Novel DNA packaging recognition in the unusual bacteriophage N15

Virology 482 (2015) 260–268

Contents lists available at ScienceDirect

Virology journal homepage: www.elsevier.com/locate/yviro

Novel DNA packaging recognition in the unusual bacteriophage N15 Michael Feiss a,n, Henriette Geyer a,b,1, Franco Klingberg a,c,2, Norma Moreno a,d,3, Amanda Forystek a,e,4, Nasib Karl Maluf a,f,5, Jean Sippy a a

Department of Microbiology, Roy J. and Lucille A. Carver College of Medicine, University of Iowa, Iowa City, IA 52242, USA Division of Viral Infections, Robert Koch Institute, Berlin, Germany c Flow Cytometry, Imaging & Microscopy, Thermo Fisher Scientific, Frankfurter Strasse 129B 64293 Darmstadt, Germany d Texas A&M University – Corpus Christi, 6300 Ocean Drive, Corpus Christi, TX 78412, United States. e Room # 2911 JPP, Dept. of Psychiatry, The University of Iowa, 200 Hawkins Drive, Iowa City, Iowa, 52242 f Alliance Protein Laboratories, Inc. 6042 Cornerstone Court West, Suite ASan Diego, CA 92121, USA. b

art ic l e i nf o

a b s t r a c t

Article history: Received 14 January 2015 Returned to author for revisions 16 February 2015 Accepted 9 March 2015 Available online 16 May 2015

Phage lambda's cosB packaging recognition site is tripartite, consisting of 3 TerS binding sites, called R sequences. TerS binding to the critical R3 site positions the TerL endonuclease for nicking cosN to λ generate cohesive ends. The N15 cos (cosN15) is closely related to cos , but whereas the cosBN15 subsite λ has R3, it lacks the R2 and R1 sites and the IHF binding site of cosB . A bioinformatic study of N15-like phages indicates that cosBN15 also has an accessory, remote rR2 site, which is proposed to increase packaging efficiency, like R2 and R1 of lambda. N15 plus five prophages all have the rR2 sequence, which is located in the TerS-encoding 1 gene, approximately 200 bp distal to R3. An additional set of four highly related prophages, exemplified by Monarch, has R3 sequence, but also has R2 and R1 sequences characteristic of cosB–λ. The DNA binding domain of TerS-N15 is a dimer. & 2015 Elsevier Inc. All rights reserved.

Keywords: Bacteriophage DNA packaging Virus DNA packaging Virus assembly Virus evolution Terminase Virus DNA recognition

Introduction6 Large DNA viruses, such as tailed bacteriophages and herpes viruses, use an ATP-powered motor to translocate viral DNA into

n

Corresponding author. E-mail addresses: [email protected] (H. Geyer), franco.klingberg@thermofisher.com (F. Klingberg), [email protected] (N. Moreno), [email protected] (A. Forystek), [email protected] (N.K. Maluf). 1 Present address: Division of Viral Infections, Robert Koch Institute, Institute, Berlin, Berlin, Germany 2 Present address: Flow Cytometry, Imaging & Microscopy, Thermo Fisher Scientific, Frankfurter Strasse 129B 64293 Darmstadt, Germany. 3 Present address: Texas A&M University-Corpus Christi, 6300 Ocean Drive, Corpus Christi, TX 78412, United States 4 Present address: Room # 2911 JPP, Dept. of Psychiatry,The University of Iowa, 200 Hawkins Drive, Iowa City, Iowa, 52242 5 Present address: Alliance Protein Laboratories, Inc. 6042 Cornerstone Court West, Suite ASan Diego, CA 92121, USA. 6 Abbreviations and nomenclature; IHF=integration host factor. wHTH=winged helix-turn-helix DNA binding motif. TerS and TerL denote the terminase small and large subunits, respectively. Hybrid phages, derivatives of λ carrying cos and terminase gene segments from N15 are designated by isolation numbers, for example, N15hy4. The two segments of recombinant genes of chimeric phages are denoted by both gene names and the hybrid identity: the small terminase subunit http://dx.doi.org/10.1016/j.virol.2015.03.027 0042-6822/& 2015 Elsevier Inc. All rights reserved.

the preformed empty shell, called the prohead or procapsid (Catalano, 2005; Newcomb et al., 1999; Rao and Feiss, 2008). Recent structural and bioinformatic studies demonstrate that the DNA packaging machinery of these viruses is descended from that of an ancient common ancestor. For example, the prohead shell is an icosahedral lattice principally constructed of many copies of a major capsid protein whose fold is conserved (Baker et al., 2005). Similarly, one of the prohead's 5-fold vertexes, the unique portal vertex, contains the radially disposed dodecameric portal protein. The portal protein contains a channel through which DNA enters and exits the shell interior (Agirrezabala et al., 2005; Lebedev et al., 2007; Simpson et al., 2000; Trus et al., 2004). Terminase, also conserved in the herpes viruses and tailed bacteriophages, is usually a hetero-oligomer of a small and a large subunit (Casjens, 2011; Feiss and Rao, 2012; Przech et al., 2003). The terminase small subunit (TerS) carries out viral DNA recognition. The large subunit contains the ATPase center that powers

(footnote continued) of λ N15hy4 is 1/Nu1hy4; the gene product is gp1/Nu1hy4. Sequence convention: for both λ and N15, bp numbers start with the first base of the left cohesive end.

M. Feiss et al. / Virology 482 (2015) 260–268

261

Fig. 1. Comparison of the coses and terminase small subunit genes of λ, 21 and N15. (A) cosλ (above) and cosN15 (below). (B) Alignment of left ends of λ (above) and N15 (below) DNAs. Vertical lines indicate sequence identity. Bp 1–15 (boxed) are cosN base pairs at the mature left chromosome ends. The large box (dotted) is I2, a sequence feature highly conserved in λ-like phages. The R3 and R2 segments of cosλ are also boxed. (C) Alignment of the DNA binding segments of the λ and N15 terminase small subunits. The winged helix-turn-helix DNA binding motif of λ's gpNu1 is formed by residues 5–12 (support helix), 13–15 (turn), 16–25 (recognition helix) and wing (residues 31–39) (de Beer et al., 2002).

