Protein kinase CK2 links extracellular growth factor signaling ... - Nature

ARTICLES

© 2008 Nature Publishing Group http://www.nature.com/naturemedicine

Protein kinase CK2 links extracellular growth factor signaling with the control of p27Kip1 stability in the heart Ludger Hauck1,2, Christoph Harms3, Junfeng An5, Jens Rohne5, Karen Gertz3, Rainer Dietz5,6, Matthias Endres3,4 & Ru¨diger von Harsdorf1,2 p27Kip1 (p27) blocks cell proliferation through the inhibition of cyclin-dependent kinase-2 (Cdk2). Despite its robust expression in the heart, little is known about both the function and regulation of p27 in this and other nonproliferative tissues, in which the expression of its main target, cyclin E–Cdk2, is known to be very low. Here we show that angiotensin II, a major cardiac growth factor, induces the proteasomal degradation of p27 through protein kinase CK2-a¢–dependent phosphorylation. Conversely, unphosphorylated p27 potently inhibits CK2-a¢. Thus, the p27–CK2-a¢ interaction is regulated by hypertrophic signaling events and represents a regulatory feedback loop in differentiated cardiomyocytes analogous to, but distinct from, the feedback loop arising from the interaction of p27 with Cdk2 that controls cell proliferation. Our data show that extracellular growth factor signaling regulates p27 stability in postmitotic cells, and that inactivation of p27 by CK2-a¢ is crucial for agonist- and stressinduced cardiac hypertrophic growth.

The tumor suppressor p27 is a potent inhibitor of cell growth and division1. Antiproliferative signals lead to accumulation and stabilization of p27, which can then inhibit Cdk2 and cause cell cycle arrest2–6. p27 protein expression is mainly regulated by posttranslational mechanisms7–9. At least three pathways can mediate its ubiquitin-dependent degradation, each of which operates during a specific time point within the cell cycle and in a specific subcellular compartment. In quiescent cells, p27 is translocated to the cytoplasm where it is ubiquitinated by the E3–ubiquitin ligase complex KPC1/2 and proteasomally degraded10–12. This process depends on growth factors and does not require Cdk2-dependent phosphorylation of p27. In the early G1 phase of the cell cycle, phosphorylation of p27 at Ser10 promotes nuclear export of p27 and its cytoplasmic degradation13–15. Of note, p27 may also be degraded in the nucleus, independently of Ser10 phosphorylation16. Sequestration of p27 into cyclin D–Cdk4 or cyclin D–Cdk6 complexes, without inhibition of the enzymatic activity of these complexes, also participates in the initial activation of cyclin E–Cdk2 (refs. 17,18). Later in the G1 and S phases, another nuclear E3–ubiquitin ligase complex containing Skp2 (refs. 19–21) recognizes p27 phosphorylated at Thr187 by cyclin E–Cdk2 (refs. 18,22,23), thereby promoting its proteasomal degradation19–21. Collectively, these mechanisms establish a positive feedback loop, amplifying Cdk2 activity through inactivation of inhibitory p27. The role of p27 in growth control in vivo is also noteworthy. Genetic deletion of p27 in mice results in multiorgan hyperplasia and tumor development24–26. In cancer patients, increased degradation of p27 correlates with aggressive tumors and poor prognosis27. Additionally,

Akt-dependent p27 phosphorylation on Thr157 impairs its nuclear translocation, leading to cytoplasmic p27 accumulation and sustained cell proliferation in human breast cancer28–30. Quiescent cells actively suppress terminal differentiation, ensuring the reversibility of cell cycle exit. In contrast, differentiated postmitotic cells are substantially refractory to reactivation of the cell cycle, exemplified by the absence of significant proliferative potential in cardiomyocytes31. The lack of regenerative capacity of adult mammalian cardiomyocytes is thought to be caused by the unavailability of G1 cyclin–Cdks, crucial positive modulators of the cell cycle, and high levels of inhibitory p27 (refs. 31–33). When exposed to aberrant growth stimuli, the heart undergoes maladaptive changes characterized by hypertrophic growth34. This process involves cell enlargement, myofibrillar disarray and reexpression of fetal genes, ultimately leading to heart failure and death35. Because p27 is highly expressed in adult myocardium despite this tissue’s lack of p27’s main target, cyclin E–Cdk2 (refs. 31–33), we hypothesized that p27 may exert a growth regulatory function in cardiomyocytes through alternative pathways linking extracellular growth stimuli to p27. Here we show that p27 is a target of hypertrophic signaling in cardiomyocytes through intracellular activation of CK2-a¢. RESULTS CK2-a¢ interacts with p27 in vitro and in cardiomyocytes We screened an adult human heart cDNA library for proteins that interact with p27 using a Gal4-based yeast two-hybrid system. One such clone contained the C-terminal amino acid residues 167–350 of

