formation of nanoparticles, which are subjectto grain growth or dissolution into the host ma-terial during operation. The method is independentof phonon properties, implying that improvementsin zT induced by reducing the lattice k value canwork in conjunction with the mechanism de-scribed here. We anticipate that deliberatelyengineered impurity-induced band-structure dis-tortions will be a generally applicable route toenhanced S and zT in all TE materials. We areoptimistic about the commercial use of suchPbTe-based materials because there is an exten-sive knowledge base among the manufacturersof thermoelectric generators about the assemblyof PbTe-based devices, in particular the abilityto make stable metallic contacts with low ther-mal and electrical resistance.

References and Notes1. G. J. Snyder, E. S. Toberer, Nat. Mater. 7, 105 (2008).2. T. C. Harman et al., Science 297, 2229 (2002).3. A. I. Hochbaum et al., Nature 451, 163 (2008).4. K. F. Hsu et al., Science 303, 818 (2004).5. B. Poudel et al., Science 320, 634 (2008)6. B. Sales et al., Science 5266, 1325 (1998).7. J. Androulakis et al., Adv. Mater. 18, 1170 (2006).8. P. F. R. Poudeu et al., Angew. Chem. Int. Ed. 45, 3835

(2006).

9. G. A. Slack, in Solid State Physics, Vol. 34, H. Ehrenreich,F. Seitz, D. Turnbull, Eds. (Academic Press, New York,1979), pp. 1–71.

10. R. H. Costescu et al., Science 303, 989 (2004).11. L. D. Hicks, M. S. Dresselhaus, Phys. Rev. B 47, 12727

(1993).12. L. D. Hicks, M. S. Dresselhaus, Phys. Rev. B 47, 16631

(1993).13. J. P. Heremans et al., Phys. Rev. Lett. 88, 216801 (2002).14. G. D. Mahan, J. O. Sofo, Proc. Natl. Acad. Sci. U.S.A. 93,

7436 (1996).15. M. Cutler, N. F. Mott, Phys. Rev. 181, 1336 (1969).16. S. Ahmad et al., Phys. Rev. Lett. 96, 056403 (2006).17. S. A. Nemov et al., Physics-Uspekhi 41, 735 (1998).18. B. A. Volkov et al., Physics-Uspekhi 45, 819 (2002).19. K. Nakayama et al., Phys. Rev. Lett. 100, 227004

(2008).20. V. Jovovic et al., J. Appl. Phys. 103, 053710 (2008).21. V. Jovovic et al., in Mater. Res. Soc. Symp. Proc. 1044,

T. P. Hogan, J. Yang, R. Funahashi, T. Tritt, Eds. (MaterialsResearch Society, Warrendale, PA, 2008), U04-09.

22. Y. Matsushita et al., Phys. Rev. B 74, 134512 (2006).23. Materials and methods are available as supporting

material on Science Online.24. R. W. Fritts, in Thermoelectric Materials and Devices,

I. B. Cadoff, E. Miller, Eds. (Reinhold, New York, 1960),pp. 143–162.

25. Yu. I. Ravich et al., Semiconducting Lead Chalcogenides(Plenum, New York, 1970).

26. We have not been able to find the carrier-densitydependence of the hole mobility for PbTe. For electrons,at 300 K, me ≈ 550 cm2/Vsat, n = 5 × 1019 cm−3 (32),whereas the ratio of electron to hole mobility at n ≈ 3 ×

1018 cm−3 and 300 K is 2.2 (33), so that we estimate thehole mobility to be about 250 cm2/Vs 5 × 1019 cm−3.

27. A. F. Ioffe, Semiconductor Thermoelements and ThermoelectricCooling, (Infosearch Limited, London, 1957).

28. J. P. Heremans et al., Phys. Rev. B 70, 115334 (2004).29. Ravich (34) suggests that the energy dependence of the

hole scattering on the resonant states (resonantscattering) increases S by increasing the second term,dm/dE, and therefore the low temperature zT.

30. H. Preier, Appl. Phys. (Berl.) 20, 189 (1979).31. H. B. Callen, Thermodynamics (Wiley, New York, 1960).32. Yu. I. Ravich et al., Phys. Status Solidi B 43, 453

(1971).33. R. S. Allgaier, W. W. Scanlon, Phys. Rev. 111, 1029 (1958).34. Yu. I. Ravich, in CRC Handbook of Thermoelectrics,

D. M. Rowe, Ed. (CRC Press, Boca Raton, FL, 1995),pp. 67–81.

35. This work was supported by the BSST Corporation, theState of Ohio Department of Development’s Center forPhotovoltaic Innovation of Commercialization (OSU), theBeckman Institute, the Swedish Bengt Lundqvist MinneFoundation, and Jet Propulsion Laboratory, NASA(Caltech). Patent protection related to this work ispending.

Supporting Online Materialwww.sciencemag.org/cgi/content/full/321/5888/554/DC1Materials and MethodsFig. S1References

28 April 2008; accepted 17 June 200810.1126/science.1159725

BSKs Mediate SignalTransduction from the ReceptorKinase BRI1 in ArabidopsisWenqiang Tang,1 Tae-Wuk Kim,1 Juan A. Oses-Prieto,2 Yu Sun,1 Zhiping Deng,1Shengwei Zhu,1,3 Ruiju Wang,1,4 Alma L. Burlingame,2 Zhi-Yong Wang1*

Brassinosteroids (BRs) bind to the extracellular domain of the receptor kinase BRI1 toactivate a signal transduction cascade that regulates nuclear gene expression and plantdevelopment. Many components of the BR signaling pathway have been identified and studiedin detail. However, the substrate of BRI1 kinase that transduces the signal to downstreamcomponents remains unknown. Proteomic studies of plasma membrane proteins lead to theidentification of three homologous BR-signaling kinases (BSK1, BSK2, and BSK3). The BSKs arephosphorylated by BRI1 in vitro and interact with BRI1 in vivo. Genetic and transgenic studiesdemonstrate that the BSKs represent a small family of kinases that activate BR signalingdownstream of BRI1. These results demonstrate that BSKs are the substrates of BRI1 kinasethat activate downstream BR signal transduction.

