Cell envelope composition of Escherichia coli K12: a comparison of the cell poles and the lateral wall

Minicells as a Damage Disposal Mechanism in Escherichia coli

Bacteria have the ability to produce minicells, or small spherical versions of themselves that lack chromosomal DNA and are unable to replicate. A minicell can constitute as much as 20% of the cell’s volume. Although molecular biology and biotechnology have used minicells as laboratory tools for several decades, it is still puzzling that bacteria should produce such costly but potentially nonfunctional structures. Here, we show that bacteria gain a benefit by producing minicells and using them as a mechanism to eliminate damaged or oxidated proteins. The elimination allows the bacteria to tolerate higher levels of stress, such as increasing levels of streptomycin. If this mechanism extends from streptomycin to other antibiotics, minicell production could be an overlooked pathway that bacteria are using to resist antimicrobials.

Many bacteria produce small, spherical minicells that lack chromosomal DNA and therefore are unable to proliferate. Although minicells have been used extensively by researchers as a molecular tool, nothing is known about why bacteria produce them. Here, we show that minicells help Escherichia coli cells to rid themselves of damaged proteins induced by antibiotic stress. By comparing the survival and growth rates of wild-type strains with the E. coliΔminC mutant, which produces excess minicells, we found that the mutant was more resistant to streptomycin. To determine the effects of producing minicells at the single-cell level, we also tracked the growth of ΔminC lineages by microscopy. We were able to show that the mutant increased the production of minicells in response to a higher level of the antibiotic. When we compared two sister cells, in which one produced minicells and the other did not, the daughters of the former had a shorter doubling time at this higher antibiotic level. Additionally, we found that minicells were more likely produced at the mother’s old pole, which is known to accumulate more aggregates. More importantly, by using a fluorescent IbpA chaperone to tag damage aggregates, we found that polar aggregates were contained by and ejected with the minicells produced by the mother bacterium. These results demonstrate for the first time the benefit to bacteria for producing minicells.

IMPORTANCE Bacteria have the ability to produce minicells, or small spherical versions of themselves that lack chromosomal DNA and are unable to replicate. A minicell can constitute as much as 20% of the cell’s volume. Although molecular biology and biotechnology have used minicells as laboratory tools for several decades, it is still puzzling that bacteria should produce such costly but potentially nonfunctional structures. Here, we show that bacteria gain a benefit by producing minicells and using them as a mechanism to eliminate damaged or oxidated proteins. The elimination allows the bacteria to tolerate higher levels of stress, such as increasing levels of streptomycin. If this mechanism extends from streptomycin to other antibiotics, minicell production could be an overlooked pathway that bacteria are using to resist antimicrobials.

Localization of anionic phospholipids in Escherichia coli cells

Cardiolipin (CL) is an anionic phospholipid with a characteristically large curvature and is of growing interest for two primary reasons: (i) it binds to and regulates many peripheral membrane proteins in bacteria and mitochondria, and (ii) it is distributed asymmetrically in rod-shaped cells and is concentrated at the poles and division septum. Despite the growing number of studies of CL, its function in bacteria remains unknown. 10-N-Nonyl acridine orange (NAO) is widely used to image CL in bacteria and mitochondria, as its interaction with CL is reported to produce a characteristic red-shifted fluorescence emission. Using a suite of biophysical techniques, we quantitatively studied the interaction of NAO with anionic phospholipids under physiologically relevant conditions. We found that NAO is promiscuous in its binding and has photophysical properties that are largely insensitive to the structure of diverse anionic phospholipids to which it binds. Being unable to rely solely on NAO to characterize the localization of CL in Escherichia coli cells, we instead used quantitative fluorescence microscopy, mass spectrometry, and mutants deficient in specific classes of anionic phospholipids. We found CL and phosphatidylglycerol (PG) concentrated in the polar regions of E. coli cell membranes; depletion of CL by genetic approaches increased the concentration of PG at the poles. Previous studies suggested that some CL-binding proteins also have a high affinity for PG and display a pattern of cellular localization that is not influenced by depletion of CL. Framed within the context of these previous experiments, our results suggest that PG may play an essential role in bacterial physiology by maintaining the anionic character of polar membranes.

