Translational recoding signals between gag and pol in diverse LTR retrotransposons

Because of their compact genomes, retroelements (including retrotransposons and retroviruses) employ a variety of translational recoding mechanisms to express Gag and Pol. To assess the diversity of recoding strategies, we surveyed gag/pol gene organization among retroelements from diverse host species, including elements exhaustively recovered from the genome sequences of Caenorhabditis elegans, Drosophila melanogaster, Schizosaccharomyces pombe, Candida albicans, and Arabidopsis thaliana. In contrast to the retroviruses, which typically encode pol in the -1 frame relative to gag, nearly half of the retroelements surveyed encode a single gag-pol open reading frame. This was particularly true for the Ty1/copia group retroelements. Most animal Ty3/gypsy retroelements, on the other hand, encode gag and pol in separate reading frames, and likely express Pol through +1 or -1 frameshifting. Conserved sequences conforming to slippery sites that specify viral ribosomal frameshifting were identified among retroelements with pol in the -1 frame. None of the plant retroelements encoded pol in the -1 frame relative to gag; however, two closely related plant Ty3/gypsy elements encode pol in the +1 frame. Interestingly, a group of plant Ty1/copia retroelements encode pol either in a +1 frame relative to gag or in two nonoverlapping reading frames. These retroelements have a conserved stem-loop at the end of gag, and likely express pol either by a novel means of internal ribosomal entry or by a bypass mechanism.     

Nucleotide sequence of AKV murine leukemia virus.

AKV is an endogenous, ecotropic murine leukemia virus that serves as one of the parents of the recombinant; oncogenic mink cell focus-forming viruses that arise in preleukemic AKR mice. I report the 8,374-nucleotide-long sequence of AKV, as determined from the infectious molecular clone AKR-623. The 5'-leader sequence of AKV extends to nucleotide 639, after which lies a long open reading frame encoding the gag and pol gene products. The reading frame is interrupted by a single amber codon separating the gag and pol genes. The pol gene overlaps the env gene within the 3' region of the AKV genome. The nucleotide sequence of the 5' region of AKV reveals the following features. (i) The 5'-leader sequence lacks any AUG codon to initiate translation of gPr80gag, suggesting that gPr80gag is not required for the replication of AKV. (ii) A short portion of the leader region diverges in sequence from the closely related Moloney murine leukemia virus and appears to be related to a sequence highly repeated in eucaryotic genomes. (iii) As in Moloney murine leukemia virus, there is a potential RNA secondary structure flanking the amber codon that separates the gag and pol genes. This structure might function as a regulatory protein binding site that controls the relative levels of synthesis of the gag and pol precursors. The nucleotide sequence of the 3' region of AKV is compared with sequences reported previously from both infectious and noninfectious molecular clones of AKV.     

Protein-coding potential of mouse mammary tumor virus genome RNA as examined by in vitro translation.

The protein-coding capacity of the mouse mammary tumor virus genome has been examined by in vitro translation of genome length and polyadenylated subgenomic fragments of viral RNA. Intact genome RNA of about 35S programmed synthesis of the Pr77gag, Pr110gag and Pr160gag/pol precursors seen in infected cells in vivo. Polyadenylated RNA fragments of 18 to 28S encoded products whose tryptic peptide maps resembled those of the nonglycosylated precursor to the envelope glycoproteins, confirming the gene order 5'-gag-pol-env-3'. Translation of polyadenylated RNA fragments smaller than 18S yielded a series of related proteins whose peptide maps bore no resemblance to any of the virion structural proteins. Thus, a region of the mouse mammary tumor virus genome distal to the env gene appears to have an open reading frame sufficient to encode at least 36,000 daltons of protein as of yet unknown function.     

A viable mutation in cauliflower mosaic virus, a retroviruslike plant virus, separates its capsid protein and polymerase genes.

A viable strain of cauliflower mosaic virus is described which arose by illegitimate recombination of two lethal parents. In this strain, the normally overlapping open reading frames IV and V, corresponding to the retrovirus gag and pol genes, are separated by a short intergenic region, suggesting that in this virus and in contrast to retroviruses, fusion of gag and pol gene products is not obligatory.     

Identification of a protease gene of human T-cell leukemia virus type I (HTLV-I) and its structural comparison

We determined the nucleotide sequence of a region between the gag and pol genes of a replication-competent proviral clone of a human T-cell leukemia virus type I (HTLV-I) from MT-2 cells. This region overlapping the gag and pol genes contains an open reading frame with a different phase from others. The deduced amino acid sequences show significant homology with the known protease gene of other retroviruses, and harbors highly conserved amino acid sequences that are well conserved in other retroviral protease domains. These results indicate that this open reading frame encodes a HTLV-I protease.     

Identification of a potential protease-coding gene in the genomes of bovine leukemia and human T-cell leukemia viruses

The genomes of bovine leukemia and human T-cell leukemia viruses both contain an unidentified region between the gag and pol genes. These regions harbor an open reading frame that is in a different phase from the reading frames of the gag and pol genes. Based on the deduced amino acid sequences, we show here that they potentially encode a gag precursor-cleaving protease, which is known to be fused to the gag and pol products of avian and murine retroviruses, respectively. This finding raises the interesting question of the expression and evolution of retroviral genes.     

