogens were grown and lysates were prepared as described for the phage yield determinations. Ampicillin-resistance (Ap^r) transducing particles were titered by diluting lysates in 10 mM MgSO₄, and diluted lysates were used to infect MF2049 (MF1427 +) in TB plus 0.2% maltose. The infected cells were incubated at 31° for 60 min and then plated on LA containing ampicillin at 31°.

	RESULTS

TOP ABSTRACT MATERIAL AND METHODS RESULTS DISCUSSION LITERATURE CITED

Saturation mutagenesis of cosQ:
The 17-bp sequence from bp 48,468 to 48,484 (Fig 1) is highly conserved in the related lambdoid phages 21, 80, and N15 (SMITH and FEISS 1993 ; RAVIN et al. 2000 ). In addition, -21 hybrid phages containing the terminase genes of 21 and the right chromosomal end of , including cosQ, are viable, suggesting the conserved 17-bp-long sequence contains cosQ (SMITH and FEISS 1993 ). Cosmid packaging experiments have shown that sequences to the left of bp 48,473 are not necessary for packaging of the -chromosome (HOHN 1983 ; MIWA and MATSUBARA 1983 ). Mutational analysis has also shown that a 14-bp deletion of bp 48,470–48,483 is lethal, resulting in a >10⁵-fold reduction in phage yield, while a point mutation at bp 48,477, termed cosQ1, resulted in inefficient packaging termination (CUE and FEISS 1993 ).

To define exactly the extent of cosQ, we constructed 48 mutant phages carrying each of the three possible point mutations for all positions in the sequence from bp 48,468 to 48,483. We used -P1:5R cI857 Kn^r nin5 as the parent phage; -P1 carries a kanamycin-resistance gene and has a plasmid prophage state. To determine the effects of the mutations on chromosome packaging, two virus yield assays were performed. For viable mutants, lysogens were induced and the numbers of plaque-forming units were determined. To analyze the virus yield of non-plaque-forming mutants, we titered progeny phage as kanamycin-resistant (Kn^r) transducing particles. Kn^r transduction eliminated the constraint that accompanies the plaque assay, which requires that the mutant phage produce enough phage per induced lysogen to form a plaque. Phages carrying mutations in bp 48,468–48,472 and bp 48,480–48,483 had little effect on the -virus yield (Table 1). In contrast, mutations in the 7-bp-long sequence extending from 48,473 to 48,479 exhibited a wide range of effects on virus yields of -P1 (Table 1; Fig 2). The following is a summary of each position of the 7-bp sequence:

• 48,473: G_48,473A resulted in a >10-fold decrease in virus yield. G_48,473T and G_48,473C were unable to form plaques and resulted in a >100-fold decrease in the ability to produce Kn^r transducing particles.

  
  

  View larger version (27K): In this window In a new window Download PPT slide   Figure 2. Relative virus yields of cosQ mutational analysis. The virus yields of the 48 cosQ mutations in the region of from bp 48,468 to 48,483 are shown relative to wild-type levels (1.0). The bold sequence at the bottom of the figure represents the 16 wild-type nucleotides of the initial sequence of cosQ (top strand) that was examined. The 3 nucleotides above represent the 3 mutations that were introduced for each position in the sequence (16 x 3 = 48). The open bars represent the virus yields of the 21 mutations in the newly defined 7-bp region of cosQ as indicated by the varying degrees of reduction in virus yield. Only 1 mutation in this region, T48,479A, has a virus yield near to that of wild type. The solid bars represent the virus yields of the mutations in the initial cosQ region that lie outside this 7-bp region and are relatively unaffected by mutations at those positions. Standard deviations are shown.

• 48,474: All mutations resulted in the inability to produce plaques and a >100-fold decrease in the production of Kn^r transducing particles.

• 48,475: G_48,475A and G_48,475T reduce the virus yield about 5-fold and 2.5-fold, respectively. G_48,475C was unable to form plaques, resulting in a >10-fold decrease in the production of Kn^r transducing particles.

• 48,476: T_48,476A and T_48,476C resulted in a >10-fold decrease in virus yield, while T_48,476G reduced virus yield 3-fold.

• 48,477: C_48,477A and C _48,477T reduced the virus yield about 3- and 8-fold, respectively, while C_48,477G resulted in a >10-fold decrease.

