Efficient doubled haploid production in microspore culture of loose-curd cauliflower (Brassica oleracea var. botrytis)
- First Online:
- Cite this article as:
- Gu, H., Zhao, Z., Sheng, X. et al. Euphytica (2014) 195: 467. doi:10.1007/s10681-013-1008-x
- 813 Downloads
We present an improved protocol for highly efficient production of doubled haploid loose-curd cauliflower plants (Brassica oleracea var. botrytis) via microspore culture. Our experiment explored factors such as donor plant treatment, flower bud pretreatment, embryo germination medium, and ploidy characterization of regenerated plants. Our technique efficiently produced embryos from both tight- and loose-curd donor plants, although the embryo yields were genotype dependent. We achieved a germination rate of around 30 % by employing a hormone combination of zeatin, indole-3-acetic acid, and 6-benzylaminopurine pretreatment culture. We also used 1–4 days of cold pretreatment of the flower buds, which were submerged into NLN-13 medium, to induce microspore embryogenesis. Analysis using an FCM Ploidy Analyzer showed that more than 50 % of regenerated plants were spontaneously doubled haploids, more than 25 % were tetraploids, and fewer than 7 % were haploid. Visual examination of plants in the field revealed that they had distinct phenotypic characteristics relating to their ploidy level. The efficient production of double haploids using our improved microspore culture technique is a promising approach that can be applied in loose-curd cauliflower breeding programmes and genetic research.
KeywordsLoose-curd cauliflower (Brassica oleracea var. botrytis) Microspore culture Doubled haploid Spontaneous doubling Cold pretreatment
Petal length/Anther length
- NLN medium
Nitsch and Nitsch medium
- MS medium
Murashige and Skoog medium
Microspore culture technology has been routinely applied in Brassica (B.) breeding programmes and genetic research since successful isolation and culture of microspores in B. napus was first reported (Lichter 1982). In the past 30 years, there have been great advances in microspore culture of major B. species, including B. napus, B. rapa, B. juncea, B. carinata (Chuong and Beversdorf 1985; Lichter 1989; Sato et al. 1989; Baillie et al. 1992; Burnett et al. 1992; Ferrie et al. 1995; Lionneton et al. 2001; Cousin et al. 2009; Takahira et al. 2011). Several new techniques for microspore culture in B. species have been developed (Ferrie and Caswell 2011; Bhowmik et al. 2011; Takahira et al. 2011; Ahmadi et al. 2012), but there are still some problems that prevent universal application of the technique in B. oleracea breeding and genetic research. B. oleracea var. botrytis is considered to be one of the most recalcitrant varieties of B. oleracea in terms of embryogenic response, although great effort has been devoted to refining techniques in this species (Takahata and Keller 1991; Duijs et al. 1992; Yang et al. 1992; Stipic and Campion 1997; Chatelet et al. 1999; da Silva Dias 1999, 2001; Gu et al. 2004b; Winarto and Teixeira da Silva 2011). The main problems are genotype dependency, a low embryogenesis and germination rate, and unknown ploidy characterization (Olmedilla 2010).
Loose-curd cauliflower (B. oleracea var. botrytis), a recently developed variety, has flowers like loose curd in appearance, a long green flower peduncle, and good flavor. It has been popular in southern China, including Taiwan, for more than 10 years, and is gaining popularity in other areas of China and Asia (Gu et al. 2012; Zhao et al. 2012). For commercial purposes, the classic breeding approach for cauliflower is to produce F1 hybrids based on inbred lines that selfed and selected through 6–8 generations of inbreeding, which is made more difficult because the plants are self-incompatible. Doubled haploids (DH) produced by microspore culture are an ideal approach because they generate these inbred lines only in one generation, and they are highly efficient at accelerating the breeding of new varieties, saving time and labor (Palmer et al. 1996b; Forster et al. 2007; Wędzony et al. 2009).
During more than 10 years of investigation, we developed a promising technique for highly efficient production of DH plants via microspore culture in loose-curd cauliflower. Our improved protocol took into consideration growing conditions for the donor plants, cold pretreatment of the flower buds, alterations to the embryo germination medium, and ploidy characterization of regenerated plants. Recently, several new varieties of loose-curd cauliflower were bred using this improved microspore culture technique (Gu et al. 2012).
