A trypta ine producing tryptophan decarboxylase gene of plant origin.

BACKGROUND OF THE INVENTION Tryptophan decarboxylase (TDC; E.C. 4.1.1.27) catalyses the conversion of L-tryptophan to tryptamine. This enzyme has been detected in numerous plant systems and it has been suggested that its primary role is to supply possible precursors for auxin biosynthesis (Baxter, C. & Slaytor, M. (1972) Phytochemistry ϋ, 2763-2766; Gibson, R.A., Barret, G. & Wightman F. (1972) J. Exp. Bot. 2., pages 775-786; Gross, W. & Klapchek, S. (1979) Z. Pflanzenphysiol. £3., pages 359-363).

In the Gramineae, TDC catalyses the synthesis of precursors for the protoalkaloids which have considerable physiological activity in higher animals (Smith, T.A. , (1977) Phytochemistry Vol. 1., pages 171- 175). It is also known that tryptophan-derived tryptamines are also precursors of the tricyclic β- carboline alkaloids formed by condensation with a one- or two- carbon moiety (Slaytor, M. , & McFarlane, I.J., (1968) Phytochemistry 7_, pages 605-610).

Furthermore, in periwinkle (Catharanthus roseus) , TDC produces tryptamine for biosynthesis of the commercially important antineoplastic monoterpenoid indole alkaloids, vinblastine and vincristine (De Luca, V., & Kurz, .G. . (1988), Cell Culture and Somatic Cell Genetics of Plants, Constabel, F. and Vasil, I.K., eds. Academic Press 5_, pages 385-401). The TDC from Catharanthus roseus has been purified to homogeneity. It occurs as a dimer consisting of 2 identical subunits of Mr 54,000 and it requires pyridoxal phosphate for activity (Noe, W. , Mollenschott, C. , & Berlin J. (1984) Plant Mol. Biol. 3., pages 281-288) .

The enzyme possesses characteristics of plant aromatic decarboxylases which usually exhibit high

substrate specificity. For example, TDC will decarboxylate L-tryptophan and 5-hydroxy-L-tryptophan but is inactive towards L-phenylalanine and L-tyrosine, while the tyrosine decarboxylases from Syrinqa vulgaris (Chappie, C.C.S., (1984) Ph.D. Thesis, University of Guelph, Guelph, Ontario, Canada), Thalictrum ruqosu and Escholtzia californica (Marques, I.A., & Brodelius, P. (1988) Plant Physiol. 88., pages 52-55), accept L- tyrosine and L-dopa as substrates but not L-tryptophan or 5-hydroxy-L-tryptophan. The aromatic L-amino acid decarboxylases (dopa decarboxylase (DDC), ED 4.1.1.28) of EL, roelanogaster (Clark, W.C., Pass, P.S., Venkatararman, B., & Hodgetts, R.B. (1978) Mol. Gen. Genet. 162, pages 287-297; Eveleth, D.D., Gietz, R.D., Spencer C.A., Nargang, F.E., Hodgetts, R.B. & Marsh, J.L. (1986) Embo. J. 5_, pages 2663-2672; Morgan B.A., Johnson, .A. & Hirsh, J. (1986) Embo. J. 5_, pages 3335- 3342) and mammals (Albert, V.R., Allen, J.M. , & Joh, T.H. (1987) J. Biol. Chem. .262., pages 9404-9411) have a broader substrate specificity with L-dopa, tyrosine, phenylalanine and possibly histidine also serving as substrates.

In animals, the role of aromatic L-amino acid decarboxylase is to produce the major neurotrans itters dopamine and serotonin and, in E . elanogaster, the DDC enzyme serves a second, inducible role, in the sclerotization of the insect cuticle (Christenson, J.G., Dairman, . & Undenfriend, S. (1972) Proc. Natl. Acad. Sci. USA 6J9., pages 343-347; Lovenberg, . , Weissbach, ., & Undenfriend S. (1962) J. Biol. Chem. 237, pages 89-93; Yuwiler, A., Geller, E. & Eiduson, S. (1954) Arch. Biochem. Biophys. jJO., pages 162-173; Brunet, P. (1980) Insect Biochem. .10., pages 467-500).

It would appear highly desirable to be able to clone the cDNA sequence of tryptophan decarboxylase from Catharanthus roseus, thus, providing the development of the cDNA sequence in a plasmid vector capable of

transforming cell lines that will produce the tryptophan decarboxylase enzyme.

