JOURNAL
OF
Vol. 140, No. 1
BACTERIOLOGY, Oct. 1979, p. 43-49
0021-9193/79/10-0043/07$02.00/0
Deoxyribonucleic Acid and Outer Membrane: Binding to Outer Membrane Involves a Specific Protein HANS WOLF-WATZ* AND ANDERS NORQVIST Department of Microbiology, University of Umea, S-901 87 Umed, Sweden Received for publication 26 June 1979
The binding of deoxyribonucleic acid (DNA) to the outer membrane of Escherichia coli was examined. The amount of DNA found to be bound to outer membrane was low and was estimated to be about 0.4% of the total DNA. Treatment of cells with chloramphenicol or rifampin caused a disassociation of the apparent DNA-outer membrane complex. The results presented here suggest that the binding between membrane and DNA is specific and involves a membrane protein having a molecular weight of 31,000.
col and rifampin cause a condensation of the nucleus to the center of the cell and a concomitant release of the DNA from the membrane (7, 8, 19, 28). This report is concerned with the association of DNA and the outer membrane of E. coli. Radioactively labeled DNA was traced to both the outer and cytoplasmic membranes, as has also been found by Gomez-Eichelman and Bastarrachea (12). These authors suggested that the binding of DNA to the outer membrane occurs during the preparation of the membranes and is nonspecific. Our results indicate that there is a specific binding of DNA to the outer membrane and that one particular outer membrane protein (molecular weight, 31,000; the 31K protein) is involved in that binding.
The cell envelope of gram-negative bacteria is a highly complex structure consisting of three morphologically and biochemically distinct layers, the cytoplasmic membrane, the peptidoglycan layer, and the outer membrane (for a review, see Freer and Salton [11]). When the bacteria grow and divide, each daughter cell must receive a full genetic complement. A strict correlation must therefore exist between envelope growth, cell division, DNA replication, and DNA segregation. Jacob et al. have suggested that the bacterial chromosome is attached to the bacterial cell membrane and that this attachment is involved in the regulation of DNA synthesis and the segregation of the chromosome to the daughter cells (14). In Escherichia coli data have been presented which indicate that both the chromosomal origin and the chromosomal replication point are associated with a membrane-like structure (1, 10, 16, 17, 20). Different techniques have been used to isolate the DNA membrane complex of E. coli (8, 12, 13, 25). Such complexes contain both cytoplasmic and outer membrane components (13, 20, 21, 24). A role for the outer membrane in DNA replication has also been proposed on the basis of one study which indicated that the origin of chromosomal replication is associated with the outer membrane (20). Recent experiments indicate that the specific binding of the origin to the membrane is mediated by a protein (5) and that this binding of DNA to the cell envelope is rapidly disrupted in vivo by rifampin and chloramphenicol treatment (6). The effect of different antibiotics on the shape and size of nuclear bodies as well as the attachment of these nuclear bodies to the membranes of E. coli has been studied to gain further information about the association of DNA and membranes. Antibiotics such as chlorampheni-
MATERIALS AND METHODS Organism, media, and growth conditions. E. coli CR34 (thy thr leu) was used throughout this study. The minimal medium used was medium E (26) supplemented with 0.2% glucose, 1 pg of thiamine per ml, 50 ,ug of the L-isomer of each required amino acid per ml, and 5 pg of thymine per ml. The Casamino Acids medium contained basal minimal medium, 0.2% casein hydrolysate, 0.2% glucose, and 5 pg of thymine per ml. Separation of outer and cytoplasmic membranes. Separation of outer and cytoplasmic membranes was performed as described by Osborn et al. (22). Normally 100 ml of a cell culture was harvested at an optical density at 450 nm of 0.5, as measured in a Zeiss PMOII spectrophotometer. Cells were lysed, after spheroplast formation, by repeated sonication (five times 10 s). The total membrane fraction was isolated, and outer and cytoplasmic membranes were separated by isopycnic sucrose gradient centrifugation in a Beckman SW40 rotor for 16 h at 38,000 rpm. Purification of DNA. Two liters of cells were grown in minimal medium supplemented with [3H]thymine (5 pg/ml, 0.2 pCi/pg) and [35S]methionine (20 43
44
J. BACTERIOL.
WOLF-WATZ ANI) NORQVIST'
ig/ml, 0.05 1iCi/ig). The l)NA was purified according to the method developed by Marmur (18). Gel electrophoresis of proteins. The gel electrophoresis procedure was that of Laemmli and Favre (15). A slab gel apparatus with a 12.5-cm separating gel of 12% acrylamide was run at 20 mA. Staining, destaining, and drying of gels were as described by Fairbanks et al. (9). Determination of radioactivity in gels. The polyacrylamide gel was dried and fractionated by cutting the gel into 0.5-cm-wide slices. Each fraction was combusted in a Packard Tri-Carb sample oxidizer, model 306. Thereafter, the samples were counted in a Nuclear Chicago Mark II liquid scintillation counter.
RESULTS In vivo labeling of DNA, followed by separation of outer membrane from cytoplasmic membrane. The thymine-requiring E. coli strain CR34 was grown for several generations in Casamino Acids medium containing ['4C]glycerol and [3H]thymine. At a cell density of about 0.5 units of optical density at 450 nm, the cells were harvested, and the outer membrane was separated from the cytoplasmic membrane according to the method developed by Osborn et al. (22). The membranes were separated into two main fractions by sucrose density gradient centrifugation (Fig. 1A). The fractions were des-
ignated OM (outer membrane) and CM (cytoplasmic membrane). As can be seen in Fig. 1A, DNA was associated with both the cytoplasmic and the outer membrane. The amount of DNA that adhered to the membranes was low (about 0.1 to 0.4% of the total DNA). Effect of antibiotic treatment on the observed binding of DNA to the membrane. Treatment of E. coli cells with antibacterial agents that interfere with the protein synthesis (e.g., chloramphenicol or rifampin) results in a condensation of the nucleoplasm and a concomitant release of DNA from its membrane sites (7, 8, 19, 28). The effect of these two drugs on the binding of DNA to the cytoplasmic and outer membrane was therefore tested. Cells were grown and labeled as described above, but chloramphenicol (50 tig/ml) or rifampin (200 jig/ml) was added to the cultures 100 min before the cells were harvested and the membrane was separated. When the cells were treated with either drug, no DNA was found to be associated with either the outer or the cytoplasmic membrane (Fig. 1B and C). When protein synthesis is inhibited by the addition of these drugs, the rate of DNA synthesis gradually declines. The time required for a complete stop of DNA synthesis ranges from 40
1000
500 E
E
c
0
E
,xI
1
(D 1000
5 00 -
FRACTION No.
