Fitness of the first backcross generations from the second to the sixth progenies of glyphosate-resistant transgenic Brassica napus and wild Brassica juncea in absence of the herbicide

Abstract

Successful introgression of a transgene from a transgenic crop into a wild or weedy relative is determined by the fitness of backcross generations carrying the transgene. To provide insight for ecological risk assessment of gene flow between transgenic Brassica napus and wild Brassica juncea, this study investigated the fitness of the first backcross generations from the second to the sixth progenies (BC₁F₂R–BC₁F₆R) between glyphosate-resistant transgenic B. napus and wild B. juncea at low density (5 plants/m²) and high density (10 plants/m²), and monoculture and mixed planting (wild B. juncea: BC₁F₂R–BC₁F₆R = 1:1). Correlations between the fitness components of backcross progeny, planting density and planting patterns were analyzed. In the monoculture at low density, compared with B. juncea, earlier generations BC₁F₂R and BC₁F₃R had low composite fitness, while later generations from BC₁F₄R to BC₁F₆R were more fit. At high density, whatever monoculture or mixed planting, all backcrossed generations had lower composite fitness than B. juncea. Correlation analysis indicated that both planting density and pattern significantly affected the fitness components of the first backcross generations from the second to the sixth progenies (BC₁F₂R–BC₁F₆R). These results indicate that the probability of transgene introgression from cultivated B. napus to weedy B. juncea is likely to be highly contingent on the specific growing conditions of their backcross descendants.

摘要

抗草甘膦转基因油菜和野芥菜回交1代子2代-子6代在无除草剂选择压下的适合度

转基因作物的抗性基因能否成功渗入野生或近缘杂草取决于携带转基因回交后代的适合度。为深入了解转基因油菜(Brassica napus)向野芥菜(wild B. juncea)基因流动的生态风险，本文研究了在低密度(5 株/m²)和高密度(10 株/m²)下，以及单种和混合种植(野芥菜：BC₁F₂R–BC₁F₆R = 1:1)下抗草甘膦转基因油菜与野芥菜回交1代子2代到子6代(BC₁F₂R–BC₁F₆R)的适合度；并分析回交后代的适合度成分、种植密度和种植方式之间的相关性。研究结果表明：在单种低密度下，与野芥菜相比，回交1代子2代(BC₁F₂R)和子3代(BC₁F₃R)具有较低的适合度，而回交1代子4代-子6代(BC₁F₄R–BC₁F₆R)具有较高的适合度；无论是单种还是混种，高密度下所有回交后代的适合度均低于野芥菜。相关性分析结果显示，种植密度和种植方式对回交后代(BC₁F₂R–BC₁F₆R)的适合度成分有显著影响。这些结果表明，转基因从栽培油菜渗入到野芥菜很大程度上取决于回交后代的特定生长环境。

INTRODUCTION

Oilseed rape (Brassica napus, 2n = 38 = AACC) is grown worldwide as the second most important source of vegetable oil (Su 2006). Weeds compromise the growth and yield of oilseed rape, and it has consequently been modified to make it herbicide resistant, especially for non-selective glyphosate or glufosinate ammonium. Since 1996, based on the data from the International Agricultural Biotechnology Application Service Organization (ISAAA), 37 single genetic modification events were approved for oilseed rape with different herbicide-resistant (HR) genes to allow selective control of various wild or weed species in oilseed rape fields. Due to its profitability, reduced labor force requirements and trait efficiency, transgenic HR oilseed rape has been the fourth major biotech crop after soybeans, maize and cotton since transgenic crops were commercialized firstly in Canada in 1996. In 2019, 10.1 million hectares of biotech oilseed rape were planted in the USA, Canada, Australia and Chile. Most of this cultivation was transgenic HR oilseed rape (ISAAA 2019; Jiang and Wen 2021).

In China, commercial cultivation of transgenic HR oilseed rape has not been authorized yet. It is noteworthy that GT73, T45, Oxy-235, Ms8Rf3, RF3, MON88302, Ms1Rf1, Ms1Rf2, Topas19/2, GT73 and DP73496 transgenic oilseed rape were awarded safety certificates allowing them to be imported as raw material for processing since 2019. Meanwhile, domestic transgenic HR oilseed rape has also been developed under the support of the government. Therefore, assessing the potential risks of gene flow from transgenic HR oilseed rape to wild or weedy relatives in China is needed (Wang et al. 2018).

