J~u~nal~`~f ~ne~ral~ Vir~l~gy .!~99%~.7.~3~!~4.9~3 !~4:~F~!n~ed~~n?e~a!~Br!!

a~

...........................................................................................................................

Multiple viral determinants affect seed transmission of pea seedborne mosaic virus in Pisum sativum
I. E. J o h a n s e n , 1 W. G. D o u g h e r t y , 2 K. E. Keller, 3 D. W a n g 4 a n d R. O. H a m p t o n 3

1Biotechnology Group, The Danish Institute of Plant and Soil Science, Lyngby, DK-2800 Denmark z Department of Microbiology and Center for Gene Research and Biotechnology, Oregon State University, Corvallis 97331-3804, USA 3Department of Botany and Plant Pathology, Oregon State University, Corvallis 97331-2902, USA 4 Department of Virus Research, John Innes Centre, Norwich NR4 7UH, UK

Two pea seedborne mosaic potyvirus (PSbMV) isolates, P-1 DPD1 (P-l), which is highly seedtransmitted, and P-4 NY (P-4), which is rarely seedtransmitted, and chimeras between P-1 and P-4 were analysed to map the viral genetic determinants of seed transmission. Infectivity of chimeric viruses was evaluated by inoculating Pisum sativum with RNA transcribed in vitro from recombinant fulllength cDNA clones. The chimeric viruses that were used demonstrated that a genomic segment encoding the 49 kDa protease and putative RNA polymerase was responsible for symptom induction. Attempts to determine transmission of the chimeric

viruses in P. sativum cultivars known to transmit P1 at high frequencies showed that seed transmission is a quantitative character influenced by multiple viral determinants. Seed transmission frequency did not correlate with accumulation of virus in vegetative tissue. The 5' 2"5 kb of the 10 kb PSbMV genome had a major influence on the seed transmission frequency and was analysed further. This showed that, while the helper-component protease was a major determinant of seed transmission, the potyviral Pl-protease exerted no measurable influence.

Introduction Transmission through seed has been described for 108 plant viruses in one or more of their hosts. For all of these, except tobacco mosaic virus, successful seed transmission depends on the virus entering and surviving in the embryo (Mink, 1993). Seed transmission is precluded when the virus is unable to infect the gametes prior to fertilization, unable to enter the embryo during development, or when the virus is inactivated in the embryo during seed maturation and storage (Maule & Wang, 1996). In virus-host combinations with potential for seed transmission, the frequency of seed transmission depends on both host and virus genotype and may range from 0% to almost 100% (Mink, 1993). The mechanisms of resistance to seed transmission are not resolved and inheritance of resistance has been investigated in only a few cases. In Hordeum vulgare cultivar Modjo, resistance was reported to be conditioned by a single recessive gene (Carroll et al., 1979), whereas resistance to seed transmission of pea seedborne mosaic potyvirus (PSbMV) in Pisum sativum is Author for correspondence: I. E. Johansen.
Fax +45 45 93 22 13. e-mail e.johansen@dips-lyngby.dk

inherited as a quantitative character controlled by the action of multiple maternal genes (Wang & Maule, 1994). In a single host cultivar, different virus isolates are seed-transmitted to different degrees. These differences may reflect differences in virus replication and movement which determine the frequency at which the virus successfully enters the gametes or the embryo (Carroll, 1981). Pseudorecombination studies with isolated viral RNAs from strains differing in seed transmissibility showed that the seed transmission phenotype was linked to RNA 1 of raspberry ringspot and tomato black, zing nepoviruses (Hanada & Harrison, I977) and to RNA I of cucumber mosaic cucumovirus (Hampton & Francki, 1992). Analyses of barley stripe mosaic hordeivirus (BSMV) chimeras facilitated mapping of the major determinants of seed transmission in barley (H. vulgare) to the 5' untranslated leader of RNAT, a 369 bp repeat in the 7a gene, and the 7b gene (Edwards, 1995). The seed transmission frequency of PSbMV is influenced by the genotype of the virus isolate (Kohnen et al., 1995). The complete nucleotide sequences of two isolates of PSbMV which differ widely in seed transmission frequency in several P. sativum cultivars have been determined (Johansen et al., 1991, 1996). However, a comparison of the primary sequences

