Vol. 31, No. 2
JOURNAL OF VIROLOGY, Aug. 1979, p. 494-505 0022-538X/79/08-0494/12$02.00/0
Further Characterization of the Replicative Complex of Vesicular Stomatitis Virus CHRISTIAN C. SIMONSEN,* VIRGINIA M. HILL, AND DONALD F. SUMMERS Department of Cellular, Viral and Molecular Biology, University of Utah Medical Center, Salt Lake City, Utah 84132 Received for publication 23 March 1979
Replicating vesicular stomatitis virus ribonucleoprotein (RNP) complexes were isolated in nonequilibrium Renografin density gradients. These nascent RNPs had the same buoyant density as virion nucleocapsids in both isopycnic Renqgrafm and CsCl gradients. Both transcribing and replicating RNP complexes were shown to be stable in sucrose gradients, whereas only replicating RNP complexes were stable in Renografin gradients. Size analysis of the 5-min-pulse-labeled RNA species from the replicating RNPs using methylmercury gels revealed that the nascent strands were primarily less than full-length molecules. Longer times of radiolabeling demonstrated that the nascent RNA accumulated as 42S RNA, which was primarily of the same sense as the virion strand when it was radiolabeled at 5 h postinfection. The percentage of this radiolabeled RNA which was plus stranded was higher at 2.5 h postinfection, reflective of the shift in plus- to minus-strand full-length 42S RNA synthesis which occurs in the cell. Addition of cycloheximide to the infected cells before the addition of the radiolabel prevented the formation of these RNP complexes. Both the change in the percentage of minus strands found in the RNP complexes at the different times postinfection and the sensitivity to cycloheximide indicate that the RNP complex which was isolated was indeed the replicative complex. During the infection of HeLa cells by vesicular stomatitis virus (VSV), both virion minus-strand and complementary plus-strand full-length 42S RNA can be isolated from intracellular ribonucleoprotein (RNP) complexes (34, 36). These RNPs are composed of 42S RNA and three VSV-specific proteins: the N, or nucleocapsid, protein; the NS protein, a phosphorylated polypeptide; and the L protein (17, 18, 36). The virion nucleocapsid is able to direct the in vitro synthesis of the five VSV-specific mRNA's which, like the in vivo VSV mRNA's, are capped at the 5' end, methylated, and polyadenylated (2, 7, 20, 25-27, 32). Reconstitution experiments have shown that both the L and NS proteins are required to form the virion-associated RNA-dependent RNA polymerase (17, 18, 28). The template for transcription has been demonstrated to be composed of minus-strand 42S RNA and N protein (11, 17). The requirements for replication, however, are not well known. The temperature-sensitive mutant tsG114, which possesses a thermolabile L protein (21), has been used to provide indirect evidence that the L protein is involved in replication (30). Several other mutants suggest a role in replication for the N and NS proteins as well (15, 16). For these reasons it has been assumed
that replication occurs on an RNP template similar to the transcriptive RNP. Fractionation techniques previously used to isolate VSV RNPs do not separate replicating RNPs from transcribing RNPs. Sedimentation velocity centrifugation analyses have indicated that intracellular VSV RNPs sediment as 140S molecules, whether or not nascent mRNA is associated with the RNP complexes (31, 35). The lack of resolution is further complicated by the fact that replication accounts for only 10% of the total VSV-specific RNA synthesized in infected cells (38); thus, sucrose gradients have not been able to separate replicating RNPs from transcribing RNPs. CsCl gradients have been used to isolate glutaraldehyde-fixed messenger RNPs (19) and intracellular nucleocapsids (34); however, the messenger RNPs are not stable in CsCl without prior fixation with glutaraldehyde. RNPs isolated with CsCl gradients do not retain the L and NS proteins, which are known to comprise part of the functional RNPs in the cell (17, 18). In addition, all of the intracellular RNPs band at the same density in CsCl. For these reasons, CsCl gradients have not been used to separate the RNP complexes. lodinated density gradient media such as metrizamide and Renografm have been employed
494
VSV REPLICATIVE COMPLEX
VOL. 31, 1979
to obtained resolution of proteins, nucleic acids, and RNPs in the same gradients (6, 22). These media are hypertonic and minimize nonspecific RNA-protein interactions, yet do not dissociate most RNPs. Several investigators have used metrizamnide and Renografin to examine chromatin (12), influenza RNPs (10), and proteins (8). We have recently described the isolation in Renografin gradients of RNP complexes from VSVinfected HeLa cells (V. M. Hill, C. C. Simonsen, and D. F. Summers, submitted for publication). Several peaks of radioactivity were observed when an infected cell extract radiolabeled from 3 to 5 h postinfection (p.i.) was centrifuged in a Renografin gradient. One of those peaks contained predominantly minus-strand 42S RNA and was termed peak I. A similar pattern of radioactivity was seen after an infected cell extract pulse-labeled for 5 min at 5 h p.i. was analyzed, except that peak I exhibited a pronounced trailing edge. Virion nucleocapsids formed only a homogeneous peak I which lacked a trailing edge. Genome length and smaller RNAs were found in the peak fractions of the pulse-labeled peak I, whereas only the smaller RNAs were found in the trailing fractions. This RNA, pulse-labeled for 5 min at 5 h p.i., was 80% minus stranded, the same percentage observed with pulse-labeled 42S RNA obtained from intracellular nucleocapsids at 5 h p.i., when replication is maximal (34). The nascent minus-strand RNA was shown to be identical to the 5' end of the VSV genome RNA. We also observed that the nascent RNA was RNase resistant, was single stranded, and was isolated at a density of 1.240 g/cm3 in Renografin. All of this suggested that peak I contained nascent replicating VSV RNA associated with protein as a replicating RNP complex. Herein we extend these observations and further characterize the complexes to show by several different criteria that these RNPs are indeed VSV replicative complexes. MATERIALS AND MErHODS virus infection, and radioactive labeling. Cells, Suspension cultures of HeLa S3 cells were grown in Joklik modified minimal essential medium (Flow Laboratories) supplemented with 2 mM glutamine plus 5% fetal calf serum (Flow Laboratories) at a concentration of 2 x 105 to 8 x 10' cells per ml. Stock preparations of VSV (Indiana serotype) were grown in HeLa cells, purified, and assayed as described previously (3, 4). Cells were collected by centrifugation, resuspended in growth medium minus serum at a concentration of 6 x 106 cells per ml, and then infected with 10 PFU of VSV per cell. At 1 h p.i. serum was added to 5% and actinomycin D (a gift from Merck, Sharp & Dohme) was added to 1 yg/ml. At times ranging from 2.5 h to 5 h p.i. ['H]uridine or ['H]-
495
adenosine (25 to 30 Ci/mmol; New England Nuclear, Corp., Boston, Mass.) or both were added to a concentration of 200 ,tCi/ml for the indicated labeling periods, after which crushed, frozen medium was added and the cells were pelleted by centrifugation for 3 min at 900 x g. The cell pellet was lysed by the addition of 1 ml of 1% Nonidet P-40 in NT buffer (0.1 M NaCl, 0.01 M Tris, pH 7.5) and kept on ice for 10 min. The nuclei were removed by centrifugation as described above and were then resuspended in 1 ml of 0.2% deoxycholate-0.3% Nonidet P-40 in NT buffer and recentrifuged. The supernatants from the 1% Nonidet P-40 treatment and the deoxycholate-Nonidet P-40 wash were then combined and layered onto preformed CsCl or Renografin gradients. Isolation of RNP complexes. RNPs were isolated by using either Renografin, CsCl, or sucrose gradients. For isolation of the replicative RNPs, the cytoplasmic extracts from above were layered onto preformed 36ml, 15 to 50% or 20 to 60% Renografin (E. R. Squibb & Sons) gradients in NT buffer and centrifuged at 23,000 rpm and 4°C in an SW27 rotor for 16 h (for nonequilibrium gradients) or for 80 h (for equilibrium gradients). The gradients were fractionated by pumping from the bottom, and radioactivity was determined by directly assaying 50-1l portions in 7 ml of scintillation cocktail in a Beckman liquid scintillation spectrometer.
