CA2112373C - Genetic sequences encoding flavonoid pathway enzymes and uses therefor - Google Patents

Abstract

The present invention relates to a nucleic acid isolate comprising a sequence of nucleotides encoding, or complementary to a sequence encoding, a dihydrokaempferol (DHK) hydroxylating enzyme or derivative or path thereof. The present invention al-so relates to transgenic plants carrying and expressing the above-mentioned nucleic acid material.

Description

WO 93/01290 3 7 3 PCr/AU92/00334 GENETIC SEQUENCES ENCODING FLAVONOID PATHWAY
ENZYMES AND USES THEREFOR

The present invention relates generally to genetic sequences encoding flavonoid pathway metabolising enzymes and their use such as in the manipulation of pigmentation in plants and other organisms.

The flower industry strives to develop new and different varieties of flowering plants. An effective way to create such novel varieties is through the manipulation of flower colour and classical breeding techniques have been used with some success to produce a wide range of colours for most of the commercial varieties of flowers. This approach has been limited, however, by the constraints of a particular species' gene pool and for this reason it is rare for a single species to have a full spectrum of coloured varieties. Indeed, due to the limited availability of blue flowers, less than five percent of cutflowers sold through the auction system in Holland in 1988 were blue. Amongst the twelve top selling flowers, only iris and freesia offer blue coloured varieties and these varieties constitute less than four percent of all flower sales. The development of blue varieties of the major cutflower species, for example, rose, chrysanthemum, carnation and gerbera would offer a significant opportunity in both the cutflower and ornamental markets.

Flower colour is predominantly due to two types of pigment: flavonoids and carotenoids. Flavonoids contribute to a range of colours from yellow to red to blue. Carotenoids impart an orange or yellow tinge and are commonly the only pigment in yellow or orange flowers. The flavonoid molecules which make the major contribution to flower colour are the anthocyanins which are glycosylated derivatives of cyanidin, delphinidin, petunidin, peonidin, malvidin and pelargonidin, and are localised in the vacuole. The different anthocyanins can produce marked differences in colour. Flower colour is also influenced by 2 PCr/AU92/00334 co-pigmentation with colourless flavonoids, metal complexation, glycosylation, acylation, methylation and vacuolar pH (Forkmann, 1991).

The biosynthetic pathway for the flavonoid pigments (hereinafter referred to as the "flavonoid pathway") is well established and is shown in Figure 1 (Ebel and Hahlbrock, 1988; Hahlbrock and Grisebach, 1979; Wiering and de Vlaming, 1984; Schram et al, 1984; Stafford, 1990). The first committed step in the pathway involves the condensation of three molecules of malonyl-CoA with one molecule of p-coumaroyl-CoA. This reaction is catalysed by the enzyme chalcone synthase (CHS). The product of this reaction, 2',4,4',6'-tetrahydroxychalcone, is normally rapidly isomerized to produce naringenin by the enzyme chalcone flavanone isomerase (CHI). Naringenin is subsequently hydroxylated at the 3 position of the central ring by flavanone 3-hydroxylase (F3H) to produce dihydrokaempferol (DHK).
The B-ring of dihydrokaempferol can be hydroxylated at either the 3', or both the 3' and 5' positions, to produce dihydroquercetin (DHQ) and dihydromyricetin (DHM), respectively. Two key enzymes involved in this pathway are flavonoid 3'-hydroxylase (hereinafter referred to as 3'-hydroxylase) and flavonoid 3',5'-hydroxylase (hereinafter referred to as 3',5'-hydroxylase). The 3'-hydroxylase acts on DHK to produce DHQ and on naringenin to produce eriodictyol. The 3',5'-hydroxylase is a broad spectrum enzyme catalyzing hydroxylation of naringenin and DHK in the 3' and 5' positions and of eriodictyol and DHQ in the 5' position (Stotz and Forkmann, 1982), in both instances producing pentahydroxyflavanone and DHM, respectively. The pattern of hydroxylation of the B-ring plays a key role in determining petal colour.

WO 93/01290 2112373 PCr/AU92/00334 Flavonoid 3'-hydroxylation in microsomal extracts requires NADPH and 02 as well as the aglycone of either naringenin or DHK. The parsley cell culture enzyme has been well studied (Hagmann et a.L, 1983). Inhibition by carbon monoxide, cytochrome c and NADP+ indicated that the enzyme is a cytochrome P450-dependent enzyme. A similar enzyme activity has been demonstrated in maize (Larson and Bussard, 1986). The 3, 5'-hydroxylase is also of the cytochrome P450 class of enzymes. Cytochrome P450 enzymes are widespread in nature and have been characterised in vertebrates, insects, yeasts, fungi, bacteria and one plant. Sequences of at least 154 cytochrome P450 genes have been determined and the genes grouped into 27 different gene families (Nebert et al, 1991). Within a single family, the P450 protein sequences are generally >40% identical whilst sequences within the same subfamily are > 46% identical (Nebert et al, 1991). Information on plant cytochromes P450 is limited.

The ability to control in plants 3' or 3',5'-hydroxylase activity, or other enzymes involved in the flavonoid pathway, would provide a means to manipulate petal__colour thereby enabling a single species to express a broader spectrum of flower colours. In accordance with the present invention, the genetic sequences encoding flavonoid pathway enzymes such as 3',5'-hydroxylase have been identified and cloned. These recombinant sequences permit the modulation of DHK metabolism as well as the metabolism of other substrates such as DHQ, naringenin and eriodictyol, thereby determining the hydroxylation pattern of the anthocyanins and providing a means to manipulate petal colour. The present invention, however, extends beyond flowers to fruit and vegetable plants and to leaves of, for example, ornamental plants.

Accordingly, one aspect of the present invention provides a nucleic acid isolate comprising a sequence of nucleotides encoding, or complementary to a sequence encoding, a dihydrokaempferol (DHK) hydroxylating enzyme, or a derivative or part thereof.

In one aspect, there is described an isolated DNA
molecule comprising: (a) a sequence of nucleotides encoding a 3',5'-hydroxylase or encoding a derivative of 31,51-hydroxylasewhich is capable of hydroxylating dihydrokaempferol (DHK); or (b) a sequence complementary to the sequence in (a); wherein the sequence in (a) comprises:
(i) a nucleotide sequence set forth in Figure 9 or 10; or (ii) a nucleotide sequence capable of hybridizing to the complement of the nucleotide sequence set forth in Figure 9 or 10 under stringency conditions of 20% v/v formamide, 6 x SSC, 1% w/v SDS at 42 C for 16 hours, followed by washing in 6 x SSC, 1% w/v SDS at 65 C for 1 hour.

In another aspect, there is described an isolated DNA
molecule comprising: (a) a sequence of nucleotides encoding a 3',5'-hydroxylase or encoding a derivative of 3',51-hydroxylase which is capable of hydroxylating dihydrokaemferol (DHK); or (b) a sequence complementary to the sequence in (a); wherein the sequence in (a) comprises:
(i) a nucleotide sequence set forth in Figure 9 or 10; or (ii) a nucleotide sequence encoding an amino acid sequence having at least 60% identity after optimal alignment to the amino acid sequence shown in Figure 9 or 10.

In another aspect, there is described a vector comprising the DNA molecule of the invention capable of transferring the DNA molecule to a plant cell or tissue.

4a In another aspect, there is described a protein having DHK hydroxylating activity having an amino acid sequence encoded by the DNA molecule of the invention.

In another aspect, there is described an isolated 3',5'-hydroxylase protein comprising the amino acid sequence as set forth in Figure 9 or 10, or an isolated derivative of 3',5'-hydroxylase which is capable of hydroxylating dihydrokaempferol (DHK), wherein the derivative is encoded by a nucleotide sequence capable of hybridizing to the complement of the nucleotide sequence shown at Figure 9 or 10 under stringency conditions of 20% v/v formamide, 6 x SSC, 1% w/v at 42 C for 16 hours, followed by washing in 6 x SSC, 1% w/v SDS at 65 C for 1 hour.

In another aspect, there is described a 3',5'-hydroxylase protein having DHK hydroxylating activity, having an amino acid sequence shown in Figure 9 or 10, or an amino acid sequence having at least 60% identity after optimal alignment to said amino acid sequence.

In another aspect, there is described a method for producing a transgenic plant capable of expressing a recombinant 3',5'-hydroxylase or which directs transcription of DNA sequence which is complementary to all or part of a mRNA molecule translatable to a 3',5'-hydroxylase, said method comprising introducing into a cell of a plant the DNA

molecule of the invention under conditions permitting the expression of said DNA molecule, regenerating a transgenic plant from the cell and growing said transgenic plant for a time and under conditions sufficient to permit the expression of the DNA molecule.

In another aspect, there is described an isolated DNA
molecule encoding a flavonoid 3',5'-hydroxylase capable of hydroxylating dihydrokaempferol (DHK), wherein said isolated 4b DNA molecule comprises the nucleotide sequence as set forth in Figure 9 or Figure 10.

In another aspect, there is described a transgenic plant cell incorporating a DNA molecule of the invention, or a vector of the invention.

In another aspect, there is described a method for cloning a nucleic acid molecule comprising a sequence of nucleotides encoding a flavonoid 3',5'-hydroxylase or derivative thereof which hydroxylates one or more of DHK and naringenin, or the sequence complementary thereto, said sequence comprising: the nucleotide sequence set forth in Figure 9 or 10, or the sequence complementary to either the nucleotide sequence in Figure 9 or 10, or a nucleotide sequence capable of hybridizing to the nucleotide sequence set forth in Figure 9 or 10, or to the sequence complementary thereto, under stringency conditions of 20%
v/v formamide, 6 x SSC, 1% w/v SDS at 42 C for 16 hours followed by washing in 6 x SSC, 1% w/v SDS at 65 C for 1 hour, said method comprising amplification of flavonoid 31,5'-hydroxylase nucleotide sequences or complementary sequences from a suitable preparation of nucleic acid molecules from cells of said plant by one or more polymerase chain reactions using one or more oligonucleotide primers based on the amino acid sequence PFGAGRRICPG.

In another aspect, there is described a method for producing a transgenic plant cell capable of expressing a recombinant 31,5'-hydroxylase or which directs transcription of a nucleic acid sequence which is complementary to all or a part of a mRNA molecule translatable to a 3',5'-hydroxylase, said method comprising introducing into a cell 4c of a suitable plant the nucleic acid isolate of the invention under conditions permitting the eventual expression of said nucleic acid isolate.

In another aspect, there is described an isolated DNA molecule for use as an antisense molecule comprising a complementary sequence, wherein the complementary sequence:
(a) is complementary to a length of the nucleotide sequence set forth in Figure 9 or 10; or (b) is a length of a nucleotide sequence capable of hybridizing to the nucleotide sequence set forth in Figure 9 or 10 under stringency conditions of 20% v/v formamide, 6 x SSC, 1% w/v SDS at 42 C
for 16 hours, followed by washing in 6 x SSC, 1% w/v SDS at 65 C for 1 hour; wherein the length is sufficient for the complementary sequence to act as an antisense molecule to down-regulate expression of a nucleotide sequence encoding a 3',5'-hydroxylase or a derivative thereof capable of hydroxylating dihydrokaempferol (DHK).

For convenience and by way of shorthand notation only, reference herein to "DHK hydroxylating enzyme"

includes flavonoid pathway hydroxylating enzymes acting on one or more of the following: DHK, DHQ, naringenin, eriodicytol.

Preferably, the DHK hydroxylating enzyme is 3',5'-hydroxylase. However, the methods employed to clone the genetic sequences encoding this enzyme could be employed to isolate other genetic sequences encoding enzymes such as the 3'-hydroxylase. Accordingly, reference herein to the isolation and cloning of 3',5'-hydroxylase should be taken to include reference to other flavonoid hydroxylating enzymes such as 31-hydroxylase.

4d By the term "nucleic acid isolate" is meant a genetic sequence in a non-naturally occurring condition.
Generally, this means isolated away from its natural state or formed by procedures not necessarily encountered in its natural environment. More specifically, it includes nucleic acid molecules formed or maintained in vitro, recombinant or synthetic molecules and nucleic acids in combination with heterologous nucleic acids. It also extends to naturally occurring sequences following at least a partial purification relative to other nucleic acid sequences.

By "genetic sequences" as used herein is meant any contiguous series of nucleotide bases specifying directly, or via a complementary series of bases, a sequence of amino acids in a DHK hydroxylating enzyme, for example, 3',51-hydroxylase. The nucleic acid or its complementary form may encode the full length enzyme or a derivative or part thereof. By "derivative" is meant any WO 93/01290 M23 ~ ~ PCT/AU92/00334 single or multiple amino acid substitutions, deletions, and/or additions relative to the naturally occurring enzyme. In this regard, the nucleic acid includes the naturally occurring nucleotide sequence encoding 3',5'-hydroxylase or may contain single or multiple nucleotide substitutions, deletions and/or additions to said naturally occurring sequence. The nucleic acid sequences contemplated herein also encompass oligonucleotides useful as genetic probes or as "anti-sense" molecules capable of regulating expression of the corresponding gene in a plant. Accordingly, when the nucleic acid or its complementary form encodes a"part" of the 3',5'-hydroxylase, then such a nucleic acid molecule may be useful as an oligonucleotide probe, primer for polymerase chain reactions or in various mutagenic techniques.

Amino acid insertional derivatives of the DHK hydroxylating enzyme and in particular the 3',5'-hydroxylase of the present invention include amino and/or carboxyl terminal fusions as well as intra-sequence insertions of single or multiple amino acids. Insertional amino acid sequence variants are those in which one or more amino acid residues are introduced into a predetermined site in the protein although random insertion is also possible with suitable screening of the resulting product. Deletional variants are characterised by the removal of one or more amino acids from the sequence. Substitutional amino acid variants are those in which at least one residue in the sequence has been removed and a different residue inserted in its place. Typical substitutions are those made in accordance with the following Table 1:

~ " - -t) Suitable residues for amino acid substitutions OriQinal Residue E.xemp~ Substitutions Ala Ser Arg Lys Asn Gln; His Asp Glu Cys Ser Gln Asn Glu Asp Gly Pro His Asn; Gln Ile Leu; Val Leu Ile; Val Lys Arg; Gln; Glu Met Leu; Ile Phe Met; Leu; Tyr Ser Thr Thr Ser Trp Tyr Tyr Trp; Phe Val Ile; Leu Where the 3',5'-hydroxylase is derivatised by amino acid substitution, the amino acids are generally replaced by other amino acids having like properties, such as hydrophobicity, hydrophilicity, electronegativity, bulky side chains and the like. Amino acid substitutions are typically of single residues. Amino acid insertions will usually be in the order of about 1-10 amino acid residues and deletions will range from about 1-20 residues. Preferably, deletions or insertions are made in adjacent pairs, i.e. a deletion of two residues or insertion of two residues.

