A Large-Scale Validation of an Improved Embryo-Rescue Protocol for the Obtainment of New Table-Grape Seedless Genotypes

The new trends in the consumption of table grapes and the growing interest in the environmental impact of this crop have pushed breeders toward the development of seedless cultivars endowed with resistance, through crossbreeding programs. To obtain seedless grapes, the use of embryo-rescue techniques is fundamental. In this research, a grape embryo-culture protocol was optimized and validated by using 39 cultivars and 41 cross-combinations carried out in the framework of a large private table grape program of the private network Italian Variety Club in the period 2017–2021 evaluating several factors, such as the improvement in embryo formation, germination and growth, and plantlet development. The embryo culture attitude of crosses between different combinations of seedless parents was assessed, and the rates of embryo development from the extracted ovules mostly ranged from 3.5 to 35.5% with 5 out of 43 genotypes outliers. Experiments conducted at different sampling times, in a range of 43–62 days after pollination (DAP), did not show significant differences between the samples analyzed, while the rate of embryos developed with the applied protocol proved its employability on multiple genotypes, although the grapevine genotype significantly influenced the technique efficiency.


Introduction
The European Union is the world's main grape producer, with an estimated production of 1.7 million tonnes of grapes for fresh consumption in 2021 [1].The market of table grapes is the most important, with an import value of EUR 1.6 billion, after bananas and avocados, in the European scenario.With a production stabilized around 1 million tons per year of table grapes, Italy is the main producer in Europe [1].In addition to volume, the strong point of Italy is the extensive area of cultivation of organic grapes (2177 hectares of organic table grapes in 2017) [2].
To improve production, Italian growers are slowly joining the seedless trend, mostly in the Puglia region, and at the same time, they are looking for late cultivars to extend the seasonal availability of produce.In the development of new cultivars, the main sought characteristics are the general berry traits (weight, color, shape, and skin thickness), flavor (aromas and sweetness), fruit shelf life, harvesting time, and tolerance to pathogens, such as Plasmopara viticola and Erysiphe necator.However, nowadays, the seedlessness and disease resistance traits are the most desirable [3].Over the last few years, the quality and nutritional composition and the breeding of new seedless cultivars have been the most significant efforts in research [4][5][6].
From a botanical point of view, seedlessness is the phenomenon whereby fruit is formed without developing seeds, and plants are sterile due to the absence of embryos.Particularly in the cultivated grapevine (Vitis vinifera subsp.sativa L.), parthenocarpy and stenospermocarpy are the two types of fruit seedlessness [7,8].In parthenocarpic cultivars (Corinth type), berry production is not preceded by ovule fecundation, while in stenospermocarpic cultivars (Sultanina type), some of the ovules possess normal embryo sacs, fertilization occurs, and embryos may develop, but their development is often stopped, and normal seeds cannot form [9,10].Despite this information, nowadays, the exact causes of embryo abortion in seedless grapevines are still not fully understood; the main hypothesis proposes that parental tissues alter the hormonal balance during the first steps of the ontogenetical cycle, causing the embryo's abortion [11].It is then obvious that conventional breeding methods cannot be applied for genetic improvement, and in vitro techniques are the only available approach.The first reported vines originated via ovule culture from the seedless grape in 1983 [11], and it was followed by several other reports [12][13][14][15][16][17][18][19][20][21][22][23]; the exploited biotechnological principle consists of the cultivation of extracted ovules or immature embryos in artificial media, to admit the development of the plantlet without the abortion caused by a natural process.Nowadays, grape breeding projects using this in vitro technique can improve the efficiency, by reducing the time required to obtain seedless cultivars, usually from 6 to 8 years [20].Moreover, by adopting seedless stenospermic instead of seeded cultivars as female parents, the frequency of seedlessness found in the progeny could be increased by up to 10-15% [24].When the rescue is performed in "seedless × seedless" crosses, where female and male parents are from a "seeded × seedless" or a "seedless × seedless" cross, a high percentage of seedless progeny with natural big berries and small seed traces is obtained [3,25].The main steps of the embryo-rescue technique are as follows: (i) in vitro culturing of ovules, (ii) embryo collection and culturing for plant development, and (iii) managing of rooted plantlets (elongation, acclimatization, and transplanting to soil) [24].The most crucial step is the first because an optimal culture medium is the key to a successful rescue of potentially abortive hybrid embryos.Indeed, an appropriate composition of growth regulators in the medium is crucial for embryo germination and growth and plantlet development.However, the number of obtained embryos also depends on the parent's genotype and harvesting [11,[26][27][28][29][30][31][32][33].On stenosternocarpic seedless grapes, collection time is important to block embryo abortion, considering that it determines the herbaceous or woody texture of seed tissues [11,31,33,34].According to previous reports, the best sampling time to achieve the maximum efficiency of the embryo-rescue protocol is 40-60 days after pollination [14,33,35,36].
The technique of the rescue of immature embryos provides important genetic gains for the seedless feature, due to the size reduction of seed traces resulting from crosses between seedless parents [26].The advantages of the technique were also evaluated in the genetic improvement of grapevine for disease resistance [37], a highly relevant issue since the 50% cut in pesticide use by 2030 is a key goal of the European Farm to Fork strategy.The aim of this work was the improvement and validation of an embryo-rescue protocol working with embryos from several crosses between seedless parents selected for features like moscato flavor and putative resistance to the most relevant pathogens, E. necator and P. viticola.

