Genetic Resistance to Viral Infection: The Molecular …Resistance to Viral Infection 527 were rinsed in 3 X SSC, air dried, baked at 82" for 4 hr. For the cloning of genomic DNA in

Copyright 0 1989 by the Genetics Society of America

Genetic Resistance to Viral Infection: The Molecular Cloning of a Drosophila Gene That Restricts Infection by the Rhabdovirus Sigma

D. Contarnine,* A.-M. Petitjean* and M. Ashburner+ *Laboratoire de Ginitique des Virus, C.N.R.S., Gifsur-Yvette, 91190 France, and +Department of Genetics,

University of Cambridge, Cambridge, England Manuscript received February 3, 1989 Accepted for publication July 27, 1989

ABSTRACT The ref(2)P gene of Drosophila melanogaster has two common alleles, r e f ( 2 ) P which permits the

infection of flies by the rhabdovirus sigma (a), and ref(2)P which is restrictive for u infection. This gene has been cloned by P element tagging and shown to code for two RNAs in adult flies. These RNAs are expressed in both males and females, but only the larger is expressed in ovaries. Both transcripts are shorter, by about 50 nucleotides, in flies carrying the ref(2)Pp allele than in those carrying ref(2)P. The dominance relationships of these two alleles, and the fact that ref(2)P""" alleles are permissive to u infection, suggest that the ref (2)P product is antimorphic to that of the ref(2)pP allele.

~.

V IRUSES are cellular parasites: their replication demands interactions between their own mac-

romolecules and those of their host. These interactions can include the use, for viral replication, of host cell proteins in novel ways. A classic example is that three of the four polypeptides of the replicase of the bacteriophage QP are encoded by the host cell. These three host proteins are the S1 ribosomal protein and the T u and Tf elongation factors required for host protein synthesis. After viral infection these proteins are "hijacked" to function in viral RNA synthesis (BLUMENTHAL, LANDERS and WEBER 1972; WAHBA et al . 1974).

Although the genetic analysis of viral-host function is relatively easy in the case of bacteriophage and their bacterial hosts this is not usually so for animal viruses. Nevertheless, genetic analysis of these interactions is important if we are to understand the role of host functions in the life cycles of particular viruses. One convenient model system for such an analysis is Dro- sophila melanogaster and its sigma virus (u), a naturally occurring rhabdovirus that is widespread in natural populations of this fly (BRUN and PLUS 1980; EMENY and LEWIS 1984).

The most dramatic consequence of sigma virus infection of D. melanogaster is that infected flies are very sensitive to the effects of carbon dioxide-in contrast to uninfected flies those carrying sigma virus are rapidly killed as a consequence of exposure to C o n (L'HERITIER 1958). Natural populations of D. mela- noguster are often polymorphic for alleles that confer resistance to sigma viral infection. These restrictive alleles map to five different loci (GAY 1978). Other

of page charges. This article must therefore be hereby marked "advertisement" The publication costs of this article were partly defrayed by the payment

in accordance with 18 U.S.C. $1734 solely to indicate this fact.

Genetics 123: 525-533 (November, 1989)

than the differences seen on sigma infection, strains that are homozygous for the permissive or restrictive alleles at any of these five loci are indistinguishable. The specificity of the restrictive alleles is also shown by the occurrence of mutations of the virus that can overcome the consequences of host restriction (CON- TAMINE 198 1; COULON and CONTAMINE 1982).

The best known of the loci at which genetic varia- tion affects the ability of sigma virus to productively infect D. melanogaster is ref(2)P (GUILLEMAIN 1953). This gene maps in the cytogenetic region 37E3-37F3 on chromosome arm 2L, between the visible loci hooked and purple (NAKAMURA, GAY and CONTAMINE 1986). Two classes of ref(2)P allele are common, ref(2)P" that is permissive for sigma and ref(2)PP that is restrictive. Heterozygotes show a reduced probabil- ity of viral infection (NAKAMURA 1978). In natural populations in France the frequency of the ref(2)PP allele is between 0.15 and 0.4 (FLEURIET 1976). Loss- of-function alleles of ref(2)P, equivalent to a deletion of the gene, can be obtained after mutagenesis with X-rays (NAKAMURA 1978) and chemicals (D. CONTAM- INE and P. GAY, unpublished data).

The dominance relationships between ref(2)PP, ref(2)P" and ref(2)P"" alleles are not straightforward. ref( 2 )P homozygotes, ref(2)P"lDf and ref( 2))P"/ ref(2)P"" heterozygotes are all equally sensitive to sigma virus. ref(2)PP/ref(2)P" heterozygotes are less sensitive than these three genotypes yet more sensitive than ref( 2)PPIDf or ref( 2)Pp/ref( 2)P""". ref(2)Pp homozygotes are the least sensitive genotype. Thus, in terms of decreasing sensitivity to sigma virus infection, the genotypes can be ranked in the following order: PIP" = P"/Pnun = P"/Df > P / P p > Pp/PUu = Pp/Df > P"/Pp/Pp > The intermediate pheno-

526 D. Contamine, A.-M. Petitjean and M. Ashburner

TABLE 1

Descriptions of the chromosomes used in this study

Deleted for":

Chromosome Cytology 1(2)37Ea T e f O P 1(2)37Fbb Reference' 1(2)37Fa and

Df(2L)TWlSS Df(2L)E55 Df(2L)ODPS Df(2L)pr21 Df(2L)OD12 Df(2L)TW12

Dp(2;Y)G-M1Zd Dp(Y;2)G-H1 Db(2:Y)G-H3d

DP(2;Y)G 37E2-F4;39C2-D1 36B5-C1;40F;Y 37EF" 37F4-38Al;39C2-39Dl 37E2-F1:40B-F

+ - - + + + + - + +

1 1 2 3 2 1 4 4 5 5

' - included within deletion, +, not included within deletion. These two lethal complementation groups are not ordered with respect to each other. References: (1) WRIGHT, HODGETTS and SHERALD (1976); (2) P. GAY and D. CONTAMINE (unpublished data); (3) BRITT-

