Sequential Headful Packaging and Fate of the Cleaved DNA Ends in Bacteriophage SPP1

J. Mol. Biol. (1996) 264, 954–967

Sequential Headful Packaging and Fate of the Cleaved DNA Ends in Bacteriophage SPP1 Paulo Tavares*, Rudi Lurz, Asita Stiege, Beate Ru¨ckert and Thomas A. Trautner Max-Planck-Institut fu¨r Molekulare Genetik Ihnestrabe 73, D-14195 Berlin, Germany

The virulent Bacillus subtilis bacteriophage SPP1 packages its DNA from a precursor concatemer by a headful mechanism. Following disruption of mature virions with chelating agents the chromosome end produced by the headful cut remains stably bound to the phage tail. Cleavage of this tail-chromosome complex with restriction endonucleases that recognize single asymmetric positions within the SPP1 genome yields several distinct classes of DNA molecules whose size reflects the packaging cycle they were generated from. A continuous decrease in the number of molecules within each class derived from successive encapsidation rounds indicates that there are several packaging series which end after each headful packaging cycle. The frequency of molecules in each packaging class follows the distribution expected for a sequential mechanism initiated unidirectionally at a defined position in the genome (pac). The heterogeneity of the DNA fragment sizes within each class reveals an imprecision in headful cleavage of 02.5 kb (5.6% of the genome size). The number of encapsidation events in a packaging series (processivity) was observed to increase with time during the infection process. DNA ejection through the tail can be induced in vitro by a variety of mild denaturing conditions. The first DNA extremity to exit the virion is invariably the same that was observed to be bound to the tail, implying that the viral chromosome is ejected with a specific polarity to penetrate the host. In mature virions a short segment of this chromosome end (55 to 67 bp equivalent to 187 to 288 Å) is fixed to the tail area proximal to the head (connector). Upon ejection this extremity is the first to move along the tail tube to exit from the virion through the region where the tail spike was attached. 7 1996 Academic Press Limited

*Corresponding author

Keywords: bacteriophage SPP1; DNA packaging; virus structure; connector; DNA ejection

Introduction The majority of tailed icosahedral bacteriophages package their chromosome from a concatemer precursor into a pre-assembled pro-capsid structure. Encapsidation is normally initiated at a specific nucleotide sequence (cos or pac) and proceeds unidirectionally. After DNA is packaged inside the viral pro-capsid either a site specific (e.g. in l, f105, T3, T7) or sequence independent (headful packaging; e.g. in T4, P22, T1, P1, SPP1) Abbreviations used: EM, electron microscopy; dsDNA, double-stranded DNA; EDAC, 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide; pfu, plaque-forming units. 0022–2836/96/500954–14 $25.00/0

terminal endonucleolytic cleavage occurs defining the mature chromosome size (reviewed by Black, 1989). Subsequent packaging events occur in a processive fashion starting from the DNA end generated by such a cut. After DNA encapsidation the filled phage head is stabilized by binding of additional proteins. One or more of these bind to the portal protein, a turbine-like oligomer with a central pore through which DNA is believed to move into and out of the capsid (Bazinet & King, 1985; Dube et al., 1993; Valpuesta & Carrascosa, 1994). This interaction leads to closure of the portal pore and creates the correct structural arrangement for tail attachment occurring at the unique vertex of the head where the portal protein is located (Casjens & Hendrix, 7 1996 Academic Press Limited

Headful Packaging and Fate of the Cleaved DNA Ends

1988). The multiprotein complex assembled at the portal vertex is named connector (Hendrix, 1978). Binding of the tail finishes morphogenesis yielding the infective virion, a metastable structure highly resistant to environmental challenges but also designed to deliver the viral chromosome in a regulated and efficient way upon interaction with the host. The virulent Bacillus subtilis bacteriophage SPP1 packages its DNA by a headful mechanism. The phenomenology of DNA encapsidation (Morelli et al., 1979; Humphreys & Trautner, 1981), characterization of the target sequence (pac) for initiation of the process (Deichelbohrer et al., 1982; Bravo et al., 1990), recognition and cleavage at pac by the terminase which is a multimer of gp1 and gp2 (Chai et al., 1992, 1994, 1995), and participation of the SPP1 portal protein (gp6) in determination of chromosome size (Tavares et al., 1992, 1995) were previously investigated. Here we describe studies on the processivity of the general packaging reaction, investigate the topology of the DNA molecule ends in the phage particle, and discuss its potential implications for chromosome ejection.

Results The DNA pac distal end is attached to the phage tail During early studies on bacteriophage SPP1 it was observed that exposure of viral particles to chelating agents leads to disassembly of the virion into partially disrupted heads and tails (Esche et al., 1975). When SPP1 particles are disrupted, one end of the viral chromosome is found tightly bound to the tail region where the connector is located (Figure 1). In order to determine whether a particular extremity of the DNA molecule is attached and, if this is the case, which one of the

955

Figure 1. SPP1 particles disrupted after treatment with EDTA. Two empty head shells (arrowheads) and one tail-chromosome complex can be observed. The DNA is bound to the head proximal end of the tail (connector, arrow) while the tail spike is present at the other extremity. The bar represents 0.5 mm. The material for EM was prepared by adsorption to mica.

two ends, we employed the strategy outlined in Figure 2A. The phage chromosome-tail complex was digested with BglII which cuts at a single position located asymmetrically within the SPP1 genome using our previous convention for its presentation (Humphreys & Trautner, 1981), approximately 13 kb from the packaging initiation sequence (pac; Bravo et al., 1990; Chai et al., 1992). This procedure allowed us to distinguish the pac proximal and distal DNA ends. To facilitate the analysis we eliminated the effect of circular permutation in the chromosome population by

Figure 2. Identification of the DNA end bound to the tail. A, Experimental strategy. The heavy bar represents a precursor concatemer from which encapsidation occurs. The origin of packaging (pac), the BglII site (B), and the direction of packaging are indicated. The thin lines below represent series of molecules generated after sequential packaging initiated at the right most pac site, and cleavage with BglII of SPP1wt and SPP1sizS DNAs. Upward thick lines symbolize tails attached to the DNA molecules extremity generated by the headful cut. B, EM measurement of individual DNA molecules associated with the phage tail after disruption of SPP1sizS with EDTA and subsequent BglII cleavage. Molecules are aligned by the end bound to the tail and ordered by increasing size. The upper thick bar represents the chromosome size (44.1 kb; Tavares et al., 1992). Complexes were prepared for EM measurements by the droplet technique.