Preformed 20 to 40% (wt/wt) CsCl (Varlacoid Chemical Co.) gradients in TNE buffer (0.025 M Trishydrochloride, pH 7.5, 0.05 M NaCl, 0.002 M EDTA) were overlayed with fractions from the Renografin gradient which had been diluted 1:3 in NT buffer. These gradients were centrifuged in an SW41 rotor at 33,000 rpm for 16 h at 4°C. After centrifugation, the gradients were fractionated, and acid-precipitable radioactivity was determined as described previously (3, 4). The densities of the fractions were determined by measuring the refractive indexes with a Bausch and Lomb refractometer. Cytoplasmic extracts were also layered onto 15 to 30% sucrose gradients in NT buffer, which were centrifuged in an SW27 rotor for 16 h at 16,000 rpm and 4°C and fractionated as described above. Phenol extraction of RNA. An equal volume of phenol saturated with NETS buffer (0.1 M NaCl, 0.001 M EDTA, 0.01 M Tris, pH 7.5, 0.2% sodium dodecyl sulfate) was added to the samples in 3 ml of NETS buffer and mixed, and the two layers were separated by centrifugation. The aqueous layer was collected, and the phenol layer was extracted once more with 3 ml of NETS buffer. The aqueous fractions were combined, sodium acetate was added to a final concentration of 0.2 M, 2 volumes of ethanol were added, and the RNA was precipitated at -70°C overnight. RNA was pelleted by centrifugation at 16,000 x g for 30 min at40C. Purification of VSV-specific RNA. Minus-strand 42S RNA was prepared from purified virions by disrupting the virus in NETS buffer and centrifuging the samples in 15 to 30% sucrose-sodium dodecyl sulfate gradients in an SW41 rotor. The 42S RNA peak was ethanol precipitated and dissolved in 2xA buffer (0.3 M NaCl, 0.02 M Tris, pH 7.4, 0.002 M EDTA). VSV mRNA was prepared from infected cells at 4.5 h p.i. A
496
SIMONSEN, HILL, AND SUMMERS
cytoplasmic extract was prepared as described above and was phenol extracted, ethanol precipitated, and chromatographed on an oligodeoxythymidylic acidcellulose column. Oligo deoxythymidylic acid-cellulose chromatography of RNA. Oligodeoxythymidylic acid-cellulose (Collaborative Research, Inc., Waltham, Mass.) columns were prepared and eluted by the method of Banerjee and Rhodes (7). The high-salt elution buffer contained 0.01 M Tris-hydroxychloride, pH 7.4, and 0.5 M NaCl, and the low-salt buffer contained 0.01 M Tris-hydrochloride, pH 7.4. RNA was loaded onto the column in high-salt buffer. Fractions containing bound material were pooled and ethanol precipitated as described above. Hybridization of VSV RNA. Hybridizations were performed in sealed capillary pipettes in 0.02-mIl volumes by using the conditions of Kolakofsky (24). To each sample of [3H]RNA isolated from nucleocapsids were added increasing amounts of unlabeled VSV virion 42S RNA or mRNA. The samples were then heat denatured by heating to 115°C for 3 min and incubated for 3 h at 730C. Each sample was treated with a 30-,ug/ml solution of RNase A (Worthington Biochemicals Corp.) in 2xA buffer for 30 min at 250C and was then assayed for acid-precipitable radioactivity as described above. The self-annealing reactions were identical, except that no unlabeled RNA was added to the samples. Methylmercury-agarose gel electrophoresis. Methylmercury-agarose gel electrophoresis was performed by a modification of the method of Bailey and Davidson (5). Seakem agarose powder was obtained from Marine- Colloids, Inc., and methylmercury (II) hydroxide was obtained as a 1 M solution from Alfa Products, Danvers, Mass. Because of the toxic and volatile nature of the mercury compound, operations were performed under a hood. Horizontal slab gels (19 by 13.3 by 0.4 cm; 100-ml bed volume) were 1% agarose in borate buffer (0.05 M boric acid, 0.005 M sodium borate, 0.01 M sodium sulfate, 0.001 M trisodium EDTA, pH 8.0) and contained 0.005 M methylmercury hydroxide. Samples were applied in 40 ,ul of 1:1 sample-sample buffer (borate buffer containing 50% glycerol and 0.1% bromophenol blue), with methylmercury hydroxide added to a concentration of 0.005 M just before loading onto the gel. The gel reservoirs contained borate buffer, and this buffer was recirculated at approximately 80 ml/h throughout the electrophoresis period. Gels were run at 75 to 100 V (50 mA) for 6 h (dye front migrated 14 cm). Fluorography of methylmercury gels. Gels contaiing [3H]RNA were fixed for 10 min in 10% acetic acid containing 0.01 M cysteine. They were dehydrated in 100% methanol for two successive 1-h periods. After they were dried to paper thinness under a vacuum, the gels were soaked in a 10% (wt/vol) solution of 2,5-diphenyloxazole (New England Nuclear Corp.). in methanol for 3 h. Gels were soaked in water for 10 min to precipitate the 2,5-diphenyloxazole, blotted dry, and mounted on a, piece of 3 MM paper (Whatman, Ltd., England). After being covered with Saran Wrap, gels were exposed to Kodak SB-5 X-ray film at -70°C and developed after appropriate periods of time.
J. VIROL.
RESULTS Kinetics of labeling of peak I. We wished to determine whether 42S RNA would accumulate in the Renografin-isolated peak I if the length of the pulse-labeling period was increased. We also wished to learn whether the material in the trailing fractions of peak I would also accumulate during a longer pulse-labeling period. VSV-infected HeLa cells were radiolabeled for 5, 10, 15, and 30 min at 4.5 h p.i. Cytoplasmic extracts were prepared from each sample and centrifuged in identical 15 to 50% Renografin gradients. The results (Fig. 1) demonstrate that the trailing fractions of the 5-min sample (fractions 26 to 31) were not as apparent in the l0-min sample and were even less noticeable in the 15- and 30-min samples. All samples were continuously labeled; therefore, the material in the trailing fractions was still present in the longer-labeled fractions, but comprised a minor proportion of the total radiolabeled material due to the increase in the amount of radiolabel in the peak fractions (fractions 26 to 31, all samples). We next pooled the fractions of the replicative complex from Fig. 1. The trailing fractions of the 5-min-pulse-labeled sample were included because the radiolabeled RNA contained in these fractions represented an appreciable amount of the total radiolabeled RNA. Only the peak fractions from the remaining samples were pooled because the material in the trailing fractions did not represent a significant proportion of the total radiolabeled RNA. The RNA was extracted from these pools and analyzed in methylmercury-agarose gels. The densitometer tracings of the autoradiographs from these gels are shown in Fig. 2. The RNA from the 5-min-pulse-labeled samples was heterogeneous in size, ranging from less than 18S to 42S, although the average size of this sample was approximately 26S. The size distribution of the 10-min sample was also heterogeneous; however, the average size of the labeled RNA increased to a value of approximately 35S, and there was an obvious 42S band. Both the 15- and 30-min samples contained primarily 42S RNA, although there were significant amounts of heterogeneous material less than 42S RNA in size. This suggests that the radiolabel accumulated as genome length RNA as the pulse time was increased. Figure 2 also implies that it took longer than 5 min for 42S RNA to begin to accumulate. It is not clear why this was the case, since one would expect that a 5-min pulse-label would uniformly label all nascent RNA. Perhaps the synthesis of VSV 42S RNA does not occur at a uniform rate throughout the template; as we have noted previously, there is an apparent
VOL. 31, 1979
VSV REPLICATIVE COMPLEX
497
T 0
FRACTION NUMBER
FIG. 1. Renografin gradient centrifugation of cytoplasmic extracts from pulse-labeled VSV-infected cells. Infected cells at 4.5 h p.i. were labeled with 100 ,uCi each of [3H]adenosine and [3H]uridine per ml. Samples were removed at 5, 10, 15, and 30 min after the addition of the labels. Cytoplasmic extracts were prepared, layered onto 15 to 50% Renografin gradients, and centrifuged for 17 h at 23,000 rpm and 4°C in an SW27 rotor. Fractions (0.5 ml) were collected by pumping out from the bottom, and the radioactivity in each fraction was determined by counting 50-,iu portions in a toluene-based scintillation fluid.
