Development of In Vitro Anther Culture for Doubled Haploid Plant Production in Indica Rice (Oryza sativa L.) Genotypes

Anther culture is an efficient biotechnological tool in modern plant breeding programs to produce new varieties and parental lines in hybrid seed productions. However, some bottlenecks—low induction rate, genotype dependency, albinism—restrict the widespread utilization of in vitro anther culture in rice breeding, especially in Oryza sativa ssp. indica (indica) genotypes, while an improved efficient protocol can shorten the process of breeding. Three different induction media (N6NDK, N6NDZ, Ali-1) and four plant regeneration media (mMSNBK1, MSNBK3, MSNBKZ1, MSNBKZ2) were tested with five indica rice genotypes to increase the efficiency of in vitro androgenesis (number of calli and regenerated green plantlets). The production of calli was more efficient on the N6NDK medium with an average 88.26 calli/100 anthers and N6NDZ medium with an average of 103.88 calli/100 anthers as compared to Ali-1 with an average of 6.96 calli/100 anthers. The production of green plantlets was greater when calli was produced on N6NDK medium (2.15 green plantlets/100 anthers) compared when produced on to N6NDZ medium (1.18 green plantlets/100 anthers). Highest green plantlets production (4.7 green plantlets/100 anthers) was achieved when mMSNBK1 plant regeneration medium was used on calli produced utilizing N6NDK induction medium. In the best overall treatment (N6NDK induction medium and mMSNBK1 plant regeneration medium), four tested genotypes produced green plantlets. However, the genotype influenced the efficiency, and the green plantlets production ranged from 0.4 green plantlets/100 anthers to 8.4 green plantlets/100 anthers. The ploidy level of 106 acclimatized indica rice plantlets were characterized with flow cytometric analyses to calculate the percentage of spontaneous chromosome doubling. Altogether, 48 haploid-, 55 diploid-, 2 tetraploid- and 1 mixoploid plantlets were identified among the regenerant plantlets, and the spontaneous chromosome doubling percentage was 51.89%. Utilization of DH plants have been integrated as a routine method in the Hungarian rice breeding program. The tetraploid lines can be explored for their potential to offer new scopes for rice research and breeding directions in the future.


Introduction
Doubled haploid (DH) plant production methods have improved and led to accelerating the breeding of new varieties and hybrids. These methods are widely used in many crop plants such as barley, rapeseed, maize etc. Anther culture (AC) techniques can produce homozygous DH lines within one generation. Thus, the long process of conventional breeding methods can be reduced by homozygosity in early generations. The recessive alleles could be obtained and selected earlier due to the homozygosity of DH lines. The

Effect of Genotype, Induction Media and Their Interactions on Androgenesis Induction in AC
Three different induction media were compared in in vitro AC of five indica rice genotypes, which were cultured at the uninucleate stage of microspores (Figure 1a). The microspore-derived calli were observed after 4 weeks of AC (Figure 1b), and later regenerated green and albino plantlets were seen on the regeneration medium (Figure 1c).

Effect of Genotype, Induction Media and Their Interactions on Androgenesis Induction in AC
Three different induction media were compared in in vitro AC of five indica rice genotypes, which were cultured at the uninucleate stage of microspores (Figure 1a). The microspore-derived calli were observed after 4 weeks of AC (Figure 1b), and later regenerated green and albino plantlets were seen on the regeneration medium (Figure 1c). In vitro androgenesis induction was successful in AC of each genotype. Based on the statistical analysis (Table 1), the effect of genotype, the induction media and their interaction were shown to influence the number of calli significantly (p < 0.001).  Table 2 shows the calli/100 anthers and genotypic differences. The number of calli/100 anthers ranged from 0 to 214.4 depending on genotype and induction medium ( Table 2). Among the three tested media, the lowest number of calli was seen on the Ali-1 induction medium, where only one ('Co 39') among the five tested genotypes produced calli (34.8 calli/100 anthers). The mean of AC-derived calli/100 anthers was 103.88 and 88.26 on the N6NDZ and N6NDK induction media, respectively. The genotype 'Co 39' was the most responsive based on the number of induced calli, but the calli production was also significant in 'PSB RC 94' and 'Mangala' genotypes. Altogether, the number of AC-derived calli was significantly higher on N6NDZ and N6NDK media than on the Ali-1 medium. In vitro androgenesis induction was successful in AC of each genotype. Based on the statistical analysis (Table 1), the effect of genotype, the induction media and their interaction were shown to influence the number of calli significantly (p < 0.001).  Table 2 shows the calli/100 anthers and genotypic differences. The number of calli/ 100 anthers ranged from 0 to 214.4 depending on genotype and induction medium ( Table 2). Among the three tested media, the lowest number of calli was seen on the Ali-1 induction medium, where only one ('Co 39') among the five tested genotypes produced calli (34.8 calli/100 anthers). The mean of AC-derived calli/100 anthers was 103.88 and 88.26 on the N 6 NDZ and N 6 NDK induction media, respectively. The genotype 'Co 39' was the most responsive based on the number of induced calli, but the calli production was also significant in 'PSB RC 94' and 'Mangala' genotypes. Altogether, the number of AC-derived calli was significantly higher on N 6 NDZ and N 6 NDK media than on the Ali-1 medium.