If the tryptophan decarboxylase gene could be inserted into living organisms by transformation to produce tryptamine and related protoalkaloids, it could supplement a neurotransmitter deficiency.

Further, the insertion of this gene in plants could be useful to alter the spectrum of tryptophan- based chemicals normally produced by the plant. For example, the insertion of constitutive expression of tryptophan decarboxylase in Brassica species could sequester the cytoplasmic tryptophan pool for the synthesis of tryptamine and related protoalkaloids and therefore repress the normal synthesis and accumulation of indole glucosinolates.

Hence, creation of plants with an altered chemical spectrum may produce novel phenotypes which have resistance to various pathogenic diseases or to insect pests. SUMMARY OF THE INVENTION

In accordance with the present invention, there is now provided the sequence of a cDNA clone which includes the complete coding region of tryptophan decarboxylase, preferably tryptophan decarboxylase (E.C. 4.1.1.27) from periwinkle (Catharanthus roseus) . The cDNA clone (1747 bp) was isolated by antibody screening of a cDNA expression library produced from poly A* RNA found in developing seedlings of C_j_ roseus. The clone hybridized to a 1.8 kb RNA from developing seedlings and from young leaves of mature plants.

Also within the scope of the present invention is a method for inserting TDC gene into living organisms by transformation. The identity of the clone was confirmed when extracts of transformed E_j_ coli expressed a protein containing tryptophan decarboxylase enzyme activity. The tryptophan decarboxylase cDNA clone encodes a protein of 500 amino acids with a calculated

molecular mass of 56,142 Da. The amino acid sequence shows a high degree of similarity with the aromatic L- a ino acid decarboxylase (dopa-decarboxylase) and the alpha-methyldopa hypersensitive protein of Drosophila melaqonaster. The tryptophan decarboxylase sequence also showed significant similarity to feline glutamic acid decarboxylase and mouse ornithine decarboxylase suggesting a possible evolutionary link between these amino acid decarboxylases. Furthermore, the protein encoded by the cDNA clone of the present invention is active in. vitro. IN THE DRAWINGS

Figure 1 (lane 2) represents the TDC enzymatic activity in extracts of pTDC5-transformed E^ coli, compared to those in control Ej_ coli (lane 1) and that in C_j_ roseus itself (lane 3).

Figure 2 represents the hybridization of the pTDC-5 clone to a 1.8 kb mRNA species isolated from periwinkle. Figure 3 shows the nucleotide sequence of the pTDC5 cDNA clone and its deduced amino acid sequence. The putative polyadenylation signal is underlined.

Figure 4 shows the amino acid sequence alignments of the protein for the I _ melanogaster alpha methyldopa hypersensitive gene (AMD), C_j_ roseus tryptophan decarboxylase (TDC), and Drosphila DOPA decarboxylase iεoenzy e 1 (DDC1).

Figure 5 shows hydropathy profile of TDC and DDC1. Other advantages of the present invention will be readily illustrated by referring to the following description.

DETAILED DESCRIPTION OF THE INVENTION cDNA synthesis and DNA sequencing. Seedlings of Cj_ roseus (L. ) G. Don cv "Little

Delicata" were germinated and grown for 5 days in the dark as described previously (De Luca, V. , Alvarez-

Fernandez, F., Campbell, D., & Kurz, W.G.W. (1988) Plant Physiol. 86., 447-450). Seedlings were harvested after 18 hours of light treatment and total RNA was isolated as described by Jones, J.D.G., Duns uir, P. & Bedrook, J. (1985) EMBO J. 4., 2411-2418.