Efficiency of the Embryo-Rescue Protocol
Data on the mean numbers of berries and ovules collected from each bunch, the numbers of ovules extracted and those developed from each berry, and the proportions of developed embryos on the in vitro settled ovules are reported in Table 1.In five years of work, a total of 43 crosses carried out with 25 female parental tablegrape cultivars were compared.In detail, 449 bunches and 39,351 berries were manipulated to obtain 88,837 in vitro settled ovules and 11,941 germinated embryos.For each bunch, the number of berries was in the range of 27 to 215, and each berry had on average two ovules.Significant differences were recorded in the six crosses in which TS28, TS25, and TS27 were used as the female parent.The ratios of ovules:berries were below 1.5:1 in the four crosses including TS28 and TS25, while the ratio values were higher than 3.2:1 in the two crosses including TS27 as female parents.
The percentage of ovules that developed embryos was in the range of 3.5% to 35.4% for 38 crosses, with 5 out of 43 outlier genotypes.The embryogenetic efficiency was indeed less than 3.5% in the cross TS04 × TS24 and all three crosses with TS25 as the female parent, while the cross TS29 × TS38 yielded the highest efficiency with a value of 41.1%.Summarizing the data from all the crosses, the percentages of ovules developed in embryos were less than 5.0% for only 2 crosses and were in the range of 6.0-10.0%for 11 crosses, 11.0-20.0%for 14 crosses, and higher than 20% for the remaining 8 crosses.The protocol efficiency, relative to the number of embryos developed by settled ovules, was between 3.0 and 14.0% in more than half of the crosses and between 8.2 and 11.8% in all the crosses involving TS27 as the female parent.
As expected, not all the obtained embryos were successful in developing plantlets (Table 2 and Figure 1).Different causes, including in vitro and in vivo microbial contaminations were responsible for losses during the acclimatization steps; the percentages of embryos that originated plants were in the range between 10.0% (TS15 × TS32) and 62.4% (TS33 × TS28), but only in six crosses were below 20.0%.More than half of the crosses had a percentage of losses in the stage of acclimatization below 60.0%.Considering the ovules, the percentage of plantlets was in the range of 0.3% (TS04 × TS24; TS25 × TS39) − 20.5% (TS29 × TS38).* % of developed embryos: the ratio was calculated by multiplying by 100 the ratio plantlets/embryos; ** % of developed ovules: the ratio was calculated by multiplying by 100 the ratio plantlets/ovules.

