The experiment was carried out at a vineyard (28°38′37.8″N, 77°09′27.8″E) of the Division of Fruits and Horticultural Technology, ICAR-Indian Agricultural Research Institute, New Delhi, in 2023. The plant materials for the study were comprised of 10 grape genotypes belonging to three species, i.e., V. vinifera L., V. champini Planc., and V. parviflora Roxb., and were selected for pollen micro-morphology and storage studies ( Table 1 ). All vines were own-rooted, 10 years old, and trained on a bower system. For cross-compatibility study, Salt Creek (V. champini Planc.), Male Hybrid (V. vinifera L.), Pusa Navrang (V. vinifera L.), and Dogridge (V. champini Planc.) were used as male parents and V. parviflora Roxb. as a female parent.

Inflorescences expected to bloom within 7–8 days were covered with perforated butter paper bags. Pollen was collected from bagged panicles of three vines of each genotype between 7 am to 9 am. These panicles were monitored regularly and excised when they reached “EL 23 Full flowering: 50 % flower hoods fallen” according to the Eichhorn and Lorenz (E-L) system (Lorenz et al., 1995). Subsequently, the collected clusters were transferred to the laboratory and dried for 24 hours at room temperature. Collected floral debris was passed through sieves to obtain pollen and transferred to 1.5ml vials by brush (Rane et al., 2023).

Dehydration of pollen was performed by placing vials with their lids open in the silica gel desiccator at 4 °C until analysis with proper labeling to maintain their identity and avoid contamination (Lukšić et al., 2022). After dehydration, the small quantities of pollen were mounted for examination using a fine brush on aluminum stubs that were coated with double-sided transparent tape. Subsequently, a 0.02-µm-thick gold layer was applied using a sputter coater (EMITECH SC7620 mini sputter coater). At each stage, contamination of samples by pollen from different genotypes was avoided. The samples were examined at 10 kV using a scanning electron microscope (TESCAN VEGA3). Measurements of 30 pollen grains from each cultivar were taken and recorded as data for three separate replicates per genotype, with each replicate consisting of 10 pollen grains. These observations were made at a magnification of ×3,000 for whole grain and ×15,000 for exine ornamentation. The pollen traits were, namely, number of apertures, length of polar axis (P), length of equatorial axis (E), P/E ratio, length and width of colpus, mesocolpium width, and exine surface ornamentation. Shapes of pollen were categorized according to Erdtman (1971).

The pollen samples for storage study were prepared separately. For this purpose, pollen was stored in sealed vials with proper labeling. Pollen of Early Perlette Selection (EPS), Pearl of Csaba (POC), PER, Pusa Navrang (PN), Beauty Seedless (BS), Flame Seedless (FS), Male Hybrid (MH), and Salt Creek (SC) was stored under four different storage conditions, viz., at room temperature, 4 °C, −20 °C, and −196 °C (in liquid N₂). In vitro pollen germination and viability were tested after the 0th day (just after fresh pollen collection), 1st day, 3rd day, 5th day, 7th day, 15th day, and 30th day of storage under each storage condition. The samples were stored in multiple sets of vials for each genotype to minimize the stress linked to thawing.

The determination of pollen viability was performed by using 1% of 2,3,5-triphenyltetrazolium chloride (TTC) diluted in a 50 % sucrose solution (Shivanna, 2003 modified). One drop of solution was placed on a labeled slide, and stored pollen grains were spread with a brush on the slide and covered with a coverslip on the top. The slides were placed in a Petri dish and placed in the dark at 35°C with 50 % relative humidity (RH) for 6 hours. Pollen viability counts were made under a light microscope (OLYMPUS CX33, Tokyo, Japan). Pollen grains were considered viable if they were stained with orange or bright red color. Three randomly chosen microscopic areas (each with a minimum of 100 pollen grains) were used to count the pollen grains on each slide.