translocation of the DNA into the prohead and the endonuclease that cuts the concatemeric DNA into unit-length virion chromosomes. For temperate bacteriophage λ, viral DNA packaging occurs in a cell containing roughly equal amounts of phage and host DNA. Virion λ chromosomes, produced during the packaging process, are linear dsDNA molecules with cohesive ends, i.e., complementary, 50 -ended, 12 base-long, single-strand extensions that anneal, circularizing the DNA, upon entry into a host cell. The DNA segment containing the annealed cohesive ends is called cos. At late times during the lytic cycle, recombination and rolling circle replication produce concatemeric DNA (Furth and Wickner, 1983). During packaging, concatemeric DNA is recognized and processed by terminase into monomeric virion chromosomes. The small and large subunits of λ terminase are gpNu1 and gpA, the gene products of the Nu1 and A genes, respectively. The simplest form of λ terminase is the protomer, a gpNu12:gpA1 heterotrimer. Protomers further assemble into tetramers of heterotrimers (Maluf et al., 2006, 2005). GpNu1's N-terminal domain, which specifically binds cos, is a tight dimer containing a winged helixturn-helix (wHTH) DNA binding motif (de Beer et al., 2002; Feiss et al., 2010). GpA contains two large domains: the N- and Cterminal halves contain the translocation ATPase and the cohesive end-generating endonuclease, respectively (Duffy and Feiss, 2002; Hang et al., 2001; Hwang and Feiss, 2000; Ortega et al., 2007). GpNu1 recognizes λ DNA by specific binding to cosB, a cos subsite adjacent to cosN, where gpA endonuclease domains introduce staggered nicks to generate the cohesive ends (Davidson and Gold, 1992; Hang et al., 2001; Higgins and Becker, 1995; Higgins et al., 1988; Hwang and Feiss, 2000). cosB binding by gpNu1 properly positions gpA endonuclease domains on cosN for introducing nicks. When cosB is deleted or re-positioned, cosN nicking is inefficient and inaccurate (Hang et al., 2001) Higgins (Higgins and Becker, 1994a, 1995; Miller and Feiss, 1988).

Packaging is initiated when terminase binds and nicks a cos of a concatemer. Following cosN nicking and cohesive end separation, terminase forms a tight complex, Complex I, on the cosB-containing chromosomal end (Becker et al., 1977; Yang et al., 1997). Complex I then docks on the portal protein, gpB, of the prohead (Dokland and Murialdo, 1993; Lander et al., 2008). Following portal docking of Complex I, gpA's ATPase is activated and ATP hydrolysis-powered translocation of the DNA into the shell ensues (Dhar and Feiss, 2005; Yang and Catalano, 2003). Terminase remains docked on the portal during translocation. When the next cos along the concatemer is encountered, terminase (1) nicks cosN, (2) dissociates from the portal, and (3) remains bound to the cosBcontaining end of the next chromosome along the concatemer (Feiss et al., 1985). By remaining bound to the next chromosome along the concatemer as a Complex I, terminase continues to package downstream chromosomes in a processive manner. The third cos sub-site, cosQ, is essential for recognition and nicking of the downstream cos during termination of packaging (Cue and Feiss, 1997; Wieczorek and Feiss, 2001). In sum, cosN and cosB are required to initiate λ DNA packaging, cosQ and cosN are required for termination, and all three subsites are required for processivity. λ The 120 bp-long cosB is complex, containing three gpNu1 binding sites, R3, R2, and R1 (Fig. 1) (Catalano et al., 1995). Between R3 and R2 is a binding site, I1, for IHF, the Escherichia coli bending protein (Bear et al., 1984; Kosturko et al., 1989; Ortega and Catalano, 2006; Xin and Feiss, 1993). IHF bends DNA into an approximately 180o hairpin (Rice et al., 1996). At cos, the IHFinduced I1 bend positions R3 and R2 such that the wHTH motifs of dimeric gpNu1 can be docked into the major grooves (de Beer et al., 2002). Complex I likely includes this nucleoprotein structure (Yang et al., 1997). Phage 21 is a λ-like phage whose head genes share strong sequence identity with λ's head genes – the two phages are descended from a common ancestor phage, as follows. The

262

M. Feiss et al. / Virology 482 (2015) 260–268

terminase subunits of λ and 21 share 4 50% sequence identity (Miller and Feiss, 1985; Smith and Feiss, 1993). Nevertheless, λ and 21 do not efficiently package the other phage's DNA (Feiss et al., 1981). Phages λ and 21 have very similar cos structures: cosQ and cosN are identical, and 21's cosB has three R sequences and the IHF binding site. Phage N15 is an unusual λ-like phage. N15's prophage is linear. The genomic sequence shows that N15's head genes share considerable sequence identity with those of λ and 21 (Ravin et al., λ 2000). N15's cosQ is identical to cosQ , and cosNN15 differs from

λ

cosN at positions 9 and 12 of the 12-base cohesive ends (Fig. 1) (Ravin et al., 1998). Similarly, based on sequence alignments, N15's small terminase subunit, gp1, is predicted to have a helix-turnhelix DNA binding motif analogous to that of gpNu1, though it is unclear whether a wing is present. Surprisingly, no repeated DNA sequence feature is found, indicating the absence of a complex λ cosB similar to cosB and cosB21 (Ravin et al., 1998). Furthermore, the start of the small terminase subunit gene, 1, at bp 89, is much closer to the left chromosome end than the corresponding start for the λ Nu1gene, at bp 191, or the phage 21 1 gene start at bp 189. Here we report studies defining cosBN15.

Results

λ-N15 hy4 and the cosmid packaging assay

Fig. 2. Cross of λ cos2 with pJM0N15 to generate recombinant phages with chimeric terminases. The lethal cos2 deletion prevents DNA packaging by λ cos2. Recovery of plaque forming recombinants required that λ cos2 pick up a functional cosN from pJMN015 by a double crossover as indicated by the dotted lines. pJM0N15 contains λ sequence from 44141 to an introduced XbaI site at 48442. To the right of the XbaI site is N15 sequence including cos and the terminase genes. To pick up cosN, the left crossover occurs in the long region of λ sequence homology to the left of the XbaI site. The right crossover takes place between segments of partial sequence identity between the λ and N15 segments of λ cos2 and pJM0N15, generating hybrid phages with a chimeric terminase gene. Recombinants were selected as KnR-transducing phages. The small black rectangle represents cosN15, and the small white box represents the cos2 mutation.

