The present invention relates to a class of « ribozymes », i. e. nucleic acid molecules having endoribonuclease activity, which are capable of conferring on plants, resistance to infection by potyviruses. The invention also relates to plants and plant cells expressing the ribozymes, and showing resistance or tolerance to potyvirus infection.

Potyviruses belong to the picorna-like supergroup of viruses, and within this supergroup, to the family of the potyviridae. Potyviruses are mostly aphid-, transmitted, although some are transmitted by mites, whiteflies or fungi. They are one of the world's most widespread plant pathogens, causing especially severe disease in cucurbits in many countries.

Potyviruses include : Zucchini yellow mosaic virus (ZYMV), Watermelon mosaic virus 2 (WMV2), Tobacco vein mottling virus (TVMV), Tobacco etch virus (TEV), Potato virus Y (PVY), Pepper mottle virus (PepMoV), Soyabean mosaic virus (SbMV), Papaya ringspot virus (PRSV), Pea seed-borne mosaic virus (PSbMV), Turnip mosaic virus (TuMV), Johnson grass mosaic virus (JGMV), Plum pox virus (PPV, Maize Dwarf Virus (MDV), Bean common mosaic virus (BCMV), Bean yellow mosaic virus (BYMV), Carnation vein mottle (CarVMV), Celery mosaic virius (CeMV), Clover yellow vein virus (C1YW), Henbane mosaic virus (HMV), Lettuce mosaic virus (LtMV), Sugarcane mosaic (SCMV).

Potyviruses regarde as strains of the above potyviruses, or of uncertain relationship to the above viruses, and included in the definition of « Potyviruses » for the purposes of the present invention, include : Clover mosaic virus, Datura mosaic virus, Desmodium mosaic virus, Freesia mosaic virus, Mungbean mosaic virus, Passionfruit ringspot virus, Pea necrosis virus, Sweet potato A virus, Wild potato mosaic virus. Possible members of the Potyvirus family include : Carrot mosaic virus, Celer latent virus, Palm mosaic virus and fern virus.

Of these viruses, WMV2, ZYMV and PRSV are responsible for severe yield reduction in Cucurbit crops. Complete resistance to infection by potyviruses, rarely occurs in economically important crops, although tolerance, in which plants are susceptible to infection but either remain symptomless or only slowly develop inconspicuous symptoms and often support restricted virus replication, is more common.

The establishment of a means to impart resistance to potyvirus infection and/or disease development would clearly have considerable economic importance.

In addition to breeding for resistance, a number of approches using genetic engineering have been explore in an attempt to provide resistance to potyviruses.

For example, Van der Vlugt (Plant Molecular Biology, 1992, vol. 20 : 631-639) produced transgenic tobacco plants expressing both translatable and non- translatable Coat Protein (CP) RNA from PVY. Full protection (i. e completely free of systemically spread virus) following mechanical inoculation of transformed tobacco plant was observe for some plants, whether or not the CP RNA was translated. The work described did not extend to testing for resistance against inoculation by aphids, or against other related and unrelated potyviruses.

Immunity to infection by ZYMV in transgenic muskmelon plants expressing the full-length ZYMV CP gene has been reporte by Fang and Grumet (Mol. Plant- Microbe Interactions, 1993, vol. 6 n°3 : 358-367).

Protection was also exhibited by plants expressing either a truncated, conserve core of the CP gene, or CP antisense RNA, although these constructs were reporte to be less efficient then the full-length sense CP gene. These results are in contrast to those obtained by Lindbo and Dougherty (Mol. Plant-Microbe Interactions, 1992, vol. 5 : 144-153) where expression of antisense and translationally deficient sense TEV CP RNA resulted in much higher protection against TEV infection than that conferred by expression of the sense version of the full-length CP gene. The effectiveness of coat protein and antisense techniques in obtaining potyvirus resistance therefore appears to have met with varying success depending upon the precise constructs and viruses involved.

Another approach has been described by Van der Vlugt et al (Virus Research, 1993, vol. 27 : 185-200) who synthesized two ribozymes against a relatively conserve region of the RNA-dependent, RNA-polymerase- encoding cistron (also known as the NIb cistron) of PVY. The target triplet of these ribozymes is GUC at position 8122-8124 of PVYn as described in Robaglia (1989). These ribozymes were tested in vitro against an artificial PVY-specific target RNA. Whilst the « long- arm » ribozyme (total arm length 106 bases) cleaved the substrate RNA efficiently, the « short-arm » ribozyme (total arm length 19 bases) failed to cleave. This short-arm ribozyme was however capable of cleaving a much shorter PVY specific RNA transcript. The authors conclue that the cleavage activity of the ribozyme is in part determined by the substrate involved, long substrate RNA's such as full-length viral RNA's being likely to possess extensive secondary structure which prevents the formation of the correct ribozyme hammerhead structure and thus prevents cleavage. No in vivo tests are described for these ribozymes.

The technical problem underlying the present invention is to provide a means for conferring protection, on plants, to potyvirus infection, or disease development. Advantageously, the means provide protection to a large number of host plants against a particular potyvirus and to serologically related potyviruses, preferably in the heterozygous state.

The present invention provides a solution to the above technical problem by providing nucleic acid molecules or « ribozymes » having endoribonuclease activity which are designed to specifically cleave full-length potyviral RNA, (either (+) sense genomic RNA and/or the (-) sense replicative strand RNA).

The ribozymes of the present invention may comprise single or multiple catalytic regions (« mono- » or « poly- » ribozymes).

The present invention relates to a nucleic acid molecule, called « ribozyme », having endoribonuclease activity, comprising at least one hybridising region and at least one catalytic region, said hybridising region being complementary to one or more target regions of the genomic (+) RNA of a potyvirus, and/or to one or more target regions of the replicative (-) strand RNA of a potyvirus. Advantageously, the said target region does not contain the GUC triplet located at position 8122-8124 of the PVYn strain or the equivalent triplet of other potyviruses.

More particularly, the invention relates to a nucleic acid molecule, called polyribozyme, having endoribonuclease activity, comprising at least first and second hybridising regions and at least two catalytic regions, said first and second hybridizing regions being complementary to first and second target regions, respectively, in the genomic (+) RNA of a potyvirus, or in the replicative (-) strand RNA of a potyvirus, said first and second target regions being either : -within or comprising different genes or untranslated regions on the same RNA strand, or -on different RNA stands, or -within the same gene on the replicative (-) strand RNA.

The resistance provided by the ribozymes of the present invention is, for a number of reasons, unexpected on the basis of previous knowledge.

Specifically, it has been shown both by the present inventors and by other authors (Edington and Nelson, Gene Regulation-Biology of Antisense RNA and DNA : Ed. Erickson and Izant, Raven Press Ltd. New York, 1992), that in-vitro cleavage results obtained with ribozymes are not always predictive of in-vivo situations. In vitro data obtained by Van der Vlugt (1993 supra) can therefore not be reliably extrapolated to in vivo situations. Furthermore, use of ribozymes on full-length viral RNA could not be expected to work efficiently in view of the secondary structure associated with such long RNA, as pointed out by Van der Vlugt et al (1993 et al supra).

Moreover, some authors have postulated that the stand RNA may be protected by a protein coat during the early stages of infection (« stripeosome » hypothesis) until replication, and therefore would not be an appropriate target for ribozymes.

Many of the present ribozymes give rise to resistance in the heterozygous state, thereby facilitating transmission of the resistance characteristic to progeny of transgenic plants.

In addition, the polymeric ribozymes of the invention are capable of conferring resistance in vivo to virus strains related to, but different from the particular targeted trains.

Finally, many of the polymeric ribozymes of the invention target sites which are either on different strands of RNA (i. e (+) and (-)), or have sites which are widely separated regions on the same strand. The ribozymes thus do not only act as a single molecule simultaneously targeting a plurality of sites, but rather can act as if the molecule were a plurality of monoribozymes each targeting a single site, thus increasing the probability that the substrate will be cleaved.

In the context of the present invention, the following terms have the following meanings : Potyvirus. A large group of PLANT VIRUSES in which the virions are flexuous filaments (ca. 680-900 x 11 mm) comprising one type of coat protein (MWt ca. 32,000- 36,000) and one molecule of linear positive-sense ssRNA (MWt ca. 3.0-3.5 x 106) comprising a single large open reading frame (ORF) coding for a polyprotein of ca.

350.000 Da which is proteolytically processed by virally encoded proteinases.

As a group, the viruses infect a wide range of host plants, including monocots and dicots, but individual members typically have a narrow host range.

Symptoms depend on plant cultivar, virus strain, and environmental conditions, but typically include mosaic or mottle in the leaves, colour-breaking in flowers, mottling and/or distortion in the fruits ; characteristic proteinaceous inclusion bodies (serologically unrelated to virus coat protein) are formed in the cytoplasm of infecte plant cells, appearing as « pinwheels » in transverse section or as « bundles » in longitudinal section. A few members also form crystalline nuclear inclusions. Potyviruses are transmitted (non-persistently) by aphids and can be transmitted mechanically; some viruses tentatively included in the group are transmitted by whitefly, mites or fungi.

The replication cycle of potyviruses involves the following successive steps : infection, uncoating, translation, polyprotein processing to yield functional, viral coded products, genome replication via a negative strand RNA template, progeny virus particle assembly, cell-to-cell movement within the infecte plant and vector transmission from plant to plant.

Gens : In the context of potyviral genomic RNA, the term « gene » as used herein is synomynous with « cistron". All potyviral proteins are generated from a single large primary translation product (polyprotein) by viral protases, produced by translation of the single genomic ORF. In the present context, « gene » signifies a region of the genomic ORF coding for a protein released as an independent molecule by viral protases. Potyviral « genes » or cistrons are referred to herein by reference to the designation accepte in the art for the type members of the potyviruses family PVY as shown in figure 20, or their functional equivalent in other members of the family. Reference to genes or cistrons in the (-) strand signifies the region corresponding to (i. e. complementary to) that gene or cistronin the (+) strand.

(+) sense RNA : In the context of potyviral RNA, « (+) sense RNA » signifies the genomic RNA of the virus as contained within the virion, which serves directly as a template for translation of a polyprotein. The (+) sense RNA may be encapsidated or not. Naked (+) sense RNA is also infectious.

Replicative or (-) sense RNA: In the context of <BR> <BR> <BR> <BR> potyviral RNA, « (-) sense or replicative RNA » signifies the full-length, non-coing negative strand template RNA produced by the copying of the entire (+) RNA genome by a viral polymerase (NIb). No sub-genomic RNAs are produced in potyviruses. In the context of the present invention, positions of nucleotides in the viral (-) sense RNA are indicated by reference to the corresponding position of the (+) strand, with a (-) in rackets. Reference to genes or cistrons in the (-) strand (which is in fact non-coding) signifies a region or a site in the (-) strand corresponding to that region or site in the (+) strand.

Potyviral RNA : In the present context, this expression encompasses both (+) sense genomic RNA and (-) sense replicative RNA.

Resistance: reduction or prevention of viral infection e. g. absence of symptoms and absence of detectable viral particles, absence or reduction or delay of apparition of symptoms. Symptoms are sometimes atypical. Recovery possible.

Hybridisinq/complementary : the « arms » of the ribozyme are said to « hybridise » to a target RNA sequence or to be « complementary » to the target region in the potyvirus. Such hybridisation and complementarity is defined herein as an extent of duplex formation sufficient to form a stable enough structure so that cleavage adjacent to the target triplet can occur. Depending on the target sequence, the ribozyme and the secondary structure of the RNA, preferably less than about 5-10, more preferably less than about 3 to 5 and even more preferably less than 1 to 3 bases would be mismatched between the target RNA sequence and one or both hybridising arms of the ribozyme. Over the whole length of the hybridising arm, not more than 10% of mismatches should occur.

Accordingly, the present invention extends to ribozymes described herein, to the polyribozymes containing them, and to any or all functional mutants, derivatives, parts, homologues or analogues thereof. Examples of such mutants include single or multiple nucleotide substitutions, deletions and/or additions to the hybridising regions and/or catalytic regions of the ribozymes provided such modifie ribozymes are still functional in cleaving the target RNA sequence, and/or conferring resistance to potyviruses. Such mutants are said to exhibit « substantial complementarity » to the viral target sequence.

The present invention therefore extends to a ribozyme having one or a plurality of hybridising domains and catalytic domains, wherein the hybridising domain comprises a sequence of nucleotides substantially complementary to a sequence of nucleotides in a potyvirus RNA in vitro under high stringency conditions and wherein said catalytic region is capable of cleaving said target RNA sequence. This aspect of the present invention is conveniently determined by the extent of cleavage of the target RNA sequence in vitro.

For the purposes of the present invention, the stringency conditions set forth in Sambrook et al, Molecular Cloning : A laboratory Manual, Cold Spring Harbor laboratories, Cold Spring Harbor, NY, USA, pages 387-389 at paragraph 11 are considered high stringent conditions. Alternatively, medium stringency may be used involving 0.25-0.5 % w/v SDS at 45°C for 2-3 hours. The alternative conditions are applicable depending on concentration, purity and source of nucleic acid molecules.

Hybridisinq reqion : the total length of hybridizing arms. Normally, in the context of the invention, when a hybridizing region is said to be complementary to a region within or comprising a gene, this is intended to signify that the site at which endoribonucleolytic cleavage occurs is within that gene.

Target region : In the context of potyviral RNA, « target region » signifies a stretch of ribonucleotides, normally having a length of from approximately 10 bases up to the length of the full potyviral RNA, to which the hybridising region of the ribozymes of the invention hybridise. The target region contains a precise target site at which cleavage can occur when acted upon by the ribozyme. Generally speaking, the extent of the « target region » is defined as the length of the sequence in the viral RNA over which hybridisation occurs.

Tarqet triplet : In the context of hammerhead ribozymes, « target triplet » signifies the three bases immediately 5' to the cleavage site in the substrate RNA. The target triplet will be noted herein as X°X°X° wherein X°-any ribonucleotide, preferably X°UX° wherein X° is any ribonucleotide and U is uracil, or X°UY wherein Y is adenine, cytosine or uracil.

Targe sites in (-) sense RNA: Ribozyme target sites in the potyviral (-) strand replicative RNA are designated by reference to the corresponding cistron (or « gene ») in the (+) strand.

The potyviruses targeted by the ribozymes of the present invention are preferably : Zucchini yellow mosaic virus (ZYMV), Watermelon mosaic virus 2 (WMV2), Tobacco vein mottling virus (TVMV), Tobacco etch virus (TEV), Potato virus Y (PVY), Pepper mottle virus (PepMoV), Soyabean mosaic virus (SbMV), Papaya ringspot virus (PRSV), Pea seed-borne mosaic virus (PSbMV), Turnip mosaic virus (TuMV), Johnson grass mosaic virus (JGMV), Plum pox virus (PPV), Maize Dwarf Virus (MDV).

Strains of ZYMV and WMV2 as well as serologically related viruses are particularly preferred.

Target regions of potyviral RNA which are particularly preferred are those cistrons (and the corresponding regions of (-) sense RNA) which give rise to proteins involved in viral processing, replication, and assembly such as NIa, NIb and CP. These cistrons are located in the 3'-half of the genomic RNA.

According to a preferred embodiment, the nucleic acid molecules of the invention comprise at least one unit « R » having the general formula : <BR> <BR> <BR> <BR> <BR> <BR> <BR> f(X)n-Q-(X')nlp<BR> <BR> <BR> <BR> <BR> wherein : -X and X' each represent, independently, any ribonucleotide ; (X) n and (X') n'represent ribonucleotide sequences substantially complementary to at least part of said target region of potyvirus genomic (+) RNA or to at least part of its (-) strand replicative RNA; -n and n'are each equal to or greater than 4 ; - Q represents the catalytic region of a ribozyme ; p > 1.