1Division of Cardiology, University Network Hospitals and Toronto General Research Institute, 2McEwen Centre for Regenerative Medicine, 200 Elizabeth Street, Toronto M5G 2C4, Canada. 3Department of Neurology, 4Center for Stroke Research Berlin, Charite´-Universitaetsmedizin Berlin, Charite´platz 1, 10117 Berlin, Germany. 5Department of Cardiology, Campus Virchow Clinic, Charite`, Humboldt University, Augustenburger Platz 1, 13353 Berlin, Germany. 6Max-Delbru¨ck Center for Molecular Medicine, Robert-Rossle-Strasse 10, 13092 Berlin, Germany. Correspondence should be addressed to R.v.H. ([email protected]).

Received 31 July 2007; accepted 22 January 2008; published online 2 March 2008; corrected after print 21 March 2008; doi:10.1038/nm1729

NATURE MEDICINE

VOLUME 14

[

NUMBER 3

[

MARCH 2008

315

ARTICLES

kDa 45 32 25 18

p27 anti-His

45 198

0.0


198

198 186 151

151 86

86 87

p27 Gal4-AD WT CK2-α′

198

186

25 18

1 1 1 1 25 81 151 186 Cdk2

CK2-α′

87 1

1

1

1

e lP

WT p27

f kDa 116

lP

66

C K2 -α ′ Ly sa t pr e elS

32

Merged

p2 7

0.4 0.2

66

CK2-α′

p27

Ly sa t pr e elS

CK2-α′ anti-His

d – Ang II

0.8 0.6

Gal4-BD

Input

GST-CK2-α′

kDa

+ Ang II

1.0

c

GST-p27

*

C K2 p2 α′ 7

b

1.2

C K2 -α Be ′ ad s In pu t

β-gal activity (absorption at 405 nm)

a

kDa 116 66

Figure 1 A cytoplasmic complex of CK2-a¢ and p27 exists in cardiomyocytes. (a) The 45 45 CK2-α′ interaction between CK2-a¢ and p27 in the yeast two-hybrid system reconstituted a functional 32 Gal4-transcriptional activator leading to expression of the b-galactosidase reporter. b-Gal, 32 p27 25 25 b-galactosidase. AD, Gal4-transactivation domain. BD, Gal4-binding domain. Data are means 18 18 ± s.e.m., n ¼ 4. *P o 0.005. (b) Recombinant CK2-a¢ and p27 proteins interacted in vitro. 14 14 His-tagged WT CK2-a¢ was incubated with GST-conjugated WT p27 immobilized on glutathione-Sepharose beads. Samples were immunoblotted with antibody to the His tag (antiWB p27 WB CK2-α′ His). (c) The cyclin-Cdk2–binding site of p27 was not required for its interaction with CK2-a¢. In vitro–translated mutants of p27 were tested for binding to WT CK2-a¢ immobilized on GST beads (left). p27 was detected by immunoblot analysis with anti-His. Input amounts of p27 are shown on the right. (d) CK2-a¢ and p27 colocalized in the cytoplasm of isolated primary rat ventricular cardiomyocytes. Cells were incubated with angiotensin II (Ang II; 100 nM) for 12 h and costained with specific antibodies as indicated. Scale bar, 20 mm. (e,f) CK2-a¢ was bound to p27 in cardiomyocytes. Cardiomyocyte cytoplasmic extracts (1.5 106 cells/lane) were subjected to immunoprecipitation with specific antibodies covalently conjugated to protein A–agarose. Alternatively, cytoplasmic extracts (Lysate; 4 105 cells/lane) were subjected to immunoblotting. Immunoblotting was performed with anti-p27 (e) and anti–CK2-a¢ (f). pre-IS, pre–immune serum. IP, immunoprecipitation. WB, western blot.

