The Conformation of Packaged Bacteriophage T7 DNA: Informative Images of Negatively Stained T7

JOURNAL OF STRUCTURAL BIOLOGY ARTICLE NO. SB973895

120, 32–43 (1997)

The Conformation of Packaged Bacteriophage T7 DNA: Informative Images of Negatively Stained T7 Philip Serwer, Saeed A. Khan, Shirley J. Hayes, Robert H. Watson, and Gary A. Griess Department of Biochemistry, The University of Texas Health Science Center, San Antonio, Texas 78284-7760 Received February 12, 1997, and in revised form May 12, 1997

and T7; illustrated in Fig. 1a) and then drawing DNA into the capsid. The procapsid is both rounder and smaller than the capsid of the mature bacteriophage (Fig. 1c). In the case of both bacteriophages T7 and T7 capsid I, but not most other bacteriophages, the internal proteins form a cylinder that projects from the tail attachment site into the region of packaged DNA (illustrated in Figs. 1a and 1c). The T7 and T3 internal cylinder, external tail, and outer shell are joined by a doughnut-shaped 12-mer called the connector (Kocsis et al., 1995; details for T3 are described in Carazo et al., 1986); connectors are also present in the capsids of the other studied doublestranded DNA bacteriophages (reviews: Bazinet and King, 1985; Steven and Trus, 1986; Wurtz, 1992; Valpuesta and Carrascosa, 1994). An axial hole in the internal cylinder–connector–tail complex is assumed to be the path through which T7 DNA is injected into a host cell (reviews: Earnshaw and Casjens, 1980; Casjens, 1985; Steven and Trus, 1986; Black, 1989; Serwer, 1989). After infecting Escherichia coli with T7, incompletely packaged DNA (ipDNA) is found primarily in capsid II (Fig. 1b), the expanded, polygonal form of the (tail-free) capsid (Khan et al., 1995). Thus, most packaging of DNA appears to occur after capsid I has converted to capsid II. For revealing the protein outer shell of doublestranded DNA bacteriophages, electron microscopy of a negatively stained specimen has been a primary technique used. However, with one exception (bacteriophage G: Sun and Serwer, 1997), none of the internal structure is revealed by negative staining of the intact bacteriophage (reviews: Steven and Trus, 1986; Wurtz, 1992). However, for several icosahedral bacteriophages, including T7, electron microscopy of negatively stained DNA expelled from its capsid has revealed wrapping of DNA segments around a common axis (Richards et al., 1973). Cryoelectron microscopy is an alternative procedure that has revealed the arrangement of subunits in the outer shells of the capsids of intact bacteriophage (examples:

Within the icosahedral protein outer shell of bacteriophage T7, a 40-kbp DNA genome occupies a cavity also occupied by a protein cylinder that projects into the DNA from the outer shell. However, neither the internal cylinder nor separately resolved DNA segments are revealed in the conventional negatively stained specimens of intact bacteriophage T7. In the present study, a procedure of negative staining is used that reveals both internal proteins and separately resolved segments of packaged DNA during electron microscopy of intact particles of a hybrid T7 bacteriophage; the hybrid is genetically T7, except for a tail fiber gene that has a segment from the T7-related bacteriophage, T3. The negatively stained packaged DNA segments of this hybrid bacteriophage are found to be wrapped around the axis of the internal cylinder. To obtain additional information about the conformation of packaged T7 DNA, electron microscopy is performed of negatively stained capsids that are incompletely filled with DNA (ipDNA-capsids); a procedure is described for improved isolation of ipDNA-capsids from lysates of hybrid bacteriophage T7-infected cells. The packaged DNA segments of ipDNA-capsids are found not to be wrapped around any axis. Images of ipDNA-capsids are explained by the hypothesis that DNA does not achieve its wrapped condition until the capsid is more than 40% full of DNA. Wrapping via folding is, therefore, proposed to explain the images of DNA packaged in bacteriophage T7. r 1997 Academic Press

INTRODUCTION

Most double-stranded DNA bacteriophages consist of a protein capsid that contains DNA packaged in a cavity enclosed by the capsid’s polygonal outer shell. Attached to the outer shell are both an external protein projection (tail) and internal proteins. These bacteriophages package DNA by first assembling a DNA-free, tail-free capsid (procapsid, also called capsid I in the case of the related bacteriophages, T3 1047-8477/97 $25.00 Copyright r 1997 by Academic Press All rights of reproduction in any form reserved.

32

PACKAGED BACTERIOPHAGE T7 DNA

33

FIG. 1. A schematic drawing of T7 capsids. The capsids shown are (a) capsid I, (b) capsid II, and (c) the capsid of mature bacteriophage T7. Proteins are indicated by p, followed by the number of the gene that encodes the protein (Studier and Dunn, 1983).

Dokland and Murialdo, 1993; Prasad et al., 1993; Thuman-Commike et al., 1996). Cryoelectron microscopy has also revealed parallel-packed DNA strands that are (1) aligned in the direction of the long axis of elongated variants of bacteriophage T4 (Lepault et al., 1987) and (2) wrapped around an axis that appears to be the axis of the internal bacteriophage T7 cylinder (Booy et al., 1992; Cerritelli et al., 1996). However, the resolution of these studies was not sufficient to trace the path of the packaged DNA double helix; wrapping via folding was not distinguished from wrapping via winding of T7 DNA. In addition, the internal cylinder was not visualized during cryoelectron microscopy of T7. The internal cylinder of T7 had been demonstrated by negative staining of either DNA-free capsids or partially disrupted particles of bacteriophage T7 (review: Steven and Trus, 1986). To obtain a more complete analysis of the conformation of packaged bacteriophage T7 DNA, procedures are needed whereby (1) both packaged DNA segments and the internal cylinder are simultaneously visualized and (2) the resolution limits of electron microscopy are either overcome or bypassed. In a recent study (Khan et al., 1997), a hybrid was constructed that is genetically T7, except for a segment from T3 in the gene (gene 17; Studier and Dunn, 1983) that encodes the protein that forms six fibers that extend from the T7 tail (Fig. 1c). During preliminary characterization of this hybrid bacteriophage, it was found to have properties useful for both observation of the internal structure of intact particles of bacteriophage and isolation of capsids that are partially filled with DNA, i.e., capsids that have ipDNA1 (called ipDNA-capsids). In the present communication, this hybrid bacteriophage is used, together with an altered procedure of negative staining, to (1) simultaneously visualize both the internal cylinder and strands of packaged T7 DNA and (2) obtain images of ipDNA-capsids that constrain the possible ways in which DNA can be wrapped around the internal T7 cylinder. 1 Abbreviations used: ipDNA, incompletely packaged DNA; ipDNA-capsid, capsid with incompletely packaged DNA.

