Differential activity of CYTOKININ OXIDASE/DEHYDROGENASE3 (CKX3) and CKX5 genes in regulating yield components in Brassica napus L
Ireen Schwarz, Christian Möllers, Thomas Schmülling
TL;DR
This study shows how specific genes in oilseed rape affect flower and seed production, which could help improve crop yields.
Contribution
The study reveals the distinct roles of BnCKX3 and BnCKX5 gene combinations in regulating yield traits in oilseed rape.
Findings
BnCKX5 plays a strong role in regulating inflorescence meristem activity and ovule number.
Specific BnCKX3 gene mutant combinations support yield component regulation.
Mutant plants maintain similar oil and fatty acid content as wild-type plants.
Abstract
Cytokinin is a plant hormone that regulates several yield-related traits in plants. Previously, it was demonstrated that in tetraploid oilseed rape (Brassica napus L.), mutation of all four cytokinin-degrading BnCKX3 and both BnCKX5 genes resulted in increased cytokinin concentration, larger and more active inflorescence meristems, and a higher number of ovules per gynoecium. This resulted in the formation of more flowers and pods on the main stem, thereby increasing seed yield from the main stem of the plants. Here, we investigated the relative contributions of distinct combinations of BnCKX3 and BnCKX5 genes of the A and C genomes to these yield components. Our analysis revealed an unexpectedly strong role for BnCKX5 in regulating these traits and identified distinct supportive BnCKX3 gene mutant combinations. These findings facilitate the selection of relevant alleles for breeding.…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6- —Freie Universität Berlin (1008)
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsPlant Molecular Biology Research · Lipid metabolism and biosynthesis · Nitrogen and Sulfur Effects on Brassica
Introduction
Increasing seed yield is an important breeding goal for many crop plants. Seed yield is a complex trait controlled by numerous genes (Ding et al. 2012; Siles et al. 2021). To achieve a steady increase in yield, both source and sink capacity must be improved (Körner 2015; Sonnewald and Fernie 2018). Studies of plants with enhanced sink strength in shoots or roots strongly support the view that plants are often not source-limited and that enhanced sink strength may be reached by altering organ growth or number without negative tradeoffs (Ashikari et al. 2005; Werner et al. 2010; Bartrina et al. 2011, 2017; Ramireddy et al. 2018; Schwarz et al. 2020).
In seed-bearing plants, enhanced sink strength and increased seed yield can be achieved by increasing the size and activity of reproductive meristems to form more flowers and more ovules, which together cause the formation of more seeds. A further option is to increase metabolic sink strength, leading to the formation of more storage compounds. One known factor to regulate the size and activity of reproductive tissues, as well as metabolic sink strength in both mono- and dicotyledonous plants, is the hormone cytokinin (CK) (Jameson and Song 2016, 2020; Kieber and Schaller 2018).
The CK content or signaling can be modified locally in a targeted fashion to alter plant growth (Werner et al. 2010, 2021; Ramireddy et al. 2018; Zeng et al. 2022). One important tool to alter the cytokinin content in plants to modify yield-related traits are genes encoding cytokinin oxidases/dehydrogenases (CKX), which are enzymes catalyzing the breakdown of CK (Schmülling et al. 2003; Werner et al. 2006). The first evidence for the role of CKX genes in regulating yield came from the discovery that the Grain number1 (Gn1) locus, which causes an increase in grain number in rice, encodes OsCKX2 (Ashikari et al. 2005).
CKX genes also play a role in regulating yield components in dicots, as was shown for Arabidopsis thaliana (Bartrina et al. 2011) and Brassica napus L. (oilseed rape) (Schwarz et al. 2020). In Arabidopsis, it was found that the combined mutation of the AtCKX3 gene, which is the Arabidopsis orthologue of the OsCKX2 gene, but whose mutation alone was ineffective, and AtCKX5, alters several yield-related traits, namely flower number, silique number, distance of ovules, and seed density in siliques. The sum of all changes caused a significant increase in seed yield of more than 50% (Bartrina et al. 2011). Subsequently, it was shown that mutation of the four BnCKX3 and two BnCKX5 genes of the tetraploid B. napus, which is the closest relative of Arabidopsis among the crop plants and globally the third most important source of vegetable oil (Weselake et al. 2024), enhances several of these yield components as well (Schwarz et al. 2020), consistent with changes caused by exogenous treatment with CK (Zuñiga-Mayo et al. 2018).