pause site near the juncture of the L- and Gprotein cistrons (Hill et al., submitted for publication). Hybridization of pulse-labeled RNA from Renografin RNPs. We have recently reported that the ratio of plus- to minus-strand 42S RNA synthesis in infected cells is greater early in the infection (2.5 h p.i.) than when replication is maxinal at 5 h p.i., at which time 80% of the newly synthesized 42S molecules are minus stranded (34). We have also shown that pulselabeled nascent RNA species which are isolated from putative replication complexes at 5 h p.i. are approximately 80% minus stranded (Hill et al., submitted for publication).
To determine whether pulse-labeled nascent RNA extracted from the replication complexes would reflect the in vivo findings, infected HeLa cells were labeled for 5 min with [3H]uridine at 2.5 or 5 h p.i. Cytoplasmic extracts were prepared from these cells and were centrifuged in 15 to .50% Renografin gradients. The nascent RNA contained in the RNP complexes (at a peak density of 1.240 g/cm3) was hybridized with unlabeled VSV 42S RNA and VSV mRNA. The RNPs labeled at 2.5 h p.i. contained significantly more plus-strand RNA than the RNPs labeled at 5 h p.i. (Table 1). The change in the ratio of plus- to minus-strand RNA contained in these RNP complexes from approximately 35:65 at 2.5
498 origin
I
J. VIROL.
SIMONSEN, HILL, AND SUMMERS 42S
28S
I
I
185 5min
_
_
10min
E 0 to
cli
15min
.o cn
by the addition of the drug. This result also demonstrates that the cycloheximide-sensitive RNP complex is a replicative complex. Isopycnic gradient centrifugation of the replicative complexes. We next examined the RNP complexes in isopycnic Renografin gradients to study the basis for the isolation of the heterogeneous peak I. A 5-min-pulse-labeled cytoplasmic extract was prepared from cells at 5 h p.i. as described above and was centrifuged in a 15 to 50% Renografm gradient for 17 h at 23,000 rpm in an SW27 rotor. Fractions from three portions of the RNP peak were pooled as indicated (Fig. 4A) and combined, and samples were recentrifuged on identical 20 to 60% gradients for 17, 40, and 80 h (Fig. 4B, C, and D). The
replicative RNP peak, which was very heterogeneous after being recentrifuged for 17 h (Fig. 4B), formed a homogeneous peak at a density of 1.248 g/cm3 in the isopycnic gradients (Fig. 4C and D). Under the same conditions of centrifu30min
gation, VSV 42S RNA was found at 1.19 g/cm3 (Fig. 4E). Poliovirus top component (nucleocap-
sid which lacks RNA) and poliovirus infectious particles were found in the equilibrium gradients at 1.325 and 1.218 g/cm3, respectively (Fig. 4F). The result indicated that the ratio of protein to RNA in the nascent RNPs was the same 2
4 6 DISTANCE MIGRATED (cm)
throughout the peak
E
and
trailing
fractions of
peak I, since the smaller RNPs contained in the FIG. 2. Densitometer tracings of fluo,rographed trailing edge of peak I had the same isopycnic methylmercury-agarose gels. The RNA cointained in density in Renografin as the RNPs in the peak the pulse-labeled RNPs was extracted from m the frac- fractions. tions pooled as indicated in Fig. 1. Appiroximately 50,000 cpm of purified RNA from each stample was electrophoresed in a methylmercury-agaro ge fluorographed as described in the text. T'he autoradiograph was analyzed with an Ortec nMoael 4MlU densitometer. The scale is identical for a,11 samples.
getoand
h p.i. to 15:85 at 5 h p.i. is similar to tihe change in the ratio of plus- to minus-strand 42S RNA synthesis we have previously observe indicating that the plus-strand RNA slpecies isolated in peak I were products of replic ation. Effect of cycloheximide on the rebplicative RNA. A further means of differentiiating the replicative process from the transcripitive process is by using the protein synthesis inhibitor cycloheximide. The synthesis of the VSV mRNA's is not affected appreciably tby the addition of cycloheximide, whereas repllication of 42S RNA is rapidly inhibited (30, 38)I. Figure 3 shows that the synthesis of the peatk I RNA complex from Renografin was inhibit;ed by the addition of cycloheximide, whereas t:he formation of the slower-sedimenting peak II, which contains exclusively plus-strand seque-nces (Hill et al., submitted for publication), was anaffected d in vivo,
i
TABLE
1.
Hybridization analysis of the RNA
isolated from replicative RNPs from cytoplasmic extracts pulse-labeled at 2.5 and 5 hp.i.a Tixme p.i. (h)
Total
Self-annealed material
Unlabeled RNA added mRNA
cpm
%
42S RNA
cpm % cpm % 776 298 38 545 70 266 34 2.5 707 277 39 2.5 488 69 247 35 407 1,175 94 225 18 5 33 1,247 572 192 33 544 95 60 10 5 a Infected cells were pulse-labeled at either 2.5 or 5 h p.i. as described in the legend to Fig. 1. The replicative RNP complex was isolated from nonequilibrium Renografin gradients, and the RNA contained in the RNP was isolated by phenol extraction. The pulse-labeled RNA was mixed with increasing amounts of unlabeled VSV mRNA or 42S RNA in 20-Il reaction volumes by using the conditions of Kolakofaky (24). The samples were heated to 115°C for 1 min, annealed for 3 h at 730C, and then treated with 30 yg of RNase A per ml in 2xA buffer for 30 min at 220C. Plateau levels of RNase resistance are expressed after being corrected for the RNaseresistant material which was present when the sample was heat denatured but not annealed. This figure varied from 2 to 4% of the input radioactivity. Self-annealing of the material was determined as described above, except that no unlabeled RNA was added to the samples.
VSV REPLICATIVE COMPLEX
VOL. 31, 1979
499
Pe E
n Number
FIG. 3. Effect of cycloheximide on the production of replicative RNP complex. A culture of infected cells was split in half at 4.5 hp.i., and to one-half of the culture cycloheximide was added to a final concentration of 100 tLg/ml. After 15 min in the presence of the drug, [3H]uridine was added to each culture to a final concentration of 100 ,aCi/ml. After a 5-min incubation, the cells were removed by centrifugation and cytoplasmic extracts were prepared as described in the text. The extracts were layered onto separate identical 15 to 50% Renografin gradients and were centrifuged for 16 h at 23,00( rpm and 4°C in an SW27 rotor. The radioactivity from the fractionated gradients was analyzed as described previously. Symbols: -4, untreated control; 0-----S cycloheximide-treated sample.