Fi(J. 1. Effect of uarious treatments on the DNA meni brane complex in strain CR.34. The str'ainllwas grown in Casamino Acids medium supplemented with [3 Hlthymine (5 tg/ml; 1 [&'i/ml) and f1'C]glycerol (1 m71M; 0.1 uCi/ml). At an optical density at 450 nm of 0.5, 100 ml of the culture was harvested, and the outer and cytoplasmic membr-anes were separated by isopycnic sucrose gradient centrifugation (A). The rest of the cultur-e was divided into three aliquots (100 ml). The cells were treated wvith chloramphenicol (200 pg/mlll) (B), rifamipin (200 pug/ml) (C), or nalidixic acid (20 ptg/ml) (D) and incubated for 100 mini before the cells were harvested and the membranes were separated.
VOL. 140, 1979
DNA BINDING TO E. COLI OUTER MEMBRANE
min to about 80 min depending upon the drug used (23). In contrast to this the addition of nalidixic acid causes an immediate cessation of DNA synthesis (4). Nalidixic acid (20 Ag/ml) treatment for 100 min caused no disassociation of the observed complexes between DNA and the two membranes (Fig. 1D). Rather, a twofold increase in the amount of DNA bound to the outer membrane was noticed. Therefore, the inhibition of DNA synthesis per se does not cause the observed release of DNA from the membranes seen after rifampin or chloramphenicol treatment. DNA binding in vivo to outer membrane involves a 31K outer membrane protein. It has been pointed out that the origin DNA is specifically bound to the membrane complex via a protein link (5) and that the DNA found to bind to the outer membrane is specifically enriched in origin DNA (20). With this in mind, we asked whether there exists an outer membrane protein which has the ability to bind DNA. Cells were therefore labeled with [3H]thymine and [35S]methionine. From these cells outer membrane as well as cytoplasmic membrane was isolated according to the method of Osborn et al. (22). The proteins of these two membranes were separated by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (Fig. 2). The gel tracks containing the outer membrane proteins and the cytoplasmic membrane proteins were sliced, and the amount of
45
[3H]thymine in each fraction was determined. As can be seen in Fig. 2, the gel track containing the outer membrane proteins gave a relatively high background level of tritium counts, but one fraction in particular contained a much larger amount of [3H]DNA. This fraction also contained a minute amount of an outer membrane protein having a molecular weight of 31,000. In sharp contrast to this, it was found that almost no [3H]DNA entered the gel when the gel tracks of the separated cytoplasmic membrane proteins were analyzed (Fig. 2), indicating that, in the latter case, the DNA bound to the cytoplasmic membrane was too big in size to pass the stacking gel (which was not sliced) and enter the 12% polyacrylamide gel. The fact that DNA comigrated with one outer membrane protein in the gel was confirmed when the outer membrane was purified from a different strain [MM318(pED620)] using another isolation technique (the Sarkosyl method) (27). Strain MM318pED620 was grown in Casamino Acids medium at 30°C to an optical density at 450 nm of about 0.4, 400 /iCi of [3H]thymidine was added to the culture, and incorporation of isotope was allowed for 30 min. The outer membrane was then isolated from the cells, and the outer membrane proteins were separated by SDS-polyacrylamide gel electrophoresis. As a control, total sonicated bulk DNA was also added to the gel. In this case too, [3H]DNA comigrated with an outer membrane
FIG. 2. Comigration between an outer membrane protein and DNA after separation of outer membrane proteins by polyacrylamide gel electrophoresis. The gel track containing the separated outer membrane proteins (A) or cytoplasmic membrane proteins (B) was fractionated. Each fraction was combusted and counted in a liquid scintillation counter. Symbols: (0) outer membrane; (0) cytoplasmic membrane.
46
\I!t~ subjectd
WOLF-WATZ ANI) NORQVIST
J. BACTERIOLJ.
protein having a molecular weight of 31,000 (Fig. 3). In contrast, total DNA did not enter the gel (data not shown). In vitro binding of DNA to outer membrane involves a 31K protein. The apparent bond between the 31K outer membrane protein and DNA is probably covalent, since boiling in SDS-sample buffer did not result in separation of DNA from the protein (Fig. 4). Because of this it could be argued that purified chromosomal DNA would contain the 31K outer membrane protein. Chromosomal DNA was therefore purified from cells prelabeled with [ 'H]thymine and r35s]methionine. The purified DNA preparation after DNase treatment was solubilized in SDS-sample buffer and subjected to polyacrylamide gel electrophoresis in SDS. This purified DNA contained one predominant protein class showing a molecular weight of 31,000 (indicated by arrows, Fig. 4). Purified outer membrane applied to the same gel contained a very minute amount of a protein with the same molecular weight as the predominating protein associated with purified DNA (Fig. 4). The celis from which the DNA was purified were also labeled with both [3H]thymine and [ 35S]methionine. This labeled, purified DNA was added to a spheroplast suspension prior to sonication and separation of outer and cytoplasmic membrane (see Materials and Methods for details). Labeled purified DNA was found to be associated with outer membrane but not with cytoplasmic membrane (Fig. 5). The ratio of
FIG. 4. .SDS-polyacrylamide slab gel electrophoresis of isolated outer membrane and purified DNA. Outer membrane was prepared from strain CR34 5grown in minimal medium supplemented with (5 yig/ml, 20 ,uCi/,ig) and [35Slmethios[1H/thymine Q_. 4_ nine (20 ,ug/ml, 0.05 ,uCi/p.g). DNA was purified according to the method of Marmur (18) from cells of ol strain CR34 grown in glucose minimal medium sup3 Il plemented uith [3H]thymine (5 ,ug/ml, 0.2 ,iCi/,ig) 7 and [3%Slmethionine (20 tig/ml, 0.05 WCi/,ig). Isolated ._ It _ outer membrane or DNase-treated purified DNA was . 2 L dissolved in SDS-containing buffer and incubated at >1I '100°C for 5 min before the samples were applied to - the gel. After electrophoresis the gel was dried and I t IE1 t \. to autoradiography. Gel track (a), Outer membrane. Gel track (b), Purified DNA. Arrow indi* ,S \ S.-. ' cates the predominating protein of purified DNA, 0 *>. 10 20 30 40 0 50 showing a molecular weight of 31,000. ,, 31 000
m
,
Fraction no. Fi(. :3. Comigration between an outer membrane pr-otein and DNA after separation of outer memgel electrophoresis eproteins bi-an branc Ouiter membrane ut'as isolated from cells of str-ain MM.3'18pED620) labeled with [1H/thymidine. The gel track containing the separ ated outer membrane proteins was sliced, and the amount of radioactivity in each fraction uas determined.
bvpo)lyacrylamide proteinseby polacsylamded gem electrophotreis.
DNA ([:'H]thymine) and protein ([3'S]methionine) in the purified DNA preparation was 40 (counts per minute/counts per minute). This ratio was considerablylowered (about20-fold) in the outer membrane-containing fractions (Fig. 5). In the top of the gradient, on the other hand, the ratio was higher and approached that of
DNA BINDING TO E. COLI OUTER MEMBRANE
VOL. 140, 1979
47
OM l I
a
-
12
10
i
E
II
E
0.