A comparison of the mutant phenotypes shows that CKX3 and CKX5 collectively control similar yield components to a different extent in Arabidopsis and oilseed rape (Bartrina et al. 2011; Schwarz et al. 2020). In oilseed rape, the size and activity of the inflorescence meristem of the Bnckx3,5 mutants are increased, and more flowers and pods are formed. Bnckx3,5 mutants have larger gynoecia, and the distance of ovules is reduced, although to a lesser extent in oilseed rape than in Arabidopsis. In contrast to Arabidopsis, where the formation of larger siliques containing more densely packed seeds contributes significantly to the increase in seed yield, in B. napus, a portion of the higher number of ovules per pod does not develop into seeds. Therefore, the seed number per pod is not increased, thus limiting the increase in seed yield in B. napus, which was in the range of 20–32% on the main stem of Bnckx3,5 mutants (Schwarz et al. 2020). The oil content of seeds from these lines was not reported, although it is a crucial factor for productivity, and high oil content is a major breeding goal for oilseed rape. Oil from oilseed rape contains a healthy lipid profile with a high proportion of mono- and polyunsaturated fatty acids (Weselake et al. 2024). Also, the content of glucosinolates, which are secondary metabolites protecting the plants from insect pests and pathogens, is as yet unknown in seeds of Bnckx3,5 mutants. Because high glucosinolate concentrations can be harmful for humans and livestock, breeding aims to produce varieties with a low glucosinolate content (Miao et al. 2021).
B. napus is an allotetraploid plant originating from natural hybridization between B. rapa and B. oleracea, which contribute the A and C subgenomes (Chalhoub et al. 2014; Gu et al. 2024). Preceding genome triplications in the diploid ancestors and chromosomal rearrangements have led to a complex genome structure with different gene numbers. For example, there are 23 BnCKX genes belonging to seven BnCKX gene families (Liu et al. 2013; Schwarz et al. 2020). Among these, there are four BnCKX3 and two BnCKX5 genes. The redundancy due to the complex evolutionary genome history and polyploidy makes it challenging to study gene function and its potential relevance to improve agronomically relevant traits.
With the combined mutation of four BnCKX3 and two BnCKX5 genes, the maximum number of potentially relevant BnCKX3/5 genes from both subgenomes had been mutated. To follow this number of genes in practical breeding is challenging, even if it is assisted by molecular markers. It would therefore be desirable to lower the number of mutated BnCKX genes required to obtain an enhancement of these yield components. It could indeed be that not all the mutated genes are required to obtain the observed effect. For example, a partial subfunctionalization or neofunctionalization may have occurred for the different BnCKX3 genes, as is indicated by their differential expression pattern (Schwarz et al. 2020). Therefore, the regulation of different yield components may be exerted (partly) by different BnCKX3 alleles of the A and C genomes. It could also be that a complete loss of gene function is not optimal, as some CKX activity might be necessary during certain developmental stages of reproductive tissues (Di Marzo et al. 2020). Furthermore, the change of CK concentration achieved by reducing its breakdown is relevant as supraoptimal concentrations of CK may have negative effects (Ferreira and Kieber 2005). Taken together, the mutation of all BnCKX3,5 genes may not result in optimal changes, at least not for all yield traits. Finally, given their differential expression and contribution of homoeologous genes to phenotypic traits, it would be interesting to explore which BnCKX3 and BnCKX5 gene copies of the A or C genomes contribute most to the phenotypic changes in the Bnckx3,5 mutant (Chalhoub et al. 2014; Wu et al. 2018).
To explore the functional relevance of gene dosage and specific combinations of BnCKX3 and BnCKX5 mutant alleles in regulating yield components, we have analyzed a variety of Bnckx3,5 mutant plants. In each of these, half of the genes of BnCKX3 and BnCKX5 are mutated in different combinations in the background of the full set of mutated alleles of the respective other gene. Thus, these eight different genotypes harbour only one mutant allele of the two BnCKX5 genes from the A and C genomes, or different combinations of two of the four different BnCKX3 genes of the A and C genomes, always in a background of a fully mutated set of BnCKX5 or BnCKX3 genes. The results show that mutation of both BnCKX5 alleles is necessary to obtain full expressivity of the mutant phenotypes. In contrast, two of the four Bnckx3 mutant alleles, preferentially those of the A genome, are sufficient to cause the formation of significantly more flowers and ovules. The results identify the functionally important alleles to enhance yield components in oilseed rape with a lower number of mutant alleles than known previously, thus facilitating their use in breeding.