Materials and methods
Donor materials, curd-cutting method, and growing conditions
Microspore embryogenesis and embryo development in loose-curd cauliflower
Mean embryo yield (Embryos/bud)
Highest embryo yield (Embryos/bud)
Mean rate of germinated embryos (%)
Xueshan No. 2
Microspore isolation and culture
The cauliflower microspore culture protocol was a modification of our technique with B. species that adjusted for local experimental conditions (Gu et al. 2003, 2004a, b). The microspore isolation work was conducted each year from January to April, when the range of ambient temperatures was 5–30 °C, which was suitable for the culture.
For each isolation and replication procedure, 9–12 flower buds at the late uninucleate to early binucleate stage of microspore development were selected. Inspection by a dissecting microscope and the naked eye confirmed that the microspores were developing normally and pollen was mature before flower buds were harvested. Normal development was defined as normal bud size (4–6 mm long) and normal ratio of petal length to anther length (P/A, 1–1.2). The chosen buds were surface sterilized in 0.1 % (w/v) mercuric chloride (HgCl2) solution containing one drop of Tween 20 per liter of detergent for 10 min, prior to washing 3–5 times with cold sterile water. The sterile buds were macerated in cold NLN-13 medium (Nitsch and Nitsch 1967; Lichter 1982) with 13 % sucrose, buffered at pH 6.0, and filtered through a 40-μm nylon filter into a 50-mL centrifuge tube. The crude microspore suspension was centrifuged at 850 rpm for 4 min, twice, and then the microspores were resuspended in the same fresh NLN-13 medium. After activated charcoal was added, the microspore suspension was then dispensed into 60 × 15 mm2 Petri dishes, 4 mL in each dish; that is, the equivalent of one flower bud was placed in each Petri dish with a microspore density of about 2 × 104 per mL. The dishes were sealed with double layers of Parafilm, incubated at 32–33 °C of heat shock in the dark for 1 day, and then moved to 25 °C, still in the dark.
Effect of cold pretreatment on microspore embryogenesis in cauliflowers
Cold pretreatment (days)
No. of embryo yield per flower bud ± SE
6.7 ± 1.1ba
10.5 ± 2.0b
26.8 ± 3.5a
24.3 ± 3.4a
20.5 ± 3.2a
29.3 ± 4.3a
10.2 ± 2.3c
13.8 ± 2.7bc
16.2 ± 3.0b
16.7 ± 3.5b
8.5 ± 1.1a
12.2 ± 1.4a
10.3 ± 1.2a
6.5 ± 0.7a
7.3 ± 0.9a
Embryo germination and plant regeneration
As soon as white embryos were visible to the naked eye, the dishes were transferred to a slow rotary shaker at 45 rpm, still in the dark at 25 °C. After about 30 days of microspore isolation, embryos at the cotyledon stage were counted. These embryos were then transferred to glass growth containers containing semi-solid MS-2 medium with 2 % sucrose, 0.4 % low-gelling agarose, 2 mg/L trans-Zeatin, 0.1 mg/L indole-3-acetic acid, (pH 5.8), and 2 mg/L 6-benzylaminopurine (Murashige and Skoog 1962). After 7–10 days’ pretreatment culture, the swelling embryos were transferred to glass growth containers containing hormone-free solid MS-2 medium with 2 % sucrose and 1 % agar at pH 5.8 for embryo germination under a photoperiod of 16 h (100 μE/m−2/s−1).
Approximately 3–4 weeks later, a cluster of shoots developed from the embryo callus. The shoots were cut off the callus or hypocotyl tissue and were transferred to glass growth containers with fresh solid MS-2 medium, where rooting took place. Over a period of several months, regenerated plants were regenerated several times, 2–4 plantlets were regenerated from one embryo in one container in order to reduce the loss rate when the plants were transferred to the field. Rooted plantlets were transferred to soil–perlite plugs and were kept for 2 weeks in a small plastic tunnel (with low light intensity and high humidity). In this environment they gradually adapted to natural conditions until October, when the mean temperature outdoors drops below 25 °C.
