Study of polyembryony and development of molecular markers for identification of zygotic and nucellar seedlings in Khasi mandarin (Citrus reticulata Blanco)

The objective of this work was to evaluate the occurrence of polyembryonic seedlings and other morphological parameters in Khasi mandarin during three harvest years and to identify zygotic (sexual) seedlings from nucellar (asexual) ones grown under in-vitro conditions using molecular markers. Embryos from 27 polyembryonic and 7 monoembryonic seeds of Khasi mandarin were grown in-vitro. DNA from seedlings and mother parent was analyzed using 16 ISSR and 5 RAPD primers, of which 4 ISSR and a set of 3 RAPD primers were effective to identify zygotic or nucellar origin of the seedlings. In-vitro culture enables maximum embryos of each seed to grow, favouring the origin of seedlings to be identified as zygotic. Among 69 tested individuals, 37 zygotic and 32 nucellar seedlings were recognized. In polyembryonic and monoembryonic seeds, 59.6% and 42.8% of the seedlings, respectively, have the sexual origin. Morphological characteristics of seeds and the seedlings generated varied significantly and were not correlated with polyembryony except for the clutch size and the number of branches. Polyembryonic seeds in the cultivar are high, ranging from 50.0%, 55.5% to 83.3% over three harvest years with more clutch size and the possibility of obtaining zygotic plants from them is high. In polyembryonic seeds not all zygotic seedlings were produced by small embryos located at the micropyle. Identification of zygotic seedlings by ISSR and RAPD markers in Khasi mandarin cultivar is efficient and reliable at an early developmental stage.


INTRODUCTION
Khasi mandarin (Citrus reticulata Blanco) is a commercially popular fruit crop of the Eastern Himalayas exhibiting the common reproductive trait polyembryony (Nakano et al., 2012). Polyembryony has serious significance in citrus breeding since nurserymen use the genetically uniform healthy offspring to revive the old clones that have lost their vigour through regular vegetative propagation (Mondal and Saha, 2014). When citrus is propagated from polyembryonic seeds, the farmers allow them to grow and produce several seedlings (including hybrid and genetically uniform ones) for a long juvenile period of 6-7 years and they normally rogue off the hybrid ones, choosing those with desirable traits and presume that this selection guarantee the origin of the seedling (Mondal et al., 2014). However, this technique is not reliable due to some unusual developments which take place during embryo maturity resulting in the formation of good proportion of twin and triplets in the population (Mondal et al., 2015). In an open-pollinated population, the morphological identification of zygotic embryos becomes even more difficult. Moreover, the orchards are suffering from decline syndrome due to genetic erosion of proper planting material. To assure the variety and good vigour, it is indispensable to identify zygotic seedlings from nucellar ones as zygotic ones are vigorous and can compete with nucellar ones (Andrade-Rodriguez et al., 2005) for establishment of new Citrus orchards. Thus, it is imperative to develop efficient methods for differentiating zygotic embryos from nucellar ones at an early stage, for breeding purpose in polyembryonic cultivars. Several PCR based systems are available for genomic DNA analysis, amongst which ISSR, as well as RAPD markers are simple, reproducible, and userfriendly for farmer populations (Rao et al., 2008). These markers enable early selection of the progeny and this quality is particularly useful when a mixed hybrid population has to be rapidly analyzed and differentiated (Goodwin et al., 1997). Though Golein et al., 2011), but no investigation have been reported using molecular approaches for identification of origin of seedlings in Khasi mandarin cultivar mainly from an open-population system and this finding may considerably contribute to orchard management programs for this cultivar in the Indian subcontinent. The objective of this work was to evaluate the occurrence of polyembryonic seedlings as well as other morphological parameters collected during three harvest years in Khasi mandarin and the hypothesis tested was to see if these harvest years influence on the morphological parameters in the cultivar. The study also aimed to identify zygotic (sexual) seedlings from nucellar (asexual) ones in Khasi mandarin grown under in-vitro conditions.