Material and methods
Plant material
An elite spring oilseed rape breeding line of Brassica napus L. was used as the wild type (WT) (Schwarz et al. 2020). Mutants carrying point mutations in the four BnCKX3 and two BnCKX5 genes were originally identified by TILLING (McCallum et al. 2000) in a mutant population generated by treatment with ethyl methanesulfonate (EMS) (Lammertyn et al. 2017). These individual mutants were backcrossed four times to wild type before their combination by crossing to generate the quadruple Bnckx3 mutant (a1/a2/c1/c2), the double Bnckx5 mutant (a1/c1), and the sextuple Bnckx3,5 mutant (a1/a2/c1/c2; a1/c1) (Schwarz et al. 2020).
Plants containing different combinations of Bnckx3 and Bnckx5 mutant alleles (genotypes G1, G2, and G7 to G12 according to Table 1) were obtained by crossing Bnckx3,5 to the wild type. The F1 generation was backcrossed to Bnckx3,5. 250 BC1 plants were genotyped using kompetitive allele-specific PCR (KASP) (Majeed et al. 2019) to identify the 64 genotypes containing all possible combinations of homozygote or heterozygote Bnckx3 and Bnckx5 mutant alleles from the A and C genomes. In the next generation, obtained by selfing, ~ 1300 BC1S1 plants of the selected genotypes G1, G2, and G7 to G12, shown in Table 1, were again genotyped to eliminate plants containing heterozygous alleles. From the selected plants, the BC1S2 was obtained and used for all further experiments. Genotyping was carried out by BASF (Gent, Belgium) (BC1) and by VHLGenetics (Wageningen, Netherlands) (BC1S1). Table 1. Different Bnckx3,5 mutant genotypes of B. napus. B. napus genotypes analyzed in this study harboring different combinations of mutated *BnCKX3 *and BnCKX5 genes of the A and C genomes are shown. Genes written in capitals and marked in light grey are wild-type, whereas small letters marked in darkgrey indicate mutant alleles
Plant growth conditions
Plants were grown in the greenhouse in pots with a diameter of 13 cm filled with 1.5 L of soil (Stender D400 from Stender AG, Schermbeck, Germany) and fertilised with Osmocote START at 19–22 °C under long-day conditions (16 h light/8 h dark) and watered every day. Before planting in soil, seeds were surface-sterilized with 0.12% H_2_O_2_ + 0.01% Triton X-100 for 10 min, washed several times with water, and germinated in vitro on wet filter paper under long-day conditions. After one week, seedlings were transferred to soil and grown in 6 cm pots for three weeks, then in 9 cm pots for an additional 1–2 weeks, and finally transferred into 13 cm pots before the reproductive phase started.
Plants grown in the field were germinated under standard greenhouse conditions and transferred after four weeks to a field trial area located at FU Berlin, Germany. Plants were grown in 12 plots of 1.2 m × 1.2 m containing one plant of each of the 12 genotypes planted in a mixed arrangement. Seeds from these plants were used to determine the content of oil and protein in seeds, the percentage of oleic acid (18:1) and linolenic acid (18:3), as well as of glucosinolate content.
Measurement of morphological yield components
To determine the inflorescence meristem activity, the number of flowers formed between day 2 and day 23 after the beginning of flowering was counted. The beginning of flowering was defined as the opening of the first flower. The number of developed pods was counted at the end of the reproductive growth period. The analysis was mostly limited to the main stem as it shows the most stable growth behaviour and is the principal source of seed yield (Diepenbrock 2000; Siles et al. 2021).
To determine the number of ovules per gynoecium, the number of developed seeds per pod, and empty positions were counted at the end of pod development using a stereomicroscope. Seeds from both halves of the pod were counted in three pods obtained from the main stem of 5–6 plants.
Analysis of seed oil content
Seed quality parameters were determined by NIRS using the FOSS monochromator model 6500 (NIRSystem Inc., Silver Spring, MD, USA). The spectra of 2 g of seeds in small ring cups were recorded from 400 to 2498 nm at 2-nm intervals. WinISI software (Version 4.12.0.15440, FOSS NIR Systems Inc., USA) was used to analyze the spectra of the harvested seeds. Calibration 'raps2024.eqa' provided by VDLUFA Qualitätssicherung NIRS GmbH (http://www.vdlufa-nirs.de) was used to determine the seed content of oil (%), protein (%) and glucosinolate (µmol/g seed). Seed oil, protein, and glucosinolate contents were determined on a 91% seed dry matter basis (Holzenkamp and Möllers 2024). The calibration ‘raps2024.eqa’ also allowed for determining the oleic acid and linolenic acid content of the seed oil as described (Velasco and Becker 1998; Velasco et al. 1998, 1999).
