Vol. 135, No. 1

JOURNAL OF BACTERIOLOGY, July 1978, P. 78-89 0021-9193/78/0135-0078$2.00/0 Copyright i 1978 American Society for Microbiology

Printed in U.S.A.

Glycolipids Stimulate DNA Polymerase Activity in a DNAMembrane Fraction and in a Partially Purified Polymerase System Extracted from Pneumococci AURELIO ZERIAL, IRWIN GELMAN, AND WILLIAM FIRSHEIN* Biology Department, Wesleyan University, Middletown, Connecticut 06457 Received for publication 28 February 1978

We have assayed the ability of various lipids to affect DNA polymerase activity in a DNA-membrane complex extracted from Streptococcus pneumoniae by the Sarkosyl-M-band technique. In addition, to determine which DNA polymerases were affected by the lipids, we partially purified three DNA polymerase activities from cell lysates, the first such demonstration outside of Escherichia coli and Bacillus subtilis. Glycolipids are unique among polar lipids in stimulating the rate and extent of DNA polymerase activity in M-bands and in Sarkosyl lysates from which the M-band is derived. It appears that they exert this stimulatory effect, in part, by removihg (neutralizing) detergent molecules which act as inhibitors, as well as by substituting for the detergent, thereby creating a favorable environment for the polymerases involved in DNA synthesis. That the stimulatory effect is not simply a detoxification of the detergent was shown by two observations. One, phospholipids, although interacting with Sarkosyl and therefore "potentially" capable of detoxifying the system, did not stimulate DNA polymerase activity in vitro. Two, glycolipids were capable of stimulating the activity of at least two DNA polymerases partially purified from cell lysates in the absence of any Sarkosyl. The stimulatory effect was greater for a polymerase that had four characteristics similar to those observed with polymerase III in other organisms. DNA-membrane complexes have been the subject of intensive investigation during the past decade in an attempt to define the site of DNA synthesis in vivo. Complexes have been extracted from a variety of bacteria (see reference 16 for a review), and some have been shown to synthesize DNA endogenously without the requirement of added enzymes (5, 9, 14). Our own studies have dealt with Streptococcus pneumoniae. A considerable amount of evidence has been amassed to indicate that in this organism the DNA-membrane complex extracted by the Sarkosyl-M-band technique (27) is the site of DNA replication in vivo. Not only is nascent DNA which sediments at 9 to 11S in alkaline sucrose gradients (Okazaki fragments) (5) detected in the fraction, but a complex of enzymes acting cooperatively to synthesize DNA is also present (5, 12). These activities remain with the complex after two different purification procedures that remove a considerable percent-

age of components present in the crude DNAmembrane fraction (5, 6). Nucleotides can be incorporated into DNA at high rates in vitro, and inhibitor studies suggested the presence of DNA polymerases that are involved in replica-

tion of the genome (6). We have also been able to separate the complex into two subcomplexes, both capable of synthesizing DNA in vitro and both possessing a different complement of DNA polymerases (7). One important area of research concerns the effect of lipids on DNA synthesis in the DNA complex. Although several investigations have linked lipids to a metabolic role in DNA synthesis (8, 21), no direct evidence for a specific role of lipids has been found. Only one report (30) has shown that a specific lipid or class of lipids (phospholipids) will enhance the activity of a specific DNA polymerase (polymerase III in Escherichia coli), but only under conditions in which the polymerase is in a multienzyme complex (polymerase III holoenzyme). Cell envelopes of gram-positive and gram-negative bacteria are fundamentally different. Besides the lack of an outer membrane in the former (20), gram-positive bacteria are rich in a class of lipids not found in gram-negative organisms, the glycolipids (25, 26). The most significant feature of the glycolipids is the occurrence of identical glycolipids in members of the same genus. For example, all of the lactobacilli so far 78

VOL. 135, 1978

DNA POLYMERASE ACTIVITY AND GLYCOLIPIDS

79

examined (of which Pneumococcus is one ex- so-called M-band from the remaining membrane and ample) contain galactosylglucosyl diglyceride. cellular components (top fraction). Extraction and purification of glycolipids and No significant explanation for the role of glycoLipids were extracted from the Mlipids in the cell has yet emerged, except that phospholipids. and the remaining cell extract (top fraction) by band they may possibly maintain the integrity of the procedure of Bligh and Dyer (2). To 0.8 volume of membranes (as phospholipids do), provide struc- the M-band or top fraction, 2 volumes of methanol tural support for membrane pores, or transfer was added, and the mixture was stirred at room temsugar residues to polysaccharide chains (26). perature for 20 min. One volume of chloroform was In preliminary studies, we found that 48% of added, and stirring was continued for an additional 45 the total content of lipids in the M-band con- min. The suspension was then filtered through Whatsisted of glycolipids, whereas 25% consisted of man no. 1 paper directly into a separatory flask. Two phospholipids (35%) and neutral lipids (18%). volumes of chloroform and 2 volumes of water were the flask was vigorously agitated and then However, there were no significant differences added, and to stand to permit separation of the two between these percentages and those present in allowed phases. The chloroform layer was removed, and the the remaining membrane fraction (top fraction). material was concentrated in rotating spherical flasks Nevertheless, our interest in the glycolipids was under vacuum with a RINCO rotary evaporator heightened by the finding that glycolipids, but (RINCO Instrument Co., Greenville, Ill.). When a not phospholipids, stimulated DNA synthesis control extract (that is, a total-cell extract obtained by five- to sixfold in vitro when a dialyzed M-band freezing the cells twice in a dry ice-acetone mixture and then thawing twice) was treated with methanol was used as the sole source of enzymes and DNA and chloroform in the right proportions, no difference template. the total recovery of lipids was noted from that This report analyzes the effects of glycolipids in obtained by extraction of the M-band plus top fraction. further and indicates that they exert two differThe procedure of Wells and Dittmer (29) was folent effects on DNA synthesis, one of protection lowed to further purify the lipids. After purification, against inhibitors and one of stimulation ofDNA the various lipid classes were separated by the followpolymerase activity, in particular, a polymerase ing procedure. Lipids were resuspended in 5 ml of resembling polymerase III in other organisms. chloroform and chromatographed on a silicic acid colTo assay the effects of glycolipids on specific umn (2 g of Unisil silicic acid, further activated at DNA polymerases, it was necessary to partially 1100C for 20 h; Clarkson Chemical Co., Williamsport, purify them from cell lysates. Three polymerase Pa.). Elution of lipids was carried out with (in order): activities were detected, including the one men- (i) 60 ml of chloroformn (ii) 30 ml of chloroform-acetone (1:1), (iii) 30 ml of acetone, (iv) 30 ml of chloroformtioned above. methanol (1:1), and finally, (v) 30 ml of chloroform. MATERIALS AND METHODS Growth of organisms. The use and maintenance of an encapsulated strain of type III S. pneumoniae (A66) and the growth of the organisms in a medium containing Casitone, tryptone, albumin, and yeast extract were all described previously (5). Extraction of DNA-membrane fraction. Cells were inoculated into 2 liters of medium with or without 10 ,uCi of [U-4C]palmitate (500 ACi/mmol; New England Nuclear Corp., Boston, Mass.) or 100 ,uCi of [methyl-3H]thymidine (20 Ci/mmol; New England Nuclear Corp.) and incubated statically for 6 h at 37°C until the middle of the log phase. They were centrifuged for 15 min in a Sorvall refrigerated centrifuge (RC2B) using a GSA rotor at 10,000 rpm and washed three times in TMK buffer [0.01 M tris(hydroxymethyl)aminomethane (Tris), 0.01 M magnesium acetate, 0.1 M potassium chloride; final pH 7.5]. A DNA-membrane complex was extracted by the Sarkosyl-M-band technique as described in detail previously (5). In this procedure, Sarkosyl (sodium lauroylsarcosinate) is added to whole pneumococcal cells in the presence of Mg2e. The cells lyse, and at low temperatures in the presence of this ion, Sarkosyl crystals are formed which apparently attach to that portion of the cell membrane to which most of the cellular DNA is complexed. Centrifugation in a biphasic sucrose solution (15% over 60%) separates the