CK2-a¢. This ubiquitous serine-threonine kinase36,37, in combination with wild-type p27 (WT p27), reconstituted Gal4-dependent transcriptional activation of the b-galactosidase reporter (Fig. 1a). Pulldown affinity assays showed that recombinant CK2-a¢ could bind to the carboxy terminus of p27 independently from the N-terminal cyclin E/A–Cdk2 interaction site of p27 (ref. 38; Fig. 1b,c). Immunocytochemical analysis revealed that CK2-a¢ colocalized with p27 in the cytoplasm of unstimulated cardiomyocytes (Fig. 1d). Addition of angiotensin II (100 nM), a potent inducer of cardiomyocyte hypertrophy34, caused downregulation of p27, whereas CK2-a¢ levels remained unaffected. Endogenous CK2-a¢ and p27 could be coimmunoprecipitated from cardiomyocyte cytoplasmic extracts (Fig. 1e,f), providing additional evidence that p27 can bind CK2-a¢. CK2-a¢ phosphorylates p27 at Ser83 and Thr187 in vitro Inspection of the p27 primary sequence with Scansite (http://scansite. mit.edu) and eukaryotic linear motif prediction algorithms (http:// elm.eu.org) revealed the presence of two consensus sites for CK2 phosphorylation at Ser83 and Thr187, the latter being a nonoptimal CK2 recognition motif. Both sites are conserved in the primary sequences of human, mouse and rat p27. The Ser83 site is also present in the related Kip family member p57Kip2, but not in p21Cip1/Waf1. To test whether p27 is phosphorylated by CK2-a¢, we incubated recombinant WT p27 or mutant versions of p27, in which either the Ser83 or Thr187 residue had been changed to an alanine residue, with recombinant active CK2-a¢ in in vitro kinase reactions. WT p27 was phosphorylated by CK2-a¢ but not catalytically inactive kinase-dead CK2-a¢ (KD CK2-a¢; Fig. 2a). The S83A and T187A mutants had a 68% and 27% reduction in phosphorylation by CK2-a¢, respectively, indicating that Ser83 is the dominant CK2-a¢ phosphorylation site in vitro. In contrast, CK2-a¢ did not phosphorylate p27DPi, in which both Ser83 and Thr187 were changed to alanines (S83A,T187A). Thus,

316

phosphorylation of p27 by CK2-a¢ in vitro requires the Ser83 and Thr187 residues in p27. Phosphorylation of p27 impairs its interaction with CK2-a¢ In proliferating cells, p27 is not only a substrate but also an inhibitor of Cdk2 (refs. 18,39,40). We were interested in finding out whether a similar feedback loop exists in postmitotic cardiomyocytes. WT p27 completely inhibited recombinant CK2-a¢ activity in vitro (Fig. 2b). To test whether phosphorylation of p27 by CK2-a¢ at residues Ser83 or Thr187 could affect the ability of p27 to inhibit CK2-a¢ activity, we used p27 variants prephosphorylated by CK2-a¢ at either or both of these residues: p27Pi-Ser83 (p27 phosphorylated only at Ser83), p27Pi-Thr187 (p27 phosphorylated only at Thr187) and p27Pi-Ser83,Pi-Thr187 (p27 phosphorylated at both Ser83 and Thr187) were generated by phosphorylating T187A p27, S83A p27 and WT p27, respectively, with CK2-a¢. p27Pi-Ser83 and p27Pi-Thr187 had less inhibitory activity toward CK2-a¢ compared to WT p27, with p27Pi-Thr187 having a more pronounced effect (Fig. 2b). Phosphorylation of both Ser83 and Thr187 in p27Pi-Ser83,Pi-Thr187 or introduction of phosphomimetic mutations into p27 (p27S83D,T187D) abolished the ability of p27 to inhibit CK2-a¢ (Fig. 2b). Next, we determined whether this effect was correlated with decreased binding of phosphorylated p27 to CK2-a¢. We consistently found that CK2-a¢ immunoprecipitates contained less p27Pi-Ser83 or p27Pi-Thr187 compared to WT p27, with a more pronounced effect on p27Pi-Thr187 (Fig. 2c). Conversely, we observed less p27Pi-Ser83– or p27Pi-Thr187–bound CK2-a¢ in immunocomplexes compared to WT p27–bound CK2-a¢, with again a more pronounced effect on p27Pi-Thr187 (Fig. 2c). We detected only minimal amounts of CK2-a¢ associated with p27Pi-Ser83, p27Pi-Thr187 or p27S83D,T187D (Fig. 2c). Thus, impaired CK2-a¢ inhibition by phosphorylated p27 seems to be due to its reduced binding affinity to CK2-a¢. These findings suggest that the phosphorylation state of p27 at Ser83 and Thr187 is crucial for p27 binding to CK2-a¢.

VOLUME 14

[

NUMBER 3

[

MARCH 2008

NATURE MEDICINE

ARTICLES

*

32

[ P]p27 (%)

100

40 20

## – + + – – –

+ – + – – –

+ + + – – – – – – + – – – + – – – +

p27Pi-T187 p27Pi-S83

e IP CK2-α′ WB p27

**

[32P]p27


c

**

60

0 CK2-α′ KD CK2-α′ WT p27 p27S83A p27T187A p27∆Pi

Input

b

#

80

CK2-α′ kinase activity (%)