The amino acid variants referred to above may readily be made using peptide synthetic techniques well known in the art, such as solid phase peptide synthesis (Merrifield, 1964) and the like, or by recombinant DNA
manipulations. Techniques for making substitution mutations at predetermined sites in DNA having known or partially known sequence are well known and include, for example, M13 mutagenesis. The manipulation of DNA sequence to produce variant proteins which manifest as substitution.al, insertional or deletional variants are conveniently described, for example, in Sambrook ec al., (1989).

Other examples of recombinant or synthetic mutants and derivatives of the 3',5'-hydroxylase of the present invention include single or multiple substitutions, deletions and/or additions of any molecule associated with the enzyme such as carbohydrates, lipids and/or proteins or polypeptides.

The terms "analogues" and "derivatives" also extend to any functional chemical equivalent of the 3',5"-hydroxylase and also to any amino acid derivative described above.

The nucleic acids of the present invention may be ribonucleic acids or deoxyribonucleic acids, single or double stranded and linear or covalently closed circular molecules. Preferably, the nucleic acid molecule is cDNA. The present invention also extends to other nucleic acid molecules which hybridise under low, preferably under medium and most preferably under high stringency conditions with the nucleic acid molecules contemplated by the present invention. Expressed in alternative terms, the present invention extends to a nucleic acid molecule having a nucleotide s4quence set forth hereinafter (e.g. the cDNA inserts from pCGP176, pCGP602 and pCGP175) or to a molecule having at least 35%, more preferably at least 45%, even more preferably at least 55%, still more preferably at least 65-70%, and yet even more preferably greater than 85% similarity at the level of WO 93/01290 21.~_~ ~ ~373 PCT/AU92/00334 nucleotide or amino acid sequence and wherein the nucleic acid encodes or is complementary to a sequence which encodes an enzyme having 3',5'-hydroxylase activity. It should be noted, however, that nucleotide or amino acid sequences may have similarities below the above given percentages and yet still encode DHK hydroxylating enzymes and such molecules may still be considered in the scope of the present invention where they have conserved regions of homology. The present invention further extends to nucleic acid molecules in the form of oligonucleotide primers capable of hybridising to a portion of the nucleic acid molecules contemplated above under low, preferably under medium and most preferably under high stringency conditions.

The nucleic acid molecules contemplated herein may exist alone or in combination with a vector molecule and preferably an expression-vector. Such vector molecules replicate and/or express in eukaryotic and/or prokaryotic cells. Preferably, the vector molecules or parts thereof are capable of integration into the plant genome. The nucleic acid molecule may additionally contain a promoter sequence capable of directing expression of the nucleic acid molecule in a plant cell. The nucleic acid molecule and promoter may be introduced into the cell by any number of means such as by electroporation or Agrobacterium mediated transfer.

The present invention is exemplified using nucleic acid sequences derived from petunia since this represents the most convenient and preferred source of material to date. However, one skilled in the art will immediately appreciate that similar sequences can be isolated from any number of sources such as other plants or certain microorganisms. The genetics of the 3'-hydroxylases is known in flowers of Antirrhinurr~ Verbena and Petunia and in seedlings and aleurone layers of seeds of maize (Heller and Forkmann, 1988). The gene CD.E
controls 3'-hydroxylase in Antirrhinum (Forkmann and Stotz, 1981), while the Hi1 and a genes control similar enzymes in Petunia (Stotz et aL, 1985) and in maize aleurone layer, respectively (Larson and Bussard, 1986). Chemogenetic studies of Verbena hybridq for example, have shown that in this plant the hydroxylation of the B-ring of anthocyanins in both the 3' and 5' position is controlled by one gene (Beale, 1940). Enzyme activity for 3',5'-hydroxylation is only present in flower extracts of delphinidin-producing strains (Stotz and Forkmann, 1982). NADPH-dependent microsomal enzyme activity for hydroxylation in the 3' and 5' position was also demonstrated in flower extracts of Callistephus and Lathyrus (Forkmann, 1991). As in V. hybridq enzyme activity for the 3',5'-hydroxylation of flavanones and dihydroflavonols was only found to be present in flower extracts of those strains which contain 3',4',5'-hydroxylated flavonoid compounds (or methylated derivatives of them) in the flowers. Thus, formation of the 3',4',5'-hydroxylated flavonoids is clearly dependent on flavonoid 3',5'-hydroxylase activity.

The genes encoding 3',5'-hydroxylase have been identified in a number of ornamental plants including Callistephus (R), Petunia (Hfl, Hf2) and Verbena (P.) by the presence of the respective mutants unable to produce delphinidin.
Moreover, the respective enzyme activities have been demonstrated (Forkmann, 1991). The 3',5'-hydroxylase was also considered to be a microsomal cytochrome P450 enzyme (Heller and Forkmann, 1988). However, there are no published reports of the cloning of a 3',5'-hydroxylase gene from these or other plant species.

Other plant species capable of producing 3',4',5'-hydroxylated flavonoids or their derivatives include hydrangea (Takeda et al, 1985), delphinium (Asen et al, 1975), lisianthus (Asen et al, 1986), tomato (von Wettstein-Knowles, 1968) and potato (Harborne and Simmonds, 1962). These species, or other plants capable of producing 3',4',5'-hydroxylated flavonoids, would also be suitable sources for the isolation of a 3',5'-hydroxylase gene. All such nucleic acid sequences encoding directly or indirectly a flavonoid pathway enzyme (e.g.
3',5'-hydroxylase) are encompassed by the present invention regardless of their source.

aa 2 ~ ~L L~~~' - 10-Likewise, the gene cloning strategy outlined here may be used to isolate a 3',5'-hydroxylase gene from other plants which produce 3',4',5'-hydroxylated flavonoids. Clones and oligonucleotides herein disclosed may be used to detect, isolate and clone similar genetic sequences using the same technology as herein described, although some minor modification(s) to the experimental procedures may be required. All such minor variations are encompassed by the present invention. Examples of other suitable sources of enzymes such as 3',5'-hydroxylase include, but are not limited to, verbena, delphinium, grape, iris, freesia, hydrangea, cyclamen, potato, pansy and eggplant.
In accordance with the present invention, a nucleic acid sequence encoding a DHK hydroxylating enzyme such as 3',5'-hydroxylase may be introduced into and expressed in a transgenic plant thereby providing a means to convert DHK
and/or other suitable substrates, if synthesized in the plant cell, ultimately into anthocyanin derivatives of anthocyanidins such as delphinidin, petunidin or malvidin. The production of these anthocyanins contributes to the production of a variety of shades of blue colour or blue-like colour. Expression of the nucleic acid sequence in the plant may be constitutive, inducible or developmental.
Accordingly, another aspect of the present invention provides a method for producing a transgenic plant capable of expressing a recombinant DHK
hydroxylating enzyme or active mutants or derivatives thereof, said method comprising introducing into a cell of a suitable plant a nucleic acid molecule which comprises a sequence of nucleotides encoding said DHK hydroxylating enzyme, under conditions permitting the eventual expression of said nucleic acid molecule, regenerating a transgenic plant from the cell and growing said transgenic plant for a time and under conditions sufficient to permit the expression of the nucleic acid.

WO 93/01290 PC.T/AU92/00334 In a preferred embodiment, the present invention contemplates a method for producing a transgenic flowering plant exhibiting altered inflorescence properties, said method comprising introducing into a cell of a suitable plant the nucleic acid sequence of the present invention under conditions permitting the eventual expression of said nucleic acid sequence, regenerating a transgenic plant from the cell and growing said transgenic plant for a time and under conditions sufficient to permit the expression of the nucleic acid sequence into the DHK hydroxylating enzyme.
Preferably, the DHK hydroxylating enzyme is 3',5'-hydroxylase, is developmentally regulated and the altered inflorescence includes the production of blue or red flowers or other colour shades depending on the physiological conditions of the recipient plant. By "suitable plant" is meant a plant capable of producing a substrate for the 3',5'-hydroxylase enzyme, and possessing the appropriate physiological properties and genotype required for the development of the colour desired. This may include but is not limited to rose, petunia, carnation, chrysanthemum and gerbera. In certain plant species it may be preferable to select a "high pH line", such being defined as a variety having a higher than average petal vacuolar pH. The origin of the recombinant 3',5'-hydroxylase or its mutants and derivatives are as hereinbefore described and include enzymes of petunia, verbena, delphinium, grape, iris, freesia, hydrangea, cyclamen, potato, pansy or eggplant origin.

One skilled in the art will immediately recognise the variations applicable to this method such as increasing or decreasing the expression of the enzyme naturally present in a target plant. This would lead to differing shades of colours such as different shades of blue or red.

In order to decrease activity of a target enzyme, such as 3',5'-hydroxylase, the nucleic acid sequence encoding this enzyme or various parts thereof could be used in the antisense orientation. Although not wishing to limit the present WO 93/01290 PCr/AU92/00334 _12_ invention to any one theory, it is probable that such an antisense nucleic acid sequence would form a duplex with all or part of the naturally occurring mRNA specified for the enzyme thus preventing translation of the mRNA into active enzyme. Alternatively, ribozymes could be used to inactivate target nucleic acid sequences.

Accordingly, the present invention extends to a method for producing a transgenic plant capable of expressing a recombinant dihydrokaempferol (DHK) hydroxylating enzyme or which directs transcription of a nucleic acid sequence which is substantially complementary to all or a part of a mRNA
molecule translatable to DHK hydroxylating enzyme, said method comprising introducing into a cell of a suitable plant the nucleic acid isolate according to claim 1 or 6 under conditions permitting the eventual expression of said nucleic acid isolate, regenerating a transgenic plant from the cell and growing said transgenic plant for a time and under conditions sufficient to permit the expression of the nucleic acid isolate. In this embodiment, suitable recipient plants extend to inter alia, iris, tulip, lily, lisianthus, freesia, delphinium, limonium and pelargonium.

The above methods of producing transgenic plants, therefore, extend to the alternative of introducing a gene or DNA fragment encoding an antisense mRNA or oligonucleotide to all or a portion or region of a sequence of nucleotides encoding, or complementary to a sequence encoding, a 3',5'-hydroxyl ase.
Consequently, the present invention extends to all transgenic plants containing all or part of the nucleic acid sequence of the present invention, or antisense forms thereof and/or any homologues or related forms thereof and in particular those transgenic plants which exhibit altered inflorescence properties. The transgenic plants, therefore, contain a stably introduced nucleic acid molecule comprising a nucleotide sequence encoding or complementary to a sequence encoding a DHK hydroxylating enzyme and in 23'199-181 particular high pH plant lines carrying such introduced nucleic acid molecules as referred to above. The invention also extends to seeds from such transgenic plants. Such seeds, especially if coloured, will be useful as proprietary tags for plants.

A further aspect of the present invention is directed to recombinant forms of the DHK hydroxylating enzymes and in particular recombinant 3',5'-hydroxylase. The recombinant forTns of the enzymes will provide a source of material for research to develop, for example, more active enzymes and may be useful in developing in vitro systems for production of coloured compounds.

Another aspect of the present in~vention contemplates a method for cloning a nucleic acid molecule comprising a'sequence of nucleotides which encode or are complementary to a sequence which encode, a cytochrome P450 molecule or like molecule from a plant, said method comprising amplification of cytochrome P450 nucleotide sequences or complementary sequences from a suitable preparation of nucleic acid molecules from cells of said plant by polymerase chain reactions using one or more oligonucleotide primers, said primers having a nucleotide sequence derived from one or more consensus sequences of known microsomal cytochrome P450 molecules.

In a related embodiment, the method for cloning the cytochrome P450 nucleic acid molecules or their complementary sequences comprises selecting from a suitable cDNA library a clone capable of hybridising to one or more oligonucleotide primers corresponding to one or more consensus sequences or known cytochrome P450 molecules.

Preferably, one of the consensus sequences.is from the haem-binding domain of cytochrome P450 molecules and is more preferably F(G,S) XGXRXCXG
(wherein X is any amino acid) or is PFGAGRRICPG . In a most preferred embodiment, the nucleotide sequences to be cloned encode or are complementary to sequences which encode, a DHK hydroxylating enzyme. and in particular 3',5'-hydroxylase. Even more preferably, the 3',S'-hydroxylaaze is as hereinbefore described and, more particularly, has an amino acid sequence or is encoded for by a sequence of nucleotides substantially as in the cDNA inserts in pCGP175, pCGP176 and pCGP602 or has similarity thereto as defined above.

The present invention is further described by reference to che following non-limiting Figures and Example.

Figures 1(A) and (B) are schematic -epresentations of the biosynthesis pathway for the flavonoid pigments. Enzymes involved in the first part of the pathway have been indicated as follows: PAL = Pheaylalanine ammonia-lyase; C4H
= Cinnamate 4-hydroxylase; 4CL = 4-coumarate: CoA ligase; CHS =
Chalcone synthase; CHI = Chalcone flavanone isomerase; F3H = Flavanone 3-hydroxylase; DFR = Dihydroflavonol-4-reductase; UFGT = UDP-glucose:
flavonoid-3-O-glucosyltransferase. The later steps correspond to conversions that occur in P. hybrida flowers and include: 1= addition of a rhamnose sugar to the glucosyl residue of cyanidin-3-glucoside and delphinidin-3-glucoside; 2= acylation and 5-0-glucosylation; 3 3' methylation; 4 = 5' methylation; 5 = 3'5' methylation.
Figure 2(A) shows 3',5'-hydroxylase activity in petal extracts of P. hybrida cv V23 (Hf1 /Hf1, HV/I3f2) and the lack of 3',5'-hydroxylase activity in P.
hybrida cv R51 (hfl/hf1, hf2 /hf2). 3',5'-hydroxylase activity was detected by conversion of 3H-naringenin to the 3'- and 3',5'-hydroxylated derivatives eriodictyol and pentahydroxyflavanone. On the left-hand side of the figure the biochemical structures of the substrate, naringenin and the product of the 3'-hydroxylase reaction, eriodictyol and the 3',5'-hydroxylase reaction, pentahydroxyflavanone are shown. The location of the substrate and the hydroxylated products on the TLC plate is indicated on the right-hand side of the Figure which shows from left to right the autoradiographs of the reaction products produced by petal extracts of flowers from P. hy6rida cv V23 and P.
hybrida cv R51 and the control showing no hydroxylation of naringenin when NADPH is omitted from the reaction mixture.