Deletion derivatives of Dp(2;Y)G. NACHER and GANETSKY (1983); (4) NAKAMURA, GAY and CONTAMINE (1986); (5) HODGETTS (1980).

e An X-ray induced derivativc \tf Dp(2;Y)G with the new order Y1(36C1-37EF)IY1(37EF-40F)IY, the lethal 1(2)37Ea is present but subject to position effect inactivation.

type of ref(2)P/ref(2)PP argues for the codominance of these alleles; yet ref(2)Pp, but not ref(2)P", is haplo- insufficient. The equal sensitivity of ref(2)P" and ref(2)P""" genotypes (e.g., when either allele is heterozygous with a deletion) points to a lack of function of ref (2)P yet this cannot be so, since r e f ( 2 ) P l ref(2)Pp is a far more sensitive genotype than r e f ( 2 ) P l Of or ref(2)P/ref(2)PnuL' (NAKAMURA 1978). These considerations make the molecular analysis of ref(2)P and of its gene products of particular interest.

We describe the molecular cloning of the ref(2)P gene by transposon tagging with the P element. Mu- tations of ref(2)P" to alleles conferring resistance to sigma infection were recovered from a PM hybrid dysgenesis screen. These were found to result from the insertion of a P element into the ref(2)P gene.

MATERIALS AND METHODS

Stocks: The chromosomes used in this work are described in Table 1. The deletions were kept balanced over Zn(ZLR)O, Cy and the duplications as stocks of the type 0~(Z;Y)/Df(PL)E55/Zn(ZLR)O, Cy. The deficiencies and duplications all derive from chromosomes that carried the permissive, ref(2)P, allele. The second chromosomes of the Canton-S and 7r2 stocks also carry this allele. Two second chromosomes carrying the restrictive allele, ref(2)Pp, were used: one was unmarked (ref(2)P) and the other was ref(2)P B1. The OM wild-type strain was used as the reference ref(2)P strain. The 0 and E chromosomes were the base chromosomes on which the OD and E series of deletions were respectively induced. The 1(2)37Ea, 1(2)37Fa and 1(2)37Fb lethal complementation groups were represented by their alleles, E124, OE92 and E146, respectively.

Assay for sigma virus sensitivity: A line of a CyO, ref(Z)P/Zn(PLR)Pm, ref(2)P stock stabilized for an infection with u virus (that is to say, a stock in which the germ- line was infected with, and would transmit, the virus) was selected for transmission of u t o all progeny. The ref(2)P B1 chromosome was then introduced and ref(2)PP Bl/CyO that were sensitive to COn and would transmit u to all

progeny were selected. With the uA3 strain of virus (GAY 1978) all ref (2)P Bl/Df and ~ef(2)P' Bl/ref(Z)P progeny of this strain are resistant to CO2 (1 5 min exposure at 12"). They recover from paralysis within 15 min of exposure to air following treatment with Con. ref(2)P Bl/ref(Z)P" progeny of this strain are CO:! sensitive under the same test conditions; they remain permanently paralyzed by the Con treatment and, eventually, die.

The symbols [u] and [MI or [PI are used to describe a stabilized infection by sigma and cytotype with respect to the PA4 system of dysgenesis. Where appropriate the genotype of the u virus is indicated, where it is not, uA3 may be assumed.

Molecular techniques: The general source for molecular methods and recipes for buffers, etc., was MANIATIS, FRITSCH and SAMBROOK (1 982).

For the extraction of total nucleic acids and RNA, flies were homogenized in lysis buffer (THIREOS, GRIFFEN-SHEA and KAFATOS 1980), using 25 flies per 0.5 ml. After two extractions with phenol (non-neutralized) and chloroform, followed by two chloroform extractions, the nucleic acids were ethanol precipitated at -20°, washed with 70% ethanol at room temperature and dissolved in TE buffer. These preparations were either used directly for electrophoresis or poly(A)+ RNA was selected from them using an oligo-(dT)-cellulose column. For Northern analysis RNA was denatured with glyoxal and dimethylsulfoxide and then electrophoresed in a 1.5% agarose-6 M urea gel, as described by BLONDEL et al. (1988).

Genomic DNA for Southern analysis was prepared by a similar method to that described above, except that neutralized phenol containing 0.1 % hydroxyquinoline was used and all mixings of solutions were done by gentle inversion of tubes, to avoid hydrodynamic shearing of the DNA. After ethanol precipitation the DNA was dissolved in 5 pl TE buffer containing ribonuclease (at 20 pg/ml) per fly equivalent and then incubated at 37" for 1 hr. This material was then used directly for enzymatic digestion. Restriction enzyme digestions were all done overnight in the presence of bovine serum albumin (at 100 pg/ml) and spermidine (at 4 mM). Digested DNA was electrophoresed on 1% agarose gels in Tris-acetate buffer. The gels were treated with 0.25 N HCI for 20 min and then transferred to nitrocellulose, essentially as described by SOUTHERN (1975). The filters

Resistance to Viral Infection 527

were rinsed in 3 X SSC, air dried, baked at 82" for 4 hr. For the cloning of genomic DNA in plasmid vectors DNA

was prepared as described above but, after ethanol precipitation, dissolved in 20 PI TE buffer, per fly equivalent, re- precipitated with ethanol, washed with 70% ethanol and then re-dissolved in 3 p1 TE buffer containing ribonuclease (20 pg/ml per fly equivalent). After complete digestion with EcoRI the DNA was electrophoresed in 1% low melting point agarose (Bethesda Research Laboratories Inc.) and a band corresponding to fragments of molecular weights 2.2- 2.8 kb was removed, melted at 65" and its DNA extracted with phenol.