956 using SPP1sizS, a mutant characterized by a smaller chromosome size which is virtually identical to that of the genome (44.6 kb; Tavares et al., 1992), implying that the DNA molecules packaged have a very short terminal redundancy. Thus, in contrast to the wild-type situation, a relatively constant topology was expected (Figure 2A). Electron microscopy (EM) measurements of the chromosome-tail complexes cleaved with BglII showed that a relatively homogeneous population of long fragments (31 to 35 kb) remained attached to the tail (Figure 2B) demonstrating that the pac distal end is bound to the tail. Processivity of headful packaging The tail attached to a specific end of SPP1 DNA provides a fixed reference to align the population of phage chromosomes. This feature enabled us to characterize several properties of the packaging reaction which preceded the interactions between SPP1 DNA and the tail. Since mature DNA molecules are generated by sequential headful encapsidation initiated at pac (Tye et al., 1974; Gill & MacHattie, 1976; Morelli et al., 1979), we expected that several classes of DNA molecules shall be present when the tail-chromosome complex is digested with a restriction enzyme cutting at a single asymmetric position within the SPP1 genome. Each class would include molecules with similar size and this size would depend on the packaging cycle from which they were generated. If the endonucleolytic cleavage occurs very distal to the pac extremity, the smallest molecules bound to the phage tail would be those originating from the first event of the encapsidation series and the next classes would be derived from successive packaging cycles (Figure 3A). Furthermore, the size step between each class would reflect the extent of terminal redundancy. Our experimental approach required only the measurement of double-stranded DNA (dsDNA). It was therefore considerably more accurate than previous size determinations which involved measurements of molecules with both single and double-stranded regions (Tye et al., 1974; Morelli et al., 1979). The tail-chromosome complexes obtained by EDTA treatment of SPP1wt virions were cleaved with SnaBI that recognizes a single sequence within fragment EcoRI-1 of the SPP1 wild-type genome (040 kb from the pac extremity, which corresponds to 4893 bp before the following pac sequence in the concatemer used as substrate for packaging (Chai et al., 1993; Figure 3A). EM measurements of DNA molecules from SPP1wt yielded a distribution of fragments attached to the tails where different size classes were discernable. The distinction between these classes, however, was hampered by the small steps between successive classes (not shown). Therefore, we repeated the experiment with a mutant carrying deletion X (3427 bp; Chai et al., 1993) located between the SnaBI site and pac of the SPP1 genome. Since reduction in SPP1delX110 chromosome size is compensated by an increase in

Headful Packaging and Fate of the Cleaved DNA Ends

terminal redundancy (Tavares et al., 1992), the spacing between the genomic location of successive headful cleavages is increased accordingly (from about 1.6 kb, which is the average terminal repetition for SPP1wt chromosomes, to approximately 5.0 kb; Figure 3A). This strategy, also used in the seminal work from Tye et al. (1974), allowed a clear distinction of five discrete classes of DNA molecules and a few extra long fragments (Figure 3B). In addition to the visual inspection procedure we also used an analytical criterion to define classes. The terminal redundancy was calculated by subtracting the distance between the SnaBI site and pac (1459 bp in SPP1delX110) from the average size of molecules in the first class of Figure 3B to D and this value (5.05 kb) was used as the step between classes. The mean value between the size expected for two successive classes (vertical arrows in the graphics of Figure 3B to D) was used as a cut-off length to assign molecules to a specific class. The distribution in Figure 3B shows that 48% of the SPP1delX110 molecules measured were derived from the first encapsidation cycle while this percentage was only 032% for SPP1wt (not shown). Classes derived from subsequent packaging cycles of SPP1delX110 are characterized by a continuous reduction in the number of elements (27, 9, 8 and 6% for classes 2, 3, 4 and 5, respectively; Figure 3B). The extent of circular permutation in the complete population of mature chromosomes can be calculated by subtracting the size of the smallest DNA fragment in the first class from the length of the largest molecule measured (29 kb (70% of the genome size) for SPP1delX110 and 16 kb (36%) for SPP1wt). This value divided by the terminal redundancy size gives the packaging series length (i.e. the number of sequential headfuls). Based on the data obtained with our phage preparations we calculated a maximum number of 6 and 12 headfuls from a single packaging series for the deletion mutant and wild-type, respectively. Comparison between SPP1wt and SPP1delX110, as well as other quantifications not shown, revealed that the frequency of SPP1 chromosomes derived from the first encapsidation cycle varied among different lysates. Therefore, we determined if this frequency changes during SPP1 infection as found in the case of bacteriophage P22 (Adams et al., 1983). Host cells infected with SPP1delX110 were lysed at 20 and 120 minutes after infection by addition of lysozyme and chloroform to culture samples. Assembled phages were purified and the topology of their chromosomes was investigated as described above. The distribution of size classes was considerably different at the two stages of infection (Figure 3C, D): while within 20 minutes the vast majority of packaged DNA molecules is derived from the first encapsidation cycle (82%) and the few remaining from the second (16%) and third (1%) cycles, at least eight classes can be distinguished and full circular permutation of the chromosome population is observed after 120 minutes. As would be expected considering the

Figure 3. Processive headful packaging in bacteriophage SPP1. A, Sequential encapsidation initiated at pac. The heavy lines in the centre represent the concatemeric substrate DNA for encapsidation from SPP1wt (top) and SPP1delX110 (bottom). Deletion end points marked in the SPP1wt sequence are connected to the site of deletion in SPP1delX110 and the deleted region is also highlighted in the SPP1wt concatemer (black boxes). SnaBI (S) and pac cleavage sequences are indicated. Vertical arrows show the position of SmaI cleavage (Santos et al., 1986) in the first genomic unit of the SPP1wt concatemer. The thicker arrow indicates three close sites. All restriction sites shown for the SPP1wt concatemer are found at identical positions in SPP1delX110 (not represented for simplification of the scheme). The pac terminal fragment generated by SmaI cleavage (SmaI-4) is represented by a bar. Ensembles of mature molecules cleaved with SnaBI are depicted above (SPP1wt) or below (SPP1delX110) the concatemeric structures using the convention of Figure 2A. B, Electron microscopic measurement of individual DNA molecules bound to the phage tail after disruption of SPP1delX110 (standard lysate) with EDTA and cleavage with SnaBI. All molecules present in a pre-defined area of the carbon grid were measured to provide an unbiased statistical distribution but only those bound to tail structures are depicted. Vertical arrows indicate the size limits used to assign molecules to a specific packaging cycle (class) using the criterion described in the text. Data presentation and methods are as in Figure 2B. C, D, Results from experiments similar to B except that the SPP1delX110 phage population was analysed 20 and 120 minutes after the initiation of infection, respectively. E, Increase in the number of packaging events per encapsidation series (average series length) during infection with SPP1wt (input multiplicity, 5 phage/bacterium). Beginning of infection is time 0. The average series length is calculated from the molar ratio between the SmaI-3 restriction fragment and the pac terminated fragment (SmaI-4; arrow on the insert) in restriction patterns (insert) as described in the text (experimental time points are represented by diamonds and the full line is a visual aid fitted by hand). The yield in viable phage progeny determined by titration with B. subtilis YB886 is also shown (triangles).

958 sequential character of the process, the number of molecules derived from successive packaging cycles decreased continuously in all the populations analysed. This reduction, however, is much less pronounced for the chromosomes of phages assembled late in infection suggesting that the encapsidation series length increases during infection due to a higher packaging processivity. The distribution observed for the SPP1 lysate described above (Figure 3B) is intermediate between the extreme cases shown in Figure 3C and D and might represent the most common situation for lysates produced routinely under laboratory conditions. One feature of the set of individual molecules of the same class is the considerable variability in size, normally within a range of 2 to 3 kb (Figure 3B to D). This variation is not only due to experimental error in the measurements but, more importantly, to inaccuracy of the headful cleavage in SPP1 (Humphreys & Trautner, 1981). To assess the contribution of the two factors we performed parallel measurements of replicative forms of phage M13mp18 chromosomes linearized with EcoRI or HindIII. Since these dsDNA molecules have a constant length (7250 bp; Yanisch-Perron et al., 1985), the variation observed gives an estimate of the accuracy of our measurements. An average size of 7.25(20.12) kb (01.7% variation) was found for a dataset of 65 M13 molecules (data not shown). In comparison the length of fragments of the first class in Figure 3C is 6.66(20.69) kb (010.3% variation). The experimental data follow a normal (gaussian) distribution. Comparable results were obtained with measurements of DNA molecules adsorbed to mica. We could thus conclude that the variation in dimensions observed within the first class of SPP1 fragments (02.5 kb; 5.6% of the genome size) was due essentially to imprecision in headful cleavage. Heterogeneity in size of molecules produced from subsequent encapsidation cycles is normally also within a range of 2 to 3 kb but increases in the case of the last classes, an effect attributable to additive imprecision from several packaging rounds. This variability in chromosome size is larger than the step between classes (terminal redundancy) in the case of SPP1wt preventing the discrimination of clear size classes. Furthermore, it raises also the probability of error to define groups of molecules generated from late packaging cycles in long encapsidation series from SPP1delX110. The imprecision of headful cleavage in SPP1 is comparable to the reports for other phages packaging their DNA by a headful mechanism (Mu: 02.7 kb, Chow & Bukhari, 1977; P22: 01.5 kb, Casjens & Hayden, 1988) and yields a random distribution of molecular ends at the level of resolution of the EM measurements. The processivity of DNA packaging increases during infection Complementary information on the processivity of packaging was obtained from quantification of