Figure 2(B) shows 3',5'-hydroxylase activity in petal extracts of P. hybrida cv Old Glory Blue (OGB) flowers at different developmental stages. From left to right the autoradiographs of the TLC plates show (1) Stage 1 flowers [Unpigmented, closed bud (<25 mm in length)]: limited conversion of naringenin to the 3',5'-hydroxylated derivative pentahydroxyfl avanone, (2) Stage 2 flowers [Pigmented, closed bud (25-35 mm in length)]: increased conversion to pentahydroxyflavanone indicative of higher 3',5'-hydroxylase levels, (3) Stage 3 flowers: [Dark purple bud with emerging corolla (>35 mm in length)]: maximal 3'5'-hydroxylase activity, (4) Stage 4 flowers [Dark purple opened flower pre-anther dehiscence (>50 mm in length)]: maximal 3',5'-hydroxylase activity (5) Stage 5 flowers [Fully opened flower with all anthers dehisced]: no detectable 3',5'-hydroxylase levels.
Figure 3(A) is a schematic representation of a mRNA molecule encoding a cytochrome P450. The shaded region indicates the relative position of sequences encoding the haem binding domain. A consensus amino acid sequence for the most conserved region of this domain has been shown using single letter code. Amino acids that are present in 100% of cytochrome P450 sequences present in the SWISS-PROT database have been boxed and X
indicates positions where there is a low level of sequence conservation.
Figure 3(B) shows the position of oligos used for PCR amplification of cytochrome P450 molecules pCGP450 and pCGP454 from cDNA library #1.
Oligos 1 and 3 covered sequences in the conserved haem binding domain while oligos 2 and 4 corresponded to the pBluescript (Stratagene) -20 and reverse primer sequences, respectively. Oligos 1 and 2 were used to synthesize the cDNA insert in pCGP450; oligos 3 and 4 were used the synthesize the cDNA insert in pCGP454. Representation of a generalized cDNA molecule is identical to that shown in Fig. 3A; vector sequences have been indicated by light shading.

Figure 4(A) is a schematic representation of DNA fragments used to probe cDNA library #1 to identify cytochrome P450 homologues, including pCGP174 and pCGP175. P450: generalized cytochrome P450 cDNA clone with the haem-binding domain (Haem) indicated by the shaded box; Fragment 1: a 900 bp fragment was obtained by PCR with oligos 5 and 6 using pCGP142 DNA as template; Fragment 2: a 1.3 kb fragment was isolated from a SZ1.I-F..-C~I
digest of pCGP147; Fragment 3: a 750 bp fragment was obtained by PCR with oligos 4 and 7 using pCGP158 DNA as template; Fragment 4: a 670 bp fragment was isolated from a PqI-F.saRV digest of pCGP160; Fragment 5: a 150 bp fragment was obtained by PCR with oligos 3 and 4 using pCGP454 DNA as template. All purified fragments were labelled with 32P-dCTP as described in the Materials and Methods.

Figures 4(B) to (H) show partial nucleotide sequences and the corresponding predicted amino acid translation products for the cDNA inserts from (i) pCGP142, (ii) pCGP147, (iii) pCGP158, (iv) pCGP160 and (v) pCGP454. The regions used to probe cDNA library #1 to isolate pCGP174 and pCGP175 have been delineated by arrowheads.

Figures 5 (A) and (B) are diagrammatic representations of plasmids pCGP174 and pCGP175, respectively. The cDNA inserts are indicated as open boxes with the region encoding the putative haem-binding domain shown as a shaded region. An F.sORI site is at the 5'-end and aXhQ-I site is at the 3'-end of both cDNA inserts.
Figure 6(A) is an autoradiograph of an RNA blot probed with the 3' region of the pCGP174 cDNA insert. Each lane contained a 20 pg sample of total RNA
isolated from the following petunia tissues- 1-5: OGB limb tissue of flowers at the five (1-5) different stages of flower development described in the Materials and Methods; T: OGB tube tissue from stage 3-4 flowers; L: leaf tissue from 6 week old OGB seedlings; IL: glucose/high light treated leaf tissue from 6 week old OGB seedlings; V23: V23 limb tissue from stage 3-4 flowers; R51:

WO 93/01290 211t' ~3i r~~' PCr/AU92/00334 d R51 corolla tissue from stage 3-4 flowers; VR: petal limb tissue from stage 3-flowers of the V23xR51 F1 hybrid; Sw63: petal limb tissue from stage 3-4 flowers of Sw63; Th7: petal limb tissue from stage 3-4 flowers of Th7.

Figure 6(B) is a representative autoradiograph from the RFLP analysis of the V23xR51 (V/R) F2 plants. XhaI digested genomic DNA was probed with the 3' region of pCGP174. The V23 fragment that hybridized strongly to the probe was detected in all F2 plants that had 3',5'-hydroxylase activity in tube tissue of the flowers (+). RFLP designation for the strongly hybridizing bands (RFLP#1) has been indicated for the various plants. V: V23-like RFLP, R:
R51-like RFLP, H: heterozygotic (VR).

Figure 7(A) is an autoradiograph of a RNA blot probed with the 3' region of the pCGP175 cDNA insert. Each lane contained a 20 pg sample of total RNA
isolated from the following - 1-5: OGB limb tissue of flowers at the five (1-5) different stages of flower development described in the Materials and Methods; T: OGB tube tissue from stage 3-4 flowers; L: leaf tissue from 6 week old OGB seedlings; IL: glucose/high light treated leaf tissue from 6 week old OGB seedlings; V23: V23 limb tissue from stage 3-4 flowers; R51: R51 corolla tissue from stage 3-4 flowers; VR: petal limb tissue from stage 3-4 flowers of the V23xR51 F1 hybrid; Sw63: petal limb tissue from stage 3-4 flowers of Sw63; Th7: petal limb tissue from stage 3-4 flowers of Th7.
Figure 7(B) is a representative autoradiograph from the RFLP analysis of the V23xR51 (V/R) F2 plants. XhaI digested genomic DNA was probed with the 3' region of pCGP175. The RFLP designation obtained using the pCGP175 probe was identical to the go- designation assigned using the chi-A probe. V:
V23-like RFLP; R: R51-like RFLP; H: heterozygotic (VR) RFLP.

Figure 8 shows a diagrammatic representation of a restriction enzyme map of pCGP602. Partial lengths of the cDNA insert are indicated by the bolder lines with solid ends (as opposed to arrows). These were subcloned into M13-mp18 and mp19 and sequenced using oligonucleotide primer sequences, as indicated, to obtain overlapping sequence information. The extent and direction of sequence information obtained from each subcloned piece is shown by lines with half arrow-heads. S1 = primer sequence 1; S2 = primer sequence 2; S3 = primer sequence 3; ATG indicates the methionine initiation codon and the total length of the clone in base pairs is also indicated.

Figures 9(A) to (D) are the nucleotide sequences and predicted amino acid sequences for the cDNA inserts from pCGP176 and pCGP602. The insert from pCGP602 includes the entire sequence shown. The 5' end of the pCGP176 insert is indicated with an arrowhead.

Figures 10(A) to (C) represent the nucleotide sequence and predicted amino acid sequence for the cDNA insert from pCGP175.

Figure 11 is a diagrammatic representation of the construction of pCGP618.
pCGP618 was constructed by cloning the pCGP175 cDNA insert in a sense orientation behind the yeast glyceraldehyde-3-phosphate dehydrogenase promoter (PGPD) in the expression vector pYGA22m. The cDNA insert from pCGP175 was ligated as an EmRI-KpnI fragment with the large fragment that resulted from an EsaRI/Kpn.I digest of pYGA22m. E = EmRI, H=HiadIIl, K = KPnI, X = XhQI, IR = Inverted repeat of 2pn plasmid, Trpl = TSp1 gene, Ap = Ampicillin resistance marker.

~..~ r ~,~

Figure 12(A) shows a 3',5'-hydroxylase assay of yeast extracts using 3H-naringenin as substrate. The autoradiographs show conversion of 3H-naringenin to the 3',5'-hydroxylated derivative pentahydroxyflavanone by extracts of yeast transformed with the plasmid pCGP618 (1 and 2). No 3',5'-hydroxylase activity was detected in untransformed yeast (C). Conversion of naringenin to pentahydroxyflavanone by OGB 3',5'-hydroxylase is also shown (OGB C).
Figure 12(B) shows a 3',5'-hydroxylase assay of yeast extracts using 3H-dihydroquercetin as substrate. The autoradiographs show conversion of 3H-dihydroquercetin (DHQ) to 3H-dihydromyricetin (DHM) by extracts of yeast transformed with the plasmid pCGP618 (1 and 2). No 3',5'-hydroxylase activity was detected in untransformed yeast (C). Conversion of DHQ to DHM.
by OGB 3',5'-hydroxylase is also shown (OGB C).
Figure 13 shows a 3',5'-hydroxylase assay of yeast extracts using 3H-naringenin as substrate. The autoradiograph shows conversion of 3H-naringenin to the 3',5'-hydroxylated derivative pentahydroxyflavanone by extracts of yeast transformed with plasmids pCGP618 and pCGP620 (1 and 2, respectively). The reaction products obtained from the pCGP620 extract also included the 3'-hydroxylated eriodictyol as well as some of the original naringenin substrate indicating incomplete conversion to the 3',5'-hydroxylated end product. No 3',5'-hydroxylase activity was detected in untransformed yeast (C).

Figure 14 is a diagrammatic representation of the plasmid pCGP90. The cDNA
insert from pCGP602 was cloned in a sense orientation behind the Mac promoter of the expression vector pCGP293 as illustrated.

1.2 3 73 -20-Figure 15 shows a 3',5'-hydroxylase assay of petunia petal extracts. The autoradiograph shows the presence of low levels of 3',5'-hydroxylase activity (conversion of 3H-naringenin to 3H-pentahydroxyflavanone) in petal limb tissue (L) of Skr4 x Sw63. Significantly higher levels of activity were detected in the limb tissue (L) of two Skr4 x Sw63/pCGP90 transgenics (T/G 1602 and T/G
1603). No 3',5'-hydroxylase activity was detected in extracts of the petal tube (T) of either the non-transgenic Skr4 x Sw63 hybrid or the two pCGP90 transgenics. Conversion of naringenin to pentahydroxyflavanone by extracts of limb (L) and tube (T) petal tissue of OGB is also shown.

Figure 16 is a photographic representation of an autoradiograph of an RNA
blot probed with 32P-labelled Hf1 cDNA. Each lane contained a 20ug sample of total RNA isolated from (1) P. hybrida cv. OGB petals; (2) pansy petals;
(3) potato stems; (4) eggplant skins; (5) Nicotiana alata flowers; (6) Ageratum flowers. The probe used for A and B was derived from a 660 bp B,L11 DNA fragment; a 1.4 kb EmRI/HindIII fragment was used for C.
Washing conditions used were: (A) 6 x SSC at 50 C; (B) 2 x SSC at 50 C;
(C) 0.2 x SSC at 65 C.
Figure 17 is a photographic representation of autoradiographs of Southern blots probed with 32P-labelled Hfl cDNA. Each lane contained lOpg of DNA
digested with EsaRI. The DNA samples were isolated from (1) eggplant, (2) dutch iris, (3) potato, (4) violets and (5) anemone. Washing conditions were:
(A) 6 x SSC at 50 C; (B) 2 x SSC at 65 C.

EXAMPLE
1,. MATERIALS AND METHODS
Chemicals Enzymes and Radioisotopes Eriodictyol and dihydroquercetin were obtained from Carl Roth KG and naringenin was obtained from Sigma. Dihydromyricetin was chemically synthesized from myricetin (Extra Synthese, France) by the method of Vercruysse et al. (1985). [3H]-naringenin (5.7 Ci/mmole) and [3H]-dihydroquercetin (12.4 Ci/mmole) were obtained from Amersham. All enzymes were obtained from commercial sources and used according to the manufacturer's recommendations.
Bacterial Strains The Escherichia coli strains used were;
DH5a 242E44, A(LacZYA-Arg,F)U169, c~801acz AM15, hsdR17 (rk-, mk +), rtcAl, endAl,_gyrA96, thi 1, re.1A1, de-QR. (Hanahan, 1983 and BRL, 1986).

XLl-Blue aupF-44, hsSfRl7 (rk-, mk+), re-cAl, =dA1, gyrA96, thi 1, rdA1, 1aC,[F'proAB, 1aCI9, 1icZ AM15, Tn10(tet`)] (Bullock et al,1987).
PLK-F' resA, hadR17 (rk-, mk +), m.crA-, m.~rB , l.a!;~, supE44, gWK2, ga1T22, m~~Bl, [F' prQAB, 1aCIq, 1jr.ZAM15, Tn10(tet`)] (Stratagene).

The disarmed Agrobacterium tumefaciens strain AGLO (Lazo et al, 1991) was obtained from R. Ludwig (Department of Biology, University of California, Santa Cruz).

The cloning vectors pBluescript and pBluescribe were obtained from Stratagene.

{
WO 93/01290 PC'T/AU92/00334 2~.12373 -22-Transformation of E. cdi and A. aunefaciens Transformation of the E. coli strain DH5 a cells was performed according to the method of Inoue et al. (1990).

The plasmid pCGP90 (Fig. 14) was introduced into the Agrobacterium tumefaciens strain AGLO by adding 5pg of plasmid DNA to 100 pL of competent AGLO cells prepared by inoculating a 50mL MG/L (Garfinkel and Nester, 1980) culture and growing for 16 hrs with shaking at 28 C. The cells were then pelleted and resuspended in 0.5 mL of 85% (v/v) 100 mM CaC12/15% (v/v) glycerol. The DNA-Agrobacterium mixture was frozen by incubation in liquid N2 for 2 minutes and then allowed to thaw by incubation at 37 C for 5 minutes. The DNA/bacterial mix was then placed on ice for a further 10 minutes. The cells were then mixed with 1 mL of MG/L media and incubated with shaking for 16 hours at 28 C. Cells of A. tumefacienscarrying pCGP90 were selected on MG/L
agar plates containing 100 pg/mL gentamycin. The presence of pCGP90 was confirmed by Southern analysis of DNA isolated from the gentamycin resistant transformants.

