University of Groningen

TomatoHille, Jacques; Koornneef, Maarten; Ramanna, M.S.; Zabel, Pim

Published in:Euphytica

DOI:10.1007/bf00042609

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite fromit. Please check the document version below.

Document VersionPublisher's PDF, also known as Version of record

Publication date:1989

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):Hille, J., Koornneef, M., Ramanna, M. S., & Zabel, P. (1989). Tomato: a crop species amenable toimprovement by cellular and molecular methods. Euphytica, 42(1). https://doi.org/10.1007/bf00042609

CopyrightOther than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of theauthor(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons thenumber of authors shown on this cover page is limited to 10 maximum.

Download date: 09-06-2021

https://doi.org/10.1007/bf00042609https://research.rug.nl/en/publications/tomato(c00bea5a-b283-48dc-adff-0dbd8f1e15e2).htmlhttps://doi.org/10.1007/bf00042609

Euphytica 42: 1 -23, 1989 .©1989 Kluwer Academic Publishers . Printed in the Netherlands .

Tomato: a crop species amenable to improvement by cellular and molecular

methods

Jacques Hille', Maarten Koornneef', M .S. Ramanna3 and Pim Zabel''Free University, Department of Genetics, De Boelelaan 1087, 1081 HV Amsterdam, The Netherlands ;'Agricultural University, Department of Genetics, Dreijenlaan 2, 6703 HA Wageningen, The Netherlands ;3Agricultural University, Department of Plant Breeding, P .O. Box 386, 6700 AJ Wageningen,The Netherlands; 'Agricultural University, Department of Molecular Biology, Dreijenlaan 3,6703 HA Wageningen, The Netherlands

Received 19 September 1988 ; accepted in revised form 15 November 1988

Key words: Lycopersicon esculentum, tomato, genetic transformation, protoplast fusion,restriction fragment length polymorphisms (RFLP's), transposons

Summary

Tomato is a crop plant with a relatively small DNA content per haploid genome and a well developedgenetics . Plant regeneration from explants and protoplasts is feasable which led to the development ofefficient transformation procedures .

In view of the current data, the isolation of useful mutants at the cellular level probably will be of limitedvalue in the genetic improvement of tomato . Protoplast fusion may lead to novel combinations of organelleand nuclear DNA (cybrids), whereas this technique also provides a means of introducing genetic informationfrom alien species into tomato . Important developments have come from molecular approaches . Followingthe construction of an RFLP map, these RFLP markers can be used in tomato to tag quantitative traits bredin from related species . Both RFLP's and transposons are in the process of being used to clone desired genesfor which no gene products are known . Cloned genes can be introduced and potentially improve specificproperties of tomato especially those controlled by single genes . Recent results suggest that, in principle,phenotypic mutants can be created for cloned and characterized genes and will prove their value in furtherimproving the cultivated tomato .

Introduction

Although tomato was cultivated by Indians in Mex-ico and introduced in Europe as early as the 16thcentury, it lasted until the second half of the nine-teenth century before Lycopersicon esculentum,became generally recognized as a highly valuableand nutritious food crop . Nowadays, tomato is oneof the major vegetable crops throughout the world .It is grown in both tropical, sub-tropical and tem-

perate areas with an annual production of approxi-mately 50 million metric tons (see Atherton &Rudich, 1986) .The commercial tomato (L . esculentum Mill .,

2n = 2x = 24) is a member of a relatively smallgenus, Lycopersicon, within the large family Sola-naceae. This genus consists of nine species withPeru as its main centre of diversity . Phylogeneticrelationships on the basis of morphological andcrossability studies reveal two species complexes,

2

viz. the 'esculentum complex' consisting of L. escu-lentum, L . pimpinellifolium, L . cheesmanii, L. hir-sutum, L. pennellii, L . chmielewskii and L. par-viflorum and the 'peruvianum complex' with itsmembers L. peruvianum and L. chilense . The spe-cies of both complexes, which show a high degreeof homosequentiality in their chromosomes, haveserved as an invaluable source of genetic variationand disease resistance genes in improving the culti-vated tomato (Rick, 1982; Stevens & Rick, 1986 ;Atherton & Rudich, 1986) .

As tomato turned out to be highly amenable tobasic genetic and cytogenetic analysis, tomatobreeding and genetics have gone hand in hand dur-ing the first half of this century. Thus, a wealth ofgenetic information and knowledge has been gath-ered on tomato that is now starting to be assimilat-ed by cell- and molecular biologists/geneticists .With the advent of DNA recombinant technology,restriction fragment length polymorphism(RFLP)-mapping, effective DNA transformationprocedures and cell- and tissue culture techniques,tomato has become also amenable to cellular andmolecular analysis and manipulation . As a result,new trends in basic tomato research emerged in theearly eighties, which promise to be of enormouspotential also to practical breeding . Unfortunately,many of the molecular and cellular data and con-cepts gathered from research laboratories fail toreach those involved in commercial tomato pro-duction and vice versa . Accordingly, the integra-tion of newly developed techniques and conceptsinto classical breeding proceeds slowly .

In this review, an attempt is made to condensethe various lines of molecular and cellular tomatoresearch that may be pertinent to breeding, butremained so far fragmented and outside the scopeof the excellent monograph `The Tomato Crop'(Atherton & Rudich, 1986) which is covering indetail the biology and cultivation of tomato .

Plant genetics

When compared to other crop plants, a relativelylarge amount of single gene determined traits havebeen described in tomato . Stevens & Rick (1986)

estimated the total number of available monogenicmutants at 1200 . These mutants include both spon-taneous genetic variants (from inter- and intraspec-ific origin), as well as those induced by irradiationand chemical treatment . Genes in tomato havebeen characterized by their stable variant pheno-type as compared with the cultivar Marglobe, rep-resenting the standard wildtype (+) allele and bymonogenic inheritance of this variant phenotype .Lists with the name, symbol and if known the mapposition of these genes are published every fewyears in `Genetic maps' (O'Brien, ed. Cold SpringHarbor Lab, 1987) and the annually appearing Re-port of the Tomato Genetics Cooperative . Themajority of the genetic stocks are being maintainedby the Tomato Genetic Stock Center, Departmentof Vegetable Crops, University of California atDavis, USA.

By repeated backcrossing of monogenic mutantswith the English cultivar Ailsa Craig near isogeniclines for many of the well known marker geneshave been developed at the Glasshouse Crops Re-search Station at Littlehampton, UK (Maxon-Smith & Ritchie, 1983) .

Several examples in which monogenic traits havecontributed to tomato breeding can be found inrecent reviews (Stevens & Rick, 1986 ; Stevens,1986; Tigchelaar, 1986) . The respective genes af-fect the growth habit of the tomato plant (e.g . sp,br, d), flower characteristics such as jointless ped-icel (j-l, j-2), they may cause male sterility (msmutants), a delay in fruit ripening (alc, nor, Nr, rin)or they affect fruit pigmentation (u, hp, gs, B, og, r,t, y) . In addition to these `useful' genes, many welldefined biochemical variants are known in tomatoe.g. for genes involved in anthocyanin synthesis,allozymes, plant hormones and photoreceptors,which are especially useful for basic plant research .

Monogenic traits are of interest because they canbe used relatively easily in breeding programs andalso because the respective genes, in principle, canbe isolated molecularly . Recently, cloning strate-gies are being developed, also for tomato, thatallow the cloning of a gene for which only a variantphenotype and a map position is known (see sec-tions Molecular markers and Transposontagging) .A special challenge for tomato molecular biologists

will be the cloning of monogenic disease resistances(Table 1), many of which are dominant and derivefrom other Lycopersicon species (Rick, 1982) .

Cytogenetics

The cultivated tomato is one of the few crop plantsin which a critical and intensive cytogenetic analy-sis has been conducted . Among several factors thathave favoured such an analysis, the following areprominent: (1) tomato is a naturally self-pollinateddiploid species ; (II) despite their rather small size,the chromosomes can be accurately identified ;(III) numerous easily recognizable genetic markersare available ; and (IV) the plants are highly prolificso that genetic analysis is easy . Utilizing these ad-vantages, cytogeneticists have built up a vastamount of information regarding the genome oftomato. Some of the salient features of this in-formation will be outlined below :

Despite the large basic chromosome number(n = 12), tomato must be considered a true diploidspecies from the following lines of evidence . First-ly, all the 12 pairs of chromosomes are morpholog-ically distinct from each other (Barton, 1950 ; Ra-manna & Prakken, 1967), thus ruling out the possi-bility of polyploidization in its recent evolutionaryhistory . Secondly, analysis of chromosome pairing

Table 1 . Diseases for which monogenic resistances are present

F: Fungus ; B: Bacterium ; V : Virus ; N: Nematode .* Capital indicates dominance of the resistant allele .

3

in tomato monohaploids has indicated the absenceof intragenomic homology (Ecochard et al ., 1969) .Thirdly, extensive genetic and molecular analysishas revealed that, with only a few exceptions, du-plicate loci are absent in the genome (Rick, 1971 ;Tanksley et al ., 1987) . In addition, there is evi-dence that the karyotypes of Lycopersicon speciesare strikingly similar to those of diploid tuberousSolanum, a genus closely related to the former(Gottschalk, 1954 ; Ramanna, 1979), indicatingthat the karyotypes of the species of these twogenera have been highly conserved during evolu-tion. In fact, Hawkes & Smith, (1965) have arguedthat the genus Solanum and its relatives haveevolved at least for 100 million years . The simi-larities between the karyotypes of Solanum andLycopersicon species are also evident from the cy-tology of intergeneric hybrids (Menzel, 1962 ;Klush & Rick, 1963) .The chromosomes of tomato have the unique

characteristic of being linearly differentiated intodistinct segments of euchromatin and heterochro-matin. As in many other organisms, heterochroma-tin in tomato is relatively darkly stained as com-pared to euchromatin . This differentiation can becytologically demonstrated in both mitotic andmeiotic cells by treating them with common nucle-ar stains such as acetocarmine or Feulgens reagent(Ramanna & Prakken, 1967) .

Disease Pathogen Type Resistance genes*

Alternaria stem canker Alternaria alternata f.sp . lycopersici F AscEarly blight Alternaria solani F adLeaf mold Cladosporium fulvum F CfseriesFusarium wilt Fusarium oxysporum F 1,1-2Late blight Phytophtora infestans F PhCorky root Pyrenochaeta lycopersici F pylVerticillium wilt Verticillium sp . F VeGray leaf spot Stemphylium sp . F SmSeptoria leaf spot Septoria lycopersici F SeBacterial Speck Pseudomonas syringae B PtoTobacco mosaic virus TMV V TM-1, Tm-2Root knot nematodes Meloidogyne spp . N Mi

4

The karyotypes can be analysed both at meta-phase stage of mitosis and pachytene stage of meio-sis . However, identification of metaphase chromo-somes of tomato is more difficult and less accuratethan those of pachytene chromosomes . For thisreason, all the cytogenetic analysis has been pos-sible in tomato only because of the ease and accu-racy with which pachytene chromosomes can beidentified . A common feature to both mitotic andmeiotic chromosomes is that the heterochromaticsegments are invariably present in the proximalpositions of the centromeres . This also means, thatthe position of euchromatic segments is invariablydistal . The only exceptional chromosome arm isthe short arm of chromosome 2 which is entirelyheterochromatic. This arm possesses the nucleolarorganizing region (NOR) which can be recognizedas a constriction in one pair of chromosomes atmitotic metaphase and as a nuclear bivalent at pa-chytene stage both of which are marked as chromo-some 2 .

