Penicillin and Cell Wall Synthesis: A Study of Bacillus cereus by Electron Microscopy

The changes in wall structure of a penicillinase micro-constitutive strain of Bacillus cereus (569/H/24), on exposure to penicillin, and after its removal by addition of penicillinase, have suggested the following model for the growth of the walls of these cylindrical cells. Longitudinal extension is by addition of material to a large and continuously increasing number of growing points uniformly distributed over the cylindrical surface. Addition is only in the longitudinal direction so that the cell diameter remains constant. Cross walls grow by addition to their inner edge, and on completion the two new rounded ends of the daughter cells are formed by splitting at the outer edge and continued addition at the center. The ends are conserved.     

Lifetime of messenger ribonucleic acid for penicillinase synthesis in several strains of bacilli

1. Previous studies of penicillinase synthesis in Bacillus licheniformis showed that enzyme synthesis after the addition of actinomycin continues for far longer in the constitutive mutant 749/C than in the parental inducible strain (Yudkin, 1966). This result was interpreted as indicating a difference in the lifetime of specific messenger RNA in the two strains. Other bacilli have now been examined in an attempt to see whether this difference is general. 2. There was no difference in the lifetime of messenger RNA for penicillinase synthesis between an inducible and a constitutive strain of Bacillus cereus. 3. Three freshly isolated constitutive mutants of B. licheniformis also had short-lived messenger RNA, like their inducible parent. 4. A reinvestigation of mutant 749/C confirmed the original finding that, on treatment with actinomycin, it continued to synthesize penicillinase far longer than did its parent. 5. An inducible revertant of mutant 749/C was indistinguishable from the original inducible strain, and appeared to have lost both constitutivity and long-lived messenger RNA in the back mutation.     

Further evidence for a partially folded intermediate in penicillinase secretion by Bacillus licheniformis.

Protoplasis of Bacillus licheniformis 749/C (a mutant constitutive for penicillinase production) continued to synthesize and release penicillinase in hypertonic growth medium in the presence of trypsin and chymotrypsin at 25 mug each per ml. When the protoplasts were stripped of about half of their membrane-bound penicillinase by pretreatment at pH 9.5 or with a higher level of trypsin, penicillinase activity no longer increased in the presence of the proteases. This effect was immediately eliminated after addition of soybean trypsin inhibitor. These proteases do not significantly inhibit general protein synthesis. Stripped protoplasts of strain 749/C and of uninduced strain 749 (unable to synthesize penicillinase) were incubated with 50 mug of chymotrypsin per ml, and the supernatent fluids were examined immunochemically for peptides derived from the penicillinase chain. Such fargments were found only with the protoplasts capable of synthesizing penicillinase (strain 749/C). The direct detection of the products of protease degradation of a susceptible form of penicillinase provides strong evidence that, in stripped protoplasts of B. licheniformis 749/C, penicillinase synthesis continues in the presence of trypsin or chymotrypsin and that, in these modified membranes, the protease is able to act on an early form of the enzyme that has not yet attained the protease-resistant conformation characteristic of the membrane-bound and exopenicillinases. This finding is discussed in terms of the current models of penicillinase secretion.     

Characteristics of Penicillinase Secretion by Growing Cells and Protoplasts of Bacillus licheniformis

Cultures of the inducible penicillinase-producing strain 749 of Bacillus licheniformis, induced with small amounts of benzylpenicillin, synthesized penicillinase at a high rate for a short period, after which the rate of synthesis slowly declined. During the period of active synthesis, the rate of secretion, as a fraction of the level of cell-bound penicillinase (which is originally high), gradually decreased to a constant level. Chloramphenicol, at a concentration (40 mug/ml) which completely inhibited synthesis of penicillinase, partially inhibited secretion if added during the period of active synthesis. During the phase of reduced synthesis, chloramphenicol was without effect on secretion. Penicillinase secretion, by actively growing cultures of the constitutive penicillinase-producing mutant 749/C, was inhibited by 75% immediately after addition of chloramphenicol. The secretion of part of the penicillinase released during active growth is probably dependent on synthesis of penicillinase, but part of the secreted penicillinase can be released in the absence of synthesis. Protoplasts were obtained from which periplasmic penicillinase has been removed, and these protoplasts were capable of substantial growth and penicillinase synthesis without lysis. At pH 7.5, there was no net incorporation of penicillinase into the cell membrane; the enzyme released was almost entirely of the exo form and was roughly equivalent to the amount of new enzyme formed. At pH 6.0, there was some incorporation of penicillinase into the plasma membrane, and approximately half of the extracellular penicillinase was in the exo form; the remainder perhaps represented membrane fragments. In the presence of chloramphenicol, a small amount of penicillinase was released at pH 7.5 as the exo form; at pH 6.0, practically none was released. We suggest that, with the removal from protoplasts of the periplasmic penicillinase-containing particles, a restriction on secretion has been lifted.     

