Two-dimensional SDS-polyacrylamide gel electrophoresisof heat-modifiable outer-membrane proteins172 R. R. B. R U S S E L L ~
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/11/14 For personal use only.
Division of Biological Sciences, National Research Council of Ccrnada, Otta~vcr,Canada K I A OR6 Accepted September 8, 1975
RUSSELL,R. R. B. 1976. Two-dimensional SDS-polyacrylamide gel electrophoresis of heatmodifiable outer membrane proteins. Can. J. Microbiol. 22: 83-91. An examination has been made of the effect which temperature of solubilization has upon the subsequent migration in SDS-polyacrylamide gel electrophoresis of proteins from the cell envelopes of Eschericlzia coli K12 and Neisseria sicca ATCC 9913. Conventional electrophoresis in tubes revealed substantial differences in the staining patterns of gels, depending upon whether the envelope samples were solubilized at 37 "C or 100 "C; in the case of N . sicca at least 6 of 13 discernible bands displayed heat-modifiable behavior. The relationship of the bands produced by each of the two temperatures was investigated by a two-dimensional electrophoresis procedure, in which a sample was solubilized at 37 "C and run in a usual cylindrical gel; the entire gel was then resolubilized at 100 OC, and laid along an acrylamide slab for electrophoresis in the second dimension. It was found that "free endotoxin" of both organisms examined contained the same major proteins as the total envelope fraction, and that these free endotoxin proteins showed the same heat-modifiable properties as when present in total envelopes. RUSSELL,R. R. B. 1976. Two-dimensional SDS-polyacrylamide gel electrophoresis of heatmodifiable outer membrane proteins. Can. J. Microbiol. 22: 83-91. Un examen a kt6 fait sur les effets que la temperature de solubilisation peut avoir sur la migration subsiquente des prottines par electrophorese sur un gel SDS-polyacrylamide a partir des enveloppes cellulaires des Escherichia c o b K12 et des Neisseria sicca ATCC 9913. L'electrophorese conventionnel en tubes rkvele des differences substantielles dans les patrons de coloration des gels, selon que les echantillons furent solubilises a 37 "C ou 100 "C; dans le cas de N . siccn au moins 6 des 13 bandes visibles montrent un comporternent modifiable a la chaleur. La relation des bandes produites par chacune des deux temperatures fut ttudiee par une procedure d'electrophor&se en deux dimensions, dans laquelle un echantillon fut solubilise a 37 "C et conduit dans un gel cylindrique usuel; le gel entierfut ensuite resolubilise a 100 "C, et eta16 le long d'une bande acrylamide pour I'electrophorese dans la deuxikme dimension. I1 a kt6 trouve que "les endotoxines libres" des 'deux organismes examines contiennent les m&mes proteines majeures dans la fraction totale de I'enveloppe, et que ces proteines endotoxines libres montrent les m h e s proprietks rnodifiables a la chaleur tels que presents dans les enveloppes totales. [Traduit par le journal]
.
.
Introduction The outer membrane of the gram-negative bacteria is a complex structure composed primarily of protein, lipopolysaccharide, and phospholipid (4). The protein composition of a variety of species has been examined by the technique of sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis, and in all cases the number of proteins detected is fairly limited, up t o 20 bands being discernible o n gels (with about two-thirds of the total protein being ac'Received July 14, 1975. 2NRCC No. 15050. 3Prosent address: Royal College of Surgeons of England, Dental Research Unit, Downe, Kent BR6 755, U.K.
counted for by only two to four polypeptides), and these proteins can be shown to be the major ones of the total cell envelope (2, 3, 12, 14, 16, 17, 20, 23, 24, 25, 26). A number of reports have described how certain major outer-membrane proteins display anomalous behavior in SDS-polyacrylamide gel electrophoresis, their mobility on electrophoresis being dependent in unexpected ways upon the conditions used for sample solubilization in SDS. Thus it has been reported that prior exposure to acidic conditions, urea, o r organic solvents can result in altered mobilities of particular proteins (10, 20) a s can the temperature selected for solubilization (2, 3, 8, 10, 12, 15, 19, 20, 22, 24, 25).
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/11/14 For personal use only.
84
C A N . J . MlCROBIl3L. VOL. 22. 1976
It is possible to investigate the effect of solubilization conditions on migration if a pure protein is available, or by recovering a particular protein band from one gel, then subjecting it t o reelectrophoresis under different conditions. The latter procedure, however, is not practicable if the band pattern is complex. The present paper describes the use of two-dimensional SDSpolyacrylamide gel electrophoresis to investigate the effect of solubilization temperature on the subsequent migration of outer-membrane proteins from Neisseria sicca and Escherichia coli K12.
Materials and Methods Orga~tisr?rs The bacteria used were N . siccn ATCC4 9913 and E. coli K I2 strain AB3282 (F- f11i-I, leu-35 1, pmA2, frp-356, his-4, iloC7, orgE3, galK2, lacY1, ma/-358, sir-704, tsx358). Note that AB3282 is not the same as strain 3282 used by Henning and his co-workers (7, 8). Growtlr of Bncieria The medium used was Difco trypticase soy broth with 0.5% Difco yeast extract added. Cultures were grown in shaking Erlenmeyer flasks and harvested in the late logarithmic phase of growth. Neisserin sicca was grown at 37 "C and E. coli at 42 "C. No difference could be detected between envelopes from cultures of E. coli grown at 32 "C, 37 "C, o r 42 "C, but the release of free endotoxin is greater at the higher temperature (6). Prepnratio~ro f E~loelopeFraclio~ls The total envelope fraction was prepared according to the method of Inouye and Guthrie (9) by sonication of whole cells followed by centrifugation steps to purify the envelopes. Free endotoxin (the term introduced by Crutchley el a/. (5) for the outer-membrane material which spontaneously blebs off the cell surface) was obtained by removing cells from a culture by centrifugation at 10 000 x g for 10 min, and centrifuging the supernatant at 200 000 x g for 1 h. The required material was obtained as a gel-like pellet. All envelope fractions were stored frozen in water at -20 OC, at a protein concentration of 1-2 rng/ml. Elec~roplroresis SDS-polyacrylamide gel electrophoresis in glass tubes was carried out using either the discontinuous Tris5-HCI buffer system of Laemmli (1 3) as described before (24), o r in the phosphate buffer system used by Henning et al. (8). The concentration of acrylamide was 10%. Samples, which contained 100-200 pg of protein, were solubilized in the appropriate buffer for 2 h at 37 "C o r for 5 min in a boiling water bath. In some cases, the sample was incorporated into a small piece of gel (cast in a 4-mrn inside diameter tube) by adding it to acrylamide before polymerization. This allowed recovery of samples for electrophoresis in a second dimension. 4ATCC, American Type Culture Collection. 5Tris, tris(hydr0xymethy1)aminomethane.
