Genetic analysis of rootstock-mediated nitrogen (N) uptake and root-to-shoot signalling at contrasting N availabilities in tomato.

Selecting rootstocks for high nitrogen acquisition ability may allow decreased N fertilizer application without reducing tomato yields, minimizing environmental nitrate pollution. A commercial hybrid tomato variety was grafted on a genotyped population of 130 recombinant inbred lines (RILs) derived from Solanum pimpinellifolium, and compared with self- and non-grafted controls under contrasting nitrate availabilities (13.8 vs 1.0mM) in the nutrient solution. Grafting itself altered xylem sap composition under N-sufficient conditions, particularly Na+ (8.75-fold increase) concentration. N deprivation decreased shoot dry weight by 72.7% across the grafted RIL population, and one RIL rootstock allowed higher total leaf N content than the best of controls, suggesting more effective N uptake. Sixty-two significant QTLs were detected by multiple QTL mapping procedure for leaf N concentration (LNC), vegetative growth, and the xylem sap concentrations of Mn and four phytohormone groups (cytokinins, gibberellins, salicylic acid and jasmonic acid). Only three LNC QTLs could be common between nitrogen treatments. Clustering of rootstock QTLs controlling LNC, leaf dry weight and xylem sap salicylic acid concentration in chromosome 9 suggests a genetic relationship between this rootstock phytohormone and N uptake efficiency. Some functional candidate genes found within 2 Mbp intervals of LNC and hormone QTLs are discussed.

Genetic dissection of tomato rootstock effects on scion traits under moderate salinity.

KEY MESSAGE: Rootstock HKT1 genotype affected fruit [Na(+)] and non-commercial fruit yield; QTL analysis of rootstock-mediated scion nutrition is a powerful forward genetic approach to identify wild genes for rootstock breeding. The present study approaches the QTL dissection of rootstock effects on a commercial hybrid variety grafted on a population of RILs derived from Solanum pimpinellifolium, genotyped for 4370 segregating SNPs from the SolCAP tomato panel and grown under moderate salinity. Results are compared to those previously obtained under high salinity. The most likely functional candidate genes controlling the scion [Na(+)] were rootstock HKT1;1 and HKT1;2 as it was previously reported for non-grafted genotypes. The higher fruit [Na(+)] found when rootstock genotype was homozygote for SpHKT1 supports the thesis that scion HKT1 is loading Na(+) into the phloem sap in leaves and unloading it in sink organs. A significant increment of small, mostly seedless, fruits was found associated with SlHKT1 homozygous rootstocks. Just grafting increased the incidence of blossom end rot and delayed fruit maturation but there were rootstock RILs that increased commercial fruit yield under moderate salinity. The heritability and number of QTLs involved were lower and different than those found under high salinity. Four large contributing (>17%) rootstock QTLs, controlling the leaf concentrations of B, K, Mg and Mo were detected whose 2 Mbp physical intervals contained B, K, Mg and Mo transporter-coding genes, respectively. Since a minimum of 3 QTLs (two of them coincident with leaf K and Ca QTLs) were also found governing rootstock-mediated soluble-solids content of the fruit under moderate salinity, grafting desirable crop varieties on stress-tolerant rootstocks tenders an opportunity to increase both salt tolerance and quality.

Genetic analysis of physiological components of salt tolerance conferred by Solanum rootstocks. What is the rootstock doing for the scion?

