Polymorphism of bacteriophage T71

J. Mol. Biol. (1997) 273, 658±667

Polymorphism of Bacteriophage T7 Irene S. Gabashvili, Saeed A. Khan, Shirley J. Hayes and Philip Serwer Department of Biochemistry The University of Texas Health Science Center, San Antonio TX 78284-7760, USA

For viruses made of nucleic acid and protein, the structure of the protein outer shell has, in the past, been found to be uniquely determined by the viral genome. However, here, non-denaturing agarose gel electrophoresis of bacteriophage T7 reveals two states of the mature T7 capsid; the conditions of growth are found to alter the population by T7 of these two electrophoretically de®ned states. Both states have been previously observed for a genetically altered T7 and they are observed here for wild-type T7. The average electrical surface charge density of a bacteriophage particle (d) determines its state; the d of particles in both states is negative. For a given condition of growth, the population of these two states is in¯uenced by the extent to which the major T7 outer shell protein, p10A, is accompanied by its minor readthrough variant, p10B. Comparison of the two electrophoretic states reveals the following. (1) No difference in radius is present in the outer shell (2%). (2) As the pH of electrophoresis is either increased or decreased from neutrality, the state becomes more highly populated for which d is greater in magnitude (state 1). By changing the pH, some T7 particles are made to change state. (3) Particles in state 1 adsorb less quickly to host cells than do the particles in the alternative state (state 2). This latter observation suggests the hypothesis that state 1 evolved to reduce the probability of re-initiating an infection when conditions are not favorable for growth. This hypothesis is supported by the observation that, as conditions of growth become apparently more unfavorable, progeny increasingly populate state 1. # 1997 Academic Press Limited

*Corresponding author

Keywords: bacteriophage evolution; bacteriophage structure; agarose gel electrophoresis; two-dimensional electrophoresis

Introduction The characteristics of a virus are usually assumed to be uniquely determined by the viral genome. Although host proteins are known to participate in the assembly of at least some viruses (and, therefore, can alter the rate of assembly), the condition of the host is not known to have a signi®cant effect on the physical characteristics of viral progeny (for reviews for bacteriophages see: Casjens, 1985; Steven & Trus, 1986; Kellenberger, 1990; Wurtz, 1992). Some viruses have a unique enough structure so that the structure of the viral capsid (though not the packaged nucleic acid) can be determined by X-ray crystallography (for Abbreviations used: d, average electrical surface charge density; m, electrophoretic mobility; RE, effective radius; moÂ, m extrapolated to a gel concentration of 0; mo, mo corrected for electro-osmosis; Z, number of surface electrical charges per particle; S2*, fraction of particles in state 2. 0022±2836/97/430658±10 $25.00/0/mb971353

reviews see: Rossmann & Johnson, 1989; Liljas, 1991, 1996; Johnson et al., 1994). Results for the single-stranded DNA bacteriophage, fX174, are given by McKenna et al. (1992, 1996) and Olson et al. (1992). However, post-assembly survival of viral progeny could, in theory, be enhanced by structural adaptations that are caused by the condition of the viral host. For example, if conditions were unfavorable for growth, survival of viral progeny would be promoted by a change in structure that (1) reduced the frequency with which the virus adsorbed to host cells, and (2) reversed if the viral progeny found themselves in a more favorable environment. In the case of bacteriophage T7, gel electrophoresis reveals a negative average electrical surface charge density (d) that varies among T7 and its relatives, T3 and fII. The tail ®ber, however, has a positive d. A positively charged extensible tail ®ber might overcome electrical charge-charge repulsion of the negatively charged bacteriophage T7 and its host during the adsorption (Serwer et al., 1983). # 1997 Academic Press Limited

Polymorphism of T7

659

During the gel electrophoretic studies by Serwer et al. (1983), fII, T3 and T7 each form a single band. However, in a subsequent study, variability of d causes the formation of two bands during the non-denaturing gel electrophoresis of a bacteriophage whose genome is a T7 genome, except for replacement of a segment of the tail ®ber gene by the partially homologous segment from T3 (Khan et al., 1997). Although the cause of the conversion from one to two gel electrophoretic bands appeared to be a change in genotype, results shown here indicate that the cause of this conversion is change in the conditions of growth. Because the variability of d appears to have some features expected of a survival-promoting characteristic, here, the following questions have been asked: Why does the number of gel electrophoretic bands vary among different preparations of the same bacteriophage? Can changing of external conditions cause a change in electrophoretic state? What is the correlation between d and adsorbability of the bacteriophage particle to a host cell? What is the dependence of d on the conditions of growth: temperature, medium and concentration of host cells? What is the dependence of d on the relative amount of two variants of the protein of the T7 outer shell (p10A and p10B; reviewed by Steven & Trus, 1986)?