0001-4205 1996 SGM

~14c~

i',i i' iiiiii!ii ii! i':i',ii:iii:ii:ii:.igii

iiiiii,iii.iiiiiiiiiiiiii',i
1992) with primers that added a DraI site at each 5' end and an XbaI site at each 3' end. The DraI site allowed precise excision of the viral 5' termini, and the XbaI site facilitated subsequent assembly of full-length clones. An XbaI linker was inserted 3' to the cDNA covering the viral poly(A) tails, thus creating a unique XbaI site for linearization of the fulllength plasmids prior to in vitro transcription. The cDNA sequences were modified by site-directed in vitro mutagenesis (Sculptor, Amersham). A HindIII site was introduced at nt 1276 of P-l, and the P-4 sequence was modified in three positions, introducing a HindIII site at nt I234, a BamHI site at nt 2259 and a PstI site at nt 5812. All nucleotide substitutions were translationally silent. Full-length clones for in vitro transcription were assembled in the vector pTTE19( + ) (Petty, 1988), which was modified to remove the SphI and HindIII sites from the polylinker. The virus sequences were cloned directly behind the T7 promoter by digesting modified pT7E19( + ) with Sad, treating with T4 DNA polymerase and then digesting with XbaI before inserting the 5' terminus of either P-1 or P-4 as a DraI-XbaI fragment. These plasmids were digested with SphI and XbaI, and cloned cDNAs (covering Sph1965 to BamH12300 of P-1 and SphI 924 to BamHI 2259 of P-4) were inserted as SphI-XbaI fragments creating plasmids pPa and pPb, respectively. Cloned cDNAs covering HindIII 8647 to the poly(A) tail of isolate P-1 and HindIII 8585 to the poly(A) tail of isolate P-4 were inserted as BamHI-XbaI fragments 3' to both BamHI 2300 of pPa and BamHI 2259 of pPb resulting in pPc, pPd, pPe and pPf. pPg (containing cDNA covering BamH12300 to HindlII 8647 of P-l) and pPh (containing cDNA covering BamHI 2259 to HindlII 8585 of P4) were digested with BamHI and PstI, and the BamHI-PstI fragment of pPg was inserted into pPh, creating pPi. Full-length plasmids pP-1, pP-1114 and pP-4111 were assembled by

revealed too m a n y differences to conclude which part of the g e n o m e determined seed transmission (Johansen et al., I996). Therefore a study was undertaken to analyse the 10 kb R N A g e n o m e of P S b M V b y creating chimeras b e t w e e n the two virus isolates a n d analysing their transmission through seed in P. sativum cultivars k n o w n to transmit P S b M V at high frequencies.

Methods Virus i s o l a t e s . PSbMV isolate P-1 DPD1 was recovered from a pea
seed sample analysed at the Danish Plant Directorate (Lyngby, Denmark). The PSbMV P-4 NY isolate was recovered from USDA Pisum P.I. accession 471128 and kindly provided by R. Provvidenti (New York State Agriculture Experiment Station, Geneva, NY, USA). For convenience P-1 DPD1 and P-4 NY will be referred to hereafter as isolates P-I and P-4; however, it should be noted that seed transmissibility is not necessarily linked to the pathotype of PSbMV. Cloning and sequencing of P-1 and P-4 were reported by Johansen el al. (1991, 1996).

cDNA modificationsand construction of full-length clones.