Grafting desirable crop varieties on stress-tolerant rootstocks provides an opportunity to increase crop salt tolerance. Here, a commercial hybrid tomato variety was grafted on two populations of recombinant inbred lines developed from a salt-sensitive genotype of Solanum lycopersicum var. cerasiforme, as female parent, and two salt-tolerant lines, as male parents, from S. pimpinellifolium, the P population, and S. cheesmaniae, the C population, to identify an easy screening method for identifying rootstocks conferring salt tolerance in terms of fruit yield. Potential physiological components of salt tolerance were assessed in the scion: leaf biomass, [Na(+)], nutrition, water relations and xylem ABA concentration. A significant correlation between scion fruit yield and scion leaf fresh weight, water potential or the ABA concentration was found in the C population under salinity, but the only detected QTL did not support this relationship. The rootstocks of the P population clearly affected seven traits related to the sodium, phosphorous and copper concentrations and water content of the scion leaf, showing heritability estimates around 0.4 or higher. According to heritability estimates in the P population, up to five QTLs were detected per trait. QTLs contributing over 15% to the total variance were found for P and Cu concentrations and water content of the scion leaf, and the proportion of fresh root weight. Correlation and QTL analysis suggests that rootstock-mediated improvement of fruit yield in the P population under salinity is mainly explained by the rootstock's ability to minimise perturbations in scion water status.

Identification of fruit yield loci controlling the salt tolerance conferred by solanum rootstocks.

The rootstock effect on the fruit yield of a grafted tomato variety was genetically analyzed under salinity using as rootstock two populations of F(9) lines developed from a salt sensitive genotype of Solanum lycopersicum var. cerasiforme, as female parent, and two salt tolerant lines, as male parents, from S. pimpinellifolium, the P population (123 lines), and S. cheesmaniae, the C population (100 lines). There were rootstock lines from the two populations (up to 65% in the P population) that raised the fruit yield of the commercial hybrid under saline conditions. It is shown that this salt tolerance rootstock effect is a heritable trait (h (2) near 0.3), governed by at least eight QTLs. The most relevant component was the number of fruits. Thus most detected QTLs correspond to this component. In general, QTL gene effects are medium-sized, with contributions from 8.5 up to 15.9% at most, and the advantageous allele comes from the wild, salt tolerant species. Only two fruit yield QTLs on chromosomes P9 and C11 might correspond to fruit yield QTLs of the non-grafted lines indicating their root system dependence. A fruit yield QTL on chromosome 3 is acting epistatically in both populations. The epistatic interactions found were dominant and they were unveiled using the associated marker as cofactor in the composite interval mapping methodology. Therefore, an efficient and profitable utilization of wild germplasm can be carried out through the improvement of rootstocks that confer salt tolerance in terms of fruit yield to the grafted variety.

Genetic analysis of Na(+) and K (+) concentrations in leaf and stem as physiological components of salt tolerance in Tomato.

The sodium and potassium concentrations in leaf and stem have been genetically studied as physiological components of the vegetative and reproductive development in two populations of F(8) lines, derived from a salt sensitive genotype of Solanum lycopersicum cv. Cerasiforme, as female parent, and two salt tolerant lines, as male parents, from S. pimpinellifolium, the P population (142 lines), and S. cheesmaniae, the C population (116 lines). Genetic parameters of ten traits under salinity and five of them under control conditions were studied by ANOVA, correlation, principal component and QTL analysis to understand the global response of the plant. Two linkage maps including some tomato flowering time and salt tolerance candidate genes encoding for SlSOS1, SlSOS2, SlSOS3, LeNHX1, LeNHX3, were used for the QTL detection. Thirteen and 20 QTLs were detected under salinity in the P and C populations, respectively, and four under control conditions. Highly significant and contributing QTLs (over 40%) for the concentrations of Na(+) and K(+) in stems and leaves have been detected on chromosome 7 in both the populations. This is the only genomic position where the concentration QTLs for both the cations locate together. The proportion of QTLs significantly affected by salinity was larger in the P population (64.3%, including all QTLs detected under control) than in the C population (21.4%), where the estimated genetic component of variance was larger for most traits. A highly significant association between the leaf area and fruit yield under salinity was found only in the C population, which is supported by the location of QTLs for these traits in a common region of chromososome C1. As far as breeding for salt tolerance is concerned, only two sodium QTLs (lnc1.1 and lnc8.1) map in genomic regions of C1 and C8 where fruit yield QTLs are also located but in both the cases the profitable allele corresponds to the salt sensitive, cultivated species. One of those QTLs, lnc1.1 might involve LeNHX3.