Cell-surface receptor kinases activate cel-lular signal transduction pathways uponperception of extracellular signals, thereby

mediating cellular responses to the environment

and to other cells. The Arabidopsis genome en-codes more than 400 receptor-like kinases (RLKs)(1). Some of these RLKs function in growthregulation and plant responses to hormonal andenvironmental signals. However, the molecularmechanism of RLK signaling to immediate down-stream components remains poorly understood,as no RLK substrate that mediates signal trans-duction has been established in Arabidopsis (2).BRI1 is an RLK that functions as the major re-ceptor for the steroid hormones brassinosteroids(BRs) (2). BRs bind the extracellular domain ofBRI1 to activate its kinase activity, initiating asignal transduction cascade that regulates nuclear

gene expression and a wide range of develop-mental and physiological processes (fig. S1) (3).Many components of the BR signaling pathwayhave been identified, and much detail has beenrevealed about how BR activates BRI1 (4–8) andhow phosphorylation by downstream GSK3-like kinase BIN2 regulates the activity of thenuclear transcription factors that mediate BR-responsive gene expression (fig. S1) (3, 9–13).However, no direct interaction has been ob-served between BRI1 and BIN2, and it remainsunclear how BRI1 kinase at the plasma mem-brane transduces the signal to cytoplasmic com-ponents of the BR pathway (14).

To identify additional components of theBR signaling pathway, we performed quantitativeproteomic studies of BR-responsive proteins usingtwo-dimensional difference gel electrophoresis(2D DIGE). Seedlings of BR-deficient det2-1 mu-tant were treated with brassinolide (BL) (themost active form of BRs) or mock solution, andproteins were labeled with Cy3 or Cy5 dyes,mixed together, and separated in the same gelby 2D gel electrophoresis (2-DE). BL-inducedBAK1 phosphorylation and BZR1 dephospho-rylation were detected in the plasma membraneand phosphoprotein fractions, respectively (15),but not in total proteins (16). Similar to BAK1,two additional rows of spots showed a BR-inducedincrease of the acidic forms and a decrease ofthe basic forms (Fig. 1, A and B), which is con-sistent with BR-induced phosphorylation. Massspectrometry analysis of these spots identified twokinases encoded by Arabidopsis genes At4g35230and At5g46570, which we named BR-signalingkinases 1 and 2 (BSK1 and BSK2) (Fig. 1B and

1Department of Plant Biology, Carnegie Institution of Wash-ington, Stanford, CA 94305, USA. 2Department of Pharma-ceutical Chemistry, University of California, San Francisco, CA94143, USA. 3Key Laboratory of Photosynthesis and Envi-ronmental Molecular Biology, Institute of Botany, ChineseAcademy of Sciences, Beijing 100093, China. 4Institute forMolecular Biology, College of Life Science, Nankai University,Tianjin 300071, China.

*To whom correspondence should be addressed. E-mail:zywang24@stanford.edu

www.sciencemag.org SCIENCE VOL 321 25 JULY 2008 557

REPORTS

on

Oct

ober

10,

200

9 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

fig. S2). BSK1 and BSK2 share 60% amino acidsequence identity (fig. S3) and are members ofthe receptor-like cytoplasmic kinase subfamilyRLCK-XII (1). The RLCK-XII subfamily includes12 Arabidopsis proteins that each contains a ki-nase domain at the N-terminal side and tetra-tricopeptide repeat (TPR) domains at the Cterminus (fig. S3) (1). TPR domains are knownto mediate protein-protein interactions and arepresent in components of steroid receptor com-plexes in animals (17). BSK1 and BSK2 do notcontain predicted transmembrane domains buthave putative N-terminal myristylation sites (gly-cine 2) that could mediate their membranelocalization (fig. S3).

The BR-induced shift of BSK1 from the basicto the acidic side in 2-DE gels was confirmed byimmunoblotting of transgenic plants expressing aBSK1–yellow fluorescence protein (YFP) fusionprotein (Fig. 1, C and E). The response wasobviously weaker in the bri1-5mutant background(Fig. 1, D and E), suggesting that BR regulation ofBSK1 is BRI1-dependent. Consistent with theiridentification in the plasma membrane frac-tions, BSK1-YFP fusion proteins showed lo-calization on the cell surface, and the localizationwas not affected by BL treatment (Fig. 1F).

The plasma membrane localization and BR-induced modification of BSKs suggest thatthey might be substrates of BRI1 or BRI1’s co-receptor kinase BAK1 (18, 19). In vitro kinaseassays demonstrated that BRI1, but not BAK1,phosphorylates BSK1 (Fig. 2A). Mass spectrom-etry analysis of BRI1-phosphorylated BSK1 iden-

tified Ser230 of BSK1 as a BRI1 phosphorylationsite (fig. S4). This same residue is also phospho-rylated in vivo (20). Whereas deletion of theC-terminal TPR domain has no effect on BSK1phosphorylation by BRI1, a Ser230 → Ala230

(S230A) mutation reduced the phosphorylation

by 82% (Fig. 2B), indicating that Ser230 is themajor site for BRI1 phosphorylation.

We demonstrated in vivo interactions withBRI1 with the use of bimolecular fluorescencecomplementation (BiFC) and coimmunoprecipi-tation assays. Whereas cells coexpressing BSK1

Fig. 1. Identification of BSK1 and BSK2 as early BR-regulatedplasma membrane proteins. (A) 2D DIGE image of plasmamembrane proteins isolated from 7-day-old det2 seedlings treatedfor 2 hours with either 100 nM BL (labeled with Cy5, red) or mocksolution (Cy3, green). (B) Zoom-in view of an area in (A) showing BR-induced (black arrows, red spots) and BR-repressed (white arrows,green spots) protein spots. The table summarizes the protein identity,the number of unique peptides, and the percentage of proteinsequence coverage of mass spectrometry data for the spots numberedin the upper panel. (C to E) 2D gel immunoblotting analysis of BRregulation of posttranslational modification of BSK1 in det2 (C) andbri1-5 (D) background. Transgenic det2 or bri1-5 mutant seedlingsexpressing BSK1-YFP fusion protein were treated for 15 min with mocksolution (-BL) or 100 nM BL (+BL). The proteins were separated by2-DE and immunoblotted with anti-YFP antibody. (E) Quantitation ofrelative spot intensity along the isoelectric focusing dimension in (C)and (D). (F) Confocal microscopy images show localization of BSK1-YFPin hypocotyl cells of 3-day-old dark-grown transgenic det2 seedlingsbefore (-BL) and 2 hours after treatment with 100 nM brassinolide(+BL). Scale bar, 10 mm.