a

#

IP p27 WB CK2-α′ WT p27 p27Pi-S83 p27Pi-T187 p27Pi-S83, Pi-T187 p27S83D, T187D

100 80

#

*

60 #

40

*

20

0 WT p27 – p27Pi-S83 – p27Pi-T187 – p27Pi-S83, Pi-T187 – p27S83D, T187D –

+ – – – –

– + – – –

## + – – – –

– + – – –

– – + – –

– – – + –

– – – – +

d Time (h) 0

[32P]histone H1

– – + – –

– – – + –

– – – – +

Control siRNA

p27 Pi-S83 Pi-T187

CK2-α′ siRNA

p27 Pi-S83 Pi-T187 Ang II CIP

DMAT – –

+ –

p27 Pi-S83 Pi-T187 p27 Pi-S83 Pi-T187 Ang II CIP

KD CK2-α′

CK2-α′ siRNA + CK2-α′ – –

+ –

+ +

+ +

Cytoplasm Nucleus Ang II Ang II 1 2 4 6 8 12 16 20 24 0 1 2 4 6 8 12 16 20 24

p27 p27Pi-T187

CK2-α′

p27Pi-S83

p27

p27

– – – – –

p27 Pi-S83 Pi-T187

CK2-α′ 32

CK2-α′

[ P]histone H1 Tubulin Lamin

Figure 2 CK2-a¢ phosphorylates p27 on Ser83 and Thr187. (a) The ability of CK2-a¢b to phosphorylate recombinant p27 protein variants was examined using in vitro kinase assays. Aliquots were blotted to a membrane, and the amount of g-32P incorporated into p27 was quantified with a PhosphorImager and ImageJ software. Input amounts of p27 and CK2-a¢ were monitored by probing the membranes with anti-p27 and anti–CK2-a¢ antibodies, respectively. The amount of phosphorylation on Thr187 or Ser83 was determined using phosphorylation site–specific antibodies to p27. Data are means ± s.e.m., n ¼ 3. *P o 0.001 versus KD CK2-a¢. **P o 0.05 versus WT p27. #P o 0.01. ##P o 0.001 versus WT p27. (b) Phosphorylation of p27 by CK2-a¢ impaired the ability of p27 to inhibit CK2-a¢. WT p27 (100 ng) or p27 protein variants were incubated with recombinant CK2-a¢b (30 ng) in the presence of [g-32P]ATP and histone H1 substrate as indicated. Input amounts of recombinant CK2-a¢ and p27 proteins were monitored by immunoblot analysis with antiHis. Data are means ± s.e.m., n ¼ 3. *P o 0.01 versus CK2-a¢ control (green). #P o 0.01. ##P o 0.001 versus CK2-a¢ control. (c) Phosphorylation of p27 impaired its interaction with CK2-a¢. CK2-a¢ kinase assays from b were subjected to IP analysis as indicated. (d) Ang II induced CK2-a¢ kinase activity, p27 phosphorylation and downregulation of p27 in cardiomyocytes. Isolated cardiomyocytes were treated with Ang II for the indicated time points prior to lysis. Biochemically fractionated extracts (4 105 cells/lane) were immunoblotted with the antibodies indicated. Anti-CK2-a¢ immunocomplex kinase assays were performed with histone H1 as substrate. (e) Ang II induced phosphorylation of p27 at Ser83 and Thr187 that was abolished by silencing or inhibition of CK2-a¢. Cardiomyocytes were transduced with the indicated lentiviruses or preincubated with DMAT (20 mM) for 30 min before Ang II treatment for 4 h. Alternatively, CK2-a¢ siRNA–expressing cardiomyocytes were transduced with TAT-conjugated CK2-a¢ before Ang II addition. Total cell extracts (50 mg/lane) were immunoblotted with anti-p27 and anti–phospho-p27. The reactivity of anti–phospho-p27 antibodies in the immunoblot was abrogated by treatment of lysates with calf intestine phosphatase (CIP) before immunoblotting.

Activation of CK2-a¢ decreases p27 protein expression We examined the kinetics of the CK2-a¢ response to angiotensin II and its relationship to p27 phosphorylation and turnover in cardiomyocytes. Cytoplasmic CK2-a¢–dependent kinase activity was induced by angiotensin II within 1 h; this overlapped with induction of p27 phosphorylation at Ser83 and Thr187, which also occurred within 1 h after angiotensin II treatment (Fig. 2d). Phosphorylation of nuclear p27 at these residues was not observed under the conditions used (Fig. 2d). Induction of CK2-a¢ activity was associated with a progressive decrease in p27 abundance over the next 12 h (Fig. 2d). These findings are compatible with the concept that an inhibited p27–CK2-a¢ complex resides in the cytoplasm awaiting angiotensin II–mediated activation and that CK2-a¢ participates in p27 turnover. Inactivation of CK2-a¢ abolishes p27 phosphorylation Silencing of CK2-a¢ by RNA interference (CK2-a¢ small interfering RNA (siRNA); Supplementary Fig. 1 online) or inhibition of CK2-a¢ function by lentivirus-mediated expression of KD CK2-a¢ or treatment with the pharmacological inhibitor DMAT (2-dimethylamino-4,5,6, 7-tetrabromo-1H-benzimidazol) abolished Ser83 and Thr187 phosphorylation on p27 (Fig. 2e). To further examine the requirement of CK2-a¢ for p27 phosphorylation, we re-introduced CK2-a¢ into CK2a¢ siRNA–expressing cardiomyocytes, employing transactivator of transcription (TAT)-mediated protein transduction41 of CK2-a¢. To ensure cytoplasmic localization, this CK2-a¢ construct was engineered with the strong nuclear export sequence of HIV-1 Rev (LPPLERLTL) at its N terminus (Supplementary Methods online). Protein transduction of CK2-a¢ rescued the detrimental impact of CK2-a¢ siRNA on p27

