key: cord-0844055-u8otoxn4
authors: Sawicki, Stanley G.; Sawicki, Dorothea L.
title: The effect of loss of regulation of minus-strand RNA synthesis on sindbis virus replication
date: 1986-06-30
journal: Virology
DOI: 10.1016/0042-6822(86)90054-1
sha: fa5d0074495d60de05c6a95a182bbc299f2fb7d5
doc_id: 844055
cord_uid: u8otoxn4

Abstract During the replication cycle of Sindbis virus minus-strand synthesis stops normally at the time that plus-strand synthesis reaches a maximum rate. We have isolated and characterized revertants of ts24, a member of the A complementation group of Sindbis HR mutants, that we had demonstrated previously to have a temperature-sensitive defect in the regulation of minus-strand synthesis. These revertants of ts24 replicated efficiently at 40° but nevertheless retained the temperature sensitive defect in the regulation of minus-strand synthesis: they continued to synthesize minus strands late in the replication cycle at 40° but not at 30° and in the presence or absence of protein synthesis. Although failure to regulate the synthesis of minus strands resulted in continuous minus-strand synthesis and in the accumulation of minus strands, the rate of plus-strand synthesis was not increased concertedly. Minus strands synthesized after the rate of plus-strand synthesis had become constant were demonstrated to be utilized as templates for 26 S mRNA synthesis. Thus, the change from an increasing to a constant rate of plus-strand synthesis during the alphavirus replication cycle cannot be governed solely by the number of minus strands that accumulate in infected cells. We present a model for the preferential utilization of minus strands as a mechanism for the cessation of minus-strand synthesis that occurs normally during alphavirus replication.

Sindbis virus (SIN) is an enveloped, plusstrand RNA virus which replicates in a wide variety of vertebrate and invertebrate cells (Kennedy, 1980; Strauss and Strauss, 1983) . Replication of SIN virus requires the 49 S genome RNA to serve initially as a messenger RNA for the translation of viral encoded nonstructural proteins which function as components of a viral RNAdependent RNA polymerase. The RNA polymerase transcribes the parental 49 S plus-strand RNA into a complementary 49 S minus-strand which is recognized in turn as a template by the RNA polymerase for the synthesis of additional plus strands. In addition to serving as a template for 49 S plus-strand RNA, the minus strand also serves as a template for the synthesis via internal initiation of a subgenomic plusstrand RNA that sediments at 26 S and is ' To whom requests for reprints should be addressed.

3'-coterminal with the 49 S plus strand and that is translated into the viral structural proteins.

Whereas the synthesis of plus strands once started continues throughout the replication cycle and does not require continued protein synthesis (Wengler and Wengler, 1975) , the synthesis of minus strands is limited to the early phase of the replication cycle and requires de novo protein synthesis (Sawicki and Sawicki, 1980) . Our studies (Sawicki et al, 1981a ) demonstrated that minus-strand RNA synthesis required a viral encoded function that was temperature sensitive in tall, the sole member of the B complementation group of the heatresistant strain (HR) of SIN. In contrast to tall, ts24 of the A complementation group was demonstrated to have a temperature-sensitive (ts) defect in the shutoff of minus-strand RNA synthesis (Sawicki et ak, 1981b) . This was the phenotype that we predicted for a mutant which was defective in the regulation of minus-strand RNA synthesis and suggested that a viral (6 X 10' PFU/ml) used in these studies was encoded polypeptide functioned to shut off prepared by infecting CEF cells at a m.o.i. the synthesis of minus strands. Because we of less than 1 PFU/cell with virus from the could turn back on minus-strand synthesis first stock of ts24R1, without any additional in ts24 infected cells by shifting to the non-plaque purification. permissive temperature late in the repli-Infection and RNA labeling. Monolayers cation cycle after minus-strand RNA syn-of CEF cells in plastic petri dishes, usually thesis had shut off and in the absence of 60 mm, were infected with a m.o.i. of 100 new protein synthesis, we concluded that with either SIN HR, one of the mutants, proteolytic cleavage of a viral polypeptide or one of the revertants. The infection prowas not the mechanism responsible for in-tocol was as described (Sawicki and Sawactivating minus-strand RNA synthesis. icki, 1985) . Actinomycin D (usually at 2 pg/ Recent studies (Sawicki and Sawicki, 1985) (Sawicki and Sawicki, 1980 ) that double-stranded core (RF RNA) of the the number of minus-strand templates dereplicative intermediates and for the determines the rate of plus-strand RNA syn-termination of minus-strand RNA synthethesis.