Plant Regeneration from AC-Derived Calli of Indica Genotypes
The AC-derived calli produced on the N 6 NDZ and N 6 NDK induction media were used in the plant regeneration experiments. The comparisons of the effects of four combinations of growth regulators were based on the number of regenerated plantlets are shown in Table 3. The Two-way ANOVA revealed that the genotype significantly (p < 0.001) influenced the number of regenerated plantlets (total, green and albinos) except the number of green plantlets regenerated from calli induced on N 6 NDK medium ( Table 3). The effect of treatment (growth regulator combinations) was also significant except for the number of green plantlets regenerated from the N 6 NDZ-derived calli. The genotype × media interaction was significant in the case of the number of green plantlets regenerated from the N 6 NDZ-derived calli. Data of plant regeneration experiments using the AC-derived calli induced on N 6 NDK medium has been shown in Tables 4 and 5. The means of plant regeneration efficiency (total, green and albino) was highest using the mMSNBK1 medium. The highest plant regeneration efficiency (18.4 plantlets/100 anthers) was observed in AC of the 'Co 39' genotype. However, this genotype produced higher levels of albinos (18 albinos/100 anthers). The highest value of green plantlets production (8.4 green plantlets/100 anthers) was achieved in AC of 'PSB RC 94'. Each of the four tested genotypes produced green plantlets in the best treatment, the green plantlets production ranged from 0.4 to 8.4 green plantlets/100 anthers depending on genotype.  Table 5 demonstrates the regeneration frequency of AC-derived callus induced on N 6 NDK medium. Altogether 4668 calli of the four tested genotypes were applied to compare the effect of four plant regeneration media. The highest plant regeneration frequency was achieved using mMSNBK1 medium and the percentage of green and albino plantlet regeneration were 4.03% and 6.34%, respectively. The percentage of green and albino plantlet regeneration was lowest on the MSNBKZ1 plant regeneration medium. 'Co 39' genotypes produced only two in vitro green plantlets, while the other genotypes ('PSB RC 94', 'IRRI 147' and 'Mangala') regenerated a few dozen green plantlets. Tables 6 and 7 show the data of the plant regeneration experiments using the ACderived calli induced on the N 6 NDZ medium. The means of total, albino and green plant regeneration efficiency were 9.08, 7.9 and 1.18 plantlets/100 anthers, respectively. The regenerated plantlets were mostly albinos in this experiment. In the best treatment, the highest mean of green plantlet regeneration (2.72 green plantlets/100 anthers) was achieved using the MSNBK3 medium containing 0.5 mg/L, 0.5 mg/L and 1.5 mg/L NAA, BAP and kinetin, respectively. However, the means of green plantlets production were significantly lower than in the first experiment. In this treatment, the green plantlets production of 'Mangala' was outstanding as an exception. Green plantlets production was successful from the AC-derived calli of only two tested genotypes.     Table 7 shows the efficiency of green and albino plant regeneration of AC-derived calli induced on the N 6 NDZ induction medium. Altogether 5276 calli of the five tested genotypes were placed on the four plant regeneration media to compare their effect on the number of green and albino plantlets. The green plantlet regeneration efficiency was highest using the MSNBK3 plant regeneration medium and the percentage of green plant regeneration was 2.58%. However, this value was lower than the green plantlets production of mMSNBK induction medium (4.03%) and calli produced on N 6 NDK induction medium ( Table 5). The percentage of green plant regeneration was lowest on the MSNBKZ2 plant regeneration medium (0.3%).
'Mangala' genotypes produced 53 in vitro green plantlets in the experiment, while 'PSB RC 94' and 'IRRI 64' regenerated only a few green plantlets. The 'Co 39' and 'IRRI 147' genotypes regenerated only albino plantlets from the calli induced on the N 6 NDZ induction medium.
The regenerated green plantlets were grown in individual tubes and later jars till the transplantation of well-developed green plantlets into the greenhouse. The AC-derived green plants acclimatized to the greenhouse conditions well.