A cistron map of PShMV is shown in Fig. 1 (a). Locations of natural and engineered restriction sites in the cDNA of isolates P-1 and P-4 are identified by the first nucleotide (nt) of the recognition sequence. Maps and names of isolates P-I and P-4 intermediary clones are shown in Fig. I (b). cDNA of the 5' termini of isolates P-1 and P-4 covering nt 1-Sph[ 965, and 1-SphI 924, respectively, were amplified by RT-PCR (Kohnen et at.,

(a)

/
c,

e'
6kl

P1Pr HC-ProI Pa II
I

IivPgl 6k~

I RdRp I CP ~poly(A)
I

%
(b) Dral Sphl BamHI Pstl Hindlll Xbal

4,

4,

4,
pPa

pPb
r-----~

pPc pPd

i i:

pPe

I ]I

pP! PPg

pPh
pPi

Fig. I. (o) Cistron map of PSbMV showing the noncoding regions (solid line), the open reading frame encoding the potyvirus polyprotein (open box) and putative proteolytic cleavage sites (vertical lines). Relevant restriction sites in the cDNA are shown above (P-I) and below (P-4) the cistron map. Restriction sites marked with an asterisk were engineered into the sequences. PI Pro, PI protease; HC-Pro, helper-component protease; P3, P3 protein; 6ki, 6 kDa protein I ; CI, cylindrical inclusion protein; 6k2, 6 kDa protein 2; 49k-Pro, 49 kDa protease; VPg, genome linked virus protein; RdRp, putative RNA dependent RNA polymerase; CP, coat protein. (b) Maps and names of intermediary clones used in the assembly of the full-length clones pP-1, pP-1114, pP-411 I, pP1144 and pP-4144. The P-I isolate sequences (light grey) and P-4 isolate sequences (dark grey) contained in the clones are shown as boxes, positioned according to the sequence in the genome they represent. Areas represented by a solid line are not present in the clones. Vector sequences are not shown.

(a)

(b)
Symptoms Virus acc. 0.66 c Inf./Tot. 116/464 1/505 136/366 93/521 28/614 44/626 12/729 % ST 25 b % Germination 94 b,c 91 c,d 96 a,b 88 d 98 a 95 a,b 95 a,b

P-1 P-4 vP-1 " ~4~ ~<~i~@~<~;~

H

Severe Mild Severe Severe Severe Mild Mild

0.37 d 0.71 b,c 0.34 d 0.93 a 0.81 a,b 0.34 d

0.2 d 37 18 a b

vP-1114 ~:<~:~-.'~'u~
B

vP-4111 ~ ; ~ : ~ & ~ ; ~ ; ~ ; ~ ; ~
P

4.6 c 7.0 c 1.6 d

vP-1144 ~t~
B

I
P

vP-4144

Fig. 2. (o) Cistron maps of isolates P-I (light grey), P-4 (dark grey), transcript derived vP-1 and chimeras vP-1114, vP-411 I, vP-1144 and vP-4144. (b) Symptoms and virus accumulation in infected P. sativum '549 ', seed transmission frequency and percentage germination of seeds from infected plants. The restriction sites B (BamHl), P (Pstl) and H (Hindlll) were used for assembly of the recombinant full length clones. Severe, P-I -like symptoms; mild, P-4-1ike symptoms. Virus acc., virus accumulation in the youngest fully expanded leaves of P. sativum 3 weeks after inoculation determined by ELISA (A4o5 values, mean of eight plants). Inf./Tot., total number of infected progeny seedlings detected/total number of viable progeny seedlings assayed (pooled from three separate experiments). % ST, percentage seed transmission detected (average of three separate experiments). Means followed by the same letter are not statistically different at the 95% level.

(a)
vP-1 P-4

(b)
Inf./Tot. 169/316 0/105 % ST 53 a

pPa with the HinclI I02-HindlII 1234 of pPb; and the HindIII 1276-BamHI 2300 of pPa with the HindIII 1234-BamHI 2259 of pPb, respectively, followed by assembly of the full-length clones as described for pP-I.

