Role of a Cytotoxic-T-Lymphocyte Epitope-Defined,
http://www.100md.com
病菌学杂志 2005年第7期
Department of Microbiology and Immunology and the Norris Cotton Cancer Center, Dartmouth Medical School, Lebanon, New Hampshire
ABSTRACT
LP-BM5 murine leukemia virus-infected C57BL/6 mice develop profound immunodeficiency and B-cell lymphomas. The LP-BM5 complex contains a mixture of defective (BM5def) and replication-competent helper viruses among which BM5def is the primary causative agent of disease. The BM5def primary open reading frame (ORF1) encodes the single gag precursor protein (Pr60gag). Our lab has recently demonstrated that a novel immunodominant cytotoxic-T-lymphocyte (CTL) epitope (SYNTGRFPPL) is expressed from a +1-nucleotide translational open reading frame of BM5def during the course of normal retrovirus expression. The SYNTGRFPPL CTL epitope may be generated from either of two initiation methionines present, ORF2a or ORF2b, located downstream of the ORF1 initiation site. This study investigates the role(s) of the alternative ORF2-derived gag protein(s) of BM5def in viral pathogenesis. We have examined the disease-inducing capabilities of mutant viruses in which the translational potential of either the initiating ORF2a or ORF2b AUG has been disrupted. Although these mutated viruses are capable of wild-type ORF1 expression, they are unable to induce disease. Our data strongly suggest the existence of a novel ORF2 product(s) that is required for LP-BM5-induced pathogenesis and have potentially broad implications for other retroviral diseases.
INTRODUCTION
Genetically susceptible C57BL/6 (B6) mice infected with the LP-BM5 retroviral isolate (18) develop immunodeficiency similar to AIDS. This disease is called murine AIDS (MAIDS), stages of which are characterized by splenomegaly, lymphadenopathy, hypergammaglobulinemia, profound immunodeficiency, and B-cell lymphomas in the context of retroviral-induced immunodeficiency, generally similar to AIDS-related non-Hodgkin's lymphomas. Among the three classes of murine leukemia viruses (MuLV) in the LP-BM5 isolate, the BM5eco (and possibly recombinant mink cell focus-forming virus) MuLV function as helper viruses for the replication-defective BM5def, which is the primary etiological agent causing disease (1, 4, 13, 18). Two molecular clones of the defective genome, Du5H and DEF27, have been characterized from different isolates, with both revealing large common regions of pol and env that have been deleted (1, 5). The BM5def primary open reading frame (ORF1) encodes the single gag precursor protein (Pr60gag). The BM5def gag sequences encoding the core proteins MA (p15), CA (p30), and NC (p10) have 80 to 90% homology to the analogous sequences in the Pr65gag of other ecotropic MuLV. However, within a 25-amino acid (aa) segment at the carboxy-terminal end of p15 and another 25-aa region in p12, BM5def displays a distinctively high diversity from the nonpathogenic BM5eco and other typical replication-competent ecotropic MuLV (1, 5) (Fig. 1A).
All of the work done to understand LP-BM5-induced pathogenesis, including yeast two-hybrid screening for host proteins that bind BM5def gag protein, has focused conceptually on the BM5def ORF1-encoded Pr60gag protein exclusively. The essential role of ORF1 in pathogenesis has been established by generating a myristylation-negative (Myr–) Pr60gag mutant of BM5def (Du5H clone) (12). Similar to other myristylated proteins, Pr60gag on myristylation attaches to the inner leaflet of the plasma membrane. Following this, myristylated Pr60gag may interact with other membrane-bound molecules leading to disease. The Myr– Pr60gag mutant BM5def is unable to induce either the expansion of target B cells or the splenomegaly characteristic of MAIDS, demonstrating that an unaltered Pr60gag ORF1 is necessary for viral pathogenesis (12). Other groups have also established the crucial roles of ORF1 coding regions of the Pr60gag gene in pathogenesis. By examining the disease-inducing potential of MuLV-rescued BM5def carrying truncated gag, it has been proposed that the MA (p15)- and p12-encoding regions of BM5def gag are sufficient for pathogenicity (24). Also, substitution of the p15 and p12 sequences of BM5def virus with BM5eco-specific sequences eliminates its ability to induce MAIDS (17). Furthermore, by comparing chimeric viruses with mutated p12 regions, the unique nucleotide sequence in the BM5def p12gag region, not homologous to BM5eco, appears crucial for MAIDS development (16, 17). Collectively, these observations (12, 16, 17, 24) have been taken as confirmation of the essential role of ORF1 in pathogenesis.
However, these studies have concentrated on modifying only nucleotide sequences and considering their resultant changes in the ORF1-derived Pr60gag polypeptide. The potential that alternative open reading frames (and consequently proteins encoded by them) are embedded within the BM5def gag gene and may play an important role in pathogenesis of BM5def has not yet been studied. In addition to studying viral pathogenesis, our lab has focused on protective immune responses in MAIDS-resistant BALB/c mice (20, 22, 26). We and others have shown that resistance in H-2a and H-2d mouse strains is primarily dependent on CD8+ T cells (19, 20, 22, 23, 26, 27). Specifically, we have demonstrated that an H-2Kd-restricted immunodominant cytotoxic-T-lymphocyte (CTL) epitope is nontraditionally expressed from an alternative translational gag ORF during the course of normal in vivo and in vitro LP-BM5 retrovirus expression (22). Thus, we have shown that in MAIDS-resistant BALB/cByJ, C57BL/KsJ, and (BALB/c x C57BL/6)F1 mice, CTLs raised against BM5eco or BM5def gag protein are directed against the conserved p30-encoding region, but in fact are specific for a novel epitope (SYNTGRFPPL) generated from a second, +1-nucleotide (nt) register defined as ORF2 (20, 22). Our lab also demonstrated that, in the context of the extended minigene constructs used for minimal epitope mapping, generation of SYNTGRFPPL depends on a naturally occurring ORF2 Met initiation codon (22). Hence, it seems likely that SYNTGRFPPL expression in the context of full-length BM5def gag is also the result of conventional AUG-directed translation. Thus, the SYNTGRFPPL epitope may be generated from either one of two potential initiation codons in BM5def gag ORF2, i.e., starting at 157 (ORF2a AUG) or 646 (ORF2b AUG) nt, respectively, downstream of the ORF1 AUG. According to the Kozak consensus sequence (15), ORF2a is in a strong, and ORF2b is in a very strong, initiation codon context (21, 22). This extended alternative gag ORF is unique for retroviral systems, especially when comparing BM5def to other ecotropic MuLV such as AKR623, Friend, and Moloney. Furthermore, given the usual frequent random occurrences of stop codons in the +1- and +2-nt registers, the extended BM5def ORF2 predicted maximum protein is unusually long at 193 aa (Fig. 1A).
Based on these observations, the present study examines the role(s), if any, of the alternative ORF2-derived gag protein(s) of BM5def in viral pathogenesis. Significantly, as the pathogenic BM5def viral component is replication incompetent, the putative ORF2 protein(s) it encodes cannot be crucial to replication. Rather, we hypothesize that in addition to the primary ORF1-encoded Pr60gag, BM5def ORF2 translation proceeding from the ORF2a and/or ORF2b AUG(s) gives rise to protein(s) of 193 or 30 aa in length (maximally), and one or both of these ORF2 products are required for LP-BM5 viral pathogenesis per se.
MATERIALS AND METHODS
Viruses. LP-BM5 was prepared in our laboratory as previously described (11). The molecular clone for the wild-type (wt) BM5def, p127/A1, generously provided by Sisir Chattopadhyay, was used to generate the BM5defORF2aAUGmut or BM5defORF2bAUGmut mutant viruses. A QuickChange kit from Stratagene was used for site-directed mutagenesis. The primers used to generate the BM5defORF2aAUGmut mutant were 5' CCACTAGACGGTACTTTTAATTTAGAC and 3' AATGTCTAAATTAAAAGTACCGTCTAG. To generate the BM5defORF2bAUGmut mutant, the primers used were 5' GTATTGTAACTGACCGTTACCCCCCAAACG and 3' CGTTTGGGGGGTAACGGTCAGTTACAATAC.