Results Electrophoretic heterogeneity of T7wt The T7-T3 hybrid bacteriophage has previously been found to form two bands during non-denaturing electrophoresis in gels formed by underivatized agarose (Khan et al., 1997). As both previously found (see Khan et al., 1997) and con®rmed with the bacteriophage preparations used here (not shown), T7wt undergoes gel adherenceinduced distortion of its migration in gels formed by underivatized agarose, but not IsoGel. In preliminary studies, the (surprising) observation was made that, when grown in the 2xLB medium previously used for growing the T7-T3 hybrid bacteriophage, T7wt also formed two bands during electrophoresis in gels formed by IsoGel, cast in T/M electrophoresis buffer (Figure 1a; lane labeled T7wt). As expected, the hybrid also formed two bands that were, however, in positions slightly different from the positions of the bands of T7wt (Figure 1a; lane labeled T7-T3). However, 2xLB medium-grown bacteriophage T3 formed only one band (Figure 1a; lane labeled T3). The state of those particles that form the origin-distal T7 band will be called state 1; the state of those particles that form the origin-proximal T7 band will be called state 2. For the T7-T3 hybrid bacteriophage, the particles in state 2 (origin-proximal) were previously shown to have a d smaller in magnitude than the d of the particles in state 1; no difference was detected in RE (Khan et al., 1997). To determine

Figure 1. Gel electrophoresis of bacteriophages. The 2xLB-grown particles indicated above a lane were subjected to non-denaturing gel electrophoresis at 2 V/cm, for the indicated time, by use of both the indicated electrophoresis buffer and 0.6% IsoGel: (a) 24 hours, T/M, (c) 20 hours, P/M, (d) 20 hours, T/M. The gel was stained with ethidium. The arrowheads indicate the origins of electrophoresis; the arrow indicates the direction of electrophoresis for a, c and d. In b, a m versus gel concentration plot is presented for the particles indicated in the Figure, subjected to electrophoresis at 2 V/cm, in 0.1 to 2.0% IsoGel cast in T/M electrophoresis buffer.

whether difference in d was also the reason for the formation of two bands in the case of T7wt, m was determined as a function of IsoGel concentration. A difference was observed in the intercept on the m axis (moÂ), but no signi®cant difference was observed in the slope of a semilogarithmic m versus gel concentration plot, when a plot for the origin-proximal band was compared to a plot for the origin-distal band (both are indicated by T7wt in Figure 1b). Thus, for T7wt, particles in state 1 have d greater in magnitude than the d of particles in state 2; no difference in RE was detected. This latter conclusion was con®rmed (2%) by two-dimensional electrophoretic analysis by Procedure 2 of Materials and Methods (not shown).

660

Polymorphism of T7

Test for change of state In preliminary experiments, the observation was made that the relative number of particles in the two T7wt states depended on the buffer used for electrophoresis. For example, based on quantitative ¯uorometry, the fraction of T7wt particles in state 2 (this fraction will be called S2*) was 0.45, when T/M electrophoresis buffer was used in Figure 1a. However, the S2* was 0.70, when P/M electrophoresis buffer was used for the same sample of T7wt (Figure 1c). This effect of buffer is qualitatively observed by inspection of Figure 1a and c. As found for T/M electrophoresis buffer, T3 formed one band in P/M electrophoresis buffer (Figure 1c). The S2* values of T7wt, however, varied with the T7wt bacteriophage preparation used; this variability is investigated in a subsequent section. To determine whether the buffer-dependent change in S2* was caused by a state 1 ! state 2 conversion when the electrophoresis buffer was changed from T/M to P/M, the following buffers were used during two-dimensional analysis by Procedure 1. The ®rst-dimensional electrophoresis was performed by use of T/M electrophoresis buffer; the second-dimensional electrophoresis was performed by use of P/M electrophoresis buffer. For T7wt, the pattern obtained after this two-dimensional electrophoresis consisted of two bands that differed in length; the lower band was 1.7 longer than the upper band (bands labeled T7wt in Figure 2a). A similar pattern was obtained for the T7-T3 hybrid (bands labeled T7-T3 in Figure 2a). As expected, a single band was observed for bacteriophage T3 (band labeled T3 in Figure 2a). The T3 band had the length of the shorter, origin-proximal bands of both T7wt and the T7-T3 hybrid. Inspection of the bands for both T7wt and the T7T3 hybrid revealed that the left end of the upper band was at the same horizontal coordinate as the left end of the lower band: However, the right end of the lower band was further to the right than the right end of the upper band. The equality of the position of the left ends indicates that some particles origin-distal in the ®rst (T/M) dimension became origin-proximal in the second (P/M) dimension. The non-equality in position of right ends indicates that other particles origin-distal in the ®rst dimension remained origin-distal in the second dimension. That is, the suspected state 1 ! state 2 change did occur for some, but not all, particles. When either P/M electrophoresis buffer (Figure 2b) or T/M electrophoresis buffer (not shown) was used for both dimensions, this conversion was reduced, and apparently eliminated in the case of T7wt. To determine whether origin-proximal (state 2) particles can be converted to origin-distal (state 1) particles by exchanging P/M electrophoresis buffer for T/M electrophoresis buffer, the two-dimensional electrophoresis of Figure 2a was repeated by use of one change: P/M electrophoresis buffer was used for the ®rst dimension; T/M electrophoresis

Figure 2. Test for buffer-induced change of state. During non-denaturing two-dimensional gel electrophoresis by Procedure 1 of Materials and Methods, the following buffers were used for the ®rst and second dimensions, respectively: (a) T/M, P/M; (b) P/M, P/M; (c) P/M, T/M. The gel was stained with ethidium. The origins of electrophoresis are identi®ed by both the arrowheads and the upper labels; the directions of the ®rst (I) and second (II) dimensional electrophoresis are identi®ed by the numbered arrows. The bands are identi®ed by the lower labels.

buffer was used for the second dimension, rather than vice versa. Analysis of the bands reveals that (1) some particles origin-proximal in the ®rst dimension became origin-distal in the second dimension, and (2) other particles origin-proximal in the ®rst dimension remained origin-proximal in the second dimension (Figure 2c). Thus, the state 2 ! state 1 change also occurred for some, but not all, particles. The effect of buffer on the population of states 1 and 2 To more extensively quantify the effect of buffer on S2*, this ratio was determined for T7wt, as a