Eluates ii and iii were combined and classified as glycolipids (13), as they contained material which gave a positive reaction to anthrone. Eluates iv and v were also combined, and on the basis of phosphorus content they were classified as phospholipids. Fraction i contained neutral lipids only. This procedure was used for the preparative fractionation of lipids to provide material for the in vitro studies of DNA synthesis. Analysis of lipids. To identify specific lipids within a class of lipids, thin-layer chromatography of radioactive lipids on Silica Gel G (Analabs Inc., North Haven, Conn.) was used (13, 28). Purified lipids were layered at the origin as a series of spots and eluted with benzene-ethyl ether-ethanol-acetic acid (50: 40:2:0.2) to remove neutral lipids (23). The material retained at the origin (polar lipids) was scraped off the plate and transferred to a small column which was then eluted with 5 column volumes of acetone and 5 column volumes of chloroform-methanol-water (60:30:4.5). The lipids were dried, resuspended in a minimal volume of chloroform-methanol (2:1), and layered at the origin of another plate, which was then developed with acetone-acetic acid-water (100:2:1) to elute glycolipids while phospholipids were still retained at the origin. All lipids were detected by staining the gels with iodine vapors or by autoradiography. Two glycolipid spots (present in a 1:1 ratio [Fig. 1]) were detected specifically by staining with orcinol (4). Unstained areas on replicate gels were scraped off the

80

~ ~ ~.

ZERIAL, GELMAN, AND FIRSHEIN

J. BACTERIOL.

..w-F. .

*

|

& GL-GL

FIG. 1. Thin-layer chromatography of lipids on Silica Gel G plates of '4C-labeled polar lipids. Two glycolipids, GL-I (Rf = 0.75) and GL-II (Rf = 0.42), were resolved by developing the plate with acetone-acetic acid-water (100:2:1). Phospholipids remained at the origin. The lipid spots were visualized by autoradiography.

plate and hydrolyzed with acid (2 N HCI at 100°C for 40 min). The hydrolysates were rechromatographed on silica gel with the appropriate sugar standards and stained for hexoses by the anthrone method of Radin et al. (22). It was found that one glycolipid contained glucose while the other contained glucose and galactose, as reported in other studies with Pneumococcus

(13).

The same procedure was carried out for phospholipids, except that the acetone washing of the column was omitted and the last plate was developed with chloroform-methanol-water (60:25:4) (3) to separate phospholipids. Almost 85% of the total pospholipids detected consisted of phosphatidylglycerol, which was identified by its reaction with the Schiff reagent after oxidation with periodate (24). Phosphorus was determined by the method of Bartlett (1). Since the two glycolipids were present in approximately the same ratio, and 3 mol of hexose is liberated after hydrolysis (13), we estimated that 1.5 ,Amol of glucose equivalent corresponds to 1.0 ,mol of lipid. In the case of phospholipids, since the great majority is represented by phosphatidylglycerol, 1 umol of lipid corresponds to 1.0 umol of phosphate. Preparation of lipid vesicles. To prepare lipids for use in the DNA synthesis experiments, the combined eluates ii and iii for glycolipids or iv and v for phospholipids from above were dried in 15-ml Corex tubes under nitrogen flow and suspended in a minimal volume of water by extensive blending in a Vortex mixer. A sonic-treatment step of 10 min with a Bran-

son Sonifier at a 40-W output was also carried out. Glycolipid and phospholipid concentrations were expressed in micromole of glucose equivalents (hexoses) per milliliter or micromoles of phosphate per milliliter, respectively. Interaction of lipid vesicles with Sarkosyl. Dried lipids were suspended at a concentration of 11.5 mg/ml of 5 mM N-2-hydroxyethyl piperazine-N'-2ethanesulfonic acid buffer, pH 7.4, containing 0.145 M NaCl. Samples of 120 pl were mixed with different concentrations of Sarkosyl. After a 5-min incubation at room temperature, the samples were chromatographed on a 3-ml Sepharose 4B (Pharmacia Fine Chemicals Inc., Piscataway, N.J.) column to separate intact vesicles surviving the lytic action of the detergent. The concentration of vesicles was measured by following absorbance at 460 nm. Assays for DNA synthesis. M-bands, top fractions, and Sarkosyl lysate samples were dialyzed for 10 to 12 h in the cold (4°C) against 0.02 M Tris (pH 8.0), containing 20% glycerol (wt/vol) and 0.1% defatted bovine serum albumin (Sigma Chemical Co., St. Louis, Mo.) to reduce salt and detergent concentration. Unless specifically indicated, each sample was adjusted by dilution with the above buffer to contain 100 jig of protein and 106 nmol of DNA. In the case of the top fraction, it was necessary to add pneumococcal DNA extracted from the M-band as described previously (7) to provide the proper concentration, because most of the cellular DNA was present in the M-band. The mixtures were either incubated at 30°C (prein-

VOL. 135, 1978

DNA POLYMERASE ACTIVITY AND GLYCOLIPIDS

cubation step) or allowed to synthesize DNA immediately by addition of 40 ,ul of a solution containing: 0.047 M Tris, pH 8.6; 4.7% glycerol; 10 mM dithiothreitol (DTT); 0.04 M MgCl2 6H20; 0.7 mM each dATP, dCTP, and dGTP; 0.45 M nicotinamide adenine dinucleotide (NAD); 0.25 mM ATP; and 0.87 IgM [methyl-3H]TTP (specific activity, 4,000 cpm/pmol); New England Nuclear Corp. The NAD was added to enhance DNA ligase activity which was also present in the M-band (12). Samples were pipetted onto Whatman 3MM filters which were transferred to a beaker containing 5% trichloroacetic acid and 1% sodium PPi (about 10 ml of this solution per filter). The filters were washed twice in 5% trichloroacetic acid and twice in 95% cold ethanol. They were then transferred to scintillation vials and dried in an oven at 60°C for 2 h before assay for radioactivity in the scintillation counter.