In order to localize genes on chromosomes, theidentification of different types of aneuploids is oneof the important pre-requisites . A significant stepin this direction was achieved through the identifi-cation of a complete primary trisomic series in to-mato (Rick & Barton, 1954 ; Rick et al ., 1964) .Subsequently, the detection and use of various sec-ondary and tertiary trisomics facilitated the local-ization of genes to specific arms of chromosomes(for detailed review see Gill, 1983) . Especially withthe help of tertiary trisomics and telotrisomics, thecentromere positions have been rather accuratelymapped on the chromosomes . However, a singularaspect of gene localization in tomato is the applica-tion of the so called 'pseudodominant' methodwhich has facilitated the most accurate delimita-tions of some of the loci in the genome (see Khush& Rick, 1968) . This technique involves the irradia-tion of the pollen of a parent that carries certainnormal genes and the subsequent pollination to aparent with known recessive homozygotes . In theF,, those recessive genes that show 'pseudo-dom-inance' are selected and analysed cytologically atpachytene stage in order to detect the deletionscorresponding to the pseudo-dominant gene . Bythis method Khush & Rick (1968) approximated 35

loci on 18 of the 24 chromosome arms . Amongthese, some of the loci could be delimited to verynarrow chromosome segments of 5 µ or less (Rick,1971) .

In addition to the classical linkage map, the de-velopment of a molecular map of the genome (seesection Molecular markers) had added a new di-mension to the cytogenetics of the tomato . Themolecular maps are developed on the basis of allel-ic differences that exist for the lengths of restrictionfragments (restriction fragment length polymor-phisms or RFLPs) for certain loci which can bedetected through DNA hybridization techniques .These analyses have been carried out by using seg-regation data of interspecific hybrids between L.esculentum andL. pennellii . Although the chromo-somes of these two species are supposed to behomosequential (Khush & Rick, 1963), it is by nomeans certain that the crossing over in the inter-specific hybrids is strictly comparable to that inpure esculentum . In fact, Rick (1969) has pointedout reduced recombination and marked deviationsin segregation for specific loci in the back-crosssegregations of L. esculentum x L . pennellii hy-brids when compared to normal tomato . In view ofthis, the map distances estimated in the case ofmolecular maps ought to be treated with cautionand quantitative comparisons between molecularand classical maps should await more critical data .

Molecular biology

It is not within the scope of this section to review allthe gene systems currently being studied at themolecular level . They range from housekeepinggenes to genes involved in reproduction and inadaptation to abiotic and biotic stress . Some ofthem are studied to understand the fundamentalbiology of tomato, while others are examined be-cause of their supposed economic importance . Thereader is referred to a book entitled `Tomato Bio-technology', the Proceedings of a Symposium heldat the University of California, Davis, August 20-22 1986 (Eds . D .J . Nevins & R .A. Jones ; Alan R .Liss Inc ., New York) for an extensive discussion ofthe various gene systems . In this section' we will

restrict ourselves in discussing the basic molecularfeatures of the chloroplast-, mitochondrial and nu-clear genome. In the following sections, specificgenes and molecular approaches as they apply topractical breeding will be discussed .

Chloroplast DNA

Chloroplast DNA molecules from a wide variety ofplants have many features in common and are re-markably conserved in size and structure . They arerelatively small (120-200 kbp), circular double-stranded molecules containing a pair of invertedrepeat sequences which are flanking large singlecopy and small single copy regions (Palmer &Thompson, 1982 ; Palmer, 1985 ; Umesono & Ore-ki, 1987) .

Recently, the complete sequence of two chloro-plast genomes has been established, viz . Nicotianatabacum (Shinozaki et al ., 1986) and Marchantiapolymorpha (a liverwort ; Ohyama et al., 1986) .Sequence comparison of the two chloroplast ge-nomes of these evolutionarily very distant plantspecies reveals a remarkably similar organization,strongly suggesting that chloroplast genomes in allgreen plants may have originated from a uniqueancestor (Umesono & Oreki, 1987) .

Though no complete nucleotide sequence is yetavailable for tomato chloroplast DNA, all the evi-dence derived from complete restriction maps andclone banks point to a molecular organizationwhich is in agreement with the picture describedabove (Phillips, 1985a, 1985b ; Piechulla, et al .,1985 ; Hanson & McClean, 1987 ; Gruissem et al .,1987). The tomato chloroplast DNA is a circularmolecule of 156.6-159.4 kbp (Phillips, 1985a ; Pie-chulla et al., 1985) with two identical sequences of23 .3-27.7 kbp which are arranged as an invertedrepeat and are located at similar positions as inother plant species . Comparison of the tomato re-striction map with those of tobacco and petuniashows a high conservation of restriction sites (Phil-lips, 1985a) . By using heterologous probes frompea, wheat and spinach and by using coupled invitro transcription/translation on tomato plastidDNA clones, 14 protein-coding genes and the

5

rDNA genes could be mapped on the tomato re-striction map (Phillips, 1985b) . The results showthat all studied tomato chloroplast genes are insimilar positions as those in spinach, tobacco andpetunia .

One of the first studies to be carried out on thephylogenetic relationships of chloroplast DNA se-quences among Lycopersicon and Solanum specieswere those by Palmer & Zamir (1982). Though alimited amount of sequence divergence was found,by using a large number of restriction enzymes,sufficient sequence variation was detected amongthe 484 restriction sites surveyed to allow the con-struction of a chloroplast DNA phylogeny . Thephylogenetic tree thus established, reflected theevolutionary relationships based on morphologyand crossability, except for L. chmielewski which iscloser to L. peruvianum and L. chilense accordingto chloroplast DNA polymorphisms than otherspecies of the esculentum complex .

The restriction site polymorphisms between thechloroplast DNA of Lycopersicon species has beenapplied successfully to identify the parental originof chloroplast DNAs found in somatic hybridplants (O'Connell & Hanson, 1985, 1986 ; Tan,1987) .

Mitochondrial DNA

Unlike chloroplast genomes, mitochondrial ge-nomes from higher plants exhibit wide variation insize, ranging from approximately 218 kb in Brassicacampestris (Palmer & Shields, 1984) via 570 kb inmaize (Lonsdale et al., 1984) to 2500kb in musk-melon (Ward et al ., 1981) . In addition, they containmany direct repeats . Restriction fragment lengthanalysis of tomato mitochondrial DNA gave anestimate of 300 kb as the minimal size (McClean &Hanson, 1986 ; Hause et al ., 1986 ; Hanson &McClean, 1987) . For Brassica campestris it wasshown, that the mitochondrial genome is organizedinto three physically distinct circular molecules(218 kb, 135 kb and 83 kb), one of which bears theentire sequence complexity of the genome . Thetwo smaller circles contain distinct subsets of the`master' chromosome (Palmer & Shields, 1984) . A

6

similar but more complex pattern exists in maize,with a master chromosome of 570 kb and multiplecircular species of DNA (47 to 503 kb) which arederived from each other by recombination (Lons-dale et al ., 1984) . So far, it is not known whethermultiple DNA species are present within the toma-to mitochondrion. Intergenomic recombinationhas been suggested to occur in generating novelgenomes in tomato somatic hybrids following fu-sion (O'Connell & Hanson, 1985, 1986) .Mitochondrial DNA sequence divergence

among Lycopersicon and closely related Solanumspecies has been studied by McClean & Hanson(1986) . As compared to chloroplast genomes, mi-tochondrial genomes of tomato species are lessconserved. They show, however, much less diver-gence than mammalian mitochondrial genomes .The mitochondrial DNA-derived phylogenetic re-lationships resemble the relationships based on`classical' taxonomic data, except for the groupingof L. pennellii with L. peruvianum and L. chmie-lewskii and the close relationship between L. hirsu-tum and L. esculentum .

Nuclear DNA

Among the higher plants, tomato has a relativelysmall genome. Using chemical methods and flowcytometric analysis of nuclei, the nuclear DNAcontent of a wide variety of plants has been deter-mined and shown to range from 0 .5 pg to over200pg per haploid genome (Bennett et al ., 1982 ;Galbraith et al ., 1983) . Within this range, tomatoranks low with 0 .74 pg per haploid nucleus. As onepg of double stranded DNA corresponds to0.965 x 10 9 nucleotide pairs, this amount repre-sents approximately 0 .71 x 10 9 nucleotide pairs,which is roughly 180 times as much as E . coli, 10times as much as Arabidopsis thaliana but 5 and 13times as less as maize and rye, respectively . Thegenetic content of the tomato genome has beentaken as approximately 1200 cM (Tanksley, 1983) .Thus, on average, one cM in tomato corresponds toabout 600 kb .

A large proportion of the nuclear DNA of highereukaryotes is composed of repetitive, non-protein

coding, sequences and tomato is, in this regard, noexception . On basis of their chromosomal distribu-tion, repeated DNA sequences can be divided intofour major classes : repeated sequences (I) locatedat or around centromeres (II) associated with telo-meres (III) clustered at specific domains on chro-mosome arms and (IV) interspersed, intra- andinterchromosomal, among other genomic (singlecopy) sequences (see Flavell, 1982 ; Ganal et al .,1988; Manuelidis, 1982 ; Zabel et al ., 1985) . Re-cently, a systematic survey has been conducted intomato on the proportion and distribution of sin-gle-copy sequences and the various repetitive se-quence classes (Ganal et al ., 1988 ; Tanksley et al .,1987; Schweizer et al., 1988 ; Zabel et al ., 1985 ;Zamir & Tanksley, 1988) . Unlike the genome ofvarious monocots (wheat, rye and maize) and otherdicots like pea, in which repeated sequences ac-count for 60-80% of the genome, the tomato ge-nome is comprised largely of single copy and lowcopy number (2-20 copies) sequences . Among thefour major classes of repetitive DNA sequencefamilies identified, the majority is specific for Ly-copersicon species and Solanum lycopersicoides .

Application of the RFLP-technology (see sec-tion Molecular markers) has been a major tool instudying the chromosomal position of protein-cod-ing genes . By linkage analysis the position of morethan 100 random cDNA clones as well as variouscloned genes of known function has been establish-ed (Mutschler et al ., 1987 ; Tanksley et al ., 1987) .From these studies, the following conclusions canbe drawn .1. The majority of genes in tomato are represented

by a single copy at a single locus (Bernatzky &Tanksley, 1986) and are dispersed throughoutthe genome .

