Solubilization and Immunoprecipitation of ... - Journal of Virology

Vol. 65, No. 3

JOURNAL OF VIROLOGY, Mar. 1991, p. 1496-1506

0022-538X/91/031496-11$02.00/0 Copyright © 1991, American Society for Microbiology

Solubilization and Immunoprecipitation of Alphavirus Replication Complexes DAVID J. BARTON, STANLEY G. SAWICKI, AND DOROTHEA L. SAWICKI*

Department of Microbiology, Medical College of Ohio, Toledo, Ohio 43699 Received 24 September 1990/Accepted 11 December 1990

Alphavirus replication complexes that are located in the mitochondrial fraction of infected cells which pellets at 15,000 x g (P15 fraction) were used for the in vitro synthesis of viral 49S genome RNA, subgenomic 26S mRNA, and replicative intermediates (RIs). Comparison of the polymerase activity in P15 fractions from Sindbis virus (SIN)- and Semliki Forest virus (SFV)-infected cells indicated that both had similar kinetics of viral RNA synthesis in vitro but the SFV fraction was twice as active and produced more labeled RIs than SIN. When assayed in vitro under conditions of high specific activity, which limits incorporation into RIs, at least 70% of the polymerase activity was recovered after detergent treatment. Treatment with Triton X-100 or with Triton X-100 plus deoxycholate (DOC) solubilized some prelabeled SFV RIs but little if any SFV or SIN RNA polymerase activity from large structures that also contained cytoskeletal components. Treatment with concentrations of DOC greater than 0.25% or with 1% Triton X-100-0.5% DOC in the presence of 0.5 M NaCl released the polymerase activity in a soluble form, i.e., it no longer pelleted at 15,000 x g. The DOC-solubilized replication complexes, identified by their polymerase activity in vitro and by the presence of prelabeled RI RNA, had a density of 1.25 g/ml, were 20S to 100S in size, and contained viral nsPl, nsP2, phosphorylated nsP3, nsP4, and possibly nsP34 proteins. Immunoprecipitation of the solubilized structures indicated that the nonstructural proteins were complexed together and that a presumed cellular protein of -120 kDa may be part of the complex. Antibodies specific for nsP3, and to a lesser extent antibodies to nsPl, precipitated native replication complexes that retained prelabeled RIs and were active in vitro in viral RNA synthesis. Thus, antibodies to nsP3 bound but did not disrupt or inhibit the polymerase activity of replication complexes in vitro. Cells infected with alphaviruses such as Sindbis virus (SIN) or Semliki Forest virus (SFV) contain a virus-specific RNA-dependent RNA polymerase activity that is responsible for the synthesis of three species of RNA: the viral positive-strand 49S genome RNA and subgenomic 26S mRNA and the 49S negative-strand RNA that is the template for synthesis of both species of positive-strand RNAs. The formation of the polymerase activity requires translation of the incoming genome RNA into nonstructural proteins (nsPs), which are numbered according to the positions of the genes from the 5' end of the genome RNA (reviewed in reference 58). The nsPs are synthesized initially as polyproteins that are subsequently cleaved to yield four mature proteins, nsPl to nsP4. For certain alphaviruses, limited readthrough of an opal codon located between nsP3 and nsP4 results in synthesis of the fusion protein nsP34 (56, 57). nsPl, nsP2, and nsP3 are present in larger amounts in infected cells than is nsP4 or nsP34 (28, 33). Recent evidence indicates that nsP4, which contains the conserved X-Gly-Asp-Asp (XGDD) sequence common to many RNA-dependent polymerases (26, 43) and an autoprotease activity (59), is an essential component of the alphavirus polymerase and functions in elongation and possibly initiation (4, 18, 47, 52). It is not known whether nsP34 has a function different from that of nsP4, as has been postulated (8, 21). nsPl has been implicated in negative-strand synthesis (19, 50, 60) and in the methylation and capping of positive strands (7, 37). nsP2 contains a conserved domain found in nucleotide-binding and helicase proteins (14), possesses an autoprotease activity (9, 23), and may interact directly with nsP4 (19). Mutations in nsP2 result in temperature-sensitive

