Fertility and Progeny Ploidy of Triploids in China Rosa Germplasm

and tea to modern roses, and played very important roles in the formation history of modern roses. In China Rosa germplasm, there are some triploids that have not been used in breeding program but possess valuable traits. A better understanding of the fertility of China Rosa triploids and their behavior when used in interploidy crosses will not only improve gene introgression from China triploid roses to modern roses, but also enhance our understanding of male and female fertility mechanism with triploids. This study will also help us clarify the ‘triploid bridge’ in modern roses formation history, and make it possible for breeders to create novel rose types in the future. The objectives of this study were to usefulness triploid China rose for breeding by measuring the pollen size, quantifying the fertility and following the ploidy transmission of interploidy crosses.


Abstract Background
Although triploid plants are often assumed to be sterile, in Rosa, triploids appear to have been important in both species evolution and breeding. A better understanding of the fertility of China Rosa triploids and their behavior when used in interploidy crosses will not only improve gene introgression from China triploid roses to modern roses, but also enhance our understanding of male and female fertility mechanism with triploids.
The objective of this study was to evaluate the usefulness of ve triploid China roses for breeding by measuring the pollen size, quantifying the fertility and following the ploidy transmission of interploidy crosses.

Results
In the ve triploids, the pollen grain size of Rosa multi ora var. cathayensis Rehd 'Fen Tuan Qiang Wei' followed a normal distribution, suggesting that 1 mainly ploidy level pollen grain were produced, while the pollen size of other 4 triploids followed skewed, attened, and slightly bimodal distributions, indicating a wide range of chromosome numbers. Although none of the 5 triploids produced hips as females, 3 gave good hip and seed production on the tetraploid mother plant when used as the pollen parents. Based on the ploidy level analysis of their progeny, 'Fen Tuan Qiang Wei' produced 1n viable pollen whereas 'Chun Shui Lü Bo' and 'Yu Shi Zhuang' produced 2n viable pollen. In the meiosis of triploids, the triads, pentads and microcyte in the meiosis products indicated the objective triploid production of euploid pollen grains.

Conclusions
In this study we reported 1 China triploid rose produces 1n viable pollen, and 2 China triploid roses produce 2n viable pollen, and no anueploid progeny was produced. Our results indicate that the triploids contribute signi cantly to tetraploids formation in the crossing system of Genus Rosa, and it is possible for breeders to create novel rose types with valuable triploids in the future. With the special triploid resources found in this experiment, Genus Rosa will be an excellent material to study the mechanism of triploids in the future. Background Polyploidy (whole-genome duplication, WGD) is a major force in the evolution of higher plants (Simillion et al., 2002;Blanc and Wolfe, 2004). Most natural polyploids are thought to origin from sexual mechanism. As the fertilization of two unreduced gametes (female gamete and male gamete) seems to be an unlikely event, natural tetraploids are most likely to origin in two steps, via a triploid intermediary. Some studies about the mechanism of polyploid origin have been reported (Ramsey and Schemske 1998;Burtonand Husband, 2001;Husband, 2004). Most triploid are assumed to be sterile because their gametes are aneuploid, but some triploids do have fertility in breeding (Zhang, 2009;Zhang, 2017). The greater variability for pollen size in triploids had been observed, which re ected different ploidy levels or aneuploid gametes (Jacob and Pierret 2000;leus 2005;Crespel et al. 2006;Zlesak, 2009). Irregular chromosome pairing and abnormal meiotic division are the reasons for variations in pollen size and fertility, as has been reported in triploid Populus (Wang et al. 2010) and lily (Zhang, 2017). The aberrations during chromosome pairing, segregation, spindle formation, or cytokinesis might produce triads, dyads or monads, which are probably to produce unbalanced gametes (Bosco et al. 1999 reported, but the ploidy levels of the progeny were not consistent. In the crosses between tetraploid females and triploid males, Leus (2005) reported about 98% (123/125) tetraploid offspring, but Zlesak (2009) obtained about half triploid (23/43) offspring and half tetraploid offspring (20/43). In the crosses between diploid females and triploid males, Barden got diploid, triploid, and tetraploid offspring ( Barden and Zlesak, 2004).
Rosa resources from China, such as Rosa chinensis and Chinese Old Garden Roses, have contributed recurrent owering and tea scent to modern roses, and played very important roles in the formation history of modern roses. In China Rosa germplasm, there are some triploids that have not been used in breeding program but possess valuable traits. A better understanding of the fertility of China Rosa triploids and their behavior when used in interploidy crosses will not only improve gene introgression from China triploid roses to modern roses, but also enhance our understanding of male and female fertility mechanism with triploids.
This study will also help us clarify the 'triploid bridge' in modern roses formation history, and make it possible for breeders to create novel rose types in the future. The objectives of this study were to evaluate the usefulness of triploid China rose for breeding by measuring the pollen size, quantifying the fertility and following the ploidy transmission of interploidy crosses.