In vitro pollen germination was observed on standardized semi-solid medium containing 10 % sucrose + 100 mg/L boric acid + 300 mg/L calcium nitrate and agar 0.6%, at pH 5.85 ( Supplementary Table 1 ). A small rectangular piece of medium was removed by a scalpel and placed on a labeled slide. Pollen grains were dusted uniformly on the surface of the medium with a fine brush. Slides were placed in a Petri dish with damp filter paper and covered with a lid to avoid drying of the medium and maintain humidity. The Petri dishes were placed in a dark environment for incubation. The growth of pollen tubes on the medium was observed under a light microscope (OLYMPUS CX33) after 24 hours of incubation. Pollen grain was considered germinated if the length of the tube exceeded the pollen (Nepi and Franchi, 2000). The germination rate was calculated by counting three fields per sample, each comprising 100 pollen grains.

Analysis of variance (ANOVA) was performed using PROG GLM of the SAS software package, version 9.4 (SAS Institute, Cary, NC, USA). Pollen dimension and shape-related parameters were analyzed through one-way ANOVA. Further, a two-way ANOVA was performed for each genotype separately, with storage temperature and duration as the factors, to study their impact on pollen germination (%) and viability. Tukey's Honest Significant Difference (HSD) test was used to identify the pairwise significant differences among the various factors and their interactions. Data were transformed using square root transformation of the form (x + 0.5), {i.e.√(x + 0.5)} for the parameters like colpus length, colpus width, mesocolpium width, pollen viability, and fruit set (%), and arcsine transformation was used for pollen germination (%) to make it amenable for ANOVA. A p < 0.05 was considered statistically significant.

Among the 10 Vitis genotypes investigated in this study, EPS, PN, POC, PER, BS, and FS from V. vinifera L. were identified as hermaphrodites, exhibiting well-developed stamens and pistils. However, two genotypes, MH from V. vinifera L. and SC from V. champini Planch., were functionally male, having a rudimentary ovary and well-developed stamens ( Figure 1 ). VP and DR (V. champini Planc.) had well-developed pistils and reflexed stamens. The pollen of genotypes with reflexed stamens (DR and VP) exhibited inaperturate or acolporate pollen, which was subprolate (P/E ratio: 1.14–1.33) in shape ( Figures 2A, B and Table 2 ). However, the remaining genotypes were tricolporate and prolate (P/E ratio: 1.33–2.00) in shape ( Figures 2C–J and Table 2 ). The EPS and MH genotypes had reticulate exine ornamentation, while BS had reticulate-striate exine ornamentation. The remaining genotypes, i.e., PN, PER, POC, FS, DR, and VP, possess foveolate-perforate exine ornamentation.

The study of pollen dimensions ( Table 2 ) showed that the length of the polar axis ranged from 23.524 µm to 31.181 µm and that the length of the equatorial axis ranged from 14.090 µm to 20.657 µm among the 10 grape genotypes. The average values for the P/E ratio varied between 1.255 and 1.962. The P/E ratio did not differ significantly in the tricolporate type of pollen. However, acolporate or inaperturate pollen had the lowest P/E ratio. The minimum colpus length (22.119 µm) was recorded in genotype BS. However, PN had the maximum colpus width (0.809 µm), which differed non-significantly from PER (0.653 µm). However, the minimum colpus width (0.378 µm) was observed in MH, which was statistically non-significant with EPS, POS, BS, FS, and SC. Considerable variation was also recorded in the mesocolpium width of all studied genotypes ranging from 9.886 µm to 13.599 µm ( Table 2 ).

In all studied Vitis genotypes, maximum germination and viability were recorded in fresh pollen ( Figure 3 ). The germination of fresh pollen ranged from 40.4% (POC) to 90.4 % (FS) ( Table 3 ), while pollen viability rate ranged from 40% (POC) to 95 % (FS). Notably, there was a higher degree of variation observed among the genotypes concerning both pollen viability and germination. However, pollen of DR and VP were inaperturate and did not germinate in vitro, but they did show viability during staining by TTC ( Figure 4 ). The pollen of these two genotypes was not studied further for storage studies.