Polymorphism of T7

function of pH at constant ionic strength, 0.14. The result was a roughly bell-shaped plot that had a maximum at approximately neutral pH. Although the shape of this plot was reproducible among different T7wt preparations, the height of the peak varied, even among preparations made by identical procedures. Plots for two T7wt preparations are shown in Figure 3 (indicated in the Figure by T7wt). In a subsequent section, variability of S2* is further explored. In comparison to the effects of pH, the effect of ionic strength at pH 7.4 was small. For three preparations of T7wt at pH 7.4, the average S2* did not signi®cantly change when the ionic strength was changed from 0.14 to either 0.07 or 0.28. For these experiments, ionic strength was changed by varying the phosphate concentration in P/M electrophoresis buffer. Adsorption to host cells Because E. coli cells are negatively charged at neutral pH (Bayer & Sloyer, 1990), electrostatic effects should bias bacteriophage-host adsorption in the direction of more rapid adsorption for particles in state 2, in comparison to particles in state 1. To quantitatively compare the amount of adsorption, T7wt bacteriophage particles (both states populated) were added to host bacteria; after incubation of this mixture for ®ve minutes, adsorbed bacteriophage particles were separated from unadsorbed bacteriophage particles by cen-

Figure 3. The effect of pH on S2*. At an ionic strength of 0.14, values of S2* were determined as a function of pH, by use of the buffers described in Materials and Methods. Electrophoresis was performed at 2 V/cm for 20 hours through 0.6% IsoGel. The values obtained are plotted as a function of pH for two preparations of 37 C, 2xLB medium-grown T7wt (indicated in the Figure by T7wt). The results are also shown for T7A.

661

Figure 4. Adsorption to host cells. Both (1) before and (2) after adsorption to cells, unadsorbed particles of 2xLB-grown (a) T7wt (109 cells/ml; multiplicity ˆ 10), (b) T7wt (4 x 108 cells/ml; multiplicity ˆ 500) and (c) a mixture of T7A and T7B (109 cells/ml; multiplicity ˆ 50) were analyzed by non-denaturing gel electrophoresisethidium staining. Electrophoresis was performed at 2 V/cm, for 22, 24 and 26 hours, respectively, through 0.6% IsoGel cast in either (a) P/M or (b), (c) T/M electrophoresis buffer. The arrowheads indicate the origins of electrophoresis; the arrow indicates the direction of electrophoresis.

trifugal pelleting of the bacteria. Bacteriophage particles in both state 1 and state 2 were separately assayed in the supernatant by non-denaturing gel electrophoresis. When the multiplicity of infection was 10, inspection of both the electrophoretic pro®le of unadsorbed bacteriophage particles in the supernatant (Figure 4a, lane 2) and the electrophoretic pro®le of bacteriophage particles before adsorption (Figure 4a, lane 1) revealed that the particles in state 2 had been selectively adsorbed. Before adsorption, S2* was 0.77; after adsorption, this ratio was indistinguishable from 0. When the multiplicity of infection was increased to 500, again selective adsorption was observed for the T7wt particles in state 2 (Figure 4b). Presumably because of increased occupancy of binding sites, the value of S2* underwent a change in Figure 4b that was smaller than the change in Figure 4a; S2* changed from 0.51 to 0.30. When the experiment of Figure 4b was repeated by use of cell concentrations between 4  108 and 40  108, S2* after adsorption varied from 0.2 to 0.7 the S2* before adsorption (®ve different experiments; average ˆ0.5  0.2). For all experiments, the adsorption of particles in state 2 was greater than the adsorption of particles in state 1. Loss of infectivity during electrophoresis prevents comparison of speci®c infectivity for T7 in each of its two states (Khan et al., 1997).

662

Polymorphism of T7

The effect of conditions of growth In a previous study performed by non-denaturing electrophoresis of intact T7wt through gels formed by IsoGel (Serwer et al., 1983), only a single band was observed. A possible cause for this previous observation of only one state is the use of M9 medium to grow the bacteriophage particles analyzed by Serwer et al. (1983). To test this hypothesis, ®ve independent crude T7wt bacteriophage preparations (Materials and Methods) were made by growth in M9 medium at 37 C; the host was infected at 4  108 cells/ml at a multiplicity of 5. The non-denaturing gel electrophoretic pro®le of bacteriophage particles from all ®ve preparations had only one band during electrophoresis in either T/M electrophoresis buffer (Figure 5a) or P/M electrophoresis buffer (Figure 5e; lanes are labeled by the name of the bacteriophage, followed by the growth medium, the temperature of growth, the state of the cells and the mole fraction of p10B, the latter in parentheses; a single lane is used for each set of conditions). When the same experiment was done by use of 2xLB medium, instead of M9 medium, bacteriophages from all preparations yielded the expected two bands (Figure 5b). As shown in Figure 5e, the single band of M9 medium-grown bacteriophage was most closely aligned with the origin-proximal of the two bands of the 2xLB medium-grown bacteriophage. Because the same high multiplicity 2xLB-grown inoculum was used for all cultures of Figure 5a and b, variability of progeny bacteriophage particles was con®rmed to have a source other than variability of the bacteriophage genomes. Further con®rmation was the observation that M9 medium-grown T7wt formed two bands after subsequent growth in 2xLB-medium (data not shown). To further investigate the effect of growth conditions on the state of T7 bacteriophage particles, the experiment performed at 37 C in Figure 5a and b was repeated at 42 C. In the case of the M9 medium-grown bacteriophage, a single band was observed (not shown). In the case of the 2xLBgrown bacteriophage, two bands were observed (Figure 5c). At 42 C, the mean S2* was 0.40, lower than it was for 37 C (0.45); when P/M electrophoresis buffer was used, the S2* values were 0.55 and 0.70, respectively (gel pro®les not shown). Because 42 C is higher than the temperature at which T7wt could have grown in the wild, these S2* values suggest the hypothesis that S2* decreases as the conditions of growth become less favorable. To further test this hypothesis, the experiment of Figure 5 was repeated at 42 C, but the bacterial cells used for growth were in late-log phase, a condition that should be even more unfavorable for growth. In support of the above hypothesis, the mean S2* was signi®cantly lower (0.30 in T/M, 0.50 in P/M electrophoresis buffer) for the 42 C, 2xLB, late-log phase host-grown T7wt (results for T/M are in Figure 5d; results for P/M are in Figure 5e) than it was for the 42 C, 2xLB, mid-log