Comparative QTL analysis of salinity tolerance in terms of fruit yield using two Solanum populations of F7 lines.

Salt tolerance has been analysed in two populations of F(7) lines developed from a salt sensitive genotype of Solanum lycopersicum var. cerasiforme, as female parent, and two salt tolerant lines, as male parents, from S. pimpinellifolium, the P population (142 lines), and S. cheesmaniae, the C population (116 lines). Salinity effects on 19 quantitative traits including fruit yield were investigated by correlation, principal component analysis, ANOVA and QTL analysis. A total of 153 and 124 markers were genotyped in the P and C populations, respectively. Some flowering time and salt tolerance candidate genes were included. Since most traits deviated from a normal distribution, results based on the Kruskal-Wallis non-parametric test were preferred. Interval mapping methodology and ANOVA were also used for QTL detection. Eight out of 15 QTLs at each population were detected for the target traits under both control and high salinity conditions, and among them, only average fruit weight (FW) and fruit number (FN) QTLs (fw1.1, fw2.1 and fn1.2) were detected in both populations. The individual contribution of QTLs were, in general, low. After leaf chloride concentration, flowering time is the trait most affected by salinity because different QTLs are detected and some of their QTLxE interactions have been found significant. Also reinforcing the interest on information provided by QTL analysis, it has been found that non-correlated traits may present QTL(s) that are associated with the same marker. A few salinity specific QTLs for fruit yield, not associated with detrimental effects, might be used to increase tomato salt tolerance. The beneficial allele at two of them, fw8.1 (in C) and tw8.1 (for total fruit weight in P) corresponds to the salt sensitive parent, suggesting that the effect of the genetic background is crucial to breed for wide adaptation using wild germplasm.

Increasing salt tolerance in the tomato.

In this paper, a number of strategies to overcome the deleterious effects of salinity on plants will be reviewed; these strategies include using molecular markers and genetic transformation as tools to develop salinity-tolerant genotypes, and some cultural techniques. For more than 12 years, QTL analysis has been attempted in order to understand the genetics of salt tolerance and to deal with component traits in breeding programmes. Despite innovations like better marker systems and improved genetic mapping strategies, the success of marker-assisted selection has been very limited because, in part, of inadequate experimental design. Since salinity is variable in time and space, experimental design must allow the study of genotype x environment interaction. Genetic transformation could become a powerful tool in plant breeding, but the growing knowledge from plant physiology must be integrated with molecular breeding techniques. It has been shown that the expression of several transgenes promotes a higher level of salt tolerance in some species. Despite this promising result, the development of a salt-tolerant cultivar by way of transgenesis has still not been achieved. Future directions in order to overcome the present limitations are proposed. Three cultural techniques have proved useful in tomato to overcome, in part, the effects of salinity: treatment of seedlings with drought or NaCl ameliorates the adaptation of adult plants to salinity; mist applied to tomato plants grown in Mediterranean conditions improves vegetative growth and yield in saline conditions; and grafting tomato cultivars onto appropriate rootstocks could reduce the effects of salinity.

Genetic analysis of citrus leafminer susceptibility.

Damage caused by the citrus leafminer (CLM), Phyllocnistis citrella, is highly dependent on the citrus flushing pattern. Chemical control is only required in young trees, both in nurseries and in newly established orchards. However, this situation is completely different in countries where the causal agent of citrus canker, the bacterium Xanthomonas axonopodis pv. citri exists. CLM infestation results in a higher incidence of citrus canker infection. Among preventive control strategies that provide environmentally sound and sustainable solutions, resistant or tolerant varieties remain the most economical means of insect control. The objective of the present study is to genetically analyse the resistance/susceptibility to CLM and two other traits that might be related, the deciduous behaviour and leaf area of the tree, in a progeny of citradias derived from the cross between two species with different CLM susceptibility--C. aurantium L. and Poncirus trifoliata--using linkage maps of each parent that include several resistance gene analogues. We detected two antibiosis and six antixenosis putative quantitative trait loci (QTLs) in a random sample of forty-two of those citradias. An important antibiosis QTL (R2=18.8-26.7%) affecting both percentage of infested leaves and number of pupal casts per leaf has been detected in P. trifoliata linkage group Pa7, which is in agreement with the CLM antibiotic character shown by this species, and independent from any segregating QTL involved in its deciduous behaviour. The maximum value for the Kruskal-Wallis statistic of the other putative antibiosis QTL coincides with marker S2-AS4_800 in sour orange linkage map. Given that the sequence of this marker is highly similar to several nucleotide binding site-leucine-rich repeat (NBS-LRR)-type resistance genes, it might be considered as a candidate gene for insect resistance in citrus.