In the above general formula (X) n and (X')n' represent the complementary sequences or « arms » of the ribozyme which hybridise to the chosen target region of the potyvirus. (X) n and (X')n' together represent the « hybridizing region » of the ribozyme.

(X) n and (X') ni normally hybridise to the sequences immediately 5'and 3'respectively of the target site in the viral RNA.

The values of n and n'are, independently, at least 4, preferably at least 8, for example from 4 to 4000. Preferred values are from 4 to 50, for example 6 to 50 or 8 to 12 nucleotides for « short arm » ribozymes. The total length of the two hybridising arms is at least 8 nucleotides, preferably at least 16. For « long arm » ribozymes, n and n' normally have values greater than 50, for example 50 to 400, particularly 60 to 300 nucleotides.

In the above general formula, p 1 for monoribozymes. For polyribozymes, p ! 2, for example 2- 20 or 4-10. Polyribozymes are particularly preferred.

Preferably, the complementary sequences (X) n and (X')n' are complementary to target regions of potyvirus RNA which are conserve between strains of a given species, and preferably also between different species of potyvirus. In the context of the present invention, « conserve » signifies that the sequences on at least one side, and preferably on both sides of the cleavage site in the substrate are essentially identical over a length of 20 bases, i. e. have at least 85% identity, and preferably at least 90% identity over 20 bases.

Generally speaking, in a conserve target region, the target triplet itself is identically conserve between two viruses. Exceptionally, since the third base in the target triplet X°X°X° remains unpaired after hybridisation of the ribozyme to the target region in the substrat, this base may vary between two viruses in a conserve region without affecting cleavage capacity.

According to yet a further embodiment of the invention, the hybridising arms of the polyribozyme may be derived from at least two different types of virus.

Such polyribozymes are referred to hereinafter as « chimeric polyribozymes ». As an example of this type of construction, mixtures of ZYMV and WMV2 derived sequences can be cited. This type of polyribozyme ensures the inactivation of both types of virus, if present together, or allows the use of a single construction in different geographical areas.

According to a preferred embodiment, the catalytic region « Q » in the above general formula is a hammerhead ribozyme catalytic region. Such ribozymes are described in European patent EP-B-0 321 201. Such a ribozyme can be represented as a molecule of formula (I): wherein, (X) n and (X) nl represent ribonucleotide sequences substantially complementary to an RNA sequence of a potyvirus (either genomic or replicative intermediate); an (*) represents a base pair between complementary ribonucleotides; Y represents any ribonucleotide and each Y may be the same or different; y'and Y"represent ribonucleotides of complementary sequence along at least part of their length to allow base pairing between the oligoribonucleotides, or Y'and Y"together form a single RNA sequence wherein at least part of said sequence comprises a stem formed by base pairing between complementary nucleotides; and optionally, an additional nucleotide selected from any one of A, G, C or U is inserted after 1A.

In another embodiment, the ribozyme is as set forth in formula (II) : wherein (X) n and (X) n I represent ribonucleotide sequences substantially complementary to an RNA of a potyvirus; Y represents any ribonucleotide and each Y residue may be the same or different; an (*) represents a base pair between complementary ribonucleotides; n and n'are as previously defined; B represents a bond, a base pair, a ribonucleotide, or an oligoribonucleotide containing at least 2 ribonucleotides; m and m'are 1 or more and may be the same or different; and optionally, an additional nucleotide selected from any one of A, G, C or U is inserted after 1A.

In a preferred embodiment, the ribozyme is as set forth in formula (III) : wherein (X) n and (X) nl represent a ribonucleotide sequence substantially complementary to an RNA of a potyvirus; Y represents any ribonucleotide and each Y residue may be the same or different; an (*) represents a base pair between complementary ribonucleotides; n and n'are as previously defined.

According to further embodiments of the invention, the catalytic region of the ribozyme, for example"Q"in the above general formula may be a hairpin domain as described in WO 95/06731 and EP-B- 360257. An example of such a hairpin domain is given below in Formula IV : TARGET SEDUENCE RIBOZYME 5. 3' U A U A U U A C A'N G U C G U G G U C U G G U N N N N N N N N N N 3' C A C C A A A G A 6 A C C A H'H'N'N'N'N'N'N'N'N'S A A G A Alternatively, the catalytic region of the ribozyme may be the catalytic domain of a Hepatitis delta virus, as described in WO 95/06731, or the catalytic domain of an RnaseP ribozyme as described in EP 546092.

Although the catalytic regions illustrated above have a conserve structure and sequence, it has been observe that some nucleotides may be deleted, inserted, substituted or modifie without prejudice to the activity of the ribozyme. The invention comprises the use of these modifie catalytic regions in the polyribozyme provided that their catalytic activity is conserve. This activity can be verified by using the tests described below.

For example, one or more nucleotides of the catalytic region II illustrated in formula I, II and III may be replace by nucleotides containing bases such as adenine, guanine, cytosine, methylcytosine, uracil, thymine, xanthine, hypoxanthine, inosine or other methylated bases. The « conserve » bases C-G which together form the first base pair of the catalytic loop, can be replace by U-A.

Other preferred variant hammerhead catalytic regions have the following sequences : Mutant ribozyme I : cugaugaGUCCugaGGACgaa Mutant ribozyme II : cugaugaGUCCcgugaGGACgaa Mutant ribozyme ICI : cugaugaGUCgugagGACgaa Mutant ribozyme IV : cugaugaGCGgugaCGCgaa.

Wild type sTobRV : cugaugaGUCCgugaGGACgaa (Helix II shown in upper case lettes) Because ribozymes act in a sequence-specific manner, it is possible for a target RNA to become resistant to ribozyme-mediated inactivation by just changing one or a few critical bases. For the present application, it was also desirable to build resistance genes which would be protect plants against infection by several strains of a virus species. Both the goals of achieving a robust resistance gene, and obtaining protection against many trains, were achieved by selecting the target sites for the ribozymes according to the following criteria : -sites are chosen which cannot easily mutate without changing the coding capacity (i. e. the amino acids corresponding to a particular codon); -sites are chosen which are located in highly conserve regions.

In a preferred embodiment of the invention, the 3' (X) n sequence of a given R unit and the 5' (X') nl sequence of the adjacent R unit are complementary to sequences which are contiguous in the potyvirus genomic stand RNA or replicative (-) strand RNA.

According to this variant, the complementary sequence may hybridise with the entire length of the target transcript. On the other hand, it may hybridise with only a portion or region of the targeted transcript. The portion or region in question must be long enough to allow the inclusion of at least 1 catalytic region in the corresponding sequence of the polyribozyme. In general, for this variant, the length of the complementary sequence, not counting the catalytic regions (the sum of the hybridising arms), may vary from about 4 to 4000 bases. Lengths of 40 to 1000 are particularly preferred.

According to another highly preferred embodiment of the invention, in the polyribozyme having the formula : R1-R2-R3-------Rn' <BR> <BR> <BR> <BR> <BR> <BR> <BR> <BR> the 3'hybridising arm (X') n of a given « R » unit and the 5'hybridising arm (Xl) nl of the adjacent « R » unit are complementary to non-contiguous portions of the genome or of the (-) sense RNA transcript. An example of this type of structure is the « short arm polyribozyme ».

The structure of these non-contiguous polyribozymes can be thought of as a sequence complementary to the target sequence in which ribozyme catalytic regions are embedded and which has undergone deletions in the complementary sequence. According to this embodiment, the distance between two catalytic regions in the polyribozyme is normally shorter than the distance between the two corresponding X°X°X° sites in the target sequence. The distance between ribozymes can be reduced significantly with respect to the distance between the corresponding target sites in the substrat, for example, if two adjacent targeted X°UX° sites in the substrate are separated by a distance of 1000 nucleotides, the distance between the catalytic regions of the « short arm ribozyme » may be as little as 30 or 40 nucleotides. Reductions of over 90% of the length of the hybridising sequence may therefore be made. The minimum length between two adjacent catalytic regions is normally about 8 nucleotides and preferably at least 20. The present inventors have shown that such shortened hybridising sequences do not, surprisingly, prevent the polyribozyme from efficiently cleaving the substrate and can, in fact, in many cases improve the efficiency of the inactivation.

For example, the reduction in the length of the complementary sequence avoids problems associated with the formation of competing or unfavourable secondary structure leading to the inactivation of the ribozyme.

Thus, for a given length of ribozyme molecule, more catalytic regions may be inserted targeting more cleavage sites in the substrat.

Furthermore, the use of the short arm polyribozymes means that efficent polyribozymes may be constructed even if only parts of the sequence of the substrate are known. It is sufficient to know the 10 or 15 bases surrounding an X°X°X° target site.

Another important avantage is the possibility to use, as the shortened hybridising arms, exclusively those sequences which are conserve in different viruses, for example viruses of the same type. In this way, different viruses may be inactivated by the same polyribozyme even if the sequences between the conserve regions are totally dissimilar. These viruses may be of the potyviridae family or species within the family or different isolates of the same species.

Finally, the smaller molecule allows an easier and biologically more economical synthesis. The functional nature of the « short arm » ribozymes of the invention proves that, in a polyribozyme, the different catalytic regions can act independently of each other.

Long-arm ribozymes, having arm lengths of at least 50 nucleotides, either contiguous or non- contiguous, are also preferred embodiments of the invention. Examples are shown in Table 8 below.

Any one polyribozyme of the invention may include a mixture of « long arm » and « short arm » ribozymes, and also of « contiguous » and « non-contiguous » ribozymes. Alternatively, it may include only long arm ribozymes, or only short arm, or only contiguous, etc... The structure of the ribozyme can therefore be optimise by testing using the in vitro-in vivo and transient assay described below.

According to a further variant of the invention, the polyribozymes may inclue, between two consecutive « R » units, a non-complementary sequence (A) which has a length of at least 1 base, for example 3 or more bases, and which is non-complementary to the (+) or (- ) sense potyviral RNA.

The non-complementary sequence of the polyribozyme may have a precise function, for example, it may be constituted by a coding sequence which can be used to select transformants or also a sequence containing a ribozyme which acts on a substrate other than the primary target of the DNA virus or which is cis-acting on a part of the polyribozyme. It typically contains a polylinker for cloning, or an adaptor to facilitate constructions of chimeric polyribozyme. The non-complementary sequence « A » usually has a length comprise between about 2 and 500 bases, for example 20 to 100 bases. When there is a plurality of non- complementary sequences, they may together constitute as much as about 90% of the length of the polyribozyme, more preferably 50%.

Preferred ribozymes of the present invention have multiple catalytic domains (« polyribozyme ») targeting different regions or different RNA strands of the potyvirus. According to their construction (long arm or short arm, contiguous or not contiguous), the polyribozyme may act as a series of individual monoribozymes or may act as a « unimolecule » simultaneously against each of its target sites.

In particular, according to a preferred variant, in a polyribozyme, two ribozyme units R1 and R2 target either, different cistrons (or « genes ») on the same RNA strand, or they target the same or different cistrons on different RNA stands. In such systems, the targeted RNAs differ from each other, either in that they comprise, or are within different cistrons, and/or in that they are within different stands.

Alternatively, the targeted sequences may be within the same cistron on the replicative (-) strand RNA.

Specifically, such a ribozyme comprises at least two of said R units, « R1 » and « R2 », wherein the complementary sequences (X) n and (Xl) nt of R1 are complementary to sections of a first target region, Zl, of the potyvirus genomic (+) RNA or of the replicative (-) strand RNA, region Zl being within or comprising a first cistron or untranslated region (UTR), and X and X'of R2 are complementary to sections of a second target region, Z2, of the potyvirus genomic RNA or (-) strand replicative intermediate, region Z2 being within or comprising a second cistron or untranslated region, Z1 and Z2 being either : -different cistrons or untranslated regions, on the same RNA strand, or -being the same or different cistrons or untranslated regions on different strands of the potyvirus RNA, or -being the same genes on the (-) strand replicative RNA.

Regions Zl and Z2 contain the target triplet for the ribozymes R1 and R2 respectively. Regions Zl and Z2 thus contain the ribozyme cleavage sites of R1 and R2 respectively.

Preferred examples of this type of polyribozyme are those comprising a plurality of targets on the (+) strand, but within different cistrons. This type of polyribozyme is exemplified by a nucleic acid wherein the complementary sequences (X) n and (Xl) nl of R1 are complementary to a region within or comprising a first gene or U. T. R. on the genomic (+) RNA, and X and X'of R2 are complementary to a region within, or comprising, a second gene or U. T. R. on the genomic (+) RNA.

Alternatively, all targets may be on the (-) strand, usually within regions corresponding to different cistrons or untranslated regions. This type of polyribozyme is exemplified by a nucleic acid wherein the complementary sequences (X) n and (Xt) n of R1 are complementary to a region within or comprising a region corresponding to a first gene or U. T. R. on the replicative (-) strand RNA, and X and X'of R2 are complementary to a region within or comprising a region corresponding to a second gene or U. T. R. on the replicative (-) strand RNA.

Another example of the ribozyme of the invention is a polyribozyme which targets regions on different RNA strands of the potyvirus. This type of polyribozyme is exemplified by a nucleic acid wherein the complementary sequences (X) n and (X') y of R1 are complementary to a region Zl within or comprising a gene or U. T. R. on the genomic (+) RNA, and X and X'of R2 are complementary to a region Z2 within or comprising a region corresponding to a gene or U. T. R. on the replicative (-) strand RNA. In this type of polyribozyme, the target regions may correspond to (+) and (-) forms of the same cistron, or they may be regions of the (+) and (-) strand which have no functional or structural relation to each other. When the target regions correspond to (+) and (-) forms of the same cistron, the polyribozyme may be defined as a nucleic acid of the invention wherein the complementary sequences (X) n and (Xl) nl of R1 are complementary to a region Zl within or comprising a cistron « C » on the (+) strand RNA and the complementary sequences of (X) n and (Xl) nl of R2 are complementary to a region Z2 on the (-) strand within or comprising the region corresponding to the same cistron « C ».

In all of the above ribozymes, the complementary sequences (X) n and (X') nt may be long or short arms as previously defined. The polyribozymes may comprise both long and short arms. Polyribozymes targeting different cistrons on the same strand are frequently designed to be long arm structures, except when the targets are all the (-) capsid protein gene in which case short arm structures are preferred. Polyribozymes targeting the same cistrons on different strands can be short arm, or a mixture of short and long arm, or alternatively long arm. Preferred polyribozymes have at least one target site in the capsid protein gene (+) or (-) strand, in association with at least one target site in a different gene.

A further preferred polyribozyme of the invention is a polyribozyme targeting at least two sites within the same cistron on the same strand. This type of polyribozyme comprises at least two R units, « R1 » and « R2 », wherein the complementary sequences (X) n and (Xl) nl of R1 are complementary to sections of a target region Zl within a given gene or untranslated region of the potyvirus genomic (+) strand RNA or of the corresponding region of the replicative (-) strand RNA, and the hybridising sequences (X) and (X') of R2 are complementary to a target region Z2 within the same gene or untranslated region as that of Zl, on the same RNA strand. Preferably, both regions Zl and Z2 are on the (-) strand. However, if, according to this variant, Zl and Z2 are both on the genomic (+) strand and both are within the capsid protein gene, the polyriboyme contains a further catalytic unit targeting a site on a different gene or on the (-) strand RNA.

Preferred target regions for the polyribozymes of the invention are regions within or comprising the cistron giving rise to the capsid protein, the NIa cistron, the NIb cistron, the 3'untranslated region of a potyvirus, or the corresponding sequences in the replicative (-) strand RNA.