pH 4 7

Spots Protein I D

Unique peptides

Sequence coverage

1 BSK1 6 1 8 % 2 BSK1 6 1 6 % 3 BSK2 2 4 6 6 % 4 BSK2 1 6 5 3 % 5 BSK2 1 0 2 9 %

3 4 5

1 2

A

B

-BL

+BL

pH

C

D-BL

+BL

D

E

0

5

10

15

20

25

30det2-det2+bri1-bri1+

Position of spots (pH)

Rel

ativ

e in

tens

ity (

%)

-BL +BL

F

D

C

B

BRI1-GFP

BL - - + - - +Input Anti-GFP IP

1 2 3 4 5 6

BSK1-myc

BSK1-cYFP+BAK1-nYFP

BSK1-cYFP+BRI1-nYFP

BSK1-cYFP+ nYFP

A

BRI1+BSK1

mBRI1+BSK1

BSK1BRI1

mBRI1BAK1+BSK1

mBAK1+BSK1

BAK1mBAK1

BRI1BRI1+BSK1

BRI1+TPR

BRI1BRI1+BSK1

BRI1+S230A

BSK1S230A

Fig. 2. BSK1 is a substrate of BRI1. (A and B) BRI1phosphorylates Ser230 of BSK1 in vitro. (A) Autoradiog-raphy of in vitro kinase assays performed with WT(BRI1 and BAK1) or kinase-dead mutant (mBRI1 andmBAK1) forms of the kinase domain of BRI1 and BAK1as glutathione S-transferase fusion proteins (B). In vi-

tro kinase assays of BRI1 phosphorylation of full-length BSK1, a truncated BSK1 with deletion of theTPR domain (DTPR) and the S230A mutant BSK1. (C) BiFC assay shows BRI1 interaction with BSK1. YFPfluorescence images of Nicotiana benthamiana leaf epidermal cells cotransformed with the indicatedconstructs. (D) Coimmunoprecipitation of BRI1 with BSK1. Arabidopsis plants expressing BSK1-myc only(lanes 1 and 4) or coexpressing BSK1-myc and BRI1-GFP (lanes 2, 3, 5, and 6) were treated with 100nM brassinolide (BL+) or mock solution (BL-) for 30 min. Microsomal proteins (lanes 1 to 3) wereimmunoprecipitated with anti-GFP antibodies (lanes 4 to 6), and the immunoblot was probed with anti-GFP antibodies or anti-myc antibodies.

25 JULY 2008 VOL 321 SCIENCE www.sciencemag.org558

REPORTS

on

Oct

ober

10,

200

9 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

fused to the C-terminal half of YFP (BSK1-cYFP)and the nonfusion N-terminal half of YFP (nYFP)or BAK1-nYFP fusion showed no or weak flu-orescence signals (Fig. 2C), cells coexpressingBRI1-nYFP and BSK1-cYFP showed strongBiFC fluorescence at the plasma membrane (Fig.2C). Antibodies to BSK1 (anti-BSK1) immunopre-cipitated the BRI1–green fluorescent protein (GFP)protein expressed from the BRI1 promoter (fig.S5), and a BSK1-myc protein was immunoprecip-itated by anti-GFP antibodies only in transgenicArabidopsis plants expressing both BRI1-GFPand BSK1-myc (Fig. 2D). BR treatment reducedthe amount of the coimmunoprecipitated BSK1-myc to 46% of the untreated sample (Fig. 2D),suggesting that BSK1 might be released fromBRI1 upon phosphorylation. These results indi-cate that BSK1 is a BRI1 kinase substrate that isphosphorylated upon BR activation of BRI1.

To determine the functions of BSKs and theirhomologs in BR signaling, T-DNA insertion mu-tants were obtained for BSK2, BSK3, BSK4,

BSK5, and BSK12 genes (21). Of these, only thebsk3-1 mutant showed an obvious phenotype(fig. S6). The bsk3-1 mutant contains a T-DNAinsertion in the 5′ untranscribed region and ex-presses a much reduced level of the BSK3 RNA(Fig. 3, A and B). The bsk3-1 mutant seedlingsgrown in the dark on regular medium or me-dium containing the BR biosynthetic inhibitorbrassinazole (BRZ) showed shorter hypocotyllength than did wild-type (WT) seedlings (Fig.3C). BL treatment increases hypocotyl elongationand inhibits root growth in WT plants grown inthe light. Compared to the wild type, the bsk3-1mutant showed reduced responses to BL inhypocotyl elongation, root inhibition, and expres-sion of the BZR1 target gene DWF4 and theBES1 target gene SAUR-Ac (Fig. 3D and fig. S7).These results demonstrate that loss-of-functionmutation of bsk3 reduces BR sensitivity, indicat-ing an essential role for BSK3 in BR signaling.Similar to BSK1, the BSK3 protein is also reg-ulated by BR (fig. S8), is phosphorylated byBRI1 kinase in vitro (fig. S9), and interacts withBRI1 in a BR-dependent manner in vivo (fig.S10). BSK1 and BSK3 are expressed in similartissues as is BRI1 (fig. S11). These resultssuggest that BSK3 and its homologs play

redundant or overlapping roles in BR signaling,which could explain the weak BR-insensitivephenotypes of bsk3-1.

When overexpressed in the BR-insensitivebri1-5 mutant (Fig. 4, A and B, and fig. S12) orBR-deficient det2-1 (fig. S13) mutant back-grounds, BSK1, BSK3, and BSK5 obviouslysuppressed the dwarf phenotypes of the mutants.Consistent with reduced BR sensitivity of thebsk3-1 mutant, overexpression of BSK3 is mosteffective in rescuing the bri1 phenotypes. Thegrowth phenotypes correlated with altered ex-pression of the BZR1 target gene DWF4 (Fig.4C and fig. S12), indicating that overexpressionof the BSKs activates downstream BR signal-ing. Overexpression of BSK3 partly suppressedthe dwarf phenotype of the null allele bri1-116(Fig. 4D), but not that of the bin2-1 mutant (Fig.4E), indicating that BSK3 functions downstreamof BRI1 but upstream of BIN2, which is con-sistent with BSK3 being a substrate of the BRI1kinase.

