escherichia coli an outer membrane protein (ompa) of ... · pdf filean outer membrane protein...

THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1986 by The American Society of Biological Chemists, Inc.

VOl. 261, No. 24, Issue of August 25, pp. 11355-11361,1986 Printed in U.S.A.

An Outer Membrane Protein (OmpA) of Escherichia coli K-12 Undergoes a Conformational Change during Export*

(Received for publication, February 20, 1986)

Roland FreudlS, Heinz SchwarzS, York-Dieter Stierhofg, Konrad GamonSll, Ingrid HindennachS, and Ulf HenningS From the $Max-Planck-Institut fur Biologie, Corrensstrasse 38 and the §Hygiene-Znstitut, Uniuersitat Tubingen, Silcherstrasse 7, 0-7400 Tubingen, Federal Republic of Germany

Pulse-chase experiments were performed to follow the export of the Escherichia coli outer membrane protein OmpA. Besides the pro-OmpA protein, which carries a 21-residue signal sequence, three species of ompA gene products were distinguishable. One probably represented an incomplete nascent chain, another the mature protein in the outer membrane, and the third, designated imp-OmpA (immature processed), a protein which was already processed but apparently was still associated with the plasma membrane. The pro- and imp-OmpA proteins could be characterized more fully by using a strain overproducing the ompA gene products; pro- and imp-OmpA accumulated in large amounts. It could be shown that the imp- and pro-OmpA proteins differ markedly in conformation from the OmpA protein. The imp-OmpA, but not the pro-OmpA, underwent a conformational change and gained phage receptor activity upon addition of lipopolysaccharide. Utilizing a difference in detergent solubility between the two polypeptides and employing immunoelectron microscopy, it could be demonstrated that the pro-OmpA protein accumulated in the cytoplasm while the imp-OmpA was present in the periplasmic space. The results suggest that the pro-OmpA protein, bound to the plasma membrane, is processed, and the resulting imp-OmpA, still associated with the plasma membrane, recognizes the lipid A moiety of the lipopolysaccharide. The resulting conformational change may then force the protein into the outer membrane.

The 325-residue OmpA protein (1) is one of the abundant proteins of the Escherichia coli outer membrane (2). It is synthesized as a precursor with a 21-residue signal sequence (3, 4). A number of experiments have indicated that in addition to the information within the signal sequence, information required for export and localization of E. coli proteins exists within the mature polypeptide. It has been suggested that this information consists of short amino acid sequences, called export or sorting signals, and that outer membrane proteins possess a common sorting signal (5 , 6). In a study designed to investigate the nature of this information in the OmpA protein, we had previously concluded that such a common sorting signal may not exist and that the information

* Financial support for this work was received from the Fonds der Chemischen Industrie. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

ll Present address: Henkel KG, Aft. Biotechnologie, Dusseldorf, Federal Republic of Germany.

for sorting could be contained within the conformation of the protein (7).

A protein destined for the outer membrane has to avoid incorporation into the plasma membrane. We have demonstrated previously that completed pro-OmpA protein is associated with the plasma membrane (8). One could, therefore, consider the possibility that the pro-OmpA protein differs in conformation from that of the OmpA protein, the latter being and the former not being “outer membrane compatible.” A search for such a conformational change revealed that a newly synthesized OmpA protein exists which has already lost the signal sequence, which still appears to be associated with the plasma membrane, and which differs markedly in conformation from the protein located in the outer membrane.

EXPERIMENTAL PROCEDURES

Bacterial Strains and Growth Conditions-The pulse-chase experiments were performed with the E. coli K-12 strain P400.6 (thi argE proA thr leu mtl xyl aragaZK lacy ompC rpsL supE) (9). Strain UH203 ( lac supF ompA recA proA or B rpsL/F’ ladQ lacZ M15 pro AB+) (10) harbors plasmid pRD87 (7) which is a derivative of pUC9 (11). The chromosomally ompA wild-type strain employed is an ompA’ derivative of UH203 not containing the plasmid. Plasmid pRD87 carries the ompA gene under the control of the lac regulatory elements of pUC9. The medium for the pulse-chase experiment was M9 (12) supplemented with the required L-amino acids (4Opg/ml), thiamine (1 pg/ml), and glucose (0.4%). For all other experiments cells were grown at 37 “C in L-broth (12). To obtain envelopes from strain UH203 with the plasmid, 10 ml of cells, grown overnight in L-broth containing 0.4% glucose and ampicillin (40 pg/ml) were centrifuged, resuspended in 20 ml of broth supplemented with IPTG’ (1 mM) and ampicillin, and shaken for 6 h at 37 “C.