MATERIALS AND METHODS Bacteriophages Wild-type bacteriophages T7 and T3 (Hausmann strain of T3) were received from Dr. F. W. Studier (Studier, 1979). The A1 tail fiber hybrid of Khan et al. (1997) is the T7–T3 hybrid used here; both the nucleotide sequence and the construction of the hybrid genome have been described in Khan et al. (1997). T7 temperaturesensitive in gene 19 (T7ts19) is described in Khan et al. (1995). The protein encoded by gene 19 is necessary for DNA packaging, but not for either capsid assembly or DNA synthesis (reviewed in Studier and Dunn, 1983). A T7–T3 hybrid that had the ts19 mutation (T7–T3ts19) was constructed by use of a standard genetic cross (Studier, 1969). The host for all bacteriophages was Escherichia coli BB/1. Bacteriophages were both grown in 2 3 LB medium (20 g tryptone, 10 g yeast extract, 5 g NaCl per liter of water) and purified by centrifugation in cesium chloride density gradients (Serwer, 1976). Purified bacteriophages were dialyzed against Tris/Mg buffer: 0.2 M NaCl, 0.01 M Tris–Cl, pH 7.4, 0.001 MgCl2. The concentration of bacteriophages was determined by measurement of optical density at a wavelength of 260 nm (39.7 µg packaged DNA per OD unit; Bancroft and Freifelder, 1970). Electron Microscopy To negatively stain particles of bacteriophage, these particles were diluted to the indicated concentration. Subsequently, the bacteriophage particles were adsorbed to a carbon substrate, washed with water, and negatively stained with uranyl acetate, by use of procedures previously described (Serwer, 1976). From the first contact with uranyl acetate until the specimen appeared dry, the standard time elapsed was 20–30 sec. When indicated, this time was extended to 80–100 sec by slowing the withdrawal of stain during drying of the specimen. Specimens were observed in a Philips 301 transmission electron microscope. The procedure used to make the carbon substrate yields a carbon substrate that is ultra-adherent (Serwer, 1976). Processing and Reproduction of Electron Micrographs To both process and reproduce images of negatively stained particles, unless otherwise indicated, electron micrographs recorded on film were digitized by use of a video camera. Both the procedure of digitization and the preexisting software are reviewed in Griess et al. (1992). To both sort digitized images for statistical analysis and manipulate digitized images for averaging, a macro program was added to the previously developed software, NIH IMAGE. To test a digitized image for rotational symmetry, rotational averaging was used (Markham et al., 1963). For any given integral value of n, n-fold rotational symmetry is assumed for a structural element, whenever this element is both reinforced during averaging of images separated by 2p/n radians, but not reinforced during averaging of images separated by either 2p/(n 1 1) or 2p/(n 2 1) radians (see Markham et al., 1963; Crowther and Amos, 1971). Because of both intraparticle variabil-

34

SERWER ET AL.

ity of staining and, primarily, the low weight of potential symmetrical elements, rotational power spectra (Crowther and Amos, 1971) were not used to test for symmetry (see also Kocsis et al., 1995). Procedures previously described (Griess et al., 1992) were used to photographically both reproduce and mark digital images. The NIH IMAGE macro program is available from the authors, on request. Purification of ipDNA Capsids In a previous study (Khan et al., 1995), the existence was demonstrated of ipDNA-capsids in lysates of bacteriophage T7infected E. coli; the relative amount of these ipDNA-capsids was increased by growing T7ts19 (instead of wild-type T7) at a permissive temperature. However, because of their aggregation, the ipDNA-capsids could not be purified sufficiently for electron microscopy; stability of these particles appeared not to be a major problem (Khan et al., 1995). Both yield and purity of ipDNAcapsids were measured (and are measured here) by gel electrophoresis of DNA released from capsids (see Fig. 2, below). In the present study, further purification of ipDNA-capsids was achieved by use of a procedure based on the following observations made by comparisons performed one at a time: (1) Use of the T7–T3 tail fiber hybrid, instead of wild-type T7, resulted in both improved yield and improved purity of ipDNA-capsids, when the procedure

of purification was that of Khan et al. (1995). (2) Both yield and purity were further improved when the nonionic detergent Triton X-100 (1.0%) was included in both the lysate (the Triton X-100 was added immediately after lysis) and all subsequently used buffers, including the buffer used for the final dialysis to remove cesium chloride from ipDNA-capsids. When Triton X-100 was omitted during dialysis, the ipDNA-capsids were lost, presumably because of adherence to the dialysis tubing. (3) As expected, use of T7–T3ts19 (instead of the hybrid without the ts19 mutation) resulted in a further improvement in yield, when the procedure used was that of Khan et al. (1995), improved by the use of Triton X-100. (4) When buoyant density centrifugation was performed for a time (40 hr) greater than the time (approximately 20 hr) needed to bring DNA-free capsids to an equilibrium density, a further improvement in purity occurred. After observation of the density gradient by use of light scattering, a band of host debris was observed that formed between 20 and 40 hr (density 5 1.24 g/ml; data not shown). Electron microscopy revealed that the use of Triton X-100 dramatically reduced the contamination of ipDNAcapsids with vesicles of host membrane. The use of the T7–T3 hybrid presumably reduced bacteriophage–capsid aggregation; this hybrid was originally constructed to reduce the adherence of bacteriophage T7 to agarose gels. Thus, the procedure used to purify ipDNA-capsids was that of Khan et al. (1995), modified by (1) the use of the T7–T3ts19 bacteriophage, (2) the addition of 1.0% Triton X-100 in both the lysate (obtained at 30°C) and all buffers subsequently used for purification, including the buffer used for the final dialysis, and (3) increase to 40 hr of the time of buoyant density centrifugation in a cesium chloride density gradient. After buoyant density centrifugation, DNase-digested ipDNA-capsids were analyzed by first expelling packaged DNA (ipDNA in the case of an ipDNA-capsid) and then performing agarose gel electrophoresis; procedures are described in the next section. In Fig. 2a, the profile of packaged DNA is revealed by ethidium staining for ipDNA from fractions whose density (g/ml) is indicated at the top of a lane (f indicates DNA from purified bacteriophage T7). This profile consisted of a band at the position of mature length T7 DNA (horizontal arrow in Fig. 2a), accompanied by shorter ipDNA, some of which formed bands (the length of a band-forming ipDNA in kilobase pairs is indicated in the middle of Fig. 2). These band-forming ipDNAs had previously been identified by probing (Khan et al., 1995), but in quantities 1–2 orders of magnitude lower than those achieved here. Agarose Gel Electrophoresis