sis by hybridization of heat denatured RF RNA with unlabeled, purified 49 S plus-

AND METHODS strand RNA were as described (Sawicki et al., 1981b) . Cells and viruses. Chicken embryo fibro-Preparation of labeled SIN polypeptides.

blast (CEF) cells were prepared from lo-The infection, labeling, and preparation of day-old embryos from the eggs of leukosiscell extracts for polyacrylamide gel elecfree (SPF-COFAL/Marek-negative) flocks trophoresis were as described (Sawicki and (Spafas, Roanoke, Ill.) and were grown as Sawicki, 1985) . The samples were run on described previously (Sawicki et al, 1981b) . 8-12s linear gradient acrylamide gels in

The heat-resistant strain of Sindbis vi-Laemmli buffer (Laemmli, 1970) and were rus (SIN HR) and the RNA-negative temprocessed for fluorography. Scanning of the perature-sensitive mutants of the A comfluorograms was with a Shimadzu CS-930 plementation group of SIN HR have been dual wavelength scanner (Shimadzu Cordescribed previously (Burge and Pfefferporation, Japan). korn, 1966; Strauss et al., 1976; Sawicki and Sawicki, 1985) . The parental SIN HR used in these studies was what we call the Ohio RESULTS strain (Sawicki and Sawicki, 1985) . The isolation of revertant viruses of ts24, tsl?, We have already reported (Sawicki and and ts133 of the A complementation group Sawicki, 1985) that revertants of ts24, ts17, and their partial characterization has also and ts133 lost their RNA-negative phenotype and the temperature-sensitive defect been reported previously (Sawicki and - ._. ~~. -. in cleavage of the viral ns230 polyprotein. Sawicki, 1985) . The second stock of ts24Rl

The revertant of ts133 retained the tsdefect in 26 S mRNA synthesis. We wanted to know if these revertants had lost or retained the temperature-sensitive defect in the regulation of minus-strand RNA synthesis. Table 1 presents the results of experiments designed to determine if ts1'7, ts24, or ts133 or their revertants would cause minus-strand RNA to be synthesized late in the infectious cycle in the presence a Cultures were infected and incubated at 30" in the absence of actinomycin D until 10 hr p.i. when one set of cultures was shifted to 40". Cycloheximide and actinomycin D were added to all cultures at 10 hr p.i. and medium containing rH]uridine, cycloheximide, and actinomycin D was added from 11 to 12 hr p.i. The RFs were isolated, denatured, and annealed in the presence of an excess of unlabeled plus strands as described under Materials and Methods. The RNase-resistant incorporation was the cpm recovered in the annealed aliquot after subtraction of the background (Sawicki et al, 1981b) .

b The percentage of the labeled viral RF RNA that was in minus-strand RNA. The values outside the parentheses represent the results from one experiment which determined the ability to synthesize minusstrand RNA by all the viruses listed, the values within the parentheses are averages of two to seven experiments. A value of zero indicates that the RNase-resistant RNA after annealing was equal to or less than the RNase-resistance in either the heated and fast cooled or the self-annealed control. (Sawicki et al, 1981b) . of cycloheximide at 30" or after shift to 40". None of the infected cultures pulse labeled late in infection at 30" synthesized significant amounts of minus-strand RNA in the presence of cycloheximide. Cultures infected with ts17, ts24, or ts133 or with the revertants of ts24, ts24R1, and ts24R2, synthesized minus-strand RNA in the presence of cycloheximide after they were shifted to 40" late in infection. The level of minusstrand RNA synthesis increased 4-to 15 fold after the shift to 40". However, neither the revertant of ts17 nor the revertant of ts133 retained this ts defect in the regulation of minus-strand RNA synthesis; rather, they resembled SIN HR and failed to synthesize minus-strand RNA in the presence of cycloheximide after shift to 40".