FCM and field ploidy level analysis
The ploidy level was estimated using an FCM Ploidy Analyzer (Partec GmbH, Germany). Then regenerated plants were transferred to the plastic tunnel field (Gu et al. 2003). Approximately 1.5 cm2 of young leaves were chopped drastically in a detergent solution by double-edged razor blade, and the leaf sample was then stained with 4′-6-diamidino-2-phenylindole solution for 1 minute. The suspension of cell nuclei and debris was filtered through 50-μm nylon gauze and the filtrate was immediately analysed with the FCM Ploidy Analyzer. The instrument was calibrated by using leaf tissue from normal diploid cauliflower plants as a standard. Each histogram was generated using at least 10,000 cells for each of three replications. In addition, the ploidy level of regenerated plants in the field was judged by visual inspection and characterization of plant type, leaf shape, flower size, and pollen quantity.
Microspore embryogenesis and plant regeneration
In all examined donor genotypes, regardless of their different origins and maturity levels, microspore embryogenesis occurred in both the tight- and loose-curd types in an average of 88.2 % of cases, although the embryo yields per flower bud varied with the particular genotype (Table 1). The tight-curd type produced an average of 10 embryos per bud, and the loose-curd type produced an average of 11.6 embryos per bud, which was not significantly different. The mean rate of germination for the embryos was 36.2 % for the tight-curd type and 33.7 % for the loose-curd type. Loose-curd genotype 3146 produced the highest mean number of embryos (58.2 embryos per flower bud) and the highest single yield in a flower bud (178 embryos), whereas the two genotypes 3038 and 3024 failed to produce embryos at all. The efficient germination rate of around 30 % was achieved by employing the hormone pretreatment culture for only a short time; this rate was much higher than the rate we achieved in our earlier experiment when we directly transferred MS-2 medium (Gu et al. 2004b).
Influence of cold pretreatment on microspore embryogenesis
The genotypes 3100, 3203, and 801, which showed a good response to microspore embryogenesis, were employed as donor plants to test the influence of cold pretreatment on microspore embryogenesis. Flower buds were collected at the mid- to late-uninucleate stage and were distributed randomly among the treatments, which varied in duration (The stage of flower buds for instant isolation were late uninucleate and early binucleate). In all tested genotypes, 1–4 days of cold pretreatment efficiently induced microspore embryogenesis to different degrees (Table 2).
In genotype 3100, embryo yield was significantly improved by cold pretreatment of flower buds; 2 days’ treatment produced a threefold increase. In genotype 3203, embryo yield was obviously reduced by cold pretreatment, but the yield was gradually increased when the duration of treatment was extended. In genotype 801, embryo yield was not affected by 1–4 days of cold pretreatment compared with the control, which was relatively stable for embryogenesis. Our results indicated that the cold pretreatment effect was genotype-dependent, but the application of cold pretreatment could effectively produce embryos because collection and sterilization of flower buds was a one-time event, saving time and labor.
Ploidy level of regenerated plants
FCM Ploidy level in microspore-derived plantlets of loose-curd cauliflower
No. of plantlets tested
Pentaploid and hexaploid (%)
Aneuploid and chimera (%)
The results showed that, on average, more than 50 % of regenerated plants were spontaneously doubled haploids, more than 25 % were tetraploids, more than 10 % were aneuploid and chimera, fewer than 7 % were haploid, and about 2 % of plantlets were other polyploids such as triploids, pentaploids, and hexaploids. The highest spontaneous DH percentage of 70 % occurred in genotype 801, and the lowest DH percentage of 32.7 % was found in genotype 3203. Interestingly, all tested genotypes had a low percentage of haploids; the result was similar to data from more than 10,000 regenerated cauliflower plants from our experiments over the past few years, implying that additional chromosome doubling is not necessary in order to enhance DH production using microspore culture techniques.
The field investigation of ploidy characterization by plant type, leaf shape, flower size, and pollen quantity at the flowering stage of regenerated plants revealed that the ploidy results mostly matched the FCM level. The field-regenerated plants were significantly characterized by their ploidy level (e.g. haploids demonstrated weak growth, lack of pollen, and small flowers; tetraploids showed more wrinkled leaves, smaller leaves, relatively less pollen, and larger flowers; the diploids had normal growth characteristics in terms of plant type, leaf shape, and flower size, which was no different from plant growth of inbred lines).
