Jan 26, 1976 - cleic acid (DNA)-membrane fraction extracted from the cells. A variety of techniques .... fluid (Nuclear-Chicago) in a Mark I Nuclear-Chi- .... previously or treated to elute the bound material. For elution, the ..... titiously entrapped. We favor the ..... for the detection, during gel electrophoresis, or sedi- mentation ...
JouRNAL o BACTRuoWoY, July 1976, p. 14-23 Copyright C 1976 American Society for Microbiology
Vol. 127, No. 1 Printed in U.S.A.
Role of Polyadenylic Acid in a Deoxyribonucleic AcidMembrane Fraction Extracted from Pneumococci WILLAM
FIRSHEIN,* BEN MEYER, ELLIOT EPNER,
AND
JANE VIGGIANI
Department ofBiology, Wesleyan University, Middletown, Connecticut 06457 Received for publication 26 January 1976
After the addition of radioactive polyadenylic acid to cell suspensions of part of the radioactivity becomes associated with a deoxyribonucleic acid (DNA)-membrane fraction extracted from the cells. A variety of techniques show that a portion of this associated radioactivity may represent oligoadenylates complexed to DNA, probably as part of a ribonucleic acid (RNA) component. Polyadenylic acid, which had previously been shown to enhance DNA synthesis in cell suspensions (Firshein and Benson, 1968), also enhances the extent of DNA synthesis by the DNA-membrane fraction in vitro under specific conditions of concentration and conformation. The mechanism of action of this enhancement may be related to the ability of oligoadenylates to increase the number of initiation sites for DNA replication by stimulating the production of an RNA primer, thus providing additional 3'-OH groups with which DNA polymerase can react. pneumococci,
mained obscure until recently, when we observed a correlation between the ability of poly(A) to enhance DNA synthess in cell suspensions and the localization of the increase in a DNA-membrane fraction extracted from the same cell population (13). This DNA-membrane fraction has been found to contain a DNA replication complex and may be the site of initiation of DNA synthesis in pneumococci (10, 11; W. Firshein, submitted for publication). A variety of possibilities exist to explain the function of poly(A). For example, poly(A) could enhance DNA replication by acting as a primer. By hybridizing with clusters of thymine nucleotides on the DNA template, poly(A) or its oligomers could provide open 3'-OH groups to permit extension of the DNA chain. However, since it has been proposed that RNA may act as a primer for DNA replication by the same mechanism (i.e., provide open 3'-OH groups) (3, 18, 29; Firshein, submitted for publication), poly(A) could be involved in the fornation of this macromolecule either as an initiator or as a structural unit. Another possibility is that poly(A), acting as a polyanion, could produce conformational changes in the cell membrane, which could act to trigger initiation of DNA replication (2). Finally, the possibility that poly(A) affects DNA synthesis through the stimulation of cAMP activity (2) or by increasing energy levels in the cells (after degradation) must also be considered. We have-concentrated on the first possibility for a variety of reasons. First, other studies (3,
The role of poly(A) in cellular metabolism has become increasingly important in the past decade (4, 8, 17, 21, 23, 25, 26). Most of the interest has centered around the finding that poly(A) is part of most classes of mRNA in eukaryotic cells (9, 21, 23), but not ribosomal or transfer RNA. It has been proposed that poly(A) plays a role in the migration of messenger RNA from the nucleus to the cytoplasn (21, 23). Poly(A) polymerases have been found in animal and plant cells including mitochondria (4, 17, 19). In bacteria, the role of poly(A) is much less clear. Poly(A) has been detected in several bacterial species (1, 6, 7, 26, 28) and inhibits RNA polymerase activity in vitro (1, 7). The effects of poly(A) in bacteria may not only be restricted to RNA metabolism. Templates containing poly(A) can be used with high efficiency by DNA polymerase I (20) and even may be used as a primer by this enzyme as well as by the RNA-dependent DNA polymerases of tumor viruses (22). We found several years ago that poly(A) and poly(C) could enhance DNA synthesis in cell suspensions of encapsulated pneumococci (12) but the effects of the forner were greater. Poly(A) was taken up by the cells, and, of the amount incorporated, at least 50 to 80% was in the form of undegraded poly(A) or its oligomers (mostly the latter) (14). Although an inducing effect of poly(A) on a specific deoxyribonucleotide kinase (dGMP) was demonstrated (12), the mechanism of action of the polynucleotide re14
VOL. 127, 1976
POLY(A) IN A PNEUMOCOCCAL DNA-MEMBRANE FRACTION
18, 29; Firshein, submitted for publication) have shown that transcription may be required for DNA synthesis. RNA-DNA hybrids have been detected during early stages of DNA synthesis, and antibiotics that inhibit transcription have been found to inhibit, in part, DNA synthesis under appropriate conditions. MATERIALS AND METHODS
Organism, preparation of cell suspensions, and extraction of the DNA membrane fraction. The organism (an encapsulated strain of type III Diplococcus pneumoniae A66,) and preparation of cell suspensions to dissociate DNA synthesis from multiplication and obtain preparative amounts of the DNAmembrane fraction, as well as the extraction of such a fraction through the use of the detergent sodium lauroyl sarkosinate (Sarkosyl, Geigy Chemical Corp.), were all described in detail previously (9, 10). DNA synthesis in the cell suspensions was enhanced by adding a special supplement consisting of poly(A) (100 ug/ml) plus all eight of the naturally occurring deoxyribonucleosides and -tides (200 ,ug! ml each of thymidine, deoxycytidine, deoxyadenosine, deoxyguanosine, and their phosphorylated derivatives). The latter provided basic precursors for DNA synthesis (9), whereas poly(A) was required for the enhancement of DNA synthesis to occur. Without the polynucleotide, no increase over control levels was observed (12). [Poly(A) refers to the undegraded native molecule of approximately 335 nucleotides in length. As will be shown, it is possible that the stimulatory effects of poly(A) are actually due to oligoadenylates of unknown molecular weight produced by nucleases in the DNA-membrane fraction during incubation in vitro. However, for simplicity, unless specified otherwise, we will refer to the term "poly(A)" in the text as including the oligoadenylates.] Macromolecular analysis. Bulk DNA and protein in the DNA-membrane fraction were assayed by the methods of Burton (5) and Lowry et al. (24), respectively, or in the case of DNA by steady-state labeling with labeled deoxycytidine as described previously (10). RNA was estimated by steady-state labeling procedures using labeled uridine. In the case of steady-state labeling, samples were treated with an equal volume of cold trichloroacetic acid (10%, wt/vol) to precipitate macromolecules, and the precipitates were transferred quantitatively to membrane filters (Millipore type HA, 0.45 Mm). The filters were washed twice with 5-ml portions of 5% trichloroacetic acid and twice with 5-ml portions of cold water. They were dried at 60 C for 2 h, placed into scintillation vials, and assayed for acid-insoluble radioactivity with a toluene-based scintillation fluid (Nuclear-Chicago) in a Mark I Nuclear-Chicago scintillation counter. In some cases, cell suspensions were pulsed with '4C-labeled 2-deoxycytidine (1.0 MCi/ml) for 15 to 30 s. Incorporation into the cells (10 ml) was stopped by adding an equal volume of ice-cold Tris buffer (0.01 M, pH 8.1) containing 0.02 M sodium cyanide (10).