FIG. 2. Analysis of packaged DNA by gel electrophoresis. For ipDNA-capsids fractionated by buoyant density centrifugation, agarose gel electrophoresis was performed of DNA released from DNase-treated capsids. Either (a) staining with ethidium or (b) probing was used to detect DNA. The stained gel of (a) is the same as the probed gel of (b). The density of the fractions used is indicated above a lane (g/ml); f indicates mature T7 DNA. The vertical arrow indicates the direction of electrophoresis; the arrowheads indicate the origins of electrophoresis. The horizontal arrow indicates the position of mature (39.936-kbp) T7 DNA; the numbers in the middle indicate the lengths (kilobase pairs) of band-forming ipDNAs.

To perform agarose gel electrophoresis of DNA expelled from bacteriophage capsids, the procedures of Khan et al. (1995) were used for both digesting capsid-external DNA with DNase I and performing electrophoresis with ethidium staining of DNA. Electrophoresis was performed at 22°C for 18 hr, at 0.7 V/cm, through a 0.7% submerged, horizontal agarose gel (Seakem LE agarose, FMC Bioproducts, Rockland, ME), cast in 0.09 M Tris–acetate, pH 8.4, 0.001 M EDTA. For both improved detection sensitivity and improved quantification, fractionated DNA was also detected by probing with a 32P-labeled oligonucleotide that was specific for the right end of T7 DNA (Khan et al., 1995). A phosphorimager (Molecular Dynamics, Sunnyvale, CA) was used for quantification of 32P (Johnston et al., 1990). RESULTS

Improved Visualization of Internal Structure for Intact Bacteriophages When the unaltered procedure of negative staining (i.e., drying within 30 sec; Materials and Methods) was used to observe T7–T3 hybrid bacteriophage

PACKAGED BACTERIOPHAGE T7 DNA

particles prepared at a concentration of 6.6 3 1013 particles/ml, the particle images were observed (Fig. 3a) to have a radius more variable than images of wild-type T7 previously observed in less concentrated specimens (Serwer, 1976; Steven and Trus, 1986). The presumed explanation for the increased variability in radius is increased variability in the dehydration-induced shrinkage and flattening of the bacteriophage particles (Serwer, 1977). In addition, in Fig. 3a, the central region of the image of many particles was denser (more intensely stained) than the central region of images obtained from specimens of either wild-type T7 or the T7–T3 hybrid prepared at 1.1 3 1013 particles/ml. When the prolonged procedure of negative staining (Materials and Methods) was used to prepare T7–T3 hybrid bacteriophages at 6.6 3 1013 particles/ml, the density of most of the central, DNA-containing region further increased for some (but never all) particles of the hybrid bacteriophage. The region of this increase in density, however, usually surrounded another region that had a comparatively low density (Fig. 3b). Sometimes this low-density region was at the center of the image of the particle. In this case, the lowdensity region had a roughly circular (often polygonal) outline; within the low-density region, often a comparatively dense center was found. This dense center presumably was the image of a hole. In Fig. 3b, some of these latter particles are indicated by arrows 1; images at higher magnification are presented below. In Fig. 3b, the particles with internal structure (to be called internally contrasted particles) appear to be located primarily in 1- to 10-µm zones that were interspersed with zones that had particles with the more conventional appearance of particles in Fig. 3a. Therefore, the internally contrasted appearance of the former particles appears to be the result of a transition in their structure that is caused by local (i.e., zones of 1–10 µm) conditions of staining. The field of Fig. 3b is typical of most fields of internally contrasted particles. Because the internally contrasted particles were not observed when either 0.5% uranyl acetate (instead of 1.0% uranyl acetate) was used as the negative stain or the specimen concentration was reduced to 1.1 3 1013 particles/ml during the prolonged procedure of negative staining (data not shown), the thickness of the stain appears to play a role in producing internally contrasted particles. The influence of at least one other factor is indicated by the proximity of conventional to informative particles (Fig. 4a). The low-density region in the internally contrasted particles of Fig. 3b and 4a is possibly the image of the T7 cylinder–connector–tail complex. By this hypothesis, the improved visibility of the cylinder–connector–tail complex is caused by an increase