Assays for the more purified DNA polymerases carried out without the addition of NAD and ATP, but with the addition of the appropriate templates. A mixture was prepared containing the following components (in 95 gIl): 10 pl of 0.05 M Tris buffer, pH 7.5 (with 10% sucrose and 40 mM DTT); 6 mM MgCl2 6H20; 0.08 mM each dATP, dCTP, and dGTP; 2.5 gM [3H]TTP (specific activity, 61.5 cpm/pmol); 10 p1 of template, containing either 53 nmol of activated calf thymus DNA, 3.06 nmol of oligodeoxythymidylic acid-polydeoxyadenylic acid [oligo(dT)-poly(dA)], or 4.3 nmol of M-band DNA; 10 iLl of enzyme (unless specified) in 0.02 M Tris (pH 7.5) containing 20% glycerol, 5 mM mercaptoethanol, and 50 mM NaCl. When oligo(dT)-poly(dA) was used as the template, labeled TTP was the only nucleotide present in the assay mixture. After incubation at 30°C for various periods of time, radioactivity was determined as described above for the M-band and top fraction. With both assays, when N-ethylmaleimide (NEM) was added, DTT and mercaptoethanol were omitted from the assay mixtures. Control assays carried out in the absence of DTT showed a 5 to 10% decrease in activity, and values were adjusted to reflect this slight drop in activity. Partial purification of two DNA polymerase activities sensitive to NEM. Cells obtained from 6 liters of cultures were washed twice in TMK buffer. They were suspended in 0.02 M Tris (pH 7.5) containing 0.3 M NaCl, washed once in the latter buffer, and finally suspended at an optical density of 17.5 at 600 nm. After 4.5 ml of the cell suspension was added to 5ml cellulose acetate centrifuge tubes, 0.092 ml of a 10% Triton X solution (Sigma Chemical Co.) was added, and the tubes were allowed to stand in an ice bucket for 30 min. They were then centrifuged at 20,000 rpm in an SW50.1 rotor using a Beckman L2 ultracentrifuge for 65 min at 3°C. The yellow supernatant containing most of the protein and little DNA was removed and dialyzed overnight in the cold against buffer A (0.02 M Tris, pH 7.5, containing 20% glycerol, 50 mM NaCl, and 5 mM mercaptoethanol). The pellet, which contained most of the DNA of the cells and little protein, was discarded because it contained DNA polymerase activity resistant to NEM. A diethylaminoethyl (DEAE)-cellulose column (42ml volume) was prepared following the instructions of

were

81

the manufacturer (DE 52; Whatman) and equilibrated in buffer A. The supernatant containing 80 mg of protein was allowed to enter the resin, which was then washed with 3 column volumes of buffer A containing 100 mM NaCl. A linear gradient of 360 ml of buffer A (0.1 to 0.43 M NaCI) was then applied to the column at a flow rate of 0.4 ml/min. The linearity of the gradient was checked by a Sigma kit (technical bulletin no. 830) for the determination of Cl- ions. A 10-ml phosphocellulose column (Whatman cellulose phosphate Pi1 washed with 1 N NaOH and then extensively with distilled water) was prepared and equilibrated in buffer B (0.02 M potassium phosphate, pH 6.8, containing 2 mM ethylenediaminetetraacetic acid, 5 mM mercaptoethanol, and 20% glycerol). The fractions obtained from DEAE-cellulose (adjusted to contain 10 mg) and possessing enzymatic activity were pooled and dialyzed in buffer B containing 0.05 M potassium phosphate. This fraction was then diluted to 0.02 M potassium phosphate just before its addition to the column. A 100-ml gradient (0.02 to 0.05 M potassium phosphate) was applied, and the linearity of the gradient was checked by measuring its refractive index. The peak fractions containing enzymatic activity that were inhibited by NEM were pooled separately and rechromatographed on a similar phosphocellulose column by the same procedures as described above. Sedimentation studies. A 150-,il quantity of the enzyme or 150 gIl of bovine catalase was layered at the top of a 5 to 30% sucrose gradient (in 0.05 M Tris, pH 7.5, containing 0.12 M KCl and 1 mM mercaptoethanol), prepared in SW50.1 centrifuge tubes. Centrifugation was carried out at 3°C at 39,000 rpm for 19 h in the ultracentrifuge. Fractions were collected by puncturing the tubes and pumping a heavy sucrose solution (60%) through the bottom. Catalase was assayed by following the disappearance of a substrate (H202) at 250 nim. The s2o, values were calculated by the method of Martin and Ames (19). Preparation of DNA templates. Activated calf thymus DNA (Sigma Chemical Co.) was prepared by the procedure of Kornberg and Gefter (15), with the exclusion of the exonuclease treatment. Such a template, therefore, corresponds to nicked doublestranded DNA. M-band DNA was extracted from dialyzed M-bands, as described previously (7). The template is probably activated through the action of Mg2' ions and/or nucleases during the dialysis step. Oligo(dT)i0-poly(dA) was obtained from Miles Research Products (Elkhart, Ind.) and was dissolved in water before use. Other analytical determinations. Protein concentration was determined by the method of Lowry et al. (17).

RESULTS Effects of lipid addition on DNA synthesis with DNA-membrane complexes. Although there were no significant differences in bulk lipid content between a DNA-membrane fraction extracted by the M-band technique and the remaining cell extract (top fraction; see above), it was still possible that the lipids could

82

ZERIAL, GELMAN, AND FIRSHEIN

affect DNA synthesis in these complexes positively. Therefore, we examined the effects of lipid addition on DNA polymerase activity detected previously in M-band complexes (5-7). We were aware that simple addition of various lipids to such complexes might not show any effect if these lipids were already present in the complex. However, we felt that a response could be obtained under conditions in which the magnitude of the effect was a function of concentration of glycolipids. In addition, it was possible that the Sarkosyl extraction procedure resulted in a redistribution or substitution of lipids in the complex, yielding an M-band which would be deficient or different with respect to specific lipid content.