• 48,478: All mutations resulted in the inability to produce plaques and a >1000-fold decrease in the yield of Kn^r transducing particles.

• 48,479: T_48,479A and T_48,479G resulted in the inability to produce plaques and a >10-fold decrease in the yield of Kn^r transducing particles. Interestingly, of the 21 mutations in this sequence, T_48,479C was the only mutation that failed to exhibit a significant reduction in virus yield.

From the data presented, cosQ resides in the 7-bp sequence extending from 48,473 to 48,479. It is interesting to note that this sequence, GGGTCCT, corresponds to the recognition sequence for the restriction enzyme EcoO109I, (G/A)GGNCC(T/C).

Sequence matching of cosQ with cosN half-sites:
Sequence similarity has been observed between cosQ and the two half-sites of cosN, cosNL and cosNR (Fig 3). CUE and FEISS 1997 based this sequence match on the previously defined 17-bp region of cosQ. Here, we have amended this match to include only the newly defined 7-bp cosQ segment. cosQ has a 4-bp match when aligned with cosNL. The cosQ1 mutation, a C:T transition at position 48,477, reduces this match to 3/7. CUE and FEISS identified cosQ2, a G:A transition at position 48,473, as a local suppressor of cosQ1. cosQ2, coupled with cosQ1, further reduces the match to 2/7. When cosQ is compared to cosNR, a 3/7-bp match is observed. cosQ1 increases the match to 4/7 bp, while cosQ1, coupled with cosQ2, restores the sequence identity between cosQ and cosNR to 3/7 bp. If terminase is transferred from cosQ to cosNR as depicted by the depolarization model, there may be an optimum sequence match between cosQ and cosNR of 3/7 bp. Terminase may be bound too strongly with a match of 4/7 due to the cosQ1 mutation, which is restored by the suppressor cosQ2 (CUE and FEISS 1997 ).

View larger version (21K): In this window In a new window Download PPT slide   Figure 3. cosQ sequence relationships. The previously defined cosQ sequence is aligned with the R boxes of cosB as well as the right (cosNR) and left (cosNL) half-sites of cosN to reveal sequence homology. The box denotes the newly defined region of cosQ and the subsequent homologous nucleotides of cosN and cosB. Nucleotide positions of the sequences are given.

We examined this possibility of an optimum sequence match between cosQ and cosNR to test this aspect of our depolarization model of cosQ. Six mutant phages were analyzed to test this model for the sequence match requirement between cosQ and cosNR. The first three consisted of the cosQ1, cosQ2, and cosQ3 mutations alone. The cosQ1, cosQ2, and cosQ3 mutations resulted in phage yields 13, 9, and 3% of wild type, respectively (Table 1 and Table 2). Thus, the cosQ2 and cosQ3 mutations reduce the phage yield; cosQ2 by itself has a negative effect in the absence of the cosQ1 mutation. Since these mutations individually reduce the match to 2/7, the potential requirement for a 3/7 match between cosQ and cosN remained valid. An absence of a discernible phenotype would have allowed us to conclude that these base pairs are not necessary for proper cosQ function and that a 3/7 match is not necessary.

Three constructs consisting of the various combinations of the three mutations, cosQ1, cosQ2, and cosQ3, were further analyzed (Table 2). The cosQ1 cosQ2 double mutant increased the phage yield 6-fold over the yields of the single mutants, while the cosQ1 cosQ3 and cosQ2 cosQ3 double mutants had phage yields reduced >1000-fold. From these data, we can conclude the following: (1) cosQ2 is not general suppressor of cosQ mutations and (2) the cosQ1 cosQ3 double mutant has the proposed optimal sequence of 3/7. Since this construct has a severe defect, the proposal that a 3/7 match between cosQ and cosNR is required for cosQ function is not supported.

Spacing and sequence requirement between cosQ and cosN:
We examined the 17-bp region between cosQ and cosN to determine whether it was necessary for chromosome packaging. Sequence comparisons between phage and 80, phage 21, and N15, three related lambdoid phages, identified two stretches of conserved bases, a 4-bp stretch consisting of ATCC and a second region consisting of CXXGTTA (see Fig 4; SMITH and FEISS 1993 ). We altered the first of these stretches in by substitution with ClaI (data not shown) and AflII restriction sites. These substitution alleles, sub1 and sub2, respectively, displayed wild-type phenotypes. We then constructed the sub3 allele, consisting of a BstEII restriction site in the second of these conserved stretches of DNA in addition to the AflII restriction site of the sub2 allele. cosQsub3 had a mild decrease in virus yield (Table 3). We concluded that nucleotides between cosQ and cosN can be altered without producing serious effects on chromosome packaging.

