Horticulture Research, 2022, 9: uhac070 https://doi.org/10.1093/hr/uhac070 Article Construction of a collection of introgression lines of “Texas” almond DNA fragments in the “Earlygold” peach genetic background Naveen Kalluri1 , Octávio Serra2 , José Manuel Donoso3 , Roger Picañol4 , Werner Howad1 , 5 , Iban Eduardo1 , 5 , * and Pere Arús1 , 5 , * 1 Centre for Research in Agricultural Genomics (CRAG) CSIC-IRTA-UAB-UB, Campus UAB, Edifici CRAG, Cerdanyola del Vallès (Bellaterra), 08193 Barcelona, Spain 2 Instituto Nacional de Investigação Agrária e Veterinária, I.P., Banco Português de Germoplasma Vegetal (BPGV), Braga, Portugal 3 Instituto de Investigaciones Agropecuarias (INIA), Centro Regional de Investigación Rayentué, Av. Salamanca s/n Sector Los Choapinos, Rengo 2940000, Chile 4 Rijk Zwaan Ibérica S.A. Finca La Marina-PJ Lo Contreras 30395, La Puebla|Cartagena (Murcia), Spain 5 IRTA, Campus UAB, Edifici CRAG, Cerdanyola del Vallès (Bellaterra), 08193 Barcelona, Spain *Corresponding authors. E-mail: pere.arus@irta.cat, iban.eduardo@irta.cat Abstract Peach [Prunus persica L. Batsch] is one of the major temperate fruit tree species, the commercial materials of which have a low level of genetic variability. Almond [P. dulcis (Mill) DA Webb], a close relative of peach cultivated for its kernels, has a much higher level of diversity. The species are inter-compatible and often produce fertile hybrids, almond being a possible source of new genes for peach that could provide biotic and abiotic stress tolerance traits. In this paper we describe the development of a collection of peach- almond introgression lines (ILs) having a single fragment of almond (cv. Texas) in the peach background (cv. Earlygold). Lines with few introgressions were selected with markers from successive generations from a “Texas” × “Earlygold” F1 hybrid, initially using a set of SSRs and later with the 18 k peach SNP chip, allowing for the final extraction of 67 lines, 39 with almond heterozygous introgressions covering 99% of the genome, and 28 with homozygous introgressions covering 83% of the genome. As a proof of concept, four major genes and four quantitative characters were examined in the selected ILs giving results generally consistent with previous information on the genetics of these characters. This collection is the first of its kind produced in a woody perennial species and promises to be a valuable tool for genetic analyses, including dissection of quantitative traits, positional cloning, epistasis and as prebreeding material to introgress almond genes of interest into the peach commercial gene pool. Introduction with different allelic dosages of one or more QTLs can Introgression line (IL), near-isogenic line (NIL) or chro- be created by crossing selected ILs between them or with mosome segment substitution line (CSSL) collections are other lines [11]. Given that ILs are usually constructed sets of lines containing a single fragment of a donor with a background of elite commercial lines, they provide genome in the background of a recurrent genome. These an optimal resource to analyze the effects of QTLs of collections consist of tens to a few hundred lines, each exotic materials in the breeding materials, which are with a different introgressed donor fragment, the sum of masked in other populations used for genetic analysis by which cover most, or ideally all, the donor genome [1, 2]. interactions with other loci of exotic origin. This makes They are usually developed by backcrossing to the recur- IL libraries a direct source of improved materials for rent parent in the initial generations, and selfing in the plant breeding as well as an invaluable tool for the advanced backcross generations, using a set of markers evaluation and introgression of useful genes from wild or with good genome coverage to identify the individuals of exotic materials in the cultivated gene pool, facilitating interest. the use of the variability stored in exotic materials that IL collections serve as valuable tools for genetics and can compensate for the loss of diversity resulting from breeding applications, as in the dissection of complex domestication1 . The first IL collection was produced from genetic traits [3–6] and the fine mapping and positional a cross between tomato and the wild species Solanum cloning of genes or QTLs [7–9]. They are suitable for the pennellii [12]. Since then, IL collections have been exten- study of the effects of specific QTLs in different envi- sively developed and used in many model and culti- ronments and different genetic backgrounds [10] and the vated species, including most staple crops and horticul- analysis of interallelic and epistatic interactions, as lines tural herbaceous species [2]. No examples exist for tree Received: 1 December 2021; Accepted: 9 March 2022; Published: 23 March 2022; Corrected and Typeset: 1 June 2022 © The Author(s) 2022. Published by Oxford University Press on behalf of Nanjing Agricultural University. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. 2 | Horticulture Research, 2022, 9: uhac070 species, due to their long intergeneration periods that makes the process of IL generation extremely long. Peach was used as the recurrent parent and almond (P. dulcis) as donor. Peach and almond are sexually compat- ible species that originated from a common ancestor in central Asia about 5 Mya [13, 14]. Peach was domesticated in China ∼5000 years ago [15], while almond domestica- tion is still unclear, although it probably occurred some- where between the steppes of central Asia and the east- ern Mediterranean shores [16]. Peaches are cultivated for their fleshy mesocarp, as are other Prunus stone fruit species such as apricot, cherry, and plum, and almonds are cultivated for their seed: both are of major economic importance. One of the key biological differences between peach and almond is that peach is self-compatible and has low levels of genetic variability, while almond is self- incompatible with a highly diverse genome [17, 18]. Almond appears to be a good source of new alleles that could provide useful variability conferring adaptation to climate change and disease and pest resistance in peach. Various studies have been undertaken to understand the inheritance of almond variability in the Figure 1. Breeding scheme for the