The nascent RNPs were further analyzed by centrifuging fractions from the peak and trailing fractions of peak I in CsCl gradients (Fig. 5). The nascent RNPs had identical equilibrium densities in CsCl. This result also suggests that the nascent RNPs have a protein-RNA ratio sunilar to that of virion nucleocapsids, since all of the radioactivity was found at a density of 1.31 g/cm3 (19). The nascent RNA was then removed from the RNPs and analyzed in sucrose-sodium dodecyl sulfate gradients, which showed that the RNA contained in the CsClbanded RNPs was shorter than full length and not just 42S RNA (Fig. 6). The demonstration that nascent RNPs are stable in CsCl is further evidence suggesting that these nascent RNPs are derived from a replicative intermediate since transcriptive products do not band in CsCl (19). It thus appears likely that the heterogeneity of the 5-min-pulse-labeled RNP complex which has been isolated is not due to a difference in density between the RNPs in the peak and the trailing fractions. Examination of the trailing fractions of peak I. The results shown in Fig. 1 demonstrated that the pulse-labeled material in the trailing fractions did not accumulate during long pulse-labeling periods. Furthermore, it has been
shown that nascent RNA is found in both the peak and the trailing fractions of peak I, but detectable newly replicated 42S molecules are found in the peak fractions only (Hill et al., submitted for publication). These results perhaps suggest that the material in the trailing fractions is nascent RNA which has dissociated from the template RNA. We have observed that the nonequilibrium Renografin gradients resolve RNA molecules on the basis of size; thus, it is possible that the trailing fractions contain smaller RNPs separated on the basis of size. At the present time there are two possible ways that this problem can be studied directly. Electron microscopic examination of the material contained in the peak and trailing fractions was performed; however, due to the large number of nonradiolabeled structures present in the same area as peak I, we could not prove that the trailing edge represented nascent strands dissociated from their templates (C. Naeve, unpublished data). We did observe possible replication forms in the peak fractions (Naeve, manuscript in preparation), but due to the high background of nonreplicating RNPs, it was impossible to deal conclusively with the problem of nascent strand attachment to template RNP. Another way of directly showing that the peak fractions
500
J. VIROL.
SIMONSEN, HILL, AND SUMMERS
7d
E
-
5
3
20
5
25
30
35
5
1015
F
E
25
20
30
35
C
~~~~~~~~~~~~8-
4-
..
2-
-1.20
4-
-1.10 25
30
40
35
45
50
5 55 FRACTION NUMBER
'15
10
20
30
25
'35
FIG. 4. Equilibrium Renografin densitygradient centrifugation of the 5-min-labeled RNP complex. Infected cells at 5 h p.i. were labeled by the addition of 100 ,LCi each of [3H]uridine and [3H]adenosine per ml. Crushed, frozen medium was added to the sample after 5 min in the presence of the labels, and the cells were removed by centrifugation. A cytoplasmic extract was prepared and layered onto a 15 to 50% Renografin gradient which was centrifuged for 16 h at 23,000 rpm and 4°C in an SW27 rotor. The radioactivity was quantitated as described in the text. The fractions indicated fiom the peak and trailing fractions in (A) were combined and diluted. Samples were layered onto identical 20 to 60% Renografin gradients and were centrifuged at 23,000 rpm and 4°C in an SW27 rotor for 17 h (B) 40 h (C), or 80 h (D). 3H-labeled 42S RNA (E) and poliovirus nucleocapsid and top component (F) were centrifuged in identical Renografin gradients at 23,000 rpm and 4°C for 80 h.
8-A
Peak H2-B
PeokI
C
D
*140
6g E
.4
|~1 2" 23': ~ ,
,
,
.
,
2
.
5
1"
5
,
130 .120
2-
5 10 1520 25
30 35
5 10
15
FRACTION NUMBER
5
8
15
5
10
15
FIG. 5. CsCl gradient centrifugation of peak I from a Renografin gradient. A cytoplasmic extract from infected cells pulse-labeled for 5 min at 5 h p.i. was layered onto a 15 to 50% Renografin gradient and centrifuged as for Fig. 1. The RNP peak was split into three pools as indicated in (A). The pools were then layered onto 20 to 40% (wt/wt) CsCl gradients in TNE buffer and centrifuged for 16 h at 33,000 rpm and 4°C in an SW41 rotor. Fractions (0.5 ml) were collected from the bottom, trichloroacetic acid-precipitable radioactivity was determined for 250 itl of each fraction, and the density was recorded by measuring the refractive indexes of the fractions with a Bausch and Lomb refractometer. (B) CsCl gradient ofpool 1 from (A). (C) Pool 2. (D) Pool 3.
VSV REPLICATIVE COMPLEX
VOL. 31, 1979
501
FRATION NUMBER
FIG. 6. Size analysis of the pulse-labeled RNA contained in CsCl-banded replicative RNPs. The pulselabeled RNPs which were banded in CsCl in Fig. 5B, C, and D were phenol extracted, and the RNA contained in these complexes was centrifuged in 15 to 30% sucrose-sodium dodecyl sulfate gradients for 16 h at 20,000 rpm in an SW41 rotor. Trichloroacetic acid-precipitable radioactivity was determined for each of the fractions, which were collected by dripping from the bottom of the tube. Symbols: 0, ['4C]uridine-labeled 42S, 28S, 18S, and 4S markers; *, [3H]uridine-labeled RNA from Fig. 5B, C, and D (panels A, B, and C, respectively).
contain nascent strands attached to their template is to complete the nascent strands in vitro; however, at the present time this has not been possible. It was therefore necessary to study this problem indirectly. A size analysis of the RNPs contained in the peak and trailing fractions of the Renografin gradient was attempted by using sucrose gradients, but due to aggregation of the RNPs when removed from Renografin, the results were inconclusive. The sizing of the RNPs was next attempted by comparing the sedimentation characteristics of RNPs from the peak and trailing fractions of peak I to those of RNPs derived from a strain of defective-interfering (DI) particles containing a genome which is only 15% of the wild-type genome. We reasoned that the DI RNPs , which also band in CsCl at 1.31 g/cm3 (M. Leppert, personal communication), should be found in the same region of a nonequilibrium Renografin gradient as nascent RDYPs similar in size to the DI RNPs if the Renografin gradients could separate RNPs on the basis of size. Figure 7 shows that the nucleocapsids derived from the DI particles are found in exactly the same region of a nonequilibrium Renografin gradient as nascent RNPs, which are about 15% replicated (Fig. 7A and B; density, 1.18 g/cm3). Virion nucleocapsids cosedimented with the peak fractions of peak I at a density of 1.22 g/ cm3 (data not shown). The DI RNPs presumably have the same structure as the nascent RNPs, since DI RNPs and nascent RNPs have identical densities in both isopycnic Renografin (Fig. 7C) and CsCl gradients. This suggests, but does not
prove, that the RNPs contained in the trailing fractions are dissociated from their templates and are being separated from the peak fractions on the basis of size. However, we cannot exclude the possibility that conformational factors cause some of the nascent RNPs to move more slowly through the Renografin gradient, thus forming the trailing edge. Effect of Renografin on transcribing complexes. The total intracellular viral RNPs consist of transcribing complexes, nonenveloped virion RNPs to be assembled into virus particles, and replicating complexes. Previous studies have shown that intracellular RNPs cosediment with virion RNPs in sucrose gradients as 120S to 140S molecules. Nascent mRNA can be detected associated with some of these intracellular RNP complexes; thus, it is likely that the association of nascent product strands to template RNA does not greatly affect the sedimentation coefficient of the transcribing complex (31, 35). We wanted to determine whether replicative complexes were found in the 140S peak, as well as to determine the effect of Renografin on transcribing complexes. A 5-min-pulse-labeled cytoplasmic extract was prepared from VSV-infected cells at 5 h p.i. The extract was divided into two samples; one was centrifuged in a 15 to 47% Renografin gradient (Fig. 8A), and the other sample was centrifuged in a 15 to 30% sucrose gradient (Fig. 8B). The peak fractions from the Renografm gradient in Fig. 8A, which contained the replicative complex, were combined as indicated and recentrifuged in a 20 to 47% Renografin gradient