Spontaneous Chromosome Doubling of AC-Derived Green Plantlets as Demonstrated by the Ploidy Levels Using Flow Cytometric Analysis
The ploidy levels of the AC-derived plantlets of five genotypes were analyzed by flow cytometric analyses. Figure 2 shows the relative DNA content of the five genotypes used in this study. The regenerated green plantlets were grown in individual tubes and later jars till the transplantation of well-developed green plantlets into the greenhouse. The AC-derived green plants acclimatized to the greenhouse conditions well.

Spontaneous Chromosome Doubling of AC-Derived Green Plantlets as Demonstrated by the Ploidy Levels Using Flow Cytometric Analysis
The ploidy levels of the AC-derived plantlets of five genotypes were analyzed by flow cytometric analyses. Figure 2 shows the relative DNA content of the five genotypes used in this study.
Altogether, 48 haploids (Figure 2b Table 8 demonstrates the ploidy level of the acclimatized plantlets by genotype. The percentage of spontaneous chromosome doubling was 51.89% among the acclimatized plantlets. Altogether, the seeds of 55 fertile DH lines were harvested, and two partially fertile tetraploid lines produced a few seeds. The produced DH lines of 'Mangala', 'IRRI 147' and 'PSB RC 94' genotypes were 27, 15 and 13 DHs, respectively.   Table 8 demonstrates the ploidy level of the acclimatized plantlets by genotype. The percentage of spontaneous chromosome doubling was 51.89% among the acclimatized plantlets. Altogether, the seeds of 55 fertile DH lines were harvested, and two partially fertile tetraploid lines produced a few seeds. The produced DH lines of 'Mangala', 'IRRI 147' and 'PSB RC 94' genotypes were 27, 15 and 13 DHs, respectively. The fertility of the DH 0 plantlets tested by flow cytometric analyses was checked based on seed production ( Figure 3). The haploid plants produced panicles without any seed (sterile), the diploid plantlets were fertile, while the panicles of tetraploid plants contained a few seeds (partially fertile).

Induction of Androgenesis in Indica Rice
The N6 basal medium is the most generally used medium for the induction of rice androgenesis [13,[30][31][32], while Ali et al. reported the superiority of the Ali-1 induction medium compared to N6 medium in AC of rice [28]. The present study compared the N6NDK induction medium to Ali-1 and N6 basal medium supplemented with combinations of different growth regulators (N6NDZ). In contrast to the report by Ali et al. [28], the induction media prepared with N6 basal medium (N6NDK and N6NDZ) outperformed the Ali-1 induction medium with regard to callus production. The number of produced calli was limited on Ali-1 induction medium [28]. N6NDK medium proved an efficient induction medium regarding callus production in AC of in indica rice genotypes.

Effect of Induction Medium on Plant Regeneration
The regeneration of green plantlets is the next critical point of DH plant production in indica rice genotypes because of the low efficiency of plant regeneration and the high number of albinos [6,8,14,15,19,21,22,25,30]. The induction medium affects the regeneration ability of calli [23]. There were differences among the plants regenerated utilizing the generated on the best two induction media (N6NDK and N6NDZ) with differences in the

Induction of Androgenesis in Indica Rice
The N 6 basal medium is the most generally used medium for the induction of rice androgenesis [13,[30][31][32], while Ali et al. reported the superiority of the Ali-1 induction medium compared to N 6 medium in AC of rice [28]. The present study compared the N 6 NDK induction medium to Ali-1 and N 6 basal medium supplemented with combinations of different growth regulators (N 6 NDZ). In contrast to the report by Ali et al. [28], the induction media prepared with N 6 basal medium (N 6 NDK and N 6 NDZ) outperformed the Ali-1 induction medium with regard to callus production. The number of produced calli was limited on Ali-1 induction medium [28]. N 6 NDK medium proved an efficient induction medium regarding callus production in AC of in indica rice genotypes.