Pulse Labeling, Sucrose Gradient Centrifugation, and Immunopre- cipitation-Cells growing exponentially at 25 “C were concentrated 10-fold, pulse labeled with 50 pCi/ml [35S]methionine (1000 Ci/mmol, Amersham Corp.), and chased at 25 “C with a final concentration of 20 mM L-methionine as described (13). The chase was ended either by pipetting 0.4-ml samples into glass tubes held at -78 ‘C in COz- methanol or, in the case of 10-ml samples (for membrane separation experiments), by pouring into an equal volume of crushed ice. For the preparation of cell envelopes the frozen pulse-chased samples (5 X lo9 cells/ml) were thawed at 0 “C, diluted with ice-cold 10 mM Tris- HCl, pH 7.5 (Tris buffer), to a final volume of 1 ml and lysed by sonication. Envelopes were obtained by centrifugation at 100,000 X g for 10 h at 4 “C. For the separation of inner and outer membranes the cells were concentrated to 5 X 10”/ml in Tris buffer, and 1 ml was broken in a French press at 4 “C. Intact cells were removed by centrifugation at 4,500 X g for 5 min, and envelopes were spun down for 10 h at 100,000 x g. Plasma and outer membranes were separated on a sucrose step gradient as described (8). Immunoprecipitation was carried out essentially as detailed previously (8) with the following modifications. Membrane pellets, derived from at most lo9 cells were

’ The abbreviations used are: IPTG, isopropylthiogalactoside; LPS, lipopolysaccharide; SDS, sodium dodecyl sulfate; Hepes, N-2-hydrox- yethylpiperazine-N’-2-ethanesulfonic acid.

11355

11356 Protein Export in E. coli

resuspended in 0.6 ml of Tris buffer containing 2% sodium dodecyl sulfate and either boiled for 5 min or heated to 50 'C for 1 h. Solubilized membrane samples (i.e. Sarkosyl-soluble inner or outer membrane fractions obtained following sucrose gradient centrifugation) were extracted in the same manner, after addition of sodium dodecyl sulfate to a final concentration of 2%. To precipitate OmpA protein, 1 ml of Tris buffer and 0.1 ml of OmpA-specific antiserum (14) were added. Immune complexes formed after overnight incubation at 4 "C were reacted with 0.04 ml of goat anti-rabbit immuno- globulin (23 mg of IgG/ml) for another 15 h. The precipitates were washed twice in saline and resuspended in 0.03 ml of electrophoresis sample buffer.

Sarkosyl Extraction-Envelopes from pulse-labeled cells were sus- pended in 0.5 ml of Tris buffer containing 0.5% Sarkosyl NL97 (Geigy), shaken for 1 h at 4 "C, and centrifuged at 100,000 X g for 1 h to separate soluble and insoluble fractions. The envelope fraction from 2 ml of cells, induced for OmpA synthesis, was resuspended in 0.1 ml of 50 mM citrate-disodium phosphate buffer, pH 6, containing 1.5% Sarkosyl and 5 mM MgCIP. Following incubation for 30 min at room temperature, the mixture was spun for 15 min in an Eppendorf centrifuge, and the extraction was repeated. Protein from the combined supernatants was precipitated with acetone (final concentration 90%).

Digestion with Trypsin-Envelopes from 4 ml of cells were sus- pended in 0.5 ml of 30 mM Tris-HCI, pH 8, and trypsin (15 pg) was added. After incubation for 2 h a t 37 'C the envelopes were recovered by centrifugation (15 min in an Eppendorf centrifuge).

SDS-Polyacrylamide Gel Electrophoresis-This was performed as described previously (8). The gels were fluorographed with EN- HANCE (New England Nuclear) and exposed on preflashed films (15). Apparent molecular weights of proteins were determined using the standards (Combitek, Boehringer Mannheim) phosphorylase b (97,400), glutamate dehydrogenase (55,400), lactate dehydrogenase (36,500), and trypsin inhibitor (20,100). Using another electrophoretic system, the molecular weights of the two forms of the OmpA protein were previously estimated to be 28,000 and 33,000 (16-18). With the electrophoresis system we are presently using the molecular weights of the heat-modified and unmodified forms of OmpA were estimated to be 36,000 and 31,000, respectively (Ref. 19 and this study).

Phage Receptor Actiuity-Proteins to be tested for their activity as a receptor for phage K3 were combined with LPS essentially as described (20). Sarkosyl-soluble (the acetone precipitate, see above) and insoluble material from envelopes obtained from 20 ml of induced cells (or 40 ml when the chromosomally ompA wild type strain was used) were dissolved in 0.4 ml of 10 mM Hepes, pH 7.4, containing 2.5 mM EDTA and 8 M urea, and 500 pg of LPS, dissolved in the same buffer, was added. The mixture was then dialyzed against this buffer without urea. None of the proteins precipitated. Controls, protein, or LPS alone were treated in the same way. Phage inactivation was measured by incubating 3.5 x IO3 plaque-forming units in 2 ml of L-broth containing 20 mM MgC12 and varying amounts of dialyzed protein (see Fig. 6) for 30 min at 37 "C; surviving phage were titrated on strain P400.6.