The Petunia hybrida varieties used are shown in Table 2.

WO 93/01290 2112373. PCT/AU92/00334 Plant Material Plant variety Properties Source/Reference Old Glory Blue(OGB) F, Hybrid Ball Seed, USA
V30 Anl, An2, An3, An4, An6 Koes et al. (1986) An8, An9, An10, Anll, Phl Ph2, Ph3, Ph4, Ph5, Hfl, Hf2, Htl, Ht2, Rt, Mtl, Mt2, mfl, po, Gf V23 Anl, An2, An3, An4, An6 Wallroth et al. (1986) An8, An9, An10, phl, Hfl, Doodeman et al. 1984 Hf2, htl, Rt, Po, B1, Fl R51 Ani, An2, An3, an4, An6, Wallroth et al. (1986) An8, An9, AnlO, Anil, van Tunen et al. (1990) Phl, hfl, hf2, Htl, rt, Doodeman et al. (1984) po, bl, fl Sw63 Anl, An2, An3, an4, An6, I.N.R.A., Dijon, Cedex, An8, An9, AnlO, Ani l, France Phi, Ph2, Ph5, hfl, hf2 ht1, ht2, rt, po, mfl, fl, Gf Th7 Anl, An2, An3, An4, An6 An9, An10, An11, Hfl, Hf2, Htl, Ht2, Phl, Ph2, Ph5, Rt, po, mfi, mf2, Gf, fl Skr4 Anl, An2, An3, An4, An6, Anli, hfl, hf2, htl, Phl, Ph2, Ph5, rt, Po, Mfl, Mf2, fl Skr4 x Sw63 Skr4 x Sw63 F1 Hybrid WO 93/01290 PC'T/AU92/00334 TABLE 2 continued...

Plant variety Properties Source/Reference Rw14 Anl, An2, An4, Phl, ph2, I.N.R.A., Dijon, Cedex Ph5, hfl, hf2, Htl, Rt, Po, France B1, Lgl, Lul, Vsl, Vs3, Vs5, la, Ygl, ws, Gf, Mtl, Mf2, fl Rp57 Anl, An2, An4, Phl, ph2, Ph5, hfl, hf2, Ht1, Rt, Po, Mt, Mf, fl, Gf, Bl, Lgl, Lul, Vsl, Vs3, Vs5, Ygi, Ws.
Rp57 x Rw14 Rp57 x Rw14 F, Hybrid Plants were grown in specialised growth rooms with a 14 hr day length at a light intensity of 10,000 lux and a temperature of 22 to 26 C. OGB flowers were harvested at developmental stages defined as follows:

Stage 1: Unpigmented, closed bud (< 25 mm in length) Stage 2: Pigmented, closed bud (25-35 mm in length).
Stage 3: Dark purple bud with emerging corolla (>35 mm in length).
Stage 4: Dark purple opened flower pre-anther dehiscence ( > 50 mm in length).
Stage 5: Fully opened flower with all anthers dehisced.

Flowers of the other varieties, as described in Table 2, were harvested prior to anther dehiscence at the stage of maximum pigment accumulation.

_~~12373 Preparation of Plant Extracts for Assay of 3',5'-Hydroxylase Activity Plant tissue was homogenised in a 2 to 5 times volume of ice-cold extraction buffer (100 mM potassium phosphate (pH 7.5), 1 mM EDTA, 0.25 M sucrose, 0.25 M mannitol, 0.1% (w/v) BSA, 100 nM Pepstatin, 100 nM Leupeptin, 0.1 ~ mg/mL PMSF, 20 mM 2-mercaptoethanol and 10 mg/mL polyclar AT). The homogenate was centrifuged at 10,000 rpm in a JA20 rotor (Beckman) for 10 min at 4 C and an aliquot of the supernatant was assayed for 3',5'-hydroxylase activity.
3 ' ,5 ' -Hydroxylase Assay 3',5'-Hydroxylase enzyme activity was measured using a modified version of the method described by Stotz and Forkmann (1982). The assay reaction mixture typically contained 100 pL of plant extract, 5pL of 50 mM NADPH in assay buffer (100 mM potassium phosphate (pH8.0), 1 mM EDTA and 20 mM
2-mercaptoethanol), 10 pCi of [3H] naringenin or 5 pCi of [3H]-dihydroquercetin and was made up to a final volume of 210 pL with the assay buffer. Following incubation at 23 C for 2-16 hours, the reaction mixture was extracted with 0.5 mL of ethylacetate. The ethyl acetate phase was dried under vacuum and then resuspended in 10 ~1L of ethyl acetate. The tritiated flavonoid molecules were separated on cellulose thin layer plates (Merck Art 5577, Germany) using a chloroform: acetic acid: water (10:9:1, v/v) solvent system. At the completion of the chromatography, the TLC plates were sprayed with 7% (v/v) 2,5-diphenyloxazol in diethyl ether. The reaction products were localised by autoradiography and identified by comparison to non-radioactive naringenin, eriodictyol, dihydroquercetin and dihydromyricetin standards which were run alongside the reaction products and visualized under UV light.

WO 93/01290 PCr/AU92/00334 Glucose/High Light Induction of Delphinidin Synthesis in Leaves Leaves were harvested from P. hybrida cv. OGB and cut into 1cm2 sections in sterile water. The leaf sections were then floated on a 2% (w/v) glucose solution and exposed to a light intensity of 24,000 lux for 96 hours.
Construction of cDNA Library #1 Twenty grams of stage 3 to 4 OGB flower limbs were homogenised in 100 mL
of PEB (200 mM Tris-HCl (pH 8.6), 60 mM KCI, 30 mM MgC12, 25 mM
EGTA) containing 10 mM vanadyl ribonucleoside complex. Cell debris was removed by filtering the homogenate through sterile Miracloth (Calbiochem).
The filtrate was layered on top of a step gradient of 6 mL of PEB containing 25% (w/v) sucrose, 250 units InhibitAce (5-Prime 3-Prime), and 6 mL of PEB
containing 50% (w/v) sucrose and 250 units InhibitAce in Ultra-ClearTm Quick-Sea1TM (Beckman) centrifuge tubes. The tubes were centrifuged for 3.5 hours at 26,000 rpm in a 70Ti rotor. Membrane-bound polysomes were collected from the 25% (w/v) sucrose /50% (w/v) sucrose interface and added to a 4 M guanidium isothiocyanate solution. RNA was isolated from the denatured polysomes by pelleting through a 5.7 M CsCI cushion, as described by Turpen and Griffith (1986).

A Uni-ZAPTM XR vector kit (Stratagene) was used to construct a directional cDNA library in ,IZAP using 25 pg of the polysomal RNA as template. The primary library, which contained 250,000 plaque forming units (pfu), was amplified by overnight growth on NZY plates (Sambrook et al, 1989) and the amplified phage stock was eluted in PSB (100 mM NaCI, 8 mM MgSO41 50 mM Tris-HCl (pH 7.5), 0.01 % (w/v) gelatin) as described by Sambrook et al, (1989).

, 27_ Construction of cDNA Library #2 Total RNA was isolated from the petal tissue of P. hybrida cv. OGB stage 3 to 4 flowers using the method of Turpen and Griffith (1986). poly(A)+ RNA was selected from the total RNA by three cycles of oligo-dT cellulose chromatography (Aviv and Leder, 1972).

Two micrograms of poly(A)+ RNA were reverse transcribed in a 20 L volume containing 1 x SuperscriptTm reaction buffer, 10 mM dithiothreitol, 500 M
dATP, 500 pM dGTP, 500 pM dTTP, 500 pM 5-methyl-dCTP, 0.75 pg oligonucleotide #8 and 2 pL SuperscriptTm reverse transcriptase (BRL). The reaction mix was incubated at 37 C for 50 minutes, 44 C for 10 minutes, then placed on ice.

Second strand reaction mix (140 L) was added to the first strand reaction.
The second strand reaction mix consisted of 21 mM Tris-HCl, 104 mM KCI, 5.3 mM MgC121171 pM R-NAD, 11.4 mM (NH4)2SO4, 214 pM dATP, 642 M
dCTP, 214 pM dGTP, 214pM dTTP, 4 mM DTT, 10 pCi 32P-dCTP (3000 Ci/mMole), 15 units E. coli DNA ligase, 40 units DNA polymerase (Boehringer) and 0.8 units RNAse H. The final mixture was incubated for 150 minutes at 16 C. To make the double-stranded cDNA blunt-ended, 10 units T4 DNA polymerase was added, and the reaction continued for a further 15 minutes at 16 C. The reaction was stopped and the cDNA purified by phenol/chloroform extraction, followed by chloroform extraction and ethanol precipitation.

EsaRI adaptors (Promega) were ligated with the cDNA and then kinased using conditions recommended by the manufacturer. The enzymes were denatured by heat (70 C, 20 minutes) and the DNA was purified by phenol/chloroform extraction and ethanol precipitation. The cDNA was digested with 50 units Xh,QI (Boehringer) in a reaction volume of 100 pL, using conditions recommended by the manufacturer. The enzyme was heat killed 2'JIZ373 -28-(70 C, 20 minutes) and the mixture passed through an S400 spun column (Pharmacia) which had been equilibrated in STE buffer (Sambrook et al, 1989). The eluate was phenol/chloroform extracted and ethanol precipitated.
After microcentrifugation at 4 C for 30 minutes the cDNA pellet was rinsed with 70% (v/v) ethanol, air dried and resuspended in 10 pL of TE buffer (10 mM Tris-HCI (pH7.5), 1 mM EDTA).

NA-45 membrane (Schleicher and Schuell) was used to isolate cDNA in the size range of 1.3 to 2.5kb from a 7.5 L sample that had been electrophoresed through a 1% (w/v) agarose gel.

The size fractionated cDNA was ligated with 1 pg XZAPII EcaRI/XhQ.I/ CIAP
treated vector (Stratagene) in 5}iI. reaction buffer consisting of 50 mM Tris-HC1 (pH 7.0), 10 mM MgC12, 10 mM dithiothreitol, 1 mM ATP and 2 units T4 DNA ligase. The reaction was performed at 4 C for 2 days.

After leaving at room temperature for two hours, the ligation reaction mixture was packaged using the Packagene system (Promega). The total number of recombinants was 270,000 pfu.
An amount of 150,000 pfu of the packaged cDNA was plated at 10,000 pfu per 15 cm diameter plate after transfecting PLK-F' cells. The plates were incubated at 37 C for eight hours, then stored overnight at 4 C. Duplicate lifts were taken onto Colony/Plaque ScreenT"^ filters (Dupont) and treated as recommended by the manufacturer.
Synthesis of Oligonucleotides Oligonucleotides were synthesized on an Applied Biosystems PCR-Mate DNA
synthesizer using methods recommended by the manufacturer. The oligonucleotides synthesized were, 5'-3':
Oligo 1: GGAAGCTTATICCITT(T/C)GGIGCIGG
Oligo 2: GGATGACTCAGTAAAACGACGGCCAGT

Oligo 3:
CCIGG(A/G)CAIATIC(G/T)(C/T)(C/T)TICCIGCICC(A/G)AAIGG
Oligo 4: GGATGACTCAAACAGCTATGACCATG
Oligo 5: GTTCAATTCGGAATGATG
Oligo 6: GCTGCACTTAATCCATAT
Oligo 7: TGCATAGCTT'I'TGGG
Oligo 8:
GAGAGAGAGAGAGAGAGAGATCTCGAG
Oligo 9: ATGTCTCCTCCAGTG
Oligo 10: CTAGACTCCAATCAC

Oligos 2 and 4 included a GCN4 binding site (indicated by underlining) which has been shown to facilitate the enrichment of double stranded PCR products (Lew and Kemp, 1989).

The basis for the design of oligo 3 was as follows: Amino acid sequences from the putative haem-binding domain of an avocado cytochrome P450 (Bozak et al, 1990) and the corresponding sequences encoded by the two petunia cytochrome P450 homologues pCGP142 and pCGP147 were aligned:
avocado P F G A G R R G C P G
pCGP142 P F G A G K R I C P G
pCGP147 P F G S G R R I C P G
The consensus amino acid sequence of the haem-binding region for the three plant cytochromes P450 could thereby be seen to be:
P F G A(S) G R(K) R I(G) C P G

1 4.2~

Possible permutations of nucleotide sequence that could encode the amino acids found in the haem-binding domain of the three cytochrome P450 molecules could then be deduced:
51- CCX TTT GGX GCX GGX AGX CGX ATX TGT CCX GGX -3' C AG CA A GG C
T
X indicates nucleotide positions where all four nucleotides (A,C,G and T) can be used. Oligo 3 was designed to complement a subset of the consensus sequence derived from the three plant cytochromes P450. Deoxyinosine (I) was used predominantly when base degeneracy was greater than three. The resulting oligonucleotide sequence was as shown above.
PCR Reactions Helper phage R408 (Stratagene) was used to excise pBluescript phagemids containing petunia cDNA inserts from 200,000 pfu of the amplified IZAP
cDNA library #1 using methods described by the manufacturer. Escherichia coli XLl- Blue were transfected with the phagemid mixture and 250,000 colonies were plated out on ampicillin-containing media. Cells were resuspended in LB (Sambrook et al, 1989) and plasmid DNA was isolated using an alkaline lysis procedure (Sambrook et al, 1989). Plasmid DNA was further purified by banding on a CsCI gradient. This DNA was used as the template for PCR.

PCR reactions for amplification of petal cytochrome P450 homologues contained 5 to lOOng of excised DNA, 10 mM Tris-HCI (pH 8.3), 50 mM KCI, 1.5 mM MgC12, 0.01% (w/v) gelatin, 0.2 mM each dNTP, 0.4 pM each primer and 1.25 units Taq polymerase (Cetus). Reaction mixes (50 L) were cycled 30 times between 94 C, 48 C and 72 C for 1 minute at each temperature. The amplified products were gel-purified using Geneclean (Bio 101 Inc.), reamplified to obtain sufficient material for cloning and then end-repaired using T4 DNA polymerase. DNA amplified using oligos 1 and 2 was digested with HindI1I and XhQI prior to cloning into pBluescript. The PCR product generated by amplification between oligos 3 and 4 was cloned directly into the ddT-tailed pBluescript vector described by Holton and Graham (1991).