15
In vitro synthesis of DNA by the DNA-membrane fraction. After extraction of the DNA-membrane fraction, approximately 5 to 15 ml was dialyzed for 24 h at 4 C against two changes of a buffer consisting of Tris (0.02 M), glycerol (final concentration, 20%), and albumin (bovine serum albumin fraction V) (final concentration, 0.1%) (final pH, 8.1). The dialysate was concentrated by rolling the bags in Sephadex G-200 for 20 to 25 min at 4 C. To 0.3 ml of the dialyzed concentrated fraction (140 to 180 ,ug of protein and 25 to 30 ug of DNA) the following were added (total, 0.4 ml): Tris (0.2 M, pH 8.6) (4 umol); TTP, dCTP, and dGTP (0.04 Amol); [8-'4C]dATP (6.5 x 105 counts/min per ,umol) (0.04 ,umol); MgCl2 * 6H20 (7 mM); dithiothreitol (1 mM); and NAD (0.17 ,umol). The NAD was added to enhance DNA ligase activity that was present in the DNAmembrane fraction (11). Poly(A) was added in concentrations of 2 to 28 ,ug/0.4 ml. In some cases [3H]poly(A) (2.0 x 105 counts/min per umol, 1.6 ,umol [in nucleotide equivalent]) was added. After incubation at 37 C for various periods of time, the reaction was stopped either by adding 0.4 ml of cold 10% trichloroacetic acid in 0.02 M sodium pyrophosphate or by adding a solution (9.6 ml) of 27% sucrose (wt/vol), 20 mM KCN, and 10 mM EDTA in 0.015 M NaCl plus 0.0015 M sodium citrate (SSC-sucrose). In the first case with trichlorooacetic acid-pyrophosphate, acid-insoluble radioactivity was determined as described above, whereas in the second case with SSC-sucrose nucleic acids were extracted as described below. Purification of nucleic acids. Nucleic acids were extracted from DNA-membrane fractions that either were not treated further or were first incubated with DNA precursors and poly(A) (labeled or unlabeled) in the in vitro assay for DNA synthesis described above. If bulk DNA was to be prepared from the DNA-membrane fraction [derived from cells incubated without poly(A)], the extract was treated with pancreatic ribonuclease (10 Ag/ml for 30 min at 37 C). Pronase (1 mg/ml, previously autodigested at 37 C for 1 h) was added, and the DNAmembrane fraction was incubated for 4 h at 37 C. An equal volume of freshly distilled water-saturated phenol was added, and the resulting solution was rolled for 30 min at 60 rpm in a rotary evaporator (Rinco). After centrifugation to clarify the layers (5,000 rpm at 4 C for 10 min), the aqueous layer was removed and the phenol procedure was repeated. Twice as much anhydrous ether was then combined with the aqueous layer and rolled for 10 min in the evaporator, and the ether sequence was repeated. The resultant phenol-free aqueous layer was dialyzed against 2 liters of a buffer containing 0.02 M Tris (pH 8.0) and EDTA (0.02 M) with several changes over a period of 3 days. The dialysate was concentrated by rolling the dialysis bags in powdered Sephadex G-200 and analyzed for nucleic acid and protein content as described above or subjected to density gradient centrifugation as described below. Equilibrium centrifugation of nucleic acids in density gradients. Cs2SO4 (3.1 g) was added to 4.2 ml of nucleic acid solution, which was either boiled for 5
16
FIRSHEIN ET AL.
min and quick-cooled or incubated with 0.3 N NaOH for 2 h at 37 C, followed by neutralization with HCl. Centrifugation was for 36 h (15 C) at 36,000 rpm in a Spinco SW50.1 rotor. Tubes were punctured from the bottom, and 0.1-ml fractions were precipitated with an equal volume of cold 10% trichloroacetic acid. Buoyant density was determined on Cs2SO4 solutions by using a Bausch and Lomb refractometer and an internal marker of pneumococcal DNA. Nitrocellulose binding of poly(A). Poly(A) can bind reversibly to cellulose acetate (Millipore) filters (23) and is dependent upon ionic strength and pH and independent of temperature. Nucleic acids extracted from DNA-membrane fractions originally derived from cell suspensions incubated with labeled poly(A) and a labeled DNA precursor ['4C]deoxycytidine, were extensively dialyzed against a buffer consisting of Tris (0.01 M, pH 7.5), MgCl2-6H2O (0.01 M), and KCI (0.2 M). After dialysis, the nucleic acid solution was diluted 10-fold with the same buffer, slowly passed through a Millipore filter (0.45 ,um) that had been soaked previously in the same buffer, and washed twice, also with this buffer. Subsequently, the filter was either assayed for acid-insoluble radioactivity as described previously or treated to elute the bound material. For elution, the filter was shaken in 0.5 to 1 mM SDS in 0.1 M Tris buffer (pH 9.0) at 0 C for 1 h. The eluate was characterized by heating and sedimentation through Cs2SO4 density gradients as described in the previous section. Detection of poly(A) sequences by hybridization. Poly(A) can hybridize to poly(U). The molecular weight of the complex is greater than that of poly(U) alone, and the sedimentation properties of the latter are therefore altered in a sucrose gradient (16). The most important variable in this technique is the concentration of poly(U) that must be added in excess of the A sequences present to prevent a spurious polydisperse distribution of radioactivity in the sucrose gradient (16). In preliminary experiments, we found that 300 ,ug of poly(U) was adequate for this purpose. A 7 to 25% (wt/vol) sucrose gradient was prepared in a solution of Tris buffer (0.05 M, pH 7.6), 0.15 M NaCl, 0.01 M EDTA, and SDS (0.5%). Purified nucleic acid (80 ,ug/0.09 ml), derived originally from a DNA-membrane fraction extracted from cell suspensions incubated with poly(A), was mixed with ['4C)poly(U) (0.67 mCi/300 ug) at room temperature and layered within 10 min onto a 5.2-ml gradient in a cellulose acetate tube. Control nucleic acids obtained from cells incubated without poly(A) were also mixed with ['4C]poly(U) and layered. The gradients were centrifuged in an SW50.1 rotor of an L-2 Spinco ultracentrifuge for 2.5 h at 45,000 rpm at 4 C. Fractions (0.1 ml) were collected and assayed for acid-insoluble radioactivity. 32p transfer experiments. The junction between an RNA and DNA base in an RNA-DNA hybrid molecule can be determined by observing the transfer of 32P from a deoxyribonucleotide to a ribonucleotide (15). The DNA-membrane fraction was extracted from cell suspensions incubated without any poly(A) supplement. In vitro DNA synthesis was performed as described in a previous section, except