Electron Microscopy and Immunolabeling-For better differentia- tion of inner and outer membranes cells were plasmolysed with 25% sucrose before fixation with 2% formaldehyde and 0.05% glutaraldehyde. For immunolabeling of ultrathin plastic sections, cells were embedded in agarose, dehydrated with ethanol, and embedded in Lowicryl K4M at low temperature (21). To prepare cryosections, blocks of cells fixed identically were soaked in 2.1 M sucrose and sectioned with a Reichert FC-4/Ultracut cryoultramicrotome. For conventional Epon embedding, cells were fixed with 2.5% glutaraldehyde and, after embedding in agar, postfixed with 1% osmium tetroxide in 0.1 M sodium cacodylate. Samples were dehydrated with ethanol a t room temperature and after infiltration with Epon 812 polymerized at 60 "C. Before immunolabeling, Lowicryl and cryosections were treated with 0.5% bovine serum albumin, 0.2% gelatin in phosphate-buffered saline to block nonspecific binding sites (22). After incubation with rabbit OmpA antiserum the sections were washed extensively and labeled with protein A-gold complexes. For this purpose colloidal gold with a diameter of 7-8 nm was prepared using citrate and tannic acid as described (23). The labeled cryosections were postfixed with 0.1% osmium tetroxide, stained with 0.5% uranyl acetate, and embedded in Epon (24). Such samples were also cut perpendicular to the original surface of the cryosection. This method allows a precise localization of gold particles on the sections and demonstrates that labeling occurs exclusively at the surface of the cryosections.

RESULTS

Assessment of a Conformational Change of the OmpA Pro- tein-A characteristic property of OmpA is its so-called heat modifiability. In SDS-polyacrylamide gels its electrophoretic mobility depends on the temperature of solubilization (16- 18). When boiled in the presence of SDS, the protein migrates as a 36-kDa species; when held at 60 "C or below, it migrates as a 31-kDa species (earlier these molecular weights had been determined as 33 and 28 kDa, respectively; see "Experimental Procedures"). It has been suggested that this behavior reflects a conformational change and t h a t the faster moving polypeptide binds an excess of SDS, because the P-structure of the protein is not unfolded below 60 "C (25). It has been shown that the isolated protein can be converted from the 36-kDa form to the 31-kDa form by the addition of LPS (20). There- fore, the possibility existed that the anomalous migration of OmpA with an apparent molecular weight of 31,000 is caused by bound LPS rather than being due to a certain conformation. We labeled cells with ["P]orthophosphate and subjected electrophoretograms of their cell envelopes to autoradiogra- phy. Both phospholipid and LPS were heavily labeled, but no trace of radioactivity was found associated with either form of the OmpA protein (data not shown). Thus, the interpreta- tion of the heat modifiability as a conformational change is correct.

Properties of Newly Synthesized Species of OmpA-The OmpA protein, like other outer membrane proteins, remains largely insoluble when cell envelopes are extracted with lauryl sarcosinate (26). However, we find that approximately 10- 20% of the protein is soluble in this detergent. The reason for this is unknown.

Proteins labeled during a pulse-chase experiment were tested for their heat modifiability and solubility i n Sarkosyl. Fig. L4 shows the appearance of the OmpA protein in the insoluble fraction; it exhibited the typical heat modifiability. In the soluble fraction (Fig. 1B) at least 4 species of the protein are discernible. Only one of them was heat modifiable, and it did not ,disappear from this fraction during the chase.

l0O0C A 50' C

0"" -36 - ""-31 1 2 3 4 5 6 7 1 2 3 4 5 6 7

B

c- -&&-."-*"" -36 .- . . a. - 0 - 0 0 - 3 1

1 2 3 4 5 6 7 1 2 3 4 5 6 7

FIG. 1. Autoradiogram resulting from a pulse-chase experiment. Cells were labeled with ["S]methionine for 10 s at 25 "C and chased with nonradioactive methionine. Following extraction of cell envelopes with Sarkosyl, the insoluble (A) and soluble ( B ) fractions were treated with SDS at 100 or 50 "C. Both extracts were treated with anti-OmpA serum, and the resultingprecipitates were solubilized with sample buffer a t either 100 or 50 "C and subjected to SDS- polyacrylamide gel electrophoresis. Chase times in lanes 1-7 were 10 s, 30 s, 50 s, 80 s, 160 s, and 30 min. Asterisk, position of the pro- OmpA protein; arrowhead, the 30-kDa species (see text); numbers, molecular weights in kDa.