*

(ts) synthesis of 26S mRNA (19, 48, 55), which suggests that a domain of nsP2 affects recognition by the polymerase of the recently mapped internal promoter in the 49S negativestrand template (15). No function has been demonstrated yet for nsP3, which is phosphorylated (32, 33, 42). However, a mutation in nsP3 of SIN ts7 is responsible for its ts RNAnegative phenotype (19). Therefore, replication complexes would be expected to contain the other nsPs, and perhaps host proteins, in addition to nsP4 or nsP34. The synthesis of viral RNA initiates with the synthesis of the first negative strand. Negative strands in turn serve as the preferred templates in replication complexes that are stable once formed and that function to synthesize genome and subgenome mRNAs (49). Increasing numbers of replication complexes are formed early in infection, coinciding with the synthesis of increased numbers of negative strands (49, 51). After this early period, synthesis of negative strands ceases, the rate of positive-strand synthesis changes from an exponentially increasing rate to a constant one, and the number of replication complexes and replicative intermediate RNAs (RIs) becomes constant. The alphavirus replication complex is associated with cytoplasmic membranous structures that copurify with the mitochondrial fraction of infected cells (10, 16, 35, 36, 38, 54). Infected cells were found also to contain cytopathic vacuoles whose numbers increased with time after infection (10, 11), but their relationship to the viral replication complexes has not been established. More recently, Clewley and Kennedy (5), Ranki and Kaariainen (46), and Gomatos et al. (13) reported the purification of SFV replication complexes after solubilization of the mitochondrial pellet (P15) fraction of infected cells with various detergents. These investigators found that there was loss of polymerase activity after detergent treatment and that some of the SFV replication complexes were

Corresponding author. 1496

VOL. 65, 1991

released as 25S to 40S structures. We have continued these investigations and have determined conditions for the solubilization of active SFV and SIN replication complexes. We also utilized antibodies that were obtained from Hardy and Strauss (22) and that are specific to the viral nsPs to identify the SIN nsPs and to demonstrate that they were present in solubilized replication complexes. We report that the antibodies to nsP3 were the most effective in immunoprecipitating solubilized replicative complexes that retained their RNA templates and polymerase activities after immunoprecipitation. MATERIALS AND METHODS Virus and cell cultures. Baby hamster kidney (BHK-21) cells were grown in Dulbecco modified Eagle minimum essential medium supplemented with 6% newborn calf serum and 5% tryptose phosphate broth (52). The heat-resistant strain of SIN (SIN HR) has been described previously (50). Infection and preparation of the P15 fraction. Monolayers of BHK cells in 150-mm-diameter plastic petri dishes (-70 x 106 cells per dish) were infected with a multiplicity of infection of 30. At the end of a 1-h adsorption period, the monolayers were refed with 37°C medium and incubated at 37°C until 5 h postinfection (p.i.), when the cells were harvested in the cold and the P15 fraction was obtained as described previously (4). Briefly, the cells were rinsed with ice-cold phosphate-buffered saline (PBS), scraped from the dish in PBS, and collected by centrifugation at 900 x g for 5 min at 4°C. The cells, typically from four 150-mm-diameter dishes, were resuspended in 15 ml of RS buffer (10 mM Tris hydrochloride, pH 7.8, 10 mM NaCI) and were allowed to swell on ice for 15 min before being broken open in a Dounce homogenizer. The nuclei were collected by centrifugation at 900 x g for 5 min at 4°C and were resuspended in 15 ml of RS buffer on ice. The postnuclear supernatant was centrifuged at 15,000 x g for 20 min at 4°C in the Beckman JS13.1 rotor, and both the supernatant (S15 fraction) and the pellet (P15 fraction) were collected. The P15 fraction was suspended in storage buffer (10 mM Tris hydrochloride, pH 7.8, 10 mM NaCl, 15% glycerol) at 1 to 4 mg of protein per ml and stored in aliquots at -80°C. In vitro transcription reaction. The amount of P15 fraction used in a typical reaction was equivalent to 2.5 x 105 cells. Equal volumes of P15 and a 2x reaction mixture (100 mM Tris hydrochloride, pH 7.8; 100 mM KCl; 7 mM MgCl2; 20 mM dithiothreitol; 20 ,ug of dactinomycin per ml; 10 mM creatine phosphate; 50 ,g of creatine phosphokinase [Calbiochem, San Diego, Calif.] per ml; 4 mM [each] ATP, GTP, and UTP and 0.4 mM CTP for low-specific-activity reactions and no unlabeled CTP for high-specific-activity reactions; 3 mCi of [ca-32P]CTP [ICN Radiochemicals, Irvine, Calif., or Dupont NEN, Wilmington, Del.] per ml; 400 to 800 U of RNasin [Promega Biotec, Madison, Wis.] per ml) were incubated at 30°C for 30 min or as indicated in the text. Reactions were terminated by the addition of 5% lithium dodecyl sulfate (LDS) containing 100,ug of proteinase K per ml when RNA products were being assayed. After phenol and chloroform extraction and ethanol precipitation of the samples, the RNA was analyzed by separation on 0.8% agarose gels in TBE (89 mM Tris base; 89 mM boric acid; 2 mM EDTA) or 0.8% agarose gels containing 2.2 M formaldehyde in morpholinepropanesulfonic acid (MOPS) buffer (20 mM MOPS, 5 mM sodium acetate, 1 mM EDTA, pH 7.0) and was visualized by autoradiography of the dried gels. Solubilization and velocity and buoyant-density centrifuga-