35

in the contrast between the internal cylinder and the surrounding uranium stain–DNA complex. The following observations support the hypothesis that the low-density region is the image of the cylinder– connector–tail complex: (1) The low-density region sometimes contacts the image of the outer shell, as though the internal cylinder were being observed perpendicular to its axis; several images of this type are in Fig. 3b (arrows 2). (2) The low-density region is, as stated above, sometimes both central and shaped like the image of a cylinder–connector–tail complex viewed in a direction parallel to its axis (more details are described below). However, if the electron transparent region was always the image of an intact cylinder–connector–tail complex, then the axis-perpendicular projection should reveal the coaxial tail attached to the outside of the capsid. These expected tails are missing in Fig. 3b (arrows 2). For the particles indicated by arrows 2 in Fig. 3b, possibly the tail has bound to the support film and either folded under the particle or become dislodged. Objects that resemble dislodged tails were observed (arrows 3 in Fig. 3b). In the following sections, a central low-density region will be assumed to be the image of at least the internal cylinder; the tail and connector may be, but are not necessarily, also present. Most internal cylinders (.80%) were viewed in a direction parallel to the cylinder’s axis; this preferential orientation is an indication that the bacteriophage’s tail adhered the bacteriophage particle to the support film. Attempts were made to obtain internally contrasted particles by applying the technique of Figs. 3b and 4a to bacteriophages T7 and T3. Although internally contrasted particles were observed, the frequency of internally contrasted particles was always 1–2 orders of magnitude lower for either T7 or T3 than it was for the T7–T3 hybrid bacteriophage (not shown). Conformation of Packaged DNA In the images of internally contrasted negatively stained particles, the internal cylinder was surrounded by comparatively electron-dense material that included uranium cations. However, for the following reasons, these cations were accompanied by packaged DNA (i.e., the DNA had not been expelled): (1) No evidence was observed of expelled DNA outside of the capsid. (2) DNA strands were sometimes observed inside of the capsid. Of 170 randomly selected internally contrasted particles observed in an orientation parallel to the axis of the internal cylinder, 127 had parallel, separately resolved strands of DNA (examples are in Fig. 5). These strands had a width of 1–2 nm, equal within experimental error to the width of a DNA double helix (2 nm). Thus, each

36

SERWER ET AL.

FIG. 3. Visualization of internal structure by electron microscopy. For a specimen of the T7–T3 hybrid bacteriophage (6.6 3 1013 particles/ml), the following times were used for negative staining, before observation by electron microscopy: (a) the standard time (20–30 sec) of exposure to uranyl acetate and (b) lengthened exposure (80–100 sec) to uranyl acetate. The length of the bar is 200 nm. The images were reproduced without a digitized intermediate. In (b), numbered arrows indicate the following particles: (1) internally contrasted bacteriophage particles that have an internal electron transparent region that has a roughly circular perimeter, (2) internally contrasted bacteriophage particles that appear to be viewed in an orthogonal direction, and (3) particles that appear to be tails broken from bacteriophage particles.

PACKAGED BACTERIOPHAGE T7 DNA

37

FIG. 4. Selected internally contrasted particles. By use of lengthened exposure to uranyl acetate, the T7–T3 hybrid bacteriophage was negatively stained for electron microscopy. (a) Three internally contrasted particles are shown, each of which is near a conventional particle. (b) In the top row, internally contrasted particles are shown that appear to have at least two fibers (indicated by white lines for the rightmost particle) radiating from a central object. In the bottom row, internally contrasted particles are shown that have either three or more bright zones in their outer shells. A rounded segment of an outer shell of the top–middle particle in (b) is indicated by an arrowhead. The length of the bar is 100 nm.

observed strand is assumed to be either a single segment of a DNA double helix or several such segments stacked in register in the direction of observation. The DNA strands were always wrapped around the internal cylinder. DNA strands were never seen in either conventional particles or internally contrasted particles not viewed parallel to the axis of the internal cylinder–connector–tail complex. Possible details of wrapping are illustrated in the

models of Fig. 6a–6c. The electron micrographs of Fig. 5 do not discriminate among these models. An alternative approach to performing this discrimination is described below. Symmetry of the Capsid The assumption has been made that the cylinder– connector–tail complex (Fig. 1) is attached to the icosahedral T7 outer shell at an axis of fivefold

38

SERWER ET AL.

FIG. 5. The packaged DNA. By the procedure of Fig. 4, electron micrographs were obtained of particles viewed parallel to the axis of the cylinder–connector–tail complex. Images are shown that have resolved strands of DNA. The length of the bar is 100 nm.

rotational symmetry (review: Steven and Trus, 1986). However, this assumption has not yet been verified directly. For the outer shell of most particles viewed in a direction parallel to the internal cylinder (99 of 134 selected at random), inspection revealed approximately fivefold rotational symmetry (see, for example, the internally contrasted particles in Figs. 4 and 5). However, these particles often had an outer shell that was rounder than the outer shell of conventional negatively stained T7 (Figs. 4 and 5). In these rounder particles, one segment of the outer shell was more rounded than the remainder of the outer shell (example: arrowhead in Fig. 4b). Rounded segments were almost never observed in particles that had a hexagonal outer shell (see Fig. 1b). During screening of images of these particles, some also appeared to have fibers that, in projection,

extended from the capsid’s outer shell to the internal cylinder. These fibers were seen only when a particle was viewed in a direction parallel to the axis of the cylinder–connector–tail complex. Unlike the T7 tail fiber (Steven et al., 1988), these fibers did not appear to be bent (first row in Fig. 4b; for the rightmost particle, fibers are indicated by white lines). Nonetheless, the possibility exists that each fiber is either (1) one of the two straight segments previously found (Steven et al., 1988) to comprise a T7 tail fiber or (2) a tail fiber that has been altered (possibly folded on itself ) during adherence to the carbon support film; the capsid-proximal tail fiber segment is 16.4 nm long; the capsid-distal segment is 15.5 nm long (Steven et al., 1988); the T7 capsid has an outer radius of 30.1 nm (Stroud et al., 1981). In any case, at the point of outer shell–fiber overlap, a zone was

FIG. 6. Models of possible conformations of packaged DNA. For all three models, the capsid has the expanded polygonal outer shell of the capsid called capsid II (review: Steven and Trus, 1986). The DNA is (a) unidirectionally wrapped, or (b) bidirectionally wrapped so that, in projection, all folds face one point on the capsid’s outer shell, or (c) bidirectionally wrapped so that folds face random points on the capsid’s outer shell. The models of (a), (b), and (c) do not require, but are compatible with, inside-to-outside polarity of the position of the DNA ends. Such polarity has been found for both T7 and other double-stranded DNA bacteriophages (reviews: Black, 1989; Serwer, 1989).