Sampling Time
Detailed results obtained by comparing cultivars characterized by a similar ripening time (grouped as middle, middle-late, and late ripening time) for the numbers of ovules extracted per berry and the percentages of settled ovules developed in acclimatized plants at different sampling times are reported in Supplementary Table S1 and Figure 2. Considering the crosses characterized by the middle ripening moment of the female parent, the mean ratio of ovules:berries showed a tendency to increase for the earlier three sampling times and decrease at 58-62 days after pollination (DAP), so that the highest values, over 2.5 ovules per berry, were recorded at 53-57 DAP.The cross TS27 × TS03 showed the highest outstanding value of 3.4 ovules:berries at 48-52 DAP.For the same cultivars, the embryogenic efficiency was between 10.0 and 12.0% for the earlier three sampling times, while the highest mean value (17.1%) was recorded at the last sampling.Almost all the data were in the range of 4.0-20.0%.The only exception was the cross TS17 × TS27 at 43-47 DAP (24.7%).
For the crosses with female parents with middle-late ripening time, the ratios of ovules:berries were in the range of 1.5 to 2.6.TS04 × TS24 was the only cross sampled at the first sampling time and yielded an ovules:berry ratio of 2.3 and a percentage of embryo development of 0.8%.The ovules:berries ratios showed the highest mean value (2.5) at 48-52 DAP; after that, the value decreased and increased again at 58-62 DAP (2.3).Ovules:berries ratios for all crosses were in the range of 1.5-2.6.The best embryogenic efficiency in this group was 20.1%, recorded at 53-57 DAP, while the highest value (27.6%) and the lowest value (1.0%) were recorded at 58-62 DAP (TS33 × TS28) and 53-57 DAP (TS04 × TS24), respectively.The sample at 53-57 DAP was characterized by two outstanding high values, 20.1% for TS23 × TS28 and 0.9% for TS04 × TS24.

Discussion
Since the first report of the usage of in vitro ovule culture for originating recombinant seedless grape genotypes [11], the usefulness of the technique has become clear, and numerous factors have been studied to improve the efficiency of applied protocols.It was proved that the number of obtained embryos depends on the culture medium [29,33], parental genotypes [27,28], and the time of berry collection [11,13,29,30,32,38,39].Several papers reported the use of Nitsch & Nitsch's (NN) salts [15,20,40,41], different hormones, such as IAA or GA3 [38,42], as well as activated charcoal [15,43,44], and mashed banana [33].Most of the factors were also studied in this research, paying particular attention to grapevine genotypes and the role of hormone concentrations in the culture medium [38].The outcome was an improved medium for the stabilization of explanted ovules and for embryo growth.The medium contained macro-and NN microelements [45], IAA (3.00 mg/L), and GA3 (2.00 mg/L) and was supplemented with activated charcoal (2 g/L) for reducing tissue browning and embryo abortion rate [41,42].The improved medium and the related protocol were validated over five years on 43 different crosses involving 39 grapevine genotypes characterized by different ripening times.Overall, the adopted protocol yielded an efficiency of embryos developed by in vitro settled ovules ranging from 3.0 to 14.0% in more than half of the assayed crosses.
The large-scale experiments on very high numbers of berries, ovules, and embryos allowed us to ascertain the adaptability of the newly developed protocol to different grapevine genotypes.The obtained efficiencies of plantlet formation were in line with previously reported data [27,33,34,38,41,46].
The results of this research evidenced a high variability of the ratios between embryo growth and ovules, in the range from 0.8% (TS04 × TS24) to 41.1% (TS29 × TS38), depending on the crossed grapevine genotypes.The influence of the genotype was also confirmed by the similar results obtained in different years using TS27 and TS25 as the female parent in crosses, with ratios around 10.0 and 2.0%, respectively.Our results align with the ones reported by Puglisi et al. [27] reporting great variability among parents, also when the same cultivar was used as the male or female parent.
The efficiency of the new protocol in terms of percentages of acclimatized ready-to-use plants from grown embryos was in the range of 10.0-62.0%,fitting well with the results obtained by Jiao et al. [28] reporting a plantlet development rate between 17 and 58%, using 15 different genotypes and 11 cross combinations [28].
An overall protocol efficiency over 20.0% was obtained by the cross between TS29 (female, seedless, early ripening) and TS38 (male, seedless, early ripening), confirming the influence of grapevine genotypes on the embryogenic ability of ovules, on the ability of plantlet development in vitro and, consequently, on their aptitude to breeding [18,27,28,31,36].
Li et al. [47] observed different abilities of embryo growth depending on the sampling time of berries in the cultivars "Thompson seedless", "Flame seedless" (TS11), and "Ruby seedless".They concluded that the length of time elapsing between pollination and berry collection was significantly and negatively correlated with the proportions of grown embryos in seedless grapes.Similar findings, but under a significant influence of grapevine genotypes, were also reported by others [30,31,33,48].We applied the new protocol to berries collected at four five-day intervals in the period from 43 to 62 DAP.Overall, data showed a tendency for a higher protocol embryogenetic efficiency with sampling at 58-62 DAP.These results agree with findings reported by Guo et al. [48] based on observations carried out on crosses in which the seedless cultivars "Jumegui", "Kyoho", and "Red globe" were used as the female parent.Nevertheless, the embryogenetic efficiency did not show significant differences at different sampling times for crosses involving grapevine genotypes with medium, medium-late, or late ripening as also reported by Kebeli et al. [29].These findings can be relevant for the application of the embryo-rescue protocol in largescale breeding projects in which a very high number of berries need to be processed.The slight, if any, influence of sampling time on the results in terms of final plant production, indeed, could reduce the laboratory workload in the earlier and time-consuming steps of the protocol.