In this study, four different temperatures, viz., room temperature, 4 °C, −20 °C, and −196 °C, were used to store pollen. For better understanding and to reduce the repetition, the results are discussed storage temperature-wise. At all the storage temperatures in every genotype, the maximum pollen germination and pollen viability was on the day of collection ( Tables 3 , 4 ).

Pollen stored at room temperature recorded a sharp reduction in germination and viability rate over time in all the genotypes. At room temperature, pollen showed germination and viability extending up to the seventh day of storage. Beyond this period, it declined to zero in all grape genotypes except POC and PN. In the case of these two genotypes, storage for only 3 days was possible. However, the acceptable pollen germination percentage of >30% was found up to only the first day of storage in PN, POC, and BS; up to the third day in EPS; up to the fifth day in PER and MH; and a maximum of up to 7 days in FS and SC ( Tables 3 , 4 ). FS and SC recorded pollen germination rates of 51.6 % and 46.7 %, respectively, even on the seventh day of storage, which is significantly higher than those of other studied genotypes under similar storage conditions. In all the genotypes, pollen germination was consistently the lowest when stored at room temperature, regardless of the duration. The decline in both pollen germination and viability occurred notably faster at room temperature compared to the other storage temperatures.

Pollen germination and pollen viability at 4 °C are presented in Tables 3 , 4 . In all genotypes, pollen germination was more than acceptable, i.e., 30 % during 30 days of storage except in POC, which had it for only up to the third day of storage. In all the genotypes, pollen viability was significantly higher at 4 °C as compared to room temperature.

In all the genotypes, pollen germination percentage had a non-significant difference during the whole storage period at −20 °C and −196°C except in POC. Pollen stored at −196 °C maintained higher pollen germination compared to other storage conditions. However, in the case of pollen viability, the non-significant difference between pollen viability at different storage durations exists only at −196°C. The pollen storage at −20 °C had intermediate pollen viability with respect to pollen viability at −4°C and −196 °C.

Palynology is primarily concerned with the study of various pollen features, which could be utilized in plant classification and phylogenetic studies (Claxton et al., 2005; Al-Hakimi et al., 2017) at the same time; the applied part of pollen-based information could be utilized for many purposes, viz., crop improvement and conservation of plant genetic resources. In this study, the basic purpose behind studying pollen-based features, i.e., the palynology of grape genotypes of three different species with asynchronous flowering ( Supplementary Figure 1 ), was to find out the possibility of their utilization in successful pollination and fertilization for grape rootstock improvement. These three grape species, V. vinifera L., V. champini Planc., and V. parviflora Roxb., cannot be differentiated based on the pollen micro-morphology, as the genotypes that belonged to them did not indicate marked variation in parameters like polar, equatorial axes and colpus dimensions. Similar to our findings, Soares et al. (2013) observed variations in Passiflora edulis and Passiflora setacea, but these variations were not significant enough to differentiate them. However, Joujeh et al. (2019) found variations in pollen color, exine ornamentation pattern, and density of spines distributed on the exine surface of six species of Centaurea genus, which can be utilized for deciding taxonomic position.