We then asked if the kinetics of plusstrand and minus-strand RNA synthesis in cells infected with the revertants of ts24 differed from that in cells infected with ts24 or SIN HR. Infected cultures were maintained at 30" or 40" from the beginning of infection and were pulse labeled with [3H]uridine for 30-min periods. At 30°, cells infected with SIN HR, ts24, and the two independently isolated revertants, ts24Rl and ts24R2, showed similar patterns of RNA synthesis (Fig. 1A ). There was a period when the rate of viral RNA synthesis increased followed by a period of a relatively constant rate of viral RNA synthesis. The increase in the labeling of RF RNA (the double-stranded core of the viral replicative intermediates) paralleled the overall increase in the rate of RNA synthesis. Minus-strand synthesis was readily detectable early in infection but was shut off late in infection. When the infection was at 40", essentially no or very little viral RNA synthesis was detectable in ts24 infected cells (data not shown). However, cultures infected with SIN HR or ts24Rl or ts24R2 showed a similar pattern of viral RNA synthesis at 40" compared to 30". The rate of plus-strand RNA and RF RNA synthesis increased early in infection and changed to a relatively constant rate late in infection (Fig. 1B) . However, whereas the synthesis of minus-strand RNA had ceased by 4-5 hr p.i. in SIN HR infected cells, minus-strand synthesis continued at a high rate in cells infected with ts24Rl and ts24R2. Although minus-strand RNA synthesis continued at a relatively constant rate late in the replication cycle of ts24Rl and ts24R2, the rate of labeling of RF RNA and the rate of viral plus-strand synthesis remained at a constant rate (Fig. 1B) . The failure of the rate of plus-strand transcription to increase might result if the new minus-strands were unstable or did not function as templates in plus-strand synthesis. If they were unstable, minusstrands would not accumulate in infected cells; if they did not function as templates, minus-strands would not be found in replicative intermediates active in viral plusstrand synthesis.

To determine whether or not minus strands continued to accumulate in ts24Rl infected cells, cells were infected with SIN HR or ts24Rl and labeled continuously with rH]uridine beginning at 1 hr p.i. and harvested at various times thereafter ( and actinomycin D was added for 30-min periods beginning at 2 hr p.i. at 30" or at 1.5 hr p.i. at 40", and the cultures were harvested at the end of the pulse period. The total acid-insoluble radioactivity in viral plusstrand RNA, RF RNA, and minus-strand RNA was determined and is expressed per 10' cells. The two stocks of ts24Rl tested are described below. Different experiments are identified by different symbols; in one experiment, virus was allowed to absorb to cells at 30" for 1 hr, and then the cultures were shifted to 40" and incubated at 40" (circles). At 1 hr pi., the inoculum was removed and medium containing ['Hluridine and actinomycin D was added to all cultures. At the times indicated, cultures from each set were harvested. The acid-insoluble radioactivity in total viral RNA (A) and in minus-strand RNA (C) was determined. The incorporation per unit time in total viral RNA (B) or minusstrand RNA (D) was obtained by dividing the incorporation at each time point by the number of minutes of the labeling period. Inset: Cultures in 35-mm petri dishes were infected as above and were pulse labeled with ['Hluridine for 30-min periods between 1.5 and 6 hr p.i. The acid-insoluble radioactivity in total viral RNA is shown for cells infected with SIN HR (0), ts24R1,6 X lo9 PFU/ml (O), or ts24R1,7 X 10' PFU/ ml (m).

60-fold increase in the incorporation of rH]uridine into viral RNA during the '7-hr labeling period. In both SIN HR and ts24Rl infected cells, the incorporation was linear with time although the rate of incorpora-tion of [3H]uridine between 1 and 4 hr p.i. continued to accumulate in ts24Rl infected was slightly greater than between 4 and 8 cells, in contrast to SIN HR infected cells hr p.i. We tested two stocks of ts24Rl: One in which minus strands did not continue to stock (6 X 10' PFU/ml) replicated to the accumulate because their synthesis ceased same extent as SIN HR; another stock (7 at about 4 hr p.i. X lo* PFU/ml) gave a total incorporation We next determined if the minus strands of rH]uridine in infected cells that was that accumulated in ts24Rl infected cells about 40% of that in SIN HR infected cells.