Embryo-Rescue Protocol
All the grape seeds used from the embryo rescue activities originated from crosses between different V. vinifera cultivars.The vines were grown in commercial vineyards located in Puglia (Southern Italy).Thirty-nine cross-combinations were set up during a period of five years (2017-2021).The parental genotypes herein used (Table 3) were selected by the Scientific Committee of the Rete Italian Variety Club, a network joining 23 private companies (www.reteivc.it(accessed on 30 March 2023)) [49], in the frame of a multi-year and extensive table large breeding program launched in 2014.Selection was mainly for the following characteristics: seedlessness, berry size, color, pulp firmness, ripening time, and putative resistance to E. necator and/or P. viticola.The cultivars used as parents are almost all patented so the lab code TS01-TS39 was used instead of the cultivar name, according to the confidentiality agreement between researchers and breeders.If known, their pedigree is reported.For the same reason, no details on the crosses are herein supplied.
Unlike a previous breeding work in which seedless cultivars were only used as pollen parents [50], the development of the ovule-culture protocol made it possible to use seedless vines as both parents [11][12][13] increasing the number of progenies marked by the seedless characteristic [51].A traditional grapevine breeding technique based on the emasculation of flowers, as proposed by Eibach [48], was used.Briefly, the inflorescences of already fertilized flowers characterized by raised calipers were removed.At a different time during May, before blooming, flowers of the female parent of the cross were emasculated and closed in paper bags containing pollen from the male parent.Details on crosses are in Table 4. Air-dried berries were dissected longitudinally with a sterile scalpel (Figure 3a) with the help of a stereo microscope under a laminar flow hood, and ovules, well cleaned by all berry tissue residues, were gently and accurately clamp-collected.The ovules were placed in Petri dishes (diameter 90 mm, height 16.2 mm) containing about 20 mL of the culture medium.Based on previous works [15,40,[43][44][45], the composition of the medium, containing active charcoal, used in this study was optimized as reported in Supplementary Table S2.In each Petri dish, 30-40 explants from the same bunch were placed (Figure 3b).Plates, sealed with Parafilm, were maintained at 24 ± 3 • C under lighting from white light LED with a luminous intensity of 3000 lux, with a photoperiod of 16/8 h.From 8 to 30 weeks, the spontaneous growth of embryos (Figure 3c) was checked by the naked eye.After 10 weeks, the embryos were manually collected from ovules showing externally brown tissues (Figure 3d).Due to the long-time growth required, ovules were individually transferred onto a fresh growth medium, especially when dehydration or microbial contamination occurred.The grown sprouts were transferred into new sterile, transparent, and airtight 60 mL-sized containers (Figure 3e), containing 15 mL of a culture medium with no hormones for settling.Details on the used media are available in Supplementary Tables S2 and S3 [52,53].The sprouts, properly oriented according to the embryo orientation, were inserted approximately 1 mm into the medium and maintained under the same above-described growing conditions until the complete plantlet development (Figure 3e).Once the plants within the substrate developed their first two true leaves and reached a height of at least 3-4 cm, they were ready for transplanting.Plantlets were individually collected from in vitro containers and transferred to previously rehydrated disks of dehydrated peat (Jiffy-7 of 44 mm, Jiffy Products International AS, Norway).Properly carded plants were placed in trays to facilitate their watering.Trays, covered with a transparent plastic film to ensure a gradual transition to the external conditions, were maintained at 24 ± 3 • C in a climatic chamber equipped with LED lights (AGRO light of 22 W) and a photoperiod of 16/8 h.About 10 days later, the plastic film was gradually removed.Well-developed plants were finally transferred into pots (9 cm × 9 cm × 10 cm) containing peat and maintained in the growth chamber until reaching a height of about 12 cm, at which point, they were transferred to the screenhouse (Figure 3f).