Protein Export in E. coli 11357

This latter species undoubtedly represents the fraction of the mature protein, mentioned above, which remains Sarkosyl soluble. The pro-OmpA protein was not heat modifiable nor was a fraction of the protein which has already been processed. Both appear to be chased into the modifiable form. Evidently, this processed polypeptide has not yet assumed the conformation of the modifiable one. We, therefore, designate it imp- OmpA (for h m a t u r e processed OmpA). I t should be noted that, while the imp-OmpA protein is clearly chased into the mature protein, the precursor-product relationship between the pro- and imp-OmpA protein is not obvious. The amounts of the two proteins appearing during the chase are under- standable, however, if the conformational change is slower than the processing reaction. At the earliest chase times still another nonmodifiable protein (about 30 kDa), which is slightly smaller than the mature OmpA protein, can be seen (arrow in Fig. 1B). It appeared earlier than the processed protein and could represent an intermediate nascent chain (see "Discussion").

The imp-OmpA was completely soluble in Sarkosyl while most of the OmpA protein was not. This suggested that the imp-OmpA may be present in the plasma membrane since proteins of this membrane are fairly selectively solubilized by the detergent (26). To test this possibility, cells were pulse labeled with ["S]methionine, chased for 40 s with methionine, and lysed in a French press. The membranes were separated by sucrose gradient centrifugation and fractions collected into six pools (of which pools I and I11 visibly contained outer and inner membranes, respectively). Each fraction was immuno- precipitated, and the resulting precipitates were solubilized in sample buffer a t 50 "C prior to SDS-polyacrylamide gel electrophoresis (Fig. 2). Three different species of OmpA protein were present in the inner membrane: pro-OmpA protein, imp- OmpA protein, and a small proportion of the 31-kDa species. The bulk of the latter species comigrated with the outer membrane. Only trace amounts of the imp-OmpA protein were detectable in the outer membrane fraction. This was most likely due to cross-contamination as was the presence

"pro-Omp A "Omp A 36

A b-Omp A 31

I I I r n I

2 4 6 8 10 12 14 16 18

F r a c t i o n number

FIG. 2. Sucrose gradient. Cells were pulse labeled with ["SI methionine a t 25 "C for 10 s and chased for 40 s with unlabeled methionine. A French press lysate was subjected to sucrose gradient centrifugation. Samples (5 pl) of each fraction were counted, pooled as indicated (I-VI), and reacted with antiserum. The precipitates were solubilized in sample buffer at 50 "C. No radioactive immuno- precipitates were detectable in pools IV-VI. Numbers at the autoradiogram, molecular weights in kDa.

of the 31-kDa polypeptide in the plasma membrane. These results leave several uncertainties. First, they did not

prove that the imp- and pro-OmpA proteins are in fact in the plasma membrane; for example, the proteins may have been periplasmic and may have become stuck to this membrane during the experimental procedures. This, a t least for the imp-OmpA protein, appears rather unlikely, as it is destined for the outer membrane and, therefore, one would expect any artificial association of this protein with a membrane to be with the outer rather than with the plasma membrane. In any event, it is clear that neither protein was located in the outer membrane. Second, and more important, we cannot be abso- lutely sure whether the imp-OmpA protein really represents an immature processed protein. Although it is difficult to imagine what else it might be we wished to characterize both the imp- and pro-OmpA proteins more fully.

Properties of the pro-OmpA and imp-OmpA Proteins-Bet- ter characterization of these two proteins using the system described above was virtually impossible due to the minute amounts of labeled products. However, we knew from earlier studies, concerning OmpA synthesis in minicells, that in such cells the protein is overproduced and that pro-OmpA is accumulated in large amounts (27, 28). We have, therefore, employed conditions under which production of the protein is greatly increased. In this system the cloned ompA gene (in plasmid pRD87, see "Experimental Procedures") is under control of t he kc operator and promoter. Fig. 3 compares the protein profiles of cell envelopes prepared from such cells induced for OmpA synthesis with those of cells which are chromosomally ompA wild type. In the former case both the pro-OmpA and the processed protein were present in large amounts. The behavior of these proteins in response to temperature is also shown in Fig. 3. As was observed in the pulse- chase experiment, the pro-OmpA protein was not heat modifiable. Only a small fraction of the processed polypeptide was heat modifiable. The amount of this fraction was the same as that present in cells which are chromosomally ompA wild type. This fraction, therefore, most likely represents OmpA protein properly assembled in the outer membrane while the bulk of the processed protein apparently consists of the imp- OmpA protein found during the pulse chase. As mentioned above, isolated OmpA protein (the 36-kDa species) can be renatured to assume the 31-kDa conformation by the addition of LPS. Addition of LPS to envelopes under conditions described under "Experimental Procedures'' converted most of the imp-OmpA to the heat-modifiable form while most of the precursor did not respond (Fig. 4). It seems that a small fraction of the precursor did become heat modifiable. We cannot exclude that this polypeptide represents still another form of the processed OmpA protein. In any event, the amount of this protein could not be increased by varying the LPS concentration.

r) =.O - pro-Omp A -0mpA 36

I - 0 -0mpA 31 ~-

- 0

5OoC 100°C 5OoC loO°C FIG. 3. Stained electrophoretogram of cell envelopes. Left

two lanes, envelopes from the chromosomally ompA wild-type strain; right two lanes, envelopes from strain UH203 carrying plasmid pRD87 after 4 h of induction of ompA expression. The envelopes were heated in sample buffer at the temperatures indicated. 36, 31, molecular weights in kDa of the respective proteins.