extraction of introgression lines from peach background using the offspring of a “Texas” (syn the almond × peach cross. “Texas Prolific”, syn. “Mission”) almond × “Earlygold” peach F1 plant (MB1.37), selfed (the T × E population) and backcrossed to “Earlygold” (the T1E population) Results [19, 20]. A large backcross progeny was initially used to Development of a peach-almond IL collection demonstrate the feasibility of a marker-based method A total of 8467 fruits from different generations of (Marker-Assisted Introgression; MAI) to produce ILs from the “Texas” × “Earlygold” cross were obtained from this interspecific hybrid only two backcross generations 2011 to 2020 and used to construct this IL collec- after [21]. This was necessary to show that introgression tion (Supplementary Table 1, Figure 1). Results to 2015 from a distant source was possible within a reasonable have been reported earlier [21]. Since that date, 4916 timeframe, considering the long intergeneration period fruits from the offspring of the E2T set, a set of of tree species. Among these lines, those carrying a gene trees from the BC2 generation of the MB1.37 hybrid for resistance to peach powdery mildew from almond to “Earlygold” with 2–3 almond introgressions, were have already been incorporated in the IRTA peach breed- used for in vitro embryo rescue, giving rise to 1276 ing program [22]. Based on the T1E selections, some ILs seedlings (Supplementary Table 1). Using the 113 SSRs were extracted and a first collection of individuals with genotyped in the E2T set (Supplementary Table 2 and two or three introgressions with the peach cytoplasm in Supplementary Figure 1), we extracted 146 ILs (97 in a second backcross with T1E individuals (the E2T set) heterozygosis and 49 in homozygosis) that, along with were developed [21]. the 137 lines (109 in heterozygosis and 28 in homozy- The objective of this study was to develop a complete gosis) previously selected [21], resulted in a full set of peach-almond IL collection and to describe its main 283 ILs covering the complete genome of almond “Texas” features. Due to the long life and ease of clonal repro- in the “Earlygold” background in heterozygosis (206 lines; duction of Prunus trees, two complete sub-collections Supplementary Figure 2) and 94% of the almond genome have been elaborated, one with the almond introgressed in homozygosis (77 lines; Supplementary Figure 3). Of fragment in heterozygosis and the other in homozygosis, the 206 heterozygous ILs, 65 had unique introgressed facilitating the analysis of interallelic interactions. We fragments and in 141 these fragments were in common have tested this collection with a set of characters of with other lines (Supplementary Figure 2), whereas simple and complex inheritance of known genetic basis for the 77 homozygous ILs, 29 were unique and in as a proof of concept of its value for genetic analysis. This the remaining 48 the introgressed fragments were in is a powerful resource to characterize almond diversity common with other lines (Supplementary Figure 3). in a peach genetic background and will be useful for The number of lines covering each linkage group in fruit tree and Prunus geneticists and breeders to identify the heterozygous ILs were G1 (20), G2 (27), G3 (21), G4 and genetically-dissect valuable traits of almond and (19), G5 (24), G6 (60), G7 (16) and G8 (19). For the 77 to rapidly integrate them in new commercial peach homozygous ILs, almond fragments were located on varieties. G1 (11), G2 (12), G3 (3), G4 (8), G5 (8), G6 (25), G7 (5) Kalluri et al. | 3 and G8 (5). The missing regions in homozygosis were the ends of G3 and G8 (Supplementary Figure 3). One of Genome age (%) cover these lines (37P18–44) had two introgressed fragments, 100 80 96 96 43 86 97 81 83 one in heterozygosis (G1) and another in homozygosis (G2). For all chromosomes in the heterozygous ILs and for chromosomes 2, 4 and 6 in homozygous ILs, at least covered 188.1 one of the ILs spanned the whole chromosome distance Mbp 36.3 29.1 26.4 11.1 18.5 26.5 21.8 18.4 covered by the markers used. Average IL fragment Selection of an IL collection with 18 k SNP genotyping (Mbp) 24.0 23.4 16.2 11.1 11.5 14.6 10.9 11.8 15.4 We selected 135 ILs from the total collection to be geno- Table 1. Heterozygous and homozygous IL collections selected with the 18 k SNP chip. Numbers of lines, fragments identified and genome coverage typed for the 18 k SNP chip, including 81 heterozygous ILs Largest IL fragment (100% almond genome coverage) and 54 homozygous ILs (Mbp) (94% coverage). A total of 6624 SNPs were identified with 36.3 29.1 26.4 11.1 14.5 26.5 19.6 15.2 the appropriate segregations. Markers with “Earlygold” and the MB1.37 hybrid in homozygosis (417) and with Smallest IL fragment missing data in either of them (85) were discarded. The (Mbp) remaining 6122 SNPs nearly covered the entire peach 17.0 19.2 11.1 6.0 6.0 5.1 2.2 8.5 genome (98.4%), with an average density of one SNP every 37 kb. Approximately half (3080) of these SNPs Fragments were homozygous for “Earlygold” and heterozygous for MB1.37 and were used to determine the positions of the 28 almond introgressions. The remaining 3042 SNPs, which 3 4 2 1 4 8 2 4 ∗ were heterozygous for “Earlygold and with any genotype ILs HOM for MB1.37, were used to establish the composition of the ILs 28 ‘Earlygold” background. 3 5 2 2 3 8 2 3 From the 135 lines analyzed with the 18 k SNP chip, Coverage of different non-overlapping fragments in which the chromosome is divided by the ILs of the collection four did not produce reliable results, three did not detect 100 100 100 100 100 100 100 any introgressed fragments and of the remaining 128 (%) 96 99 lines, 39 were selected for the IL collection in heterozy- covered gosis, 38 with a single DNA introgressed fragment from 224.4 Mbp almond and one with two fragments, a major one on 47.8 29.1 27.4 25.8 18.5 30.8 22.4 22.6 G1 and the other on G8. The genome coverage was 99%: including all but one small almond DNA fragment at the Average IL fragment proximal end of chromosome 2 (1.35 Mbp) that could not (Mbp) be detected only with the SSRs used (Figure 2, Table 1 and 24.4 16.8 16.1 18.3 21.4 18.7 12.4 17.1 8.7 Supplementary Table 3). Twenty-eight additional lines with introgressed frag- Largest IL fragment ments in homozygosis were also