INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Mouse mammary tumor virus (MMTV) is transmitted from the milk of infected mothers to the gut of susceptible offspring (33). MMTV infects B cells in the guts of newborn mice, and these cells express the virally encoded superantigen (Sag) at the plasma membrane in association with the major histocompatibility complex (MHC) class II protein (1, 28). Sag is a type II transmembrane protein that is required for efficient transmission of milk-borne MMTV from the gut to the mammary gland (12, 17, 23). The Sag-MHC complex is recognized by entire classes of T cells bearing particular T-cell receptor (TCR)  chains (18, 28). Recognition of Sag by specific T cells leads to cell proliferation and/or release of cytokines, and the released cytokines recruit additional B and T lymphocytes that are infected by MMTV (27). B-cell-deficient mice or mice lacking Sag-reactive T cells cannot be efficiently infected by MMTV (4, 12). Ultimately, viral infection of the mammary gland is necessary to allow MMTV release into the milk. Mice that lack B cells or Sag-specific T cells also are defective in the spread of MMTV within the mammary gland (14). Thus, Sag is required for generation of a reservoir of virally infected B and T lymphocytes that are involved in MMTV transmission and viral spread to the mammary tissues.

The regulation of MMTV sag expression is controversial. Early studies indicated that Sag is translated from a singly spliced mRNA that initiates at the U3/R border of the viral long terminal repeat (LTR) from the predominant U3 promoter (21, 41). This U3 promoter also drives expression of the viral structural genes, gag, pol, and env (for a review, see reference 10) (Fig. 1). Cloning and sequencing revealed that this sag mRNA uses a single splice donor site in the leader region that also is used for the generation of spliced envelope mRNA (32). The splice acceptor site for this sag mRNA is located in the envelope region just upstream of the 3' LTR (21). Multiple start codons are located near the 5' end of this mRNA, and mutagenesis experiments have suggested that the first or second codons can suffice for functional Sag production in cell culture (8). Subsequently, at least three other potential sag mRNAs have been described (Fig. 1). One of these mRNAs uses the same splice donor and acceptor sites as those described earlier, except that this transcript initiates approximately 500 bp upstream of the standard viral RNAs (16). Another transcript initially was identified as a phorbol ester-inducible, cyclosporine-suppressible RNA in a T-cell lymphoma (11, 31, 38). This sag-specific RNA is initiated from an intragenic promoter within the envelope-coding region; the transcript uses a unique splice donor site within the envelope region and the same splice acceptor site as the other two sag RNAs. Most recently, a different sag promoter has been described within the envelope region (2). This promoter was shown to be active in transient transfection experiments with reporter gene constructs. The resulting sag mRNA appears to be unspliced and initiated within 100 bp of the 3' LTR (2).

Using PCR assays, our experiments have shown that all of the spliced sag transcripts are detectable in lymphocytes of BALB/c mice that contain three endogenous MMTVs, Mtv-6, Mtv-8, and Mtv-9 (9, 24). However, only the spliced sag transcript from the envelope promoter was detectable in cells infected in vitro or in vivo with milk-borne C3H MMTV. Similarly, deletion mutants of a C3H MMTV-derived infectious molecular clone (39) suggested that an intragenic envelope promoter and an enhancer in the pol gene were responsible for sag expression (34, 35). To test whether the spliced mRNA from the env promoter is required for C3H MMTV sag expression, we used the C3H-derived infectious molecular clone to construct a mutant with alterations in the splice donor site within the envelope region that is unique to this transcript. Stable transfections of this mutant or a control frameshift mutant with an alteration within the sag coding region that has been shown to abolish Sag function (13) produced virus-expressing cell lines. Both the splice donor and frameshift mutants induced some mammary tumors with long latency and sporadic T-cell deletion after direct injection of transfected cells. However, neither the frameshift nor the splice donor mutant was transmissible to susceptible third-litter progeny through the milk-borne route. These experiments show that the spliced mRNA from the envelope region is necessary and sufficient for efficient C3H MMTV milk-borne transmission.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Mice. BALB/cJ mice were purchased from Jackson Laboratories (Bar Harbor, Maine). All animals were bred and maintained at the University of Texas at Austin Animal Resources Center. The animals were tested at periodic intervals and were free of common bacterial and viral pathogens, including mouse hepatitis virus. Four- to 5-week-old weanlings were injected with a total of 2 × 10⁷ XC cells expressing MMTV proviral constructs and divided among five sites, four subcutaneous injections near the mammary glands proximal to each leg and one intraperitoneal injection as described by Shackleford and Varmus (39). All injected females were bred continuously to stimulate lactogenic hormones and MMTV production. Animals were palpated weekly for the appearance of mammary tumors.