NATURE MEDICINE

VOLUME 14

[

NUMBER 3

[

MARCH 2008

phosphorylation in angiotensin II–treated cardiomyocytes (Fig. 2e). These findings support the view that p27 phosphorylation in vitro is largely dependent on CK2-a¢ activity. Phosphorylation-resistant p27 inhibits CK2-a¢ in cardiomyocytes To determine the influence of p27 phosphorylation on CK2-a¢ activity in cardiomyocytes, we generated TAT-conjugated p27 constructs carrying a nuclear export sequence to ensure their cytoplasmic localization (Supplementary Methods). p27DPi, which is refractory to CK2-a¢–dependent phosphorylation, completely blocked CK2-a¢ activation by angiotensin II (Fig. 3a). Transduced WT p27 inhibited CK2-a¢ less effectively compared to p27DPi, associated with a decreased steady-state abundance of WT p27 (Fig. 3a). The phosphomimetic mutant p27Ser83D,S187D was unstable and did not have an inhibitory effect on CK2-a¢ (Fig. 3a). We conclude that phosphorylation of p27 impairs its ability to inhibit CK2-a¢ activity. CK2-a¢ mediates angiotensin II–triggered p27 degradation Growth factor–activated Cdk2 phosphorylates p27, which is necessary for p27’s ubiquitination and proteasomal degradation19,20. This led us to investigate whether CK2-a¢ regulates p27 turnover in cardiomyocytes. The angiotensin II–triggered decrease in p27 protein abundance was sensitive to proteasomal inhibition by lactacystin (Fig. 3b). In lactacystin-treated cells, endogenous CK2-a¢ remained catalytically active despite p27 accumulation, consistent with the hypothesis that Ser83 and Thr187 phosphorylation on p27 impairs its capacity to inhibit CK2-a¢. In the presence of cycloheximide, a protein synthesis inhibitor, the half-life of p27 was markedly reduced in angiotensin

317

ARTICLES siRNA–expressing cardiomyocytes, we observed Ser83 phosphorylation of p27 at 4 h after angiotensin II treatment and drastically reduced p27 levels at 24 h after angiotensin II treatment (Fig. 3d). These effects were abrogated in angiotensin II–treated CK2-a¢ siRNA– expressing cardiomyocytes (Fig. 3d), showing that both of these effects require CK2-a¢. Next, we analyzed whether overexpression of CK2-a¢ or expression of p27 siRNA could acutely induce hypertrophy in cultured cardiomyocytes. Overexpression of siRNA-resistant CK2-a¢ (SR CK2-a¢) or

II–treated cells (Fig. 3c). Silencing of CK2-a¢ by siRNA or preincubation of cardiomyocytes with DMAT prevented p27 destabilization (Fig. 3c). Thus, the integrity of CK2-a¢ function is crucial for p27 degradation in response to angiotensin II. CK2-a¢ siRNA abrogates cardiomyocyte hypertrophy To elucidate the function of p27 in cardiomyocyte hypertrophy, we examined phosphorylation of p27 on Ser83 and the correlation of Ser83 phosphorylation with p27 protein abundance. In control