View larger version (17K): In this window In a new window Download PPT slide   Figure 4. Changing base pairs in the cosQ–cosN spacer. Shown are the substitution alleles in the cosQ–cosN spacing region. The wild-type 17-nucleotide region (top strand) is given with homologous sequences between , phage 21, 80, and N15 shown in boldface. The boxed regions indicate the sequence substitutions that were constructed and the corresponding restriction enzyme sites that were introduced. The cosQsub2 allele was used for further manipulations involving the introduction of additional spacing mutations between cosQ and cosN (see text).

In the looping/hop version of the depolarization model, it was proposed that the region of DNA between cosQ and cosN bends to form a loop in order for cosQ and cosN to be aligned in the same orientation for transfer of the gpA subunit from cosQ to cosN (CUE and FEISS 1998 ). If the looping aspect of the model is correct, the addition of half-integral helical turns of +5 and +15 bp would not allow cosQ and cosN to align properly, resulting in severe defects in chromosome packaging. The addition of integral helical turns of +11, +21, and +31 bp would allow alignment and result in efficient chromosome packaging. We modified the sub2 allele by the addition of these integral and half-integral helical turns of DNA between cosQ and cosN to test the looping/hop version of the depolarization model. Plasmids containing an ampicillin-resistance gene in addition to the various spacing mutations were transformed into a cosQ lysogen. The prophages were induced and the efficiency of packaging the plasmids into ampicillin-resistant (Ap^r) transducing particles was determined. The yields of the spacing mutants were normalized to that of the phage containing the sub2 allele through which the mutants were generated. This served as a positive control. All plasmids containing the alterations of 5 bp or greater were unable to be packaged efficiently (Table 3).

We then constructed more subtle spacing alterations between cosQ and cosN consisting of the addition of 1 bp ( cosQadd1) to 4 bp ( cosQadd4). We performed the kanamycin transduction assay as a measure of packaging efficiency by counting the number of Kn^r transducing particles produced per infected cell (Table 4). The number of Kn^r transducing particles per infected cell for the sub2 allele was set at 100%. As the spacing was increased from +1 to +5, the values were 101, 62, 19, 1, and <0.001%, respectively. In addition, a plaque-forming unit assay was also performed for the plaque-forming spacing mutants, and the virus yields relative to the sub2 allele were in close agreement with the Kn^r transduction results (data not shown). These results indicate that as the distance between cosQ and cosN is increased, the severity of the packaging defect also increases. Proper spacing between cosQ and cosN is required.

	DISCUSSION

TOP ABSTRACT MATERIAL AND METHODS RESULTS DISCUSSION LITERATURE CITED

Genetic structure of cosQ:
To accurately define cosQ, we constructed 48 mutant phages carrying each of the three point mutations for all positions in the sequence from bp 48,468 to 48,483. It became evident after examining the effects the various mutations had on virus yields that a 7-bp region from bp 48,473 to 48,479 was required for cosQ function. All mutations flanking cosQ, i.e., from bp 48,468 to 48,472 and from bp 48,480 to 48,483, had little or no effect on phage production (Table 1; Fig 2). In contrast, mutations in the 7-bp region from bp 48,473 to 48,479 resulted in a wide range of effects on phage production. Of the 21 mutations in this region, only one mutation, T_48,478C, had a burst size near that of wild-type -P1. Only five other mutations resulted in virus yields of >10% of wild type: G_48,475A (30%), G_48,475T (56%), T_48,476G (35%), C_48,477A (32%), and C_48,477T (13%). All other mutations in this region resulted in a >10-fold decrease in phage yield.

The 7-bp cosQ segment is quite small. CUE and FEISS 1998 showed that cosQ does not act alone. Rather, cosQ acts with other elements of cos. Top strand cosN nicking is inaccurate without I2 and cosB, and cosQ is required for efficient N2 nicking. Bottom strand cosN nicking at N1 requires the presence of the proper cosNR sequence, at least in the absence of cosB. Under the conditions we have used, cosQ is clearly a 7-bp-long segment, but it is possible that under different conditions additional base pairs may be involved in cosQ function. A case in point is cosB: Certain point mutations in the R sequences, while having little effect on -development in cells with integration host factor (IHF), have strong phenotypic effects in cells lacking IHF (CUE and FEISS 1992 ).