One of the concerns surrounding the import of transgenic HR oilseed rape is that, through seed spillage during transportation and processing, the HR trait can escape into agricultural or semi-natural habitats. Thus, transgenic HR oilseed rape varieties might form volunteer populations. Many reports document the recruitment of HR volunteer seedlings in countries either commercializing or importing transgenic HR oilseed rape (Devos et al. 2004; Liu et al. 2013a, 2013b; Schulze et al. 2014). These volunteer populations might be important reservoirs from which a HR transgene could be passed into the genome of weedy relatives (Stewart et al. 2003). Once this happens, wild or weedy relatives carrying the HR transgene would no longer be controlled by those herbicides to which they are tolerant. Introgression of the HR transgene from crop to weed could thereby result in a more problematic weed.

Introgression describes the transfer of alleles from one taxon to another via backcrossing. The initial stage of introgression is the formation of a hybrid. Following this first step, the initial hybridization of transgenic crops and wild or weedy relatives, the successful introgression of transgenes into wild or weedy relatives depends on the fitness of the hybrid and backcross progenies carrying transgenes (Chapman and Burke 2007; Gueritaine et al. 2002). Fitness is reflected by vegetative and reproductive growth indicators and determines whether the hybrids can survive and transmit their genes to successive generations (Warwick et al. 2009).

Brassica napus is a predominantly self-pollinated crop (Beckie et al. 2003). However, its pollen can be dispersed by wind and insects, spontaneously hybridizing with the most cross-compatible recipient weed species, B. rapa (2n = 20, AA) (Jorgensen et al. 1996; Mikkelsen et al. 1996) and more distantly relative weed Raphanus raphanistrum (Chevre et al. 2000; Darmency et al. 1998). The fertile F₁ hybrids or subsequent hybrid progenies can be produced from B. napus to B. rapa (Hauser et al.1998a, 1998b).

After Brassica rapa, Brassica juncea (2n = 36 = AABB) is the second most cross-compatible recipient of the gene flow from B. napus. Brassica juncea is widely cultivated in Asia, the USA and Canada, especially for oil and mustard production (Sun et al. 2018; Tsuda et al. 2014). Self-compatible wild B. juncea originates from northwest China, and has spread eastward along the Yellow River and Yangtze River, becoming a noxious weed of cropping systems, roadsides and even empty lots across China (Sun et al. 2018).

The spontaneous hybridization of conventional and transgenic oilseed rape with cultivated or wild B. juncea has been reported since shortly after the introduction of transgenic technology (Bing et al. 1996; Frello et al. 1995). However, the sexual fertility of the F₁ hybrid is low (Bing et al. 1996; Frello et al. 1995; Mikhaylova and Kuluev 2018; Tang et al. 2018). Despite this, the first backcross using F₁ hybrids as maternal or paternal plants crossed with wild B. juncea successfully produces a BC₁ generation (Liu et al. 2010; Song et al. 2010). Fertility of BC₁ generations from these reciprocal crosses of F₁ (B. juncea × glyphosate-resistant B. napus) and wild B. juncea was also reported by Song (Song et al. 2010). The average seed per silique of BC₁ obtained from wild B. juncea × F₁ and F₁ × wild B. juncea were less than 3 and 2, respectively, which was significantly lower than wild B. juncea (average 13 seeds per silique). However, fitness of self-pollinated progeny of BC₁ was not reported. The process of introgression is not as simple as a single BC₁ generation and actually occurs in many steps that involve several hybrid generations (Stewart et al. 2003).

This study aimed to investigate the fitness of the first backcross generations from the second to the sixth progenies of glyphosate-resistant transgenic B. napus and wild B. juncea at two different planting densities (5 plants/m² and 10 plants/m²) and two planting patterns (monoculture and mixed planting). The results may be helpful in assessing potential introgression from transgenic HR oilseed rape to wild B. juncea under realistic agricultural conditions.

