Genetic evidence for an indispensable role of somatic embryogenesis receptor kinases in brassinosteroid signaling

Abstract

The Arabidopsis thaliana somatic embryogenesis receptor kinases (SERKs) consist of five members, SERK1 to SERK5, of the leucine-rich repeat receptor-like kinase subfamily II (LRR-RLK II). SERK3 was named BRI1-Associated Receptor Kinase 1 (BAK1) due to its direct interaction with the brassinosteroid (BR) receptor BRI1 in vivo, while SERK4 has also been designated as BAK1-Like 1 (BKK1) for its functionally redundant role with BAK1. Here we provide genetic and biochemical evidence to demonstrate that SERKs are absolutely required for early steps in BR signaling. Overexpression of four of the five SERKs-SERK1, SERK2, SERK3/BAK1, and SERK4/BKK1-suppressed the phenotypes of an intermediate BRI1 mutant, bri1-5. Overexpression of the kinase-dead versions of these four genes in the bri1-5 background, on the other hand, resulted in typical dominant negative phenotypes, resembling those of null BRI1 mutants. We isolated and generated single, double, triple, and quadruple mutants and analyzed their phenotypes in detail. While the quadruple mutant is embryo-lethal, the serk1 bak1 bkk1 triple null mutant exhibits an extreme de-etiolated phenotype similar to a null bri1 mutant. While overexpression of BRI1 can drastically increase hypocotyl growth of wild-type plants, overexpression of BRI1 does not alter hypocotyl growth of the serk1 bak1 bkk1 triple mutant. Biochemical analysis indicated that the phosphorylation level of BRI1 in serk1 bak1 bkk1 is incapable of sensing exogenously applied BR. As a result, the unphosphorylated level of BES1 has lost its sensitivity to the BR treatment in the triple mutant, indicating that the BR signaling pathway has been completely abolished in the triple mutant. These data clearly demonstrate that SERKs are essential to the early events of BR signaling.

Figure 1. SERK1, SERK2, BAK1, BKK1 are the only four genes in the LRR-RLK II subfamily playing redundant roles in mediating BR signal transduction.

A. Phenotypes of wild-type Arabidopsis (WS2), an intermediate BRI1 mutant, bri1-5, and a null BRI1 mutant, bri1-4. B. Overexpression of SERK1, SERK2, BAK1, and BKK1 can partially suppress the defective phenotypes of bri1-5. Overexpression of SERK5 and other members in LRR-RLK II subfamily can not suppress the bri1-5 phenotypes. C. Overexpression of kinase-inactive mutants, mSERK1 (K330E), mSERK2 (K333E), mBAK1 (K317E), and mBKK1 (K322E) dramatically enhances bri1-5 defective phenotypes. Overexpression of mSERK5 (K303E) and other kinase-inactive mutants in LRR-RLK II subfamily does not enhance the bri1-5 phenotypes. Representative 18-day-old plants were photographed.

Figure 2. Two independent sets of T-DNA insertion null mutants used in this study.

Exons are indicated with filled black boxes. Lines between boxes represent introns. The insertion sites are shown with triangles and the T-DNA orientations are indicated with arrows. The T-DNA insertion mutants labeled with blue were the 1^st set of single null mutants used to create double, triple and quadruple mutants. The ones labeled with green were the 2^nd set of mutants used to generate double, triple, and quadruple mutants.

Figure 3. Representative loss-of-function mutant phenotypes of SERKs.

A. Representative loss-of-function phenotypes of 20-day-old SERK mutants in the light. Only bak1-4 shows a subtle bri1-like phenotype among the single knock-out mutants. The double knock-out mutant serk1-8 bak1-4 shows phenotypes similar to the bri1 weak allele, bri1-301; and bak1-4 bkk1-1 shows a seedling-lethality phenotype at the early developmental stage . The triple knock-out mutant serk1-8 bak1-4 bkk1-1 shows phenotypes similar to the bak1-4 bkk1-1 mutant plants. B. Representative loss-of-function phenotypes of SERK mutants grown in the dark for 5 days. The mutant bri1-701, a T-DNA insertion mutant of BRI1, shows a typical null bri1-like phenotype in the dark with opened cotyledons, short and swollen hypocotyls. The double knock-out mutant serk1-8 bak1-4 shows partially de-etiolated phenotypes with opened cotyledons and semi-dwarfed hypocotyls. The triple knock-out mutant serk1-8 bak1-4 bkk1-1 shows a de-etiolated phenotype almost identical to that of the BRI1 null mutant, bri1-701. C. Measurements of the dark-grown seedlings shown in B. Error bars represent standard deviation (SD).

Figure 4. The phenotypes of SERK mutants can be complemented by each of the SERKs.