We used quantitative proteomics to identifyBSKs as previously unrecognized BR signal trans-duction components. This study demonstratesthat sample prefractionation followed by 2D DIGEis a powerful proteomic approach for dissecting

Col bsk3 Col bsk3-BRZ +BRZ

Col bsk3 Col bsk3-BL +BL

Hyp

oco

tyl l

eng

th (

mm

)

+BRZ-BRZ0

2

4

6

8

10

12

D

C

Salk_096500 500 bp ol bsk3BSK3UBC

B

BL (nM)

Hyp

oco

tyl

00.

1 1 10 100

500

0

2

4bsk3 Col

Ro

ot

(m

m)

A

0

6

12

18

Colbsk3

Fig. 3. The bsk3-1 mutant has reduced BR sensi-tivity. (A) T-DNA insertion site of bsk3-1 knockoutmutant (T-DNA line SALK_096500). bp, base pair.(B) Reverse transcription polymerase chain reac-tion (RT-PCR) analysis of BSK3 RNA expression inseedlings of WT Columbia ecotype (Col) and thebsk3-1 mutant, with UBC RNA as control. (C) WT(Col) and bsk3-1 seedlings grown in the dark for4 days on a regular medium (-BRZ) or a mediumcontaining 1 mM brassinazole (+BRZ). Averagehypocotyl length of at least 25 seedlings is shownat right. Error bars indicate SE. (D) The bsk3-1mutant shows reduced sensitivity to BL. The leftpanel shows representative seedlings of Col orbsk3-1 grown in the absence (-BL) or presence(+BL) of 50 nM BL for 7 days under constant light.The right panel shows hypocotyl and root lengths(average of at least 60 seedlings) of WT (Col) andbsk3-1 seedlings grown on various concentrationsof BL under continuous light. Error bars indicate SE.

WT BSK3 BSK5 BSK1 bri1-5

A

B0

2

4

6

8

DW

F4

RN

A le

vel

WT BSK3 BSK5 BSK1 bri1-5

C

bri1-116 bri1-116/BSK3-ox

bin2-1bin2-1/

BSK3-ox

D E

F BRI1 BRBAK1

BIN2

BZR1/2pBZR1/2

BSU114-3-3

BSKp p

BSKpp p

BIN2

BR-targetgenes

BSU1

BRI1

BAK1

?

+BR-BR

BKI1

BKI1

p

Fig. 4. BSKs function downstream of BRI1 but upstream of BIN2 in the BR signaling pathway. Phenotypeof light-grown 3-week-old (A) or 5-week-old (B) WT, bri1-5, or transgenic bri1-5 overexpressing BSK3,BSK5, or BSK1 as YFP fusion proteins. (C) Quantitative RT-PCR analysis of DWF4 RNA expression in plantsrepresented in (A). Error bars indicate SD. (D) Overexpression of BSK3 partly suppresses the bri1-116mutant. (E) Overexpression of BSK3 cannot suppress the bin2-1 mutant. (F) A model of BR signaltransduction. Components in the inactive and active states are shown in blue and red, respectively. In theabsence of BR (-BR), BRI1 associates with BSKs in an inactive state; BIN2 phosphorylates BZR1 and BZR2to inhibit their DNA binding activity and promote their cytoplasmic retention by the 14-3-3 proteins. BR-binding (+BR) to BRI1 induces its dimerization with BAK1 and activation of BRI1 kinase, whichphosphorylates BSKs. Phosphorylated (p) BSKs dissociate from BRI1 and presumably inhibit BIN2 kinaseand/or activate BSU1 phosphatase through yet unknown mechanisms, leading to dephosphorylation ofBZR1 and BZR2, which regulate BR-responsive gene expression.

www.sciencemag.org SCIENCE VOL 321 25 JULY 2008 559

REPORTS

on

Oct

ober

10,

200

9 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

signaling pathways. Although only BSK1 andBSK2 were identified in the proteomic study,additional members (BSK3 and BSK5) of thisfamily of RLCKs appear to play a similar role inBR signaling. Our results support a model forthe function of BSKs in BR signaling (Fig. 4F).In the absence of BR, BSKs are associated withBRI1. Upon BR activation of BRI1, BSKs arephosphorylated and then disassociate from thereceptor complex to activate downstream sig-naling. Such ligand-induced disassociation froma preexisting receptor complex potentially pro-vides faster signaling than does ligand-inducedrecruitment of a free component into the recep-tor complex.

Both BSKs and BAK1 are substrates of theBRI1 kinase, but several lines of evidence indi-cate that they play distinct roles in BR signaling.First, BR induces BRI1-BAK1 interactions (6) butreduces BRI1-BSK1 and BRI1-BSK3 interactions.Second, overexpression of BSK3 suppressesthe bri1-116 null allele, whereas overexpres-sion of BAK1 only suppresses weak alleles butnot a strong allele of bri1 nor a double mutantcontaining the weak bri1-5 allele and the BR-biosynthetic mutation det2-1 (19). This suggeststhat BSK3 functions downstream of BRI1, whereasBAK1’s action on the downstream BR responserequires a functional BRI1. BAK1 and its homolog

BKK1 are required in additional signaling path-ways, and BAK1 is also a co-receptor for theFLS2 receptor kinase (a receptor for flagelin),suggesting that BAK1 is not a specific compo-nent of the BR pathway (22–25). BAK1 mostlikely mediates activation of BRI1 kinase ratherthan signal transduction to specific downstreamcomponents in the BR signaling pathway. In con-trast, the BSKs directly mediate signal trans-duction from BRI1 to downstream BR responses(Fig. 4F). Identification of the downstream directtargets of BSKs will be the key to fully under-standing how the BR signal is transduced from thecell surface to the nuclear transcription factors.

References and Notes1. S. H. Shiu et al., Plant Cell 16, 1220 (2004).2. K. L. Johnson, G. C. Ingram, Curr. Opin. Plant Biol. 8,

648 (2005).3. G. Vert, J. L. Nemhauser, N. Geldner, F. Hong, J. Chory,

Annu. Rev. Cell Dev. Biol. 21, 177 (2005).4. Z. Y. Wang, H. Seto, S. Fujioka, S. Yoshida, J. Chory,

Nature 410, 380 (2001).5. T. Kinoshita et al., Nature 433, 167 (2005).6. X. Wang et al., Plant Cell 17, 1685 (2005).7. X. Wang, J. Chory, Science 313, 1118 (2006); published

online 19 July 2006 (10.1126/science.1127593).8. X. Wang et al., Dev. Cell 8, 855 (2005).9. Z. Y. Wang et al., Dev. Cell 2, 505 (2002).10. J.-X. He et al., Science 307, 1634 (2005); published

online 27 January 2005 (10.1126/science.1107580).11. Y. Yin et al., Cell 120, 249 (2005).