0.0 d 29 55 b a

vP-1 (p:4 5'UTR)

fi H
vP-1 (P-4 Pl pro)

54/184 57/103

H a vP-I(P-4 HCpro N) 18/210 8.6 c

Fig. 3. (a) Cistron maps of vP-1 (P-4 5'UTR), vP-1 (P-4 P/pro) and vP-1 (P-4 HCproN). The restriction sites B (BamHl), H (Hindlll) and Hi (Hincll) were used to exchange the 5' UTR, P/pro and HCproN in the recombinant full-length clones. (b) Seed transmission in P. sativum 'Vedette'. Inf./Tot., total number of infected progeny seedlings detected/total number of viable progeny seedlings assayed (pooled from two separate experiments). % ST, percentage seed transmission detected (average of two separate experiments). Means followed by the same letter are not statistically different at the 95% level.

In vitro transcription and inoculation of RNA. Full-length PSbMV cDNAs were cloned and amplified in Escherichia coli strain SURE (Stratagene). Caesium chloride gradient-purified plasmids were linearized with XbaI, and 2-3 lag of linearized plasmid was used for in vitro synthesis of capped transcripts (mMESSAGE MACHINE, Ambion). Approximately 2 lag of transcript was inoculated mechanically onto two carborundumdusted leaves of each P. sativum plant. Two weeks after inoculation, infectivity of the in vitro-synthesized transcripts was assayed by doubleantibody sandwich (DAS)-ELISA (Converse & Martin, 1990) using an antiserum specific for PSbMV coat protein (Hampton & Mink, 1989). Virus derived from plasmid pP-1 is referred to as vP-1, and chimeric viruses are referred to as vP-1114, vP-4111 etc. Accumulation of viruses in young, fully expanded, systemically infected leaves was compared by measuring (by DAS-ELISA) the accumulation of coat protein in plants infected with native isolates P-1 or P-4, or transcript-derived viruses, respectively. For each virus, eight plants were tested and confidence intervals (95%) were calculated, assuming a standard normal distribution of A40~ values. Determination of seed transmission frequency. The pea cultivar '549' was used for the first three seed transmission experiments under greenhouse conditions described by Kohnen et al. (1995). ' Vedette' was used for the fourth and fifth experiment because this cultivar matures early and transmits PSbMV P-1 through seed at very high frequencies (Wang et aI., 1993). lnoculum was raised in leaves of P. sativum '549' or 'Vedette' infected with in vitro-synthesized transcripts. Infection was confirmed by DAS-ELISA 2 weeks after inoculation. Leaves of infected plants were homogenized in buffer (50 raM-sodium phosphate pH 7"0) and used to inoculate 2-3-week-old P. sativum '549" or 'Vedette'.

inserting the BamHI-Hindi[I fragment of pPg into BamHI/HindIIIdigested pPc, pPe and pPf, respectively, pP-II44 and pP-4144 were assembled by inserting the BarnHI-HindIII fragment of pPi into BamHI/HindlII-digested pPe and pPd. The full-length clones are shown in Fig. 2 (a). Plasmids pP-I(P-4 5'UTR), pP-I(P-4 Plpro) and pP-I(P-4 HCpro ~) (Fig. 3 a) were created by substituting the HincI[ 102-BamH12259 of pPb with the HinctI I46--BamHr 2300 of pPa; the HinclI I4&-HindlII 1276 of

iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii!iii!!iii!iiii iii iiiiiii!ii!iiiiiiiiiiiiii iiii

Infectedplants were grown to maturityand seeds were harvested as the seeds matured. Seeds from infected plants were sown and seedling infection was determined by DAS-ELISA 3 weeks after emergence. Confidence intervals (95 %) were calculated for the probability of the observed seed transmissionfrequency,assuminga binomial distribution of the data. interference with host growth regulator metabolism and gene regulation (Fraser et al., 1986; Wang & Maule, 1995).