Apertures are key pollen micro-morphology-based features for identifying pollen grains. According to Moore and Webb (1978), these apertures are the primary characteristics utilized in distinguishing variations among pollen grains or spore fossils. In this study, two categories of pollen were found based on the number of apertures. The first one comprises acolporated pollen found in DR (V. champini Planc.) and VP (V. parviflora Roxb.). The second group characterized by tricolporate pollen consists of eight grape genotypes, including genotypes from V. vinifera and V. champini Planc. The presence of an aperture determines the functionality of pollen. The presence of two forms of pollen (acolporate and tricolporate), i.e., pollen dimorphism, is documented in various species of Vitis, namely, V. riparia Michx. (Kevan et al., 1985), Vitis aestivalis Michx. (Kevan et al., 1988), and Vitis coignetiae Pulliat. (Kimura et al., 1988), and represents a form of functional dioecy. The presence of an aperture is also linked to the type of flower. Genotypes having inaperturate (acolporate) pollen typically exhibit well-developed pistils surrounded by reflexed and short stamens (Vasconcelos et al., 2009). This floral structure is linked with the production of inaperturate pollen, which is generally considered sterile or non-functional (Meneghetti et al., 2006; Barbieri et al., 2012). In contrast, tricolporate pollen is linked to hermaphrodite flowers (Pratt, 1971; Punt et al., 2003; Vasconcelos et al., 2009). Acolporate pollen lacks furrows (Furness, 2007); as a result, it cannot germinate (Abreu et al., 2006). Meanwhile, tricolporate pollen has three furrows, is fertile, and can germinate (Padureanu and Patras, 2018). This was further reflected in the fertility of the pollen of these genotypes ( Tables 3 , 4 ; Figures 5 – 9 ). The presence of acolporate/inaperturate pollen within the same genus or species is linked with their different development from the normal one. A gene, INAPERTURATE POLLEN1 (INP1), specifically involved in the formation of pollen apertures, has been recently isolated and characterized in Arabidopsis thaliana; the loss of INP1 entails the lack of furrows and is responsible for the formation of inaperturate pollen (Dobritsa and Coerper, 2012). It further results in the uniform thickening of cell walls of pollen, lack of colpi, and an abnormal round shape. This could be the probable cause of acolporate pollen in DR and VP in this study.

Acolporated pollen in DR (V. champini Planc.) and V. parviflora Roxb. was unable to germinate on in vitro medium but was stained with TTC ( Figure 4 ). The observed color change is attributed to the reduction of a colorless, soluble tetrazolium salt to form a visually detectable, reddish insoluble material known as formazan. This reduction is facilitated by dehydrogenase enzymes present specifically in metabolically active cells. However, there were no furrows on the pollen, so pollen tubes did not form (Caporali et al., 2003). A similar finding was reported in other species like Solanum appendiculatum (Zavada and Anderson, 1997) and Arabidopsis (Dobritsa and Coerper, 2012). Additionally, other morphological features, including shape and wall ornamentation, play a crucial role in pollen identification. However, dimension-based pollen parameters like polar axis, equatorial axis, mesocolpium width, and colpus length and width showed variation in different grape genotypes but did not follow any specific pattern with respect to the species they belong to (Lukšić et al., 2022). All the grape genotypes except EPS and BS exhibited foveolate-perforate exine ornamentation (Marasali et al., 2005; Abreu et al., 2006; Lukšić et al., 2022). EPS and BS both are from V. vinifera L., and they showed reticulate (Padureanu and Patras, 2018) and reticulate striate exine ornamentation, respectively.

Fresh pollen of each grape genotype exhibited a distinct pollen germination percentage. For instance, FS recorded a germination rate of 90.4 %, while POC showed a lower rate of 40.4 %. This suggests that the pollen from each genotype possesses inherent characteristics influencing its ability to germinate under specific conditions as observed by Kelen and Demirtas (2003) in grapes. Similarly, significant variation in fresh pollen germination has been reported within the same genus or species, as reported in the case of the Serbian Autochthon Apple (Ćalić et al., 2021) and crape myrtle (Lagerstromia spp.) (Akond et al., 2012). Probably, this could be a reason behind the fact that some grape genotypes are good pollinators as compared to others (Pereira et al., 2018; Wang et al., 2022), similar findings have been reported in Annona (Pereira et al., 2014) and plum (Đorđević et al., 2022).