selected, covering 83% (Mbp) 36.9 29.1 26.4 24.3 14.0 30.8 22.4 19.5 of the genome with a large gap on G4 (Figure 3, Table 1 and Supplementary Table 4) and smaller ones on all Smallest IL groups except G5. While most of these plants had a single fragment major fragment in homozygosis, two had two fragments, (Mbp) 14.0 11.0 one in homozygosis and the other in heterozygosis. Nine 4.5 6.0 4.0 6.0 8.5 6.4 of the selected lines had part of their introgression in homozygosis and part in heterozygosis, indicating their Fragments origin from gametes with recombinations in the almond fragment. These lines would require an additional round 35 5 5 5 4 3 4 4 5 ∗ of selection in their selfed offspring to obtain completely single-fragment homozygous ILs. The remaining lines Mbp total ILs 39 genotyped with the SNP chip (60) were discarded as they 4 7 5 4 3 9 4 3 contained additional almond DNA fragments. The SNP analysis provided a much more detailed pic- 225.7 47.8 30.4 27.4 25.8 18.5 30.8 22.4 22.6 ture of the genome and allowed identification of eight small fragments, <1% of the genome (<2.5 Mbp), that ∗ Number Linkage ILs HET group were not detected by the SSRs: one on G1 (0.97 Mbp), Total G1 G2 G3 G4 G5 G6 G7 G8 G2 (0.82 Mbp), G3 (0.52 Mbp), G5 (1.40 Mbp), and G7 4 | Horticulture Research, 2022, 9: uhac070 Figure 2. Graphical genotype of the 39 heterozygous ILs of almond (“Texas”) in the peach (“Earlygold”) background based on the 18 k SNP chip and the physical positions of the SNPs. G1 to G8 are the eight linkage groups of Prunus. Figure 3. Graphical genotype of the 28 homozygous ILs of almond (“Texas”) in the peach (“Earlygold”) background based on the 18 k SNP chip and the physical positions of the SNPs. Red fragments are heterozygous almond introgressions. G1 to G8 are the eight linkage groups of Prunus. (2.17 Mbp), and three on G8 (0.91, 1.29 and 1.60 Mbp). used to obtain them, ranging from 4.0–36.9 Mbp (average Their positions are indicated in Figures 2 and 3. They 17.1 Mbp) in the heterozygous ILs and from 2.2–36.3 Mbp were not considered when selecting or discarding ILs, (average 15.4) in the homozygous ILs. Several linkage although one of the homozygous ILs contains only the groups were completely or almost completely (>95%) largest fragment, the one on G7, in homozygosis. Intro- covered with a single introgression: these were G2, G3, gressed fragments were generally large (see Table 1), as G4, G6, and G7, the two former in both homozygosis and expected considering the low number of generations heterozygosis and the rest in heterozygosis. Kalluri et al. | 5 Figure 4. Graphical genotype of the “Earlygold” background in the selected heterozygous ILs using the 18 k SNP chip and the physical positions of the SNPs. Segregation in the “Earlygold” background was also PAILE2–1625, PAILO2–0123, PAILO2–0125a, PAILO2–0125b, studied with the SNP chip. Results (Figures 4 and 5) show PAILO2–0130∗, and PAILE2–0625). that there were large regions without segregating mark- Maturity date (MD), a gene located in the central region ers, covering approximately half of the genome (50.5%) of chromosome 4 (ref. [24]), was studied in the spring/- as previously observed [19, 23]. The rest of the genome summer of 2021. This trait has sometimes been scored of each IL segregated in large blocks for the two alleles as qualitative and sometimes as quantitative, depending of “Earlygold”. Knowledge of the “Earlygold” genotype for on the alleles present [20, 23, 24]. In the case of the each specific IL is important to incorporate the possible T × E, T1E and E × E populations (Supplementary Table 6) effects of a segregating background in the genetic analy- there are three alleles, one inherited from “Texas” to sis of trait variability. the MB1.37 hybrid (T), and the other two (E1 and E2) from Earlygold, where the T allele produces the latest maturity date, while it is intermediate with E1 and early Major genes and QTLs analyzed in the IL with E2, with gene action predominantly additive. In this collection case, the average number of Julian days to maturity for The data were obtained from single trees of different ILs without the fragment from almond including MD ages, some too young to produce fruit, some grown on (28 ILs with genotypes involving only alleles E1 and E2), their own roots and others grafted. For these reasons, and “Earlygold”, was 167.7 (± 12.4), whilst, as expected, and as a preliminary attempt to understand the potential the two ILs (PAILE4-0226b and PAILE4–0526) that con- of ILs for genetic analysis, here we studied traits that tained the late almond T allele matured later (the 214th were determined by major genes and could be analyzed day, more than one month later). However, PAILE4-0226a as qualitative, as well as certain quantitative characters also containing the T allele matured on the 179th day, previously studied for QTLs in the T × E and T1E popula- i.e. within the range of “Earlygold” and the early and tions [20]. Three of the major genes expected to segregate intermediate maturing ILs. This can be explained by the in the ILs, juiciness (Jui), blood flesh (DBF2) and resistance “Earlygold” background, where PAILE4-0226b and PAILE4– to powdery mildew (Vr3), had the expected phenotypes in 0526 are TE1 and have a late phenotype, whereas PAILE4- all ILs that could be studied, 28, 28 and 63, respectively 0226a is TE2 (see Figure 4), resulting in an individual of (Supplementary Table 5). For the two linked genes Jui and intermediate maturity date. DBF2 on chromosome 1, all ILs had the juicy, yellow Quantitative measurements were taken for three flesh phenotype except for PAILE1–2348 and PAILE1–3448 fruit traits, weight (FW), soluble solid contents (SSC) that were both non-juicy and red-fleshed. For Vr3, only and titratable acidity (TA) (Supplementary Tables 5 ILs with almond introgressions in the region of chro- and 7). With the criteria used to consider a signif- mosome 2 that contains the gene [22] were resistant icant QTL (see Materials and methods section and (PAILE2–0119, PAILE2–0125, PAILE2–0130a, PAILE2–0130b, Supplementary Table 5), two QTLs were identified for FW, 6 | Horticulture Research, 2022, 9: uhac070 Figure 5. Graphical genotype of the “Earlygold” background in the selected homozygous ILs using the 18 k SNP chip and the physical positions of the SNPs. one with