A. Rescued mutant phenotypes of 21-day-old plants in the light. The pS1::SERK1-GFP construct restored the serk1-8 bak1-4 mutant phenotype to a typical bak1-4 like mutant phenotype, and the pS3::BAK1-GFP construct restored the serk1-8 bak1-4 mutant phenotype to a wild-type like plant. The pS3::BAK1-GFP and pS3::BKK1-GFP constructs restored the serk1-8 bak1-4 bkk1-1 mutant phenotype to almost a wild-type like plant. However, pS1::SERK1-GFP can not rescue the serk1-8 bak1-4 bkk1-1 phenotype in the light. B. Rescued mutant phenotypes of 5-day-old seedlings in the dark. The pS1::SERK1-GFP and pS3::BAK1-GFP constructs restored the serk1-8 bak1-4 mutant phenotypes. The pS1::SERK1-GFP construct partially restored the serk1-8 bak1-4 bkk1-1 mutant phenotypes to the bak1-4 bkk1-1 double mutant phenotypes with opened cotyledons and taller hypocotyls than the triple mutant. The pS3::BAK1-GFP and pS3::BKK1-GFP constructs restored the serk1-8 bak1-4 bkk1-1 mutant phenotype to a wild-type like seedling. C. Measurements of the dark-grown seedlings shown in B. Error bars represent SD.

Figure 5. serk1-8 bak1-4 bkk1-1 triple null mutant is insensitive to exogenous BR treatment.

A, B. Root and hypocotyl growth analyses of wild-type and mutant plant seedlings grown on medium containing different 24-epiBL concentrations. A. Seven-day-old seedlings grown in the light for root growth analysis. B. Five-day-old seedlings grown in the dark for hypocotyl growth analysis. The double mutant serk1-8 bak1-4 shows reduced sensitivity to BR treatment and the triple mutant serk1-8 bak1-4 bkk1-1 is completely insensitive to BR treatment. Error bars represent SD. C, D. Expression of CPD and DWF4 in serk1-8 bak1-4 bkk1-1 shows responses to exogenously applied BR similarly to the null bri1-701 mutant but differently to wild-type seedlings. Seven-day-old seedlings were treated with or without 1 µM 24-epiBL. Relative expression level of CPD and DWF4 was measured by quantitative RT-PCR. The feedback inhibition of CPD in serk1-8 bak1-4 is also reduced. ACTIN2 was used as the reference gene. The used primers are listed in Table S1. Error bars represent SD (n = 3). E, F. The phosphorylation level of BES1 is not responsive to BR treatment in the serk1-8 bak1-4 bkk1-1 triple null mutant. E. Nine-day-old seedlings of wild-type and mutants grown in the light were treated with 0 or 1 µM 24-epiBL for 4 h. Total proteins were analyzed by an immuo-blotting assay with a specific anti-BES1 antibody. F. Coomassie blue staining showing equally loaded proteins between each pair of samples. −, without 24-epiBL treatment; +, with 24-epiBL treatment. BES1-P, phosphorylated BES1; BES1, unphosphorylated BES1.

Figure 6. Overexpression of BRI1 dramatically increases hypocotyl growth of wild type but does not promote hypocotyl growth of serk1-8 bak1-4 bkk1-1.

A. 35S::BRI1-GFP can not suppress the mutant phenotype of serk1-8 bak1-4 bkk1-1. Five-day-old dark grown seedlings were photographed. B. Measurements of the dark-grown seedlings shown in A. Error bars represent SD. C. Western hybridization indicates the expressed BRI-GFP fusion protein in transgenic plants. D. Phosphorylation of BRI1 does not respond to exogenously applied BR in the mutant serk1-8 bak1-4 bkk1-1. The membrane proteins were extracted from seven-day-old light-grown seedlings and immunoprecipitated with an anti-GFP antibody. The threonine phosphorylation levels of BRI1 were detected with a phosphoThr antibody by immuno-blotting analysis. The immunoprecipitated BRI1-GFP was immuno-blotted with an anti-GFP antibody to show equal loading. −BL, samples not treated with 24-epiBL; +BL, samples treated with 100 nM 24-epiBL for 90 min.

Figure 7. A current model showing SERKs are indispensable to BR signaling.

Left: In wild-type plants, upon perception of BR by the receptor BRI1 and coreceptor BAK1 or its functionally redundant proteins, the phosphorylation levels of BRI1 respond to BR, triggering downstream BR signaling cascade, resulting the accumulation of unphosphorylated BES1, and ultimately leading to the expression of BR responsive genes. Right: In serk1 bak1 bkk1 mutant plants, the phosphorylation levels of BRI1 remain at an almost undetectable basal level and do not alter regardless of elevated concentrations of BR. As a result, most of the BES1 protein exists as an inactive phosphorylated form, which cannot enter to the nuclei and is incapable to regulate the expression of BR responsive genes.