12. G. Vert, J. Chory, Nature 441, 96 (2006).13. S. S. Gampala et al., Dev. Cell 13, 177 (2007).14. J. M. Gendron, Z. Y. Wang, Curr. Opin. Plant Biol. 10,

436 (2007).15. W. Tang et al., Mol. Cell. Proteomics 7, 728 (2008).16. Z. Deng et al., Mol. Cell. Proteomics 6, 2058 (2007).17. D. F. Smith, Cell Stress Chaperones 9, 109 (2004).18. K. H. Nam, J. Li, Cell 110, 203 (2002).19. J. Li et al., Cell 110, 213 (2002).20. T. Niittyla, A. T. Fuglsang, M. G. Palmgren, W. B. Frommer,

W. X. Schulze, Mol. Cell. Proteomics 6, 1711 (2007).21. J. M. Alonso et al., Science 301, 653 (2003).22. K. He et al., Curr. Biol. 17, 1109 (2007).23. B. Kemmerling et al., Curr. Biol. 17, 1116 (2007).24. D. Chinchilla et al., Nature 448, 497 (2007).25. A. Heese et al., Proc. Natl. Acad. Sci. U.S.A. 104, 12217

(2007).26. We thank W. Briggs for commenting on the manuscript.

Our research was supported by grants from NSF (NSF0724688), the U.S. Department of Energy (DE-FG02-04ER15525), and NIH (R01GM066258). S.Z. and R.W.are supported by the Chinese Scholarship Council. TheUniversity of California, San Francisco, MassSpectrometry Facility (A. L. Burlingame, Director) issupported by the Biomedical Research TechnologyProgram of the National Center for Research Resourcesand grants NIH NCRR RR01614, RR012961, andRR019934.

Supporting Online Materialwww.sciencemag.org/cgi/content/full/321/5888/557/DC1Materials and MethodsFigs. S1 to S13References

25 February 2008; accepted 24 June 200810.1126/science.1156973

One-Third of Reef-Building CoralsFace Elevated Extinction Risk fromClimate Change and Local ImpactsKent E. Carpenter,1* Muhammad Abrar,2 Greta Aeby,3 Richard B. Aronson,4 Stuart Banks,5Andrew Bruckner,6 Angel Chiriboga,7 Jorge Cortés,8 J. Charles Delbeek,9 Lyndon DeVantier,10Graham J. Edgar,11,12 Alasdair J. Edwards,13 Douglas Fenner,14 Héctor M. Guzmán,15Bert W. Hoeksema,16 Gregor Hodgson,17 Ofri Johan,18 Wilfredo Y. Licuanan,19Suzanne R. Livingstone,1 Edward R. Lovell,20 Jennifer A. Moore,21 David O. Obura,22Domingo Ochavillo,23 Beth A. Polidoro,1 William F. Precht,24 Miledel C. Quibilan,25Clarissa Reboton,26 Zoe T. Richards,27 Alex D. Rogers,28 Jonnell Sanciangco,1Anne Sheppard,29 Charles Sheppard,29 Jennifer Smith,1 Simon Stuart,30 Emre Turak,10John E. N. Veron,10 Carden Wallace,31 Ernesto Weil,32 Elizabeth Wood33

The conservation status of 845 zooxanthellate reef-building coral species was assessed by usingInternational Union for Conservation of Nature Red List Criteria. Of the 704 species thatcould be assigned conservation status, 32.8% are in categories with elevated risk of extinction.Declines in abundance are associated with bleaching and diseases driven by elevated sea surfacetemperatures, with extinction risk further exacerbated by local-scale anthropogenic disturbances.The proportion of corals threatened with extinction has increased dramatically in recent decadesand exceeds that of most terrestrial groups. The Caribbean has the largest proportion of corals inhigh extinction risk categories, whereas the Coral Triangle (western Pacific) has the highestproportion of species in all categories of elevated extinction risk. Our results emphasize thewidespread plight of coral reefs and the urgent need to enact conservation measures.

Coral reefs harbor the highest concentra-tion of marine biodiversity. They havehigh aesthetic, recreational, and resource

values that have prompted close scientificscrutiny, including long-term monitoring (1, 2),and face increasing threats at local and global

scales. Globally, rapid buildup of carbon dioxide(and other greenhouse gases) in the atmosphere isleading to both rising sea surface temperatures(with an increased likelihood of mass coralbleaching and mortality) and acidification (3).Ocean acidification is reducing ocean carbonate

ion concentrations and the ability of corals tobuild skeletons (4). Local threats include humandisturbances such as increased coastal develop-ment, sedimentation resulting from poor land-useand watershed management, sewage discharges,nutrient loading and eutrophication from agro-chemicals, coral mining, and overfishing (1, 2, 5–9).Local anthropogenic impacts reduce the resil-ience of corals to withstand global threats, re-sulting in a global deterioration of reef structureand ability of these ecosystems to sustain theircharacteristic complex ecological interactions(1–3, 5–9).

In view of this ecosystem-level decline, weused International Union for Conservation ofNature (IUCN) Red List Categories and Criteriato determine the extinction risk of reef-buildingcoral species. These criteria have been widelyused and rely primarily on population sizereduction and geographic range information toclassify, in an objective framework, the extinc-tion risk of a broad range of species (10). Cate-gories range from Least Concern, with very littleprobability of extinction, to high risk, CriticallyEndangered (Table 1). The threatened categories(Vulnerable, Endangered, and Critically Endan-gered) are intended to serve as one means of set-ting prioritymeasures for biodiversity conservation.