Mutational analysis of human immunodeficiency virus type 1 protease suggests functional homology with aspartic proteinases.

Processing of the retroviral gag and pol gene products is mediated by a viral protease. Bacterial expression systems have been developed which permit genetic analysis of the human immunodeficiency virus type 1 protease as measured by cleavage of the pol protein precursor. Deletion analysis of the pol reading frame locates the sequences required to encode a protein with appropriate proteolytic activity near the left end of the pol reading frame but largely outside the gag-pol overlap region, which is at the extreme left end of pol. Most missense mutations within an 11-amino-acid domain highly conserved among retroviral proteases and with sequence similarity to the active site of aspartic proteinases abolish appropriate processing, suggesting that the retrovirus proteases share a catalytic mechanism with aspartic proteinases. Substitution of the amino acids flanking the scissile bond at three of the processing sites encoded by pol demonstrates distinct sequence requirements for cleavage at these different sites. The inclusion of a charged amino acid at the processing site blocks cleavage. A subset of these substitutions also inhibits processing at the nonmutated sites.     

Characterization of ribosomal frameshifting for expression of pol gene products of human T-cell leukemia virus type I.

For study of the pol gene expression of human T-cell leukemia virus type I (HTLV-I), RNA was transcribed in vitro from proviral DNA and translated in rabbit reticulocyte lysates. This cell-free translation resulted in two major translation products representing the Gag and Gag-Pro polyproteins. By contrast, the Gag-Pro-Pol polyprotein could be readily observed only when translation was performed with mutant mRNA in which the protease (pro) reading frame was aligned to gag to eliminate the frameshifting event in the gag-pro overlap. The results indicated that two independent ribosomal frameshifting events are required for expression of the HTLV-I pol gene product. Studies with mutant DNAs facilitated the characterization of the primary structure of the HTLV-I mRNA responsible for the ribosomal frameshift in the pro-pol overlap and demonstrated that the frameshift occurs at the signal sequence UUUAAAC. Direct amino acid sequencing of the transframe protein localized the site of the frameshift to the asparagine codon AAC.     

Nucleotide sequence of the gag gene and gag-pol junction of feline leukemia virus.

The nucleotide sequence of the gag gene of feline leukemia virus and its flanking sequences were determined and compared with the corresponding sequences of two strains of feline sarcoma virus and with that of the Moloney strain of murine leukemia virus. A high degree of nucleotide sequence homology between the feline leukemia virus and murine leukemia virus gag genes was observed, suggesting that retroviruses of domestic cats and laboratory mice have a common, proximal evolutionary progenitor. The predicted structure of the complete feline leukemia virus gag gene precursor suggests that the translation of nonglycosylated and glycosylated gag gene polypeptides is initiated at two different AUG codons. These initiator codons fall in the same reading frame and are separated by a 222-base-pair segment which encodes an amino terminal signal peptide. The nucleotide sequence predicts the order of amino acids in each of the individual gag-coded proteins (p15, p12, p30, p10), all of which derive from the gag gene precursor. Stable stem-and-loop secondary structures are proposed for two regions of viral RNA. The first falls within sequences at the 5' end of the viral genome, together with adjacent palindromic sequences which may play a role in dimer linkage of RNA subunits. The second includes coding sequences at the gag-pol junction and is proposed to be involved in translation of the pol gene product. Sequence analysis of the latter region shows that the gag and pol genes are translated in different reading frames. Classical consensus splice donor and acceptor sequences could not be localized to regions which would permit synthesis of the expected gag-pol precursor protein. Alternatively, we suggest that the pol gene product (RNA-dependent DNA polymerase) could be translated by a frameshift suppressing mechanism which could involve cleavage modification of stems and loops in a manner similar to that observed in tRNA processing.     

Requirement of the Pr55gag precursor for incorporation of the Vpr product into human immunodeficiency virus type 1 viral particles.

The human immunodeficiency virus type 1 (HIV-1) particles consists of two molecules of genomic RNA as well as molecules originating from gag, pol, and env products, all synthesized as precursor proteins. The 96-amino-acid Vpr protein, the only virion-associated HIV-1 regulatory protein, is not part of the virus polyprotein precursors, and its incorporation into virus particles must occur by way of an interaction with a component normally found in virions. To investigate the mechanism of incorporation of Vpr into the HIV-1 virion, Vpr- proviral DNA constructs harboring mutations or deletions in specific virion-associated gene products were cotransfected with Vpr expressor plasmids in COS cells. Virus released from the transfected cells was tested for the presence of Vpr by immunoprecipitation with Vpr-specific antibodies. The results of these experiments show that Vpr is trans-incorporated into virions but at a lower efficiency than when Vpr is expressed from a proviral construct. The minimal viral genetic information necessary for Vpr incorporation was a deleted provirus encoding only the pr55gag polyprotein precursor. Incorporation of Vpr requires the expression but not the processing of gag products and is independent of pol and env expression. Direct interaction of Vpr with the Pr55gag precursor protein was demonstrated by coprecipitation experiments with gag product-specific antibodies. Overall, these results indicate that HIV-1 Vpr is incorporated into the nascent virion through an interaction with the Gag precursor polyprotein and demonstrate a novel mechanism by which viral protein can be incorporated into virus particles.     