b

[32P]histone H1 CK2-α′

WB endog. p27

2

**

1

WB + – + –

Input: anti-His p27 Ang II p27∆Pi

+ – – +

WT p27 p27S83D, T187D

d

10

Control siRNA MHC DNA p27

8 p27

6 4

– Ang II Mock + Ang II Mock

DMAT Time (h) 0

1 2 4 6 12

24 h + Ang II

CK2-α′ siRNA Control siRNA

p27

+ – – +

p27

+ – + –

p27Pi-S83

+ + – –

4 h + Ang II

+ – – –

p27Pi-S83

– – – –

p27Pi-S83

0 Ang II CK2-α′ siRNA Control siRNA DMAT

0h

2

WB anti-His p27 – – – –

+ – – –

+ + – –

+ – + –

+ – – +

6

* * *

5

Anti–His-p27∆Pi

*

4

Endog. p27 p27Pi-T187 p27Pi-S83

#

3

*

WB

Anti–His-p27∆Pi

2

**

1

0 Ang II DMSO LLM Lactacystin p27∆Pi

– – – – – CK2-α′ –

CK2-α′ siRNA MHC DNA p27

#

+ – – – – –

e

+ + – – – –

+ – + – – –

+ – – + – –

10

Endog. CK2-α′ Anti–His-CK2-α′

+ – – + + –

+ – – + – +

CK2-α′

– Ang II

+ Ang II

#

8 6

#

*

*

#

*

* *

4

**

**

2

5

#

8

#

6

#

#

4

* * + – – – –

+ + – – –

#

* *

* **

2

– – – – Control siRNA + CK2-α′ SR CK2-α′ p27 SR p27∆Pi

4

+ + – – – –

+ – + – – –

+ – – + – –

+ – – + + –

+ – – + – + + Ang II

– Ang II Cell size

#

# #

3

*

2

*

#

*

*

*

**

**

1 0 10 Protein synthesis

0 10

0 CK2-α′ siRNA SR CK2-α′ p27 siRNA SR p27∆Pi

+ – – – – –

– – – – – –

Ang II DMSO LLM Lactacystin p27∆Pi

Surface area 2 3 (µm × 10 )

Anti–His-p27

*

3

p27

– – + – –

– – + + –

– – – – +

+ – – – –

** + + – – –

– – + – –

[35S]Methionine incorporation (c.p.m. × 103)

IP CK2-α′ WB

#

4

0 Ang II – + + p27∆Pi – – + WT p27 – – – p27S83D, T187D – – –

c

CK2-α′

*

Relative mRNA levels (fold induction) BNP/GAPDH ANF/GAPDH

#

*

5

CK2-α′ kinase activity 3 (c.p.m. × 10 )


6

p27 p27Pi-T187 p27Pi-S83

IP CK2-α′ WB

[32P]histone H1

p27 half-life (h)


a

*

4

*

* **

2

– – – – Control siRNA +

*

#

*

#

6

0 CK2-α′ siRNA SR CK2-α′ p27 siRNA SR p27∆Pi

– – + + –

#

#

8

+ – – – –

+ + – – –

– – + – –

– – + + –

– – – – +

+ – – – –

** + + – – –

– – + – –

– – + + –

Figure 3 CK2-a¢-mediated p27 degradation is essential for cardiomyocyte hypertrophy. (a) Mutant p27DPi lacking both CK2-a¢ phosphorylation sites was metabolically stable and inhibited CK2-a¢ in cardiomyocytes. Cells were transduced with TAT-conjugated p27 protein variants for 1 h before Ang II addition for 4 h. A quantitative analysis of CK2-a¢ kinase activities is shown at left. The association of WT p27 and p27 protein variants with endogenous CK2-a¢ was analyzed by immunoprecipitation and western blotting (right). Data are means ± s.e.m., n ¼ 3. *P o 0.01 versus unstimulated cells. **P o 0.01 versus Ang II. #P o 0.01. Endog., endogenous; c.p.m., counts per minute. (b) Pharmacological inhibition of the 26S proteasome abrogated Ang II–triggered p27 degradation. Cardiomyocytes were pretreated with the proteasome inhibitor lactacystin (10 mM) or proteasomally-inactive LLM (50 mM) for 30 min and then incubated with Ang II for 6 h before lysis. Alternatively, lactacystin-treated cardiomyocytes were transduced with TAT-conjugated CK2-a¢ or p27DPi for 1 h before Ang II addition for 4 h. A quantitative analysis of CK2-a¢ kinase activities is shown at left. Binding of WT p27 and phosphorylation-resistant p27DPi to CK2-a¢ was analyzed by immunoprecipitation and western blotting (right). (c) Inhibition of CK2-a¢ prevents Ang II–induced degradation of p27. For determination of p27 half-life, transduced cardiomyocytes were labeled with [35S]methionine-cysteine and then chased with excess cold methionine in the presence of Ang II and cycloheximide. p27 was immunoprecipitated with anti-p27 covalently linked to protein A–agarose. Aliquots were resolved on SDSPAGE, transferred to nitrocellulose membranes, autoradiographed (bottom) and quantified with ImageJ software (top). (d) Elimination of CK2-a¢ function abolished p27 phosphorylation at Ser83 and its downregulation in Ang II–stimulated cardiomyocytes, as analyzed by immunocytochemistry. Cells were transduced with lentiviruses encoding CK2-a¢ siRNA or control siRNA for 72 h before Ang II treatment for the indicated times. Fixed cells were stained for indirect immunofluorescence microscopy with rabbit polyclonal antibody to p27, antibody to p27Pi-Ser83 (green), antibody to tropomyosin (red) to identify cardiomyocytes, and Hoechst 33342 (blue) for DNA. Scale bar, 20 mm. (e) Analysis of hypertrophic marker gene expression (ANF and BNP), de novo protein synthesis and cell size. Cells were lentivirally infected for 72 h in the presence or absence of Ang II. Quantitative RT-PCR analysis of ANF and BNP mRNA is shown at left. De novo protein synthesis was determined by in vivo labeling of cardiomyocytes with [35S]methionine (right). Cell size was determined by indirect immunofluorescence microscopy with antibody to tropomyosin and analyzed by ImageJ. Expression of endogenous CK2-a¢ and p27 was determined by immunoblotting of total cell extracts. Ectopic SR CK2-a¢ and SR p27 were detected with anti-His. Data are means ± s.e.m., n ¼ 4. *P o 0.005 versus control siRNA alone. **P o 0.005 versus control siRNA + Ang II. #P o 0.005.