Headful Packaging and Fate of the Cleaved DNA Ends

the pac terminated restriction fragment in encapsidated DNA (Figure 3A, E). This segment of DNA has one end derived from cleavage at pac and the other from digestion with the restriction enzyme used. It is produced, therefore, only once per packaging series and its molar ratio to other fragments reveals the average frequency of packaging cycles initiated at pac relative to the total number of encapsidation events (Ba¨chi & Arber, 1977; Jackson et al., 1978; Ratcliff et al., 1979). This ratio is inversely proportional to the average number of encapsidation events per packaging series (average series length = 1/pac fragment ratio). Densitometric analysis of restriction patterns obtained by SmaI cleavage of packaged viral DNA showed that the ratio between the pac terminal fragment (SmaI-4; nomenclature according to Santos et al., 1986; Figure 3A) and equimolar fragments derived from the SmaI digest (SmaI-3, 5 or 6) varied among different phage lysates independently of the SPP1 strain. Since the average series length was observed to increase with time after P22 infection (Adams et al., 1983) and our EM data also demonstrated a temporal variation in the case of SPP1 we followed the rise of this parameter after infection with SPP1wt (Figure 3E). A value of 1.8 (equivalent to a pac ratio of 0.56) was determined at 20 minutes post-infection, indicating that most packaged chromosomes were derived from the first encapsidation cycle. The series length then increased continuously reaching 5.4 (pac ratio 0.19) for the last time point tested in the experiment of Figure 3E (120 minutes). Early after infection the number of progeny virions increased rapidly until starting to stabilize around 40 minutes post-infection (Figure 3E), as would be expected for an intracellular growth curve (Klotz & Spatz, 1971). Interestingly, between 40 and 120 minutes post-infection, an interval characterized by a lower rise in virion number, there is a significant increase in series length. This observation suggests that a majority of phages assembled during this period carry chromosomes derived from long packaging series. The effect of the increase in packaging series length with time on the full population of progeny chromosomes can best be observed by comparing the class distributions of SPP1delX110 molecules shown in Figure 3B to D. Determination of the average series length based on the pac ratio for this mutant, however, is biased by the large circular permutation which implies that headful cleavages occur spread within most of the genome reducing also the representativity of ‘‘true’’ restriction fragments. The monotonic increase in the average packaging series length is very similar to the one reported by Adams et al. (1983) for P22, although, in contrast to the experimental design of these authors, we could not prevent normal lysis of the host cell (starting after 30 minutes under the infection conditions used) and consequent stabilization of the progeny particle titre (Figure 3E). Thus, for interpretation of

Headful Packaging and Fate of the Cleaved DNA Ends

959

the data we have to take into consideration that the number of host bacteria is reduced during the experiment and that, at late time points, the rise in average series length in the entire population of viral chromosomes is due to phage particles produced by only a fraction of the initial infected cells. Consequently, the values presented are, in fact, an underestimation of the series length for late packaging events. We did not observe considerable re-infection, another phenomenon that could complicate data analysis, as confirmed by the modest rise in progeny after the first cycle of phage growth (Figure 3E). The DNA end to be packaged last is ejected first In order to distinguish whether the DNA end fixed to the tail structure is the first or last to exit the virion upon infection we took advantage of the observation that DNA ejection through the phage tail can be triggered in vitro by several mild denaturing conditions (e.g. 2.5 M sodium iodide; see below) leaving, in a large number of cases, one end of the chromosome still bound to the tail region distal to the head (Figure 4A). This end was identified by restriction with endonucleases (BglII or SnaBI) and measurement of the DNA molecules remaining attached to the phage tail, a strategy similar to the one employed in the experiment of Figure 2. SPP1sizS was used due to the virtual absence of circular permutation, a feature which simplifies the analysis (see Figure 2A). DNA ejection from SPP1sizS particles was triggered by NaI treatment. After cross-linking with glutaraldehyde and digestion with the endonuclease, the material was prepared for EM. DNA bound to the tails of empty phages was then measured. The size distributions obtained show that the chromosome extremity associated with the tail is more distal to the SnaBI site and closer to the BglII cleavage site (Figure 4B). Thus, the first end to exit the tail is the one generated by headful cleavage. The finding of a minor population of molecules considerably smaller than expected is most probably due to partial ejection and not to heterogeneity in chromosome size as it was not detected when the same phage preparation was used for EDTA treatment (Figure 2; data not shown). Absence of any molecules significantly larger than expected in the case of the BglII experiment (13 kb) demonstrates that the chromosome extremity attached to the connector (generated by headful cleavage; Figures 1, 2) is invariably the first to exit the virion upon ejection. Size of the DNA segment protected by the tail To measure the size of the chromosome fragment associated to the tail of virions disrupted with EDTA, we performed a DNAase protection assay. Following incubation with the chelating agent, the tail-DNA complex generated was exhaustively

Figure 4. Chromosome ejection from SPP1sizS particles. A, Ejection triggered in vitro by NaI (2.5 M). Two ghosts can be observed with DNA leaving the tail end distal from the head (arrowheads). The bar represents 0.5 mm. The material for EM was prepared by adsorption to mica. B, Identification of the first chromosome end that is ejected. The upper thick line represents the SPP1 chromosome. Cleavage positions for the endonucleases BglII (B) and SnaBI (S), and the initial cut at pac performed by the terminase are represented by vertical arrows. Packaging is from right to left. The lower part shows results from EM measurements of individual DNA molecules associated with the phage tail after ejection from SPP1sizS triggered by NaI. Measurements of intact and restricted molecules (SnaBI and BglII) are depicted. Data presentation and methods are as in Figure 2B.