-u
E
I I
Q.
8
0)-
cL
w
c
U
c
c
A
L
A
._
I
~~~fry K~~~~~~~~~~I jh
c
6
0
E
A A~~~~~~~ AA
Ec >
A~~~~~~~~~~~
a)
200
r
4
u
m
m
LO 0
100
2 a
10 20 25 30 15 FRACTION No. FIG. 5. Association ofpurified DNA to the outer and cytoplasmic membrane. DNA was purified from cells of strain CR34 grown in glucose minimal medium supplemented with [3H]thymine (5 ig/ml, 0.2 yCi/lig) and [3"Slmethionine (20 fLg/ml, 0.05 1iCi/ltg). A 100-ml culture of strain CR34 was grown in unlabeled glucose minimal medium. At an optical density at 450 nm of 0.5, the cells were harvested and the membranes were separated. Prior to the sonication step, purified labeled DNA was added. Separation of the membrane fractions was carried out in a 13-ml discontinuous sucrose gradient. Outer membrane was found in fractions 1 to 9, whereas cytoplasmic membrane was found in fractions 16 to 24. 5
purified DNA (Fig. 5). This indicates that there is a specific enrichment of protein in the DNA that is bound to the outer membrane in vitro. Thus, the in vitro binding of purified DNA to outer membrane proceeds with a concomitant binding of protein copurifying with purified DNA. DISCUSSION We found DNA associated with both outer and cytoplasmic membrane after separation of the membranes by isopycnic sucrose density gradient centrifugation. The DNA recovered in the outer membrane fraction was low and amounted to about 0.4% of the total DNA. After treatment of the cells with chloramphenicol or rifampin, the observed binding of DNA to the outer membrane was greatly reduced (Fig. 1). On the basis of these results we argue that if the interaction between outer membrane and DNA is due to nonspecific aggregation during the preparation procedure, the amount of labeled DNA bound to the outer membrane fraction should be the same irrespective of the treatment to which the cells were subjected prior to harvest. Thus, it is likely that the observed interaction between outer membrane and DNA is not artificial in nature. When the bond between DNA and the outer membrane fraction was examined more carefully, it was found that labeled outer membrane DNA comigrated with one protein having
a molecular weight of 31,000 (the 31K protein) after separation of the outer membrane proteins by SDS-polyacrylamide gel electrophoresis (Fig. 2 and 3). How can these results be interpreted? There are three possible explanations: (i) a metabolic effect, i.e. the isotopes are degraded and incorporated specifically into the 31K protein; (ii) a certain fraction of DNA moves to a position in the gel corresponding in molecular weight to a 31,000-molecular-weight protein; (iii) the outer membrane protein, the 31K protein, binds specifically to DNA. Our data are most consistent with the last hypothesis. A metabolic effect is highly unlikely, since (i) we have used three different isotopes, all with the same result, and (ii) a novel metabolic pathway has to be suggested to explain how one protein in particular can be labeled with DNA-specific isotopes. The fact that neither total bulk DNA nor cytoplasmic membrane DNA entered the gel rules out the possibility that a certain fraction of DNA moves to a position in the acrylamide gel corresponding in molecular weight to a 31,000-molecular-weight protein. Thus, the most likely explanation is that a specific protein, the 31K protein, is involved in the binding of DNA to the outer membrane fraction. The bond between the 31K protein and DNA is probably covalent, since heating to 100°C in
48
WOLF-WATZ AND
NORQNQVIST
SDS-containing buffer did not separate the 31K protein from DNA (Fig. 3). Therefore it was of interest that one protein having a molecular weight of 31,000 copurified with DNA (Fig. 4). This protein seems to have affinity for the outer membrane, since almost the total amount of '"Slabeled protein that copurified with DNA was specifically bound in vitro to the outer membrane (Fig. 5), indicating that the protein that copurifies with DNA has a specific affinity for the outer membrane. It is therefore possible that the protein associated with purified DNA and the 31K outer membrane-associated protein found to bind DNA are one and the same. The amount of DNA found to comigrate with the 31K protein is much smaller than that of sonicated DNA found to be bound to the outer membrane fraction. Based on recovery data, the fraction of DNA comigrating with the 31K protein is about 0.003% of total DNA. If we assume that the 31K protein is involved in the origin specific binding to the outer membrane (see accompanying paper [27]) and that there are a mean of four replication origins per cell, the piece of DNA that comigrates with the 31K protein can roughly be estimated to be about 50 to 75 nucleotides in length. The question is whether or not this polynucleotide affects the mobility of the 31K protein in the gel system used. To our knowledge there are in the literature no available data on the migration behavior of a DNA-protein complex in an SDS-polyacrylamide gel system. It must be assumed that a large percentage of DNA in such a complex would cause an abnormal migration behavior of this molecule. However, Baily and Apirin (2) have observed the migration behavior of RNA (4S, 5S) in the SDS-polyacrylamide gel system developed by Laemmli (15). Using their data, it can be calculated that a single-strand polynucleotide exceeding 400 bases would not enter a 12% gel, whereas a 40-bases-long polynucleotide would migrate with the front (2). It is therefore possible that the piece of DNA found to comigrate with the 31K protein does not affect the migration behavior of this protein. In support of this are the following observations: (i) DNase treatment of purified outer membrane does not alter the position of the 31K protein in the gel; (ii) after DNase treatment of purified DNA a 31K protein copurifying with DNA is detected; (iii) induction of a Aasn transducing phage causes an increase of the 31K protein in the outer membrane (27), which most likely is not accompanied with a corresponding increase of DNA found to be bound to this protein. Thus, it is possible that the true molecular weight of this 31K protein-DNA complex is
J BAC TERIOL.
higher than 31,000. How can such a small piece of DNA be generated? One possible mechanism is that the 31K protein recognizes and binds to a certain DNA sequence. The release of the chromosome from the outer membrane could in this case be mediated by a site-specific endonuclease. As a consequence of this enzyme action a small polynucleotide is still bound to the 31K protein. Alternatively, the DNA bound to this protein is already very small in vivo. A small specific polynucleotide which is covalently bound to the 31K protein is synthesized. This DNA-protein complex may thereafter base pair to a specific part of the chromosome. In this latter case the DNA acts as helper to direct the 31K outer membrane protein to its target on the chromosome. In conclusion, we have here demonstrated that DNA is bound to the outer membrane and that this link probably is mediated by a protein having a molecular weight of 31,000. In addition, it has been shown that the origin of replication is specifically bound to the membrane via a protein (5) and that outer membrane DNA is enriched in origin DNA (20). Moreover, the binding between the cell envelope and the replication origin is rapidly disrupted in vivo by rifampin and chloramphenicol (6). We therefore speculate that the 31K protein described here is involved in the origin-specific binding to the outer membrane fraction. It must be emphasized, however, that outer membrane is here defined as a membrane of E. coli moving into a certain position in an isopycnic sucrose gradient, or a membrane structure which is insoluble in the detergent sodium lauryl sarcosinate. These two definitions do not rule out the possibility that DNA is bound to the outer membrane fraction through the observed zones of adhesion between inner and outer membrane (3), as has been pointed out by others (13, 20, 24). However, the ultimate answer to this question must await until these adhesion zones have been isolated and characterized. ACKNOWLEDGMENTS This work was supported by grants from The Swedish Natural Science Research Council (no. B3557-002 to H. WolfWatz and no. B3373-008 to S. Normark) and a grant from Harald and Greta Jeanssons Foundation to H. Wolf-Watz.