PACKAGED BACTERIOPHAGE T7 DNA

usually observed that was considerably lighter than the remainder of the capsid’s outer shell (these zones will be called light zones; first row in Fig. 4b). Sometimes, the light zones were observed without any observable fiber (second row in Fig. 4b). Presumably, in this latter case, the fiber was present, but not visible. By inspection, both the fibers and the light zones appeared to be at sixfold symmetric positions, even though both fibers and light zones were sometimes absent from some positions that were expected for sixfold symmetry. To test more rigorously for symmetry, rotational averaging was applied to images of bacteriophage particles whose periphery appeared minimally distorted (i.e., the periphery had not acquired elliptical character) by drying during specimen preparation. Images were selected that also had an axially viewed cylinder–connector–tail complex at the center. The images were not selected for symmetry of the outer shell. Initially, images of six different particles were aligned so that the most well-revealed fiber was at the left; these six images were averaged (Fig. 7a). Then, rotational averaging of the image in Fig. 7a was performed. The result was the finding that some (but not all) of the polygonal character of the outer

39

shell was maintained for fivefold averaging (Fig. 7c), but not for fourfold averaging (Fig. 7b), sixfold averaging (Fig. 7d), or sevenfold averaging (Fig. 7e). That is, the averaged outer shell does have fivefold symmetry, but this symmetry is not perfect. This imperfection is caused, at least in part, by the partial rounding of the outer shell. It is presumably also caused by either imperfect alignment or comparatively subtle specimen preparation-induced distortion of particles. In the case of both the fibers and the light zones, reinforcement occurred during sixfold averaging (Fig. 7d), but not either fivefold (Fig. 7c) or sevenfold (Fig. 7e) averaging. Thus, within the limitations of the procedure used, the symmetry mismatch assumed for both T7 (review: Steven and Trus, 1986) and other bacteriophages (reviews: Murialdo and Becker, 1978; Bazinet and King, 1985; Valpuesta and Carrascosa, 1994) has been confirmed. Identification of ipDNA-Capsids Although wrapping of T7 DNA around the internal cylinder has been observed both here by negative staining and in Cerritelli et al. (1996) by cryoelectron microscopy (see also Booy et al., 1992), the details resolvable by these procedures have not been suffi-

FIG. 7. Rotational averaging of images of particles viewed parallel to the axis of the cylinder–connector–tail complex. By the procedure of Fig. 4, electron micrographs were obtained of hybrid bacteriophage particles viewed in a direction parallel to the axis of the cylinder–connector–tail complex. (a) An image is reproduced that is the average of six images of different particles aligned so that the best-revealed fiber is at the left. This averaged image was then subjected to rotational averaging by use of an n that was (b) 4, (c) 5, (d) 6, or (e) 7. The length of the bar is 50 nm.

40

SERWER ET AL.

cient to discriminate among models, like those in Fig. 6, for the conformation of packaged T7 DNA. In an attempt to discriminate among these models, electron microscopy was performed of the ipDNA within ipDNA-capsids. When observed by electron microscopy after negative staining, some particles in either the 1.388 or the 1.353 g/ml fraction of Figs. 2a and 2b were unaggregated; others were aggregated. The aggregates were smaller (usually 5–30 particles) than those previously observed (Khan et al., 1995). The particles observed had the following types of appearance: (1) Some particles had the rounded, comparatively small outer shell that is a characteristic of capsid I; some capsid I-like particles had an internal cylinder, some did not (arrow 1 in Fig. 8). (2) Other particles had the more angular, larger outer shell of capsid II, but had either no or comparatively little visible internal material (arrow 2 in Fig. 8). (3) Additional particles were indistinguishable from particles of mature bacteriophage T7 (arrow 3 in Fig. 8). (4) The most abundant particles had capsid II-like outer shells that enclosed a cavity that contained something that was not the internal cylinder (arrows 4 in Fig. 8). Possibly, the internal content of the type

FIG. 8. Electron microscopy of potential ipDNA-capsids. After negative staining without prolonged exposure to uranyl acetate, particles in the 1.388 g/ml fraction of Fig. 2 were observed by electron microscopy. The image was reproduced without a digitized intermediate. The length of the bar is 100 nm. The images of capsids are, on average, smaller than they are in previous figures. Because hybrid bacteriophage particles appear comparably small when added to 1.5% Triton X-100 just before specimen preparation, the presumed reason is reduced flattening caused by the presence of the nonionic detergent Triton X-100.

4 particles was ipDNA. In addition to the particles of types 1–4 in Fig. 8, a few (,1% of the capsid-like particles) tubular particles made of T7 outer shell protein (review: Steven and Trus, 1986) were among the particles observed in aggregates (not shown). To determine whether the type 4 particles were present in the relative amount expected of ipDNAcapsids, the molar ratio of ipDNA to mature-length DNA was determined from the gel profiles in Fig. 2b; this ratio was compared to the ratio of type 4 particles (possible ipDNA-capsids) to type 3 particles (bacteriophage) observed by electron microscopy. After categorization of 260 randomly selected particles from electron micrographs of the 1.388 g/ml fraction, this latter ratio was 1.32, not significantly different from the gel electrophoresis-determined molar ratio of ipDNA to mature length DNA, 1.38. The field of Fig. 8 was selected to have a fraction of type 4 particles that was higher than average. Thus, although the ipDNA-capsids were significantly contaminated with other particles, they were sufficiently purified so that they are identifiable in electron micrographs. The Features of ipDNA-Capsids In a previous study, the capsid component of all ipDNA-capsids made in vivo was found by nondenaturing gel electrophoretic analysis to be capsid II (Khan et al., 1995). This analysis was performed without concentrating particles and, therefore, without high concentration-induced aggregation of ipDNA-capsids. In agreement, electron microscopy revealed that the shape and thickness of the outer shell of type 4 particles were both those of capsid II. In contrast, no precedent appears to exist for the appearance of the packaged ipDNA of any virus. If DNA segments were wrapped around a common axis in the type 4 particles (i.e., ipDNA-capsids), the reduced DNA packing density, and therefore the increased DNA–DNA spacing, in these particles should make wrapping more visible than it was for the completed bacteriophage (see, for example, Richards et al., 1973). However, in a random sample of over 2000 type 4 particles, wrapping of the DNA, such as that observed in Fig. 5 for bacteriophage particles, was not observed. Instead, the internal material of type 4 particles appeared more heterogeneous, sometimes fibrous and sometimes less asymmetric, as though fibers were being observed in several different directions in relation to the long axis (Fig. 8). The internal fibrous objects were 2–6 nm wide, sometimes wider than a single DNA double helix (2 nm). Although some forms of DNA-free capsid II do contain the T7 internal cylinder after isolation (Serwer, 1980), most (.95%) type 4 particles were full enough so that protein from the internal cylinder, if visible at all, could account for no more than roughly 20% of