after the rate of viral RNA synthesis had The insert in Fig. 2A shows that in all cul-became constant were utilized as templates tures viral RNA synthesis reached a max-for plus-strand RNA synthesis. The 49 S imum rate at about the same time after minus-strand RNA would be expected to infection. And, if the amount of [3H]uridine be synthesized by a replicative structure that accumulated at any time is divided by with the same relative resistance to ribothe elapsed time of labeling (Fig. 2B) , 1976) . Therefore, the presviral RNA synthesis attained by one stock ence of labeled minus-strand RNA assoof ts24Rl was only half of that attained by ciated with RF1 molecules would reflect SIN HR or the second stock of ts24R1, we nascent or newly completed minus-strand would argue that the leveling in the rate RNA, as well as minus strands that had of viral RNA synthesis was not due to the been synthesized during the labeling period depletion of some cellular component. In and functioned as templates for 49 S plus-SIN HR infected cells, [3H]uridine accu-strand synthesis. The presence of newly mulated into minus-strand RNA until synthesized minus strands in RF111 moleabout 4 hr p.i. (Fig. 2C ). After this time cules would specifically indicate that the there was no further accumulation of label newly synthesized minus strands funcinto minus strands. In contrast to SIN HR, tioned as templates in plus-strand synthecells infected with either stock of ts24Rl sis because RF11 and RF111 are derived continued to accumulate rH]uridine into from RIs, the so-called Rib, which are acminus strands until at least 8 hr p.i. (Fig. tive in 26 S plus-strand RNA synthesis 2C). The level of accumulation of minus (Simmons and Strauss, 1972; strands in cells infected with the two stocks 1978). RF111 is separable from RF1 moleof ts24Rl was proportional to the overall cules by velocity sedimentation. Cells inincorporation in total viral RNA observed fected with ts24Rl were labeled for 2-hr for these stocks. Figure 2D shows the result periods early or late in infection. RFI, RFII, of dividing the amount of [3H]uridine that and RF111 were generated from the viral accumulated into minus strands by the RIs by limited digestion with RNase and elapsed time of labeling. In SIN HR in-were separated by centrifugation on sufected cells the [3H]uridine incorporated crose gradients. As was expected, newly per unit time into minus strands began to synthesized minus strands were present in decrease about 5 hr p.i., whereas in ts24Rl viral RFs isolated from infected cells early infected cells the [3H]uridine incorporated after infection (Table 2 ). More significantly, into minus strands per unit time did not all three RFs isolated from infected cells decrease but remained constant between 4 that had been labeled from 6 to 8 hr p.i. and 8 hr p.i. The higher titer stock of ts24Rl also contained newly synthesized minusgave a similar result, but the incorporation strand RNA, although the newly syntheper unit time in minus-strand RNA was sized minus strands appeared to accumu-2.5 times higher than that obtained for the late preferentially in RF1 molecules. lower titer stock (data not shown). We Therefore, minus strands synthesized in conclude from this data that minus strands excessive numbers in ts24Rl infected cells continued to accumulate in viral RIs, and at least some of these minus strands entered transcription complexes and served as templates for 26 S mRNA, and presumably also 49 S plus-strand RNA, synthesis.

Because minus strands were utilized as templates for plus-strand synthesis after the rate of plus-strand synthesis had become constant, we wanted to know whether a transcription complex needed to be newly formed in order for newly synthesized minus strands to be utilized as templates or whether newly synthesized minus strands would replace old minus strands as templates for transcription.

Late in infection viral RNA synthesis continued in the absence of protein synthesis. The rate of viral RNA synthesis became constant at about 4 hr p.i. and the addition of cycloheximide at 4 hr p.i. resulted in the same rate of viral RNA synthesis as in the absence of cycloheximide (data not shown). In ts24Rl infected cells the minus strands synthesized in the absence of protein synthesis between 4 and 6 hr p.i. did not accumulate as singlestranded 49 S molecules: no labeled minus strands could be detected in the 49 S region of a sucrose gradient (data not shown). However, labeled minus strands that were synthesized in the presence of cycloheximide were found in RF cores of the viral ' CEF cells in loo-mm petri dishes were infected with ts24Rl at 40" in the presence of actinomycin D. Cultures were labeled for the times indicated with medium containing ['Hluridine (200 jXi/ml, 5 ml/ culture) and actinomycin D. The viral RFs, the RNase-A resistant core of the RIs, were obtained (Sawicki et al, 1981b) , and the presence of labeled minus-strand RNA was determined by hybridization.