For two-dimensional electrophoresis, gels were removed from their tubes after completion of the firstdimension run, and soaked in Laemmli solubilizing buffer for 30 min, before being placed in a boiling water bath for 10 min. The gels were then placed on top of a slab electrophoresis apparatus similar to that described by Kaltschmidt and Wittman ( l l ) , which was kindly provided by L. Visentin. The size of the gel slab was 8 x 8 x 0.3 crn, topped by an 8 x 1 x 0.3 cm stacking gel. The conditions of electrophoresis were adjusted to be the same as those used for the first dimension. The relationship of the cylindrical and slab gels is illustrated in Fig. 6. T o detect protein bands after electrophoresis, gels were fixed in 12.5% trichloracetic acid at 60°C for IS min, then in Coomassie blue stain ( 2 g Coomassie brilliant blue in 450 ml water, 450 ml ethanol, 100 ml glacial acetic acid) at 60 "C for 15 min. Destaining was achieved by exhaustive washing of the gels in 10% acetic acid.
Results Heat-modifiable Proteins of N. sicca We recently described the protein composition of the outer membrane of N. sicca and reported the effect of temperature of solubilization on the staining pattern of gels of the free endotoxin fraction, which contains nearly all the proteins of the outer membrane (24). Figure l a shows the pattern seen when solubilization is carried out a t 37 "C, and Fig. Ib the pattern when 1 0 0 ° C is used. It is readily apparent that band 7 is not found when the higher temperature is used, whereas new bands 1, 5, a n d 6 are formed. Use of the split-gel technique (Fig. Ic) reveals what appear to be slight shifts in the mobilities of several of the minor bands, and differences in intensity of staining of bands can also be detected by spectrophotometric scanning of gels. Comparison of the two scans in Fig. 2, shows the band changes mentioned above, and also shows that the increase in solubilization temperature causes less intense staining of band 8 but increased staining at the position of band 10. T o investigate the behavior of the heat-modifiable proteins further, a two-dimensional electrophoresis technique was used, in which a sample solubilized at 37 "C was electrophoresed on a cylindrical gel as usual, the entire gel then immersed in solubilizing buffer, heated a t 100 O C , and placed on top of a gel slab for electrophoresis in the second dimension. If the gel was run in the second dimension without the 100 OC treatment, the protein bands all migrated with the same mobility as they did in the first-dimension gels and so lay o n a diagonal line across the slab (not illustrated).
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/11/14 For personal use only.
RUSSELL: SDS-POLYACRYLAMIDE GEL ELECTROPHORESIS
FIG.1. SDS-polyacrylamide gel electrophoresis (Laemmli system) of free endotoxin of N. sicca: (a) sample solubilized a t 37 ' C , (b) sample solubilized at 100 ' C , (c) split gel with sample solubilized at 37 "C and sample solubilized at 100 "C. For the split gel experiment, the sample compartment was divided by a small strip of Teflon.
Figure 3 shows the pattern obtained when a sample is run in the first dimension after solubilization at 37 "C. and then heated to 100 OC before the run in the second dimension. The most obvious effect of the heat treatment is that a large proportion of the protein which was present in band 8 in the first dimension now travels with a reduced mobility, and is located at a position corresponding to that of band 6 (i.e. it migrates with an apparent molecular weight of
50 000 instead of 38 000). No spot could be detected at the position corresponding to band 7 on the diagonal, but there was one (very faint in Fig. 3) directly above, with the mobility predicted for band 5 in the second dimension. This indicates that the exposure to 100 "C completely converted a protein to a more slowly migrating form. An explanation for the change in intensity of band 10 is provided by the fact that two proteins from the region of bands 11 and 12
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/11/14 For personal use only.
86
C A N . J. MlCROBll3 L . VOL. 22, 1976
100 "C (Fig. 4b), two or more strong bands in the 30 000 - 40 000 molecular weight range are found, and we previously reported the resolution of bands 30K and 33K (with apparent molecular weights 30 000 and 33 000, respectively). Figures 4c and d show the protein composition of the free endotoxin released by growing cells of E. coli K12. As has been previously reported for Neisseria (24, 26), free endotoxin contains the same major proteins as the cell envelope from which it is derived. The same dependence of the staining pattern on temperature of solubilization is also seen, 33K being insoluble at 37 "C but soluble at 100 "C (Fig. 4e). Our previous experiments showed (24), and the two-dimensional method confirms (Fig. 5), that the 30K band is formed by the 24K protein migrating more slowly after 100 "C treatment, while the 33K band is not solubilized at all at 37"C, but moves out of the sample gel after 100 "C solubilization. The other major band, 18K, is not heat-modifiable but a number of minor envelope proteins lie off the diagonal in Fig. 5.