502
SIMONSEN, HILL, AND SUMMERS
,
JOURNAL OF VIROLOGY, Aug. 1979, p. 494-505 0022-538X/79/08-0494/12$02.00/0
Further Characterization of the Replicative Complex of Vesicular Stomatitis Virus CHRISTIAN C. SIMONSEN,* VIRGINIA M. HILL, AND DONALD F. SUMMERS Department of Cellular, Viral and Molecular Biology, University of Utah Medical Center, Salt Lake City, Utah 84132 Received for publication 23 March 1979
Replicating vesicular stomatitis virus ribonucleoprotein (RNP) complexes were isolated in nonequilibrium Renografin density gradients. These nascent RNPs had the same buoyant density as virion nucleocapsids in both isopycnic Renqgrafm and CsCl gradients. Both transcribing and replicating RNP complexes were shown to be stable in sucrose gradients, whereas only replicating RNP complexes were stable in Renografin gradients. Size analysis of the 5-min-pulse-labeled RNA species from the replicating RNPs using methylmercury gels revealed that the nascent strands were primarily less than full-length molecules. Longer times of radiolabeling demonstrated that the nascent RNA accumulated as 42S RNA, which was primarily of the same sense as the virion strand when it was radiolabeled at 5 h postinfection. The percentage of this radiolabeled RNA which was plus stranded was higher at 2.5 h postinfection, reflective of the shift in plus- to minus-strand full-length 42S RNA synthesis which occurs in the cell. Addition of cycloheximide to the infected cells before the addition of the radiolabel prevented the formation of these RNP complexes. Both the change in the percentage of minus strands found in the RNP complexes at the different times postinfection and the sensitivity to cycloheximide indicate that the RNP complex which was isolated was indeed the replicative complex. During the infection of HeLa cells by vesicular stomatitis virus (VSV), both virion minus-strand and complementary plus-strand full-length 42S RNA can be isolated from intracellular ribonucleoprotein (RNP) complexes (34, 36). These RNPs are composed of 42S RNA and three VSV-specific proteins: the N, or nucleocapsid, protein; the NS protein, a phosphorylated polypeptide; and the L protein (17, 18, 36). The virion nucleocapsid is able to direct the in vitro synthesis of the five VSV-specific mRNA's which, like the in vivo VSV mRNA's, are capped at the 5' end, methylated, and polyadenylated (2, 7, 20, 25-27, 32). Reconstitution experiments have shown that both the L and NS proteins are required to form the virion-associated RNA-dependent RNA polymerase (17, 18, 28). The template for transcription has been demonstrated to be composed of minus-strand 42S RNA and N protein (11, 17). The requirements for replication, however, are not well known. The temperature-sensitive mutant tsG114, which possesses a thermolabile L protein (21), has been used to provide indirect evidence that the L protein is involved in replication (30). Several other mutants suggest a role in replication for the N and NS proteins as well (15, 16). For these reasons it has been assumed
that replication occurs on an RNP template similar to the transcriptive RNP. Fractionation techniques previously used to isolate VSV RNPs do not separate replicating RNPs from transcribing RNPs. Sedimentation velocity centrifugation analyses have indicated that intracellular VSV RNPs sediment as 140S molecules, whether or not nascent mRNA is associated with the RNP complexes (31, 35). The lack of resolution is further complicated by the fact that replication accounts for only 10% of the total VSV-specific RNA synthesized in infected cells (38); thus, sucrose gradients have not been able to separate replicating RNPs from transcribing RNPs. CsCl gradients have been used to isolate glutaraldehyde-fixed messenger RNPs (19) and intracellular nucleocapsids (34); however, the messenger RNPs are not stable in CsCl without prior fixation with glutaraldehyde. RNPs isolated with CsCl gradients do not retain the L and NS proteins, which are known to comprise part of the functional RNPs in the cell (17, 18). In addition, all of the intracellular RNPs band at the same density in CsCl. For these reasons, CsCl gradients have not been used to separate the RNP complexes. lodinated density gradient media such as metrizamide and Renografm have been employed
494
VSV REPLICATIVE COMPLEX
VOL. 31, 1979
to obtained resolution of proteins, nucleic acids, and RNPs in the same gradients (6, 22). These media are hypertonic and minimize nonspecific RNA-protein interactions, yet do not dissociate most RNPs. Several investigators have used metrizamnide and Renografin to examine chromatin (12), influenza RNPs (10), and proteins (8). We have recently described the isolation in Renografin gradients of RNP complexes from VSVinfected HeLa cells (V. M. Hill, C. C. Simonsen, and D. F. Summers, submitted for publication). Several peaks of radioactivity were observed when an infected cell extract radiolabeled from 3 to 5 h postinfection (p.i.) was centrifuged in a Renografin gradient. One of those peaks contained predominantly minus-strand 42S RNA and was termed peak I. A similar pattern of radioactivity was seen after an infected cell extract pulse-labeled for 5 min at 5 h p.i. was analyzed, except that peak I exhibited a pronounced trailing edge. Virion nucleocapsids formed only a homogeneous peak I which lacked a trailing edge. Genome length and smaller RNAs were found in the peak fractions of the pulse-labeled peak I, whereas only the smaller RNAs were found in the trailing fractions. This RNA, pulse-labeled for 5 min at 5 h p.i., was 80% minus stranded, the same percentage observed with pulse-labeled 42S RNA obtained from intracellular nucleocapsids at 5 h p.i., when replication is maximal (34). The nascent minus-strand RNA was shown to be identical to the 5' end of the VSV genome RNA. We also observed that the nascent RNA was RNase resistant, was single stranded, and was isolated at a density of 1.240 g/cm3 in Renografin. All of this suggested that peak I contained nascent replicating VSV RNA associated with protein as a replicating RNP complex. Herein we extend these observations and further characterize the complexes to show by several different criteria that these RNPs are indeed VSV replicative complexes. MATERIALS AND MErHODS virus infection, and radioactive labeling. Cells, Suspension cultures of HeLa S3 cells were grown in Joklik modified minimal essential medium (Flow Laboratories) supplemented with 2 mM glutamine plus 5% fetal calf serum (Flow Laboratories) at a concentration of 2 x 105 to 8 x 10' cells per ml. Stock preparations of VSV (Indiana serotype) were grown in HeLa cells, purified, and assayed as described previously (3, 4). Cells were collected by centrifugation, resuspended in growth medium minus serum at a concentration of 6 x 106 cells per ml, and then infected with 10 PFU of VSV per cell. At 1 h p.i. serum was added to 5% and actinomycin D (a gift from Merck, Sharp & Dohme) was added to 1 yg/ml. At times ranging from 2.5 h to 5 h p.i. ['H]uridine or ['H]-
495
adenosine (25 to 30 Ci/mmol; New England Nuclear, Corp., Boston, Mass.) or both were added to a concentration of 200 ,tCi/ml for the indicated labeling periods, after which crushed, frozen medium was added and the cells were pelleted by centrifugation for 3 min at 900 x g. The cell pellet was lysed by the addition of 1 ml of 1% Nonidet P-40 in NT buffer (0.1 M NaCl, 0.01 M Tris, pH 7.5) and kept on ice for 10 min. The nuclei were removed by centrifugation as described above and were then resuspended in 1 ml of 0.2% deoxycholate-0.3% Nonidet P-40 in NT buffer and recentrifuged. The supernatants from the 1% Nonidet P-40 treatment and the deoxycholate-Nonidet P-40 wash were then combined and layered onto preformed CsCl or Renografin gradients. Isolation of RNP complexes. RNPs were isolated by using either Renografin, CsCl, or sucrose gradients. For isolation of the replicative RNPs, the cytoplasmic extracts from above were layered onto preformed 36ml, 15 to 50% or 20 to 60% Renografin (E. R. Squibb & Sons) gradients in NT buffer and centrifuged at 23,000 rpm and 4°C in an SW27 rotor for 16 h (for nonequilibrium gradients) or for 80 h (for equilibrium gradients). The gradients were fractionated by pumping from the bottom, and radioactivity was determined by directly assaying 50-1l portions in 7 ml of scintillation cocktail in a Beckman liquid scintillation spectrometer.