J. BACTERIOL.
that [32P]dCTP replaced nonlabeled dCTP and labeled dATP was replaced by nonlabeled dATP. Poly(A) was added at a concentration of 7 jug. After incubation for 5 min at 37 C, an equal volume of cold 10% trichloroacetic acid was added and the mixture was kept cold for 2 h. It was then minced thoroughly and washed twice with cold 5% trichloroacetic acid and twice with ethanol. The precipitate was treated with 0.3 ml of 0.3 N KOH for 16 h at 37 C to hydrolyze RNA [and poly(A) or its oligomers] to 2',3'ribonucleotides. The hydrolysate was neutralized with 6 u.l of 60% (vol/vol) perchloric acid, and the precipitate that formed (containing DNA) was removed by centrifugation. The supernatant was analyzed for 32p transfer by descending paper chromatography together with the four 2',3'-ribonucleotides as standards. Chromatography proceeded for 24 h with Whatman no. 1 paper and isobutyric acidwater-NH,OH-0.2 M EDTA (100:56:4:8). Strips were cut out and analyzed with an ultraviolet lamp (253.8 nm) and a radioactive chromatogram scanner (Vangard). Degradation of poly(A) in the DNA-membrane fraction. Labeled (3H) poly(A) (105 counts/min per 0.1 ml) was added to 0.3 ml of the DNA-membrane fraction and incubated for 5 or 10 min at 37 C in a solution containing HEPES buffer (0.01 M, pH 7.4) and albumin (0.01%). Trichloroacetic acid (equal volume of 10%) was added, the supernatants and precipitates were separated by centrifugation, and the latter were assayed for acid-insoluble radioactivity as described before. A control sample containing [3H]poly(A) was prepared but treated immediately with trichloroacetic acid. Activity was measured as the percentage of acid-insoluble radioactivity rendered soluble. The supernatants were evaporated under reduced pressure (to 0.2 ml), and 0.1-ml samples were chromatographed for 48 h on Whatman 3MM chromatography paper (descending) in a solvent consisting of propanol, NH4OH, and water (55:10:35). Labeled standards of poly(A) and adenylic acid were also chromatographed. Poly(A) does not migrate in this solvent system, whereas small oligomers of A will migrate according to their chain length. Adenylic acid will migrate with the solvent front. After chromatography, ultraviolet quench spots were located and paper strips were cut out as described in the previous section. One-inch (ca. 2.54cm) square sections were cut from the strips and placed into scintillation vials, and radioactivity was determined as described previously. Radioactive and other substances. All radioactive materials were purchased from Schwarz/Mann. These included '4C- and 3H-labeled precursors for DNA and RNA synthesis ([2-'4C]deoxycytidine, 20 mCi/mmol; [2-14C]uridine, 35 mCi/mmol), [14C]dATP (35 mCi/mmol), [3H]poly(A) (10 mCi/mmol), and [14CJpoly(U) (0.2 mCi/mmol of P). [a-32P]dCTP (4.0 mCi/mmol) was also obtained from this source. All nonlabeled compounds, including poly(A), all eight of the naturally occurring deoxyribonucleosides and -tides, their triphosphate derivatives, and NAD were obtained from either Sigma Chemical Co., Calbiochem, or Miles Laboratories. Abbreviations. The abbreviations used through-
POLY(A) IN A PNEUMOCOCCAL DNA-MEMBRANE FRACTION
VOL. 127, 1976
out are: poly(A), polyadenylic acid; poly(C), polycytidylic acid; poly(U), polyuridylic acid; RNA, ribonucleic acid; DNA, deoxyribonucleic acid; dGMP, deoxyguanosine 5'-monophosphate; cAMP, cyclic adenosine 3',5'-monophosphate; dTTP, deoxythymidine 5'-triphosphate; dCTP, deoxycytidine 5'-triphosphate; dGTP, deoxyguanosine 5'-triphosphate; dATP, deoxyadenosine 5'-triphosphate; Tris, tris(hydroxymethyl)aminomethane; NAD, nicotinamide adenine dinucleotide; EDTA, ethylenediaminetetraacetate; SDS, sodium dodecyl sulfate; and HEPES, N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid. RESULTS
Effects of poly(A) on DNA synthesis in the DNA-membrane fraction. To ascertain whether poly(A) had any effect on DNA synthesis initiated by the DNA-membrane fraction in vitro, various concentrations of poly(A) were added to solutions of the DNA-membrane fraction containing substrates and cofactors for DNA synthesis as described above. In addition, samples of poly(A) were boiled for 5 min before being added to the assay solution. The results (Fig. 1A and B) show that poly(A) had a stimulatory effect on the rate and extent of DNA synthesis in the DNA-membrane fraction but that the effect depended on the concentration and probable length and conformation of the added poly(A). Thus, there was a critical range of concentration that stimulated DNA synthesis (between 2.0 and 15 ,ug/ml), and it was necessary to subject the higher concentrations of poly(A) (over 8 jAg/ml) to boiling for 5 min for stimulation to occur (Fig. 1A). This latter treatment resulted in a significant decrease in vis6
6
A
B
5
5-
~~~4-
4-
x~~~~~~~ 00
2W
2~~~~
L)
1,g
poly AMTi
TIME (min.)
