We have cloned and sequenced four pollen-specific cDNAs. None of the clones are complete at their 5' ends. One of the clones shows significant homology to the tomato fruit-ripening polygalacturonase and to a pollen-specific polygalacturonase from Oenothera. The other three clones have no significant homologies to any reported sequence.

The 2a isoenzyme of tomato polygalacturonase was purified from ripe fruit and characterised. The N-terminal amino acid sequence of the protein was determined in order to identify polygalacturonase cDNA clones. The nucleotide sequence of a ripening-related cDNA (pTOM 6) was determined and found to encode the N-terminal sequence of mature polygalacturonase 2a. The complete open reading frame encodes a polypeptide of molecular weight 50,051, including a putative pre-sequence of 71 amino acids.

We have amplified a 1.5-kb PG cDNA including its complete open reading frame, using two primers specific to PG gene and PCR technique. Its fidelity and correctness were confirmed by restriction mapping and partial DNA sequencing. The PG gene was then inserted into a binary vector, pBin437, in an inverted orientation between the 35S promoter with duplicated enhancers and the Nos3' transcriptional termination sequence, and the resulting plasmid pBPG was introduced into LBA 4404. The A. tumefaciens harboring pBPG was used to transform tomato plants. Sixty kanamycin-resistant transformants were selected. PCR detection showed two-third of the transformed plants tested (30 plants) contain the exogenous PG gene. The PG activity assay and SDS-PAGE analysis of proteins from ripening fruit cell wall showed that the PG activity in transgenic tomato plants were inhibited in different extent. One of them, plant 3, gave only 7% PG activity compared with the control (100%). These results suggest that the expression of the anti-PG gene has effectively inhibited the expression of endogenous PG gene. The effect of low-level PG activity on tomato fruit ripening and storage is being investigated.