Figure 5. The effect of conditions of growth. After use of the more limited procedure (Materials and Methods) for purifying bacteriophage particles from lysates of infected cells, electrophoresis of the lysates was performed. The gels were stained with ethidium. For each condition in a to d, the pro®les are shown for several lysates, each of which was independently obtained by use of the following conditions: (a) M9 medium, 37 C, mid-log phase cells, (b) 2xLB medium, 37 C, mid-log phase cells, (c) 2xLB medium, 42 C, mid-log phase cells, (d) 2xLB medium, 42 C, late-log phase cells, (e) indicated above a lane in the Figure. The conditions of electrophoresis were (a to d) 2 V/cm, 0.6% IsoGel cast in T/M electrophoresis buffer, (e) 2 V/cm, 0.6% IsoGel cast in P/M electrophoresis buffer. For e, the lanes are labeled by the name of the bacteriophage, followed by the medium used for growth, the temperature of growth, the state of host cells, and the mole fraction of p10B in the outer shell. For a given condition in b to d, all lysates were derived from the same master culture, within ten minutes of each other. The arrowheads indicate the origins of electrophoresis; the arrow indicates the direction of electrophoresis.

phase host-grown T7wt in Figure 5c. In further support, when T7wt was grown in M9 medium on a late-log phase host, two bands were formed by the progeny T7wt bacteriophage particles in T/M electrophoresis buffer (average S2* ˆ 0.86; gel pro®les not shown). The effect on S2* of changing the state of the cells was also observed when the temperature was either 30 or 37 C (Figure 5e). Effect of the mole fraction of p10B Because the two-state electrophoretic polymorphism of bacteriophage T7 is derived from a

663

Polymorphism of T7 Table 1. Electrophoretic characteristics of bacteriophagesa Bufferb

Particlec

ÿmod (cm2/V.S  10ÿ4)

ÿd (ESU/cm2  103)

ÿZ (102)

T/M

T7wt (1) T7wt (2) T7B T3 T7wt (1) Twwt (2) T7B T3

0.49 0.42 0.86 0.77 0.79 0.69 1.04 0.99

1.6 1.4 2.8 2.5 2.6 2.3 3.4 3.3

3.8 3.3 6.7 6.0 6.1 5.4 8.1 7.6

P/Md

a Values of d and Z were calculated from mo by use of procedures described by Serwer & Pichler (1978). Experimental error is 10%. b Electrophoresis buffer used to quantify mo. c Bacteriophage characterized. The numbers 1 and 2 refer to states 1 and 2. d Values of mo for T7wt in P/M electrophoresis buffer are signi®cantly higher than given by Serwer & Pichler (1978), because in this previous study correction was not made for the electro-osmosis induced by electrically charged groups on the agarose gel.

change in the d of the T7 outer shell, a possible factor in controlling this polymorphism is the relative amount of p10B in the T7 outer shell. To determine whether the presence of p10B is necessary for this polymorphism, 2xLB-grown T7A was subjected to electrophoresis in both P/M electrophoresis buffer (pH 7.4) and T/M electrophoresis buffer (pH 8.4). For 2xLB, 37 C-grown T7A, both states were clearly occupied in T/M electrophoresis buffer (average S2*, 0.69; Figure 1d,) but only state 2 was occupied in P/M electrophoresis buffer (Figure 5e), unless late log phase cells were used (gel pro®le not shown). These results indicate that p10B is not necessary for the two-state electrophoretic polymorphism. Like T7wt, M9 medium-grown T7A only populated state 2 (gel pro®le not shown). The 2xLB medium-grown T7A underwent pH-dependent changes in S2* that were similar to the changes undergone by T7wt (Figure 3). To determine whether the presence of p10A is necessary for the two-state T7 electrophoretic polymorphism, 2xLB-grown T7B was also subjected to electrophoresis in both T/M electrophoresis buffer and P/M electrophoresis buffer. In both cases, only one band was observed (illustrated for T/M electrophoresis buffer in Figure 1d). Electrophoresis in both H/M electrophoresis buffer (pH 9.5) and Pt/M electrophoresis buffer (pH 5.7) also yielded only one band (data not shown). Thus, the conclusion is drawn that T7B occupies only one electrophoretic state. In T/M electrophoresis buffer, the d of this state of T7B is greater in magnitude than the d of either state 1 or state 2, or even bacteriophage T3 (Table 1). Thus, the value of m does not determine which (if either) of the two T7wt states is occupied by T7B. For T7B, a possible cause of the increased negative surface charge is addition to the outer bacteriophage surface of negative charge that is on the C-terminal region of p10B, a region that is not present in p10A. The Cterminal region of p10B has a net charge of negative 3 (0 Asp, seven Glu, three Arg, one Lys); the three most C-terminal charged amino acids are Glu (Dunn & Studier, 1983). Assuming 420 p10B sub-