FIG. 1. Poly(A) effects on DNA synthesis in vitro by the DNA-membrane fraction as a function of concentration and time. (A) Assay solutions were incubated for 10 min at 37 C; *, poly(A) added without boiling; 0, poly(A) boiled for 5 min prior to addition. (B) 0, 8 pg ofpoly(A); *, no poly(A); 0, 22 pg of poly(A).
17
cosity of the added poly(A) and some degradation (not shown). Concentrations of poly(A) higher than 15 ,ig resulted in either no enhancement of DNA synthesis or an actual inhibition from control values. Maximum stimulation of DNA synthesis by the appropriate concentration of poly(A) (8 ,ug/ml) occurred at about 10 min of incubation after which time the stimulation decreased to levels that were only slightly higher than those of unsupplemented assay solutions (Fig. 1B). A higher concentration of poly(A) (22 ,ug/ml) significantly inhibited DNA synthesis from the control value even at early incubation times of between 2 and 4 min. Degradation of poly(A) by the DNA-membrane fraction in vitro and in vivo. Although the results suggested that the stimulatory effects on DNA synthesis were due to poly(A), it was not certain whether the effects were due to undegraded poly(A) or to oligoadenylates (high or low molecular weight) that could be produced during the 5- to 10-min incubation period by ribonucleases present in the DNA-membrane fraction. We found previously (14) that poly(A) was degraded in vivo (mostly to oligoadenylates) after incorporation. (i) In vitro. To ascertain the level of degradation in vitro, experiments were performed as described in Materials and Methods. After incubation of the DNA-membrane fraction and the in vitro synthesizing system with [3H]poly(A) for 5 and 10 min, trichloroacetic acid was added and acid-insoluble radioactivity was determined. The degree of radioactivity lost as compared with that of unincubated controls to which labeled poly(A) was added was taken as a measure of the degradation of poly(A). The results (Table 1) showed that approximately 38% of the initial acid-insoluble radioactivity was solubilized after incubation for 10 min. Chromatography of the acid-soluble supernatant with a control marker of adenylic acid and one of undegraded poly(A) showed that 55% of the degradation products were lowmolecular-weight oligoadenylates and 35% were adenylic acid. Thus, a total of approximately 87% of the added poly(A) either remained as such or was degraded to high- and low-molecular-weight oligomers of adenylic acid. Only 13% appeared as adenylic acid. When this latter concentration was added to in vitro synthesizing systems, alone or in combination with poly(A), no effects were noted (data not shown). Therefore, it can be concluded that the effects of poly(A) are most probably due either to undegraded poly(A) or to high- or lowmolecular-weight oligomers of adenylic acid.
18
J. BACTERIOL.
FIRSHEIN ET AL.
(ii) In vivo. The level of degradation of poly(A) in vivo after incorporation was determined previously (14). However, it was not known how much of the remaining undegraded poly(A) or its acid-insoluble oligomers was present in the DNA-membrane fraction. Table 2 shows the distribution of radioactivity derived from poly(A) in cell extracts after incorporation. It can be seen that, of the total radioactivity added to the cell suspension, approximately 12% was incorporated by the cells. Of this amount, approximately 35% consisted of acidinsoluble counts per minute, with 25% associated with the DNA-membrane fraction and 10% with the top fraction (the cell extract remaining
TABLE 1. Degradation of['4C]poly(A) by the DNA membrane fraction in vitro, Incubation time Expt no.
(min)
Acid-insoluble
(A) radio['4CJpoly activity (counts/ min)
% Degradation from
time zero
0
1 2
105,271 115,227
5
1 2
90,772 90,223
15.9 27.7
10
1 2
77,128 81,837
36.4 40.8
a After [14C]poly (A) was added to the DNA-membrane fraction, an equal volume of cold 10% trichloroacetic acid was added at 0, 5, and 10 min of incubation. Acid-insoluble radioactivity was then determined.
TABLE 2. Degradation of ['4C]poly (A) in vivo and distribution in the DNA membrane fraction" Recovery Total acid-insolState of cells
from cells (
uble radioactiv-
ity (counts/min) 950,460 115,201
Unwashed cells Thoroughly washed cells
100
DNA-membrane fraction
25
28,744
Top fraction
10
10,832
53 Acid-soluble extract 61,008b a ['4C]poly(A) (1.4 ,uCi/30 ,tg) was added to the cell suspensions and incubated for 15 min at 37 C. The cells were centrifuged, washed five times with cold TMK buffer (0.01 M Tris [pH 7.41, 0.01 M MgCl2, 0.1 M KCl), and either subjected to trichloroacetic acid extraction as described previously (14) or treated with Sarkosyl to extract the DNA membrane and top fractions. b Acid-soluble counts per minute.
after formation of the DNA-membrane fraction). The remaining radioactivity (53%) consisted of acid-soluble fragments derived from poly(A) or adenylic acid. It should be pointed out that some of those acid-soluble oligoadenylates could have been associated with the DNAmembrane or top fraction prior to extraction and washing with trichloroacetic acid. Association of poly(A) with nascent DNA in the DNA-membrane fraction in vivo and in vitro. As discussed in the Introduction, poly(A) could act either as a primer for initiation of DNA replication or as a component of the RNA, which is the actual primer. In either case, it should be possible to detect poly(A) bound to newly synthesized DNA. (i) In vivo. To test for in vivo binding, replicate cell suspensions were incubated with [3H]poly(A) for 15 min and then pulsed for 20 s with ['4Cldeoxycytidine. The DNA-membrane fraction was extracted from the cells, and nucleic acids were extracted from the complex with phenol as described previously. Nucleic acids were also extracted from a DNA-membrane sample that was steady-state labeled with [14C]deoxycytidine only. After extraction, the nucleic acid solution from the pulse-labeled sample was divided in half; one half was treated with alkali and the other half was subjected to boiling for 5 min as described previously. Both methods denature DNA in the extract. These two samples plus the steady-statelabeled sample (which was also heated for 5 min at boiling) were then centrifuged to equilibrium in three separate Cs2S04 gradients: one for the untreated pulse-labeled sample, one for the alkali-treated sample, and one for the steady-state sample. The profile of the steadystate sample was superimposed over the profiles of the other two samples (Fig. 2A and B). The experiment was repeated several times to ascertain the validity of this approach. In no instance was there any significant deviation from the observed patterns of centrifugation. According to other studies (Firshein, submitted for publication; 18, 27), the possible existence of an RNA [poly(A)I-DNA single-strand hybrid molecule can be demonstrated by the above method, providing that the molecular weight of the nascent DNA portion is sufficiently small. Under these conditions, the density of the hybrid is greater than that of single-stranded, long-term-labeled DNA or the hybrid pretreated with alkali to degrade the RNA [poly(A)I portion of the macromolecule. Figure 2 shows that pulse-labeled DNA and radioactivity derived from poly(A) banded in the gradient at a slightly heavier density than
POLY(A) IN A PNEUMOCOCCAL DNA-MEMBRANE FRACTION
VOL. 127, 1976 28r 241
28r A
24p B
201 - x>
20 .