FIG.2. Densitometer scan of gels similar to those illustrated in Fig. l a and b. Gels were scanned at a wavelength of 570 nm with a Gilford 240 spectrophotometer with a gel-scanning attachment.
now migrate more slowly, to lie close to band 10. There are thus three distinct components contributing to the staining of band 10 when solubilization has been at 100°C. No spot corresponding to band 1 could be resolved in gels such as that shown in Fig. 3, and this is in agreement with our previous finding that this protein fails to enter the gel at all when a solubilization temperature of 37 "C is used (24). Heat-modifiable Proteins of E. coli The major protein of the E. coli K12 cell envelope detected by the Laemmli gel system when samples are solubilized at 37 "C (Fig. 4a) has a molecular weight of about 24000 (2, 12, 24); using the nomenclature of Ames (1) we refer to this band as 24K. When solubilization is at
Comparison of Two Gel Systems As the mobility, and hence the apparent molecular weight of envelope proteins is dependent upon the conditions of sample solubilization, one cannot place any confidence in estimations of molecular weight derived from SDS-polyacrylamide gel electrophoresis, and must rely on other methods of molecular weight calculation. Henning and his co-workers have purified several of the proteins from E. coli envelopes, and determined their molecular weights by chemical means (7, 8). As the apparent molecular weights Henning et a/. obtained with their electrophoresis system were in good agreement with the chemical estimations, it was of interest to compare their phosphate-buffered electrophoresis system with the discontinuous Tris buffer system of Laemmli. At the top of Fig. 6 the pattern of envelope proteins obtained with Henning's method is shown; the sample is the same as that in Fig. 4a and b. The pattern, while more complex than that of the 'ghost' preparation of Henning, shows the same major bands with apparent molecular weights of 40 000, 18 000, and about 10 000. The 40 000 molecular weight band contains two proteins, named I and 11* by Henning et al. (8), and
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/11/14 For personal use only.
RUSSELL: SDS-POLYACRYLAMIDE GEL ELECTROPHORESIS
87
FIG.3. Two-dimensional electrophoresis of N. sicca free endotoxin. The sample was solubilized at 37°C for electrophoresis in the first dimension, then the entire gel resolubilized at 10O0C for the
second-dimension run. Both electrophoresis runs were in the Laemmli system. Spots are numbered as in Fig. 1. The band at point X is an artefact introduced during handling of the gel.
the 18 000 and 10 000 bands correspond to bands I11 and IV. Band 1V has been identified as a lipoprotein (8), and with our staining procedure it turns purple upon storage in acetic acid, while all other bands are blue. The Laemmli gel system only gives good resolution of proteins of molecular weight greater than about 15 000. Nevertheless, it was possible to detect the presence of the lipoprotein on Laemmli gels of both total envelope
and free endotoxin fractions of E. coli by the characteristic purple color. Figure 6 shows the result of an experiment in which an E. coli envelope sample was run on a Henning-type gel, then in the second dimension on the Laemmli buffer system. The second-dimension conditions clearly give a much improved separation of the two proteins which had apparent molecular weight of 40 000 in the first dimension, and which allow the conclusion that
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/11/14 For personal use only.
88
CAN. J. MLCROBIOL. VOL.
22, 1976
FIG.4. SDS-polyacrylamide gel electrophoresis (Laemmli system) of E. coli cell envelope proteins. Total envelope was solubilized a t (a) 37 "C and (b) 100 "C. Free endotoxin solubilized a t (c) 37 "C and (d) 100 "C; ( e ) shows the fraction of free endotoxin which is insoluble a t 37 "C but soluble at 100 "C, and was obtained by recovering the gel-embedded sample after the electrophoresis shown in (c).
protein I is identical with 33K, while II* is identical with 30K. Proteins 111 and IV show similar mobility in both dimensions.
Discussion SDS-polyacrylamide gel electrophoresis has been widely used in studies of envelope proteins of bacteria but it has become clear, particularly from the numerous reports dealing with E. coli, that both the number of polypeptides resolved, and their apparent molecular weights, are ex-
tremely dependent on the conditions of electrophoresis and the method used for solubilization of the sample. The choice of temperature of solubilization has been repeatedly demonstrated t o alter the electrophoretic pattern found, and raising the temperature of solubilization may either increase o r decrease the rate of migration of individual heat-modifiable proteins(2,8, 10, 18, 19, 20, 24). The use of electrophoresis in two dimensions, with the introduction of a heat treatment of the sample between the first- and second-
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/11/14 For personal use only.
RUSSELL: SDS-POLYACRYLAMIDE GEL ELECTROPHORESIS
89
FIG.5. Two-dimensional electrophoresis of E. colicell envelope. Thesample was solubilized at 37 "C for electrophoresis in the first dimension, then both sample and separating gels resolubilized at 100 OC for the second-dimension run. The sample gel was placed at the top left side of the second-dimension slab. Both electrophoresis runs were in the Laernrnli system.
dimension runs, has provided a convenient method for examining the effect of temperature on the envelope proteins of N. sicca and E. coli and has confirmed and extended the results presented earlier (24). In N. sicca six of the outer membrane bands staining with Coomassie blue (bands I, 5, 6, 7, 8, and 10) show substantial changes in their migration rate or staining intensity when solubilized at 100 "C as compared with 37 "C, and there appear to be slight changes in the mobility of several other bands (Figs. 1, 2). The two-dimensional approach has allowed elucidation of the relationship of the bands resulting from solubilization at the two temperatures, and shows itself of particular value in cases such as band 10,
which clearly contains at least three separate polypeptides after 100 "C solubilization (Fig. 3). No attempt will be made to establish correspondence between the protein bands of the E. coli envelope identified here and by others, as there is evidence that considerable variation exists between different strains of E. coli (1, 19, 21). It is clear, however, that most strains contain a major envelope protein which exhibits a faster rate of migration when a high temperature has been used for sample preparation. The poor solubilization by SDS at 37 "C of this type of heat-modifiable protein can be used as a first step in its purification (2, 3, 19). Rosenbusch (19) has provided evidence that "matrix protein," which displays behavior analogous to 33K,
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/11/14 For personal use only.
CAN. J. MICROBIOL. VOL. 22, 1976
FIG.6. Electrophoresis of E. colicell envelope solubilized at 100 "C and run in the Henning system in the first dimension and the Laemmli system in the second dimension. The protein-staining pattern of a Henning-system gel is shown across the top.
will only bind SDS to the maximum extent when exposed to 100 'C. As migration of proteins in SDS-polyacrylamide gel electrophoresis is directly related to the amount of SDS bound, samples solubilized at 37 "C will clearly have an
anomalously slow migration. It is not clear whether the low capacity of unheated protein for SDS is due to some intrinsic properties of the protein, perhaps related to the high content of p-structure (15, 19) or to its association with the
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/11/14 For personal use only.