Accession numbers
The accession numbers of the B. napus CKX3 genes are BnaA10g28940D (CKX3_A1), BnaA02g08420D (CKX3_A2), BnaC09g33450D (CKX3_C1), and BnaCnng41060D (CKX3_C2). The accession numbers of the B. napus CKX5 genes are BnaAnng09190D (CKX5_A1) and BnaC06g36220D (CKX5_C1).
Statistical analysis
Data were statistically evaluated using GraphPad Prism, version 8 (GraphPad Software, La Jolla, CA, United States). Statistical tests used are indicated in the figure and table legends. A p-value < 0.05 was considered to indicate a significant difference.
Results
Generation of different Bnckx3,5 mutant genotypes
Mutation of either all four BnCKX3 genes or both BnCKX5 genes alone did not have any or only slight phenotypic consequences on yield-related traits, while the mutation of all six genes caused distinct and strong phenotypic changes (Schwarz et al. 2020). To study the functional relevance of BnCKX3 and BnCKX5 genes from the A and C genomes and their combinations, we constructed lines harbouring different numbers and combinations of mutated genes. To obtain these lines, Bnckx3,5 was first crossed to the wild type. The resulting F1 generation was backcrossed to Bnckx3,5, and the BC1 plants were genotyped to identify the 64 possible different Bnckx3 Bnckx5 mutant combinations. Selected genotypes were propagated, self-fertilized, and the resulting BC1S1 again genotyped to eliminate plants containing heterozygous alleles. BC1S2 plants homozygous for the selected individual mutant alleles were used for all further experiments.
Table 1 shows the different Bnckx3 Bnckx5 mutant allele combinations analyzed in this study. These lines harbour different mutant alleles of BnCKX3 or BnCKX5 in the background of the full set of mutated alleles of the respective other gene. Two of the lines (genotypes G1, G2) are homozygous for all four Bnckx3 mutant alleles and contain, in addition, a single Bnckx5_a1 or Bnckx5_c1 mutant allele. Six other lines (G7 to G12) contain both Bnckx5 mutant alleles and all six possible combinations of the additional two Bnckx3 mutant alleles. Thus, each of the eight lines contains a full set of mutated Bnckx3 or Bnckx5 genes plus half the gene dosage of the respective other genes in different allele combinations. During the vegetative phase, growth and development of all mutants, including the Bnckx3,5 sextuple mutant, were comparable to wild type (see also Schwarz et al. 2020). The focus of this work has been on changes during the reproductive phase as outlined below.
Activity of the inflorescence meristem
First, we studied the activity of the inflorescence meristem by counting the number of flowers formed between days 2 and 23 after the beginning of flowering as the readout. In three independent greenhouse trials, WT plants formed during that time span about 80 to 90 flowers on the main stem, while Bnckx3,5 mutants formed between 100 and 120 flowers, i.e., on average one third more than the WT (Figs. 1, S1; see also Schwarz et al. 2020). In comparison, plants harboring mutations in only the BnCKX3 or BnCKX5 genes showed only a slight increase of 6–7% in the number of flowers at the main stem compared with their wild-type segregants (Schwarz et al. 2020).Fig. 1. Inflorescence meristem activity*.* The number of flowers formed between 2 and 23 days after the opening of the first flower at the main stem of plants grown in a greenhouse. Data shown are mean values ± SD, n = 6. The statistical significance of differences compared to WT was calculated using one-way ANOVA (Kruskal–Wallis test). * p ≤ 0.05, ** p ≤ 0.01
Mutation of either one of the two BnCKX5 gene copies in a Bnckx3 mutant background (G1, G2) almost completely abolished the enhanced flower formation found in Bnckx3,5, revealing the functional importance of BnCKX5 for this trait. This was a reproducible result in all three independent trials (Figs. 1, S1).