MATERIALS AND METHODS

Plant species

Glyphosate-tolerant (Roundup Ready, event RT73) oilseed rape (B. napus, AACC) was provided by Dr. Wenming Zhang, Agriculture and Agri-Food Canada (Ottawa Research and Development Centre). RT73 (carrying cp4-epsps and gox transgenes that confer glyphosate tolerance) were produced from spring-type B. napus cv. Westar. The cp4-epsps and gox transgenes in RT73 are located on chromosomes in the A-genome (Song et al. 2021). Wild B. juncea seedlings were identified in a winter wheat field in Jiangpu (32.01° N, 118.62° E), Nanjing, China, using SSR markers (Sun et al. 2018).

Experimental material

F₁ interspecific hybrids were obtained by hand pollination, using at least twenty wild B. juncea plants (maternal) crossed with the pollen of twenty glyphosate-tolerant oilseed rape plants (paternal), as described previously (Song et al. 2010). The mature seeds of the F₁ produced by the different maternal plants, were harvested separately, and stored in dark conditions in a refrigerator at 4 °C until further use. Seeds were planted and emerging seedlings sprayed with glyphosate at the fourth to fifth leaf stage, as described in Methods. The surviving F₁ (referred to as F₁R) seedlings were used for backcrossing.

The inflorescences of the first backcrossed plants were bagged before anthesis and allowed to self-pollinate to produce the first-generation progeny, named BC₁mF₁ and BC₁pF₁. The plants of BC₁mF₁ and BC₁pF₁ that survived after being sprayed with glyphosate were named BC₁mF₁R and BC₁pF₁R. The second to sixth progeny of the first backcross generation were obtained in the same way. A schematic diagram of hybrid and backcross progeny is shown in Fig. 1.

Methods

At least 100 filled seeds from each of the 10 mother plants of the first backcross progeny displaying the highest seed fertility were selected randomly; individual seeds were directly sown into single plastic pots (6 cm in diameter) containing growth media (a mixture of garden soil and peat at 1:1 (v/v) ratio) for the duration of the investigation. Plants were germinated and grown as described previously by Song et al. (2010).

Once the plants reached the fourth- to fifth-leaf stage (after 4 weeks), they were sprayed with 41% glyphosate isopropylammonium AS (Roundup Ultra; Monsanto, St. Louis, MO, USA) at 1037 g (a.i)/hm² as described by Song et al. (2021). Fifty plants of each type of BC₁ progeny were randomly selected for PCR testing as described by Song et al. (2021).

Uniformly sized progeny of BC₁ plants that survived treatment with glyphosate were transplanted individually into field plots according to the experimental design. The area of the monoculture planting plot at both low and high densities was 3 m² (2.5 m × 1.2 m). It contained a 3 × 5 grid (15 individuals) and 6 × 5 grid (30 individuals) with a 50-cm spacing at low (5 plants/m²) and high density (10 plants/m²), respectively. The area of the mixed planting plot at low density was 4 m² (2.5 m × 1.6 m) and at high density was 3 m² (2.5 m × 1.2 m). It contained a 4 × 5 grid (20 individuals) and 6 × 5 grid (30 individuals) with a 50-cm spacing at low (5 plants/m²) and high density (10 plants/m²), respectively (Fig. 2). The distance between adjacent plots was 50 cm. Plots were established in a randomized block design with three replications. Cultivation practices, including fertilization, were uniform across all experimental plots. Before transplanting, compound fertilizer (active ingredient content of 75% at 555 kg/ha, Monsanto) was supplied.

The experimental area was isolated by a 100-m-wide corn crop and surrounded by a fence to prevent human or animal damage.

Measurement of fitness composition

Each plant was harvested individually in May of the following year. The harvested plants were sun dried to complete desiccation. Fitness-associated traits are evaluated as described in Table 1.

Statistical analysis

All observed data were subjected to one-way ANOVA, and either the t-test or Duncan multiple range test was used to compare between wild B. juncea and progenies. Pearson correlation analysis was conducted between fitness components. All statistical analyses were performed using SPSS 17.0.

To estimate relative fitness, the fitness-associated traits of B. juncea were defined as ‘1’, and every other type was assigned a fitness value based on its ratio to the value assigned to progenies of the transgenic oilseed rape and wild B. juncea. The composite fitness was the mean of seven relative fitness estimates from the vegetative to mature stages, including plant height, stem diameter, dry-aboveground biomass per plant, silique number per plant, seed weight per plant, silique length and seed number per silique.