Plasmid construction. Construction of the sag frameshift mutation at the ClaI site in the 3' LTR of the HYB MTV provirus has been described previously (13). The splice donor sequence in the C3H MMTV envelope gene (nucleotide [nt] 7339) was mutated using a PCR-based method. The choice of mutations was based on the conserved sequences encompassing the splice donor site. Since the splice donor overlaps the coding sequences in the env gene, only the third position in each codon was changed, except in one case, where a conservative valine-to-leucine change was made (Fig. 1). The method for mutagenesis of the splice donor site was essentially that described by Hoguchi (19). The first PCR was performed using the sense oligonucleotide C3Hpol6361(+) (5' ATC TCA CGT CAC GGG GAT CCC TTA CAA TCC 3') and the mutant antisense oligonucleotide C3HenvSD7352() (5' GGA GAA AAt gag Agt CCc TGG TCA GGG AAG GCG CAA GGC AAC 3') (with mutant sequences lowercased and boldfaced). (The primer numbering system corresponds to that for the complete BR6 provirus [32]). The second PCR was performed using the mutant sense oligonucleotide C3HenvSD7326(+) (5' CCT GAC CAg GGa cTc tca TTT TCT CCA AAA GGG GCC CTT GGG 3') and the antisense oligonucleotide C3Henv7519() (5' CTC TAT CAT TGG GAT CCT TAG GAG AAT TTT CCC 3'). The final 1.3-kb product was gel purified, digested with BamHI, and ligated to a 15-kb BamHI fragment from pHYB MTV, an infectious molecular clone of MMTV (39), to generate pHYB SD. The sequence of the clone was confirmed by automated fluorescent DNA sequencing.

Transfections. Stable cell lines of rat XC fibroblasts were generated by using 10 µl of DMRIE-C (GIBCO BRL, Gaithersburg, Md.), 5 µg of wild-type or mutant CsCl-purified plasmid DNA, and 0.05 µg of DNA expressing the hygromycin expression cassette pTR174 (36). Transfections were performed in triplicate using six-well plates, and the cells were selected in Dulbecco's modified Eagle's medium containing 7.5% fetal bovine serum (HyClone Laboratories, Inc. Logan, Utah), 50 µg of streptomycin/ml, 100 U of penicillin/ml, 200 mM glutamine, and 0.5 mg of hygromycin (GIBCO BRL)/ml until discrete colonies were established. The colonies in the three wells were pooled and expanded. The pooled clones were induced for MMTV expression using 10⁶ M dexamethasone (DEX) (Sigma Chemical Laboratories, St. Louis, Mo.), and a portion of the pooled population was used to make RNA, DNA, and proteins to assess MMTV expression.

Immunoprecipitation and Western blot analysis. Whole-cell protein extracts were prepared from DEX-induced XC stable cell lines essentially as described previously (26). The precleared lysates were incubated overnight with 1 µl of anti-p27^gag MMTV polyclonal antiserum (National Cancer Institutes/Biological Carcinogenesis Branch [NCI/BCB] Repository, National Institutes of Health) at 4°C. The immune complexes were precipitated using 40 µl of anti-goat immunoglobulin G (IgG)-agarose (Sigma) and separated on SDS-polyacrylamide gels (10 to 12% polyacrylamide). Separated immune complexes were transferred to Optitran nitrocellulose membranes (Schleicher and Schuell, Keene, N.H.) overnight at 4°C, blocked with 10% dried milk in a high-salt phosphate-buffered saline solution (0.2 M NaCl, 10 mM Na₂HPO₄, 1.7 mM KH₂PO₄, and 2.6 mM KCl) containing 0.1% Tween 20, and incubated with anti-MMTV goat polyclonal antiserum (1:500 dilution) or monoclonal murine anti-SU (1:10 dilution of Blue5; kindly provided by T. Golovkina, Jackson Labs). MMTV-specific proteins were detected by using biotinylated anti-goat serum (Sigma) and horseradish peroxidase-conjugated streptavidin (Calbiochem, Cambridge, Mass.) or horseradish peroxidase-conjugated anti-mouse serum (Amersham Pharmacia Biotech Limited, Little Chalfont, United Kingdom). Proteins were visualized using the enhanced chemiluminescence kit as described by the manufacturer (Amersham).