Mice. Seven-week-old male C57BL/6NCr mice were purchased from the National Cancer Institute, housed at the Dartmouth Medical School animal facility, and infected intraperitoneally with 2.5 x 104 PFU of LP-BM5 or BM5eco-rescued wt BM5def or BM5defORF2aAUGmut or BM5defORF2bAUGmut viruses at 8 to 10 weeks of age.
Activational and immunosuppressive parameters of MAIDS pathogenesis. Enzyme-linked immunosorbent assay determinations of serum immunoglobulin (Ig), splenomegaly measurements, allo-CTL and 51Cr release assays, and proliferation assays in response to mitogen were carried out as previously described (11).
BM5def gag GFP-tagged fusion proteins. We used the enhanced green fluorescent protein (EGFP) to generate the wt or mutant ORF1AUGEGFP or ORF2AUGEGFP fusion proteins. The GFP fragment was subcloned out of the pEGFP-N1 plasmid (BD Biosciences, Clontech) and in frame with the AUGs of the BM5def ORF1, BM5def ORF2a, or BM5def ORF2b.
Viral load. By using a real-time SYBR-Green/iCycler iQ-based quantitative assay that our lab has recently developed (7), quantitative reverse transcription-PCR (qRT-PCR) assays were used to quantify BM5def and BM5eco gag RNA levels. Primers used in this qRT-PCR assay were designed to specifically distinguish between BM5def gag and BM5eco gag and have been described previously (7). It is important to note that primers for BM5eco gag were generated by using unique BM5eco sequences found at base pairs 1188 and 1370 relative to the BM5eco gag initiating ATG and these defining nucleotides specific to BM5eco (and not to BM5def) were positioned at the 3' end.
RESULTS AND DISCUSSION
Presence of alternative reading frames in BM5def and generation of BM5defORF2aAUGmut and BM5defORF2bAUGmut mutant viruses. By using site-directed mutagenesis, we introduced a series of nucleotide changes in the BM5def molecular clone, Def27, to disrupt the translational potential of either the ORF2a or ORF2b AUG-defined reading frames. These mutants carrying disrupted ORF2a or ORF2b AUGs are referred to as either BM5defORF2aAUGmut or BM5defORF2bAUGmut viruses, respectively. The mutations in ORF2a or ORF2b were carefully designed so as to alter neither the sequence integrity nor the preferred mammalian codon usage (8) of ORF1. The BM5defORF2aAUGmut virus was generated by mutating the most-5' ORF2 start AUG (nt 157 to 159) to ACG and additionally introducing a stop codon at position 169 to 171 (Fig. 1A and B). The BM5defORF2bAUGmut virus was generated by mutating the more-3' ORF2 AUG (nt 646 to 648) to ACG, as it was impossible to engineer an adjacent stop codon without altering the amino acid coding sequence in ORF1 (Fig. 1B). Viral stocks were produced by generating stable transfectants of the wt BM5def (as a transfection control) or the mutated viruses in SC-1eco cells that are chronically infected with the helper BM5eco virus. BM5eco-rescued viral stocks, prior to injection into B6 mice, were confirmed for the presence of BM5eco and BM5def (wt or mutants) by RT-PCR and real-time qRT-PCR assays (7) (data not shown). In addition, the mutations were verified by sequencing. Viral stocks were adjusted for delivery of an equivalent infectious titer of 2.5 x 104 PFU, as determined by the XC-plaque assay (25) (data not shown). Additionally, viral genomic RNA was isolated from the stocks and converted to cDNA, and viral RNA copy numbers were determined by qRT-PCR by using standard curves. We determined that our standard infectious viral titer of 2.5 x 104 PFU of LP-BM5 averaged approximately 1.3 million (±0.2 million) copies of BM5eco RNA and 1.93 million (±0.05 million) copies of BM5def RNA. Similarly, the number of viral genomes per 2.5 x 104 PFU for the BM5eco-rescued viral preparations were as follows: (i) wt BM5def (transfection control), 1.14 million (±0.14 million) copies of BM5eco RNA and 1.97 million (±0.04 million) copies of BM5def RNA; (ii) BM5defORF2aAUGmut, 1.12 million (±0.2 million) copies of BM5eco RNA and 1.82 million (±0.04 million) copies of BM5def RNA; (iii) BM5defORF2bAUGmut, 1.32 million (±0.2 million) copies of BM5eco RNA and 1.9 million (±0.03 million) copies of BM5def RNA.
Ability of BM5def ORF2a and ORF2b AUGs to initiate translation of GFP fusion proteins. In order to determine the translational potential of the alternative ORF2a and ORF2b AUGs, we first generated BM5defORF1AUGEGFP (as a positive control) and BM5defORF2aAUGEGFP and BM5defORF2bAUGEGFP tagged viruses. These constructs allow for the expression of the ORF1 or ORF2a or ORF2b AUG-defined reading frames as fusions to the N terminus of EGFP (Fig. 2A). Thus, the EGFP gene, including its endogenous stop codon, was cloned at the C terminus of, and in the same reading frame as, either the ORF1, ORF2a, or ORF2b AUGs, with no intervening in-frame stop codons. For the BM5defORF2aAUGEGFP construct, EGFP was cloned 5' of the ORF2b initiation AUG to ensure that all EGFP expression would necessarily emanate only from the ORF2a initiation AUG. Stable transfectants of the EGFP-tagged BM5def viruses were generated in SC-1eco cells and analyzed for EGFP fusion protein expression by fluorescence-activated cell sorting (FACS). GFP expression was detected in cells carrying BM5defORF1AUGEGFP as expected and also by BM5defORF2aAUGEGFP- or BM5defORF2bAUGEGFP-transfected cells (Fig. 2A). Importantly, BM5defORF2a/2b AUG-directed EGFP expression was detected at somewhat lower (24% lower total mean fluorescence intensity) but still very substantial levels compared to the conventional BM5def ORF1-directed levels. Additionally, to ascertain how much EGFP expression in the BM5defORF2bAUGEGFP construct originated from the ORF2b AUG specifically (and not as a read-through from the upstream intact ORF2a AUG), we also generated a construct wherein the ORF2a AUG was mutated and the EGFP gene was cloned 5' of and in the same reading frame as the ORF2b (BM5defORF2aAUGmut;ORF2bAUGEGFP, Fig. 2B). SC-1eco cells transfected with this BM5defORF2aAUGmut;ORF2bAUGEGFP virus expressed considerable GFP (only 37% lower total mean fluorescence intensity than the BM5defORF2bAUG carrying the unmutated ORF2a AUG), confirming that the ORF2b initiation AUG was independently capable of translating a fusion protein without any read-through from the upstream intact ORF2a AUG (Fig. 2B).
To verify that the GFP was being translated from the initiation AUGs of ORF2a or ORF2b and that the introduced mutations were effective in disrupting ORF2a or ORF2b AUG-initiated translation, EGFP was similarly engineered into the BM5defORF2aAUGmut and BM5defORF2bAUGmut viruses (Fig. 2C). To generate the BM5defORF2aAUGmutEGFP virus, EGFP was cloned 5' of the ORF2b initiation AUG. Compared to the BM5defORF2aAUGEGFP, the total mean fluorescence intensity of BM5defORF2aAUGmutEGFP was consistently lowered in three independent experiments by >99%; essentially this mutant did not appear to express detectable GFP (Fig. 2C). These data clearly show that the ORF2a AUG is capable of initiating translation of a fusion protein, and the introduced mutations inhibit ORF2a AUG-initiated translation. In contrast, mutation of the ORF2b initiation codon did not completely inhibit the EGFP signal (Fig. 2C), as predicted from the results in Fig. 2B. Rather, the mutated BM5defORF2bAUGmutEGFP expressed substantial GFP (only 30% lower total mean fluorescence intensity than BM5defORF2bAUG), most probably due to the read-through from the intact ORF2a AUG (also see below).