Figure 2 shows that addition of glycolipids to cell lysates obtained by Sarkosyl treatment or from M-bands separated from such lysates resulted in a stimulation of DNA synthesis in vitro and that this stimulation was a linear function of the glycolipid concentration. Table 1 shows that phospholipids had no effect on DNA synthesis when added to Sarkosyl lysates or Mbands. Table 1 also shows that glycolipids affected DNA synthesis because the addition of deoxyribonuclease prevented enhancement and that, when glycolipids were treated with alkali,

2.0

1.5

0.

TABLE 1. Effect of lipids on DNA synthesis in the M-band and top fractionsa Fraction

Condition

M-band

+ Pancreatic deoxyribonuclease + Glycolipids + Glycolipids + pancreatic deoxyribonuclease + Glycolipids + alkali treat-

DNA sis synthe-

100 9.6 198 3

97

ment

+ Phospholipids + Phosphatidylglycerol

102 98

Top

100 + Glycolipids 116 + Phospholipids 100 a Lipids, when added, were preincubated with the M-band (250,ug of protein per ml) or top fraction (535 ,ug of protein per ml) for 24 min. A 500 nmol concentration of pneumococcal DNA was added to the top fractions to equalize the concentration of DNA present in the M-band. Glycolipids were used at a final concentration of 1.9 ,umol of glucose equivalent per ml; phospholipids were added at 2.7 ,umol of phosphate per ml; and phosphatidylglycerol was added at 3.21 Mmol/ml. Pancreatic deoxyribonuclease was used at a final concentration of 200 ,g/ml, and the treatment was carried out from the beginning of the DNA synthesis period. Glycolipids were treated with 0.1 N NaOH at 100°C for 1 h under reflux, and the hydrolysate was neutralized and added to the assay mixture at the same concentration of glucose equivalents as in the untreated samples. When a control sample containing only neutralized KOH was tested, no effect on polymerase activity was detected. The conditions of the experiment are described in the text. The amount of TMP incorporated into DNA in 35 min in the case of untreated samples (0.74 pmol for the M-band and 0.2 pmol for the top fraction) was normalized to 100.

.

0.51

1.0

2.0

3.0

moles gluc. eQ./mi FIG. 2. Stimulation of in vitro DNA synthesis in Sarkosyl lysates (0) or in M-bands (0) byglycolipids. Quantities of 100 Ml' of M-bands or of Sarkosyl lysates adjusted to 350 and 5X0 jg ofprotein per ml, respec-

tively,

J. BACTERIOL.

were preincubated

at 300C for 20 min with 20

of lipids. At the end of the preincubation step, 40 Mul of the DNA synthesis mixture described in the text, Mi

containing 64.5 pmol of /H]TTP, was added, and the incubation was carried out for an additional 20 min. Samples of 30 il were pipetted onto Whatman 3MM filters, and then processed as described in the text for determination of acid-insoluble radioactivity.

the stimulating activity disappeared. There was some DNA polymerase activity in the top fraction which was also stimulated by glycolipids, but the effect was far less than that seen in the M-band. Thus, it appeared as though the enhancement by the glycolipids was specific for the activity associated with the DNA-membrane complex, at least in its magnitude. To further characterize the glycolipid effect, M-bands were preincubated with or without glycolipids before assay for polymerase activity. In addition, glycolipids were added to assay samples preincubated for various periods without glycolipids. The results are shown in Fig. 3. Preincubation without glycolipids for 30 min resulted in a decay of the ability of M-bands to synthesize DNA (about 30 to 40%). However, the presence of glycolipids throughout preincubation not only prevented this loss in activity,

DNA POLYMERASE ACTIVITY AND GLYCOLIPIDS

VOL. 135, 1978

A

4-

3-

aS\ I

0

10

20

30

FIG. 3. Effect of preincubation of M-bands on in vitro DNA synthesis in the presence and absence of

100) ,Il of M-bands at a plg/ml were incubated or 30 min at 30°C. The samples were then

glycolipids.

Quantities of

protein concentration of 580 for 0.5, 15,

assayed for DNA synthesis as described in the text

lmzol

without (0) or with (A) glycolipids present (6.5 of glucose equivalent per ml).

(A)

Value of DNA

synthesis obtained when the samples were incubated for 30

min,

but with glycolipids present throughout

preincubation. Samples of 30

p1l

were transferred to

Whatman 3MM paper filters and processed as described in the text for

83

cubation without glycolipids results in an Mband less capable of synthesizing DNA than non-preincubated samples. In attempts to ascertain whether glycolipids affect more than one polymerase present in the M-band, the sulfhydryl inhibitor NEM was added in the presence and absence of glycolipids, and in the absence of DTT (see above). This compound inhibits DNA polymerase activity involved in replication rather than in repair (11, 15). Figure 5 shows that with increasing concentrations of NEM (up to 9 mM), there was an increasing inhibition of DNA polymerase activity in the M-band in the presence or absence of glycolipids. However, in the presence of glycolipids, there was a more drastic inhibition of polymerase activity between 0 and 2 mM NEM than in their absence (58.7 versus 28.5%). It is interesting to note in this respect that 2 mM NEM completely inhibits DNA polymerase III in Bacillus subtilis (11). Nevertheless, it also appears that glycolipids stimulate DNA polymerase activity that is insensitive to NEM. This is shown by the fact that even at a concentration of 9 mM NEM, DNA polymerase activity could still be detected, and in the presence of glycolipids, the activity was significantly greater. Thus, it can be concluded that glycolipids probably stimulate or

determnination of acid-insoluble

radioactivity.

but the subsequent synthesis (Fig. 3) was almost as high as that of non-preincubated

containing glycolipids (about 400%

samples

stimulation).

This almost-complete protection and stimulation were partially lost the longer the M-band

0.

samples were preincubated without glycolipids. Under these conditions, the addition of glycolipids failed to restore the activity of the M-band

completely to levels seen in the zero-time control with glycolipids. However, the extent of synthesis with glycolipids was still much higher (e.g., after a 30-min preincubation) than the zero-time

control without glycolipids (approximately 100 to 150% stimulation). Figure 4 shows the kinetics of DNA synthesis in M-band samples under various conditions of

preincubation and non-preincubation, with and without glycolipids. It can be seen that both the rate and the extent of DNA synthesis were enhanced

in

M-bands

when

the

samples

were

preincubated with glycolipids, as compared with samples

preincubated

without

glycolipids.