The 7-bp cosQ segment is coincident with a recognition site for the EcoO109I restriction-modification system. EcoO109I is a type II restriction endonuclease isolated from E. coli with the heptanucleotide recognition sequence 5'-PuGGNCCPy-3' (MISE and NAKAJIMA 1985 ). Whereas the EcoO109I site has twofold rotational symmetry and is presumably recognized by symmetrically disposed subunits of the restriction enzyme and methylase, some of the cosQ mutations we have studied behave asymmetrically. The first position (48,473) requires a G residue for cosQ function, whereas the seventh position (48,479) can be a T or C. It may be that this asymmetry is indirectly imposed by the effects of base pairs flanking cosQ or by the other cos subsites that function with cosQ. An example of such indirect effects is that symmetrically disposed mutations in cosN behave asymmetrically due in large part to interactions of terminase with cosB (J. Q. HANG, C. E. C. CATALANO and M. FEISS, unpublished observations). In contrast, the mutations affecting the symmetrically disposed cosQ bp 48,474 and 48,478 have symmetric severe lethal phenotypes for all changes. It is interesting that cosQ and the EcoO109I site are coincident. Since cosQ is essential, -like phages cannot easily mutate to inactivate the EcoO109I target site, so E. coli strains that have the EcoO109I system have a defense against many -like phages that is not easily circumvented. We have recently found that methylation of cosQ by the EcoO109I methylase mildly interferes with growth of cos⁺ and severely interferes with the growth of phages bearing mild cosQ mutations that retain the EcoO109I recognition site (D. WIECZOREK and M. FEISS, unpublished observations).

We also examined the 17-bp region between cosQ and cosN to determine whether it was necessary for chromosome packaging. Sequence comparisons between phages , 21, 80, and N15 identified two stretches of conserved bases, a 4-bp stretch consisting of ATCC and a second region consisting of CXXGTTA (see Fig 4). Altering the first of these stretches in by substituting ClaI and AflII restriction sites resulted in a wild-type phenotype. The resulting wild-type levels of phage production further indicate that sequence outside of the newly defined 7-bp region is not essential for proper cosQ function. A third allele composed of a BstEII restriction site in the second of these conserved stretches of DNA in addition to the AflII restriction site resulted in a mild decrease in the burst size. This additional substitution alters the 7-bp immediately adjacent to cosN. This mild reduction in virus yield may be an effect of an alteration of cosN and have no bearing on the role of the cosQ site. XU and FEISS 1991A , XU and FEISS 1991B reported that a lethal transversion mutation affecting bp 48,502 (-1), termed cosNG-1T, causes defects in both DNA packaging and DNA injection. It is possible that the changes associated with the BstEII substitution from bp 48,494 to 48,500 (-9 to -3) may result in mild defects in DNA injection or cosN cleavage. We conclude that nucleotides between cosQ and cosN can be altered without producing serious effects on chromosome packaging.

Depolarization model:
Sequence similarities between cosQ and the R boxes of cosB as well as the half-sites of cosN have been observed, raising the possibility that cosQ is recognized by either gpNu1 or gpA, the small and large subunits of terminase. Molecular and genetic evidence against a role for gpNu1 in the recognition of cosQ has previously been presented. SHINDER and GOLD 1988 used DNase I footprinting to show that gpNu1 binds cosB but not cosQ. CUE and FEISS 1992 introduced analogous mutations in cosQ and each of the R boxes of cosB and showed that the mutations in R1, R2, and R3 made dependent on IHF for plaque formation, while the presence of IHF was not able to suppress the defect in the R4 sequence. CUE and FEISS 1997 later showed that trans-acting suppressors of the cosB R box mutations that alter gpNu1 were unable to suppress a point mutation in cosQ, termed cosQ1, while suppressors of cosQ are unable to suppress the defects of cosB mutants. The fact that cosQ and cosB are suppressed by two different types of suppressors suggests that they are recognized by different determinants.