Comparative microsatellite linkage analysis and genetic structure of two populations of F6 lines derived from Lycopersicon pimpinellifolium and L. cheesmanii.

A population of recombinant inbred lines (RILs) has several advantages over its F2 population counterpart with respect to quantitative trait loci (QTLs) and genomic studies. The objective of the investigation reported here was the comparative characterization by simple sequence repeat (SSR) and sequence characterized amplified region (SCAR) markers of two populations of F6 lines derived from Lycopersicon pimpinellifolium (P population, consisting of 142 lines) and L. cheesmanii (C population, consisting of 115 lines) and sharing the female parent, L. esculentum var. cerasiforme. Almost the same percentage of polymorphic markers was found for each population although a different set of markers was involved. The proportion of SSR primer pairs (93 in total) that resulted in polymorphism for the main band was larger (55-56%) than for SCAR ones (13-16%). The C population showed the largest proportion of markers with zygotic and gametic segregation distortion, which is in agreement with the larger genetic distance reported between L. esculentum and L. cheesmanii than with the former and L. pimpinellifolium. Zygotic distortion corresponded primarily to an excess of heterozygotes in both populations, suggesting that the increment of homozygosity was the main factor limiting viability/self-fertility of the lines. Despite both populations sharing the female parent, P alleles were slightly favored in the P population while E alleles were the most frequently fixed in the C population. A linkage map for each population was obtained, with the average distances between consecutive markers being 3.8 cM or 3.4 cM depending on the population. Discrepancy between the maps for the location of only four markers on chromosomes 3, 6 and 10 was observed. Two possible causes of this discrepancy were investigated and can not be discarded: (1) the presence of duplicated markers and (2) segregation distortion caused by the selective advantage of gametes carrying one of the two alleles. This marker characterization of both populations will continue and will enable the comparative QTLs and candidate gene analysis of complex traits towards a more efficient utilization of genetic resources and breeding strategies.

QTL analysis of citrus tristeza virus-citradia interaction.

Citrus tristeza virus (CTV) has caused the death of millions of trees grafted on sour orange ( Citrus aurantium). However, this rootstock is very well adapted to the Mediterranean, semi-arid conditions. The aim of the present research is to genetically analyze the accumulation of CTV in a progeny derived from the cross between C. aurantium and Poncirus trifoliata, both resistant to CTV isolate T-346. Graft propagation of 104 hybrids was done on healthy sweet orange as a rootstock. Three months later, each rootstock was graft inoculated with two patches of infected tissue (isolate T-346). One, 2, and sometimes, 3 and 4 years after inoculation, hybrids and infected patches were tested for CTV by tissue-blot immuno-assay. Additionally, CTV multiplication was evaluated every year as the optical density of double-antibody sandwich enzyme-linked immuno-sorbent assay reactions. Linkage maps for P. trifoliata based on 63 markers, and for C. aurantium based on 157 markers, were used. Most molecular markers were microsatellites and IRAP (inter-retrotransposon amplified polymorphisms). Some analogues of resistance and expressed sequences were also included for candidate gene analysis. Resistance against CTV was analyzed as a quantitative trait (CTV accumulation) by QTL (quantitative trait loci) analysis to avoid the assumption of monogenic control. Three major resistance QTLs were detected where the P. trifoliata resistance gene, Ctv-R, had been previously located in other progenies. Up to five minor QTLs were detected ( Ctv-A(1) to Ctv-A(5)). A significant epistatic interaction involving Ctv-R(1) and Ctv-A(1) was also found. An analogue of a resistance gene is a candidate for Ctv-A(3), and two expressed sequences are candidates for Ctv-A(1) and Ctv-A(5). Single-strand conformational polymorphism analysis of CTV genes QTL P20 and P25 (coat protein) in susceptible hybrids, was carried out to test whether or not any QTL accumulation was a defeated resistance gene. Since the same haplotype of the virus was visualized independently on the CTV titer, differences in the amount of virions are not explained through the selection of CTV genotypes by the host, but through differences among citradias in CTV replication and/or movement.