Our assessments of extinction risk cover allknown zooxanthellate reef-building corals andinclude 845 species from the Scleractinia plusreef-building octocorals and hydrocorals (fami-lies Helioporidae, Tubiporidae, and Millepori-dae). Corals have persisted for tens of millions of

25 JULY 2008 VOL 321 SCIENCE www.sciencemag.org560

REPORTS

on

Oct

ober

10,

200

9 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

www.sciencemag.org/cgi/content/full/321/5888/557/DC1

Supporting Online Material for

BSKs Mediate Signal Transduction from the Receptor Kinase BRI1 in Arabidopsis

Wenqiang Tang, Tae-Wuk Kim, Juan A. Oses-Prieto, Yu Sun, Zhiping Deng, Shengwei Zhu, Ruiju Wang, Alma L. Burlingame, Zhi-Yong Wang*

*To whom correspondence should be addressed. E-mail: zywang24@stanford.edu

Published 25 July 2008, Science 321, 557 (2008)

DOI: 10.1126/science.1156973

This PDF file includes:

Materials and Methods Figs. S1 to S13 References

1

SUPPLEMENTAL ONLINE MATERIALS This PDF file includes: Materials and Methods Figs. S1 to S13 References MATERIALS AND METHODS Materials The det2-1 and bzr1-1D mutants are in Arabidopsis thaliana Columbia ecotype background, and bri1-5 is in the WS ecotype. For hypocotyl and root growth assay, seedlings were grown on vertical phytoagar plate containing ½ MS medium and 1% sucrose (pH 5.7) in the dark for 4 days or under continuous light for 7 days. Confocal miscroscopy and BiFC analysis were performed according to Gampala et al (S1). Plasma membrane isolation, 2-D DIGE, and mass spectrometry Growth and BR treatment of det2 mutant seedlings, plasma membrane protein isolation, and 2D-DIGE and LC-MS/MS analyses were performed as described previously Tang et al. (S2). Immunoblotting of 2-DE Leaves from 3-week-old transgenic det2 or bri1-5 mutant seedlings expressing the BSK1-YFP fusion protein were cut into two pieces along the middle vein. Each half was treated for 15 min with mock solution (0.01% ethanol) or 100 nM BL. Total protein was extracted and separated by 2-DE using a 7cm or 24 cm pH 4-7 immobilized pH gradient gel (IPG) strip, and a 10% SDS PAGE gel. The proteins were blotted to a nitrocellulose membrane and detected using anti-YFP antibodies. GST tagged protein purification and in vitro phosphorylation assays GST tagged BSK1, BSK1-ΔTPR, BSK1-S230A, BSK3, BRI1 kinase domain (BRI1-KD), BAK1 kinase domain (BAK1-KD), kinase dead BRI1 (mBRI1-KD) or BAK1 kinase domain (mBAK1-KD), were expressed in E. coli and purified by standard procedure using glutathione agarose beads. In brief, 10 ml overnight grown BL 21 cells expression the desire constructs were transferred into 500 ml of LB and grown at 37 degree for three more hours. 0.5 mM IPTG was then added into the culture to induce the protein expression for two more hours. The bacteria cells were sonicated in PBS with 1% Triton X-100 and centrifuge at 20,000 g for 15 min to remove the insoluble cell debris. Suppernatant was incubated with PBS pre-equilibrated glutathione agarose beads and rotate at cold room for 2 hours. After wash with PBS for 5 times, the GST tagged proteins were eluted using 5 mM Glutathione. Eluted protein was concentrated by ultrafiltration using centricon centrifuge tubes with a 10 kD molecular weight cut-off. In vitro phosphorylation assays were performed in a mixture containing 20 mM Tris-Cl, pH 7.5, 10 mM MgCl2, 100 mM NaCl, 1 mM DTT, 100 uM ATP, 10 μCi 32P-γATP and 0.5 μg GST-BRI1-KD or GST-BAK1-KD with or without 15 μg GST tagged BSKs. One

2

microgram of GST tagged mBRI1-KD or mBAK1-KD was used. The reaction mixtures were incubated at 30 °C for 3 hr with constant shaking. Kinase reactions were stopped by adding equal volume of 2x SDS sample buffer and boiling for 5 minutes. The protein phosphorylation was analyzed by SDS-PAGE and autoradiography. Anti-BSK1 antibodies. N terminal 87 amino acid of BSK1 were purified from E coli as a GST-fusion protein (GST-BSK1N87) following the procedure described above. Purified protein was separated using 12% SDS-PAGE. After Commassie Blue staining, GST-BSK1N87 protein was cutted from the gel and used to immunize rabbits. The antibodies were affinity purified using an antigen column containing immobilized GST-BSK1N87 protein. The antibodies specifically detect GST-BSK1 but not GST-BSK3. Co-Immunoprecipitation: All procedures were performed at 4 °C. Rosette leaves of 3-week-old Arabidopsis plants were homogenized with mortar and pestle in the grinding buffer (20 mM HEPES, pH 7.5, 40 mM KCl, 250 mM sucrose, 5 mM MgCl2 and protein inhibitors) at a ratio of 3 mL buffer per gram of tissue. Homogenate was filtered through 2 layers of Miracloth (CalBiochem, La Jolla, CA), and centrifuged at 10,000 g for 10 minutes. The supernatant was further centrifuged at 100,000 g for 60 min to pellet the microsomal fraction. Microsomes were re-suspended in 0.5 mL grinding buffer and sonicated for 10 sec. Triton X-100 was added to a final concentration of 1%. After mixing, the extracts were centrifuged at 20,000 g for 10 min. Supernatant was diluted 5 times with grinding buffer to reduce Triton X-100 to 0.2%. Affinity-purified anti-YFP or anti-BSK1 antibodies bound to Protein A Sepharose beads were added, and the mixture was rotated in a cold room for 30 min. Beads were washed 6 times with 0.5 mL grinding buffer each and eluted by boiling directly with SDS sample buffer for 5 min. Samples were analyzed by SDS-PAGE and immunoblotting. Knockout mutants of BSKs. T-DNA knockout mutants Salk_001600 (bsk2-3), Salk_105500 (bsk2-2), Salk_096500 (bsk3-1), Salk_032845 (bsk4), Salk_074467 (bsk5), Salk_051462 (bsk12) were obtained from the Arabidopsis Biological Resource Center (ABRC, www.arabidopsis.org) (S3). Mutants were first genotyped with gene specific primers and T-DNA primer LBb1 5’-GCGTGGACCGCTTGCTGCAACT-3’ to confirm the position of T-DNA insertion site. The primers used for different bsk knockouts were: BSK2-2GR: CAAACTCAGGGACGCTTATTCT; BSK2-F: CACCATGGGTTGTTTACATTCCAAAACTGC; BSK1-3GF: CGAATCGCAGACTATATTGCAG; BSK1-3GR: ATGTGTTTGCCGCTTAAAAGAT; BSK3-1GF: TGCTTCGATTCCCAAAATTTAC; BSK3-1GR: ATATACTCCACGGCAAAACCAG; BSK4-1GF: AATAGTCGGGATGGGAAAAGTT; BSK4-1GR: ACTTTGAGGCAGACCCATTAGA; BSK5-2GF: CATTTCTCAAACGATGATGGAA; BSK5-2GR: ATTTGCTACACCCTCGTCATCT;