Poly(A) ⁺ RNA was isolated by chromatography on oligo (dT) ^" cellulose (Aviv, H. & Leder, P. (1972) Proc. Natl. Acad. Sci. USA 69., 1408-1412) and double-stranded cDNAs were prepared according to the procedure of Gubler and Hoffman (1983, Gene 25., 263-269). Following ligation with Eco RI linker, the cDNA was inserted into the Eco RI site of the expression vector ZAP (Stratagene, San Diego, Short, J.M., Fernandez, J.M., Sorge, J.A. & Huse, W.D. (1988) Nucl. Acids Res. 16., 7583-7600). A library containing 3.1 X 10 ⁵ recombinant phages was obtained and after amplification, 2 X 10 plaques were screened with specific polyclonal antiserum raised against-TDC. Plasmids (pBluescript) containing a TDC cDNA insert were rescued using the R408 fl helper phage (Short, J.M., Fernandez, J.M., Sorge, J.A. & Huse, W.D. (1988) Nucl. Acids Res. ϋ, 7583-7600) and the nucleotide sequence of a full-length cDNA clone (pTDC5) was determined on both strands by the dideoxy-chain- termination method (Sanger, F., Nicklen, S. & Coulson, A.R. (1977) Proc. Natl. Acad. Sci. USA 74., 5463-5467). The sequencing strategy included subcloning of restriction fragments and the use of oligonucleotide primers. The sequence for all restriction sites used for the subcloning was determined on at least one strand. Comparisons of the pTDC5 cDNA nucleotide sequence and of the deduced amino acid sequence with Genbank and NBRF sequence libraries were performed using the FASTA program package (Pearson, W.R. & Lipman, D.J. (1988) Proc. Natl. Acad. Sci. USA 85 _. , 2444-2448). RNA blot hybridization.

Poly(A) ⁺ RNA was isolated from 6 day old developing seedlings and from young leaves of mature

plants as described above. These tissues were chosen as a likely source of TDC poly(A) ⁺ RNA based on the presence of high levels of TDC enzyme activity (De Luca, V., Alvarez-Fernandez, F., Campbell, D., & Kurz, W.G.W. (1988) Plant Physiol. 86., 447-450). RNA was denatured, fractioned by electrophoresis on formaldehyde/agarose gels, and then transferred to nitrocellulose filters (Maniatiε, T., Fritsch, E.F. & Sambrook, J. (1982) In: Molecular Cloning, A Laboratory Manual. Cold Spring Harbor, New York). Blotted RNA was hybridized to [ ³² PJ- labelled pTDC5 DNA and autoradiography was performed using Kodak XAR-5 films. TDC activity in extracts of E^_ coli.

A culture (100 ml) of the E_^ coli strain ZL1- blue containing pTDC5 or pBluescript was incubated at 37°C for 2 hours before adding the IPTG inducer at a final concentration of 1 mM. Incubation was continued for an additional 2 hours. Cells were harvested, washed in TE buffer, resuspended and lysed in 3 ml of a buffer containing 0.1 M Hepes, pH 7.5, 1 mM DTT. Debris was removed by centrifugation and the supernatant was desalted by passage over a Sephadex 6-25** column. TDC enzymatic activity in bacterial supernatants was determined by monitoring the conversion of L-[roethylene- ¹⁴ C]-tryptophan to [ ¹⁴ C]- ryptamine (De Luca, V., Alvarez- Fernandez, F. , Campbell, D., & Kurz, W.G.W. (1988) Plant Physiol. 86., 447-450). Supernatants (30 pi ) were incubated in the presence of 0.1 μCi of ["C]-tryptophan (sp. act. 59 mCi/mmol. ) for 30 minutes and reactions were stopped with 100 jul NaOH. Radioactive tryptamine was extracted from the reaction mixture with ethyl acetate and was analyzed by silica gel thin layer chromatography and autoradiography. Determination of TDC enzyme activity in leaves was performed as described previously (De Luca, V., Alvarez-Fernandez, F., Campbell, D., & Kurz, W.G.W. (1988) Plant Physiol. 86, 447-450).

TDC enzymatic activity in E_;_ coli.

A tryptophan decarboxylase cDNA clone of C. roseus was isolated by the use of antibody screening of an expression library. The antigenicity and enzymatic activity (Figure 1) of the encoded protein established the identity of the TDC cDNA.

When the original cDNA library was screened with the anti-TDC antibody, 27 clones were identified. Six clones were selected and submitted to further analysis. Partial sequence analysis revealed no difference among these clones, except for their length. Therefore, the clone having the longest cDNA insert (pTDC5) was selected for further characterization. To confirm that this cDNA clone corresponded to TDC, enzymatic activity was measured in cell extracts from Ej _, coli. Figure 1 shows that [ ¹⁴ C]-tryptamine was produced with extracts from cells transformed with pTDC5, and with extracts from Cj_ roseus leaves (lane 3), but not with extracts from cells containing only the vector (lane 1). The conversion of [ ^l *C]-tryptophan to [ ¹⁴ C]- tryptamine was monitored in extracts of E^ coli and C. roseus leaves. [ ¹⁴ C]-tryptophan (sp. act. 50 Ci/mmol) for 30 minutes. After addition of base, [ ¹⁴ C]-tryptamine was extracted fro the reaction mixture with ethyl acetate and reaction products were analyzed by thin layer chromatography on silica gel (solvent CHCl _j MeOH: 25% NH, (5s :1) and autoradiography. In Figure 1, TDC enzymatic activity is shown; lane 1, E^_ coli is transformed by the pBluescript vector, lane 2, Ej_ coli is transformed by pTDC5 and lane 3, C_j_ roseus extract is shown.

