Cytokinin oxidase/dehydrogenase family genes exhibit functional divergence and overlap in ricegrowth and development, especially in control of tillering

Cytokinins play key roles in plant growth and development; hence, cytokinin biosynthesis and degradation have been extensively studied. Cytokinin oxidase/dehydrogenases (CKXs) are a group of enzymes that regulate oxidative cleavage to maintain cytokinin homeostasis. In rice, 11 OsCKX genes have been identified to date; however, most of their functions remain unknown. Here, we comprehensively analyzed the expression patterns and functions of OsCKX genes. Using CRISPR/Cas9 technology, we constructed mutants of all OsCKX genes to determine the functions of OsCKXs in rice development. The results revealed that the single osckx and higher-order osckx4 osckx9 mutant lines showed functional overlap and subfunctionalization. Notably, the osckx1 osckx2 and osckx4 osckx9 double mutants displayed contrasting phenotypic changes in tiller number and panicle size compared to the wild type. Moreover, we identified several genes with significantly altered expression in osckx4 and osckx9 single and double mutant plants. Many differentially expressed genes were found to be associated with auxin and cytokinin pathways. Additionally, the cytokinins in osckx4 osckx9 mutants were increased compared to the wild type. Overall, our findings provide new insights into the functions of OsCKX genes in rice growth and may be used as a foundation for future studies aimed at improving rice yield. Key words: Cytokinin, expression pattern, OsCKX, panicle, phenotype, rice, tiller Highlight The osckx4 osckx9 double mutant had a significantly greater number of tillers, whereas the osckx1 osckx2 double mutant showed the opposite phenotypic change, compared to the wild type


8 9
The third clade was composed of OsCKX4, OsCKX5, and OsCKX9. OsCKX4, which is generally 1 9 0 strongly expressed in the vegetative organs, displayed extremely special expression in the roots.

9 1
OsCKX9 was generally expressed at low levels in all the tissues analyzed but showed special 1 9 2 expression in the leaf blade and whole axillary buds. OsCKX5, usually highly expressed in all the 1 9 3 checked tissues, also exhibited especial expression in the roots and leaves.

9 4
The fourth clade consisted of OsCKX3 and OsCKX8. OsCKX3, which is expected to have the 1 9 5 highest expression in the stem and young panicles, was found to have particular expression in the shoot 1 9 6 base and young panicle. OsCKX8, commonly expressed at lower levels than OsCKX3 in all vegetative 1 9 7 organs, was extraordinarily expressed in the shoot base, flag leaf primordia, and inflorescence.

9 8
OsCKX11 showed a different sequence from other OsCKX genes; it was generally expressed at a 1 9 9 higher level in all tissues, specifically in the reproductive-stage root and inflorescence, displaying 2 0 0 special expression in the root, shoot base, and young inflorescence.  osckx2, osckx8, and osckx11 mutants were significantly wider than those in the WT, while the osckx3, 2 2 5 osckx4, and osckx5 mutants had thinner basal internodes (Supplemental Figure S5, E and F).

8
The panicle number is determined by the tiller number per plant and has a significant effect on 2 2 9 rice yield. In the field and pot experiments, the osckx4 and osckx9 mutants developed more tillers than  Table S2). In contrast, the osckx2 mutant showed significantly reduced 2 3 4 tiller number in both the 2019 and 2020 field and pot experiments (Supplemental Tables S2 and S3).

3 5
However, no significant differences were observed in the tiller numbers of the other osckx mutants.

3 6
Additionally, there was no significant change in the yield per plant of the osckx mutants, except for 2 3 7 osckx1 and osckx11-1, which had significantly improved production. In contrast,

4 6
The osckx1 osckx2 double mutants have significantly reduced tiller numbers 2 4 7 OsCKX1 and OsCKX2 displayed similar trends in the phenotypes of flag leaf length and width, culm 2 4 8 diameter, and 1000-grain weight, and the two genes were grouped into the same clade. Based on this, ( Figure 5A). To verify the accuracy and reproducibility of the RNA-seq results, we randomly selected 3 1 2 seven previously studied genes associated with the phenotypes observed in this study for quantitative 3 1 3 reverse-transcription PCR analysis. The expression profiles of these genes were found to corroborate 3 1 4 the RNA-seq results (Supplemental Figure S12).