RESULTS

The composite fitness of the progeny of the first backcross generations in monocultures

Low density

In the low-density monoculture planting experiment, BC₁F₂R and BC₁F₃R were shorter and less productive than B. juncea (Table 2). The composite fitness of BC₁F₂R and BC₁F₃R was 0.74 and 0.86, respectively, which was significantly lower than B. juncea. However, all fitness components of BC₁F₄R, BC₁F₅R and BC₁F₆R were similar to B. juncea. The composite fitness of BC₁F₄R, BC₁F₅R and BC₁F₆R was 0.94–1.0. It was not significantly different compared with wild B. juncea (Fig. 3).

High density

In the high-density monoculture planting experiment, all BC generations were shorter and less productive than wild B. juncea (Table 3). The composite fitness of the first backcross generations from the second to the sixth progenies was 0.68–0.95 at high density, which was significantly lower than wild B. juncea (Fig. 3).

The composite fitness of the first backcross generations and wild B. juncea in mixed planting

Low density

In the low-density mixed planting experiment, all fitness components of BC generations were lower than wild B. juncea (Table 4). The composite fitness of the first backcross generations from the second to the sixth progenies was 0.72–0.93, which was significantly lower than wild B. juncea (Fig. 4).

High density

In the high-density mixed planting experiment, all BC generations were shorter and less productive than wild B. juncea (Table 5). The composite fitness of the first backcross generations from the second to the sixth progenies was 0.65–0.85, which was significantly lower than wild B. juncea (Fig. 4).

Correlation analysis between fitness components of wild B. juncea and the first backcross generations

Correlation analysis across wild B. juncea and B. napus showed that under the experimental conditions of the study, vegetative components and reproductive components were significantly positively correlated with the planting density and planting patterns, and not significantly correlated with their interaction (Table 6).

DISCUSSION

Transgenic oilseed rape (B. napus, AACC), which is one of the most important transgenic crops in the world, outcrosses with wild or weedy relatives such as B. rapa (AA = 20) (Allainguillaume et al. 2006; Johannessen et al. 2006) and B. juncea (AABB = 36) (Dong et al. 2018; Huangfu et al. 2011; Tsuda et al. 2012) via either wind or insects. Spontaneous hybridization, which is the first step for transgene introgression from transgenic crops to wild relatives, is mainly affected by sexual compatibility (Chevre et al. 1997). Sexual compatibility is determined by genetic backgrounds, which is related to the homology of genome, including ploidy; i.e. the higher the homology of the chromosome, the higher the seed-setting rate of the hybrids (Chèvre et al. 2008; Ellstrand et al. 1999).

Besides the initial crop-wild or weedy hybridization, fitness of the subsequent backcross progeny is also an important determinant of introgression likelihood (Liu et al. 2013a, 2013b). Fitness is defined as the reproductive success or the proportion of genes an individual contributes to the gene pool of a population, and it is measured by vegetative and reproductive growth variables; it determines whether the hybrids can survive and produce progeny (Warwick et al. 2009). Although the sexual fertility of the F₁ hybrid between B. napus and wild B. juncea carrying the transgene was low, the fitness-associated traits and composite fitness were restored with the self-pollination of the BC₁ generation. The composite fitness of the glyphosate-resistant second-generation progenies from B. napus was similar to the fitness of wild B. juncea in greenhouse (Song et al. 2021). BC₁ and BC₂ between F₁ (transgenic glyphosate- or glufosinate-resistant oilseed rape as male parent) and wild B. juncea produced fewer seeds per plant than the wild B. juncea, BC₃ and BC₄ produced a number of seeds per silique that was similar to that of self-pollinated wild B. juncea in greenhouse (Song et al. 2010).