2.1
Plant material The seeds were collected from an open-pollinated Khasi mandarin (Citrus reticulata) trees (around fifteen years old) growing in a citrus farm located in Boko (25.97 ○ N and 91.23 ○ E), Assam, India. Matured and healthy fruits of Khasi mandarin were randomly harvested from six clonal mother trees located in close proximity to each other were collected during three harvest years of November in 2013, 2014 and 2015. Five fruits per plant were collected, amounting to a total of 30 fruits at each time of harvest. Five seeds per fruit were randomly selected to complete a set of 150 seeds and three replications of 150 seeds were used for each harvest year.

Study of polyembryony and morphological characteristics
To study polyembryony and morphological characteristics, the seeds were surface sterilized as described by Yun et al. (2007). The endocarp and seed coat (testa with tegmen covering all embryos) from each seed were removed and the size (length, mm) of the embryo (measured from the tip of the radicle to the opposite tip of the largest cotyledon) were recorded. The percentage of polyembryonic seeds, the number of embryos per polyembryonic seed and the average number of embryos or clutch size (Kishore et al., 2012) in polyembryonic seeds were recorded. The calculation of the average number of polymebryonic seeds were done by dividing the number of polyembryonic seeds by the total number of seeds and as a whole was calculated by dividing the total number of embryos by total number of seeds (both polyembryonic and monoembryonic) at each harvest. For polyembryony studies, embryos from 27 polyembryonic and 7 monoembryonic seeds was grown in-vitro following Andrade-Rodriguez et al. (2005). Invitro germination data were observed and recorded for three harvest years. The emergence of different seedlings was noted and the number of leaves/seedling, number of branches/seedling, and shoot length/seedling were recorded. Representative embryos from the open pollinated seed population were evaluated to find suitable markers for identifying the zygotic seedlings.

2.3
Plant sample for polymorphism study Fully expanded leaves from the six mother plants (from which the fruits were collected initially), were collected in plastic bags and kept at -80 °C until DNA extraction. The leaves from in-vitro germinated seedlings were also collected and placed in 2 ml eppendorf tubes, labelled and placed on ice until DNA extraction. Six mother plants, 62 seedlings from polyembryonic seeds and 7 seedlings from monoembryonic seeds were used for identification of zygotic seedlings using ISSR and RAPD markers.

2.4
DNA extraction Total genomic DNA was extracted according to the method described by Doyle and Doyle (1990). DNA concentration and purity were quantified using a Nanodrop 1000 spectrophotometer (Invitrogen, Waltham, Massachusetts, USA) at 260/280 nm absorbance.

2.5
PCR amplification PCR amplification was carried out in a Veriti thermal cycler (Applied Biosystems, USA) in a final volume of 25 µl reactions containing 25 ng of template DNA, 0.1 mM total dNTPs, 0.3 µM primer, 2.5 µl of 1X PCR buffer with 15 mM of magnesium chloride and 0.5 unit of Taq DNA polymerase (Bangalore genei, Bangalore, India). A total of 50 ISSR primers developed by University of British Columbia, Canada and 10 RAPD primers were custom synthesized from Eurofins Genomics (Bangalore, India). Initially two representative samples each from mothers and progenies were analyzed by PCR amplification. About 10 µl of PCR-amplified product (with 2 µl of 6X loading buffer) was analyzed on a 2% agarose gel in 1X TAE buffer stained with 10 mg/ml of ethidium bromide for visualization of bands for 3 h at 80 V and examined under UV transilluminator using the UVitech gel documentation system (Bangalore genei, Bangalore, India). Molecular weights of the PCR products were estimated by comparing them with 100 bp (100 µg/ml) and ØX174 DNA /HaeIII digest (500 µg/ml) DNA ladders (Bangalore genei, Bangalore, India).