units in the outer shell of T7B (Steven & Trus, 1986), the C terminus of p10B is capable of contributing 1260 negative charges per bacteriophage particle to d. Of the negative charges in the outer shell of T7B in P/M electrophoresis buffer, approximately 810 contribute to d (Table 1). This latter number is 64% of the negative charge that is potentially added by the C terminus of p10B, if the entire C terminus is on the outer surface of T7. Thus, the negative charge of the C terminus of p10B is more than suf®cient to cause the increase in d observed for T7B. Although the presence of p10B is not necessary for the occupying of two states by T7wt, a variability in the position of both origin-proximal and origin-distal T7wt bands was observed. This variability does appear correlated with the mole fraction of p10B in the T7 outer shell; the higher this mole fraction, the greater the magnitude of m (examples are in Figure 5e). Thus, the variable presence of p10B appears to have a secondary effect on m, even for T7wt. A clear correlation of this secondary effect with growth conditions has not been observed. To determine whether the adsorption to a host cell differed between T7A and T7B, the experiment of Figure 4a was repeated by use of a mixture of T7A and T7B. As previously found in Figure 4a and b, adsorption was greatest for T7A in state 2; all detectable particles in state 2 were adsorbed. In contrast, less than 15% of the T7A particles in state 1 were adsorbed; 65% of the T7B particles were adsorbed (Figure 4c). Thus, the adsorption of T7B was intermediate to the adsorption of T7A in state 1 and T7A in state 2.

Discussion For the electrophoretic polymorphism of T7wt, the data presented here reveal two causes that are not derived from variability of the T7 genome: (1) the variable presence of negative surface charge from p10B, and (2) the occupying by the capsid of two structural states that differ in d. Because a par-

664 ticle can change state, the assumption is made that the two states are not differentiated by composition. The occupying of two states by T7A supports this assumption. In contrast, compositionbased electrophoretic polymorphism has previously been described for a procapsid of bacteriophage T3 (Serwer et al., 1985). The structural basis for the two-state electrophoretic polymorphism is, however, not known. State 1 is favored by pH values on either side of neutrality. In the wild, the deviation of pH from neutrality presumably signi®es improbability of the presence of a healthy host. Thus, in terms of evolution, one possible explanation for state 1 is the observed reduction in host-adsorption of T7, when conditions are not favorable for infection, i.e. the adaptation discussed in the Introduction. Because of the inert character of state 1, this hypothesis will be called the inert state hypothesis. The following observation supports the inert state hypothesis. For 2xLB medium-grown T7wt, growth conditions that appeared unfavorable promoted occupancy of state 1. This hypothesis proposes a form of intergenerational communication that appears plausible for both plant and animal viruses, as well as bacteriophages. However, to the authors' knowledge, the polymorphism described here has not been described for other viruses. In the case of spherical (icosahedral) single-stranded RNA plant viruses, change in state of the capsid has been caused by varying the concentration of divalent cations. However, this change includes a change in radius, i.e. swelling in the presence of chelating agents (for reviews see: Rossmann & Johnson, 1989; Liljas, 1991). In contrast, the data presented here indicate no change in RE greater than 2%. To the authors' knowledge, the most analogous polymorphism is a pH-induced change in structure that was accompanied by a change in d, in the case of ®lamentous bacteriophages (Bhattacharjee et al., 1992). The absence of the twostate polymorphism in the case of bacteriophage T3 is not understood. Whatever the structural basis of the two-state T7 polymorphism, change of state can be caused by a change in pH. However, when the pH is constant, most bacteriophage particles must not change state during the time of electrophoresis. If most particles did change state during electrophoresis, then the formation of two distinct bands would not have been observed. Instead, a more continuous, averaged distribution would have been observed. Changing the pH changes the state of some (though not all) T7wt particles in a time less than that of the electrophoresis. These observations indicate that substates of the T7 capsid exist that differ in the pH at which change in state occurs. The structural difference between T7 particles in states 1 and 2 is not known. However, for the following reasons, the difference in d probably resides in the T7 outer shell, rather than the T7 tail. (1) During T7 morphogenesis, a DNA-free capsid (procapsid) is assembled that subsequently both

Polymorphism of T7

packages DNA and changes in structure. Among the changes in procapsid structure is a change in d that is larger than the difference in d between states 1 and 2 of the mature T7 capsid (Serwer & Pichler, 1978); p10A and p10B are the outermost proteins of both the T7 procapsid and the mature T7 capsid (Serwer et al., 1982). Thus, the outer T7 shell has the capacity for a change in d comparable to the change that differentiates states 1 and 2 of the mature T7 capsid. (2) In contrast, addition of the entire tail to T7 causes a comparatively small change in d, roughly equal to the change that differentiates states 1 and 2 (Serwer & Pichler, 1978); no change in d of the tail is known to exist. That the difference in d resides in the outer shell is also the most direct explanation for the observation that, for 2xLB medium-grown bacteriophages in P/M electrophoresis buffer, the electrophoretic polymorphism was eliminated by removing p10B from the outer shell. Although most difference in d appears not to reside in the tail, the conformation of the tail may still be different in state 1 than it is in state 2. Difference in the conformation of the tail appears to be the best explanation for the observation that T7B host-adsorbs better than T7wt particles in state 1, even though host-bacteriophage electrostatic repulsion is greater for T7B than it is for T7wt particles in state 1. Possibly, changes in the structure of the outer shell are linked to changes in the structure of the tail.