Particularly preferred polyribozymes include amongst their target sites, at least one within the capsid protein cistron, or within the corresponding cistron of the (-) strand RNA. Such ribozymes include nucleic acid molecules wherein the complementary sequences (X) n and (Xl) nl of at least one R unit are complementary to regions within or comprising the capsid protein (CP) gene, or the corresponding sequences in the replicative (-) strand RNA. Such polyribozymes may also include ribozyme sequences targeting other cistrons or non-translated regions of a potyvirus RNA or its replicative (-) strand RNA.

Polyribozymes targeting at least one site in the region of the (-) strand RNA corresponding to the capsid protein are especially preferred, particularly in association with a target site in a different gene. The (+) or (-) strand polyribozymes targeting the capsid protein may have any combination of other targes, arm lengths and arm configurations described above.

Preferred examples of capsid protein ribozymes have either just one target in the capsid protein in the (+) or (-) strand, or alternatively three or four targets in the capsid protein (+) strand RNA. Capsid protein ribozymes according to this variant of the invention are especially advantageous when complementary to WMV2.

Specific examples of ribozymes of the invention are the following : -ribozymes comprising at least two R units, R1 and R2, wherein the complementary sequences (X) n and (Xl) nl of R1 and of R2 are respectively complementary to regions within the CP cistron (either (+) or (-) strand) and the 3'non-translated region of the genomic (+) strand RNA; -ribozymes comprising at least two R units, R1 and R2, wherein the complementary sequences (X) n and (Xl) nl of R1 and R2 are respectively complementary to regions within the region of the replicative (-) strand RNA corresponding to the CP cistron and the NIb cistron, or to the NIb cistron and the NIa cistron; -ribozymes comprising at least two R units, R1 and R2, wherein the complementary sequences (X) n and (Xl) nl of R1 are complementary to the CP cistron of the genomic (+) strand RNA, and the hybridising regions of R2 are complementary to all or a portion of the region corresponding to the CP cistron on the replicative (-) strand RNA; -ribozymes comprising at least three R units, R1, R2 and R3, wherein the complementary sequences (X) n and (Xl) nl of R1 are complementary to regions within the CP cistron or the 3'NT region of the genomic (+) strand RNA, the complementary sequences (X) n and (Xl) nl of R2 are complementary to regions within the region corresponding to the NIb cistron on the replicative (-) strand RNA and the complementary sequences (X) n and (Xl) nl of R3 are complementary to regions within the region corresponding to the NIa or CP cistrons on the replicative (-) strand RNA; -ribozymes wherein the complementary sequences (X) n and (Xl) nl of R1 and R2 are complementary to regions within the CP gene on the replicative (-) strand RNA.

Further examples of preferred ribozymes are indicated in table 8 below.

In addition to the choice of target site (s), target region (s) and catalytic region (s), a further variant can be obtained by incuding in the ribozyme construction, all or part of a 3'non-coding region of a potyvirus, preferably WMV2, to control intracellular localisation of the ribozyme gene. This may help to direct the ribozyme RNA to the cellular compartment where the potyviral RNA might be. The enhanced functioning of ribozymes carrying the 3'non-coding region is suspecte by analogy to the observations in animal cells (Sullenger and Cech 1993, Science vol. 262 p 1566-1569) that a ribozyme against a retrovirus was much more efficient when it was co-localised with the target, using a retroviral vector to express it.

The 3'non-translated sequence usually contains all the untranslated region of the virus and may also contain the last 10 to 20 codons of the protein and the first 10 to 15 residues of the viral poly A tail.

The ribozymes of the present invention can be tested for their capacity to cleave a substrate RNA, and/or to confer resistance to potyvirus infection, by in vitro, in vivo or transient assays as described below in examples 8,9 and 10 respectively.

The in vitro test involves incubating a suitable RNA substrat, uniformly labeled with a radioactive label, with the ribozyme under test, prepared by in vitro transcription. Incubation is carried out at varying temeratures as described in example 8. The capacity of the ribozyme to cleave the substrate at the expected target sites is shown by carrying out an autoradiogram on a polyacrylamide gel.

The ribozymes and genetic constructs of the present invention can be conveniently tested in vivo using virus-sensitive plants. The analysis may be qualitative such as by screening for severity of symptoms, or quantitative where samples are analyse, for example, by measuring the level of virus replicating in infecte plants using procedures such as ELISA, or alternatively, the level of ribozyme RNA, or the level of target RNA. A combination of qualitative or quantitative analyses may also be employed.

The genetic nature of the virus-resistant transgenic plants can be determined by carrying out a segregation analysis. The goal of this test is to determine genetic linkage of the ribozyme gene and virus resistance. Self-fertilisation, or a cross with a non transformed genotype can be carried out on a primary descendant to obtain T1. Subsequently, T plants are inoculated with the potyvirus by mechanical inoculaion or natural inoculation or by using aphids fed on infecte plants. Plants are scored for resistance to infection. A segregation ratio of 3: 1 (resistant : susceptible) indicates a single locus acting as a heterozygous trait (« dominant »); a segregation ratio of 1: 3 indicates a single locus acting as a heterozygous trait (« recessive »), and a segregation ratio of >3: 1 indicates a plurality of loci, acting independently of each other, and conferring resistance as a single copy as a heterozygous trait.

The ribozymes of the invention can also be tested using a transient transformation assay as desrcibed below in Example 10. The target is linked to a reporter gene such as LUC. DNA encoding the substrate-reporet gene hybrid is co-introduced into a plant cell along with a plasmid encoding the ribozyme under test. The level of expression of the reporter gene is measured.

Expression is decreased when the ribozyme interacts with the target.

The present invention also extends to genetic constructs comprising the ribozymes herein described.

These genetic constructs generally comprise a vector sequence as a vehicle for transferring the ribozyme to a host cell and a promoter sequence to direct transcription of the ribozyme and a terminator to signal termination of transcription. A preferred promoter inclues but is not limited to a 35S Cauliflower Mosaic Virus (CaMV) or to a double 35S CaMV promoter. The CaMV promoter is described by Odell et al, Nature 313 : 810-812,1985. The ribozymes are generally in the complementary form such that upon transcription, the ribozyme is produced. The genetic constructs of the present invention are used to generate transgenic plants capable of constitutively, developmentally or inducibly (e. g. in response to certain stimuli) expressing the ribozymes herein described. Most preferred transgenic cells and plants are rendered substantially resistant to potyvirus infection.

The invention also extends to a method for producing transgenic plants which exhibit at least some resistance to infection by potyviruses, and to the plants which can be obtained thereby. According to this aspect of the present invention, there is provided a method of generating trangenic plants exhibiting reduced symptoms following exposure to potyviruses compare to non-transgenic plants, said method comprising transforming plant cells with a genetic construct encoding a ribozyme as hereinbefore described capable of cleaving potyvirus RNA, regenerating plants from said transformed cells, self-fertilising said regenerated plants and obtaining seeds there from and then growing transgenic plants from said seeds.

All of the known means for introducing foreign DNA into plants may be used, for example Aqrobacterium, electroporation, protoplast fusion, bombardment with a particle gun, or penetration of DNA into cells such as pollen, the microspore, the seed and the immature embryo, viral vectors such as the Geminiviruses or the satellite viruses. Agrobacterium tumefaciens and rhizogenes constitute the preferred means. In this case, the sequence coding for the polyribozyme is introduced into a suitable vector together with all of the regulatory sequences necessary such as promoters, terminators, etc... as well as any sequence necessary for selecting the transformants.

The invention also extends to progeny, reproductive material and seeds from the transgenic plants. The ribozymes may be expressed constitutively, developmentally or inducibly in the whole plant, or in specific tissues of the plant.

As examples of plants which can be rendered resistant to potyvirus infection by the present invention, mention can be made of plants belonging to the familier : Cucurbitaceae (e. g. courgette melon), Chenopodiaceae (e. g. beet), Cruciferaceae, Graminaceae (e. g. oat, wheat, sugar cane, barley, ryegrass), Leguminoseae (e. g. bean, pea, soya, peanut), Ombelliferaceae (e. g. carrot, celery, parsnip), Rosaceae (e. g. plum), Solanaceae (e. g. potato, tomato, tobacco, pepper).

The present invention also extends to the use of the claimed ribozymes in plant cells or plants, in association with other virus resistance genes or systems, either to potyviruses or to other viruses, for example, anti-CMV ribozymes (CMV signifies « cucumber mosaic virus), or coat-protein mediated systems for virus resistance such as coat protein antisense or sense (translated, or translation deficient) RNA.

According to this variant of the invention, the plant may comprise cells which are transformed both by the ribozyme sequences of the invention and by the sequences providing associated protection. Expression of the different sequences providing protection may be simultaneous or otherwise.

According to a further embodiment of the invention, there is provided a nucleic acid molecule which consists of a fragment of the WMV2 viral replicative (-) RNA, which has been shown to be particularly amenable to cleavage or inactivation by ribozymes, and can thus be used as an artificial target molecule. Indeed, the unexpected discovery that the sequence from the negative strand of the virus around site 1747 is specially susceptible to gene inactivation opens useful possibilities to use this sequence as a ribozyme or antisense target not only for protecting plants against viruses but for other uses as well.

This molecule, and its derivatives will be referred to hereinafter as the « artificial target » molecule.

The artificial target molecules of the invention consist of either nucleotides 1704 to 1873 of the (-) strand of the sequence illustrated in figure 18, or of at least 20, for example between 40 and 170, consecutive nucleotides of the 1704-1873 fragment, including the ATG shown at position 1748-1750 in Figure 18. The artificial target molecules may also consist of a sequence having at least 75% identity and preferably at least 85% or even 95% identity with the underlined sequence of Figure 18, or a fragment thereof hving at least 20 nucleotides.

The artififcial target sequence present on pBPL 0186 as described in figure 15, is made of approximately 180 bases of viral sequence from the minus strand of the viral genome, in the region coding on the plus strand for the replicase NIb gene. The sequence is made of 125 bases in 5'of the A of the GUA ribozyme cleavage site and 44 bases in 3'of the A. The A is the complementary base to the T shown as base nô 1748 on figure 18. Part of this sequence is present on many RNA viruses. For example around this cleavage site in WMV2 : Potato Virus Y has 20 out of 22 bases identical to WMV2 (90.9%) Potato Virus A has 22 out of 23 bases identical to WMV2 (95.6%) Pea Seedborne Mosaic Virus has 20 out of 23 bases identical to WMV2 (86.95%) Human Cytomegalovirus has 17 out of 23 bases identical to WMV2 (73.9%) Influenza Virus has 17 out of 19 bases identical to WMV2 (89.5%).

This region is therefore a good target for molecules designed against those viruses.

The invention also relates to larger nucleic acid molecules containing the artificial target sequences of the invention with the exception of known full length viral (-) RNA which naturally contains the sequence.

Preferred examples of molecules comprising the artificial target sequences are molecules in which the artificial target sequence is transcriptionally or translationally fused to a second nucleic acid molecule.

The artificial target sequence and its derivatives is thus appende to a gene of interest (hereafter called « the useful gene » or « gene of interest »), in 5'or in 3' of the open reading frame or coding sequence of the useful gene, so that it does not prevent the function of the useful gene and makes this gene easily controllable by a ribozyme (such as those described earlier), or antisense, directe against the artificial target.

The term « coding sequence in this context inclues those sequences active as RNA like tRNA, rRNA, etc. in addition to those sequences coding for proteins. The appende sequence may also be a regulatory sequences, sequences encoding functional RNA such as rRNA, tRNA, mRNA etc.

The 5'fusion can be transcriptional, or translational. Similarly the 3'fusion can be transcriptional or translational. The precise nature of the fusion depends on the functionality of the useful gene. For example, in the case of a gene coding for a protein, a translational fusion can be used if the aminoacids added to the protein (in NH2 or COOH terminus as the case may be) do not prevent its function.

The artificial target sequence can be smaller than the sequence described above, as demonstrated by the inhibitory activity of short arm ribozymes pairing with only a very small stretch of that sequence. For example the artificial target sequence can be reduced to 40 bases on each side of the A of the GUA, or to 10 bases in 5'rand 12 bases in 3'of the A, or even to 8 bases on each side of the A, if a slightly lower efficiency of the ribozyme is acceptable.

Optimal length can be found by routine experiment.

The high activity of ribozymes against the artificial target sequence in the experiments described can probably be explained in part by its accessibility to inhibitor RNA molecules. Therefore, it is possible to use not only ribozymes but also antisense RNA against the artificial target, albeit with a lower efficiency.

The fusion between the gene of interest and the artificial target can now be used as a target for ribozymes or antisense RNA. The inactivation of the useful gene containing the artificial target can be caused by the ribozyme gene brought in through a cross, or by the tissue specific expression of the ribozyme gene if the ribozyme gene is under control of a tissue specific promoter. Alternatively, it can be caused by the inducible expression of the ribozyme gene if the ribozyme gene is under control of an inducible promoter. By inducible is meant chemically inducible (by a natural or an artificial molecule) or inducible by environmental factors such as light, drought, stress, pathogen attack and other stimuli. Inducible and tissue specific promoters are well known.

The avantage of controlling gene expression this way is that it makes use of a known ribozyme gene, already shown to function properly. Therefore the need to expriment to find the best molecule for inactivating the useful gene is reduced.

An example of a useful gene is the gene A9-PR- Glucanase (Worrall et al, The Plant Cell, 1992, vol. 4, p 759-771). This gene is used to cause artificial male sterility (AMS) in transgenic plants. AMS plants are used as the female parent in the cross to produce hybrid seeds. When the hybrid plant obtained is grown for its fruits, it is often desirable to restore its fertility to guarantee the optimal fruit set. To facilitate the restoration of the AMS caused by the A9- PR-Glucanase gene, the gene can be engineered by fusing to it the artificial target sequence described above so that the gene function is not affecte. It is then possible to use a ribozyme against the artificial target as a restorer of the male sterility. The ribozyme gene is introduced in the male parent where it has no effect (since there is no target sequence in that plant). The hybrid plant resulting from the cross between thé. AMS female plant and the male restorer plant will be fully fertile since the ribozyme brought in by the male parent will inactivate the AMS gene brought in by the female parent. Other AMS gene can be used similarly. The use of the artificial target is not limited to the area of plant improvement. The artificial target could be used with any artificial gene when an efficient inhibition of the gene is necessary in some circumstances. For example, the artificial target could be appende to a therapeutic gene used in a gene therapy treatment in humans or animals. A tissue specific ribozyme gene against the artificial target could then be used to protect some specific tissue against the activity of the therapeutic gene if such activity was deemed undesirable in said tissue. An inducible ribozyme gene against the artificial target could be used to give the clinician the ability to stop the effect of the therapeutic gene if it was necessary, just by inducing the ribozyme gene. Indeed such systems might be absolutely required for certain gene therapy procedures, perhaps wwhere the expression of toxin genes is involved. A synthetic ribozyme against the artificial target could also be used through extraneaous delivery to control activity of the therapeutic gene in the appropriate tissue or at the appropriate time.

Similarly the artificial target could be used in association with genes used in animal improvement. It could also be used in association with genes used in the food areas for example to improve microorganisms used in food production.

Different aspects of the invention are illustrated in the figures : Figure 1: Location of the primers used to amplify the viral sequences. The amino acid sequence used to design the primers is indicated (see figure2,3, and 4 below).

Figure 2: Conserve Sequence in NIb coding region, used to design primer 2. Viruses are mentioned with the reference of the publication for their partial (WMV2 and ZYMV) or complete (all others) genomic sequence. Position on the published sequence is indicated. The standard one letter amino-acid code is used.

Figure 3: Conserve Sequence in NIb coding region, used to design primer 3.

Notations are as in Figure 2. CP is Coat Protein.

Figure 4: Conserve Sequence at end of NIa coding region, used to design primer 4. Notations are as in Figure 2.