LITERATURE CITEI) I. Abe, M., C. Brown, W. G. Henrickson, D. H. Boyet, N. Glifford, R. H. Corte, and M. Schaechter. 1977. Release of Escherichia coli DNA from membrane complexes by single-strand endonucleases. Proc. Natl. Acad. Sci. U.S.A. 74:2756-2760. 2. Baily, S. C., and D. Apirin. 1977. Identification of lipopolysaccharides and phospholipids of Escherichhia coli in polyacrylamide gels. J. Bacteriol. 131:347-355.
VOL. 140, 1979
DNA BINDING TO E. COLI OUTER MEMBRANE
3. Bayer, M. E. 1968. Adsorption of bacteriophages to adhesions between wall and membrane of Escherichia coli. J. Virol. 2:346-356. 4. Cozzarelli, N. R. 1977. The mechanism of action of inhibitors of DNA synthesis. Annu. Rev. Biochem. 46: 641-668. 5. Craine, B. L., and C. S. Rupert. 1978. Identification of a biochemically unique DNA membrane interaction involving the Escherichia coli origin of replication. J. Bacteriol. 134:193-199. 6. Craine, B. L., and C. S. Rupert. 1979. Deoxyribonucleic acid-membrane interactions near the origin of replication and initiation of deoxyribonucleic acid synthesis in Escherichia coli. J. Bacteriol. 137:740-745. 7. Dworsky, P. 1974. Einfluss von Inhibitoren der Proteinund Nucleinsauresynthese auf die Gestalt des Nucleoids von Escherichia coli. Z. Allg. Mikrobiol. 14:3-28. 8. Dworsky, P., and M. Schaechter. 1973. Effect of rifampin on the structure and membrane attachment of the nucleoid of Escherichia coli. J. Bacteriol. 116:13541374. 9. Fairbanks, G., T. L. Steck, and D. F. H. Wallach. 1971. Electrophoretic analysis of the major polypeptides of the human erythrocyte membrane. Biochemistry 10: 2606-2617. 10. Fielding, P., and C. F. Fox. 1970. Evidence for stable attachment of DNA to membrane at the replication origin of Escherichia coli. Biochem. Biophys. Res. Commun. 41:157-162. 11. Freer, J. H., and M. R. J. Salton. 1972. The anatomy and chemistry of Gram negative cell envelopes, p. 67129. In G. Weinbaum, S. Kadis, and J. S. Aje (ed.), Microbial toxins, vol. 4. Academic Press Inc., New York. 12. Gomez-Eichelmann, M. C., and F. Bastarrachea. 1975. Progress in the resolution of the cytoplasmic membrane DNA initiation complex of Escherichia coli. Biochem. Biophys. Acta 407:273-282. 13. Heidrich, H. G., and W. L. Olsen. 1975. Deoxyribonucleic acid envelope complexes from Escherichia coli. J. Cell. Biol. 67:444-460. 14. Jacob, F., S. Brenner, and F. Cuzin. 1963. On the regulation of DNA replication in bacteria. Cold Spring Harbor Symp. Quant. Biol. 28:329-347. 15. Laemmli, U. K., and M. Favre. 1976. Maturation of the head of bacteriophage T4. J. Mol. Biol. 80:575-599. 16. Lundqvist-Parker, P., and D. A. Glaser. 1974. Chromosomal sites of DNA membrane attachment in Escherichia coli. J. Mol. Biol. 87:153-168.
49
17. Lundqvist-Parker, P., and D. A. Glaser. 1975. Effect of growth conditions on DNA membrane attachment in Escherichia coli. Proc. Natl. Acad. Sci. U.S.A. 72:24462459. 18. Marmur, J. 1961. A procedure for the isolation of deoxyribonucleic acid from microorganisms. J. Mol. Biol. 3: 208-218. 19. Morgan, C., H. S. Rosenkranz, H. S. Can, and H. M. Rose. 1967. Electron microscopy of chloramphenicoltreated Escherichia coli. J. Bacteriol. 93:1987-2002. 20. Nicolaides, T., and B. Holland. 1978. Evidence for the specific association of chromosomal origin with outer membrane fractions isolated from Escherichia coli. J. Bacteriol. 135:178-189. 21. Olsen, W. L., H. G. Heidrich, K. Haning, and P. H. Hofschneider. 1974. Deoxyribonucleic acid envelope complexes isolated from Escherichia coli by free flow electrophoresis. Biochemical and electron microscope characterization. J. Bacteriol. 118:646-653. 22. Osborn, M., J. E. Gander, E. Parisi, and J. Carson. 1972. Mechanism and assembly of the outer membrane of Salmonella typhimurium. Isolation and characterization of cytoplasmic and outer membrane. J. Biol. Chem. 247:3961-3972. 23. Pato, M. L. 1975. Alterations of the rate of movement of deoxyribonucleic acid replication forks. J. Bacteriol. 123:272-277. 24. Portailer, R., and A. Worcel. 1975. Association of the folded chromosome with the cell envelope of Escherichia coli characterization of the proteins at the DNA membrane attachment site. Cell 8:245-255. 25. Stonington, 0. G., and D. E. Pettijohn. 1971. The folded genome of Escherichia coli isolated in protein DNA-RNA complex. Proc. Natl. Acad. Sci. U.S.A. 68: 6-9. 26. Vogel, H. J., and D. M. Bonner. 1966. Acetylornithinase of Escherichia coli: partial purification and some properties. J. Biol. Chem. 218:97-106. 27. Wolf-Watz, H., and M. Masters. 1979. Deoxyribonucleic acid and outer membrane: strains diploid for the oriC region show elevated levels of deoxyribonucleic acidbinding protein and evidence for specific binding of the oriC region to outer membrane. J. Bacteriol. 140:50-58. 28. Zusman, D. F., A. Carbonell, and J. Y. Haga. 1973. Nucleoid condensation and cell division in Escherichia coli MX74T2 ts52 after inhibition of protein synthesis. J. Bacteriol. 115:1167-1178.