Cell wall turnover in growing and nongrowing cultures of Bacillus subtilis

Cell wall turnover was studied in cultures of Bacillus subtilis in which growth was inhibited by nutrient starvation or by the addition of antibiotics. Concomitantly, the synthesis of wall, as measured by the incorporation of radioactively labeled N-acetylglucosamine, was followed in some of these cultures. In potassium- or phosphate-starved cultures, growth stopped, but wall turnover continued at a rate slightly lower than that in the control cultures. Lysis of cells did not occur. In glucose-starved cultures, continued wall turnover caused lysis of cells, since wall synthesis apparently was inhibited. The same phenomenon was observed after growth arrest by the addition of wall synthesis inhibitors such as fosfomycin, cycloserine, penicillin G, and vancomycin. Growth arrest by the addition of chloramphenicol allowed the continuation of wall synthesis; therefore, the observed turnover generally did not cause cell lysis.     

Characteristics of penicillinase release by washed cells of Bacillus licheniformis

Saline-washed cells of Bacillus licheniformis strain 749/C (constitutive for penicillinase) were able to release exopenicillinase in the presence of concentrations of chloramphenicol that prevented protein synthesis completely. The release reaction was strongly pH-dependent, occurring at a faster rate at alkaline pH in anionic or cationic buffers than at neutral pH. A strongly pH-dependent release reaction was noted in growing cells also. The reaction in washed cells can be stopped completely by changing the pH to 6.0. Within 30 min at pH 9.0, about 55% of the cell-bound penicillinase was released; thereafter, release continued at a greatly reduced rate. Suspensions of washed cells retained their capacity to release penicillinase at pH 9.0 for 90 min. Penicillinase released at pH 9.0 from either cells or protoplasts was not readsorbed over a 60-min period after changing the pH to 6.0. The release reaction was strongly temperature-dependent. We examined the effect of a large number of metabolic inhibitors and other compounds on the pH-dependent release phenomenon. Quinacrine hydrochloride, chloroquine diphosphate, and chlorpromazine hydrochloride reduced secretion substantially at 10(-4)m. Deoxycholate and Triton X-100 were active at 10(-3)m, but tungstate, arsenate, and molybdate had small effects at 10(-1)m. The rate of exopenicillinase release at pH 9.0 from fully stabilized protoplasts was one-half that of intact cells. Protoplasts lysed in hypotonic media or detergents showed even greater reduction in releasing activity. Penicillinase released from washed cells at pH 7.5 or 9.0 appeared to be derived from the periplasmic tubule and vesicle fraction that was released by protoplast formation.     

Peptidoglycan biosynthesis in Neisseria gonorrhoeae strains sensitive and intrinsically resistant to beta-lactam antibiotics.

Treatment of penicillin-sensitive and intrinsically resistant Neisseria gonorrhoeae strains with their respective inhibitory concentrations of penicillin caused rapid cell death. When the peptidoglycan syntheses of these two strains were examined in the presence of penicillin, the sensitive strain continued to make this cell wall polymer for an extended time, whereas the resistant strain underwent a rapid and marked depression in synthesis. Examination of the labeled sodium dodecyl sulfate-insoluble peptidoglycan made in the presence of inhibitory concentrations of penicillin revealed further differences. The primary effect on the penicillin-sensitive gonococcus was a slight change in peptide cross-linking and a sharp decline in the degree of O-acetylation. In contrast, the resistant strain exhibited a substantial decline in cross-linking, with a very moderate change in O-acetylation. The degree of saturation of the individual penicillin-binding proteins (PBPs) was assessed under these conditions. PBP 2, which exhibits a reduced affinity for penicillin in the resistant strain, appeared to be related to O-acetylation, whereas PBP 1 was implicated in the transpeptidation reaction.     

