WO2000068403A2 - Tapetum-specific promoters - Google Patents

Abstract

The present invention relates to nucleic acid sequences encoding tapetum specific promoters, for use in Artificial Male Sterility systems in plants. In particular, the promoters may be the pMAC2 promoter; the pMAC20 promoter; or promoter sequences which naturally controls the expression of a coding sequence substantially homologous to the MAC2 or MAC20 coding sequences. Also provided are regulatory elements of the promoters; plant cells and plants transformed with the promoter sequences.

Description

Tapetum-Specific Promoters

This invention relates to the application of recombinant DNA technology to plants, for the purpose of achieving male sterility.

The production of hybrids via sexual hybridisation of parents with differing genetic backgrounds is an important practice in modern agriculture. Due to the manifestation of hybrid vigour the offspring are superior to the parents in such characters as yield and disease resistance. In addition, where the parents are extensively homozygous, the resulting offspring are genetically very uniform, and therefore the crop behaves in an equally uniform manner in such important characteristics as germination time, height of growth, susceptibility to disease, flowering time, seed ripening time etc, which greatly improves the efficiency of crop management. For these reasons hybrid seed is attractive to the farmer.

In nature, self-fertilisation is favoured with the production of non-hybrid offspring. Therefore, in order to produce hybrid seed free from contamination with selfed seed, cross-fertilisation is carried out using a variety of mechanical, chemical and genetic methods that prevent self-pollination. This can be achieved in a number of different ways:

(a) by mechanically removing or chemically inactivating the pollen-producing organs of the female parent before they reach maturity; this method has been used for example in maize (corn) and tomato; (b) by using cytoplasmic male sterile (CMS) mutant plants; this method has been used for example in oilseed rape and sunflower;

(c) by using a recessive nuclear male sterile mutant plant; and

(d) by using a dominant nuclear male sterile genetically engineered plant (artificial male sterility or AMS) as described for example in Mariani et al, Nature 347 737- 741 (1990) or in Worrall et al, The Plant Cell 4 759-771 (1992). There are practical difficulties with all of the above. Mechanical male sterilisation is labour intensive, costly and prone to human error, giving a problem of the quality of hybrid seeds. It is practical only for the species where the flower is big enough to be emasculated manually; it is not practical therefore for most cereals. An attempt to overcome this difficulty and reduce costs uses chemical instead of mechanical emasculation. The efficiency of this technique is very dependent on environmental conditions at the time of spraying the gametocide, and leads the seed producer to take a considerable risk each season.

Cytoplasmic male sterility is very convenient, but its use is limited by the availability of the appropriate mutant plant in each species of interest. The loss of cytoplasmic genetic diversity when all breeders use the same cytoplasm in their breeding program can be a serious problem as seen in the US in maize in the 1970's.

The use of recessive nuclear male sterile mutants is not practical. Because the male sterility gene is recessive, maintenance of the male sterile line involves screening the VΛ of male plants out of the ³/ fertile in the selfed progeny of an heterozygous plant. In the absence of a tightly linked selectable or easily screenable marker this is practically impossible.

The use of AMS systems provides a means of avoiding the problems associated with the other methods. AMS gene systems are potentially universal, being limited only to genetically transformable species. It does not rely on the existence of a mutant as in CMS. The maintenance of the male sterile line may be obtained by engineering a dominant male sterility gene linked to a marker gene that allows selection of AMS plants in a population segregating Vz AMS plants. To be practical, this marker is often a herbicide resistance gene.

AMS systems generally make use of tissue specific expression, for instance by utilising promoters/regulatory sequences which drive expression in one or more of those tissues involved in the development of male fertility. For example, the tapetum, which is a specialised cell layer within the anther and which plays a crucial role in the supply of nutrients to the developing microspores. Malfunction of the tapetum is the cause of many types of natural male sterility.

Certain tapetum-specific genes and their promoters have been previously isolated from both dicots and monocots. For example, WO 92/11379 discloses pA3 and pA9, which probably represent the earliest expressed tapetum-specific promoters isolated to date. Monocot genes which are A9-like have also been disclosed. These are sequences whose coding regions, when translated, putatively encode a protein with homology to

A9. Examples of these include the Maize promoter Msfl4 (Wright SY, et al., Plant J. 1993 (1): 41-9.), which is almost identical to, and is therefore probably the same as Ca444 (WO 92/13957); Osg4 from rice (Tsuchiya et al, Plant Mol. Biol., 26(6): 1737- 46); and LH6 and LH7 from lily. In addition, there are several monocot tapetum- specific cDNAs or promoters isolated from monocots that are not A9-like. These are

Ca455 and its promoter pa55 from maize (WO 92/13957); pEl and ρT72 from rice (WO 92/13956); and pOSG6B from rice (Tsuchiya et al, supra).