1 5
However, although there were numerous DEGs between the WT and mutants, there were 71 DEGs 3 1 6 in the roots and 90 DEGs in the shoot bases that overlapped between the osckx4 vs. WT and osckx9 vs.

1 7
WT comparisons and occupied only a small fraction of the total DEGs ( Figure 5B). These included

2 9
Since we observed that osckx4 osckx9 exhibited more severe phenotypes and possessed more 'biosynthesis of amino acids', and those related to carbon and nitrogen metabolism. Since most of the 3 3 4 DEGs were not functionally annotated, the relationship between the DEGs and the enriched pathways 3 3 5 and their influence on the growth and development of the mutant plants were not fully understood.

6
However, we identified several genes that may be associated with the phenotypes observed in this study.

7
OsCKX4 and OsCKX9 regulate cytokinin levels 3 3 8 As OsCKX4 and OsCKX9 are involved in biodegradation of cytokinins, we first analyzed the DEGs

4 2
Type-A OsRR genes are a group of genes that can be induced by cytokinins. However, the expression osckx4 osckx9 mutants ( Figure 6A). In contrast, in the axillary bud, the expression of detectable type-A

4 8
Although type-A OsRR genes can indicate the content of cytokinins indirectly, we sampled the roots 3 4 9 and shoot base of NIP and osckx4 osckx9 to detect the content of cytokinins. In the root, compared to 3 5 0 NIP, the content of cZ (cis-zeatin), iP, tZ, IPA (isopentenyl adenosine), and tZR (trans-zeatin riboside) 3 5 1 showed a significant increase, while 6-BA and DHZR (dihydrozeatin riboside) contents significantly 3 5 2 decreased in osckx4 osckx9 ( Figure 6C). However, in the shoot base, we found a higher content of iP 3 5 3 and lower content of tZR and DHZR in osckx4 osckx9 ( Figure 6D). Additionally, to explain the reason for reduced plant height of osckx4 osckx9, we identified genes 3 7 8 related to gibberellin metabolism, as gibberellins can regulate plant height (Sasaki et al., 2002). We 3 7 9 found that the expression of OsKS1 and OsKS2 was significantly downregulated in the roots of osckx4 functions (summarized in Figure 9). Loss of function of some OsCKX genes cause changes to the yield.

9 7
First, researchers elucidated the function of OsCKX2 after observing that null mutants or plants with between OsCKX4 and the root system instead of the shoot system. Here, we discovered that osckx4 4 1 0 mutants produced more tillers, especially at the early stages, but has fewer spikelets per panicle.

1 1
Additionally, osckx9 mutants have been reported to develop more tillers, fewer spikelets per panicle, displayed more tillers, larger panicles, and lighter 1,000-grain weight (Zhang et al., 2020); these were 4 2 0 also observed in our study.

2 1
In contrast to our expectations, none of the single mutants showed a stable significant increase in 4 2 2 yield, while the double mutants osckx1 osckx2 and osckx4 osckx9 showed lower yields. The lower yield 4 2 3 is correlated with some of the disadvantageous phenotypes in these mutants. The discoveries in this 4 2 4 study pointed out that single loss-of-function mutants were not good for breeding. If the expression of 4 2 5 OsCKX gene in different tissues or at different developmental stages can be regulated in breeding, 4 2 6 some favorable phenotypes may be utilized.