Penicillin-Binding Component of Bacillus cereus

(14)C-penicillin is irreversibly bound by Bacillus cereus 569. Incubation of penicillin-treated cells in a cell wall digestion medium results in solubilization of approximately 60% of the irreversibly bound lable. The extent of the solubilization is the same when cells are prepared by either a cold or 37 C treatment procedure. However, spheroplasts prepared by the cold treatment are leaky. When the resulting spheroplasts are incubated in supplemented medium, reduced rates and levels of penicillinase synthesis, relative to induced whole-cell controls, are observed. Spheroplasts from both cold and 37 C prepared cells exhibit this phenomena, although the spheroplasts from 37 C prepared cells synthesized approximately sixfold higher levels of penicillinase. The size distribution of the label solubilized during the preparation of spheroplasts was examined by using Bio Gel P-150 columns. Although no label appeared in the exclusion volume fractions when the cell wall digest of the 37 C treated cells was chromatographed, approximately 10% of the label from cold-treated cells did appear. These results suggest that the presence of irreversibly bound penicillin is required for the synthesis of induced levels of penicillinase and that the irreversibly bound penicillin can be solubilized as a labile complex with material which is excluded from BioGel P-150. It may be concluded that the penicillin-binding lipoprotein complex which has been previously observed is the penicillin-specific binding site. However, the location of this complex in relation to the cell membrane could not be determined.     

Characteristics of Secretion of Penicillinase, Alkaline Phosphatase, and Nuclease by Bacillus Species

The distribution of alkaline phosphatase and nuclease activity between cells and medium was examined in one strain of Bacillus licheniformis and four strains of B. subtilis. Over 95% of both activities was found in the medium of the B. licheniformis culture, but in the B. subtilis cultures the amount of enzyme activity found in the medium varied with the strain and the enzyme considered. B. licheniformis 749 and its penicillinase magnoconstitutive mutant 749/C were grown in continuous culture with phosphorous as the growth-limiting factor, and the kinetics of penicillinase formation and secretion were examined. Nutrient arrest halted secretion (usually after a lag of about 30 min) in both the inducible and constitutive strains. Chloramphenicol did not eliminate secretion, but under certain circumstances reduced its rate. In the inducible strain treated with a low level of inducer, the rate of secretion was more affected by the rate of synthesis than by the level of cell-bound enzyme. During induction, the onset of accretion of cell-bound penicillinase and secretion of the exoenzyme were nearly simultaneous. It seems unlikely that a long-lived, membrane- or cell-bound intermediate is mandatory in the secretion of the three enzymes by Bacillus species. In the case of penicillinase secretion, there are at least two different phases. When penicillinase synthesis is proceeding rapidly, the rate of secretion is five to six times greater at equivalent concentrations of membrane-bound penicillinase than it is when penicillinase synthesis is reduced. The data require that any membrane-bound intermediate in the formation of exoenzyme be much shorter-lived in cells with a high rate of synthesis than in cells with a low rate. Either there are two separate routes for the secretion of penicillinase or the characteristics of the process vary substantially between the early stages and the declining phase of induction.     

Resistance of Escherichia coli to penicillins. V. Physiological comparison of two isogenic strains, one with chromosomally and one with episomally mediated ampicillin resistance

Two essentially isogenic strains of Escherichia coli K-12 were compared: D31 had chromosomally and D1-R1 episomally mediated resistance to ampicillin. The two strains had the same ability to form colonies on ampicillin plates, but in other tests they were quite different. In serial dilution tests as well as in exponentially growing cultures, D1-R1 was far more resistant to ampicillin than was D31. The inoculum effect with D1-R1 was large and with D31 was rather small. On plates, D31 was more resistant to penicillin G than was D1-R1. The penicillinase activity of buffer suspended cells against dl-ampicillin was 15 times higher for D1-R1 than for D31, but the two strains showed about the same rate of hydrolysis of penicillin G. With dl-ampicillin as substrate, for D1-R1 the apparent K(m) was 1.7 x 10(-4)m, whereas D31 gave a slightly sigmoid curve with a half-saturation concentration of about 5 x 10(-3)m. No induction of penicillinase activity was found. When the growth rate was varied by a factor of four, the amount of penicillinase per cell mass was constant in both D1-R1 and D31, whereas in two wild-type strains the amounts of penicillinase increased with increasing growth rates. With exponentially growing D1-R1, ampicillin disappearance started within 3 min, but at low ampicillin concentrations the rate was less than 10% of the rate of hydrolysis by buffer-suspended cells. Before D31 started hydrolysis, there was a lag period that lasted at least one generation and depended on the concentration of ampicillin. After this lag period, the rate of hydrolysis was 10 times higher than that observed with buffer-suspended cells. These differences between growing and nongrowing cells indicate that both the chromosomally and the episomally mediated penicillinases are controlled by some products present in growing cells.     