a strong increase in fruit weight compared to genomics of crop species. Introgression line sets are opti- “Earlygold” in two overlapping homozygous ILs (PAILO6– mal for the dissection of quantitative traits and to facil- 0005 and PAILO6–0008) with average values of 144 and itate genetic analysis and gene cloning, particularly in 148 g/fruit compared to 83 g of “Earlygold” (73–78% the background of lines of agricultural interest, so avoid- weight increase). Another IL in this region, PAILO6–0308, ing problems related with donor-donor epistatic effects had FW values similar to “Earlygold”, indicating that the and making possible the detection of subtle pleiotropic QTL was located at the extreme of the chromosome effects, usually difficult to identify in other population (0–4 Mbp). When looking at heterozygous ILs in this types such as F2, BC1 or RILs [25]. However, the construc- region (PAILE6–0008b, PAILE6–0009, PAILE6-0031a and tion of such populations is especially difficult in woody PAILE6-0031b), we observed an inconsistent pattern with perennials, as it represents an enormous investment some lines having values similar to “Earlygold” (PAILE6- in time and resources, a consequence of the large size 008b and PAILE6-0031a), and others with significantly of the individuals and the long intergeneration time of lower (PAILE6–0009 and PAILE60031b) values. For SSC, these species. In this paper we present a peach-almond only one highly significant QTL was observed in the introgression collection that we believe is the first to proximal end of G6 (PAILO6–0308) with the almond allele have been produced in a perennial species. One advan- producing an increase of the value of this trait, and tage of certain tree crops is that they can be clonally for TA no QTLs with sufficient effects were identified reproduced, which has given us the possibility of con- (Supplementary Table 5). structing two collections, one with 39 ILs that contain Petiole length was a variable trait in the crosses introgressed almond fragments in heterozygosity and between almond and peach studied before [20], where covers 99% of the almond genome, and another with 28 typically almond had a long petiole, peach a short one, homozygous ILs with 83% of genome coverage. This con- intermediate in the hybrid, and the progeny segregated tinues research begun in 2006 in our lab with the initial between these two extremes. In the IL collection, all objective of finding fast approaches to introgress useful plants had a short petiole like in peach except for PAILE8– genes from the highly variable Prunus wild or cultivated 0516 and PAILO8–1018 that had a highly significantly species, almond in this case, to enrich the much narrower longer petiole than ‘Earlygold, similar to that of the variability of cultivated peach. This initial research led interspecific hybrid (Supplementary Tables 5 and 7). to an incomplete first set of introgression lines [21] with One more line (PAILE2–2630), had a significantly shorter partial genome coverage (64% in heterozygosis and 14% petiole than “Earlygold”. in homozygosis). Here, we selected the ILs from the selfed progeny of a set of individuals with a low number of introgressions (2–3), which has considerably simplified Discussion our work. Plant collections adequate for efficient genetic analysis Using a set of SSRs we selected 283 ILs, 135 of which are a key resource for the progress in the genetics and were genotyped in depth with the 18 k SNP chip. SNP Kalluri et al. | 7 results were particularly informative as they increased when studying fruit or color-related traits. The genotypes the marker density more than 50-fold (from an average for the “Earlygold” background in the IL collection are of one SSR every 2015 kb to one SNP every 37 kb) allowing given in Figures 4 and 5 and may be useful to identify the us to identify and discard plants having more than one genotypes of the recurrent parent and consider or discard introgression. The final collection of 67 individuals with possible effects of the background on the expression of a single introgression (with a few exceptions with two their character of interest. introgressions) could therefore be selected with great Using the plants of the IL collections, we examined the certainty, and with a precise position of the boundaries segregation of certain major genes previously described of each almond fragment. The accuracy of this analysis in “Texas” × “Earlygold” progenies, and for four of them also made possible the identification of several smaller (Jui, DBF2, Vr3 and MD) we observed the expected phe- introgressions (<1% of the genome) that were undetected notypes. Other genes that segregate in the IL collection, by the SSR markers, and of some missing fragments at such as two independent male fertility restorer genes the end of certain linkage groups. Our initial expectations (Rf1/rf1 and Rf2/rf2) [19] were not expected to produce were that these fragments would occur only very sporad- any phenotypes as all E2T individuals had the peach ically because of the relatively low level of recombination cytoplasm. The almond fruit (Alf ) gene that determines of peach chromosomes and specifically of this interspe- the formation of the thick mesocarp [20] typical of peach, cific cross (1.2 crossovers per chromosome on average would produce only almond fruit types in the individuals [19]), but the fact that we recovered them suggests that homozygous for the recessive almond (alf ) allele, and a high-density genotyping step before elaborating a final the ILs homozygous at the corresponding region of G4 set of ILs is an advisable option. were not available in homozygosity. Other lines having One aspect to consider in the analysis of this IL collec- this fragment in homozygosity, along with additional tion is that a commercial cultivar (“Earlygold”) was used introgressed fragments, were too young to produce fruit. as a recurrent parent. “Earlygold” is not an inbred line, The case of one IL (PAILE4-0226a) with an intermediate which has the drawback that segregation for chromoso- maturity date phenotype while it was expected to mature mal fragments heterozygous in this variety may interfere late as it had the almond introgression, can be explained with the interpretation of genes/QTLs from the intro- by the interaction of the almond