BM5def ORF1-encoded gag is expressed in the BM5defORF2aAUGmut or BM5defORF2bAUGmut mutants. To confirm that the mutations introduced in BM5def ORF2a or ORF2b AUGs are in no way detrimental to the translation of ORF1, which is required for viral pathogenesis, we also generated viruses with the ORF1AUGEGFP fusion construct in both the BM5defORF2aAUGmut and BM5defORF2bAUGmut viruses. The EGFP fragment was cloned in frame with the initiating AUG, and at the C terminus, of ORF1 in both mutants (Fig. 3). The wt BM5def carrying the ORF1AUGEGFP fusion protein defined above was used as a positive control in analogous FACS experiments. ORF1AUGEGFP expression was detected in the BM5defORF2aAUGmut or BM5defORF2bAUGmut viruses at levels indistinguishable from this positive control (Fig. 3). These data confirm that ORF1 AUG-initiated translation is not disrupted in the BM5defORF2aAUGmut or BM5defORF2bAUGmut mutants, thus allowing us to address the potential need for ORF2 translation for pathogenesis under conditions of normal levels of required ORF1 product.
Activational and immunosuppressive parameters of MAIDS pathogenesis are not induced by the BM5defORF2aAUGmut or BM5defORF2bAUGmut mutants. To assess the pathogenicity of BM5def viruses mutated to inhibit translation from their alternative ORF2 AUGs, B6 mice were infected with rescued preparations carrying BM5eco virus and either the BM5defORF2aAUGmut or BM5defORF2bAUGmut mutant virus or with positive (LP-BM5 and wt BM5def rescued with BM5eco) or negative (phosphate-buffered saline) controls. Splenomegaly, as a standard activational parameter characteristic of MAIDS (10, 11), was determined 11 weeks postinfection (wpi) by spleen weight (Fig. 4A). At 5, 9, and 11 wpi, serum IgG2a and IgM levels were determined (11 wpi; Fig. 4B), as a measure of the generalized hyper-Ig of MAIDS. Compared to the LP-BM5 or wt BM5def rescued control-infected mice, B6 mice infected with either the BM5defORF2aAUGmut (P < 0.001) or BM5defORF2bAUGmut (P < 0.001) virus did not induce measurable splenomegaly or hypergammaglobulinemia. This pattern was replicated in three additional independent experiments.
In addition, a standard panel of immunosuppressive parameters was measured at termination 11 wpi: responses to mitogens by T (ConA) versus B (lipopolysaccharide) cells, as well as the generation of alloantigen-specific (anti-H-2d) CTL activity. Again, in contrast to LP-BM5 or wt BM5def rescued control-infected mice, B6 mice infected with the BM5eco-rescued BM5defORF2aAUGmut or BM5defORF2bAUGmut mutants did not develop disease, i.e., they had normal mitogen responses (Fig. 4C; data representative of four independent experiments, P < 0.001) and generation of CTL activity (Fig. 4D; data representative of three independent experiments). These results were further confirmed by flow cytometric analyses for the characteristic spleen cell phenotypic changes that are observed in uninfected versus LP-BM5-infected B6 mice (9, 14). LP-BM5 or wt BM5def rescued control-infected mice showed the expected expanded populations of CD19+ CD23– and CD4+ Thy1.2– spleen cells, but these changes were not observed in mice infected with the BM5defORF2aAUGmut or BM5defORF2bAUGmut mutant (data not shown).
Determination of BM5def and BM5eco gag RNA levels in infected mice. On termination of each experiment, we verified the presence of the appropriate control and experimental viruses in infected mice and quantified their expression by qRT-PCR assays for BM5eco and BM5def (7). Although there was no evidence of MAIDS induction by the BM5defORF2aAUGmut or BM5defORF2bAUGmut preps, mice infected with these mutants expressed substantial levels of both BM5eco and BM5def (Fig. 5). BM5def gag RNA was more abundant than BM5eco gag RNA in spleen cells from both control and viral mutant-infected mice. These data are similar to what our lab has observed previously (7) as well as to BM5def and BM5eco gag RNA levels determined in the viral stocks that were generated in vitro and used for infecting mice in this study (data not shown). BM5eco gag and BM5def gag RNA levels from mice infected with the BM5defORF2aAUGmut or BM5defORF2bAUGmut mutant viruses were only about two- to threefold lower than the viral loads for wt rescued control and LP-BM5-infected mice (Fig. 5). It seems very unlikely that this degree of difference in viral load would be sufficient to account for the absence of disease induction after infection by the mutated BM5def preparations. The higher viral loads in the positive controls, in this case infection with nonmutated viral preps, seem to simply reflect the fact that in full-blown MAIDS the characteristic extensive lymphoproliferation is a major contributor to the spread of the virus, in addition to multiple rounds of infection by free virions. This concept that B6 background mouse strains resistant to MAIDS pathogenesis still exhibit a substantial viral load indicative of normal initial infection and early viral spread has been further supported by both viral titration experiments and the study of several MAIDS-resistant versus -susceptible B6 strains, with either of the genes knocked out (CD154, CD40, CD80/86) or as chimeric CD40 transgenics (9, 10, 11). As we have previously discussed in some detail (9), these comparisons across mouse strains have shown that distinctly measurable disease can be observed under conditions that result in a few-fold lower BM5def viral load at termination of the experiment. In addition, on termination of each experiment, by using RT-PCR we isolated the BM5def viruses from spleen cells and sequenced them. Sequencing data confirmed the presence of only the wt BM5def in the mice infected with the positive control viral preps, versus only the appropriate mutant sequence in the mice infected with the experimental viruses. An inherent advantage of this system is that any reversion of the mutated to the wt sequence would have resulted in a defective virus that would have caused disease. These data clearly suggest that a virus different from what was initially injected into the mice had not arisen to detectable levels as a result of reversion during the 11-week course of infection.
In summary, based on (i) both activational and immunodeficiency parameters of MAIDS (Fig. 4), (ii) viral load data (Fig. 5), and (iii) sequence analysis of viruses isolated from mice 11 wpi, and ORF1/ORF2 expression data (Fig. 2 and 3), it is evident that the viral mixes consisting of either wt BM5eco-rescued ORF2aAUGmut or ORF2bAUGmut BM5def virus, each with intact wt ORF1 expression capability, are essentially unable to induce disease. These data strongly suggest the existence of a novel ORF2 product(s) that is required for LP-BM5-induced pathogenesis. The ORF2 product(s) may be initiated from either the ORF2a or ORF2b AUG (Fig. 2), as disruption of either one of their translational potentials leads to inhibition of disease (Fig. 4). One interpretation of these findings is that there is a requirement for both protein products of 193 and 30 aa, produced as separate entities from the ORF2a and ORF2b AUGs, respectively. These proteins may be independently essential in targeting different specific host cellular proteins. Alternatively or in addition, the two alternative ORF2 translational products may function together to cause disease by independently interacting with the same host cell protein target or combining to form a single ORF2a/ORF2b heterodimer.
As an option to two separate ORF2 protein products, it is also possible that only a single protein initiating from the ORF2a AUG is required. Thus, the data from the BM5defORF2bAUGmutEGFP mutant virus (Fig. 2C) show that, despite disrupting the initiation AUG of ORF2b, GFP is still expressed (albeit at a reduced level) as a fusion protein, apparently from the unaltered ORF2a AUG which is in the same +1-nt reading frame as ORF2b (Fig. 2B). Importantly, however, the BM5defORF2bAUGmut virus is unable to cause disease. This finding could be explained based not on the ORF2b AUG mutation as a block to translation initiation at this site but rather on the basis of the introduction of a missense mutation at the ORF2b site in the ORF2a AUG-initiated maximal ORF2 protein. Thus, the presence of threonine instead of methionine may be deleterious to the structure of the extended protein produced from the ORF2a AUG.
The results presented in this paper support the existence of an alternative translation initiation mechanism in BM5def for the expression of an ORF2-encoded gag product(s) and its requirement in LP-BM5-induced pathogenesis. These data have obvious implications for pursuing the molecular mechanisms by which BM5def causes disease and studies are currently under way to identify the nature and function of the necessary ORF2 product(s). Our data are interesting to compare to the results of other labs that have also studied atypically encoded CTL epitopes, including in human immunodeficiency virus (2, 3, 6, 28). In particular, and comparable to our novel BM5def ORF2a and/or ORF2b AUG(s)-encoded 193 or 30 aa protein(s), a previously unobserved, conserved 87-residue protein encoded by influenza A virus has recently been identified by first defining a CTL epitope encoded in an alternative translational ORF (6). This immunogenic influenza virus protein, PB1-F2, recognized by CD8+ T cells, is derived from a +1 alternative reading frame and may play a critical role in pathogenesis by inducing apoptosis in host immune cells that respond to influenza virus infection (6). Our identification of an immunodominant CTL epitope novelly expressed from an alternative translational gag ORF (20, 22), combined with our present data that the ORF2-encoded extended gag product(s) have a functional role in disease induction, raise the possibility that alternative reading frame-derived proteins may also be crucial in other disease models.
