~EVIEWS Tie tomato has long been cultivated as a crop plant in its native country of Peru, and yellow, golden, red and white varieties were known as early as 1623 (Ref. 1), indicating a tradition of selective breeding by the indigenous fam~ers. Among the higher plants, the tomato has a relatively small genome (0.74 pg DNA per haploid nucleus) and tomato breeding, genetics and cytogenetics advanced rapidly during the first half of this century, when a large number of fruit and plant characteristics were used to test Mendel's rediscovered laws of inheritance z. Tomato plants can be regenerated readily in tissue culture and the introduction of trans- genes using transformation methods mediated by Agrobacterium tumefaciens has become routine. The total number of available monogenic nmtants has been estimated at 1200 (Ref. 3). Over 1000 RFLPs have been mapped ~ and transgenic lines containing active Ac/Ds transposable elements from maize ~.6 have been produced. The tomato fruit has evolved to perform two basic functions: to protect the developing immature seeds from predators, and to attract many of the same poten- tial predators to consume the fruit and disperse the seeds when mature. Ripening, which involves changes in colour, flavour, texture and aroma, can be con- sidered as a transition from the fomaer to the latter role. The ripening process requires the expression of new genes and affects all cellular compartments. Textural modifications are due to metabolism of polymers in the cell wall, and alteraticms in colour occur as chloroplast constituents are degraded and these organelles are converted to chromoplasts by the accumulation of carotenoids. Many biochemical pathways contribute to flavot,r and aroma, and the whole process is ac- companied by major changes in the nucleus, cytosol and nlitoclaondria. The various facets of ripening are stimulated by ethylene, which is synthesized by mature fruit cells. This p,'ogranmled developmental process makes fruit ripening an excellent system for studying gene regulation. Analysis is not complicated by cell division, and the availability of non-ripening mutants .~ facilitates dissection of the control mechanisms, Construction of cDNA libraries from ripening fruit, together with various screening procedures and the generation of targeted mutations, has led to the iso- lation and identification of novel genes that encode enzymes such as ACC synthase and ACC oxidase (required for ethylene synthesis), polygalacturonase ,which metabolizes cell-wall pectin) and phytoene syn- thase (involved in the production of coloured carotenoids) (reviewed in Ref. 7). This has stimulated new transgenic approaches to the study of gene control in plants, strategies that can also be used in directed breeding programmes. Control of cell.wall metabolism and texture change using antisense genes Expression of antisense RNA in plants, generated by using part of the gene of interest to construct a trans- gene in which the non-coding DNA strand is tran- scribed, has proved a particularly successful mechanism for down-regulating a specific gene, or gene familyS,9. Although the exact mechanism of inhibition remains t Its)3 El',esicr Science I'ubliqmr~ Lid (l'KI tll{~4 - ',)'~2~ 03 S(~.O0 Molecular genetics of tomato fruit ripening RUPERT G. FRAY AND DONALDGRIERSON Tomato ripening is an excellent system for studying control of gene expression in plants. A multiplicity of well-defined biochemical and genetic ct,anges occur in a precise sequence, regulated by a gaseous hormone. Thegeneration of targeted mutations using sense and antisense genes provides a means of manipulating endogenousgene expression, bothfor answeringfundamental questions and for crop improvement. unclear (see below), it provides a means of making directed mutations in transgenic plants, to test the func- tion of specific gene products. One of the earliest uses of antisense technology was in inhibition of the tomato polygalacturonase (PG) en- zyme, which degrades cell wallss.n0. During ripening, PG catalyses the hydrolytic cleavage of 0t-l,4 linkages be- tween adjacent demethylated galactumnic acid residues in cell-wall pectin. PG was believed to play a major role in fruit softening, as its activity correlated with softness of different tomato cultivars and non-ripening mutant# n. A single PG gene appears to be expressed in tomato fruit, but variable glycosylation and association with a 'polygalacturonase converter' protein result in the pro- duction of three isoforms n2. PG expression was inhibited in transgenic plants generated with either a full-length sequence or a 730 bp fragment of the eDNA inverted relative to a constitutive promoter 8.L0, Plants containing a single antisense transgene showed reduced levels of both PG mRNA and enzyme activity; different trans- formants had varying degrees of inhibition, which prob- ably depended on the position of the inserted gene s,L~. While the hemizygous primary transformant had levels of PG enzyme activity as low as 10% those of normal plants, homozygous transgenic progeny produced by selling had less than 1% the activity of the wild type. Progeny that did not inherit the transgene had normal PG levels (Fig. 1). Inhibition specifically affected the tar- get gene; other aspects of ripening, such as production of the red pigment lycopene or the phytohomlone ethylene, were unaffected. The recovery of wild-type progeny by selling primary transformants indicated that no pemlanent alteration or inactivation of the PG gene had occurred. Surprisingly, some control plants that carried the same PG DNA fragment in the sense orientation also had substantially reduced PG mRNA and enzyme levelsVL This phenomenon, which has also been observed in other systems, such as inhibition of flower colour in petunias 9A~, was originally termed co-suppression, but is better described as sense-suppression. A number of mechanisms have been proposed to explain it16,17 and will be discussed in more detail below. In antisense plants that had PG levels less than 1% of nomlal, all isoforms of PG were reduced and the polyuronides derived from cell-wall pectin during ripen- ing remained in their high molecular we'ght form 13. TIGDECEMBER 1993 VOL. 9 NO. 12