OF
Vol. 140, No. 1
BACTERIOLOGY, Oct. 1979, p. 43-49
0021-9193/79/10-0043/07$02.00/0
Deoxyribonucleic Acid and Outer Membrane: Binding to Outer Membrane Involves a Specific Protein HANS WOLF-WATZ* AND ANDERS NORQVIST Department of Microbiology, University of Umea, S-901 87 Umed, Sweden Received for publication 26 June 1979
The binding of deoxyribonucleic acid (DNA) to the outer membrane of Escherichia coli was examined. The amount of DNA found to be bound to outer membrane was low and was estimated to be about 0.4% of the total DNA. Treatment of cells with chloramphenicol or rifampin caused a disassociation of the apparent DNA-outer membrane complex. The results presented here suggest that the binding between membrane and DNA is specific and involves a membrane protein having a molecular weight of 31,000.
col and rifampin cause a condensation of the nucleus to the center of the cell and a concomitant release of the DNA from the membrane (7, 8, 19, 28). This report is concerned with the association of DNA and the outer membrane of E. coli. Radioactively labeled DNA was traced to both the outer and cytoplasmic membranes, as has also been found by Gomez-Eichelman and Bastarrachea (12). These authors suggested that the binding of DNA to the outer membrane occurs during the preparation of the membranes and is nonspecific. Our results indicate that there is a specific binding of DNA to the outer membrane and that one particular outer membrane protein (molecular weight, 31,000; the 31K protein) is involved in that binding.
The cell envelope of gram-negative bacteria is a highly complex structure consisting of three morphologically and biochemically distinct layers, the cytoplasmic membrane, the peptidoglycan layer, and the outer membrane (for a review, see Freer and Salton [11]). When the bacteria grow and divide, each daughter cell must receive a full genetic complement. A strict correlation must therefore exist between envelope growth, cell division, DNA replication, and DNA segregation. Jacob et al. have suggested that the bacterial chromosome is attached to the bacterial cell membrane and that this attachment is involved in the regulation of DNA synthesis and the segregation of the chromosome to the daughter cells (14). In Escherichia coli data have been presented which indicate that both the chromosomal origin and the chromosomal replication point are associated with a membrane-like structure (1, 10, 16, 17, 20). Different techniques have been used to isolate the DNA membrane complex of E. coli (8, 12, 13, 25). Such complexes contain both cytoplasmic and outer membrane components (13, 20, 21, 24). A role for the outer membrane in DNA replication has also been proposed on the basis of one study which indicated that the origin of chromosomal replication is associated with the outer membrane (20). Recent experiments indicate that the specific binding of the origin to the membrane is mediated by a protein (5) and that this binding of DNA to the cell envelope is rapidly disrupted in vivo by rifampin and chloramphenicol treatment (6). The effect of different antibiotics on the shape and size of nuclear bodies as well as the attachment of these nuclear bodies to the membranes of E. coli has been studied to gain further information about the association of DNA and membranes. Antibiotics such as chlorampheni-
MATERIALS AND METHODS Organism, media, and growth conditions. E. coli CR34 (thy thr leu) was used throughout this study. The minimal medium used was medium E (26) supplemented with 0.2% glucose, 1 pg of thiamine per ml, 50 ,ug of the L-isomer of each required amino acid per ml, and 5 pg of thymine per ml. The Casamino Acids medium contained basal minimal medium, 0.2% casein hydrolysate, 0.2% glucose, and 5 pg of thymine per ml. Separation of outer and cytoplasmic membranes. Separation of outer and cytoplasmic membranes was performed as described by Osborn et al. (22). Normally 100 ml of a cell culture was harvested at an optical density at 450 nm of 0.5, as measured in a Zeiss PMOII spectrophotometer. Cells were lysed, after spheroplast formation, by repeated sonication (five times 10 s). The total membrane fraction was isolated, and outer and cytoplasmic membranes were separated by isopycnic sucrose gradient centrifugation in a Beckman SW40 rotor for 16 h at 38,000 rpm. Purification of DNA. Two liters of cells were grown in minimal medium supplemented with [3H]thymine (5 pg/ml, 0.2 pCi/pg) and [35S]methionine (20 43
44
J. BACTERIOL.
WOLF-WATZ ANI) NORQVIST'
ig/ml, 0.05 1iCi/ig). The l)NA was purified according to the method developed by Marmur (18). Gel electrophoresis of proteins. The gel electrophoresis procedure was that of Laemmli and Favre (15). A slab gel apparatus with a 12.5-cm separating gel of 12% acrylamide was run at 20 mA. Staining, destaining, and drying of gels were as described by Fairbanks et al. (9). Determination of radioactivity in gels. The polyacrylamide gel was dried and fractionated by cutting the gel into 0.5-cm-wide slices. Each fraction was combusted in a Packard Tri-Carb sample oxidizer, model 306. Thereafter, the samples were counted in a Nuclear Chicago Mark II liquid scintillation counter.
RESULTS In vivo labeling of DNA, followed by separation of outer membrane from cytoplasmic membrane. The thymine-requiring E. coli strain CR34 was grown for several generations in Casamino Acids medium containing ['4C]glycerol and [3H]thymine. At a cell density of about 0.5 units of optical density at 450 nm, the cells were harvested, and the outer membrane was separated from the cytoplasmic membrane according to the method developed by Osborn et al. (22). The membranes were separated into two main fractions by sucrose density gradient centrifugation (Fig. 1A). The fractions were des-
ignated OM (outer membrane) and CM (cytoplasmic membrane). As can be seen in Fig. 1A, DNA was associated with both the cytoplasmic and the outer membrane. The amount of DNA that adhered to the membranes was low (about 0.1 to 0.4% of the total DNA). Effect of antibiotic treatment on the observed binding of DNA to the membrane. Treatment of E. coli cells with antibacterial agents that interfere with the protein synthesis (e.g., chloramphenicol or rifampin) results in a condensation of the nucleoplasm and a concomitant release of DNA from its membrane sites (7, 8, 19, 28). The effect of these two drugs on the binding of DNA to the cytoplasmic and outer membrane was therefore tested. Cells were grown and labeled as described above, but chloramphenicol (50 tig/ml) or rifampin (200 jig/ml) was added to the cultures 100 min before the cells were harvested and the membrane was separated. When the cells were treated with either drug, no DNA was found to be associated with either the outer or the cytoplasmic membrane (Fig. 1B and C). When protein synthesis is inhibited by the addition of these drugs, the rate of DNA synthesis gradually declines. The time required for a complete stop of DNA synthesis ranges from 40
1000
500 E
E
c
0
E
,xI
1
(D 1000
5 00 -
FRACTION No.