The purified DNA fragments, within the size range 2.2- 2.8 kb, were ligated with 0.25 pg of pUC9 plasmid DNA (in 180 pl of ligation mix) that had been cut with EcoRI and alkaline phosphatase treated. After ligation, competent Esch- erichia coli DH 1 cells (prepared by the protocol of HANAHAN 1985) were transformed (54 pl ligation suspension with 600 pl of cell suspension) and plated onto ten plates. Replica colonies were lifted onto nitrocellulose filters and screened as described by MASON and WILLIAMS (1985), with slight modifications.

A library of D. melanogaster Oregon-R genomic DNA cloned in the phage vector XEMBL4 was kindly made avail- able to us by M.-C. MARIOL. DNA was prepared from phage as described by MARIOL, PREAT and LIMBOURG-BOUCHON (1987).

Nitrocellulose filters were prehybridized, for at least 6 hr at 45", in a mixture of 50% formamide (deionized), 5 X Denhardt's solution, 0.5% SDS, 3 X SSCP (0.36 M NaCI, 75 mM Na-citrate, 100 mM Na-phosphate, pH 7.2) and 0.1 mg/ ml denatured salmon sperm DNA. Filters were hybridized with radiolabeled probes overnight at 45 " in the prehybrid- ization solution plus 10 mM EDTA, 10% (w/v) dextran sulfate and 10 ng/ml nick-translated probe (usually 2 X 1 Os cpm/pg). The filters were then washed with agitation for six 20-min periods and various stringency conditions-the lowest stringency used was that in 1 X SSC, 0.5% SDS at 65" (for Southern blots) or at 55" (for Northern blots). Filters were autoradiographed with X-ray film without pre- flashing. For reuse the filters were de-hybridized by incu- bation for at least 4 hr at 75" in T E buffer containing 0.2% SDS. In situ hybridization: In situ hybridization to polytene

chromosomes was done as described by GUBB et al. (1 984).

RESULTS

P element induced ref(2)P mutations: PM dysgenic males homozygous for re f (2)P were crossed to ref(2)Pp Bl/CyO females that were stabilized for u virus infection (Figure la). All the progeny of this cross will be infected by u, due to the maternal inheritance of the virus, and all will be sensitive to C 0 2 unless heterozygous for ref(2)P' and either a newly induced ref(2)P""" allele or a deletion of the ref(2)P gene (NAKAMURA 1978; NAKAMURA, GAY and CONTAMINE 1986). Of approximately 30,000 ref(2)P' Bl/ref(2)P* progeny (where ref(2)P* indicates the re f (2 )P allele subjected to dysgenesis) 75 C 0 2 resistant flies were recovered. These were backcrossed to ref(2)P' Bl/ Cy0 [a], and subjected to a second screen with carbon dioxide. After the second screen three independent mutations of ref(2)P were retained (ref(2)phd', ref(2)Phd2 and ref(2)Phd3). Two of these ( h d l and hd3)

rcffZ)P'/rcffZ)P' IMI 99 x rcffZ)P1/reff2JP" IPI dd

a Y L

CyO/reffZ)PP El Iol lMl 9 9 x n.ffZJ~/reflZ)P' lPMl dd

cot CyO/reffZ)P' lo1 reffZJPP Bl/rcffZJP* lo1 reffZ)f%' El/rel(Z)P' 101

sensitive Sensltlve resistant 7 + CyO/ref(ZJPP 81 IollMl x single d or 9

7 CO,: CyO/reff2)PP El lo1 CyO/reffZJP' lo1 reffZJPP BllreflZJP' lo1

sensltlve resistant

stock

b CyO/DffZL)E55 /MI 99 x CyO/ref(Z)phd IPI dd +

CyO/Df(ZL)E55 IMI 99 x CyO/reffZ)phd l P M l dd

CVJ/WJPM' single d X CyO/DffZL)E55 Dfl2LJE55/rcfcZJP"

I d fertile ?

ref(2,phd*/DffZLJE55

viable ? d fertile ?

FIGURE 1 .-The crosses used to generate PM dysgenic u resistant alleles of ref(2)P and their revertants. The P strain was 7r2 and the M strains were either Canton-S or OM. All three of these strains are ref(2)P. a, The screen for PM induced ref(2)phd alleles, dysgenic ref(2)P males were crossed to ref (2)P Bl/CyO [MI females stably infected by u. From their progeny ref(2)P*/ref(Z)P Bl flies resistant to the effects of CO:, were selected and crossed singly to u-infected ref(2)PP Bl/CyO males or females. The progeny of this cross was subjected to a second round of selection by exposure to COn and resistant ref(Z)P*/ref(Z)P Bl flies recovered. Their ref(2)P* chromosomes were then balanced over CyO. Forty vials (12,000 chromosomes) were tested using the Canton-S M strain and 90 vials (16,000 chromosomes) with the OM M strain. b, To select revertants of ref(2)P' alleles ref(Z)phd/CyO [PI males were crossed to Df(ZL)E55/CyO [MI females and their dysgenic ref(Z)P/CyO sons backcrossed to females of the same genotype as their mothers. The ref(2)phd/Df(2L)E55 sons were tested for fertility by crossing to homozygous ebony females; the ref(2)phd/Cy0 sons were tested for lethal mutations within Df(ZL)E55 by crossing to Df(2L)E55/CyO females. Viable ref(2)phd/Df(2L)E55 males from this cross were also tested for their fertility.

appeared as clusters, of six and seven flies, respectively, in the first screen.

The PM induced alleles of ref(2)P have similar genetic properties to ref(2)Pnu" alleles (NAKAMURA 1978): thus, ref(2)phd/lref(2)P', like ref(B)P"""/ ref(2)P', are restrictive to u and ref(2)phd/ref(2)P, like ref(2)Pn""/ref(2)P, are permissive to u infection. All three dysgenic alleles are viable when heterozygous with a deletion for ref(2)P, Df(2L)E55. However, ref(2)Phd/Df(2L)E55, like ref(2)PU"/Df(2L)E55, are all male sterile. These males possess sperm, but these are amotile. T h e male sterility of ref(2)P""'l alleles is conditional: on some genetic backgrounds ref(2)Pn""/ Dfmales are fertile. This has been found to be due to a dominant suppressor of their sterility that maps just proximal to scarlet on chromosome arm 3L. Labora- tory stocks, and natural populations, are polymorphic for dominant suppressor and recessive nonsuppressor alleles of this gene (P. GAY, unpublished data).