ACKNOWLEDGMENTS
We especially thank On Ho for her contributions in determining disease parameters as well as helpful discussions during the course of these experiments. We also thank Amanda Birdsey for generating the BM5defORF2bAUGmutEGFP construct and the stable line expressing this construct and Kathy Green for generating the LP-BM5 viral stocks as well as for helpful discussions.
A.G. was supported by NIH Institutional Training Grant T32 A017363. This work was supported in part by U.S. Public Health Service grant CA50157. The flow cytometer was the generous gift of the Fannie E. Rippel Foundation and is partially supported by the core grant of the Norris Cotton Cancer Center (CA23108).
REFERENCES
Aziz, D. C., Z. Hanna, and P. Jolicoeur. 1989. Severe immunodeficiency disease induced by a defective murine leukemia virus. Nature 338:505-508.
Bullock, T. N., and L. C. Eisenlohr. 1996. Ribosomal scanning past the primary initiation codon as a mechanism for expression of CTL epitopes encoded in alternative reading frames. J. Exp. Med. 184:1319-1329.
Cardinaud, S., A. Moris, M. Fevrier, P. S. Rohrlich, L. Weiss, P. Langlade-Demoyen, F. A. Lemonnier, O. Schwartz, and A. Habel. 2004. Identification of cryptic MHC I-restricted epitopes encoded by HIV-1 alternative reading frames. J. Exp. Med. 199:1053-1063.
Chattopadhyay, S. K., H. C. Morse III, M. Makino, S. K. Ruscetti, and J. W. Hartley. 1989. Defective virus is associated with induction of murine retrovirus-induced immunodeficiency syndrome. Proc. Natl. Acad. Sci. USA 86:3862-3869.
Chattopadhyay, S. K., D. N. Sengupta, T. N Fredrickson, H. C. Morse III, and J. W. Hartley. 1991. Characteristics and contributions of defective, ecotropic, and mink cell focus-inducing viruses involved in a retrovirus-induced immunodeficiency syndrome of mice. J. Virol. 65:4232-4241.
Chen, W., P. A. Calvo, D. Malide, J. Gibbs, U. Schubert, I. Bacik, S. Basta, R. O'Neill, J. Schickli, P. Palese, P. Henklein, J. R. Bennink, and J. W. Yewdell. 2001. A novel influenza A virus mitochondrial protein that induces cell death. Nat. Med. 7:1306-1312.
Cook, W. J., K. A. Green, J. J. Obar, and W. R. Green. 2003. Quantitative analysis of LP-BM5 murine leukemia retrovirus RNA using real-time RT-PCR. J. Virol. Methods 108:49-58.
Fedorov, A., S. Saxonov, and W. Gilbert. 2002. Regularities of context-dependent codon bias in eukaryotic genes. Nucleic Acids Res. 30:1192-1197.
Green, K. A., C. L. Ahonen, W. J. Cook, and W. R. Green. 2004. CD40-associated TRAF6 signaling is required for disease induction in a retrovirus-induced murine immunodeficiency. J. Virol. 78:6055-6060.
Green, K. A., W. J. Cook, A. H. Sharpe, and W. R. Green. 2002. The CD154/CD40 interaction required for retrovirus-induced murine immunodeficiency syndrome is not mediated by upregulation of the CD80/CD86 costimulatory molecules. J. Virol. 76:13106-13110.
Green, K. A., R. J. Noelle, B. G. Durell, and W. R. Green. 2001. Characterization of the CD154-positive and CD40-positive cellular subsets required for pathogenesis in retrovirus-induced murine immunodeficiency. J. Virol. 75:3581-3589.
Huang, M., and P. Jolicoeur. 1994. Myristylation of Pr60gag of the murine AIDS-defective virus is required to induce disease and notably for the expansion of its target cells. J. Virol. 68:5648-5655.
Huang, M., C. Simard, and P. Jolicoeur. 1989. Immunodeficiency and clonal growth of target cells induced by helper-free defective retrovirus. Science 246:1614-1617.
Knoetig, S. M., T. A. Torrey, Z. Naghashfar, T. McCarty, and H. C. Morse III. 2002. CD19 signaling pathways play a major role for murine AIDS induction and progression. J. Immunol. 169:5607-5610.
Kozak, M. 1989. Context effects and inefficient initiation at non-AUG codons in eucaryotic cell-free translation systems. Mol. Cell. Biol. 9:5073-5080.
Kubo, Y., K. Kakimi, K. Higo, H. Kobayashi, T. Ono, Y. Iwama, K. Kuribayashi, H. Hiai, A. Adachi, and A. Ishimoto. 1996. Possible origin of murine AIDS (MAIDS) virus: conversion of an endogenous retroviral p12gag sequence to a MAIDS-inducing sequence by frameshift mutations. J. Virol. 70:6405-6409.
Kubo, Y., K. Kakimi, K. Higo, L. Wang, H. Kobayashi, K. Kuribayashi, T. Masuda, T. Hirama, and A. Ishimoto. 1994. The p15gag and p12gag regions are both necessary for the pathogenicity of the murine AIDS virus. J. Virol. 68:5532-5537.
Laterjet, R., and J. F. Duplan. 1962. Experiment and discussion on leukemogenesis by cell-free extracts of radiation-induced leukemia in mice. Int. J. Radiat. Biol. 5:223-233.
Makino, M., S. K. Chattopadhyay, J. W. Hartley, and H. C. Morse III. 1992. Analysis of role of CD8+ T cells in resistance to murine AIDS in A/J mice. J. Immunol. 149:1702-1706.
Mayrand, S. M., P. A. Healy, B. E Torbett, and W. R. Green. 2000. Anti-Gag cytolytic T lymphocytes specific for an alternative translational reading frame-derived epitope and resistance versus susceptibility to retrovirus-induced murine AIDS in F1 mice. Virology 272:438-449.
Mayrand, S. M., and W. R. Green. 1998. Non-traditionally derived CTL epitopes: exceptions that prove the rules? Immunol. Today 19:551-556.
Mayrand, S. M., D. A. Schwarz, and W. R. Green. 1998. An alternative translational reading frame encodes an immunodominant retroviral CTL determinant expressed by an immunodeficiency-causing retrovirus. J Immunol. 160:39-50.
Pavlovitch, J., E. Hulier, M. Rizk-Rabin, M. Marussig, D. Mazier, M. Joffret, S. Hoos, and M. Papiernik. 1996. Resistance to murine AIDS in offspring of mice infected with LP-BM5: role of CD8 T cells. J. Immunol. 156:4757-4765.
Pozsgay, J. M., M. W. Beilharz, B. D. Wines, A. D. Hess, and P. M. Pitha. 1993. The MA (p15) and p12 regions of the gag gene are sufficient for the pathogenicity of the murine AIDS virus. J. Virol. 67:5989-5999.
Rowe, W. P., W. E. Pugh, and J. W. Hartley. 1970. Plaque assay techniques for murine leukemia viruses. Virology 42:1136-1139.
Schwarz, D. A., and W. R. Green. 1994. CTL responses to the gag polyprotein encoded by the murine AIDS defective retrovirus are strain dependent. J. Immunol. 153:436-442.
Tang, Y., A. W. Hugin, N. A. Giese, L. Gabriele, S. K Chattopadhyay, T. N. Fredrickson, D. Kagi, J. W. Hartley, and H. C. Morse III. 1997. Control of immunodeficiency and lymphoproliferation in mouse AIDS: studies of mice deficient in CD8+ T cells or perforin. J. Virol. 71:1808-1815.
