S. D. Tanksley

Advanced backcross QTL analysis: a method for the simultaneous discovery and transfer of valuable QTLs from unadapted germplasm into elite breeding lines

Advanced backcross QTL analysis is proposed as a method of combining QTL analysis with variety development. It is tailored for the discovery and transfer of valuable QTL alleles from unadapted donor lines (e.g., land races, wild species) into established elite inbred lines. Following this strategy, QTL analysis is delayed until the BC2 or BC3 generation and, during the development of these populations, negative selection is exercised to reduce the frequency of deleterious donor alleles. Simulations suggest that advanced backcross QTL analysis will be effective in detecting additive, dominant, partially dominant, or overdominant QTLs. Epistatic QTLs or QTLs with gene actions ranging from recessive to additive will be detected with less power than in selfing generations. QTL-NILs can be derived from advanced backcross populations in one or two additional generations and utilized to verify QTL activity. These same QTL-NILs also represent commercial inbreds improved (over the original recurrent inbred line) for one or more quantitative traits. The time lapse from QTL discovery to construction and testing of improved QTL-NILs is minimal (1–2 years). If successfully employed, advanced backcross QTL analysis can open the door to exploiting unadapted and exotic germplasm for the quantitative trait improvement of a number of crop plants.

RFLP analysis of phylogenetic relationships and genetic variation in the genus Lycopersicon

SummaryForty single-copy, nuclear probes of known chromosomal position were used to examine restriction fragment length polymorphism in the tomato genus Lycopersion. The probes were from three libraries: one cDNA, and two genomic libraries ⦓ne genomic made with EcoRI and the other with PstI. Total DNA from 156 plants representing eight species was cut with five different restriction enzymes and scored in 198 probe-enzyme combinations. Genetic distances between accessions (populations) and species were calculated from the resultant restriction patterns and proportion of shared bands. Accessions belonging to the same species largely clustered together, confirming their current classification. However, one mountain accession, classified as L. peruvianum var. humifusum (LA2150), was sufficiently distinct from the other accessions of L. peruvianum that it may qualify as a separate species L. esculentum and L. pimpinellifolium were the least clearly differentiated, possibly reflecting introgressive hybridization, known to have been promoted by man in recent history. Dendrograms constructed from cDNA versus genomic clones were nearly identical in their general grouping of species. The dendrograms revealed two major dichotomies in the genus: one corresponding to mating behavior [self-compatible (SC) versus self-incompatible (SI) species] and the other corresponding to fruit color (red versus green-fruited species). The ratio of withinversus between-accession diversity was much lower for SC species, indicating that most of the diversity within these species exists between populations, rather than within populations. Overall, the amount of genetic variation in the SI species far exceeded that found in SC species. This result is exemplified by the fact that more genetic variation could be found within a single accession of one of the SI species (e.g., L. peruvianum) than among all accessions tested of any one of the SC species (e.g., L. esculentum or L. pimpinellifolium). Results from this study are discussed in relationship to germ plasm collection/utilization and with regard to the use of RFLPs in tomato breeding and genetics.

Survey of plant short tandem DNA repeats

Length variations in simple sequence tandem repeats are being given increased attention in plant genetics. Some short tandem repeats (STRs) from a few plant species, mainly those at the dinucleotide level, have been demonstrated to show polymorphisms and Mendelian inheritance. In the study reported here a search for all of the possible STRs ranging from mononucleotide up to tetranucleotide repeats was carried out on EMBL and GenBank DNA sequence databases of 3026 kb nuclear DNA and 1268 kb organelle DNA in 54 and 28 plant species (plus algae), respectively. An extreme rareness of STRs (4 STRs in 1268 kb DNA) was detected in organelle compared with nuclear DNA sequences. In nuclear DNA sequences, (AT)n sequences were the most abundant followed by (A)n · (T)n, (AG)n · (CT)n, (AAT)n · (ATT)n, (AAC)n · (GTT), (AGC)n · (GCT)n, (AAG)n · (CTT)n, (AATT)n · (TTAA)n, (AAAT)n · (ATTT)n and (AC)n · (GT)n sequences. A total of 130 STRs were found, including 49 (AT)n sequences in 31 species, giving an average of 1 STR every 23.3 kb and 1 (AT)n STR every 62 kb. An abundance comparable to that for the dinucleotide repeat was observed for the tri- and tetranucleotide repeats together. On average, there was 1 STR every 64.6 kb DNA in monocotyledons versus 1 every 21.2 kb DNA in dicotyledons. The fraction of STRs that contained G-C basepairs increased as the G+C contents went up from dicotyledons, monocotyledons to algae. While STRs of mono-, di- and tetranucleotide repeats were all located in non coding regions, 57% of the trinucleotide STRs containing G-C basepairs resided in coding regions.