RUSSELL: SDS-POLYACRYLAMIDE GEL ELECTROPHORESIS
peptidoglycan component of the envelope. The results presented here would argue against an involvement of peptidoglycan as proteins insoluble a t 37 "C (band 1 of N. sicca and 33K of E. coli) show the same heat-modifiable behavior in samples of both whole cell envelopes and free endotoxin, despite the fact that free endotoxin contains no macromolecular peptidoglycan. The second prominent type of heat-modifiable envelope protein is exemplified by 24K, which is converted to 30K by boiling. The results of Schnaitman (20) and Reithmaier and Bragg (18) have indicated that this reduction in mobility after boiling- in SDS mav be attributable to the protein assuming a more extended configuration, which is also manifested in temperature-dependent behavior in viscosity measurements and gel-permeation chromatography (20, 18). Heatmodifiable behavior such as this is not confined to the major proteins, as a t least four proteins of the N. sicca envelope show reduced mobility with heat treatment.
91
10. INOUYE, M., and M.-L. YEE. 1973. Homogeneity of envelope proteins of Escltericl~iacoli separated by gel electrophoresis in sodium dodecyl sulphate. J. Bacteriol. 113: 304-3 12. 11. KALTSCHMIDT, E., and H. G. WITTMAN. 1969. Ribosomal proteins. VII. Two-dimensional polyacrylamide gel electrophoresis for fingerprinting of ribosomal proteins. Anal. Biochem. 36: 401-412. 12. KOPLOW,J., and J. GOLDFINE. 1974. Alterations in the outer membrane of the cell envelope of heotose-deficient mutants of Escl~erichiocolj. J . ~ a c t k riol. 117: 527-543. 13. L A E M M LU. I , K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 277: 680-685. 14. LEE,N., and M. INOUYE. 1974. Outer membraneproteins of Escherichia coli: biosvnthesis and assemblv. FEBS Lett. 39: 167-170. 15. MIZUSHIMA, S . 1974. Effect of sodium dodecylsulfate and heating on protein conformation in outer and cytoplasmic membranes from Escl~erichicr coli. Biochem. Biophys. Res. Commun. 61: 1221-1226. 1973. 16. O L T M A N NL., F., and A. H. STOUTHAMER. Purification of cytoplasmic membranes and outer membranes from Proterrs tnirahilis. Arch. Mikrobiol. 93: 31 1-325. 17. OSBORN,M. J., J. E. G A N D E RE., P A R I S Iand , J. CARSON.1972. Mechanism of assembly of the outer Acknowledgments membrane ofSolmor~ell'otyphirn~rrirrm.J. Biol. Chem. 247: 3962-3972. I am grateful to A. T. Matheson and L. 18. R E I T H M E I ER. R , A. F., and P. D. BRAGG.1974. Purification and characterisation of a heat-modifiable Visentin for lending me equipment, and to K. G. protein from the outer membrane of Esclzerichicr coli. Johnson and I. J. McDonald, in whose laboraFEBS Lett. 41: 195-198. tories this work was done. 19. ROSENBUSCH, J. P. 1974. Characterisation of the coli. J . Biol. major envelope protein from Escl~ericl~icr 1. AMES,G. F.-L. 1974. Resolution of bacterial proteins Chem. 249: 8019-8029. by polyacrylamide gel electrophoresis on slabs. J. 20. S C H N A I T M AC.NA. , 1973. Outer membrane proteins Biol. Chem. 249: 634-644. of Escl~erichiocoli. I. Effect of preparative conditions 2. AMES,G. F.-L., E. N. SPUDICH, and H. N I K A I D O . on the migr-ation of protein in polyacrylamide gels. 1974. Protein composition of the outer membrane of Arch. Biochem. Biophys. 157: 541-552. Salmor~c~lla rypl~itnr~ri~rtn : effect of lipopolysaccharide 21. S C H N A I T M A C.NA. , 1974. Outer membrane proteins mutations. J. Bacteriol. 117: 406-416. coli IV. Differences in outer membrane of Escl~c~ricltin 3. BRAGG,P. D., and C. Hou. 1972. Organization of proteins due to strain and cultural differences. J . Bacproteins in the native and reformed outer membrane of teriol. 118: 454464. Escl~erichiacoli. Biochim. Biophys. Acta, 274: 47822. S I C C A R DA.I ,G., A. L A Z D U N S and K I , B. M. SHAPIRO. 488. 1972. Interrelationship between membrane protein 4. COSTERTON, J. W., J. M. INGRAM, and K.-J. CHENG. composition and deoxyribonucleic acid synthesis in 1974. Structure and function of the cell envelope of Escl~ericl~icr coli. Biochemistry, 11: 1573-1582. gram-negative bacteria. Bacteriol. Rev. 38: 87-100. 23. S T I N N E TJ., D., and R. G. EAGON.1973. Outer (cell 5. CRUTCHLEY, M. J., D. G. MARSH,and J. CAMERON. wall) membrane proteins of Pse~rdotnonnsaer~rgi1967. Free endotoxin. Nature, 214: 1052. nosn. Can. J. Microbiol. 19: 1469-1471. 6. EGAN,A. F., and R. R. B. RUSSELL.1974. Relaxed 24. RUSSELL,R. R. B., K . G . JOHNSON,and I. J. control of outer membrane synthesis in Escllerichia MCDONALD.1975. Envelope proteins in Neisseria. coli K12. Proc. Aust. Biochem. Soc. 7: 63. Can. J. Microbiol. 21: 15 19-1534. 7. GAKTEN, W., and U . HENNING. 1974. Cell envelope 25. WU, H. C. 1972. Isolation and characterisation of an and shape of Eschericl~incoli K12. Isolation and preEscherichia coli mutant with alteration in the outer liminary characterisation of the major ghost-memmembrane proteins of the cell envelope. Biochim. brane proteins. Eur. J. Biochem. 47: 343-352. Biophys. Acta, 290: 274-289. 8. H E N N I N G U., , B. HOHN,and I. SONNTAG. 1973. Cell 26. ZOLLINGER, W. D., D. L. KASPER, B. J. VEI.TRI,and envelope and shape of Escherichia coli. The ghost 1972. Isolation and characterisaM. S. ARTENSTEIN. membrane. Eur. J. Biochem. 39: 27-36. tion of a native cell wall complex from Neisseria 9. INOUYE, M., and J. P. GUTHRIE. 1969. A mutation meningitidis. Infect. Imrnun. 6: 835-851. which changes a membrane protein of E. coli. Proc. Nat. Acad. Sci. U.S.A. 64: 957-961.