This result indicated that TDC enzymatic activity was retained by the protein produced using a TDC cDNA clone under the control of the Lac promoter of the pBluescript vector. No attempts were made to quantify the level of activity of TDC in E. coli.

Sequence analysis of a TDC cDNA clone.

DNA sequence analysis of pTDC5 revealed the presence of an open reading frame coding for a protein of 500 amino acids, which corresponded to a molecular mass of 56,142 Da (Figure 2). The 5'-nontranslated region of pTDC5 contained 69 nucleotides and included, near its beginning, a long stretch of alternating pyriraidines. Sequence around the ethionine initiation codon (AAUAAUGGG) matched closely the consensus sequence for plant gene initiation codons (AACAAUGGC) (Lϋtcke, H.A. , Chow, K.C., Mickel, F.S., Moss, K.A., Kerm, H.F. and Scheele, G.A. (1987) EMBO J. 6., 43-48). The 3'nontranslated region consisted of 168 nucleotides up to the poly(A) tail and contained an AAUAAA putative poly(A) ⁺ addition signal 17 nucleotides upstream from the start of the poly(A)* tail. Examination of the predicted amino acid sequence did not reveal the presence of a signal sequence (Watson, M.E.E. (1984) Nucl. Acids Res. 12, 5145-5164), which is consistent with the proposed cytoplasmic location of TDC within the cell (De Luca, V., Alvarez-Fernandez, F., Campbell, D., & Kurz, W.G.W. (1988) Plant Physiol. 8£ _r 4474-50).

Comparison of TDC-cDNA nucleotide and deduced amino acid sequences with nucleotide sequences in the Genbank DNA sequence database and with amino acid sequences in the NBRF protein sequence database revealed surprising similarity (40% amino acid identity) with the dopa-decarboxylase isoenzyme 1(DDC1) from D. melanogaster (Eveleth, D.D., Gietz, R.D., Spencer, C.A., Nargang, F.E., Hodgetts, R.B. & Marsh, J.L. (1986) EMBO J. 5., 2663-2672; Morgan, B.A., Johnson, W.A. & Hirsh, J. (1986) EMBO J. 5., 3335-3342), and with the protein corresponding to the D_;_ melanogaster alpha-methyldopa hypersensitive gene (AMD, 35% amino acid identity) (Eveleth, D.D. & Marsh, J.L. (1986) Genetics 114. 469- 483) (Figure 3). In Figure 3, the boxes show TDC residues present in AMD and/or DDC1 sequences. Amino

acids are numbered for TDC (top) and DDC1 (bottom). The areas of amino acid similarity extended throughout the protein and were not restricted to a particular portion of either structure. Furthermore, the 39% amino acid sequence similarity could be extended to the predicted distribution of potential alpha helices and beta sheets. This indicated that the amino acid differences between the two proteins did not significantly alter their secondary structures, and may indicate the importance of such conserved domains to mediate subunit assembly, as well as catalytic function and substrate specificity.

Limited proteolysis of pig kidney dopa decarboxylase and amino acid sequencing of a tryptic fragment produced a sequence for 50 amino acid residues one third of the distance from the COOH terminus of this protein. (Tancini, B., Dominici, P., Simmaco, M., Schinina, M.E., Barra, D. , & Voltatormi, CD. (1988) Arch. Biochem. Biophys. 260, 569-576). Comparison of this 50 amino acid sequence with periwinkle TDC and D. melanogaster DDCI gave 20 and 32 identical amino acids, respectively. Furthermore, comparison of Cj_ roseus TDC to feline glutamic acid decarboxylase (Kobayashi, Y., Kaufman, D.L. & Tobin, A.J. (1987) J. Neurosci. l_ _r 2768- 2772) showed that 10% of the amino acid residues were identical between these two proteins. This similarity could be extended to 25% on a 396 aa stretch. Mouse ornithine decarboxylase (Kahana, C. & Nathans, D. (1985) Proc. Natl. Acad. Sci. USA 82., 1673-1677) showed a statistically significant (Pearson, W.R. & Lip an, D.J. (1988) Proc. Natl. Acad. Sci. USA 8_5, 2444-2448) 12% amino acid sequence similarity to the plant TDC which also extended throughout the protein sequence. We also found that the sequence Pro-His-Lys, beginning at position 317 in TDC, was identical to the sequence at the pyridoxal phosphate binding sites of D.^ melanogaster DDC (Marques, I.A., & Brodelius, P. (1988) Plant

