Structure of the Filamentous Bacteriophage fl LOCATION OF THE A, C, AND D MINOR COAT PROTEINS*

The location within the virion of the A, C, and D minor coat proteins of the filamentous bacteriophage fl has been analyzed. The A protein is present in -5 copies/particle and is located at the tip of normal length phage, miniphage, and fl/pBR322 chimeric phage, a longer than normal length phage. The mole ratios of the A, C, and D proteins are the same for each type of particle, consistent with a model of phage or- ganization in which the minor coat proteins are clus-tered near or at the ends of the phage. Normal length phage were fragmented by passing them through a French press, and those fragments that contained the A protein were separated from those that did not by treating the mixture with anti-A protein antibody. Analysis of the protein compositions of the two popu-lations of fragments showed that the A and D proteins were found together in one population of fragments and that most, it not all, of the C protein was found in the other. These results show that the D protein is located near or at the A protein end of the phage and that the C protein is located in a region near or at the opposite end. Treatment of the virion with proteases which lowered the infectivity of the phage resulted in particles in which only the A protein was cleaved to any detectable extent. These particles remained resist- ant to the action of micrococcal nuclease.

Studies in many laboratories on the structure of the fiiamentous bacteriophage fl, fd, and M13 have established that the bulk of the phage particle consists of a circular singlestranded DNA molecule coated with -2710 molecules of the major coat protein (product of phage gene VIII, B protein) (1). The virion also contains minor amounts of three other phage-encoded proteins which have been shown to be the products of gene 1 1 1 (A protein), gene IX (C protein), and gene VI (D protein) (2-5). The A, C, and D proteins are present in -5, 10, and 5 copies/phage particle, respectively (5). The A protein is necessary for adsorption of these virions to their host (6) and has been located at one end of the phage particle (2, 3). Neither the function nor the location within the virion of the C and D proteins has been established as yet.
The filamentous bacteriophage are unique among bacteriophage in that the size of the DNA which can be packaged into a particle is highly variable. Miniphage ranging in length from 0.2 to 0.5 times the normal particle length have been isolated (7)(8)(9). These particles contain no intact genes and require the presence of helper phage for infection (10). In addition, the ability to clone long stretches of foreign DNA into the fiiamentous phage genome has led to the production of phage particles which are longer than normal (maxiphage) (For example see . The exact length of such particles is directly proportional to the size of the DNA insert. Although both miniphage and maxiphage have a filamentous phage appearance by electron microscopy, little is known about the amounts or location of the minor A, C, and D proteins in these particles.
In this paper, we show that the A and D minor coat proteins are located near or at one end of the normal length phage particle and that the C protein is located near or a t the opposite end. In addition, we present evidence that the protein organization of miniphage and maxiphage is probably the same as that for normal length bacteriophage.
Growth a n d Purification of Bacteriophage-Escherichia coli strains K38 (15) and JMlOl (A[lac, pro], F'lac obtained from J. Messing through D. Bastia of Duke University) were used for the growth of wild type fl, fl miniphage, and R208, a fl/pBR322 chimeric phage (16) obtained from J. Boeke (The Rockefeller University).
Wild type (normal length) fl was grown on K38 in supplemented MTPA media, labeled with ["'S]cystine or ['Hlleucine, and purified as previously described (5). Phage containing radioactive DNA were prepared in the same manner except that 1 mCi of [32P]orthophosphate was added 15 min after infection in place of cystine. Miniphage were grown on K38, labeled with ["%]cystine or ['HJeucine, precipitated with polyethylene glycol, washed with 0.1% sarcosyl, and precipitated again with polyethylene glycol as described (5). The miniphage were then isolated by two successive sucrose gradient centrifugations and were further purified by cesium chloride gradient centrifugation (9).
To grow R208 (maxiphage), strain JMlOl was grown in supplemented MTPA medium to a density of 2 X IOH bacteria/ml and infected with R208 at a multiplicity of infection of 10. After 40 min, the bacteria were plated on tryptone plates containing 50 pg/ml of ampicillin. An ampicillin-resistant colony was then picked into supplemented MTPA media lacking cystine or leucine but containing 50 pg/ml of ampicillin, and grown to a density of 2 X 10" bacteria/ml. ["'SjCystine or ["Hlleucine was added, and the phage were grown, harvested, and purified as described (5).