318

VOLUME 14

[

NUMBER 3

[

MARCH 2008

NATURE MEDICINE

ARTICLES

0

2 4 8 Age (months)

*

10

n.s.

n.s.

TAB

150

*

100 50 Sham TAB Sham TAB

WT p27 KO p27 KO WT

[32P]histone

5 IP CK2-α′ WB

4 3

p27 CK2-α′

WB

2

p27 Actin

1 Sham

TAB

Sham

Sham + – + – TAB – + – +

TAB

4 2

#

*

3

*

2

*

1

n.s.

n.s.

h p27 KO WT #

5

*

4 3

*

2 1 0

Sham TAB Sham TAB

8 6

4 2 0 8 6 4 2 2 4 8 Age (months)

p27 KO WT

*

*

*

*

4 2 0 8 6 4 2 0

Sham TAB Sham TAB

α-MHC/GAPDH

BNP/GAPDH

f

6

0

2 4 8 Age (months)

WT

H1

CK2-α′

6

p27 KO WT

4

0 6 5 4 3 2 1 0

2

0

TAB

Sham

Long axis/ short axis

TAB

*

+

Sham

#

TUNEL nuclei (%)

0

n.s.

4

0 8

2 4 8 Age (months)

0 2

g

0

*

20

0

6

0 6 5 4 3 2 1 0

5

*

30

4

1

p27 KO # WT

40

Short axis (µm)

#

*

6

50

Long axis (µm)

Heart weight / body weight (mg/g)

e 8

*

50

4 8 Age (months)

p27 KO WT

#

100

d 10

*

2

6

β-MHC/GAPDH

#

150

*

Relative mRNA levels (fold induction)

#

# #

p27 KO WT

8

Sham TAB Sham TAB

p27 KO p27 KO WT Caspase-3 activity (nmol pNA/min/ mg protein)

10

3

p27 KO WT

8

BNP/GAPDH

2

4

c

*

ANF/GAPDH

Long axis (µm)

WT

*

20

0

p27 KO

*

#


8 2 4 Age (months)

#

30

p27 KO WT

ANF/GAPDH

2

40

5


4

#

Sham

Short axis (µm)

6

0

p27 KO WT #

50

GATA4 TUNEL

b

p27 KO WT

GATA4 TUNEL

*

p27-knockout mice develop age-dependent cardiac hypertrophy Mice with targeted homozygous disruption of the gene encoding p27 (p27-knockout mice) develop cardiac hyperplasia, consistent with a role for p27 in both cell differentiation and proliferation24–26. At 2 months of age, p27 loss led to markedly smaller cardiomyocytes without significant differences in cardiac mass and echocardiographic parameters as compared to WT controls (Fig. 4a,b and Supplementary Table 1 online). By 4 months of age, p27-knockout mice developed cardiac hypertrophy with an average increase of 45% in heart/body weight ratio and a marked increase in both the length and width of cardiomyocytes as compared to controls (Fig. 4a,b). The length/width ratio remained unaltered, indicating that the change in

Myocyte area 2 3 (µm × 10 )

*

Myocyte area 2 3 (µm × 10 )

Heart weight / body weight (mg/g)

#

8

Long axis/ Short axis

10

a



silencing of p27 by RNA interference (Supplementary Fig. 2 online) induced cardiomyocyte hypertrophic growth in the absence of angiotensin II stimulation, as evaluated by mRNA expression of hypertrophic brain natriuretic peptide (BNP) and atrial natriuretic factor (ANF), protein synthesis and cell size (Fig. 3e). Cardiomyocytes expressing CK2-a¢ siRNA were completely refractory to angiotensin II action. Cotransduction of SR CK2-a¢ re-induced hypertrophic growth, confirming that the effects of CK2-a¢ siRNA on hypertrophy were a result of silencing the intended CK2-a¢ target (Fig. 3e). Because cotransduction of SR p27DPi and p27 siRNA effectively prevented hypertrophy, we conclude that CK2-a¢ mediates angiotensin II responses, including p27 inactivation.