Escherichia coli minicell membranes are enriched in cardiolipin

The phospholipid composition of Escherichia coli minicells has been studied as a model for the cell division site. Minicells appeared to be enriched in cardiolipin at the expense of phosphatidylglycerol. Mass spectrometry showed no differences between the gross acyl chain compositions of minicells and wild-type cells.

Hypothesis: membrane domains and hyperstructures control bacterial division

The mechanism responsible for creating the division site in the right place at the right time in bacteria is unknown. It has been attributed to the formation of proteolipid domains in the cytoplasmic membrane surrounding the nucleoids. We interpret the growing evidence for this hypothesis by invoking hyperstructures, which exist at a level of organization intermediate between macromolecules and genes. Non-equilibrium hyperstructures comprise the genes, mRNA proteins and lipids required for a particular function such as cell division, and assemble and disassemble according to the needs of the cell.

Division-associated changes in membrane viscosity of Escherichia coli

Septum formation is initiated by the FtsZ ring assembly in the middle of rod-shape bacteria. The mechanism which determines the division site in the membrane and makes it recognizable by FtsZ is still unknown. We have recently demonstrated that the putative division membrane domains can be visualized by a fluorescent membrane probe (Fishov and Woldring, Mol. Microbiol., 1999) and that these domains can be dissipated by interrupting the process of coupled transcription and translation of proteins (Binenbaum et al., Mol. Microbiol., 1999). Here, we examined the membrane dynamics of Escherichia coli during division and after a reversible division arrest. Anisotropy of DPH fluorescence, used as an indicator of membrane dynamics (viscosity), correlated with the rate of division in synchronous cells. It decreased during filamentation caused by drugs or by temperature, but not in the ftsZ mutant and when DNA replication was blocked by nalidixic acid. Based on previous data, we incline to interpret these results as reflecting formation and dissipation of putative membrane domains marking the division sites; domains are formed by partitioning nucleoids and dissipate while used for constriction or after the nucleoids have been segregated too far in a filament.

Visualization of membrane domains in Escherichia coli

Bacterial membrane and nucleoids were stained concurrently by the lipophilic styryl dye FM 4-64 [N-(3-triethylammoniumpropyl)-4-(6-(4-(diethylamino)phenyl) hexatrienyl)pyridinium dibromide] and 4',6-diamidino-2-phenylindole (DAPI), respectively, and studied using fluorescence microscopy imaging. Observation of plasmolysed cells indicated that FM 4-64 stained the inner membrane preferentially. In live Escherichia coli pbpB cells and filaments, prepared on wet agar slabs, an FM 4-64 staining pattern developed in the form of dark bands. In dividing cells, the bands occurred mainly at the constriction sites and, in filaments, between partitioning nucleoids. The FM 4-64 pattern of dark bands in filaments was abolished after inhibiting protein synthesis with chloramphenicol. It is proposed that the staining patterns reflect putative membrane domains formed by DNA-membrane interactions and have functional implications in cell division.

Hypothesis: chromosome separation in Escherichia coli involves autocatalytic gene expression, transertion and membrane-domain formation

To explain how daughter chromosomes are separated into discrete nucleoids and why chromosomes are partitioned with pole preferences, I propose that differential gene expression occurs during DNA replication in Escherichia coli. This differential gene expression means that the daughter chromosomes have different patterns of gene expression and that cell division is not a simple process of binary fission. Differential gene expression arises from autocatalytic gene expression and creates a separate proteolipid domain around each developing chromosome via the coupled transcription-translation-insertion of proteins into membranes (transertion). As these domains are immiscible, daughter chromosomes are simultaneously replicated and separated into discrete nucleoids. I also propose that the partitioning relationship between chromosome age and cell age arises because the poles of cells have a proteolipid composition that favours transertion from one nucleoid rather than from the other. This hypothesis forms part of an ensemble of related hypotheses which attempt to explain cell division, differentiation and wall growth in bacteria in terms of the physical properties and interactions of the principal constituents of cells.