Results and Discussion

Infectivity of RNA transcripts from native and recombinant full-length cDNA clones Inoculation of P. sativum with in vitro-synthesized transcripts of pP-1, pP-1114, pP-4111, pP-1144 or pP-4144 (Fig. 2a), and pP-I(P-4 5'UTR), pP-I(P-4 Plpro) or pP-I(P-4 HCpro N) (Fig. 3 a) all resulted in infection of 50 to 100 % of the inoculated plants. In subsequent mechanical transfers, using infected plant material as inoculum, the infectivity of vP-1 and the chimeric viruses was comparable to native isolates P-I and P-4. In contrast, transcripts of pP-4 and reciprocal chimeras to those shown in Fig. 2 (a) were not infectious. All these plasmids contained cDNA covering nts 2259-5812 of P-4, suggesting that this cDNA segment was defective. However, substitution of this region with cDNA from a new cDNA synthesis reaction did not result in the generation of biologically active transcripts. Symptom induction and accumulation of transcriptderived viruses Symptoms induced by vP-1, vF-1114, vP-4111, vP-I(P-4 5'UTR), vP-I(P-4 Plpro) and vP-I(P-4 HCpro ) were similar to those of native P-1 virus (transient vein clearing, downward rolling of leaflets and shortened internodes). Symptoms of vP1144 and vP-4144 were mild, causing only a slight growth reduction similar to plants inoculated with native P-4 virus (Fig. 2 b). These observations suggest that the region of the PSbMV genome between PstI 5874/5812 and HindIII 8647/8585 has a major influence on symptom severity in P. salivum. This region contains the potyvira149 kDa protease and the putative RNA dependent RNA polymerase, both proposed to be involved in replication of the potyvirus genome (Riechmann et al., 1992). To determine whether the differences in symptom induction correlated with differences in virus accumulation, the relative accumulation of virus in systemically infected leaves was determined 3 weeks after inoculation (Fig. 2 b). vP-4111 accumulated to a higher concentration than P-l, vP-1 and vP-1144, which, in turn, accumulated to higher concentrations than P-4, vP-1114 and vP-4144. These data demonstrate that P-l, P-4 and the chimeras accumulated to different levels in vegetative tissues, but that the accumulation of virus was not correlated with symptom induction. Therefore, the symptoms induced were not merely a result of competitive inhibition of host growth, but rather might have been due to an

Seed transmission of vP-1114, vP-1144, vP-4144 and vP-4111 in P. sotivum ~549' The seed transmission frequencies of vP-1114, vP-1144, vP-4144 and vP-4111 were determined in three experiments. The absolute seed transmission frequency varied between experiments, which was probably a result of seasonal changes affecting greenhouse conditions. However, in all three experiments, vP-1 was transmitted at the highest frequency, followed by P-l, vP-1114, vP-1144, vP-4111 and vP-4144. P-4 was transmitted in only one seed. Results of the combined data are shown in Fig. 2 (b). The seed transmission frequency of vP-1 was higher than P-1 in all three experiments. The difference, which was significant (P < 0"05), could be a result of the maintenance of P-1 by repeated mechanical transfers without selection for seed transmissibility. In contrast, vP-1 was derived from a cDNA of viral RNA isolated from P-1 shortly after the virus isolate was first recovered from an infected seed sample. The seed transmission frequencies of the chimeric viruses were intermediate between P-1/vP-1 and P-4. Of the chimeras, vP-4144 was transmitted at the lowest frequency (slightly, but not significantly, higher than P-4). The increasing contribution of sequences from P-1 found in vP-4144, vP-1144 and vP-1114 correlated with increasing seed transmission, demonstrating that seed transmission of PSbMV is influenced by determinants contained in each of the exchanged genome fragments. The 5' region (nt 1-2259) of P-4 appeared to have the strongest influence on the seed transmission frequency, reducing the seed transmission frequency of vP-4111 to 4"6%, compared with 37% for vP-I. The seed transmission frequency was not correlated to the virus concentration in infected vegetative tissues (Fig. 2 b). This is in agreement with results obtained by Wang et al. (1993) who found no obvious relationship between virus content and the efficiency of seed transmission of PSbMV in different pea cultivars. Also, Ligat & Randles (1993) observed that repeated transfers through seed in the cultivar 'Dundale' resulted in a gradual reduction of PSbMV accumulation in vegetative tissue to a level where it was not detectable by ELISA, while the seed transmission frequency exceeded 90 %. The percentage germination of seeds from infected plants ranged from 88% (pP-1114) to 98% (pP-4II1). However, there was no apparent correlation of seed viability with symptom severity, virus accumulation or seed transmission frequency. Seed transmission of vP-1 (P-4 5'UTR), vP-1 (P-4 P1 pro) and vP-I (P-4 HCpro N) in P. sotivum 'Vedette'
As described above, the region of isolate P-4 from nt 1-2259, which covers the 5'-untranslated leader (UTR), the P1 protease (Plpro) and the N-terminal two-thirds of the helper-component protease (HCpro ~) appeared to have a