In the background of the Bnckx5 mutant, the contribution of distinct Bnckx3 mutant alleles became apparent. Among the Bnckx3 mutant gene combinations, the Bnckx3a1a2 mutant combination (G7) showed in all three trials a significant increase (24% to 35%) in flower number compared to WT (Fig. 1, S1). The increase was in all three trials close to or even similar to the one obtained with the Bnckx3,5 mutant. In contrast, the rate of flower formation in plants carrying mutations in both BnCKX3 gene copies from the C genome (G8) was in all three trials not statistically significantly different from WT. Mutants carrying combinations of mutated BnCKX3 from both genomes (G9 to G12) mostly also showed a tendency to a higher frequency of flower formation than WT. The increase in the rate of flower formation of up to 29% in comparison to WT was statistically significant in one of the three trials for each of the genotypes G9, G10, and G11 (Fig. 1, S1). Taken together, this comparison highlights the functional importance of the BnCKX3 gene copies from the A genome as compared to those of the C genome. The rate of flower formation in Bnckx3a1a2 ckx5 (G7) mutants carrying four mutant alleles could be as high as the rate of flower formation in the sextuple Bnckx3,5 mutant.
Ovule number
The number of ovules formed per gynoecium, calculated as the total number of aborted and developed seeds, was another yield-related trait that was increased in Bnckx3,5 mutants. The increase in ovule number was due to a reduced distance between individual ovules that had formed on the placenta and longer gynoecia (Schwarz et al. 2020). In four independent trials, Bnckx3,5 mutants showed an increase in the ovule number per gynoecium between 16 and 32% (Figs. 2, S2), which is consistent with previous data (Schwarz et al. 2020). Of the Bnckx3 and Bnckx5 mutants, which were not analysed by Schwarz et al. (2020), only Bnckx5 showed an increased ovule number of 21% (Fig. 2a), which was almost in the range of the increase of 31% shown by Bnckx3,5. In this trial, Bnckx5 mutants had, on average, 44 ovules per gynoecium, while WT had 36, and Bnckx3,5 had 48. In contrast, the ovule number in gynoecia of the Bnckx3 mutant was not significantly different from WT (Fig. 2a).Fig. 2. Number of ovules per gynoecium. a The number of ovules in wild type (WT) and the mutant lines Bnckx3, Bnckx5, and Bnckx3,5.b The number of ovules in genotypes with different combinations of Bnckx3 and Bnckx5 mutant alleles. Plants were grown in a greenhouse. Data shown are mean values ± SD, n = 15–18. The statistical significance of differences compared to WT was calculated using one-way ANOVA (Tukey´s test) in (A) and one-way ANOVA (Kruskal–Wallis test) in (B). * p ≤ 0.05, ** p ≤ 0.01
Mutation of only one of the two BnCKX5 gene copies in a Bnckx3 mutant background reduced the number of ovules per gynoecium to WT level, highlighting the relevance of BnCKX5 to control the ovule number in the placenta tissue (Figs. 2b, S2). All mutants carrying two mutated BnCKX3 gene copies in the Bnckx5 mutant background (G7 to G12) showed an increase in ovule number per gynoecium, which is consistent with an important role for BnCKX5. The highest increase was shown by the genotypes G7 and G10, which harbour the a1a2 and a2c2 mutant combinations. They showed in all three independent trials an increase of ovule number between 17 and 28%, which is in the same range as the increase recorded for the Bnckx3,5 mutant (Figs. 2b, S2). Also, the increase in ovule number of G12 (a2c1) was in all three trials highly significant, around 15%. In comparison, G8 (c1c2), G9 (a1c1), and G11 (a1c2) showed a lower increase in ovule number (Figs. 2b, S2). These results indicate a relatively higher importance of the BnCKX3-A2 allele, which might be combined with one of the other three mutated BnCKX3 alleles and mutated BnCKX5 to achieve the highest increase in ovule number per gynoecium.