From 40 to 60% of the total labeled RF RNA annealed to an excess of unlabeled virion plus strands. (Fig. 3) . Labeled minus strands were detected in RF1 and RF111 as early as 30 min after the addition of rH]uridine to ts24Rl infected cells that had been either treated or not treated with cycloheximide at 4 hr p.i., and the accumulation of rH]uridine into minus strands was proportional to the length of the labeling period. More [3H]uridine-labeled minus strands accumulated in RF1 than in RF111 in both cycloheximide and untreated cultures. The rate of accumulation of labeled minus-strand RNA in RF1 and RF111 cores in the presence of cycloheximide was reduced by 40 and 17%, respectively, compared to that detected in the absence of cycloheximide.

Therefore, the accumulation of newly synthesized minus strands was affected to a slight degree by the presence of the drug. These results are in marked contrast to cells infected with SIN HR: almost no [3H]uridine was incorporated into minus strands after 4 hr p.i. at 40" during a 2-hr labeling period either in the absence or presence of cycloheximide. Therefore, in ts24Rl infected cells newly made minus strands replaced at 40" previously synthesized minus strands. This suggested that old minus strands accumulated as RF1 structures and was reminiscent of the turnover of poliovirus minusstrand RNA which accumulate as RF structures (Baltimore, 1968) . The rate of flow of newly synthesized minus strands into poliovirus RIs was similar to that observed for alphavirus minus strands early in infection (Sawicki and Sawicki, 1980) . Figure 4 shows Fig. 3 ) and harvested at 30,60,90, and 120 min later or beginning at 6 hr pi. and harvested after 120 min that was in minus-strand RNA. SFV infected cells were pulse labeled for 5,10, 15,20, and 30 min with PH]uridine beginning at either 2 hr pi. (0) or at 3 hr p.i. (0) (Sawicki and Sawicki, 1980) . The dashed line is the average value (53% ,range 4064%)obtained from 13 determinations of the percentage of label in RF RNA that was in minus-strand RNA when SIN HR infected cells were labeled with [BH]uridine from 1 to 5 hr p.i. strands. In contrast to SIN HR infected cells, when cells infected with ts24Rl were labeled with rH]uridine at 4 hr p.i. or later, labeled minus strands accumulated in RF RNA. However, instead of taking only 10 min, it took 60-90 min before 35% of the label in RF RNA was in minus strands. Therefore, the minus strands synthesized late in ts24Rl infected cells may enter a pool of previously synthesized minus strands before being utilized as templates for plus-strand synthesis or only some of the newly synthesized minus strands are utilized as templates for plus-strand synthesis. Thus, the ts defect in ts24 affecting minus-strand synthesis may reside in a function normally responsible for the stable association of alphavirus minus strands with an active replication complex.

Both ts24Rl and ts24R2 replicated as efficiently at 40" as at 30" and continued to synthesize minus strands at 40" in the absence of protein synthesis if viral RNA and protein synthesis were allowed to start first before adding cycloheximide.

Therefore, using these revertants, we were able to ask directly if there was a pool of viral proteins present early in the alphavirus replication cycle that would assemble into transcription complexes with newly synthesized minus-strand templates. The results of an experiment designed to answer this question is shown in Fig. 5 . The addition of cycloheximide to cells infected with ts24Rl at half-hour intervals beginning at 1.5 hr p.i. resulted in the failure to attain the level of viral RNA synthesis attained by the untreated infected cells. Under the conditions described in Fig. 5 petri dishes were treated with cycloheximide (100 pg/ ml) at various times after infection at 40". Cultures were pulse labeled with PHjuridine for 30-min periods between 1 and 5 hr p.i. in the absence of cycloheximide (filled symbol) or were pulse labeled for 30-min periods in the presence of cycloheximide (empty symbols) beginning at the time of addition of the drug (indicated in the figure) and at two other intervals during the next 2 hr. Acid-insoluble radioactivity in each culture is shown.