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/11/14 For personal use only.
Division of Biological Sciences, National Research Council of Ccrnada, Otta~vcr,Canada K I A OR6 Accepted September 8, 1975
RUSSELL,R. R. B. 1976. Two-dimensional SDS-polyacrylamide gel electrophoresis of heatmodifiable outer membrane proteins. Can. J. Microbiol. 22: 83-91. An examination has been made of the effect which temperature of solubilization has upon the subsequent migration in SDS-polyacrylamide gel electrophoresis of proteins from the cell envelopes of Eschericlzia coli K12 and Neisseria sicca ATCC 9913. Conventional electrophoresis in tubes revealed substantial differences in the staining patterns of gels, depending upon whether the envelope samples were solubilized at 37 "C or 100 "C; in the case of N . sicca at least 6 of 13 discernible bands displayed heat-modifiable behavior. The relationship of the bands produced by each of the two temperatures was investigated by a two-dimensional electrophoresis procedure, in which a sample was solubilized at 37 "C and run in a usual cylindrical gel; the entire gel was then resolubilized at 100 OC, and laid along an acrylamide slab for electrophoresis in the second dimension. It was found that "free endotoxin" of both organisms examined contained the same major proteins as the total envelope fraction, and that these free endotoxin proteins showed the same heat-modifiable properties as when present in total envelopes. RUSSELL,R. R. B. 1976. Two-dimensional SDS-polyacrylamide gel electrophoresis of heatmodifiable outer membrane proteins. Can. J. Microbiol. 22: 83-91. Un examen a kt6 fait sur les effets que la temperature de solubilisation peut avoir sur la migration subsiquente des prottines par electrophorese sur un gel SDS-polyacrylamide a partir des enveloppes cellulaires des Escherichia c o b K12 et des Neisseria sicca ATCC 9913. L'electrophorese conventionnel en tubes rkvele des differences substantielles dans les patrons de coloration des gels, selon que les echantillons furent solubilises a 37 "C ou 100 "C; dans le cas de N . siccn au moins 6 des 13 bandes visibles montrent un comporternent modifiable a la chaleur. La relation des bandes produites par chacune des deux temperatures fut ttudiee par une procedure d'electrophor&se en deux dimensions, dans laquelle un echantillon fut solubilise a 37 "C et conduit dans un gel cylindrique usuel; le gel entierfut ensuite resolubilise a 100 "C, et eta16 le long d'une bande acrylamide pour I'electrophorese dans la deuxikme dimension. I1 a kt6 trouve que "les endotoxines libres" des 'deux organismes examines contiennent les m&mes proteines majeures dans la fraction totale de I'enveloppe, et que ces proteines endotoxines libres montrent les m h e s proprietks rnodifiables a la chaleur tels que presents dans les enveloppes totales. [Traduit par le journal]
.
.
Introduction The outer membrane of the gram-negative bacteria is a complex structure composed primarily of protein, lipopolysaccharide, and phospholipid (4). The protein composition of a variety of species has been examined by the technique of sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis, and in all cases the number of proteins detected is fairly limited, up t o 20 bands being discernible o n gels (with about two-thirds of the total protein being ac'Received July 14, 1975. 2NRCC No. 15050. 3Prosent address: Royal College of Surgeons of England, Dental Research Unit, Downe, Kent BR6 755, U.K.
counted for by only two to four polypeptides), and these proteins can be shown to be the major ones of the total cell envelope (2, 3, 12, 14, 16, 17, 20, 23, 24, 25, 26). A number of reports have described how certain major outer-membrane proteins display anomalous behavior in SDS-polyacrylamide gel electrophoresis, their mobility on electrophoresis being dependent in unexpected ways upon the conditions used for sample solubilization in SDS. Thus it has been reported that prior exposure to acidic conditions, urea, o r organic solvents can result in altered mobilities of particular proteins (10, 20) a s can the temperature selected for solubilization (2, 3, 8, 10, 12, 15, 19, 20, 22, 24, 25).
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/11/14 For personal use only.
84
C A N . J . MlCROBIl3L. VOL. 22. 1976
It is possible to investigate the effect of solubilization conditions on migration if a pure protein is available, or by recovering a particular protein band from one gel, then subjecting it t o reelectrophoresis under different conditions. The latter procedure, however, is not practicable if the band pattern is complex. The present paper describes the use of two-dimensional SDSpolyacrylamide gel electrophoresis to investigate the effect of solubilization temperature on the subsequent migration of outer-membrane proteins from Neisseria sicca and Escherichia coli K12.
Materials and Methods Orga~tisr?rs The bacteria used were N . siccn ATCC4 9913 and E. coli K I2 strain AB3282 (F- f11i-I, leu-35 1, pmA2, frp-356, his-4, iloC7, orgE3, galK2, lacY1, ma/-358, sir-704, tsx358). Note that AB3282 is not the same as strain 3282 used by Henning and his co-workers (7, 8). Growtlr of Bncieria The medium used was Difco trypticase soy broth with 0.5% Difco yeast extract added. Cultures were grown in shaking Erlenmeyer flasks and harvested in the late logarithmic phase of growth. Neisserin sicca was grown at 37 "C and E. coli at 42 "C. No difference could be detected between envelopes from cultures of E. coli grown at 32 "C, 37 "C, o r 42 "C, but the release of free endotoxin is greater at the higher temperature (6). Prepnratio~ro f E~loelopeFraclio~ls The total envelope fraction was prepared according to the method of Inouye and Guthrie (9) by sonication of whole cells followed by centrifugation steps to purify the envelopes. Free endotoxin (the term introduced by Crutchley el a/. (5) for the outer-membrane material which spontaneously blebs off the cell surface) was obtained by removing cells from a culture by centrifugation at 10 000 x g for 10 min, and centrifuging the supernatant at 200 000 x g for 1 h. The required material was obtained as a gel-like pellet. All envelope fractions were stored frozen in water at -20 OC, at a protein concentration of 1-2 rng/ml. Elec~roplroresis SDS-polyacrylamide gel electrophoresis in glass tubes was carried out using either the discontinuous Tris5-HCI buffer system of Laemmli (1 3) as described before (24), o r in the phosphate buffer system used by Henning et al. (8). The concentration of acrylamide was 10%. Samples, which contained 100-200 pg of protein, were solubilized in the appropriate buffer for 2 h at 37 "C o r for 5 min in a boiling water bath. In some cases, the sample was incorporated into a small piece of gel (cast in a 4-mrn inside diameter tube) by adding it to acrylamide before polymerization. This allowed recovery of samples for electrophoresis in a second dimension. 4ATCC, American Type Culture Collection. 5Tris, tris(hydr0xymethy1)aminomethane.