Expressed sequence enrichment for candidate gene analysis of citrus tristeza virus resistance.

Several studies have reported markers linked to a putative resistance gene from Poncirus trifoliata ( Ctv-R) located at linkage group 4 that confers resistance against one of the most important citrus pathogens, citrus tristeza virus (CTV). To be successful in both marker-assisted selection and transformation experiments, its accurate mapping is needed. Several factors may affect its localization, among them two are considered here: the definition of resistance and the genetic background of progeny. Two progenies derived from P. trifoliata, by self-pollination and by crossing with sour orange ( Citrus aurantium), a citrus rootstock well-adapted to arid and semi-arid areas, were used for linkage group-4 marker enrichment. Two new methodologies were used to enrich this region with expressed sequences. The enrichment of group 4 resulted in the fusion of several C. aurantium linkage groups. The new one A(7+3+4) is now saturated with 48 markers including expressed sequences. Surprisingly, sour orange was as resistant to the CTV isolate tested as was P. trifoliata, and three hybrids that carry Ctv-R, as deduced from its flanking markers, are susceptible to CTV. The new linkage maps were used to map Ctv-R under the hypothesis of monogenic inheritance. Its position on linkage group 4 of P. trifoliata differs from the location previously reported in other progenies. The genetic analysis of virus-plant interaction in the family derived from C. aurantium after a CTV chronic infection showed the segregation of five types of interaction, which is not compatible with the hypothesis of a single gene controlling resistance. Two major issues are discussed: another type of genetic analysis of CTV resistance is needed to avoid the assumption of monogenic inheritance, and transferring Ctv-R from P. trifoliata to sour orange might not avoid the CTV decline of sweet orange trees.

Identification and genomic distribution of gypsy like retrotransposons in Citrus and Poncirus.

Transposable elements might be importantly involved in citrus genetic instability and genome evolution. The presence of gypsy like retrotransposons, their heterogeneity and genomic distribution in Citrus and Poncirus, have been investigated. Eight clones containing part of the POL coding region of gypsy like retrotransposons have been isolated from a commercial variety of Citrus clementina, one of the few sexual species in Citrus. Four of the eight clones might correspond to active elements given that they present all the conserved motifs described in the literature as essential for activity, no in-frame stop codon and no frame-shift mutation. High homology has been found between some of these citrus elements and retroelements within a resistance-gene cluster from potato, another from Poncirus trifoliata and two putative resistance polyproteins from rice. Nested copies of gypsy like elements are scattered along the Citrus and Poncirus genomes. The results on genomic distribution show that these elements were introduced before the divergence of both genera and evolved separately thereafter. IRAPs based on gypsy and copia types of retrotransposons seem to distribute differently, therefore gypsy based IRAPs prove a new, complementary set of molecular markers in Citrus to study and map genetic variability, especially for disease resistance. Similarly to copia-derived IRAPs, the number of copies and heterozygosity values found for gypsy derived IRAPs are lower in Poncirus than in Citrus aurantium, which is less apomictic and the most usual rootstock for clementines until 1970.

Comparison between Poncirus and Citrus genetic linkage maps.