Vassilaki, N., and P. Mavromara. 2003. Two alternative translation mechanisms are responsible for the expression of the HCV ARFP/F/core+1 coding open reading frame. J. Biol. Chem. 278:40503-40513.(Arti Gaur and William R. )
ABSTRACT
LP-BM5 murine leukemia virus-infected C57BL/6 mice develop profound immunodeficiency and B-cell lymphomas. The LP-BM5 complex contains a mixture of defective (BM5def) and replication-competent helper viruses among which BM5def is the primary causative agent of disease. The BM5def primary open reading frame (ORF1) encodes the single gag precursor protein (Pr60gag). Our lab has recently demonstrated that a novel immunodominant cytotoxic-T-lymphocyte (CTL) epitope (SYNTGRFPPL) is expressed from a +1-nucleotide translational open reading frame of BM5def during the course of normal retrovirus expression. The SYNTGRFPPL CTL epitope may be generated from either of two initiation methionines present, ORF2a or ORF2b, located downstream of the ORF1 initiation site. This study investigates the role(s) of the alternative ORF2-derived gag protein(s) of BM5def in viral pathogenesis. We have examined the disease-inducing capabilities of mutant viruses in which the translational potential of either the initiating ORF2a or ORF2b AUG has been disrupted. Although these mutated viruses are capable of wild-type ORF1 expression, they are unable to induce disease. Our data strongly suggest the existence of a novel ORF2 product(s) that is required for LP-BM5-induced pathogenesis and have potentially broad implications for other retroviral diseases.
INTRODUCTION
Genetically susceptible C57BL/6 (B6) mice infected with the LP-BM5 retroviral isolate (18) develop immunodeficiency similar to AIDS. This disease is called murine AIDS (MAIDS), stages of which are characterized by splenomegaly, lymphadenopathy, hypergammaglobulinemia, profound immunodeficiency, and B-cell lymphomas in the context of retroviral-induced immunodeficiency, generally similar to AIDS-related non-Hodgkin's lymphomas. Among the three classes of murine leukemia viruses (MuLV) in the LP-BM5 isolate, the BM5eco (and possibly recombinant mink cell focus-forming virus) MuLV function as helper viruses for the replication-defective BM5def, which is the primary etiological agent causing disease (1, 4, 13, 18). Two molecular clones of the defective genome, Du5H and DEF27, have been characterized from different isolates, with both revealing large common regions of pol and env that have been deleted (1, 5). The BM5def primary open reading frame (ORF1) encodes the single gag precursor protein (Pr60gag). The BM5def gag sequences encoding the core proteins MA (p15), CA (p30), and NC (p10) have 80 to 90% homology to the analogous sequences in the Pr65gag of other ecotropic MuLV. However, within a 25-amino acid (aa) segment at the carboxy-terminal end of p15 and another 25-aa region in p12, BM5def displays a distinctively high diversity from the nonpathogenic BM5eco and other typical replication-competent ecotropic MuLV (1, 5) (Fig. 1A).
All of the work done to understand LP-BM5-induced pathogenesis, including yeast two-hybrid screening for host proteins that bind BM5def gag protein, has focused conceptually on the BM5def ORF1-encoded Pr60gag protein exclusively. The essential role of ORF1 in pathogenesis has been established by generating a myristylation-negative (Myr–) Pr60gag mutant of BM5def (Du5H clone) (12). Similar to other myristylated proteins, Pr60gag on myristylation attaches to the inner leaflet of the plasma membrane. Following this, myristylated Pr60gag may interact with other membrane-bound molecules leading to disease. The Myr– Pr60gag mutant BM5def is unable to induce either the expansion of target B cells or the splenomegaly characteristic of MAIDS, demonstrating that an unaltered Pr60gag ORF1 is necessary for viral pathogenesis (12). Other groups have also established the crucial roles of ORF1 coding regions of the Pr60gag gene in pathogenesis. By examining the disease-inducing potential of MuLV-rescued BM5def carrying truncated gag, it has been proposed that the MA (p15)- and p12-encoding regions of BM5def gag are sufficient for pathogenicity (24). Also, substitution of the p15 and p12 sequences of BM5def virus with BM5eco-specific sequences eliminates its ability to induce MAIDS (17). Furthermore, by comparing chimeric viruses with mutated p12 regions, the unique nucleotide sequence in the BM5def p12gag region, not homologous to BM5eco, appears crucial for MAIDS development (16, 17). Collectively, these observations (12, 16, 17, 24) have been taken as confirmation of the essential role of ORF1 in pathogenesis.
However, these studies have concentrated on modifying only nucleotide sequences and considering their resultant changes in the ORF1-derived Pr60gag polypeptide. The potential that alternative open reading frames (and consequently proteins encoded by them) are embedded within the BM5def gag gene and may play an important role in pathogenesis of BM5def has not yet been studied. In addition to studying viral pathogenesis, our lab has focused on protective immune responses in MAIDS-resistant BALB/c mice (20, 22, 26). We and others have shown that resistance in H-2a and H-2d mouse strains is primarily dependent on CD8+ T cells (19, 20, 22, 23, 26, 27). Specifically, we have demonstrated that an H-2Kd-restricted immunodominant cytotoxic-T-lymphocyte (CTL) epitope is nontraditionally expressed from an alternative translational gag ORF during the course of normal in vivo and in vitro LP-BM5 retrovirus expression (22). Thus, we have shown that in MAIDS-resistant BALB/cByJ, C57BL/KsJ, and (BALB/c x C57BL/6)F1 mice, CTLs raised against BM5eco or BM5def gag protein are directed against the conserved p30-encoding region, but in fact are specific for a novel epitope (SYNTGRFPPL) generated from a second, +1-nucleotide (nt) register defined as ORF2 (20, 22). Our lab also demonstrated that, in the context of the extended minigene constructs used for minimal epitope mapping, generation of SYNTGRFPPL depends on a naturally occurring ORF2 Met initiation codon (22). Hence, it seems likely that SYNTGRFPPL expression in the context of full-length BM5def gag is also the result of conventional AUG-directed translation. Thus, the SYNTGRFPPL epitope may be generated from either one of two potential initiation codons in BM5def gag ORF2, i.e., starting at 157 (ORF2a AUG) or 646 (ORF2b AUG) nt, respectively, downstream of the ORF1 AUG. According to the Kozak consensus sequence (15), ORF2a is in a strong, and ORF2b is in a very strong, initiation codon context (21, 22). This extended alternative gag ORF is unique for retroviral systems, especially when comparing BM5def to other ecotropic MuLV such as AKR623, Friend, and Moloney. Furthermore, given the usual frequent random occurrences of stop codons in the +1- and +2-nt registers, the extended BM5def ORF2 predicted maximum protein is unusually long at 193 aa (Fig. 1A).
Based on these observations, the present study examines the role(s), if any, of the alternative ORF2-derived gag protein(s) of BM5def in viral pathogenesis. Significantly, as the pathogenic BM5def viral component is replication incompetent, the putative ORF2 protein(s) it encodes cannot be crucial to replication. Rather, we hypothesize that in addition to the primary ORF1-encoded Pr60gag, BM5def ORF2 translation proceeding from the ORF2a and/or ORF2b AUG(s) gives rise to protein(s) of 193 or 30 aa in length (maximally), and one or both of these ORF2 products are required for LP-BM5 viral pathogenesis per se.
MATERIALS AND METHODS
Viruses. LP-BM5 was prepared in our laboratory as previously described (11). The molecular clone for the wild-type (wt) BM5def, p127/A1, generously provided by Sisir Chattopadhyay, was used to generate the BM5defORF2aAUGmut or BM5defORF2bAUGmut mutant viruses. A QuickChange kit from Stratagene was used for site-directed mutagenesis. The primers used to generate the BM5defORF2aAUGmut mutant were 5' CCACTAGACGGTACTTTTAATTTAGAC and 3' AATGTCTAAATTAAAAGTACCGTCTAG. To generate the BM5defORF2bAUGmut mutant, the primers used were 5' GTATTGTAACTGACCGTTACCCCCCAAACG and 3' CGTTTGGGGGGTAACGGTCAGTTACAATAC.