2 7
Potential causes of the functional differences and redundancy in OsCKX genes

2 8
Each OsCKX gene has its distinct functions during rice growth and development, but why different 4 2 9 phenotypes arise is worthy of discussion. The main function of CKX enzymes is the biodegradation of 4 3 0 cytokinins, with previous studies reporting that CKXs catalyze the degradation of different cytokinin 4 3 1 compounds. In maize, the ZmCKX enzyme activity has been systematically analyzed; for example,

5 0
These three CKXs potentially regulate the cytokinin balance in other cellular organelles. However, the 4 5 1 locations of the remaining OsCKXs remain unknown. Hence, we hypothesize that the subcellular 4 5 2 localization of OsCKXs may also influence their functional differentiation.

5 3
In this study, the OsCKX genes exhibited different expression patterns. For example, OsCKX1 and 4 5 4 OsCKX2 were expressed in the flower and grains, but OsCKX1 showed lower expressed in the 4 5 5 inflorescence meristem, which is the same as the previous study (Ashikari et al., 2005). As OsCKX4 2 5 several genes that are regulated by OsCKX4 and OsCKX9 individually. In summary, the similar amino 5 0 1 acid sequences but different expression patterns and different subcellular localization of OsCKX4 and 5 0 2 OsCKX9 enable them to functionally overlap in the regulation of some pathways, while also 5 0 3 individually regulating specific pathways.

0 4
In conclusion, our systematic analysis of 11 genes in the OsCKX family reveals their complex 5 0 5 expression patterns in rice. A total of 27 OsCKX mutant lines were produced using CRISPR/Cas9 5 0 6 technology. By examining the phenotypes of the mutants throughout the rice growth period, we 5 0 7 determined the functions of specific OsCKX genes in plant development. Specifically, we discovered 5 0 8 that OsCKX4 and OsCKX9 inhibited tillering, while OsCKX1 and OsCKX2 promoted tillering.

0 9
Considered together, our findings establish a community resource for fully elucidating the function of 5 1 0 OsCKXs, providing new insights that may be used for future studies to improve rice yield. Supplemental data 5 1 4 The following supplemental materials are available.

1 5
Supplemental Figure S1. Gene structures and mutation details of OsCKXs.

1 7
Supplemental Figure S3. Mutation details and chromatograms of double and triple mutants 5 1 8 Supplemental Figure S4. The osckx mutants of Zhonghua11 from the 2020 field experiment.

1 9
Supplemental Figure S5. Phenotypic characterization of the vegetative organs in osckx mutants from 5 2 0 the 2020 field experiment.

2 1
Supplemental Figure S6. Phenotypic characterization of the second and third top leaves in osckx 5 2 2 mutants from the 2020 field experiment.

3 8
Supplemental Table S1. Information of primers used in the study.

3 9
Supplemental Table S2. Characterization of the vegetative organs and yield-related phenotypes in 5 4 0 osckx mutants from the 2019 field experiment.

4 1
Supplemental Table S3. Tiller numbers in osckx mutants from the 2019 field experiment.

4 8
All data supporting the findings of this study are available within the paper and within its 5 4 9 supplementary data published online.

0
Acknowledgments 5 5 1 We thank Jiankang Zhu and Caixia Gao for providing the vectors of the CRISPR-Cas9 system. We also 5 5 2 thank Biogle and Biorun genome editing center for producing transgenic rice.

7
Conflict of interest

8
The authors declare no competing interests in relation to this work.

Figure7. OsCKX4 and OsCKX9 are associated with auxin metabolism and signaling transduction.
Differentially expressed genes (DEGs) related to auxin metabolism and signaling transduction in the root (R) and shoot base (BP) of different mutants compared with Nipponbare (NIP). The fold change was based on log 2 3 7 (FPKM mutant /FPKM WT ) values. * Q < 0.05, ** Q < 0.01.  The function of each OsCKX gene in regulating the growth and development of each organ in rice summarized from the phenotypes of all mutants. "+" means positive regulation, while "-" means negative regulation.