Materials and Methods Bacteriophages and bacteria The following related wild-type bacteriophages were received from Dr F. W. Studier: fII, T3 and T7 (Studier, 1979; Studier & Dunn, 1983); wild-type T7 will be indicated by T7wt. The T7-T3 hybrid is the A1 tail ®ber hybrid constructed by Khan et al. (1997). All these bacteriophages were grown in Escherichia coli BB/1. The outer shell of T7 consists of p10A (344 amino acid residues) and p10B (397 residues), the latter produced by a translational frame-shift at residue 341 of p10A. To produce bacteriophage T7 that had an outer shell without p10B (i.e. only p10A present), a C-terminally deleted T7 gene 10 mutant (nucleotides 24,005 to 24,169; Rosenberg, A. and Studier, F. W., unpublished work) was grown on E. coli BL21. To produce bacteriophage T7 that had an outer shell without p10A, a completely deleted gene 10 mutant (nucleotides 22,873 to 24,168; Rosenberg, A. and Studier, F. W., unpublished work) was grown on a strain of E. coli BL26 that carried a plasmid with a copy of p10B cloned behind a lac promoter that was fully induced 30 minutes before infection (Studier et al., 1990). Both deletion mutants were received from Drs A. H. Rosenberg and F. W. Studier. To indicate T7 that has an outer shell with a controlled mole fraction of p10B, the following nomenclature will be used: T7 that has an outer shell made only of p10A will be called T7A; T7 that has an outer shell made only of P10B will be called T7B. Unless otherwise indicated, growth of all bacteriophages was performed at 37 C on mid-log phase cultures (2  108 to 5  108 cells/ml) aerated by forced air in either a glucose-based synthetic medium (M9

Polymorphism of T7 medium; Serwer et al., 1983) or 2xLB medium: 20 g tryptone, 10 g yeast extract, 5 g NaCl per liter. The growth medium used will be indicated in Results. If not otherwise indicated, bacteriophages were puri®ed by centrifugation in cesium chloride density gradients (Serwer et al., 1983); puri®ed bacteriophages were dialyzed against 0.2 M NaCl, 0.01 M Tris-HCl (pH 7.4), 0.001 M MgCl2. The concentration of bacteriophages was determined by measurement of absorbance at 260 nm (Bancroft & Freifelder, 1970). By use of sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (see below), the mole fraction of p10B in the T7 outer shell was determined for each bacteriophage preparation. Although this mole fraction varies among different T7wt preparations (Steven & Trus, 1986), it is usually not more than 0.1 (Condron et al., 1991). The composition of the outer shell of both T7B and T7A was con®rmed by SDS-polyacrylamide gel electrophoresis. The difference in mole fraction of p10B was the only observed difference in protein composition among T7wt , T7A and T7B. For determining the effect of growth conditions on their non-denaturing gel electrophoretic pro®le, bacteriophages were puri®ed by use of the following, more limited procedure: After spontaneous lysis of cells infected at a multiplicity of 5, sodium chloride was added to the lysate (50 g/liter of lysate). Subsequently, the lysate was clari®ed by centrifugation at 10,000 rpm, 4 C, for ten minutes in a Beckman J21 rotor. Next, bacteriophages were pelleted by centrifugation at 18,000 rpm, 4 C, for 1.5 hours in a Beckman J21 rotor; the pellet was resuspended in 0.5 M NaCl, 0.01 M Tris-HCl (pH 7.4), 0.001 M MgCl2. Finally, the concentrated bacteriophage preparation was clari®ed a second time. Unless otherwise indicated, the cultures used for these experiments were in mid-log phase (4  108 bacteria/ml; indicated by mid in Figure 5). Cultures in late-log phase (2  109 bacteria/ml; indicated by late in Figure 5) were also sometimes used. Non-denaturing one-dimensional agarose gel electrophoresis To perform non-denaturing agarose gel electrophoresis, bacteriophage particles were diluted into a solution of sucrose (®nal concentration 4%), bromophenol blue (®nal concentration 400 mg/ml) and electrophoresis buffer. This mixture was layered in the sample well of a submerged horizontal slab gel cast in one of the electrophoresis buffers described in the next paragraph. The gel was formed by either underivatized agarose (Seakem LE) or IsoGel. Both gel-forming compounds were obtained from FMC Bioproducts (Rockland, ME). Gels formed by IsoGel have been used in the past to avoid the adherence of T7wt to the gel, during non-denaturing gel electrophoresis; this adherence is caused by tail ®bers (Serwer et al., 1983; Khan et al., 1997). Unless otherwise indicated, electrophoresis with buffer recirculation (minimum rate, 100 ml/min) was performed at room temperature (22( 3) C), by use of the indicated time and electrical potential gradient. When indicated, improved temperature control (0.5 deg. C) was achieved by procedures previously described (Khan et al., 1997). For quantitative studies of electrophoretic migration as a function of gel concentration, the gels used (running gels) were embedded in a frame of Seakem LE agarose (see Serwer et al., 1983). After electrophoresis, a gel was stained and destained in electrophoresis buffer with 1 mg/ml ethidium; the gel was photographed through a Tiffen 3A (orange) ®lter, during illumination with a