Fi(J. 1. Effect of uarious treatments on the DNA meni brane complex in strain CR.34. The str'ainllwas grown in Casamino Acids medium supplemented with [3 Hlthymine (5 tg/ml; 1 [&'i/ml) and f1'C]glycerol (1 m71M; 0.1 uCi/ml). At an optical density at 450 nm of 0.5, 100 ml of the culture was harvested, and the outer and cytoplasmic membr-anes were separated by isopycnic sucrose gradient centrifugation (A). The rest of the cultur-e was divided into three aliquots (100 ml). The cells were treated wvith chloramphenicol (200 pg/mlll) (B), rifamipin (200 pug/ml) (C), or nalidixic acid (20 ptg/ml) (D) and incubated for 100 mini before the cells were harvested and the membranes were separated.
VOL. 140, 1979
DNA BINDING TO E. COLI OUTER MEMBRANE
min to about 80 min depending upon the drug used (23). In contrast to this the addition of nalidixic acid causes an immediate cessation of DNA synthesis (4). Nalidixic acid (20 Ag/ml) treatment for 100 min caused no disassociation of the observed complexes between DNA and the two membranes (Fig. 1D). Rather, a twofold increase in the amount of DNA bound to the outer membrane was noticed. Therefore, the inhibition of DNA synthesis per se does not cause the observed release of DNA from the membranes seen after rifampin or chloramphenicol treatment. DNA binding in vivo to outer membrane involves a 31K outer membrane protein. It has been pointed out that the origin DNA is specifically bound to the membrane complex via a protein link (5) and that the DNA found to bind to the outer membrane is specifically enriched in origin DNA (20). With this in mind, we asked whether there exists an outer membrane protein which has the ability to bind DNA. Cells were therefore labeled with [3H]thymine and [35S]methionine. From these cells outer membrane as well as cytoplasmic membrane was isolated according to the method of Osborn et al. (22). The proteins of these two membranes were separated by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (Fig. 2). The gel tracks containing the outer membrane proteins and the cytoplasmic membrane proteins were sliced, and the amount of
45
[3H]thymine in each fraction was determined. As can be seen in Fig. 2, the gel track containing the outer membrane proteins gave a relatively high background level of tritium counts, but one fraction in particular contained a much larger amount of [3H]DNA. This fraction also contained a minute amount of an outer membrane protein having a molecular weight of 31,000. In sharp contrast to this, it was found that almost no [3H]DNA entered the gel when the gel tracks of the separated cytoplasmic membrane proteins were analyzed (Fig. 2), indicating that, in the latter case, the DNA bound to the cytoplasmic membrane was too big in size to pass the stacking gel (which was not sliced) and enter the 12% polyacrylamide gel. The fact that DNA comigrated with one outer membrane protein in the gel was confirmed when the outer membrane was purified from a different strain [MM318(pED620)] using another isolation technique (the Sarkosyl method) (27). Strain MM318pED620 was grown in Casamino Acids medium at 30°C to an optical density at 450 nm of about 0.4, 400 /iCi of [3H]thymidine was added to the culture, and incorporation of isotope was allowed for 30 min. The outer membrane was then isolated from the cells, and the outer membrane proteins were separated by SDS-polyacrylamide gel electrophoresis. As a control, total sonicated bulk DNA was also added to the gel. In this case too, [3H]DNA comigrated with an outer membrane
FIG. 2. Comigration between an outer membrane protein and DNA after separation of outer membrane proteins by polyacrylamide gel electrophoresis. The gel track containing the separated outer membrane proteins (A) or cytoplasmic membrane proteins (B) was fractionated. Each fraction was combusted and counted in a liquid scintillation counter. Symbols: (0) outer membrane; (0) cytoplasmic membrane.
46
\I!t~ subjectd
WOLF-WATZ ANI) NORQVIST
J. BACTERIOLJ.
protein having a molecular weight of 31,000 (Fig. 3). In contrast, total DNA did not enter the gel (data not shown). In vitro binding of DNA to outer membrane involves a 31K protein. The apparent bond between the 31K outer membrane protein and DNA is probably covalent, since boiling in SDS-sample buffer did not result in separation of DNA from the protein (Fig. 4). Because of this it could be argued that purified chromosomal DNA would contain the 31K outer membrane protein. Chromosomal DNA was therefore purified from cells prelabeled with [ 'H]thymine and r35s]methionine. The purified DNA preparation after DNase treatment was solubilized in SDS-sample buffer and subjected to polyacrylamide gel electrophoresis in SDS. This purified DNA contained one predominant protein class showing a molecular weight of 31,000 (indicated by arrows, Fig. 4). Purified outer membrane applied to the same gel contained a very minute amount of a protein with the same molecular weight as the predominating protein associated with purified DNA (Fig. 4). The celis from which the DNA was purified were also labeled with both [3H]thymine and [ 35S]methionine. This labeled, purified DNA was added to a spheroplast suspension prior to sonication and separation of outer and cytoplasmic membrane (see Materials and Methods for details). Labeled purified DNA was found to be associated with outer membrane but not with cytoplasmic membrane (Fig. 5). The ratio of
FIG. 4. .SDS-polyacrylamide slab gel electrophoresis of isolated outer membrane and purified DNA. Outer membrane was prepared from strain CR34 5grown in minimal medium supplemented with (5 yig/ml, 20 ,uCi/,ig) and [35Slmethios[1H/thymine Q_. 4_ nine (20 ,ug/ml, 0.05 ,uCi/p.g). DNA was purified according to the method of Marmur (18) from cells of ol strain CR34 grown in glucose minimal medium sup3 Il plemented uith [3H]thymine (5 ,ug/ml, 0.2 ,iCi/,ig) 7 and [3%Slmethionine (20 tig/ml, 0.05 WCi/,ig). Isolated ._ It _ outer membrane or DNase-treated purified DNA was . 2 L dissolved in SDS-containing buffer and incubated at >1I '100°C for 5 min before the samples were applied to - the gel. After electrophoresis the gel was dried and I t IE1 t \. to autoradiography. Gel track (a), Outer membrane. Gel track (b), Purified DNA. Arrow indi* ,S \ S.-. ' cates the predominating protein of purified DNA, 0 *>. 10 20 30 40 0 50 showing a molecular weight of 31,000. ,, 31 000
m
,
Fraction no. Fi(. :3. Comigration between an outer membrane pr-otein and DNA after separation of outer memgel electrophoresis eproteins bi-an branc Ouiter membrane ut'as isolated from cells of str-ain MM.3'18pED620) labeled with [1H/thymidine. The gel track containing the separ ated outer membrane proteins was sliced, and the amount of radioactivity in each fraction uas determined.
bvpo)lyacrylamide proteinseby polacsylamded gem electrophotreis.
DNA ([:'H]thymine) and protein ([3'S]methionine) in the purified DNA preparation was 40 (counts per minute/counts per minute). This ratio was considerablylowered (about20-fold) in the outer membrane-containing fractions (Fig. 5). In the top of the gradient, on the other hand, the ratio was higher and approached that of
DNA BINDING TO E. COLI OUTER MEMBRANE
VOL. 140, 1979
47
OM l I
a
-
12
10
i
E
II
E
0.