allele with an allele gressed almond fragments in the ILs. At the beginning of “Earlygold” that confers early maturity [20, 23], as of this project we considered, but discarded, the option opposed to other ILs that had the late allele of “Earlygold”. of switching to a more homozygous line. The first main This would be a clear case of segregation in the “Ear- reason was that the availability of homozygous lines lygold” background that would result in an unexpected is scarce in peach, and most of them are either weak phenotype in an IL. Considering the importance of this individuals, probably as a consequence of inbreeding, or region for phenological and fruit related characters, it genotypes, such as the Spanish non-melting flesh vari- would be interesting to fix one of the alleles of “Earlygold” eties, that are genetically distant from the major com- at this locus in all lines to have a uniform phenotype mercial peach gene pool [26]. Secondly, including a new for maturity date which may avoid interference with recurrent parent would have delayed the construction other characters, particularly fruit characters, that may of this IL collection by two generations, equivalent to be associated with this region. 7–9 years. A third reason is that we expected that the Four quantitative characters were studied, three in variability detected by the almond genome would often fruits of some ILs that were already fruiting, and one in be of a sufficiently different nature and genome position the leaf petiole, all with previous inheritance studies and compared to that of peach, as shown in previous results with QTLs identified for some of them (FW and petiole based on the analysis of “Texas” × “Earlygold” progenies length) in the offspring of crosses involving “Texas” and [19, 20], making interferences between almond/peach vs. “Earlygold” [20, 23], the main features of which are sum- peach/peach easily detectable and interpretable. Addi- marized in Supplementary Table 6. For one character, tionally, “Earlygold” is heterozygous in less than half of fruit weight, a major QTL was previously found at the its genome [20, 23], because of the existence of large proximal end of G6 in almond × peach progenies [20], DNA fragments identical by descent, as in many other where the allele of almond produced an increase in fruit modern peach varieties [27], due to the recent history weight with a high percentage of explained phenotypic of co-ancestry of modern peach breeding programs [28], variance (R2 = 19–21%). A major QTL for fruit weight was so we were only expecting a relatively narrow window also detected at the same chromosomal region by com- of segregation to occur in its offspring. The inheritance bining linkage mapping and association studies in a col- analysis of a large set of characters of the F2 of “Early- lection of peach genotypes [29]. We found a high increase gold” is available [23], and relevant for this work as it in fruit weight at this region in the homozygous ILs, but identified a limited number of QTLs. Only one of these, heterozygous ILs had a similar or even significantly lower a major QTL located on G4 in the region of MD that, weight than “Earlygold”, while, based on previous results, in addition to maturity date, partially affects other fruit an increased size was expected (Supplementary Table 6). characters such as fruit weight, soluble solid contents A more detailed analysis with uniform and replicated and leaf color at senescence, appears to be of concern materials is necessary to fully understand the genetics of 8 | Horticulture Research, 2022, 9: uhac070 this especially interesting QTL, as fruit size is a character two key aspects of the divergence between these two of great commercial value in peach and the use of the species: the differences between a dry and a fleshy edible almond allele, alone or in combination with other peach fruit from morphological, physiological and metabolic alleles may have an immediate application in breeding. standpoints, and the better adaptation of almonds For SSC and TA, for which no QTLs have previously been to abiotic stress, particularly drought, of great impor- detected in these materials (Supplementary Table 6), we tance and urgency considering the menace of climate identified a QTL in the proximal end of G6 for SSC and change. none for TA. A QTL for SSC at the beginning of chro- mosome 6 has also been previously reported using a multi-progeny approach that included peach × peach Materials and methods and peach × other species (with almond one of them) Plant materials [30]. Finally, petiole length was significantly longer in the In previous research, the F1 of the cross between almond lines PAILE8–0516 and PAILO8–1018 than in “Earlygold” “Texas” used as female parent crossed with peach “Early- (43–70% longer) and the other ILs, coinciding with the gold” as pollen donor, named “MB1.37”, was backcrossed position of a QTL with large effects (R2 = 15–39%) and to “Earlygold” as male parent, in the winter of 2006, dominance for the almond allele previously identified 2007 and 2008 (ref. [21]). A large offspring (N = 1095) [20] for this character. Another IL with petiole length was obtained, referred to as the T1E population, and a shorter than “Earlygold” was also found in G2, suggest- few (N = 18) individuals, the pre-introgression line (prIL) ing the existence of alleles that may determine shorter set, carrying 2–4 almond introgressions were selected petioles in the species (almond) that has the long petiole with simple-sequence repeat (SSR) markers covering the phenotype. whole peach genome [21]. The inheritance of various The results on the trait analysis reported here were agronomic traits with a group of N = 190 T1E plants was obtained on single plants grown on their own roots previously analyzed [20], most selected at random from and from seedlings obtained in various years (see the T1E initial population, but also including those of Supplementary Tables 3 and 4), some of which were the prIL set. Due to the cytoplasmic male sterility con- still in their juvenile period. While they can be accurate ferred by the almond cytoplasm detected in the T1E for traits with clear alternative phenotypes that can be population [19], an additional backcross was performed scored qualitatively, these conditions are not optimal using “Earlygold” as the female parent (E2T; N = 160), for the analysis of metric traits. The next step in this to ensure that the plants obtained were pollen-fertile. research is the production of grafted replicates from The T1E plants used as staminate parents to generate the complete IL collection, to be planted in two sites. the E2T offspring were chosen to contain a low number Once these plants reach the production stage, we will of introgressions and to provide full almond genome start phenotyping them for the traits that we consider coverage, including mostly plants from the prIL set. A of interest. We also plan to continue the production of subset of 37 marker-selected plants of the E2T progeny to new ILs to complete the collection, especially that of contain only two or three almond introgressions, the E2T homozygous ILs where there are still some gaps. We are set, was selected [21]. In this work, the selfed (E2TS1) or also expanding this collection to include ILs with wild backcrossed to “Earlygold” (E3T) progeny of trees from the crop relatives of peach, particularly P. davidiana and P. E2T set were used for this last step of IL extraction, where mira. One direction of our current research is to fine map, our objective was to obtain a collection of individuals and eventually clone, the main major genes detected so with a single almond introgression, in both homozygosis far in peach × almond crosses. We have already started and heterozygosis, and covering the full almond genome. with the Vr3 gene that confers resistance to powdery The breeding scheme for IL extraction is given in Figure 1. mildew [22], and are advancing with others such as (DBF2, Plants from parents, hybrid and the T1E generation Jui and Alf ). are kept in the experimental fields of IRTA at Cabrils and Gimenells (Catalonia, Spain) as described before [21]. The E2T set was grown at two different IRTA stations: Conclusion in Caldes de Montbui grafted to “Garnem” rootstock and We report a nearly-complete a collection of ILs that, in Mollerussa on their own roots. The selected ILs were given the length of the peach intergeneration period, has planted at Caldes de Montbui on their own roots. For the required a 15-year process of crossing and marker selec- E2T set and ILs the spacing between trees and rows is tion along three backcross generations. The collection 2.5 × 4.0 m. has been studied in detail with a high-density SNP chip and is ready to be used by members of the scientific com- In vitro embryo rescue munity to which we offer any plants in this collection, As the recurrent parent “Earlygold” is an early maturing or the full collection, for research purposes. Apart from variety (May–June), most of the E2T individuals ripened its use in enriching the narrow peach variability, and early (June–July). The embryos of these individuals facilitating gene cloning, we believe that this collection were not completely developed at fruit maturity time is specially adequate for understanding the genetics of and in vitro embryo rescue was necessary to be able Kalluri et al. | 9 to germinate all seeds collected. The method used for 5 min and at the final step, the PCR product was held was based on previous work [31]. The fruits were first at 4◦ C to complete. For capillary electrophoresis (Applied immersed in a disinfectant solution of water and NaOCl Biosystems ABI PRISM® 3130xl Genetic Analyzer), 2 μl (3.7%), thoroughly washed with water, then opened using of PCR product was mixed with 12 μl of formamide, a nutcracker to collect the seeds. The seeds were surface 0.35 μl GeneScan500 LIZ (Applied Biosystems, USA) and sterilized in a solution of 300 ml of NaOCl (3.7%), 700 ml denatured at 94◦ C for 3 min. Allele sizes were identified H2 O and two drops of tween for 10 min, then washed with the GeneMapper 5.0 software. with autoclaved water in a laminar flow hood. Seeds A selected set of 135 lines with a single introgression were opened and the embryos extracted and placed in based on their SSR genotype, to cover the largest test tubes with a solution of sucrose (30 g/l), Duchefa possible region of the genome, and prioritizing those M0220 (2.46 g/l) and plant agar (8 g/l), with the addition lines covering whole chromosomes and lines that were of 10 ml of 100 μM benzyl aminopurine for seeds <5 mm. already producing fruits, were genotyped with the 18 k These were kept in a cold chamber at 4◦ C for 8–12 weeks Illumina chip [34]. For that, 1000 ng genomic DNA or until the onset of radical growth, then transferred to a was dried using a speed vacuum and eluted in 20 μl dark chamber (closed in carton box) at 23◦ C for one week. water to a final concentration of 50 ng/μl. DNA purity After a week in the dark, the embryos were exposed to was checked by Nanodrop absorbance values, with gradually increasing light conditions. Those plants with ratios of 1.8 for 260/280 and 2.0–2.2 for 260/230 being well-formed shoots and roots were moved to trays in the considered pure. The samples were genotyped at Fon- greenhouse, and finally to the field. dazione Edmund Mach facility, San Michelle all ‘Adige, Italy. Genotyping and introgression line extraction The raw data from the genotyping platform was pro- Genomic DNA was extracted from young leaves using the cessed using Genome Studio Illumina 2.0. The two out- CTAB method [32] in 96 well plates. Genotyping data from put files, final report and DNA report from the Genome the E2T set were available for 113 SSRs with almost full Studio analysis were used as input files for genetic anal- genome coverage of the peach genome, spanning 212.2 ysis using the ASSisT software [35], an automatic SNP Mbp (359.3 cM), with the average interval between the scoring tool, along with a generated pedigree file and markers being ∼1.9 Mbp (3.2 cM) (Serra et al. 2016; see map file. The population was set to germplasm, allowed Supplementary Table 2 and Supplementary Figure 1). missing data range to 1.0 and all the other parameters Additionally, data from the same SSRs and the 9 k (call rate tolerance, p-value segregation distortion, unex- peach SNP chip [33] had been previously obtained [19] pected genotype threshold per individual and frequency in the parents, the F1 and the T1E individuals used for rare allele) set to zero. This way, 41.3% of the SNPs were mapping. To select the ILs in