Figure 5: Sequence of the cDNA of the 3'region of WMV2 strain Mar6. Bases 1-10 were added during cloning. Viral sequence starts at base 11 and ends at base 2971 where the polyA tail starts.

Figure 5 bis: Sequence of the cDNA of the 3' region of WMV2 strain Val3. Bases 1-10 were added during cloning. Viral sequence starts at base 11 and ends at base 2970 where the polyA tail starts.

Figure 6: Sequence of the cDNA of the 3'region of ZYMV strain E15. Bases 1-6 were added during cloning. Viral sequence starts at base 7 and ends at base 2922 where the polyA tail starts.

Figure 7 Sequences used to build the short arm ribozyme gene against WMV2.

A: Short arm ribozyme against the minus strand, cloned into BlueScript SK+ (pBPL4741).

Lower case letters are used for the ribozyme catalytic sequence and the bases from polylinkers. Relevant restriction sites are in bold lettes. The sequence complementary to the ribozyme RNA is shown.

B: Short arm ribozyme against the plus strand, cloned into BlueScript SK+ (pBPL4742). Notations as above. The sequence complementary to the ribozyme RNA is shown.

C : Sequence of the cloned 3' UTR of WMV2 Val3 used in the short arm ribozyme gene. Lower case letter show polylinker regions from BlueScript or the BamHI cloning site added by PCR. Cloning sites destroyed during cloning are shown between rackets. Numbering refers to the original sequence determined.

Figure 8: Sequences used to build short arm ribozymes against ZYMV.

A : First combination against the plus strand, cloned in BlueScript SK+: pBPL4746. Notations as in figure 7. The sequence complementary to the ribozyme RNA is shown.

B : First combination against the minus strand, cloned in BlueScript SK+: pBPL4747. Notations as above.

The sequence complementary to the ribozyme RNA is shown.

C : Second combination, against both stands, cloned in BlueScriptSK+: pBPL4750. Notations as above.

The sequence complementary to the ribozyme RNA is shown.

D : Third combination, against minus strand, cloned in BlueScriptSK+: pBPL4751. Notations as above.

The sequence complementary to the ribozyme RNA is shown.

Figure 9: Sequence of the long arm ribozyme gene against ZYMV: pBPL4760.

Base numbers corresponding to the sequence of the + strand shown in figure 6 are indicated with a * below the base.

The first 486 bases are targeting the minus strand (between base 2383 and base 1961), the junction is shown by ** under the bases, and the last 568 bases are targeting the + strand (between base 2388 and base 2871).

Notations as above. The sequence complementary to the ribozyme RNA is shown.

Figure 10 : In vitro assay of ribozyme from pBPL4734.

Figure 11 : In vitro assay of ribozyme from pBPL4736.

Figure 12 : In vitro assay of ribozyme from pBPL4741.

Figure 13 : In vitro assay of ribozyme from pBPL4742.

Figure 14 : In vitro assay of ribozyme from pBPL4750 and pBPL4751.

Figure 15 : Plasmids used in the transient assay.

A : Target plasmids pBPL0184, 0185,0186 and 0187 contain 190 bp, 210 bp, 170 bp and 180 bp respectively of minus strand sequence of WMV2 cloned in sites NsiI- NcoI of pGLFuz (Sawyer and Birch 1996).

Target plasmid pBPL0188 contain 540 bp of the plus strand of WMV2 cloned into sites NsiI-NcoI of pGLFuz.

B : Ribozyme plasmid pBPL0189 contains a 400 bp insert cloned into the BglII-XhoI sites of a derivative of pGLFuz, encoding an octameric ribozyme against WMV2, targeting sites 696 (-), 996 (-), 1747 (-), 2067 (-), 2259 (+), 2403 (+), 2520 (+) and 2720 (+).

Ribozyme plasmid pBPL0190 contains a 220 bp insert cloned into the BglII-ShoI sites of a derivative of pGLFuz, encoding a pentameric ribozyme against the minus strand of ZYMV, targeting sites 1243,1743 (equivalent to 1747 in WMV2), 1929,2442 and 2564.

Figure 16 : Transient LUC activity after co- bonbardment of either an octanmeric ribozyme (pBPL0189) or a pentameric ribozyme (pBPL0190) with various plasmids containing different target regions fused 5' to the luciferase gene. Each co-bombardment consiste of 2 ug of ribozyme plasmid (pBPL0189 or pBPL0190) and 2 ug of the target plasmid as indicated. Bars represent standard errors of the mean of three independent experiments.

Figure 17 : Construction of pGLFuz vector (Sawyer and Birch 1996). The 1.8 Kb PCR modified uida gene was inserted as a NsiI-NcoI DNA fragment upstream of the luc gene of pZorz to yield pGLfuzTAG. The luc gene of pGLfuzTAG was removed (by digestion with NotI, treatment with T4 DNA polymerase, then digestion with XhoI) and replace by the 1.9 Kb luc gene from pSPNF+luc (prepared by digestion with KpnI, treatment with T4 DNA polymerase and digestion with XhoI), yielding the guida: luc fusion construct pGLFuz. B, GglII; Bl, blund end; K, KpnI; Luc, luc gene from pZorz; Mod. GUS, PCR modifie uidA gene; Nc, NcoI; Ne, NheI; Ni, NsiI; No, NotI; P, osa artificial promoter <BR> <BR> <BR> <BR> <BR> region containing 4ocs-#35S-Ac leader regions ; pSPNF+Luc, luc gene from pSPNF+luc ; T, ocs 3' polyadenylation region; TAG, stop codon; X, ShoI.

Figure 18 : Sequence of the region of WMV2 strain Mar6 (top line = (+) strand, bottom line = (-) strand) around ribozyme cleavage site 1747 in the (-) strand (indicated in bold type). Underlined sequence (positions 1704 # 1873 in (-) strand) represents the full length artificial target sequence which can be linked to a reporter gene in transient ribozyme assays, or to a gene of interest as a ribozyme-regulatable element. Arrow (T) indicates nucleotide in (-) strand which is immediately 5' of the ribozyme cleavage site.

(G/T) at position 1863 in the (-) strand indicates that the naturally occurring G can be replace with a T to mutate the stop codon at 1864-1862 (TGA). This facilitates in-frame fusion, and can be introduced using appropriate PCR primers.

Figure 19 : shows the DNA sequence of pBPL3817 (long arm ribozyme against WMV2 NIb and CP cistrons).

The sequence is written from 5'tao 3'. The first base corresponds to base number 279 in figure 5. The last viral base is G, 34 bases before the end of the sequence. The ribozyme catalytic regions are underlined, in italics Figure 20 : shows the genome organisation of the type member of the potyirus family, PVY (from Chapter 4, « The Potyviridae », Ed. Shukla, D. D., Ward, C. W., Brunt, A. A, 1994 CAB International). P1 : First Protein ; HC-Pro: Helper component ; P3 : Third Protein; 6K1, 6K2 : additional 6kb proteins CI : Cylindrical Inclusion Protein NIa : Small Nuclear Inclusion protein ; NIb : large Nuclear Inclusion Protein; CP : Coat Protein Figure 21 : shows the determination of the viral epidemiology in Season 2 of field trials (described in Example 10 below). The technique used was ELISA analysis on virus susceptible control, taken randomly.

The notations were made over a period of one month at intervals of one to two weeks.

WMV2 = Watermelon Mosaic Virus 2 ; <BR> <BR> CMV = Cucumber Mosaic ;<BR> <BR> ZYMV = Zucchini Yellow Mosaic Virus ; PRSV = Papaya Ringspot Virus.

EXAMPLES Molecular Biology techniques used are essentially those described in Current Protocols in Molecular Biology (Ausubel F. M. et al. Eds, 1994-1997, John Wiley & Sons, Inc., New York.) except where otherwise mentioned.

1. Cloning and Sequencing the 3'region of genomic RNA of target viruses.

Target viruses for this work are Watermelon Mosaic Virus 2 (WMV2) and Zucchini Yellow Mosaic Virus (ZYMV).

To design ribozymes against these viruses, a partial sequence of the genome of field isolates of these viruses from the Mediterranean region was determined as follows: a-Infected plant material was harvested, and total RNA extracted. b-The RNA was used as the template for a Reverse Transcription-Polymerase Chain Rection (RT- PCR) using primers specific for the 3'terminal region of the potyviral genome. c-The PCR products were cloned and sequenced.

These three steps are detailed below: a-Zucchini plants were mechanically inoculated with strains Mar6 or Val3 of WMV2 or strains E9 or E15 of ZYMV. These strains are field isolates from the Mediterranean region. Leaves were harvested and total RNA was extracted. b-A sequence comparison of all published potyvirus sequences allowed the identification of regions of sequence conservation among potyviruses.

Figure 1 shows the location of relevant conserve sequences identifie on the genome of several potyviruses.

Figure 2 shows the conserve sequence identifie in the region coding for protein NIb used to design primer 2.

Figure 3 shows the conserve sequence identifie in the region coding for protein NIb used to design primer 3.

Figure 4 shows the conserve sequence identifie at the end of the region coding for protein Nia, and used to design primer 4.

Reverse transcription was performed using the primer Poly-T-anchor: Poly-T-anchor: 5'GGGAGCTCGCTAGCGGATCCTl7 Primer 1, used in the PCR step, is homologous to the anchor region, with extra bases added to facilitate cloning of the PCR products: Primer 1: 5'AATTGGGAGCTCGCTAGCGGAT The other primers were designed to match the consensus sequence of the conserve regions identifie as described above, and some bases were added to facilitate cloning. The primers were moderately degenerate to take into account polymorphism. Multiple bases at a given position are indicated in the sequence by a/ : T/G means that both T and G could be present in the primer at a given position. N means that any base could be present.

Primer 2: 5'AATTGCTAGCAGTT/GGTT/GGACAATACACTT/AATGG Primer 3: 5'AATTCTAGAATA/CAT/GATCATCACCATTG/TGC Primer 4: 5'AATTCTAGATTCTGGAT/AACATTGGATAA/TC/GNAC c-PCR products of the predicted size were obtained with primers pairs 1-2 (1.7 kb) and 3-4 (1.3 kb). The PCR products were purifie on gel, cut with the appropriate restriction enzymes, cloned into plasmid BlueScript SK+ (commercially available from Stratagene) and the complete nucleotide sequence of the viral fragment was determined using standard methods known in the art. Most of the sequence was determined from two independent PCR products from each RNA template. For WMV2, a PCR product was also obtained with primer pair 1-4. It was used subsequently to build long arm ribozyme genes (see Example 3 below).

The sequence of the 3'region of WMV2 strain Mar6 is described in Figure 5 and the sequence of the 3' region of WMV2 strain Val3 is described in figure 5 bis. The two strains are more than 99% identical.

The sequence of the 3'region of ZYMV strain E15 is described in Figure 6. The E9 strain was more than 99% identical to E15.

2. Design of ribozymes against WMV2 Both the plus (+, coding, or genomic) RNA strand of the virus and the minus replication intermediate) RNA strand of the virus were targeted by ribozymes.

The following criteria were used when choosing cleavage sites on the target RNA: Triplets known to be well cleaved in vitro were preferred (Perriman et al., 1992): GUC, GUA, GUU, UUC, UUU.

Triplets where the codons had limited degeneracy were preferred.

This increases the chance that the target sequence will not evolve easily so that the virus overcomes the resistance. Some examples are shown below for illustration, with the cleaved triplet in Bold.

Many more combinations are possible.

For the plus strand: Aminoacid sequence: MetSer MetTyr MetPhe Phe Nucleotide sequence: AUGUCN AUGUAU/C AUGUUU/C UUU/C For the minus strand: Aminoacid sequence (+): MetLys MetThr GluLys Tyr Nucleotide sequence (+): 5'AUGAAA/G AUGACN GAA/GAAA/G UAU/C Nucleotide sequence (-): 3'UACUUU/C UACUGN CUU/CUUU/C AUA/G Triplets located in a region of sequence conservation were preferred. This would increase the chance that the ribozyme could target multiple strains of the virus or even multiple species of related viruses. It would also decrease the chance that a virus resistant to cleavage arises.

Four sites were selected on the plus strand and four sites were selected on the minus strand. The number used to describe a given ribozyme refers to the base underlined, on the + strand in the following tables, according to the numbering of the sequence presented in figure 5. Cleavage site is in Bold and the cleaved bond is indicated by a-.

Sites selected on the minus strand: Number Target sequence + strand Aminoacid sequence sequence 696 5'WC-A 3'5'UGAA 3'Met Lys 996 5'GUC-A 3'5'UGAC 3'Met Thr 1747 -UGUA AUAC3'Tyr5' 2067 -G3'5'CAAA3'ThrLysUUU Sites selected on the plus strand: Number Target sequence Aminoacid sequence 2259 5'UUU-A 3'Phe 2403 5'WC-U 3'Phe 2520 5'WU-U 3'Phe 2720 -A3'GlnOchGUA Of these sites, site 2259 was conserve in ZYMV (at base 2249 on figure 6).

Site 1747 was highly conserve among most potyviruses: a comparison of a 23 base segment surrounding the site shows the following identity: ZYMV: 23/23 Potato Virus Y: 20/22 Sugar Cane Mosaic Virus: 19/23 Pea Seed-borne Mosaic Virus: 20/23 Plum Pox Virus: 20/23 Maize Dwarf Mosaic Virus: 20/23 The structure and properties of the ribozymes exemplifying the present invention are indicated below in Table 8.

TABLE 8 n.'t :' :.. v 3 n : T.'. \. v TANCE ..: STAN r RESI . IN V1V0 ., .... c < > ., , k...,,. TRUCTURE OF RIBOZY zs. vk a: s ycs tt. $5P<tc3'f5, 2 r: c > c STRUCTURE OF RIBOZYME IN VIVO RESIS f t £ <'K oY? Su vs: I. v. u< r: . x F f'E. a-. 2 w < 3 j -Y S ri ): i Yv . y'y =strfr'. 're . Y v.: % M wl. R7IIIG pBPL4738 WMV2 Nib+CP (-) 4 long-No Variant Resistan (heteroygous) Not tested (pBPL 3817) PBPL 4739 WMV2 CP+3'NT (+) 4 iong No Variant Resistan (heteroygous) Not tested {I + III) (pBPL380'10) PBPL 4743 WMV2 Nlb+CP+3'NT (+) and 8 short not contiguous Yes Variant Resistance (homozygous) Not tested (pBPL37s4) (-) (IV) ? BPL4748 ZYMV (E15) Nla+CP+Nlb (+) and 10 short not contiguous Yes Variant E-15 13-18 (pBPL 3906) (-) (IV) No resistance Resistance (homorygous) PBPL 4750 ZYMV (E15) CP (+) and 5 short some No variant E-15 13-18 (pBPL 4186) (-) contiguous (IV) No resistance Resistance (homoygous) pBP14757 ZYMV (E15) CP (-) 2 short contiguous No Variant E-15 13-18 (pBPL 4194) (IVy No resistance Resistance (homozygous) pBPL4760 ZYMV (E15) CP (+) and 7 short and contiguous No Natural E-15 13-18 (pBPL4771) (-) long Resistan (heteroygous) Resistan (homoygous) 1Construct name in brackets indicates designation of ribozyme in plant transformation vector<BR> 2« 3'UTR » signifies presence, on 3' end of ribozyme construct, of the 3' untranslated region of WMV2 strain Val3 (approximately 300 bases)<BR> 3Catalytic regions: « natural » signifies sTobRV catalytic domain, « variant » signifies one or more of the catalytic regions comprise mutations @<BR> in brackets indicate the type of mutation, as defined above in description 3. Construction of a long arm multimeric ribozyme gene against the minus strand of WMV2 (pBPL 4738) A subclone of the cDNA for the 3'region of WMV2 Mar6, extending from base 269 (EcoRI site) to base 216 8 (XhoI site) (see figure 5) was built by internal deletions of a plasmid containing the PCR product obtained with primers 1 and 4 as described in 1. above.