is that there is a 15-to 30-min lag between when minus strands associate with a viral polymerase and when they become active as templates for plus-strand synthesis. This was the same result obtained with Semliki Forest virus (Sawicki and Sawicki, 1980) and SIN HR (Sawicki et aL, 1981a) which do not continue to synthesize minus strands in the presence of cycloheximide. Therefore, both the continued synthesis of minus strands and of viral proteins are required to increase the rate of viral RNA synthesis early in infection. Figure 6 presents the results of an analysis of the ability of cells infected with SIN HR or ts24Rl to synthesize viral nonstructural proteins at early and late times. The synthesis of viral proteins at 40" was clearly apparent, in the absence of hypertonic treatment, between 1.5 and 2 hr p.i., reaching maximal rates of synthesis at 4.5 to 5 hr p.i. In cells infected with both SIN HR or ts24R1, the synthesis of viral nonstructural proteins continued at a high rate late in infection. Furthermore, cells in-fected with ts24Rl synthesized greater amounts of the viral nonstructural proteins than did cells infected with SIN HR. The increased rate of synthesis of nonstructural proteins was related to a high level of 49 S RNA synthesis in the ts24Rl infected cells. However, in cells infected with SIN HR and in cells infected with ts24Rl the maximum rate of viral RNA synthesis was about the same and occurred at the same time (insert, Fig. 2A ). The large molecular weight precursors seen in the ts24Rl infected cells were also present, but in reduced amounts, in SIN HR infected cells. A similar increased production of the viral nonstructural proteins was also observed in ts24R2 infected cells (Sawicki and Sawicki, 1985) . DISCUSSION A summary of our analysis of the revertants of the three A complementation group mutants which possessed defects in the regulation of minus-strand synthesis 1 2 3 4 5 6 7 8 9 loll 1213 FIG. 6. Protein synthesis in SIN HR and SIN HR ts24Rl infected cells. CEF cells were infected with 100 PFU/cell of SIN HR or ts24Rl and maintained at 40" in the presence of actinomycin D. The cells were pulse labeled for 30 min with [%]methionine at 1.5 hr (lane 2), at 2.5 hr (lanes 3 and 8), 3.5 hr (lanes 4 and S), 4.5 hr (lanes 5 and lo), 5.5 hr (lanes 6 and ll), and 6.5 hr (lanes 7 and 12). Lanes 1 and 13 are of mock-infected cells. Lanes 2-7 are of SIN HR infected cells and lanes 8-12 are of SIN HR ts24Rl infected cells. The gel shown is an 8 to 12% Laemmli (1970) gel. is presented in Table 3 . All the revertants lost both the RNA-negative phenotype and the ts defect in cleavage of the ns230 polyprotein. The ts defects in synthesis of 26 S mRNA and in regulation of minus-strand synthesis were segregated independently from both the RNA-negative phenotype and the ts defect in ns230 cleavage (this report; Sawicki and Sawicki, 1985) . The ts133 revertant retained the ts defect in 26 S mRNA synthesis but regained the capacity to regulate minus-strand synthesis at 40'; the revertants of ts24 lost the ts defect in 26 S mRNA synthesis but retained the ts defect in the regulation of minusstrand synthesis. Thus, only in the revertants of ts24 was the ts defect in the regulation of minus-strand synthesis independent of all other ts phenotypes. Also, because the revertants of ts24 replicated efficiently at 40", the inability to regulate alphavirus minus-strand synthesis was not of itself conditionally lethal.

Our characterization of minus-strand synthesis in cells infected with SIN HR Ls24 (Sawicki et aL, 1981b) and in cells infected with revertants of ts24 demonstrate that previously formed replication complexes are able to associate with newly synthesized minus strands and synthesize plus strands. Continued synthesis of minus strands must have resulted either from reactivation of inactive replication com-plexes or from the replication complex exchanging a minus-strand template for a plus-strand template. Our results favor the latter interpretation because the rate of plus-strand synthesis does not increase in ts24Rl infected cells even though they accumulated two to four times more minus strands than SIN HR infected cells and continued to synthesize high levels of the nonstructural proteins. Friedman and Grimley (1969) , Scheele and Pfefferkorn (1969) , Ranki and Kaariainen (1970) , Bruton and Kennedy (19'75) , Wengler and Wengler (19'75) , and Sawicki and Sawicki (1980) demonstrated that the alphavirus replication complex, once formed, was stable and continued to synthesize plus strands in the absence of new protein synthesis. However, in ts24Rl or ts24R2 infected cells the continued synthesis and utilization of minus strands late in infection and in the absence of continued protein synthesis demonstrated that newly made minus strands replaced old minus strands. Normally minus strands would not be replaced by newly formed minus strands because minus-strand synthesis ceased late in infection or in the absence of protein synthesis.

