Ref(2)P has been described as one of the Drosophila proteins that interacts with the sigma virus cycle. We generated alleles to identify critical residues involved in the restrictive (inhibiting viral multiplication) or permissive (allowing viral multiplication) character of Ref(2)P. We demonstrate that permissive alleles increase the ability of the sigma virus to infect Drosophila when compared to null alleles and we confirm that restrictive alleles decrease this capacity. Moreover, we have created alleles unfunctional in viral cycling while functional for Ref(2)P fly functions. This type of allele had never been observed before and shows that fly- and virus-related activities of Ref(2)P are separable. The viral status of Ref(2)P variants is determined by the amino-terminal PB1 domain polymorphism. In addition, an isolated PB1 domain mimics virus-related functions even if it is similar to a loss of function toward fly-related activities. The evolutionary tree of the Ref(2)P PB1 domain that we could build on the basis of the natural allele sequences is in agreement with an evolution of PB1 domain due to successive transient selection waves.

ABSTRACT The UL25 gene of pseudorabies virus (PrV) can encode a protein of about 57 kDa which is well conserved among herpesviruses. The UL25 protein of herpes simplex virus type 1 is a capsid constituent involved in virus penetration and capsid maturation. To identify and characterize the UL25 gene product of PrV, polyclonal mouse anti-UL25 antibodies were raised to a bacterially expressed fusion protein. In immunoblotting and immunoprecipitation assays of PrV-infected cell lysates, these anti-UL25 antisera specifically recognized a protein of the expected size with late expression kinetics. This 57-kDa product was also present in purified virions and was found to be associated with all types of capsids. Synthesis of a protein migrating at the same size point was directed from the eukaryotic expression plasmid pCG-UL25. To determine the subcellular localization of UL25, immunofluorescence studies with anti-UL25 antisera were performed on Nonidet P-40-extracted COS-7 cells infected with PrV or transfected with pCG-UL25. In PrV-infected cells, newly synthesized UL25 is directed mainly to distinct nuclear compartments, whereas UL25 expressed in the absence of other viral proteins is distributed more uniformly in the nucleus and colocalizes also with microtubules. To study the fate of UL25 at very early stages of infection, immunofluorescence experiments were performed on invading PrV particles in the presence or absence of drugs that specifically depolymerize components of the cytoskeleton. We found that the incoming nucleocapsids colocalize with microtubules during their transport to the nucleus and that UL25 remains associated with nucleocapsids during this transport.

The genome of pseudorabies virus (PrV) is collinear with the herpes simplex virus type 1 (HSV1) genome, except for an inversion in the unique long region, the right extremity of which resides within the BamHI fragment 9 and the left within the BamHI fragment 1. We previously sequenced the right border of the inversion which is situated next to the UL44-gC gene and found that it encodes the UL24, UL25, UL26 and UL26.5 gene counterparts of HSV1. We have now sequenced 5317 base pairs of the BamHI fragment 1, upstream of the UL27-gB gene. We found two open reading frames homologous to UL46 and UL47 of HSV1 yet UL45 was absent and replaced by a set of strictly repeated sequences. PrV UL46 and UL47 are transcribed into two 3′ co-terminal messenger RNAs with early and late kinetics, respectively. Comparison of the PrV UL46 and UL47 protein sequences with their counterparts from alphaherpesviruses indicated a strong similarity. The genome is rearranged in this region with respect to HSV1 and the inversion must have taken place, on the left side, within the UL46-UL27 intergenic region. Thus, the inversion should include genes UL27 to UL44.

The genomes of pseudorabies virus (PrV) and of herpes simplex virus type 1 (HSV1) are colinear, excepting an inversion in the unique long region, of which one extremity resides within the BamHI fragment 9. This fragment (4088 bp) encodes the counterparts of HSV1 UL24, UL25, UL26 and UL26.5 that are transcribed into four 3′-coterminal mRNAs. Multiple alignments of UL24, UL25 and UL26 protein homologs from α-, β-, and γ-herpesviruses were performed. The PrV UL24 protein is shorter than its counterparts, missing the non-conserved COOH-terminal region. The region which is common to all viruses contains a basic NH2-terminus and a hydrophobic COOH-end, suggesting that UL24 may function as a matrix protein. The UL25 proteins are well conserved, particularly among the α-herpesviruses. All the domains involved in the proteolytic activity of the UL26 protein are highly conserved, as well as the two cleavage sites. Thus, its function and processing may be similar in PrV as in other herpesviruses. Due to the fact that in PrV the UL26 and UL44 genes are adjacent and their ends are conserved, the right border of the inversion must lie within their intergenic region.