In

non-preincubated samples, the effect of glycolipids was primarily on the extent of synthesis; the

initial rate with glycolipids was only slightly greater than that of the control sample without glycolipids. These results are in keeping with the previous observations shown in Fig. 3 that prein-

min. FIG. 4. Kinetics of DNA synthesis in M-bands incubated with or without glycolipids. Quantities of 300 1d of M-bands, adjusted to a protein concentration of 250 pg/ml, were mixed with 60 ,ul of glycolipids or 60 ,ul of water. They were immediately assayed for DNA synthesis by addition of 120 iLl of the DNA synthesis mixture containing 31.5 pmol of [3HJTTP or preincubated before assay. Final lipid concentration was 1.9 pmol of glucose equivalent per ml. Samples of 10 All were pipetted onto Whatman 3MM filters and processed as described in the text for determination of acid-insoluble radioactivity. Symbols: 0, preincubated, without glycolipids; A, preincubated, with glycolipids; 0, not preincubated, without glycolipids; A, not preincubated, with glycolipids.

84

ZERIAL, GELMAN, AND FIRSHEIN 5

4.

I-1a-

\

E 2

0 2 4 6 8 NEM (mM) FIG. 5. Effects of NEM on in vitro DNA synthesis ofM-bands. Quantities of 1 OO,ul ofM-bands, adjusted to a protein concentration of 580 pg/ml, were preincubated with (A) or without (0) glycolipids for 30 min and then assayed for DNA synthesis in vitro in the presence of NEM, as described in the legend to Fig. 3, except that DTT was omitted from the assays. In controlpreparations, the omission of DTT resulted in a 5 to 10% decrease in activity which was adjusted for in this experiment. Glycolipid concentration was 6.53 umol of glucose equivalent per ml.

protect the activity of more than one DNA polymerase present in the M-band. What do glycolipids protect against in the Mband during preincubation and incubation? We examined two possibilities: (i) loss of endogenous template by nucleases and (ii) inactivation of DNA polymerases by residual Sarkosyl. The first possibility, that glycolipids protect the endogenous template in the M-band from nucleases, was examined by preparing M-bands from cells grown in [3H]thymidine as described above. Samples containing the labeled DNA were mixed with glycolipids in water or with water alone and were incubated for 1 h at 30°C (the temperature at which synthesis is carried out). Very little loss of acid insolubility was detected for both untreated and glycolipidtreated samples after incubation (about 5%). Although the significance of this slight loss in template by nuclease action is unknown, glycolipids do not seem to be involved in the phenomenon. It should be pointed out that previous studies (6) did reveal a loss of template (about 20 to 25%), but the temperature of incubation was 370C, which could have accelerated the rate of degradation. The second possibility was examined by reconstruction experiments with Sarkosyl. Figure 6 shows the effect of Sarkosyl addition on DNA synthesis carried out by the M-band in vitro in the presence and in the absence of glycolipids.

J. BACTrERIOL.

It can be seen that Sarkosyl exerted inhibitory effects on activity and that glycolipids overcame these effects of Sarkosyl at lower concentrations. However, as the concentration of Sarkosyl increased, the protective effect decreased and was almost completely lost at the highest Sarkosyl concentration used. It should be pointed out that there was a certain amount of residual Sarkosyl already present in the M-band. The exact amount is unknown because dialysis removes a portion of the detergent. Another point is that glycolipids were also present in the "endogenous" M-band, which could be important in explaining why DNA synthesis could be detected in M-bands in vitro despite the presence of this residual Sarkosyl. Is the- only effect of glycolipids that of detoxifying the system by removing Sarkosyl? One important observation with the M-band system argues against this simple explanation. Phospholipids, which do not affect DNA synthesis in pneumococci even at high concentrations (see Table 1), also interact with Sarkosyl, although to a slightly lesser extent. This is shown in Fig. 7, where the lysis of phospholipid and glycolipid vesicles was followed as a function of the Sarkosyl concentration. It can be seen that glycolipid vesicles were lysed approximately 30% more efficiently than were phospholipids at the highest concentration 4

.3 IL

a

I.-

2 a

0Q23 046 OA 0os2 155 sarkosyl added (mg/mi) FIG. 6. Effects of Sarkosyl and glycolipid addition on DNA synthesis in M-bands. Quantities of 100 y1 of M-band (420 pg ofprotein per ml), obtained by lysing the cells with 0.075% Sarkosyl, were preincubated with (A) or without (0) glycolipids (6.53 pmol of glucose equivalent per ml), for 25 min at 30°C. At the end of preincubation, 40 ul of the DNA synthesis assay mixture described in the text containing 43.5 pmol of [3HJTTP was added with different amounts

of Sarkosyl. The samples were incubated at 300C for 20 min. Samples of 30 Ml were pipetted onto Whatman 3MM filters and then processed as described in the text for determination of acid-insoluble radioactivity.

VOL. 135, 1978 Or

25

-

c

I S

61

.%

85

DNA POLYMERASE ACTIVITY AND GLYCOLIPIDS

50

75

100

2 6 4 sarkosyl (mgIml) FIG. 7. Lysis of phospholipid (0) and glycolipid (A) vesicles as a function of Sarkosyl concentration. Lysis is expressed as the percent loss of the total optical density at 450 nm of the vesicle preparation emerging in the void volume of a Sepharose column as described in the text.

of Sarkosyl added. However, at the concentration used to lyse the cells (1.5 mg/ml), there was only a 10% difference in efficiency. Effects of glycolipids on NEM-sensitive DNA polymerases partially purified from cell lysates. Because of the structural complexity of the M-band and the possibility of indirect effects of glycolipids, an attempt was made to show a specific effect of glycolipids in a purified DNA-synthesizing system. Because it appeared from Fig. 5 that glycolipids were affecting DNA polymerase activity both sensitive and insensitive to the sulfhydryl inhibitor NEM, we attempted to isolate the former activity because of its possible relationship to replication rather than to repair (11, 15). It was recognized that due to the lack of mutants deficient in repair activity (presumably the most abundant DNA polymerase activity), it might be difficult to separate the NEM-sensitive activity, because in other organisms these activities are present in low amounts. Nevertheless, procedures have been devised by others to separate all of the polymerases from other wild-type organisms (11, 15), and these procedures, modified for pneumococci, were followed in this study. When cells were lysed with 0.2% Triton X in high salt and the lysate was centrifuged at 20,000 rpm for 1 h as described above, it was possible to obtain a supernatant that contained activity which was inhibited by the sulfhydryl inhibitor NEM and which had very little endogenous activity (i.e., activity with no added template). NEM inhibits the activity of polymerase HI- and HI-type enzymes but not polymerase I in B. subtilis and E. coli (11, 15). Table 2 shows that addition of template [activated calf thymus