The homogeneity of the radioactive phage preparations used in 539 Proteins this study was determined by agarose gel electrophoresis of whole phage as described by Nelson et al. (17) and by electron microscopy. Normal length phage preparations contained 90 to 95% normal length phage with small amounts of contaminating diploid, triploid, and miniphage particles. Miniphage preparations contained less than 10% normal length phage. Maxiphage preparations were greater than 90% homogeneous, the remaining species consisting of what appeared to be higher molecular weight multimers of single length particles. Purification of Antibodies and Conjugation with Ferritin-Nonimmune serum was obtained from a white male New Zealand rabbit. Rabbit antiserum against purified fd A protein was from the same batch of serum used in previous studies (2). Purified IgG was obtained by passing sera down a protein A-Sepharose column (18) prepared using the procedure of March et al. (19). Affinity purified goat antirabbit IgG was prepared (20) and conjugated with ferritin using glutaraldehyde (21).
Fragmentation of Bacteriophage-Wild type (normal length) fl phage labeled with either [32P]orthophosphate or ["5S]cysteine were diluted to a concentration of 5 X phage particles/ml in 0.10 M NaCl, 0.001 M EDTA, 0.05 M Tris-HC1 (pH 7.8) containing 1 mg/ml of bovine serum albumin. Between 0.6 and 2 ml of this phage solution were passed through an Aminco French pressure cell (3-ml capacity) at a pressure of 8000 p.s.i. at 4°C. The cell was rinsed thoroughly with buffer to remove phage which did not pass through the press the first time. The fragmented phage were then passed through the press a second time to assure that all phage had been fragmented. Recovery of radioactive phage was greater than 90%. whereas phage titer was reduced to less than 0.1% of the initial titer.
Separation of Fragments Originating from Different Ends of the Bacteriophage-Approximately 700 pg of anti-A protein IgG and 0.7 ml of freshly fragmented phage were incubated together for 1 h on ice, followed by I-h incubations at room temperature and 37°C. The mixture was cooled and immediately layered onto 10 ml of a 5 to 20% (w/w) linear sucrose gradient formed over a 1.4-ml shelf of 408 (w/ w) metrizamide and 30% (w/w) sucrose (all solutions prepared in 0.10 M NaCl, 0.001 M EDTA, 0.05 M Tris-HC1 (pH 7.8) containing 1 mg/ml of bovine serum albumin). The sample was centrifuged at 150,000 X g for 3.5 h in a Beckman SW41 rotor at 4'C. The gradient was collected from the bottom and the radioactivity in 5 p1 of each 0.2-ml fraction was determined by scintillation counting. The material that sedimented onto the metrizamide shelf (Peak X) or the material that remained in the middle of the gradient (Peak Y) was diluted 6-fold with 0.10 M NaCl, 0.001 M EDTA, 0.01 M Tris-HC1 (pH 7.6) containing 1 mg/ml of bovine serum albumin, and 0.1 mg of unlabeled phage was added as carrier. Peak X material was pelleted by centrifugation at 161,000 X g for 90 min, and Peak Y material was pelleted at 272,000 x g for 6 h in a Beckman 65 rotor at 4°C. Pellets were resuspended at room temperature in 0.4 ml of 0.10 M NaC1, 0.001 M EDTA, 0.01 M Tris-HC1 (pH 7.6) containing 2% SDS' and 2.5% mercaptoethanol.
In experiments where it was not necessary to obtain large amounts of material or in control experiments, 0.06 ml of intact or fragmented phage was incubated with nonimmune IgG or anti-A protein IgG as described for the large scale preparations. The reacted material was amide shelf and was centrifuged at 164,000 X g for 3 h in a Beckman layered onto a 3.6-ml sucrose gradient formed over a 0.55-ml metriz-SW60 rotor at 4°C. Gradients were collected from the bottom, and the radioactivity was determined by scintillation counting.