This subclonewas used as a template for the next step.

Inverse PCR (IPCR) rections were used to insert a ribozyme catalytic sequence in replacement of the third base of each target triplet, essentially as described in Unit 3.17.3 of Current Protocols in Molecular Biology. The ribozyme catalytic sequence from the satellite of Tobacco Ringspot Virus (sTobRV, Haseloff and Gerlach, 1988) was used.

The primers used are listed below.

Bases in lower case are encoding the ribozyme catalytic sequence. Bases in upper case are complementary to the target region and used to prime the plasmid template. Bases paired with the first 2 bases of the cleaved triplet are in bold. For the first primer pair the sequence of the final product around site 696 is shown, to demonstrate that ligation of the IPCR product will produce a molecule with a functional catalytic sequence inserted in replacement of the third base of the target site. <BR> <BR> <BR> <BR> <BR> <P>Rbz696-3' : 5'gtgaggacgaaAAAGCTGCTGTCGGTGCTC3'<BR> <BR> <BR> Rbz696-5' : 3'CTAAATACCTAGGGACTTGTAgactactcagg5' Result for 696: 5'... GAACAUcugaugaguccgugaggacgaaAAAGC... 3' Rbz996-3': 5'gtgaggacgaaACTAAGTTTTATGGGGGTTGG3' Rbz996-5': 3'CAGGTACCTGTCAACCGTAgactactcagg5'<BR> <BR> <BR> <BR> <BR> <BR> <BR> <BR> <BR> <BR> <BR> Rbz1747-3' : 5'gtgaggacgaaACATAGCTGAGACAGCATTG<BR> <BR> <BR> <BR> <BR> Rbz1747-5': 3'CGTAGGCCTTTTCGAGGTgactactcagg5' Rbz2067-3' : 5'gtgaggacgaaAAGAAAATGAACCTTCCAACAG3' Rbz2067-5': 3'GGGTGCTGATGTTTTCTATTGgactactcagg5' The plasmid template was amplifie by IPCR using the first pair of primers. The full length PCR product was purifie on a gel, and circularised by ligation.

The circula product was transformed into bacteria, and the resulting plasmid was extracted from a bacterial colony and submitted to the second IPCR rection using the second pair of primers. The procedure was repeated with the third and fourth pair, and the final product was analyse by sequencing the modifie regions. A plasmid called pBPL4734 was obtained; it had the correct sequence except for ribozyme 696 which had one extra base in the loop, and ribozyme 996, where a small deletion had taken place. Attempts were made to repair errors in 996 by a new round of IPCR. The final product was called pBPL4738 and had the expected sequence except at sites 696 and 996 where the catalytic sequence is mutated in the ribozyme loop (upper case): Mutant ribozyme 696. cugaugaguccCGUGAggacgaa Mutant ribozyme 996. cugaugaguccUGAggacgaa Wild type sTobRV: cugaugaguccGUGAggacgaa.

To build a functional ribozyme gene in a plant transformation vector, the multimeric long arm ribozyme sequence was cloned in the plant expression cassette from pBIOS512, made of the enhanced 35S promoter and the nos terminator, and the gene was then inserted in the plant transformation vector pSCK.

Description of pBIOS512: this plasmid was built by replacing the 35S promoter region of plasmid pBIOS3 (Perez et al., 1989) by the enhanced 35S promoter region from pJIT163, a plasmid derived from pJIT60 (described in Guerineau and Mullineaux, 1993) by inserting an NcoI site in the polylinker. The enhanced 35S promoter was isolated from pJIT163 as a SstI- BamHI fragment and cloned into sites SstI-BamHI of pBIOS3 in replacement of the 35S promoter.

Description of pSCK: this plasmid was built by inserting a kanamycin resistance gene into the plant transformation vector pSCV1 (Firek et al., 1993). The kanamycin resistance gene was built by inserting the nptII coding sequence from pCaMVNEO (Fromm et al., 1986) as a BamHI fragment in the BamHI site of pBIOS1 (Perez et al., 1989), resulting in pBIOSlK. The gene was then isolated from pBIOSlK as a EcoRI (Klenow) fragment and inserted into site EcoRV of pSCV1, resulting in pSCK.

The ribozyme sequence was cloned in the expression cassette of pBIOS512, as a EcoRV-SmaI- HindIII fragment in the HindIII (Klenow) site of pBIOS512. The gene was then excised as a EcoICRI- EcoRI (Klenow) fragment and inserted into the SmaI site of pSCK.

The resulting plasmid was called pBPL3817. It contains a T-DNA with a marker gene conferring resistance to kanamycin and the long arm ribozyme gene expressing an RNA with approximately 1900 bases complementary to the minus strand of the virus (the target RNA) and targeting four sites on the target RNA.

4. Construction of a long arm multimeric ribozyme gene against the plus strand of WMV2 (pBPL 4739) A procedure similar to that described in 3. was used to build a ribozyme gene against the + strand. The plasmid template used in the IPCR rections extended from base 2164 (site XhoI) to the end of the cDNA (base 2988, including 18 A of the polyA tail) and was obtained by internal deletions of a partial cDNA clone of the 3'region of WMV2.

Inverse PCR rections were used as described above in 3. using the following primer pairs: (same notations as above) <BR> <BR> <BR> Rbz2259+3': 5'gtgaggacgaaAAACCATTCATAATCACACCC3'<BR> <BR> <BR> <BR> Rbz2259+5': 3'CAATAGCTATGTGGTTTGGTAgactactcagg5'<BR> <BR> <BR> <BR> <BR> <BR> <BR> Rbz2403+3' : S'gtgaggacgaaAAATGGTGCATGATTTGTC3' Rbz2403+5' :3'CGAAGACGACGCAGACTgactactcagg5' Rbz2520+3': 5'gtgaggacgaaAAGTCAAAAGCATAGCGTGC3' Rbz2520+5':3'-CACAAAACCTTCATTGGAGTATgactactcagg5' Rbz2720+3' : 5'gtgaggacgaaACTGCGGTGGACCCATACC3' Rbz2720+5':3'-CACTGGTCAAATGGATCAGAAgactactcagg5' The plasmid obtained was called pBPL4736, and had half of ribozyme 2259 missing, and a one base deletion in the 3 other ribozymes'loop. New rounds of IPCR inserted a correct ribozyme for 2259 and correcte ribozyme 2403. The final product still contained some errors which were not considered serious problems. The ribozyme sequence for sites 2259 and 2403 is correct (wild type as described). The ribozyme sequences for sites 2520 and 2720 each have one base deleted: Mutant ribozyme in 2520: cugaugaguccUGAggacgaa Mutant ribozyme in 2720: cugaugagucGUGAGgacgaa Wild type sTobRV: cugaugaguccGUGAggacgaa These mutations are known not to abolish ribozyme activity. This was confirme by in vitro assays (see below). This plasmid was called pBPL4739.

The sequence from pBPL4739 was cloned in the expression cassette of pBIOS512 as a BamHI (Klenow)- EcoRI fragment in the HindIII (Klenow) site of pBIOS512.

The gene was then transferred to the plant transformation vector pSCK as described above in 3.

The resulting plasmid was called pBPL38010. It contains a T-DNA with a marker gene conferring resistance to kanamycin and the ribozyme gene expressing an RNA with approximately 830 bases complementary to the plus strand of the virus (the target RNA) and targeting four sites on the target RNA.

5. Construction of a short arm multimeric ribozyme gene against both strands of WMV2 (pBPL 4743) In order to build a short arm multimeric ribozyme targeting the sites described above for the long arm ribozymes, synthetic oligonucleotides encompassing the entire sequence of the desired ribozyme gene were purchased. The oligonucleotides were assemble and the desired full length product was obtained using Ligation based PCR (LbPCR) as described by Betzner et al.

(1997). The short arm multimeric ribozyme genes against the minus strand and the plus strand were synthesised independently and assemble using restriction sites built at the end of the synthetic sequences.

The ribozyme catalytic sequence used for these constructs was a variant of the sTobRV ribozyme described above, with a 3 base-pair helix II instead of the natural 4 base-pair. The ribozyme used for the short multimeric ribozyme was : cugaugaGCGgugaCGCgaa.

(Helix II shown here in upper case lettes) The pairing regions between the ribozymes and the target (helix I and helix III) was made of 2 arms of 10 to 12 bases (see Haseloff and Gerlach, 1988 for a description of helices I, II, and III).

The sequence of the oligonucleotides used is given below; lower case letters are the ribozyme catalytic sequence, lower italic letters are linker regions, upper case letters are target sequences. All sequences are shown 5'to 3'. The sequence synthesised (shown) is actually the strand complementary to the ribozyme RNA.

Oligonucleotides used for the ribozymes against minus strand: <BR> <BR> <BR> <BR> <BR> WM2067S :<BR> <BR> <BR> <BR> <BR> <BR> <BR> gcagctgctggatccatgTCATTTTCTTttcgcgtcaccgctcatcagGTTATCT TTTG WM1747S TCAGCTATGTttcgcgtcaccgctcatcagTGGAGCTTTTC <BR> <BR> <BR> <BR> <BR> <BR> <BR> WM996S :<BR> <BR> <BR> <BR> <BR> <BR> CagatctATAAAACTTAGTttcgcgtcaccgctcatcagATGCCAACTG WM696S: CAGCAGCTTTttcgcgtcaccgctcatcagATGTTCAGGGAcctgcatctgatca agcttg Guide oligonucleotides used to align the full length sequence for the ligation rection: WMGU4S: assemble WM2067S with WM1747S GAAACATAGCTGACAAAAGATAACCTG WMGU5S: assemble WM1747S with WM996S GTTTTATAGATCTGGAAAAGCTCCAC WMGU6S: assemble WM996S with WM696S GAAAAAGCTGCTGCAGTTGGCATC Oligonucleotides used for the ribozymes against plus strand: WM2259S: <BR> <BR> <BR> <BR> <BR> gcagctgctggatccATGAATGGTTTttcgcgtcaccgctcatcagATGGTTTGG WM2403S: TGCACCATTTttcgcgtcaccgctcatcagTCAGACGCAGat WM2520S: CTTTTGACTTttcgcgtcaccgctcatcagTATGAGGTTAC WM2720S: CACCGCAGTttcgcgtcaccgctcatcagAAGACTAGGTAAcctgcatctgatca agcttg Guide oligonucleotides used to align the full length sequence for the ligation rection: WMGU1S : assemble WM2259S with WM2403S GAAAAATGGTGCACCAAACCATCTG WMGU2S: assemble WM2403S with WM2520S CGAAAAGTCAA.A.AGATCTGCGTCTGAC WMGU3S: assemble WM2520S with WM2720S AAACTGCGGTGGTAACCTCATACTG Primers used to amplify the assemble ligated oligonucleotides by PCR: WMPC1S:AATTGCAGCTGCTGGATCCATG WMPC2S: AATTCAAGCTTGATCAGATGCAG The PCR products were digeste with BamHI and HindIII and cloned into BlueScript SK+. The ribozyme against the minus strand gave plasmid pBPL4741, and the ribozyme against the plus strand gave pBPL4742. The sequences of the relevant regions are on figure 7.

Restriction sites used for further cloning are indicated.

The 2 ribozymes were combine by inserting the BstXI-BspMI fragment of pBPL4742 in sites BstXI- BamHI of pBPL4741, yielding pBPL4743.

The 3'non coding region of WMV2 strain Val3 (3' UTR) was then cloned in 3'of the multimeric ribozyme sequence. This was done as an attempt to increase the efficacy of the ribozyme by co-compartmentalisation with the target viral RNA. If 3'end sequences of RNA are involved in cellular localisation of RNA, as is the case with many ORNAS, it was hoped that by providing the ribozyme RNA with a 3'sequence similar to that of its target, the 2 molecules would be present in the same cellular compartment.

The 3'UTR region of WMV2 Val3 was cloned from the original cDNA (see sequence in figure 5 bis) as a HinfI (at base 2678)-BamHI (from the anchor region in primer 1, see 1. above) fragment into EcoRI-BamHI of BlueScript SK+. The fragment contains bases 2678 to 2970 of strain Val3 cDNA, which contains the last 14 codons of the cDNA followed by the complete untranslated region. This sequence (shown in figure 7) was then inserted as an EcoRV-NotI fragment into BamHI (Klenow)-NotI sites of pBPL4743, giving pBPL4744.

The now complete sequence was then transferred as a BamHI-Hind III fragment in sites BamHI-HindIII of pBIOS512 to create a functional gene, as described above in 3. This gene was then cloned in the SmaI site of pSCK as described above in 3., yielding pBPL3754. pBPL3754 contains a T-DNA with a marker gene conferring resistance to kanamycin and the ribozyme gene expressing an RNA approximately 700 bases long.

This RNA contains 8 regions of homology to the plus or minus strand of the virus. Each region is between 20 and 22 bases long, and is split in the middle by the ribozyme catalytic sequence. In addition the ribozyme RNA has a 3'region containing approximately 300 bases identical to the viral 3'end.

6. Construction of several short arm multimeric ribozyme genes against ZYMV Site selection will be described using the same notations as in 2. above. The numbering refers to the sequence shown on figure 6.

Sites selected on the minus strand: Number Target sequence + strand sequence Aminoacid sequence 1245 5'GW-G 3'5'CAAC 3'Asn 1743 5'GUA-U 3' 5' AUAC 3' Tyr 1929 -A3'5'UGAC3'AspGUC 2009 -A3'5'UGAA3'ValAsnUUC 2027 -C3'5'GGAA3'GlyLysUUC 2354 -U3'5'AAAA3'GluAsnUUU 2442 -U3'5'AUAC3'TyrGUA 2564 -A3'5'UGAA3'MetLysUUC Sites selected on the plus strand: Number Target sequence Aminoacid sequence 110 -U3'ValGUC 2161 -U3'AlaSerGUC 2255 5'GW-U 3'Val 2455 -U3'ArgTyrGUA 2466 5'WC-G 3'LeuArg 2590 5'WC-U 3'ValSer 2603 5'UUU-G 3 Phe Several combinations of short arm ribozymes were built using synthetic oligonucleotides. The procedure was similar to the procedure described in 5. above.

6.1. Construction of pBPL 4748 The first combination contained 10 ribozyme activities, 5 against each strand: 1245-1743-1929-2442-2564 against the minus strand (plasmid pBPL4747) 110-2161-2255-2455-2590 against the plus strand (plasmid pBPL4746) The ribozyme catalytic sequence used for these constructs was the same variant of the sTobRV ribozyme as described in 5. above, with a 3 base-pair GCG helix II instead of the natural 4 base-pair.

The pairing regions between the ribozymes and the target (helix I and helix III) was made of 2 arms of between 8 and 11 bases.

Sequence of the relevant region of pBPL4746 and pBPL4747 is shown on figure 8.

The sequences were assemble together with the 3'UTR of WMV2 Val3 (shown on figure 7), by a tripartite ligation: the EcoRV (de phosphorylated)-NotI fragment containing the 3'UTR and the BamHI (Klenow)- HindIII (Klenow) fragment from pBPL4747 were ligated together in pBPL4746 cut with BamHI (Klenow) and NotI (both sites de phosphorylated). The resulting plasmid was called pBPL4748 and contained (from HindIII to BamHI) the sequence for the ribozymes against the plus strand, followed by the ribozymes against the minus strand followed by the 3'UTR. The complete sequence was then cloned as a BamHI-HindIII fragment in the expression cassette of pBIOS512 as described in 3. above, and the functional gene was then cloned into the plant transformation vector pSCK as described in 3. above yielding pBPL3906.