2 . Approximately 30-35% of the cDNA clonescorrespond to duplicated genes, the majority ofwhich are at genetically independent loci . Aminority (10-15%) of the clones are representedby 3-5 loci which reside on different chromo-somes. So far, at least 20 sets of unlinked, dupli-cate genes have been detected .

3. Two of the proteins involved in photosynthesis -small subunit of ribulose-bis-phosphate carbox-ylase (RBCS) and a chlorophyl alb binding poly-

peptide (CAB) - are encoded by multigene fam-ilies . The RBCS - multigene family consists offive genes which are distributed among threeloci at chromosome 2 and chromosome 3 (Valle-jos et al ., 1986; Pichersky et al., 1987). TheCAB-gene family consists of 11-12 genes dis-tributed among five loci at chromosomes 2, 8, 7and 12 respectively (Vallejos et al ., 1986 ; Pi-chersky et al ., 1987) .

4. No internal duplication of chromosome seg-ments has been detected except for chromo-some 2, which contains a duplication of a chro-mosomal fragment with an RBCS-gene and twoanonymous cDNA sequences .

In vitro culture methods

Like other Solanaceae, such as Nicotiana and petu-nia, tomato is relatively favourable for in vitro cul-ture of various types of tissues and organs . Proce-dures have been developed for the in vitro cultureof embryos, shoots, roots, leaf explants, and pro-toplasts. These methods have opened the possibil-ities for various applications, e .g. embryo rescue,in vitro mutant selection, basic biochemical andvirus research, the isolation of mitotic chromo-somes, protoplast fusion and genetic transforma-tion. In this section some of the in vitro methodswill be briefly outlined .

Callus and suspension cultures

Tomato tissue can develop callus or shoots depend-ing upon the composition of the medium (especial-ly hormonal balance), culture conditions and fac-tors such as age, vigour and the genotype of thedonor plant . In order to initiate the culture, anyone of the following tissues or organs have beenused: roots, cotyledons, hypocotyls, stems, leaves,anthers and embryos .

Callus is generally induced on media with rela-tively high auxin contents combined with moderatelevels of cytokinins . When callus of most L. escu-lentum cultivars is repeatedly transferred on to thesame medium, the ability to regenerate plants from

such calli upon transfer to cytokinin-containingmedia is very limited (Morgan & Cocking, 1982) .In contrast, callus of L, peruvianum maintains itsregeneration capacity for a much longer time(Muhlbach,1980 ; Thomas & Pratt, 1982) . This 're-generation capacity' has been bred into L. esculen-tum genotypes by backcrossing hybrids of L. peru-vianum x L. esculentum with L. esculentum(Koornneef et al ., 1986 ; Koornneef et al., 1987) .These genotypes facilitate the application of cellu-lar techniques that aim at obtaining geneticallymodified plants .

Shoot regeneration on explants

Direct regeneration of adventitious shoots on ex-plants placed on media with a high cytokinin/auxinratio has been studied by many authors (Padma-nabhan et al ., 1974 ; Kartha et al ., 1976; Behki &Lesley, 1976 ; Frankenberger et al ., 1981a; Locky1983 ; Zelcer et al ., 1984) . In general, the observa-tion of Kartha et al ., 1976 that media with ben-zyladenine (BA) or zeatin (Z) at concentrations of1-5 mgl-l combined with indoleacetic acid (0.1-0.5 mgl-1) are most effective for shoot regenerationhas been confirmed . Shoot formation by L. escu-lentum genotypes is often accompanied by callusformation on the explant . However on L. peruvia-num explants, shoots often appear without the si-multaneous appearance of callus tissue . Also with-in L. esculentum genotypes, genetic differences forshoot forming capacity have been described (Ohkiet al ., 1978; Frankenberger et al ., 1981b ; Tan et al .,1987b) .

Plant regeneration from protoplasts

Unlike L. peruvianum protoplasts, those of L. es-culentum were found recalcitrant to regenerate(Miihlbach, 1980 ; Morgan & Cocking, 1982) . Re-cently, however, (Shahin, 1985 ; Niedz et al ., 1985 ;Tan et al., 1987a) procedures have been developedwhich allow a more efficient regeneration of toma-to protoplasts from a wide range of genotypes . Theimportance of a proper pretreatment of the plant

7

8

material to be used as a source for protoplasts(Tabaeizadeh et al ., 1984; Shahin, 1985 ; Tan et al .,1987a) and a relatively low ammonium ion contentin the media (Zapata et al ., 1981) are generallyrecognized as being critical .

Transformation

In this section methods are reviewed to transferwell characterized DNA sequences into the plantgenome and to select for transformed plant cells .Two fundamentally different delivery systems aredescribed. Firstly, direct (physical) DNA deliverysystems to plant cells will be discussed . Secondly,transformation methods based on the natural inter-action between the soil bacterium Agrobacteriumtumefaciens and plants will be in focus .

Selection markers

Upon introduction of new genetic information inplant cells, it is crucial to be able to distinguishbetween transformed and non-transformed plantcells. To this end, vectors have been developed inwhich the gene of interest is carried along with agene which confers resistance to an antibiotic thatis lethal to plant cells . Following introduction of thevector into a population of plant cells, transformedplant cells can be selected for by growth in thepresence of the antibiotic . Genes coding for anti-biotic resistance have been isolated from bacterialstrains and brought under control of plant geneexpression signals to allow proper expression inplant cells . Thus, dominant selectable markershave been constructed which confer resistance tothe antibiotics kanamycin (Bevan et al., 1983; Her-rera-Estrella et al ., 1983a), chloramphenicol (Her-rera-Estrella et al., 1983b), hygromycin (van denElzen et al ., 1985b ; Waldron et al ., 1985) and bleo-mycin (Hille et al ., 1986), to the herbicide phosphi-notricine (PTT ; De Block et al ., 1987) and to thedrug methotrexate (Herrera-Estrella et al ., 1983b),an inhibitor of dihydrofolate reductase . Amongthese markers, kanamycin resistance is the mostpopular in tomato, although hygromycin resistance

is equally effective (Weide, pers . comm .) .

Direct gene transfer

In the past few years methods have been developedto directly deliver characterized DNA sequencesinto plant cells . These procedures have commonlybe called direct gene transfer . In general, plantprotoplasts are incubated with calciumphosphate-DNA precipitates in conjunction with polymerslike polyethylene glycol (Paszkowski et al ., 1984 ;Negrutiu et al ., 1987) or polyvinylalcohol (Hain etal., 1985) . More recently, conditions have beenestablished for electrically introducing DNA intoplant protoplasts (so-called electroporation, Shilli-to et al., 1985), and also microinjection of DNAinto plant protoplasts, either cytoplasmic or in-tranuclear, has been successfully applied (Cross-way et al ., 1986 ; Miki et al ., 1987). A new devel-opment is the application of high velocity microprojectiles, coated with DNA, in transformation ofplant cells (Klein et al ., 1988). This approach hasthe advantage of not being limited to protoplasts .

Transformation of tomato protoplasts has beenperformed using the calciumphosphate DNAtransformation procedure (Koornneef et al ., 1986 ;Pieterse & Koornneef, 1988) . Selection for kana-mycin resistance (50 mg/1) encoded by the plasmidDNA used in transformation, was started twoweeks after protoplasts transformation . Using thisprocedure, kanamycin resistant calli were obtainedat frequencies up to 10-3 .

Six tomato transformants were chosen for fur-ther detailed molecular analysis (Jongsma et al .,1987). By Southern blot analyses it was shown thatthe transformants contain up to three copies of thekanamycin resistance gene integrated into the to-mato nuclear DNA. Less than full length plasmidcopies were found to be integrated into the nucleargenome suggesting that nuclease activity had linea-rized and degraded the plasmid DNA to differentextents prior to integration. Evidence was foundfor physical linkage of integrated plasmid DNAfragments in transformants containing more thanone integrated fragment. A similar linkage wasobserved in tobacco transformants obtained fol-

lowing electroporation (Riggs & Bates, 1986) .

Agrobacterium tumefaciens - mediatedtransformation methods

A remarkable example of naturally occurring plantgenetic engineering is provided by the soil bacteri-um Agrobacterium tumefaciens . This phytopatho-gen transforms susceptible plant cells to causecrown gall, a neoplastic disease of dicotyledonousplants . The transfer, integration and expression ofT (transferred)-DNA from its Ti (tumor-inducing)-plasmid into the nuclear genome of plant cells isnow known to be the molecular basis for pathogen-icity of A . tumefaciens (for reviews see : Hille et al .,1984 ; Schell et al ., 1984; Koukolilova-Nicola et al .,1987) .Two kinds of plant gene vectors have been devel-oped to introduce new genetic information in plantcells (for a recent review : see Rogers & Klee,1987). In the so-called cointegrate type T-DNAvector, all oncogenic functions located on the T-DNA of the Ti-plasmid have been deleted andreplaced by a DNA sequence, e.g. a copy of plas-mid pBR322 which functions as a target for homo-logous recombination . All sequences, cloned in E.coli on a pBR322 like plasmid, can then be trans-ferred toA . tumefaciens and placed in the T-DNAof the Ti-plasmid following homologous recombi-nation between the pBR322 sequences (Zambryskiet al., 1983; Van Haute et al., 1983) .

In the binary type T-DNA vector, the Ti-plasmidhas been split into two independently replicatingplasmids: one carrying virulence functions and theother the T-DNA . The plasmid containing the T-DNA (from which the oncogenes have been delet-ed) is now sufficiently small and can be used in E.coli for cloning the desired genes . This plasmid isthen transferred back to A . tumefaciens harboringthe other plasmid containing the virulence func-tions and the resulting strain can be used for plantcell transformation (Hoekema et al., 1983 ; De Fra-mond et al ., 1983) .

Currently, the most widely used techniques forthe efficient introduction of new genetic informa-tion into plant cells mediated by A . tumefaciens are

9

the cocultivation method (Marton et al ., 1979) andthe leaf disc method (Horsch et al ., 1985) . In thefirst method, purified plant protoplasts are incu-bated for 2-4 days to allow cell wall regeneration tooccur. Then, as the first cell divisions are visible, A.tumefaciens is added in a ratio of 100 bacteria perprotoplast and the mixture is incubated (cocultivat-ed) for 2-4 days . Subsequently, the bacteria areremoved by washing and the plant cells are furthercultured in the presence of antibiotics to kill theremaining bacteria . (Fraley et al ., 1984; Van denElzen et al ., 1985a ; Hille et al ., 1986) . Using thisprocedure, transformation frequencies up to 50%of the regenerating micro-calli have been reported .However, both the cocultivation method and thedirect gene transfer methodology have their limita-tions in genetic manipulation of plants, since theyare both based on protoplast regeneration . Farfrom all plant species can be regenerated fromprotoplasts and plants derived from protoplasts of-ten show variation due to mutations and/or chro-mosomal abnormalities (so-called somaclonal vari-ation) .