PACKAGED BACTERIOPHAGE T7 DNA

the internal material; no type 4 particles had internal material that resembled the internal cylinder. Thus, the internal material of the type 4 particles in Fig. 8 will be assumed to be ipDNA. The thicker fibers formed by the packaged ipDNA are potentially hairpin-like condensates. DISCUSSION

In the present study, the T7 cylinder–connector– tail complex has been observed, for the first time, in mature bacteriophage particles viewed parallel to the axis of this complex. Because the cylinder– connector–tail complex is viewed in projection, the contribution of the cylinder to the image cannot be separated from the contribution of either the tail or the connector. In any case, in this orientation, the strongest symmetry of this complex is sixfold (Serwer, 1976; Results); the strongest symmetry of the outer shell was found here to be fivefold. This symmetry mismatch, originally found for bacteriophage T4 (Moody, 1965; Eiserling, 1983), has been assumed for all double-stranded DNA bacteriophages (Hendrix, 1978; Bazinet and King, 1985; Valpuesta and Carrascosa, 1994). The data obtained here constitute the most conclusive experimental support for this assumption in the case of T7. For most negatively stained T7 bacteriophage particles in which the cylinder–connector–tail complex was both visible and parallel to the direction of observation, strands of DNA were also visible; the DNA appeared to be wrapped around the cylinder. This observation supports the observation of DNA wrapping by cryoelectron microscopy (Booy et al., 1992; Cerritelli et al., 1996). In Figs. 6a–6c, three possible forms of wrapping are illustrated. To distinguish these three forms of wrapping directly by use of electron microscopy, resolution of single DNA duplexes is required in all three dimensions; in only two dimensions, overlapping DNA strands prevent the resolving of folds such as those in Figs. 6b and 6c. Because this degree of resolution is not yet possible, a more indirect approach was used here to distinguish among models for the conformation of packaged T7 DNA: observation of DNA packaged in ipDNA-capsids. In the past, electron microscopy of ipDNA-capsids was not possible because ipDNAcapsids could not be isolated in sufficient purity– quantity for electron microscopy. For the ipDNAcapsids that have an ipDNA length greater than 10% of the length of mature T7 DNA, previously presented data (Khan et al., 1995) indicate that their formation is a result of the premature cleavage of DNA during DNA packaging in a T7-infected cell. Thus, although these particles are assumed to be unmaturable, they appear to accurately reflect the state of ipDNA during DNA packaging.

41

In the case of ipDNA-capsids from the 1.388 g/ml fraction of Fig. 2, the length of ipDNA will be assumed to be equal to the mean length determined from the data in Fig. 1b, 16 kbp. Because mature T7 DNA (39.936 kbp; Dunn and Studier, 1983) occupies 0.453 the volume of the cavity in which the DNA is packaged (Stroud et al., 1981), the average ipDNA of this fraction occupies 0.183 the volume in which it is packaged. This comparatively low DNA packing density in ipDNA-capsids provides space to separate ipDNA segments sufficiently so at least some were independently resolved, whether or not they formed an ordered array. The visualization of ipDNA segments appears to have been assisted by contrast enhancement provided by the negative stain. Thus, if ipDNA were wrapped around a common axis in the ipDNA-capsids, this wrapping should have been more visible than DNA wrapping was in T7–T3 hybrid bacteriophages. The complete absence of wrapping in images of ipDNA-capsids implies, therefore, that the ipDNA was not wrapped. In addition to assisting the visualization of packaged ipDNA strands, the comparatively low DNA packing density of ipDNA-capsids could conceivably make possible the occurrance of a change in the mode of DNA compaction, during either purification of ipDNA-capsids or preparation of ipDNA-capsids for electron microscopy. However, conversion of the DNA conformation of Fig. 6a to that of either Fig. 6b or 6c would require complete unraveling, followed by recompaction of the DNA. This process appears improbable in the confines of a bacteriophage capsid. The assumption is made here that this conversion did not occur during either purification or preparation of ipDNA-capsids for electron microscopy. The assumption is also made that this conversion did not occur while ipDNA-capsids were in a T7-infected cell. If so, then the absence of observed DNA wrapping in the ipDNA-capsids implies that the correct mode of wrapping in the mature bacteriophage has DNA strands in a conformation that is interconvertible with a conformation that is unwrapped. If so, then, during DNA packaging in an infected cell, wrapping did not occur until more than 40% of the DNA was packaged. Any model in which folding of the DNA repeatedly occurs, including those in both Figs. 6b and 6c, has DNA strands that can unwrap locally (i.e., between folds), without disturbing DNA segments further along the double helix. Specifically, wrapping of ipDNA in the conformation of either Fig. 6b or 6c is, if sufficient space is available inside of the capsid, interconvertible with a conformation that includes straightened DNA strands that are separated by hairpin-like folds. This type of conformation is also consistent with images of ipDNAs (i.e., Fig. 8),

42

SERWER ET AL.

