SHORT COMMUNICATION. RNA Splice Site Utilization by Simian Immunodeficiency Viruses Derived from Sooty Mangabey Monkeys

Transcription

1 VIROLOGY 224, (1996) ARTICLE NO SHORT COMMUNICATION RNA Splice Site Utilization by Simian Immunodeficiency Viruses Derived from Sooty Mangabey Monkeys TODD A. REINHART, MICHAEL J. ROGAN, and ASHLEY T. HAASE 1 Department of Microbiology, University of Minnesota Medical School, Minneapolis, Minnesota Received April 23, 1996; accepted July 23, 1996 Alternative splicing of the full-length, primary transcript into numerous subgenomic mrnas is one way that lentiviral gene expression is regulated. Because the behaviors of different viral isolates might reflect in part differences in splicing, we investigated the patterns of splice site utilization by simian immunodeficiency viruses (SIVs) originally isolated from sooty mangabey monkeys (Cercocebus atys). We used reverse transcription polymerase chain reaction (RT PCR), molecular cloning, and DNA sequencing approaches to characterize SIVdeltaB670, a pathogenic and neurovirulent isolate, and SIVsmmH4, a related molecular clone. The majority of randomly selected SIVdeltaB670 and SIVsmmH4 partial cdnas contained tat, rev, nef, and long terminal repeat (LTR) intron splice donor and acceptor sites positioned as expected based on the proviral sequence of SIVsmmH4. Nearly all (87%) of the partial cdnas analyzed contained a spliced LTR intron. A greater number of partial cdnas derived from SIVdeltaB670 infected cells contained putative alternatively spliced introns in comparison to SIVsmmH4, including two previously undocumented splice junctions involving the LTR intron splice donor. These data provide the first comprehensive analysis of splice site utilization by an isolate of SIV in comparison to a related molecular clone and the first characterization of SIVsmm splice site utilization Academic Press, Inc. Infection by lentiviruses can lead to a spectrum of viral and models of HIV-1 latency (17, 18), and although some replication states ranging from latent to fully productive in vivo studies have shown a similar shift (18, 19) other infections, both in the host organism and in individual studies have not supported this conclusion (20, 21). The cells (1 4). The regulation of lentiviral gene expression patterns of splicing exhibited by HIV-1 and SIVmac are is dependent on many host cell and viral factors (5) and complex and characterizations of these patterns have once a provirus is transcriptionally active, alternative indicated that more than 20 different mrnas can be pro- splicing of viral RNA is one mechanism of regulation. duced from the primary transcript (9, 10). The human immunodeficiency viruses (HIV-1 and HIV- Such studies have thus far not included SIVs (SIVsmm) 2) and simian immunodeficiency viruses (SIV) encode a derived from sooty mangabey monkeys (Cercocebus single full-length transcript that is processed by the cellular atys), which have been molecularly cloned and se- splicing machinery and associated cellular and viral quenced and are highly related to SIVmac and HIV-2 factors to give rise to intermediate and short size classes (22 24). Included in the SIVsmm group of viruses are of viral mrnas (6 15). Included among the short class both pathogenic and nonpathogenic molecular clones of viral transcripts are the tat and rev mrnas, which are and pathogenic isolates, such as the acutely lethal essential for viral replication, and nef, which is required SIVsmmPBj (25). We have used reverse transcription for infection and disease induction by SIVmac in adult polymerase chain reaction (RT PCR), cloning and se- rhesus macaques (16). The regulation of gene expres- quence analyses to compare the patterns of splice site sion through alternative splicing of the primary transcript utilization by the uncloned, pathogenic isolate of is not completely understood, although the contributions SIVsmm, SIVdeltaB670 (26, 27), to the minimally pathogenic of Rev to increased accumulation of singly spliced and SIVsmmH4 molecular clone for which the entire unspliced viral RNAs are well characterized (for a review proviral sequence has been determined (24). see Ref. 5). Temporal shifts in the pattern of viral RNA As a first step toward characterizing splice site utiliza- splicing from predominantly short (tat, rev, and nef ) tion by SIVdeltaB670, we PCR amplified, subcloned, and mrnas to intermediate (env, vif, vpr, and vpu) and fulllength DNA sequenced the region of proviral DNA surrounding mrnas have been described in in vitro infections the major subgenomic splice donor. Determination of these sequences was necessary to eventually define 1 To whom correspondence and reprint requests should be addressed. exon/exon junctions, particularly because there are no Fax: (612) published seqence data for full-length viruses contained /96 $18.00 Copyright 1996 by Academic Press, Inc. All rights of reproduction in any form reserved. 338