For two-dimensional electrophoresis, gels were removed from their tubes after completion of the firstdimension run, and soaked in Laemmli solubilizing buffer for 30 min, before being placed in a boiling water bath for 10 min. The gels were then placed on top of a slab electrophoresis apparatus similar to that described by Kaltschmidt and Wittman ( l l ) , which was kindly provided by L. Visentin. The size of the gel slab was 8 x 8 x 0.3 crn, topped by an 8 x 1 x 0.3 cm stacking gel. The conditions of electrophoresis were adjusted to be the same as those used for the first dimension. The relationship of the cylindrical and slab gels is illustrated in Fig. 6. T o detect protein bands after electrophoresis, gels were fixed in 12.5% trichloracetic acid at 60°C for IS min, then in Coomassie blue stain ( 2 g Coomassie brilliant blue in 450 ml water, 450 ml ethanol, 100 ml glacial acetic acid) at 60 "C for 15 min. Destaining was achieved by exhaustive washing of the gels in 10% acetic acid.
Results Heat-modifiable Proteins of N. sicca We recently described the protein composition of the outer membrane of N. sicca and reported the effect of temperature of solubilization on the staining pattern of gels of the free endotoxin fraction, which contains nearly all the proteins of the outer membrane (24). Figure l a shows the pattern seen when solubilization is carried out a t 37 "C, and Fig. Ib the pattern when 1 0 0 ° C is used. It is readily apparent that band 7 is not found when the higher temperature is used, whereas new bands 1, 5, a n d 6 are formed. Use of the split-gel technique (Fig. Ic) reveals what appear to be slight shifts in the mobilities of several of the minor bands, and differences in intensity of staining of bands can also be detected by spectrophotometric scanning of gels. Comparison of the two scans in Fig. 2, shows the band changes mentioned above, and also shows that the increase in solubilization temperature causes less intense staining of band 8 but increased staining at the position of band 10. T o investigate the behavior of the heat-modifiable proteins further, a two-dimensional electrophoresis technique was used, in which a sample solubilized at 37 "C was electrophoresed on a cylindrical gel as usual, the entire gel then immersed in solubilizing buffer, heated a t 100 O C , and placed on top of a gel slab for electrophoresis in the second dimension. If the gel was run in the second dimension without the 100 OC treatment, the protein bands all migrated with the same mobility as they did in the first-dimension gels and so lay o n a diagonal line across the slab (not illustrated).
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/11/14 For personal use only.
RUSSELL: SDS-POLYACRYLAMIDE GEL ELECTROPHORESIS
FIG.1. SDS-polyacrylamide gel electrophoresis (Laemmli system) of free endotoxin of N. sicca: (a) sample solubilized a t 37 ' C , (b) sample solubilized at 100 ' C , (c) split gel with sample solubilized at 37 "C and sample solubilized at 100 "C. For the split gel experiment, the sample compartment was divided by a small strip of Teflon.
Figure 3 shows the pattern obtained when a sample is run in the first dimension after solubilization at 37 "C. and then heated to 100 OC before the run in the second dimension. The most obvious effect of the heat treatment is that a large proportion of the protein which was present in band 8 in the first dimension now travels with a reduced mobility, and is located at a position corresponding to that of band 6 (i.e. it migrates with an apparent molecular weight of
50 000 instead of 38 000). No spot could be detected at the position corresponding to band 7 on the diagonal, but there was one (very faint in Fig. 3) directly above, with the mobility predicted for band 5 in the second dimension. This indicates that the exposure to 100 "C completely converted a protein to a more slowly migrating form. An explanation for the change in intensity of band 10 is provided by the fact that two proteins from the region of bands 11 and 12
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/11/14 For personal use only.
86
C A N . J. MlCROBll3 L . VOL. 22, 1976
100 "C (Fig. 4b), two or more strong bands in the 30 000 - 40 000 molecular weight range are found, and we previously reported the resolution of bands 30K and 33K (with apparent molecular weights 30 000 and 33 000, respectively). Figures 4c and d show the protein composition of the free endotoxin released by growing cells of E. coli K12. As has been previously reported for Neisseria (24, 26), free endotoxin contains the same major proteins as the cell envelope from which it is derived. The same dependence of the staining pattern on temperature of solubilization is also seen, 33K being insoluble at 37 "C but soluble at 100 "C (Fig. 4e). Our previous experiments showed (24), and the two-dimensional method confirms (Fig. 5), that the 30K band is formed by the 24K protein migrating more slowly after 100 "C treatment, while the 33K band is not solubilized at all at 37"C, but moves out of the sample gel after 100 "C solubilization. The other major band, 18K, is not heat-modifiable but a number of minor envelope proteins lie off the diagonal in Fig. 5.
FIG.2. Densitometer scan of gels similar to those illustrated in Fig. l a and b. Gels were scanned at a wavelength of 570 nm with a Gilford 240 spectrophotometer with a gel-scanning attachment.
now migrate more slowly, to lie close to band 10. There are thus three distinct components contributing to the staining of band 10 when solubilization has been at 100°C. No spot corresponding to band 1 could be resolved in gels such as that shown in Fig. 3, and this is in agreement with our previous finding that this protein fails to enter the gel at all when a solubilization temperature of 37 "C is used (24). Heat-modifiable Proteins of E. coli The major protein of the E. coli K12 cell envelope detected by the Laemmli gel system when samples are solubilized at 37 "C (Fig. 4a) has a molecular weight of about 24000 (2, 12, 24); using the nomenclature of Ames (1) we refer to this band as 24K. When solubilization is at
Comparison of Two Gel Systems As the mobility, and hence the apparent molecular weight of envelope proteins is dependent upon the conditions of sample solubilization, one cannot place any confidence in estimations of molecular weight derived from SDS-polyacrylamide gel electrophoresis, and must rely on other methods of molecular weight calculation. Henning and his co-workers have purified several of the proteins from E. coli envelopes, and determined their molecular weights by chemical means (7, 8). As the apparent molecular weights Henning et a/. obtained with their electrophoresis system were in good agreement with the chemical estimations, it was of interest to compare their phosphate-buffered electrophoresis system with the discontinuous Tris buffer system of Laemmli. At the top of Fig. 6 the pattern of envelope proteins obtained with Henning's method is shown; the sample is the same as that in Fig. 4a and b. The pattern, while more complex than that of the 'ghost' preparation of Henning, shows the same major bands with apparent molecular weights of 40 000, 18 000, and about 10 000. The 40 000 molecular weight band contains two proteins, named I and 11* by Henning et al. (8), and
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/11/14 For personal use only.