A method that bypasses most of these limitationsis the leaf disc method as developed by Horsch etal ., (1985), which combines the efficient plant celltransformation capacity ofA. tumefaciens with re-generation capacity of leaf explants . To this end,leaf discs are cut from leaves, dipped in a culture ofA. tumefaciens and incubated for two days . Sub-sequently, the leaf explants are transferred to ashoot-inducing medium containing antibiotics tokill the bacteria and to select for transformed plantcells. Using the leaf disc transformation procedureall plant species, which can be regenerated fromleaf explants and which can be infected by A . tume-faciens, can be transformed and regenerated intotransgenic plants. In case of the cultivated tomatoall genotypes tested, including some cultivars,could be transformed and regenerated into fertileplants . The newly introduced kanamycin resistancegene was shown to transmit through meiosis in aMendelian manner (Koornneef et al., 1986 ;McCormick et al., 1986) .

10

Molecular markers

Over the past decades, a linkage map has beenconstructed containing loci for morphologicalmarkers, resistance genes and mutants affectingphysiological functions . Though this map hasserved well in genetic studies, its applications inplant breeding programs is rather limited. There-fore, more recently, particular attention is paid tothe construction of a detailed molecular map basedon protein and restriction fragment length poly-morphism (RFLP) markers which do not affect thephenotype. Protein markers are loci encoding pro-teins (isozymes) that can be separated by electro-phoresis and subsequently visualized by specificstaining to establish the presence or absence ofspecific alleles . The term restriction fragmentlength polymorphism has been introduced to de-scribe the variation in length of DNA restrictionfragments from two genetically distinct individualsthat are homologous to a labeled cloned DNAsequence. Actually, both types of markers, proteinand RFLP, are based on polymorphisms which aredue to differences in the primary sequence of ge-nomic DNA. With protein markers, the polymor-phisms are usually detected as electrophoretic orantigenic polypeptide variants, whereas DNAmarkers directly reflect the variation at the DNAlevel, being either in a coding or noncoding region .In the human genome the variation in restrictionfragment length is mainly due to single basepairchanges (Barker et al ., 1984; see also Gusella,1986) while in plants, like maize, changes by in-sertions and deletions seem to be common (Helent-jaris et al ., 1985) .

As discussed by Tanksley (1983) and manyothers (Beckmann & Soller, 1983, 1986 ; Soller &Beckmann, 1983 ; Burr et al ., 1983 ; Helentjaris etal., 1986; Landry & Michelmore, 1987b) molecularmarkers possess several distinct advantages overmorphological and other conventional `classical'markers . The following discussion was given byTanksley (1983) .1. While phenotypes of most morphological mark-

ers can only be recognized at the plant level,genotypes of molecular loci can be establishedat cellular, tissue and plant levels .

2. The number of distinguishable alleles at morph-ological or biochemical loci is rather limited andoften dependent on the application of exog-enous mutagens. RFLPs, however, provide avirtually unlimited number of markers, as thepolymorphism does not need to alter the pheno-type or the charge of a protein to become detec-table. With markers for all chromosomes at in-tervals of 10-15 map units, the flow of any geneof interest through segregating generations in abreeding program can be monitored, using alinked (anonymous) RFLP-marker as `tag',

3 . Morphological markers are usually associatedwith unwanted phenotypic effects, which is notthe case with molecular markers .

4. As alleles of most molecular markers are codo-minant, all possible genotypes can be distin-guished in any segregating generation .

5. Only a few epistatic or pleiotropic effects occurwith molecular markers, allowing a virtually un-limited number of segregating markers to bemonitored in a single population .

Among the molecular markers, isozyme markershave proven their usefulness already in geneticstudies and breeding programs of various crops(Tanksley & Rick, 1980 ; Tanksley & Orton, 1983) .For tomato, an isozyme linkage map has been con-structed containing 41 isozymic genes correspond-ing to 15 unique enzyme reactions. Thirty six ofthese genes have been mapped to their respectivechromosomes (Tanksley & Bernatzky, 1987) .Thus, tomato isozymes have served many purposes(see Tanksley & Bernatzky, 1987), like e .g ., verify-ing the purity of hybrid seed (Tanksley & Jones,1979), tagging genes of economic importance (Rick& Fobes, 1977 ; Tanksley et al ., 1984), detectinggenes and chromosomes from wild species follow-ing introgression (Rick et al ., 1986), identifyingsomatic hybrids and mapping genes underlyingquantitative variation (Vallejos & Tanksley, 1983) .The great potential power of using isozyme mark-ers might best be inferred from the publication of atwo volume manual entitled `Isozymes in Plant Ge-netics and Breeding' (Tanksley & Orton, Eds .1983) which comprises an extensive discussion ofboth theoretical and practical aspects for manycrops .

Nevertheless, isozymic loci suffer several limita-tions which are not inherent to the use of DNAmarkers (RFLPs) . The great potential usefulnessof RFLPs in basic plant genetic studies and plantbreeding programs has been advocated in the earlyeighties by Beckman & Soller (1983), Burr et al .,(1983), Helentjaris et al ., (1985) and Tanksley(1983), following a thorough description of the the-oretical basis of using RFLPs as a new source ofmarkers for the diagnosis of human diseases (Bot-stein et al., 1980 ; Little et al ., 1980 ; Phillips et al .,1980; Gusella et al ., 1983 ; Gusella, 1986) . Sincethen, RFLP markers have found their way as pow-erful tools in human genetics and cloning geneswhich are linked to them and would otherwise nothave been clonable (see Gusella, 1986) . In plants,the integration of RFLP-technology in basic andapplied research proceeds more slowly but shouldhave a similar impact . Thus far, genetic linkagemaps based on RFLPs have been constructed fortomato (Bernatzky & Tanksley, 1986a, b ; Helent-j aris et al ., 1985 . Mutschler et al ., 1987; Tanksley etal ., 1987), lettuce (Landry et al., 1987a, b) maize(Helentjaris et al ., 1986 ; Helentjaris, 1987) andArabidopsis thaliana (Chang et al., 1988) .

For an elaborate discussion of the methodologyand the prospects of the applications of RFLP anal-ysis in plant systems, the reader is referred to re-views by Landry & Michelmore (1987a, b) andTanksley et al. (1987) .

In a first series of experiments on RFLPs intomato, random cDNA clones, derived from leafmRNAs, were used successfully in genetic map-ping (Bernatzky & Tanksley, 1986a, 1986b ; He-lentjaris et al ., 1985) . These studies have been ex-tended more recently with genomic clones repre-senting single/low copy number sequences (He-lentjaris et al ., 1986 ; Tanksley et al ., 1987; Younget al., 1987) .

By linkage analysis, the chromosomal position ofmore than 200 DNA markers, including random(anonymous) cDNA clones, genomic clones andvarious cloned genes of known function, has nowbeen established (Mutschler et al ., 1987; Tanksleyet al ., 1987, Young et al., 1987) . Thus, a molecularlinkage map of tomato has been constructed inwhich each of the 12 chromosomes is represented

11

with a characteristic set of DNA markers .With a molecular linkage map containing more

than 200 DNA markers now available, how manymore markers are needed or desirable to saturatethe tomato genome? The answer to this questiondepends on the type of application being pursued,namely whether the RFLP markers are consideredto be used as genetic markers to tag chromosomesegments in breeding programs or as moleculartools for isolating nearby genes of interest . Formost purposes in plant breeding, the current map isalready sufficient in tagging genes or chromosomesof interest or detecting quantitative trait loci(QTLs), as was shown for the loci controlling fruitmass (Paterson et al., 1988), fruit pH (Paterson etal., 1988) and the soluble solid content (Osborn etal ., 1987; Paterson et al ., 1988 Tanksley & Hewitt,1988). Another example of the successful applica-tion of current RFLP markers in tomato concernsthe determination of the chromosomal location offoreign DNA sequences introduced in the genomebyA . tumefaciens (Chyi et al ., 1986) . For the othertype of application - RFLP markers as moleculartools for isolating genes linked to them - moremarkers should be mapped to fill in the gaps . Analternative to mapping RFLP's in the proximity ofthe desired gene is to identify DNA markers re-vealing polymorphisms between near isogenic linesas was recently described for the Tm-2a (TMV-resistance) gene (Young et al ., 1988) .

Though gaps are certainly not the plant molecul-ar biologist's wish-dream, the development of newtypes of clone libraries (chromosome jumping andlinking libraries, Poustka & Lehrach, 1986) andnew techniques for generating and fractionatingrestriction fragments up to 4000 kb . (Barlow &Lehrach, 1987) makes even these distances nolonger elusive. Currently already, the gap in reso-lution between classical genetics and molecularDNA techniques is being bridged in mammaliansystems (Barlow & Lehrach, 1987 ; Poustka & Leh-rach, 1986) . Genetic loci which were only knownfrom their mutant phenotype and genetic map posi-tion have become amenable to molecular analysisand cloning, following the construction of physicalmaps at the cM-level . Application of RFLP-mark-ers in conjunction with these long-range cloning

12

and mapping techniques could have a similar im-pact on tomato genetics . Disease resistance loci areobvious candidates for such a `reverse genetic' ap-proach, since their protein products are as yet un-known and identification of coding sequences cor-responding to the locus remains otherwise elusive .As a prelude to constructing a physical genetic mapof the nematode resistance Mi region on chromo-some 6 of tomato, we have recently developed aprocedure for the isolation of megabase-sized chro-mosomal DNA and the subsequent generation andfractionation of restriction fragments in the range100-1500 kb (van Daelen et al ., 1989) . These condi-tions should allow the construction of long-rangerestriction maps and the isolation of genes in closeproximity to RFLP markers . Given the pace ofthese new mapping and cloning techniques, it isanticipated, that physical genetic maps of variouschromosomal regions of tomato will be establishedwith the next couple of years .

Transposon tagging

Isolation of plant genes and their introduction incrops gives a new dimension in crop improvement .However, it is often difficult to isolate these usefulgenes molecularly, since, in most cases, gene prod-ucts are not known . Two general approaches arecurrently being developed for their molecular iso-lation: RFLP tagging and transposon tagging .

As described earlier, a close linkage betweenknown genes and RFLPs can be used as startingpoints for the molecular isolation of certain genes .An inherent problem associated with this RFLP-tagging approach is that it is difficult to associatethe appropriate DNA sequences with the corre-sponding phenotype of the desired gene . A solu-tion to this problem can be found in some cases .For example, if the cloned DNA sequences repre-sent the wild type, such sequences might be in-troduced into a recessive mutant through genetictransformation . In this case, the sequence or se-quences that restore wild type in the transformantscan be easily associated with the correspondingmutant gene .