Plant materials
Five triploid China roses, including 4 Chinese old garden rose cultivars 'Chun Shui Lü Bo', 'Yu Shi Zhuang', 'Hu Zhong Yue', 'Si Mian Jing' and 1 species R. multi ora var. cathayensis Rehd 'Fen Tuan Qiang Wei' were used as triploid material (Table 1). Eight tetraploid modern roses and one diploid Chinese old garden rose were used as hybrid parents. Crosses between the triploids and tetraploids or diploids were conducted to produce populations with ploidy level segregation. Ploidy level analysis was conducted on the 190 randomly selected seedlings from the hybrid population.

Pollen viability and size observation
Fresh pollen sample was collected (Zlesak, 2009), then stored in refrigerator at -20℃. The sample was then mounted in a drop of 1% aceto-carmine on a microscope slide and viewed under the Olympus microscope (DP74, Japan). Empty and shrunken pollen grains were scored. The diameter of 500 pollen grains were measured using the attached calibrated micrometer of Olympus microscope.
Pollen viability was evaluated using the method proposed by Feng et al. (2017). A pollen grain was considered to have germinated if the pollen tube length was equal to or greater than the pollen diameter. Five microscopic eld areas were randomly selected, and 5 plant samples were observed.
Crossing experiment

Analysis of ploidy levels
Counts of somatic chromosome number was done in shoot tip tissue following Wang's method . The preparation was observed by using an Olympus DP75 microscope.
The ploidy level of progeny was also determined by ow cytometry (Feng et al., 2017).Samples were analyzed with a BD Company Accuri C6 ow cytometer. 5000-10,000 particles were detected and collected, 'Old Blush' was used as diploid standard. The somatic chromosome numbers of some seedlings were also checked by counting chromosomes of cells from shoot tips as described above.
Cytological observations of microsporogenesis

Results
Pollen morphology, size-frequency distribution and viability Pollen grains of the ve triploids had 3 colporates that annularly distributed in equal space, and extended to the two poles. They were long elliptic shaped, and their outer wall was coarse, microperforate, and decorated with streaks ( Fig. 1; N3P4C5 type, see Erdtman, 1971). Some anomalous pollens were found. The pollen grain size of the 5 triploids were analyzed by using 1 diploid and 1 tetraploid as controls (Table 2). According to P/E value, compared with 'Old Blush', the ve triploids had more long spherical shaped pollens, and less extra long spherical shaped pollens. Compared with the tetraploid 'Dee Dee Bridgwater', 'Yu Shi Zhuang' and 'QingliangXueshi' had similar quantity of long spherical shaped pollens, and the other 3 triploids had fewer long spherical shaped pollen. Gaussian distribution tests were conducted on frequency distribution with polar axis length of pollen grains (Table 3). Only 'Old Blush', 'Dee Dee Bridgewater ' and 'Fen Tuan Qiang Wei' showed a normal distribution of pollen size, indicating only one ploidy level pollens were produced. The other four triploids had skewed, attened, and slightly bimodal distributions, indicating the wide range of pollen size in triploid (Fig. 1). Pollen germination rate of 5 triploids on arti cial liquid medium were very different, with 'Fen Tuan Qiang Wei' had the highest germination rate of 32.00%, and 'Huzhong Yue' has the smallest germination rate of 4.00% (Table 4).