The nucleotide sequence of the genes 2 and 3 of theDrosophilarhabdovirus sigma was determined from cDNAs to viral genome and poly(A)+mRNAs. Gene 2 comprises 1032 nucleotides and contains a long ORF encoding a molecular weight 35,208 polypeptide present in infected cells and in virions which migrates in SDS–PAGE as a doublet ofMrabout 60 kDa. The distribution of acidic charges as well as the electrophoretic properties of the protein are characteristic of the rhabdovirus P proteins. Gene 3 comprises 923 nucleotides and contains a long ORF capable of coding a polypeptide of 298 amino acids of MW 33,790. The putative protein (PP3) is similar in size to a minor component of the virions. Computer analysis shows that the sequence of PP3 contains three motifs related to the conserved motifs of reverse transcriptases.

The genome of the sigma rhabdovirus of Drosophila melanogaster consists of six genes in the order 3' N-2-3-4-G-L 5'. The nucleotide sequences of the N and of the fourth genes were determined from cDNA clones. Each gene contained a single long open reading frame encoding polypeptides with predicted MW of 50 and 25 kDa, respectively. Evidence that these genes encode the nucleocapsid N and the matrix M proteins were obtained using antibodies raised against recombinant proteins derived from the cloned genes and expressed in Escherichia coli. The M and N predicted amino acid sequences were compared with those of other rhabdoviruses. The M protein of sigma virus shared a similar domain arrangement to the other M proteins, but it showed very little sequence conservation. The N protein of sigma virus showed no significant homology with its counterpart in SYNV or IHNV and VSHV; it did, however, show sequence homology with the N of four vesiculoviruses and two lyssaviruses. The extent of amino acid identity suggests that sigma virus occupies an intermediate evolutionary position between these two genera. Some conserved motifs in the M and N proteins were deduced from the comparisons.

The ref(2)P gene of Drosophila melanogaster was identified by the discovery of two alleles, Po and Pp, respectively, permissive and restrictive for sigma rhabdovirus multiplication. A surprising variability of this gene was first noticed by the observation of size differences between the transcripts of permissive and restrictive alleles. In this paper, another restrictive allele, Pn, clearly distinct from Pp, is described: it exhibits a weaker antiviral effect than Pp and differs from Pp by its molecular structure. Five types of alleles were distinguished on the basis of their molecular structure, as revealed by S1 nuclease analysis of 17 D. melanogaster strains; three alleles were permissive and two restrictive. Comparison of the sequences of four haplotypes revealed numerous point mutations, two deletions (21 and 24 bp) and a complex event involving a 3-bp deletion, all affected the coding region. The unusual variability of the ref(2)P locus was confirmed by the high ratio of amino acid replacements to synonymous mutations (7:1), as compared to that of other genes, such as the Adh (2:42). Nevertheless, nucleotide sequence comparison with the Drosophila erecta ref(2)P gene shows that selective pressures are exerted to maintain the existence of a functional protein. The effects of this high variability on the ref(2)P protein are discussed in relation to its specific antiviral properties and to its function in D. melanogaster, where it is required for male fertility.

The sequence of the sigma virus glycoprotein gene has been reported. We report here the cloning of the cDNAs to four other genes, their physical map on the genome, the sequencing of the intergenic regions, and the determination of the gene junction signals. This analysis reveals unusual traits for a rhabdovirus: four genes map upstream of the glycoprotein gene and the glycoprotein gene overlaps the upstream gene by 33 nucleotides. Northern analysis did not show bicistronic transcripts of the overlapping genes, whereas it revealed bicistronic transcripts of two genes which are separated by a stretch of six nucleotides. This leads us to reconsider the significance of rhabdovirus consensus gene-end sequences and to integrate these data into the models proposed for polymerase functioning.