A role for gpA in the recognition of cosQ remains a possibility. CUE and FEISS 1998 have shown that cosQ is required for the nicking of the bottom strand of DNA in the right half-site of cosN, cosNR, and the subsequent terminal cleavage at cosN. From these observations, a depolarization model has been proposed in which cosQ delivers gpA to cosNR for bottom strand nicking (Fig 5; CUE and FEISS 1998 ). In the looping/hop version of the depolarization model, it has been proposed that the region of DNA between cosQ and cosN bends to form a loop in order for cosQ and cosN to be aligned in the same orientation for transfer of the gpA subunit from cosQ to cosN. A second version of the depolarization model, the pause/recruitment version, proposes that cosQ is needed for a pause in the packaging process in order to recruit the second of two gpA subunits from solution to the cosN site for the nicking of the bottom strand of DNA. If terminase is transferred from cosQ to cosNR, an optimum sequence match between cosQ and cosNR of 3/7 bp may be necessary. Terminase may be bound too strongly with a match of 4/7 due to the cosQ1 mutation, which is restored by the suppressor cosQ2 (CUE and FEISS 1997 ). We examined the possibility of an optimum sequence match between cosQ and cosNR to test our depolarization model of cosQ.

View larger version (28K): In this window In a new window Download PPT slide   Figure 5. Depolarization model of cosQ. gpA protomers (represented by the ovals) in the translocation complex are aligned in a polarized fashion (indicated by the arrows) due to interactions between the portal protein and the carboxy terminus of gpA. The observation that cosN exhibits twofold rotational symmetry suggests that a gpA protomer of the opposite orientation may be required to cut the bottom strand (cosNR). In the looping/hop version, a gpA protomer bound to cosQ may be reoriented, or depolarized, and transferred to cosNR by the introduction of a bend in the region between cosQ and cosN allowing for the alignment of these two sites. In the pause/recruitment version, the binding of gpA to cosQ may result in a pause in the packaging process and allow for the recruitment of a gpA protomer from solution.

Three constructs consisting of combinations of the cosQ1, cosQ2, and cosQ3 mutations were analyzed to determine whether an optimum 3/7-bp match between cosQ and cosNR was necessary for efficient DNA packaging. While the cosQ1 cosQ2 double mutant, with a 3/7 match, was healthy, the cosQ2 cosQ3 double mutant (1/7 match) and cosQ1 cosQ3 double mutant (3/7 match) reduced the phage yield >1000-fold. The cosQ1 cosQ3 double mutant restored the sequence match between cosQ and cosNR to the proposed optimal match of 3/7. Since this construct results in inefficient chromosome packaging, we can conclude that a 3/7 match between cosQ and cosNR is not sufficient for cosQ function, and this evidence fails to lend support to this version of the depolarization model of cosQ.

An alternative explanation for the behavior of the cosQ1cosQ2 double mutant is that one of the mutations simply increases gpA's binding affinity to cosQ so that transfer of gpA to cosNR does not take place, while the second mutation decreases gpA's binding affinity. Thus, a combination of the two would cancel each other out. In this view, cosQ3 might either increase or decrease gpA's binding affinity to cosQ and therefore be able to suppress the defect of either the cosQ1 or cosQ2 mutation, resulting in a healthy phage. The fact that both the cosQ1cosQ3 and cosQ2cosQ3 double mutants are unhealthy argues against such a simple scenario. We note that the virus yields for cosQ1, cosQ2, and cosQ3 are similar (Table 2).

MILLER and FEISS 1988 showed that a defined spacing requirement exists between cosN and cosB. While a deletion of 1 bp between cosN and cosB resulted in a phage yield indistinguishable from that of wild type, further deletions of 2 and 3 bp resulted in phage yields of 79 and 17% of wild type, while 7-, 11-, and 26-bp deletions all were lethal (FEISS et al. 1983 ). An insertion of 2 bp between cosN and cosB resulted in a phage yield of 79% of wild type, similar to that of the 2-bp deletion. They suggest that the effects of the spacing changes might be due to an alteration in a DNA bending site or to a specific spacing requirement between cosB and cosN. They noted that the lethality of the 11 deletion suggests that the phenotypes of the spacing mutations are not due to changes in helix face positions of sites flanking the spacing mutations. HIGGINS and BECKER 1994 later demonstrated that terminase requires a defined distance of 47 ± 2 bp between cosN and cosB for proper nicking at cos to occur. We wished to test spacing requirements for cosQ and cosN. As was observed for cosN-cosB spacing, when the spacing between cosQ and cosN was increased, the severity of the packaging defect increased as well. In addition, all of our integral helical and half-helical insertions of 5 bp or greater were lethal, indicating that the phenotypes of these spacing mutations are not simply due to changes in helix face positions of cosQ in relation to cosN as proposed in the depolarization model. It remains a possibility that the various insertions disrupt a possible DNA bending site in the region between cosQ and cosN, but the high G-C content of the region and the short distance between cosQ and cosN (21 bp) makes the presence of a static bend unlikely. Furthermore, YEO et al. 1990 searched for sequence-induced bend sites in the cos region, but none were found in the sequence between cosQ and cosN. We cannot rule out the possibility of a packaging-induced transient bend that may occur in this region.