There is no evidence that any of these promoters can form the basis of an efficient AMS system in monocots, utilising a preferred AMS sterility gene such as PR- glucanase. Moreover, certain promoters such as pA3 and pA9, although efficient in certain dicots such as tomato (WO97/38116), when linked to PR-glucanase only produce a low frequency of complete male sterility in other dicots such as tobacco (Worrall et al, The Plant Cell, 4:759-771 (1992)). We have now identified additional promoters that are more efficient in generating male sterile monocot and dicot plants using a preferred sterility gene such as PR-glucanase, than other promoters previously described.

Thus, in a first aspect, the present invention provides a recombinant or isolated nucleic acid molecule comprising or consisting of a promoter which is: (i) the pMAC2 promoter sequence as shown in figure 7; (ii) the pMAC20 promoter sequence as shown in figure 12; (iii) a promoter controlling expression of a coding sequence which is substantially homologous to those shown in figure 3 or figure 6; or

(iv) a sequence capable of hybridising under stringent conditions to any one of (i), (ii) or (iii).

Such promoters are tapetum specific. That is to say, that in the context of the present invention these promoters primarily drive expression in the tapetum.

In the context of the present invention the term "substantially homologous" means that said sequence has a greater degree of homology with any of the sequences described herein than with prior art nucleic acid sequences.

When comparing nucleic acid sequences for the purposes of determining the degree of homology one can use programs such as BESTFIT and GAP (from the Wisconsin Package™, Genetics Computer Group (GCG) Madison, Wise. USA). BESTFIT, for example, compares two sequences and produces an optimal alignment of the most similar segments using the algorithm of Smith and Waterman (Advances in Applied

Mathematics 2: 482-489, 1981). GAP enables sequences to be aligned along their whole length and finds the optimal alignment by inserting spaces in either sequence as appropriate, using the algorithm of Needleman and Wunsch (J. Mol. Biol. 48: 443- 453, 1970). Suitably, in the context of the present invention when discussing homology of nucleic acid sequences, the comparison is made by alignment of the sequences along their whole length.

Preferably, sequences which have substantial homology have at least 50% sequence homology, desirably at least 75% sequence homology and more desirably at least 90 or at least 95% sequence homology with said sequences. In some cases the sequence homology may be 99% or above.

The skilled person will appreciate that what is important is that any sequence functions as a promoter and will drive expression primarily in the tapetum.

In the context of the present invention, suitable "stringent conditions" are defined as those given in Plant Genetic Transformation and Gene Expression: A laboratory manual, Ed. Draper, J. et al, 1988, Blackwell Scientific Publications, pp252-255, modified as follows: prehybridization, hybridization and washes at 55-65°C, final washes (with 0.5X SSC, 0.1% SDS) omitted.

In addition, it is possible to derive essential regulatory elements from the promoters provided herein. Thus, those elements of the promoter sequence responsible for both its function as a promoter and, more importantly, its tapetum specificity, can be isolated and incoφorated into nucleic acid molecules which, although not falling within the definitions (i) to (iv) above, nonetheless still function in an equivalent manner.

Therefore, in a second aspect, the present invention provides a recombinant or isolated nucleic acid molecule comprising or consisting of one or more regulatory elements derived from any one of the sequences (i) to (iv) capable of driving expression in a tapetum specific manner.

In addition, the pMac2 putative protein possesses a signal peptide which targets the protein for secretion in the endoplasmic reticulum (see example 5), preventing access of MAC2 to the tapetal cell ribosomes. Removal of the signal peptide allows the Mac 2 protein to accumulate in the cytosol and inactivate the ribosomes causing cell death. Thus, a MAC2 protein lacking a signal peptide can be used as a cellular ablator. Thus, in a further aspect, the present invention provides a recombinant or isolated nucleic acid molecule encoding a MAC2 protein lacking its natural signal peptide.

Such a nucleic acid can be obtained by means of PCR amplification of sequence shown in figure 3 using suitable primers having the sequence: 5 ' CCCATGGCCTCCACCGCCTATCC 3 '

5' GCCGCGGTAATTACCAGTATCTACTTCC 3'

or a sequence which hybridises thereto under stringent conditions. Such primers form an additional aspect of the invention.

Suitably, the nucleic acid molecule in the above-noted aspects of the invention is a DNA molecule.

The tapetum specific promoters of the invention find use in AMS systems. Thus, they may be used to drive the expression of a variety of sterility DNA sequences which code for RNAs proteins or polypeptides which bring about the failure of mechanisms to produce viable male gametes. A number of classes and particular examples of male- sterility sequences are preferred.

For example, the male sterility DNA may encode a lytic enzyme. The lytic enzyme may cause lysis of one or more biologically important molecules such as macromolecules including nucleic acid, protein (or glycoprotein), carbohydrate and in some circumstances lipid.