2 SHORT COMMUNICATION 339 FIG. 1. Amplification strategy and sequence characterization of SIVdeltaB670 viral DNA sequences surrounding the major subgenomic splice donor. (A) Shown is a schematic of the SIV genome and spliced nef, rev, and tat mrnas. Primers used for amplification of SIV viral DNA were a1597 [5 -GAAGACCCTGGTCTGTTAGGA-3, positions of SIVsmmH4 (24)] and a1601 [5 -CTGTTCCTGTTTCCACCACTA-3, positions ]. Primers used for amplification of reverse-transcribed SIV RNA were z3086, z3087, PCRAR, PCR25, and PCRTAT5. (B) Sequences contained in four representative PCR clones generated with primer pair a1597 and a1601 using DNA isolated from CEM1174 cells 7 days after infection with SIVdeltaB670 are shown in comparison to the same sequences from SIVsmmH4, SIVsmmPBj14, and SIVsmm9. Computer alignments of DNA sequences were performed using the Molecular Biology Information Resource (MBIR; Baylor College of Medicine, Houston, TX) software package. The relative frequencies of the 24 SIVdeltaB670 clones are shown on the right and putative intron sequences are boxed. Dots represent sequence identity with the SIVsmmH4 reference sequence and dashes represent sequence gaps. Nucleotide positions follow that for SIVsmmH4 (24). in the SIVdeltaB670 isolate (26, 27). PCR was performed but one between SIVdeltaB670 and SIVsmmH4 in this on genomic DNA purified soon (7 days) after infection region were located in the putative intron. In this 130- of CEM1174 cells (28) with the SIVdeltaB670 isolate to nucleotide region a 2-nucleotide insertion (position 994) reduce the possibility of selecting for variants present was unique to SIVdeltaB670, whereas 2 nucleotide sub- among the SIVdeltaB670 quasispecies. SIV-specific stitutions (positions 911 and 997) were present in all primers (Fig. 1A) a1597 and a1601 were used to PCR SIVsmm sequences except SIVsmmH4. These results in- amplify a 600-bp product that included the major subgen- dicate that the proviral DNA sequences surrounding the omic splice donor and gel purified products were subcloned SIVdeltaB670 major subgenomic splice donor are highly and sequenced (29). As shown in Fig. 1B, the 130 conserved and that the sequences comprising the puta- nucleotides surrounding the putative major subgenomic tive major subgenomic splice donor are identical to se- splice donor were highly related to the same sequences quenced SIVsmm proviruses. in SIVsmmH4, SIVsmm9, and SIVsmPBj14. The SIVdeltaB670 The generation of partial cdnas expected to contain clones were placed into four groups, repre- splice junctions specific for tat, rev, and nef mrnas of sented by clones pb6.gag.2, pb6.gag.3, pb6.gag.6, and SIVdeltaB670 and SIVsmmH4 was accomplished using pb6.gag.12, based on the presence or absence of either nested RT-PCR with primer pairs that selectively amplified 2 nucleotide substitutions (positions 1007 and 1025) or a across junctions between adjoined exons (Fig. 1A). single nucleotide insertion (position 1029). All differences RNA templates were purified either from CEM1174 cells

3 340 SHORT COMMUNICATION

4 SHORT COMMUNICATION 341 infected with an aliquot of SIVdeltaB670 or from the intron leads to increased translation of an indicator gene chronically infected H9/smmsmH-4 cell line (24), both in vitro (33) and its removal from a large proportion of obtained directly from the AIDS Reference and Reagent SIVsmm mrnas suggests that these viruses utilize this Repository. The primers for first round amplification were as a mechanism to regulate viral gene expression. SIVz3086 and z3087 and the primers for second-round ampli- deltab670 clones N5 and N6 represent newly characterized fications were PCRAR, PCR25, and PCRTAT5, which SIV splice junctions generated by either the removal have been described (10). The second round PCRs with of a previously undescribed LTR intron or the removal of primer pair PCRAR/PCRTAT5 were performed to increase a single 8.2-kb intron bounded by the LTR intron splice the sensitivity of generating partial cdnas con- donor (position 575) and the nef splice acceptor (position taining tat splice junctions because tat mrnas have 8806). This is the largest intron that could be removed been shown to be expressed at low levels (10, 12). cdna from a SIV mrna by utilization of documented splice clones were randomly selected and analyzed either by donors and acceptors, and its removal would be ex- DNA sequencing or colony hybridization with putative tat pected to disrupt the secondary structure of the SIV TAR [5 -AGGGTGTCTCCATGTCTGCAACCGGAGGCC-3, positions element (33). A similar splice junction between the LTR , of SIVmacBK28 (31)] or LTR intron splice donor and rev splice acceptor has been intron (5 -TGACCAGGCGGCGACTGCTAGTGCTGGAG- documented for SIVmac251 (34). 