Screening of cDNA Libraries Duplicate plaque lifts were hybridised and washed as follows: High stringency conditions (hybridization: 50% (v/v) formamide, 6 x SSC, 1% (w/v) SDS at 42 C for 16 hrs and washing: 2 x SSC, 1% (w/v) SDS at 65 C for 2 x 15 minutes followed by 0.2 x SSC, 1% (w/v) SDS at 65 C for 2 x 15 minutes) were used to detect sibling clones and low stringency conditions (hybridisation:
20% (v/v) formamide, 6 x SSC, 1% (w/v) SDS at 42 C for 16 hrs and washing:

6 x SSC, 1%(w/v) SDS at 65 C for 1 hour) were used to detect related sequences.

Northern Analysis Total RNA was isolated from tissue that had been frozen in liquid N2 and ground to a fine powder using a mortar and pestle. An extraction buffer of 4 M guanidium isothiocyanate, 50 mM Tris-HCl (pH 8.0), 20 mM EDTA, 0.1%
(v/v) Sarkosyl, was added to the tissue and the mixture was homogenized for 1 minute using a polytron at maximum speed. The suspension was filtered through Miracloth (Calbiochem) and centrifuged in a JA20 rotor for 10 minutes at 10,000 rpm. The supernatant was collected and made to 0.2 g/ mL
CsCI (w/v). Samples were then layered over a 10 mL cushion of 5.7 M CsCI, 50 mM EDTA (pH 7.0) in 38.5 mL Quick-seal centrifuge tubes (Beckman) and centrifuged at 42,000 rpm for 12-16 hours at 23 C in a Ti-70 rotor. Pellets were resuspended in TE/SDS (10 mM Tris-HCl (pH 7.5), 1 mM EDTA, 0.1%
(w/v) SDS) and extracted with phenol:chloroform:isoamyl alcohol (25:24:1) saturated in 10 mM EDTA (pH 7.5). Following ethanol precipitation the RNA pellets were resuspended in TE/SDS.

WO 93/01290 PCr/AU92/00334 ~~ ~ ~3'73 -32-RNA samples were electrophoresed through 2.2 M formaldehyde/1.2% (w/v) agarose gels using running buffer containing 40 mM
morpholinopropanesulphonic acid (pH 7.0), 5 mM sodium acetate, 0.1 mM
EDTA (pH 8.0). The RNA was transferred to Hybond-N filters (Amersham) as described by the manufacturer and probed with 32P-labelled cDNA fragment (108 cpm/pg, 2 x 106 cpm/mL). Prehybridization (1 hr at 42 C) and hybridization (16 hr at 42 C) was carried out in 50% (v/v) formamide, 1 M
NaC1, 1% (w/v) SDS, 10% (w/v) dextran sulphate. Degraded salmon sperm DNA (100 pg/mL) was added with the 32P-labelled probe for the hybridization step.

Filters were washed in 2 x SSC/ 1% (w/v) SDS at 65 C for 1 to 2 hours and then 0.2 x SSC/ 1% (w/v) SDS at 65 C for 0.5 to 1 hour. Filters were exposed to Kodak XAR film with an intensifying screen at -70 C for 48 hours.

RFLP Analysis a. Isolation of Genomic DNA
DNA was isolated from leaf tissue essentially as described by Dellaporta et al, (1983). The DNA preparations were further purified by CsCI buoyant density centrifugation (Sambrook et al, 1989).

b. Southern blots The genomic DNA (10 g) was digested for 16 hours with 60 units of Xb I
and electrophoresed through a 0.7% (w/v) agarose gel in a running buffer of TAE (40 mM Tris-acetate, 50 mM EDTA). The DNA was then denatured in denaturing solution (1.5 M NaCI/0.5 M NaOH) for 1 to 1.5 hours, neutralized in 0.5 M Tris-HCl (pH 7.5)/ 1.5 M NaCI for 2 to 3 hours and the DNA was then transferred to a Hybond N (Amersham) filter in 20 x SSC.

WO 93/01290 PCr/AU92/00334 :~.~3~3 c. Isolation of chi A probe A cDNA clone of chi A(van Tunen et a1.,1988) was synthesized by PCR using cDNA template made from OGB stage 3 petal RNA and two oligonucleotide primers: #9, which covered nucleotides 6-20 and #10, which was complementary to nucleotides 711-725 of the published chi-A cDNA sequence (van Tunen et al, 1988). The resulting PCR product was ligated into the S1nal site of pBluescribe M13- (Stratagene) and sequenced to confirm that the cloned fragment corresponded to the published sequence.
32P-Labelling of DNA Probes DNA fragments (50 to 100 ng) were radioactively labelled with 50 pCi of [ a 32P]-dCTP using an oligolabelling kit (Bresatec). Unincorporated [a 32P]-dCTP*
was removed by chromatography on a Sephadex G-50 (Fine) column.
DNA Sequence Analysis DNA sequencing was performed essentially by the method of Sanger et al.
(1977) using the Sequenase enzyme (USB, version 2.1). The complete sequence of clones pCGP602, pCGP176 and pCGP175 was determined by compilation of sequence from different M13 -mpl8 and -mp19 (Norrander et al, 1983; Yanisch-Perron, 1985) subclones obtained using standard cloning procedures (Sambrook et al, 1989). For some regions it was necessary to synthesize specific oligonucleotide primers to obtain overlapping sequence data. The following six primers were synthesized for that purpose:
5' CGTGCCAATGAGCTAGG 3' primer sequence 1 5' GATGTTGGTTGTACTGAG 3' primer sequence 2 5' GGAAACCAGATTTTCTTG 3' primer sequence 3 5' TTTTTTTTTTTTTTTTT(AGC) 3' primer sequence 4 5' GTTTTCCCAGTCACGAC 3' primer -40 5. AACAGCTATGACCATG 3' reverse primer A restriction map of pCGP602 showing the position of several of these sequences may be seen in Figure 8.

Homology searches against Genbank, SWISS-PROT and EMBL databases were performed using the FASTA and TFASTA programs (Pearson and Lipman, 1988).

Construction of pCGP293 The expression binary vector pCGP293 was derived from the Ti binary vector pCGN1559 (McBride and Summerfelt, 1990). Plasmid pCGN1559 was digested with KgnI and the overhanging 3' ends were removed with T4 DNA
polymerase according to standard protocols (Sambrook et al,1989). The vector was then further digested with XbaI and the resulting 5' overhang was repaired using the Klenow fragment of DNA polymerase I. The vector was then re-ligated to give pCGP67. A 1.97 kb Pai.I fragment containing the Mac promoter, mas terminator and various cloning sites (Comai et al, 1990) was isolated from pCGP40 and inserted into the psi,1 site of pCGP67 to give pCGP293.
Plasmid pCGP40 was constructed by removing the GUS gene (Jefferson et al, 1987) as aB_=HI-SzQI fragment from pCGN7334 and replacing it with the B=HI-S_ar,I fragment from pBluescribe M13- that includes the multicloning site. Plasmid pCGN7334 (obtained from Calgene, Inc. CA, USA), was constructed by inserting the fragment containing the Mac-GUS-mas gene fusion into the XhQ1 site of pCGN7329 (Comai et al, 1990).
Construction of pCGP90 Plasmid pCGP90 was constructed by cloning the cDNA insert from pCGP602 in a sense orientation behind the Mac promoter (Comai et al, 1990) of pCGP293. The BamHI-KpaI fragment containing the cDNA insert was isolated from pCGP602 and ligated with aB-atttHI/KpnI digest of pCGP293.

WO 93/01290 PCr/AU92/00334 _35_ Correct insertion of the insert in pCGP90 was established by restriction analysis of DNA isolated from gentamycin resistant transformants.
Construction of the Yeast Expression Vector pYGA22m M13-mp18 was digested with EcaRI and BgJ.II to produce a 700 bp fragment that contained a multicloning site. This fragment was ligated with the 9 kb E-caRI-U1I fragment from pYGA2269 (Ashikari et al, 1989). The resulting construct, designated pYGA22m, contained the multicloning site inserted downstream of the yeast glyceraldehyde-3-phosphate dehydrogenase promoter (Fig. 11).

Construction of pCGP618 A 1.8 kb FmRI-Kg.u.I fragment that included the entire cDNA insert from pCGP175 was ligated with the 9 kb F_cQ.RI - Kp.nI fragment from pYGA22m.
The resulting plasmid, pCGP618, contained the pCGP175 cDNA fragment ligated in a sense orientation behind the glyceraldehyde-3-phosphate dehydrogenase promoter (Figure 11).