528 D. Contamine, A.-M. Petitjean and M. Ashburner

I f the resistance to u shown by the flies heterozygous for ref(2)Phd is due to the insertion of a P element into the ref(2)P gene, then the mutant alleles should revert to CO:! sensitivity in the presence of u, by further dysgenesis. Dysgenic ref(2)Phd/Cy0 males were obtained from a cross of ref(2)Phd/Cy0 [PI males to I)f(2L)E55/CyO [MI females and backcrossed to Df(2L)E55/CyO females (Figure Ib). The male ref(2)phd/Df(2L)E55 progeny were tested for their fertility and the ref(2)Phd/Cy0 progeny tested for le- thals mapping within Df(2L)E55. From approximately 1500 individual males 12 independent revertants of ref(2)P""' and three revertants of ref(2)Pnd2 were recovered, by the criterion of the fertility of males heterozygous with Df(2L)E55. All of these revertants also recovered the full characteristics of a ref(2)P" allele, that is to say they were all active and permissive for u virus. One lethal revertant of ref(2)phd2 was recovered-this chromosome carries a deletion, Df(2L)CPPBR (= Df(2L)37F4;38A5.6 inclusive) that overlaps Df(2L)E55. This deletion does not include ref(2)P, but does include the two lethal complementation groups [1(2)37Fu and 1(2)37Fb] that map im- mediately proximal to this gene.

ref(2)phd' has a P element in 37F One of the three mutations, ref(2)Phd', and one of its revertants, ref(2)phd'"' were selected for further study. Using a cloned P element as a probe this mutation was shown to have a P element inserted into region 37F, as well as at about 30 other sites. Its revertant lacked the insertion at 37F (data not shown).

To reduce the number of P elements in the ref(2)phd' and ref(2)Phd'"' stocks the X and third chromosomes were replaced with those of an M strain. In addition recombination between these second chromosomes and those from an M strain, using pr as a closely linked marker, was used to reduce the numbers of P elements in these stocks. Several independent ref(2)phd' and ref(2)Phd'"' recombinant chromosomes were retained for molecular analysis.

Cloning strategy: Genomic DNA was extracted from several independent recombinant lines of ref(2)phd' and, after digestion with the restriction enzyme EcoRI, electrophoresis and Southern blotting, this was probed with a labeled probe prepared from the 540-bp HindIII-PvuII fragment of the P element. As is evident from the data shown in Figure 2 all the lines show many EcoRI fragments with homology to the P element. Nevertheless all of the ref(2)phd' lines, but none of the ref(2)phd'"' lines, show hybridization of the probe to an approximately 2.5-kb EcoRI fragment. This is the pattern of hybridization expected to a restriction fragment from the ref(2)P gene carrying an inserted P element. These results were confirmed by probing EcoRI-Sal1 double digests of DNA from these stocks (data not shown).

DNA from one of the ref(2)Phd' recombinant lines

kb

9.4-

4.3-

+ 2.0-

FIGURE 2.-Southern blot hybridimtion of genomic DNA digested with EcoRl from six different ref(2)pd and five different ref(2)pdm strains probed with the 540-bp HindllI-Puull fragment of the P element. The 2.5-kb RcoRl fragment (drrowed) is seen in all strains with ref(2)pd chromosomes but in none with the revertant chromosomes.

was digested to completion with EcoRI and fragments in the size range of 2.2-2.8 kb were isolated by gel purification on a preparative agarose gel. These were cloned into the EcoRI site of the plasmid pUC9. From approximately 20,000 colonies of E. coli transformed with these constructs and probed with P element sequences one positive clone was recovered (P2). This clone hybridized to a 2.5-kb EcoRI fragment from ref(2)phd'"' DNA but to a 1.3-kb EcoRI fragment from ref(2)phd'"' DNA (data not shown). DNA from two different ref(2)phd'/Df genotypes, ref(2)phd'/ Df(2L)E55 and ref(2)phd'/Df(2L)ODP5, digested with EcoRI and probed with P2 shows only the 2.5-kb band. DNA from ref(2)phd'/+ genotypes shows both the 2.5- kb and 1.3-kb bands while DNA from control genotypes shows only the 1.3-kb band (Figure 3). These data indicate that the 2.5-kb EcoRI fragment includes both DNA from the ref(2)P region and approximately 1.2-kb of P element DNA: the element had, in the ref(2)phd' mutation, inserted into an approximately 1.3-kb genomic EcoRI fragment.

Using probes prepared from the P2 plasmid nine X phage were selected from a library of Oregon-R DNA cloned in the vector XEMBL4. These phage include about 34-kb of genomic DNA from which EcoRI subclones in pUC9 were made for further characterization of the ref(2)P region (Figure 4).