Seeds per pod
The increased number of ovules per gynoecium may eventually result in an increased number of seeds per pod. However, in the Bnckx3,5 mutant, about one-third more ovules per gynoecium did not lead to a higher number of seeds per pod due to increased seed abortion (Schwarz et al. 2020). Also, in the four trials shown here, Bnckx3,5 has either a similar or a lower number of seeds per pod compared to WT (Fig. 3, S3). Also, Bnckx3 and Bnckx5 mutants had a similar number of seeds per pod like WT, with a relatively large variability in all genotypes (Fig. 3a). The comparison of the eight mixed genotypes yielded a similar result. On average, the number of seeds per pod was comparable in all lines, including WT, and again the mutant genotypes showed a relatively large variability (Fig. 3b). In summary, this confirms the result of Schwarz et al. (2020) that not all of the additional ovules found in the mutants develop into mature seeds.Fig. 3. Number of seeds per pod. a The number of seeds per pod in wild type (WT) and the mutant lines Bnckx3, Bnckx5, and Bnckx3,5. b The number of seeds per pod in different genotypes carrying various combinations of Bnckx3 and Bnckx5 mutant alleles. Plants were grown in a greenhouse. Data shown are mean values ± SD, n = 15–18. The statistical significance of differences compared to WT was calculated using one-way ANOVA (Tukey´s test). None of the mutant genotypes shows a difference from WT with p ≤ 0.05
Seed oil and protein content
Important yield components that have not yet been analysed in Bnckx3,5 mutants are the oil content of the seeds and the proportion of unsaturated fatty acids. Cytokinin is a factor regulating sink strength of tissues (Guivarc´h et al. 2002; Werner et al. 2008), raising the possibility that the oil content might change if CK metabolism genes are mutated. The analysis of the seeds of plants grown in the field by near-infrared spectroscopy (NIRS) detected an oil content between 37 and 40% and a protein content of 21–23% (Fig. 4).Fig. 4. Oil and protein content of seeds. a Percent of oil content in seeds. b Percent of protein content in seeds at 9% moisture content. Seeds were harvested from plants grown in the field, and data were recorded using NIRS. Data shown are mean values ± SD, n = 12. The statistical significance of differences compared to WT was calculated using one-way ANOVA (Tukey´s test). None of the mutant genotypes shows a difference from WT with p ≤ 0.05
Generally, seeds having a higher oil content had a lower protein content. In this trial, seeds from Bnckx3,5 contained 6% more oil than seeds from WT, i.e., 40% compared to 37%. The percentage of unsaturated fatty acids was around 60% for oleic acid (18:1) and around 12% for linoleic acid (18:3) for all genotypes (Fig. 5). The content of glucosinolate was between 16 and 20 µmol per g seed for the different genotypes, with a significant increase compared to WT only in the Bnckx5 mutant (Fig. S4).Fig. 5. Content of oleic acid and linolenic acid in seed oil. a Percent of oleic acid (18:1). b Percent of linolenic acid (18:3). Seeds were harvested from plants grown in the field, and data were recorded using NIRS. Data are given as percentages of total fatty acids and are mean values ± SD, n = 12. The statistical significance of differences compared to WT was calculated using one-way ANOVA (Tukey´s test). * p ≤ 0.05
The oil and protein content of seeds from plants of the same genotypes grown in a greenhouse was about 44% and 19%, respectively, with only small differences between the genotypes and irrespective of whether the seeds were harvested from the main stem or lateral stems (Fig. S5). The proportion of oleic acid was between ca. 58–62% in seeds from all genotypes and comparable for main stem and lateral branches. Seeds derived from Bnckx3,5 and G10 showed a significant increase in the proportion of oleic acid compared to WT from both main stem and lateral branches, while this was limited to the main stem for other genotypes with mutated BnCKX3 alleles (Fig. S6a,b). In this trial, the proportion of linolenic acid was lowest in the genotypes with the highest content of oleic acid, but the differences from WT were mostly not significant (Fig. S6c–d). The concentration of glucosinolate tended to be increased in several mutant genotypes, but also these changes were mostly not significant (Fig. S6e–f). It is noteworthy that the glucosinolate content of plants grown in the greenhouse was considerably lower than that of plants grown in the field (Fig. S4), highlighting the impact of growth conditions and underpinning the need for further field trials.
Discussion
The genetic analysis of the roles of individual BnCKX3 and BnCKX5 genes of the tetraploid B. napus in regulating yield components has extended earlier work (Schwarz et al. 2020) and revealed distinct contributions of these genes to two important yield components, the activity of the inflorescence meristem and the number of ovules. Not all BnCKX gene copies contributed equally to these traits; instead, a nuanced functional redundancy was found. A striking result has been the unexpectedly strong role of BnCKX5 in regulating yield components. For BnCKX3, genetic analysis revealed a stronger contribution of genes from the A genome compared to the C genome of B. napus. Finally, increasing the activity of the inflorescence meristem and the number of ovules did not negatively influence seed oil content and fatty acid profile.