Ribonuclease (such as Rnase Tl) and barnase are examples of enzymes which cause lysis of RNA. Glucanase is an example of an enzyme which causes lysis of a carbohydrate. The enzyme callase (a β(l,3)-glucanase) is naturally produced in anthers where it functions to release the young microspores from a protective coat of poly- glucan (callose) laid down before meiosis. The appearance of the enzyme activity is developmentally regulated to coincide with the correct stage of microspore development.

One advantage of using glucanase as a male sterility DNA is that it is less prone to potential problems of ectopic expression. In certain environmental conditions or at a particular developmental stage it is possible that transgenes will be expressed at low levels ectopically. This expression may be due to the activation of genes and promoters surrounding the transgene or the expression of transactivators that bind in the vicinity of the transgene (position effect). It is not predictable whether a particular transgene will be ectopically activated in a given environmental condition or developmental stage given that the genomic sequence surrounding each transgene may be unique. This is since current transformation technologies result in an unpredictable integration of the transgene into the genome. Such problems of unpredictable transgene expression are particularly serious with highly active non cell-specific cytotoxic transgenes such as barnase. It has been claimed that a single barnase protein is sufficient to cause cell death. Extensive field trialling of barnase transformants will eliminate the majority of transforments where such ectopic expression occurs. However this is laborious and there is always a chance that conditions that cause ectopic expression are not encountered prior to commercialisation of the plant line.

Such a problem of eco topic barnase expression can minimised by 'constitutive' expression of the inhibitor of barnase, barstar such that it is expressed in all cell types apart from the target cell type (eg the anther tapetum). However constitutive expression of barstar may be undesirable since all plant parts consumed now contain barstar protein. Also, not all cell types may have sufficient expression of barstar to be protected.

Glucanase and barnase represent preferred embodiments of a lytic enzyme for use with the nucleic acid molecules of the invention.

A further advantage is that the PR glucanase system is more 'natural' than the barnase system. Premature expression of PR-glucanase mimics or phenocopies natural male sterile sorghum and petunia mutants (Worrall et al, (1992) Plant Cell. 4, 759-771).

Male sterility DNA does not have to encode a lytic enzyme. Other examples of male sterility DNA encode enzymes which catalyse the synthesis of phytohormones, such as isopentyl tranferase, which is involved in cytokinin synthesis, and one or more of the enzymes involved in the synthesis of auxin. A further example of a male sterility DNA encodes an RNA enzyme (known as a ribozyme) capable of highly specific cleavage against a given target sequence (Haseloff and Gerlach, Nature 334 585-591 (1988)).

Other male sterility DNAs include antisense sequences. Introducing the coding region of a gene in the reverse orientation to that found in nature can result in the down- regulation of the gene and hence the production of less or indeed none of the gene product. The RNA transcribed from antisense DNA is capable of binding to, and destroying the function of, a sense RNA version of the sequence normally found in the cell, thereby disrupting function.

It is not crucial for antisense DNA to be solely transcribed at the time when the natural sense transcript is being produced. Antisense RNA will in general only bind when its sense complementary strand is present, so will only have its toxic effect when the sense strand is transcribed.

In a further aspect, the present invention provides a set of primers suitable for PCR amplification of the promoter region of the maize MAC2 gene and having the following sequence:

3' GGTCGACTTGGAATAATTTAAGTTGT 5' 3' GATCACCATGGTACTACTCCAC 5'

or having a sequence which hybridises thereto under stringent conditions. The primers used may be used to amplify a promoter from maize genomic DNA. A person skilled in the art will appreciate though that the same primers may be suitable for PCR amplification from other monocots such as rice, wheat and lily.

DNA in accordance with the invention may be in the form of a vector. Such vectors form an additional aspect of the invention. The vector may be, for example, a plasmid, cosmid or phage. Vectors will frequently include one or more selectable markers to enable selection of cells transfected or transformed and to enable the selection of cells harbouring vectors incorporating heterologous DNA. Examples of such a marker gene include antibiotic resistance genes (EP-A-0242246) and glucuronidase (GUS) expression genes (EP-A-0344029). Expression of the marker gene is preferably controlled by a second promoter which allows expression in cells other than the tapetum, thus allowing selection of cells or tissue containing the marker at any stage of regeneration of the plant. The preferred second promoter is derived from the gene which encodes the 35S subunit of Cauliflower Mosiac Virus (CaMV) coat protein.

However, any other suitable second promoter could be used.

Cloning vectors may be introduced into E. Coli or another suitable host which facilitate their manipulation. Nucleic acid sequences in accordance with the invention may be introduced into plant cells by any suitable means. Thus, according to yet a further aspect of the invention, there is provided a plant cell including a nucleic acid molecule in accordance with the invention. Preferably, the plant cell with be transgenic.