3, positions , of SIVmacBK28) splice Among the splice acceptors utilized in nef clones N1 junction probes, or both. These data are summarized in N7, N9, and in rev clones R1 R5 were two alternative Fig. 2. splice acceptors separated by three nucleotides (positions Individual clone types identified as having been derived 8803 and 8806) for the removal of the single intron from tat, rev, and nef mrnas were given designa- in nef transcripts and the second intron in rev transcripts. tions consisting of a letter T, R, or N referring to tat, rev,or A third, alternative splice acceptor in this region was not nef, respectively, and a number. The structures of nef detected, as has been described for SIVmac (10, 34) but clones N1 through N4, rev clones R1 through R4, and tat examination of a larger number of clones might demon- clones T1 and T2 were predicted to contain the exons strate utilization of such a site. The significance of the shown (Fig. 2), based on comparison of previously published utilization of these alternative splice acceptors remains SIVmac splice junctions (6, 10, 12, 13) and the unclear. However, it is a highly conserved feature of SIV SIVsmmH4 proviral sequence (24). At least one clone splicing as all comprehensive analyses of SIV splicing was obtained in each of these 10 classes, except for R2 have demonstrated such alternative splicing. and T2, and in each class, the putative major subgenomic Despite the similarities between the patterns of splice splice donor (Fig. 1B) was utilized. The tat, rev, nef, and site utilization by SIVdeltaB670 and SIVsmmH4, differences LTR intron splice sites utilized by SIVdeltaB670 and were also observed and consisted of either the SIVsmmH4 were also positioned as expected (10, 24) absence of a predominant splice junction for a given and were found in consensus splice sites (Fig. 3A; Ref. transcript class, the presence of additional exons, or the 32). In addition, the splice donors and splice acceptors utilization of putative alternative splice donor and/or acceptor comprising SIVdeltaB670 and SIVsmmH4 LTR intron and sites (Fig. 2). For example, the predominant SIVcomprising tat splice junctions were not only at expected positions, deltab670 nef/rev clone (clone R1) was represented in but were identical in sequence to SIVmac splice junctions only 4 out of 20 clones, whereas the majority of (10). SIVsmmH4 nef/rev clones were represented in a single Interestingly, 96% (120 out of 125) of the clones in rev clone (R3) obtained as 23 out of 34 clones. In addition, these 10 groups had the same intron removed from the SIVdeltaB670 expressed a larger proportion of tat mrnas 5 end of the mrna that is also removed from a subset (T4 T7) that contained additional exons. The extra, small of SIVmac and HIV-2 mrnas (10, 14). Splicing of the LTR exons present in clones T3 T7 (except clone T6) and FIG. 2. Structure of SIVdeltaB670 and SIVsmmH4 partial cdnas containing splice junctions. Adjoined sequences contained in nested RT-PCR products generated with primer pair PCRAR/PCR25 or PCRAR/PCRTAT5 in the second round of amplification are shown as black boxes. The remainder of the sequences expected to be present in full-length cdnas are shown as shaded boxes. Total cellular RNA was isolated from 10 6 cells and 15 mg was reverse-transcribed using Moloney murine leukemia virus RT and oligo(dt) as cdna primer (Perkin Elmer Corp.). PCRs contained 1 mm each primer, 1.5 mm MgCl 2,50mMKCl, 10 mm Tris ph 8.3, and 2.5 U of Taq polymerase per 100-ml reaction and were cycled 30 times at 94 for 30 sec, at 55 for 45 sec, and at 72 for 75 sec. Of the first-round reactions 4 ml was placed in second-round reactions (100 ml). Ethanol-precipitated PCR products were restriction digested with EcoRI and SphI, agarose gel purified, and ligated to EcoRI/SphI-digested puc19 following standard recombinant DNA methodologies (30). Nucleotide positions follow that of SIVsmmH4 (24). The sequences of the primers used in the initial PCR were z3086 (5 -CTGTATTCAGTCGCTCTG-3, positions of SIVsmmH4) and z3087 (5 -TGGTTGGAGGATCTGGTA-3, positions ). The numbers of each type of clone obtained from each of the four separate PCRs are shown to the right. The percentage represented by each clone is given in parentheses., Putative splice donors and acceptors, which when aligned with SIVsmmH4, SIVsmmPBj14, and SIVsmm9, are not predicted to be followed by a GU or preceded by an AG, respectively., Clones containing previously undocumented splice junctions involving the LTR intron splice donor.