Mapping breakpoints in the ref(2)P region: Ge- netically ref(2)P maps between the proximal limit of Df(2t )TW158 and the distal limits of Df(2L)CPP2R and the deletion associated with Dp(Y;2)G-H3. A breakpoint of Dp(Y;2)G-M15 is associated with a


1 2 3 4 5

t3kb-

FIGURE 3.- DNA, digested with EcoRI, from various genotypes probed with the plasmid P2 selected with a P element probe (the 540-bp HindIII-Puull fragment) from a pUC9 plasmid library of sized DNA from r e f ( 2 ) p d ' . DNA from flies carrying the ref(2)phd' allele shows hybridization to a 2.5-kb EcoRI fragment, DNA from flies not carrying this dysgenic ref(2)P allele shows hybridization to a 1.3-kb fragment. DNA from flies heterozygous for this ref(2)phd allele and a nondeleted second chromosome shows hybridization to both fragments. The tracks are: ( I ) ref(2)P"d'/Df(2L)E55, (2) re f (2 )pd ' /E , r e f (2 )P . (3) ref(2)phdl/Df(2L)ODP5,(4) ref(2)P"d'/0, ref(2)P". (5) ref(2)P"'"/CyO, r e f ( 2 ) P . (6) Df(PL)ODPS/CyO, ref(2)P. (7) E , re f (P)P/CyO, re f (2)P. and (8) Df(2L)E55/Cy0, re f (2)P. The 0 and E chromosomes are control chromosomes for Df(2L)ODP5 and Df(2L)E55, respectively.

ref(2)P""" mutation. DNA from flies carrying these and other aberrations in this chromosome region, as well as DNA from flies carrying suitable control ( i e . , progenitor) chromosomes (see Table 1 and MATERIALS AND METHODS), was digested with various restriction enzymes and, after electrophoresis, probed with various subclones of the X phage. The data are summa- rized in Figure 4.

The proximal limit of Df(2L)TwZ58 is at about + 17 kb (data not shown). The distal breakpoints of Df(2L)CPPBR and the deletion of Dp(Y;2)G-H3 are at approximately +8 kb and + 10 kb, respectively (Figure 5b the data for Dp(Y;2)G-H3). The region 37EF breakpoint, associated with a ref(2)P""" mutation, of Dp(Y;2)G-MZS is within the 1.3-kb EcoRI fragment that spans coordinate 0 (Figure 5a). The three independent dysgenic alleles of ref(2)P all carry insertions within this 1.3-kb region. The size (and orientation) of the insertion is not the same in any two of these chromosomes. Two different sublines of ref(2)phd2, both carrying null alleles, have been found to carry P element insertions that differ in their size.

These data are all consistent with the conclusion

that the ref(2)P gene includes sequences within the 1.3-kb EcoRI fragment that spans coordinate 0.

Identification of ref(2)P transcripts: Northern hybridization to total RNA, extracted from ref(2)P" (the OM strain) and ref(2)phd adult flies, with subclones from the region identified as including the ref(2)P gene as probes revealed several different transcripts from between coordinates - 13 and +9 (Figure 6) (see Figure 4 for the limits of the EcoRI subclones). With the 23E probe an abundant 2.3-kb transcript is seen in both ref(2)P" and ref(2)phd" adults. This RNA is very much less abundant in RNA extracted from all three of the dysgenic alleles. In flies carrying each allele it is apparently replaced by RNAs of higher molecular weight, these differ in size among the three mutant strains (e.g., Figure 7). The obvious difference in abundance of the 2.3-kb RNA between males and females of ref(2)phd'"' (Figure 6a) is mostly due to the fact that this RNA is synthesized in ovaries (see below). It can be less obvious in some experiments, due to the relative overexposure and nonlinearity of the X-ray film.

The 4 1 E-A probe hybridizes weakly and, presumably nonspecifically, to the major ribosomal RNAs (Figure 6b). This hybridization is not seen in North- erns using poly(A)+ selected RNA (data not shown). In addition this probe hybridizes to a small (1.4 kb) RNA whose expression in adult flies may be female specific. The 17E probe hybridizes to an abundant 4- kb RNA. Neither the 1.4-kb RNA detected with probe 4 1 E-A nor the 4-kb RNA detected with probe 17E are affected by mutations at the ref(2)P gene. These data strongly implicate the 2.3-kb RNA seen with probe 23E as being the product of ref(2)P.

In Figure 7 is shown the result of hybridizing the 23E probe to poly(A)+ selected RNA from various ref(2)P genotypes. These data confirm that flies carrying the dysgenic loss-of-function alleles have much less of the 2.3-kb RNA than ref(2)P" or ref(2)PP. These data also illustrate allele specific novel transcripts associated with the dysgenic alleles.

The 2.3-kb RNA from ref(2)P can be resolved into two RNA species by electrophoresis, these differ in their size by about 100 nucleotides (Figure 8). Both RNAs are seen in adult males and females but only the larger is expressed (at any abundance) in ovaries. That both RNA species are the products of the same gene is known from the fact that the ref(2)P region, as identified here, is unique in the genome and that both RNAs are smaller, by about 50 nucleotides, in ref(2)PP flies than in ref(2)P" flies (Figure 8).

DISCUSSION

The interest in the ref(2)P gene comes from the fact that its alleles differ in whether or not they permit a productive infection of D. melanogaster by the rhab-

530 D. Contarnine, A.-M. Petitjean and M. Ashburner

2 3 phage

9 E' H' -20 -10 0 10 tf 20 kb I I I I I I

I S s' s' ;ti ;r X

- Df(2i) TWl 58 * Df(2L)CPPZR

hdl Dp(2;Y)G-H3 - Df(2L)pr21

I I I I I 1 I I I 1 19E 35E 14E 72E 13E 23E 41EA 41EB 17E 31 E

n n f?v RI subclones - 3-kb

"- 2.3-kb 1.4-kbY 4-kb

polyA(+) RNA

FIGURE 4.-A molecular map of the ref(2)P region showing the extent of the recombinant X phage and the positions of chromosome aberrations mapped to this DNA. The 0 coordinate is the position of the P element inserted in ref(2)Pd' and the values of the coordinates are negative toward the telomere and positive toward the centromere. Only the sites for the restriction enzymes Sal1 ( S ) , Hind111 ( H ) and Xhol ( X ) are shown over the entire walk: BamHI ( B ) sites are only shown proximal to coordinate 0. Sites marked with an asterisk are only found on the Cy0 chromosome. The small insertion at about +16.5 is about 1 kb in size and is also only in the Cy0 chromosome. The two insertions are not drawn to scale. Beneath the map are shown the limits of the subcloned EcoRI fragments used to analyze both the DNA of mutant strains and the transcripts of this region. Subclone 41E is of a single EcoRI-EcoRI fragment; subclones 41EA and 41EB are the EcoRI- PvuII and PvuII-EcoRI regions of 41E, respectively. The order of EcoRI sites between coordinates +7 kb and +10 kb is uncertain, therefore the precise position of the 17E EcoRI subclone is ambiguous. The approximate regions from which transcripts were detected in adult flies are also indicated. The distal breakpoints of Df(2L)TWlZ and Df(2L)OD12 were not found on the cloned DNA-they are presumably located to the right of the region cloned.