arL)r
16
0
I
1
-X X
I'
x 12 z
I
-X-X
4 8 12 16 20 24 28 4 8 12 6 2024 TOP FRACTION NO.
FIG. 2. Sedimentation of denatured, pulse-labeled DNA in CsgSO4 density gradients after extraction from DNA-membrane fractions derived from cells incubated with labeled poly(A). (A) 0, Radioactivity from [3H]poly(A); 0, radioactivity from pulse-labeled [14CJDNA; x, radioactivity from steady-state-labeled ['4C]DNA. (B) 0, Radioactivity from pulse-labeled ['4C]DNA as in (A) but treated with alkali prior to sedimentation in the Cs2SO4gradient; x, radioactivity from steady-state-labeled [ IC]DNA.
19
tually is seen in the inset of Fig. 3. [3H]poly(A) added to the DNA extract prior to centrifugation did not band at the same density in the gradient. Almost all of the counts were recovered at the bottom of the tube. Although these results suggest that radioactivity derived from added poly(A) is somehow linked to newly synthesized DNA (in vivo and in vitro), two possibilities (of artifacts) exist that make further experimentation necessary. First, it is possible that poly(A) or RNA could complex with nascent DNA even under conditions ofheat denaturation and sedimentation in Cs2SO4, thus producing the results observed in Fig. 2 and 3. Second, it is possible that the radioactivity derived from poly(A) represented labeled adenylic acid incorporated into RNA after degradation of poly(A) in vivo rather than poly(A) or oligomers of A. The second possibility is assessed in the next two sections, and the first possibility is investigated in the last section. Origin of radioactivity from poly(A). Since poly(A) was degraded by nucleases both in vivo and in vitro, labeled adenosine or AMP could be incorporated into RNA, thus producing the banding pattern seen in Fig. 2 and 3. Two types of experiments explored these possibilities: (i) the reversible binding of poly(A) to nitrocellulose filters, and (ii) hybridization of poly(A) to poly(U). For nitrocellulose binding studies, the DNAmembrane fraction was extracted from cells in-
long-term-labeled DNA (Fig. 2A) or alkalitreated nasoent DNA (Fig. 2B). In addition, radioactivity from poly(A) was not detected in the alkali-treated sample. (ii) In vitro. To determine whether the association between poly(A) and DNA could be de32 tected in vitro, the DNA-membrane fraction 28R was incubated with [3HIpoly(A) in the synthesizing system described previously (including 24~ [14C]dATP). After incubation, the nucleic acids were extracted, denatured, and centrifuged to equilibrium in the Cs2SO4 density gradient. 0 The results (Fig. 3) show the same banding x 4 k! 1V~2 : characteristics of radioactivity from poly(A) z and DNA synthesized aftpr 5 min. The peak is very broad and not much different from bulk 16 24 32 40 X () 2 , y DNA but, as shown in a previous study (Firshein, submitted for publication), it is necessary to incubate only a very short time (30 s) in vitro a a a a . I . to detect the heavier-density RNA-DNA hy4 8 12 6 20 24 28 32 36 40 brid. After 60 s, enough DNA has been synthesized to shift the density significantly back to *TOP FRACTION NO. that of bulk DNA. However, the important obFIG. 3. Sedimentation of denatured DNA in servation is the banding of a significant amount Cs8204 density gradients synthesized in vitro in the of radioactivity derived from poly(A) with syn- presence labeled poly(A). Symbols: 0, radioactiv60% of the added ity derivedoffrom poly(A); thesized DNA. Approximately 0, DNA. Inset: Sedimentapoly(A) sedimented at the bottom of the gra- tion of denatured DNA in Cs2SO4 density gradients dient as free poly(A). That the complex be- synthesized in vitro in the absence ofpoly(A) but to tween poly(A) and DNA is not formed artifac- which poly(A) was added prior to centrifugation.
~8
I
I
20
FIRSHEIN ET AL.