Pollen tube growth observation and male fertility of ve triploids
The pollen of 'Chun Shui Lü Bo' on the stigma of 'Carefree Beauty'germinated 8-12 hours after pollination ( Fig. 3a, b, c). By 24 hours the pollen tubes grew through the tissue of the stigma and reached the middle or the upper part of the style (Fig. 3d), by 48 hours the pollen tubes reached the lower middle of the style (Fig. 3e), and by 72 hours after pollination, a few pollen tubes reached the ovary (Fig. 3f).
Higher fruit set and number of seeds per hip were observed in interploidy crosses with the tetraploid 'Dee Dee Bridgewater' as a seed parent and triploid 'Chun Shui Lü Bo', 'Yu Shi Zhuang' and 'Fen Tuan Qiang Wei'as pollen parents (Table 5). The cross between diploid 'Old Blush'( ) and triploid'Fen Tuan Qiang Wei' ( ) had lower fruit set rate and germination rate than the cross between tetraploid 'Dee Dee Bridgewater' ( ) and Page 8/16 triploid 'Fen Tuan Qiang Wei' ( ). Crosses with triploid 'Qin Lian Xue Shi' and 'Hu Zhong Yue' as pollen parents did not set any hips. Stigma receptivity analysis and female fertility of triploids 'Carefree Beauty' stigmas turned to blue with high bubbling and 'Old Blush' stigmas turned yellow with high bubbling indicating high and moderate pistil receptivity respectively. In contrast, the stigmas of the ve triploids showed little bubbling indicating low pistil receptivity (Fig. 4c, d, e, f, g). This agrees with the lack of set among the triploids when used as females (Table 6). Progeny ploidy levels screening by ow cytometric analysis and chromosome counting Flow cytometric analyses of the 138 progenies, which were from the crosses with 'Chun Shui Lü Bo' or 'Yu Shi Zhuang' as male parents and tetraploids as female parents, indicated that 129 seedlings were tetraploid, 6 seedlings were pentaploid and 3 seedlings were triploid (Table 7 ). Chromosome counts of the pentaploid and triploid seedlings showed that these were tetraploids with 28 chromosomes (Fig. 5). Other check of the ow cytometric results with chromosome counts con rmed the ploidy designation. Table 7 Ploidy level evaluation of rose F 1 interploidy progeny by ow cytometry analysis.  (Table 7). Chromosome counting were also conducted on 6 seedlings which were randomly selected from the cross of 'Gold Mary' and 'Fen Tuan Qiang Wei', and the results showed they were all triploid (Fig. 5), coincident with the ow cytometric analysis.

Abnormal meiotic chromosome behaviors of 'Chun Shui Lü Bo' and 'Fen Tuan Qiang Wei'
Abnormal chromosomal behaviors were recorded (Table 8) in all stages of meiosis, indicating complex chromosome pairing and unbalanced chromosome segregation in the triploid rose genotypes examined.