Digestion of Phage with Proteases-One hundred microliters of ["5S]cysteine-labeled normal length phage at a concentration of 10 mg/ml in 0.1 M NaCl, 0.01 Tris-HC1 (pH 7.1) were incubated with 100 pg of trypsin, chymotrypsin, thermolysin, or subtilisin. After incubation at 37°C for 10 h, phage were precipitated from the reaction mixtures by the addition of 10 pl of 5 M NaCl and 20 pl of 25% (w/w) polyethylene glycol. Proteases and labeled peptides released by digestion were not precipitated under these conditions. Phage pellets were collected by centrifugation and were resuspended in 100 p1 of 0.1 M NaCl, 0.01 M Tris-HC1 (pH 7.1) and stored at -10°C.
Quantitation of Proteins A, C, and D-To determine the amount of A protein/phage particle, [3H]leucine-labeled phage in 0.06 ml of 0.125 M Tris-HCI (pH 6.8) were incubated for 5 min at 42°C with 0.005 ml of chloroform. The solution was made 2% in SDS and 10% in glycerol, heated at 100OC for 5 min, and subjected to SDS-polyacrylamide gel electrophoresis as described by Lin et al. (5). Gels were sliced and radioactivity present in the A and B proteins was determined (5). The amount of A protein/phage was calculated using values of 19 residues of leucine/molecule of A protein, 2 residues of ' The abbreviation used is: SDS, sodium dodecyl sulfate leucine/molecule of B protein, and 2710 copies of B protein/phage particle (22). The number of leucine residues/A or B protein molecule was derived from the fl DNA sequence. ' To determine the mole ratios of the A, C, and D proteins, whole or fragmented phage particles labeled with ["Slcysteine were solubilized, subjected to gel filtration on A-5m agarose, and electrophoresed on SDS-polyacrylamide gels as described by Lin et al. (5). The procedures used for slicing gels, for extracting radioactivity from gel slices, and for the determination of the mole ratios of the minor coat proteins have been given previously (5).
Protease-treated phage in 0.7% SDS, 0.1 M NaC1, 0.01 M Tris-HC1 (pH 7.1) were heated at 100°C for 2 min and subjected to SDSpolyacrylamide gel electrophoresis according to Laemmli (23) using a separating gel of 16.25% acrylamide and 0.43% bisacrylamide. After electrophoresis, gels were either stained with Coomassie blue or cut into 5-mm slices. Gel slices were placed in scintillation vials and incubated with 0.4 ml of Hz0 and 1.0 ml of Protosol at 37°C for 1 day. Another 1.0 ml of Protosol was added, and incubation was continued for 2 more days. Finally, 10 ml of scintillation counting fluid were added to each vial, and after incubation for 1 more day, the amount of radioactivity in each vial was determined.
Electron Microscopy-Samples for electron microscopy were taken directly from gradients or were diluted with 0.1 M ammonium acetate. A glow-discharged, carbon-coated, 400-mesh copper grid was floated on a drop of sample for 30 s, washed 3 to 4 times by floating on H20 or 0.1 M ammonium acetate, and stained for 15 to 40 s with 2% uranyl acetate in H20. Samples were viewed at 80 kV in a JEOL l00C electron microscope and photographed at a magnification of 20,000 to 33,000. Actual magnifications were determined by reference to a grating replica. Particle lengths for normal length phage, miniphage, and maxiphage were determined by tracing micrographs that were enlarged 3 times.

Proteins A, C, and D are Present in Specific Regions of
Normal Length Phage, Miniphage, and Maxiphage-We have shown previously that the average length phage particle contains 5 molecules of the A protein (2, 3, 5 ) . This was done by determining the percentage of various radioactive amino acids in the A and B proteins and then calculating the number of A protein molecules/particle, knowing the sequence of the proteins and assuming that each phage contains 2710 molecules of B protein (22). Goldsmith and Konigsberg (2) and Woolford et al.
(3) showed that only one end of a normal length phage particle was labeled when phage were reacted with anti-A protein IgG and ferritin-conjugated goat antirabbit IgG. We have now done similar experiments using a miniphage and a maxiphage (R208) in addition to normal length phage. All three types of phage contain -5 molecules of A protein/virion (Table I). Only one end of each type of phage particle appeared t o react with anti-A protein IgG ( Fig.  1). When nonimmune IgG was substituted for anti-A protein IgG, no interaction between phage and antibody could be demonstrated (data not shown).