Nucleic acid may be transformed into plant cells using a disarmed Ti-plasmid vector and carried by agrobacterium by procedures known in the art, for example as described in EP-A0117618 and EP-A-0270822. Alternatively the foreign nucleic acid could be introduced directly into plant cells using a particle gun. This method may be preferred for example when the recipient plant is a monocot. A whole plant can be regenerated from a single transformed plant cell. Thus, in a further aspect the present invention provides transgenic plants (or parts of them, such as propagating material, which may also be transgenic) including nucleic acid sequences in accordance with the invention. The regeneration can proceed by known methods. When the transformed plant flowers it can be seen to be male sterile by the inability to produce viable pollen. Where pollen is produced it can be confirmed to be non- viable by the inability to effect seed set on a recipient plant.

In final aspects, the present invention provides:

(a) the use of the nucleic acid molecules of the invention in transforming a host cell, preferably a plant cell, and more preferably a monocot plant cell; and

(b) the use of the nucleic acid molecules of the invention in the production of a male sterile plant.

Preferred features for each aspect are as for each other aspect mutatis mutandis.

The invention will now be described by the way of the following examples, which should not be construed as in any way limiting the scope of the invention. The examples refer to the accompanying drawings in which:

Figure 1 - shows the RT-PCR data for Mac2. The maize RNA source used in each reaction is indicated below the gel. -RT reactions are control reactions lacking reverse transcriptase. +RT reactions contain reverse transcriptase. lkb Ladder (Gibco BRL) is loaded as a DNA size marker either side of lanes containing RT-PCR products.

Figure 2 - shows the in situ data for Mac2. Sections of a male floret were probed with labelled sense Mac2 as a control (A) or with labelled antisense Mac2 (B). The tapetum can been seen to be heavily labelled (black circles) only with the sense probe. Figure 3 - shows the Mac2 cDNA sequence. The predicted amino acid sequence of the Mac2 protein is shown underneath the DNA sequence. Primers used for TAIL-PCR are shown above the cDNA sequence.

Figure 4 - shows the alignment of Mac2 with type 1-RIPs. Genebank accession numbers for sequences are:- Maize RIP3 (M83926), Barley RIPl (M62905, M36990) and Wheat Tritin (D13795).

Figure 5 - shows the RT-PCR data for Mac20. The maize RNA source used in each reaction is indicated below the gel. -RT reactions are control reactions lacking reverse transcriptase. +RT reactions contain reverse transcriptase. lkb Ladder (Gibco BRL) is loaded as a DNA size marker either side of lanes containing RT-PCR products.

Figure 6 - shows the Mac20 cDNA sequence. The predicted amino acid sequence of the Mac20 protein is shown underneath the DNA sequence.

Figure 7 - shows the TAIL-PCR sequence of Mac2T-3. The AD1 and Mac2t3 primers used to PCR this sequence are shown above the DNA sequence. Nucleotide differences compared to Mac2T-l and other TAIL-PCR products are shown beneath the DNA sequence ('-'= missing nucleotide). The amino acid sequence of the putative Mac2 peptide is shown underneath the DNA sequence. The sequences of primers used to PCR the promoter are shown in lower case above the TAIL-PCR sequence. Two sequence changes in the PCR product compared to the TAIL-PCR sequences are indicated above the TAIL-PCR sequence.

Figure 8 - is a schematic diagram of pMac2-GUSbin

Figure 9 A - shows a schematic diagram of pMac2 -barnase bin. Figures 9B-D show thin transverse sections of tobacco pMac2 -barnase anthers, viewed by light microscopy. Figure 9B shows a single pollen sac from a wild type plant; Figure 9C shows a single pollen sac from a plant carrying the Mac2-barnase construct; and Figure 9D shows a low magnification view of anther section from a Mac2 -barnase plant in which all four pollen sacs are visibly collapsed.

Figure 10 A - shows a schematic diagram of the pMac2 -PR-glucanase bin. Figures 10B and IOC show fluorescence micrographs of wild type and transformed tobacco microspore tetrads respectively, stained for callose with aniline blue. Figures 10D and E show scanning electron micrographs of tobacco microspores/pollen from a wild type and transformed plants, respectively.

Figure 11 - shows A) Partial DNA sequence of pMac2 showing where the ΔMac2F and ΔMac2R primers bind; and B) a schematic diagram of pA9-ΔMac2.

Figure 12 - shows the pMac20 genomic sequence with the putative Mac20 peptide sequence shown below the DNA sequence.

Example 1- Isolation of maize cDNAs representing transcripts of tapetum- specific genes expressed prior to microspore release

In most plant species the stage of microsporogenesis within anthers is strongly correlated with the length of floral buds and anthers (Scott et al. (1991) Plant Mol Biol. 17, 195-207). Microscopic analysis of anthers from the maize variety A188 established the following correlation between anther length and developmental stage:-

0-2 mm = prior to tetrad stage. 2-4 mm = tetrads, microspore release and free microspores. 4-6 mm = free microspores.