Slot blot analysis. RNA was extracted from cultured cells or mouse tissues using the TRI Reagent as recommended by the manufacturer (Molecular Research Center, Cincinnati, Ohio). Dilutions of RNA were made and denatured in a slot blot RNA denaturation cocktail (0.69× SSC [1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate], 8.9% formaldehyde, and 69% formamide) and blotted onto a Zeta-Probe nylon membrane (Bio-Rad) using the Bio-Dot SF microfiltration apparatus from Bio-Rad. The blotted filter was rinsed in 2× SSC and UV cross-linked twice at 1,200 µJ using a Stratalinker (Stratagene). The filter was hybridized to a 1.4-kb probe detecting all MMTV mRNAs (a PstI fragment from the 3' C3H LTR), while the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) probe was a PCR fragment containing sequences between nt 427 and 983 of the murine GAPDH gene (37, 43). Probes had specific activities of ca. 10⁸ cpm/µg of DNA. The blots were washed and subjected to autoradiography.

RT-PCR. Twenty to forty micrograms of total cellular or tissue RNA was DNase treated using 3 to 5 U of amplification grade DNase I (GIBCO BRL) at 37°C for 1 h. DNase I was heat inactivated after the addition of EDTA to a final concentration of 2.5 mM and incubation at 70°C for 10 min. Five to ten micrograms of DNase-treated RNA was used in the reverse transcription (RT) reaction with poly(dT₁₇) primer. The primer was mixed with the RNA and boiled for 5 min, followed by quick cooling on ice for 5 min. The denatured RNA was reverse transcribed in a 50-µl reaction volume using 2 µl of Moloney murine leukemia virus (M-MLV) reverse transcriptase (GIBCO BRL) for 1 h at 37°C. Five microliters of cDNA was used in reactions containing 45 µl of SuperMix (GIBCO BRL) and 100 ng (~15 pmol) of each of the appropriate primers. Primer C3H LTR 420(), 5' GAT TCA TTT CTT AAC ATA GTA AC 3', was designed to specifically discriminate C3H-specific sequences from the endogenous MMTVs, Mtv-6, Mtv-8, and Mtv-9. The C3H LTR 420() primer was used in combination with various sense primers to amplify (i) all MMTV-specific mRNAs (155+, 5' GGC ATA GCT CTG CTT TGC 3'), (ii) sag mRNAs from the U3 promoters (230+, 5' GTG AAT TCC ATC ACA AGA GCG GAA CGG AC 3') (43), and (iii) sag mRNA from the 5' intragenic env promoter (7255+, 5' ATC GCC TTT AAG AAG GAC GCC TTC TTC T 3'). Twenty microliters of each PCR product was analyzed by electrophoresis on 2% agarose gels and stained with ethidium bromide prior to photography.

Antibodies and flow cytometry analysis. Injected mice and their progeny were bled from the retro-orbital sinus at appropriate intervals. Peripheral blood lymphocytes were purified using Histopaque (Sigma) and subjected to dual staining for CD4 and V14 as described by Wrona et al. (42). Antibodies were obtained from PharMingen (San Diego, Calif.).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Construction of a splice donor mutant. At least four different RNAs have been described for the production of MMTV Sag protein (2, 11, 16, 41) (Fig. 1A). Our previous work indicated that the spliced RNA initiated within the envelope region was the only C3H sag-specific RNA detectable in cells infected by C3H MMTV (44). However, Sag is expressed in extremely small amounts (29), and it is possible that small amounts of sag mRNA from alternative promoters would be sufficient to allow Sag function in vivo. To determine if the spliced RNA that initiated in the envelope region was solely responsible for C3H MMTV Sag expression, we prepared a mutant of the infectious MMTV clone, HYB MTV (39). This mutant disrupted the splice donor site that is unique to the spliced sag mRNA from the envelope promoter (Fig. 1B). If the spliced RNA from the envelope region is necessary for Sag production from C3H MMTV, mutation of this splice donor site should eliminate C3H MMTV Sag function as measured by deletion of reactive T cells and milk-borne transmission of MMTV.