665 300 nm (peak wavelength) ultraviolet transilluminator. Subsequently, the gel was destained in 0.001 M sodium EDTA (pH 7.4), and re-photographed. In the case of bacteriophage particles, but not unpackaged DNA, a fourto ®vefold increase in band intensity occurred during the second destaining, because of expulsion of DNA from the T7 capsid. This increase was used to con®rm that all bacteriophage bands were formed by bacteriophage particles, not DNA expelled from bacteriophages. To quantify DNA in-gel after expulsion, a previously described procedure of video ¯uorometry was used (Griess et al., 1995). Photographs were both digitized and reproduced by procedures previously described (Khan et al., 1997). To determine the effect of electrophoresis buffer on the gel electrophoretic pro®le of bacteriophage particles, the following electrophoresis buffers were used: Ta/M, 0.05 M sodium tartrate (pH 4.5), 0.001 M MgCl2; C/M, 0.05 M sodium citrate (pH 5.0), 0.001 M MgCl2; Pt/M, 0.05 M sodium phthalate (pH 5.7), 0.001 M MgCl2; P/M, 0.05 M sodium phosphate (pH 7.4), 0.001 M MgCl2; T/M, 0.09 M Tris-acetate (pH 8.4), 0.001 MgCl2; H/M, 0.05 M sodium parahydroxybenzoate (pH 9.5), 0.001 M MgCl2. The ionic strength for all buffers was 0.14. When indicated in Results, the concentration of sodium phosphate in P/M buffer was changed. The same procedure of electrophoresis was used for all electrophoresis buffers. During gel electrophoresis, the electrophoretic mobility (m ˆ velocity/electrical potential gradient) of a particle is determined by two characteristics of a particle, the effective radius (RE) and the average electrical surface charge density (d). The value of d is directly proportional to m that has been both extrapolated to a gel concentration of 0 (moÂ) and corrected for electro-osmosis (mo). The slope of a m versus gel concentration plot increases in magnitude as the RE of the particle increases (for reviews see: Shaw, 1969; Serwer, 1983; Chrambach, 1985). To obtain values of d, values of m as a function of gel concentration were determined by use of bacteriophage T3 as an internal standard whose m values had been determined at temperature controlled with an accuracy of 0.5 deg. C. In the case of gels formed by IsoGel, no correction for electro-osmosis (Serwer, 1983) was necessary. Values of d were calculated from mo by use of procedures previously described (Shaw, 1969; Serwer & Pichler, 1978). From d, the total number of surface electrical charges per particle (Z) was calculated from the known radius (30.1 nm; Serwer et al., 1983) of T7.

Two-dimensional non-denaturing gel electrophoresis To determine the effect of changing electrophoresis buffer on gel electrophoretic pro®le, ®rst, gel electrophoresis was performed in a ®rst dimension at 2 V/cm, for 18 hours, at room temperature, by use of 0.75% IsoGel cast in the indicated electrophoresis buffer. Subsequently, the gel was rotated by p/2 radians; the electrophoresis buffer was exchanged for a second electrophoresis buffer. This exchange was performed by equilibration of the gel against the second electrophoresis buffer, once renewed; the time for exchange was two to three hours. Finally, electrophoresis was performed in the second dimension, by use of the same time, temperature and electrical potential gradient used for the ®rst dimension. Procedures of staining were the same as those used for one-dimensional non-denaturing gel electrophoresis. This procedure of two-dimensional gel electrophoresis will be called Procedure 1.

666 To compare the RE of one particle to the RE of another, a second procedure of two-dimensional non-denaturing gel electrophoresis was used (Procedure 2). Electrophoresis in the ®rst dimension was performed at 2 V/cm, for 22 hours, at room temperature, through a comparatively dilute (0.15%) gel formed by IsoGel cast in T/M electrophoresis buffer. Orthogonally oriented electrophoresis in the second dimension was performed at 2 V/cm, for 58 hours, at room temperature, through 1.8% IsoGel cast in T/M electrophoresis buffer. Both procedural details and interpretation are reviewed by Tietz (1987). Sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis To determine the proteins present in a bacteriophage preparation, SDS-polyacrylamide gel electrophoresis was performed. A 10% (w/v) polyacrylamide gel was used. Gels were stained with Coomassie blue (Serwer et al., 1983). By densitometry of a photograph, the mole fraction of p10B in the T7 outer shell was determined (Serwer et al., 1983). Quantification of adsorption to host cells To quantify the adsorption of bacteriophage particles to host cells, 30 C-2xLB-grown-log phase E. coli BB/1 at 4  108/ml were, ®rst, concentrated by centrifugation (10,000 rpm, 2 C, in a Beckman TL-100 centrifuge). These cells were resuspended in 0.5 M NaCl, 0.01 M Tris-Cl (pH 7.4), 0.001 M MgCl2, at the indicated ®nal concentration. At the indicated multiplicity of infection, bacteriophage were added to resuspended cells at 37 C. This bacteriophage-host mixture was incubated for ®ve minutes before pelleting bacteriophage-host complexes by centrifugation at 10,000 rpm, 2 C, for ten minutes, in a Beckman TL-100 centrifuge. Unadsorbed bacteriophage particles in the supernatant were both detected and characterized by non-denaturing agarose gel electrophoresis-ethidium staining. To quantify the particles that formed a band, ethidium ¯uorescence was quanti®ed by video ¯uorometry.

Acknowledgements We thank Drs Alan H. Rosenberg and F. William Studier for both providing T7 deletion mutants and making helpful comments, Karen Lieman and Michele Gates for technical assistance and Linda C. Winchester for typing this manuscript. We gratefully acknowledge support form the National Institutes of Health (GM24365) and the Robert A. Welch Foundation (AQ-764).