J. BACTERIOL.
cubated with [3H]poly(A) and long-term-labeled [14CIDNA. Nucleic acids were extracted from the complex and bound to Millipore filters as described above. During this procedure, poly(A) will bind to the filters whereas other 0 kinds of RNA will not (23). After replicate filX 4 ters were assayed for acid-insoluble radioactivz ity, some were treated to elute the bound nucleic acids, and the eluates were centrifuged to equilibrium in a Cs2SO4 density gradient after heat denaturation. The results (Table 3) show first that both acid-insoluble 14C and 3H counts were bound to filters, suggesting that poly(A) 4 8 12 16 20 24 2B 32 36 40 was indeed present in the samples. Second, FRACTION NO. -Top eluted DNA and radioactivity derived from poly(A) banded at the same density in the FIG. 4. Cosedimentation ofpoly(A) with material Cs2SO4 density gradient, suggesting that the eluted from nitrocellulose filters in CsSO4 density eluted material contained sequences of adenylic gradient. Symbols: 0, radioactivity from poly(A); 0, acid (Fig. 4). However, not all of the 3H label DNA. banded with the DNA; some was found free at the bottom ofthe gradient, suggesting that part ity since the results of Fig. 2A showed that no of the complex between the putative poly(A) "free" poly(A) was found at the bottom of a and DNA was dissociated. Such dissociation Cs2S04 gradient after centrifugation of a nucould be due to treatment of the filters with cleic acid extract similar to that used in the SDS prior to centrifugation to elute bound mac- present experiment. The only difference beromolecules (see Materials and Methods) or to tween the two experiments was the filtering centrifugation in Cs2SO4, which could remove and elution of the nucleic acid from the filters poly(A) that was either weakly bound or adven- by SDS. This relatively harsh elution procedure titiously entrapped. We favor the first possibil- may also be responsible for the degradation of some of the DNA to fragments of different denTABLE 3. Binding of[3H]poly(A) to Millipore filtersa sity, producing the relatively broad main DNA peak and smaller peaks seen in Fig. 4. Input/filter Bound/filter Hybridization of poly(A) in the DNA-mem(counts/min) (counts/min) Conditions fraction with poly(A). In another apbrane 3H 3H 14C 14C proach, the presence of adenylic acid sequences 845 Nucleic acids ex- 5,800 1,010 2,700 in the DNA-membrane fraction was detertracted from DNAmined by hybridization experiments with ramembrane fraction dioactive poly(U). Replicate cell suspensions cells derived from were incubated for 15 min in the presence or with incubated and [3H]poly(A) absence of unlabeled poly(A) and in the pres['4C]deoxycytidine ence of [3H]deoxycytidine. After the DNAto label DNA membrane fraction from each cell suspension 940 was extracted, nucleic acids were extracted [3H]poly(A) control 1,100 from each of the complexes, mixed with [14C]DNA control 5,000 1,870 ['4C]poly(U), and centrifuged through a linear sucrose gradient. The results (Fig. 5) showed 720 7,400 (14C]RNA control that part of the ['4C]poly(U) added to the nua After the nucleic acids were extracted from DNA memcleic acid extract derived from cells incubated brane fraction and filtered slowly through Millipore filters, they were dried and assayed for acid-insoluble radioactiv- with poly(A) sedimented more rapidly than ity. For the [PHlpoly(A) control, the same amount of free ['4CIpoly(U) added to nucleic acids derived from [3H]poly(A) radioactivity as that detected in the DNA-mem- cells lacking poly(A). The difference in molecubrane fraction after extraction from the cells was added to lar size is presumably due to the presence of the filter. The [T4C]DNA control was obtained by extracting DNA from cell suspensions incubated with (14C]deoxy- adenylic acid sequences in the DNA-membrane cytidine only. The PT4CJRNA control was obtained by ex- fraction of "poly(A)" cells that complexed with tracting RNA from cell suspensions incubated with ["4C]- the poly(U). Moreover, the distribution of some uridine only. We consistently obtaned binding of DNA to Millipore filters regardles of what concentration KCI of the DNA in the former extract is also shifted was used. The data represent the average of four slightly "downward" in the gradient, implyexperiments. ing that DNA was part of the complex between
POLY(A) IN A PNEUMOCOCCAL DNA-MEMBRANE FRACTION
VOL. 127, 1976 RPly A Cells N
TABLE 4. Distribution of radioactivity in 2',3'ribonucleotides after alkaline hydrolysis ofproducts formed in a 5-min reaction with [a-32P]dCTP and
24- Control
1E
wfl\%
18
x 12 12
12
z
'I
, 6
6 .
.
a
a
a
a
A
a
4 8 e E 20 24 28 32 4 8 12 FRACTION NQ
21
6
~~~t I 2D 24 28
32
*Top
FIG. 5. Detection of poly(A) sequences in vivo by hybridization with poly(U). Poly(A) cells: 0, ['4C]poly(U); 0, DNA. Control: 0, ['4C]poly(U); *, DNA; -, ['4C]poly(U) marker. -
poly(A) and poly(U). Whether the DNA is bound specifically to both polynucleotides in the complex is unknown, but it is more likely that poly(A) sequences are involved since, in controls, no downward shift of DNA is observed despite the rather broad peak of poly(U) present in heavier regions of the sucrose gradient. A question also may be raised as to how the putative poly(A) is bound to DNA since the latter was not subjected to denaturation prior to centrifugation and, even if it were, a doublestranded complex between poly(A) and singlestranded DNA does not readily form in vitro (if at all). One possibility is that poly(A) is already bound to DNA prior to extraction from the cells via a phosphodiester bond (see below). Finally, it should be noted that the distribution of labeled poly(U) in the gradient after mixing with nucleic acids from cells lacking poly(A) is much broader than that of labeled poly(U) alone, suggesting the possible existence of adenylic acid sequences under4hatural conditions. Detection of c4valent links between DNA and RNA by 32P transfer. If poly(A) either acts as a primer for DNA synthesis or is part of the primer, it might be possible to detect the binding of poly(A) to DNA via a phosphodiester bond. The detection of the bond would help rule out the possibility of artifactual aggregation of poly(A) (or RNA) to nascent DNA in the Cs0SO4 gradient as seen in Fig. 2 and 3. Experimentally, such a transfer can be elucidated by the transfer of 32P as outlined in Materials and Methods. Therefore, we investigated the possible role of poly(A) and [a-32P]dCTP in the phosphodiester linkage. This deoxyribonucleoside triphosphate was chosen as the labeled substrate in the in vitro assay for DNA synthesis because a number of investigators (15, 18) found that it is frequently incorporated into the RNA-DNA junctions. The results (Table 4) showed that, in the absence of poly(A), 32p
poly(A) Poly(A)
(Qglml)
in the in vitro
system"
Counts/min of 32P transferred to: AMP
CMP
GMP
UMP
34 38 0 160 336 304 7.0 864 2,032 4,672 After the DNA-membrane fraction was incubated with [32PJdCTP, poly(A), and the rest of the assay solution, the reaction was stopped with trichloroacetic acid and ethanol. The precipitate was hydrolyzed in 0.3 M KOH for 16 h at 37 C and neutralized with HCIO4, and the hydrolysate (after removal of the DNA precipitate) was chromatographed and analyzed for radioactivity. Results are from duplicate reactions.