-u
E
I I
Q.
8
0)-
cL
w
c
U
c
c
A
L
A
._
I
~~~fry K~~~~~~~~~~I jh
c
6
0
E
A A~~~~~~~ AA
Ec >
A~~~~~~~~~~~
a)
200
r
4
u
m
m
LO 0
100
2 a
10 20 25 30 15 FRACTION No. FIG. 5. Association ofpurified DNA to the outer and cytoplasmic membrane. DNA was purified from cells of strain CR34 grown in glucose minimal medium supplemented with [3H]thymine (5 ig/ml, 0.2 yCi/lig) and [3"Slmethionine (20 fLg/ml, 0.05 1iCi/ltg). A 100-ml culture of strain CR34 was grown in unlabeled glucose minimal medium. At an optical density at 450 nm of 0.5, the cells were harvested and the membranes were separated. Prior to the sonication step, purified labeled DNA was added. Separation of the membrane fractions was carried out in a 13-ml discontinuous sucrose gradient. Outer membrane was found in fractions 1 to 9, whereas cytoplasmic membrane was found in fractions 16 to 24. 5
purified DNA (Fig. 5). This indicates that there is a specific enrichment of protein in the DNA that is bound to the outer membrane in vitro. Thus, the in vitro binding of purified DNA to outer membrane proceeds with a concomitant binding of protein copurifying with purified DNA. DISCUSSION We found DNA associated with both outer and cytoplasmic membrane after separation of the membranes by isopycnic sucrose density gradient centrifugation. The DNA recovered in the outer membrane fraction was low and amounted to about 0.4% of the total DNA. After treatment of the cells with chloramphenicol or rifampin, the observed binding of DNA to the outer membrane was greatly reduced (Fig. 1). On the basis of these results we argue that if the interaction between outer membrane and DNA is due to nonspecific aggregation during the preparation procedure, the amount of labeled DNA bound to the outer membrane fraction should be the same irrespective of the treatment to which the cells were subjected prior to harvest. Thus, it is likely that the observed interaction between outer membrane and DNA is not artificial in nature. When the bond between DNA and the outer membrane fraction was examined more carefully, it was found that labeled outer membrane DNA comigrated with one protein having
a molecular weight of 31,000 (the 31K protein) after separation of the outer membrane proteins by SDS-polyacrylamide gel electrophoresis (Fig. 2 and 3). How can these results be interpreted? There are three possible explanations: (i) a metabolic effect, i.e. the isotopes are degraded and incorporated specifically into the 31K protein; (ii) a certain fraction of DNA moves to a position in the gel corresponding in molecular weight to a 31,000-molecular-weight protein; (iii) the outer membrane protein, the 31K protein, binds specifically to DNA. Our data are most consistent with the last hypothesis. A metabolic effect is highly unlikely, since (i) we have used three different isotopes, all with the same result, and (ii) a novel metabolic pathway has to be suggested to explain how one protein in particular can be labeled with DNA-specific isotopes. The fact that neither total bulk DNA nor cytoplasmic membrane DNA entered the gel rules out the possibility that a certain fraction of DNA moves to a position in the acrylamide gel corresponding in molecular weight to a 31,000-molecular-weight protein. Thus, the most likely explanation is that a specific protein, the 31K protein, is involved in the binding of DNA to the outer membrane fraction. The bond between the 31K protein and DNA is probably covalent, since heating to 100°C in
48
WOLF-WATZ AND
NORQNQVIST
SDS-containing buffer did not separate the 31K protein from DNA (Fig. 3). Therefore it was of interest that one protein having a molecular weight of 31,000 copurified with DNA (Fig. 4). This protein seems to have affinity for the outer membrane, since almost the total amount of '"Slabeled protein that copurified with DNA was specifically bound in vitro to the outer membrane (Fig. 5), indicating that the protein that copurifies with DNA has a specific affinity for the outer membrane. It is therefore possible that the protein associated with purified DNA and the 31K outer membrane-associated protein found to bind DNA are one and the same. The amount of DNA found to comigrate with the 31K protein is much smaller than that of sonicated DNA found to be bound to the outer membrane fraction. Based on recovery data, the fraction of DNA comigrating with the 31K protein is about 0.003% of total DNA. If we assume that the 31K protein is involved in the origin specific binding to the outer membrane (see accompanying paper [27]) and that there are a mean of four replication origins per cell, the piece of DNA that comigrates with the 31K protein can roughly be estimated to be about 50 to 75 nucleotides in length. The question is whether or not this polynucleotide affects the mobility of the 31K protein in the gel system used. To our knowledge there are in the literature no available data on the migration behavior of a DNA-protein complex in an SDS-polyacrylamide gel system. It must be assumed that a large percentage of DNA in such a complex would cause an abnormal migration behavior of this molecule. However, Baily and Apirin (2) have observed the migration behavior of RNA (4S, 5S) in the SDS-polyacrylamide gel system developed by Laemmli (15). Using their data, it can be calculated that a single-strand polynucleotide exceeding 400 bases would not enter a 12% gel, whereas a 40-bases-long polynucleotide would migrate with the front (2). It is therefore possible that the piece of DNA found to comigrate with the 31K protein does not affect the migration behavior of this protein. In support of this are the following observations: (i) DNase treatment of purified outer membrane does not alter the position of the 31K protein in the gel; (ii) after DNase treatment of purified DNA a 31K protein copurifying with DNA is detected; (iii) induction of a Aasn transducing phage causes an increase of the 31K protein in the outer membrane (27), which most likely is not accompanied with a corresponding increase of DNA found to be bound to this protein. Thus, it is possible that the true molecular weight of this 31K protein-DNA complex is
J BAC TERIOL.
higher than 31,000. How can such a small piece of DNA be generated? One possible mechanism is that the 31K protein recognizes and binds to a certain DNA sequence. The release of the chromosome from the outer membrane could in this case be mediated by a site-specific endonuclease. As a consequence of this enzyme action a small polynucleotide is still bound to the 31K protein. Alternatively, the DNA bound to this protein is already very small in vivo. A small specific polynucleotide which is covalently bound to the 31K protein is synthesized. This DNA-protein complex may thereafter base pair to a specific part of the chromosome. In this latter case the DNA acts as helper to direct the 31K outer membrane protein to its target on the chromosome. In conclusion, we have here demonstrated that DNA is bound to the outer membrane and that this link probably is mediated by a protein having a molecular weight of 31,000. In addition, it has been shown that the origin of replication is specifically bound to the membrane via a protein (5) and that outer membrane DNA is enriched in origin DNA (20). Moreover, the binding between the cell envelope and the replication origin is rapidly disrupted in vivo by rifampin and chloramphenicol (6). We therefore speculate that the 31K protein described here is involved in the origin-specific binding to the outer membrane fraction. It must be emphasized, however, that outer membrane is here defined as a membrane of E. coli moving into a certain position in an isopycnic sucrose gradient, or a membrane structure which is insoluble in the detergent sodium lauryl sarcosinate. These two definitions do not rule out the possibility that DNA is bound to the outer membrane fraction through the observed zones of adhesion between inner and outer membrane (3), as has been pointed out by others (13, 20, 24). However, the ultimate answer to this question must await until these adhesion zones have been isolated and characterized. ACKNOWLEDGMENTS This work was supported by grants from The Swedish Natural Science Research Council (no. B3557-002 to H. WolfWatz and no. B3373-008 to S. Normark) and a grant from Harald and Greta Jeanssons Foundation to H. Wolf-Watz.