RUSSELL: SDS-POLYACRYLAMIDE GEL ELECTROPHORESIS
87
FIG.3. Two-dimensional electrophoresis of N. sicca free endotoxin. The sample was solubilized at 37°C for electrophoresis in the first dimension, then the entire gel resolubilized at 10O0C for the
second-dimension run. Both electrophoresis runs were in the Laemmli system. Spots are numbered as in Fig. 1. The band at point X is an artefact introduced during handling of the gel.
the 18 000 and 10 000 bands correspond to bands I11 and IV. Band 1V has been identified as a lipoprotein (8), and with our staining procedure it turns purple upon storage in acetic acid, while all other bands are blue. The Laemmli gel system only gives good resolution of proteins of molecular weight greater than about 15 000. Nevertheless, it was possible to detect the presence of the lipoprotein on Laemmli gels of both total envelope
and free endotoxin fractions of E. coli by the characteristic purple color. Figure 6 shows the result of an experiment in which an E. coli envelope sample was run on a Henning-type gel, then in the second dimension on the Laemmli buffer system. The second-dimension conditions clearly give a much improved separation of the two proteins which had apparent molecular weight of 40 000 in the first dimension, and which allow the conclusion that
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/11/14 For personal use only.
88
CAN. J. MLCROBIOL. VOL.
22, 1976
FIG.4. SDS-polyacrylamide gel electrophoresis (Laemmli system) of E. coli cell envelope proteins. Total envelope was solubilized a t (a) 37 "C and (b) 100 "C. Free endotoxin solubilized a t (c) 37 "C and (d) 100 "C; ( e ) shows the fraction of free endotoxin which is insoluble a t 37 "C but soluble at 100 "C, and was obtained by recovering the gel-embedded sample after the electrophoresis shown in (c).
protein I is identical with 33K, while II* is identical with 30K. Proteins 111 and IV show similar mobility in both dimensions.
Discussion SDS-polyacrylamide gel electrophoresis has been widely used in studies of envelope proteins of bacteria but it has become clear, particularly from the numerous reports dealing with E. coli, that both the number of polypeptides resolved, and their apparent molecular weights, are ex-
tremely dependent on the conditions of electrophoresis and the method used for solubilization of the sample. The choice of temperature of solubilization has been repeatedly demonstrated t o alter the electrophoretic pattern found, and raising the temperature of solubilization may either increase o r decrease the rate of migration of individual heat-modifiable proteins(2,8, 10, 18, 19, 20, 24). The use of electrophoresis in two dimensions, with the introduction of a heat treatment of the sample between the first- and second-
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/11/14 For personal use only.
RUSSELL: SDS-POLYACRYLAMIDE GEL ELECTROPHORESIS
89
FIG.5. Two-dimensional electrophoresis of E. colicell envelope. Thesample was solubilized at 37 "C for electrophoresis in the first dimension, then both sample and separating gels resolubilized at 100 OC for the second-dimension run. The sample gel was placed at the top left side of the second-dimension slab. Both electrophoresis runs were in the Laernrnli system.
dimension runs, has provided a convenient method for examining the effect of temperature on the envelope proteins of N. sicca and E. coli and has confirmed and extended the results presented earlier (24). In N. sicca six of the outer membrane bands staining with Coomassie blue (bands I, 5, 6, 7, 8, and 10) show substantial changes in their migration rate or staining intensity when solubilized at 100 "C as compared with 37 "C, and there appear to be slight changes in the mobility of several other bands (Figs. 1, 2). The two-dimensional approach has allowed elucidation of the relationship of the bands resulting from solubilization at the two temperatures, and shows itself of particular value in cases such as band 10,
which clearly contains at least three separate polypeptides after 100 "C solubilization (Fig. 3). No attempt will be made to establish correspondence between the protein bands of the E. coli envelope identified here and by others, as there is evidence that considerable variation exists between different strains of E. coli (1, 19, 21). It is clear, however, that most strains contain a major envelope protein which exhibits a faster rate of migration when a high temperature has been used for sample preparation. The poor solubilization by SDS at 37 "C of this type of heat-modifiable protein can be used as a first step in its purification (2, 3, 19). Rosenbusch (19) has provided evidence that "matrix protein," which displays behavior analogous to 33K,
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/11/14 For personal use only.
CAN. J. MICROBIOL. VOL. 22, 1976
FIG.6. Electrophoresis of E. colicell envelope solubilized at 100 "C and run in the Henning system in the first dimension and the Laemmli system in the second dimension. The protein-staining pattern of a Henning-system gel is shown across the top.
will only bind SDS to the maximum extent when exposed to 100 'C. As migration of proteins in SDS-polyacrylamide gel electrophoresis is directly related to the amount of SDS bound, samples solubilized at 37 "C will clearly have an
anomalously slow migration. It is not clear whether the low capacity of unheated protein for SDS is due to some intrinsic properties of the protein, perhaps related to the high content of p-structure (15, 19) or to its association with the
Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 11/11/14 For personal use only.