5 342 SHORT COMMUNICATION FIG. 3. SIVsmm splice donor and splice acceptor sequences and deduced partial amino acid sequences for Rev and Tat. (A) SIVdeltaB670 and SIVsmmH4 splice junction sequences contained in the clones shown in Fig. 2. Intron sequences for SIVsmmH4 (24) and intron sequences for SIVdeltaB670 are from clone T9 (for the tat splice acceptor) or a subset of SIVdeltaB670 PCR products containing sequences surrounding the nef splice acceptor (data not shown). Lowercase letters represent intron sequences extrapolated from the SIVsmmH4 proviral sequence (24). (B) Alignment of the first coding exon of rev and deduced partial amino acid sequences of SIVdeltaB670 and SIVsmmH4 Rev and Tat proteins encoded by this exon and resulting from utilization of alternative splice acceptors at positions 8803 or Underlined are potential enhancers of splicing of exons, ESEs (35, 36). Dots represent nucleotide or amino acid sequence identity, whereas dashes represent gaps in nucleotide or amino acid sequences. N9 are positioned identically to small exons detected in (32). We therefore compared the sequences of all putative SIVmacBK28 mrnas (10). The small exon in clone T6, alternative splice donors and splice acceptors to available however, ends at the same position in clone T5 and full-length SIVsmm sequences and those sites that, by in a subset of SIVmacBK28 clones (10). Therefore, the extrapolation, did not have predicted GT or AG dinucleo- generation of mrnas containing these exons is a con- tides as their immediate 3 or 5 neighboring nucleotides, served feature of SIV RNA processing, particularly for tat respectively, have been marked with a dagger in Fig. 2. mrnas. Similar exons derived from the 3 end of the pol Of the 20 clones with putative alternative splice sites, 7 gene are present in HIV-1 mrnas, and their positions are (N5, N6, N9, T3 T5) had introns removed that were pre- also conserved when splice site utilization by different dicted to be limited by GT and AG sequences and utilized strains is compared (8, 9). 1 or more splice sites that were observed in legitimate In addition to the partial cdna clones generated by tat, rev, or nef cdnas. An additional 7 clones (N7, T6 T8) RNA splicing at predicted splice donors and acceptors, a utilized at least 1 splice site found in legitimate tat, rev, subset of the cdnas derived from SIVdeltaB670-infected or nef cdnas but contained other putative splice sites not cells (18 out of 71) appeared to be generated by utilization predicted to have GU or AG sequences bordering the of putative alternative splice donors and splice acceptors introns. The last 6 of these clones (N8, N10, N11, R5, and or by the inclusion of additional exons, such as for tat T9) did not contain previously observed splice sites and clones T3 T9. The set of 74 SIVsmmH4 partial cdnas did not have predicted GU or AG sequences bordering contained only 2 of these types of clones, N9 and T3. the introns and we cannot rule out the possibility that Introns removed from RNA Pol II pre-mrnas contain GU they were derived from deleted proviruses rather than and AG dinucleotides at their 3 and 5 limits, respectively alternative splice site utilization.