These observations fail to support aspects of our looping/hop version of the depolarization model as presented by CUE and FEISS 1998 . It does not appear that either looping or the sequence match between cosQ and cosN is necessary. Nevertheless, we cannot rule out the depolarization model for the transfer of a gpA subunit from cosQ to cosN for terminal cutting. A revised model not invoking the sequence match or DNA looping, but instead involving the wrapping of DNA around a protein is plausible. DNA looping would allow for flexibility in the distance between the two sites with the addition of helical turns, as long as proper facing between the two sites remains the same. This flexibility was not observed. With wrapping, on the other hand, a fixed distance between the two sites would be required since any change in distance would result in the misalignment of the two sites. In addition, we cannot rule out a role for gpA in the recognition of the cosQ site. It remains a possibility that cosQ simply causes a pause in packaging allowing for the recruitment of a gpA subunit from solution to the cosN site (Fig 5). However, CUE and FEISS 1998 found that the presence of cosQ, either singly or in multiple copies, was insufficient to arrest DNA translocation when a second cosQ site was placed 1 kb upstream of the normal cosQ site, indicating that cosQ by itself is not a packaging stop signal.

Recognition of cosQ:
Our ultimate goals in this study of cosQ are to identify the factor that binds or recognizes this site and to determine how cosQ recognition contributes to packaging termination. We are confident that the factor we are seeking is involved in the translocation complex; candidates include gpNu1, gpA, and the portal protein gpB. BLAST searches involving the EcoO109I restriction endonuclease, the EcoO109I methylase, and the known -proteins involved in the translocation complex, gpNu1, gpA, and gpB, have shown no significant homology regarding potential protein-cosQ DNA recognition motifs. We cannot rule out the possibility that a host protein may in fact be involved in the recognition of the cosQ site. Because most in vitro packaging studies to date (HWANG and FEISS 1995 ) have employed mature DNA, a requirement for a host protein involving cosQ recognition may have been bypassed.

It has been reported that the product of gp49 (endonuclease VII) of bacteriophage T4 is a Holliday-structure resolvase (X-solvase) responsible for clearing branched replicative DNA prior to packaging (FRANKEL et al. 1971 ; LUFTIG et al. 1971 ; MIZUUCHI et al. 1982 ; GOLZ and KEMPER 1999 ). GOLZ and KEMPER 1999 also showed that gp49 binds as dimer to the portal protein, gp20, and is an integral part of the packaging machine of T4. In phage , suppressors of cosQ mutations have been localized to the portal protein, gpB, indicating the packaging complex involves terminase, the portal vertex, cos-containing DNA, and potentially other factors necessary for the termination of DNA packaging (CUE and FEISS 1997 ; D. WIECZOREK and M. FEISS, unpublished observations). T4 recombines at a much higher rate than , and it is not clear that DNA needs to be debranched prior to packaging. Whether a host enzyme fulfills a similar or alternate role of gp49 for phage in E. coli remains to be seen. Pseudorevertant analysis is underway in hopes of identifying allele-specific trans-acting suppressors of our cosQ mutations.

	ACKNOWLEDGMENTS

We thank our co-workers, Carol Duffy, Qi Hang, Jenny Meyer, and Jean Sippy for advice and interest during the course of this work. This work was supported by National Institutes of Health (NIH) Research Grant GM-51611 and NIH Genetics Research Training Grant T32GM08629 (D.W.).

Manuscript received January 29, 2001; Accepted for publication March 16, 2001.

	LITERATURE CITED

TOP ABSTRACT MATERIAL AND METHODS RESULTS DISCUSSION 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, NY.