References Bancroft, F. C. & Freifelder, D. (1970). Molecular weights of coliphages and coliphage DNA. I. Measurement of the molecular weight of bacteriophage T7 by high-speed equlibrium centrifugation. J. Mol. Bol. 54, 537 ± 546. Bayer, M. E. & Sloyer, J. L., Jr (1990). The electrophoretic mobility of Gram-negative and Gram-positive bacteria: an electrokinetic analysis. J. Gen. Microbiol. 136, 867 ± 874. Bhattacharjee, S., Glucksman, M. J. & Makowski, L. (1992). Structural polymorphism correlated to sur-

Polymorphism of T7 face charge in ®lamentous bacteriophages. Biophys. J. 61, 725± 735. Casjens, S. (1985). Nucleic acid packaging by viruses. In Virus Structure and Assembly (Casjens, S., ed.), pp. 75±147, Jones and Bartlett, Boston. Chrambach, A. (1985). The Practice of Quantitative Gel Electrophoresis. VCH, Weinheim. Condron, B. G., Atkins, J. F. & Gesteland, R. F. (1991). Frameshifting in gene 10 of bacteriophage T7. J. Bacteriol. 173, 6998± 7003. Dunn, J. J. & Studier, F. W. (1983). Complete nucleotide sequence of bacteriophage T7 DNA and the locations of T7 genetic elements. J. Mol. Biol. 166, 477± 535. Griess, G. A., Louie, D. & Serwer, P. (1995). A desktop, low-cost video ¯uorometer for quantitation of macromolecules after gel electrophoresis. Appl. Theor. Electrophor, 4, 175± 177. Johnson, J. E., Munshi, S., Liljas, L., Agrawal, D., Olson, N., Reddy, V., Fisher, A., McKinney, B., Schmidt, T. & Baker, T. S. (1994). Comparative studies of T ˆ 3 and T ˆ 4 icosahedral RNA insect viruses. Arch. Virol. 9, 497± 512. Kellenberger, E. (1990). Form determination of the heads of bacteriophages. Eur. J. Biochem. 190, 233± 248. Khan, S. A., Watson, R. H., Hayes, S. J. & Serwer, P. (1997). PCR-directed formation of viral hybrids in vitro. Virology, 227, 409± 419. Liljas, L. (1991). Structure of spherical viruses. Int. J. Biol. Macromolecules, 13, 273± 280. Liljas, L. (1996). Viruses. Curr. Opin. Struct. Biol. 6, 151 ± 156. McKenna, R., Xia, D., Willingmann, P., Ilag, L. L., Krishnaswamy, S., Rossmann, M. G., Olson, N. H., Baker, T. S. & Incardona, N. L. (1992). Atomic structure of single-stranded DNA bacteriophage fX174 and its functional implications. Nature, 355, 137 ± 143. McKenna, R., Bowman, B. R., Ilag, L. L., Rossman, M. G. & Fane, B. A. (1996). Atomic structure of the degraded procapsid particle of the bacteriophage G4: Induced structural changes in the presence of calcium ions and functional implications. J. Mol. Bol. 256, 736± 750. Olson, N. H., Baker, T. S., Willingmann, P. & Incardona, N. L. (1992). The three-dimensional structure of frozen-hydrated bacteriophage fX174. J. Struct. Biol. 108, 168± 175. Rossmann, M. G. & Johnson, J. E. (1989). Icosahedral RNA virus structure. Annu. Rev. Biochem. 58, 533 ± 573. Serwer, P. (1983). Agarose gels: properties and use for electrophoresis. Electrophoresis, 4, 375± 382. Serwer, P. & Pichler, M. E. (1978). Electrophoresis of bacteriophage T7 and T7 capsids in agarose gels. J. Virol. 28, 917± 928. Serwer, P., Hayes, S. J. & Watson, R. H. (1982). The structure of a bacteriophage T7 procapsid and its in vivo conversion product probed by digestion with trypsin. Virology, 122, 392± 401. Serwer, P., Watson, R. H., Hayes, S. J. & Allen, J. L. (1983). Comparison of the physical properties and assembly pathways of the related bacteriophages T7, T3 and fII. J. Mol. Biol. 4, 375± 382. Serwer, P., Watson, R. H. & Hayes, S. J. (1985). Heterogeneity of the procapsid of bacteriophage T3. J. Virol. 55, 232± 237. Shaw, D. (1969). Electrophoresis. Academic Press, London.

667

Polymorphism of T7 Steven, A. C. & Trus, B. L. (1986). The structure of bacteriophage T7. In Electron Microscopy of Proteins (Harris, J. R. & Horne, R. W., eds), vol. 5, pp. 1 ± 35, Academic Press, London. Studier, F. W. (1979). Relationships among different strains of T7 and among T7± related bacteriophages. Virology, 95, 70± 84. Studier, F. W. & Dunn, J. J. (1983). Organization and expression of bacteriophage T7 DNA. Cold Spring Harbor Symp. Quant. Biol. 47, 999±1007.

Studier, F. W., Rosenberg, A. H., Dunn, J. J. & Dubendorff, J. W. (1990). Use of T7 RNA polymerase to direct expression of cloned genes. Methods Enzymol. 185, 60 ± 89. Tietz, D. (1987). Gel electrophoresis of intact subcellular particles. J. Chromatog. 418, 305 ± 344. Wurtz, M. (1992). Bacteriophage structure. Electron Microsc. Rev. 5, 283± 309.

Edited by M. Yanagida (Received 12 May 1997; received in revised form 11 August 1997; accepted 13 August 1997)