6 SHORT COMMUNICATION 343 Although the positions of rev splice acceptors were splice acceptors for nef as well as for the third exon of as expected, the nucleotide and deduced amino acid rev (and presumably tat) mrnas (Fig. 2). HIV-1 is unique, sequences of SIVdeltaB670 in the first rev coding exon though, in that multiple splice acceptors are utilized in were unique (Fig. 3B). These sequence differences also generating rev splice junctions (8, 9, 15). By characteriz- are predicted to change the deduced Tat amino acid ing SIVsmm RNA splicing by uncloned SIVdeltaB670 and sequence which is translated from a /1 reading frame molecularly cloned SIVsmmH4, we have demonstrated relative to Rev. In all tat cdnas the predicted initiation that the patterns of splice site utilization by SIVs are codon was the first one following the tat splice acceptor, highly conserved between genetically and pathogenet- whereas in all rev cdnas, there was an AUG codon 5 ically distinct SIVsmm and SIVmac viruses, and in some bases upstream of the predicted initiation codon for both aspects, between these viruses and HIV-1. SIVdeltaB670 and SIVsmmH4, similar to SIVmac RNAs (10, 34). Comparison of these sequences to the same ACKNOWLEDGMENTS regions from SIVsmm9 and SIVsmmPBj14 sequences in- We thank Dr. Gregory Viglianti for the kind gift of primers PCRAR, dicated that SIVdeltaB670 is highly related to character- PCR25, and PCRTAT5 and for a critical reading of the manuscript. ized SIVsmm viruses except for a 6-nucleotide insertion The following reagents were obtained through the AIDS Research and at the 5 end of rev which was unique to SIVdeltaB670. Reference Reagent Program, Division of AIDS, NIAID, NIH: CEM1174 Two sequence elements (GAAGAAGAA and GAA- cell line from Dr. Peter Creswell, H9/smmsm-H4 cell line from Dr. Philip Johnson, and SIVdeltaB670 virus from Dr. Michael Murphey-Corb. We GAAG[A/G][C/T]), were identified that are identical and also thank Dr. Ernest Retzel for his generous help in synthesizing similar, respectively, to short sequences that are en- oligonucleotides and Dr. Jim List and Steve Wietgrefe for helpful discushancers of splicing of exons (ESEs) which function as sions. T.A.R. is a Pediatric AIDS Foundation Scholar. cis-acting, positive regulators of HIV-1 splicing (35, 36). Because the second of these potential ESEs contains 2 REFERENCES nucleotide differences between SIVsmmH4 and SIVdel- 1. Haase, A. T., Stowring, L. S., Narayan, O., Griffin, D., and Price, D., tab670, and the majority (68%) of SIVsmmH4 cdnas gen- Science 195, (1977). erated with PCRAR/PCR25 utilized the splice acceptor 2. Haase, A. T., Nature 322, (1986). just 5 to these sequence elements, it is conceivable 3. Embretson, J., Zupancic, M., Ribas, J. L., Burke, A., Racz, P., Tenner- that these sequences specify efficient splicing at position Racz, K., and Haase, A. T., Nature 362, (1993). 4. Pantaleo, G., Graziosi, C., Demarest, J. F., Butini, L., Montroni, M., 6501 of the SIVsmmH4 primary transcript. Fox, C. H., Orenstein, J. M., Kotler, D. P., and Fauci, A. S., Nature Studies of lentiviral splicing patterns have thus far fo- 362, (1993). cused on molecularly cloned viruses. Despite the clon- 5. Cullen, B., Annu. Rev. Microbiol. 45, (1991). ality of the viruses in studies of HIV-1 and SIVmac splicand 6. Colombini, S., Arya, S. K., Reitz, M. S., Jagodzinski, L., Beaver, B., ing, complex patterns of splice site utilization have been Wong-Staal, F., Proc. Natl. Acad. Sci. USA 86, (1989). observed. We have extended analyses of SIV splicing by 7. Guatelli, J. C., Gingeras, T. R., and Richman, D. D., J. Virol. 64, 4093 determining and comparing the patterns of splice site 4098 (1990). utilization of a SIVsmm isolate and a related molecular 8. Robert-Guroff, M., Popovic, M., Gartner, S., Markham, P., Gallo, clone. The similarities in splicing patterns observed be- R. C., and Reitz, M. S., J. Virol. 