ALPHAVIRUS REPLICATION COMPLEX

1497

tion. Aliquots of the P15 fraction in RS buffer were mixed with an equal volume of 2x detergent (either 2% Triton X-100 [Pierce Chemical Co., Rockford, Ill.] plus 1% sodium deoxycholate [DOC; Sigma Chemical Co., St. Louis, Mo.] in RS buffer or 2% DOC in RS buffer), and the samples were immediately processed for centrifugation. To determine the sizes of the detergent-treated replication complexes, the P15 was layered onto 15 to 30% glycerol gradients in 50 mM Tris-HCl, pH 7.8, 150 mM NaCI, 1 mM EDTA, and 0. 1% Triton X-100 or 0.5% DOC, respectively, and centrifuged at 4°C for 18 h at 25,500 rpm in the Beckman SW 28.1 rotor. Gradient fractions were collected from the bottom and analyzed through a UV monitor, with 60S and 40S ribosomal subunits serving as sedimentation markers. The buoyant density of the replication complexes was determined by adjusting the sample to 30% sucrose and to 0.6 ml and layering it on top of 0.5 ml of 50% sucrose and 1 ml of 60% sucrose (gradients without detergent were made with sucrose in RS buffer, while detergent-treated samples were analyzed on gradients of sucrose in RS and 0.1% Triton X-100 or 0.5% DOC). Centrifugation was in the Beckman TLS-55 rotor (2.2-ml tubes) for 259,000 x g for 20 h at 4°C. Fractions (0.1 ml) were collected, and their densities were determined from the refractive indices. Protein labeling and immunoprecipitation. Cultures of BHK-21 cells were infected with SIN HR at 37°C. Beginning at 2 h p.i., the cells were given a 2-h pulse-label in Dulbecco modified Eagle minimum essential medium with 1% of the methionine and cysteine, 100 ,XCi of Trans 3"S-label per ml (1,150 Ci/mmole; ICN Radiochemicals), 2% dialyzed fetal bovine serum, 2 ,ug of dactinomycin per ml, and 22 mM N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid, pH 7.4, after which they were chased for 2 h in Dulbecco modified Eagle minimum essential medium supplemented with 3% newborn calf serum and 5% tryptose phosphate broth before harvest and fractionation. For direct analysis of the labeled proteins, aliquots of the cellular fractions were adjusted to 1% LDS, heated at 100°C for 3 min, and rapidly cooled before electrophoresis on 5 to 10% polyacrylamide gels with buffers described by Laemmli (30). For immunoprecipitation, the denatured proteins in 1% LDS were diluted to 0.1% LDS in RIPA by the addition of 9 volumes of modified radioimmunoprecipitation (RIPA) buffer (1) that contained 1% Triton X-100, 1% DOC, 10 mM Tris hydrochloride, pH 8.1, and 150 mM NaCl but lacked LDS. Equal aliquots of cellular fractions that had not been denatured with LDS and heat were diluted in modified RIPA buffer. The solutions were preabsorbed with group G streptococcal cells (100 ,ul of a 10% solution of heat-killed streptococcal G cells per solubilized P15 from 5 x 105 infected cells), followed by end-over-end mixing for 20 min at room temperature. The bacterial cells were pelleted at 10,000 x g for 5 min, and the supernatant was divided into 600-pu aliquots for immunoprecipitation, essentially as described previously (22). The immunoprecipitates were resolved on S to 10% polyacrylamide gels in Laemmli buffer. RESULTS Characterization of the in vitro polymerase activity. The P15 fraction from SIN- or SFV-infected cells was enriched in viral replication complexes and contained about 80% of the polymerase activity (Fig. 1 and Table 1). The nuclear and the cytoplasmic supernatant (S15) fractions each contained about 10% of the polymerase activity (Table 1). RNA synthesized in vitro migrated on agarose gels at the positions

1498

J. VIROL.

BARTON ET AL.