Transposon tagging might be a more straight

forward approach for the molecular isolation ofuseful genes. Application of endogenous planttransposable elements has shown to be a usefulmethod to obtain plant mutants . The insertion of atransposable element into a genetic locus can giverise to mutations . Using a transposable elementfrom maize, the Ac-element, Fedoroff et al .(1984), were able to clone a locus which until thenwas only characterised genetically . Their resultsindicate that in principle any locus in maize, identi-fiable with a transposable element, can be clonedand analysed at the molecular level . This approach,called transposon tagging, has successfully beenapplied for the isolation of various maize genes(Wienand & Saedler, 1987) .

The subsequent demonstration that the maizetransposable element Ac can transpose in dicotshas opened the possibility to use this well-charac-terised element for gene-isolation in plants otherthan maize (Baker et al ., 1986, 1987 ; Van Sluys etal ., 1987) . Using the A. tumefaciens mediated leafexplant transformation procedure, the transpos-able element Ac has been introduced into varioustomato lines and shown to transpose within theselines (Yoder et al ., 1988 ; Haring et al ., 1989) .

In maize and also in tobacco it has been shownthat Ac has a tendency to transpose in the vicinityrather than to far away positions (Greenblat,1984 ;Dellaporta & Chomet, 1986). Therefore, to obtaina transposon insertion in a particular gene, it isadvantageous to start with a plant containing thetransposable element close to the position of thedesired gene . Currently, the utility of this trans-poson tagging system for tomato gene isolation isunder study in various research groups .

Finally, it must be pointed out that in spite ofextensive genetic and cytogenetic investigations,there has not been any indication so far for thepresence of transposable elements in tomato . Itwould be highly attractive to detect and character-ize transposable elements, if any, in tomato for tworeasons: Firstly, they could be useful for enhancingthe knowledge regarding the genome organizationat the molecular level, and secondly, they might beuseful as tools in isolating desirable genes through'transposon tagging' (Federoff et al ., 1984) . In thisconnection, attempts are underway (Ramanna et

al., 1985) to induce genetic instability in tomatothrough 'breakage-fusion-bridge-cycle' (BFBC), amethod that was originally used by McClintock inmaize (for a review, see McClintock, 1984) . Forthis purpose, the iso-chromosome line of tomatothat was described by Moens (1965) has been espe-cially useful . The iso-chromosomes in these linesare entirely heterochromatic and many copies(1-9) can survive in plants without affecting vigourand fertility. A remarkable feature of these extrachromosomes is that they are structurally unstableand generate chromosomes with `sticky ends'which cause BFBC. Plants with such unstable chro-mosomes can give rise to progeny that contain alow frequency of genetically unstable plants as welas with chromosomal abnormalities (Ramanna etal ., 1985 and unpublished) .

Applications

Anther and pollen culture

Although the production of a tomato haploid byanther culture has been reported as early as 1972(Greshoff & Doy, 1972) and the application ofpollen culture was reported in the same year (Sharpet al ., 1972), haploids are still difficult to obtain .The response to anther culture was found to begenotype dependent (Zagorska et al., 1982) andespecially the positive effect of genotypes homozy-gous recessive for the ms-1035 allele has been re-ported by several laboratories (Zagorska et al .,1982 ; Zamir et al ., 1980 ; Ziv et al ., 1984) . How-ever, from such cultures haploid plantlets wererarely recovered . Probably the rapid polyploidiza-tion of haploid callus cells of tomato and the out-growth of anther wall tissue (Levenko et al ., 1977,Zamir et al ., 1981) are the main reasons for thisfailure .

Haploid plants originating by parthenogenesiscan be found spontaneously at frequencies ofapproximately 2 x 10 -4 (Ecochard et al., 1969 ;Koornneef et al., 1989) and can be propagatedvegetatively . When protoplasts from such haploidscan be induced to divide and regenerated intoplants, they might be used for the selection of re-

cessive mutants at the cell level .

Embryo culture

In some of the interspecific and intergeneric cross-es, the embryos abort and prevent the formation ofhybrid seeds . Such premature embryos can be res-cued by cultivating them in vitro (embryo rescue) .One early example is the interspecific cross L. escu-lentum x L. peruvianum which was instrumentalfor the introduction of resistance against root knotnematode into the cultivars (Smith, 1944) . Theother examples are the hybrids of L. esculentum xL. chilense (Rick & Smith, 1953 ; Rick, 1963), andthe intergeneric hybrid L. esculentum x Solanumlycopersicoides (Rick, 1951; De Verna et al ., 1987) .

An improvement of embryo culture techniquewas published by Thomas & Pratt (1981) who in-duced callus formation on very young embryos sub-sequently followed by shoot induction .

Somaclonal variation

Variation among plants derived from tissue culture(called somaclonal variation) is considered a draw-back by those interested in the uniform multiplica-tion of a particular genotype. In contrast, plantbreeders looking for more variation consider soma-clonal variation as an advantage of tissue culture .However, one should realize that most variantsfound will not be useful . Plants with a higher ploidylevel than in the original plant material occur de-pending on the type of tissue culture system em-ployed. Plants derived from diploid explants byadventitous shoot formation show predominantlythe original ploidy level. However, plants regener-ated from established callus cultures or protoplastsare in majority tetraploid (O'Connell et al ., 1986 ;Koornneef et al ., 1989) .

Sibi (1982) reported for the first time in tomatothe occurrence of `epigenetic variation' among re-generated plants derived from tissue culture .Evans & Sharp (1983) studied the progeny of 230plants derived from leaf explants and found 13tentative nuclear mutations . Four of these mutants

13

14

showed allelism with mutants that were describedearlier (Evans & Sharp, 1986) . Buiatti et al., (1985)found that 17% of 88 progenies segregated mono-genic mutants. Another striking example of soma-clonal variation in tomato has been published byShahin & Spivey (1986) who found 5 out of 100plants that were resistant to Fusarium oxysporumrace 2 (a dominant resistance) following regener-ation of protoplasts . Darden et al . (1986) reportedthe isolation of somaclonal variants with maternal-ly inherited TMV resistance . Sibi et al., (1984)regenerated plants from cotyledons of F t hybridsheterozygous for several linked genes and found anincrease in recombination frequencies as comparedwith control F, hybrids . The increase was 5-7% forgenes located at approximately 20 cM from eachother. A comparison of chemical mutagenesis withtissue culture induced mutants indicated differenc-es in both mutant spectrum and mutant frequency(Gavazzi et al., 1987) .

Mutant selection at the cellular level

Selection at the cellular level for resistance againsttoxic compounds seems a feasible and attractiveapproach to obtain mutants . Indeed, various varia-nts have been described (Table 2) . There are, how-ever, several major problems associated with thistype of mutant selection. First, variant plant cells

Table 2 . Variants isolated at the cell level in Lycopersicon

are often found difficult to regenerate . Secondly,chromosomal instability can occur in the regenerat-ed plants . Thirdly, there is a lack of haploid plantsor cell cultures which can be efficiently regeneratedin order to facilitate the isolation of recessivemutants .

Protoplast fusion

In comparison to sexual crossing, protoplast fusionoffers several unique possibilities including : (1)sexual barriers can be overcome, (2) different cy-toplasms can be mixed which cannot normally beaccomplished, and (3) parts of a genome can betransferred through asymmetric fusion .

The demonstration of protoplast fusion as a tech-nique to hybridize species that cannot be crossedsexually was reported for the first time with toma-to/potato hybrids (Melchers et al ., 1978) . Recently,somatic hybrids have been produced betweenother Lycopersicon species and related Solanumspecies (Table 3) .

The efficiency for the selection of hybrids is sofar limited, however, mainly due to the lack ofsuitable cell selection markers . Selection criteriathus far applied in fusion experiments are the ca-pacity pf hybrids to regenerate (Adams & Quiros,1985) ; Wijbrandi et al., 1988), kanamycin resist-ance (Adams & Quiros, 1985), the ability to survive

* Resistance was expressed at the callus level (also on callus derived from regenerated plants) but not at the plant level .

Variants phenotype resistant to Parental genotype Regeneration ofvariant plants

References

Glyphosate L. peruvianum x L. esculentum Smith et al . (1986)(Herbicide) hybridParaquat L. peruvianum x L. esculentum + * Thomas and Pratt (1982)(Herbicide) hybrid L2Cadmium L. peruvianum suspension Bennetzen and Adams

(1984)

Aluminum L. esculentum `Marglobe' Meredith (1978)Polyethylene glycol L. esculentum Bressan et al . (1981)Fusaric acid L. esculentum + Shahin and Spivey (1986)Fusarium elicitor L. esculentum Buiatti et al . (1987)

particular media and treatments (Handley et al .,1986) and morphological selection (O'Connell &Hanson, 1985 ; Kinshara et al ., 1986 ; Tabaeizadehet al ., 1985) . Recently, nuclear markers selectableat the cell level, such as antibiotic resistances in-troduced by transformation, became available andallow a more efficient selection of hybrids. Also,chloroplast-encoded albino mutants (Hosticka &

Table 3 . Somatic hybrids obtained with Lycopersicon species

Parental species

L. esculentum (+) L . peruvianum

L. esculentum (+) S . tuberosum

L. esculentum (+) S. lycopersicoides

L . esculentum (+) S. rickiiL. esculentum (+) L . pennellii

L. peruvianum (+) L . pennellii

L. peruvianum (+) Petunia hybrida

L. esculentum (+) S . nigrum

Table 4 . Genes introduced into the tomato genome by genetic transformation

Trait

direct gene transferkanamycin resistance

A. tumefaciens mediatedkanamycin resistance

herbicide tolerance

virus resistance

insect resistancefruit ripeningtransposon activity

protein compositionin chloroplasts

References

Kinshara et al ., 1986 ;Wijbrandi et al . 1988Melchers et al ., 1978,Shepard et al ., 1983

Handley et al ., 1986 ;Tan, 1987O'Connell et al ., 1986O'Connell et al ., 1985 ;1987Adams & Quiros,1985 ; Tan, 1987Tabaeizadeh et al .,1985Guri et al ., 1988

Specification

glyphosateL-phosphinothricinTMV-coatproteinAMV-coatproteinlepdidopteran insectsdown regulation of polygalacturonase activityActivator (Ac) element from maize

plastocyanin importedin chloroplasts

Hanson, 1984) and chloroplast-encoded antibioticresistances (Jansen et al., in preparation) are nowavailable in tomato . Thus far, use of the somatichybrids in breeding programs has not been report-ed .

In order to transfer a limited amount of desirablegenetic variation from a wild species into cultivatedtomato, asymmetric protoplast fusion is a usefultool .Asymmetric hybrids were obtained by fusing irra-diated protoplasts of L. peruvianum with untreatedL. esculentum protoplasts and have been isolatedby Wijbrandi et al . (1988) by selecting the hybridson the basis of their better regeneration capacity ascompared with L. esculentum .