3

BSK12-GR: TGTTGCAATAATCCAATGCTTC; BSK12-F: CACCATGGGTTGTTGTTACTCACTATCTTC. Expression of BSK3 RNA in wild type and bsk3-1 mutant was analyzed by RT-PCR (30 cycles) using gene specific primer pairs: BSK3-F, ATGGGAGGTCAATGCTCTAGCCTGAGT and BSK3-R, CTTCACTCGGGGAACTCCATTCATCTTTG. Overexpression of BSKs Full-length cDNA coding sequences of BSKs without the stop codon were amplified by PCR using gene-specific primers and full-length cDNA clones as templates (BSK1: U14968 , BSK2: U67754, BSK3: U16452, BSK5: U85779, BSK6: U16452) (S4). The cDNAs were cloned into pENTR/SD/D-TOPO vectors (Invitrogen), and then sub-cloned between the 35S promoter and YFP coding sequence in the destination plasmid pEarleyGate 101 (S5), or between the 35S promoter and the 4xMyc tag in the pGWB17 vector, using Gateway recombination cloning reaction (Invitrogen). The fusion constructs were transformed into agrobacterium strain GV3101 and then transformed into Arabidopsis plants by floral dipping. Quantitative reverse-transcription PCR Total RNA extraction and quantitative real-time RT-PCR analysis of DWF4 expression was performed as described by He et al (S7).

4

SUPPLEMENTAL REFERENCES S1. S. S. Gampala et al., Dev Cell 13, 177 (Aug, 2007). S2. W. Tang et al., Mol Cell Proteomics Epub ahead of print (Jan 8, 2008). S3. J. M. Alonso et al., Science 301, 653 (Aug 1, 2003). S4. K. Yamada et al., Science 302, 842 (Oct 31, 2003). S5. K. W. Earley et al., Plant J 45, 616 (Feb, 2006). S6. T. Nakagawa et al., J Biosci Bioeng 104, 34 (Jul, 2007). S7. J. X. He et al., Science 307, 1634 (Mar 11, 2005). S8. D. Winter et al., PLoS One 2(8): e718 (2007).

Fig. S1. A model of the BR signal transduction pathway. Components in active statesare in red color and inactive state are in blue color. Arrows show positive action, and barends show inhibitory action. In the absence of BR (-BR, left side), both BRI1 and BAK1 areinactive; BKI1 binds BRI1 and inhibits its interaction with BAK1; BIN2 phosphorylates BZR1and BZR2 (also named BES1) (BZR1/2). Phosphorylated BZR1 and BZR2 (p-BZR1/2)cannot bind DNA and are retained in the cytoplasm by the 14-3-3 proteins. In the presenceof BR (+BR, right side), BR-binding to the extracellular domain of BRI1 activates BRI1kinase by inducing dimerization and inter-phosphorylation with BAK1 and dissociation ofBKI1. Activation of BRI1 and BAK1 kinases leads to dephosphorylation and activation ofBZR1 and BZR2, presumably through activation of the BSU1 phosphatase or inhibition ofBIN2. How the receptor kinases regulate the downstream components has remained anoutstanding question. In this study, we show that BSKs are the substrates of BRI1 kinasethat transduce the signal from BRI1 to downstream components. BSKs associate with BRI1in the absence of BR, and are phosphorylated by BRI1 and released from the BRI1 complexupon BR activation of BRI1 kinase. How BSKs regulate downstream components such asBSU1 or BIN2 remains to be further studied.

BIN2

p-BZR1/2 BZR1/2

BZR1-target genes

BSU1

BRI1 BRBAK1

BIN2

BSKsp p

ppBSKs pp p

p

BSU1

BRI1

BAK1

?

+BR-BR

BKI1BKI1

pp

pp

p

p

pp14-3-3

BSU?BIN2

Growth responses

A100

Rel

ativ

e in

tens

ity

C

B

0

100

0

100

0

m/z

Fig. S2. Representative tandem mass spectra obtained from precursor ions with m/z value A)593.8805+2, B) 635.3947+2, and C) 789.4522+2, corresponding respectively to: A), peptideDLVATLAPLQTK, spanning residues D322 to K333 of protein BSK1 (At4g35230); B), peptideLLVAEFmPNDTLAK, spanning residues L146 to K159 of protein BSK1 (At4g35230); C), peptideFLLSAVAPLQK, spanning residues F302 to K312 of protein BSK2 (At5g46570). The observedsequence ions are displayed. Water or ammonium losses are marked with stars. Int, internalions. m, oxidized methionine.

Unconserved Conserved

300

400

500

Fig. S3. Alignment of amino acid sequences of BSKs. Solid underlines show the putativekinase domain, and the dashed lines show the tetratricopeptide repeat (TPR) domain. Thenames of different BSKs are: BSK1 (At4g35230), BSK2 (At5g46570), BSK3 (At4g00710),BSK4 (At1g01740), BSK5 (At5g59010), At3g54030 (BSK6), At1g63500 (BSK7), At5g41260(BSK8), At3g09240 (BSK9) At5g01060 (BSK10), At1g50990 (BSK11), and At2g17090(BSK12). Sequences are aligned using the PRALINE multiple sequence alignment tool withstandard progressive strategy (http://zeus.cs.vu.nl/programs/pralinewww).

BSK1BSK11BSK2BSK3BSK4BSK5BSK6BSK7BSK8BSK9

BSK10BSK12

Consistency

BSK1BSK11BSK2BSK3BSK4BSK5BSK6BSK7BSK8BSK9

BSK10BSK12

Consistency

BSK1BSK11BSK2BSK3BSK4BSK5BSK6BSK7BSK8BSK9

BSK10BSK12

Consistency

BSK1BSK11BSK2BSK3BSK4BSK5BSK6BSK7BSK8BSK9

BSK10BSK12

Consistency

BSK1BSK11BSK2BSK3BSK4BSK5BSK6BSK7BSK8BSK9

BSK10BSK12

Consistency

BSK1BSK11BSK2BSK3BSK4BSK5BSK6BSK7BSK8BSK9

BSK10BSK12

Consistency

526

200

Fig. S4. Mass spectrometry analysis of BRI1 phosphorylation of BSK1. GST-BRI1-KDand GST-BSK1 proteins were expressed and purified in E. coli, and mixed in an in vitrokinase reaction. The proteins were then digested by trypsin and analyzed by LC-MS/MS.Shown is a representative tandem mass spectrum obtained from precursor ions forpeptide SYSTNLAYTPPEYLR, spanning the residues Ser-228 to Arg-242 of BSK1. Thephosphorylated residue (S230) is indicated in the peptide sequence as Sp. Phosphatelosses are marked with dots.