although not proven by these images. The presence of folds in the models of Figs. 6b and 6c also explains the finding of the perturbed DNA secondary structure observed by chemical probing of packaged T7 DNA (reviews: Black, 1989; Serwer, 1989). In contrast, local unwrapping is impossible if the DNA is packaged in the conformation of Fig. 6a. Therefore, the above assumptions yield the conclusion that neither the model in Fig. 6a nor any other model that incorporates unidirectional wrapping explains the electron micrographs of ipDNA-capsids. The data obtained here do not rigorously discriminate between the models of Figs. 6b and 6c. However, the model of Fig. 6b provides the best explanation for the observation made here of rounding of one region of an internally contrasted particle that was viewed parallel to the axis of the cylinder–connector–tail complex. That is, an explanation of these images is that the fold-containing region of the packaged DNA is adjacent to the region distal to the rounded region of the capsid’s outer shell. The wrapping-by-folding of the model in Fig. 6b has been previously proposed for bacteriophage T7 (Serwer, 1986); however, the previously proposed axis of wrapping is wrong by p/2 radians. In the case of the comparatively large bacteriophage, G, evidence supports a model in which packaged DNA partitions to form several Fig. 6b-like wrapping-by-folding units (Sun and Serwer, 1997). In the case of bacteriophage T4, folding without wrapping has been proposed (Black et al., 1985). The use of prolonged staining to improve visualization of internal components appears to be a new approach to determining the structure of bacteriophages. Its effectiveness raises the question: How does this procedure work? Although the answer to this question is not rigorously known, the following observations appear relevant: (1) This procedure is more effective when used with the T7–T3 hybrid bacteriophage than it is when used with either T7 or T3 bacteriophage. (2) Effectiveness decreases as the concentration of uranyl acetate negative stain decreases below 1%; it increases as the concentration of bacteriophage particles increases. Because the T7–T3 hybrid proteins differ in amino acid composition from those of T7 only in a part of the tail fiber, these observations are explained by the following hypothesis: During drying of the specimen, the tail fiber triggers a transition that makes the internal structure more visible when bacteriophage particles are sufficiently exposed to concentrated uranyl acetate solution. This hypothesis is supported by the evidence under Results that internally contrasted particles have adsorbed by their tail to the substrate. That is, possibly the tail fiber-induced transition requires binding of tail fibers to the substrate. By this hypothesis, the transitionary state is unstable

enough so that it survives staining only if achieved while the bacteriophage particle is in a region of the uranyl acetate solution that is concentrated enough to preserve (fix) the particle in its unstable transitionary state. In addition, the transitionary state is either more accessible or more stable for the T7–T3 hybrid than it is for either T7 or T3 bacteriophages. To explain the enhanced contrast between the internal proteins and the packaged DNA of internally contrasted particles, the characteristics of the proposed unstable transitionary state must include either the exit of (contrast-producing) uranium cations from fluid-containing regions near proteins or the entry of uranium cations into regions near the DNA (or both). When the image of an internally contrasted particle is compared to the image of a neighboring conventional particle (see Fig. 4a), the image of the cylinder–connector–tail complex of the internally contrasted particle usually has an intensity indistinguishable from the intensity of the image of the conventional particle. However, the image of the DNA of the internally contrasted particle is significantly more dark than most of the image of the DNA of the conventional particle. Thus, entry of uranium cations into the DNA-containing region appears to be the primary, possibly the only, explanation. If so, then tail fibers (i.e., fibers that are external) should be less visible in the image of an internally contrasted particle than they are in the image of a conventional particle. Because the opposite is the case for the fibers observed in Figs. 4b and 7, these fibers most probably are internal (i.e., embedded in the DNA). Possibly, they are the fibers previously observed (Serwer, 1979) in gluteraldehyde-cross-linked capsid I viewed in a direction that was either parallel or perpendicular to the axis of the internal cylinder. We thank Dr. Alasdair C. Steven for critical comments, Karen Lieman and Azalea Mendez for technical assistance, Linda C. Winchester for typing the manuscript, and the Department of Pathology, The University of Texas Health Science Center at San Antonio, for providing facilities for electron microscopy. We gratefully acknowledge support from the U.S. National Science Foundation (MCB-9316660), the National Institutes of Health (GM24365), and the Robert A. Welch Foundation (AQ-764). REFERENCES Bancroft, F. C., and Freifelder, D. (1970) Molecular weights of coliphages and coliphage DNA. I. Measurement of the molecular weight of bacteriophage T7 by high-speed equilibrium centrifugation, J. Mol. Biol. 54, 537–546. Bazinet, C., and King, J. (1985) The DNA translocating vertex of dsDNA bacteriophage, Annu. Rev. Microbiol. 39, 109–129. Black, L. W. (1989) DNA packaging in dsDNA bacteriophages, Annu. Rev. Microbiol. 43, 267–292. Black, L. W., Newcomb, W. W., Boring, J. W., and Brown, J. C. (1985) Ion etching of bacteriophage T4: Support for a spiral-fold model of packaged DNA, Proc. Natl. Acad. Sci., USA 82, 7960–7964.