DNA, M-band DNA, or oligo(dT)jo-poly(dA)] stimulated the level of DNA synthesis in the supernatant considerably. NEM at 9 mM inhibited the activity of the supematant toward activated calf thymus DNA and M-band DNA but had no effect on activity when the artificial template was used. These data suggest that the supernatant is enriched in DNA polymerases inhibited by NEM and that this activity is displayed mainly with activated calf thymus DNA and M-band DNA, whereas oligo(dT)-poly(dA) can be used more efficiently by a polymerase that is insensitive to NEM. The pellet contained high endogenous activity that was not inhibited by NEM. Therefore, it was not investigated further due to the interest in the other polymerases. Nevertheless, it should be pointed out that some NEMinsensitive activity, which had to be separated from the NEM-sensitive activity, was still present in the supematant. Table 3 shows the partial purification of DNA polymerase activities from cell lysates of pneumococci. After the Triton X supernatant from the cell lysate was chromatographed on DEAEcellulose as described above, two overlapping peaks for DNA polymerase activity were observed, using activated calf thymus DNA and oligo(dT)-poly(dA) as templates. These templates were chosen to differentiate between DNA polymerase activity that was largely inhibited by NEM (calf thymus DNA) and activity that was largely insensitive to the NEM [(oligo(dT)-poly(dA)]. The main peaks of the TABLE 2. Distribution of DNA polymerase activity in the supernatant and pellet after treatment of cells with Triton X and KCla Template

NEM +

Activated DNA Activated DNA M-band DNA M-band DNA

+ + +

pmol of [3HITTP incorporated in:

Supernatant

Pellet

0.28 0.29 8.2 6.9 5.11 4.02 0 9.44 9.41

2.6 2.6 3.63 3.64

0.53 0.58 0.63 a A 50-jil quantity of enzyme (supernatant or pellet adjusted to 500 ,ug of protein per ml) was diluted to 95 1p in the assay mixture for DNA synthesis; 20-pl samples were precipitated on Whatman 3MM filters after a 15-min incubation at 30°C. Endogenous template activity was determined with all the deoxynucleoside triphosphates (rows 1 to 6) and with labeled TTP only (rows 7 to 9). NEM was added at a final concentration of 9 mM under conditions on which DTT was omitted. As stated in the text, omission of DTT in controls reduced activity approximately 5 to 10%. The values have been normalized for this slight loss of activity.

Oligo(dT)-poly(dA) Oligo(dT)-poly(dA)

86

ZERIAL, GELMAN, AND FIRSHEIN

J. BACTERIOL.

TABLE 3. Partial purification ofDNA polymerase activities from cell lysates ofpneumococcia Enzyme activity assayed with Fraction

Total protein

(,Lg)

Activated calf thymus DNA Total activity (units) b

100,000

152,000

DEAE-cellulose Fractions 28 to 32 Fractions 33 to 38

10,000 20,000

20,000 33,200

Phosphocellulose I Enzyme a Enzyme b Enzyme c

130 50 1,200

Lysate

6,975.8 2,010 2,736

Sp act

(units/pg) 1.52

Oligo(dT),o-poly(dA) Total activity (units)

Sp act

179,000

(units/ug) 1.79

2.0 1.66

56,100 465,200c

5.61 23.26

53.66 40.2 2.28

2,678 1,750 88,200

20.6 35.0 73.5

Phosphocellulose II 33.3 1,040 31 170 5,230 Enzyme a 70 845 108 12 1,295 Enzyme b The procedures are described in the text. b One unit is defined as picomoles of TMP incorporated into DNA in 10 min at 30°C. The values for the first phosphocellulose run are derived from a composite analysis of the two runs shown in Fig. 8, one obtained from DEAE-cellulose fractions 28 to 32 and the other obtained from DEAE-cellulose fractions 33 to 38. 'The dramatic increase in activity for this template probably reflects the removal of an inhibitor from the fractions. a

first activity (fractions 28 to 32) were eluted at about 0.17 M NaCI, whereas the main peaks of the second activity (fractions 33 to 38) were eluted at about 0.23 NaCl. It can be seen that, although both templates were used in each fraction, there was a preference for activated calf thymus DNA by polymerases present in the first fraction, whereas polymerases from the second fraction preferred the oligo(dT)-poly(dA) template to a greater extent (in terms of specific activity). Phosphocellulose chromatography was used to further resolve the DNA polymerases eluted from the DEAE-cellulose columns. Each fraction was chromatographed on two separate phosphocellulose columns as described above. Three peaks of activity were present in both eluates, but there were significant quantitative differences with respect to the efficiency for each template (Fig. 8). From a composite analysis of two runs, it was determined that the first peak (enzyme a), which was eluted at a low potassium phosphate concentration, preferred activated calf thymus DNA as a template but used oligo(dT)-poly(dA) to a lesser extent. The second peak (enzyme b) also used these two templates, but the calf thymus DNA template was only slightly more efficient than the artificial one. Finally, the last activity (enzyme c) used the artificial template 10 times more efficiently than it used activated calf thymus DNA. Table 4 shows the effects of NEM on the activities of the three polymerase peaks eluted

from the phosphocellulose column. When activated calf-thymus DNA was used as a template, it can be seen that enzyme a was inhibited by 60%, enzyme b was inhibited by 40%, and no inhibition was observed for enzyme c. From these results, the peak fractions of enzymes a and b (derived originally from both DEAE-cellulose fractions) were pooled and rechromatographed on a second phosphocellulose column (Table 3) to achieve almost a threefold increase in specific activity for enzymes a and b when activated calf thymus DNA was used as the template. However, there was only a 1.5-fold increase in specific activity in enzyme a when oligo(dT)-poly(dA) was used as the template and a 2-fold increase when enzyme b was assayed with the artificial template. Because enzyme a was purified to the greatest extent, we performed additional experiments to characterize it further. The sedimentation coefficient of enzyme a in comparison to the sedimentation properties of bovine liver catalase (82o,w = 11.4) was 6.95. The concentration of KCl yielding optimum activity with activated calf thymus DNA was 20 mM (data not shown). On the basis of four different characteristics (inhibition by NEM, early elution from phosphocellulose, sedimentation characteristics, and optimum KCI concentration), we tentatively conclude that DNA polymerase a from pneumococci has properties similar to those of DNA polymerase III from other organisms (11, 15). Figure 9 shows the effects of glycolipids on