Fig. S5. Co-immunoprecipitation of BRI1 with BSK1. Microsomal proteins(Input, lane 1) from det2 mutant plants expressing BRI1-GFP under thecontrol of BRI1 promoter were immunoprecipitated (IP) using polyclonal anti-BSK1 antibodies affinity-purified using an BSK1 antigen column (lane 3).Protein A bead alone was used as IP control (lane 2). Immunoblots wereprobed using anti-GFP antibodies. BRI1-GFP can be pulled down by anti-BRI1 antibody but not by protein A bead.

1 2 3

Fig. S6. The T-DNA insertion mutant bsk3-1 shows short-hypocotyl phenotypes when grown in the dark. Seedlings ofwild type (Col) and mutants were grown in the dark on regularmedium (-BRZ) or medium containing 2 µM brassinazole(+BRZ). Error bars show standard errors. Stars markstatistically significant differences from Col.

0

2

4

6

8

10

12

14

bsk4 bsk2-3 bsk2-2 bsk3-1 bsk12 bsk5 Col

Hypo

coty

l leng

th (m

m) -BRZ +BRZ

*

*

Fig. S7. BL responsiveness of SAUR-AC and DWF4 expression is reduced in bsk3-1 knockout. Col (A) and bsk3 (B) seedlings were grown on 1/2x MS medium for 5days under continuous light and treated with mock solution or 100 nM BL for 2 hrs.mRNA was extracted and the expression level of SAUR-AC and DWF4 genes wereanalyzed by real time RT-PCR. The data represent the average from 4 independentbiological repeats. Error bars represent standard deviation.

A B

Fig. S8. BL regulates the post-translational modification of BSK3. One-week-old det2seedlings expressing a BSK3-YFP under the control of 35S promoter were treated with 100nM BL or mock solution for 1hr. Microsomal protein were isolated and separated by IEFusing a 7 cm pH 4-7 IPG strip, followed by 10% SDS-PAGE. After transferring tonitrocellulose membrane, BSK3-YFP protein were detected using GFP antibody. Bluearrows show new BSK3-YFP spots induced after BL treatment. Red arrows showed spotsthat show reduced intensity after BL treatment.

pH

Fig. S9. BSK3 can be phosphorylated by BRI1, but not by BAK1 in vitro. Invitro kinase assays were performed by incubating the full-length BSK3,BRI1 kinase domain and BAK1 kinase domain, individually or incombination as shown. All three proteins were purified from E. coli as GSTfusion proteins. Upper gel image shows autoradiography of 32P labeling,and the lower image shows total protein stained by Coomassie BrilliantBlue.

BSK3 + + + - - + + - - BRI1 + - - + - - - - - mBRI1 - + - - + - - - -BAK1 - - - - - + - + -mBAK1 - - - - - - + - +

Fig. S10. Co-immunoprecipitation of BRI1 with BSK3. Microsomal proteins(Input) from det2 mutant plants (lane 1 and 4) or det2 expressing BSK3-YFP(lanes 2, 3, 5 and 6) were immunoprecipitated (IP) using anti-YFP antibodies.Immunoblots were probed using anti-BRI1 or anti-GFP antibodies. Seedlingswere either treated with mock solution (BL-) or with 100 nM brassinolide(BL+) for 30 minutes. Band of BRI1 in IP of the mock-treated sample (lane 5)is marked with a star, and this band is not detected in IP of the BL-treatedsample (lane 6).

*

BL - - + - - +

BRI1

BSK3-YFP

1 2 3 4 5 6

Input IP

Fig. S11. Comparison of tissue specific expression between BSK1 and BRI1 (A), orbetween BSK3 and BRI1 (B). Figures were generated on line using e-FP browser from(http://bbc.botany.utoronto.ca/efp/cgi-bin/efpWeb.cgi) (Winter et al., 2007. PLoS One 2(8):e718). Relative expression levels of BSK vs BRI1 in various organs are shown by a colorscale - red color indicate higher BSK expression, blue for higher expression of BRI1, andyellow shows similar expression level.

A

B

Fig. S12. Overexpression of BSK1 and BSK3 suppresses the bri1-5mutant. A. Phenotypes of three-week-old wild type (WS, left), bri1-5(right), and bri1-5 overexpressing BSK1-YFP (middle two). B.Expression of the DWF4 RNA determined by quantitative RT-PCR.Insert shows immunoblot of BSK1-YFP and staining of Rubisco in thetwo transgenic lines. C and D, same as A and B, but foroverexpression of BSK3-YFP. Bars in A and C are 5 mm. E.Overexpression of BSK3 (line D4 and G9) suppresses dark-grownphenotypes of bri1-5. Seedlings were grown on MS medium in thedark for 4 days.

0

1

2

3

4

5

6

7

8

WS BSK1-OXB BSK1-OXE bri1-5

Rel

ativ

e D

WF4

RN

A le

vel

BSK1-YFP

Rbcs

A

B

0

1

2

3

4

5

6

7

8

WS BSK3-OXA BSK3-OXC bri1-5

Rel

ativ

e D

WF4

RN

A le

vel

BSK3-YFPRbcs

C

D

bri1-5 WS D4 G9

Hypo

coty

l leng

th (m

m)

02468

1012

bri1-5 WS D4 G9

E

Fig. S13. Overexpression of BSK3 suppresses the det2-1 mutant. A. Three-week-old wild type (Col), det2-1 mutant overexpressing BSK3-YFP (BSK3-ox), and det2-1. B and C, Seedlings of det2-1, Col, and three independentlines (CE, GC, AD) of det2-1 transformed with 35S-BSK3-YFP were grown inthe dark for 4 days. Bar in panel B is 5 mm.

Col BSK3-OX det2

0

2

4

6

8

10

det2 Col CE GC AD

Hyp

ocot

yl le

ngth

(mm

)

A

B

C