960

Headful Packaging and Fate of the Cleaved DNA Ends

that it is buried by protein (Figure 5). Evidence for the association between this fragment and the tail was obtained from sedimentation of the nuclease treated material in a sucrose gradient. Tail and the small DNA segment co-sedimented while the bulk of the disrupted heads moved considerably faster and disaggregated material remained at the top of the gradient as monitored by several techniques (extraction with phenol and DNA end-labelling, SDS-PAGE, Western blot, electron microscopy; data not shown). The size of the protected fragment (55 ˚ for DNA to 67 bp, equivalent to 187 to 228 A conformation B) is considerably shorter than the tail length. Thus we hypothesize that it is located only in the connector region, most probably ˚ extending from the portal protein pore (0105 A height; Dube et al., 1993; Tavares et al., 1995) to the lower connector part bound to the helical tail (unpublished results). In the case of bacteriophages T4 and T7, Zachary & Black (1992) observed also the protection of a short segment of DNA (040 bp for T4) in portal protein-DNA complexes isolated from acid treated virions. Disruption of the mature phage particle and DNA ejection Figure 5. Size of the DNA fragment protected by the tail after disruption with EDTA. Nuclease treated complexes, not deproteinized or extracted with phenol, were radioactively labelled and separated in a 8% native gel PAGE together with appropriate molecular weight markers (Mr ; pBR322 digested with HaeIII).

treated with nucleases after which part of it was extracted with phenol. Both samples, with or without deproteinization, were then radioactively labelled using T4 polynucleotide kinase to identify free 5' DNA ends. A small segment of labelled DNA, ranging in size from 55 to 67 bp (minor variations within these values were observed in independent experiments), was found exclusively in samples extracted with phenol, demonstrating

The experiments described above were based on the availability of reproducible methods to disassemble virions. We found two distinct types of partial disruption. First, treatment with chelating agents (e.g. EDTA, citrate) caused separation of heads from tails or head disruption (Figure 1; data not shown), an effect that was proportional to the mass of DNA present inside the virion head (Table 1). No further disaggregation of the structures was observed when the concentration of the chemical was increased. Second, incubation with a variety of denaturing agents (KSCN [e2 M], NaI [e2.5 M], guanidinium hydrochloride [e3 M], formamide [e35%]) led to the appearance of empty phage particles (ghosts) the majority of which ejected their DNA through the tail (Figure 4A,

Table 1. Characterization of virion disruption triggered by EDTA or NaI Electron microscopy Tails with DNA bound to the region proximal to Disruption conditions 100 mM EDTA 2.5 M NaI

Phage

Phage viability

Intact phages

Empty heads or ghosts

Tails

Head

Tail spike

Both sides

+Glu

X110 S X X110 S X

− ++ ++++ + + +

+ +++ ++++ − − −

+ − + ++ ++ ++

+ − − − − −

+++ + + − − −

− − − ++ ++ ++

− − − + ++ −

N ND ND P ND ND

Phages with different amounts of packaged DNA (SPP1delX110 [45.6 kb], SPP1sizS [44.1 kb] and SPP1delX [43.1 kb] Tavares et al., 1992) were treated with 100 mM EDTA for 30 minutes at 37°C or with 2.5 M NaI for ten minutes at 30°C. The effect of cross-linking the virions with glutaraldehyde ( + Glu) before chemical treatment is also shown. Phage viability, expressed as percentage of the titre from the initial lysate, was evaluated by titration with YB886. EM quantifications were based on samples prepared by mica adsorption (100 to 200 structures from each preparation were classified): + + + +, more than 75%; + + +, more than 50%; + +, more than 20%; +, more than 5%; −, below 5%. P, indicates that pre-treatment with glutaraldehyde prevents disruption, N, reveals no effect, and ND was not determined. The cross-linking reaction causes a severe reduction in phage viability (<10−5 %).

961

Headful Packaging and Fate of the Cleaved DNA Ends

Figure 6. DNA ejection triggered by NaI (2.5 M). Kinetics of DNA ejection from SPP1 mutants with various chromosome sizes (SPP1wt [45.9 kb], SPP1sizS [44.1 kb], SPP1delX [43.1 kb] and SPP1delX110 [45.6 kb]; Tavares et al., 1992). Ejection was stopped by cross-linking with EDAC at the time points shown above each gel lane and released DNA was probed by restriction with EcoRI. Digested material (equivalent to 02 × 108 pfu/well) was resolved in a 1% agarose gel and stained with ethidium bromide. The material retained in the slot and the smear in the upper part of the gel observed for short time points likely reveals DNA present in intact phages. S, SPP1wt purified DNA digested with EcoRI; P, phage sample not exposed to NaI but processed like the other samples; 0', sample taken and cross-linked immediatelly after manual mixing with NaI (<15 seconds).

Table 1 and data not shown). In this case, raising the concentration of denaturant caused disassembly of the head and ultimately of the tail. Chemical cross-linking of the phage particle with glutaraldehyde or 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDAC) prior to exposure to denaturing agents (e.g. 2.5 M NaI) prevented DNA from exiting while it did not avoid disruption by EDTA (Table 1). Increase in temperature from 30 to 37°C to 45 to 50°C, which by itself does not affect SPP1 viability, was observed to have a synergistic effect in both disruption methods (not shown). When chromosome ejection from phage particles was stopped by cross-linking at various times after addition of NaI and released DNA was probed with endonuclease EcoRI, we observed an increase in the amount of digested DNA after long periods of exposure to the chemical (Figure 6). However, the stoichiometry between fragments in the restriction pattern remained constant, being similar to ratios found for purified DNA. Thus, ejection seems to be an all or nothing process which, after being initiated, leads to release of virtually the full chromosome (Figures 4, 6). Cross-linking can stop the process by preventing the trigger and/or initiation but is most probably not able to freeze the movement of DNA through the tail tube once it has started. The level of headfilling apparently does not have a major effect on the kinetics of chromosome release (at least in the minute time range) as judged by the near identical amounts of DNA ejected from various SPP1 mutants with different chromosome size (Figure 6).

The majority of particles treated with NaI ejected most of the chromosome except for the end packaged last (Figure 4). This association between the tail extremity distal to the head and SPP1 DNA is relatively strong as a significant population of the protein-DNA complexes are maintained during equilibrium isopycnic centrifugation in CsCl gradients. It is also further stabilized by glutaraldehyde cross-linking (not shown). However, the interaction does not promote any detectable protection of the attached DNA against nuclease treatment (data not shown). Binding of DNA to the tail tip after ejection thus seems to not be specific, either because it is weaker or because the stretch of DNA associated to the tail has a heterogeneous size. The nature of this interaction and its potential physiological role, if any, remain to be determined.

Discussion Organization of the packaged DNA in the virion and functional implications SPP1 virions are highly resistant structures. Their disruption by artificial treatments involves either separation of head and tail (Figure 1) or ejection of DNA through the tail tube (Figure 4). Each effect is triggered by distinct chemicals which target different components of the phage particle. Incubation with chelating agents leads to the first type of disassembly. Its severity, as observed also for other bacteriophages, is proportional to the amount of DNA inside the phage head (l:

962 Parkinson & Huskey, 1971; P22: Casjens et al., 1992; SPP1: Tavares et al., 1992). Such correlation between the level of headfilling and the requirement for divalent cations can be interpreted as if these small molecules could act to shield the repulsive forces between the phosphate backbones closely packed in the confines of the phage capsid. A tenfold increase in the local Mg2+ concentration in the DNA packaged state was indeed reported for bacteriophage P22 (Aubrey et al., 1992). SPP1 DNA ejection through the tail can be triggered by a variety of chemicals whose unique common feature appears to be to act as denaturing agents. We believe that, instead of a particular effect, their destabilizing action mimics the natural signal initiated by adsorption of SPP1 to the B. subtilis cell and induces the conformational changes leading to DNA exit by the ‘‘legitimate way’’. It is interesting to note that, at least in vivo, the signal which is initiated at the phage adsorption apparatus (tail spike?) has to be communicated to the opposite end of the tail (connector) where the first DNA end to exit is bound (Figures 1, 5). Chromosome movement out of the virion is believed to be entropically driven by the high concentration of DNA inside the phage capsid (cf. Earnshaw & Casjens, 1980) even though we could not detect an evident correlation between the level of headfilling and the kinetics of ejection in vitro (Figure 6). DNA ejection leads to loss of the tail spike, the tail tube becomes empty if all DNA is released (stain penetrates in negative staining preparations), and the phage head appears partly collapsed indicating that presence of nucleic acid on its interior is an important requirement for structural stability (Figure 4A and EM observations not shown). The DNA end created by headful cleavage (last to be packaged) during SPP1 morphogenesis is invariably associated to the phage tail region which binds the icosahedral head (connector; Figure 1) showing that the two DNA extremities of the viral chromosome have a distinct topology in the virion. This feature was also reported for other phage systems using chemical cross-linking (Thomas, 1974; Chattoraj & Inman, 1974) suggesting that it may play an important role in viral physiology, the most obvious being to facilitate the polar exit of DNA through the tail during viral infection. The findings that the DNA end attached to the tail is the first to exit when ejection is triggered in vitro (Thomas, 1974; Figure 4) and that l mutants defective in this interaction are not able to eject their DNA (Thomas et al., 1978) favour such interpretation. Since the extremities of the SPP1 chromosome are permuted, the relevant structural feature might be that one of these ends is positioned for ejection and the choice between the two is dictated by the way DNA is packaged (cf. Chattoraj & Inman, 1974; Earnshaw & Casjens, 1980). The interaction between tail and DNA is apparently stronger in SPP1 than in other phages as it does not require any cross-linking to be stably conserved

Headful Packaging and Fate of the Cleaved DNA Ends

upon virion disruption. The absence of any observable structure bound to the first DNA end encapsidated (Figure 1) suggests that it might remain free inside the phage head. We presently do not know whether this is the case or if some loose association fixes this extremity during DNA encapsidation. The distinction between both possibilities is a relevant issue to understand how SPP1 DNA is being packed while it is translocated into the pro-capsid. Sequential headful packaging Mature SPP1 phages have the DNA end generated by headful cleavage bound to the tail. In tail-DNA complexes produced by disruption of virions with chelating agents (Figure 1) this association provides a useful physical marker to characterize the topology of individual mature chromosomes by EM, particularly when it is used in combination with cleavage at unique positions of the genome by restriction endonucleases (Figures 2, 3). The alignment of restricted tail-DNA complexes allows one to distinguish size classes whose order reflects the sequential packaging cycles from which the population of partially circularly permuted chromosomes was generated (Figure 3B). Thus, these studies provide quantitative information on the mechanism of sequential headful packaging by SPP1. The distribution of the viral chromosome population according to their original encapsidation cycle (packaging classes) shown in Figure 3 supports the concept that DNA packaging initiates at a unique position within the SPP1 genome (pac) and proceeds unidirectionally in a sequential fashion (Morelli et al., 1979; Bravo et al., 1990). This fact implies that the relative frequency of molecules generated after the first packaging cycle (classes C2 , C3 , . . . , Cn in Figure 7A) depends on the frequencies of molecules derived from encapsidation cycles occurring before in the packaging series and is, ultimately, a function of the percentage of mature chromosomes derived from the first encapsidation event (C1 ). It is thus possible to predict the full distribution of chromosomes if it is dictated by a strictly sequential process and no major biases are introduced by factors inherent to the encapsidation mechanism or infection conditions that would favour, a priori, the occurrence of packaging series with a defined size. In this case the probability (P) that after each encapsidation round another packaging cycle will follow is identical for any cycle in the packaging series (i.e. P = P1 = P2 = P3 = . . . = Pn ; see also Casjens & Hayden, 1988). P can be calculated from the percentage of molecules (C1 ) generated in the first packaging cycle (P = 1 − C1 ; see formulation in Figure 7), a value easily derived from the frequency of pac terminal fragments in endonuclease restriction profiles of phage DNA (Results and Figure 3E). The distribution for the full population of chromosomes can then be determined based on P.

963

Headful Packaging and Fate of the Cleaved DNA Ends

Figure 7. Processivity of sequential headful packaging. A, Sequential encapsidation initiated at pac. The upper heavy bar represents a substrate concatemer for packaging. Classes of mature molecules (C) are presented below and numbered according to the packaging cycle in which they were generated. B, Distribution of the different classes of chromosomes based on a theoretical model (thin lines) and on the data from Figure 3B to D (top right and bottom panels; experimental time points are represented by diamonds and the thick line is a visual aid fitted by hand). The model assumes that in a packaging series there is a constant probability (P) that one packaging round will follow another, independently of which encapsidation event is considered. The full population of chromosomes generated by sequential encapsidation can be described by the formula: C1 + C2 + C3 + · · · + Cn = 1

0ECE1

(1)

or, considering that all packaging series initiate with C1 and proceed sequentially, C1 + C1 P1 + C1 P1 P2 + · · · + C1 P1 P2 . . . Pn−1 = 1 0EPE1

(2)

in which Cn represents the number of elements in class n (see A) and Pn−1 is the probability that sequential packaging will proceed from class n − 1 to n. If we consider that P = P1 = P2 = P3 = . . . = Pn −1 then (2) can be simplified to the geometric row: C1 + C1 P + C1 P2 + · · · + C1 P n−1 = 1

(3)

C1 (P n − 1)/(P − 1) = 1

(4)

or if we do not impose a limit for the packaging series length (n : a) and P < 1 then C1 /(1 − P) = 1 (5)

or P = 1 − C1

Thus, if we know the percentage of molecules generated in the first packaging cycle (C1 ) we can easily calculate P and then the number of elements in all classes by using equation (3). This procedure was used to generate the theoretical distributions (thin lines in the Figure graphics) according to different percentual values of C1 . In the graphics, C1 is the ordinate value corresponding to the intercept of each curve with C = 1.

Figure 7 shows a comparison between the theoretical curves calculated for different values of C1 with the three distributions from Figure 3B to D. The experimental data fit in general to the distributions expected from the sequential model showing that packaging of SPP1 chromo-

somes is dictated primarily by a sequential mechanism. According to the above formulation, the value of P determined for each distribution indicates the frequency of processive headfuls (C2 + C3 + · · · + Cn ) relative to the number of packaging initiation events (C1 ). For P to give a

964 measurement of the absolute processivity of the packaging machinery, however, the duration of a complete encapsidation cycle would have to be negligible relative to the duration of the infectious cycle. This condition is not fulfilled during normal infections. A main limiting factor is cell lysis that causes premature termination of the packaging series. Long infection periods are indeed associated to an increase in the frequency of processive versus initiation packaging events that implies the rise of P from 0.16 (20 minutes) to 0.65 (120 minutes) for the distributions shown in Figure 7B. Thus if the distribution of viral chromosomes would be dictated uniquely by the processivity of the packaging apparatus higher P values should be observed. Some deviations of the experimental distributions relative to the theoretical curves from Figure 7B are observed in the case of the two datasets representative of late stages in infection. Since P increases as the infectious cycle proceeds, we attribute this effect to asynchronous lysis of part of the infected host cells. A sub-population of progeny phages would thus have chromosomes generated from packaging series with different P values. Their contribution, however, is apparently not very significant as it does not bias much the total distribution of mature DNA molecules, particularly in the case of virion populations derived from cultures lysed artificially (20 and 120 minutes in Figure 7B). Sequential headful packaging requires series initiation (pac cleavage and first encapsidation cycle) and extension events (processive headfuls). As proposed by Adams et al. (1983) based on studies with bacteriophage P22, the size of the encapsidation series is dictated by the relative frequency of each of the two processes. The frequency of initiation events is high at early stages during SPP1 infection while later extension events predominate as revealed by the rise from 1.8 to more than five average sequential headfuls per packaging series (Figure 3E). Our data do not allow us to discriminate whether this increase is due to variation on the physical length of the substrate concatemer, to alterations in the properties of the packaging machinery or to the kinetics of packaging (i.e. the time required to complete a full encapsidation round). It is also possible that all these factors contribute to some extent to the balance between initiation at pac and processivity. It is presently not known how DNA concatemers are generated (rolling circle replication and/or recombination) during SPP1 infection. Independently of the mechanism, their formation and packaging shall occur, at least partly, in parallel. The concatemer sizes would thus be defined by competition between both processes. Terminase cleavage at pac, the first event in DNA encapsidation, is one of the factors that reduces the size of the concatemer. Chai et al. (1992) demonstrated that the terminase endonucleolytic activity is regulated since the frequency of all pac sequences cut