Phospholipid domains determine the spatial organization of the Escherichia coli cell cycle: the membrane tectonics model

Escherichia coli normally divides at its equator between segregated nucleoids. Such division is inhibited during perturbations of chromosome replication (even in the absence of inducible division inhibitors); eventually, division resumes at sites which are not at this equator. Escherichia coli will also divide at its poles to generate minicells following overproduction of the FtsZ or MinE proteins. The mechanisms underlying the division inhibition and the positioning of the division sites are unknown. In the membrane tectonics model, I propose that the formation of phospholipid domains within the cytoplasmic membrane positions division sites. The particular phospholipid composition of a domain attracts particular proteins and determines their activity; conversely, particular proteins change the composition of domains. Principally via such proteins, the interaction of the chromosome with the membrane creates a chromosomal domain. The development of chromosomal domains during replication and nucleoid formation contributes to the formation and positioning of a septal domain between them. During septation (cell division), this septal domain matures into a polar domain. Each domain attracts and activates different enzymes. The septal domain attracts and activates enzymes necessary for septation. Preventing the formation of the septal domain by preventing chromosome replication prevents normal division. Altering the composition of the polar domain may allow septation enzymes to function there and generate minicells. A corollary of the model explains how the formation of an origin domain by the attachment of hemi-methylated origin DNA to the membrane may underlie the creation and migration of structures within the envelope, the periseptal annuli.

Synthesis of the cell surface during the division cycle of rod-shaped, gram-negative bacteria

When the growth of the gram-negative bacterial cell wall is considered in relation to the synthesis of the other components of the cell, a new understanding of the pattern of wall synthesis emerges. Rather than a switch in synthesis between the side wall and pole, there is a partitioning of synthesis such that the volume of the cell increases exponentially and thus perfectly encloses the exponentially increasing cytoplasm. This allows the density of the cell to remain constant during the division cycle. This model is explored at both the cellular and molecular levels to give a unified description of wall synthesis which has the following components: (i) there is no demonstrable turnover of peptidoglycan during cell growth, (ii) the side wall grows by diffuse intercalation, (iii) pole synthesis starts by some mechanism and is preferentially synthesized compared with side wall, and (iv) the combined side wall and pole syntheses enclose the newly synthesized cytoplasm at a constant cell density. The central role of the surface stress model in wall growth is distinguished from, and preferred to, models that propose cell-cycle-specific signals as triggers of changes in the rate of wall synthesis. The actual rate of wall synthesis during the division cycle is neither exponential nor linear, but is close to exponential when compared with protein synthesis during the division cycle.

Murein chemistry of cell division in Escherichia coli

The length distribution of the glycan strands of murein has been analysed with a novel method in filamentous and spherical cells of Escherichia coli, as well as during septum formation and cell separation. A shift to the longer glycan strands was observed in the murein of furazlocillin-induced filaments. In contrast, shorter glycan strands were increased in the murein of mecillinam-induced spherical cells. During septum formation in a chain-forming envA mutant that is defective in the splitting process of the septum, a shift to the shorter glycan strands was detected that was not seen in wild type E. coli cells. It is concluded that septum-specific murein structures of rather short glycan strands are released during splitting of the septum. This intermediate material remains present in the septum of the envA mutant. The splitting process of the septum was investigated by analysing the murein during penicillin-induced bacteriolysis, which is known to take place by strictly localized murein degradation in the equatorial zone of the cell. No changes in the length distribution of the glycan strands could be detected during penicillin-induced lysis, with the exception of an increase in disaccharides, the shortest glycan strands possible. This is explained by the action of exo-muramidases progressively digesting glycan strands, leaving disaccharide units covalently linked to the remaining murein at the sites of murein cross-linkage. It is proposed that this "zipper-like" mechanism represents the normal cutting process of the septum during cell separation.

Isogenic variants of Escherichia coli with altered morphology have peptidoglycan with identical muropeptide composition

The peptidoglycan compositions of three isogenic morphological mutants of Escherichia coli were determined by high-pressure liquid chromatography analysis. The muropeptide compositions of the peptidoglycan of these mutants were the same, indicating that the shape of E. coli is not (solely) determined by the chemical composition of the peptidoglycan. Furthermore, it appeared that the muropeptide composition of the peptidoglycan was not affected by growth temperature.

Spatial dependence of stress distribution for rod-shaped bacteria

The stress distribution in the cylindrical portion of the cell envelope of a rod-shaped bacterial cell was compared with that at its polar ends. Using a symmetry argument it is shown that the critical internal pressure for the initiation of yielding of the envelope material has a non-uniform distribution and is significantly higher for the polar regions.