LITERATURE CITEI) I. Abe, M., C. Brown, W. G. Henrickson, D. H. Boyet, N. Glifford, R. H. Corte, and M. Schaechter. 1977. Release of Escherichia coli DNA from membrane complexes by single-strand endonucleases. Proc. Natl. Acad. Sci. U.S.A. 74:2756-2760. 2. Baily, S. C., and D. Apirin. 1977. Identification of lipopolysaccharides and phospholipids of Escherichhia coli in polyacrylamide gels. J. Bacteriol. 131:347-355.
VOL. 140, 1979
DNA BINDING TO E. COLI OUTER MEMBRANE
3. Bayer, M. E. 1968. Adsorption of bacteriophages to adhesions between wall and membrane of Escherichia coli. J. Virol. 2:346-356. 4. Cozzarelli, N. R. 1977. The mechanism of action of inhibitors of DNA synthesis. Annu. Rev. Biochem. 46: 641-668. 5. Craine, B. L., and C. S. Rupert. 1978. Identification of a biochemically unique DNA membrane interaction involving the Escherichia coli origin of replication. J. Bacteriol. 134:193-199. 6. Craine, B. L., and C. S. Rupert. 1979. Deoxyribonucleic acid-membrane interactions near the origin of replication and initiation of deoxyribonucleic acid synthesis in Escherichia coli. J. Bacteriol. 137:740-745. 7. Dworsky, P. 1974. Einfluss von Inhibitoren der Proteinund Nucleinsauresynthese auf die Gestalt des Nucleoids von Escherichia coli. Z. Allg. Mikrobiol. 14:3-28. 8. Dworsky, P., and M. Schaechter. 1973. Effect of rifampin on the structure and membrane attachment of the nucleoid of Escherichia coli. J. Bacteriol. 116:13541374. 9. Fairbanks, G., T. L. Steck, and D. F. H. Wallach. 1971. Electrophoretic analysis of the major polypeptides of the human erythrocyte membrane. Biochemistry 10: 2606-2617. 10. Fielding, P., and C. F. Fox. 1970. Evidence for stable attachment of DNA to membrane at the replication origin of Escherichia coli. Biochem. Biophys. Res. Commun. 41:157-162. 11. Freer, J. H., and M. R. J. Salton. 1972. The anatomy and chemistry of Gram negative cell envelopes, p. 67129. In G. Weinbaum, S. Kadis, and J. S. Aje (ed.), Microbial toxins, vol. 4. Academic Press Inc., New York. 12. Gomez-Eichelmann, M. C., and F. Bastarrachea. 1975. Progress in the resolution of the cytoplasmic membrane DNA initiation complex of Escherichia coli. Biochem. Biophys. Acta 407:273-282. 13. Heidrich, H. G., and W. L. Olsen. 1975. Deoxyribonucleic acid envelope complexes from Escherichia coli. J. Cell. Biol. 67:444-460. 14. Jacob, F., S. Brenner, and F. Cuzin. 1963. On the regulation of DNA replication in bacteria. Cold Spring Harbor Symp. Quant. Biol. 28:329-347. 15. Laemmli, U. K., and M. Favre. 1976. Maturation of the head of bacteriophage T4. J. Mol. Biol. 80:575-599. 16. Lundqvist-Parker, P., and D. A. Glaser. 1974. Chromosomal sites of DNA membrane attachment in Escherichia coli. J. Mol. Biol. 87:153-168.
49
17. Lundqvist-Parker, P., and D. A. Glaser. 1975. Effect of growth conditions on DNA membrane attachment in Escherichia coli. Proc. Natl. Acad. Sci. U.S.A. 72:24462459. 18. Marmur, J. 1961. A procedure for the isolation of deoxyribonucleic acid from microorganisms. J. Mol. Biol. 3: 208-218. 19. Morgan, C., H. S. Rosenkranz, H. S. Can, and H. M. Rose. 1967. Electron microscopy of chloramphenicoltreated Escherichia coli. J. Bacteriol. 93:1987-2002. 20. Nicolaides, T., and B. Holland. 1978. Evidence for the specific association of chromosomal origin with outer membrane fractions isolated from Escherichia coli. J. Bacteriol. 135:178-189. 21. Olsen, W. L., H. G. Heidrich, K. Haning, and P. H. Hofschneider. 1974. Deoxyribonucleic acid envelope complexes isolated from Escherichia coli by free flow electrophoresis. Biochemical and electron microscope characterization. J. Bacteriol. 118:646-653. 22. Osborn, M., J. E. Gander, E. Parisi, and J. Carson. 1972. Mechanism and assembly of the outer membrane of Salmonella typhimurium. Isolation and characterization of cytoplasmic and outer membrane. J. Biol. Chem. 247:3961-3972. 23. Pato, M. L. 1975. Alterations of the rate of movement of deoxyribonucleic acid replication forks. J. Bacteriol. 123:272-277. 24. Portailer, R., and A. Worcel. 1975. Association of the folded chromosome with the cell envelope of Escherichia coli characterization of the proteins at the DNA membrane attachment site. Cell 8:245-255. 25. Stonington, 0. G., and D. E. Pettijohn. 1971. The folded genome of Escherichia coli isolated in protein DNA-RNA complex. Proc. Natl. Acad. Sci. U.S.A. 68: 6-9. 26. Vogel, H. J., and D. M. Bonner. 1966. Acetylornithinase of Escherichia coli: partial purification and some properties. J. Biol. Chem. 218:97-106. 27. Wolf-Watz, H., and M. Masters. 1979. Deoxyribonucleic acid and outer membrane: strains diploid for the oriC region show elevated levels of deoxyribonucleic acidbinding protein and evidence for specific binding of the oriC region to outer membrane. J. Bacteriol. 140:50-58. 28. Zusman, D. F., A. Carbonell, and J. Y. Haga. 1973. Nucleoid condensation and cell division in Escherichia coli MX74T2 ts52 after inhibition of protein synthesis. J. Bacteriol. 115:1167-1178.