It was not possible to locate the positions of the C and D proteins using this approach due to the present lack of antibodies directed against these proteins. Since in all three phage types the same amount of A protein is located only at one end of the virion, comparison of the amounts of C and D proteins relative to t h e A protein should give some information as to the relative location of these proteins in the phage particle. If the C and/or D proteins are located in specific regions, such as the ends of the phage particles, then their amounts relative to the A protein should remain the same in each type of phage. If the C and/or D proteins are distributed along the lengths of the particles, then their amounts relative to the A protein would be expected to vary with the length of the phage particle. The predicted mole ratios for both possibilities are given in Table I   and their lengths were determined by electron microscopy ( Table I). The amounts of the A, C, and D proteins were determined by the combined agarose gel filtration and SDSpolyacrylamide gel electrophoresis procedure described by Lin et al. (5). The phage were solubilized in SDS, and the DNA and proteins were partially separated on a Bio-Gel A-5m column. All three types of phage particles gave an elution profie identical with that shown in Fig. 5B. The amount of A, C, and D proteins in each fraction of the A-5m column was determined by polyacrylamide gel electrophoresis and the mole ratios were calculated. The observed mole ratios shown in Table I for normal length phage, miniphage, and maxiphage are virtually indistinguishable from one another and are in excellent agreement with those ratios predicted by a model in which all of the minor phage coat proteins are located in specific regions, possibly near or at the ends of the phage.
Preparation and Separation of the Ends of the Phage-It was necessary to isolate each end of the phage to determine the location of the C and D proteins. Anti-A protein IgG was used to separate the two ends of fragmented phage using the procedure outlined in Fig. 2. Normal length phage were fragmented by passing them through a French press. The random fragments generated by such treatment were incubated with anti-A protein IgG as described under "Experimental Procedures." Phage fragments that originated from the end of the phage containing the A protein formed rosette-like aggregates due to the cross-linking effect of anti-A protein IgG under these conditions. These rosette-like structures were then separated from fragments derived from the opposite end and from regions of the particle initially located between the ends of the phage on a sucrose velocity gradient layered above a dense shelf of metrizamide. The aggregates composed of fragments derived from the A protein end of the phage collected on the metrizamide shelf (Peak X , Fig. 2) while the remaining fragments migrated more slowly in the sucrose gradient (Peak Y , Fig. 2).
T o test the efficacy of the procedure, normal length phage containing "P-labeled DNA were fragmented, incubated with nonimmune or anti-A protein IgG, and subjected to sucrose velocity centrifugation as outlined above. As a control, unfragmented phage were subjected to the same procedure. Inspection of Fig. 3A clearly shows that passage of the phage through the French press produced fragments smaller than the intact phage. (The small amount of material on the metrizamide shelf may result from side-by-side clustering of phage which was observed by electron microscopy.) Treatment of intact phage with anti-A protein IgG resulted in the aggregation of all particles, which upon centrifugation migrated to the metrizamide shelf (Fractions 4 to 7, Fig. 3B). This showed that the amount of antibody used would interact with enough A protein to aggregate all of the phage present in the solution. However, treatment of fragmented phage under the same conditions (Fig. 3B) resulted in only one-third of the total nzP radioactivity sedimenting to the metrizamide shelf (Fractions 4 to 7, Peak X material), while the remaining twothirds of the radioactivity (fractions 9 to 22, Peak Y material) migrated to the same position in the gradients as the fragments treated with nonimmune IgG (Fig. 3A). These results suggested that the fragments were approximately one-third of the original phage length. This was confirmed by an exami- nation of electron micrographs of the rosettes in Peak X material and the fragments in Peak Y material where the length of the fragments ranged in size from 1/5 to 1/2 of the original phage length (Fig. 4).
The same experiment was repeated using phage labeled with [35S]cysteine (Fig. 3, C and D). In this case, most of the radioactivity should be at the A protein end of the phage since the A protein contains 8 cysteines and the C and D proteins only 1 cysteine each.' (The major coat, or B protein, does not contain cysteine.) Fig. 3 0 shows that -80% of the radioactivity sedimented to the metrizamide shelf when fragmented phage containing this label were treated with anti-A protein IgG. These data show that the material in Peak X contains those fragments derived from the A protein end of the phage and that the material in Peak Y is made up of the middle and opposite end of the phage particle.