Five genetic linkage maps were constructed for the parents of three progenies: Citrus aurantium (A) x Poncirus trifoliata var. Flying Dragon (Pa), C. volkameriana (V) x P. trifoliata var. Rubidoux (Pv) and a self-pollination of P. trifoliata var. Flying Dragon (Pp). The number of polymorphic markers assayed ranged from 48 for Pa to 120 for A according to the heterozygosity of each parental. As our focus was on genome comparison, most of the markers were newly generated simple sequence repeats. Inter-retrotransposon amplified polymorphisms based on four retrotransposon sequences isolated from Citrus spp were also used to saturate the maps. These polymorphisms were much more frequent in A (53) than in Pa (15) and randomly distributed throughout both genomes. Since comparative genomics and quantitative trait locus analysis applicability depends on the reliability of marker ordering, the causes of variation in marker order were investigated. Around 25% of the markers showed gametal segregation distortions. Segregation distortions were also observed at the zygotic level towards a reduction in the observed frequency of homozygotes from that expected in linkage groups 5 and 7. The presence of balanced lethal factors or gametal incompatibility genes in those genomic regions would explain a zygotic advantage of heterozygotes at these specific regions. Four differences in genomic organization were observed; three are putative translocations and affect homeologous linkage groups 3, 7 and 11, where highly distorted markers are found. Other causes of variation in marker order are also discussed: the introduction of new markers in the map, lowering the LOD score and the mapping software. These results represent the first comparative mapping analysis among Citrus and Poncirus species.

The diversification of Citrus clementina Hort. ex Tan., a vegetatively propagated crop species.

Clementines, due to their high quality, are one of the most important cultivated citrus mandarins. As in the case of sweet orange and satsuma mandarins, genetic variability within this species is minimal when analyzed by molecular markers, because the existing varieties have not been obtained through hybridization, but through the selection of spontaneous mutations affecting traits of agronomic interest. This would explain, at least in part, the greater diversity for agronomic traits when compared to the variability for molecular markers. Another possible (nonexclusive) reason is that the types of molecular marker used are not focused on the kind of molecular change mainly involved in the origination of new clementine cultivars; i.e., are all sources of variation equally involved in the diversification of these plants? To answer this question, different kinds of markers based on primers of random sequence, simple sequence repeats, and retrotransposon sequences that may reveal point mutations, and somatic recombination and transposon activity, respectively, were used to compare the level of variability among 24 clementine varieties. Their ISSR, RAPD, and AFLP analysis provided only two polymorphic bands, distinguishing just two varieties. No variability was found by SSRs, i.e., no new allele arising through somatic recombination was detected. Instead, the amplification of sequences adjacent to retrotransposons yielded a higher number of polymorphisms (14.6 vs 2.4% for the previous mentioned marker types). Two geographical distant groups, one from North Africa and the other from Spain, have evolved in agreement with polymorphisms based on IRAP markers anchored to, at least, two different Copia-like retrotransposon sequences. Therefore, this study suggests that the DNA of this type of mobile elements is evolving faster than the DNA of other markers in this clonal lineage.

Citrus and Prunuscopia-like retrotransposons.

Many of the world's most important citrus cultivars ("Washington Navel", satsumas, clementines) have arisen through somatic mutation. This phenomenon occurs fairly often in the various species and varieties of the genus.The presence of copia-like retrotransposons has been investigated in fruit trees, especially citrus, by using a PCR assay designed to detect copia-like reverse transcriptase (RT) sequences. Amplification products from a genotype of each the following species Citrus sinensis, Citrus grandis, Citrus clementina, Prunus armeniaca and Prunus amygdalus, were cloned and some of them sequenced. Southern-blot hybridization using RT clones as probes showed that multiple copies are integrated throughout the citrus genome, while only 1-3 copies are detected in the P. armeniaca genome, which is in accordance with the Citrus and Prunus genome sizes. Sequence analysis of RT clones allowed a search for homologous sequences within three gene banks. The most similar ones correspond to RT domains of copia-like retrotransposons from unrelated plant species. Cluster analysis of these sequences has shown a great heterogeneity among RT domains cloned from the same genotype. This finding supports the hypothesis that horizontal transmission of retrotransposons has occurred in the past. The species presenting a RT sequence most similar to citrus RT clones is Gnetum montanum, a gymnosperm whose distribution area coincides with two of the main centers of origin of Citrus spp. A new C-methylated restriction DNA fragment containing a RT sequence is present in navel sweet oranges, but not in Valencia oranges from which the former originated suggesting, that retrotransposon activity might be, at least in part, involved in the genetic variability among sweet orange cultivars. Given that retrotransposons are quite abundant throughout the citrus genome, their activity should be investigated thoroughly before commercializing any transgenic citrus plant where the transgene(s) is part of a viral genome in order to avoid its possible recombination with an active retroelement. Focusing on other strategies to control virus diseases is recommended in citrus.