Construction of pCGP620 A 1.8 kb ES.QItI-L{pnI fragment that included the entire cDNA insert from pCGP176 was ligated with the 9kb EmRI-Kp.aI fragment from pYGA22m (as described for the construction of pCGP618). The resulting plasmid, pCGP620, contained the pCGP176 cDNA fragment ligated in a sense orientation behind the yeast glyceraldehyde-3-phosphate dehydrogenase promoter.
Yeast Transformation The yeast strain G-1315 (MU a, trpl) (Ashikari et al, 1989) was transformed with pCGP618 and pCGP620 according to Ito et al, (1983). The transformants were selected by their ability to restore G-1315 to tryptophan prototrophy.

~~~~373 -36-Preparation of Yeast Extracts for Assay of 3',5'-Hydroxylase Activity a. G-1315/pCGP618 Single isolates of G-1315/pCGP618 and a G-1315 revertant that grew on media lacking tryptophan were used to inoculate 50 mL of YNBC [1.2% (w/v) yeast nitrogen base without amino acids (Difco), 2% (w/v) glucose and 0.3%
(w/v) Casamino acid (Difco)] and incubated with shaking for 2 days at 30 C.
Cells were pelleted by centrifugation and a microsomal fraction was obtained according to Oeda et al. (1985) except that the spheroplasts were disrupted in the extraction buffer used for the assay of 3',5'-hydroxylase activity in plant tissue. Microsomal pellets were suspended in 400 uL. of buffer A (10 mM
Tris-HCl (pH 7.5), 0.65 M sorbitol, 0.1 mM DTT, 0.1 mM EDTA) and a 100 pL sample was assayed for 3',5'-hydroxylase activity.

b. G-1315/pCGP620 A single isolate of G-1315/pCGP620 was used to inoculate 20 ml of YNBC
which was subsequently incubated for 2 days at 30 C. Cells were collected by centrifugation, washed once with TE, once with buffer A, and then resuspended in buffer B (10 mM Tris-HCI (pH7.5), 1.2 M sorbitol, 0.1 mM
DTT, 0.1 mM EDTA) containing zymolyase (0.1 mg/mL) (Seikagakukogyo, Japan). Following incubation for 1 hour at 30 C the cells were pelleted by centrifugation and resuspended in 400 l of buffer A. The cell suspension was then vortexed with glass beads (diameter = 0.4 mm) for 2 minutes and a 100 pl sample was assayed for 3',5'-hydroxylase activity.
Pelunia Transformation a. Plant Material Petunia hybrida (Skr4 x Sw63 and Rp57 x Rw14) seeds were sterilized in 1.25%
(w/v) sodium hypochlorite for 10 min and rinsed three times in sterile water.
Sterilized seeds were soaked in 100 mg/L gibberellic acid (GA3) solution for 16 to 20 hours. They were then germinated for 2 weeks on 10% (w/v) MS
(Murashige and Skoog, 1962) medium supplemented with 1% (v/v) sucrose and 0.8% (w/v) Difco Bacto agar. Young seedlings were transferred to MS
medium supplemented with 3% (w/v) sucrose for 3 weeks before being transferred to Jiffy peat pellets (Jiffy Products Ltd, Norway), kept under mist and illuminated (135 pE. mercury halide light, 22 C) for 2 to 3 weeks. These young plants were then transferred to a growth cabinet (68 }i.E. cool white fluorescent light, 25 C). For co-cultivation, young leaves were harvested and sterilized in 1.35% (w/v) sodium hypochlorite for 2 min followed by rinsing three times in sterile water. Leaf tissue was then cut into 25 mm2 squares and precultured on MS media supplemented with 0.05 mg/L kinetin and 1.0 mg/L
2,4-dichlorophenoxyacetic acid (2,4-D) for 24 hours.

b. Co-cultivation of A.grobacterium and Petunia Tissue Agrobacterium tumefaciens strain AGLO (Lazo et al, 1991) containing the binary vector pCGP90 (Fig. 14) was maintained at 4 C on MG/L (Garfinkel and Nester, 1980) agar plates with 100mg/L gentamycin. A single colony was grown overnight in liquid medium containing 1% (w/v) Bacto-peptone, 0.5%
(w/v) Bacto-yeast extract and 1% (w/v) NaCI. A final concentration of 5 x 108 cells/mL was prepared the next day by dilution in liquid MS medium containing 3% (w/v) sucrose (BPM). Leaf discs were dipped for 5 minutes into BPM containing AGLO/pCGP90. The leaf discs were then blotted dry and placed on co-cultivation media for 4 days. The co-cultivation medium consisted of SH medium (Schenk and Hildebrandt, 1972) supplemented with 0.05 mg/L kinetin and 1.0 mg/L 2,4-D and included a feeder layer of tobacco cell suspension spread over the co-cultivation medium with a filter paper placed on top of the tobacco cell suspension.
c. Recovery of transgenic petunia plants After co-cultivation, the leaf discs were transferred to the following selection media: Skr4 x Sw63 discs to fresh MS medium supplemented with 3% (w/v) sucrose, 2 mg/L a benzylaminopurine (BAP), 100 mg/L kanamycin, 350 mg/L
cefotaxime, 0.3% (w/v) Gelrite Gellan Gum (Schweizerhall); Rp57 x Rw14 discs to the same medium, containing 0.5 mg/L BAP and a-naphthalene acetic acid (NAA) instead of 2 mg/L BAP. After 3 weeks, regenerating explants were transferred to fresh medium. Adventitious shoots which survived the kanamycin selection were isolated and transferred to BPM containing 100 mg/L kanamycin and 350 mg/L cefotaxime for root induction. All cultures were maintained under a 16 hr photoperiod (60 E. cool white fluorescent light) at 23 2 C. When roots reached 2-3 cm in length the transgenic petunia plantlets were transferred to autoclaved Debco 51410/2 potting mix in 8 cm tubes. After 4 weeks plants were replanted into 15 cm pots using the same potting mix and maintained at 23 C under a 14 hour photoperiod (300 pE.
mercury halide light).
Tobacco Transformation a. Plant Material Nicotiana tabacum (cv. Xanthi) stock plants were maintained on MS medium supplemented with 1 mg/L indolebutyric acid (IBA) and solidified with 0.25%
(w/v) Gelrite. Leaf tissue was cut into 25 mm2 squares and placed onto MS
medium containing 1 mg/L BAP and 0.5 mg/L indoleacetic acid (IAA) for 24 hours.

b. Co-cultivation of Agrobacterium and tobacco tissue Co-cultivation was carried out as previously described for petunia.
c. Recovery of transgenic tobacco plants After co-cultivation, leaf discs were transferred to MS medium supplemented with 1 mg/L BAP, 0.5 mg/L IAA, 100 mg/L kanamycin and 350 mg/L
cefotaxime (selection medium). Regenerating explants were transferred to fresh selection medium after 2-3 weeks. Adventitious shoots which survived the kanamycin selection were isolated and transferred to MS medium containing 1 mg/L IBA, 100 mg/L kanamycin and 350 mg/L cefotaxime for root induction. When roots reached 2-3 .cm in length the transgenic tobacco plantlets were transplanted to soil as described for petunia.

WO 93/01290 2112373 PCr/AU92/00334 Anthocyanidin analysis Prior to HPLC analysis the anthocyanin molecules present in petal extracts were acid hydrolysed to remove glycosyl moieties from the anthocyanidin core.
The hydroxylation pattern on the B ring of the anthocyanin pigments was determined by HPLC analysis of the anthocyanidin core molecule. The HPLC
system used in this analysis was a Hewlett-Packard 1050 equipped with a multiwavelength detector (MWD). Reversed phase chromatographic separations were performed on a Spherisorb S5 ODS2 cartridge column, 250 mm x 4 mm ID.
a. Extraction of anthocyanins and flavonoids Flower pigments were extracted from petal segments (ca. 50 mg) with 5 ml of methanol containing 1% (v/v) of aqueous 6M hydrochloric acid. Extracts were diluted with water (1:9) and filtered (Millex HV, 0.45p) prior to injection into the HPLC system.

b. Hydrolysis of anthocyanins Crude methanolic extracts (100 pL) obtained in a. above were evaporated to dryness in Pierce Reacti-Vials using a stream of dry nitrogen at room temperature. The residues were dissolved in 200pL 2M HCI, vials were capped and then heated at 100 C for thirty minutes. Hydrolysis mixtures were diluted with water (1:9) and filtered (Millex HV, 0.45p) prior to HPLC
analysis.

c. Chromatography Separation of flower pigments was effected via gradient elution using the following system:
Solvent A: (triethylamine: conc. H3PO4 : H20) (3:2.5:1000) Solvent B: acetonitrile Gradient Conditions: 5% B to 40% B over 20 minutes Flow Rate: 1 ml/min Temperature: 35 C

~1,1.~~. 373 -40-Detection: MWD with simultaneous data acquisition at 280, 350 and 546nm.

The anthocyanidin peaks were identified by reference to known standards.
2. CLONING AND ANALYSIS OF 3',5'-HYDROXYLASE
Characterisation of the 3',5'-Hydroxylase Enzyme a. Developmental Regulation Extracts of P. hybrida cv. OGB petals harvested from flowers at the different stages of development defined above were assayed for 3',5'-hydroxylase activity.

3',5'-Hydroxylase enzyme activity in OGB petals was found to be developmentally regulated during maturation of the corolla (Figure 2B). This developmental profile paralleled the expression of other genes involved in flavonoid biosynthesis. Activity of the 3',5'-hydroxylase enzyme and expression of chalcone synthase (CHS), chalcone flavanone isomerase (CHI), dihydroflavonol reductase (DFR) genes peaked around stages 3 to 4 of flower development.

b. Induction of 3',5'-Hydroxylase Activity in Leaf Tissue Genes of the flavonoid pigment biosynthetic pathway are not normally expressed in leaf tissue. However, synthesis of delphinidin pigments was induced in OGB leaves by incubation in a 2% (w/v) glucose solution in high light. Under these conditions, 3',5'-hydroxylase enzyme activity can be detected in OGB leaf tissue. Maximal induction of enzyme activity was shown to occur after 96 hours of the glucose/high light treatment. Under these conditions the expression of several other pigment biosynthesis genes was also induced to levels comparable to those observed in emerging petals. It was concluded from these results that the Hfl and/or Hf2 genes are induced in glucose/high light treated leaf tissue.

WO 93/01290 ~ ~ 3 PCT/AU92/00334 c. Evidence that the 3',5'-Hydroxylase Belongs to the Cytochrome P450 Class of Enzymes 3',5'-Hydroxylase activity in OGB petals was shown to be associated with the microsomal fraction and dependent upon the presence of NADPH. Activity could be inhibited by treatment of the microsomes with carbon monoxide and by two inhibitors that specifically inactivate cytochrome P450 enzymes:
tetcyclasis and 1-aminobenzotriazine (Taton et al, 1988; Matthews et al, 1985;
Rademacher et al, 1987).

Construction of a cDNA Library Enriched for Cytochrome P450 Sequences Translation of cytochrome P450 mRNAs occurs on membrane-bound polysomes (Takemori and Kominami, 1989). Therefore, in order to enrich for cytochrome P450 sequences (including 3',5'-hydroxylase sequences) a cDNA
library was constructed using membrane-bound polysomal RNA isolated from OGB petals of stage 3 to 4 flowers. Isolation of the petal RNA from stage 3 to 4 flowers, ensured that 3',5'-hydroxylase sequences were maximally represented in the library since 3',5'-hydroxylase activity had been shown to be maximal at this stage of development (see above and Figure 2B). The resultant library, designated cDNA library #1, contained 250,000 primary recombinants.

PCR Amplification of a Petunia Petal Cytochrome P450 cDNA
A large number of cytochromes P450 have been sequenced, from organisms as diverse as vertebrates, fungi, insects, bacteria and one plant (Nebert et al, 1991, Bozak et al, 1990). A characteristic of all these enzymes is the existence of a number of small regions of sequence conservation, especially around the cysteine residue involved in haem binding. The amino acid sequence F(G,S)XGXRXCXG is present in the haem-binding domain of nearly all microsomal cytochromes P450 sequenced to date, where X can be any amino acid (Figure 3). This consensus sequence was compared with the NBRF
protein database, using the FASTA program (Pearson and Lipman, 1988), to determine the frequency of occurrence of amino acids around this area for all WO 93/01290 ~ ~ ~ ~ ,373 PCT/AU92/00334 of the microsomal cytochrome P450 sequences in the database. This analysis showed that the most common amino acid sequence for each position around the haem-binding domain was:
FMPFGAGXRXCLG
An oligonucleotide was designed to hybridize to a gene encoding the underlined sequence and similar sequences. This oligonucleotide, designated oligo 1, is shown below:
5'-GGAAGCTTATICCITT(T/ C)GGIGCIGG-3' The underlined portion is additional sequence which includes a HitldlIl recognition site to facilitate directional cloning of PCR products. The inclusion of deoxyinosine (I) covered the different possibilities for codon usage where more than two codons could encode the same amino acid. Deoxyinosine base-pairs with similar efficiency to A, T, G and C (Martin et al, 1985;
Ohtsuka et al, 1985).

Plasmid DNA obtained from cDNA library #1 as described in the Materials and Methods was used as a template for the amplification of a 360 bp cytochrome P450 related sequence using oligos 1 and 2 (Figure 3). Oligo 2 corresponded to the -20 primer (Stratagene) plus a GCN4 binding site (Lew and Kemp, 1989) at the 5' end. The PCR fragment was cloned into pBluescript and the resulting plasmid was designated pCGP450. The 5' region of pCGP450 encoded a polypeptide sequence with significant homology to previously sequenced cytochrome P450 molecules.

Isolation of Cytochrome P450 Homologues from a Petunia Petal cDNA Library Plasmid pCGP450 was used to screen cDNA library #1 (60,000 plaques) for related clones. Two consecutive hybridizations under conditions of high and low stringency were used to detect both sibling clones of pCGP450 and a second group of cytochrome P450 cDNAs. A representative cDNA clone of each of the sibling groups was selected for subsequent analyses. The pCGP450 WO 93/01290 PCr/AU92/00334 ~112373 sibling was designated pCGP142 and the representative of the second group was designated pCGP147. A Szil - Ej2Q.RI fragment that included only the coding sequences of pCGP147 was then used to re-probe 16,000 plaques from cDNA library #1 at low stringency. A total of 20 clones that hybridized to the probe were sequenced, allowing two further cytochrome P450 homologues to be identified: pCGP158 and pCGP160 (Figure 4A).

Isolation of an Additional Petal Cytochrome P450 Homologue by PCR
Sequence information from around the putative haem-binding domain of the petunia clones pCGP142, pCGP147 and a previously sequenced avocado cytochrome P450 sequence (O'Keefe and Leto, 1989; Bozak et al, 1990) was used, as described in the Materials and Methods, to design a second degenerate oligonucleotide (oligo 3) which covered amino acid sequences encoded by at least two of the three cytochrome P450 clones. This oligonucleotide was used to amplify related sequences by PCR using cDNA
library #1 as the template and oligo 4 as the second primer (Fig. 3B).
Reaction products in the size range 250-500 bp were isolated as described in Materials and Methods and cloned into the ddT-tailed pBluescript vector described by Holton and Graham (1991). The cloned PCR fragments were sequenced and shown to encode a fifth cytochrome P450 homologue. One clone, designated pCGP454, was chosen for further analysis.

Isolation of Further Cytochrome P450 Homologues from cDNA Library #1 A mixed probe of 32P-labelled DNA fragments that included the coding regions of the cytochrome P450 homologues pCGP142, pCGP147, pCGP158 and pCGP160 and the cDNA insert from pCGP454 (Figures 4B to 4H) was used to screen 50,000 recombinants from cDNA library #1 for related sequences. A
total of 152 hybridizing clones were detected under low stringency hybridization and washing conditions. A further 13 different cytochrome P450 homologues were identified by sequence analysis of DNA isolated from the hybridizing clones. Two closely related sibling groups were distinguished amongst these clones. The coding regions of each of the two groups showed 94% homology or similarity at the DNA level. Two representatives of one sibling group, pCGP174 (Figure 5A) and pCGP176, and one representative of the other sibling group, pCGP175 (Figure 5B), were chosen for further study.
Northern and RFLP Analysis of the Cytochrome P450 Homologues Northern and RFLP analyses were used to distinguish which cytochrome P450 homologues had the molecular characteristics of a cDNA encoding a 3',5'-hydroxylase. There are two genetic loci in P. hybrida, Hfi and Hf2, that control 3',5'-hydroxylase activity (de Vlaming et al, 1984; Wiering, 1974). Hfl is expressed in both the limb and the tube of P. hybrida flowers and gives rise to much higher levels of 3',5'-hydroxylase activity than does Hf2 which is only expressed in the limb. Petunia 3',5'-hydroxylase activity is also developmentally and spatially regulated. Under normal growth conditions, the enzyme can only be detected in petal tissues, increasing to maximal levels around stages 3-4 of flower development and then declining in the fully open flower (stage 5; see Figure 2(B)). Activity can also be induced in leaf tissue under certain stress conditions such as the glucose/high light treatment described above. Accordingly, a cDNA clone encoding a 3',5'-hydroxylase was expected to have an expression profile on RNA blots that paralleled the enzyme activity profile. It was also expected that a cDNA clone encoding a P.
hybrida 3',5'-hydroxylase would map to either the Hf 1 or Hf2 loci. Hfl has been mapped to chromosome I of the P. hybrida genome and is linked to the phl locus (Cornu, 1984; Cornu et al, 1990) while Hf2 is closely linked to Pst on chromosome V(Wallroth et al, 1986). RFLP analysis of DNA isolated from a F2 population of plants derived from a cross between the inbred lines V23 (Hfl /I3f1, 1~If2/hIf2) and R51 (hfl /hfl, hf2/hf2) was used to obtain linkage data for the various cytochrome P450 homologues. An 1=if1 /- genotype was assigned to F2 plants that had 3',5'-hydroxylase activity in the flower tube.
In addition, it was possible to assign an Hf.l /Hfl genotype to plants in the population based on linkage to the phl gene which influences the pH of the petal vacuole (Wiering and de Vlaming, 1984).

The V23 parent line (Hfl/Hfi) also had a phl/phl genotype which results in a petal homogenate pH of approximately 6.2. Since Phl /- plants have a petal homogenate pH of 5.3 it was possible to distinguish ph.l/phi (Hfi/I3.f1) plants within the R51xV23 F2 population by assaying the pH of petal homogenates.
The linkage between the Hf2 and PQ loci was used to distinguish candidate Hf2 clones. The PQ locus has been shown to correspond to the P. hybrida chi A gene which encodes the enzyme chalcone flavanone isomerase (van Tunen et al, 1991). A cDNA clone of chiA could therefore be used in the RFLP
analysis to assign a PQ or pQ genotype to individuals in the F2 population.
Since V23 has an Hf2/Hf2, p4/pQ. genotype it was possible to determine linkage to the Hf2 locus by co-segregation of the V23-like and R51-like RFLP
patterns with the pQ and PQ patterns detected by the chi-A probe.

cDNA fragments that corresponded to the 3' untranslated region of the cytochrome P450 homologues were used to probe RNA blots and Southern blots of genomic DNA isolated from individual plants in V23 x R51 F2 population. By this analysis it was shown that the genes corresponding to cDNA clones pCGP174 and pCGP175 were expressed in a manner that paralleled 3',5'-hydroxylase activity. Furthermore, the gene corresponding to pCGP174 was shown to be closely linked to the Hfl locus and pCGP175 was linked to the Hf2 locus.

a. pCGP 174 A 330bp EIit1dIII - Kp1tI 3' fragment from clone pCGP174 (Fig. 5A) gave a pattern of hybridization on both RNA and DNA blots that suggested that this clone corresponded to the Hfl locus (Figure 6). The gene was expressed in both limb and tube tissues and had a developmental profile that paralleled 3',5'-hydroxylase activity, peaking in Stage 3 petal limbs. No expression was detected in leaf, but was induced in this tissue by the glucose/high light treatment. Furthermore, there was no detectable expression of the gene in the petal tissue of the hfl /hfl mutant lines R51 or Sw63. By contrast, relatively 2 112 ~7 -3 -46-high levels of expression were detected in the hlfl /);ifl lines V23 and Th7 and the V23 x R51 hybrid (Fig. 6A).

On Southern blots of genomic DNA digested with 20mI, the 330bp HindIII -KpnI 3' fragment from pCGP174 detected two RFLPs that segregated independently in the V23 x R51 F2 population. RFLP #1 corresponded to strongly hybridizing DNA bands while RFLP #2 corresponded to weakly hybridizing bands (see Fig. 6B). Eleven out of 12 plants that had been assigned a phl /phl genotype had a V23- like pattern for RFLP #1 and 49 out of 49 plants that had 3',5'-hydroxylase activity in the tube had either a V23 -or VR - like pattern for RFLP #1. In addition, for a total of 32 plants, there was complete co-segregation of the V23, VR and R51 RFLP patterns for chi A
(po) with the corresponding patterns of RFLP #2.

These data provided strong evidence that pCGP174 encoded a 3',5'-hydroxylase and corresponded to the Hfl locus (RFLP #1) and that the 3' probe cross-hybridized to the Hf2 locus (RFLP #2).

b. pCGP175 A 320 bp HindIII/XhQJ 3' fragment from clone pCGP175 (Fig. 5B) gave a pattern of hybridization on both RNA and DNA blots that suggested that this clone corresponded to the Hf2 locus (Fig. 7). The Northern analysis showed that the gene was developmentally regulated in a similar way to pCGP174, with maximal expression in Stage 3 OGB petal limbs, however no expression was detected in OGB tube tissue. The gene was also expressed in the petal tissue of V23 (Hf2/Hf2), Th7 (Hf2/Hf2) and the V23 x R51 hybrid (Fig. 7A).
On Southern blots, the 320bp H.indIll/XhaI fragment from pCGP175 hybridized to the same genomic fragments produced by XbaI digestion of V23 and R51 genomic DNA that weakly hybridized to the pCGP174 3' probe (RFLP #2). There was complete co-segregation of the V23-, VR- and R51 -like RFLP patterns detected by the pCGP175 3' probe and the corresponding WO 93/01290 2 2 7:3 PCT/AU92/00334 RFLP patterns for chi A(PQ) (Fig. 7B).

Yeast expression experiments (see below) subsequently confirmed that pCGP175 and a sibling of pCGP174 (pCGP176) both encoded a 3',5'-hydroxylase. In addition, expression of a full-length version of pCGP174 in an hfl /hfl, hf2/hf2 Petunia mutant resulted in increased 3',5'-hydroxylase activity and production of 3',5'-hydroxylated anthocyanins above the low basal levels normally found in the non-transgenic plant. Taken together with the RFLP results, it was concluded from these data that pCGP174 correspond to the Hfl locus and pCGP175 correspond to the Hf2 locus.
Isolation of Full-Length Hfl cDNA Clones and Sequence Analysis From preliminary sequence analysis it was shown that pCGP174 did not represent a full-length clone of the corresponding transcript while pCGP175 included a putative initiation codon and was presumed to be a full-length cDNA. Sequence analysis also showed that pCGP176 was a longer version of pCGP174 and included an ATG codon 17bp from the 5' end. However, from this analysis alone it was not possible to confidently predict whether pCGP176 included the entire coding region of this gene. Accordingly, cDNA library #2 was screened for longer clones of the pCGP174/pCGP176 sibling group.
Approximately 1.5 x 105 recombinants from cDNA library #2 were screened for clones that hybridized to the 0.33 kb HindIII - KpaI 3' fragment from pCGP174. Two hybridizing clones, designated pCGP601 and pCGP602 were chosen for further analysis. Both pCGP601 and pCGP602 included presumptive translation initiation codons, but pCGP602 encoded a longer 5' untranslated region.

A restriction enzyme map of pCGP602, indicating the methodology adopted for sequencing the clone and the oligonucleotide primer sequences used to obtain overlapping sequence information, is shown in Figure 8.

WO 93/01290 PCr/AU92/00334 The nucleotide sequence and deduced amino acid sequence of the sibling clones pCGP176 and pCGP602 are shown in Figure 9. Similarly, Figure 10 shows the nucleotide sequence and deduced translation product of pCGP175.
Using an alignment generated by the LFASTA program (Pearson and Lipman, 1988), the amino acid sequences encoded by the petunia 3',5'-hydroxylase genes were found to share 94% positional identity. The nucleotide sequences are 94% identical. Based on the classification scheme for cytochromes P450, this sequence similarily would place both genes in the same family/sub-family.
Because the 3',S'-hydroxylase amino acid sequences share less than 40%
identity with any previously characterised member of the cytochrome P450 superfamily, the corresponding genes belong to a new P450 family separate from all other P450 genes.
Expression of pCGP175 cDNA in Yeast The cDNA insert from pCGP175 was ligated in a sense orientation behind the glyceraldehyde-3-phosphate dehydrogenase promoter in the yeast vector pYGA22m. The resulting construct, designated pCGP618 (Fig. 11) was transformed into the yeast strain G-1315 (Ashikari et al, 1989). A single transformant was grown in 50 mL of YNBC at 30'C for 2 days. A microsomal fraction prepared from this culture was shown to have 3',5'-hydroxylase activity while an equivalent fraction prepared from non-transformed yeast had no activity (Fig. 12). From this it was concluded that the cDNA insert from pCGP175 encoded a 3',5'-hydroxylase.
Expression of pCGP176 cDNA in Yeast The cDNA insert from pCGP176 was ligated in a sense orientation behind the glyceraldehyde-3-phosphate dehydrogenase promoter in the yeast vector pYGA22m. The resulting construct, designated pCGP620 was transformed into the yeast strain G-1315. An extract prepared from the transformed yeast was shown to have 3',S'-hydroxylase activity while an equivalent fraction a~ ~.
- ~~~ ~