Plasmid pBPL3906 contains a T-DNA with a marker gene conferring resistance to kanamycin and the ribozyme gene expressing an RNA approximately 750 bases long. This RNA contains 10 regions of homology to the plus or minus strand of ZYMV. Each region is between 16 and 20 bases long, and is split in the middle by the ribozyme catalytic sequence. In addition the ribozyme RNA has a 3'region containing approximately 300 bases identical to the non-translated 31 zend of WMV2 Val3.

6.2 Construction of pBPL4750 A second combination was made using the following ribozymes: 2354 2455- 2466(+)-2590(+)-2603(+).- The synthetic sequence cloned into BlueScript SK+ yielded pBPL4750.

This was an attempt to capitalise from the clustering of sites 2455-2466 and 2590-2603.

The ribozyme catalytic sequence used for these construct was the natural sTobRV hammerhead ribozyme except for 2354 and 2603 where the same variant of the sTobRV ribozyme as described in 5. above, with a 3 base-pair GCG helix II instead of the natural 4 base- pair was used.

The pairing regions between the ribozymes and the target (helix I and helix III) was made of 2 arms of between 8 and 12 bases. Adjacent ribozymes 2455-2466 and 2590-2603 had one arm in common.

The relevant sequence of pBPL4750 is shown on figure 8. The ribozyme sequence was cloned directly as a BamHI-HindIII fragment in the expression cassette of pBIOS512 as described in 3. above, and the functional gene was then cloned into the plant transformation vector pSCK as described in 3. above yielding pBPL4186.

Plasmid pBPL4186 contains a T-DNA with a marker gene conferring resistance to kanamycin and the ribozyme gene expressing an RNA approximately 250 bases long. This RNA contains 3 regions of homology to the plus and minus strand of ZYMV. One region of 21 bases is split in the middle by one ribozyme catalytic sequence. The other 2 regions of 31 and 32 bases are each split in three parts by two ribozyme sequences.

6.3. Construction of pBPL 4751 A third combination was made using a double ribozyme against sites 2009-2027 (both minus stand).

The synthetic sequence cloned into BlueScript SK+ yielded pBPL4751.

The ribozyme catalytic sequence used for this construct was the same variant of the sTobRV ribozyme as described in 5. above, with a 3 base-pair GCG helix II instead of the natural 4 base-pair.

The pairing regions between the ribozymes and the target (helix I and helix III) was made of 2 arms of 10 and 17 bases for each ribozyme. The 17 base arm was shared between the 2 ribozymes. In effect the total pairing region was 37 bases (10 + 17 + 10), interrupted in 2 places by the 2 ribozyme sequences.

The relevant sequence of pBPL4751 is shown on figure 8. The ribozyme sequence was cloned directly as a BamHI-HindIII fragment in the expression cassette of pBIOS512 as described in 3. above, and the functional gene was then cloned into the plant transformation vector pSCK as described in 3. above yielding pBPL4194.

Plasmid pBPL4194 contains a T-DNA with a marker gene conferring resistance to kanamycin and the ribozyme gene expressing an RNA approximately 100 bases long. This RNA contains 1 region of homology to the minus strand of ZYMV. The region is 37 bases long, and is split in three parts by the 2 ribozyme catalytic sequences.

7. Construction of a long arm multimeric ribozyme gene zains ZYMV A single gene, expressing one polyribozyme RNA targeting several sites on both strands was made.

The sites chosen were (see 6. above for numbering) : 2027(-),2354(-),2455(+),2466(+),2009(-), 2590 (+) and 2603 (+).

To create the appropriate template, two fragments from the cDNA of the 3'region of ZYMV were cloned in opposite orientation as follows (see figure 6 for sequence).

The cDNA clone obtained from ZYMV E15 with PCR primers 1 and 2 (as described in 1. above) was cut with NsiI (at base 2383) and SacII (in the polylinker, next to 5'end of the cDNA), then treated with Klenow and de phosphorylated. The large fragment was then purifie and used as a vector in the subsequent ligation. The same plasmid was cut with NsiI (at base 2383) and AlwNI (at base 1955), then treated with Klenow, and the 470 base fragment was purifie and used as the insert in the subsequent ligation. The clones obtained were screened for the appropriate orientation, and the selected plasmid was called pBPL0157. This plasmid contains the sequence between bases 2383 and 1961 linked to the sequence between bases 2388 and 2922. The ligation places side by side the region 3' of 1961 and the region 3'of 2383. When a promoter is placed in the appropriate orientation, the transcription product of this sequence will be complementary to the minus strand of ZYMV between base 1961 and 2383, and complementary to the plus strand of ZYMV between base 2388 and 2922.

Ribozyme catalytic sequences were inserted at the positions mentioned above using a sequence of IPCR rections on template plasmid pBPL0157, following the procedure described for WMV2 in 2. above. The ribozyme catalytic sequence used for this construct was the natural sTobRV hammerhead ribozyme as described previously.

The insertion of the catalytic sequence targeting site 2354 (-) proved difficult, as if this ribozyme sequence was toxic to E. coli or affecte plasmid replication.

The plasmid containing the 7 ribozyme sequences was called pBPL4760. Sequence of the ribozyme gene in pBPL4760 is shown in figure 9.

The complete sequence was recovered using HindIII (which cleaves at base 2866) and SacI (Klenow) (from the polylinker) and the fragment was cloned in the expression cassette of pBIOS512 in sites HindIII- SalI (Klenow). This had the effect to remove from the region targeting the plus strand the last 51 nucleotides. The functional gene was then transferred to the plant transformation vector pSCK as described previously, yielding plasmid pBPL4771.

Plasmid pBPL4771 contains a T-DNA with a marker gene conferring resistance to kanamycin and the long arm ribozyme gene expressing an RNA with approximately 900 bases complementary to the plus and minus strand of ZYMV RNA and targeting seven sites on the target ORNAS.

8. In vitro assay of the ribozymes Ribozyme activity was assayed in vitro by incubating an appropriate target RNA, prepared by in vitro transcription and uniformly labelle with radioactive UTP, with the ribozyme RNA prepared by in vitro transcription. The protocol is described in detail below and the result for each ribozyme are also shown.

Target RNA preparation.

The plasmids used for target RNA synthesis were all derived from the plasmids containing the original PCR amplifie viral sequences described in 1. above.

The T3 and T7 promoters present in the BlueScript vector were used to transcribe the plus or the minus strand of the cloned fragment as appropriate. In vitro transcription using T3 RNA Polymerase or T7 RNA Polymerase from Promega or New England Biolabs was done following the suppliers instructions (for example in Promega Protocols and Application Guide, available from the Promega company, Maison, Wisconsin, USA), in the presence of radioactive a32P-UTP. The template plasmid was linearised with an appropriate restriction enzyme prior to transcription. The target RNA was quantitated by measuring the incorporation of radiolabel using the DE81 filter binding assay as described in the Promega Protocols. Target RNA was stored as a precipitate under ethanol at-20°C, and resuspended in water at between 0.01 and 0.1 pmol/ul when needed. Specific activity was usually between 20,000 and 200,000 cpm/pmol.

Ribozyme RNA preparation The plasmids described above in sections 3,4,5, 6 and 7, made of the various polyribozyme sequences cloned into BlueScript, were used as templates for in vitro transcription rections as described above, but using only non radioactive nucleotides. A pilot rection with a small amount of radioactive tracer was done separately to quantitate the synthesised RNA, using the DE81 filter binding assay. Ribozyme RNA was stored as a precipitate under ethanol at-20°C and resuspended in water at between 0.1 and 1 pmol/pl when needed.

Ribozyme rection The target RNA and ribozyme RNA were mixed in T7 RNA Polymerase buffer (from New Englands Biolabs) and incubated at various temperature (22°C, 37°C or 50°C).

The standard rection was in 4 ul and contained 0.005 to 0.1 pmol of target, and between 1 and 10 fold molar excess of ribozyme. The rection was stopped by adding an equal volume of stop-loading buffer (Formamide containing lOmM EDTA and lmg/ml each Xylene Cyanol and Bromophenol Blue) to the rection and freezing the sample at-20°C until all samples were ready. The rections were then boiled for 30 seconds, placed on ice, and loaded on a denaturing polyacrylamide gel in TBE buffer with 7M urea. Polyacrylamide concentration was between 4 and 6 % depending on the expected length of the products. The gel was transferred to a Whatmann 3MM paper, dried and exposed to a PhosphorImager screen. An autoradiogram of the gel was obtained by scanning the screen using a PhosphorImager.

Long arm ribozyme against the minus strand of WMV2 The in vitro assays of long arm ribozymes proved difficult since the target and the ribozymes do not readily dissociate prior to loading and/or reassociate in the gel. The data are therefore of poor quality and difficult to interprete.

The ribozyme used was obtained from pBPL4734 (described in 3. above). The target contained 1898 bases of viral sequence, complementary to the sequence from base 2168 (site XhoI) to base 270 (site EcoRI) of WMV2 Mar6, with some sequence from the polylinker of BlueScript KS+ (between the T3 promoter and the XhoI site).

Figure 10 shows the result of one assay. There was 4 fold excess ribozyme and the rection was done at 50°C for 30,60, and 90 mn. Products clearly identifiable include the fragments generated by cleaving all four sites (shown by arrows on the gel).

However it seems that sites 996 and 1747 were preferentially cleaved.

Long arm ribozyme against the plus strand of WMV2 The ribozyme used was obtained from pBPL4736 (described in 4. above). The target contained 807 bases of viral sequence, between base 2163 (site XhoI) and base 2970 (start of polyA tail), with the polyA tail and polylinker sequence from BlueScript KS+ (between the T7 promoter and the EcoRI site).

Figure 11 shows the result of one assay.

Incubation of target with 4 fold excess ribozyme at three different temperatures showed that site 2259 was not cleaved as expected because of the mutation in that ribozyme (see 4. above). Site 2720 was the fastest cleaver and gave products at 22°C. Site 2403 was next and site 2520 appears the slowest.

When the ribozyme from plasmid pBPL4739 was tested, evidence of cleavage at site 2259 was seen (data not shown).

Short arm ribozyme against the minus strand of WMV2 The ribozyme used was obtained from pBPL4741 (described in 5. above). The target contained 1880 bases of viral sequence complementary to the sequence bewteen base 2278 (ClaI site) and base 398 (SpeI site), with polylinker sequence in 5'from BlueScript KS+ (betwen the T3 promoter and the ClaI site).

Figure 12 shows the result of an assay. All four sites are cleaved at high temperature. At 22°C, all sites except 696 are cleaved but cleavage is very slow as demonstrated by the very low abundance of double cleaved products.

Short arm ribozyme against the plus strand of WMV2 The ribozyme used was obtained from pBPL4742 (described in 5. above). The target was the same as described above for the long arm ribozyme.

Figure 13 shows the result of an assay. Note that the molar excess ribozyme over the substrate was 4.5 (+), 9 (++) and 13.5 (+++). All four sites are cleaved at 50°C; with four fold excess ribozyme, cleavage is essentially complete after 1.0 to 1.5 hour. At 22°C, only sites 2259 and 2403 are cleaved, with ribozyme 2259 being the fastest claver.

Short arm ribozyme against ZYMV: first combination (pBPL4746 and pBPL4747) The 2 polyribozymes showed some cleavage activity at 50°C, but very little at 37°C and none detectable at 22°C.

At 50°C, for the minus strand, all sites were cleaved; site 2442,1743, and 1245 were cleaved better than 2564 and 1929.

For the plus strand, all sites were cleaved ; sites 110,2590 and 2255 were cleaved faster than 2455 and 2161.

Short arm ribozyme against ZYMV: second combination (pBPL4750) The polyribozyme showed cleavage activity on the plus strand at all 4 predicted sites only at 50°C. No activity was detected at 22°C. The same polyribozyme however showed good cleavage activity on the only site present on the minus strand (site 2354) both at 50°C and 22°C.

Figure 14 shows the result of an assay of this ribozyme on the minus strand target. The target contained 1058 bases of viral sequence, complementary to the sequence between base 2633 (SpeI site) and base 1575 (XbaI site) as shown on figure 6, with 39 bases added in 5' (polylinker sequence between the T3 promoter and the XbaI site of BlueScript SK+) and 11 bases added in 3' (polylinker sequence between the XbaI and the BamHI site of SK+; BamHI was used to linearise the plasmid prior to in vitro transcription). For this expriment the labelle ribozyme from the pilot rection used to quantitate was mixed with the unlabelled ribozyme. The position of the ribozyme band on the gel is shown by an arrow. The ribozyme was used at 5 fold molar excess over the substrat.

Short arm ribozyme against ZYMV: third combination (pBPL4751) The ribozyme from pBPL4751 (targeting sites 2009 and 2027 on the minus strand) was assayed on the same target as the ribozyme from pBPL4750.

Figure 14 shows the result of this assay. Both sites were cleaved efficiently at both 22°C and 50°C.

The presence of a two bands of equal intensity for the products 5'-2 and 5'-3 (see figure 14) suggest that site 2009 is cleaved more readily or faster than 2027: about 50% of cleaved products are cleaved only at 2009.

9. Determination of resistance in transgenic plants 9.1. Genetic transformation of melon The method used is described in the following patent applications : French application n°8910848 of 11/08/89, European application n°904022829 of 09/08/90 and United States patent n° 5,422,259 of 06/06/95.

Genetically transformed plants regenerated, checked by PCR amplification for the presence of nptII gene, were transferred into pots containing a mixture of compost and coarse sand (2: 1) and placed in culture in a greenhouse. During the first seven days, the plants are protected by a plastic cover in order to maintain a humid atmosphere. The plants are watered daily. The plants are induced to flower, followed by self-fertilization. The seeds (T1 generation) are collecte.

9.2. Artificial inoculation of WMV2 and ZYMV on melon to detect resistant plants Virus trains.

The strain of WMV2 used is called « MAR ».

Two strains of ZYMV pathotype 0 are used: -one called « 13-18 » (pathotype F) is a wilting strain because of the symptoms induced on susceptible cultivars which bear the Fn gene (Flaccida necrosis) (Lecocq and Pitrat, 1984).

-one called « E15 » (pathotype NF) induces vein clearing symptoms for all the susceptible cultivars (Lecocq and Pitrat, 1984) independent of the presence or not of the Fn gene.

Virus storage : Leaves showing symptoms (vein clearing) are thinly cut with a razor blade and placed in boxes with calcium chloride to dry them. The boxes are stored at 6°C during a maximum of one year (Bos, 1969).

Virus multiplication : Melon plantlets of a susceptible variety at the two leaf stage are inoculated by rubbing the two cotyledons with the virus solution (Marrov, 1967).

Virus solution : One gram of leaves showing symptoms or one gram of the stored virus is ground in 4 ml of extraction buffer (1.075 g of disodium hydrogen phosphate, added with 0.2 g of sodium diethyldithiocarbamate for 100 ml of distille water) added with activated charcoal (0.1 g) and carborundum (0.0875 g) (Marrov, 1967). This virus solution has to be kept at cool temperature.

Virus inoculation : Forty-six to forty-eight seeds (T1 progeny) per genetic transformation event are sown in 10 cm square pots filled with a potting soil. The temperature is maintained from 18 to 25 °C with a photoperiod of 12 hours. Mechanical inoculation on melon plantlets is carried out at the first leaf stage (approximately 15 days after sowing) (Marrov, 1967). The plantlets are inoculated in the same way as for virus multiplication.

Observation of the symptoms : The symptoms are scored between 2 and 4 weeks after inoculation.

For WMV2: -susceptible plant: vein banding -tolerant plant: light vein banding or symptoms appearing late -resistant plant: no symptoms For ZYMV: -susceptible plant (absence of the Fn gene): vein clearing with the E15 or the 13-18 trains.

-susceptible plant (presence of the Fn gene): vein clearing with E15 strain and wilting rection with the 13-18 strain.