transfer occurred primarily with adenylic acid and cytidylic acid. However, when poly(A) was added, two effects were noted: an enhancement of 32P transfer to adenylic acid but a much greater enhancement of transfer to uridylic and guanylic acids. Thus, it appeared that poly(A) caused a generalized stimulation of the formation of an RNA-DNA hybrid macromolecule. In addition, although the DNA base in the junction could be cytosine, (although other "deoxytriphosphates" should be tested as well), it appeared as if the RNA base could be any one of the four constituents of RNA, but primarily uracil. This latter result is in agreement with that of Hirose et al. (18) and Flugel et al. (15), who found that predominant structure of the RNA-DNA junction in vitro to be ... p(rU)p(dC).... DISCUSSION In previous experiments (Firshein, submitted for publication), the presence of an RNADNA hybrid macromolecule during early stages of DNA synthesis in vitro was detected in a subfraction of the DNA-membrane complex, using both techniques described in this report (Fig. 2 and Table 4). The present experiments not only repeat the results with the whole DNA-membrane fraction, but show that poly(A) can enhance the formation of these RNA-DNA hybrid molecules. Thus, if such a hybrid is an important intermediate in the synthesis of DNA, then the enhancement of DNA synthesis by poly(A) both in vivo and in vitro could be due to its ability to increase the number of initiation sites for DNA polymerization. However, it is important to note that the enhancement of DNA synthesis in pneumococci by poly(A) in vitro is variable depending upon
22
FIRSHEIN ET AL.
concentration and probable length or conformation of the molecule in solution. Thus, there is a narrow range of enhancement between 2 and 8 ,ug of poly(A) per ml. Concentrations higher than 16 ,Lg/ml actually inhibit DNA synthesis from control values unless poly(A) is subjected to boiling for 5 min, in which case the range of stimulation can be extended to 16 ,ug/ml (but not higher). These differential effects may be related to the relative concentrations of poly(A) and oligoadenylates produced in the DNA membrane fraction as a result of nuclease action during the synthesis of the postulated RNA primer. Although we have only indirect evidence for this possibility, oligoadenylates could enhance the synthesis of the RNA primer, whereas undegraded poly(A) could inhibit its synthesis. At concentrations greater than 8 j,g/ml, a sufficient amount of undegraded poly(A) remains after incubation of the DNA-membrane fraction to inhibit the synthesis of the primer, despite the possible presence of a sufficient concentration of stimulatory oligoadenylates. However, at lower concentrations of poly(A), enough is degraded to permit the oligoadenylates to exhibit their stimulatory effect. When poly(A) is boiled for a short time, a greater percentage of the native molecule is degraded to smaller-molecular-weight fragments so that an apparent greater concentration of poly(A) can be added to enhance DNA synthesis (Fig. 1). In support of this hypothesis, it has been found that, whereas poly(A) inhibits RNA polymerase activity in Escherichia coli (1, 7), oligoadenylates stimulate a poly(A) polymerase activity closely associated with RNA polymerase in E. coli (27). Although in the latter case, it might appear as if an enhancement of poly(A) polymerase activity would be detrimental for the synthesis of RNA, it is entirely possible that such activity is essential to supply oligoadenylates of the proper size for synthesis of the RNA primer (see latter part of the Discussion). Regardless of whether the above explanation is valid, it seems that the oligoadenylates are probably part of the priming RNA molecule but not directly connected to nascent DNA by covalent linkage. However, other 3P-labeled deoxyribonucleoside triphosphates should be tested in a transfer experiment to ascertain whether poly(A) cannot be bound to these DNA bases as well. The priming RNA molecule probably does not have any specific base sequence since any one of the four bases (but mostly uracil) of RNA can be part of the RNA-DNA junction. Similar results were found by Reichard et al. (27) in studies of discontinuous polyoma DNA synthe-
J. BACTERIOL.
sis. They found that the size, rather than the sequence, of the priming RNA molecule was important. The length of the oligoadenylate segment in the priming RNA molecule is not yet certain. The molecular weight of the added poly(A) was 100,000 or approximately 335 nucleotides in length. Since poly(A) is degraded to small- and large-molecular-weight oligomers in vivo, it is likely that the length of the active oligomer (or oligomers) is less than that. The priming RNA segment ofE. coli has been estimated by Hirose et al. (18) to be approximately 50 to 100 nucleotides, although studies by Reichard et al. (27) have found a priming segment for polyoma DNA synthesis to be only 10 nucleotides in length. One important question is whether such a model involving pool(A) or its oligomers could still be valid without the addition of external poly(A), which, except for an unusual permeability by the pneumococcal bacterium, is not readily incorporated by many other bacterial species. At least one observation with pneumococci suggests the presence of poly(A) in vivo under natural conditions. Poly(U) sedimented in a broad peak in the sucrose gradient after mixing with nucleic acids extracted from a "control" DNA membrane fraction [i.e., derived from cells incubated without poly(A)], whereas poly(U) alone sedimented in a rather sharp peak nearer the top of the gradient. This could suggest.that poly(U) complexed with adenylic acid sequences present in the nucleic acid of the control extract. As stated earlier, a variety of investigations with other organisms has demonstrated the presence of either poly(A) in vivo or a poly(A) polymerase. For example, primer-dependent and primer-independent poly(A) polymerases have been detected in E. coli (1, 7, 26) and Shigella dysenteriae (6). One interesting observation is that the primer-independent enzyme is associated with DNA-dependent RNA polymerase and that oligoadenylates enhance the polymerase reaction by 50 to 70% (26). If a similar poly(A) polymerase exists in pneumococci, then the enzyme could provide oligoadenylates, which, according to our model, would be incorporated into the priming RNA segment. It is interesting that poly(A) segments have been detected in pulse-labeled RNA of E. coli (28). Moreover, exogenous poly(A) could stimulate the activity of the poly(A) polymerase (after its degradation in vivo), thus increasing the pool of oligoadenylates and, ultimately, the fornation of the priming RNA segment. We are now attempting to determine whether a poly(A) polymerase exists in pneumococci and particu-
POLY(A) IN A PNEUMOCOCCAL DNA-MEMBRANE FRACTION
VOL. 127, 1976
23
bonucleotides of known composition on deoxycytidyllarly in the DNA membrane fraction. This latate and deoxyguanylate kinase activity in pneumoter point is extremely important because the cocci. J. Biol. Chem. 243:3301-3311. DNA-membrane fraction can not only synthe- 13. Firshein, W., and R. C. Gilmore. 1970. DNA-membrane complex in pneumococci. Macromolecular consize DNA in vitro without the addition of tent and stimulation of enzyme activity by polyadepoly(A), but, as stated previously, an RNA-
DNA hybrid
can
be detected during early
stages of DNA synthesis in the absence of the polynucleotide (Firshein, submitted for publication). It would be expected that if poly(A) were involved in initiation of DNA replication
by the model proposed above, adenylic acid sequences should be detected in the RNA-DNA hybrid. ACKNOWLEDGMENT This research was supported by National Science Foundation grant GB 34155. LITERATURE CITED 1. August, J. T., R. J. Ortiz, and J. Hurwitz. 1962. Ribonucleic acid-dependent ribonucleotide incorporation. J.