the progenies derived from approved, which was the highest compared to the default plants of the E2T set we used the same 113 SSRs. The and all other settings we tried. Given that peach varieties two markers at the extremes of the introgressions of are known to have large regions identical by descent, each parent (Supplementary Table 2) were genotyped chromosomal fragments ≥2.5 Mbp long (approximately in the corresponding E2TS1 and E3T progenies, and 1% of the peach genome) with >3 in total or > 2 con- the individuals with a single introgression in homo- secutive SNPs in heterozygosis, were considered iden- or heterozygosis were selected. For plants having a tical by descent and consequently with two identical recombination within one of the introgressed fragments, haplotypes for “Earlygold”. SNP data from IL genotyping additional SSRs within this fragment were genotyped have been graphically represented with GGT2 (ref. [36]), until its extremes were identified. Selected individuals using the physical location of the markers in the Lovell were then genotyped with the additional SSRs with peach genome reference sequence v2.0 (ref [37]) as their coverage of the complete genome. positional reference. SSRs were analyzed in the ABI PRISM® 3130xl Genetic To identify the ILs obtained individually, we used an Analyzer (Applied Biosystems) by capillary electrophore- internal code to select an initial collection with SSRs sis following the PCR amplification method described (see Supplementary Tables 3 and 4 and Supplementary earlier [26]. For the PCR for SSR amplification, the reac- Figures 2 and 3). For the final IL collection using the tion mix (10 μl) contained 2 μl genomic DNA (200 ng), 1 μl data of the 18 k SNP chip, we coined a new terminol- of 10x NH4 reaction buffer, 0.3 μl of 50 mM NH4 MgCl2 , ogy: first a group of five letters and a number (the first 0.2 μl of 10 mM dNTP, 0.2 μl of 10 μM forward primer four letters “PAIL” to indicate Peach-Almond Introgres- labeled with fluorochrome (FAM, NED, VIC and PET), sion Line, the fifth letter is E or O indicating heterozygous 0.2 μl of 10 μM reverse primer, 0.2 μl of 5 U lab Taq and or homozygous ILs, respectively) and a digit 1–8 corre- 5.9 μl HPLC water. PCR was performed in a thermocycler sponding to the chromosome number where the intro- (Applied Biosystems, USA) with the following conditions: gression is located. The second part of the IL name is a a single cycle of initial denaturation at 94◦ C for 1 min, dash followed by four digits corresponding to the position 35 cycles of denaturation at 94◦ C for 15 s, annealing step in Mbp of the two extremes of the introgressed frag- between 50 to 65◦ C for 15 s and an extension step at ments. For example, line PAILO5–1219 is a homozygous 72◦ C for 30 s. One cycle of final elongation was at 72◦ C peach-almond introgression line located on chromosome 10 | Horticulture Research, 2022, 9: uhac070 5 and in the region spanning from the 12th to the 19th Mb. Author Contributions A small letter at the end of the IL indicates lines with the NK, OS, JMD and RP were responsible for data acquisition same fragment, e.g. PAILE2–0130a and PAILE2–0130b. and interpretation of the results in different steps of the research. NK was also responsible for compiling the available information and drafting the manuscript. Phenotyping and statistical analysis WH was involved in designing experiments and co- The IL collection was evaluated for traits known to have supervising the research of RP and OS. IE was involved in different phenotypes based on the previous information experimental design, data acquisition, co-supervision of of “Texas” × “Earlygold” progenies [20, 23]. These include NK and JMD, and reviewed the manuscript. PA conceived several major genes and a few major quantitative trait the research work, co-supervised NK, OS, JMD and RP loci (QTLs). The major genes were: juiciness (Jui/jui) as and reviewed and approved the final version of the the presence and absence of juice in the fruits at matu- manuscript. rity; blood flesh color (DBF2/dbf2) as red or yellow flesh color at maturity; maturity date (MD/md) as early or late maturing based on the number of Julian days when more Data availability statement than half of the fruits reach ripening stage, estimated on The data underlying this article are available in the arti- parameters such as fruit firmness and visual observation cle and in its online supplementary material. of fruit skin color change; and powdery mildew resis- tance (Vr3/vr3), scored as resistant or susceptible based on the absence or presence of fungal infection on the Conflict of interests leaves. A quantitative leaf character, petiole length (PL), The authors declare no competing financial interests. was measured in mm with a ruler in three to six average size leaves of each IL. We phenotyped three additional quantitative fruit traits, taking data from three samples Supplementary data (average of 3–4 fruits per sample) per tree: fruit weight Supplementary data is available at Horticulture Research (FW), titratable acidity (TA) and soluble solids content online. (SSC), measured as described before [20]. Statistical analysis of the data for quantitative traits was performed using the DescTools package of R. The References overall data were studied with a one-way analysis of 1. Zamir D. Improving plant breeding with exotic genetic libraries. variance, and Dunnett’s test used to compare the mean Nat Rev Genet. 2001;2:983–9. of each IL with that of the “Earlygold” control. An addi- 2. Balakrishnan D, Surapaneni M, Mesapogu S et al. Development tional Dunnett’s test was used to integrate the possible and use of chromosome segment substitution lines as a genetic effects of traits segregating in the “Earlygold” background resource for crop improvement. Theor Appl Genet. 2019;132:1–25. in the analysis, comparing the mean of each IL with that 3. Eshed Y, Zamir D. An introgression line population of Lycopersi- “Earlygold” and all the other ILs, excluding those that had con pennellii in the cultivated tomato enables the identification the same or part of the same introgressed fragment. Only and fine mapping of yield-associated QTL. Genetics. 1995;141: ILs showing highly significant values (P < 0.01) in both 1147–62. 