The Protein Compositions of the Separated Fragments-Large amounts of Peak X and Peak Y material (Fig. 2

) from
[3sS]cysteine-labeled phage were prepared using the procedure described above. Phage proteins were solubilized in SDS, and the DNA and labeled proteins were partially separated by gel filtration on a Bio-Gel A-5m column (5). The elution profde of solubilized Peak X material (Fig. 2) showed three ['''SIcysteine-containing peaks (Fig. 5A). The indicated fractions were pooled, lyophilized, and analyzed for the presence of the A, C , and D proteins by SDS-polyacrylamide gel electrophoresis (Fig. 6). The material in the first two peaks that eluted from the column (Fig. 5A) contained primarily A protein as shown in Fig. 6, Lanes 3 and 4 and Lanes 5 and 6, respectively:' The third peak consisted mainly of D protein and very small amounts of C protein as shown in Fig. 6, Lanes 7 and 8. Peak Y material (Fig. 2) showed only one major peak when sub-The first peak that eluted from the column (Fig. 5A) was presumably A protein complexed with anti-A protein IgG since 1) it disappeared upon heating the sample to 100°C before application to the column and 2) it was present only after fragmented phage had been treated with anti-A protein IgG. The bands seen in Lanes 3 through 6 of Fig. 6 that migrate slightly ahead of the A protein are thought to be either cleavage products of the A protein or the A protein itself, but in a different conformation ( 5 ) . The amount of radioactivity found in these bands varies from one preparation of phage to another and accounts for 1 to 9% of the total [""Slcysteine radioactivity found in the phage. or Peak Y (0--0) material (Fig. 2) was isolated, solubilized in SDS, and chromatographed on Bio-Gel A-5m, 8s described under "Experimental Procedures." Material originating from an equal number of phage particles was applied to the column in each case. Letters beneath the figure designate fractions that were pooled, lyophilized, and electrophoresed on the gel shown in Fig. 6. B, unfragmented [35S]cysteine-labeled phage from the same batch of phage were solubilized and chromatographed on the same A-5m column. Letters beneath the figure designate fractions that were pooled, lyophilized, electrophoresed, and quantitated to give the data shown in the bottom half of FIG. 6. SDS-polyacrylamide gel electrophoresis of Peak X and Peak Y A-5m fractions. Peak X and Peak Y material (Fig. 2) were subjected to gel filtration on a Bio-Gel A-5m column, and the fractions indicated in Fig. 5A were pooled and concentrated as described under "Experimental Procedures." A portion of each fraction was electrophoresed on SDS-polyacrylamide gels, and the gels were prepared for fluorography as described (5). Samples in Lanes I and 2 were taken from Peak Y material, Fractions E and F, respectively (Fig. 5A). Samples in Lanes 3 to 8 were taken from Peak X material, Fractions A to F, respectively (Fig. 5A) jected to gel filtration (Fig. 5A). When this material was analyzed by SDS-polyacrylamide gel electrophoresis, only C protein was present (Fig. 6, Lanes 1 and 2).
T o accurately determine the amount of the A, C , and D proteins in the separated phage fragments, the fractions indicated in Fig. 5A (A to F) were run on SDS-polyacrylamide gels, the radioactivity in each gel slice determined, and the amount of the A, C , and D proteins in each sample calculated as described under "Experimental Procedures." The results of this analysis are presented in Table 11. Peak X material (Fig.  2) contained 94% of the total A protein, 98% of the total D protein, and 14% of the total C protein. Peak Y material

TABLE I1
Mole ratios of the A, C, and D proteins in Peak X and Peak Y material (Fig. 2)

Mole ratio
A protein c protein D protein The mole ratios are based on the cysteine content of the A, C, and D proteins given in Table I. The counts per minute of each protein refer to the sum of the radioactivity in each protein from all fractions of the gel filtration columns in Fig. 5.
Approximately 95% of the ['%]cysteine counts initially incubated with anti-A protein IgG were recovered from the gels.
<'The fragmented and unfragmented phage originated from the same batch of [:'"S]cysteine-labeled phage as that used for the preparation of the Peak X and Peak Y samples. contained 86% of the total C protein and only minimal amounts of the A and D proteins.