dovirus sigma. This negative strand RNA virus is very common in natural populations of D. melanogaster, where it is vertically transmitted via eggs. The virus is normally benign, although some laboratory strains of the virus reduce fecundity (e.g., JUPIN, PLUS and FLEURIET 1968), unless the flies are unlucky enough to be exposed to carbon dioxide: then they are rapidly killed as a consequence of infection. Stable infection of Drosophila by u, viral replication and viral matura- tion require host functions. Some of these have been identified by genes at which alleles resistant to u map.

Five ref genes are known (GAY 1978). None have any discernible effect on the phenotype of flies, other than that which can be assayed after infection with u virus and the male sterility of null alleles of one, ref(2)P. ref(2)PnUlL alleles, whether induced by X-rays (NAKAMURA, GAY and CONTAMINE 1986), diepoxybutane (D. CONTAMINE, UNPUBLISHED DATA) OR PM hybrid dysgenesis (this paper), are male sterile. The two widespread alleles of ref(2)P, ref(2)P" and r e f ( 2 ) P , are both male fertile, The sterility of ref(2)P""" males is a consequence of the amotility of their sperm. This has no obvious connection with the biology of the u life cycle. Moreover, this sterility is complex since alleles of a gene that maps to the third chromosome act as dominant suppressors of it. These alleles are apparently loss-of-function, since they are readily induced by diepoxybutane.

The ref(2)P gene maps to the chromosome region

37E3 to 37F3 on 2L. It is proximal to the left-hand end of Df (2L)Tw158 but distal to the right-hand limits of Df(2L)CPP2R, Df(ZL)pr21, Df (2L)OD12 and the deletion associated with Dp(2;Y)G-H3. The smallest interval (between the endpoints of Df(ZL)Tw158 and Df(2L)CPP2R) is about 25 kb. The entire region is included within a much longer deletion, Df(2L)E55. This deletion (of approximately 18 polytene chromosome bands) has been reasonably well "saturated" for EMS and diepoxybutane induced lethal mutations. The 145 mutations recovered map to 14 different lethal complementation groups (P. GAY and D. CON- TAMINE unpublished data). None of these 14 loci maps to the interval between the endpoints of Df(ZL)Twl58 and Df(2L)CPP2R. Yet, in addition to ref(2)P, this interval includes at least three different transcripts that can readily be detected in RNA from adult flies.

ref(2)P has been cloned by P element tagging, using the difference in COS sensitivity between ref(Z)P"/ r e f ( 2 ) P and ref(Z)P"""/ref(Z)P genotypes (stably infected with u) as the basis for the mutant screen. All three mutations show both the COS sensitivity and male sterility characteristic of ref(2)P"" alleles; both of these phenotypes revert when the P element is lost from 37EF. All three mutations lack, or have in a greatly reduced amount, transcripts of approximately 2.3 kb in size that originate from the genomic region into which the P elements have, in each case, inte-

kb

21-

5 -

3.5-

2.

0.8

Resistance to Viral Infection

E X 41E 31E

kb

21 -

5 -

35-

2 -

53 1

FIGURE 5.”Southern blots of genomic DNA cut with restriction enzymes and then probed to illustrate the mapping of the breakpoints associated with Dp(P;Y)G-M15 (a) and Dp(2;Y)G-H3 (b). a, DNA from male Dp(2;Y)C/Df(2L)E55/Cy0 (tracks 1 and 3) and from male Dp(2;Y)G- MJ5/Df(2L)E55/Cy0 (tracks 2 and 4) was digested with either EcoRl (tracks 1 and 2) or Xhol (tracks 3 and 4) and, after electrophoresis and blotting, probed with subclone 13E. Fusion fragments (of 22 kb and 5 kb) are seen in the M15 tracks, but not in those from their progenitor chromosome. b, DNA from male DP(2;Y)G-Hl/Df(PL)E55/CyO (tracks 1 , 3, 5 , 7, 9 and 1 1 ) or male Dp(2;Y)C-Hjr/Df(2L)E55/CyO (tracks 2, 4 , 6, 8, 10 and 12) digested with BamHl (tracks 1 , 2, 7 and 8). EcoRI (tracks 3, 4, 9 and 10) or Sal1 (tracks 5, 6, 1 1 and 12) and, after electrophoresis and blotting, probed with subclone 41 E (tracks 1-6) or subclone 31E (tracks 7-12). With the 31E probe all Dp(Y;2)C-H3 tracks show the absence of fragments found in the progenitor. In tracks 9 and 10 the smaller of the EcoRI fragments is from the Cy0 homolog. The faint larger EcoRI fragment in track 10 probably results from a polymorphism among Cy0 chromosomes for the presence of the insertion at approximately +15 kb, a minority of Cy0 chromosomes lacking this insertion. With the 41E probe a 16-kb fusion fragment is seen when the DNA had been digested with BamHI. Note that the smaller of the BnmHI fragments in tracks 1 and 2 is from the Cy0 chromosomes, and results from the extra EamHI site at +9 kb unique to Cy0. All fusion bands are flanked by small black dots.