We assumed that for N15, like λ and 21, virus DNA recognition for packaging would involve interactions between cos and the putative HTH DNA binding motif in the N-terminus of the small terminase subunit. As a proxy for N15, we used a viable, recombinant phage in which the cos region and much of the TerS-encoding λ gene were from N15. In this phage, λ-N15hy4, cos and much of λ's TerS-encoding Nu1 gene were replaced by the analogous segment from N15, as described in Material and Methods. To characterize cosBN15, a cosmid packaging assay was used in which a lytically growing, N15-specific helper phage, λ-N15hy4, packages cosmid DNA, producing cosmid transducing particles that can be assayed using the cosmid's antibiotic resistance marker. During the helper phage's lytic growth, a small fraction of the cosmid molecules are converted to rolling circle replication, providing concatemeric cosmid DNA as the packaging substrate. A linear cosmid multimer is packaged from the concatemer, where the multimer length is consistent with the minimum DNA size limit for λ (Feiss et al., 1982). In preliminary experiments, we found that λ-N15hy4, when used as a helper phage, could package a

Fig. 3. Defining cosBN15: effects of scramble mutations on N15 cosmid packaging. (A) Sequence of the N15 chromosome showing locations of the scramble mutations. The gene 1 ATG start codon is underlined. See text for additional details. (B) Effects of scramble mutations on packaging of N15 cosmids by helper phage λ–P1 N15hy4. Numbers along the x axis are N15 base pair positions. The large blue rectangle represents the N15 1 gene. The y axis is the level of cosmid-containing ApR transducing phages/induced lysogen, normalized to 100% the yield the wild type cosmid (dotted line). Dashed lines represent the standard error of the mean (SEM) limits for yields of the wild type cosmids for the two sets of experiments. Vertical bars represent the yields of scramble mutations. Experiment 1 included the sc42–47, sc48–53, sc54–59, sc60–65, sc66–71 mutations, and the yield of ApR-transducing phages for the cos þ cosmid was 7.5 [SEM¼ 0.72]/induced cell. Experiment 2 included the sc71–91 and sc92–199 mutations, and the yield of ApR-transducing phages for the cos þ cosmid was 2.6  10  3 [SEM ¼ 9.5  10  4]. Results are from three independent experiments.

M. Feiss et al. / Virology 482 (2015) 260–268

cosmid carrying cosN15 roughly as efficiently as a λ aged an analogous cos cosmid.

λ helper pack-

Cosmid/helper packaging experiments define a critical cosBN15 segment The bp 42-70 segment To ask if cosN15 has a cosB, we determined the effects of 6 bp-long block mutations on helper packaging of a cosN15-containing cosmid. Each cosmid contained a 150 bp-long insert of N15 DNA including cosQN15 and cosNN15, ending at N15 bp 90, in the gene 1 start codon. Because of strong sequence identity (34/41; see Fig. 1) between λ and N15 in the cosN- and I2-containing segment from bp 1 to bp 41, the scramble mutations were designed to survey bp to the right of bp 41. The scramble mutations were introduced into a derivative of pUC19 containing cosN15 and ApR. The cosmids were used to transform a lysogen of the helper phage λ N15hy4. Prophage induction produced helper phage lysates that were assayed for cosmid-containing, ApRtransducing phage particles. The yields for the wild type cosmid and cosmids carrying the mutations sc42–47, sc60–65 and sc66–71 were within experimental error of the average for the wild type control cosmid, but the yields for the sc48–53 and sc54–59 mutants were strongly reduced and differed significantly from the wild type (p¼ 5  10  4 and 1  10  3 respectively, two-tailed Student's t test) (Fig. 3). Because the sc48–53 and sc54–59 scramble mutations severely reduce cosmid packaging, we conclude the wild type bp 48–59 segment is required or efficient cosmid packaging. The results indicate that cosN15 contains a small, essential site, R3N15, at bp 48– λ 59. Although the positions of R3N15 and R3 at approximately λ bp 50–65, have not been precisely defined, the two sites occupy similar positions with respect to distance from cosN. The sequences of the two R3 sequences share little identity

We note that alternative alignments giving greater sequence identity are possible. In the alignment shown, the sequences are equidistant from bp1 of the left cohesive end, i.e., λ bp 50 is aligned with N15 bp 50. This alignment was chosen because, in a study to be reported separately, we find that λ-N15hy4, in addition to packaging N15-specific DNA, also packages λ DNA at somewhat reduced efficiency. We reason that, since the cosN-to-R3 distance in λ is critical (Higgins and Becker, 1994b; Miller and Feiss, 1988), λ the structure formed by N15hy4 terminase at cos will be topologically similar to that formed by λ terminase, and that the N15hy4 terminase-cosN15 architecture will also be similar. The bp 71–199 segment For λ and 21, cosB extends to about bp 170. The experiment of the preceding section examined N15 DNA out to N15 bp 71. To make sure that there were no specific terminase recognition sites distal to N15 bp 71, an additional cosmid packaging experiment was carried out to ask if there was any specific DNA binding site between bp 71–199. The positive control plasmid, pNXV1, contained wild type N15 sequence out to N15 bp 199. For the experimental plasmid pNXV4, bp from 71–91 were scrambled, such that all bp were changed. For the experimental plasmid pNXV3, the gene 1 codons from bp 92–199 were scrambled so that just 18/109 bp matched between the wild type and scrambled DNA segments. A negative control plasmid, pNXV2, contained the same wild type N15 sequence as pNXV1, except that the critical segment, N15 bp 54– 59, was changed from 50 –GTTGTT-30 to 50 –CGGCCG-30 . As expected, changing the critical bp 54–59 segment reduced cosmid packaging about 30-fold, from 2.6  10  3 [SEM¼9.5  10  4] for the cos þ

263

cosmid pNXV1, to 8.2  10  5 [6.7  10  5] for pNXV2 (p¼7.6  10  5, two-tailed Student's t test). Scrambling the N15 DNA segment from 92–199 did not significantly reduce cosmid packaging by N15hy4 terminase (p¼0.44) (Fig. 3). The cosmid packaging yield of wild type pNXV1 was about as expected from previous studies. Similarly, packaging of pNXV4 was not significantly different from that of the wild type plasmid pNXV1. In sum, the only DNA segment showing sequence specificity for DNA packaging, between residues 41 and 199, is bp 48–59. An interesting question is whether the TerS of N15hy4 (and gp1 of N15) forms a structure analogous to the λ proposed structure of dimeric gpNu1 docked at R3 and R2 of cos , λ λ that is, the proposed architecture of Complex I . Complex I formation is facilitated by (1) a static bend at I1, and (2) enhancement of the I1 bend by IHF. Since cosN15 lacks a TerS binding site analogous in location to R2, a Complex IN15 analogous in architecλ ture to Complex I would require a non-sequence specific contact λ by the TerS, and a DNA bend analogous to the I1 bend in cos . λ N15hy4 forms healthy plaques on hosts lacking IHF, FIS, and HU (Geyer and Feiss, unpublished). The IHF consensus binding site is AT-rich. Inspection indicates that the cos sequence to the gene 1 side of R3N15 is not AT-rich. To ask about the presence of an IHF site near R3N15, the DNA site search application Possum (http://zlab.bu. edu/mfrith/cgi-bin/possum.cgi) (Fu et al., 2004) was used with the IHF binding site matrix of Goodrich, et al. (Goodrich et al., 1990) to examine the λ and N15 DNA sequences from bp 1–600. In λ, the overlapping strong I1A and weak I1B sites were identified, with scores of 20.4 and 5.37, respectively, but no site was identified in N15. Of course it is possible that N15 uses an alternative accessory protein, of host or phage origin, to induce a sharp bend in N15 DNA to facilitate formation of a Complex I analogous to that of λ. In the Appendix, we show that the DNA binding domain of N15's gp1, like that of λ's gpNu1, is a dimer. A report from the Ravin group identified an IHF binding site to the left of cosNN15 (Ravin et al., 1998). The Possum program failed to find a significant IHF site in the final 535 bp of N15's right chromosome end. It is possible that a different search program would identify such a site. We note that, in λ, no sequence important for packaging has been found upstream of cosQ. Since a λ–N15 hybrid has been used in our study, the presence of possible packaging signals to the left of the XbaI site near cosQN15 has not been explored in our study. Bioinformatics of N15-like DNA packaging recognition systems Comparing the left end sequences of N15 and λ, we found a segment with poor sequence identity from bp 42 to 289, which