Unequal distribution of penicillin-binding proteins among inner membrane vesicles of Escherichia coli

Escherichia coli penicillin-binding proteins (PBPs) were associated only with inner membrane vesicles when separated on 30 to 65% or 19 to 49% (wt/wt) sucrose gradients. Fractionation of vesicles through the low-density gradient revealed at least two classes of PBP-inner membrane associations. The first class consisted of PBPs 1 through 4, and the second class consisted of PBPs 5 through 8. These classes were distinguished by the density of vesicles with which they were associated; class 1 PBPs migrated with vesicles of higher density than did class 2 PBPs. Such combinations suggest that PBPs are nonrandomly distributed within the inner membrane, implying potential functional relationships among the PBPs themselves and with particular membrane domains. In addition, in cell lysates and in vesicle fractions, a 60,000-dalton aztreonam-insensitive PBP or protein fragment was observed which could potentially be confused with PBP3.

Effect of growth rate and cell shape on the peptidoglycan composition in Escherichia coli

The muropeptide composition of peptidoglycan from Escherichia coli W7 cultivated at different growth rates in chemostat cultures was compared by using high-pressure liquid chromatography. At a low growth rate (D = 0.1 h-1), about 40% more covalently bound lipoprotein and at least twofold more diaminopimelyl-diaminopimelic acid cross-bridges were found than at a high growth rate (D = 0.8 h-1). The total degree of cross-linkage was only slightly increased, and the fraction of trimeric muropeptides and the average length of the glycan chains were not changed significantly. Analysis of the peptidoglycan from a morphological variant strain of W7 revealed that the altered peptidoglycan composition in slowly growing W7 cells was not correlated with the observation that these cells, due to their decreased cell length, were relatively enriched in polar material. In fact, our results suggested that peptidoglycan forming cell poles is chemically identical to that forming lateral wall.

Biochemical characterization of HgCl2-inducible polypeptides encoded by the mer operon of plasmid R100

Minicells carrying the subcloned mer operon from plasmid R100 were pulse-labeled with [35S]methionine, and the labeled polypeptides were analyzed at various subsequent times by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The Hg(II) reductase monomer encoded by plasmid R100 occurred as two proteins of 69 and 66 kilodaltons (kd). The minor 66-kd protein is a modified form of the 69-kd protein. This modification occurs in vivo. Both of these mer proteins are found in the soluble fraction of the cell; however, the 66-kd protein appears to have a slight affinity for the cellular envelope. Both the 69- and 66-kd mer proteins have pI values greater (pI = 5.8) than that reported (pI = 5.3) for the analogous monomer encoded by plasmid R831. The 15.1- and 14-kd mer proteins are localized in the inner membrane and are probably elements of the mer-determined Hg(II) uptake system. These two mer membrane proteins, which are antigenically unrelated to the Hg(II) reductase monomer, are quite basic (pI values greater than 7.8). The 12-kd mer protein is also a basic polypeptide that is present in the soluble fraction of the cell. Unlike the two membrane-bound mer proteins, the 12-kd mer protein is processed from a 13-kd precursor.

Activity of N-acetylmuramoyl-L-alanine amidase in phospholipidic environments

A purified preparation of N-acetylmuramoyl-L-alanine amidase (EC 3.5.1.28), a murein hydrolase from Escherichia coli, was found to lose its activity during incubation in the presence of bacterial phospholipid suspensions. Whether it was co-dispersed with the phospholipids or added to sonicated phospholipid suspension, the enzyme was inhibited (or inactivated) from the first minutes of incubation at 37 degree C. As phosphatidylglycerol/cardiolipin ratio of the phospholipid suspension as increased (all other things being equal), a further decrease of amidase activity was observed. The highest losses of activity were found after co-dispersion of the enzyme and the substrate together with the phospholipids, the resulting suspension being formed of larger multilayered vesicles, as revealed by electron microscopy. In these conditions, the effect on enzyme activity was only partially accounted for by the proportion of the enzyme that was entrapped in the vesicles. The entrapment capacity of the enzyme (using a 35S-labelled enzyme preparation) and of the substrate (3H-labelled) by the multilamellar phospholipidic vesicles did not significantly change as a function of their relative content of phosphatidylglycerol and cardiolipin. The possible physiological meaning of the results is discussed is connection with our previous data and with other works related to membranous phospholipid distribution and to septum formation control in bacteria.