” ,-. - - - ”pro-Omp A - -0mpA 36

- -0mpA 31

5OoC 100°C 5OoC 100°C FIG. 4. Influence of LPS. A stained electrophoretogram is

shown. The cell envelopes were the same as those shown in Fig. 3 obtained from the OmpA overproducer. Right half, envelopes treated with LPS; left half, envelopes carried through the same procedure but without added LPS. The samples were heated in sample buffer at the temperatures indicated.

pro-Omp A. ”

Omp A* “ Omp A

- - “TF

1 2 3 4 5 6

FIG. 5. Treatment of cell envelopes with trypsin or Sarko- syl. Lunes 1-4, preparations from the OmpA overproducer (see Fig. 3); lanes 5, 6, preparations from the chromosomally ompA wild-type strain; lanes 3, 6, cell envelopes; lanes I, 2, insoluble and soluble fractions, respectively, obtained by treatment with Sarkosyl; [ones 4 , 5, envelopes after digestion with trypsin. TF, tryptic fragment of the OmpA protein.

Trypsin acting on cell envelopes removes a C02H-terminal part of the OmpA protein, whereas the NH2-terminal membrane moiety of the protein remains undigested (1, 29). Cell envelopes obtained from cells chromosomally ompA wild type and from cells accumulating imp- and pro-OmpA proteins were treated with trypsin. The results are shown in Fig. 5. Lanes 3 and 6 show the profiles of cell envelope preparations of the OmpA overproducer and the chromosomally ompA wild-type strain, respectively. Lanes 4 and 5 display the pat- terns of these envelopes after digestion with trypsin. The data are entirely consistent with those shown in Fig. 3. The amount of the tryptic fragment generated was identical in both cases. Apparently, only OmpA protein properly assembled in the outer membrane gave rise to the fragment, and the pro-OmpA and the imp-OmpA proteins were completely digested.

We have tested the solubilities of the different OmpA species in various detergents. It was found that at pH 6 accumulated imp-OmpA was almost completely solubilized in Sarkosyl while the precursor remained insoluble (Fig. 5, lanes 1 and 2). This behavior of the pro-OmpA is different from that of the precursor present in the pulse-chase experiment (Fig. 1). This is probably due to the fact that the accumulated precursor is in a different physical state than the one being chased into the mature protein (see below).

Phage Receptor Activity of pro-OmpA, imp-OmpA, and OmpA Proteins-Since the Sarkosyl extraction effected an efficient separation of the pro- and imp-OmpA proteins it was possible to test the activity of these proteins as phage recep- tors; isolated OmpA protein in combination with LPS inac- tivates phages which use the protein as a receptor (20). Cell envelopes from the strain which was chromosomally ompA wild type and from cells carrying plasmid pRD87, induced for OmpA synthesis, were extracted with Sarkosyl. Protein from the soluble fraction of the latter envelopes was precipitated

with acetone. Samples of all three preparations corresponding to equal amounts (as judged by stained electrophoretograms) of the OmpA, imp- and pro-OmpA proteins were combined with LPS. Fig. 6 shows that the OmpA and the imp-OmpA proteins exhibit the same phage-inactivating activity. The pro-OmpA protein is either much less active or, more likely, totally inactive as a phage receptor; the low level of activity of this preparation can probably be attributed to the presence of residual OmpA protein (Fig. 5).

Cellular Location of the Proteins-We have located the ompA products by electron microscopy of immuno-gold-labeled thin sections of cells (Fig. 7). As shown before (7), in chromosomally ompA wild-type cells the label was found associated only with the outer membrane. In cells overproducing ompA gene products the label was present in large clumps in the cytoplasm, in the outer membrane, and, in addition, in a dense layer in the periplasmic space (Fig. 7, D and E ) . It appeared likely that the latter protein represents the imp- OmpA and the cytoplasmic species the pro-OmpA protein, as it has generally been found that proteins, destined for export but which remain cytoplasmic (for example, those with an altered signal sequence), are not processed (6, 30). Proof of the above assumption was obtained by taking advantage of the different detergent solubilities of pro-OmpA and imp- OmpA. Cells induced for the expression of the ompA gene begin to lyse after about a 4-h incubation in the presence of IPTG. Such empty cells were extracted with Sarkosyl. Elec- tron microscopy revealed that the large clumps in the cytoplasm were not affected but that the dense layer between the inner and outer membranes had disappeared (Fig. 8). An electrophoretic analysis showed that, as with cell envelopes, the imp-0mpA was soluble and both OmpA and pro-OmpA proteins remained insoluble (data not shown, the electrophoretic pattern of the fractions was identical to that shown in lanes 1 and 2 of Fig. 5). Clearly, therefore, the accumulated imp-OmpA protein was present in the periplasm while the pro-OmpA protein had precipitated in the cytoplasm.