Fig. 4. Cosmid packaging experiment: effects of R3 and rR2 mutations on cosmid packaging by N15hy4 terminase.

264

M. Feiss et al. / Virology 482 (2015) 260–268

includes (1) part of I2, (2) N15's critical R3-like segment from bp 48–59, and (3) the beginning of the gp1 gene. A BLAST (http:// www.ncbi.nlm.nih.gov/BLAST/) search with the N15 42–289 sequence, in August 2014, identified eight enteric bacterial prophages with identity scores 4 80, including three apparently intact and nearly identical E. coli prophages: Monarch, Neon, and Skimmer (The accession numbers and proposed TerS binding site sequences of the prophages mentioned in this paper are presented in Table S1). A fourth, closely related, partial prophage, Wapsipinicon, was identified in Escherichia albertii KF1. Wapsipinicon has an intact cos and the 50 half of the 1 gene. All four have the same recognition helix sequence as N15. We noticed that in the annotated genomic sequence the predicted Monarch 1 gene start codon was at bp 228, a position even farther from the left cohesive end than the bp 191 Nu1 start in λ. To ask about the possibility of repeated R sequences in Monarch's likely cosB-containing region, the sequence was submitted to the Palindromes application [http://www.emboss.bioinformatics.nl/cgibin/emboss/palindrome]. Repeated sequences matching or complementary to the putative gp1 binding segment [N15 bp 48-to-59] were identified; giving an arrangement analogous to that of the λ cosB R sequences (Fig. 4A). The 8 bp R3-R2 and 11 bp R2-R1 matches of Monarch's R sequences are shown as follows:

A Possum search for an IHF binding site identified a site [bp 60– 107], positioned analogously to λ I1A [bp 61–108] with an identity score of 7.8. This score is similar to those of the functional, low affinity binding sites attP H0 (10.9) and hag (7.0) Catalano and coworkers have shown that gpNu1 and IHF cooperatively assemble λ on cosB (Ortega and Catalano, 2006). We note that phages λ and 21 both have strong IHF binding sites, yet λ's yield is only reduced by about 2/3 in the absence of IHF, but 21's yield is severely reduced. λ-like phages may show various degrees of dependence on IHF. Neon, Skimmer, Wapsipinicon and Monarch, being nearly identical, have the same cosB structure. In contrast, the N15-related prophage Puccoon was identified from the enteric pathogen Edwardsiella piscicida strain. The proposed gene 1 start codon of Puccoon is located at bp 64, a location close to the left cohesive end, like that of N15's 1 gene. [Note that the 50 -termini of these N15-like 1 genes contain several methionine codons, making assignment of the initiation codon uncertain.] The Palindromes application did not find any analogs of the λ cos R2 or R1 sequences. Remarkably, though, a possible R sequence was picked up in bp 254–265, within the 1 gene. This potential site, dubbed rR2 for putative remote R2-like site, has 10 contiguous bases complementary to the R3Puccoon site at bp 48 to λ 59 and, like R2 , is in opposite orientation to the R3Puccoon as shown as follows:

i.e., at bp 249-to-260. In sum, both N15 and Puccoon have putative bipartite cosB structures, with a critical site analogous λ in position and extent to R3 of cosB . Neither has the I1, R1, or R2 λ features of cosB , but both have a possible remote rR2 sequence in the 1 gene. Another match to N15 bp 42–289 was Scrap, a partial prophage in Serratia fonticola RB-25, which contains a good match to R3N15, lacks I1, R2 and R1, and has a good match to rR2Puccoon and rR2N15. In another search, when the predicted amino acid sequence of N15's gp1 was used as the BLAST query, [(http://www.ncbi.nlm.nih.gov/BLAST/)], three matches were obtained: the partial prophages Geode of Serratia liquifaseans FK01 and Nishnabotna of Pseudomonas stutzeri, and VD1, a complete prophage of Vibrio diazotrophicus. In all six examples, the rR2 sequence was found (Table S1). See discussion for further commentary. We propose that cosBN15 and the cosBs of the five additional examples are bipartite, consisting of a critical R3 site and an accessory rR2 site. Testing the role of rR2 in DNA packaging A possible functional role for N15's remote rR2 site was tested using the cosmid packaging assay. The four cosmids used contained the cos segment extending from cosQ to N15 bp 300, and were either wild type or mutant at R3 and rR2. The same 50 – GTTGTT-30 to 50 –CGGCCG-30 substitution change, used above in pNXV2, was used in R3 and rR2 sequences (Fig. S1). Although the cosmid packaging results (Fig. 4) clearly showed that changing R3N15 reduced cosmid packaging, changing the rR2N15 sequence, either in the presence or absence of the R3 mutation, did not significantly change cosmid packaging. Though giving effects directionally consistent with the rR2N15 hypothesis, the results were not statistically significant. The results clearly exclude a critical role for rR2 in N15 DNA packaging. Due to the sensitivity limitations of virus yield studies, the results do not rule out an accessory role: that rR2N15 acts to increase virus fitness to a minor extent we cannot detect with the assay we used. It is useful to recall a similar situation involving the R sequences of λ cosB , as follows. In studying the effects of a single bp mutation λ affecting a conserved bp in the R sites of cosB , the mutational effects were marginal when the phage was grown in IHF þ cells. λ λ For R2 and R1 , statistically significant mutational effects were only found in host cells lacking IHF (Cue and Feiss, 1992). Since N15 does not utilize IHF for terminase assembly, we do not have an analogous experimental condition in which to study rR2N15 mutational effects. There is a one-to-one correlation between the N15-like cos structure and the presence of an rR2 sequence in the 1 gene, we propose that the rR2 sequences function in cos recognition by the corresponding TerS subunits. A possible alternative interpretation is that the rR2-encoded amino acid sequences are important for terminase function. We note however that there are frequently used alternative codons available to specify the KXQP sequence that would not result in complementarity between R3N15 and rR2N15. Discussion A critical DNA packaging recognition site in cosN15