major effect on seed transmission. Three recombinant, fulllength clones were therefore created in which the 5' UTR, the Plpro and HCpro N of P-1 were exchanged with the corresponding region of P-4. Seed transmission of the resulting chimeric viruses was assayed in 'Vedette', and the combined data from experiments four and five are shown in Fig. 3 (b). The seed transmission frequencies of vP-I(P-4 5'UTR) and vP-I(P-4 HCpro N) were reduced to 50% and 20% of vP-1, respectively, while vP-I(P-4 Plpro) was seed transmitted at the same frequency as vP-1. The 5' UTRs of P-1 and P-4 are 143 and 99 nt in length, respectively, and there are 32 amino acid differences in the HCpro N region between the two viruses. The 5' UTR is likely to affect both translation and replication of the potyvirus genome (Riechmann et al., 1992). The region of HCpro contained in HCpro N affects aphid transmission, replication (Atreya et al., 1992) and long-distance movement (Cronin et al., 1995). Wang & Maule (1994) demonstrated that the suspensor can act as a channel for transmission of PSbMV to the embryo proper. A prerequisite for embryo infection is, therefore, that virus reaches the micropylar region of the testa before disintegration of the suspensor. This suggests that the HCpro ~ region may affect replication and/or long-distance movement in reproductive tissues and thus influence the number of embryos which become infected and hence, the final seed transmission frequency. The observation that seed transmission is affected by multiple determinants in the viral genome agrees well with the results of W a n g & Maule (1994) who demonstrated that resistance to seed transmission in P. sativum involves several nuclear genes. This suggests that several virus-host factor interactions determine the spread and accumulation of PSbMV in the maternal testa tissue. Whereas PSbMV can enter the embryo after fertilization (Wang & Maule, 1992), seed transmission of BSMV depends mainly on the ability of the virus to reach the reproductive tissues prior to fertilization (Edwards, 1995). Despite this difference, replication and movement also play central roles in seed transmission of BSMV, the major determinants being the RNA~ 5' UTR, a 369 repeat in the 7a gene, and the 7b gene. Therefore, seed transmission of both PSbMV and BSMV appears to be determined by the ability to replicate and move in the reproductive tissues and to reach the suspensor or megagametophyte, respectively, before a certain critical point in development. We thank I. T. D. Petty for pT7E19(+) and Kristian Kristensen for statistical advice and analysis.

Carroll, T.W. (1981). Seedbome viruses: virus-host interactions. In

Plant Diseases and Vectors: Ecology and Epidemiology, pp. 293-317. Edited by K. Maramorosch & K. F. Harris. New York: Academic Press. Carroll, T. W., Gossel, P. L. & Hockett, E.A. (1979). Inheritance of resistance to seed transmission of barley stripe mosaic virus in barley. Phytopathology 69, 431-433. Converse, R.H. & Martin, R.R. (1990). ELISA methods for plant viruses. In Serological Methods for Detection and Identification of Viral and Bacterial Plant Pathogens, pp. 179-I96. Edited by R. Hampton, E. Ball & S. DeBoer. St Paul, Minn. : The American Phytopathological Society.
Cronin, S., Verchot, J, Haldeman-Cahill, R., Schaad, H.C. & Carrington, J. C. (1995). Long-distance movement factor: a transport