grated. The revertants recover these transcripts. The ref(2)P gene is transcribed into two RNAs that

differ by only 100 nucleotides or so in their size. This difference is due to transcription initiating at two different sites, about 100 bp apart (S. D~ZELEE and F. BRAS unpublished data). These two RNAs are of approximately equal abundance in both males and females. However transcription of ref(2)P does show tissue specificity, since only the longer of the RNAs is expressed in the ovaries. The ovarian expression of ref(2)P is expected on two grounds: one is that the alleles of ref(2)P affect the ability of the u virus to stably infect the germ line, and to be transmitted via the eggs; the other is that ref(2)P shows a maternal effect. Sensitivity to COn is seen in ref(2)P/ref(2)P progeny of ref(2)P“/ref(2)P mothers crossed to ref(2)P/ref(2)P [u] males but never in genetically similar progeny of ref(2)PP/ref(2)P mothers crossed to the same males (GAY 1968; P. GAY, personal com- munication).The dominance relationships between the three types of allele known at ref(2)P are not simple. By one criterion the ref(2)P and ref(2)P alleles are codominant, since heterozygotes are intermediate between either homozygote in their sensitiv-

ity to u infection. This would argue that ref(2)P is not a null allele. This is confirmed by the fact that the sensitivity to u of ref(2)P homozygotes is increased by the addition of a duplication carrying ref(2)P (NAKAMURA 1978). The ref(2)P allele displays haplo- insufficiency, since when heterozygous with a deletion (or with a ref(2)P“”” allele) its sensitivity to u infection is greater than that when homozygous. The ref(2)P allele is not dosage sensitive, ref(2)P homozygotes and ref(2)P/Df(or ref(2)P/ref(2)PU”) heterozygotes have the same sensitivity to u infection. Homozygous, or hemizygous, re f (2 )F” flies have the same sensitivity to u as these genotypes.

One obvious model is that ref(2)P and ref(2)P make different gene products, that compete for some target, the ref(2)P being required for the nonpermissive state. It is of some considerable interest, therefore, that the transcripts from these genes differ in their size. This difference is due, at least in part, to a deletion within the coding region of the ref(2)P allele (D. CONTAMINE, unpublished data).

On a competitive model we can envisage the product of ref(2)P interacting, perhaps stoichiometrically (to account for this allele’s dosage sensitivity), with the

532 D. Contarnine, A.-M. Petitjean and M. Ashburner

23E 41EA 17E

kb 1 2

7.6-

4.1-

2.3-

1.3-

PO P P hd3 phdl

1 2 3 4 5 6 7 8 9 1 0 1 1 1 2

kb

2.7-

21)-

FIGURE 7.-Northern analysis of the ref(2)P transcript in poly(A)+ selected R N A of wild-type and mutant strains. Poly(A)+ selected R N A ( I O &track) from r e f ( 2 ) P (tracks 1-3), r e f ( 2 ) P (tracks 4-6). reJ(2)f‘”‘’ (tracks 7-9) and ref(2)Pd’ (tracks 10- 12). Tracks I , 4, 7, and I O are from females: tracks 2, 5.8 and 1 1 from males: and tracks 3, 6, 9 and 12 from mixed sexes. Probed with clone 23E. In the lower part of the figure the same blot is shown. but after hybridization with a D. melunogusler Act5C gene probe, as a control for the loading of the tracks.

product of some other gene (perhaps the product of another ref gene or of the viral genome, see below). This interaction would be required for the establish- ment of the condition of being nonpermissive to u infection. The product of the ref(2)P“ allele would

FIGURE &-Northern analysis of transcripts from the ref(2)P region. Total R N A (equivalent to that from three adult flies for each track) was electrophoresed and transferred to a Hybond filter. Track 1, r e f ( 2 ) P (mixed sexes), track 2 re f (2)Pd3 (mixed sexes), track 3 ref(2)Pd’”’l (females), track 4 ref(2)Pd’“’ (males), track 5 ref(2)Pd’ (females) and track 6 r4(2)Pd’ (males). The filter was successively hybridized with three different probes: in a with probe 23E, in b with probe 4 1 E-A and in c with probe 17E. See text for discussion. The 1.4-kb R N A seen with probe 4 1 E-A is not detected by probe 4 1 E- B. The broad band marked Y in b is due to hybridi~ation to rRNA. Tracks 4, 5 and 6 have been over- exposed, relative to tracks 1, 2 and 3, by approximately tenfold in a, had they not been the signal intensities in tracks 2 and 5 would have been a p proximately the same.

PO PP k b 1 2 3 4 5 6

FIGURE 8.-The tissue specificity of ref(2)P transcripts. Poly(A)+ selected R N A (1-3 pg) from handdissected ovaries (tracks 1 and 4), total adult females (tracks 2 and 5) and total adult males (tracks 3 and 6) from ref(2)P“ (tracks 1-3) and r e f ( 2 ) P (tracks 4-6) strains, probed. after electrophoresis and blotting, with clone 23E.

compete for this interaction, and hence promote a permissive state. The ref(2)P”” mutations would result in the permissive state simply because the absence of a gene product would prevent any opportunity for the interaction required for nonpermissiveness to oc- cur. Since temperature sensitive mutations of u that can infect ref(Z)P’/ref(2)P’ hosts can be selected (the haP mutations, CONTAMINE 198 l ) , the viral genome, or its products, presumably interacts directly with a host product related to the activity of ref(2)P. Indeed


it may be just this interaction which is targeted by the products of the different ref(2)P alleles: if there is a viral gene product, V, that is required for viral replication then the complex of V with the ref(2)Pp product may be inactive, but V alone or V complexed with the r e f ( 2 ) P product active. It is interesting that the restrictive effects of ref(2)Pp are stronger in the early stages of the viral cycle than after viral genome replication (D. CONTAMINE, unpublished data). Since replication is followed by an increase in transcription, then we would expect the concentration of V to increase as a consequence of replication. This would ameliorate the effects of a presumably constant concentration of the ref(2)PP gene product.

If substantiated, this model would mean that the ref(2)P" allele, that usually most common in natural populations, is an antimorph with respect to nonpermissiveness for u infection. The product of this allele, or that of ref(2)Pp is, however, required for the normal development of sperm.