Discussion
The positive correlation between DNA content and pollen size makes it possible to decide the ploidy level by measuring the pollen size (Dewitte et al. 2012). For normal diploid plants, the frequency distribution of pollen grain size is unimodal, but for unreduced pollen producers, it is bimodal (Tondini et al.1993). In this research, pollen grains size of 'Old Blush'(diploid) 'Dee Dee Bridgewater ' (tetraploid) and 'Fen Tuan Qiang Wei' (triploid) followed a unimodal distribution, suggesting that 1 ploidy level pollen grain were produced. For triploid 'Chun Shui Lü Bo' and 'Yu Shi Zhuang', the size of pollen grains had higher variation coe cient (from 18.8%to 24.3%) and followed a skewed, attened, and slightly bimodal distributions, suggesting the wide range of pollen size in triploid, which may represent the various chromosome numbers of pollen grains due to irregular chromosome pairing and unbalanced chromosome segregation.
For the 5 triploid China rose cultivars used as pollen parents in this study, 2 of them did not set any hips, 3 of them gave good hip and seed production on the tetraploid mother plant. This result indicated that some of the triploid China Rosa germplasm do have good fertility ability, and can be used as gene introgression bridge in future rose breeding program.
In this study, when triploid 'Fen Tuan Qiang Wei' was used as male parent, it produced 1n pollen as indicated by the unimodal pollen size distribution, the triploid seedlings from the 4x ( ) × 3x ( ) crosses, and the diploid seedlings from the 2x ( ) × 3x ( ) crosses. This result is different from Leus's, Zlesak's and Barden's work (Leus, 2005;Zlesak, 2009; Barden and Zlesak, 2004), and also different from the crosses in current study when 'Chun Shui Lü Bo' or 'Yu Shi Zhuang' were used as male parents. This is the rst report that triploid rose producing mostly1n viable pollen, and 1n pollen is more vigor than 2n pollen. It is a special result. To my knowledge, the only similar report I can nd is about Hieraciumechioides, Peckert reported he got about 80% triploids in 4x × 3x crosses (Peckert, 2006 In current study, when triploids were used as male parents, only euploid progeny were obtained, the same as Leus's and Zlesak's work. It seems anueploid progeny are rarely produced when triploid Rosa are used as male parents. This is different from the studies about apple (Zhang, 2009), lily (Zhang, 2017) and populus , in which more aneuploid than euploid progeny were obtained. This is also different from Ramsey's report. In Ramsey's summary, the most common pollen chromosome number was 3x/2, followed by 3x/2 − 1, and the lowest pollen chromosome number was 1 × (3%) and diploid2 × (2%) (Ramsey, 1998). In current study, the wide range of pollen size of 'Chun Shui Lü Bo' and 'Yu Shi Zhuang' suggested the various chromosome numbers of pollen grains. So probably some of the pollens were anueploid. The reason why all the progeny obtained in this study were euploid may be that euploid gametes have greater vigor than the aneuploidy gametes in fertility. Though the aneuploidy pollen is not vigorous, in crosses with triploid Rosa as male parents, the fruit set and the average number of seeds per hip were higher compared with most other triploids (Zhang, 2017;Zhang, 2009;. This result indicates that the Rosa triploids have some kind of mechanism to make sure they can produce much more vigorous euploid progeny than other triploids. In this research, the triads, pentads and microcyte in the meiosis products indicated the objective triploid production of euploid pollen grains. Microcyte formation likely resulted in the loss of chromosomes from microspores, which may increase the variation in pollen size. According to Ramsey, euploid gametes are very rare, because haploid-diploid chromosome assortments are only produced when the separation of multivalents and unpaired chromosomes at the rst meiotic division in a spore mother cell is very unequal (Ramsey and Schemske1998). Few studies had been done about the unique mechanism of meiosis in triploid roses. The reasons why triploid roses can produce so many euploid gametes, and why 'Fen Tuan Qiang Wei' produced 1n viable pollen whereas 'Chun Shui Lü Bo' and 'Yu Shi Zhuang' produced 2n viable pollen are still uncertain. More researches need to be done to clarify the mechanism in the future.

Conclusion
In this study we reported 1 China triploid rose produces 1n viable pollen, and 2 China triploid roses produce 2n viable pollen, and no anueploid progeny was produced. Our results indicate that the triploids contribute signi cantly to tetraploids formation in the crossing system of Genus Rosa, and it is possible for breeders to create novel rose types with valuable triploids in the future. With the special triploid resources found in this experiment, Genus Rosa will be an excellent material to study the mechanism of triploids in the future.

Abbreviations
Not applicable.

Declarations
Ethics approval and consent to participate: Not applicable.
Consent for publication: Not applicable.
Availability of data and material: All the data can be available, and it is according to China law that if the material can be available.