A single BnCKX5 allele is sufficient to maintain the activity of the inflorescence meristem at the WT level, even if, in addition, all BnCKX3 alleles are mutated (Fig. 1). Further mutation of the second BnCKX5 allele caused a strong increase in the frequency of flower formation, illustrating the relevance of BnCKX5. Moreover, the sole mutation of BnCKX5 is sufficient to cause a significant increase in the number of ovules per gynoecium (Fig. 2), further underpinning the important role of this gene. Previously, BnCKX3 had been suspected to be functionally more relevant as it is the closest homologue to both the yield genes OsCKX2 of rice (Ashikari et al. 2005) and AtCKX3 of Arabidopsis (Bartrina et al. 2011; Schwarz et al. 2020). For the four BnCKX3 alleles, genetic analysis in the background of mutated BnCKX5 has demonstrated a stronger role for the two gene copies of the A genome in regulating inflorescence meristem activity, as well as for the BnCKX3-A2 allele in regulating ovule number. It would be interesting to study the consequences of individual BnCKX3 gene mutations on the expression of the other BnCKX3 genes, in particular on the respective other BnCKX3-A and BnCKX3-C alleles, characterized by distinct expression profiles (Schwarz et al. 2020). The different relative functional importance of CKX genes in Arabidopsis and B. napus highlights that this can differ between species and even between a model plant and the closely related crop plant, underpinning the assessment that Arabidopsis is not an infallible model for gene expression in the Brassicaceae (Song et al. 2015).
What could be the reason(s) for BnCKX5 being the major regulator of CK activity in the inflorescence meristem and the placenta? RNA-seq analysis revealed that all six BnCKX gene copies are expressed at a low level in vegetative tissues and exhibit distinct expression patterns in different reproductive tissues. However, a conspicuous, generally stronger expression of BnCKX5 as compared to BnCKX3 was not found (Schwarz et al. 2020). In situ hybridization detected expression of BnCKX3 and BnCKX5 in the organizing centers of the inflorescence and floral meristems, with BnCKX5 covering larger domains (Schwarz et al. 2020). This is relevant as the regulation of meristem activity by cytokinin is exerted through the WUS-CLV circuit (Leibfried et al. 2005; Gordon et al. 2009; Bartrina et al. 2011). The larger expression area of BnCKX5 covers most of the CLV1 expression domain. Lack of BnCKX5 activity, resulting in a higher concentration of cytokinin, could lead to a lower expression of the CLV1 gene, which is negatively regulated by CK and functions to restrict the WUS domain, thus limiting meristem activity and size (Brand et al. 2000; Schoof et al. 2000; Lindsay et al. 2006; Gordon et al. 2009).
Furthermore, in situ hybridization has shown that BnCKX5 is expressed in ovules and placenta of young gynoecia, where no expression of BnCKX3 was detected (Schwarz et al. 2020). This is consistent with the main role of BnCKX5 in regulating the activity of meristematic cells in the placenta, thus influencing ovule primordia formation. Cytokinin acts here as a positional cue regulating the distance between ovule primordia on the placental tissue. We hypothesize that loss of BnCKX5 activity alters a cytokinin gradient, such that upon mutation of this gene, the cytokinin threshold concentration required for primordia formation is reached at a closer distance. Taken together, it is likely that the functional dominance of BnCKX5 is at least partially due to the differences in gene expression, at least in the case of ovule formation.
Additional reasons for the dominance of BnCKX5 could lie in differences in subcellular localisation of CKX proteins (Werner et al. 2003; Köllmer et al. 2014; Niemann et al. 2015; Liu et al. 2018), providing access to different cytokinin pools, and differences in substrate specificity (Galuszka et al. 2007; Kowalska et al. 2010). Interestingly, BnCKX5-A1 is predicted to be transported into the endoplasmic reticulum (ER) like the Arabidopsis orthologue, which may be expelled to the apoplast (Niemann et al. 2015). In contrast, BnCKX5-C1 is predicted to be a soluble cytoplasmic protein (Ødum et al. 2024). This difference in BnCKX5 proteins encoded by the two Brassica subgenomes requires experimental verification. Different subcellular localization of the proteins may provide access to different intra- and extracellular cytokinin pools feeding into distinct cytokinin signaling pathways (Romanov et al. 2018). Enhanced CK signaling through different pathways could be the basis for a stronger role of BnCKX5 in oilseed rape as compared to Arabidopsis.
The expression of all four BnCKX3 genes from the A and C genomes is mostly very similar and strongest in reproductive tissues (Chao et al. 2020; Schwarz et al. 2020). BnCKX3-A1/C1 are most strongly expressed in open flowers, while BnCKX3-A2/C2 expression is also found in young flowers, developing seeds, and pods (Schwarz et al. 2020). The dominant role for BnCKX3-A2 in regulating ovule number could be due to its stronger expression in comparison to BnCKX3-C2, in young developing flower buds (Schwarz et al. 2020), where most of the ovules are formed (Qadir et al. 2021; Yu et al. 2022). Indeed, a large proportion of effective mutations are located in the cis-regulatory regions of genes, and even subtle changes in gene dosage can affect yield traits significantly (Eshed and Lippman 2019). A different functional contribution of genes from the A and C genomes, as we found for BnCKX3, has also been reported for other homoeologous Brassica genes showing functional divergence (Chalhoub et al. 2014; Wu et al. 2018; Gu et al. 2024).