VOL. 135, 1978

DNA POLYMERASE ACTIVITY AND GLYCOLIPIDS

the activities of enzymes a and b. It can be seen that the former polymerase was stimulated to a far greater extent than the latter polymerase, reaching a plateau at a concentration of 1.4 ,mol of glucose equivalent per ml. The activity of enzyme b was stimulated only slightly, even at high concentrations of the glycolipids. It should be noted, however, that the stimulation by the glycolipids of the purified polymerases was less than that seen in the M-band (compare Fig. 2),

87

TABLE 4. Effects of NEM on DNA polymerase activities eluted from phosphocellulose columnsa Polymerase

NEM

pmolrated of TMP in 10incorpomin

a

+

0.25 0.67

b

+

0.60 1.01

0.23 0.21 a A 10-,ul quantity of NEM (at a final concentration of 5 mM) was added to 50 ,lA of enzyme a, b, or c (containing, respectively, 0.3, 0.6, and 2.0 ,Lg of protein per ml) in 0.02 M Tris, pH 7.5-20% glycerol, containing the remaining assay components described in the text except for DTT, which was omitted. Activated calf thymus DNA was used as the template. After a 10min incubation at 30°C, a 40-,lI sample was precipitated onto Whatman filters, and radioactivity was assayed as described in the text. c

+

suggesting the involvement of other factors in the latter stimulation. Finally, when the glycolipids were treated with alkali, the stimulatory effects on both polymerases disappeared completely as in the M-band experiments, suggesting that the structural integrity of the glycolipids was important in inducing the positive response. DISCUSSION The present results have shown that glycolipids stimulate the rate and extent of DNA synthesis in vitro in Sarkosyl lysates and in Mbands derived from these lysates. The stimulation occurs mainly in the DNA-membrane complex and only slightly in the remaining cell extract (top fraction). Phospholipids do not exert any such stimulatory effects. It appears that the i DC gZO frOf Ion glycolipids affect more than one DNA polymerFIG. 8. Phosphocellulose chromatography of ase activity in the M-band and, in particular, a DNA polymerases eluted from the DEAE-cellulose polymerase that exhibits a number of charactercolumn. (A) Chromatography ofpooled DEAE-cellu- istics of polymerase III described in other orgalose fractions 28 to 32 (10 mg of protein) on a 10-ml nisms. This was shown by experiments in which phosphocellulose column, 100-ml phosphate gradient (0.02 to 0.5 M). Samples of 10 jil of each fraction were glycolipids stimulated DNA polymerase activity diluted to 95 fld in the assay mixture for DNA synthe- partially purified from cell lysates in the absence sis. A 75-,ul portion was precipitated onto Whatman of any Sarkosyl. As to why glycolipids are not 3MM filters after a 10-min incubation and assayed able to stimulate polymerase activity to any for radioactivity as described in the text. (B) Chro- great extent in the top fraction, it may be (i) matography ofpooled DEAE-cellulose fractions 33 to that inhibitors are present or (ii) that polymer38 (18 mg of protein) on a 20-ml phosphocellulose ases (such as enzyme b) present in the top fraccolumn, 200-ml phosphate gradient (0.02 to 0.5 M). tion are not enhanced to any great extent by Samples of 10 til of each fraction were diluted to 95 glycolipids. We are currently investigating these ,ul in the assay mixture for DNA synthesis. A 75-jil portion was precipitated onto Whatman 3MM filters possibilities. It is interesting to point out that three DNA after a 10-min incubation and assayed for radioactivity as described in the text. Symbols: 0, activity polymerase activities were extracted from pneutoward activated calf thymus DNA; 0, activity to- mococci, the first such demonstration outside of ward oligo(dT)-poly(dA). E. coli and B. subtilis. Thus far, only the polym-

88

ZERIAL, GELMAN, AND FIRSHEIN

J. BACTERIOL.

pected that if the decay in activity of M-bands preincubated without glycolipids was due to re9 12.0 sidual Sarkosyl, the presence of glycolipids throughout preincubation would simply permit J 100 a restoration of the amount lost without glycolipids (about 40%), not the 400% stimulation -80 observed. It is possible that glycolipids create an envi60 ronment in the M-band favorable for the enzymes and factors involved in DNA synthesis. L 40 They do this by substitution of residual Sarkosyl molecules that remain in the M-band after dik! 20 Up-o " -~~~~~o alysis and, as a result, restore lipid domains favorable for the activity of DNA polymerases OM L16 174 2.32 Z9 that may exist in vivo, but are altered or repM Glucose q /ml moved by Sarkosyl. There are a variety of enFIG. 9. Effect ofglycolipid addition on DNA synzymatic systems associated with membranes thesis with enzymes a and b. Quantities of 10 u1l of that are inactivated or inhibited when lipids are glycolipid vesicles or of glycolipid vesicles treated removed by addition of detergents (10, 18). Only with alkali as described in Table 1, footnote a, were the appropriate lipid or lipids will restore or added to 50 u1 of enzyme a or b (0.08 and 0.18 pg of stimulate activity, showing that the effect is not protein per ml, respectively) from the second phosphocellulose chromatography (see Table 3) in 0.02 M due simply to detoxification by the lipids but to Tris, pH 7.5-2 mM mercaptoethanol-20% glycerol, at the restoration or repletion of the proper lipid the concentrations shown in the figure. Activated calf environment (18). Nevertheless, a question may still be raised. thymus DNA was used as template. After 30 min of DNA synthesis (the reaction is linear for at least 60 Why does the presumed restoration of the apmin), a sample of 40 Id was precipitated onto What- propriate lipid domains by glycolipids not enman 3MM filters (see text). The results are expressed hance DNA polymerase activity as much in the as percent stimulation. When no glycolipid vesicles purified replication system? It is possible that were present, 0.75 pmol of TMP or 1.58 pmol of TMP environment in the DNA-membrane comthe was incorporated into DNA per filter with enzyme a plex is more favorable to such activity than the or b, respectively. The concentration ofglycolipids is expressed as micromole of glucose equivalents per environment in the purified sample. It is well milliliter. Symbols: 0, enzyme a; 0, enzyme b; E, known that less-purified enzyme preparations can contain factors that stimulate enzyme activenzyme a or b with alkali-treated glycolipids. ity which are removed by purification techniques. In particular, if the DNA-membrane erase III-like enzyme a has been characterized to any great extent. Further work is needed to complex with which we are working represents, ascertain the identity of the other polymerases at least in part, the natural configuration of the replication complex in vivo, it is reasonable tfiat (bandc). The enhancement of DNA polymerase activ- events affecting DNA synthesis should occur to ity in the DNA-membrane complex by the gly- a greater extent in this complex than in a puricolipids is greater than that observed in the fied replication system which does not contain partially purified system, suggesting an interac- all of the factors required to provide an optimum tion of glycolipids with other factors in the M- environment for such synthesis. What these factors are, in addition to glycolipband complex. One of these factors is Sarkosyl, whose inhibitory activities were overcome by ids, will be the subject of future reports. glycolipids to a significant extent, as shown in ACKNOWLEDGMENT Fig. 6. However, the interaction is more than a This work was supported by National Science Foundation simple neutralization of a detergent by a lipid because (i) phospholipids, although interacting grant PCM 76-81351. with Sarkosyl and therefore potentially capable LITERATURE CITED of detoxifying the system, do not stimulate DNA 1. Bartlett, G. RI 1959. Phosphorus assay in column chrosynthesis, and (ii) as Fig. 3 shows, the decay in matography. J. Biol. Chem. 234:466-468. subsequent synthetic activity of M-bands prein- 2. Bligh, E. G., and W. J. Dyer. 1959. A rapid method of total lipid extraction and purification. Can. J. Biochem. cubated without glycolipids is not only prePhysiol. 37:911-917. vented if preincubation occurs with glycolipids, 3. Brundish, D. E., N. Shaw, and J. Baddiley. 1967. The but the subsequent enhancement is almost 400% phospholipids of Pneumococcus I-192R, A.T.C.C. 12213. greater than that observed in non-preincubated Some structural rearrangements occurring under mild conditions. Biochem. J. 104:205-211. controls without glycolipids. It would be ex940