View larger version (24K): [in this window] [in a new window]   FIG. 1.   Diagram of the MMTV sag-specific mRNAs and the location of the splice donor mutation. (A) Schematic representation of the MMTV genome and reported MMTV sag mRNAs. The boxes on the MMTV proviral genome show the positions of the indicated open reading frames and the LTRs. The reported sag transcripts are indicated below the provirus; introns are indicated by dotted lines, and exons are represented by dashed lines with arrows. The figure also shows the locations of the reported MMTV promoters and splice sites used for sag gene expression, as well as primers used to differentiate among sag mRNAs. SD, splice donor; SA, splice acceptor; CFS, ClaI frameshift. (B) Comparison of the mutant splice donor (mSD) site in env with the sequence of wild-type (WT) C3H MMTV. The boxed area highlights the mutated region of the splice donor. The underlined letters reflect the canonical GT of the splice donor, while the bold amino acids reflect the conservative change made by the mutation.

View larger version (41K): [in this window] [in a new window]   FIG. 2.   The splice donor mutation abrogates expression of sag mRNA from the 5' intragenic env promoter. RT-PCR analysis was performed using RNA extracted from XC cells transfected with the wild-type or SD mutant proviruses. (A) sag mRNAs expressed from the 5' env promoter; (B) sag mRNAs expressed from U3 promoters; (C) all MMTV mRNAs; (D) GAPDH mRNA as a control for RNA and cDNA integrity. Lane 1, XC control RNA from untransfected XC cells; lane 2, RNA extracted from a pool of wild-type HYB MTV transfectants; lane 3, RNA extracted from a pool of HYB SD transfectants.

View larger version (55K): [in this window] [in a new window]   FIG. 3.   Total MMTV RNA expression from different XC cell transfectants is equivalent. RNAs extracted from pools of transfected XC cells were diluted and subjected to slot blot analysis. Blots were hybridized to 32P-labeled probes for the MMTV LTR (A) or GAPDH (B). All RNAs were treated with DNase I prior to blotting. Relative expression of MMTV RNAs in mutant-transfected cells was calculated by using values that were obtained by phosphorimager analysis using hybridization of 0.2 and 2.0 µg of RNA to the LTR probe and then normalized for RNA loading by using the values obtained with the GAPDH probe.

View larger version (25K): [in this window] [in a new window]   FIG. 4.   HYB MTV- and mutant-transfected cell lines express MMTV-specific Gag and Env proteins. (A) Expression of MMTV Gag proteins using immunoprecipitation and Western blot analysis. A polyclonal antibody against MMTV CA protein was used for immunoprecipitation, followed by detection with anti-MMTV polyclonal antibody. There is cross-reactivity of the antibodies directed against MMTV-specific proteins with XC cell proteins. Some minor bands were observed inconsistently in different extracts, presumably due to some protein degradation or processing, but major MMTV protein levels were similar in mutant and wild-type transfectants. The positions of MMTV-specific proteins and precursors are shown on the right, and molecular weight markers are shown on the left. (B) Expression of MMTV Env protein using Western blot analysis with a monoclonal gp52 (SU)-specific antibody.

In vivo infection with sag mutants. Our data indicated that a splice donor mutation in the envelope region eliminated the ability of an infectious MMTV provirus to produce the spliced sag-specific RNA from the envelope promoter. However, this same mutation did not eliminate the ability of the virus to produce MMTV RNA and Gag or Env proteins. Therefore, wild-type HYB MTV transfectants, sag frameshift transfectants, and three different pools of HYB SD transfectants were assessed for the ability to transmit MMTV in vivo after injection into weanling BALB/c mice.

View larger version (24K): [in this window] [in a new window]   FIG. 5.   Sporadic deletion of V14+ CD4+ T cells in mutant-injected BALB/c mice. Deletion of cognate T cells determined by fluorescence-activated cell sorter (FACS) analysis is shown at various times after injection of XC cell transfectants. Each time point represents FACS analysis of the peripheral blood lymphocytes from one to three mice. Standard deviations from the means are indicated for each time point measured. Heavy arrows indicate the average latency of mammary tumors induced by the wild-type or mutant viruses.

View this table: [in this window] [in a new window]   TABLE 1.   V14+ T-cell deletion and mammary tumor development in injected mice

View larger version (34K): [in this window] [in a new window]   FIG. 6.   Spread of MMTV infection in mice injected with XC cells expressing the splice donor mutant virus. (A) Detection of MMTV-specific RNA in the various organs of injected mice using RT-PCR. (Panel I) RT-PCR with primers specific for C3H MMTV LTR sequences; (panel II) RT-PCR with primers specific for the splice donor mutation; (panel III) RT-PCR specific for GAPDH RNA. Tissues analyzed: LN, lymph node; SP, spleen; MG, mammary gland; MT, mammary tumor; SG, salivary gland. In some cases, results from two individual mice are shown. Spontaneous mammary tumors in our BALB/c colony are rare (<1%). (B) Expression of sag mRNAs from the intragenic env promoter in the mammary tumors from wild-type (HYB MTV) (lanes 1 and 2)- and SD mutant (lanes 3 and 4)-injected mice. Results from two individual mice are shown. Because a large number of cycles was used for PCR assays to detect sag mRNA, these results are not quantitative. In addition, use of the primer pair for the splice donor mutation appears to be more sensitive for PCR assays than the C3H LTR primer pair used in panel A. Expression of GAPDH was monitored by RT-PCR as a control for RNA and cDNA integrity. RT-PCR products were separated on 2% agarose gels and visualized by ethidium bromide staining prior to photography.