This study is the first report of efficient DH production achieved through an improved protocol of microspore culture in loose-curd cauliflower, cultivated under natural field conditions, employing the novel methods of cold pretreatment, hormone pretreatment, and ploidy analysis. Cauliflower has long been considered to be recalcitrant to DH technology because it shows a poor embryogenic response to microspore culture (Duijs et al. 1992; Olmedilla 2010). Microspore culture of cauliflower was successfully reported only by Duijs et al. (1992), Chatelet et al. (1999), and da Silva Dias (1999), but only Chatelet et al. applied it to cauliflower breeding practice. Prior to beginning our study, we obtained a large number of cauliflower DH plants, which we used to accelerate our plant breeding programmes, to breed new varieties, and for basic genetic research (Gu et al. 2012).
The success of microspore embryogenesis and plant regeneration depends on many factors, such as plant genotype, donor plant physiology, pollen developmental stage, pretreatments, culture conditions, and more (Palmer et al. 1996a; Ferrie and Caswell 2011; Olmedilla 2010). In previous studies of B. oleracea using microspore culture, many efforts were made to understand the influence of factors such as genotype differences, the application of activated charcoal, temperature regime, bud stage and width, microspore density, and pretreatment (Takahata and Keller 1991; Duijs et al. 1992; Chatelet et al. 1999; da Silva Dias 1999, 2001; Gu et al. 2004b; Winarto and Teixeira da Silva 2011; Yuan et al. 2011).
In our experiment, the rates of embryogenesis, embryo yield, and embryo quality were quite different for different genotypes. Of 16 donor plants, 2 failed to yield any embryos and some yielded very few embryos, which implies that the bottleneck problem in cauliflower microspore culture is still genotype. Early maturity genotype 3146, which showed the best rate of embryogenesis, will be a model donor genotype for microspore culture studies of loose-curd cauliflower.
Donor plant physiology or donor growth condition are considered to be the most important factors in achieving success in microspore culture, especially in cauliflower and under natural growth conditions (Palmer et al. 1996a). Cauliflower curd, which is composed of indefinitely undeveloped flower branches, is difficult to bolt if growing freely. Therefore, in contrast with other B. species, the cauliflower curd of this variety should be drastically cut back to guarantee the quality and nutrient content of flower buds, which is vital for success in microspore culture.
In a previous study (Fletcher et al. 1998), donor plants grew under controlled low-temperature conditions and a noticeable difference in day/night temperature, which promote embryogenesis. However, under our experimental conditions, all donor plants were cultivated in a field, where the temperature was out of our control and changing with the seasons. We applied some modifications from our previous protocol (Gu et al. 2004a) to the cold pretreatment of immature flower buds submerged into culture medium, which is a standard part of our microspore culture method. It improves embryo production and saves labor and time by allowing one-time collection and sterilization of a large number of flower buds before the isolation step. Cold pretreatment in microspore culture has been successfully employed in broccoli (Yuan et al. 2011), and appears to be useful in recalcitrant B. oleracea microspore culture.
Unlike our earlier work on other microspore culture of B. napus and B. rapa, in which the culture did not contain hormones (Gu et al. 2003, 2004a), and unlike our microspore culture of cabbage and broccoli, in which the hormones did not affect results (data not shown), in the present cauliflower microspore culture work, the novel approach of pretreatment with a combination of hormones enhanced the germination rate of embryos. The callus embryos produced more shoots, improving the survival rate of regenerated plants when they were transferred to the field (Ahmadi et al. 2012). Embryos derived from cauliflower microspores were very sensitive to hormones, and calli were easily introduced, which makes this an excellent regeneration system for gene transfer (Kumlehn 2009).
A good understanding of the ploidy characterization of microspore-derived plants is necessary for accelerating cauliflower breeding and applying genetic engineering to cauliflower. Our microspore culture technique produced a high percentage of spontaneous DH and tetraploids without employing chromosome doubling, supporting the results of previous work (Chatelet et al. 1999). Our results, however, contrasted strongly with those of other studies of B. napus and even B. oleracea (Duijs et al. 1992, da Silva Dias 1999; Gu et al. 2004a), in which a high percentage of the plants produced were haploid. This phenomenon occurs in other B. species, such as in B. rapa and B. oleracea (Zhang et al. 2001; Gu et al. 2003), and there still remains lack of knowledge. We suppose whether is the relative to the microspore culture method or B. species or isolated microspore developmental stage, because we found there were a high percentage of early binucleate microspores in cauliflower cultures.
This work was financially supported by Zhejiang Provincial Department of Science and Technology, and Zhejiang Academy of Agricultural Sciences. The authors also thank Dr. Linping Wang and Xiaohui Zhang for their kind assistance during the experiment.