Advanced backcross QTL analysis in a cross between an elite processing line of tomato and its wild relative L. pimpinellifolium.

Approximately 170 BC2 plants from a cross between an elite processing inbred (recurrent parent) and the wild species Lycopersicon pimpinellifolium LA1589 (donor parent) were analyzed with segregating molecular markers covering the entire tomato genome. Marker data were used to identify QTLs controlling a battery of horticultural traits measured on BC2F1 and BC3 families derived from the BC2 individuals. Despite its overall inferior appearance, L. pimpinellifolium was shown to possess QTL alleles capable of enhancing most traits important in processing tomato production. QTL-NIL lines, containing specific QTLs modifying fruit size and shape, were subsequently constructed and shown to display the transgressive phenotypes predicted from the original BC2 QTL analysis. The potential of exploiting unadapted and wild germplasm via advanced backcross QTL analysis for the enhancement of elite crop varieties is discussed.

Identification of QTLs affecting traits of agronomic importance in a recombinant inbred population derived from a subspecific rice cross

To detect QTLs controlling traits of agronomic importance in rice, two elite homozygous lines 9024 and LH422, which represent the indica and japonica subspecies of rice (Oryza sativa), were crossed. Subsequently a modified single-seed-descent procedure was employed to produce 194 recombinant inbred lines (F8). The 194 lines were genotyped at 141 RFLP marker loci and evaluated in a field trial for 13 quantitative traits including grain yield. Transgressive segregants were observed for all traits examined. The number of significant QTLs (LOD ⩾ 2.0) detected affecting each trait ranged from one to six. The percentage of phenotypic variance explained by each QTL ranged from 5.1% to 73.7%. For those traits for which two or more QTLs were detected, increases in the traits were conditioned by indica alleles at some QTLs Japonica alleles at others. No significant evidence was found for epistasis between markers associated with QTLs and all the other markers. Pleitropic effects of single QTLs on different traits are suggested by the observation of clustering of QTLs. No QTL for traits was found to map to the vicinity of major gene loci governing the same traits qualitatively. Evidence for putative orthologous QTLs across rice, maize, oat, and barley is discussed.

Polymorphism and phylogenetic relationships among species in the genus Oryza as determined by analysis of nuclear RFLPs

SummaryNinety-three accessions representing 21 species from the genus Oryza were examined for restriction fragment length polymorphism. The majority (78%) of the accessions, for which five individuals were tested, were found to be monomorphic. Most of the polymorphic accessions segregated for only one or two probes and appeared to be mixed pure lines. For most of the Oryza species tested, the majority of the genetic variation (83%) was found between accessions from different species with only 17% between accessions within species. Tetraploid species were found to have, on average, nearly 50% more alleles (unique fragments) per individual than diploid species reflecting the allopolyploid nature of their genomes.Classification of Oryza species based on RFLPs matches remarkably well previous classifications based on morphology, hybridization and isozymes. In the current study, four species complexes could be identified corresponding to those proposed by Vaughan (1989): the O. ridleyi complex, the O. meyeriana complex, the O. officinalis complex and the O. sativa complex. Within the O. sativa complex, accessions of O. rufipogon from Asia (including O. nivara) and perennial forms of O. rufipogon from Australia clustered together with accessions of cultivated rice O. sativa. Surprisingly, indica and japonica (the two major subspecies of cultivated rice) showed closer affinity with different accessions of wild O. Rufipogon than to each other, supporting a hypothesis of independent domestication events for these two types of rice. Australian annual wild rice O. meridionalis (previously classified as O. rufipogon) was clearly distinct from all other O. rufipogon accessions supporting its recent reclassification as O. meridionalis (Ng et al. 1981). Using genetic relatedness as a criterion, it was possible to identify the closest living diploid relatives of the currently known tetraploid rice species. Results from these analyses suggest that BBCC tetraploids (O. malampuzhaensis, O. punctata and O. minuta) are either of independent origins or have experienced introgression from sympatric C-genome diploid rice species. CCDD tetraploid species from America (O. latifolia, O. alta and O. grandiglumis) may be of ancient origin since they show a closer affinity to each other than to any known diploid species. Their closest living diploid relatives belong to C genome (O. eichingeri) and E genome (O. Australiensis) species. Comparisons among African, Australian and Asian rice species suggest that Oryza species in Africa and Australia are of polyphyletic origin and probably migrated to these regions at different times in the past.Finally, on a practical note, the majority of probes used in this study detected polymorphism between cultivated rice and its wild relatives. Hence, RFLP markers and maps based on such markers are likely to be very useful in monitoring and aiding introgression of genes from wild rice into modern cultivars.