Headful Packaging and Fate of the Cleaved DNA Ends

(packaged or free in the host cytoplasm) is kept relatively constant between 12 and 30 minutes after initiation of SPP1 infection even when no DNA packaging occurs. Maintenance of a low steadystate level of cleaved pac sites is essential to ensure sequential packaging. However, even when only 22 to 27% of all the pac sequences are cut (Chai et al., 1992), the size of the substrate DNA will be considerably limited if cleavage occurs at pac sites distributed randomly within the SPP1 DNA concatemers. Random cuts would especially prevent the occurrence of long packaging series as those observed in Figure 3B, D as well as transduction of plasmid concatemers carrying the pac sequence (Bravo & Alonso, 1990). It is therefore possible that the frequency of pac cleavage drops late during infection and a mechanism could be operative which prevents further pac cleavage on concatemers where sequential packaging was initiated. In the case of coliphage P1, Sternberg & Coulby (1987) demonstrated that the substrate dimension is indeed one main factor to limit the number of sequential packaging events. These authors inserted pac in a molecule of ‘‘infinite size’’, the host chromosome, and observed that such a construct leads to the unidirectional encapsidation of five to ten consecutive headfuls of host DNA instead of the three to four headfuls normally observed for P1 DNA. Burger & Trautner (1978) observed that after SPP1 encapsidation was initiated (from 16 minutes post-infection on) the majority of phage DNA molecules present in the infected cell have a size similar to mature chromosomes. To conciliate this result with the occurrence of long packaging series requiring substrates larger than eight to tenfold the size of the SPP1 chromosome, we propose that after the first cut at pac the packaging apparatus follows closely the replication machinery and most of the new pac sites being generated would be encapsidated before attack by the terminase. A related strategy would also be possible if formation of SPP1 concatemers involves recombination events, especially if cross-talk occurs between the recombination machinery and the packaging apparatus (see Wu et al., 1995 for a terminase-dependent recombination event). Long encapsidation series at late stages of infection would be a natural consequence of the mechanism proposed. Differences in packaging processivity during infection can also be due to variations in the relative intracellular pools of terminase, substrate DNA and pro-capsids. These could affect the length of the encapsidation series in different ways. If cleavage at pac is a rate limiting step, initiation would require higher concentrations of terminase than processive packaging. Accordingly, Adams et al. (1983) showed that a reduction in terminase levels, particularly of its major subunit, affects the frequency of packaging initiation rather than processivity during bacteriophage P22 encapsidation. An opposite effect would be expected if pro-capsids are present in limiting amounts

965

Headful Packaging and Fate of the Cleaved DNA Ends

implying that molecules cleaved at pac are encapsidated but subsequent packaging cycles would be rare due to the lack of pro-capsids. Furthermore, the rate of sequential packaging events is limited by the speed of DNA translocation into the pro-capsid and the time of formation of a new complex between the substrate DNA and another pro-capsid. In summary, DNA packaging is a dynamic process occurring pari passu with the generation of substrate concatemers. Various factors contribute to define the distribution of progeny chromosomes. Under conditions favouring SPP1 fast multiplication the encapsidation series are short due to a predominance of packaging events initiated at pac and to early host lysis. If infection lasts a long time a large number of sequential headfuls is ensured by the high processivity of the packaging machinery.

Material and Methods Bacterial and phage strains B. subtilis and SPP1 strains were as described (Tavares et al., 1992). Materials and standard methods Endonucleases SphI, SmaI and SnaBI were purchased from New England Biolabs and BglII and EcoRI were from Boehringer-Mannheim. T4 polynucleotide kinase was obtained from New England Biolabs. Micrococcal nuclease was obtained from Pharmacia Biotech and Benzonase from Eurogentec. Glutaraldehyde and EDAC were purchased from Fluka and Sigma, respectively. EDTA and NaI were obtained from Merck. All chemicals were of analytical grade. Titration and amplification of bacteriophage SPP1 wild-type and mutants, viral particle purification and phage DNA extraction were as described (Chai et al., 1992; Tavares et al., 1992). All phage preparations used were purified by centrifugation through a CsCl step gradient with the exception of the lysates used for the experiment of Figure 3E. Anti-gp6 and anti-SPP1 polyclonal sera were obtained by immunization of rabbits with purified gp6 (Tavares, 1992; Dube et al., 1993) and caesium chloride purified SPP1 particles, respectively. Determination of the pac fragment molar ratio in DNA packaged during SPP1 infection Exponentially growing YB886 were infected with SPP1wt (input multiplicity = 5) following our standard procedure (Tavares et al., 1992). Samples were taken at defined times and lysis was induced by addition of 2% (v/v) chloroform and 0.5 mg/ml lysozyme, shaken for five additional minutes at 37°C, and kept on ice until the end of the kinetics. Lysate processing and DNA extraction were as described by Tavares et al. (1992). DNA digested with SmaI was resolved in agarose gels (0.8%, w/v) stained with ethidium bromide or Syber GreenTM (Molecular Probes), scanned in a FluorImager (Molecular Dynamics), and quantified using the ImageQuant software (Molecular Dynamics). Representativity of the pac fragment (SmaI-4) was determined by comparison of the corresponding band signal with the