Effect of Induction Medium on Plant Regeneration
The regeneration of green plantlets is the next critical point of DH plant production in indica rice genotypes because of the low efficiency of plant regeneration and the high number of albinos [6,8,14,15,19,21,22,25,30]. The induction medium affects the regeneration ability of calli [23]. There were differences among the plants regenerated utilizing the generated on the best two induction media (N 6 NDK and N 6 NDZ) with differences in the combination of growth regulators. The number of total regenerated plantlets was 6.406 plantlets/100 anthers and 9.08 plantlets/100 anthers from calli produced on N 6 NDK and N 6 NDZ induction media, respectively.
The number of albinos from calli induced on N 6 NDZ was 7.9 albinos/100 anthers while the number of green plantlets was 1.18 green plantlets/100 anthers on average. The calli produced on N 6 NDK induction medium regenerated 4.256 albino/100 anthers and 2.15 green plantlets/100 anthers on average. The number of regenerated green plantlets was almost two times higher using the N 6 NDK induction medium, indicating that this induction medium is the best medium to be utilized to ensure sufficient green plant generation. Zeatin in combination with 2,4-D and NAA was applied for androgenesis induction in rice [28]. Although the number of produced calli was higher using this combination of growth regulators, adverse effects of this induction media (N 6 NDZ) were observed on plant regeneration. Interestingly, the N 6 NDK medium was efficient in the induction of in vitro androgenesis of japonica genotypes as well [37].
The percentage of green plantlet regeneration of callus was also higher from callus induced on N 6 NDK (on average 1.84%) than N 6 NDZ (on average 1.1%), these values ranged from 0% to 47.06%, which were influenced by genotype and plant regeneration medium. According to some relevant publications, green plantlet production ranged from 0% to 58.25% depending on genotype and culture conditions in indica rice AC [13,[29][30][31][32].

Effect of Exogenous Growth Regulators on the Regeneration of Green Plantlets
The compositions of plant regeneration medium influence the efficiency of green plantlets production in AC of rice. The MS basal media have been frequently applied for plant regeneration in rice AC [8,11,12,38]. Sucrose is a general carbon source during plant regeneration, while a wide range of growth regulators (BAP, IAA, NAA, kinetin and zeatin), and their combinations were applied for plant regeneration of AC-derived rice calli [11,12,21,23,24,26,28,38].
In the plant regeneration experiments, the genotypes, combination of growth regulators and quality of calli (see above) significantly influenced the plant regeneration efficiency. Zeatin was applied as an exogenous growth regulator in combination with other hormones in rice AC experiments [26,28,35], while added zeatin had an adverse effect on both callus induction and plant regeneration in AC of indica rice genotypes. The different combinations of NAA, BAP and Kinetin were applied for plant regeneration in rice AC [11,12,30]. Rout et al. reported that the combination of 0.5 NAA, 1.5 BAP and 0.5 Kinetin was efficient for green plantlets production in AC of indica rice hybrid [11]. However, the combination of 1 mg/L NAA, 1 mg/L BAP and 1 mg/L Kinetin was more efficient for green plant production than the combination of 0.5 mg/L NAA, 1.5 mg/L BAP and 0.5 mg/L Kinetin in this study. The green plantlet production was highest (4.7 green plantlets/100 anthers, average of genotypes) using calli derived from the N 6 NDK induction medium and mMSNBK1 plant regeneration medium containing 1 mg/L NAA:1 mg/L BAP:1 mg/L Kinetin; the number of regenerated green plantlets ranged from 0.4 to 8.4 green plantlets/100 anthers depending on genotype. In the treatment, the percentage of green plantlets regeneration of calli was on average 4.03%, ranging from 0.21% to 47.06% depending on genotype. The green plantlet productions are influenced by genotype (ranging from 0% to 58.25%) in indica rice AC [13,[29][30][31][32], and higher efficiency of AC was reported in indica rice hybrids [29,30,32].
This indicated that this medium and growth regulator combination was probably the best combination suited for DH green plant production for rice plant breeding as per this study. It should be noted that these media and growth regulator combinations were also effective for green plantlets production in AC of japonica genotypes [37].