RUSSELL: SDS-POLYACRYLAMIDE GEL ELECTROPHORESIS
peptidoglycan component of the envelope. The results presented here would argue against an involvement of peptidoglycan as proteins insoluble a t 37 "C (band 1 of N. sicca and 33K of E. coli) show the same heat-modifiable behavior in samples of both whole cell envelopes and free endotoxin, despite the fact that free endotoxin contains no macromolecular peptidoglycan. The second prominent type of heat-modifiable envelope protein is exemplified by 24K, which is converted to 30K by boiling. The results of Schnaitman (20) and Reithmaier and Bragg (18) have indicated that this reduction in mobility after boiling- in SDS mav be attributable to the protein assuming a more extended configuration, which is also manifested in temperature-dependent behavior in viscosity measurements and gel-permeation chromatography (20, 18). Heatmodifiable behavior such as this is not confined to the major proteins, as a t least four proteins of the N. sicca envelope show reduced mobility with heat treatment.
91
10. INOUYE, M., and M.-L. YEE. 1973. Homogeneity of envelope proteins of Escltericl~iacoli separated by gel electrophoresis in sodium dodecyl sulphate. J. Bacteriol. 113: 304-3 12. 11. KALTSCHMIDT, E., and H. G. WITTMAN. 1969. Ribosomal proteins. VII. Two-dimensional polyacrylamide gel electrophoresis for fingerprinting of ribosomal proteins. Anal. Biochem. 36: 401-412. 12. KOPLOW,J., and J. GOLDFINE. 1974. Alterations in the outer membrane of the cell envelope of heotose-deficient mutants of Escl~erichiocolj. J . ~ a c t k riol. 117: 527-543. 13. L A E M M LU. I , K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 277: 680-685. 14. LEE,N., and M. INOUYE. 1974. Outer membraneproteins of Escherichia coli: biosvnthesis and assemblv. FEBS Lett. 39: 167-170. 15. MIZUSHIMA, S . 1974. Effect of sodium dodecylsulfate and heating on protein conformation in outer and cytoplasmic membranes from Escl~erichicr coli. Biochem. Biophys. Res. Commun. 61: 1221-1226. 1973. 16. O L T M A N NL., F., and A. H. STOUTHAMER. Purification of cytoplasmic membranes and outer membranes from Proterrs tnirahilis. Arch. Mikrobiol. 93: 31 1-325. 17. OSBORN,M. J., J. E. G A N D E RE., P A R I S Iand , J. CARSON.1972. Mechanism of assembly of the outer Acknowledgments membrane ofSolmor~ell'otyphirn~rrirrm.J. Biol. Chem. 247: 3962-3972. I am grateful to A. T. Matheson and L. 18. R E I T H M E I ER. R , A. F., and P. D. BRAGG.1974. Purification and characterisation of a heat-modifiable Visentin for lending me equipment, and to K. G. protein from the outer membrane of Esclzerichicr coli. Johnson and I. J. McDonald, in whose laboraFEBS Lett. 41: 195-198. tories this work was done. 19. ROSENBUSCH, J. P. 1974. Characterisation of the coli. J . Biol. major envelope protein from Escl~ericl~icr 1. AMES,G. F.-L. 1974. Resolution of bacterial proteins Chem. 249: 8019-8029. by polyacrylamide gel electrophoresis on slabs. J. 20. S C H N A I T M AC.NA. , 1973. Outer membrane proteins Biol. Chem. 249: 634-644. of Escl~erichiocoli. I. Effect of preparative conditions 2. AMES,G. F.-L., E. N. SPUDICH, and H. N I K A I D O . on the migr-ation of protein in polyacrylamide gels. 1974. Protein composition of the outer membrane of Arch. Biochem. Biophys. 157: 541-552. Salmor~c~lla rypl~itnr~ri~rtn : effect of lipopolysaccharide 21. S C H N A I T M A C.NA. , 1974. Outer membrane proteins mutations. J. Bacteriol. 117: 406-416. coli IV. Differences in outer membrane of Escl~c~ricltin 3. BRAGG,P. D., and C. Hou. 1972. Organization of proteins due to strain and cultural differences. J . Bacproteins in the native and reformed outer membrane of teriol. 118: 454464. Escl~erichiacoli. Biochim. Biophys. Acta, 274: 47822. S I C C A R DA.I ,G., A. L A Z D U N S and K I , B. M. SHAPIRO. 488. 1972. Interrelationship between membrane protein 4. COSTERTON, J. W., J. M. INGRAM, and K.-J. CHENG. composition and deoxyribonucleic acid synthesis in 1974. Structure and function of the cell envelope of Escl~ericl~icr coli. Biochemistry, 11: 1573-1582. gram-negative bacteria. Bacteriol. Rev. 38: 87-100. 23. S T I N N E TJ., D., and R. G. EAGON.1973. Outer (cell 5. CRUTCHLEY, M. J., D. G. MARSH,and J. CAMERON. wall) membrane proteins of Pse~rdotnonnsaer~rgi1967. Free endotoxin. Nature, 214: 1052. nosn. Can. J. Microbiol. 19: 1469-1471. 6. EGAN,A. F., and R. R. B. RUSSELL.1974. Relaxed 24. RUSSELL,R. R. B., K . G . JOHNSON,and I. J. control of outer membrane synthesis in Escllerichia MCDONALD.1975. Envelope proteins in Neisseria. coli K12. Proc. Aust. Biochem. Soc. 7: 63. Can. J. Microbiol. 21: 15 19-1534. 7. GAKTEN, W., and U . HENNING. 1974. Cell envelope 25. WU, H. C. 1972. Isolation and characterisation of an and shape of Eschericl~incoli K12. Isolation and preEscherichia coli mutant with alteration in the outer liminary characterisation of the major ghost-memmembrane proteins of the cell envelope. Biochim. brane proteins. Eur. J. Biochem. 47: 343-352. Biophys. Acta, 290: 274-289. 8. H E N N I N G U., , B. HOHN,and I. SONNTAG. 1973. Cell 26. ZOLLINGER, W. D., D. L. KASPER, B. J. VEI.TRI,and envelope and shape of Escherichia coli. The ghost 1972. Isolation and characterisaM. S. ARTENSTEIN. membrane. Eur. J. Biochem. 39: 27-36. tion of a native cell wall complex from Neisseria 9. INOUYE, M., and J. P. GUTHRIE. 1969. A mutation meningitidis. Infect. Imrnun. 6: 835-851. which changes a membrane protein of E. coli. Proc. Nat. Acad. Sci. U.S.A. 64: 957-961.