The Possum application did not detect an IHF site in Puccoon. We wondered if N15 also has a possible rR2 site in gene 1, analogous to that of Puccoon. When the N15 sequence was examined with the Palindromes application, a possible rR2N15 site was found at approximately the same location as rR2Puccoon,

Surveying N15 bp 41 to 199 showed that there is a single critical segment extending from bp 48–59. Though not tested λ directly, by analogy with studies on cosB and cosB21, this site is in all likelihood the gp1 binding site. N15 clearly lacks the I1, R2 and λ R1 elements of cosB and cosB21. Extensive genetic and biochemical work shows that a major role for I1, R2 and R1 is to assist in the positioning of a gpNu1 subunit at R3, which positions gpA for efficient and accurate introduction of nicks at cosN (Cue and Feiss,

M. Feiss et al. / Virology 482 (2015) 260–268

265

Fig. 6. Unequal crossing over by a phage with a tripartite, λ-like cosB generates a cosB with a single recombinant R3/R1 sequence. See text for discussion. Fig. 5. Summary of proposed cosB structures for λ, Monarch, N15, and Puccoon. Zones highlighted in green are segments of significant sequence identity,  50% or more. Blue zone is the region of structural, but not sequence, similarity between λ and Monarch. TerS binding sites (R sequences) are indicated by blue boxes, and IHF binding sites (I1) by red boxes. Alignments of cos and TerS sequences for N15-like phages and prophages with bipartite cosBs are in Figure S3A and S3B, respectively. Alignments of cos and TerS sequences for Monarch-like prophages with tripartite cosBs are in Figure S4A and S4B, respectively.

prophage, is a strong sequence match to the reverse complement of the R3 sequence, suggesting that each has a remote “rR2” gp1 binding site (details of all the N15-like R sequences are presented in Table S1 and Fig. S4A). The possible role for rR2

1992; Hang et al., 2001; Higgins and Becker, 1994b; Ortega and Catalano, 2006). Binding of terminase to cos, including the interactions of dimeric gpNu1 to R3 and R2, assisted by IHF, forms a sharp hairpin that appears to make a very stable binary complex, λ λ Complex I (Becker et al., 1977; Yang et al., 1997) Complex I is observed following cos cleavage, and it is thought (1) protect the newly created, cosB-containing chromosome end from intracellular nuclease digestion, and (2) dock on the portal vertex of the prohead shell, leading to translocation. It is interesting that N15 apparently has a different strategy that either does not require the formation of the hairpin structure, or assembles such a structure in spite of the differences of cosBN15. Because the size and position of λ λ N15's critical site are similar to R3 of cosB , we designate N15's N15 hy4' site as R3 . A report on N15 s packaging specificity is in preparation. Proposed tripartite (λ-like) and bipartite cosB structures in the N15like bacteriophages Searches for phages with N15-like DNA packaging elements turned up a small set of prophages that have N15's gp1 recognition helix, and a sequence-match to R3N15. Remarkably, these prophages fall into two distinct cos structure classes (Fig. 5), one a group of four prophages, exemplified by Monarch, has the triparλ tite R sequence arrangement of cosB , including a predicted IHF binding site. The three R sequences are placed analogously to λ those of cosB , except that the R1Monarch site is roughly three λ λ helical turns farther from R2Monarch than R1 is from R2 (Figs. 5 and S4). In previous work, we showed that adding 11 bp, i.e., a λ λ helical turn, between R2 and R1 inactivated R1's ability to assist packaging in IHF  cells (Cue and Feiss, 1992). Whether increasing λ the distance further, i.e., by three helical turns, would allow R1 function is an open question. The other group, N15 plus five prophages, including Puccoon (Figs. 5and S3), are proposed to have a bipartite cosB, as follows. Each of these prophages has the N15 recognition helix and an N15like R3, positioned similarly to R3s of λ, 21, and Monarch. Surprisingly, at about bp 250–260 and within the 1 gene of each

Although the genetic experiment failed to show a statistically significant effect of an rR2 block mutation, there are three arguments suggesting there might be an accessory function for rR2. As discussed in results above, in the analogous situation in λ, statistically significant effects of R2 and R1 mutations, were only found in host cells lacking IHF. In the presence of IHF, only minor effects were found, and the interpretation is that R2 and R1 do affect phage fitness to a minor extent that is difficult to demonstrate experimentally. If a mutation were to reduce virus yield by 5%, for example, one would be hard pressed to demonstrate that by virus yield experiments, but the 5% yield boost provided by the accessory function would be highly significant in nature. A competition experiment would be a stronger experiment, but execution of such an experiment is complicated because rR2 is in a coding sequence. The second argument is that that all six examples of prophages with an N15-like cosB contain the rR2 sequence, and none of the four prophages with Monarch-like cosBs have the feature. The third argument is that the complementarity of the N15-like prophages' rR2s to the R3s varies from 8 to 10 bp. For comparison, the R2-to-R3 complementarity of the Monarchlike prophages is 8 in all for cases, though we note that these four examples are close relatives, unlike the more diverse N15-like examples (Table S1). If we assume that the R2 sequences of the Monarch-like prophages represent TerS binding sites, then the assumption that the rR2 sites are functional TerS binding sites is a reasonable speculation. To demonstrate rR2 is a functional TerS binding site requires a molecular experiment. Natural history of the N15-like packaging systems A very interesting question is whether the terminases of N15 and Monarch could act on each other's cos sites. Because R3 of cosBMonarch is located at the same place, relative to cosN, as R3N15, one can confidently predict that N15 terminase will package Monarch DNA. On the other hand, if Monarch terminase requires the complex tripartite cosB and perhaps IHF, then Monarch terminase may well fail to package N15 DNA. These interesting issues are amenable to genetic and biochemical studies. On the