Lack of milk-borne transmission by sag mutants. Because Sag function appears to be required for efficient transmission of milk-borne MMTV (13), we tested the progeny of HYB MTV-, HYB SD-, and sag frameshift mutant-injected animals for evidence of MMTV infection. As anticipated, offspring of HYB MTV-injected mice showed approximately 20% deletion of C3H Sag-specific T cells at 2.5 months of age (Fig. 7), and this deletion increased with age. Furthermore, offspring of HYB MTV-injected animals (three of three females) developed mammary tumors with an average latency of 9.5 months (Table 2). In contrast, third-litter progeny mice injected with HYB SD or sag frameshift mutants did not delete CD4⁺ V14⁺ T cells at any age tested (up to 1 year). No mammary tumors have developed in the offspring of animals injected with HYB SD mutants (0 of 5 females; Table 2). Together these results indicate that elimination of the spliced sag-specific RNA from the envelope promoter is sufficient to abolish C3H MMTV Sag function.

View larger version (20K): [in this window] [in a new window]   FIG. 7.   The splice donor mutant viruses are defective in milk-borne transmission of virus from mothers to offspring. The kinetics of V14+ CD4+ T-cell deletion is shown for the third litters of the injected mice. Each time point represents fluorescence-activated cell sorter analysis of the peripheral blood lymphocytes from two to six mice. Standard deviations from the means are indicated. The average latency of mammary tumors induced by the wild-type HYB MTV-derived virus is shown.

View this table: [in this window] [in a new window]   TABLE 2.   V14+ T-cell deletion and mammary tumor development in the third litters of injected mice

Appearance of MMTV recombinants that regenerate the splice donor site in the envelope gene. Because we observed sporadic deletion of Sag-reactive T cells and long-latency tumors in injected mice, we suspected that recombinants between the injected mutant viruses and endogenous MMTVs were being generated to correct the defective sag genes. Therefore, we used RNA extracted from injected mice or their third-litter progeny to perform RT-PCR with primers that would amplify the sag RNA splice junction and simultaneously detect sequences specific for the C3H MMTV LTR. The resulting product spanned a ClaI site in the C3H MMTV U3 region (Fig. 8A). Because the endogenous MMTVs of BALB/c mice (Mtv-6, -8, and -9) lack a ClaI site at this position, failure to completely digest the PCR product would indicate the presence of recombinants. As expected, ClaI digestion of PCR products from mammary tumors of HYB MTV-injected mice or their third-litter progeny showed complete digestion products of 487 and 111 bp, consistent with the presence of spliced sag RNA from the envelope promoter (Fig. 8B, lanes 2, 3, and 11). A mammary tumor from a mouse injected with the ClaI frameshift mutant expressed the spliced sag mRNA, but the RT-PCR product was not digested with ClaI, as anticipated (lane 9). RNA from mammary tumors of the HYB SD-injected mice gave the product from the gag/pol and env mRNAs, but not the product from the spliced sag mRNA (lanes 4, 5, 6, and 7). The mammary tumor shown in lane 8 also expresses recombinant MMTVs, but these recombinants are not detectable by the C3H-specific primer used (data not shown). Third-litter offspring of HYB SD-injected mice had no C3H-specific products detectable in salivary gland RNA (the mammary gland was not available) (lanes 12 to 14), but RNA from the mammary gland (or salivary gland) of a fifth-litter female that showed deletion of C3H Sag-specific T cells allowed detection of an RT-PCR product of the size expected for sag mRNA. However, most of the product was not digested with ClaI, although a control plasmid in the same reaction was digested completely (lanes 17 and 18). Sequencing analysis confirmed that this PCR product was derived from a recombination between C3H MMTV and Mtv-9 that regenerated a functional splice donor site in the envelope region, while retaining C3H sequences in the Sag region controlling TCR interaction (Fig. 8C). The ClaI-digested products probably result from a different recombinant that repaired the splice donor mutation but retains a functional ClaI site in the LTR. Together with previous experiments, these results suggest that recombinants generated in injected mice are selected during milk-borne transmission for regeneration of the splice donor site in the envelope gene.