Some representative examples of tomato geneticmanipulation

Chromosomal location of introduced genes

The genetically well characterized tomato has beenused as a plant species to test whether T-DNAinserts can be used as chromosome specific mark-ers and whether there is site-specificity in T-DNA

References

Koornneef et al . 1986Jongsma et al . 1987

Chyi et al . 1986Koornneef et al . 1986McCormick et al . 1986Filatti et al . 1987De Block et al . 1987Nelson et al. 1988Turner et al . 1987Vaeck et al . 1987Smith et al . 1988Yoder et al . 1988Haring et al . 1989De Boer et al . 1988

1 5

16

insertion . To this end, the site of T-DNA insertionshas been determined for 7 tomato transformantsusing a molecular genetic approach (Chyi et al .,1986) and for 8 tomato transformants using a genet-ic approach (Weide et al., in preparation) . The 15mapped T-DNA inserts were located on 8 differentchromosomes. It was concluded that the T-DNAdoes integrate into various tomato chromosomesand can indeed be used as a chromosome specificmarker. Similar results have been presented forT-DNA inserts in Petunia chromosomes (Wallrothet al., 1986) and in Crepis capillaris chromosomes(Ambros et al., 1986) . With respect to a particularchromosome, the T-DNA can be integrated at vari-ous positions, as was shown for tomato, petuniaand Crepis capillaris .

Herbicide tolerance

In modern agriculture the use of herbicides hasbecome essential for weed control . However, mostof the new herbicides, that combine effectivenesswith safety for animals and environment do notdistinguish between weeds and crops . Therefore,three independent strategies have been employedto obtain plants tolerant to these herbicides .(1) Overproduction of the sensitive target enzyme .

It has been demonstrated in petunia that over-expression of the protein, on which the herbi-cide acts, confers herbicide tolerance (Shah etal ., 1986) .

(2) Modification of the sensitive target enzyme .The protein on which the herbicide acts is mod-ified in the engineered tomato plants so as torender its activity insensitive to the herbicide .(Fillatti et al., 1987) .

(3) Detoxification of the herbicide . A gene is in-troduced into tomato encoding a protein whichdetoxifies the herbicide (De Block et al ., 1987) .

All three strategies have both their advantages anddisadvantages . However, it might be expected thatespecially the third strategy will prove most usefulsince quite often, the detoxifying enzymes can beisolated from bacteria and transformed crops dis-play a relatively high level of herbicide resistance(for a recent review see : Botterman & Leemans,1988) .

Virus resistance

Cross-protection, in which plants are infected witha mild virus strain to protect against infection withvirulent strains, has since long been used to controlyield of plants . Following the development of ge-netic manipulation procedures, it has been testedwhether virus resistance can be induced in alterna-tive ways. Several strategies have been followed,including (1) expressing viral coat protein in trans-genic plants (Turner et al ., 1987; Nelson et al .,1988), (2) expressing antisense viral RNA in trans-genic plants (Cuozzo et al ., 1988) and (3) express-ing satellite RNA's in transgenic plants (Harrisonet al., 1987; Gerlach et al ., 1987) . The coat proteinstrategy results in symptom retardation and is de-pendent on the amount of coat protein expressed inthe transgenic plants and on the strengh of the viralinoculum . Apparently, the effect of the expressedcoat protein can be overcome when the viral RNAaccumulates above a certain concentration . Theantisense RNA approach, using antisense coat pro-tein message, results in a low protection levelagainst virus infection, that is less effective com-pared to the coat protein mediated protection. Thesatellite RNA strategy depends on amplification ofthe satellite RNA molecules by the incoming virus,resulting in suppression of symptom development .Therefore, protection is independent of thestrengh of the viral inoculum and insensitive to theamount of satellite transcripts in the transgenicplants, but limited, however, to those viruses whichcontain a satellite RNA .

Insect resistance

In agriculture a wide variety of insecticides is usedto control insect damage . Since chemical control isexpensive and sometimes undesirable, effort is fo-cused on improving plant defences against insectattack .A single gene has been identified in the bacteriumBacillus thuringiensis which codes for a polypep-tide which is specifically toxic to a variety of insectspecies. This gene has been cloned, made suitablefor expression in plant cells and introduced intoplants. Transgenic plants, producing this new pro-

tein, were shown to be protected against feedingdamage by larvae of the insect species (Vaeck etal., 1987) .

In nature several mechanisms exist by whichplants protect themselves against insect damage .One such mechanism involves trypsin inhibitorswhich act at the catalytic site of enzymes, therebyinterfering with digestion in the insect gut . A tryp-sin inhibitor gene has been cloned from cowpea,manipulated in such a way that the expression ofthe gene was constitutive and relatively high, andintroduced into tobacco . Transgenic tobacco plantswere shown to have an enhanced resistance to oneof its major insect pests (Hilder et al ., 1987) .

Down-regulation of specific plantgene expression

Genes are universally expressed via RNA tran-scripts giving rise to either messenger RNA's whichare subsequently translated into proteins, or struc-tural RNA's. Recently, strategies have been devel-oped in plants to down-regulate endogeneous geneexpression by interfering at the RNA level . Theseapproaches involve the introduction into plants ofeither an antisense gene or a very specific endori-bonuclease activity .

Antisense RNA inhibition has been shown inpetunia at the flower colour level (van der Krol etal., 1988) and in tomato at the level of polyga-lacturonase activity in ripe fruits (Smith et al .,1988) . General rules have been established for thedesign of RNA enzymes (so-called ribozymes) ca-pable of cleaving RNA highly specific . This hasbeen applied against a model target sequence andshown to be successful in vitro (Haseloff & Ger-lach, 1988). Ribozymes have not yet been tested inplants .Both approaches will enable to create mutant

phenotypes for any molecularly characterized geneand might prove valuable in e .g. manipulating bio-chemical pathways in plants in both a quantitativeand a qualitative manner .

17

Acknowledgements

We are grateful to Dr . A. Zelcer for critically re-viewing this manuscript and to Mrs . H. Bartelsonand Mrs . K. Uyldert for typing this manuscript .

References

Adams, T.L. & C.F . Quiros, 1985 . Somatic hybridization be-tween Lycopersicon peruvianum and L. pennellii : regenerat-ing ability and antibiotic resistance as selection systems . PlantScience 40 : 209-219 .

Ambros, P.F ., A .J.M. Matzke & M .A. Matzke, 1986 . Local-ization of Agrobacterium rhizogenes T-DNA in plant chromo-somes by in situ hybridization. EMBO J . 5 : 2073-2077 .

Atherton, J . G . & J . Rudich, 1986 . The Tomato Crop . Chapmanand Hall, London, New York, 661 pp .

Baker, B ., J . Schell, H . Ldrz & N . Federoff, 1986 . Trans-position of the maize controlling element 'Activator' in tobac-co. Proc . Nati. Acad . Sci . USA 83: 4844-4848 .

Baker, B ., G . Coupland, N . Federoff, P . Starlinger & J . Schell,1987 . Phenotypic assay for excision of the maize controllingelement Ac in tobacco . EMBO J . 6 : 1547-1554 .

Barker, D.M., M . Schafer & R. White, 1984 . Restriction sitescontaining CpG show a higher frequency of polymorphism inhuman DNA. Ceti 86 : 131-138 .

Barlow, D.P. & H . Lehrach, 1987 . Genetics by gelelectrophore-sis : the impact of pulsed field gel electrophoresis on mamma-lian genetics . Trends in Genetics 3 : 167-171 .

Barton, D.W.,1950 . Pachytene morphology of the tomato chro-mosome complement . Amer . J. Bot. 37 : 639-643 .

Beckmann, J .S. & M. Soller, 1983 . Restriction fragment lengthpolymorphisms in genetic improvement: methodologies,mapping and costs . Theor . Appl . Genet 67 : 35-43 .

Beckmann, J .S. & M. Soller, 1986 . Restriction fragment lengthpolymorphisms and genetic improvement of agricultural spe-cies . Euphytica 35 : 111-124 .

Behki, R.M. & S .M. Lesley, 1976. In vitro plant regenerationfrom leaf explants of Lycopersicon esculentum (tomato) .Can . J . Bot . : 2409-2414 .

Bennett, M.D., J .B . Smith & J .S . Heslop-Harrison, 1982 . Nu-clear DNA amounts in angiosperms . Proc . R. Soc . London B,216 : 179-192 .

Bennetzen, J .L . & T.L. Adams, 1984 . Selection and character-ization of cadmium-resistant suspension cultures of the wildtomato Lycopersicon peruvianum . Plant Cell Rep . 3 : 258-261 .

Bernatzky, R . & S.D. Tanksley, 1986a . Toward a saturatedlinkage map in tomato based on isozymes and random cDNAsequences . Genetics 112 : 887-898 .

Bernatzky, R . & S.D. Tanksley, 1986b . Majority of randomcDNA clones correspond to single loci in the tomato genome .Mol. Gen. Genet . 203 : 8-14 .

18

Bevan, M.W., R.B . Flavell & M.D. Chilton, 1983 . A chimaericantibiotic resistance gene as a selectable marker for plant celltransformation . Nature 304 : 184-187 .

Botstein, D ., R .L. White, M . Sholnick & R .W. Davis, 1980.Construction of a genetic linkage map in man using restrictionfragment length polymorphisms . Am . J. Hum. Genet . 32:314-331 .

Botterman, J . & J . Leemans, 1988 . Engineering herbicide re-sistance in plants . Trends in Genetics . 4 : 219-222.

Bressan, R.A ., P.M. Hasegawa & A.K. Handa, 1981 . Resist-ance of cultured higher plant cells to polyethylene glycol-induced water stress . Plant Science Lett . 21 : 23-30 .

Buiatti, M ., G . Marcheschi, F . Tognoni, M . Lipucci di Paola, T.Collina Grenci & G . Martini, 1985 . Genetic variability in-duced by tissue culture in the tomato (Lycopersicon esculen-tum) . Z. Pflanzenzuchtung 94 : 162-165 .

Buiatti, M ., C . Simeti, S . Vannini, G . Marcheschi, A . Scala, P.Bottini, P . Bogani & M.G. Pellegrini, 1987 . Isolation of to-mato cell lines with altered response to Fusarium cell wandcomponents . Theor. Appl . Genet . 75 : 37-40.

Burr, B ., S .V. Evola, F .A. Burr & J .S . Beckmann, 1983 . Theapplication of restriction fragment length polymorphism toplant breeding . In : J .K . Setlow & A . Hollander (Eds) . Ge-netic engineering : principles and methods, vol . 5, pp . 45-59,Plenum Publ . Corp .