prepared from non-transformed yeast had no activity (Fig., 13): From this it was concluded that the cDNA insert from pCGP176 encoded a 3',5'-hydroxlase.

Expression of an Hfl cDNA
a. Expression in the hfl /I1f1, hf2/hf2 P. hybrida Fl hybrid Skr4 x Sw63 The pCGP602 cDNA insert was ligated behind the Mac promoter of the Ti-binary vector pCGP293. The resulting construct, designated pCGP90 (Fig. 14), was introduced into the F1 Petunia hybrid Skr4 x Sw63 using Agrobacteriurrr mediated gene transfer. Leaf discs of Skr4 x Sw63 were co-cultivated with AGLO/pCGP90 and integration of the pCGP602 cDNA insert in the Skr4 x Sw63 genome was confirmed by Southern analysis of plants obtained after kanamycin selection.

The transgenic plants had significantly higher levels of both 3',5'-hydroxylase enzyme activity (Fig. 15) and 3',5'-hydroxylated anthocyanins (Table 3A) than the non-transgenic Skr4 x Sw63 hybrid. Although Skr4 x Sw63 is homozygous recessive for both the Hfl and Hf2 genes these mutations do not completely block enzyme production as low levels of 3',5'-hydroxylase activity were detected in Skr4 x Sw63 petal extracts (Fig. 15). In addition, low levels (100 g/gm) of malvidin were detected in acid hydrolysed Skr4 x Sw63 petal extracts (Table 3A). Introduction of the Hfl cDNA significantly increased the level of 3',5'-hydroxylase activity in petal limb tissue (Fig. 15) and acid hydrolysed extracts of petals from the transgenic plants had four times the level of malvidin detected in the non-transgenic control (Table 3A).
b. Expression in Nicotiana tabacum cultivar Xanthi Tobacco (N. tabacum cv Xanthi) flowers produce cyanidin as the sole anthocyanidin. Transformation of tobacco with pCGP90 led to the accumulation of significant amounts of delphinidin in the flowers, in addition to cyanidin (shown in Table 3A).

Pigment Analysis Anthocyanidin levels found in acid hydrolysed petal extracts Plant Malvidin Cyanidin Delphinidin (pg/gm petal) (pg/gm petal) (pg/gm petal) Petunia Skr4 x Sw63 100 nd' nd Skr4 x Sw63/pCGP90 410 nd nd Tobacco non-co-cultivated nd 272 nd control Transgenic tobacco nd 229 36 1 not detected c. Expression in the hfl/hfl, hf2/hf2 P. hybrida Fl hybrid Rp57 x Rw14 The petunia line Rp57 x Rw14 was transformed using pCGP90 and a procedure similar to that used for Skr4 x Sw63. Transgenic flowers produced considerable amounts of petunidin and malvidin which were not detectable in the non-transformed plants (Table 3B). Petunidin and malvidin are both methylated derivatives of delphinidin.

Pigment Analysis of high pH line Rp57xRw14.
Percentages of anthocyanidins found in acid hydrolysed petal extracts Plant Cyanidin Peonidin Petunidin Malvidin (%) (%) (%) (%) Petunia Rp57xRwl4 5.0 95.0 0 0 Rp57xRw14/pCGP90 0 45.2 7.8 47.0 The expression of the introduced Hfl cDNA in the Skr4 x Sw63 hybrid had a marked effect on flower colour. The carpel and stamen tissue of the non-transgenic flowers are white whereas these same tissues in the transgenic plants were a blue/purple colour. In addition, expression of the Hfl cDNA in the Skr4 x Sw63 hybrid conferred a deep pink/violet hue to the corolla which is normally very pale pink. In the case of tobacco, the production of delphinidin derivatives led to a minor blueing of senescing flowers.
Expression of the Ht1 cDNA in the Rp57 x Rw14 hybrid again had a marked effect on flower colour. Non-transgenic Rp57 x Rw14 flowers are pink, with peonidin being the major anthocyanidin present (see Table 3B). Transformation with Hf1 cDNA led to a marked blueing of flower colour.

The colour changes observed may also be described in terms of numbers from the Royal Horticultural Society's Colour Chart. In general, the changes can be described as moving the colour from the pale-to-mid pink hues of 60C/D -65C/D, to the darker bluer/purpler hues represented by many, but not all, of ~11-the colour squares between 70 and 85. Although not wishing to limit the possible colour changes which may be achieved, some of the colours observed in the Skr4 x Sw63 hybrid could be described, approximately, as having changed from 65B (untransformed) to 70B and to 74B (both transformed).
Likewise, several in the Rp57 x Rw14 hybrid might be described as moving from 64C to 72B to 77B and to 82B. It should be remembered that other biochemical and physiological conditions will affect the individual outcome and the citing of specific colours achieved should not be interpreted as defining the possible range.
Detection of Putative 3',5'-hydroxylase Gene Sequences in Other Plant Species The presence of 3',4',5'-hydroxylated flavonoids is correlated with the presence of 3',5'-hydroxylase activity and therefore the 3',5'-hydroxylase gene. It would be expected that these genes from other species would hybridize with the petunia 3',5'-hydroxylase gene under conditions of low stringency. RNA (Figure 16) and/or DNA (Figure 17) was isolated from a number of delphinidin-producing plants, probed with 32P-labelled Hfl cDNA
and washed under different conditions of stringency. Hybridizing bands were detected in every example. Therefore, the isolation of 3',5'-hydroxylase genes from other delphinidin-producing plants should be possible using a petunia 3',5'-hydroxylase gene as a probe.

Those skilled in the art will appreciate that the invention described herein is susceptible to variations and modifications other than those specifically described. It is to be understood that the invention includes all such variations and modifications. The invention also includes all of the steps, features, compositions and compounds referred to or indicated in this specification, individually or collectively, and any and all combinations of any two or more of said steps or features.

WO 93/01290 2112373 PCr/AU92/00334 REFERENCES:
Asen, S., Stewart, R.N. and Norris, K.H. Phytochemistry 14: 2677-2682, 1975.
Asen, S., Griesbach, R.J., Norris, K.H. and Leonhardt, B.A. Phytochemistry25:
2509-2513, 1986.

Ashikari, T., Kiuchi-Goto, N., Tanaka, Y., Shibano, Y., Amachi, T., and Yoshizumi, H. Appl. Microbiol. Biotechnol. 30: 515-520, 1989.

Aviv, H. and Leder, P., Proc. Natl. Acad. Sci. USA 69: 1408-1412, 1972.
Beale, G.H. Journal of Genetics 40(3): 337-358, 1940.

Bethesda Research Laboratories. BRL pUC host: E. coli DH5aTm competent cells. Bethesda Res. Lab. Focus. 8(2): 9, 1986.

Bozak, K.R., Yu, H., Sirevag, R. and Christoffersen, R.E., Proc. Natl. Acad.
Sci.
USA 87: 3904-3908, 1990.

Bullock, W. 0., Fernandez, J.M. and Short, J.M. Biotechniques 5: 376, 1987.
Comai, L., Moran, P. and Maslyar, D., Plant Molecular Biology 15: 373-381, 1990.

Cornu, A., Genetics. In: Petunia Sink, K.C. (Ed). Springer-Verlag, Berlin, Germany pp 35-47, 1984.

Cornu, A., Farcy, E., Maizonnier, D., Haring, M., Veerman, W. and Gerats, A.G.M. In Genetic maps - Locus maps of complex genomes 5th edition, Stephen J. O'Brien (Ed), Cold Spring Harbor Laboratory Press, USA, 1990.

z 1~~~ -54-Dellaporta, S.J., Wood, J. and Hick, J.B. Plant Mol. Biol. Rep. 1: 19-21, 1983.
De Vlaming, P., Gerats, A.G.M., Wiering, H. and Wijsman, H.J.W. Plant Mol.
Biol. Rep. 2(2): 21-42, 1984.

Doodeman, M., Gerats, A.G.M., Schram, A.W., de Vlaming, P. and Bianchi, F.
Theor. Appl. Genet. 67: 357-366, 1984.

Ebel, J. and Hahlbrock, K., In The Flavonoids: Advances in Research Since 1980. Harborne, J.B. (Ed.), Academic Press, New York, USA, 641-679, 1988.
Forkmann, G. Plant Breeding 106: 1-26, 1991.

Forkmann, G. and Stotz, G. Z. Naturforsch 36c: 411-416, 1981.
Garfinkel, D.J. and Nester, E.W. J. Bacteriol.144:732-743, 1980.

Hagmann, M., Heller, W. and Grisebach, H. Eur. J. Biochem. 134: 547-554, 1983.

Hahlbrock, K. and Grisebach, H. Annu. Rev. Plant Physiol. 30: 105-130, 1979.
Hanahan, D. J. Mol. Biol. 166: 557, 1983.

Harborne, J.B. and Simmonds, N.W. Annu. Rep. John Innes Inst. 53: 29-30, 1962.

Heller, W. and Forkmann, G. In: The Flavonoids, Advances in Research Since 1980. Harborne, J.B. (Ed.), Academic Press, New York, 1988.