Influence of Sampling Time
For some crosses, bunches were sampled at two to three time points in the range 43-62 DAP to evaluate the possible influence of the female parent genotype on the best collection time [13,30,31].The sampling times were classified into four groups each of 5 days: 43-47, 48-52, 53-57, and 58-62 DAP (Table 5).Ovules were processed as described above.

Statistical Analysis
For each experiment, numbers of ovules extracted from each berry, embryo growth rates, and percentages of ovules yielding acclimatized plants were determined as follows: -Number of ovules extracted from each berry = total number of ovules extracted and settled in vitro/total number of sampled berries.-Embryo germination rate (%) = number of embryos grown from in vitro settled ovules ×100/number of in vitro settled ovules.-Acclimatized plant rate (%) = number of plants acclimatized from in vitro settled embryos ×100/number of in vitro settled ovules.Single bunches were considered biological replicates when appropriate, and standard error was calculated per each cross.
For the statistical analysis of the data, Microsoft Excel of Microsoft 365 was used.

Conclusions
A new improved embryo-rescue protocol for the obtainment of new seedless grapevine genotypes was validated in a large-scale experiment on a set of 43 crosses involving 39 parental seedless genotypes over 5 years.The feasibility of the new protocol was proved for all parental genotypes although these influenced the development rates of in vitro settled ovules.The sampling time of grapes in vineyards was evaluated in the range from 43 to 62 days after pollination, and it had little influence on the protocol efficiency, although better results were generally obtained when it was carried out from 53 to 62 days after pollination.The little influence of sampling time could be due to the main usage of crosses involving middle-to late-ripening cultivars.Research on this hypothesis is in progress.The obtained results highlight the relevance of a careful selection of parental genotypes in crosses by breeders due to their influence on the success of the economic investment required for the genetic improvement of the grapevine.

Figure 2 .
Figure 2. Comparison of ratio ovules:berries (A) and percentage of embryos developed from in vitro settled ovules (B) for each group of crosses (characterized by female's parent middle (M), middle-late (ML), and late (L) ripening time) at different sampling times (43-47 DAP, 48-52 DAP, 53-57 DAP, 58-62 DAP).For each sampling time and group of crosses, mean value (X), quartiles (-), and outstanding values ( • ) are reported; these data are condensed in X when there is only one sample.

Figure 3 .
Figure 3. (a) Berry dissection for ovule extraction; (b) settling of ovules in Petri dishes on embryo growing medium; (c) spontaneous growth from ovules settled on embryo growth medium; (d) embryo extracted manually with a scalpel from mature ovule; (e) grown embryos, settled on S.O.R. medium; (f) acclimatized plants in pots located in screenhouses.

Table 1 .
Data on plant material used in each cross and on the obtained embryos.

Mean Number per Bunch a No. of Ovules per Berry
a Mean values ± Standard Error.

Table 2 .
Efficiency of the protocol in terms of acclimatized plantlets.

Table 4 .
Numbers of sampled bunches and berries and in vitro settled ovules per cross.

Table 5 .
Numbers of bunches, berries, and ovules analyzed per cross at different sampling time points.