Chang, C ., J .L. Bowman, A.W. DeJohn, E .S . Lander & E.M .Meyerowitz, 1988 . An RFLP linkage map to facilitate genecloning in Arabidopsis thaliana . Proc . Natl . Acad . Sci . USA .85 : 6856-6860 .

Chyi, Y.S ., R.A. Jorgensen, D . Goldstein, S .D . Tanksley & F.Loaiza-Figueroa, 1986 . Localization and stability of Agrobac-terium-mediated T-DNA insertions in the Lycopersicon ge-nome. Mol . Gen. Genet . 204 : 64-69 .

Crossway, A ., J .V. Oakes, J .M. Irvine, B . Ward, V.C . Knauf &C.K. Shewmaker, 1986 . Integration of foreign DNA follow-ing microinjection of tobacco mesophyll protoplasts . Mol .Gen. Genet . 202 : 179-185 .

Cuozzo, M ., K.M. O'Connell, W. Kaniewski, R .X . Fang, N .H.Chua & N.E. Turner, 1988 . Viral protection in transgenictobacco plants expressing the cucumber mosaic virus coatprotein or its antisense RNA . Biotechnology 6 : 549-557 .

Darden, K.A ., S .S . Smith & N .H. Murakishi, 1986 . Regener-ation and screening of tomato somaclones for resistance totobacco mosaic virus . Plant Science 45 : 209-213 .

De Block, M ., J . Botterman, M. van de Wiele, J . Dockx, C.Thoen, V. Gossele, N. Rao Movva, C . Thompson, M. vanMontagu & J. Leemans, 1987 . Engineering herbicide resist-ance in plants by expression of a detoxifying enzyme . EMBOJ . 6 : 2513-2518,

De Boer, D ., F . Cremers, R. Teertstra, L . Smits, J . Hille, S.Smeekens & P . Weisbeek, 1988 . In vivo import of plastocya-nin and a fusion protein into developmentally different plas-tids of transgenic plants . EMBO J . 7 : 2631-2635 .

De Framond, A.J ., K .A. Barton & M.D. Chilton, 1983 . Mini-Ti : A new vector strategy for plantgenetic engineering . Bio-technology 1 : 262-269 .

Dellaporta, S . & P . Chomet, 1986 . In : B . Hohn & E .S . Dennis(Eds). Genetic flux in plants . Springer Verlag, Vienna, NewYork, pp . 169-216 .

De Verna, J .W, R.T. Chetelat & C.M. Rick, 1987 . Cytogenet-ic, electrophoretic and morphological analysis of progeny ofsequidiploid Lycopersicon esculentum - Solanum lycopersi-coides hybrids x L. pennellii. Biol . Zent . bI . 106 : 417-428 .

Ecochard, R., M .S. Ramanna & D . De Nettancourt, 1969 .Detection and cytological analysis of tomato haploids . Genet-ica 40 : 181-190 .

Evans, D .A. & W.R. Sharp, 1983 . Single gene mutations intomato plants regenerated from tissue culture . Science 221 :949-951 .

Evans, D .A. & W.R. Sharp, 1986 . Applications of somaclonalvariation . Biotechnology 4 : 528-532 .

Federoff, N.V., D.B. Furtek & O.E . Nelson, 1984 . Cloning ofthe bronze locus in maize by a simple and generalizable proce-dure using the transposable controlling element Activator(Ac) . Proc. Natl Acad . Sci . USA 81 : 3825-3829 .

Fillatti, J .J ., J . Kiser, B . Rose & L . Comai, 1987. Efficienttransformation of tomato and the introduction and expressionof a gene for herbicide tolerance . In : D .J . Nevins & R .A .Jones (Eds) . Tomato Biotechnology, pp. 199-210, Alan R.Liss, Inc. New York .

Flavell, R.B ., 1982 . Chromosomal DNA sequences and theirorganisation. In : B . Parthier & D . Boulter (Eds) . NucleicAcids and Proteins in Plants II . Structure, Biochemistry andPhysiology of Nucleic Acids . pp . 46-74 . Springer Verlag,Berlin, Heidelberg and New York .

Fraley, R .T., R.B. Horsch, A . Matzke, M.D. Chilton, W.S .Chilton & P.R. Sanders, 1984 . In vitro transformation ofpetunia cells by an improved method of co-cultivation with A .tumefaciens strains . Plant . Molec. Biol . 3 : 371-378 .

Frankenberger, E.A ., P.M. Hasegawa & E .C. Tigchelaar,1981a . Influence of environmental and developmental stateon the shoot-forming capacity of tomato genotypes . Z. Pflan-zenphysiol . 102 : 221-232 .

Frankenberger, E.A ., P.M. Hasegawa & E .C. Tigchelaar,1981b. Diallel analysis of shootforming capacity among se-lected tomato genotypes . Z . Pflanzenphysiol . 102 : 233-242 .

Galbraith, D.W., K.R. Harkins, J.M. Maddox, N.M. Ayres,E. D . Sharma & E. Firoozabady, 1983 . Rapid flow cytometricanalysis of the cell cycle in intact plant tissues . Science 220 :1049-1051 .

Ganal, M.W., N.L .V. Lapitan & S .D. Tanksley, 1988 . A mole-cular and cytogenic survey of major repeated DNA sequencesin tomato (Lycopersicon esculentum) . Mol. Gen . Genet . 213 :262-268 .

Gavazzi, G ., C. Tonelli, G . Todesco, E. Arreghim, F . Raffaldi,F. Vecchio, S . Barbuzzi, M .S . Biasini & F . Sala, 1987 . Soma-clonal variation versus chemically induced mutagenesis intomato (Lycopersicon esculentum L.) Theor . Appl . Genet .74:733-738 .

Gerlach, W. L ., D . Llewellyn & J . Haseloff, 1987 . Constructionof a plant disease resistance gene from the satellite RNA oftobacco ringspot virus. Nature 328 : 802-805 .

Gill, B .S ., 1983 . Tomato cytogenetics - A search for new fron-tiers . In : M .S . Swaminathan, P .K. Gupta & U . Sinha (Eds) .Cytogenetics of Crop Plants pp . 457-480 . MacMillan IndiaLtd., New Delhi .

Gottschalk, W ., 1954 . Die Chromosomenstruktur der Solana-ceen unter Berficksichtigung Phytogenetischer Fragestellun-gen. Chromosoma 6 : 539-625 .

Greenblat, LM ., 1984 . A chromosome replication pattern de-duced from pericarp phenotypes resulting from movements ofthe transposable element modulator in maize . Genetics 108 :471-485 .

Gresshoff, P .M. & C.H. Doy, 1972 . Development and differ-entiation of haploid Lycopersicon esculentum (tomato) . Plan-ta 107 : 161-170 .

Gruissem, W., K . Callan, J . Lynch, T. Manzara, M . Meighan,B. Narita, B . Piechulla, M . Sugita, M. Thelander & L . Wan-ner, 1987 . Plastid and nuclear gene expression during tomatofruit formation . In : D .J . Nevins & R .A. Jones (Eds .) . Toma-to Biotechnology, pp . 239-249 . Alan R. Liss, Inc ., NewYork .

Guri, A ., A . Levi & K .C. Sink, 1988 . Morphological and mole-cular characterization of somatic hybrid plants between Lyco-persicon esculentum and Solanum nigura . Mol. Gen. Genet .212 : 191-198 .

Gusella, J .F ., 1986. DNA polymorphism and human disease .Ann. Rev . Biochem. 55 : 831-854 .

Gusella, J .F ., N.S. Wesxler, P.M. Conneally, S .L . Nayeor,M.A. Anderson, R .E. Tanzi, P.C. Watkins, K . Ottima,M.R. Wallace, A .Y. Sakaguicki, A.B.P.C. Young, I . Shoul-son, E . Bonilla & J .B . Martin, 1983 . A polymorphic DNAmarker linked to Huntington's disease . Nature 306 : 238-244 .

Hain, R., P . Stahel, A .P . Czernilofsky, H .H. Steinbiss, L .Herrera-Estrella & J . Schell, 1985 . Uptake, integration, ex-pression and genetic transmission of a selectable chimaericgene by plant protoplasts . Mol . Gen. Genet . 199 : 161-168 .

Handley, L.W., R.L . Nickels, M .W. Cameron, P .P. Moore &K.L . Sink, 1986. Somatic hybrid plants between Lycopersi-con esculentum and Solanum lycopersicofdes . Theor. Appl .Genet . 71 : 691-697 .

Hanson, M.R. & P.M. MacClean, 1987 . Tomato plastid andmitochondrial genomes : change and conservation . In : D .J .Nevins & R .A. Jones (Eds) . Tomato Biotechnology, pp .215-228, Alan R . Liss, Inc . New York.

Haring, M.A ., J . Gao, T. Volbeda, C .T.M. Rommens, H .J .J .Nijkamp & J . Hille, 1988 . A comparative study of Tam3 andAc transposition in trangenic tobacco and petunia plants .Plant Molec . Biol . in press .

Harrison, B.D ., M.A. Mayo & D .C. Baulcombe, 1987 . Virusresistance in transgenic plants that express cucumber mosaicvirus satellite RNA . Nature 328 : 799-802 .

Haseloff, J . & W.L . Gerlach, 1988 . Simple RNA enzymes withnew and highly specific endoribonuclease activities . Nature334: 585-591 .

Hause, B ., F. Baldauf, K . Stock, C . Wasternak & M . Metzlaff,1986. Molecular analysis of mitochondrial DNA from tomatocell suspension cultures . Curr . Genet . 10 : 785-790 .

19

Hawkes, J .H. & P . Smith, 1%5 . Continental drift and the age ofangiosperm genera . Nature (Lond) 207 : 48-50 .

Helentjaris, T., 1987 . A genetic linkage map for maize based onRFLPs . Trends in Genetics 3, 217-221 .

Helentjaris, T., G. King, M. Slocum, C . Siedenstrang & S .Wegman, 1985 . Restriction fragment polymorphisms asprobes for plant diversity and their development as tools forapplied breeding . Plant Molec . Biol . 5 : 109-118 .

Helentjaris, T., M . Slocum, S . Wright, A . Schaefer & J . Nien-huis, 1986 . Construction of genetic linkage maps in maize andtomato using restriction fragment length polymorphisms .Theor. Appl. Genet . 72 : 761-769 .

Herrera-Estrella, L ., M. de Block, E . Messens, J .P . Hernal-steens, M . van Montagu & J . Schell, 1983a. Chimeric genes asdominant selectable markers in plant cells . EMBO J . 2 : 987-995 .

Herrera-Estrella, L ., A. Depicker, M . van Montagu & J .Schell, 1983b . Expression of chimaeric genes transferred intoplant cells using a Ti-plasmid derived vector . Nature 303 :209-213 .

Hilder, V.A ., A.M.R. Gatehouse, S .E. Sheerman, R .F . Barker& D. Boulter, 1987 . A novel mechanism of insect resistanceengineered into tobacco . Nature 330 : 160-163 .