4. Szalma SJ, Hostert BM, LeDeaux JR et al. QTL mapping with near- tests were considered as detecting a QTL for the trait isogenic lines in maize. Zuchter. 2007;114:1211–28. studied. 5. Szymański J, Bocobza S, Panda S et al. Analysis of wild tomato introgression lines elucidates the genetic basis of transcriptome and metabolome variation underlying fruit traits and pathogen response. Nat Genet. 2020;52:1111–21. Acknowledgments 6. Zhang G. Target chromosome-segment substitution: a way to We acknowledge financial support through “Severo breeding by design in rice. Crop J. 2021;9:658–68. Ochoa Programme for Centres of Excellence in R&D” SEV- 7. Frary A, Nesbitt TC, Grandillo S et al. fw2.2 a quantitative trait 2015-0533 and CEX2019-000902-S funded by MCIN/AEI locus key to the evolution of tomato fruit size. Science. 2000;289: 10.13039/501100011033, projects PID2019-110599RR-I00 85–8. and PCI2019-103670 funded by MCIN/AEI 10.13039/ 8. Fridman E, Carrari F, Liu Y-S et al. Zooming in on a quantitative 501100011033 and by the “European Union”, project trait for tomato yield using interspecific introgressions. Science. 2004;305:1786–9. RTA2015-00050-00-00 funded by MINECO and INIA, 9. Xu Q, Xu X, Shi Y et al. Elucidation of the molecular responses projects AGL2015-68329-R, RTI2018-100795-B-I00 and of a cucumber segment substitution line carrying Pm5.1 and its AGL2012-40228 funded by MCIN/AEI/10.13039/ recurrent parent triggered by powdery mildew by comparative 501100011033 and by “ERDF A way of making Europe”, transcriptome profiling. BMC Genomics. 2017;18:21. and project AGL2009-07305 funded by MICINN. Naveen 10. Balakrishnan D, Surapaneni M, Mesapogu S et al. Detect- Kalluri Grant BES-2016-077104 was funded by MCIN/AEI/ ing CSSLs and yield QTLs with additive, epistatic and QTL× 10.13039/501100011033 and by “ESF Investing in your environment interaction effects from Oryza sativa × O. nivara future”. IRGC81832 cross. Sci Rep. 2020;10:7766. Kalluri et al. | 11 11. Gur A, Zamir D. Mendelizing all components of a pyra- 25. Tanksley SD, Nelson JC. Advanced backcross QTL analysis: a mid of three yield QTL in tomato. Front Plant Sci. 2015;6: method for the simultaneous discovery and transfer of valuable 1096. QTLs from unadapted germplasm into elite breeding lines. Theor 12. Eshed Y, Abu-Abied M, Saranga Y et al. Lycopersicon esculentum Appl Genet. 1996;92:191–203. lines containing small overlapping introgressions from L. pennel- 26. Aranzana MJ, Carbo J, Arús P. Microsatellite variability lii. Theor Appl Genet. 1992;83:1027–34. in peach Prunus persica (L.) Batsch: cultivar identification, 13. Yu Y, Fu J, Xu Y et al. Genome re-sequencing reveals the evo- marker mutation, pedigree inferences and population structure. lutionary history of peach fruit edibility. Nat Commun. 2018;9: Theor Appl Genet. 2003;106:1341–52. 5404. 27. Micheletti D, Dettori MT, Micali S et al. Whole-genome analysis of 14. Alioto T, Alexiou KG, Bardil A et al. Transposons played a major diversity and SNP-major Gene Association in peach Germplasm. role in the diversification between the closely related almond PLoS One. 2015;10:e0136803. and peach genomes: results from the almond genome sequence. 28. Scorza R, Mehlenbacher SA, Lightner GW. Inbreeding and Plant J. 2020;101:455–72. coancestry of freestone peach cultivars of the eastern-United- 15. Faust M, Timon B. Horticultural Reviews Vol. 17. In: Janick J (ed.), States and implications for peach germplasm improvement. Ch. 10. New York: Wiley, 1995. J Amer Soc Hort Sci. 1985;110:547–52. 16. Ladizinsky G. On the origin of almond. Genet Resour Crop Evol. 29. Cirilli M, Baccichet I, Chiozzotto R et al. Genetic and phenotypic 1999;46:143–7.. analyses reveal major quantitative loci associated to fruit size 17. Byrne DH. Isozyme variability in four diploid stone fruits and shape traits in a non-flat peach collection (P. persica L. compared with other woody perennial plants. J Hered. Batsch). Hortic Res. 2021;8:232. 1990;81:68–71. 30. Hernández Mora JR, Micheletti D, Bink M et al. Integrated QTL 18. Velasco D, Hough J, Aradhya M et al. Evolutionary genomics of detection for key breeding traits in multiple peach progenies. peach and almond domestication. G three. 2016;6:3985–93. BMC Genomics. 2017;18:404. 19. Donoso JM, Eduardo I, Picañol R et al. High-density mapping 31. Ramming DW. The use of embryo culture in fruit breeding. suggests cytoplasmic male sterility with two restorer genes in HortScience. 1990;25:393–8. almond × peach progenies. Hortic Res. 2015;2:15016. 32. Doyle JJ, Doyle JL. A rapid Total DNA preparation procedure for 20. Donoso JM, Picañol R, Serra O et al. Exploring almond genetic fresh plant tissue. Focus. 1990, 1990;12:13–5. variability useful for peach improvement: mapping major genes 33. Verde I, Bassil N, Scalabrin S et al. Development and evaluation and QTLs in two interspecific almond × peach populations. Mol of a 9K SNP Array for peach by internationally coordinated Breeding. 2016;36:16. SNP detection and validation in breeding Germplasm. PLoS One. 21. Serra O, Donoso JM, Picañol R et al. Marker-assisted introgression 2012;7:e35668. (MAI) of almond genes into the peach background: a fast method 34. Gasic K, Linge CS, Bianco L et al. Development and evaluation of to mine and integrate novel variation from exotic sources in long a 9K SNP addition to the peach IPSC 9K SNP array v1. HortScience. intergeneration species. Tree Genet Genomes. 2016;12:96. 2019;54:S188–8. 22. Marimon N, Luque J, Arús P et al. Fine mapping and identification 35. Di Guardo M, Micheletti D, Bianco L et al. ASSIsT: an automatic of candidate genes for the peach powdery mildew resistance SNP scoring tool for in- and outbreeding species. Bioinformatics. gene Vr3. Hortic Res. 2020;7:175. 2015;31:3873–4. 23. Kalluri N, Eduardo I, Arús P. Comparative QTL analysis 36. van Berloo RGGT. 2.0: versatile software for visualization and in peach ‘Earlygold’ F2 and backcross progenies. Sci Hortic. analysis of genetic data. J Hered. 2008;99:232–6. 2022;293:110726. 37. Verde I, Jenkins J, Dondini L et al. The peach v2.0 release: high- 24. Pirona R, Eduardo I, Pacheco I et al. Fine mapping and identifica- resolution linkage mapping and deep resequencing improve tion of a candidate gene for a major locus controlling maturity chromosome-scale assembly and contiguity. BMC Genomics. date in peach. BMC Plant Biol. 2013;13:166. 2017;18:225.