2.6
Data analysis Mean and standard error values were calculated to reveal statistical significance for the morphological parameters evaluated per harvest year. The data (means of replications) was carried out in triplicates and statistical comparison of data was performed by ANOVA. The Pearson correlations analysis was done using R software 3.2V. A probability value of p ≤ 0.05 was adopted as the criteria for significant differences. Duncan's multiple range (DMR) test was used for comparison between the means of variables which was generated by the SAS program (version SAS 9.3, SAS Institute Inc., Cary, NC, USA). Scoring for polymorphism of the markers was carried out using a band-based method where bands were unambiguously scored as 1 and 0 for present and absent alleles respectively (Bonin et al., 2007). Both polymorphic and monomorphic loci were analyzed. Bands that resolved poorly on the gel were treated as missing ones. Polyembryony varies depending on the ecological region and cultivar suggesting that it should be specific for each variety with respect to the region. The study indicated that polyembryony does not seem to be a limiting factor for germination capacity. Scalon et al. (2003) also reported that polyembryony was not a limiting factor for seed germination and plant emergence in B. Glabra. The number of embryos in polyembryonic seed ranged from 2 to 14 (Table 1), characterized by variation in the partitioning of embryos followed by both synchronous and asynchronous development of seedlings ( Fig. 1a and  1b). The average number of embryos (clutch size) per polyembryonic seed ranged from 2.27 ± 0.03 to 2.70 ± 0.10 and per total seed ranged from 1.13 ± 0.19 to 2.30 ± 0.12. One of the causes in the differences of clutch size during harvest years might be due to the presence of dominant trait for polyembryony, movement of auxins and ploidy level which plays prominent roles in the development of extra numerous embryo (Jaskani et al., 2005). Improper endosperm growth due to lack of nutrients and growth factors from the maternal tissue to the embryo might also lead to the early breakdown of nucellar embryos (Kishore et al., 2012). Our results corroborate the findings of Kishore et al. (2012), where the number of embryos per seed reported in citrus (including mandarin, sweet orange, rough lemon, and lime) ranged from 2 to 14 while, it was different in terms of number of embryos per polyembryonic and total seed as reported by them (3.48 and 3.26). This significant disparity in the occurrence of clutch size evidently points out the impact of location on occurrence of numerous embryos. This suggests that the difference might be due to stress-induced changes in the environment leading to change in the genetic development of cells or their hormonal behavior, thereby affecting embryo development and surrounding seed organization (Batygina and Vinogradova, 2007). The number of leaves and shoot length were maximum in November 2015 and were minimum in November 2013, respectively, while the number of branches was the highest in November 2013 and the lowest in 2015 ( Table 1). The variation in the emergence of the morphological traits suggests that polyembryony was only correlated with the number of branches indicating that it can be treated as a reliable indicator of the occurrence of polyembryony in the seeds. The variation could be due to the fact that the embryos of each seed were grown in-vitro using conical flasks during the harvest years for each seed which enabled development of a seedling from every embryo as compared to other studies which utilized seedlings grown in greenhouse, pots, etc (Golein et al., 2011;Yun et al., 2011). Moreover, during seed formation, embryos adapt to different dynamics especially when adventitious polyembryony is considered (Batygina and Vinogradova, 2007). One way Anova analysis was carried out at 0.05% significance where the probability Pr value 0.97 is greater than 0.05 (Table 2) which resulted in the rejection of null hypothesis suggesting that there is no significant variation in time of harvest with respect to the morphological trait. This was justified by Pearson correlation test which resulted in positive correlation among total morphological traits against time of harvest. The most significant correlation was observed between November 2013 and November 2014 followed by November 2015 and November 2014 as shown in Table 3 and Fig. 2. This result suggested that there is gradual increase in morphological traits over time of harvest in three consecutive years.