BEAR, S., D. COURT, and D. I. FRIEDMAN, 1983  An accessory role for integration host factor: characterization of a lambda mutant dependent upon integration host factor for DNA packaging. J. Virol. 52:966-972.

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

BECKER, A., M. MARKO, and M. GOLD, 1977  Early events in the in vitro packaging of bacteriophage DNA. Virology 78:291-305[Medline].

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

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

CUE, D. and M. FEISS, 1993  A site required for the termination of packaging of the phage chromosome. Proc. Natl. Acad. Sci. USA 90:9290-9294[Abstract/Free Full Text].

CUE, D. and M. FEISS, 1997  Genetic evidence that recognition of cosQ, the signal for termination of phage DNA packaging, depends on the extent of head filling. Genetics 147:7-17[Abstract].

CUE, D. and M. FEISS, 1998  Termination of packaging of the bacteriophage chromosome: cosQ is required for nicking the bottom strand of cosN.. J. Mol. Biol. 280:11-29[Medline].

DANIELS, D., J. SCHROEDER, W. SZYBALSKI, F. SANGER, A. COULSEN et al., 1983 Completed 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, NY.

DAVIDSON, A. and M. GOLD, 1987  A novel in vitro DNA packaging system demonstrating a direct role for the bacteriophage FI gene product. Virology 161:305-315[Medline].

DAVIDSON, A. and M. GOLD, 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 189:21-30[Medline].

EMMONS, S. W., 1974  Bacteriophage lambda derivatives carrying two copies of the cohesive end site. J. Mol. Biol. 83:511-525[Medline].

FEISS, M., 1986  Terminase and the recognition, cutting, and packaging of 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.

FEISS, M., W. WIDNER, G. MILLER, G. JOHNSON, and S. CHRISTIANSEN, 1983  Structure of the bacteriophage lambda cohesive end site: location of the sites of terminase binding (cosB) and nicking (cosN). Gene 24:207-218[Medline].

FEISS, M., J. SIPPY, and G. MILLER, 1985  Processive action of terminase during sequential packaging of bacteriophage chromosomes. J. Mol. Biol. 186:759-771[Medline].

FRACKMAN, S., D. A. SIEGELE, and M. FEISS, 1984  A functional domain of bacteriophage terminase for prohead binding. J. Mol. Biol. 180:283-300[Medline].

FRANKEL, F. R., M. L. BATCHELER, and C. K. CLARK, 1971  The role of gene 49 in DNA replication and head morphogenesis in bacteriophage T4. J. Mol. Biol. 62:439-463[Medline].

FURTH, M. E., and S. H. WICKNER, 1983 Lambda DNA replication, pp. 145–173 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.

GOLZ, S. and B. KEMPER, 1999  Association of Holliday-structure resolving endonuclease VII with gp20 from the packaging machine of T4. J. Mol. Biol. 285:1131-1144[Medline].

HANAHAN, D., 1983  Studies on transformation of Escherichia coli with plasmids. J. Mol. Biol. 166:557-580[Medline].

HIGGINS, R. R. and A. BECKER, 1994  The 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. 13:6162-6171[Medline].

HIGGINS, R. R. and A. BECKER, 1995  Interaction of terminase, the DNA packaging enzyme of phage , with its cos DNA substrate. J. Mol. Biol. 252:31-46[Medline].

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

HOHN, B., 1983  DNA sequences necessary for packaging of bacteriophage DNA. Proc. Natl. Acad. Sci. USA 80:7456-7460[Abstract/Free Full Text].

HWANG, Y. and M. FEISS, 1995  A defined system for in vitro DNA packaging. Virology 211:367-376[Medline].

HWANG, Y. and M. FEISS, 1996  Mutations affecting the high affinity ATPase center of gpA, the large subunit of bacteriophage terminase, inactivate the endonuclease activity of terminase. J. Mol. Biol. 261:524-535[Medline].

LUFTIG, R. B., W. B. WOOD, and R. OKINARA, 1971  On the nature of gene 49-defective heads and their role as intermediates. J. Mol. Biol. 57:555-573[Medline].

MILLER, G. and M. FEISS, 1988  The bacteriophage cohesive end site: isolation of spacing/substitution mutations that result in dependence on Escherichia coli integration host factor. Mol. Gen. Genet. 212:157-165[Medline].