Identification of the Ploidy Level of AC-Derived Plantlets by Flow Cytometric Analyses
Flow cytometric analyses, the most accurate method to measure the ploidy levels of the microspore-derived plantlets, were used on 106 AC-derived acclimatized plantlets generated in this study. Overall, 48 haploids, 55 diploids, 2 tetraploids and a single mixoploid plantlet were identified among them. Based on these data, the spontaneous chromosome doubling was 51.89% average across genotypes. A significant difference was observed among the tested genotypes because of the different numbers of regenerated plantlets per genotype and, in turn, genotype dependency of chromosome doubling. Results recorded in this study were similar results (42.24-69.00%) to those recorded in most studies [2,5,6,21,24,37], but Cha-um et al. [27] observed lower values of spontaneous chromosome doubling (21.5-31.9%).
Since the performance of DH lines should be uniform in the following generations, stability of the ploidy of DH is also critical in plant breeding. as in some cases green plants can result from anther walls instead of the microspores. Hence it is critical that the uniformity of the regenerated diploid plants can be checked in the nursery in the DH 1 and DH 2 generations.
Two tetraploid plantlets (1.89%) with partially fertile panicles were identified among acclimatized plantlets in this study. Mishra et al. [32] reported a range of 6.3-16.6% in their study. The irregular development of pollen grains was mentioned among the causes of incomplete fertility [32].
It should be noted that the 48 haploids after the confirmation using flow cytometric analyses, the chromosome number of haploids can be doubled by chemical treatment (colchicine), which is feasible before (in vitro) [14,15] or after acclimatization (in vivo) [2]. This will further improve the efficiency of green DH plant regeneration in rice. Nevertheless, tetraploid rice plants are of interest in polyploid research and breeding programs [39]. The standard rice cultivation method was used to grow donor plants in the nursery. After direct seeding, pre-emergent herbicide spraying (active ingredient: pendimethalin) was applied. After 5 weeks of seeding, permanent flooding was applied with a 5-20 cm irrigation level depending on the development of the plants. Furthermore, 60-60 kg/ha (total amount of 120 kg N/ha) of nitrogen fertilizer was applied two times (38 and 70 days after sowing). During the plant growing season, weeds were controlled manually.

Collection and Pre-Treatment of Donor Tillers
The tillers of donor genotypes were harvested at the booting stage when the panicles were covered with flag leaves and the distance was 2-5 cm between the flag leave and the second leaf. The donor tillers were placed into an Erlenmeyer flask containing tap water and were wrapped with PVC bags. The cold pre-treatment was carried out at 10 • C for three to seven days.

Sterile Technic and Isolation of Anthers
After pre-treatment, the developmental stages of microspores were identified by evaluation with Olympus CK-2 inverted microscope (Olympus Ltd., Southend-on-Sea, UK). The florets of panicles containing early-and mid-uninucleate microspores were surface-sterilized for 20 min in 2% sodium hypochlorite solution with 2 drops of Tween-80. After sterilization, the donor materials were rinsed three times with sterile distilled water. The anthers (100/Petri dish) were isolated from the donor florets and placed in the 60 mm diameter plastic Petri dishes which contained different induction media. At least 3000 anthers were isolated from each genotype to test the efficiency in vitro AC with three induction media and five genotypes.

Condition of Anther Culture for Induction of Calli
The effects of three different induction media were tested on the number of induced calli and regenerated plantlets in indica rice AC (Table 9). These media were Ali-1 induction medium [28], N 6 NDZ and N 6 NDK. For the N 6, media combinations, the N 6 basal medium [40] was supplemented with two different combinations of growth regulators. They were induction media named here as N 6 NDZ (1 mg/L NAA, 1 mg/L 2,4-D and 0.1 mg/L Zeatin), and N 6 NDK (2.5 mg/L NAA, 1 mg/L 2,4-D and 0.5 mg/L kinetin). Both media contained 60 g maltose, 500 mg/L L-Proline and 500 mg/L L-Glutamine, pH was adjusted at 5.8. The medium was solidified with 2.8 g/L Phytagel. Furthermore, the Petri dishes with anthers cultured were incubated at 28 • C in a dark thermostat.