ISSR and RAPD markers for identification of zygotic and nucellar seedlings
ISSR markers generated the highest number of amplified fragments with an average of 7.25 bands obtained per markers (Table 4). From the total markers, only 16 ISSR and 5 RAPD primers were selected based on the polymorphic and reproducible banding patterns. Of these, finally 4 ISSRs primers namely UBC 810, UBC 835,  (Fig. 3a). The three RAPD primers generated differentiating fragments of varying sizes; OPAA 10 -650 bp, OPA 04 -710 bp, and OPA 18 -600 bp (Fig. 3b). In addition, a distinguishable ISSR amplification pattern, related to the mother and progeny, was obtained using the UBC855 and OPAA 10 primers. The amplified UBC 855-590 bp fragments was not from zygotic progeny but from the mother plant (Fig. 4a) and the amplified OPAA 10-650 bp fragment was present in the zygotic progeny but not in the mother plant as indicated by arrows (Fig. 4b) Of 62 polyembryonic seedlings studied, 37 seedlings showed a banding pattern different from that of the mother plant using both ISSR and RAPD primers (Table  5) indicating a zygotic origin; primer UBC 855 helped to identify 18 of these zygotic seedlings (Fig. 4a) and OPAA 10 helped to identify 26 of these zygotic seedlings (Fig.  4b). A finding reported by Vilarinhos et al. (2000) identified 12 zygotic seedlings from Volkamerian lemon × Cravo lemon using six of 20 RAPD primers in a combination. Of the 7 seedlings from monoembryonic seeds, primers UBC 835 and UBC 810 identified 14.2% (1/7) and primers OPA 04, OPAA 10 and OPA 18 identified 42.8% (3/7), 28.5% (2/7) and 14.2% (1/7), respectively to be zygotic in nature (Table 5). Additionally, among both polyembryonic and monoembryonic seedlings, 53.6% of them were zygotic in nature and 46.3% nucellar, which are in close proximity to the value of 53% and 47% of zygotic and nucellar seedlings as reported by Mondal et al. (2014) in Citrus reticulata using RAPD. In the present study, seedlings genetically identical to the mother plant were generated at a low incidence in polyembryonic seeds produced by the highly polyembryonic population. As reported, the percentage of nucellar seedlings declines with an increase in the proportion of polyembryonic seed generated (Soost and Roose, 1996; Andrade-Rodriguez et al., 2004). Thus, it is likely that the lower proportion of nucellar population is related to a higher range of polyembryony. In the present work, 66.6% of the Khasi mandarin seeds had 2 to 3 zygotic embryos (Table 5). This coincides with Das et al. (2007) that reports when more than one zygotic embryo per seed for citrus was observed it suggests the chances of different microgametes induced fertilization.
Our result indicated that zygotic embryos were located in the micropylar region in 40% (2 of 5) of the polyembryonic seeds; while in the remaining 60% seeds, they were located near the micropyle region but not in the micropyle (Table 5). This observations suggests that embryo distribution inside seed does not follow any specific pattern and might be genetically controlled since zygotic embryos cannot be visually distinguished based on their morphological traits such as shape, size etc (Yun et al., 2007). Similar results were also observed where it was reported that the maximum of the zygotic seedlings were obtained from embryos located close to the micropylar region (Andrade-Rodriguez et al., 2005). The reason for positioning of zygotic embryos at micropylar area could be related to their growth habit which depends on the endosperm since its presence stimulates the formation of adventitious embryos in the micropylar region but suppresses their initiation towards the chalazal part based on the distance of embryo from the micropylar end (Kishore et al., 2012).
The size of embryos in polyembryonic seeds ranged from 2-6 mm, while in monoembryonic seeds, it varied from 6-8 mm ( Table 5). The study indicated that embryos of even small size (2-3 mm) were able to develop seedlings suggesting the availability of food reserve as well as hydrated condition of the mature embryos during in-vitro seed germination. The embryo size is an important attribute in polyembryony since smaller ones are generally found to be inviable due to insufficient food reserve (Andrade- Rodriguez et al., 2005). In citrus, as the number of embryos per seed increases, its size decreases (Soares Filho et al., 2003). In orange, it was reported that the embryo size decreases as it approaches near the micropylar region of the seed (Villegas and Andrade, 2008).

IV.
CONCLUSION A simple outlook that emerges from the study is that the popular mandarin cultivar, Khasi mandarin is highly polyembryonic, exhibiting polyembryony in more than 80% of their seeds with more clutch size and the possibility of obtaining zygotic plants from them is high. A population of 69 individual seedlings were generated, out of which 62 individuals were tested as polyembryonic plantlets. Out of tested plantlets, 37 hybrids and 25 nucellars were recognized through molecular screening by ISSR and RAPD markers. ISSR-UBC 855 and RAPD-OPAA 10 can be used as efficient markers for identifying hybrid seedlings at an early developmental stage confirming their potential for early selection in citrus breeding. The ISSR and RAPD markers used in the present study showed 15-45% efficiency in the identification of the origin of seedlings. Observation of embryo size and position allowed us to find that zygotic seedlings were obtained from embryos located close to the micropylar region and not at the micropylar region. Since polyembryony is species and region-specific, this attempt may aid in the development of uniform population for the farmer community of the region. This work could be elaborated with a large population size or a hybrid system for overall citrus germplasm management in India.

DISCLOSURE STATEMENT
The authors declare no conflict of interest.