500

000

500

0

a 2 5 10 I I I

100 FIG. 6. Phage inactivation. Ordinate, survivingphage in plaque-

forming units per ml. Protein combined with LPS was incubated with phage K3 as detailed under “Experimental Procedures.” 0, A, Sar- kosyl-insoluble material from cells of the OmpA overproducer and the chromosomally ompA wild-type strains, i.e. pro-0mpA and OmpA proteins, respectively. 0, Sarkosyl-soluble material from the former cells representing the imp-0mpA protein.


FIG. 7. Localization of ompA gene products by immunoelectron microscopy. A, chromosomally ompA wild-type strain; B, strain UH203 carrying pRD87 after induction with IPTG for 5 h. A and B, Lowicryl K4M-meth- acrylate sections. C, the occurrence of the aggregated pro-OmpA protein (cf. Fig. 8) in the cytoplasm could best be demonstrated in unlabeled Epon sections of glutaraldehyde-osmium tetroxide fixed bacteria. D and E, cryosections of the strain overproducing the OmpA proteins cut perpendicular to the original immuno-gold-labeled surface of the sections; the arrowheads show the gold label in the periplasmic space. The bar represents 0.5 pm.

f r:

. ' . .. ' . I' , . . ., . ... . ..

:" . '. . .. ... .

4

r

DISCUSSION

At least four ompA gene products could be distinguished as a result of pulse-chase experiments. The first, visible only after the shortest chase times and exhibiting a molecular weight of about 30,000, may represent an intermediate nascent chain. Nascent intermediates have been observed during the synthesis of the maltose-binding protein of E. coli (31). The mRNA for this protein may form several stable stem- loop structures, and it has been suggested that some of them could account for the apparent translational pauses (32). The mRNA for the OmpA protein also has the potential to form extensive stem-loop structures (4). Codons corresponding to minor tRNAs, possibly causing translational pauses, do not exist in any of the ompA genes sequenced (33); hence a mRNA secondary structure may be responsible for the product observed. Alternatively or in addition, it is possible that the presumed halt in translation after reaching the size of 30 kDa is connected with the translocation of the protein across the plasma membrane. Indeed, Josefsson and Randall (34) have reported that processing of the pro-OmpA protein begins only after about 86% of the polypeptide has been synthesized. Depending on whether or not the 30-kDa protein still pos- sesses the signal sequence it would represent an OmpA chain with a length of about 82 or 78%, respectively, i.e. roughly the size a t which processing begins.

The next two intermediates are the pro- and imp-OmpA proteins. Neither are heat modifiable, and hence both differ in conformation from the mature OmpA protein. Both are soluble in Sarkosyl and are, in all probability, associated with

the plasma membrane. This location is consistent with the fact that processing is a late event in the synthesis of the protein and that about 40% of the precursor is processed post- translationally (34).

Expression of the ompA gene at a much increased rate resulted in the presence of three protein species. One, a minor component, was heat modifiable and exhibited the partial trypsin resistance, typical of an OmpA protein assembled in the outer membrane. The other two, which accumulated in massive amounts, were not heat modifiable and were completely digested by trypsin. One, the pro-OmpA protein, remained largely nonmodifiable upon addition of LPS while the other, representing the imp-OmpA protein (see also below), became heat modifiable when LPS was added. Obviously, the outer membrane has a limited capacity to incorporate OmpA protein, as only a fraction of the overproduced protein can be assembled in this membrane and the pro-OmpA and imp- OmpA proteins accumulate elsewhere in the cell.

The accumulated pro-OmpA and imp-OmpA proteins could be separated almost quantitatively because the latter was soluble in Sarkosyl while the former was not. Addition of LPS to the imp-OmpA protein resulted not only in conversion to the heat-modifiable form but also in acquisition of phage receptor activity. The degree of this activity was the same as that of the mature protein obtained from the outer membrane. This imp-OmpA protein thus clearly represents a precursor of the mature protein and certainly is the same imp-OmpA protein as that detected in the pulse-chase experiments.

LPS could not convert the pro-OmpA protein to a heat-


..&. i

FIG. 8. Selective solubilization of the imp-OmpA protein. A, unlabeled Epon section of a cell carrying pRD87 after induction with IPTG for 5 h; this cell had undergone lysis during induction of OmpA synthesis. E , immuno-gold-labeled Lowicryl section of such a cell. C, same as B but after extraction with Sar- kosyl. Note the difference in density of gold beads, associated with the outer membrane, between B and C.

.’ . , . . ..