intensities of SmaI-3, 5 and 6, taking into account the ratio between the Mr s of the species being analysed. Phage particles disruption and cross-linking The chemicals to be tested were diluted from a stock solution in TBT buffer (Biswal et al., 1967) to the concentration required (8 ml total volume), equilibrated briefly at 30°C and the reaction was initiated by addition of 2 ml from SPP1wt phages (01012 pfu/ml). Incubation was for 30 minutes at 30°C. The reaction was stopped by 50-fold dilution with TBT or phosphate buffer (for cross-linking, see below) followed by characterization: DNA loss from the viral capsid was tested by spotting 10 ml of the suspension in a 2 ml drop of ethidium bromide (10 mg/ml) and visualization under ultraviolet light, and phage viability was evaluated by spotting serial dilutions of the suspension (done in microtitre plates) on a lawn of B. subtilis YB886. After this preliminary screening, selected samples were observed in the electron microscope using negative staining, direct adsorption to mica or cytochrome c techniques (see EM methods). For cross-linking, phage particles or disrupted material (01012 pfu/ml) in phosphate buffer (10 mM sodium phosphate (pH 7.5), 10 mM MgCl2 ) were mixed, with an identical volume of a 1% (w/v) solution of glutaraldehyde or 50 mM freshly prepared EDAC. After incubation for five minutes at 30°C the reaction was quenched with 100 mM glycine or 50-fold dilution in TBT. Aliquots were then further treated with NaI or EDTA, or characterized as described above. Buffer changes were done by dialysis in 0.025 mm VS filters (Millipore) when required. Results obtained with either glutaraldehyde or EDAC cross-linking were essentially identical. Complexes of phage DNA bound to the connector region of the tail for electron microscopy were obtained by treatment of virions (1010 to 1011 pfu) with 100 mM EDTA for 45 minutes at 45°C. The chelating agent was eliminated by dialysis against bi-distilled water for 90 minutes in 0.025 mm pore size VS filters (Millipore) and the sample was then used for EM studies or digested with an endonuclease (BglII, SphI or SnaBI), normally overnight, before preparation for EM measurements to determine the sizes of DNA molecules. Complexes for measurement of the size of the DNA fragment protected by the tail structure were prepared identically except that higher amounts of phages were used (01011 pfu), and the sample was then treated overnight with Benzonase (150 units/ml) and micrococcal nuclease (75 units/ml). After DNA digestion, half of the material was extracted twice with phenol and once with chloroform. Residual organic solvents were eliminated by dialysis against TE (10 mM Tris-HCl (pH 7.5), 1 mM EDTA) in 0.025 mm filters. DNA present in both deproteinized and non-treated samples was end-labelled with [g-32P]ATP using T4 polynucleotide nuclease (Sambrook et al., 1989). Non-incorporated label was removed by filter dialysis. The size of the radioactive species was estimated by running the samples and appropriate Mr standards in native or denaturing PAGE (Sambrook et al., 1989) followed by analysis of the dried gels using a PhosphorImager (Molecular Dynamics). When required, the nucleasestreated disrupted phages (1012 pfu) were concentrated in a Centricon-50 (Amicon) and applied to 5% to 20% sucrose (w/v) gradients in 0.5 × TBT buffer prepared with a Biocomp apparatus. Centrifugation was at 35,000 rpm for 90 minutes in a SW50.1Ti rotor (4°C). Individual fractions were dialysed against 0.1 × TBT, concentrated tenfold under vacuum and an aliquot was

966 processed as above. Additionally, all fractions were characterized by SDS-PAGE (Laemmli, 1970), Western blot with anti-gp6 and anti-SPP1 sera (Tavares et al., 1995), and electron microscopy. Complexes of phage DNA ejected through the tail structure were prepared from SPP1sizS. Ejection was triggered by incubation with 2.5 M NaI for ten minutes at 30°C and the complexes were normally cross-linked with glutaraldehyde or EDAC. The ghost-DNA complexes were then used for EM measurements (see below) or analysed by equilibrium isopycnic centrifugation. For the latter experiment the particle suspension was mixed with a CsCl solution (1.533 g/cm3 ) and centrifuged for 40 hours at 35,000 rpm in a SW50.1Ti rotor (20°C). Individual fractions were dialysed against 10 mM Tris-HCl (pH 7.6), 10 mM MgCl2 and characterized by EM observation and/or restriction analysis. Kinetics of DNA ejection from various SPP1 strains (01010 pfu/reaction) triggered by NaI were performed as described above except that the reaction was stopped by cross-linking at the desired time points (see Figure 6). After quenching with glycine and dialysis against 10 mM Tris-HCl (pH 7.6), 10 mM MgCl2 , the DNA released was probed by restriction with EcoRI. Digestion products were resolved in 1% (w/v) agarose gels. Quantifications of bands in the restriction patterns were done from direct scans of the ethidium bromide stained gels (Fluorimager) or from negatives of photographed gels (densitometer [Molecular Dynamics]) using the ImageQuant software. Electron microscopy Negative staining with 1% (w/v) uranyl acetate was performed as described by Valentine et al. (1968). Preparation of the SPP1 phages for EM after treatment with different chemicals to release the DNA was done by direct adsorption to mica (Portmann et al., 1974) and by cytochrome c spreading with the droplet technique (Lang & Mitani, 1970) essentially as described by Spiess & Lurz (1988). For length measurements we preferred samples prepared by the droplet technique because most of the molecules could be traced easily due to unfolding of the DNA in the cytochrome c surface film. Except for some control experiments done to measure all DNA of a given area, we selected only those DNA molecules which had a tail or an empty phage attached at one end. Data analysis and graphics Statistical analysis was done using the program Statistica (StatSoft). The graphics derived from the polynomial formulation presented in Figure 7 and all data graphics were constructed using Micrografx Charisma or Microsoft Excel.

Acknowledgements We are thankful to Gerhild Lu¨der for performing part of the electron microscopy experiments. We thank Elmar Dro¨ge (Institut fu¨r Festko¨rperphysik, TU Berlin) for help with the mathematical formulation and generation of the theoretical curves from Figure 7, Anja Dro¨ge for discussions, and Angie Hofmann for critical reading of the manuscript. We also acknowledge Mark Achtman for statistical analysis. P.T. was partly supported by an EC fellowship (Contract ERBBIOTCL1923104).

Headful Packaging and Fate of the Cleaved DNA Ends

References Adams, M. B., Hayden, M. & Casjens, S. (1983). On the sequential packaging of bacteriophage P22 DNA. J. Virol. 46, 673–677. Aubrey, K. L., Casjens, S. R. & Thomas, G. J., Jr (1992). Secondary structure and interactions of the packaged dsDNA genome of bacteriophage P22 investigated by Raman difference spectroscopy. Biochemistry, 31, 11835–11842. Ba¨chi, B. & Arber, W. (1977). Physical mapping of BglII, BamHI, EcoRI, HindIII, and PstI restriction fragments of bacteriophage P1 DNA. Mol. Gen. Genet. 153, 311–324. Bazinet, C. & King, J. (1985). The DNA translocating vertex of dsDNA bacteriophage. Annu. Rev. Microbiol. 39, 109–129. Biswal, N., Kleinschmidt, A. K., Spatz, H. C. & Trautner, T. A. (1967). Physical properties of the DNA of bacteriophage SP50. Mol. Gen. Genet. 100, 39–55. Black, L. W. (1989). DNA packaging in dsDNA bacteriophages. Annu. Rev. Microbiol. 43, 267–292. Bravo, A. & Alonso, J. C. (1990). The generation of concatemeric plasmid DNA in Bacillus subtilis as a consequence of bacteriophage SPP1 infection. Nucl. Acids Res. 18, 4651–4657. Bravo, A., Alonso, J. C. & Trautner, T. A. (1990). Functional analysis of the Bacillus subtilis bacteriophage SPP1 pac site. Nucl. Acids Res. 18, 2881–2886. Burger, K. J. & Trautner, T. A. (1978). Specific labelling of replicating SPP1 DNA. Analysis of viral DNA synthesis and identification of phage dna-genes. Mol. Gen. Genet. 166, 277–285. Casjens, S. & Hayden, M. (1988). Analysis in vivo of the bacteriophage P22 headful nuclease. J. Mol. Biol. 199, 467–474. Casjens, S. & Hendrix, R. (1988). Control mechanisms in dsDNA bacteriophage assembly. In The Bacteriophages, (Calendar, R., ed.), vol. 1, pp. 15–91, Plenum Press, New York. Casjens S., Wyckoff, E., Hayden, M., Sampson, L., Eppler, K., Randall, S., Moreno, E. & Serwer, P. (1992). Bacteriophage P22 portal protein is part of the gauge that regulates packing density of intravirion DNA. J. Mol. Biol. 224, 1055–1074. Chai, S., Bravo, A., Lu¨der, G., Trautner, T. A. & Alonso, J. C. (1992). Molecular analysis of the B. subtilis bacteriophage SPP1 region encompassing genes 1 to 6. The products of gene 1 and gene 2 are required for pac cleavage. J. Mol. Biol. 224, 87–102. Chai, S., Szepan, U., Lu¨der, G., Trautner, T. A. & Alonso, J. C. (1993). Sequence analysis of the left end of the Bacillus subtilis bacteriophage SPP1 genome. Gene, 129, 41–49. Chai, S., Kruft, V. & Alonso, J. C. (1994). Analysis of the Bacillus subtilis bacteriophages SPP1 and SF6 gene 1 product: a protein involved in the initiation of headful packaging. Virology, 202, 930–939. Chai, S., Lurz, R. & Alonso, J. C. (1995). The small subunit of the terminase enzyme of Bacillus subtilis bacteriophage SPP1 forms a specialized nucleoprotein complex with the packaging initiation region. J. Mol. Biol. 252, 386–398. Chattoraj, D. K. & Inman, R. B. (1974). Location of DNA ends in P2, 186, P4 and lambda bacteriophage heads. J. Mol. Biol. 87, 11–22. Chow, L. T. & Bukhari, A. I. (1977). Bacteriophage Mu genome: structural studies on Mu DNA and Mu mutants carrying insertions. In DNA Insertion