Alterations in peptidoglycan chemical composition associated with rod-to-sphere transition in a conditional mutant of Klebsiella pneumoniae

Klebsiella pneumoniae Mir M7 is a spontaneous parentless morphology mutant which grows as cocci at pH 7 and as rods at pH 5.8. This strain has been characterized as defective in lateral wall formation (at pH7). Data suggest that the cell wall is mainly made up of poles of the rods (G. Satta, R. Fontana, P. Canepari, and G. Botta, J. Bacteriol. 137:727--734, 1979). In this work the isolation and the biochemical properties of the peptidoglycan of both Mir M7 rods and cocci and a nonconditional rod-shaped Mir M7 revertant (strain Mir A12) are described. The peptidoglycan of Mir M7 (both rods and cocci) and Mir A12 strains carried covalently bound proteins which could be easily removed by pronase treatment in Mir M7 rods and Mir A12 cells, but not in Mir M7 round cells. However, when the sodium dodecyl sulfate-insoluble residues of Mir M7 cocci were pretreated with ethylenediaminetetraacetic acid (EDTA), pronase digestion removed the covalently bound proteins, and pure peptidoglycan was obtained. EDTA treatment of the rigid layer of Mir M7 cocci removed amounts of Mg2+ and Ca2+, which were 10- and 50-fold higher, respectively, than the amount liberated from the rigid layer of Mir M7 rods and Mir A12 cells. Amino acid composition was qualitatively similar in both strains, but Mir M7 cocci contained a higher amount of alanine and glucosamine. Mir M7 cocci contained approximately 50% less peptidoglycan than rods. Under electron microscopy, the rigid layer of the Mir M7 rods and Mir A12 cells appeared to be rod-shaped and their shape remained unchanged after EDTA and pronase treatment. On the contrary, the Mir M7 cocci rigid layer appeared to be round, and after EDTA treatment it collapsed and lost any definite morphology. In spite of these alterations, the peptidoglycan of Mir M7 cocci still appeared able to determine the shape of the cell and protect it from osmotic shock and mechanical damages. The accumluation of divalent cations appeared necessary for the peptidoglycan to acquire sufficient rigidity for shape determination and cell protection. We concluded that the coccal shape in Mir M7 cells is not due to loss of cell wall rigidity but is a consequence of the formation of a round peptidoglycan molecule. The possibility that the alterations found in the Mir M7 cocci rigid layer may reflect natural differences in the biochemical composition of the septa and lateral wall of normally shaped bacteria is discussed.

Phospholipid synthesis during the cell division cycle of Escherichia coli

Stepwise changes in the rate of phosphatidylethanolamine and phospholipid synthesis during the cell division cycle of Escherichia coli B/r were observed. The cell ages at the increases were found to be a function of the growth rate. At each growth rate, the increase occurred around the time new rounds of chromosome replication were inaugurated in the cycle.

Structural heterogeneity of the cytoplasmic and outer membranes of Escherichia coli

The cytoplasmic and outer membranes of gram-negative bacteria can be isolated from spheroplasts, and separated on sucrose density gradients. Lysis of spheroplasts causes extensive membrane fragmentation and since the characteristics of the fragments obtained by different lysis procedures need not be identical, the influence of the disruption method on membrane composition has been examined. Spheroplasts of Escherichia coli J5 were lysed by osmotic shock, which did not significantly separate the cytoplasmic and outer membranes, but resulted in mixed membrane vesicles. Lysis in the French press and by sonication caused extensive membrane fragmentation and separation. Sonication, however, also caused some fusion between fragments of the outer and the cytoplasmic membranes; this intermembrane fusion increased with sonication time. When the cytoplasmic and outer membranes were well separated and intermembrane fusion was minimal or absent, the cytoplasmic and outer membrane fragments were heterogeneous with respect to density and ovarll phospholipid, protein and lipopolysaccharide composition. In addition, cytoplasmic, but not outer, membrane fragments were also heterogeneous with respect to protein composition. It is concluded, therefore, that membrane fragments obtained from the cytoplasmic and outer membranes are heterogeneous independently of the lysis procedures used to obtain these fragments. Possible reasons for this heterogeneity are discussed.