, .

modifiable form; also, very little, if any, activity as phage receptor was effected with LPS. In addition, this precursor was insoluble in Sarkosyl. Do these facts indicate that the pro-OmpA protein has a conformation different from that of the imp-OmpA protein? They may but need not. Inability to assume the conformation resulting in heat modifiability may be caused by the signal sequence. The lipid A moiety of LPS suffices to convert the 36-kDa OmpA species to the 31-kDa form (20). An affinity of the lipophilic signal sequence to lipid A could interfere with the normal OmpA-LPS interaction without which the overall conformation of the pro-OmpA protein needs to be different from that of the imp-OmpA protein. The insolubility of overproduced pro-OmpA in Sar- kosyl, in contrast to the solubility of the pro-OmpA protein found in the pulse-chase experiments, is probably due to the fact that the accumulated precursor is present in a precipitated state. In these precipitates the signal sequences may interact with each other, and this lipophilic interaction may resist the detergent. In summary, an export intermediate, the imp-OmpA protein, exists which differs in conformation from the OmpA protein in the outer membrane and which may, but need not, differ conformationally from the pro-OmpA protein.

In immuno-gold-labeled thin sections of cells which had accumulated the pro- and the imp-OmpA proteins label was found mainly in two areas. One in the cytoplasm, where the labeled protein was present in large clumps and the other as a dense layer in the periplasmic space. I t could be shown that the latter protein represents the imp-OmpA and the former the pro-OmpA protein. This location of the imp-OmpA protein is entirely consistent with an association of the pulse- labeled imp-OmpA with the cytoplasmic membrane. Thus, as more and more of the polypeptide appears on the periplasmic side of the plasma membrane the protein will be pushed into the periplasmic space.

We envisage the following events leading from synthesis of the pro-OmpA protein to localization of the OmpA protein in the outer membrane. The completed precursor, located in the plasma membrane, is processed but remains associated with

this membrane. This imp-OmpA protein is exposed at the periplasmic side of the plasma membrane and is ready to “search” for the lipid A moiety of the LPS. As mentioned above, it is possible that the pro-OmpA protein is unable to do so either because it may differ in conformation from the imp-OmpA protein or because it may not have the right orientation for this search. Recognition of lipid A causes the conformational change described, and this change forces the protein into the outer membrane. Concerning this hypothet- ical scheme, it should be noted that the experiments reported cannot, of course, tell us whether or not the observed conformational change is in fact causally related to the protein’s incorporation into the outer membrane. We are presently searching for mutants in which this incorporation is blocked.

It has been shown that another outer membrane protein, OmpF, passes through a Sarkosyl-soluble state on the way to its eventual Sarkosyl-insoluble state in the outer membrane (35). Moreover, for two such proteins, OmpC and OmpF, it has been demonstrated that they interact specifically with LPS (36). In view of the high degree of amino acid sequence homology between these proteins and the outer membrane protein PhoE (37-39) it would seem likely that PhoE also interacts in a similar way with LPS. Thus, it is possible that a whole group of proteins is “attracted” to this membrane by LPS. If so, we might end up with the question of the chicken and the egg; how does LPS find its membrane? This is unknown (40), but it is conceivable that this substance recognizes a protein in the outer membrane. It is unfortunate that so far true protoplasts cannot be prepared with E. coli. If components of the outer membrane require other components of it as tags for incorporation, one would predict that a cell devoid of an outer membrane could never regenerate this membrane.

Finally, the existence of the imp-OmpA protein has some bearing on the process of translocation across the plasma membrane. We have recently shown that a unique export signal does not exist in the OmpA protein. Overlapping inter- nal deletions were constructed in the ompA gene, and it was found that all products were exported (7). If information for


export, in addition to that contained in the signal sequence, is required and present in the protein, then these results show that it must exist repeatedly in the polypeptide chain. A model has been proposed in which the protein crosses the outer membrane eight times in antiparallel P-sheet conformation, i.e. the protein would be arranged in the membrane by forming four loops (41). Such a conformation is of a repetitive char- acter. We, therefore, considered the possibility that each of the four loops may be able to dictate translocation. The data presented in this communication, however, indicate that the protein assumes the p-sheet conformation only in the outer membrane and that the loops in question do not exist in the imp-OmpA protein. As also discussed previously (7) we, therefore, feel that it remains rather questionable whether another export signal, in addition to that represented by the signal sequence, is required by this or other secreted proteins of E. coli.

Acknowledgments-We thank Ian Crowlesmith for advice and dis- cussions regarding the pulse-chase experiments, Sheila MacIntyre for critically editing the manuscript, Barbara Breidenbend for excellent technical assistance, and Volkmar Braun for suggesting to test the detergent solubility of the proteins.

1.

2.

3. 4.

5.

6.

7.

8.

9.

10.

11. 12.

REFERENCES

Chen, R., Schmidmayr, W., Kramer, C., Chen-Schmeisser, U., and Henning, U. (1980) Proc. Natl. Acad. Sci. U. S. A. 77,

Lugtenberg, B., and van Alphen, L. (1983) Biochim. Biophys. Acta

Beck, E., and Bremer, E. (1980) Nucleic Acids Res. 8, 3011-3027 Mowa, N. R., Nakamura, K., and Inouye, M. (1980) J. Mol. Biol.