Effect of β-Lactamase Location in Escherichia coli on Penicillin Synergy

Resistance to ampicillin in Escherichia coli is due generally to the presence of a beta-lactamase (penicillinase). Resistant strains have been found to fall into two groups: those with high-level resistance (1,000 mug/ml or greater) and those with low-level resistance (8 to 250 mug/ml). Most of the high-level resistant organisms posses beta-lactamases whose synthesis is episomally mediated. These strains release penicillinase from the cell when they are subjected to osmotic shock. Low-level resistant strains do not release the enzyme with osmotic shock. High-level resistant strains are not susceptible to the synergistic action of a penicillinase-resistant penicillin with ampicillin. Seventy eight per cent of low-level resistant strains are susceptible to the synergistic action of ampicillin and oxacillin. The two types of beta-lactamases are similar in regard to most properties; both enzymes are subject to competitive inhibition by penicillinase-resistant penicillins. The difference in location in the cell might explain why only some strains of E. coli are susceptible to the synergistic action of penicillin combinations.     

Lysis of Saccharomyces cerevisiae with 2-deoxy-2-fluoro-D-glucose, an inhibitor of the cell wall glucan synthesis

The effect of a synthetic glucose analogue, 2-deoxy-2-fluoro-d-glucose (FG) on growth and glucose metabolism of Saccharomyces cerevisiae was studied. The addition of FG (0.005-0.05%) to a 2% glucose medium resulted in reduction of the initial growth rate and, after several hours, in a complete cessation of the culture growth. These two events were due to extensive lysis of the population which continued long after the period when no more growth was recorded. Electron microscope examination of lysed cells showed that the lysis was a consequence of a dissolution of the cell walls. FG inhibited to a similar extent the initial growth rate and the incorporation of radioactivity from labeled glucose into growing population. The inhibition of radioactivity incorporation from glucose by growing protoplasts was much less. The yeast was found to be extremely FG sensitive whenever the synthesis of new cell wall material was involved. All observations imply that FG interferes mainly with the cell wall formation of S. cerevisiae. A comparison of the FG effects on metabolic activity of protoplasts, simultaneous secretion of mannan-proteins into the growth medium, and the formation of glucan fibrils on the surface of protoplasts demonstrated that the cell wall glucan synthesis is the most FG-sensitive process and evidently the growth-limiting factor in intact cells. FG-resistant cells were selected during growth experiments. They exhibited an altered mode of cell division when grown in the presence of FG.     

Interaction of Bacteriophage Infection and Low Penicillin Concentrations on the Performance of Yogurt Cultures

Bacteriophage infection of a mixed-strain Streptococcus thermophilus culture, one strain of which is phage sensitive and the other phage resistant, may induce lysis of both strains. Experiments were carried out with three different phage-resistant strains. One such strain lysed in penicillin-free growth medium and another needed penicillin G (0.005 IU/ml) for lysis, while the third strain continued to grow in the presence of this concentration of antibiotic. Growth of the latter strain was inhibited when the medium contained a relatively high concentration of phage lysin. The different penicillin concentrations required to induce “lysis from without” of these phage-resistant strains correlated with their individual sensitivities to the antibiotic. The apparent relationship between the sensitivities of these strains to penicillin and to phage lysin could be explained by a difference in the degree of polymerization of the cell wall peptidoglycan.     