266

M. Feiss et al. / Virology 482 (2015) 260–268

one hand, being able to package the DNA of a similar but different virus has the disadvantage of reducing the yield of the virus doing the packaging, and provides a small advantage to a competitor. On the other hand, packaging a relative's DNA may lead genetic exchange in a subsequent infection, increasing genetic diversity. There are now enough prophage sequences and experimental studies of packaging systems of the λ-like phages to show that there are many more viruses than packaging systems. The relatively small set of packaging systems known so far may reflect genetic advantages to shared packaging specificities, and/or the difficulty in evolving new ones How might the relatively simple cosBN15 have come to be? Unequal crossing over between the R3 and R1 sequences of two chromosomes, of a phage with a tripartite cosB like that of Monarch, generates a phage with a single recombinant R3/R1 sequence at the R3 position (Fig. 6). The crossover would be accompanied by loss of I1, R2 and the R1/R3. The position of the gp1 start codon would be shifted closer to cosN, just as in N15. For example should such a crossover occur in Monarch, the gene 1 start codon would be at bp 91, near to bp 89 of N15's gene 1 start. The evolution of rR2 might occur thru (1) genetic drift during lytic phage growth, with an accompanying increase in fitness, or (2) in the prophage state. Monarch's cosB is not a recent ancestor of the cosBN15or its simple-cos relatives, as the sequence between R3N15 and the gene 1 start does not match the R1-to-gene 1 sequence of Monarch. Perhaps when the sequence data base has more N15-like examples, it will be possible to construct a detailed evolutionary pathway for cosN15. In the case of the P22-related phages, Casjens and coworkers have been able to find phages with chimeric TerS genes, and to also identify suitable parental TerS genes in related prophages (Casjens, 2011). Finally, finding that the DNA binding domain of N15's gp1 is a dimer raises further interesting questions. Is the dimeric structure required for N15 DNA binding, even though there is no R2 sequence at the position analogous to λ R2 ? If N15's simpler cos is derived from a phage with a λ-like tripartite cosB, might the dimeric structure of N15's DNA binding domain be a relic from its parent phage? Or has N15 retained its dimeric DNA binding domain so that N15 terminase retains the ability to package chromosomes of its more complex relatives like Monarch?

Materials and methods Media, bacteria, phages, plasmids, phage manipulations Luria broth (LB), LB agar, tryptone broth (TB), tryptone broth agar (TA), and tryptone broth soft agar (TBSA) were prepared as described (Arber and Murray, 1983), except TB, TA, and TBSA were supplemented with 0.01 M MgSO4. For phage infections, TB was supplemented with 0.2% maltose. When required, ampicillin, chloramphenicol, and kanamycin were added at 100 μg/mL, 10 μg/ml, and 50 μg/mL, respectively. Bacteria, phages and plasmids are listed (Appendix). Cosmid DNA inserts were purchased from IDT, Coralville IA USA or Blue Heron Biotechnology, Inc, Bothell WA USA. Construction of helper phage λ–P1 N15hy4 is described in Results. Phage crosses and yield determinations, and antibiotic transduction assays for titering cosmid transducing phages were done using standard protocols (Arber and Murray, 1983).

λ–N15 hybrid phages In previous work comparing the packaging specificities of phages λ and 21, λ  21 hybrids carrying cosB21 and various extents of the 21 terminase genes were useful for defining specificity

(Feiss, 1986). To generate analogous λ–N15 hybrids, λ cos2, a mutant deleted for cosN, was crossed with pJM0N15 (Fig. 2). pJM0N15 contains a λ DNA segment extending from the late promotor through the lysis genes to an introduced XbaI site near cosQ. From the XbaI site is N15 DNA extending from cosN15 and gene 1 to an introduced BamHI site near the 30 end of the 2 gene. From the λ cos2  pJM0N15 cross, plaque-forming recombinants were produced by double crossovers that replace cos2 with cosN þ from pJM0N15. Several rare plaque-forming recombinants were sequenced. In each recombinant, the left crossover occurred in the long λ DNA homology segment. The right crossovers occurred in regions of partial sequence identity between the terminase genes 1 and 2 of N15 in pJM0N15, and Nu1 and A in λ cos2. λ-N15hy4 was selected for further study. λ-N15hy4 contained the shortest N15 DNA segment: the right crossover occurred between Nu1 and 1 in a short stretch of sequence identity from λ bp 614 to 624. The resulting chimeric 1/Nu1 gene encodes gp1/Nu1hy4, a chimeric TerS with the amino-terminal 122 amino acids of gp1, 11 aa in common, and the carboxy-terminal 35 residues of gpNu1. The gp1derived segment of the chimera completely replaces the entire λ cosB -interacting segment of gpNu1, which is the wHTH DNA binding motif (de Beer et al., 2002; Feiss et al., 2010). We assume that the terminase of λ-N15hy4 has the specificity of phage N15 terminase.

Cosmid packaging Cosmid packaging experiments were carried out by growing cosmid-containing lysogens of the helper phage, λ N15hy4, in LB to a density of about 5  107 cells/ml. The helper phage was induced to lytic growth by shifting cultures to 42 C for 15 min, followed by shaking at 37 C for 70 min, by which time the cells has lysed. Cultures were chloroformed and clarified in a clinical centrifuge, followed by titering the helper and cosmid transducing (ApR) phages. Three N15 cosmid sets were studied by helper packaging. We carried out these experiments over several years and details of experimental design varied, as follows. The first cosmid set, pCOS90 plasmids, contained a cosN15 segment extending from cosQ through cosN to N15 bp 90 in pUCAmp. These cosmids, obtained from Blue Heron Biotechnology, Inc. (Bothell WA USA), were identical except for mutated 6 bp segments to the right of cosN. The host and helper phage were the Rec þ strain MF1427 and the N15-specific helper phage, respectively. Analysis showed that cosmid-helper phage co-integrants were not a significant fraction of the ApR transducing phages. The second set of four cosmids, the pNXV1-pNXV4 cosmids, contained a cosN15 segment similar to that of the pCOS90 plasmids, except that the N15 DNA extended farther to the right, ending in gene 1 at N15 bp 199. In these pNXV cosmids, λ sequence extending from λ bp 48442 through cosQ and cosN to λ bp 12 is followed by N15 sequence extending from bp 13 to 199, followed λ λ in turn by a stop codon (Fig. S1). The use of the cosQ to cosN segment followed experiments showing that cosQ and cosN were interchangeable between λ and N15 (Feiss et al., in manuscript form). For the positive control plasmid, pNXV1, the N15 segment was wild type. For the experimental cosmid pNXV3, gene 1 codons 2 through 36 were scrambled. Of the 108 bp comprising codons 2 through 36, just 18 bp were identical between pNXV1 and pNXV3. For the negative control plasmid pNXV2, the N15 segment was wild type except that the segment N15 bp 54-GTTGTT-59 was changed to 54-CGGCCG-59. In the pNXV4 cosmid, bp 71 to 91 were scrambled such that no bp matched the wild type bp. The cos segments were purchased from Integrated DNA Technologies (IDT, Coralville IA, USA), and inserted using the XbaI and BamHI sites of