-resistant plant: no symptoms either with E15 or 13-18 strain.

In some cases, the control plants do not show symptoms of virus attack due to bad inoculation of the cotyledons. The behaviour of these control plants enables the determination of the value of the pathology test. « Escape » plants are considered to have escaped from a « correct » virus inoculation.

9.3. Resistance of transformed plants to WMV2 and ZYMV The transformed plants are demonstrated to be resistant to WMV2 and to ZYMV.

A. Resistance to WMV2 of melon plants genetically transformed with pBPL3754 With construct pBPL3754, eight genetic transformation events were analyse for WMV2 resistance behaviour (see table 1).

For six genetic transformation events, a resistance to WMV2 was observe. The segregation ratio (susceptible: resistant) was 3: 1, as expected for a single locus being introduced.

B. Resistance to WMV2 of melon plants genetically transformed with pBPL38010 With construct pBPL38010, fifteen genetic transformation events were analyse for WMV2 resistance behaviour (see table 2).

For twelve genetic transformation events, a resistance to WMV2 was observe. The segregation ratio (susceptible: resistant) was 3: 1 for nine genetic transformation events. For one genetic transformation event, the segregation ratio (susceptible: resistant) was 1: 3.

C. Resistance to WMV2 of melon plants genetically transformed with pBPL3817 With construct pBPL3817, twenty-one genetic transformation events were analyse for WMV2 resistance behaviour (see table 3).

For fourteen genetic transformation events, the segregation ratio (susceptible: resistant) was 1: 3 and for three genetic transformation events, the segregation ratio (susceptible: resistant) was 3: 1.

D. Resistance to ZYMV of melon plants genetically transformed with pBPL3906 With construct pBPL3906, thirteen genetic transformation events were analyse for ZYMV resistance behaviour (see table 4).

With the strain 13.18, the progeny of ten genetic transformed plants was tested. Four genetic transformation events showed a resistance segregation with a ratio (susceptible: resistant) of 3: 1.

With the strain E15, all the progeny tested from nine genetic transformation events were susceptible.

E. Resistance to ZYMV of melon plants genetically transformed with pBPL4186 With construct pBPL4186, five genetic transformation events were analyse for ZYMV resistance behaviour (see table 5).

One genetic transformation event from the five tested showed a segregation resistance (ratio susceptible: resistant of 3: 1) when tested with the strain 13.18.

With the strain E15, the progeny from the three independent genetic transformation events tested were susceptible.

F. Resistance to ZYMV of melon plants genetically transformed with pBPL4194 With construct pBPL4194, eight genetic transformation events were analyse for ZYMV resistance behaviour (see table 6).

With the strain 13.18, the progeny of four genetic transformation events from the seven tested showed a resistance segregation with a ratio (susceptible: resistant) of 3: 1.

With the strain E15, the progeny of the six genetic transformation events tested were susceptible.

G. Resistance to ZYMV of melon plants genetically transformed with pBPL4771 With construct pBPL4771, six genetic transformation events were analyse for ZYMV resistance behaviour (see table 7).

With the strain 13.18, two genetic transformation events were studied. The progeny of one genetic transformation event showed a segregation ratio (susceptible: resistant) of 3: 1 and the other one a ratio of 1: 3.

With the strain E15, the progeny of two genetic transformation events from the six tested showed a resistance segregation with a ratio (susceptible: resistant) of 1: 3.

Example 10-WMV2 Resistance in the Field Field trials of genetically transformed melon for virus resistance have been carried out.

10.1-Season n° l : concernedthegenetictransformationThistrial (insertionofpolyribozymeofpBPL3817inaeventW38 calledRZ33).Molecularanalysisofitsmelonline showedasinglelocusintegrationoftheT-genomicDNA DNA. weregrown:Twelvevarieties 1:generationT1oftheeventW38-variety 2:generationT2oftheeventW38-variety 3:progenyobtainedfromthecross-variety T2oftheeventW38andthebetweengeneration RZ33untransformedline 4to11:F1hybridsrealizedwithW38-variety andotherparentallines(T3generation) 12:untransformedline,RZ33-variety 3testsforeachvarietyandforeachTherewere plantswerecultured.Plantsweregrownintest,9 normalagronomicalpracticeandscoredaccordancewith perseasonforvirussymptoms.threetimes ofsymptomswerenoted0to9asTheobservations followed: --0 virussymptom,no --5 virussymptoms,severe --9 plantisnearlydead,coveredbyvirusthe symptoms. samplesweretakenforELISAanalysis.Someleaf dataandELISAvaluesaresetoutinTheobservation below.Table9 parentplantsandcontrolplantsNontransgenic attackedbyWMV2,showingthatthewereheavily washighthatyearinthisarea.infectionpressure mosaicsymptomswereobservedontheTypicalfoliar non-transformedmaterial.

By theendoftheseason,allat (varieties1to11)hadremainedtransgenicplants immune to WMV2. This can be seen by comparison of the optical density for the plant under test with those of the safe control x2 and inoculated control.

Legend of Table 9 : Safe control : non-transformed, non inoculated plant.

Inoculated control : non-transformed, inoculated plant.

% of viral infection : natural infection pressure for the indicated virus.

??? : indicates samping error.

"-" : indicates optical density below that of safe control.

Blank espace. indicates no sample taken.

10.2-Season n°2 The results obtained from the field trial in Season n° 2 are set out below.

This trial concerne the genetic transformation event W38 (insertion of polyribozyme of pBPL3817 in a melon line called RZ33).

Six varieties were groin -variety 1 : untransformed line, RZ33 -variety 2 : generation T2 of the event W38 -variety 3 : generation T3 of the event W38 -variety 4 : generation T4 of the event W38 -variety 5 fol hybrid realized with W38 (T4 generation) -variety 6 : control F1 hybrid realized with RZ33 There were 3 tests for each variety and for each test, 8 plants were culture. Plants were grown according to normal agronomical practice and scored five times per season for virus symptoms.

The observations of symptoms were noted 0 to 9 as followed : 0 : no virus symptom 5 : severe virus symptoms 9 : the plant is nearly dead, covered by virus symptoms Some leaf samples were taken for ELISA analysis.

The observation data and ELISA values are presented in Table 10 below.

Non transgenic parent plants and control plants were heavily attacked by WMV2 and CMV, showing that the infection pressure was high that year in this area.

Figure 21 shows the viral epidemiology in the field trial of Season n° 2, for WMV2, CMV, ZYMV, and PRSV.

Infection pressure for WMV2 was high during Season n° 2 (see Table 10"% of viral infection"). Typical foliar mosaic symptoms were observe on the non-transformed material. The virus symptom development expanded to the fruit as raised green"warts"and would therefore be considered non commercial. The identity of the virus was checked by ELISA.

By contras, at the end of the season, all transgenic plants (varieties 2 to 5) had remained immune to WMV2. Transfenic fruits were safe and of high quality.

Legend of Table 10 See legend of Table 9.

TABLE 1. RESULTS OF THE RESISTANCE TEST AGAINST WMV2 on T1 PROGENY FROM PLANTS GENETICAL<BR> TRANSFORMED PLANTS WITH pBPL3754<BR> Name Number of Number of Number of Number of Number of Most probable genetic deter<BR> of the T1 plants non-scorable susceptible tolerant resistant of resistance<BR> T0 plant tested plants plants plants plants<BR> 2.1 46 0 44 2 0 escape?<BR> 4.4 45 0 33 6 6 insertion at 1 locus of a resista@<BR> efficient in the homozygous @<BR> 4.10 43 0 21 8 14 insertion at 1 locus of a resist@<BR> efficient in the homozygous @<BR> 4.12 24 0 16 4 4 insertion at 1 locus of a resist@<BR> efficient in the homozygous @<BR> 8.1 47 0 36 0 11 insertion at 1 locus of a resist@<BR> efficient in the homozygous @<BR> 93.16 45 5 37 1 2 escape?<BR> 93.42 45 1 34 4 6 insertion at 1 locus of a resist@<BR> efficient in the homozygous@<BR> 93.47 26 0 19 2 5 insertion at 1 locus of a resist@<BR> efficient in the homozygous SULTS OF THE RESISTANCE TEST AGAINST WMV2 on T1 PROGENY FROM PLANTS GENETICALLY TRANSFORMED<BR> WITH pBPL38010<BR> Number Number Number Number Number Most probable genetic determinism<BR> of T1 of non- of of of of resistance<BR> plants scorable susceptible tolerant resistant<BR> tested plants plants plants plants<BR> 45 0 43 2 0 /<BR> 45 1 43 1 0 /<BR> 33 0 14 0 19 insertion at one locus ?<BR> 23 1 17 0 5 insertion at 1 locus of a resistance gene<BR> efficient in the homozygous condition<BR> 46 0 46 0 0 /<BR> 20 0 16 0 4 insertion at 1 locus of a resistance gene<BR> efficient in the homozygous condition<BR> 37 0 28 3 6 insertion at 1 locus of a resistance gene<BR> efficient in the homozygous condition<BR> 46 11 16 5 14 insertion at one locus ?<BR> 46 0 34 0 12 insertion at 1 locus of a resistance gene<BR> efficient in the homozygous condition<BR> 12 0 9 0 3 insertion at 1 locus of a resistance gene<BR> efficient in the homozygous condition<BR> (germination problem)<BR> 13 0 5 2 6 insertion at one locus ? of a resistant gene<BR> (germination problem)<BR> 45 0 19 13 13 insertion at 1 locus of a resistance gene<BR> efficient in the homozygous condition<BR> 44 1 17 17 9 insertion at 1 locus of a resistance gene<BR> efficient in the homozygous condition<BR> 46 1 18 10 17 insertion at 1 locus of a resistance gene<BR> efficient in the homozygous condition ?<BR> 17 1 5 2 9 insertion at 1 locus of a resistance gene<BR> efficient in the heterozygous condition TABLE 3. RESULTS OF THE RESISTANCE TEST AGAINST WMV2 on T1 PROGENY FROM PLANTS GENETICALL@<BR> TRANSFORMED WITH pBPL3817<BR> Name Number of Number of Number of Number of Number of Most probable genetic determini@<BR> of the T1 plants non- susceptible tolerant resistant of resistance<BR> T0 plant tested scorable plants plants plants<BR> plants<BR> 30.1 7 6 1 0 0 look of an haploïd plant<BR> 38.2 46 0 0 0 46 insertion at multiloci<BR> 44.2 18 8 4 1 5 insertion at 1 locus of a resistance<BR> efficient in the heterozygous cond<BR> (germination problem)<BR> 46.1 46 7 2 2 35 insertion at 2 loci of a resistance g<BR> efficient in the heterozygous cond<BR> 46.2 46 0 45 0 1 escape? /<BR> 46.3 44 2 9 4 29 insertion at 1 locus of a resistance<BR> efficient in the heterozygous cond<BR> 48.1 46 5 7 1 33 insertion at 1 locus of a resistance<BR> efficient in the heterozygous cond<BR> 72.3 46 0 10 11 25 insertion at 1 locus of a resistance<BR> efficient in the heterozygous cond<BR> 72.6 27 0 3 0 24 insertion at 1 or 2 loci of a resistanc@<BR> efficient in the heterozygous cond<BR> 72.7 36 1 3 5 27 insertion at 1 or 2 loci of a resistanc@<BR> efficient in the heterozygous cond<BR> 74.11 5 0 0 0 5 germination problem (table 3 - continued)<BR> RESULTS OF THE RESISTANCE TEST AGAINST WMV2 on T1 PROGENY FROM PLANTS GENETICALLY TRANSFOR<BR> WITH pBPL3817<BR> Name Number of Number of Number of Number of Number of Most probable genetic determini<BR> of the T1 plants non- susceptible tolerant resistant of resistance<BR> T0 plant tested scorable plants plants plants<BR> plants<BR> 77.1 41 0 10 7 24 insertion at 1 locus of a resistance<BR> efficient in the heterozygous cond<BR> 77.3 44 1 6 3 34 insertion at 1 locus of a resistance<BR> efficient in the heterozygous cond<BR> 77.4 31 0 2 0 29 insertion at 2 loci of a resistance g<BR> efficient in the heterozygous cond<BR> 77.5 44 0 7 0 37 insertion at 1 locus of a resistance<BR> efficient in the heterozygous cond<BR> 77.9 39 0 18 6 15 insertion at 1 locus of a resistance<BR> efficient in the homozygous cond<BR> 77.11 37 0 24 7 6 insertion at 1 locus of a resistance<BR> efficient in the homozygous cond<BR> 85.2 41 0 17 4 20 ??<BR> 87.9 46 0 35 1 10 insertion at 1 locus of a resistance<BR> efficient in the homozygous cond<BR> 87.13 40 0 5 18 17 insertion at 1 or 2 loci of a resistanc@<BR> efficient in the heterozygous condi<BR> 87.15 11 0 4 4 3 insertion at 1 locus of a resistant @ TABLE 4. RESULTS OF THE RESISTANCE TEST AGAINST ZYMV (strains 13.18 FN and E15) on T1 PROGENY FROM @<BR> GENETICALLY TRANSFORMED WITH pBPL3906<BR> Name Number Number of "FN reaction" Number<BR> of the of T1 non- or of Most probable genetic deter@<BR> T0 plant plants scorable susceptible resistant of resistance<BR> tested plants plants plants<BR> Strain 13.18, FN<BR> Z6.2 37 0 37 0 /<BR> Z6.4 43 0 43 0 /<BR> Z6.5 33 0 33 0 /<BR> Z6.6 18 0 18 0 /<BR> Z6.7 15 0 15 0 /<BR> Z11.41 38 0 34 4 escape?<BR> Z11.42 23 0 16 7 insertion at 1 locus of a resista@<BR> efficient in the homozygous c@<BR> Z19.12 30 0 20 10 insertion at 1 locus of a resista@<BR> efficient in the homozygous c@<BR> Z19.20 10 0 8 2 insertion at 1 locus of a resista@<BR> efficient in the homozygous c@<BR> Z19.21 19 0 11 8 insertion at 1 locus of a resista@<BR> efficient in the homozygous c@<BR> Strain E15<BR> Z6.3 36 0 36 0 /<BR> Z6.4 45 0 45 0 /<BR> Z6.5 17 0 17 0 /<BR> Z6.6 8 0 8 0 /<BR> Z6.7 10 0 10 0 /<BR> Z19.4 8 0 5 0 /<BR> Z19.20 11 0 11 0 /<BR> Z19.23 22 0 12 1 /<BR> Z20.1 23 12 9 2 / TABLE 5. RESULTS OF THE RESISTANCE TEST AGAINST ZYMV (strains 13.18, FN and E15) on T1 PROGENY FROM<BR> PLANTS GENETICALLY TRANSFORMED WITH pBPL4186<BR> Name Number Number "FN reaction" Number Most probable genetic determinism<BR> of of T1 of or of of resistance<BR> the plants non- susceptible resistant<BR> T0 plant tested scorable plants plants<BR> plants<BR> Strain 13.18, FN<BR> Z4.1 40 4 36 0 /<BR> Z5.1 38 0 38 0 /<BR> Z5.2 27 16 11 0 /<BR> Z12.7 40 0 32 8 insertion at 1 locus of a resistance gene<BR> efficient in the homozygous condition<BR> Z12.8 28 0 11 17 inoculation problem<BR> Strain E15<BR> Z4.1 46 0 46 0 /<BR> Z5.1 34 0 34 0 /<BR> Z12.8 19 0 19 0 / TABLE 6. RESULTS OF THE RESISTANCE TEST AGAINST ZYMV (strains 13.18,FN and E15) on T1 PROGENY FR<BR> GENETICALLY TRANSFORMED WITH pBPL4194<BR> Name of Number of T1 Number of "FN reaction" Number of Most probable gen@<BR> the T0 plant plants tested non-scorable or susceptible resistant plants of resist<BR> plants plants<BR> Strain 13.18, FN<BR> Z1.9 36 0 36 0 /<BR> Z2.5 41 0 41 0 /<BR> Z2.6 44 0 40 4 insertion at 1 locus of<BR> efficient in the homo<BR> Z2.7 47 4 43 0 /<BR> Z2.23 46 0 40 6 insertion at 1 locus of<BR> efficient in the homo<BR> Z18.6 45 0 29 16 insertion at 1 locus of<BR> efficient in the homo<BR> Z22.9 20 0 15 5 insertion at 1 locus of<BR> efficient in the homo<BR> Strain E15<BR> Z1.9 40 0 40 0 /<BR> Z2.5 41 0 41 0 /<BR> Z2.6 9 0 9 0 /<BR> Z2.7 38 0 38 0 /<BR> Z22.9 19 0 19 0 /<BR> Z22.13 12 8 4 0 / TABLE 7. RESULTS OF THE RESISTANCE TEST AGAINST ZYMV (strains 13.18, FN and E 15) on T1 PROGENY FROM<BR> GENETICALLY TRANSFORMED WITH pBPL4771<BR> Name of Number of T1 Number of "FN reaction" Number of Most probable genetic de@<BR> the T0 plant plants tested non-scorable or susceptible resistant plants of resistance<BR> plants plants<BR> Strain 13.18, FN<BR> Z23.15 28 5 15 8 insertion at 1 locus of a resis<BR> efficient in the homozygou@<BR> Z28.3 43 3 20 20 insertion at 1 locus of a resi@<BR> efficient in the heterozygou<BR> Strain E15<BR> Z23.15 39 6 20 13 insertion at 1 locus of a resis<BR> efficient in the homozygou@<BR> Z23.22 36 3 8 25 insertion at 1 locus of a resis<BR> efficient in the heterozygou@<BR> Z25.1 12 1 10 1 /<BR> Z25.14 11 6 3 2 /<BR> Z28.3 40 11 5 24 insertion at 1 locus of a resis<BR> efficient in the heterozygou@<BR> Z31.2 22 1 18 3 / TABLE 10<BR> Observation data and ELISA values from transformed melon field trial (Season N° 2)<BR> (pBPL 3817) Notation n° I Notation n° 2 Notation n'3 Notation n'4 Notation n'5 w c z w c z w Z w Z w z c c c c c VARI ETY N°BLOC Virus M M M Virus w M M Virus m Ni y Virus M M M Virus M M v v m v v m v v m v v m v v m 2 v 2 v 2 v 2 v 2 v i B1 0 3 1,36 6 1,20 6 2 41 2, 13-NJ 1 B2 2 3 4 5 NJ ; :" : : : > : : : £.." : ; : : : 1 83 2 0, 85--3 4 5 NJ 2 B1 0 == 0 = = = 0 = = 0 i lZ, : 1, 99 2 (pc) . w 2 B2 0 0 0 2 plt 5 2, 55-5 1,0-2,31 2 B3 0 0 0 fl, Q6 1, 49-2 NJ Q, 15- w 3 B I 0 0 0 0 a. p, 11 1, 87-2 (pc)..- sx ie. 3 B2 0 0 0 0,09 1,34 2 2, 59 1 pit 4 75-1, 06 3 B3 0 0 1,loiQ2, 0, 08 0 0 NJ 0, 23 4 Bl 0 0 0 0 Q ; 1. 7, :' : 1, 56-2 (pc) y 31 (1, 14 : :-- 4 B2 0 1 0 0 1, 98 2 (pc) S F 0, 16 4 B3 0 1 0 2 plt 3 0, 09 : : 0, 95-NJ x 1, 7- 5 B I 0 0 0 0 0 t, 0,05, ;' 0, 81 1 (PC) 0, 9 5 B2 0 0 0 0 2 (pc) 0, 18 5 B3 0 1 Q ; 04 0, 22-0 0 2 (pc) 4 58-1,4- 6 Bl 2 0,93 0,76 2 1 0, 98 4 1,08--5 1,62 0, 97 6 _ 6. __. _ B-_ _ O8_ _ _ 3 4 56 6 B3 0 4 5 1,05 I, 51-4 5 Optic density Safe'cbmtrol X 2 (2x =' : : 0, 04 : 0 ; 12 0,04 0,12 O, pr 0,65 0,13 0,67 0, 02 0 ; 45. ; O, I r' value Inoculated control 2, 17 2, 13 2, 17 2,13 1,56 2,37 1,56 2,37 1,37 2,23 0,61 % of viral infection j 9y z, ooooo o ioooo ooooo 0 The samples with optic density higer than the value of the safe control X2 were considered as positive<BR> @: samples considered as negative<BR> pc : chlorotic point<BR> NJ : can not be judged<BR> plt : plant TABLE 9 Observation data and ELISA values from transformed melon field trial (Season N° 1) (pBPL3817) NotationNotationn°2Notationn°31 WZWZWZ C C C M Y M Y M Y VARJETY N°BLOC Virus MVirus M Virus M VMVMVM 2 V2V2 00B10 B2 5 0, 1 0,4 0 4 0,02 1, 0 0 4 0,01 1, 9 0 B3 0 0 O O O s B2 0 1 0 0 0 3 0,01 1,4 0 B3 1 0 0 0 4 0,03 0,9 0 0 3 B1 0 0 0 B2 0 0 2 0,05 0,8 0 B3 0 1 1. 6??? 0,7 0 1 0,02 0,7 0 4 B1 0 0 3 0,03 1,6 0 B2 0 0 0 B3 0 5 0,09 I, 1 0 0 5 BI 0 0 0 B2 0 1 0,02 1, 8 3 0,03 1,4 0 B3 4 0 0,8 0 5 0,03 1, 4 5 0,03 0,7 0 6 B1 0 0 0 B2 0 3 0,04 1, 9 0 NJ 0,01 0,7 0 B3 0 0 7 B11 0 0 0 0 0 B2 0 2 0,03 1,7 0 NJ B3 0 0.. N) 8 B1 0 0 3 0,04 1, 5 0 B2 0 4 0,17 0,9 0 0 0009B1 B2 0 0 B3 1 0 0 0 0 0 10 B1 0 0 0 0,01 0,9 0 B2 0 4 0,03 0,6 0 5 0,09 0, 1 0 B3 0 3 0,19 0,6 0 5 11 B1 1 0 0 0 0 5 0,02 1,0 0 B2 0 1 051,4060,020,900, B3 6 0 0,5 0 6 0, 02 1 0 6 0,01 0,4 0 12 B1 5 1 0 0 4 2,1 0, 15 0 0 1, 86 0,15 0 B2 1 2, 1 0 0 3 2,2 0,17 0 2 2,16 1, 67 0 B3 5 2,2 0 0 4 2,1 0,8 0 4 1, 77 1, 28 0 OpticOpticdensity Safe control X2 (x2) 0,050,050,020,150,02 Value 1,62,21,82,41,462,36control % of viral infection 92 24 0., = ll. Transient assay of anti-viral ribozyme genes active 11.1. Description of the system The transient assay system for in vivo activity of ribozyme or antisense genes was described by Sawyer and Birch (1996). A target reporter gene and the ribozyme gene to be tested are cloned on two separate plasmids and co- introduced into Nicotiana suspension culture cells by particle bombardment. The target reporter gene is an in- frame fusion between the target sequence and the LUC sequence. This in frame fusion encodes a chimeric protein with luciferase activity. 48 hours after the plasmids delivery, the cells are placed under a camera which mesures light emission; light emission represents the level of expression of the target reporter gene. The ribozyme gene tested expresses transiently a ribozyme which interacts with the target sequence and decreases the quantity of the chimeric target-LUC MARNA which is detected as a reduced luciferase activity.