Morphological Phenomena Associated with Penicillinase Induction and Secretion in Bacillus licheniformis

Cells of uninduced Bacillus licheniformis (strain 749) in mid-logarithmic phase have no extensive intracytoplasmic membrane. After induction with cephalosporin C, characteristic organelles that contain tubules and vesicles with single-layered membranes and no visible internal substructure can be seen in thin sections in the periplasm. A magnoconstitutive penicillinase producer (749/C) contains similar structures. It is suggested that they represent a penicillinase secretory apparatus. In the first 15 min after induction, negatively stained preparations of induced 749 show large intracellular vesicles without individual contact with the cell surface. Negatively stained 749/C and fully induced 749 contain invaginations comparable to the structures seen in thin sections. When protoplasts of induced 749 and of 749/C are prepared, vesicles and tubules similar to those seen in thin sections of whole cells are liberated from the cell. Growing protoplasts of induced 749 show massive convolutions of the peripheral membrane, multiple layers of membrane, and characteristic long, slender tubules extending from the protoplast surface. These phenomena are not observed in uninduced 749 except for the production of a relatively small number of tubules. In 749/C, there were fewer convolutions than in induced 749, although tubule production was similar. Multiple layers of membrane were not observed in 749/C. The relation of the penicillinase secretory structures to mesosomes and to secretory structures of other organisms is discussed.     

Regulation of Staphylococcal Penicillinase Synthesis

5-Methyl tryptophan was found to be an efficient inducer of penicillinase synthesis in Staphylococcus aureus. Addition of actinomycin D or tryptophan to the culture medium shuts off the 5-methyl tryptophan-induced synthesis of penicillinase with an apparent half-life of approximately 1 to 2 min, respectively. Hence, in the induction of penicillinase synthesis, 5-methyl tryptophan seems to function as a structural analogue of penicillin rather than by becoming incorporated in proteins and thereby creating faulty penicillinase repressor or antirepressor. This conclusion is supported by similarities in the structures of the two compounds as revealed by solid atomic models. The fact that S. aureus exposed to (14)C-penicillin in the absence of protein synthesis failed to synthesize penicillinase at an increased level when cell growth was resumed strongly suggests that a protein involved in the regulation of penicillinase synthesis must be synthesized in the presence of the penicillinase inducer. In turn, this observation suggests that the penicillinase inducer promotes penicillinase synthesis by directing the penicillinase regulatory protein (i.e., the penicillinase antirepressor) to acquire a different conformation when it is synthesized in the presence of the penicillinase inducer. A working model for the regulation of penicillinase synthesis based on these and other data has been constructed and is presented.     

Vesicle penicillinase of Bacillus licheniformis: existence of periplasmic-releasing factor(s).

In earlier studies of the membrane-bound penicillinase of Bacillus licheniformis 749/C, the enzyme present in the vesicles that were released during protoplast formation and the enzyme retained in the plasma membrane of protoplasts appeared to differ (i) in their behavior on gel permeation chromatography in the presence or absence of deoxycholate and (ii) in their tendency to convert to the hydrophilic exoenzyme (Sargent and Lampen, 1970). We have now shown that these vesicle preparations contain a soluble, heat-sensitive enzyme(s) that is released along with the vesicles during protoplast formation. The enzyme will convert the vesicle penicillinase to a form that resembles exopenicillinase, and this conversion can be inhibited by deoxycholate under certain circumstances. Sedimentation of such vesicle preparations at 100,000 X g produces vesicles which contain penicillinase that behaves as the plasma membrane enzyme obtained from protoplasts. Exopenicillinases released by growing cells at pH 6.5 and by washed cells or protoplasts at pH 9.0 have the same NH2-terminal residues (lysine and some glutamic acid); in addition, the various release systems show a parallel sensitivity to inhibition by deoxycholate, quinacrine, chloroquine, and o-phenanthroline. The formation of exopenicillinase (by cleavage of the membrane-bound enzyme) may well be dependent on the action of the releasing enzyme.     

Periplasmic Structure of Frozen-Etched and Negatively Stained Cells of Bacillus licheniformis as Correlated with Penicillinase Formation

