Search

Kunjin virus-infected cells were lysed and the cytoplasmic extract was subjected to sedimentation analysis. After centrifugation at 16,000 x g for 10 min about 70% of the original RNA-dependent RNA polymerase (RDRP) was recovered in the pellet; most of this enzymic activity was recovered in the soluble fraction after treatment with NP40 detergent. Membrane fractions were prepared from cytoplasmic extracts by centrifugation in discontinuous density gradients comprising w/w or w/v sucrose solutions, either for 3 h (top-loaded on 4 ml 20-60% sucrose) or for 19 h (centre-loaded in 37 ml 0-60% sucrose). Similar separations of bands of light membranes were obtained in all gradients. Multi-layered heavy membrane bands obtained with w/w sucrose gradients were resolved into two well-separated bands (F4 and F5) using w/v sucrose gradients. Thin-section electron microscopy of embedded membrane fractions, gel analysis of intracellular RNA, and RDRP assays showed that the w/w centre loading method and the w/v top-loading (short spin) method produced similar recoveries and distributions of smooth and rough membranes, intact virus particles and RDRP activity. The distribution of intracellular viral RNA and proteins was coincident with the RDRP, all being located in the F4 and F5 bands which contained the characteristic membrane structures induced during flavivirus infection. Significant advantages of the preferred method (w/v sucrose, top loading and short spin) were its rapidity, good preservation of membranes and RDRP, and the concentrations of RDRP achieved in the small volume fractions collected from a total of 4.5 ml.

In subcellular extracts of Kunjin virus-infected cells prepared by lysis and differential centrifugation, the viral RNA polymerase, RNA and proteins were associated mainly with cytoplasm. When the cytoplasmic extract (500 g supernate) of infected cells labelled for 3 h from 24 h post-infection was further fractionated by rapid centrifugation through a sucrose density gradient, all viral products were located only in dense or "heavy membrane" fractions, which contained three types of virus-induced morphologically distinct membrane structures. These dense fractions were treated with 0.5% NP40 and the soluble material was again centrifuged through a sucrose gradient for analyses as before. Viral RNA polymerase activity was retained and was associated with replicative intermediate RNA and some replicative form RNA in the peak enzyme fractions sedimenting at 20S to 40S. Enrichment of NS3 and of the small nonstructural proteins NS2A and NS2B/NS4A was apparent in these fractions which were well separated from the slow sedimenting structural proteins. No detergent-resistant structures in the "heavy membrane" fractions other than ribosome-like particles were visible. The data show that the RNA polymerase complex cosedimented with virus-induced membrane structures and remained associated with specific nonstructural proteins and replicative intermediate RNA after detergent treatment.

Only three forms of Kunjin virus-specified RNA were isolated from cytoplasm early after the latent period (about 15 hr) viz., 44 S genomic-sized single-stranded RNA, 20 S double-stranded "replicative form" (RF), and 20-28 S partially ribonuclease-resistant (about 70%) "replicative intermediate" (RI). The RF and RI were resolved by electrophoresis in aqueous-agarose gel only following LiCl fractionation. The RI did not enter urea-polyacrylamide gels. After denaturation of untreated or RNase-treated RI and RF, only 44 S RNA was present in electropherograms. RNA polymerase activity at 8 hr postinfection was detected by in vitro assays of cytoplasmic extracts and reached a maximum at 24 hr, the only major labeled product being RF; a trace amount of free 44 S RNA was also produced. These results, and the kinetics of incorporation of [3H]uridine into RI, RF, and 44 S RNA in pulse and pulse-chase experiments, formed the basis of a model in which flavivirus RF functions as a recycling template for semiconservative and (mainly) asymmetric replication, on which only one nascent strand is synthesized per cycle.

RNA-dependent RNA polymerase (RDRP) activity was characterized in a cytoplasmic extract of Kunjin virus-infected Vero cells at 24 hr. The activity was influenced, possibly indirectly, by the length of prior treatment of infected cells with actinomycin D; however, 6 micrograms/ml actinomycin D and 10(-5) M alpha-amanitin in the RDRP assay had no effect. The replication complex was membrane-bound and Mg2+ was essential for RDRP activity. Incorporation was more dependent on exogenous UTP and GTP than ATP or CTP. The specific activity was low, and rate of incorporation of GMP decreased as the period of assay was increased; however, incorporation of label lasted for at least 60 min. RNA products were fractionated by LiCl precipitation, and kinetic studies showed that the sequence of accumulation of label was the same as that observed in vivo, viz., RI----RF----44 S RNA; limited reinitiation was also observed. This sequence of labeling also indicated that the in vitro RDRP activity was due to an enzyme capable of elongation, release, and reinitiation of Kunjin RNA synthesis and not merely end labeling or elongating preexisting RNA molecules. No labeled bands in urea-polyacrylamide gels were observed using extracts from mock-infected cells and hence the three RNA products of assays were readily identified in a single gel. The replication complex was still active after treatment with nonionic detergent, but no labeled 44 S RNA was detected in gels, even in the presence of RNasin in the assay which inhibited some nuclease activity. Antibodies to flavivirus-specific nonstructural proteins were preincubated with infected cell extracts in the presence and absence of detergent but no inhibition of RDRP activity was observed. However, anti-dsRNA plus detergent blocked activity by as much as 78% and label was found only in RF.