Hille, J ., A . Hoekema, P . Hooykaas & R. Schilperoort, 1984 .Gene organization of the Ti-plasmid . In : Verma, D.P .S .,Hohn, Th. (Eds) . Plant gene research : genes involved inmicrobe-plant interactions, pp . 287-309. Springer, Berlin .

Hille, J ., F . Verheggen, P. Roelvink, H. Franssen, A . vanKammen & P. Zabel, 1986 . Bleomycin resistance : a newdominant selectable marker for plant cell transformation .Plant Molec . Biol . 7 : 171-176 .

Hoekema, A ., P .R. Hirsch, P .J .J. Hooykaas & R.A. Schil-peroort, 1983 . A binary plant vector strategy based on sep-aration of the Vir- and T-region of the Agrobacterium tumefa-ciens Ti-plasmid, Nature 303 : 179-180 .

Horsch, R.B ., J .E . Fry, N.L. Hoffmann, M . Wallroth, D .Eichholtz, S .G. Rogers & R .T. Fraley, 1985 . A simple andgeneral method for transferring genes into plants . Science227:1229-1232 .

Hosticka, L.P. & M. Hanson, 1984 . Induction of plastid muta-tions in tomatoes by nitrosomethylurea . J . Heredity 75 : 242-246 .

Jongsma, M ., M . Koornneef, P . Zabel & J . Hille, 1987 . Tomatoprotoplast DNA transformation : physical linkage and recom-bination of exogenous DNA sequences . Plant Molec . Biol . 8 :383-394 .

Kartha, K.K ., O.L. Gamborg, J .P. Shyluk & F . Constabel,1976. Morphogenic investigations on in vitro leaf culture oftomato (Lycopersicon esculentum Mill cv Starfire) and highfrequency plant regeneration . Z. Pflanzenphysiol . 77 : 292-301 .

Khush, G .S. & C.M. Rick, 1963 . Meiosis in hybrids betweenLycopersicon esculentum and Solanum pennellii. Genetics 33 :167-183 .

Khush, G .S. & C.M. Rick, 1968 . Cytogenetic analysis of thetomato genome by means of induced deficiencies . Chromoso-ma 23 : 452-484 .

20

Kinshara, A., S .N. Patnaik, E .C. Cocking & J .B . Power, 1986 .Somatic hybrid plants of Lycopersicon esculentum Mill . andLycopersicon peruvianum Mill . J . Plant Physiol . 125 : 225-234 .

Klein, T.M ., T. Gradziel, M.E. Fromm & J.C. Sanford, 1988 .Factors influencing gene delivery into Zea mays cells by high-velocity microprojectiles . Biotechnology 6: 559-563 .

Koornneef, M., C .J. Hanhart, M . Jongsma, I . Toma, R .Weide, P . Zabel & J . Hille, 1986. Breeding of a tomatogenotype readily accessible to genetic manipulation . PlantScience 45 : 201-208 .

Koornneef, M ., C .J . Hanhart & L . Martinelli, 1987 . A geneticanalysis of cell culture traits in tomato . Theor . Appl. Genet .74: 633-641 .

Koornneef, M ., J .A.M. van Diepen, C.J. Hanhart, A.C. Kie-boom-de Waart, L. Martinelli, H .C.H. Schoenmakers & J .Wijbrandi, 1989 . Chromosomal instability of cell and tissuecultures of tomato haploids and diploids . Euphytica (inpress) .

Koukolikova-Nicola, Z ., L. Albright & B . Hohn, 1987 . Themechanism of T-DNA transfer from Agrobacterium tumefa-ciens to the plant cell . In: Th. Hohn and J . Schell (Eds) . PlantDNA Infectious Agents . pp . 109-148 . Springer Verlag, Wien,New York .

Landry, B.S . & R .W. Michelmore, 1987a . Selection of probes ofrestriction fragment length analysis from plant genomicclones . Plant Molecular Biology Reporter 3 : 174-179 .

Landry, B.S . & R .W. Michelmore, 1987b . Methods and applica-tions of restriction fragment length polymorphism analysis toplants . In : G . Bruening, J . Harada, T. Kosuge & A . Hollaen-ders (Eds). Tailoring genes for crop improvement . pp . 25-44., Plenum Publ . Corp .

Landry, B.S ., R.V. Kesseli, B . Farrara & R.W. Michelmore,1987a. A genetic map of lettuce (Lactuca sativa L.) withrestriction fragment length polymorphism, isozyme, diseaseresistance and morphological markers . Genetics 116 : 331-337 .

Landry, B .S ., R. Kesseli, H . Leung & R .W. Michelmore,1987b. Comparison of restriction endonucleases and sourcesof probes for their efficiency in detecting restriction fragmentlength polymorphisms in lettuce (Lactuca sativa L .) Theor .Appl. Genet . 74: 646-653 .

Levenko, B.A., V.A. Kunalki & G.M. Yurkova, 1977 . Studieson callus tissue from anthers I . Tomato. Phytomorphology27:377-383 .

Little, P ., G . Annison, S . Darling, R . Williamson, T. Cambar &B. Model, 1980 . Model for antenatal diagnosis of B-thalasse-mia and other monogenic disorders by molecular analysis oflinked DNA polymorphisms. Nature 285 : 144-147 .

Locky, R.D ., 1983. Callus formation and organogenesis byexplants of six Lycopersicon species . Can . J . Bot 61 : 1072-1079 .

Lonsdale, D.M ., T.P. Hodge & CM-R. Fauron, 1984 . Thephysical map and organization of the mitochondrial genomeof the fertile cytoplasm of maize . Nucleic Acids Rec . 12 :9249-9261 .

Manuelidis, L., 1982 . Repeated DNA sequences and nuclear .structure . In : G.A. Dover & R.B . Flavell (Eds) . Genomeevolution, pp . 263-185. Acad Press .

Marton, L., G .J . Wullems, L . Molendijk & R .A. Schilperoort,1979 In vitro transformation of cultured cells from Nicotianatabacum byAgrobacterium tumefaciens . Nature 277 : 129-131 .

Maxon-Smith, J .W. & E.D.B. Ritchie, 1983 . A collection ofnear isogenic lines of tomato . Research tool of the future?Plant Molec . Biol . 1 : 41-45 .

McClean, P.E. & M.R. Hanson, 1986 . Mitochondrial DNAsequence divergence among Lycopersicon and related Sola-num species . Genetics 112: 649-667 .

McClintock, B ., 1984. The significance of responses to chal-lenge . Science 226 : 792-801 .

McCormick, S ., J . Niedermeyer, A. Barnason, R . Horsch & R.Fraley, 1986 . Leaf disc transformation of cultivated tomato(L. esculentum) using Agrobacterium tumefaciens . Plant CellRep. 5 : 81-84 .

Melchers, G ., M .D. Sacristan & A.A. Holder, 1978 . Somatichybrid plants of potato and tomato regenerated from fusedprotoplasts . Carlsberg Res . Commun . 43 : 203-208 .

Menzel, M .Y., 1962. Pachytene chromosomes of the intergener-ic hybrid Lycopersicon esculentum x Solanum lycopersi-coides . Amer. J . Bot. 49 : 605-615 .

Meredith, C.P ., 1978. Selection and characterization of alumin-ium-resistant variants from tomato cell cultures . Plant Sci-ence Lett . 12 : 35-34 .

Miki, L.A ., T.J. Reich & V.N . Iyer, 1987. Microinjection : anexperimental tool for studying and modifying plant cells . In :Th. Hohn & J . Schell (Eds) . Plant DNA Infectious Agents .pp. 249-265 . Springer-Verlag, Wien, New York .

Moens, B.P., 1965 . The transmission of a heterochromatic iso-chromosome in Lycopersicon esculentum . Can. J . Genet .Cytol. 7 : 296-303 .

Morgan, A. & E.C. Cocking, 1982 . Plant regeneration fromprotoplasts of Lycopersicon esculentum Mill . Z . Pflanzen-physiol. 106 : 97-104 .

Mdhlbach, H .P ., 1980. Different regeneration potentials ofmesophyll protoplasts from cultivated and a wild species oftomato. Planta 148 : 89-96.

Mutschler, M., S .D. Tanksley & C.M. Rick, 1987 . Linkagemaps of the tomato (Lycopersicon esculentum) . Tomato Ge-net. Coop . 37 : 5-34 .

Negrutiu, I ., R . Shillito, I. Potrykus, G . Biasini & F . Sala, 1987 .Hybrid genes in the analysis of transformation conditions ISetting up a simple method for direct gene transfer in plantprotoplasts . Plant Molec . Biol . 8 : 363-373 .

Nelson, R.S ., S.M. McCormick, X . Delannay, P . Dubre, J .Layton, E .J. Anderson, M . Kanieska, R .K. Proksch, R.B .Horsch, S.G. Rogers, R.T. Fraley & R .N. Beachy, 1988 .Virus tolerance, plant growth and field performance of trans-genic tomato plants expressing coat protein from tobaccomosaic virus. Biotechnology . 6: 403-409 .

Niedz, R ., S.M. Rutter, L.W. Handley & K .C . Sink, 1985 . Plantregeneration from leaf protoplasts of six tomato cultivars .Plant Science 39, 199-204 .

O'Brien, S .J ., 1987 . Genetic maps . Cold Spring Harbor .O'Connell, M.A. & M.R. Hanson, 1985 . Somatic hybridizationbetween Lycopersicon esculentum and Lycopersicon pennel-lii . Theor. Appl . Genet . 70 : 1-12 .

O'Connell, M.A. & M.R. Hanson, 1986. Regeneration of so-matic hybrid plants formed between Lycopersicon esculen-tum and Solanum rickii . Theor. Appl. Genet 72 : 59-65 .

O'Connell, M.A. & M.R. Hanson, 1987 . Regeneration of so-matic hybrid plants formed between Lycopersicon esculen-tum and L. pennellii . Theor. Appl . Genet . 75 : 83-89 .

O'Connell, M.A ., L.P. Hosticka & M.R. Hanson, 1986 . Exam-ination of genome stability in cultured Lycopersicon . PlantCell Reports 5 : 276-279 .

Ohki, S., C . Bigot & J . Mouseau, 1978 . Analysis of shoot-forming capacity in vitro in two lines of tomato (Lycopersiconesculentum Mill .) and their hybrids . Plant & Cell Physiol 19 :27-42 .

Ohyama, K ., H . Fukuzawa, T. Kolchi, H . Shizai, T. Sano, S .Sano, K. Umesono, Y. Shiki, M . Takeuchi, Z . Chang, S .Aota, H. Inokuchi & H . Ozeki, 1986 . Chloroplast gene orga-nization deduced from complete sequence of liverwort Mar-chantia polymorpha chloroplast DNA . Nature