Vesicular stomatitis virus (VSV) establishes a non-cytopathic persistent infection in Drosophila melanogaster cells. The synthesis of the viral glycoprotein G was specifically inhibited during a post-transcriptional step, whereas the synthesis and turnover of its mRNA were not modified compared with the other viral mRNAs. Another viral glycoprotein, migrating slightly faster than G protein on an SDS-polyacrylamide gel, was detected in infected Drosophila cells. This protein showed most of the characteristics of the intracellular Gs protein found in infected vertebrate cells. The amounts of G protein integrated into mature virions and of soluble Gs protein secreted into the culture medium were reduced greatly during VSV infection in Drosophila cells.

The ref(2)P gene of Drosophila melanogaster is implicated in sigma rhabdovirus multiplication. A permissive allele was cloned and sequenced. The structural gene (3.1 kbp) is divided into three exons. The mRNAs are heterogeneous in size. They differ only in the 5′ end of the first exon. The sequence upstream of the short mRNAs contains classical promoter elements. No TATA and CAAT boxes are appropriately positioned upstream of the initiation sites of the long mRNAs, but several repeats, palindromic sequences and inverted CAAT boxes are present. These observations, together with the tissue‐dependent distribution of short and long transcripts, support the hypothesis of the existence of at least two classes of genuine initiation sites. The long size of the untranslated leader RNA region suggests a control of gene expression at the translation level. The same translation product of 599 amino acids (76.3 kd) is predicted for all mRNAs, but the in vitro translation product migrates in SDS‐PAGE with a higher apparent mol. wt (115‐125 kd). The putative ref(2)P protein contains internal repeats, PEST regions which may be signals for protein degradation, and interesting structural motifs such as zinc finger and amphiphilic helices. These later motifs could be mitochondrial pre‐sequences. The degeneration of mitochondria is observed in the spermatids of sterile male flies homozygous for the loss‐of‐function alleles. The amino acid sequence of the ref(2)P product shows no homology with any known protein from the data banks.

Vesicular stomatitis virus RNA synthesis was investigated during the establishment of persistent infection in Drosophila melanogaster cells. The transcription rate declined as early as 5 h after infection and was strongly inhibited after 7 h, leading to a decrease in viral mRNA levels and in viral protein synthesis rates. Full-length plus-strand antigenomes and minus-strand genomes were detected after a 3-h lag time and accumulated until 15 h after infection. Short encapsidated plus-strand molecules were also generated corresponding to the 5' end of viral defective antigenomes. Assembly and release of virions were not restricted, but their infectivity was extremely reduced. In persistently infected cells, an equilibrium was reached where the level of intracellular genomes maintained was constant and maximal even after the rate of all viral syntheses had decreased. These results are discussed with regard to the establishment of persistent infection.

In cultured Drosophila melanogaster cells, vesicular stomatitis virus (VSV) establishes a persistent, noncytopathic infection. No inhibition of host macromolecular synthesis occurs. We studied the synthesis of VSV plus-strand leader RNA, which may be directly involved in vertebrate host synthesis shut-off. Leader RNA accumulated in Drosophila cell cytoplasm, but in low amounts, it was either free or associated to structures larger than the leader RNA-N protein complexes found in vertebrate cells. Only a few leader RNA copies migrated into the cell nucleus; no increase of this transport was observed at any time during the virus cycle. Viral RNAs complementary to the 3' end of the genome and ranging in size from the leader to several hundred nucleotides were found to accumulate in Drosophila cell cytoplasm. Their synthesis was inhibited in the presence of cycloheximide, which blocks all protein synthesis and VSV replication. Correlation between the absence of VSV cytopathogenicity in Drosophila cells and the lack of leader RNA transport into their nuclei is discussed, as well as the possible relationship between the restriction of viral synthesis and the frequent initiation of an abortive replication step.