Holton, T.A. and Graham, M.W. Nucleic Acids Research 19: 1156, 1991.

~112 37 3 Inoue, H., Nojima, H. and Okayama, H. Gene 96: 23-28, 1990.

Ito, H., Fukuda, Y., Murata, K. and Kimura, A. J. Bacteriol., 153: 163-168, 1983.

Jefferson, R.A., Kavanagh, T.A. and Bevan, M.W. EMBO J. 6(13): 3901-3907, 1987.

Koes, R.E., Spelt, C.E., Reif, H.J., van den Elzen, P.J.M., Veltkamp, E. and Mol, J.N.M. Nucl. Acids Res. 14(13): 5229-5239, 1986.

Larson, R.L. and Bussard, J.B. Plant Physiol. 80: 483-486, 1986.

Lazo, G.R., Pascal, A.S. and Ludwig, R.A. Bioltechnology 9: 963-967, 1991.
Lew, A.M. and Kemp, D.J. Nucl. Acids Res. 17(14): 5859-5860, 1989.

McBride, K.E. and Su.mmerfelt, K.R. Plant Molecular Biology 14: 269-276 1990.
Martin, F.M., Castro, N.M., Aboula-ela, F., Tinoco, I. Nucl. Acids Res. 13:
8927-8938, 1985.

Matthews, J.M., Dostal, L.A. and Bend, J.R. J. Pharmacology and Experimental Therapeutics 235(1) 186-190, 1985.

Merrifield, J. Am. Chem. Soc. 85: 2149, 1964.

Murashige, T. and Skoog, F. Physiol. Plan115: 73-97, 1962.

cz 1!-56-Nebert, D.W., Nelson, D.R., Coon, M.J., Estabrook, R.W., Feyereisen, R., Fujii-Kuriyama, Y., Gonzalez, F.J., Guengerich, F.P., Gunsalus, I.C. Johnson, E.F., Loper, J.C., Sato, R., Waterman, M.R., Waxman, D.J. DNA and Cell Biology 10: 1-14, 1991.

Norrander, J., Kemp, T. and Messing, J. Gene 26: 101, 1983.
Oeda, K., Sakaki, T., and Ohkawa, H. DNA 4: 203-210, 1985.
O'Keefe, D.P. and Leto, K.J. Plant Physiol. 89: 1141-1149, 1989.

Ohtsuka, E., Matsuki, S., Ikehara, M., Takahashi, Y. and Matsubara, K. J.
Biol.
Chem. 260(5): 2605-2608, 1985.

Pearson, W.R. and Lipman, D.J. Proc. Natl. Acad. Sci. USA 85: 2444-2448, 1988.

Rademacher, W., Fritsch, H., Graebe, J.E., Sauter, H. and Jung, J. Pesticide Science 21: 241-252, 1987.

Sambrook, J., Fritsch, E.F. and Maniatis, T. Molecular Cloning: A Laboratory Manual (2nd edition). Cold Spring Harbor Laboratory Press, USA, 1989.
Sanger, F., Nicklen, S. and Coulson, A. Proc. Natl. Acad. Sci. USA 74: 5463-5467, 1977.

Schenk, R.U. and Hilderbrandt, A.C. Can. J. Bot. 50: 199-204, 1972.

Schram, A.W., Jonsson, L.M.V. and Bennink, G.J.H. Biochemistry of flavonoid synthesis in Petunia hybrida. In: Petunia Sink, K.C. (ed.) Springer-Verlag, Berlin, Germany pp 68-75, 1984.

WO 93/01290 ~ I I Z 3 7 ~ PCT/AU92/00334 Stafford, H.A. Flavonoid Metabolism CRC Press, Inc. Boca Raton, Florida, USA, 1990.

Stotz, G. and Forkmann, G. Z. Naturforsch 37c: 19-23, 1982.

Stotz, G., de Vlaming, P., Wiering, H. and Forkmann, G. Theor. Apppl. Genet.
70: 300, 1985.

Takeda, K., Kubota, R. and Yagioka, C. Phytochemistry 24: 1207, 1985.
Takemori, S. and Kominami, S. Cytochrome P450. Tokyo University Press, Japan, 1989.

Taton, M., Ullman, P., Beneveniste, P. and Rahier, A. Pesticide Biochemistry and Physiology 30: 178-189, 1988.

Turpen, T.H. and Griffith, O.M. BioTechniques 4: 11-15, 1986.

van Tunen, A.J., Gerats, A.G.M. and Mol, J.N.M. Plant Mol. Biol. Rep. 8: 50-59, 1990.

van Tunen, A.J., Koes, R.E., Spelt, C.E. van der Krol, A.R., Stuitje, A.R. and Mol., J.N.M. EMBO J. 7(5): 1257-1263, 1988.

van Tunen, A.J., Mur, L.A., Recourt, K., Gerats, A.G.M. and Mol., J.N.M. The Plant Cell 3: 39-48, 1991.

von Wettstein-Knowles, P. Hereditas 60: 317-346, 1968.

Vercruysse, S.A.R., Delcour, J.A. and Dondeyne, P. J. Chromatography 324:
495-497, 1985.

Wallroth, M., Gerats, A.G.M., Rogers, S.G., Fraley, R.T. and Horsch, R.B.
Mol. Gen. Genet. 202: 6-15, 1986.

Wiering, H. and De Vlaming, P. Inheritance and Biochemistry of Pigments.
In: Petunia Sink, K.C. (Ed.), Springer-Verlag, Berlin, Germany pp 49-65, 1984.

Wiering, H. Genen Phaenen 17(1-2~ 117-134, 1974 Yanisch-Perron, C., Vieira, J. and Messing, J. Gene 33: 103, 1985.

Claims (37)

CLAIMS:

1. An isolated DNA molecule comprising:

(a) a sequence of nucleotides encoding a 3',5'-hydroxylase or encoding a derivative of 3',5'-hydroxylase which is capable of hydroxylating dihydrokaempferol (DHK); or (b) a sequence complementary to the sequence in (a);

wherein the sequence in (a) comprises:

(i) a nucleotide sequence set forth in Figure 9 or 10; or (ii) a nucleotide sequence capable of hybridizing to the complement of the nucleotide sequence set forth in Figure 9 or 10 under stringency conditions of 20% v/v formamide, 6 x SSC, 1% w/v SDS at 42°C for 16 hours, followed by washing in 6 x SSC, 1% w/v SDS at 65°C for 1 hour.

2. An isolated DNA molecule comprising:

(a) a sequence of nucleotides encoding a 3',5'-hydroxylase or encoding a derivative of 3',5'-hydroxylase which is capable of hydroxylating dihydrokaemferol (DHK); or (b) a sequence complementary to the sequence in (a);

wherein the sequence in (a) comprises:

(i) a nucleotide sequence set forth in Figure 9 or 10; or (ii) a nucleotide sequence encoding an amino acid sequence having at least 60% identity after optimal alignment to the amino acid sequence shown in Figure 9 or 10.

3. The isolated DNA molecule according to Claim 1, wherein the protein having DHK hydroxylating activity is of petunia, verbena, delphinium, grape, iris, freesia, hydrangea, cyclamen, potato, pansy or eggplant origin.

4. The isolated DNA molecule according to Claim 2, wherein the protein having DHK hydroxylating activity is of petunia, verbena, delphinium, grape, iris, freesia, hydrangea, cyclamen, potato, pansy or eggplant origin.

5. The isolated DNA molecule according to Claim 3 or 4, wherein the protein is of petunia origin.

6. A vector comprising the DNA molecule according to any one of Claims 1 to 5 capable of transferring said DNA
molecule to a plant cell or tissue.

7. The vector according to Claim 6 which is an Agrobacterium vector.

8. The vector according to Claim 6, which is pCGP90.

9. The vector according to Claim 7, which is pCGP90.

10. A 3',5'-hydroxylase protein having DHK

hydroxylating activity having an amino acid sequence encoded by the DNA molecule according to any one of Claims 1 to 5.

11. An isolated 3',5'-hydroxylase protein comprising the amino acid sequence as set forth in Figure 9 or 10, or an isolated derivative of 3',5'-hydroxylase which is capable of hydroxylating dihydrokaempferol (DHK), wherein the derivative is encoded by a nucleotide sequence capable of hybridizing to the complement of the nucleotide sequence shown at Figure 9 or 10 under stringency conditions of 20%
v/v formamide, 6 x SSC, 1% w/v at 42°C for 16 hours, followed by washing in 6 x SSC, 1% w/v SDS at 65°C for 1 hour.

12. A 3',5'-hydroxylase protein having DHK
hydroxylating activity, having an amino acid sequence shown in Figure 9 or 10, or an amino acid sequence having at least 60% identity after optimal alignment to said amino acid sequence.

13. The 3',5'-hydroxylase protein according to Claim 10, wherein said protein is of petunia, verbena, delphinium, grape, iris, freesia, hydrangea, cyclamen, potato, pansy or eggplant variety.

14. The 3',5'-hydroxylase protein according to Claim 11, wherein said protein is of petunia, verbena, delphinium, grape, iris, freesia, hydrangea, cyclamen, potato, pansy or eggplant origin.

15. The 3',5'-hydroxylase protein according to Claim 13, wherein the protein is of petunia origin.

16. A method for producing a transgenic plant capable of expressing a recombinant 3',5'-hydroxylase or which directs transcription of DNA sequence which is complementary to all or part of a mRNA molecule translatable to a 3',5'-hydroxylase, said method comprising introducing into a cell of a plant the DNA molecule according to any one of Claims 1 to 5 under conditions permitting the expression of said DNA
molecule, regenerating a transgenic plant from the cell and growing said transgenic plant for a time and under conditions sufficient to permit the expression of the DNA
molecule.

17. The method according to Claim 16, wherein the transfer of DNA molecule is carried out using a vector according to any one of Claims 6 to 9.

18. The method according to Claim 16, wherein the expression of the DNA molecule is developmentally regulated.

19. The method according to Claim 17, wherein the expression of the DNA molecule is developmentally regulated.

20. The method according to any one of Claims 16 to 19, wherein the 3',5'-hydroxylase is of petunia, verbena, delphinium, grape, iris, freesia, hydrangea, cyclamen, potato, pansy or eggplant origin.

21. The method according to Claim 20, wherein the 3',5'-hydroxylase is of petunia origin.

22. The method according to any one of Claims 16 to 19, wherein the transgenic plant is a rose, petunia, chrysanthemum, carnation, gerbera, iris, tulip, lily, lisianthus, freesia, delphinium, limonium or pelargonium.

23. The method according to Claim 20, wherein the transgenic plant is rose, petunia, chrysanthemum, carnation, gerbera, iris, tulip, lily, lisianthus, freesia, delphinium, limonium or pelargonium.

24. The method according to Claim 22 or 23, wherein the transgenic plant is a rose or petunia or carnation.

25. An isolated DNA molecule encoding a flavonoid 3',5'-hydroxylase capable of hydroxylating dihydrokaempferol (DHK), wherein said isolated DNA molecule comprises the nucleotide sequence as set forth in Figure 9 or Figure 10.

26. A transgenic plant cell incorporating the DNA
molecule according to any one of Claims 1 to 5 and 25, or the vector according to any one of Claims 6 to 9.

27. The transgenic plant cell according to Claim 26, wherein the protein is a petunia, verbena, delphinium, grape, iris, freesia, hydrangea, cyclamen, potato, pansy or eggplant origin.

28. The transgenic plant cell according to Claim 27, wherein the protein is of petunia origin.

29. The transgenic plant cell according to Claim 26, wherein the transgenic plant is rose, petunia, chrysanthemum, carnation, gerbera, iris, tulip, lily, lisianthus, freesia, delphinium, limonium or pelargonium.

30. The transgenic plant cell according to Claim 27 or 28, wherein the transgenic plant is rose, petunia, chrysanthemum, carnation, gerbera, iris, tulip, lily, lisianthus, freesia, delphinium, limonium or pelargonium.

31. The transgenic plant cell according to Claim 29, wherein the transgenic plant is a rose or petunia or carnation.

32. The transgenic plant cell according to Claim 26, wherein the cell is a flower-derived cell.

33. The transgenic plant cell according to any one of Claims 27 to 29, wherein the cell is a flower-derived cell.

34. A method for cloning a nucleic acid molecule comprising a sequence of nucleotides encoding a flavonoid 3',5'-hydroxylase or derivative thereof which hydroxylates one or more of DHK and naringenin, or the sequence complementary thereto, said sequence comprising:

the nucleotide sequence set forth in Figure 9 or 10, or the sequence complementary to either the nucleotide sequence in Figure 9 or 10, or a nucleotide sequence capable of hybridizing to the nucleotide sequence set forth in Figure 9 or 10, or to the sequence complementary thereto, under stringency conditions of 20% v/v formamide, 6 x SSC, 1% w/v SDS at 42°C
for 16 hours followed by washing in 6 x SSC, 1% w/v SDS at 65°C for 1 hour, said method comprising amplification of flavonoid 3',5'-hydroxylase nucleotide sequences or complementary sequences from a suitable preparation of nucleic acid molecules from cells of said plant by one or more polymerase chain reactions using one or more oligonucleotide primers based on the amino acid sequence PFGAGRRICPG.

35. A transgenic plant cell obtained by the method of any one of Claims 16 to 24.

36. Method for producing a transgenic plant cell capable of expressing a recombinant 3',5'-hydroxylase or which directs transcription of a nucleic acid sequence which is complementary to all or a part of a mRNA molecule translatable to a 3',5'-hydroxylase, said method comprising introducing into a cell of a suitable plant the nucleic acid isolate according to any one of Claims 1 to 5 under conditions permitting the eventual expression of said nucleic acid isolate.

37. An isolated DNA molecule for use as an antisense molecule comprising a complementary sequence, wherein the complementary sequence:

(a) is complementary to a length of the nucleotide sequence set forth in Figure 9 or 10; or (b) is a length of a nucleotide sequence capable of hybridizing to the nucleotide sequence set forth in Figure 9 or 10 under stringency conditions of 20% v/v formamide, 6 x SSC, 1% w/v SDS at 42°C for 16 hours, followed by washing in 6 x SSC, 1% w/v SDS at 65°C for 1 hour;

wherein the length is sufficient for the complementary sequence to act as an antisense molecule to down-regulate expression of a nucleotide sequence encoding a 3',5'-hydroxylase or a derivative thereof capable of hydroxylating dihydrokaempferol (DHK).