The effect of different storage temperatures and durations on in vitro pollen germination revealed that the initial pollen germination was the highest on the day of collection for all genotypes and reduced as storage duration was prolonged. Storage of pollen at lower temperatures (4 °C, −20° C, and −80 °C) has been reported to increase its longevity in various plant species, as reported in Pistachio (Aldahadha et al., 2020), hazel (Novara et al., 2017), mango (Dutta et al., 2013), litchi (Wang et al., 2015), and kiwi (Borghezan et al., 2011). The reduction in pollen germination percentage was significant at room temperature but minimal at −196 °C. As pollen ages, the following occurs: a deterioration in its intracellular structure, a decline in enzyme activity, an increase in the presence of free radicals, and processes such as de-esterification and lipid peroxidation. These factors collectively result in increased cellular component leakage when the pollen is rehydrated (Taylor and Hepler, 1997). When pollen is stored at low temperatures, the metabolic processes within the pollen grains may be diminished or suppressed, leading to an extension of the pollen grain lifespan (Ćalić et al., 2021; Đorđević et al., 2022). This phenomenon explains the minimal alterations in pollen germination and viability observed throughout the entire storage duration at −196 °C. In the study, the logic behind the selection of these four temperatures was to check the possibility of successful pollen storage under available storage facilities to support the grape hybridization program. If pollen needs to be stored for next-day pollination use, it can be stored at room temperature. For a 30-day storage period, pollen can be effectively stored at 4 °C for each genotype, excluding POC. Laboratories commonly have refrigerators, making this a practical option. However, for genotypes involved in grape hybridization and with access to storage facilities at −20 °C and −196 °C (liquid nitrogen), it is advisable to utilize these options for optimal pollen storage within the 30-day timeframe of the flowering period ( Supplementary Figure 1 ). This ensures successful pollen preservation for use in the grape hybridization program.

Our findings indicated that VP and DR had acolporated pollen grains lacking apertures ( Table 1 ), preventing germination on in vitro pollen germination medi (Anderson and Symon, 1989; Penny and Steven, 2009). Similarly, self-pollination experiments with VP and VP/DG confirmed the absence of pollen germination on the stigma ( Figures 8 , 9 ). This proves that VP and DR are male sterile or functional female and can be utilized as female parents without the need for emasculation, thus saving time and effort (Kaul, 1988).

Four cross-combinations belonged to V. parviflora Roxb. × V. vinifera L. (VP/MH and VP/PN), V. parviflora Roxb. × V. champini Planc. (VP/SC and VP/DG), and V. parviflora Roxb. × V. parviflora Roxb. VP/VP was studied to find out cross-compatibility ( Figures 5 – 9 ). It was evident from the pollen tube growth inside the pistil in VP/SC and VP/PN that it had a high degree of compatibility in pollen–pistil interaction. Aniline blue fluorescence microscopy revealed a greater number of germinating pollen on the stigmatic surface, with pollen tubes reaching the base of the stylar canal and even entering the ovules through the micropyle within 24 hours of pollination in VP/SC and VP/PN combinations. This strong compatibility was reflected in higher fruit set percentages, an increased average number of seeds per berry, and compatibility indices in VP/SC and VP/PN cross-combinations. More pollen tube growth (García-Breijo et al., 2020) and more seeds per berry indicate more compatibility among parents (Békefi and Halász, 2005; Cerovic et al., 2021).

While in VP/MH, pollen tubes did not reach the base of the stylar canal; most of the pollen tubes reached more than half of the stylar canal after 24 hours of pollination. Despite that, slower pollen tube growth for successful fertilization occurred, resulting in the production of seeds. Thus, the pollen tubes must have continued to grow for more than 24 hours after pollination and then fertilized the ovules. As a result, lower berry retention (23.25%) and number of seeds per berry (1.986) were recorded. The slower pollen tube growth could have led to reduced fertilization and subsequent fruit set as reported by Adachi et al. (2009) in autotetraploid cultivars of apple. Similarly, Pasonen et al. (2001) reported a significant positive correlation between pollen-tube growth rate and seed mass in Betula pendula. Conversely in VP/MH, although the in vitro pollen germination was higher (71.7 %) (at par with SC), the fruit set (23.25 %) was significantly less as compared to VP/SC along with the lowest compatibility index (0.511) among all successful cross-combinations (Badiger et al., 2023).