To measure the activity of various ribozyme genes protecting transgenic melon plants against potyvirus infection, we have built various target reporter genes with viral sequences fused to LUC.

11.2. Genes used (see figure 15) Target sites on the positive strand of the virus WMV2 (pBPL0188): A 540 bases fragment containing four sites targeted by the polyribozyme (2259,2403,2520,2720) was amplifie by PCR and cloned in-frame in front of LUC, under control of a modifie version of the 35S promoter, using the plasmid pGLFuz (see figure 17) as described by Sawyer and Birch (1996). Because the most 3'site is located close to the stop codon, it was also necessary to change one base of the viral sequence to remove this stop codon and provide a flanking sequence to this site.

Target sites on the negative strand of the virus WMV2 (pBPL0184 to 0187): Because the negative strand is not coding, it was impossible to clone a large sequence containing several ribozyme sites in frame with LUC. Four shorter sequences (between 180 and 220 bases in length) each containing one ribozyme site (696,996,1747,2067) were amplifie by PCR and cloned in-frame in front of LUC, under control of the modifie version of the 35S promoter in pGLFuz as above.

Ribozyme gene against WMV2 (pBPL0189): The short arm polyribozyme targeting four sites on the plus (2259,2403,2520 and 2720) and four sites on the minus strand (696,996,1747 and 2067) of WMV2 from pBPL4743 was cloned under control of the modifie version of the 35S promoter in pGLFuz.

Ribozyme gene against ZYMV (pBPLO190): One site on the minus strand (hereafter referred to as 1747) of the virus is present on both WMV2 and ZYMV. The pentameric ribozyme against the minus strand of ZYMV in pBPL4747 is targeting site 1747 amongst others. In this sequence, the ribozyme against site 1747 is identical to the ribozyme present in the construct against WMV2, but the arms are shorter: 8 bases each side. The pentameric ribozyme was cloned under control of the modifie version of the 35S promoter in pGLFuz.

11.3. Assay and results (see figure 16) The assays were carried out as described by Sawyer and Birch (1996). For each assay, target and ribozyme plasmid DNA (2xug each, approx. 25% molar excess of ribozyme) were co-precipitated onto tungsten particles and shot onto Nicotiana plumagnifolia cells on filter paper discs (harvested from suspension culture). After bombardment, the cells were cultured on the discs for 48 hours at 28° C in the dark. For luciferase assay, the cells were placed into a petri dish containing D-luciferin for 30 minutes, then imaged using a liquid nitrogen cooled CCD camera and AT1 software (Wright Instruments Ltd, Enfield, UK). Results are shown in figure 16.

The WMV2 polyribozyme (pBPL0189) could inhibit the target containing the four sites on the plus strand (pBPL0188) to an extent (70%) not seen in previous uses of the system with full length antisense when an equal amount of target and inhibitor gene was used (Sawyer and Birch, 1996).

Both ribozymes (pBPL0189 and pBPL0190) could inhibit the target containing site 1747 (pBPL0186) to an even greater extent (75%). Site 696 (pBPL0184) seemed not amenable to inhibition ; this absence of effect on one target shows that the effect observe is not simply caused by an interaction between the 2 plasmids or their promoter. Sites 996 (pBPL0185) and 2067 (pBPL0187) were slightly inhibited.

11.4. Comparison with in vitro data obtained previously Minus strand target (see figure 12): Site 696 was the site least cleaved in vitro by the WMV2 ribozyme, and was not cleaved at all at 22°C. This is in agreement with the transient assay results.

Sites 1747 and 996 seemed equally cleaved in vitro and 2067 was the site most amenable to cleavage in vitro, particularly at 22°C. This is not fully consistent with the transient assay results suggesting that 1747 is clearly more amenable to cleavage than the 2 others. The comparison of the in vitro and transient in vivo results suggests that the in vitro activity is not a perfect predictor of the transient activity in vivo (unable to rank sites 996,1747,2067), but it has some predictive capacity (it could predict that 696 would not be a good targe).

Plus strand target (see figure 13): When the ribozyme was tried in vitro on a target of about 900 bases, site 2259 was somewhat cleaved at 22°C and 37°C (with 4.5-13 fold excess ribozyme) and to a lesser extent site 2403; cleavage was far from complet.

At 50°C all four sites were cleaved, 2259 then 2403 being apparently the fastes. The transient in vivo assay shows that this ribozyme is active despite the low level of activity observe in vitro at low temperature.

ZYMV Ribozyme: When the ZYMV ribozyme was assayed in vitro on a ZYMV target, site 1747 was not cleaved at 22°C. This result cannot be compare directly with the in vitro assay on WMV2 since the target sequence is significantly different, except immediately around the site itself; it suggests however that shorter arms reduce the activity in vitro. Results of the transient assay suggest that at least for this particular site 1747, the arm length can be reduced with no significant effect on activity in vivo.

The predictive capacity of the in vitro assay for this system is therefore not good.

11.5. Construction of pGLFuz (see Figure 17): To create a guida :: luc translational fusion vector (pGLFuz) the guida gene was first modifie by PCR amplification to introduce novel XhoI and NsiI sites upstream of the start codon, replace the stop codon with an alanine codon (TGA-> GCG) and introduce a novel NotI site downstream of this codon. The modifie uidA gene PCR product was ligated into pGemT (Promega). The appropriate recombinant plasmid was designated pGemGUS.

The modifie guida gene from pGemGUS was ligated into pZorz vector as a NsiI-NcoI fragment. In the resulting plasmid clone, pGLFuzTAG, the guida coding region is flanked by an osa promoter and ocs polyadenylation region.

Due to the presence of a stop codon between the guida and luc genes, preventing a translational fusion, the luc gene in pGLFuzTAG was removed (by digestion with NotI, treatment with T4 DNA polymerase to create blunt ends, then digestion with XhoI) and replace by the luc gene from pSPNF+luc (Promega) (prepared by digestion with KpnI, treatment with T4 DNA polymerase and digestion with XhoI).

The appropriate recombinant plasmid was designated pGLFuz.

For construction of target genes, the guida coding sequence is replace by viral sequences. For construction of ribozyme genes, both the guida and luc coding sequences are replace by the ribozyme sequences. (see figure 15) REFERENCES Betzner et al., 1997, The Plant Journal, vol. 11, n° 3, p. 101-109.

Bos L (1969). Experience with a collection of plant viruses in leaf material dried and stored over calcium chloride and a discussion of litterature on virus preservation. Medec. Fac. Landbouwwet. Rijksuniv. Gent.

34:875-887 Firek et al., 1993, A wound-induced promoter driving nptII expression limited to dedifferentiated cells at wound sites is sufficient to allow selection of transgenic shoots. Plant Mol. Biol., vol. 22, p 129- 142.

Fromm M. E., Taylor L. P. and Walbot V., 1986. Nature, vol. 319, p. 791.

Guerineau F. and Mullineaux P., 1993, Chapter 4"Plant transformation and expression vectors"in Plant Molecular Biology LABFAX, ed. Croy R. R. D., BIOS Scientific Publishers, Blackwell Scientific Publications.

Haseloff and Gerlach, 1988, Simple RNA enzymes with new and highly specific endoribonuclease activities.

Nature, vol. 334, p. 585-591.

Lecocq H and Pitrat M (1984). Strains of ZYMV in Muskmelon (C. melo L.). Phytopath. Z., 11: 165-173 Marrov J (1967). Amélioration des méthodes de transmission mécanique des virus par absorption des inhibiteurs d'infection sur le charbon végétal. C. R.

Acad. Agric. Fr. 53: 972-981 Pascual Perez, Gérard Tiraby, Jean Kallerhoff and Joel Perret, 1989, Phleomycin resistance as a dominant selectable marker for plant cell transformation. Plant Molecular Biology, vol. 13, p 365-373.

Perriman et al, 1992, Extended target-site specificity for a hammerhead ribozyme. Gene, vol. 113, p. 157-163.

Sawyer B. and Birch R., 1996, A generic transient assay for antisense and ribozyme inhibition in vivo, vol 28, 1996, Proceedings of the Australian Society for Biochemistry and Molecular Biology.