In order to be certain that passage of phage through the French press did not change the relative amounts of the A, C , and D proteins, we did control experiments using ["S]cysteine-labeled phage from the same preparation as above. Either fragmented or unfragmented phage that were not treated with antibody were solubilized and subjected to gel fdtration on a Bio-Gel A-5m column (Fig. 5B). Analysis of appropriate fractions (Fig. 5B, A to D ) on SDS-polyacrylamide gels showed that the first major peak eluting from the column contained all the A protein, while most of the second peak was comprised of 3sS-labeled C and D proteins. Quantitation of the amount of A, C , and D proteins as described above showed that the mole ratios of these proteins in fragmented or unfragmented phage that were not treated with anti-A protein IgG were identical with each other and to the ratio of these proteins in the combined Peak X and Peak Y samples (Table 11). Analysis of the amount of the A, C, and D proteins in Peak X and Peak Y material (Fig. 2) prepared from four separate [35S]cysteine-labeled phage preparations gave values that were in good agreement with those presented here. It therefore appears that the D protein is located near or at the A protein end of the phage and that the C protein is located in a region near or at the opposite end: Susceptibility of the A, C, and D Proteins to Protease Digestion and Sulfhydryl Reducing Agents-In an attempt to ascertain the role of the C and D proteins in the infective process, [3'S]cysteine-labeled phage were treated with various proteases as described under "Experimental Procedures." After freeing the resulting particles from the protease and phage protein fragments derived from the protease digestion, a small aliquot of the resulting particles was titered to determine the percentage of surviving phage. Trypsin showed no effect, while chymotrypsin, thermolysin, and subtilisin lowered the titer from 6.6 X lOI3 to 2.7 X lo", 3.0 X lo'', and 3.3 'We believe that the C protein found in Peak X material results from contamination with Peak Y material. However, we cannot unambiguously rule out the possibility that 1 or 2 molecules of C protein reside at the same end as the A and D proteins. X lo9 phage/ml, respectively. To determine which viral components were affected by these protease treatments, we subjected the various protease-digested [3'S]cysteine-labeled phage to SDS-polyacrylamide gel electrophoresis. The only minor phage protein susceptible to cleavage, as detected by SDS-polyacrylamide gel electrophoresis, was the A protein (Fig. 7). There appears to be a good correlation between the amount of intact A protein left after digestion and the percentage of the remaining infectivity. The migration position of the B protein on SDS-polyacrylamide gels, as detected by Coomassie blue stain, appeared unchanged after treatment of phage with any of the proteases tested, and amino acid analysis of B protein from subtilisin-digested phage showed that the B protein was not susceptible to cleavage (data not shown). These results suggest that limited cleavage of the A protein is sufficient to cause the observed loss in activity. This loss in activity is probably not due to exposure of the DNA to solvent since the DNA in phage treated with thermolysin or subtilisin is resistant to the action of micrococcal nuclease.
Since the A and D proteins are located at the same end of the phage and both contain cysteine, it is possible that they might be associated via disulfide bonds. To test this, ["'SIcysteine-labeled phage were solubilized with 0.1% SDS overnight at 37OC and subjected to A-5m gel filtration in the absence of any sulfhydryl reducing agents. A pattern indistinguishable from that in Fig. 5B was obtained. When Fractions A to D (Fig. 5B) were analyzed by SDS-polyacrylamide gel electrophoresis in the absence of sulfhydryl reducing agents, the A, C, and D proteins all migrated at the expected positions (Fig. 8, part and were present in the expected mole ratios. This implies that any interaction between the A and D proteins or among different molecules of the same minor coat protein species is not mediated by disulfide bonds in the phage particle. However, if a fraction which contained both B and D proteins was heated in the presence of a high concentration of mercaptoethanol (1. 1 M), the D protein did not enter the separating gel (Fig. 8, compare Lane 7 of parts I and 14. The D protein remained at the top of the gel only in fractions that contained large amounts of the B protein (Fig. 8, part 11, compare Lanes 2 and 3 and Lanes 6 and 7). At B protein concentrations less than -1 mg/ml, this phenomenon was not observed. Formation of these high molecular weight complexes was promoted by the presence of 1.1 M propanol but not by 0.1 M mercaptoethanol, propanol, or dithiothreitol. These observations suggest that, under the conditions of sample preparation described, there may be some sort of interaction between the B and D proteins. This interaction is possibly hydrophobic in nature, a suggestion consistent with the known and predicted properties of these proteins (24). DISCUSSION A complete understanding about the morphogenesis of the fiamentous bacteriophage requires as much information as possible about the structure of the phage particle. Only recently has it been shown that each particle contains -5, 10, and 5 molecules of the minor A, C, and D proteins, respectively, in addition to the 2710 molecules of the major coat or B protein (4, 5). Based on the data in this paper, we believe that this is an accurate representation of the number of molecules of these proteins/particle. I n filamentous phage of all lengths, the mole ratio of the C to D proteins is always 2:l. The number of moles of the A protein is more variable, usually slightly lower than one when normalized to 1 mol of D protein in different phage preparations. This is consistent with the notion that some A protein may be lost during purification of the phage. Since comparisons of A to B protein consistently yield 4 to 5 molecules of A protein/particle (Refs. 2, 3, and this paper), then 10 C and 5 D proteins is the best estimate for the number of these proteins/average particle.