I-

0

0

0

_r

..o

._.

0

n

VOL. 135, 1978

DNA POLYMERASE ACTIVITY AND GLYCOLIPIDS

4. Dittmer, J. C., and M. A. Wells. 1969. Quantitative and

qualitative analysis of lipids and lipids components. Methods Enzymol. 14:545-546. 5. Firshein, W. 1972. The DNA-membrane fraction of Pneumococcus contains a DNA replication complex. J. Mol. Biol. 70:383-398. 6. Firshein, W. 1973. In situ activity of enzymes on polyacrylamide gels of a deoxyribonucleic acid-membrane fraction extracted from pneumococci. J. Bacteriol. 118: 1101-1110. 7. Firshein, W. 1976. Two membrane sites for DNA synthesis in Pneumococcus. Mol. Gen. Genet. 148:323-325. 8. Fralick, J. A., and K. G. Lark. 1973. Evidence for the involvement of unsaturated fatty acids in initiating chromosome replication in Escherichia coli. J. Mol. Biol. 80:459-475. 9. Ganesan, A. T., and J. Lederberg. 1965. A cell membrane bound fraction of bacterial DNA. Biochem. Biophys. Res. Commun. 18:824-35. 10. Garland, R. C., C. F. Con, and H. W. Chang. 1974. Relipidation of phospholipid-depleted microsomal particles with high glucose 6-phosphatase activity. Proc. Natl. Acad. Sci. U.S.A. 71:3805-3809. 11. Gass, K. B., R. L. Low, and N. R. Cozzarelli. 1973. Inhibition of the synthesis of deoxyribonucleic acid in bacteria by hydroxyphenylazopyrimidines. Proc. Natl. Acad. Sci. U.S.A. 70:103-107. 12. Greene, M., and W. Firshein. 1976. Role of deoxyribonucleic acid ligase in a deoxyribonucleic acid membrane fraction extracted from pneumococci. J. Bacteriol. 126:777-784. 13. Ishizuka, I., and T. Yamakawa. 1968. On the lipids of Diplococcus pneumoniae type I-R/44. Separation of glycolipids from phospholipids by silicic acid chromatography. Jpn. J. Exp. Med. 38:75-79. 14. Knippers, R., and W. Stratling. 1970. The DNA replicating capacity of isolated E. coli cell wall membrane complexes. Nature (London) 226:713-717. 15. Kornberg, T., and M. Gefter. 1974. Deoxyribonucleic acid polymerase III. Methods Enzymol. 29:22-26. 16. Liebowitz, P. S., and M. Schaechter. 1975. The attachment of the bacterial chromosome to the cell membrane. Int. Rev. Cytol. 141:1-28.

89

17. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall. 1951. Protein measurements with the Folin phenol reagent. J. Biol. Chem. 193:265-275. 18. Lu, A. Y. M., and W. Levin. 1975. The resolution and reconstitution of the liver microsomal hydroxylase system. Biochim. Biophys. Acta 344:205-210. 19. Martin, R. G., and B. N. Ames. 1961. A method for determining the sedimentation behaviour of enzymes: application to protein mixtures. J. Biol. Chem. 236:1372-1379. 20. Osborn, M. J. 1971. The role of membranes in the synthesis of macromolecules, p. 343-393. In L. Rothfield (ed.), Structure and function of biological membranes. Academic Press Inc., New York. 21. Parker, D. L, and D. A. Glazer. 1974. Chromosomal site of DNA-membrane attachment in Escherichia coli. J. Mol. Biol. 87:153-168. 22. Radln, N. S., F. B. Lavin, and J. R. Brown. 1955. Detennination of cerebrosides. J. Biol. Chem. 217: 789-796. 23. Rottem, S., and S. Razin. 1973. Membrane lipids of Mycoplasma hominis. J. Bacteriol. 113:565-571. 24. Shaw, N. 1968. The detection of lipids on thin-layer chromatograms with the periodate-Schiff reagents. Biochim. Biophys. Acta 164:435-436. 25. Shaw, N. 1970. Bacterial glycolipids. Bacteriol. Rev. 34:365-377. 26. Shaw, N., and J. Baddiley. 1968. Structure and distribution of glycosyl diglycerides in bacteria. Nature (London) 217:142-144. 27. Tremblay, G. Y., M. J. Daniels, and M. Schaechter. 1969. Isolation of a cell membrane-DNA-nascent RNA complex from bacteria. J. Mol. Biol. 40:65-76. 28. Vorbeck, M. L, and G. V. Marnetti. 1965. Separation of glycosyl diglycerides from phosphatides using silicic acid column chromatography. J. Lipid Res. 6:3-6. 29. Wells, M. A., and J. C. Dittmer. 1963. The use of Sephadex for the removal of non-lipid contaminants from lipid extracts. Biochemistry 2:1259-1263. 30. Wickner, W., and A. Kornberg. 1974. A holoenzyme form of deoxyribonucleic acid polymerase HI. Isolation and properties. J. Biol. Chem. 249:6244-6249.