View larger version (36K): [in this window] [in a new window]   FIG. 8.   Recombinants from fifth-litter progeny of HYB SD-injected mice repair the splice donor mutation. (A) Diagram showing the primers used and positions of ClaI cleavage sites in RT-PCR products from MMTV-infected mice. The sizes of digested and undigested products are given. (B) Cleavage of RT-PCR products generated using a primer just upstream of the splice donor site in the envelope region [7255(+)] and a C3H-specific primer in the LTR [420()]. RNA from mice inoculated with XC transfectants of wild-type (HYB MTV) or mutant proviruses (HYB SD and HYB CFS) was used for RT-PCRs shown in lanes 2 to 9. Of the splice donor mutant-injected mice, only the mouse in lane 4 showed C3H Sag-specific T-cell deletion. RNA from the progeny of mice inoculated with wild-type or mutant transfectants was used in lanes 11 to 18. None of the third litters tested (total, 13 mice) showed any C3H Sag-specific T-cell deletion, whereas one fifth-litter female showed >50% deletion. The undigested bands representing spliced sag mRNA from the envelope promoter in the lactating mammary gland of the fifth-litter female (lanes 17 and 18) were excised and subjected to sequencing. RT-PCR products were purified using Micro Bio-Spin P-30 chromatography columns (Bio-Rad) prior to incubation with ClaI in the presence of a plasmid digestion control (see arrows). The integrity of cDNA samples was assessed using GAPDH primers for PCRs (lower panel). MT, mammary tumor; CFS, ClaI frameshift virus; SG, salivary gland; SP, spleen; LMG, lactating mammary gland. (C) Comparison of sequences from the recombinant (REC), C3H MMTV, and endogenous Mtv-9. The sequence of the recombinant virus was obtained after isolation of the 598-bp band shown in lanes 17 and 18 of panel B. Only the portion of the splice donor mutation that is present in the spliced sag RNA is shown. The positions of AvrII, ClaI, and C3H-specific oligonucleotide primers are shown. Note that the recombinant analyzed appears to be the result of multiple recombination events, since the REC sequence is Mtv-9-like around the splice junction, C3H-like around the AvrII site, Mtv-9-like around the ClaI site, and C3H-like at the 3' end due to the primer used for PCR.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Requirement for the splice donor site in the envelope gene for C3H Sag function. Previous experiments indicate that Sag is required for efficient milk-borne MMTV expression as well as for viral spread within the mammary gland (12, 13, 17). Because four different MMTV promoters have been implicated in sag mRNA expression (2, 11, 16, 41), we have mutated the splice donor site unique to the spliced sag-specific mRNA from the intragenic envelope promoter in the context of an infectious MMTV provirus (see Fig. 1). Injection of HYB SD-transfected XC cells into weanling BALB/c mice revealed delayed and sporadic MMTV infection as assessed by deletion of Sag-cognate T cells, infection of mammary glands and other tissues, and the appearance of mammary tumors compared to findings for mice injected with wild type-transfected cells. The discrepancies between the wild-type and mutant infections were not due to differences in overall MMTV RNA production or splicing patterns in the original injected XC cells, but were correlated specifically with the absence of the spliced sag mRNA from the 5' intragenic envelope promoter.

Expression of other sag mRNAs. In previous RT-PCR experiments, we were unable to detect spliced sag mRNAs from the C3H MMTV LTR in tissues infected by the virus (44). However, using more sensitive conditions, we have been able to detect sag mRNAs from the LTR promoters in XC cells transfected with the HYB MTV provirus (Fig. 2). Sequencing of these products revealed a single splice donor site identical to that used for spliced env mRNA (11) (data not shown). The singly spliced sag mRNA from the LTR promoters also shares a splice acceptor site with the spliced sag mRNA from the envelope promoter (40, 41). These results indicate that the sag mRNAs from the LTR promoters contain functional splice donor and acceptor sites. Clearly, functional sag mRNAs from the LTR promoter can be synthesized, since Mtv-6 encodes a Sag protein that causes intrathymic deletion of CD4⁺ V3⁺ and V5⁺ T cells (3, 7, 42). This Sag expression must originate from the LTR promoters because the Mtv-6 provirus has a large internal deletion that includes most of the gag, pol, and env sequences (7).