B

A Low Specific Activity

High Specific Activity _

LO

O

,

I-

z

(In

EL

o -

C)

,

7 z

LO

L

RI -u-W 49S-~46ss- _

26S

--o

FIG. 1. Active replication complexes were enriched in the P15 fraction, and transcription in vitro required all four nucleoside triphosphates. (A) SIN-infected BHK cells were harvested at 5 h p.i. and fractionated as described in the text. Equal proportions of the total-cell lysate and the nuclear, S15, and P15 fractions (equivalent to 2.5 x 105 cells) were incubated in vitro at 30°C for 30 min in either low-specific-activity or high-specific-activity (no supplementation of the radioactive CTP) reaction conditions before the products were analyzed on 0.8% agarose gels in TBE buffer. (B) Aliquots of the P15 fraction from SIN-infected BHK cells were incubated in vitro for 30 min at 30°C in either the complete low-specific-activity reaction mixture containing 32P-CTP and all four unlabeled ribonucleoside triphosphates or in reaction mixtures lacking (W/O) one or three of the unlabeled nucleoside triphosphates. The products were deproteinized and analyzed by electrophoresis on 0.8% agarose gels in MOPS-formaldehyde buffer.

of the viral RIs and the two viral single-stranded RNAs, the genomic 49S RNA and the subgenomic 26S mRNA (Fig. 1A). As reported recently (3), radiolabel appears after 1 min of incubation in the viral RIs and subsequently appears and accumulates in the two single-stranded RNAs. Treatment of the radiolabeled RNAs with pancreatic RNase before electrophoresis yielded the expected RNase-resistant replicative forms (RFs) of the RIs, the RFI, RFII, and RFIII molecules (data not shown) (3, 53). The polymerase activity required all four nucleoside triphosphates and specifically incorporated radiolabeled CTP into all viral RNA species. Little or no label was incorporated into viral RNAs under conditions in which one or more of the four triphosphates were omitted (Fig. 1B), whereas labeling of tRNA occurred maximally TABLE 1. P15 fraction enriched in SIN polymerase activity Fraction

Total Nuclear

S15 P15

pmol of CMP incorporated/mg of proteina Low specific activity

High specific activity

22 9 5 81

0.4 0.3 0.1 3.0

a Incorporation of radiolabel into viral RNA was determined by Cerenkov counting the RI, 49S, and 26S RNAs after electrophoresis on an agarose gel. The amount of protein in the cell fractions was determined with the BCA reagent (Pierce Chemical Co., Rockford, Ill.). The specific activity of CMP in the low-specific-activity reaction was 17,000 cpm/pmol and in the high-specific activity reaction was 6.8 x 106 cpm/pmol.

when one or more of the unlabeled nucleoside triphosphates were omitted (data not shown). The small amount of labeling of RIs and 26S RNA molecules in reactions lacking one or more triphosphates was probably due to the limited elongation or release of almost-completed chains, because the amounts varied inversely with the base composition (1 G, 15 A, and 28 U residues) of the 3' heteropolymeric sequence of 26S and 49S RNA. In addition to the usual low-specificactivity labeling conditions (in which the radiolabeled CTP was supplemented with 200 ,uM unlabeled CTP), reactions were carried out also at high specific activity (the radiolabeled CTP was not supplemented with unlabeled CTP). The high-specific-activity reactions yielded predominantly radiolabeled RI RNA and little or no radiolabeled 49S and 26S RNAs. Some single-stranded RNA, mostly 26S mRNA, was released in the whole (total) lysates, probably because of endogenous unlabeled nucleoside triphosphates. About 30 times more RNA was synthesized under low-specific-activity labeling conditions than under conditions of limited CTP concentrations (Table 1); however, more 32P-CMP was found in the RI and RF RNA under high-specific-activity conditions. We have shown previously (4) that the RNA pulse-labeled under conditions of high specific activity is chased into single-stranded 49S and 26S viral RNA after the addition of unlabeled nucleoside triphosphates. We also monitored nucleoside triphosphatase and RNase activities in the P15 fractions. The amount of radiolabeled CTP that was converted to CDP, CMP, and Pi during the reactions was analyzed by thin-layer chromatography and