Genetic analysis of apomixis in Citrus and Poncirus by molecular markers.

Propagation of citrus rootstocks depends upon the production of clonal plants from nucellar seedlings. This makes apomixis one of the host important traits in breeding programs for citrus rootstocks. The genetic control of apomixis was studied in a 50-tree progeny derived from the cross C. volkameriana×P. trifoliata using 69 molecular markers and bulked segregant analysis. The proportion of nucellar seedlings was estimated by isoenzymatic analysis of 25 seedlings per tree for 2 consecutive years. The type of embryony (polyembryonic versus monoembryonic seeds) was also determined for fruit-yielding trees. Separate genetic maps for each parental species were developed. The integration and comparison of these maps could be accomplished using common multiallelic segregant loci. Differences in gene synteny between the two species-specific genetic maps were shown. Important distortions in the segregation of markers at several genomic regions, some of them also involving differences in the C-methylation pattern, have been observed, especially for the pollen parent. Analysis of quantitative trait loci (QTLs) revealed the presence of six genomic positions (two in P. trifoliata and four in C. volkameriana) contributing individually up to 24% of the total variation for apomixis. Within the same species, QTLs with positive and negative allele effects were present, even in the same linkage group. One of the markers associated to apomixis (Apo2) is also associated to embryony type. Therefore, the genetic control of apomictic reproduction found in citrus (nucellar embryony) is quite complex compared to what has been reported for gametophytic apomixis. Molecular markers linked to QTLs governing apomixis will be useful to assist selection of future apomictic rootstocks for citrus varieties.

Salt tolerance in Lycopersicon species. IV. Efficiency of marker-assisted selection for salt tolerance improvement.

The usefulness of marker-assisted selection (MAS) to develop salt-tolerant breeding lines from a F2 derived from L. esculentum x L. pimpinellifolium has been studied. Interval mapping methodology of quantitative trait locus (QTL) analysis was used to locate more precisely previously detected salt tolerance QTLs. A new QTL for total fruit weight under salinity (TW) near TG24 was detected. Most of the detected QTLs [3 for TW, 5 for fruit number, (FN) and 4 for fruit weight (FW)] had low R (2) values, except the FW QTL in the TG180-TG48 interval, which explains 36.6% of the total variance. Dominant and overdominant effects were detected at the QTLs for TW, whereas gene effects at the QTLs for FJV and FW ranged from additive to partial dominance. Phenotypic selection of F2 familes and marker-assisted selection of F3 families were carried out. Yield under salinity decreased in the F2 generation. F3 means were similar to those of the F1 as a consequence of phentoypic selection. The most important selection response for every trait was obtained from the F3 to F4 where MAS was applied. While F3 variation was mainly due to the within-family component, in the F4 the FN and FW between-family component was larger than the within-family one, indicating an efficient compartmentalization and fixation of QTLs into the F4 families. Comparison of the yield of these families under control versus saline conditions showed that fruit weight is a key trait to success in tomato salt-tolerance improvement using wild Lycopersicon germplasm. The QTLs we have detected under salinity seem to be also working under control conditions, although the interaction family x treatment was significant for TW, thereby explaining the fact that the selected families responded differently to salinity.

Genetic diversity in the orange subfamily Aurantioideae. II. Genetic relationships among genera and species.