This work was supported in part by an SERC grant to M.A. We would like to thank C. SAVAKIS and S. MCGILL for their help in the early stages of this work. We also thank M.-C. MARIOL for the gift of the genomic library of Drosophila DNA, and P. LASKO for his comments on the manuscript.

LITERATURE CITED

BLONDEL, D., A. M. PETITJEAN, S. D ~ Z E L ~ E and F. WYERS, 1988 Vesicular stomatitis virus in Drosophila melanogaster cells: regulation of transcription and replication. J. Virol. 62: 277-284.

BLUMENTHAL, T. , T. A. LANDERS and K. WEBER, 1972 Bacteriophage QB replicase contains the protein biosynthesis elongation factors EFTu and EFTS. Proc. Natl. Acad. Sci. USA 6 9 1313-1317.

BRITTNACHER, J. G., and B. GANETSKY, 1983 On the components of segregation distortion in Drosophila melanogaster. 11. Dele- tion mapping and dosage analysis of the SD locus. Genetics

BRUN, G., and N. PLUS, 1980 The viruses ofDrosophila, pp. 625- 702 in The Genetics and Biology of Drosophila, Vol. 2d, edited by M. ASHBURNER and T. R. F. WRIGHT. Academic Press, New York.

CONTAMINE, D., 1981 Role of the Drosophila genome in sigma virus multiplication. I. Role of the ref(2)P gene: selection of host-adapted mutants at the non-permissive allele ref(2))P. Virology 114: 474-488.

COULON, P., and D. CONTAMINE, 1982 Role of the Drosophila genome in sigma virus multiplication. 11. Host spectrum var- iants among the hap mutants. Virology 123: 381-392.

EMENY, J. M., and M. J. LEWIS, 1984 Sigma virus of Drosophila as a vector model. Spec. Publ. SOC. Gen. Microbiol. 12: 93-1 12.

103: 659-673.

FLEURIET, A,, 1976 Presence of the hereditary rhabdovirus sigma and polymorphism for a gene for resistance to this virus in natural populations of Drosophila melanogaster. Evolution 3 0 735-739.

GAY, P., 1968 Adaptation d'une population virale i se multiplier chez un h&e rifractaire. Ann. Genet. 11: 98-1 10.

GAY, P., 1978 Les gi.nes de Ia Drosophile qui interviennent dans Ia multiplication du virus sigma. Mol. Gen. Genet. 159 269- 283.

GUBB, D., M. SHELTON, J. ROOTE, S. MCGILL and M. ASHBURNER, 1984 The genetic analysis of a large transposing element of Drosophila melanogaster. The insertion of a w'rst' TE into the ck locus. Chromosoma 91: 54-64.

GUILLEMAIN, A., 1953 Dicouverte et localisation d'un g b e em- pCchant Ia multiplication du virus de la sensibiliti hiriditaire au Con. C. R. Acad. Sci. Paris 236: 1085-1086.

HANAHAN, D., 1985 Techniques for transformation of E. coli, pp. 109-135 in DNA Cloning. A Practical Approach, Vol. 1 , edited by D. M. GLOVER. IRL Press, Oxford.

HODGETTS, R. B., 1980 A cytogenetic description of three duplications in which portions of proximal 2L have been inserted into the Y-chromosome. Drosophila Inform. Serv. 55: 63.

JUPIN, N., N. PLUS, and A. FLEURIET, 1968 Action d'une souche de virus sigma sur la fertiliti des Drosophiles femelles. Ann. Inst. Pasteur 114: 577-594.

L'HERITIER, P., 1958 The hereditary virus of Drosophila. Adv.

MANIATIS, T., E. FRITSCH and J. SAMBROOK, 1982 Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.

MARIOL, M.-C., T . PREAT and B. LIMBOURG-BOUCHON, 1987 Molecular cloning of fused, a gene required for normal seg- mentation in the Drosophila melanogaster embryo. Mol. Cell. Biol. 7: 3244-3251.

MASON, P., and J. G . WILLIAMS, 1985 Hybridisation in the analysis of recombinant DNA, pp. 1 13-1 37 in Nucleic Acid Hybridisa- tion. A Practical Approach, edited by B. D. HAMES and S. J. HIGGINS. IRL Press, Oxford.

NAKAMURA, N., 1978 Influence du dosage de l'allde non per- missif du g b e ref(2)P de la drosophile sur les souches sensibles du virus sigma. Mol. Gen. Genet. 159: 285-292.

NAKAMURA, N., P. GAY and D. CONTAMINE, 1986 Etude du locus ref(2)P de Drosophila melanogaster. 1. Localisation cytogini- tique de ref(2)P. Biol. Cell. 5 6 227-238.

SOUTHERN, E., 1975 Detection of specific sequences among DNA fragments separated by gel electrophoresis. J. Mol. Biol. 98: 503-5 17.

THIREOS, G., R. GRIFFEN-SHEA and F. C. KAFATOS, 1980 Untranslated mRNA for a chorion protein of Drosophila melanogaster accumulates transiently at the onset of specific gene amplification. Proc. Natl. Acad. Sci. USA 77: 5789-5795.

WAHBA, A. J., M. J. MILLER, A. NIVELEAU, T. A. LANDERS, G. G. CARMICHAEL, K. WEBER, D. A. HAWLEY and L. I. SLOBIN, 1974 Subunit I of QB replicase and 30s ribosomal protein S1 of Echerischia coli. J. Biol. Chem. 249 33 14-33 16.

WRIGHT, T. R.F., R. B. HODGETTS and A. F. SHERALD, 1976 The genetics of dopa decarboxylase in Drosophila melanogaster. I. Isolation and characterization of deficiencies of the structural locus and the alpha methyl dopa hypersensitive locus. Genetics 84: 267-285.

Communicating editor: A. SPRADLING

Virus Res. 5: 195-245.

Documents

Genetic Resistance to Viral Infection: The Molecular …Resistance to Viral Infection 527 were rinsed in 3 X SSC, air dried, baked at 82" for 4 hr. For the cloning of genomic DNA in