Headful Packaging and Fate of the Cleaved DNA Ends

Elements, Plasmids, and Episomes (Bukhari, A. I., Shapiro, J. A. & Adhya, S. L., eds), pp. 295–306, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Deichelbohrer, I., Messer, W. & Trautner, T. A. (1982). Genome of Bacillus subtilis bacteriophage SPP1: structure and nucleotide sequence of pac, the origin of DNA packaging. J. Virol. 42, 83–90. Dube, P., Tavares., P., Lurz, R. & van Heel, M. (1993). Bacteriophage SPP1 portal protein: a DNA pump with 13-fold symmetry. EMBO J. 12, 1303–1309. Earnshaw, W. C. & Casjens, S. (1980). DNA packaging by the double-stranded DNA bacteriophages. Cell, 21, 319–331. Esche, H., Schweiger, M. & Trautner, T. A. (1975). Gene expression of bacteriophage SPP1. I. Phage directed protein synthesis. Mol. Gen. Genet. 142, 45–55. Gill, G. S. & MacHattie, L. A. (1976). Limited permutations of the nucleotide sequence in bacteriophage T1 DNA. J. Mol. Biol. 104, 505–515. Hendrix, R. W. (1978). Symmetry mismatch and DNA packaging in large bacteriophages. Proc. Natl Acad. Sci. USA, 75, 4779–4783. Humphreys, G. O. & Trautner, T. A. (1981). Maturation of bacteriophage SPP1 DNA: limited precision in the sizing of mature bacteriophage genomes. J. Virol. 37, 832–835. Jackson, E. N., Jackson, D. A. & Deans, R. J. (1978). EcoRI analysis of bacteriophage P22 DNA packaging. J. Mol. Biol. 118, 365–388. Klotz, G. & Spatz, H. Ch. (1971). A biological assay for intracellular SPP1 DNA. Mol. Gen. Genet. 110, 367–373. Laemmli, U. K. (1970). Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 227, 680–685. Lang, D. & Mitani, M. (1970). Simplified quantitative electron microscopy of biopolymers. Biopolymers, 9, 373–379. Morelli, G., Fisseau, C., Behrens, B., Trautner, T. A., Luh, J., Ratcliff, S. W., Allison, D. P. & Ganesan, A. T. (1979). The genome of B. subtilis phage SPP1: the topology of DNA molecules. Mol. Gen. Genet. 168, 153–164. Parkinson, J. S. & Huskey, R. J. (1971). Deletion mutants of bacteriophage lambda. I. Isolation and initial characterization. J. Mol. Biol. 56, 369–384. Portmann, R., Sogo, J. M., Koller, T. & Zillig, W. (1974). Binding sites of E. coli RNA polymerase on T7 DNA as determined by electron microscopy. FEBS Letters, 45, 64–67. Ratcliff, S. W., Luh, J., Ganesan, A. T., Behrens, B., Thompson, R., Montenegro, M. A., Morelli, G. & Trautner, T. A. (1979). The genome of Bacillus subtilis phage SPP1: the arrangement of restriction endonu-

967 clease generated fragments. Mol. Gen. Genet. 168, 165–172. Santos, M. A., Almeida, J., Lencastre, H., Morelli, G., Kamke, M. & Trautner, T. A. (1986). Genomic organization of the related Bacillus subtilis bacteriophages SPP1, 41c, r15 and SF6. J. Virol. 60, 702–707. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). Molecular Cloning. A Laboratory Manual, 2nd edit., Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Spiess, E. & Lurz, R. (1988). Electron microscopic analysis of nucleic acids and nucleic acid-protein complexes. Methods Microbiol. 20, 293–323. Sternberg, N. & Coulby, J. (1987). Recognition and cleavage of the bacteriophage P1 packaging site (pac). I. Differential processing of the cleaved ends in vivo. J. Mol. Biol. 194, 453–468. Tavares, P. (1992). Func a˜o da proteı´na portal na encapsidac a˜o do DNA pelo bacterio´fago SPP1. PhD thesis, Universidade de Coimbra, Portugal. Tavares, P., Santos, M. A., Lurz, R., Morelli, G., Lencastre, H. L. & Trautner, T. A. (1992). Identification of a gene in Bacillus subtilis bacteriophage SPP1 determining the amount of packaged DNA. J. Mol. Biol. 225, 81–92. Tavares, P. Dro¨ge, A., Lurz, R., Graeber, I., Orlova, E., Dube, P. & van Heel, M. (1995). The SPP1 connection. FEMS. Microbiol. Rev. 17, 47–56. Thomas, J. O. (1974). Chemical linkage of the tail to the right-hand end of bacteriophage lambda DNA. J. Mol. Biol. 87, 1–9. Thomas, J. O., Sternberg, N. & Weisberg, R. (1978). Altered arrangement of the DNA in injection-defective lambda bacteriophage. J. Mol. Biol. 123, 149–161. Tye, B. K., Huberman, J. A. & Botstein, D. (1974). Non-random circular permutation of phage P22 DNA. J. Mol. Biol. 85, 501–532. Valentine, R. C., Shapiro, B. M. & Stadtman, E. R. (1968). Regulation of glutamine synthetase. XII. Electron microscopy of the enzyme from Escherichia coli. Biochemistry, 7, 2143–2152. Valpuesta, J. M. & Carrascosa, J. L. (1994). Structure of viral connectors and their function in bacteriophage assembly and DNA packaging. Quart. Rev. Biophys. 27, 107–155. Wu, C. H. H., Lin, H. & Black, L. W. (1995). Bacteriophage T4 gene 17 amplification mutants: evidence for initiation by the T4 terminase subunit gp16. J. Mol. Biol. 247, 523–528. Yanisch-Perron, C., Vieira, J. & Messing, J. (1985). Improved M13 phage cloning vectors and host strains: nucleotide sequences of the M13mp18 and pUC19 vectors. Gene, 33, 103–119. Zachary, A. & Black, L. W. (1992). Isolation and characterization of a portal protein-DNA complex from dsDNA bacteriophage. Intervirology, 33, 6–16.

Edited by J. Karn (Received 7 August 1996; accepted 4 October 1996)