Genetic relationships were studied by means of ten isoenzymatic systems, at the genus and species level, using two distances and four methods of aggregation in a germplasm collection of 198 cultivars and accessions of 54 species belonging to Citrus and 13 related genera. The most consistent results were obtained by the chord distance and the neighbor-joining clustering method. Citrus species were distributed in two main groups: the orange-mandarin group and the lime lemon-citron-pummelo group. The species C. halimii and C. tachibana are not included in these groups. Mandarin species fall into three main subgroups: one includes C. sinensis; the second, C. aurantium, the third, small-fruit species. The citron, the pummelo and the ancient lemon subgroups form a cluster to which the species belonging to subgenus Papeda and the cultivated limes, lemons and bergamots are related. Microcitrus spp, to which Severinia buxifolia and Atalantia ceylanica seem to be related, cluster with the lime lemon-citron-pummelo group while Fortunella is close to the orange-mandarin group. Poncirus trifoliata, the most important species for citrus rootstock improvement is located far from Citrus but connected to it through Fortunella spp. A broad distribution of species has been found that should be taken into account to sample new genotypes in the search of desired characters in order to fully and efficiently use genetic resources for citrus improvement.

Genetic diversity in the orange subfamily Aurantioideae. I. Intraspecies and intragenus genetic variability.

Despite the great economic importance of citrus, its phylogeny and taxonomy remain a matter of controversy. Moreover pathogens of increased virulence and dramatic environmental changes are currently spreading or emerging. The objectives of the present paper, measuring genetic variability and studying its pattern of distribution, are crucial steps to optimize sampling strategies in the search of genotypes that tolerate or resist these threatening factors within the huge array of Citrus and Citrus related species. Their intraspecific and intrageneric variability was studied comparatively by means of ten enzymatic systems using eight different measures. The analysis of ten enzymatic systems allowed us to distinguish all the species and all but one artificial hybrid. The species with the lowest genotypic variability are C. myrtifolia, C. deliciosa (willow leaf mandarin), C. paradisi (grapefruit), C. limon (lemon) and C. sinensis (sweet orange), while Severinia buxifolia shows the highest value. A broad spectrum of heterozygosity values was found in the collection. Lemons (C. limon, C. meyeri, C. karna, C. volkameriana), limes (C. aurantifolia, C. limettioides, C. lattifolia) and C. bergamia show a very high percentage of heterozygosity which indicates an origin through interspecific hybridization. Two main factors limit genetic intraspecific variability: apomictic reproduction, where nucellar embryos are much more vigorous than the zygotic ones, and nurserymen selecting against variability in the seedling stage of the rootstocks or in propagating the scion cultivars vegetatively. Additionally, self-pollination appears in some species mainly used as rootstocks which would explain their low heterozygosity values. Genetic differences between species and genera are in general high, which suggests that adaptation might have played an important role during the evolution of the orange subfamily.

Factors affecting Citrus tree isozyme-gene expression.

Ten enzymatic systems of Citrus species and cultivars have been evaluated for identification purposes and for genetic variability studies. The following factors that could affect their expression were studied: season of sampling, location, rootstock, position of the branch, infection, and age of the tree. Differences involving the presence-absence of the Cu/Zn SOD within the same tree were found. This difference is mainly related to the position of the leaf relative to the sunlight. No change was observed at any of the ten enzymatic systems assayed regarding the location, the rootstock, the growing condition, the season, or the infection with most virus and virus-like pathogens. Viroids induced noticeable changes on 6PG and PRXa zymograms in C. medica. A new peroxidase (not present in healthy plants) was detected that could be related to appearance of symptoms. This may induce errors when trees without sanitary control are characterized by this enzymatic system. On the other hand, it provides a new possibility for studying the plant response to the presence of viroids. An effect of age, from 3 months up to 12 years, was observed on citrange Troyer and mandarin Cleopatra PRX, MDH and 6PG patterns. An important change occurs around the first year, most likely related to the end of the seedling stage. This is followed by a long transition phase, the end of which (around 9 years later) coincides with a change in the PRX pattern. These age-related changes seem to involve post-translational modifications of pre-existing isozymes.