Bacillus licheniformis strain 749/C (constitutive for penicillinase formation) and uninduced cells of strain 749 (penicillinase-inducible) were examined after freezeetching. In the early stationary phase, strain 749/C organisms had clusters of vesicles (30 to 40 nm in diameter) on the outer surface of the plasma membrane. These are randomly distributed on the membrane, including the region of septum formation. The vesicles are not intimately associated with the plasma membrane, and their inner and outer surfaces are devoid of particles. Periplasmic vesicles were not detected by freeze-etching in strain 749 (uninduced) or in young cells of 749/C; however, the membrane of mid-logarithmic phase 749/C cells had a corrugated appearance. Negatively stained 749/C cells (logarithmic phase) also showed many vesicular and tubular bodies in the periplasm as well as septal and cytoplasmic mesosomes of typical morphology. The periplasmic structures appear to be formed either by evagination of plasma membrane or by migration of vesicular bodies from the membranous pockets of the cytoplasm. Stationary phase cells of 749/C still have many periplasmic vesicular bodies; however, the mesosomes are greatly reduced both in number and size. In sharp contrast, strain 749 organisms have very few structures similar to the periplasmic bodies of strain 749/C. These findings support our previous view that penicillinase-producing cells of 749/C have periplasmic membranous structures that are rare in the uninduced strain 749, though there is some lack of correspondence between freeze-etching, negative staining, and thin section data. These structures may be important for the retention or storage of penicillinase in the cell.     

Solute production and net wall synthesis in the growing and non-growing cells of gravistimulated sunflower hypocotyls

Solute generation and cell wall synthesis were examined in sunflower hypocotyl peripheral layers, the growth rate of which had been altered by gravistimulation. Measurements of both the concentrations of the major solutes and the osmotic potential showed that although upper cells stopped growing, the solute levels in these cells continued to increase at rates comparable to those in lower cells. This indicated that altered growth rates, generated during gravicurvature, are not based on solute generation but must result from cell wall changes. Gravimetric and precursor incorporation studies showed that net wall synthesis continued in upper cells despite their lack of growth. An ultrastructural study of the epidermal cells on the uppermost (non-elongating) and lowermost (elongating) surfaces of horizontal cucumber hypocotyls showed that the relative amounts of the various membrane fractions were similar in upper and lower cells despite their very different growth rates.     

Cell wall and cytoskeleton reorganization as the response to hyperosmotic shock in Saccharomyces cerevisiae

Transfer of exponentially growing cells of the yeast Saccharomyces cerevisiae to hyperosmotic growth medium containing 0.7-1 M KCl, 1 M mannitol, and/or 1 M glycerol caused cessation of yeast growth for about 2 h; thereafter, growth resumed at almost the original rate. During this time, formation of fluorescent patches on the inner surface of cell walls stained with Primulin or Calcofluor white was observed. The fluorescent patches also formed in solutions of KCl or when synthesis of the cell wall was blocked with cycloheximide and/or 2-deoxyglucose. The patches gradually disappeared as the cells resumed growth, and the new buds had smooth cell walls. Electron microscopy of freeze-etched replicas of osmotically stressed cells revealed deep plasma membrane invaginations filled from the periplasmic side with an amorphous cell wall material that appeared to correspond to the fluorescent patches on the cell surface. The rate of incorporation of D-[U-14C]glucose from the growth medium into the individual cell wall polysaccharides during osmotic shock followed the growth kinetics. No differences in cell wall composition between osmotically stressed yeast and control cells were found. Hyperosmotic shock caused changes in cytoskeletal elements, as demonstrated by the disappearance of microtubules and actin microfilaments. After 2-3 h in hyperosmotic medium, both microtubules and microfilaments regenerated to their original polarized forms and the actin patches resumed their positions at the apices of growing buds. The response of S. cerevisiae strains with mutations in the osmosensing pathway genes hog1 and pbs2 to hyperosmotic shock was similar to that of the wild-type strain. We conclude that, besides causing a temporary disassembling of the cytoskeleton, hyperosmotic shock induces a change in the organization of the cell wall, apparently resulting from the displacement of periplasmic and cell wall matrix material into invaginations of the plasma membrane created by the plasmolysis.     

Immunoelectron microscopic localization of penicillinase in Bacillus licheniformis.

Penicillinase was localized in log-phase cells of Bacillus licheniformis 749/C by labeling with ferritin-anti-penicillinase immunoglobulin G conjugate. Mildly fixed homogenized cells, isolated subcellular fractions, and frozen thin sections were labeled. The label was distributed in discrete patches in the cell envelope. The patches extended from the inside part of the membrane to the outside part of the wall. The inside part of the membrane was labeled more extensively than the outside part. The cytoplasm also bound some ferritin-immunoglobulin G conjugate. Immunoelectrophoresis and biochemical assay of cytosol material suggest that the cytoplasmic antigenic sites are a protease-sensitive form of penicillinase.