Mice. Seven-week-old male C57BL/6NCr mice were purchased from the National Cancer Institute, housed at the Dartmouth Medical School animal facility, and infected intraperitoneally with 2.5 x 104 PFU of LP-BM5 or BM5eco-rescued wt BM5def or BM5defORF2aAUGmut or BM5defORF2bAUGmut viruses at 8 to 10 weeks of age.
Activational and immunosuppressive parameters of MAIDS pathogenesis. Enzyme-linked immunosorbent assay determinations of serum immunoglobulin (Ig), splenomegaly measurements, allo-CTL and 51Cr release assays, and proliferation assays in response to mitogen were carried out as previously described (11).
BM5def gag GFP-tagged fusion proteins. We used the enhanced green fluorescent protein (EGFP) to generate the wt or mutant ORF1AUGEGFP or ORF2AUGEGFP fusion proteins. The GFP fragment was subcloned out of the pEGFP-N1 plasmid (BD Biosciences, Clontech) and in frame with the AUGs of the BM5def ORF1, BM5def ORF2a, or BM5def ORF2b.
Viral load. By using a real-time SYBR-Green/iCycler iQ-based quantitative assay that our lab has recently developed (7), quantitative reverse transcription-PCR (qRT-PCR) assays were used to quantify BM5def and BM5eco gag RNA levels. Primers used in this qRT-PCR assay were designed to specifically distinguish between BM5def gag and BM5eco gag and have been described previously (7). It is important to note that primers for BM5eco gag were generated by using unique BM5eco sequences found at base pairs 1188 and 1370 relative to the BM5eco gag initiating ATG and these defining nucleotides specific to BM5eco (and not to BM5def) were positioned at the 3' end.
RESULTS AND DISCUSSION
Presence of alternative reading frames in BM5def and generation of BM5defORF2aAUGmut and BM5defORF2bAUGmut mutant viruses. By using site-directed mutagenesis, we introduced a series of nucleotide changes in the BM5def molecular clone, Def27, to disrupt the translational potential of either the ORF2a or ORF2b AUG-defined reading frames. These mutants carrying disrupted ORF2a or ORF2b AUGs are referred to as either BM5defORF2aAUGmut or BM5defORF2bAUGmut viruses, respectively. The mutations in ORF2a or ORF2b were carefully designed so as to alter neither the sequence integrity nor the preferred mammalian codon usage (8) of ORF1. The BM5defORF2aAUGmut virus was generated by mutating the most-5' ORF2 start AUG (nt 157 to 159) to ACG and additionally introducing a stop codon at position 169 to 171 (Fig. 1A and B). The BM5defORF2bAUGmut virus was generated by mutating the more-3' ORF2 AUG (nt 646 to 648) to ACG, as it was impossible to engineer an adjacent stop codon without altering the amino acid coding sequence in ORF1 (Fig. 1B). Viral stocks were produced by generating stable transfectants of the wt BM5def (as a transfection control) or the mutated viruses in SC-1eco cells that are chronically infected with the helper BM5eco virus. BM5eco-rescued viral stocks, prior to injection into B6 mice, were confirmed for the presence of BM5eco and BM5def (wt or mutants) by RT-PCR and real-time qRT-PCR assays (7) (data not shown). In addition, the mutations were verified by sequencing. Viral stocks were adjusted for delivery of an equivalent infectious titer of 2.5 x 104 PFU, as determined by the XC-plaque assay (25) (data not shown). Additionally, viral genomic RNA was isolated from the stocks and converted to cDNA, and viral RNA copy numbers were determined by qRT-PCR by using standard curves. We determined that our standard infectious viral titer of 2.5 x 104 PFU of LP-BM5 averaged approximately 1.3 million (±0.2 million) copies of BM5eco RNA and 1.93 million (±0.05 million) copies of BM5def RNA. Similarly, the number of viral genomes per 2.5 x 104 PFU for the BM5eco-rescued viral preparations were as follows: (i) wt BM5def (transfection control), 1.14 million (±0.14 million) copies of BM5eco RNA and 1.97 million (±0.04 million) copies of BM5def RNA; (ii) BM5defORF2aAUGmut, 1.12 million (±0.2 million) copies of BM5eco RNA and 1.82 million (±0.04 million) copies of BM5def RNA; (iii) BM5defORF2bAUGmut, 1.32 million (±0.2 million) copies of BM5eco RNA and 1.9 million (±0.03 million) copies of BM5def RNA.
Ability of BM5def ORF2a and ORF2b AUGs to initiate translation of GFP fusion proteins. In order to determine the translational potential of the alternative ORF2a and ORF2b AUGs, we first generated BM5defORF1AUGEGFP (as a positive control) and BM5defORF2aAUGEGFP and BM5defORF2bAUGEGFP tagged viruses. These constructs allow for the expression of the ORF1 or ORF2a or ORF2b AUG-defined reading frames as fusions to the N terminus of EGFP (Fig. 2A). Thus, the EGFP gene, including its endogenous stop codon, was cloned at the C terminus of, and in the same reading frame as, either the ORF1, ORF2a, or ORF2b AUGs, with no intervening in-frame stop codons. For the BM5defORF2aAUGEGFP construct, EGFP was cloned 5' of the ORF2b initiation AUG to ensure that all EGFP expression would necessarily emanate only from the ORF2a initiation AUG. Stable transfectants of the EGFP-tagged BM5def viruses were generated in SC-1eco cells and analyzed for EGFP fusion protein expression by fluorescence-activated cell sorting (FACS). GFP expression was detected in cells carrying BM5defORF1AUGEGFP as expected and also by BM5defORF2aAUGEGFP- or BM5defORF2bAUGEGFP-transfected cells (Fig. 2A). Importantly, BM5defORF2a/2b AUG-directed EGFP expression was detected at somewhat lower (24% lower total mean fluorescence intensity) but still very substantial levels compared to the conventional BM5def ORF1-directed levels. Additionally, to ascertain how much EGFP expression in the BM5defORF2bAUGEGFP construct originated from the ORF2b AUG specifically (and not as a read-through from the upstream intact ORF2a AUG), we also generated a construct wherein the ORF2a AUG was mutated and the EGFP gene was cloned 5' of and in the same reading frame as the ORF2b (BM5defORF2aAUGmut;ORF2bAUGEGFP, Fig. 2B). SC-1eco cells transfected with this BM5defORF2aAUGmut;ORF2bAUGEGFP virus expressed considerable GFP (only 37% lower total mean fluorescence intensity than the BM5defORF2bAUG carrying the unmutated ORF2a AUG), confirming that the ORF2b initiation AUG was independently capable of translating a fusion protein without any read-through from the upstream intact ORF2a AUG (Fig. 2B).
To verify that the GFP was being translated from the initiation AUGs of ORF2a or ORF2b and that the introduced mutations were effective in disrupting ORF2a or ORF2b AUG-initiated translation, EGFP was similarly engineered into the BM5defORF2aAUGmut and BM5defORF2bAUGmut viruses (Fig. 2C). To generate the BM5defORF2aAUGmutEGFP virus, EGFP was cloned 5' of the ORF2b initiation AUG. Compared to the BM5defORF2aAUGEGFP, the total mean fluorescence intensity of BM5defORF2aAUGmutEGFP was consistently lowered in three independent experiments by >99%; essentially this mutant did not appear to express detectable GFP (Fig. 2C). These data clearly show that the ORF2a AUG is capable of initiating translation of a fusion protein, and the introduced mutations inhibit ORF2a AUG-initiated translation. In contrast, mutation of the ORF2b initiation codon did not completely inhibit the EGFP signal (Fig. 2C), as predicted from the results in Fig. 2B. Rather, the mutated BM5defORF2bAUGmutEGFP expressed substantial GFP (only 30% lower total mean fluorescence intensity than BM5defORF2bAUG), most probably due to the read-through from the intact ORF2a AUG (also see below).