Previous research has demonstrated that planting density, planting patterns and environmental factors can have a significant effect on the fitness between transgenic oilseed rape and wild or weedy relatives (Hauser et al. 2003; Johannessen et al. 2006). However, no studies have examined the fitness of the backcross generations under different conditions in the field. In this paper, we studied the fitness of the first backcross generations from the second to the sixth progenies (BC₁F₂R–BC₁F₆R) at low density (5 plants/m²) and high density (10 plants/m²), and in monocultures and mixed planting (wild B. juncea:BC₁F₂R–BC₁F₆R = 1:1). It was found that, compared with wild B. juncea in monocultures, BC₁F₄R–BC₁F₆R displayed similar fitness at low density, while BC₁F₂R–BC₁F₃R displayed lower fitness. In mixed planting, the fitness of the first backcross generations from the second to the sixth progenies all displayed lower fitness.

Song et al. recently documented that BC₁F₁R–BC₁F₄R between transgenic B. napus and wild B. juncea were mixoploids, with differing chromosome constitution in different cells (Song et al. 2021). Although mixoploidy results in the cytological instability of the BC₁ progeny from the transgenic oilseed rape and wild B. juncea, the progeny progressively displayed increased fecundity and fitness, with the narrowing of the chromosome variation across the self-pollinating generations of the BC₁ (BC₁F₄R < BC₁F₃R < BC₁F₂R < BC₁F₁R). In this study, compared with wild B. juncea, the second- to third-generation progeny of BC₁ between B. napus and wild B. juncea were less fit at low density in monocultures because of their reduced silique number, seed number and seed weight. However, the fourth- to sixth-generation progeny of BC₁ had similar fitness compared with wild B. juncea at low density in monocultures due to the stable chromosome numbers (Song et al. 2021). This body of research suggests that both environmental growing conditions and the chromosome constitution of the backcross generations can have effects on their fitness (Lu and Kato 2001; Song et al. 2021).

Planting density has a major effect on the competitiveness of plants (Hauser et al. 2003; Johannessen et al. 2006). The seed-setting rate of the BC₁ and BC₂ generation between B.napus and B. rapa was decreased at high density compared with at low density (Hauser et al. 2003). A previous study found that the silique number, aboveground dry biomass and seed number of the F₁ (B. napus × B. rapa) decreased when grown at 100 plants/m² compared with those grown at 16 plants/m² and 45 plants/m² (Johannessen et al. 2006). The current results showed that the fitness of the first backcross generations from the second to the sixth progenies were less fit at high density in monocultures compared with wild B. juncea. Correlation analysis also corroborated that planting density significantly correlated with the fitness components, particularly with seed weight, silique number and seed number.

Hovick et al. discovered that hybrid populations between wild radish and cultivated radish (Raphanus sativus) produced at least three times more seeds than did wild radish populations, a distinction that was driven by greater hybrid seedling emergence, earlier hybrid emergence, and more hybrid seedlings surviving to flower, rather than by greater individual fecundity in competitive environments (Hovick et al. 2012). Thus, competition has an effect on fitness. In this study, the first backcross generations from the second to the sixth progenies in mixed planting were less fit than wild B. juncea at both low and high densities. Variable chromosome numbers and low hereditary stability of genome composition were also responsible for this result (Song et al. 2021). Chromosome composition of wild B. juncea is 2n = 36 = AABB which is stable, while the first backcross generation between F₁ hybrids and wild B. juncea have a varying chromosome number of 20(AA) + 8(B) + 0–8(B) + 0–9(C) due to ploidy imbalance. Moreover, abnormal meiosis was observed in the HR fourth-generation progenies of BC₁, indicating unstable inheritance. Thus, wild B. juncea suppress the vegetative and reproductive growth of the backcross generations, and the progenies of BC₁ in mixed planting displayed lower fitness compared with wild B. juncea.

However, the first backcross generations from the second to the sixth progenies could produce nearly 9000–16 000 seeds. Because of their strong shattering and dormancy the seeds of oilseed rape can form a persistent seedbank and become a serious volunteer in succeeding crops (Gruber et al. 2009; Soltani et al. 2019). Therefore, ecological risk of transgene introgression should not be ignored.

Funding

This research was supported by the National Natural Science Foundation of China (32071656).

Acknowledgements

The authors would like to express their deep appreciation to Prof. Bernal E. Valverde and Prof. Kenneth M. Olsen for their useful discussion and comments, and for revising the language and style of the manuscript.

Conflict of interest statement. The authors declare that they have no conflict of interest.

REFERENCES