Murine mammary tumor virus pol-related sequences in human DNA: characterization and sequence comparison with the complete murine mammary tumor virus pol gene.

Sequences in the human genome with homology to the murine mammary tumor virus (MMTV) pol gene were isolated from a human phage library. Ten clones with extensive pol homology were shown to define five separate loci. These loci share common sequences immediately adjacent to the pol-like segments and, in addition, contain a related repeat element which bounds this region. This organization is suggestive of a proviral structure. We estimate that the human genome contains 30 to 40 copies of these pol-related sequences. The pol region of one of the cloned segments (HM16) and the complete MMTV pol gene were sequenced and compared. The nucleotide homology between these pol sequences is 52% and is concentrated in the terminal regions. The MMTV pol gene contains a single long open reading frame encoding 899 amino acids and is demarcated from the partially overlapping putative gag gene by termination codons and a shift in translational reading frame. The pol sequence of HM16 is multiply terminated but does contain open reading frames which encode 370, 105, and 112 amino acid residues in separate reading frames. We deduced a composite pol protein sequence for HM16 by aligning it to the MMTV pol gene and then compared these sequences with other retroviral pol protein sequences. Conserved sequences occur in both the amino and carboxyl regions which lie within the polymerase and endonuclease domains of pol, respectively.     

Two autonomous myc oncogenes in avian carcinoma virus OK10.

The oncogenic avian retrovirus OK10 has the genetic structure gag-delta pol-myc-delta-env. The myc sequence is transduced from a cellular gene, proto-myc, while gag, pol, and env are essential retrovirus genes. By analogy with other directly oncogenic retroviruses, the specific myc sequence of OK10 is thought to be essential for transforming function. However, unlike the specific sequences of all other transforming retroviruses that encode unique transforming proteins, the myc sequence of OK10 encodes two potential transforming proteins, p58 and p200. p200 is translated from the gag-delta pol-myc region of genomic RNA, while p58 is thought to be translated from the gag leader and the open reading frame of myc via a subgenomic mRNA. In this paper, we ask whether both myc genes of OK10 are autonomous transforming genes. By differentially inactivating the p200 myc gene of OK10 provirus in vitro and analyzing transforming function in quail embryo cells, it was found that mutants expressing only p58 transformed like wild-type OK10. Further, it was shown that p58 with and without the gag leader had transforming function and that p58 of wild-type OK10 is initiated from the gag leader. Mutants expressing only p200 were also transforming but less efficiently than mutants that express only p58. A mutant OK10 virus in which the native frameshift of retroviruses between gag and pol was deleted expressed a shortened p200 (delta p200). Although this virus expressed more delta p200 than wild-type OK10 did, it transformed cells less efficiently. It follows that OK10 expresses two autonomous transforming genes, which is unique among retroviruses with onc genes.     

Autoprocessing of Drosophila copia gag precursor to generate a unique laminate structure in Escherichia coli.

Drosophila copia protease is likely to be encoded in the gag gene. We have expressed copia gag polyprotein precursor in E. coli. The gag precursor was correctly processed to generate a unique laminate structure in E. coli. The processing was almost completely blocked by a mutation at the putative active site of copia protease, and resulted in accumulation of the precursor. Furthermore, the laminate structure was not found in E. coli expressing the mutant precursor. These results indicate that the protease is involved in cleaving the gag precursor itself. Also, the assembly of copia gag protein should correlate to the autoprocessing of copia gag polyprotein precursor.     

Identification of posttranscriptionally active inhibitory sequences in human immunodeficiency virus type 1 RNA: novel level of gene regulation.

cis-acting inhibitory region (IR) sequences were identified within the gag/pol gene of the human immunodeficiency virus type 1 (HIV-1) by using a novel feedback-stimulated, rev-independent tat reporter gene to screen HIV-1 sequences in transient expression assays. Two regions, a 1,295-nucleotide segment in the gag gene (IR-1) and a 1,932-nucleotide segment of the pol gene (IR-2), each inhibited reporter gene expression 10- to 20-fold. IR-1 and IR-2 both contained subsequences which inhibited reporter gene expression. Introduction of IR sequences into a heterologous reporter plasmid, pCMV-CAT, resulted in decreased chloramphenicol acetyltransferase expression, suggesting that the inhibitory effect was not restricted to a reporter gene under the control of the HIV-1 promoter. The presence of HIV IR sequences in cis did not alter relative levels of reporter gene RNA; however, fractionation studies revealed IR-containing RNA accumulated in the nucleus. These findings demonstrate that IR sequences within the gag/pol region affect gene expression by altering the cellular distribution of viral RNA.