64, (1990). tween SIVdeltaB670 and SIVsmmH4 as well as between 9. Schwartz, S., Felber, B. K., Benko, D. M., Fenyo, E.-M., and Pavlakis, G. N., J. Virol. 64, (1990). these SIVs and SIVmac (6, 10, 12, 13, 34) indicate that 10. Viglianti, G. A., Sharma, P. L., and Mullins, J. I., J. Virol. 64, 4207 there are likely strong selective pressures for the mainte (1990). nance of sequence elements that direct appropriate 11. Furtado, M. R., Balachandran, R., Gupta, P., and Wolinsky, S. M., splicing events. The patterns of alternative splice site Virology 185, (1991). utilization by the primate lentiviruses SIVmac, HIV-2, HIV- 12. Park, I.-W., Steen, R., and Li, Y., J. Virol. 65, (1991). 1, and shown here, SIVsmm, are complex and there are 13. Unger, R. E., Stout, M. W., and Luciw, P. A., Virology 182, (1991). aspects of RNA splicing that are conserved between 14. Chatterjee, P., Garzino-Demo, A., Swinney, P., and Arya, S. K., AIDS these viruses and other aspects that distinguish them. Res. Human Retroviruses 9, (1993). Both SIV and HIV-1 show great diversity in the repertoire 15. Purcell, D. F. J., and Martin, M. A., J. Virol. 67, (1993). of splice site utilization including splice sites that define 16. Kestler, H. W., III, Ringler, D. J., Mori, K., Panicali, D. L., Sehgal, variably present small exons, derived from the 3 end of P. K., Daniel, M. D., and Desrosiers, R. C., Cell 65, (1991). pol, and utilized by many viral strains (7, 8, 14), though 17. Kim, S., Byrn, R., Groopman, J., and Baltimore, D., J. Virol. 63, 3708 their functional significance remains unclear. In addition, 3713 (1989). SIV and HIV-1 tat mrnas are rare transcripts and cdnas 18. Pomerantz, R. J., Trono, D., Feinberg, M. B., and Baltimore, D., Cell containing tat splice junctions can only be obtained with 61, (1990). primers that selectively amplify across the tat splice junc- 19. Michael, N. L., Mo, T., Merzouki, A., O Shaughnessy, M., Oster, C., Burke, D. S., Redfield, R. R., Birx, D. L., and Cassol, S. A., J. Virol. tion (Fig. 2; Refs. 8, 10, 12). In contrast, conserved fea- 69, (1995). tures of SIV splicing that distinguish SIV from HIV-1 are 20. Furtado, M. R., Kingsley, L. A., and Wolinsky, S. M., J. Virol. 69, the removal of the LTR intron and utilization of alternative (1995).

7 344 SHORT COMMUNICATION 21. Saksela, K., Stevens, C. E., Rubinstein, P., Taylor, P. E., and Balti- Laboratory Manual, 2nd ed. pp , , and more, D., Annu. Internal Med. 123, (1995) Cold Spring Harbor Laboratory Press, Cold Spring 22. Courgnaud, V., Laure, F., Fultz, P. N., Montagnier, L., Brechot, C., Harbor, NY and Sonigo, P., J. Virol. 66, (1992). 30. Chomczynski, P., and Sacchi, N., Anal. Biochem. 162, Dewhurst, S., Embretson, J. E., Anderson, D. C., Mullins, J. I., and (1987). Fultz, P. N., Nature 345, (1990). 31. Kornfeld, H., Riedel, N., Viglianti, G. A., Hirsch, V., and Mullins, 24. Hirsch, V. M., Olmsted, R. A., Murphey-Corb, M., Purcell, R. H., and J. I., Nature 326, (1987). Johnson, P. R., Nature 339, (1989). 32. Mount, S., Nucleic Acids Res. 10, (1982). 25. Fultz, P. N., and Zack, P. M., Virus Res. 32, (1994). 33. Viglianti, G. A., Rubinstein, E. P., and Graves, K. L., J. Virol. 66, 26. Baskin, G. B., Murphey-Corb, M., Watson, E. A., and Martin, L. N., (1992). Vet. Pathol. 25, (1988). 34. Szotyori, Z., Almond, N., and Kitchin, P., AIDS Res. Hum. Retroviruses 27. Murphey-Corb, M., Martin, L. N., Rangan, S. R. S., Baskin, G. B., 10, (1994). Gormus, B. J., Wolf, R. H., Andes, W. A., West, M., and Montelaro, 35. Amendt, B. A., Si, Z.-H., and Stoltzfus, C. M., Mol. Cell. Biol. 15, R. C., Nature 321, (1986) (1995). 28. Salter, R. D., Howell, D. N., and Cresswell, P., Immunogenetics 21, Aubert, C., and Ozden, S., J. Histochem. Cytochem. 41, (1985). (1993). 29. Sambrook, J., Fritsch, E. F., and Maniatis, T., Molecular Cloning: A 36. Staffa, A., and Cochrane, A., Mol. Cell. Biol. 8, (1995).