Plant Regeneration
The collections of AC-derived calli were finished on the 8th week of cultivation. The AC-derived calli were transferred to the 90 mm diameter plastic Petri dishes containing plant regeneration media. In the plant regeneration experiment, the AC-derived calli produced on the N 6 NDZ and N 6 NDK media were used to compare the effect of four different combinations of growth regulators on plant regeneration efficiency (green and albino plantlets).
During the plant regeneration period, the effects of different combinations of growth regulators (NAA, BAP, Kinetin and Zeatin) were tested on the regeneration of green and albino plantlets. The MS [41] medium was supplemented with growth regulator combinations, 30 g/L sucrose and pH was adjusted at 5.8 (Table 10). The plant regeneration media were solidified with an increased quantity (4 g/L) of Phytagel compared to earlier published MSNKB1 plant regeneration medium [37]. The plantlets with shoots were transferred into the glass tubes containing 1 /2MS medium supplemented with 30 g/L sucrose, pH was adjusted at 5.8 [37]. The rooting medium was solidified with 2.8 g/L Phytagel. The green plantlets were cultured in glass tubes till the end of rooting. The well-rooted and tillered green plantlets were transplanted into 720 mL glass jars containing the same 1 /2MS medium till transplantation of plantlets to the greenhouse. During the plant regeneration period, 16 h artificial light and constant temperature (24 • C) were kept in a tissue culture growing chamber.

Acclimatization of Plantlets
In the first half of March, the well-rooted green plantlets were transferred to the greenhouse. The green plants were transplanted individually into plastic pots (volume 1000 mL) containing 1:1:1 sand, peat and humus mixture, and the plantlets were covered with PVC bags to retain high humidity. After the acclimatization period (5 days), the PVC bags were removed and the acclimatized green plants were grown till harvest in the greenhouse. Ploidy levels of regenerated plants were checked by flow cytometric analyses and seed production.

Flow Cytometric Analyses
The ploidy levels of AC-derived plants and seed-derived control plants were determined by flow cytometric analysis using CytoFLEX Flow Cytometer (Beckman Coulter International S.A., Nyon, Switzerland). The samples (100 mg/plant) of young leaves were collected from the acclimatized plantlets grown in the greenhouse. The samples were destroyed in Eppendorf tubes containing 1 mL Galbraith buffer and two stainless steel beads using TissueLyser II (Qiagen, Hilden, Germany) at 20 Hz for 2 min to isolate the nuclei from the young leaves of plantlets [42]. The suspensions were purified using 20 µm sieves and 10 µL RNase solution was added to each sample at room temperature for 60 min to degrade the RNA content in the samples. DNA content was painted with 40 µL Propidium Iodide (PI) solution (1 mg/L) for 30 min at room temperature. After the preparation, the DNA content of samples was measured by flow cytometer, and the ploidy levels of samples were determined based on histograms.

Statistical Analyses
The experiments were carried out with at least ten replications (ten AC/induction medium/genotype). Data of AC response (number of calli, regenerated plantlets, albinoand green plantlets) were analyzed by 2-way ANOVA. The statistical analyses were carried out by Microsoft ® excel 2019 software developed by Microsoft (Redmond, WA, USA).

Conclusions
DH plant production methods are widely used in crop breeding and research programs because of their ability to produce genetically pure lines in one generation [3,[43][44][45][46][47][48]. In vitro androgenesis was induced in AC of indica rice genotypes. Significant green plantlets production was achieved in the best treatment (4.7 green plantlets/100 anthers on average) using N 6 NBK induction medium and mMSNBK plant regeneration medium, while an adverse effect of exogenous zeatin (induction and plant regeneration media) was observed on the number of regenerated green plantlets. The methodological improvements mitigate the genotype dependency of AC step by step to enhance the practical utilization of the method. The mean spontaneous chromosome doubling was 51.89% based on flow cytometric analyses. Although the genotype influenced significantly the efficiency of AC, this method can be used for DH plant production in indica genotypes and their F 1 combinations. Utilization of DH rice plants with genes for climate resilience will be highly useful to reduce the cycle of breeding time and adding new effective genetic combinations for climate resilience while maintaining superior yield and pest and disease tolerances for mitigation and management of climate change.