Biol. Chem. 237:3786-3793. 2. Braun, W. 1974. Regulatory factors in the immune response: analysis and perspective, p. 4-23. In W. Braun, L. M. Lichtenstein, and C. W. Parker (ed.), Cyclic AMP, cell growth and the immune response.
Springer-Verlag, New York. 3. Brutlag, D., R. Schekman, and A. Kornberg. 1971. A possible role for RNA polymerase in the initiation of M13 DNA synthesis. Proc. Natl. Acad. Sci. U.S.A. 68:2826-2829. 4. Burkard, B., and E. B. Keller. 1974. Poly(A) polymerase and poly(G) polymerase in wheat chloroplasts. Proc. Natl. Acad. Sci. U.S.A. 71:389-393. 5. Burton, K. 1968. Determination of DNA concentration with diphenylamine, p. 163-166. In L. Grossman and K. Moldave (ed.), Methods in enzymology, vol. 12, part B. Academic Press Inc., New York. 6. Colvill, A. J. E., and M. Terzi. 1968. Polyriboadenylate polymerase and its inhibition in T4 infected Escherichia coli and Shigella dysenteriae. Biochim. Biophys. Acta 155:394400. 7. DeRobertis, M., P. M. Ezcurra, N. D. Judewicz, P. R. Pucci, and H. N. Torres. 1972. Inhibition of E. coli RNA polymerase by polyadenylic acid. FEBS Lett. 25:175-178. 8. Edmonds, M., M. H. Vaughan, Jr., and H. Nakazato. 1971. Polyadenylic acid sequences in the heterogeneous nuclear RNA and rapidly-labeled polyribosomal RNA of HeLa cells: possible evidence for a precursor relationship. Proc. Natl. Acad. Sci. U.S.A. 68:1336-1340. 9. Firshein, W. 1965. Influence of deoxyribonucleic acid degradation products and orthophosphate on deoxyribonucleotide kinase activity and deoxyribonucleic acid synthesis in pneumococcus type III. J. Bacteriol. 90:327-336. 10. Firshein, W. 1972. The DNA membrane fraction of Pneumococcus contains a DNA replication complex. J. Mol. Biol. 70:383-397. 11. Firshein, W. 1974. In situ activity of enzymes on polyacrylamide gels of a deoxyribonucleic acid membrane fraction extracted from pneumococci. J. Bacteriol. 118:1101-1110. 12. Firshein, W., and R. C. Benson. 1968. Effects of polyri-
nylic acid. Science 169:1106-1113. 14. Firshein, W., and C. W. Schwenzfeier. 1969. Characterization of excess deoxyribonucleic acid synthesized by pneumococci in the presence of polyadenylic acid and deoxyribonucleic acid precursors. J. Bacteriol. 97:1106-1113.
15. Flugel, R. M., J. E. Larson, P. F. Schendel, R. W. Sweet, T. R. Tamblyn, and R. D. Wells. 1973. RNADNA bonds formed by DNA polymerases from bacteria and RNA tumor viruses, p. 309-320. In R. D. Wells and R. B. Inman (ed.), DNA synthesis in vitro. University Park Press, Baltimore. 16. Fraser, R. S. S., and U. E. Loenig. 1972. Binding of radioactive polynucleotides to RNAn A simple method for the detection, during gel electrophoresis, or sedimentation of messenger RNAs which contain poly(A) sequences. Eur. J. Biochem. 34:153-158. 17. Haff, L. A., and E. B. Keller. 1973. Two distinct poly(A) polymerases in yeast nuclei. Biochem. Biophys. Res. Commun. 51:704-710. 18. Hirose, S., R. Okazaki, and T. Tamanoi. 1973. Mechanism of DNA chain growth. XI. Structure of RNAlinked DNA fragments of Escherichia coli. J. Mol. Biol. 77:501-217. 19. Jacob, S. T., and D. G. Schindler. 1972. Polyriboadenylate polymerase solubilized from rat liver mitochondria. Biochem. Biophys. Res. Commun. 48:126-134. 20. Karkas, J. D., J. G. Stavrianopoulos, and E. Chargaff. 1972. Action of DNA polymerase I of Escherichia coli with DNA-RNA hybrids as templates. Proc. Natl. Acad. Sci. U.S.A. 69:398-402. 21. Kwan, S. W., and G. Brawerman. 1972. A particle associated with the polyadenylate segment in mammalian messenger-RNA. Proc. Natl. Acad. Sci. U.S.A. 69:3247-3250. 22. Lai, M. M. C., and P. H. Duesberg. 1972. Adenylic acid rich sequences in RNA's of Rous sarcoma virus and Rauscher mouse leukaemic virus. Nature (London) 235:383-386. 23. Lee, S. Y., J. Mendecki, and G. Brawerman. 1971. A polynucleotide segment rich in adenylic acid in the rapidly-labeled polyribosomal RNA component of mouse sarcoma 180 ascites cells. Proc. Natl. Acad. Sci. U.S.A. 68:1331-1335. 24. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall. 1951. Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193:265-275. 25. Milcarek, C., and S. Penman. 1974. Membrane bound polyribosomes in HeLa cells. Association of polyadenylic acid with membranes. J. Mol. Biol. 89:327-338. 26. Modak, M. J., and P. R. Srinivasan. 1973. Purification and properties of a ribonucleic acid primer independent polyriboadenylate polymerase. J. Biol. Chem. 248:6904-6910. 27. Reichard, P., R. Eliasson, and G. Soderman. 1974. Initiator RNA in discontinuous polyoma DNA synthesis. Proc. Natl. Acad. Sci. U.S.A. 71:49014905. 28. Srinivasan, P. R., M. Ramanarayanan, and E. Rabbani. 1975. Presence of polyriboadenylate sequences in pulse labeled RNA of Escherichia coli. Proc. Natl. Acad. Sci. U.S.A. 72:2910-2914. 29. Sugino, A., and R. Okazaki. 1973. RNA-linked DNA fragments in vitro. Proc. Natl. Acad. Sci. U.S.A. 70:88-92.