Benson, S. A., Bremer, E., and Silhavy, T. J. (1984) Proc. Natl.

Benson, S. A., Hall, M. N., and Silhavy, T. J. (1985) Annu. Reu.

Freudl, R., Schwarz, H., Klose, M., Mowa, R. N., and Henning,

Crowlesmith, I., Gamon, K., and Henning, U. (1981) Eur. J.

Schmitges, C. J., and Henning, U. (1976) Eur. J. Biochern. 6 3 ,

Freudl, R., Braun, G., Hindennach, I., and Henning, U. (1985)

Vieira, J., and Messing, J. (1982) Gene 1 9 , 259-268 Miller, J. H. (1972) Experiments in Molecular Genetics, p. 431,

4592-4596

737,51-115

143,317-328

Acud. Sci. U. S. A . 81,3830-3834

Biochem. 5 4 , 101-134

U. (1985) EMBO J. 4 , 3593-3598

Biochem. 113,375-380

47-52

Mol. Gen. Genet. 201,76-81

Cold Spring Harbor Laboratory, Cold Spring Harbor, NY 13. Crowlesmith, I., and Gamon, K. (1982) Eur. J. Biochem. 124 ,

14. Henning, U., Schwarz, H., and Chen, R. (1979) Anal. Biochem.

15. Laskey, R. A., and Mills, A. D. (1975) Eur. J. Biochem. 56,335-

16. Reithmeier, R. A. F., and Bragg, P. D. (1974) FEBS Lett. 4 1 ,

17. Schnaitman, C. A. (1974) J. Bacteriol. 118, 442-453 18. Hindennach, I., and Henning, U. (1975) Eur. J. Biochern. 5 9 ,

19. Morona, R., Tommassen, J., and Henning, U. (1985) Eur. J.

20. Schweizer, M., Hindennach, I., Garten, W., and Henning, U.

21. Carlemalm, E., Garavito, R. M., and Villiger, W. (1982) J. Mi-

22. Tommassen, J., Leunissen, J., van Damme-Jongsten, M., and

23. Muhlpfordt, H. (1982) Enperientia 3 8 , 1127-1128 24. Keller, G.-A., Tokuyasu, K. T., Dutton, A. H., and Singer, S. J.

25. Heller, K. B. (1978) J. Bucteriol. 134 , 1181-1183 26. Filip, C., Fletcher, G., Wulff, J. L., and Earhart, C. F. (1973) J.

27. Henning, U., Royer, H.-D., Teather, R. M., Hindennach, I., and

577-583

97,153-157

341

195-198

207-213

Biochem. 150,161-169

(1978) Eur. J. Biochem. 82, 211-217

crosc. (Oxf.) 126 , 123-143

Overduin, P. (1985) EMBO J. 4 , 1041-1047

(1984) Proc. Natl. Acud. Sci. U. S. A. 8 1 , 5744-5747

Bucteriol. 1 1 5 , 717-722

28.

29.

30. 31.

32.

33. 34. 35. 36.

37.

38.

39.

40. 41.

Hollenberg, C. P. (1979) Proc. Natl. Acud. Sci. U. S. A. 7 6 ,

Gamon, K., Schmidmayr, W., and Henning, U. (1980) Nucleic

Henning, U., Hohn, B., and Sonntag, I. (1973) Eur. J. Biochem.

Oliver, D. (1985) Annu. Reu. Microbiol. 39,615-648 Randall, L. L., Josefsson, L.-G., and Hardy, S. J. S. (1980) Eur.

Duplay, P., Bedoelle, H., Fowler, A., Zabin, I. Saurin, W., and

Braun, G., and Cole, S. T. (1984) Mol. Gen. Genet. 195,321-328 Josefsson, L.-G., and Randall, L. L. (1981) Cell 25, 151-157 Boyd, A., and Holland, I. B. (1980) J. Bacteriol. 143, 1538-1541 Datta, D. B., Arden, B., and Henning, U. (1977) J. Bacteriol.

Inokuchi, K., Mutoh, N., Matsuyama, S., and Mizushima, S.

Overbeeke, N., Bergmans, H., van Mansfeld, F., and Lugtenberg,

Mizuno, T., Chou, M.-Y., and Inouye, M. (1983) J. Biol. Chem.

Osborn, M. J. (1984) Harvey Lect. 78,87-103 Morona, R., Klose, M., and Henning, U. (1984) J. Bacteriol. 159,

4360-4364

Acids Res. 8,3025-3027

39,27-36

J. Biochem. 107, 375-379

Hofnung, M. (1984) J. Biol. Chem. 259,10606-10613

131,821-829

(1982) Nucleic Acids Res. 10 , 6957-6968

B. (1983) J. Mol. Biol. 163, 513-532

258,6932-6940

570-578

escherichia coli an outer membrane protein (ompa) of ... · pdf filean outer membrane protein...

Documents