In order to utilise a male sterility system based on the premature degradation of callose (β(l,3)-linked glucan) formed in meiosis it is necessary to isolate the promoter of a gene that is highly expressed during the developmental phase where callose surrounds the microsporocytes, microsporocytes undergoing meiosis and the tetrads. Thus it is essential that this promoter is expressed prior to microspore release (MR). The promoter should also be expressed in the tapetum and / or microsporocytes. Such promoters are, of course, potentially useful in male sterility systems that are not based on premature callose degradation.

Consequently, lOOmg of anthers containing microsporocytes prior to microspore release (<MR) were dissected from male florets of maize variety A188, and 13μg of total RNA isolated using a Rneasy Plant Mini Kit (Qiagen) according to manufacturers instructions. 840 ng of total RNA was used to construct a cDNA library in Lambda gtl 1 using the Capfmder PCR kit (Clonetech) according to manufacturers instructions.

To provide probes to differentially screen this <MR cDNA library, cDNA was prepared from <MR anther, 0-2 mm anther, A9-barnase anther and seedling RNA and labelled with DIG using the DIG-High Prime kit (Boehringer Mannheim). This cDNA was used to screen plaque lifts from the library according to methods provided in the

Dig User's Guide supplied by Boehringer Mannheim. First the <MR cDNA library was differentially screened with maize seedling and <MR anther probes. 42 plaques that potentially represented anther-specific messages were rescreened against seedling, anther, and also against A9-Barnase anther probes (transformation of maize with the chimeric gene A9-Barnase (Paul et al, Plant Molecular Biology 19 611-622 (1992)) results in male sterile plants due to the ablation of the anther tapetum. Thus if the plaque represents a message expressed in the anther tapetum the hybridisation signal should be absent or reduced. This secondary screening showed that 26 plaques represented messages present in fertile anther RNA but absent in seedling and barnase anther RNA.

The 26 cDNAs were PCRed out from Lambda gtl 1, recloned into pGEM-T (Promega) and the DNA sequence determined. This analysis showed that the 26 cDNAs comprised of 8 groups. Two primers were designed to each of the 8 groups and RT- PCR used to determine the spatial pattern of expression and confirm the spatial expression of the clones. The RNA used in this analysis was from <MR, A9 barnase, 0-2mm, 2-4mm and 4-6mm anthers. From this screening two clones, Mac2 and Mac20 were selected as being potentially representing strongly expressed tapetum-specific and / or microsporocyte-specific genes expressed prior to microspore release.

ϊ) Mac2

This cDNA was represented 8 times in the 26 cDNAs resulting from secondary screening. RT-PRC data (figure 1) shows Mac2 mRNA is abundant in 0-2mm anthers, is also present at reduced levels in older anthers (2-4mm) which are largely 'post microspore release' but is absent in older 4-6mm anthers. Surprisingly Mac2 mRNA is also present in A9-Barnase anthers (figure 1, lane 8) in which the tapetum is ablated.

In situ analysis of sectioned anthers was performed performed essentially as described in the Boehringer Manheim Non-Radioactive In Situ Hybridisation Manual. 15μm and 30 μm sections were cut using a cryostat (Shandon). RNA probes were labelled using the DIG RNA labelling kit (Boehringer Manheim) according to manufacturers instructions and hybridisation was performed overnight at 42°C. Results (figure 2) show that Mac2 mRNA is present in the tapetum of maize anthers and is absent in the anther wall. Given that Mac2 is tapetum-specific the finding that Mac2 mRNA is also present in A9-Barnase anthers indicates that in maize the Arabidopsis thaliana A9 - tapetum-specific promoter is expressed after the appearance of Mac2 mRNA. This suggests that the promoter of Mac2 will be expressed earlier than that of pA9 and thus will be superior for the premature expression of β(l-3) glucanase.

DNA sequence analysis shows that the longest Mac2 cDNA putatively encodes a 297 amino acid protein with a putative signal peptide predicted by the program Signal P

(Neilson et al., (1997) Protein engineering 10 1,6) (figure 3). Database searches show that the Mac2 putative protein shows low homology to type 1-ribosome inactivating proteins (RIPs) from maize, wheat and barley that are expressed in seeds (figure 4). The best homology is with Maize RIP3 with 17% identity at the protein level (Clustal V score of 14.1). Southern analysis showed that the Mac2 cDNA hybridises to 3 or 4 bands in maize genomic DNA cut with EcoRI. Hybridisation was perfomed at 65 °C using a Digoxigenin-labelled Mac2 probe as described in Protocols for Nucleic Acid

Analysis by Nonradioactive Probes , Methods in Molecular Biology Vol 28 (1994) Ed. Isaac PG Humana Press Inc.

in Mac20 This cDNA was represented once in the 26 cDNAs resulting from secondary screening. RT-PCR data (figure 5) shows Mac2 mRNA is abundant in 0-2mm anthers, is also present at very reduced levels in older anthers (2-4mm) which are largely 'post microspore release' and is absent in older 4-6mm anthers. A faint signal is also present in A9-Barnase anthers (figure 5, lane 8) again indicating that Mac20 mRNA could be expressed prior to activity of pA9 in maize.