MISE, K. and K. NAKAJIMA, 1985  Restriction endonuclease EcoO109I from Escherichia coli H709c with heptanucleotide recognition site 5'-PuGGNCCPy. Gene 36:363-367[Medline].

MIWA, T. and K. MATSUBARA, 1983  Lambda phage DNA sequences affecting the packaging process. Gene 24:199-206[Medline].

MIZUUCHI, K., B. KEMPER, J. HAYS, and R. A. WEISBERG, 1982  T4 endonuclease VII cleaves Holliday structures. Cell 29:357-365[Medline].

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.

RAVIN, V., N. RAVIN, S. CASJENS, M. E. FORD, and G. F. HATFULL et al., 2000  Genomic sequence and analysis of the atypical temperate bacteriophage N15. J. Mol. Biol. 299:53-73[Medline].

RUBINCHIK, S., W. PARRIS, and M. GOLD, 1994a  The in vitro endonuclease activity of gene product A, the large subunit of the bacteriophage terminase, and its relationship to the endonuclease activity of the holoenzyme. J. Biol. Chem. 269:13575-13585[Abstract/Free Full Text].

RUBINCHIK, S., W. PARRIS, and M. GOLD, 1994b  The in vitro ATPases of bacteriophage terminase and its large subunit, gene product A. J. Biol. Chem. 269:13586-13593[Abstract/Free Full Text].

SAMBROOK, J., E. F. FRITSCH and T. MANIATIS, 1989 Molecular Cloning; A Laboratory 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. PETERSEN, 1982  Nucleotide sequence of bacteriophage lambda DNA. J. Mol. Biol. 162:729-773[Medline].

SHINDER, G. and M. GOLD, 1988  The Nu1 subunit of bacteriophage lambda binds to specific sites in cos DNA. J. Virol. 62:387-392[Abstract/Free Full Text].

SIPPY, J. and M. FEISS, 1992  Analysis of a mutation affecting the specificity domain for prohead binding of the bacteriophage terminase. J. Bacteriol. 174:850-856[Abstract/Free Full Text].

SIPPY ARENS, J., Q. HANG, Y. HWANG, B. TUMA, and S. MAX et al., 1999  Mutations that extend the specificity of the endonuclease activity of terminase. J. Bacteriol. 181:218-224[Abstract/Free Full Text].

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

SMITH, M. P. and M. FEISS, 1993  Sites and gene products involved in lambdoid phage DNA packaging. J. Bacteriol. 175:2393-2399[Abstract/Free Full Text].

STERNBERG, N. and S. AUSTIN, 1983  Isolation and characterization of P1 minireplicons, l-P1:5R and l-P1:5L. J. Bacteriol. 153:800-812[Abstract/Free Full Text].

WU, W. F., S. CHRISTIANSEN, and M. FEISS, 1988  Domains for protein-protein interactions at the N- and C-termini of the large subunit of bacteriophage terminase. Genetics 119:477-484[Abstract/Free Full Text].

XU, S. and M. FEISS, 1991a  Structure of the bacteriophage cohesive end site: Genetic analysis of the site (cosN) at which nicks are introduced by terminase. J. Mol. Biol. 220:281-292[Medline].

XU, S. and M. FEISS, 1991b  The last duplex base-pair of the phage chromosome: Involvement in packaging, ejection, and routing of DNA. J. Mol. Biol. 220:293-306[Medline].

YEO, A. and M. FEISS, 1995a  Mutational analysis of the prohead binding domain of the large subunit of terminase, the bacteriophage DNA packaging enzyme. J. Mol. Biol. 245:126-140[Medline].

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

YEO, A., L. KOSTURKO, and M. FEISS, 1990  Structure of the bacteriophage cohesive end site: bent DNA on both sides of the site, cosN, at which terminase introduces nicks during chromosome maturation. Virology 174:329-334[Medline].

This article has been cited by other articles:

				J. E. Cronan Cosmid-Based System for Transient Expression and Absolute Off-to-On Transcriptional Control of Escherichia coli Genes J. Bacteriol., November 15, 2003; 185(22): 6522 - 6529. [Abstract] [Full Text] [PDF]

D. J. Wieczorek and M. Feiss Genetics of cosQ, the DNA-Packaging Termination Site of Phage {lambda}: Local Suppressors and Methylation Effects Genetics, September 1, 2003; 165(1): 11 - 21. [Abstract] [Full Text] [PDF]