Preformed 20 to 40% (wt/wt) CsCl (Varlacoid Chemical Co.) gradients in TNE buffer (0.025 M Trishydrochloride, pH 7.5, 0.05 M NaCl, 0.002 M EDTA) were overlayed with fractions from the Renografin gradient which had been diluted 1:3 in NT buffer. These gradients were centrifuged in an SW41 rotor at 33,000 rpm for 16 h at 4°C. After centrifugation, the gradients were fractionated, and acid-precipitable radioactivity was determined as described previously (3, 4). The densities of the fractions were determined by measuring the refractive indexes with a Bausch and Lomb refractometer. Cytoplasmic extracts were also layered onto 15 to 30% sucrose gradients in NT buffer, which were centrifuged in an SW27 rotor for 16 h at 16,000 rpm and 4°C and fractionated as described above. Phenol extraction of RNA. An equal volume of phenol saturated with NETS buffer (0.1 M NaCl, 0.001 M EDTA, 0.01 M Tris, pH 7.5, 0.2% sodium dodecyl sulfate) was added to the samples in 3 ml of NETS buffer and mixed, and the two layers were separated by centrifugation. The aqueous layer was collected, and the phenol layer was extracted once more with 3 ml of NETS buffer. The aqueous fractions were combined, sodium acetate was added to a final concentration of 0.2 M, 2 volumes of ethanol were added, and the RNA was precipitated at -70°C overnight. RNA was pelleted by centrifugation at 16,000 x g for 30 min at40C. Purification of VSV-specific RNA. Minus-strand 42S RNA was prepared from purified virions by disrupting the virus in NETS buffer and centrifuging the samples in 15 to 30% sucrose-sodium dodecyl sulfate gradients in an SW41 rotor. The 42S RNA peak was ethanol precipitated and dissolved in 2xA buffer (0.3 M NaCl, 0.02 M Tris, pH 7.4, 0.002 M EDTA). VSV mRNA was prepared from infected cells at 4.5 h p.i. A
496
SIMONSEN, HILL, AND SUMMERS
cytoplasmic extract was prepared as described above and was phenol extracted, ethanol precipitated, and chromatographed on an oligodeoxythymidylic acidcellulose column. Oligo deoxythymidylic acid-cellulose chromatography of RNA. Oligodeoxythymidylic acid-cellulose (Collaborative Research, Inc., Waltham, Mass.) columns were prepared and eluted by the method of Banerjee and Rhodes (7). The high-salt elution buffer contained 0.01 M Tris-hydroxychloride, pH 7.4, and 0.5 M NaCl, and the low-salt buffer contained 0.01 M Tris-hydrochloride, pH 7.4. RNA was loaded onto the column in high-salt buffer. Fractions containing bound material were pooled and ethanol precipitated as described above. Hybridization of VSV RNA. Hybridizations were performed in sealed capillary pipettes in 0.02-mIl volumes by using the conditions of Kolakofsky (24). To each sample of [3H]RNA isolated from nucleocapsids were added increasing amounts of unlabeled VSV virion 42S RNA or mRNA. The samples were then heat denatured by heating to 115°C for 3 min and incubated for 3 h at 730C. Each sample was treated with a 30-,ug/ml solution of RNase A (Worthington Biochemicals Corp.) in 2xA buffer for 30 min at 250C and was then assayed for acid-precipitable radioactivity as described above. The self-annealing reactions were identical, except that no unlabeled RNA was added to the samples. Methylmercury-agarose gel electrophoresis. Methylmercury-agarose gel electrophoresis was performed by a modification of the method of Bailey and Davidson (5). Seakem agarose powder was obtained from Marine- Colloids, Inc., and methylmercury (II) hydroxide was obtained as a 1 M solution from Alfa Products, Danvers, Mass. Because of the toxic and volatile nature of the mercury compound, operations were performed under a hood. Horizontal slab gels (19 by 13.3 by 0.4 cm; 100-ml bed volume) were 1% agarose in borate buffer (0.05 M boric acid, 0.005 M sodium borate, 0.01 M sodium sulfate, 0.001 M trisodium EDTA, pH 8.0) and contained 0.005 M methylmercury hydroxide. Samples were applied in 40 ,ul of 1:1 sample-sample buffer (borate buffer containing 50% glycerol and 0.1% bromophenol blue), with methylmercury hydroxide added to a concentration of 0.005 M just before loading onto the gel. The gel reservoirs contained borate buffer, and this buffer was recirculated at approximately 80 ml/h throughout the electrophoresis period. Gels were run at 75 to 100 V (50 mA) for 6 h (dye front migrated 14 cm). Fluorography of methylmercury gels. Gels contaiing [3H]RNA were fixed for 10 min in 10% acetic acid containing 0.01 M cysteine. They were dehydrated in 100% methanol for two successive 1-h periods. After they were dried to paper thinness under a vacuum, the gels were soaked in a 10% (wt/vol) solution of 2,5-diphenyloxazole (New England Nuclear Corp.). in methanol for 3 h. Gels were soaked in water for 10 min to precipitate the 2,5-diphenyloxazole, blotted dry, and mounted on a, piece of 3 MM paper (Whatman, Ltd., England). After being covered with Saran Wrap, gels were exposed to Kodak SB-5 X-ray film at -70°C and developed after appropriate periods of time.
J. VIROL.
RESULTS Kinetics of labeling of peak I. We wished to determine whether 42S RNA would accumulate in the Renografin-isolated peak I if the length of the pulse-labeling period was increased. We also wished to learn whether the material in the trailing fractions of peak I would also accumulate during a longer pulse-labeling period. VSV-infected HeLa cells were radiolabeled for 5, 10, 15, and 30 min at 4.5 h p.i. Cytoplasmic extracts were prepared from each sample and centrifuged in identical 15 to 50% Renografin gradients. The results (Fig. 1) demonstrate that the trailing fractions of the 5-min sample (fractions 26 to 31) were not as apparent in the l0-min sample and were even less noticeable in the 15- and 30-min samples. All samples were continuously labeled; therefore, the material in the trailing fractions was still present in the longer-labeled fractions, but comprised a minor proportion of the total radiolabeled material due to the increase in the amount of radiolabel in the peak fractions (fractions 26 to 31, all samples). We next pooled the fractions of the replicative complex from Fig. 1. The trailing fractions of the 5-min-pulse-labeled sample were included because the radiolabeled RNA contained in these fractions represented an appreciable amount of the total radiolabeled RNA. Only the peak fractions from the remaining samples were pooled because the material in the trailing fractions did not represent a significant proportion of the total radiolabeled RNA. The RNA was extracted from these pools and analyzed in methylmercury-agarose gels. The densitometer tracings of the autoradiographs from these gels are shown in Fig. 2. The RNA from the 5-min-pulse-labeled samples was heterogeneous in size, ranging from less than 18S to 42S, although the average size of this sample was approximately 26S. The size distribution of the 10-min sample was also heterogeneous; however, the average size of the labeled RNA increased to a value of approximately 35S, and there was an obvious 42S band. Both the 15- and 30-min samples contained primarily 42S RNA, although there were significant amounts of heterogeneous material less than 42S RNA in size. This suggests that the radiolabel accumulated as genome length RNA as the pulse time was increased. Figure 2 also implies that it took longer than 5 min for 42S RNA to begin to accumulate. It is not clear why this was the case, since one would expect that a 5-min pulse-label would uniformly label all nascent RNA. Perhaps the synthesis of VSV 42S RNA does not occur at a uniform rate throughout the template; as we have noted previously, there is an apparent
VOL. 31, 1979
VSV REPLICATIVE COMPLEX
497
T 0
FRACTION NUMBER
FIG. 1. Renografin gradient centrifugation of cytoplasmic extracts from pulse-labeled VSV-infected cells. Infected cells at 4.5 h p.i. were labeled with 100 ,uCi each of [3H]adenosine and [3H]uridine per ml. Samples were removed at 5, 10, 15, and 30 min after the addition of the labels. Cytoplasmic extracts were prepared, layered onto 15 to 50% Renografin gradients, and centrifuged for 17 h at 23,000 rpm and 4°C in an SW27 rotor. Fractions (0.5 ml) were collected by pumping out from the bottom, and the radioactivity in each fraction was determined by counting 50-,iu portions in a toluene-based scintillation fluid.