VOL. 65, 1991

A

ALPHAVIRUS REPLICATION COMPLEX

B

Prelabel 16

Pre-label

SIN

SIN

cJ

60

SFV

) 0~~~~~~~~~

55

° 12

Activity E

50 soU

X_

F

cJ )

1499

SFV v 0

.FNa.CL oo-IE ,

45O3 cn

UL8 0 8

RI -_

40 35

30

49S -.-

Fraction

26S

Activity 8 0

,

1.25g/cc

o-

60 55 50

(A 0 UM L-

45 U) 40 35 30

2

4

6

8 10 12 14 16 18

Fraction FIG. 2. SIN and SFV replication complexes have a density of 1.25 g/ml after treatment with 1% DOC or with 1% Triton X-100-0.5% DOC. Representative results are shown. (A) The P15 fraction from SIN-infected cells was incubated in vitro under high-specific-activity labeling conditions (Prelabel) and was repelleted before treatment with 1% DOC and buoyant-density centrifugation as described in Materials and Methods. An aliquot of each gradient fraction was acid precipitated directly. To identify polymerase activity (Activity), P15 fractions were treated directly with 1% DOC before centrifugation. Every second gradient fraction was incubated in vitro at 30°C for 30 min in high-specific-activity reactions, and the products were deproteinized before electrophoresis on 0.8% agarose-TBE gels. The incorporation in viral RIs was determined by cutting and counting that area of the gel. (B) Fluorogram of a 0.8% agarose gel in TBE showing that the RNA polymerase activity causing the radiolabeling of RI RNA was in the 1.25-g/ml fractions from SIN- and SFV-infected cells and that the prelabeled RIs were also in the same fraction.

indicated that the P15 fractions prepared from mock- or virus-infected cells contained similar levels of phosphatase activity and that this activity was partially responsible for the cessation of incorporation (data not shown). Viral RNAs continued to be synthesized in vitro as long as the nucleoside triphosphate concentration did not decrease to less than 50% of the initial concentration, i.e., less than 0.05 to 0.1 mM. Use of less concentrated P15 or the addition of more nucleoside triphosphates increased the time over which incorporation continued. Endogenous RNases also contributed to a failure to continue linear incorporation beyond 90 min and to a loss of previously synthesized single-stranded RNA. RNase activity was detected by incubating radiolabeled transcripts of the infectious clone of SIN in various reactions and determining the amount of intact labeled RNA at the end of the incubation period (data not shown). These nucleases were not inhibited by RNasin (Promega, Madison Wis.). Lower incubation temperatures (25 to 30°C) and shorter incubation times (15 to 30 min) gave the highest recovery of polymerase products in vitro. DOC efficiently solubilized the SIN and SFV replication complexes. We determined the extent to which polymerase

activity was affected by treatment of the P15 fraction with Triton X-100 and DOC, used singly or in combination and at concentrations of each detergent above its critical micelle concentration. Detergent-treated fractions had increased levels of RNase (data not shown) that most likely were responsible for the observed low recovery of single-stranded RNAs when the polymerase activity was assayed under low-specific-activity labeling conditions. Under high-specific-activity labeling conditions, RI RNAs were radiolabeled to high levels and at least 70% of the original polymerase activity remained after exposure to the highest concentrations of detergent (data not shown). In the absence of detergent treatment, the polymerase activity in SIN and SFV replication complexes in the P15 fraction was associated with 1.16 g/ml smooth membranes. After treatment with either 1% DOC alone (Fig. 2A) or 1% Triton X-100 plus 0.5% DOC (Fig. 2B), which solubilized lipids and 80% of the membrane proteins, active SIN replication complexes had a density of 1.25 g/ml. The radiolabel incorporated by the SIN and SFV polymerases was in viral RIs and RFs (Fig. 2B). Prelabeled SIN and SFV RIs, which were obtained by incubation in vitro and repelleting of the

1500

J. VIROL.

BARTON ET AL.

41310 800X 0~

600 > w

400 -0 _

m

-

200