BM5def ORF1-encoded gag is expressed in the BM5defORF2aAUGmut or BM5defORF2bAUGmut mutants. To confirm that the mutations introduced in BM5def ORF2a or ORF2b AUGs are in no way detrimental to the translation of ORF1, which is required for viral pathogenesis, we also generated viruses with the ORF1AUGEGFP fusion construct in both the BM5defORF2aAUGmut and BM5defORF2bAUGmut viruses. The EGFP fragment was cloned in frame with the initiating AUG, and at the C terminus, of ORF1 in both mutants (Fig. 3). The wt BM5def carrying the ORF1AUGEGFP fusion protein defined above was used as a positive control in analogous FACS experiments. ORF1AUGEGFP expression was detected in the BM5defORF2aAUGmut or BM5defORF2bAUGmut viruses at levels indistinguishable from this positive control (Fig. 3). These data confirm that ORF1 AUG-initiated translation is not disrupted in the BM5defORF2aAUGmut or BM5defORF2bAUGmut mutants, thus allowing us to address the potential need for ORF2 translation for pathogenesis under conditions of normal levels of required ORF1 product.
Activational and immunosuppressive parameters of MAIDS pathogenesis are not induced by the BM5defORF2aAUGmut or BM5defORF2bAUGmut mutants. To assess the pathogenicity of BM5def viruses mutated to inhibit translation from their alternative ORF2 AUGs, B6 mice were infected with rescued preparations carrying BM5eco virus and either the BM5defORF2aAUGmut or BM5defORF2bAUGmut mutant virus or with positive (LP-BM5 and wt BM5def rescued with BM5eco) or negative (phosphate-buffered saline) controls. Splenomegaly, as a standard activational parameter characteristic of MAIDS (10, 11), was determined 11 weeks postinfection (wpi) by spleen weight (Fig. 4A). At 5, 9, and 11 wpi, serum IgG2a and IgM levels were determined (11 wpi; Fig. 4B), as a measure of the generalized hyper-Ig of MAIDS. Compared to the LP-BM5 or wt BM5def rescued control-infected mice, B6 mice infected with either the BM5defORF2aAUGmut (P < 0.001) or BM5defORF2bAUGmut (P < 0.001) virus did not induce measurable splenomegaly or hypergammaglobulinemia. This pattern was replicated in three additional independent experiments.
In addition, a standard panel of immunosuppressive parameters was measured at termination 11 wpi: responses to mitogens by T (ConA) versus B (lipopolysaccharide) cells, as well as the generation of alloantigen-specific (anti-H-2d) CTL activity. Again, in contrast to LP-BM5 or wt BM5def rescued control-infected mice, B6 mice infected with the BM5eco-rescued BM5defORF2aAUGmut or BM5defORF2bAUGmut mutants did not develop disease, i.e., they had normal mitogen responses (Fig. 4C; data representative of four independent experiments, P < 0.001) and generation of CTL activity (Fig. 4D; data representative of three independent experiments). These results were further confirmed by flow cytometric analyses for the characteristic spleen cell phenotypic changes that are observed in uninfected versus LP-BM5-infected B6 mice (9, 14). LP-BM5 or wt BM5def rescued control-infected mice showed the expected expanded populations of CD19+ CD23– and CD4+ Thy1.2– spleen cells, but these changes were not observed in mice infected with the BM5defORF2aAUGmut or BM5defORF2bAUGmut mutant (data not shown).
Determination of BM5def and BM5eco gag RNA levels in infected mice. On termination of each experiment, we verified the presence of the appropriate control and experimental viruses in infected mice and quantified their expression by qRT-PCR assays for BM5eco and BM5def (7). Although there was no evidence of MAIDS induction by the BM5defORF2aAUGmut or BM5defORF2bAUGmut preps, mice infected with these mutants expressed substantial levels of both BM5eco and BM5def (Fig. 5). BM5def gag RNA was more abundant than BM5eco gag RNA in spleen cells from both control and viral mutant-infected mice. These data are similar to what our lab has observed previously (7) as well as to BM5def and BM5eco gag RNA levels determined in the viral stocks that were generated in vitro and used for infecting mice in this study (data not shown). BM5eco gag and BM5def gag RNA levels from mice infected with the BM5defORF2aAUGmut or BM5defORF2bAUGmut mutant viruses were only about two- to threefold lower than the viral loads for wt rescued control and LP-BM5-infected mice (Fig. 5). It seems very unlikely that this degree of difference in viral load would be sufficient to account for the absence of disease induction after infection by the mutated BM5def preparations. The higher viral loads in the positive controls, in this case infection with nonmutated viral preps, seem to simply reflect the fact that in full-blown MAIDS the characteristic extensive lymphoproliferation is a major contributor to the spread of the virus, in addition to multiple rounds of infection by free virions. This concept that B6 background mouse strains resistant to MAIDS pathogenesis still exhibit a substantial viral load indicative of normal initial infection and early viral spread has been further supported by both viral titration experiments and the study of several MAIDS-resistant versus -susceptible B6 strains, with either of the genes knocked out (CD154, CD40, CD80/86) or as chimeric CD40 transgenics (9, 10, 11). As we have previously discussed in some detail (9), these comparisons across mouse strains have shown that distinctly measurable disease can be observed under conditions that result in a few-fold lower BM5def viral load at termination of the experiment. In addition, on termination of each experiment, by using RT-PCR we isolated the BM5def viruses from spleen cells and sequenced them. Sequencing data confirmed the presence of only the wt BM5def in the mice infected with the positive control viral preps, versus only the appropriate mutant sequence in the mice infected with the experimental viruses. An inherent advantage of this system is that any reversion of the mutated to the wt sequence would have resulted in a defective virus that would have caused disease. These data clearly suggest that a virus different from what was initially injected into the mice had not arisen to detectable levels as a result of reversion during the 11-week course of infection.
In summary, based on (i) both activational and immunodeficiency parameters of MAIDS (Fig. 4), (ii) viral load data (Fig. 5), and (iii) sequence analysis of viruses isolated from mice 11 wpi, and ORF1/ORF2 expression data (Fig. 2 and 3), it is evident that the viral mixes consisting of either wt BM5eco-rescued ORF2aAUGmut or ORF2bAUGmut BM5def virus, each with intact wt ORF1 expression capability, are essentially unable to induce disease. These data strongly suggest the existence of a novel ORF2 product(s) that is required for LP-BM5-induced pathogenesis. The ORF2 product(s) may be initiated from either the ORF2a or ORF2b AUG (Fig. 2), as disruption of either one of their translational potentials leads to inhibition of disease (Fig. 4). One interpretation of these findings is that there is a requirement for both protein products of 193 and 30 aa, produced as separate entities from the ORF2a and ORF2b AUGs, respectively. These proteins may be independently essential in targeting different specific host cellular proteins. Alternatively or in addition, the two alternative ORF2 translational products may function together to cause disease by independently interacting with the same host cell protein target or combining to form a single ORF2a/ORF2b heterodimer.
As an option to two separate ORF2 protein products, it is also possible that only a single protein initiating from the ORF2a AUG is required. Thus, the data from the BM5defORF2bAUGmutEGFP mutant virus (Fig. 2C) show that, despite disrupting the initiation AUG of ORF2b, GFP is still expressed (albeit at a reduced level) as a fusion protein, apparently from the unaltered ORF2a AUG which is in the same +1-nt reading frame as ORF2b (Fig. 2B). Importantly, however, the BM5defORF2bAUGmut virus is unable to cause disease. This finding could be explained based not on the ORF2b AUG mutation as a block to translation initiation at this site but rather on the basis of the introduction of a missense mutation at the ORF2b site in the ORF2a AUG-initiated maximal ORF2 protein. Thus, the presence of threonine instead of methionine may be deleterious to the structure of the extended protein produced from the ORF2a AUG.
The results presented in this paper support the existence of an alternative translation initiation mechanism in BM5def for the expression of an ORF2-encoded gag product(s) and its requirement in LP-BM5-induced pathogenesis. These data have obvious implications for pursuing the molecular mechanisms by which BM5def causes disease and studies are currently under way to identify the nature and function of the necessary ORF2 product(s). Our data are interesting to compare to the results of other labs that have also studied atypically encoded CTL epitopes, including in human immunodeficiency virus (2, 3, 6, 28). In particular, and comparable to our novel BM5def ORF2a and/or ORF2b AUG(s)-encoded 193 or 30 aa protein(s), a previously unobserved, conserved 87-residue protein encoded by influenza A virus has recently been identified by first defining a CTL epitope encoded in an alternative translational ORF (6). This immunogenic influenza virus protein, PB1-F2, recognized by CD8+ T cells, is derived from a +1 alternative reading frame and may play a critical role in pathogenesis by inducing apoptosis in host immune cells that respond to influenza virus infection (6). Our identification of an immunodominant CTL epitope novelly expressed from an alternative translational gag ORF (20, 22), combined with our present data that the ORF2-encoded extended gag product(s) have a functional role in disease induction, raise the possibility that alternative reading frame-derived proteins may also be crucial in other disease models.