All four BnCKX3 proteins contain an N-terminal sequence predicting their import to the ER like their Arabidopsis counterparts (Schmülling et al. 2003; Werner et al. 2006; Niemann et al. 2015). Sequence analysis predicted with the highest probability that these are soluble ER proteins (Ødum et al. 2024). BnCKX3-A1/C1 also have a high score for membrane association, again similar to Arabidopsis CKX3 (Niemann et al. 2015), which is not the case for BnCKX3-A2/C2. Notably, CKX proteins show a dynamic glycosylation pattern, which may impact their subcellular localization (Motyka et al. 1996), enzymatic activity, or substrate preferences. Taken together, different possible factors may contribute to the differential activity of BnCKX3 proteins encoded by the A and C genomes.
Importantly, we have found that the seed oil content as well as the composition and concentrations of unsaturated fatty acids in seeds from BnCKX3,5 mutant plants were, in most cases, comparable to WT. There were, generally, also no major changes in the content of glucosinolates. This addresses the concern that yield gains might come at the expense of seed quality. The maintenance of oil content and unsaturated fatty acid levels indicates that enhancement of yield components is not necessarily offset by a reduction in seed quality, which is encouraging for breeding programs targeting both yield and quality (Weselake et al. 2024).
Our study has yielded information that is important for use in targeted breeding. Only a subset of BnCKX3/5 alleles needs to be mutated to enhance yield components. The reduction from six to four genes required to enhance yield components simplifies the path for marker-assisted or genome editing-based breeding approaches, particularly the introgression of the mutated loci in new germplasm. Targeting both BnCKX5 alleles is critical for maximizing inflorescence meristem activity and ovule number, while mutating in addition only a subset of BnCKX3 alleles is sufficient. Testing a Bnckx5 double mutant with only one additional mutated BnCKX3 gene of the A genome would be the logical next step. Of course, because of the limitations of greenhouse and small-scale field experiments (Khaipho-Burch et al. 2023; Flavell et al. 2025), it is necessary to test the performance of the different BnCKX3,5 mutants in more detail in field trials at multiple locations and under different environmental conditions, and to study whether enhancement of the studied traits translates into higher seed yield.
Finally, we would like to mention that there are additional options to achieve yield enhancement in oilseed rape using BnCKX and other cytokinin genes. Recently, BnCKX2 was shown to regulate seed size (Yan et al. 2023), and BnCKX1 was identified as a candidate gene affecting seed yield and weight (Pal et al. 2021). Earlier, the expression of BnCKX2 and BnCKX4 suggested their role in seed development (Song et al. 2015). Transgenic expression of cytokinin biosynthesis genes has led to a higher seed yield in oilseed rape (Roeckel et al. 1998; Kant et al. 2015). Together, this illustrates that there are numerous opportunities to use targeted manipulation of the cytokinin system to improve crop yield (Jameson and Song 2016, 2020). The enhancement of source strength could also be tested in Bnckx3,5 mutants if their overall yield is not increased due to source limitation.
Supplementary Information
Below is the link to the electronic supplementary material.Supplementary file1 (PDF 1206 kb)
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Kowalska M, Galuszka P, Frébortová J, Šebela M, Béres T, Hluska T, Šmehilová M, Bilyeu KD, Frébort I (2010) Vacuolar and cytosolic cytokinin dehydrogenases of Arabidopsis thaliana: heterologous expression, purification and properties. Phytochem 71:1970–1978. 10.1016/j.phytochem.2010.08.013
- 2Lammertyn F, Bots M, Laga B, Schmidt RC, Schmidt J, Mouchel C (2017) Brassica plants with altered properties in seed production. In. International Publication Number: WO 2017/064173
- 3Liu Z, Lv Y, Zhang M, Liu Y, Kong L, Zou M, Lu G, Cao J, Yu X (2013) Identification, expression, and comparative genomic analysis of the IPT and CKX gene families in Chinese cabbage (Brassica rapa ssp. pekinensis). BMC Genomics 14:594. 10.1186/1471-2164-14-594