function of the potyvirus helper component proteinase. Plant Cell 7, 549-559. Edwards, M. C. (1995). Mapping of the seed transmission determinants of barley stripe mosaic virus. Molecular Plant-Microbe Interactions 8, 906-915. Fraser, R.S.S., GerwJtz, A. & Morris, G. E. L. (1986). Multiple regression analysis of the relationships between tobacco mosaic virus multiplication, the severity of mosaic symptoms, and the growth of tobacco and tomato. Physiological and Molecular Plant Pathology 29, 239-249. Hampton, R. O. & Mink, G. I. (1989). Investigation of purported nonELISA-detectable latent pea seedborne mosaic virus infection in selected U.S. commercial seedlots. Pisum Newsletter 21, 26-28. Hampton, R.O. & Francki, R. I. B. (1992). RNA-I dependent seed transmissibility of cucumber mosaic virus in Phaseolus vulgaris. Phytopathology 82, 127-I30. Hanada, K. & Harrison, B. D. (1977). Effects of virus genotype and temperature on seed transmission of nepoviruses. Annals of Applied Biology 85, 79-92. Johansen, E., Rasmussen, O. F., Heide, M. & Borkhardt, B. (1991). The complete nucleotide sequence of pea seed-borne mosaic virus RNA. Journal of General Virology 72, 2625-2632.
Johansen, I.E., Keller, K.E., Dougherty, W.G. & Hampton, R.O. (1996). Biological and molecular properties of a pathotype P-1 and a

References
Atreya, C. D., Atreya, P. L., Thornbury, D. W. & Pirone, T. P. (1992).

Site-directed mutations in the potyvirus HC-PRO gene affect helper component activity, virus accumulation, and symptom expression in infected tobacco plants. Virology 191, 106-111.

pathotype P-4 isolate of pea seed-borne mosaic virus. Journal of General Virology 77, 1329-1333. Kohnen, P. D., Dougherty, W. G. & Hampton, R. O. (1992). Detection of pea seedborne mosaic potyvirus by sequence specific enzymatic amplification. Journal of Virological Methods 37, 253-258. Kohnen, P.D., Johansen, I.E. & Hampton, R.O. (1995). Characterization and molecular detection of the P4 pathotype of pea seedborne mosaic potyvirus. Phytopathology 85, 789-793. Ligat, L S. & Randles, J. W. (1993). An eclipse of pea seed-borne mosaic virus in vegetative tissue of pea following repeated transmission through the seed. Annals of Applied Biology 122, 39-47. Haule, A. J. & Wang, D. (1996). Seed transmission of plant viruses: a lesson in biological complexity. Trends in Microbiology 4, 153-158. Mink, G. I. (1993). Pollen- and seed-transmitted viruses and viroids, Annual Review of Plant Pathology 31, 375-402. Petty, I. T. D. (1988). A plasmid vector for doning directly at the transcription initiation site of a bacteriophage T7 promoter. Nucleic Acids Research 16, 8738. Riechmann, J.L., Lafn, S. & Garcfa, J.A. (1992). Highlights and prospects of potyvirus molecular biology. Journal of General Virology 73, 1-16. Wang, D. & Maule, A. J. (1992). Early embryo invasion as a determinant

}15~

in pea of the seed transmission of pea seed-borne mosaic virus. Journal of General Virology 73, I615-I620. Wang, D. & Maule, A. J. (1994). A model for seed transmission of a plant virus: genetic and structural analysis of pea embryo invasion by pea seedborne mosaic virus. Plant Ceil 6, 777-787. Wang, D. & Maule, A.J. (1995). Inhibition of host gene expression associated with plant virus replication. Science 267, 229-231.

Wang, D., Woods, R. D., Cockbain, A. J., Maule, A. J. & Biddle, A. J. (1993). The susceptibility of pea cultivars to pea seed-borne mosaic virus infection and seed transmission in the UK. Plant Pathology 42, 42-47.

Received 13 June 1996; Accepted 22 August 1996