Sequence analysis shows that the Mac20 cDNA putatively encodes a 103 amino acid protein with a putative signal peptide predicted by the program signal P (Neilson et al, (1997) Protein engineering 10 1,6) (figure 6). The Mac20 DNA sequence and the putative Mac20 protein show no significant matches in DNA and protein databases.

Southern analysis showed that the Mac20 cDNA hybridises to approximately 5 bands in maize BamHI and EcoRI-cut genomic DNA and with 3 bands in Hindlll-cut wheat genomic DNA. Hybridisation was perfomed at 65°C using a Digoxigenin-labelled probe as described in Protocols for Nucleic Acid Analysis by Nonradioactive Probes , Methods in Molecular Biology Vol 28 (1994) Ed. Isaac PG Humana Press Inc.

Example 2 - Isolation and characterisation of the promoter region of the maize Mac2 gene.

TAIL-PCR was used to isolate sequence 5' of the Mac2 gene from maize genomic DNA. TAIL-PCR was performed according to the method of Liu et al., Plant Journal 8 457-463 (1995). Three Mac2 specific primers were designed:-

Mac2tl (5'-AGT CAT CAA TGG CTA TGG CCA G-3*), which binds at positions 343-322bp of the Mac2 cDNA (figure 3); Mac2t2 (5'-CGT ATC TTT GCA TGA CCT CTT GG-3'), which binds at 232-210bp (figure 3) and Mac2t3 (5'-GTG GAG GTG

CAA AAC AGC AGG T-3') which binds at 103-82 bp (figure 3). These primers were used individually with the degenerate primer ADl (5'-NTCGASTWTSGWGTT-3') in three rounds of PCR starting with the combination of ADl with the most 3' Mac2 primer and finishing with ADl plus the most 5' Mac2 primer.

1.3 kb TAIL-PCR products were cloned into pGEM-T. 2 clones, Mac2T-l and Mac2T-3, were completely sequenced and 6 clones Mac2T-2/4/6/7/8/9 partially sequenced. The consensus sequence obtained contains uncertainties at 5 positions:- 86bp (T or C), 347bp (7 or 8 A residues), 555bp (8 or 9 A residues), 665bp (G or A) and at 754bp (G or A) (figure 7). Given this sequence two primers (figure 7) were designed to PCR out a 1.2kb putative promoter region from maize genomic DNA:- 3' GGTCGACTTGGAATAATTTAAGTTGT 5' = new Mac2P5' 3' GATCACCATGGTACTACTCCAC 5' = Mac2P3'N The 5' primer introduced a Sal site and the 3' primer an Ncol site (around the initiating ' ATG' codon of the putative Mac2 protein) to facilitate subsequent cloning.

PCR products were cloned into pGEM-T (Promega) and sequenced. The clone containing the sequence most similar to the concensus TAIL-PCR sequence was named pMac2Prom. This sequence was identical to that of the TAIL-PCR concensus except for a T to C change at position 762bp and the deletion of a T residue at position 893bp (figure 7).

To characterise the spatial and temporal expression pattern directed by the 1.2kb Mac2 promoter region (pMac2) this region was linked to the reporter gene β-glucuronidase

(Jefferson et al, EMBO J 6 3901 (1987)) and transformed into tobacco and maize. The Sail pMac2 fragment was excised from pMac2Prom and cloned into the Sail site of pBluescript KS+ (Stratagene) forming pKSMac2Prom, such that the Ncol site of pMac2 was adjacent to the Xhol site of pBluescript KS+. An Ncol-Xhol fragment containing a GUS intron+CaMV polyadenylation sequence was cloned from pDH68 (WO99/13089) between the Ncol and Xhol sites of pKSMac2Prom forming pMac2- GUS. The pMac2-GUS-CaMVpolyA region was then excised from pMac2-GUS as a

Hindlll, Xhol fragment and cloned between the Hindlll, Sail sites of the binary vector pBinl9 (Bevan MW, (1984) Nucleic Acids Research 12, 8711-8721) forming pMac2- GUS bin (figure 8).

pMac2-GUSbin was transfened into the agrobacterial strain LBA4404 and transformed into N.tabacum using a leaf-based agrobacterial technique. Transformed N.tabacum plants show GUS expression in the anther tapetum. Expression commences prior to microspore release.

pMac2-GUS was also transformed into maize using a standard particle bombardment method. Transformed maize plants exhibit GUS expression in anthers of length 0- 2mm with expression localised to the tapetum. Expression of pMac2-GUS in the maize tapetum is much stronger than observed than for pA3-GUS, pA6-GUS and pA9-GUS transformed maize. This suggests that pMac2 is superior in maize to the A.thaliana A3 (WO 92/1179), A9 (WO 92/1179) and A6 (WO 93/02197) promoters for applications that require high-level expression in the tapetum.