Our original model (Sawicki et aL, 1981b) for the mechanism by which alphaviruses temporally regulate minus-strand synthesis involved a regulator whose activity re- --+ -sulted in the selective inactivation of replication complexes engaged in minusstrand synthesis and suggested that t&4 contained a ts mutation in the regulator. Although these studies do not rule out this model, another model which is equally attractive predicts that the preferred template of the viral polymerase is the minus strand. In the preferred template model, the accumulation of minus strands would lead eventually to replication complexes synthesizing exclusively plus strands because new replication complexes stop being formed late in infection or after treatment with cycloheximide. Mutations in ts17, t&4, ts133, and the revertants of ts24 allow for the continuation of minus-strand synthesis because at 40", but not at 30", the replication complex switches templates from a minus strand to a plus strand. This results in minus-strand synthesis with the newly synthesized minus strand being used preferentially as a template for plus-strand synthesis and the old minus strand becoming inactive. It will be important to determine if the mutations responsible for the ts regulation of minus-strand RNA synthesis are located in the viral polymerase (tram-acting) or in a &s-acting sequence of the viral RNA. Why did the overproduction of potentially functional minus strands in cells that were also synthesizing nonstructural proteins not result in increased rates of viral RNA synthesis? The overall rate of RNA synthesis might not be expected to increase if newly synthesized minus strands merely replaced old minus strands in replication complexes. This would occur if the number of replication complexes became fixed early during infection. Factors in addition to the nonstructural proteins and the number of minus strands appear to limit the rate of transcription late in infection. Host factors have been implicated as playing a role in alphavirus replication (Kowal and Stollar, 1981; Scheefers-Borchel et a& 1981; Baric et al, 1933) . However, because we find ( Fig.  2A, insert) that different stocks of a virus isolate caused different levels of viral RNA synthesis in the same host cells, we argue that host factors alone cannot be the determinant which sets the level of tran-scription in alphavirus infected cells. The results of our studies reemphasize the uniqueness of alphavirus replication when compared with that of picornaviruses (Baltimore, 1969) and coronaviruses (Sawicki and Sawicki, 1986) , both of which require continued protein synthesis for plusstrand as well as minus-strand RNA synthesis.

We acknowledge the excellent technical assistance of Mrs. Debora Bruner and the secretarial assistance of Miss Cheryl Zimmerman. This investigation was supported by the United States-Japan Medical Sciences Program through Public Health Service Grant AI-15123 and by Research Career Development Award AI-00510 (D.L.S.) from the National Institutes of Health.

Structure of the poliovirus replicative intermediate RNA

Biochemistry of Viruses

Requirement for host transcription in the replication of Sindbis virus

Semliki Forest virus intracellular RNA: Properties of the multistranded RNA species and kinetics of positive and negative strand synthesis

Complementation between temperature-sensitive mutants of Sindbis virus

Inhibition of arbovirus assembly by cycloheximide

Synthesis of alphavirus RNA

Temperaturesensitive host dependent mutants of Sindbis virus

Cleavage of structural proteins during the assembly of the head of bacteriophage T4

The effect of canavanine on Semliki Forest virus RNA synthesis

Replication of Semliki Forest virus: Polyadenylate in plus strand RNA and polyuridylate in minus strand RNA

Mechanism for control of synthesis of Semliki Forest virus 26s and 42s RNA

Short-lived minus-strand polymerase for Semliki Forest virus

Functional analysis of the A complementation group mutants of Sindbis HR virus

Specific Sindbis virus-coded function for minus strand RNA synthesis

Coronavirus minus strand RNA synthesis and effect of cycloheximide on coronavirus RNA synthesis

A Sindbis virus mutant temperature sensitive in the regulation of minus strand RNA synthesis

Inhibition of interjacent (26s) ribonucleic acid synthesis in cells infected with Sindbis virus

Sindbis virus maturation in cultured mosquito cells is sensitive to actinomycin D. lrirology 110

Replication of Sindbis virus. II. Multiple forms of doublestranded RNA isolated from infected cells

Mutants of Sindbis virus. I. Isolation and partial characterization of 89 new temperaturesensitive mutants

Replication strategies of the single stranded RNA viruses of eukaryotes

Studies on the synthesis of viral RNA-polymerase template complexes in BHK-21 cells infected with Semliki Forest virus