Advanced backcross QTL analysis in tomato. I. Identification of QTLs for traits of agronomic importance from Lycopersicon hirsutum

Abstract Advanced backcross QTL (AB-QTL) analysis is a new strategy for studying the effect of unadapted alleles on the agronomic performance of elite cultivated lines. In this paper we report results from the application of the AB-QTL strategy to cultivated tomato using the wild species Lycopersicon hirsutum LA1777 as the donor parent. RFLP genomic fingerprints were determined for 315 BC2 plants and phenotypic data were collected for 19 agronomic traits from approximately 200 derived BC3 lines which were grown in replicated field trials in three locations worldwide. Between 1 and 12 significant QTLs were identified for each of the 19 traits evaluated, with a total of 121 QTLs identified for all traits. For 25 of the QTLs (20%) corresponding to 12 traits (60%), the L. hirsutum allele was associated with an improvement of the trait from a horticultural perspective, despite the fact that L. hirsutum is overall phenotypically inferior to the elite parent. For example, L. hirsutum has fruit that remains green when ripe (lack of red pigment) yet alleles were found in this species that significantly increase red color when transferred into cultivated tomatoes. Wild alleles were also associated with increases in total yield and soluble solids (up to 15%) and brix×red yield (up to 41%). These results support the idea that one cannot predict the genetic potential of exotic germplasm based on phenotype alone and that marker-based methods, such as the AB-QTL strategy, should be applied to fully exploit exotic germplasm.

Identifying the loci responsible for natural variation in fruit size and shape in tomato

Abstract Fruit size and shape are two major factors determining yield, quality and consumer acceptability for many crops. Like most traits important to agriculture, both are quantitatively inherited. Despite their economic importance none of the genes controlling either of these traits have been cloned, and little is known about the control of the size and shape of domesticated fruit. Tomato represents a model fruit-bearing domesticated species characterized by a wide morphological diversity of fruits. The many genetic and genomic tools available for this crop can be used to unraveal the molecular bases of the developmental stages which presumably influence fruit architecture, size and shape. The goal of this review is to summarize data from the tomato QTL studies conducted over the past 15 years, which together allow the identification of the major QTLs responsible for fruit domestication in tomato. These results provide the starting point for the isolation of the genes involved in fruit-size/shape determination in tomato and potentially other fruit-bearing plants.

QTL analysis of an advanced backcross of Lycopersicon peruvianum to the cultivated tomato and comparisons with QTLs found in other wild species

Abstract A BC3 population previously developed from a backcross of Lycopersicon peruvianum, a wild relative of tomato, into the cultivated variety L. esculentum was analyzed for QTLs. Approximately 200 BC4 families were scored for 35 traits in four locations worldwide. One hundred and sixty-six QTLs were detected for 29 of those traits. For more than half of those 29 traits at least 1 QTL was detected for which the presence of the wild allele was associated with an agronomically beneficial effect despite the inferior phenotype of the wild parent. Eight QTLs for fruit weight could be followed through the BC2, BC3, and BC4, generations, supporting the authenticity of these QTLs. Comparisons were made between the QTLs found in this study and those found in studies involving two other wild species; the results showed that while some of these QTLs can be presumed to be allelic, most of the QTLs detected in this study are ones not previously discovered.

Tagging genes for blast resistance in rice via linkage to RFLP markers.

SummaryBoth Pi-2(t) and Pi-4(t) genes of rice confer complete resistance to the blast fungal pathogen Pyricularia oryzae Cav. As economically important plant genes, they have been recently characterized phenotypically, yet nothing is known about their classical linkage associations and gene products. We report here the isolation of DNA markers closely linked to these blast resistance genes in rice. The DNA markers were identified by testing 142 mapped rice genomic clones as hybridization probes against Southern blots, consisting of DNA from pairs of nearly isogenic lines (NILs) with or without the target genes. Chromosomal segments introgressed from donor genomes were distinguished by restriction fragment length polymorphisms (RFLPs) between the NILs. Linkage associations of the clones with Pi-2(t) and Pi4(t) were verified using F3 segregating populations of known blast reaction. Cosegregation of the resistant genotype and donor-derived allele indicated the presence of linkage between the DNA marker and a blast resistance gene. RFLP analysis showed that Pi-2(t) is closely linked to a single-copy DNA clone RG64 on chromosome 6, with a distance of 2.8+1.4(SE) cMorgans. Another blast resistance gene, Pi-4(t), is 15.3+4.2(SE) cMorgans away from a DNA clone RG869 on chromosome 12. These chromosomal regions can now be examined with additional markers to define the precise locations of Pi-2(t) and Pi-4(t). Tightly linked DNA markers may facilitate early selection for blast resistance genes in breeding programs. These markers may also be useful to map new genes for resistance to blast isolates. They may ultimately lead to the cloning of those genes via chromosome walking. The gene tagging approach demonstrated in this paper may apply to other genes of interest for both monogenic and polygenic traits.