Example 3 - Construction of a chimeric pMac2-barnase gene and its expression in transgenic plants

To demonstrate the utility of the pMac2 promoter in the production of male sterile plants and to further characterise the spatial and temporal pMac2 expression pattern, pMac2 was linked to the ribonuclease barnase (Hartley, RW J.Mol.Biol (1988) 202, 913-915). The Sail, Ncol 1224bp Mac2 promoter fragment was excised from pMac2Prom and cloned between the Sail and Ncol sites of pWP127 (Paul et al, Plant Molecular Biology 19 611-622 (1992)). The resulting plasmid (pMac2-barnase) was digested with Sstl and EcoRV and the pMac2-barnase-CaMVpolyA chimeric gene transferred into Sstl, Smal sites of the binary vector pBinl9 forming pMac2 -barnase bin (figure 9).

pMac2-barnase bin was transferred into the agrobacterial strain LBA4404 and transformed into N.tabacum. Transformed N.tabacum plants are phenotypically wild type apart from male sterility due to the ablation of the tapetum. Figure 9 show the results of microscopic analysis of anthers from wild type and transformed tabacco plants. In the wild type (Figure 9B) the pollen sacs are seen to have a well developed tapetum (Figure 9B, T) and a tetrad stage microspores (Figure 9B, M). In the transformed plant, however, the pollen sac is collapsed and lacks a clearly defined tapetum and microspores (Figure 9C). This plant was male sterile but female fertile.

pMac2-barnase was also transformed into maize using a standard particle bombardment method. All 8 transformed plants were transformed maize plants are phenotypically wild type apart from male sterility due to the ablation of the tapetum.

Example 4 - Construction of a chimeric pMac2-PR-glucanase gene and its expression in transgenic plants

In order to create male sterile plants by the premature dissolution of callose pMac2 was linked to the PR glucanase gene (Worrall et al, (1992) Plant Cell. 4, 759- 771). The Sail, Ncol 1224bp Mac2 promoter fragment was excised from pMac2Prom and cloned between the Sail and Ncol sites of pDW80PR (Worrall et al, (1992) Plant Cell. 4, 759-771). The resulting plasmid (pMac2-PRG) was digested with Sstl and EcoRV and the pMac2-PR Glucanase-CaMVpolyA chimeric gene transferred into Sstl, Smal - cut pBin 19 forming pMac2-PRG bin (figure 10). pMac2-PRG bin was transfened into the agrobacterial strain LBA4404 and transformed into N.tabacum. Transformed N.tabacum plants are phenotypically wild type apart from male sterility. The results of microscopic examination of anthers from transformed plants is shown in Figure 10. In the wild type, (Figure 10B), the tetrads of the anthers are separate and have a regular morphology. Each tetrad has highly fluorescent callosic cross walls and callosic outer walls. In contrast, the tetrads extruded from anthers of transformed plants are clumped together, as judged by the aniline blue induced fluorescence. The tetrads lack both callosic cross walls and outer walls (Figure IOC).

The scanning electron micrographs of tobacco microspores/pollen from both wild type and pMAC2-PR glucanase plants are shown in figures 10D and E, respectively. The wild type is seen to have well developed pollen grains with uniform morphology, each grain being separate and plump with a smooth pollen wall with frequent small holes. In contrast, the microspores of the transformed plant are very irregular, and appear to be fused together in small clumps. They also appear collapsed with very irregular pollen walls.

The severely reduced levels of callose sunounding cells undergoing meiosis and sunounding tetrads leads to subsequent death of microspores and male sterility. The frequency of complete male sterility was much higher than that observed using pA9- PR glucanase (Worcall et al, supra). Cytologically the phenotype of fused tetrads resembles that produced by pA9-PR glucanase in tomato. In this species pA9-PR glucanase produces a high frequency of complete sterility.

pMac2-PRG was transformed into maize using a standard particle bombardment method. Transformed maize plants are phenotypically wild type apart from male sterility due to reduced callose levels in the anther prior to microspore release.