Enzymes synthesizing and hydrolyzing murein in Escherichia coli. Topographical distribution over the cell envelope

Envelopes from regions of the cell which in vivo show very little, if any, murein synthesis were isolated using the minicell-producing strain P678-54. Envelopes from minicells, representing in fact cell ends, were able to synthesize murein and to carry out transpeptidation in vitro; also all four murein hydrolase activities tested, carboxypeptidase, endopeptidase, amidase and transglycosylase, were found to be present. The specific activities of the murein synthesizing and degrading enzymes in envelopes derived from cell poles and from actively growing cells were similar. The topological distribution of murein-synthesizing enzymes and of murein hydrolases over the cell envelope is discussed.

Quantitative determination of N-acetylglucosamine residues at the non-reducing ends of peptidoglycan chains by enzymic attachment of [14C]-D-galactose

The ability of human milk galactosyltransferase to attach D-galactose residues quantitatively to the C-4 of N-acetylglucosamine moieties at the ends of oligosaccharides has been utilized for the specific labeling and quantitative determination of the chain length of the glycan moiety of the bacterial cell wall. The average polysaccharide chain length of the soluble, uncrosslinked peptidoglycan secreted by Micrococcus luteus cells on incubation with penicillin G was studied with this technique and found to be approximately 70 hexosamines long. Furthermore, the peptidoglycan chain length of Escherichia coli sacculi of different cell shapes and dimensions was determined both in rod-shaped cells and in filaments induced by temperature shift of a division mutant or by addition of cephalexin or nalidixic acid. The average chain length found in most of these sacculi was between 70 and 100 hexosamines long. Small spherical 'mini' cells had chain lengths similar to those of the isogenic rod-like cells.

Septation deficiency and phosphilipid perturbation in Escherichia coli genetically constitutive for the beta oxidation pathway

Mutants of Escherichia coli defective in the regulation of the fatty acids beta oxidation pathway show an ultrastructural deficiency in septum formation at high growth rate. Several independent pairs of parent and mutant strains have been analyzed biochemically. Each parent strain displays a well-defined pattern of cellular phospholipids, which varies with the growth conditions. High ratios of phosphatidylglycerol to cardiolipin characterize fast-growth conditions. None of the mutant strains, although they grow in mass nearly as rapidly as their respective parents, can reach these high ratios. The beta oxidation pathway regulatory mutation leads to an increased turnover of the glycerol moieties of these phospholipids in the inner as well as in the outer cell membrane.

A novel electrophoretic fractionation of Escherichia coli envelopes

Particulate fractions of Escherichia coli have been submitted to electrophoretic fractionation in a buffer stabilized by sucrose gradient. Inner membrane and outer membrane were readily resolved. A combination of electrophoresis, fractional centrifugation and gel filtration can remove remaining contamination by ribosomes and cytoplasm. The presence of particles containing no phospholipids was detected after differential centrifugation. The nature of this fraction is unknown. The inner membrane exhibited heterogeneity on electrophoresis.

Isolation of a small-cell mutant in the blue-green bacterium Agmenellum quadruplicatum

A new type of high-temperature conditional cell division mutant has been isolated in Agmenellum quadruplicatum strain BG1 in which the process of cell division is uncoupled from that of growth at 39 C. This mutant produces abnormally small cells under conditions of nutrient limitation and forms multinucleoid filaments under normal growth conditions.

Very stable prokaryotic messenger RNA in chromosomeless Escherichia coli minicells

E. coli minicells lack DNA, yet they make protein, the synthesis of which is sensitive to chloramphenicol but insensitive to rifamycin. This protein is coded for by very stable cellular mRNA with an estimated half-life of 40-80 min. In an R factor-containing minicell, two very different species of mRNA are observed: (i) R factor-specific mRNA with a short half-life whose synthesis is rifamycin-sensitive and (ii) cellular mRNA with a long half-life whose synthesis is rifamycin-insensitive. These findings indicate that minicells contain normal degradative mechanisms for mRNA and point out the existence of a unique class of very stable cellular mRNA. Greater than 80% of the rifamycin-insensitive protein synthesized goes into the outer minicell membrane. Relatively stable mRNA, half-life 5.5-11.5 min, for outer membrane protein in whole cells has been reported [Hirashima et al. (1973) J. Mol. Biol. 79, 373-389]. The stability of minicell mRNA is significantly greater. This and other observations suggest that there are two functional species of mRNA for outer membrane protein perhaps in different sites in the cell. Furthermore, these studies suggest that a class of cellular proteins is synthesized in bacteria without concomitant transcription and in the absence of association with chromosomal DNA.