M. Feiss et al. / Virology 482 (2015) 260–268

vector pJB4. For the NXV1-4 cosmid experiments, the host was the recA1 strain MF3510, and the helper was the N15-specific helper. The third cosmid set, pNXV14 to pNXV17, was used to test the significance of rR2. These plasmids were derivatives of pJM4 in which the XbaI-to BamHI segment was replaced by analogous segments extending from the XbaI site at λ bp 48442 to an artificially introduced BamHI site at N15 bp 300. For R3 and rR2 mutations the 6 bp-long mutation sequence, used above in pNXV2, was used. The insert sequences are in Fig. S2. For this experiment, the host was the RecA  strain MF3510 and the helper phage was λ N15hy4. Note that the yield of the cos þ cosmid in the first cosmid set was 7.5 Ap-transducing phages/induced cell, and the yield for the analogous cos þ control in the other sets was 2.6  10  3 Aptransducing phages/induced cell. In the first set, the host was Rec þ , which permitted efficient production of cosmid concatemers. Because the cosmids in the first set carried small inserts of phage DNA, recombination producing cosmid þ helper phage cointegrants was negligible. In the second and third sets, the inserts were large and so a recA1 host was used. As a consequence, the production of cosmid concatemers was severely reduced.

Acknowledgments We thank Sherwood Casjens for N15 and thank Sherwood and Caitlin Murphy for advice on bioinformatic analysis. Norma Moreno was supported by National Science Foundation, United StatesResearch Experience for Undergraduates award DBI-0097361 (www.nsf.gov/). Henriette Geyer and Franco Klingberg were practical training students from the University of Applied Science, Lausitz, Germany. This extended project was supported by funding from National Science Foundation awards MCB-0717620 and MCB1158495, National Institutes of Health, United States awards 5R01GM088186 and 5R01GM051611 (www.nih.gov/), and the University of Iowa Department of Microbiology.

Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.virol.2015.03.027.

References Agirrezabala, X., Martin-Benito, J., Valle, M., Gonzalez, J.M., Valencia, A., Valpuesta, J.M., Carrascosa, J.L., 2005. Structure of the connector of bacteriophage T7 at 8A resolution: structural homologies of a basic component of a DNA translocating machinery. J. Mol. Biol. 347, 895. Arber, W.E.L.H.B.M.N.E., Murray, K., 1983. In: Hendrix, R.W.R.J.W.S.F.W., Weisberg, R. A. (Eds.), Experimental Methods For Use With Lambda. Lambda II. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, p. 433. Baker, M.L., Jiang, W., Rixon, F.J., Chiu, W., 2005. Common ancestry of herpesviruses and tailed DNA bacteriophages. J. Virol. 79, 14967. Bear, S., Court, D., Friedman, D., 1984. An accessory role for Escherichia coli integration host factor: characterization of a lambda mutant dependent upon integration host factor for DNA packaging. J. Virol. 52, 966. Becker, A., Murialdo, H., Gold, M., 1977. Early events in the in vitro packaging of bacteriophage DNA. Virology 78, 291. Casjens, S.R., 2011. The DNA-packaging nanomotor of tailed bacteriophages. Nat. Rev. Microbiol. 9, 647. Catalano, C.E., 2005. Viral Genome Packaging Machines: Genetics, Structure and Mechanism. Landes Bioscience, Georgetown TX. Catalano, C.E., Cue, D., Feiss, M., 1995. Virus DNA packaging: the strategy used by phage lambda. Mol. Microbiol. 16, 1075. Cue, D., Feiss, M., 1992. Genetic analysis of cosB, the binding site for terminase, the DNA packaging enzyme of bacteriophage lambda. Mol. Biol. 228, 58. Cue, D., Feiss, M., 1997. Genetic evidence that recognition of cosQ, the signal for termination of phage lambda DNA packaging, depends on the extent of head filling. Genetics 147, 7.