Example 5 - Expression of a modified Mac2 protein in transgenic plants. The Mac2 putative protein shows homology to type 1-RIPs (Example 1). The possession of a signal peptide in the Mac2 protein targets the Mac2 protein for secretion into the endoplastic reticulum thus preventing access of Mac2 to the tapetal cell ribosomes. Removal of this signal peptide will allow the Mac2 protein to accumulate in the cytosol and inactivate the ribosomes causing cell death. Thus a Mac2 protein lacking a signal peptide (ΔMac2) is generally useful as a cellular ablator and provides a plant derived alternative to the bacterial barnase protein. To demonstrate the utility of this modified Mac2 protein it is used to generate male sterile plants by tapetal cell ablation. Tapetal-specific promoters that could be used include pMac2 itself or promoters isolated from Arabidopsis such as pA9 (Paul et al, Plant Molecular Biology 19 611-622 (1992)). A pA9-ΔMac2 fusion is constructed as follows. The following primers were used to PCR a Mac2 region encoding a Mac2 protein lacking the signal peptide (figurel la):- 5' CCCATGGCCTCCACCGCCTATCC 3' ΔMac2F and

5' GCCGCGGTAATTACCAGTATCTACTTCC 3' ΔMac2R

The 843bp PCR product was digested with Ncol and Sstll and cloned between the Ncol and Sstll sites of pWP112 (WO92/11379 ) forming pA9-ΔMac2 (figure l ib). The pA9-ΔMac2-CaMV polyA region of pA9-ΔMac2 was excised as a Hindlll, EcoRV fragment and cloned between the Hindlll and Smal sites of pBinl9 forming pA9-ΔMac2bin.

pA9-ΔMac2bin was transfened into the agrobacterial strain LBA4404 and transformed into N.tabacum. Transformed N.tabacum plants are phenotypically wild type apart from male sterility due to the ablation of the tapetum.

pA9-ΔMac2 was also transformed into maize using a standard particle bombardment method. Transformed maize plants are phenotypically wild type apart from male sterility due to the ablation of the tapetum. Example 6 - Isolation and characterisation of the promoter region of a maize Mac20 gene.

Inverse PCR (IPCR) was used to isolate the promoter region of a Mac20 gene. Maize genomic DNA was digested with Ball, which cleaves inside of Mac20 and recircularised with T4 DNA ligase. Divergent primer pairs which bind within Mac20 were used to PCR out a Mac20 fragment. Sequence analysis (figure 12) showed that this fragment was 96% identical to Mac20 within the region of overlap. RT-PCR analysis, with primers specific to the coding regions of the IPCR Mac20 gene, showed that the IPCR Mac20 gene has the same temporal and spatial expression pattern as the

Mac2 gene. A longer promoter region is then obtained from maize genomic DNA by TAIL PCR essentially as described in Example 2.

Claims

1. A recombinant or isolated nucleic acid molecule comprising or consisting of a promoter which is:

(i) the pMAC2 promoter sequence as shown in figure 7; (ii) the pMAC20 promoter sequence as shown in figure 12; (iii) a promoter controlling expression of a coding sequence which is substantially homologous to those shown in figure 3 or figure 6; or

(iv) a sequence capable of hybridising under stringent conditions to any one of (i), (ii) or (iii).

2. A recombinant or isolated nucleic acid molecule comprising or consisting of one or more regulatory elements derived from any one of the sequences (i) to (iv) capable of driving expression in a tapetum specific manner.

3. A recombinant or isolated nucleic acid molecule encoding a MAC2 protein lacking its natural signal peptide.

4. A pair of primers having the sequence: 5' CCCATGGCCTCCACCGCCTATCC 3'

5' GCCGCGGTAATTACCAGTATCTACTTCC 3'

or a sequence which hybridises thereto under stringent conditions.

5. A nucleic acid molecule as claimed in any one of claims 1 to 3 which is a DNA molecule.

6. A nucleic acid molecule as claimed in claim 1, claim2 or claim 5 which further comprises nucleic acid, which when expressed results in male sterility in a plant.

7. A nucleic acid molecule as claimed in claim 6 wherein the male sterility nucleic acid codes for a lytic enzyme.

8. A nucleic acid molecule as claimed in claim 7 wherein the lytic enzyme is glucanase or barnase.

9. A nucleic acid as claimed in claim 8 wherein the glucanase is PR-glucanase.

10. A pair of primers having the following sequence: 3' GGTCGACTTGGAATAATTTAAGTTGT 5'

3' GATCACCATGGTACTACTCCAC 5'

or having a sequence which hybridises thereto under stringent conditions.

11. A vector comprising a nucleic acid molecule as defined in any one of claims 1 to 3 or claims 5 to 9.

12. A host cell transformed with a vector as defined in claim 11.

13. A host cell as claimed in claim 12 which is a plant cell.

14. A host cell as claimed in claim 13 which is a monocot plant cell.

15. A plant comprising one or more plant cells as defined in claim 13 or claim 14.

16. The use of a nucleic acid molecule as defined in any one of claims 1 to 9 in transforming a host cell, preferably a plant cell and more preferably a monocot plant cell.

17. The use of a nucleic acid molecule as defined in any one of claims 1 to 9 in the production of a male sterile plant.

18. The use as claimed in claim 17 wherein the male sterile plant is a monocot, eg maize, rice, lily or wheat.