pause site near the juncture of the L- and Gprotein cistrons (Hill et al., submitted for publication). Hybridization of pulse-labeled RNA from Renografin RNPs. We have recently reported that the ratio of plus- to minus-strand 42S RNA synthesis in infected cells is greater early in the infection (2.5 h p.i.) than when replication is maxinal at 5 h p.i., at which time 80% of the newly synthesized 42S molecules are minus stranded (34). We have also shown that pulselabeled nascent RNA species which are isolated from putative replication complexes at 5 h p.i. are approximately 80% minus stranded (Hill et al., submitted for publication).
To determine whether pulse-labeled nascent RNA extracted from the replication complexes would reflect the in vivo findings, infected HeLa cells were labeled for 5 min with [3H]uridine at 2.5 or 5 h p.i. Cytoplasmic extracts were prepared from these cells and were centrifuged in 15 to .50% Renografin gradients. The nascent RNA contained in the RNP complexes (at a peak density of 1.240 g/cm3) was hybridized with unlabeled VSV 42S RNA and VSV mRNA. The RNPs labeled at 2.5 h p.i. contained significantly more plus-strand RNA than the RNPs labeled at 5 h p.i. (Table 1). The change in the ratio of plus- to minus-strand RNA contained in these RNP complexes from approximately 35:65 at 2.5
498 origin
I
J. VIROL.
SIMONSEN, HILL, AND SUMMERS 42S
28S
I
I
185 5min
_
_
10min
E 0 to
cli
15min
.o cn
by the addition of the drug. This result also demonstrates that the cycloheximide-sensitive RNP complex is a replicative complex. Isopycnic gradient centrifugation of the replicative complexes. We next examined the RNP complexes in isopycnic Renografin gradients to study the basis for the isolation of the heterogeneous peak I. A 5-min-pulse-labeled cytoplasmic extract was prepared from cells at 5 h p.i. as described above and was centrifuged in a 15 to 50% Renografm gradient for 17 h at 23,000 rpm in an SW27 rotor. Fractions from three portions of the RNP peak were pooled as indicated (Fig. 4A) and combined, and samples were recentrifuged on identical 20 to 60% gradients for 17, 40, and 80 h (Fig. 4B, C, and D). The
replicative RNP peak, which was very heterogeneous after being recentrifuged for 17 h (Fig. 4B), formed a homogeneous peak at a density of 1.248 g/cm3 in the isopycnic gradients (Fig. 4C and D). Under the same conditions of centrifu30min
gation, VSV 42S RNA was found at 1.19 g/cm3 (Fig. 4E). Poliovirus top component (nucleocap-
sid which lacks RNA) and poliovirus infectious particles were found in the equilibrium gradients at 1.325 and 1.218 g/cm3, respectively (Fig. 4F). The result indicated that the ratio of protein to RNA in the nascent RNPs was the same 2
4 6 DISTANCE MIGRATED (cm)
throughout the peak
E
and
trailing
fractions of
peak I, since the smaller RNPs contained in the FIG. 2. Densitometer tracings of fluo,rographed trailing edge of peak I had the same isopycnic methylmercury-agarose gels. The RNA cointained in density in Renografin as the RNPs in the peak the pulse-labeled RNPs was extracted from m the frac- fractions. tions pooled as indicated in Fig. 1. Appiroximately 50,000 cpm of purified RNA from each stample was electrophoresed in a methylmercury-agaro ge fluorographed as described in the text. T'he autoradiograph was analyzed with an Ortec nMoael 4MlU densitometer. The scale is identical for a,11 samples.
getoand
h p.i. to 15:85 at 5 h p.i. is similar to tihe change in the ratio of plus- to minus-strand 42S RNA synthesis we have previously observe indicating that the plus-strand RNA slpecies isolated in peak I were products of replic ation. Effect of cycloheximide on the rebplicative RNA. A further means of differentiiating the replicative process from the transcripitive process is by using the protein synthesis inhibitor cycloheximide. The synthesis of the VSV mRNA's is not affected appreciably tby the addition of cycloheximide, whereas repllication of 42S RNA is rapidly inhibited (30, 38)I. Figure 3 shows that the synthesis of the peatk I RNA complex from Renografin was inhibit;ed by the addition of cycloheximide, whereas t:he formation of the slower-sedimenting peak II, which contains exclusively plus-strand seque-nces (Hill et al., submitted for publication), was anaffected d in vivo,
i
TABLE
1.
Hybridization analysis of the RNA
isolated from replicative RNPs from cytoplasmic extracts pulse-labeled at 2.5 and 5 hp.i.a Tixme p.i. (h)
Total
Self-annealed material
Unlabeled RNA added mRNA
cpm
%
42S RNA
cpm % cpm % 776 298 38 545 70 266 34 2.5 707 277 39 2.5 488 69 247 35 407 1,175 94 225 18 5 33 1,247 572 192 33 544 95 60 10 5 a Infected cells were pulse-labeled at either 2.5 or 5 h p.i. as described in the legend to Fig. 1. The replicative RNP complex was isolated from nonequilibrium Renografin gradients, and the RNA contained in the RNP was isolated by phenol extraction. The pulse-labeled RNA was mixed with increasing amounts of unlabeled VSV mRNA or 42S RNA in 20-Il reaction volumes by using the conditions of Kolakofaky (24). The samples were heated to 115°C for 1 min, annealed for 3 h at 730C, and then treated with 30 yg of RNase A per ml in 2xA buffer for 30 min at 220C. Plateau levels of RNase resistance are expressed after being corrected for the RNaseresistant material which was present when the sample was heat denatured but not annealed. This figure varied from 2 to 4% of the input radioactivity. Self-annealing of the material was determined as described above, except that no unlabeled RNA was added to the samples.
VSV REPLICATIVE COMPLEX
VOL. 31, 1979
499
Pe E
n Number
FIG. 3. Effect of cycloheximide on the production of replicative RNP complex. A culture of infected cells was split in half at 4.5 hp.i., and to one-half of the culture cycloheximide was added to a final concentration of 100 tLg/ml. After 15 min in the presence of the drug, [3H]uridine was added to each culture to a final concentration of 100 ,aCi/ml. After a 5-min incubation, the cells were removed by centrifugation and cytoplasmic extracts were prepared as described in the text. The extracts were layered onto separate identical 15 to 50% Renografin gradients and were centrifuged for 16 h at 23,00( rpm and 4°C in an SW27 rotor. The radioactivity from the fractionated gradients was analyzed as described previously. Symbols: -4, untreated control; 0-----S cycloheximide-treated sample.