ACKNOWLEDGMENTS
We especially thank On Ho for her contributions in determining disease parameters as well as helpful discussions during the course of these experiments. We also thank Amanda Birdsey for generating the BM5defORF2bAUGmutEGFP construct and the stable line expressing this construct and Kathy Green for generating the LP-BM5 viral stocks as well as for helpful discussions.
A.G. was supported by NIH Institutional Training Grant T32 A017363. This work was supported in part by U.S. Public Health Service grant CA50157. The flow cytometer was the generous gift of the Fannie E. Rippel Foundation and is partially supported by the core grant of the Norris Cotton Cancer Center (CA23108).
REFERENCES
Aziz, D. C., Z. Hanna, and P. Jolicoeur. 1989. Severe immunodeficiency disease induced by a defective murine leukemia virus. Nature 338:505-508.
Bullock, T. N., and L. C. Eisenlohr. 1996. Ribosomal scanning past the primary initiation codon as a mechanism for expression of CTL epitopes encoded in alternative reading frames. J. Exp. Med. 184:1319-1329.
Cardinaud, S., A. Moris, M. Fevrier, P. S. Rohrlich, L. Weiss, P. Langlade-Demoyen, F. A. Lemonnier, O. Schwartz, and A. Habel. 2004. Identification of cryptic MHC I-restricted epitopes encoded by HIV-1 alternative reading frames. J. Exp. Med. 199:1053-1063.
Chattopadhyay, S. K., H. C. Morse III, M. Makino, S. K. Ruscetti, and J. W. Hartley. 1989. Defective virus is associated with induction of murine retrovirus-induced immunodeficiency syndrome. Proc. Natl. Acad. Sci. USA 86:3862-3869.
Chattopadhyay, S. K., D. N. Sengupta, T. N Fredrickson, H. C. Morse III, and J. W. Hartley. 1991. Characteristics and contributions of defective, ecotropic, and mink cell focus-inducing viruses involved in a retrovirus-induced immunodeficiency syndrome of mice. J. Virol. 65:4232-4241.
Chen, W., P. A. Calvo, D. Malide, J. Gibbs, U. Schubert, I. Bacik, S. Basta, R. O'Neill, J. Schickli, P. Palese, P. Henklein, J. R. Bennink, and J. W. Yewdell. 2001. A novel influenza A virus mitochondrial protein that induces cell death. Nat. Med. 7:1306-1312.
Cook, W. J., K. A. Green, J. J. Obar, and W. R. Green. 2003. Quantitative analysis of LP-BM5 murine leukemia retrovirus RNA using real-time RT-PCR. J. Virol. Methods 108:49-58.
Fedorov, A., S. Saxonov, and W. Gilbert. 2002. Regularities of context-dependent codon bias in eukaryotic genes. Nucleic Acids Res. 30:1192-1197.
Green, K. A., C. L. Ahonen, W. J. Cook, and W. R. Green. 2004. CD40-associated TRAF6 signaling is required for disease induction in a retrovirus-induced murine immunodeficiency. J. Virol. 78:6055-6060.
Green, K. A., W. J. Cook, A. H. Sharpe, and W. R. Green. 2002. The CD154/CD40 interaction required for retrovirus-induced murine immunodeficiency syndrome is not mediated by upregulation of the CD80/CD86 costimulatory molecules. J. Virol. 76:13106-13110.
Green, K. A., R. J. Noelle, B. G. Durell, and W. R. Green. 2001. Characterization of the CD154-positive and CD40-positive cellular subsets required for pathogenesis in retrovirus-induced murine immunodeficiency. J. Virol. 75:3581-3589.
Huang, M., and P. Jolicoeur. 1994. Myristylation of Pr60gag of the murine AIDS-defective virus is required to induce disease and notably for the expansion of its target cells. J. Virol. 68:5648-5655.
Huang, M., C. Simard, and P. Jolicoeur. 1989. Immunodeficiency and clonal growth of target cells induced by helper-free defective retrovirus. Science 246:1614-1617.
Knoetig, S. M., T. A. Torrey, Z. Naghashfar, T. McCarty, and H. C. Morse III. 2002. CD19 signaling pathways play a major role for murine AIDS induction and progression. J. Immunol. 169:5607-5610.
Kozak, M. 1989. Context effects and inefficient initiation at non-AUG codons in eucaryotic cell-free translation systems. Mol. Cell. Biol. 9:5073-5080.
Kubo, Y., K. Kakimi, K. Higo, H. Kobayashi, T. Ono, Y. Iwama, K. Kuribayashi, H. Hiai, A. Adachi, and A. Ishimoto. 1996. Possible origin of murine AIDS (MAIDS) virus: conversion of an endogenous retroviral p12gag sequence to a MAIDS-inducing sequence by frameshift mutations. J. Virol. 70:6405-6409.
Kubo, Y., K. Kakimi, K. Higo, L. Wang, H. Kobayashi, K. Kuribayashi, T. Masuda, T. Hirama, and A. Ishimoto. 1994. The p15gag and p12gag regions are both necessary for the pathogenicity of the murine AIDS virus. J. Virol. 68:5532-5537.
Laterjet, R., and J. F. Duplan. 1962. Experiment and discussion on leukemogenesis by cell-free extracts of radiation-induced leukemia in mice. Int. J. Radiat. Biol. 5:223-233.
Makino, M., S. K. Chattopadhyay, J. W. Hartley, and H. C. Morse III. 1992. Analysis of role of CD8+ T cells in resistance to murine AIDS in A/J mice. J. Immunol. 149:1702-1706.
Mayrand, S. M., P. A. Healy, B. E Torbett, and W. R. Green. 2000. Anti-Gag cytolytic T lymphocytes specific for an alternative translational reading frame-derived epitope and resistance versus susceptibility to retrovirus-induced murine AIDS in F1 mice. Virology 272:438-449.
Mayrand, S. M., and W. R. Green. 1998. Non-traditionally derived CTL epitopes: exceptions that prove the rules? Immunol. Today 19:551-556.
Mayrand, S. M., D. A. Schwarz, and W. R. Green. 1998. An alternative translational reading frame encodes an immunodominant retroviral CTL determinant expressed by an immunodeficiency-causing retrovirus. J Immunol. 160:39-50.
Pavlovitch, J., E. Hulier, M. Rizk-Rabin, M. Marussig, D. Mazier, M. Joffret, S. Hoos, and M. Papiernik. 1996. Resistance to murine AIDS in offspring of mice infected with LP-BM5: role of CD8 T cells. J. Immunol. 156:4757-4765.
Pozsgay, J. M., M. W. Beilharz, B. D. Wines, A. D. Hess, and P. M. Pitha. 1993. The MA (p15) and p12 regions of the gag gene are sufficient for the pathogenicity of the murine AIDS virus. J. Virol. 67:5989-5999.
Rowe, W. P., W. E. Pugh, and J. W. Hartley. 1970. Plaque assay techniques for murine leukemia viruses. Virology 42:1136-1139.
Schwarz, D. A., and W. R. Green. 1994. CTL responses to the gag polyprotein encoded by the murine AIDS defective retrovirus are strain dependent. J. Immunol. 153:436-442.
Tang, Y., A. W. Hugin, N. A. Giese, L. Gabriele, S. K Chattopadhyay, T. N. Fredrickson, D. Kagi, J. W. Hartley, and H. C. Morse III. 1997. Control of immunodeficiency and lymphoproliferation in mouse AIDS: studies of mice deficient in CD8+ T cells or perforin. J. Virol. 71:1808-1815.
Vassilaki, N., and P. Mavromara. 2003. Two alternative translation mechanisms are responsible for the expression of the HCV ARFP/F/core+1 coding open reading frame. J. Biol. Chem. 278:40503-40513.(Arti Gaur and William R. )