267

Davidson, A., Gold, M., 1992. Mutations abolishing the endonuclease activity of bacteriophage lambda terminase lie in two distinct regions of the A gene, one of which may encode a leucine zipper DNA binding domain. Virology 161, 305. de Beer, T., Fang, J., Ortega, M., Yang, Q., Maes, L., Duffy, C., Berton, N., Sippy, J., Overduin, M., Feiss, M., Catalano, C.E., 2002. Insights into specific DNA recognition during the assembly of a viral genome packaging machine. Mol. Cell 9, 981. Dhar, A., Feiss, M., 2005. Bacteriophage lambda terminase: alterations of the highaffinity ATPase affect viral DNA packaging. J. Mol. Biol. 347, 71. Dokland, T., Murialdo, H., 1993. Structural transitions during maturation of bacteriophage lambda capsids. J. Mol. Biol. 233, 682. Duffy, C., Feiss, M., 2002. The large subunit of bacteriophage lambda's terminase plays a role in DNA translocation and packaging termination. J. Mol. Biol. 316, 547. Feiss, M., 1986. Terminase and the recognition, cutting and packaging of lambda chromosomes. Trends Genet. 2, 100. Feiss, M., Fisher, R., Siegele, D.A., Widner, W., 1981. Bacteriophage lambda and 21 packaging specificities. In: DuBow, M.S. (Ed.), Bacteriophage Assembly. Alan R. Liss Inc., New York, p. 213. Feiss, M., Rao, V.B., 2012. The bacteriophage DNA packaging machine. Adv. Exp. Med. Biol. 726, 489–509. Feiss, M., Reynolds, E., Schrock, M., Sippy, J., 2010. DNA packaging by lambda-like bacteriophages: mutations broadening the packaging specificity of terminase, the lambda-packaging enzyme. Genetics 184, 43. Feiss, M., Siegele, D.A., Rudolph, C.F., Frackman, S., 1982. Cosmid DNA packaging in vivo. Gene 17, 123. Feiss, M., Sippy, J., Miller, G., 1985. Processive action of terminase during sequential packaging of bacteriophage lambda chromosomes. J. Mol. Biol. 186, 759. Feiss et al., in manuscript form. Fu, Y., Frith, M.C., Haverty, P.M., Weng, Z., 2004. MotifViz: an analysis and visualization tool for motif discovery. Nucleic Acids Res. 32, W420. Furth, M.E., Wickner, S.H., 1983. Lambda DNA replication. In: Hendrix, R.W., Roberts, J.W., Stahl, F.W., Weisberg, R.A. (Eds.), Lambda II. Cold Spring Harbor Press, Cold Spring Harbor NY, p. 145. Geyer , Feiss, unpublished. Goodrich, J.A., Schwartz, M.L., McClure, W.R., 1990. Searching for and predicting the activity of sites for DNA binding proteins: compilation and analysis of the binding sites for Escherichia coli integration host factor (IHF). Nucleic Acids Res. 18, 4993. Hang, J.Q., Catalano, C.E., Feiss, M., 2001. The functional asymmetry of cosN, the nicking site for bacteriophage lambda DNA packaging, is dependent on the terminase binding site, cosB. Biochemistry 40, 13370. Higgins, R.R., Becker, A., 1994a. Chromosome end formation in phage lambda, catalyzed by terminase, is controlled by two DNA elements of cos, cosN and R3, and by ATP. EMBO J. 13, 6152. Higgins, R.R., Becker, A., 1994b. The lambda terminase enzyme measures the point of its endonucleolytic attack 47 þ /- 2bp away from its site of specific DNA binding, the R site. EMBO J. 13, 6162. Higgins, R.R., Becker, A., 1995. Interaction of terminase, the DNA packaging enzyme of phage lambda, with its cos DNA substrate. J. Mol. Biol. 252, 31. Higgins, R.R., Lucko, H.J., Becker, A., 1988. Mechanism of cos DNA cleavage by bacteriophage lambda terminase: multiple roles of ATP. Cell 54, 765. Hwang, Y., Feiss, M., 2000. The Endonuclease and helicase activities of bacteriophage lambda terminase: changing nearby residue 515 restores activity to the gpA K497D mutant enzyme. Virology 277, 204. Kosturko, L., Daub, E., Murialdo, H., 1989. The interaction of E. coli integration host factor and lambda cos DNA multicomplex formation and protein-induced bending. Nucleic Acids Res. 17, 329. Lander, G.C., Evilevitch, A., Jeembaeva, M., Potter, C.S., Carragher, B., Johnson, J.E., 2008. Bacteriophage lambda stabilization by auxiliary protein gpD: timing, location, and mechanism of attachment determined by cryo-EM. Structure 16, 1399. Lebedev, A.A., Krause, M.H., Isidro, A.L., Vagin, A.A., Orlova, E.V., Turner, J., Dodson, E.J., Tavares, P., Antson, A.A., 2007. Structural framework for DNA translocation via the viral portal protein. EMBO J. 26, 1984. Maluf, N., Gaussier, H., Bogner, E., Feiss, M., Catalano, C., 2006. Assembly of bacteriophage lambda terminase into a viral DNA maturation and packaging machine. Biochemistry 45, 15259. Maluf, N., Yang, Q., Catalano, C., 2005. Self-association properties of the bacteriophage lambda terminase holoenzyme: implications for the DNA packaging motor. J. Mol. Biol. 347, 523. Miller, G., Feiss, M., 1985. Sequence of the left end of phage 21 DNA. J. Mol. Biol. 183, 246. Miller, G., Feiss, M., 1988. The bacteriophage lambda cohesive end site: isolation of spacing/substitution mutations that result in dependence on Escherichia coli integration host factor. Mol. Gen. Genet. 212, 157. Newcomb, W.W., Homa, F.L., Thomsen, D.R., Trus, B.L., Cheng, N., Steven, A., Booy, F., Brown, J.C., 1999. Assembly of the herpes simplex virus procapsid from purified components and identification of small complexes containing the major capsid and scaffolding proteins. J. Virol. 73, 4239. Ortega, M.E., Catalano, C.E., 2006. Bacteriophage lambda gpNu1 and Escherichia coli IHF proteins cooperatively bind and bend viral DNA: implications for the assembly of a genome-packaging motor. Biochemistry 45, 5180. Ortega, M.E., Gaussier, H., Catalano, C.E., 2007. The DNA maturation domain of gpA, the DNA packaging motor protein of bacteriophage lambda, contains an ATPase site associated with endonuclease activity. J. Mol. Biol. 373, 851.

268

M. Feiss et al. / Virology 482 (2015) 260–268

Przech, A.J., Yu, D., Weller, S.K., 2003. Point mutations in exon I of the herpes simplex virus putative terminase subunit, UL15, indicate that the most conserved residues are essential for cleavage and packaging. J. Virol. 77, 9613. Rao, V.B., Feiss, M., 2008. The bacteriophage DNA packaging motor Annual Reviews Palo Alto CA. Ann. Rev. Genet. 42, 647. Ravin, N.V., Doroshenko, O.I., Ravin, V.K., 1998. [COS-region of temperate coliphage N15]. Mol. Genet. Mikrobiol. Virusol. 2, 17–20. Ravin, V., Ravin, N., Casjens, S., Ford, M.E., Hatfull, G.F., Hendrix, R.W., 2000. Genomic sequence and analysis of the atypical temperate bacteriophage N15. J. Mol. Biol. 299, 53. Rice, P.A., Yang, S., Mizuuchi, K., Nash, H.A., 1996. Crystal structure of an IHF-DNA complex: a protein-induced DNA U-turn. Cell 87, 1295. Simpson, A., Tao, Y., Leiman, P., Badasso, M., He, Y., Jardine, P., Olson, N., Morais, M., Grimes, S., Anderson, D., Baker, T., Rossmann, M., 2000. Structure of the bacteriophage phi29 DNA packaging motor. Nature 408, 745.

Smith, M.P., Feiss, M., 1993. Sequence analysis of the phage 21 genes for prohead assembly and head completion. Gene 126, 1. Trus, B.L., Cheng, N., Newcomb, W.W., Homa, F.L., Brown, J.C., Steven, A.C., 2004. Structure and polymorphism of the UL6 portal protein of herpes simplex virus type 1. J. Virol. 78, 12668. Wieczorek, D.J., Feiss, M., 2001. Defining cosQ, the site required for termination of bacteriophage lambda DNA packaging. Genetics 158, 495. Xin, W., Feiss, M., 1993. Function of IHF in lambda DNA packaging. I. Identification of the strong binding site for integration host factor and the locus for intrinsic bending in cosB. J. Mol. Biol. 230, 492. Yang, Q., Catalano, C.E., 2003. Biochemical characterization of bacteriophage lambda genome packaging in vitro. Virology 305, 276. Yang, Q., Hanagan, A., Catalano, C.E., 1997. Assembly of a nucleoprotein complex required for DNA packaging by bacteriophage lambda. Biochemistry 36, 2744.