Fig 1.
Expression of CSN gene products and CUL1 modification status in various viable csn mutants.
(A) Apical apex tissues from indicated mutants were extracted and the proteins examined by immunoblotting using indicated antibodies. Lower levels of the CSN (CSN3 and CSN5A) and hyper-neddylation of CUL1 were found in csn5a-1, csn5a-2, csn1-10, and csn2-5. (B) Light-grown seedlings (4-day) of indicated genotypes were used for immunoblotting with anti-CSN5B. The antibody strongly reacted with CSN5B, but also cross-reacted with CSN5A. Endogenous CSN5B migrated slower than CSN5A on the SDS-PAGE. See also S1 Fig.
Fig 2.
Seed germination phenotype of the csn mutants.
(A) Representative images showing seed germination status of indicated mutants. Seeds were cold stratified for 4 days (Cold Str) or had not stratified (Un-str). Photographs were taken at day 6. (B) Germination rates of unstratified (solid lines) or cold stratified (broken lines) seeds of the csn mutants. csn5a-1, csn5a-2, csn5b-1, csn1-10, csn3-3 were compared to Col wild type, while csn2-5 to Ler wild type. With the exception of csn5b-1, all other csn mutants tested exhibited slower and lower germination rates compared to Col, when not cold stratified. (C) Germination rates of Col, csn1-10 and csn5a-1 seeds that had been in dry storage for indicated period of time. csn1-10 and csn5a-1 seeds required extended after-ripening time than Col to release dormancy. csn5a-1 additionally showed delayed germination. (D) Germination rates of freshly collected, partially after-ripened (2 weeks), or fully after-ripened (11 weeks) seeds of Col, csn1-10, and csn5a-1. The germination deficiency of csn1-10 and csn5a-1 were most readily observed with the seeds of two weeks after seed-collection. csn5a-1 seeds exhibited a delay in germination even in fully after-ripened seeds. In all panels, error bars represent standard deviation from 4 repeats (n = 4). See also S2 and S3 Figs.
Fig 3.
The germination defects of csn1 and csn5a are seed-coat dependent.
Intact seeds or embryos dissected from the same seed population were germinated on water-agar plates. (A) csn5a-1, or (B) csn1-10, with the corresponding wild type Col seeds as the controls. (C) csn1-1 (fus6-1) mutant seeds were picked out from the heterozygous (+/-) seed population based on the dark purple seed color, and normal colored seeds were used as the positive control (WT). While the seeds of the csn mutant could not germinate on time, their embryos showed normal progression of growth, with radicle elongation on day-2 and cotyledon greening and expansion on day-3. csn1-1 mutants arrested the growth after germination.
Fig 4.
csn1-10 and csn5a-1 showed defects in timely down-regulation of RGL2 and ABI5 proteins following seed imbibition.
(A and B) Time-course analysis of RGL2 levels by immunoblots showing the deficiency of csn1-10 (A) and csn5a-1 (B) in timely decline of RGL2 protein compared to wild type (Col). Anti-RPN6 blot or a background band were shown as the loading reference. (C) Time course analysis of ABI5 level by immunoblotting. csn1-10, csn5a-1, csn5a-2, but not csn5b-1, showed defect in timely decline of ABI5 protein following seed imbibition. Anti-histone H3 was used as an internal loading reference. The lower panel shows the densitometry quantification of ABI5 relative to the corresponding Histone H3 in each sample. Samples were collected at indicated time points in hours (hr) post imbibition, as described in Materials and Methods.
Fig 5.
csn1-10 and csn5a-1 show different genetic interactions with rgl2-13, and different responses to PAC.