The experiments in this paper strongly suggest that the A and D proteins reside near or a t one end of the phage particle and that the C protein is in a region near or at the opposite end. This conclusion is based on the following two observations. First, the location and number of A protein molecules and the mole ratio of the A, C, and D proteins remain constant, independent of the length of the phage particle. Second, when the phage are fragmented into smaller pieces, the A and D proteins are always found on the same fragment. Most, if not all, of the C protein is present on fragments devoid of the A and D proteins. The location of the A and D proteins at the same end of the virion is consistent with the earlier observations that nonpermissive bacteria infected with gene 111 (A protein-deficient) or gene VI (D protein-deficient) mutant phage produce primarily noninfectious polyphage containing multiple copies of unit length DNA (25). Presumably, these two proteins have some role in determining the formation of one end of the phage particle during morphogenesis.
Treatment of the phage with subtilisin renders the phage noninfective but does not alter the buoyant density of the virions in CsCl (26). Gray et aL5 have shown that this treatment removes knob-like structures that are present at one end of the virion and have suggested that they are part of the A protein. We have found that treatment of the intact phage particle with chymotrypsin, thermolysin, or subtilisin results in a noninfectious particle in which only the A protein has ' C. W. Gray, R. S. Brown, and D. A. Marvin, personal communication.
been cleaved to any detectable extent. This supports the identification of the knobs as being part of the A protein. Gray et aL5 have proposed a two domain theory for the structure of the adsorption complex in which one domain (knob) is responsible for mediating the interaction of the phage with the pilus during the process of infection and the other domain (stem) may be essential for anchoring the A protein to the phage particle itself. It is possible that the stem might be composed of part of the A protein since the knobless phage produced by treatment with subtilisin still retain a 15,000dalton portion of the A protein. The fact that the A and D proteins appear to be at the same end of the phage might indicate that there is some interaction between these two proteins. However, any such association cannot be due to disulfide bridges since the individual proteins can be readily dissociated from the phage and from each other in the absence of sulfhydryl reducing agents.
If any of the minor proteins recognize a specific region of the DNA, then the DNA would have to be present in a unique orientation in the phage pzrticle with respect to these proteins since they are located near or at the ends of the phage. To date, a number of groups have addressed the question of DNA orientation within the phage and have reached somewhat conflicting conclusions (27)(28)(29)(30). The approach presented here should allow us to determine whether a specific region of the DNA is present at the A and D protein-containing end of the phage.
Acknowledgments-We thank D. F. Hill, G. B. Peterson, and C. W. Gray for communicating results prior to publication. We also thank Ms. Lucy Hamilton and Ms. Gerda Michalsky for expert technical assistance.
Addendum-We have recently learned that J. Schoenmakers and G. Simons have results which suggest that the low molecular weight band (C protein band) obtained after subjecting disrupted MI3 phage to SDS-polyacrylamide gel electrophoresis may contain some gene VI1 product as well as the protein coded by phage gene IX. Both of these proteins have nearly identical molecular weights, and each contains 1 residue of cysteine per molecule. Therefore, the value of 10 copies of C protein per phage particle presented in this paper would represent the sum of the number of copies of both the gene VI1 and gene IX products in the phage.