Physiol. 88., 52-55; Clark, W.C., Pass, P.S., Venkataraman, B. , & Hodgetts, R.B. (1978) Mol. Gen. Genet. 162, 287-297), feline gluta ic acid decarboxylase (Kobayaεhi, Y. , Kaufman, D.L. & Tobin, A.J. (1987) J. Neurosci. 7_, 2768-2772) and pig dopa-decarboxylase (Bossa, F., Martini, F., Barra, D., Borri Voltatorni, C, Minelli, A. & Turano, C, (1977) Biochem. Biophys. Res. Commun. 28., 177-183). In contrast, the AMD protein, whose enzymatic function is unknown, contained the sequence Leu-Hiε-Lys at the pyridoxal phosphate binding domain. The sequence similarity observed between TDC, feline glutamic acid decarboxylase and mouse ornithine decarboxylase also suggests an evolutionary link between these three amino acid decarboxylases.

Structural similarities between TDC and D. melanogaster DDCl proteins were further revealed by comparing their hydropathy profiles (Figure 4). Each value was calculated as the average hydropathic index of a sequence of 9 amino acids and plotted to the middle residue of each sequence. Positive and negative values indicate hydrophobic and hydrophillic regions of the proteins, respectively. Close examination of the alignment of hydrophobic and hydrophillic regions for the two proteins showed a striking match between them, except for the area near the N terminus and the region around TDC residue 225.

Most decarboxylases require for their activity a pyridoxal phosphate co-factor linked to the C amino group of a lysine residue. The observed similarities around the pyridoxal binding site of pig kidney dopa- decarboxylase, I . melanogaster dopa-decarboxylase and feline glutamate decarboxylase with that of periwinkle TDC strongly suggests that lysine 319 of TDC binds pyridoxal phosphate.

The aromatic amino acid decarboxylases of plants, insects and mammals are remarkably similar in

subunit structure, molecular mass and kinetic properties (Maneckjee, R. , & Baylin, S.B. (1983) Biochemistry 22, 6058-6063). Plant aromatic amino acid decarboxylases (Noe, W., Mollenschott, C. & Berline J. (1984) Plant Mol. Biol. 3., pages 281-288; Chappie, C.C.S., (1984) Ph.D. Thesis, University of Guelph, Guelph, Ontario, Canada; Marques, I.A., & Brodelius, P. (1988) Plant Physiol. 88., pages 52-55), in contrast to those from animals, display high substrate specificity for indole or phenolic substrates but not to both. The strong similarity observed between periwinkle TDC and DDCl of D. melanogaster suggests that plant aromatic amino acid decarboxylase specific for tyrosine, phenylalanine or dihydroxyphenylalanine may be structurally similar to TDC and may, therefore, also be evolutionarily related. The recent purification of specific L-tyrosine decarboxylases (Marques, I.A., & Brodelius, P. (1988) Plant Physiol. 88, pages 52-55) to homogeneity should allow cloning of these genes and direct testing of this hypothesis.

TDC BRNA accumulation.

Total poly(A) ⁺ RNAs (1 μg) from six day old C. roseus seedlings and from young leaves of mature plants were run on an agarose/formaldehyde gel and were transferred to nitrocellulose paper. Hybridization was performed with [ ³² P]-labelled pTDC5 insert (sp. act. 1.2 X 10 ⁸ cpm/μG). When total poly(A) ⁺ RNA isolated from six day old.seedlings was probed with a 1.6 kb cDNA fragment isolated from pTDC5, a 1.8 kb mRNA was detected (Figure 5, lane 1). Young leaves from the mature plant also contained a 1.8 kb mRNA (Figure 5, lane 2). A fainter signal corresponding to a transcript of 3.2 kb was also present in both the lanes. This signal could be a precursor form of the TDC mRNA or an unrelated transcript having some sequence similarity to TDC.