Summary The growth cycle of sigma virus in Drosophila melanogaster cells was studied: optimal virus production was reached about 40 h after infection; virus release declined thereafter and then remained approximately constant (carrier state). In the presence of DEAE-dextran during virus adsorption, more cells became infected and sigma virus production was enhanced. Sigma virus was partially purified by a gentle procedure. Five presumptive virus-specific proteins with molecular weights 210K, 68K, 57K, 44K and 25K were observed. The p68 polypeptide was glycosylated and formed the spikes of the virion particles. The nucleocapsid contained a single major protein, p44, and one or two minor proteins (p210 and probably p57); another protein, p25, was more loosely associated with the nucleocapsid. None of these proteins was found to be phosphorylated.

The phosphoproteins of vesicular stomatitis virus released from infected Drosophila melanogaster cells were examined. The membrane (M) protein was more phosphorylated than after multiplication in chicken embryo cells, even in Drosophila cell cytoplasm before its association with cellular membranes. Analysis of phosphopeptides generated after partial proteolysis and of phosphoamino acids obtained after complete acid hydrolysis showed that M phosphorylation was quantitatively and qualitatively changed, while NS protein phosphorylation was only slightly modified.

In cultured Drosophila melanogaster cells, vesicular stomatitis virus (VSV) established a persistent, noncytopathic infection. No inhibition of host protein synthesis occurred even though all cells were initially infected. No defective interfering particles were detected, which would explain the establishment of the carrier state. In studies of the time course of viral protein synthesis in Drosophila cells, N, NS, and M viral polypeptides were readily detected within 1 h of infection. The yield of G protein and one of its precursors; G1, was very low at any time of the virus cycle; the released viruses always contained four to five times less G than those produced by chicken embryo cells, whatever the VSV strain or serotype used for infection and whatever the Drosophila cell line used as host. Actinomycin D added to the cells before infection enhanced VSV growth up to eight times. G and G1 synthesis increased much more than that of the other viral proteins when the cells were pretreated with the drug; nevertheless, the released viruses exhibited the same deficiency in G protein as the VSV released from untreated cells. Host cell control on both G-protein maturation process and synthesis at traduction level is discussed in relation to G biological properties.

Two forms of yeast RNA polymerase A are resolved by phosphocellulose chromatography. One of these, called RNA polymerase A*, is lacking two polypeptide chains of 48,000 and 37,000 daltons. The properties of the two enzymes are compared in the present paper. RNA polymerase A* transcribes d(A-T)n with a similar efficiency as the complete enzyme, but it is comparatively much less active with native DNA. The two enzymes can also be differentiated on the basis of their ionic strength and divalent cation requirements. RNA polymerase A* has a particularly low activity at high salt and low Mg2+ concentrations. Thermal inactivation curves of the two enzymes are different when residual activity is assayed with native DNA. In contrast with d(A-T)n as template the apparent inactivation curves of the two enzymes are identical. The data suggest that the two dissociable polypeptide chains play an important role in transcription.

Special care to prevent proteolysis during yeast RNA polymerase B purification leads to the appearance of two forms of enzymes, BI and BII, with different molecular weight (465 000) and 435 000, respectively). The two forms of enzyme can be separated by ion-exchange chromatography or polyacrylamide gel electrophoresis. Their subunit structures were compared by sodium dodecylsulfate gel electrophoresis, the only observed difference between the two enzymes is in the molecular weight of the heaviest subunit which is 220 000 for enzyme BI and 180 000 for enzyme BII. Otherwise, the two enzymes have seven common subunits of molecular weights 150 000, 45 000, 26 000, 22 500, 14 500, 12 500 and 9000. Two additional polypeptide chains of 32 000 and 16 500 Mr are dissociated from the enzyme upon polyacrylamide gel electrophoresis or DEAE Sephadex chromatography. The largest subunit of enzyme BI (Mr 220 000) can be specifically cleaved in vitro by a yeast protease extract, generating a polypeptide chain indistinguishable from the largest subunit of enzyme BII. This proteolytic cleavage of enzyme BI in vitro is inhibited by phenylmethylsulfonyl fluoride and does not significantly change the activity of the enzyme with single-stranded or double-stranded DNA as template. The precursor-product relationship of the different forms of class B RNA polymerases in eukaryotic cells is discussed.