The nascent RNPs were further analyzed by centrifuging fractions from the peak and trailing fractions of peak I in CsCl gradients (Fig. 5). The nascent RNPs had identical equilibrium densities in CsCl. This result also suggests that the nascent RNPs have a protein-RNA ratio sunilar to that of virion nucleocapsids, since all of the radioactivity was found at a density of 1.31 g/cm3 (19). The nascent RNA was then removed from the RNPs and analyzed in sucrose-sodium dodecyl sulfate gradients, which showed that the RNA contained in the CsClbanded RNPs was shorter than full length and not just 42S RNA (Fig. 6). The demonstration that nascent RNPs are stable in CsCl is further evidence suggesting that these nascent RNPs are derived from a replicative intermediate since transcriptive products do not band in CsCl (19). It thus appears likely that the heterogeneity of the 5-min-pulse-labeled RNP complex which has been isolated is not due to a difference in density between the RNPs in the peak and the trailing fractions. Examination of the trailing fractions of peak I. The results shown in Fig. 1 demonstrated that the pulse-labeled material in the trailing fractions did not accumulate during long pulse-labeling periods. Furthermore, it has been
shown that nascent RNA is found in both the peak and the trailing fractions of peak I, but detectable newly replicated 42S molecules are found in the peak fractions only (Hill et al., submitted for publication). These results perhaps suggest that the material in the trailing fractions is nascent RNA which has dissociated from the template RNA. We have observed that the nonequilibrium Renografin gradients resolve RNA molecules on the basis of size; thus, it is possible that the trailing fractions contain smaller RNPs separated on the basis of size. At the present time there are two possible ways that this problem can be studied directly. Electron microscopic examination of the material contained in the peak and trailing fractions was performed; however, due to the large number of nonradiolabeled structures present in the same area as peak I, we could not prove that the trailing edge represented nascent strands dissociated from their templates (C. Naeve, unpublished data). We did observe possible replication forms in the peak fractions (Naeve, manuscript in preparation), but due to the high background of nonreplicating RNPs, it was impossible to deal conclusively with the problem of nascent strand attachment to template RNP. Another way of directly showing that the peak fractions
500
J. VIROL.
SIMONSEN, HILL, AND SUMMERS
7d
E
-
5
3
20
5
25
30
35
5
1015
F
E
25
20
30
35
C
~~~~~~~~~~~~8-
4-
..
2-
-1.20
4-
-1.10 25
30
40
35
45
50
5 55 FRACTION NUMBER
'15
10
20
30
25
'35
FIG. 4. Equilibrium Renografin densitygradient centrifugation of the 5-min-labeled RNP complex. Infected cells at 5 h p.i. were labeled by the addition of 100 ,LCi each of [3H]uridine and [3H]adenosine per ml. Crushed, frozen medium was added to the sample after 5 min in the presence of the labels, and the cells were removed by centrifugation. A cytoplasmic extract was prepared and layered onto a 15 to 50% Renografin gradient which was centrifuged for 16 h at 23,000 rpm and 4°C in an SW27 rotor. The radioactivity was quantitated as described in the text. The fractions indicated fiom the peak and trailing fractions in (A) were combined and diluted. Samples were layered onto identical 20 to 60% Renografin gradients and were centrifuged at 23,000 rpm and 4°C in an SW27 rotor for 17 h (B) 40 h (C), or 80 h (D). 3H-labeled 42S RNA (E) and poliovirus nucleocapsid and top component (F) were centrifuged in identical Renografin gradients at 23,000 rpm and 4°C for 80 h.
8-A
Peak H2-B
PeokI
C
D
*140
6g E
.4
|~1 2" 23': ~ ,
,
,
.
,
2
.
5
1"
5
,
130 .120
2-
5 10 1520 25
30 35
5 10
15
FRACTION NUMBER
5
8
15
5
10
15
FIG. 5. CsCl gradient centrifugation of peak I from a Renografin gradient. A cytoplasmic extract from infected cells pulse-labeled for 5 min at 5 h p.i. was layered onto a 15 to 50% Renografin gradient and centrifuged as for Fig. 1. The RNP peak was split into three pools as indicated in (A). The pools were then layered onto 20 to 40% (wt/wt) CsCl gradients in TNE buffer and centrifuged for 16 h at 33,000 rpm and 4°C in an SW41 rotor. Fractions (0.5 ml) were collected from the bottom, trichloroacetic acid-precipitable radioactivity was determined for 250 itl of each fraction, and the density was recorded by measuring the refractive indexes of the fractions with a Bausch and Lomb refractometer. (B) CsCl gradient ofpool 1 from (A). (C) Pool 2. (D) Pool 3.
VSV REPLICATIVE COMPLEX
VOL. 31, 1979
501
FRATION NUMBER
FIG. 6. Size analysis of the pulse-labeled RNA contained in CsCl-banded replicative RNPs. The pulselabeled RNPs which were banded in CsCl in Fig. 5B, C, and D were phenol extracted, and the RNA contained in these complexes was centrifuged in 15 to 30% sucrose-sodium dodecyl sulfate gradients for 16 h at 20,000 rpm in an SW41 rotor. Trichloroacetic acid-precipitable radioactivity was determined for each of the fractions, which were collected by dripping from the bottom of the tube. Symbols: 0, ['4C]uridine-labeled 42S, 28S, 18S, and 4S markers; *, [3H]uridine-labeled RNA from Fig. 5B, C, and D (panels A, B, and C, respectively).
contain nascent strands attached to their template is to complete the nascent strands in vitro; however, at the present time this has not been possible. It was therefore necessary to study this problem indirectly. A size analysis of the RNPs contained in the peak and trailing fractions of the Renografin gradient was attempted by using sucrose gradients, but due to aggregation of the RNPs when removed from Renografin, the results were inconclusive. The sizing of the RNPs was next attempted by comparing the sedimentation characteristics of RNPs from the peak and trailing fractions of peak I to those of RNPs derived from a strain of defective-interfering (DI) particles containing a genome which is only 15% of the wild-type genome. We reasoned that the DI RNPs , which also band in CsCl at 1.31 g/cm3 (M. Leppert, personal communication), should be found in the same region of a nonequilibrium Renografin gradient as nascent RDYPs similar in size to the DI RNPs if the Renografin gradients could separate RNPs on the basis of size. Figure 7 shows that the nucleocapsids derived from the DI particles are found in exactly the same region of a nonequilibrium Renografin gradient as nascent RNPs, which are about 15% replicated (Fig. 7A and B; density, 1.18 g/cm3). Virion nucleocapsids cosedimented with the peak fractions of peak I at a density of 1.22 g/ cm3 (data not shown). The DI RNPs presumably have the same structure as the nascent RNPs, since DI RNPs and nascent RNPs have identical densities in both isopycnic Renografin (Fig. 7C) and CsCl gradients. This suggests, but does not
prove, that the RNPs contained in the trailing fractions are dissociated from their templates and are being separated from the peak fractions on the basis of size. However, we cannot exclude the possibility that conformational factors cause some of the nascent RNPs to move more slowly through the Renografin gradient, thus forming the trailing edge. Effect of Renografin on transcribing complexes. The total intracellular viral RNPs consist of transcribing complexes, nonenveloped virion RNPs to be assembled into virus particles, and replicating complexes. Previous studies have shown that intracellular RNPs cosediment with virion RNPs in sucrose gradients as 120S to 140S molecules. Nascent mRNA can be detected associated with some of these intracellular RNP complexes; thus, it is likely that the association of nascent product strands to template RNA does not greatly affect the sedimentation coefficient of the transcribing complex (31, 35). We wanted to determine whether replicative complexes were found in the 140S peak, as well as to determine the effect of Renografin on transcribing complexes. A 5-min-pulse-labeled cytoplasmic extract was prepared from VSV-infected cells at 5 h p.i. The extract was divided into two samples; one was centrifuged in a 15 to 47% Renografin gradient (Fig. 8A), and the other sample was centrifuged in a 15 to 30% sucrose gradient (Fig. 8B). The peak fractions from the Renografm gradient in Fig. 8A, which contained the replicative complex, were combined as indicated and recentrifuged in a 20 to 47% Renografin gradient
502
SIMONSEN, HILL, AND SUMMERS
,