PACKAGED BACTERIOPHAGE T7 DNA Booy, F. P., Trus, B. L., Newcomb, W. W., Brown, J. C., Serwer, P., and Steven, A. C. (1992) Organization of dsDNA in icosahedral virus capsids, Proc. Electron Micros. Soc. Am. 50, 452–453. Carazo, J. M., Fujisawa, H., Nakasu, S., and Carrascosa, J. L. (1986) Bacteriophage T3 gene 8 product oligomer structure, J. Ultrastruct. Mol. Struct. Res. 94, 105–113. Casjens, S. (1985) Nucleic acid packaging by viruses, in Casjens, S. (Ed.), Virus Structure and Assembly, pp. 75–147, Jones and Bartlett, Boston. Cerritelli, M. E., Cheng, N., Rosenberg, A. H., Simon, M. N., and Steven, A. C. (1996) Organization of dsDNA in bacteriophage T7 capsids: Support for the coaxial-spool model, Biophys. J. 70, A363. Crowther, R. A., and Amos, L. A. (1971) Harmonic analysis of electron microscope images with rotational symmetry, J. Mol. Biol. 60, 123–130. Dokland, T., and Murialdo, H. (1993) Structural transitions during maturation of bacteriophage lambda capsids, J. Mol. Biol. 233, 682–694. Dunn, J. J., and Studier, F. W. (1983) Complete nucleotide sequence of bacteriophage T7 DNA and the locations of T7 genetic elements, J. Mol. Biol. 166, 477–535. Earnshaw, W., and Casjens, S. (1980) DNA packaging by the double-stranded DNA bacteriophages, Cell 21, 319–331. Eiserling, F. A. (1983) Structure of the T7 virion, in Mathews, C. K., Kutter, E. M., Mosig, G., and Berget, P. B. (Eds.), Bacteriophage T4, pp. 11–24, Am. Soc. for Microbiol., Washington, DC. Griess, G. A., Moreno, E. T., and Serwer, P. (1992) Desktop digital imaging: Application to detection of length heterogeneity after hyperresonant pulsed-field gel electrophoresis of mature bacteriophage P22 DNA, in Burmeister, M., and Ulanovsky, L. (Eds.), Methods in Molecular Biology, Vol. 12: Pulsed-Field Gel Electrophoresis, pp. 173–181, Humana Press, Totowa, NJ. Hendrix, R. W. (1978) Symmetry mismatch and DNA packaging in large bacteriophages, Proc. Natl. Acad. Sci. USA 75, 4779– 4783. Johnston, R. F., Pickett, S. C., and Barker, D. L. (1990) Autoradiography using storage phosphor technology, Electrophoresis 11, 355–360. Khan, S. A., Hayes, S. J., Watson, R. H., and Serwer, P. (1995) Specific, nonproductive cleavage of packaged bacteriophage T7 DNA in vivo, Virology 210, 409–420. Khan, S. A., Watson, R. H., Hayes, S. J., and Serwer, P. (1997) PCR-directed formation of viral hybrids in vitro, Virology 227, 409–419. Kocsis, E., Cerritelli, M. E., Trus, B. L., Cheng, N., and Steven, A. C. (1995) Improved methods for determination of rotational symmetries in macromolecules, Ultramicroscopy 60, 219–228. Lepault, J., Dubochet, J., Baschong, W., and Kellenberger, E. (1987) Organization of double-stranded DNA in bacteriophages: A study by cryo-electron microscopy of vitrified samples, EMBO J. 6, 1507–1512. Markham, R., Frey, S., and Hills, G. J. (1963) Methods for the enhancement of image detail and accentuation of structure in electron microscopy, Virology 20, 88–102. Moody, M. F. (1965) The shape of the T-even bacteriophage head, Virology 26, 567–576.

43

Murialdo, H., and Becker, A. (1978) Head morphogenesis of complex double-stranded DNA bacteriophages, Microbiol. Rev. 42, 529–576. Prasad, B. V., Prevelige, P. E., Marietta, E., Chen, R. O., Thomas, D., King, J., and Chiu, W. (1993) Three-dimensional transformation of capsids associated with genome packaging in a bacterial virus, J. Mol. Biol. 231, 65–74. Richards, K. E., Williams, R. C., and Calendar, R. (1973) Mode of DNA packing within bacteriophage heads, J. Mol. Biol. 78, 255–259. Serwer, P. (1976) Internal proteins of bacteriophage T7, J. Mol. Biol. 107, 271–291. Serwer, P. (1977) Flattening and shrinkage of bacteriophage T7 after preparation for electron microscopy by negative staining, J. Ultrastruct. Res. 58, 235–243. Serwer, P. (1979) Fibrous projections from the core of a bacteriophage T7 procapsid, J. Supramol. Struct. 11, 321–326. Serwer, P. (1980) A metrizamide-impermeable capsid in the DNA packaging pathway of bacteriophage T7, J. Mol. Biol. 138, 65–91. Serwer, P. (1986) Arrangement of double-stranded DNA packaged in bacteriophage capsids: An alternative model, J. Mol. Biol. 190, 509–512. Serwer, P. (1989) Double-stranded DNA packaged in bacteriophages: Conformation, energetics and packaging pathway, in Adolph, K. (Ed.), Chromosomes: Eukaryotic, Prokaryotic and Viral, Vol. 3, pp. 203–223, CRC Press, Boca Raton. Steven, A. C., and Trus, B. L. (1986) The structure of bacteriophage T7, in Harris, J. R., and Horne, R. W. (Eds.), Electron Microscopy of Proteins, Vol. 5, pp. 1–35, Academic Press, London. Steven, A. C., Trus, B. L., Maizel, J. V., Unser, M., Parry, D. A. D., Wall, J. S., Hainfeld, J. F., and Studier, F. W. (1988) Molecular substructure of a viral receptor-recognition protein: The gp17 tail-fiber of bacteriophage T7, J. Mol. Biol. 200, 351–365. Stroud, R. M., Serwer, P., and Ross, M. J. (1981) Assembly of bacteriophage T7: Dimensions of the bacteriophage and its capsids, Biophys. J. 36, 743–757. Studier, F. W. (1969) The genetics and physiology of bacteriophage T7, Virology 39, 562–574. Studier, F. W. (1979) Relationships among different strains of T7 and among T7-related bacteriophages, Virology 95, 70–84. Studier, F. W., and Dunn, J. J. (1983) Organization and expression of bacteriophage T7 DNA, Cold Spring Harbor Symp. Quant. Biol. 47, 999–1007. Sun, M., and Serwer, P. (1997) The conformation of packaged bacteriophage G DNA, Biophys. J. 72, 958–963. Thuman-Commike, P. A., Greene, B., Jakana, J., Prasad, B. V. V., King, J., Prevelige, P. E., Jr., and Chiu, W. (1996) Threedimensional structure of scaffolding-containing phage P22 procapsids by electron cryo-microscopy, J. Mol. Biol. 260, 85–98. Valpuesta, J. M., and Carrascosa, J. L. (1994) Structure of viral connectors and their function in bacteriophage assembly and DNA packaging, Q. Rev. Biophys. 27, 107–155. Wurtz, M. (1992) Bacteriophage structure, Electron Microsc. Rev. 5, 283–309.