(A and B) rgl2-13 rescued the germination deficiency of csn1-10 but not csn5a-1. (A) The csn1-10 rgl2-13 double mutants displayed normal germination rate as wild type Col and rgl2-13, in contrast to low germination of csn1-10. Double mutants db26 and db35 are two different lines of the csn1-10 rgl2-13. (B) Germination rate of csn5a-1 rgl2-13 double mutants displayed low germination rate similar to csn5a-1 single mutant, compared to wild type Col and rgl2-13. Double mutants db19 and db20 are two different lines of the csn5a-1 rgl2-13. (C) csn5a-1 was hypersensitive to PAC. Cold stratified seeds were germinated on regular GM plate, GM plate containing PAC (1 μM), or PAC (1 μM) plus GA3 (50 μM). Seed germination rates were shown in graphs. (D) rgl2-13 conferred resistance to PAC (1 μM) in germination of rgl2-13 and double mutant csn1-10 rgl2-13, but not double mutant csn5a-1 rgl2-13. (E) The csn5a-1 rgl2-13 double mutant showed abnormally prolonged accumulation of ABI5, similar to csn5a-1 single mutant. Samples were taken at indicated times after imbibition of the seeds, and proteins were probed with antibodies as indicated on the left. (A-D) Error bars represent standard deviation from 4 repeats (n = 4). See also S4 Fig.
Fig 6.
abi5-4 mutant rescued the germination of csn1-10 as well as csn5a-1.
(A and B) Germination of csn1-10 abi5-4 (A) or csn5a-1 abi5-4 (B) double mutants displayed normal rate, in contrast to slow germination of the corresponding single mutant csn1-10 and csn5a-1, respectively. All of the csn1-10 abi5-4 lines including double mutant db-5, wild type (WT), csn1-10, and abi5-4 are segregating sibling lines from the same genetic cross. Likewise, all of the csn5a-1 abi5-4 lines including the double mutant csn5a-1 abi5-4, wild type (WT), csn5a-1, and abi5-4 are segregating sibling lines from the same genetic cross. Error bars represent standard deviation from 4 repeats. (C and D) Anti-RGL2 immunoblots showing that timely removal of RGL2 was restored in csn1-10 abi5-4 double mutant (C) as well as in csn5a-1 abi5-4 double mutant (D). Samples of indicated lines were taken at indicated times after seed imbibition. See also S5 Fig.
Fig 7.
Effects of csn1-10 and csn5a-1 mutations on transcriptome and ABA sensitivity.
(A) The diagrams showing the number of SSTF (statistically significant two-fold change) genes in csn5a-1 and csn1-10 mutants compared to wild type Col in dry seeds, or after two-day imbibition. csn1-10 affected more genes than csn5a-1 in dry seeds, while csn5a-1 affected more genes than csn1-10 in 2-day imbibed seeds. The areas of the squares were drawn in proportion to the number of the genes in the indicated category. (B) Real-time qPCR analysis of representative ABA related genes on day2 and day3 imbibed seeds of Col, csn1-10, and csn5a-1. The expression level of each sample was normalized to IPA-like1 (AT1G17210). (C) Sensitivity to ABA-mediated inhibition of germination is enhanced in csn1-10, csn5a-1, and csn5a-1rgl2 double mutant. The csn5a-1abi5-4 double mutant showed resistance to ABA. Cold stratified seeds were used. Error bars represent standard deviation from 4 repeats. (D) ABA biosynthetic inhibitor norflurazon (5μM) rescued the low germination rates of both csn1-10 and csn5a-1, but did not rescue the delayed germination in csn5a-1. Error bars represent standard deviation from 4 repeats. (E) Yeast-two-hybrid sassy showing interactions of ABI5 with CSN5A and CSN5B. AD, activation-domain fusion vectors; BD, DNA-binding domain fusion vectors; EV, empty vector. See also S6 and S7 Figs.
Fig 8.
CSN functions in the seed germination pathway.
A model indicating the targets of the CSN and CSN5A in the GA- and ABA- regulated seed germination pathway. The CSN complex (or CSN1) promotes germination by facilitating SCF-mediated degradation of RGL2 DELLA repressor. CSN5A, independently from the CSN complex, is involved in the turnover of ABI5, which acts downstream of RGL2. ABI5 appears to have a role in stabilizing RGL2, while RGL2 has been reported to activate ABI5 expression directly and by inducing ABA synthesis.