10 8 6

*

*

4 2 0 Sham TAB Sham TAB

Figure 4 Mice with targeted homozygous deletion of the p27 gene (p27 KO) develop age-dependent cardiac hypertrophy. (a) Heart weight corrected for body weight and Masson stain of myocardial sections. Data are means ± s.e.m., n ¼ 8–12. #P o 0.01, *P o 0.01 versus sham. Scale bar, 2 mm. (b) Cardiomyocyte cell size isolated from WT and p27 KO mice were measured at the indicated ages. One hundred cells from three different mice were analyzed per group. Data are means ± s.e.m. #P o 0.01. *P o 0.01 versus p27 KO at 2 months. n.s., not significant. (c) Determination of mRNA levels of hypertrophic marker genes in cardiomyocytes by quantitative RT-PCR. Data are means ± s.e.m., n ¼ 4–6. (d) Susceptibility of p27 KO mice to pressure overload–induced cardiac dysfunction. Six-week-old mice were subjected to TAB for 3 weeks. Data are means ± s.e.m. n ¼ 8–10. #P o 0.01. *P o 0.01 versus sham. (e) Morphometric analysis of isolated adult cardiac myocytes. Data are means ± s.e.m., n ¼ 100. #P o 0.01. *P o 0.01 versus sham. (f) Determination of mRNA levels of hypertrophic marker genes in cardiomyocytes by quantitative RT-PCR. Data are means ± s.e.m., n ¼ 4–6. *P o 0.01 versus sham. (g) TAB induced CK2-a¢ kinase activity and downregulation of p27 protein expression as determined in left ventricular tissue samples. Left ventricular extracts (2–3 mg) were subjected to anti–CK2-a¢ immunocomplex kinase assays using histone H1 as substrate (left). Levels of endogenously expressed proteins in total left ventricular heart tissue samples (60 mg) were analyzed by immunoblotting (right). (h) Analysis of apoptosis in left ventricular cardiac myocytes (white arrowheads) by immunocytochemical and TUNEL assays (left, middle) and by colorimetric determination of caspase-3 proteolytic activity (right) in total left ventricular heart tissue samples. Data are means ± s.e.m., n ¼ 200 GATA4-positive nuclei. GATA4 staining was used as a marker for the nuclei of cardiomyocytes. *P o 0.01 versus sham. pNA, Ac-DEVD-para-nitroaniline. Unless otherwise indicated, mice were 9 weeks old.

NATURE MEDICINE

VOLUME 14

[

NUMBER 3

[

MARCH 2008

319

ARTICLES Figure 5 CK2-a¢ is important for Ang II action 10 p27∆Pi Saline p27∆C KD CK2-α′ CK2-α′ Saline Ang II in vivo: protein transduction of KD CK2-a¢ or # 1 2 5 3 4 8 * * * p27DPi abrogates cardiac hypertrophy in response 6 ** ** to Ang II treatment in WT mice, and transduction 4 6 7 8 9 10 of recombinant active CK2-a¢ evokes cardiac 2 hypertrophy in the absence of Ang II. (a) Six0 1 2 3 4 5 6 7 8 9 10 week-old WT mice were transduced with the p27∆Pi – + – – – – + – – – indicated constructs. Heart weight corrected for p27∆C – – + – – – – + – – KD CK2-α′ – – – + – – – – + – body weight (left) and Masson stain of myocardial CK2-α′ – – – – + – – – – + cross sections (right) are shown. Data are means Saline Ang II Saline Ang II Saline Ang II 50 5 8 ± s.e.m., n ¼ 6–8. #P o 0.01 versus sham # * * * 4 40 saline. *P o 0.01 versus sham Ang II. * # 6 * * * * * # 30 3 **P o 0.01 versus Ang II. Scale bar, ** ** ** ** 4 2 20 2 mm. (b) Cell size analysis of adult cardiac ** ** 2 10 1 myocytes isolated from 8-week-old WT mice 0 0 0 transduced with the indicated constructs. 8 # 6 * * * n.s. n.s. Data are means ± s.e.m., n ¼ 100. #P o 0.01 150 6 5 * * * ** ** 4 # versus sham saline. *P o 0.01 versus sham 4 100 3 Ang II. **P o 0.01 versus Ang II. 2 2 50 ** ** 1 (c) Determination of ANF and BNP mRNA levels 0 0 0 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 by quantitative RT-PCR in cardiomyocytes of p27∆Pi – + – – – – + – – – p27∆Pi – + – – – – + – – – p27∆Pi – + – – – – + – – – 8-week-old WT mice transduced with the p27∆C – – + – – – – + – – p27∆C – – + – – – – + – – p27∆C – – + – – – – + – – KD CK2-α′ – – – + – – – – + – KD CK2-α′ – – – + – – – – + – KD CK2-α′ – – – + – – – – + – indicated constructs. Data are means ± s.e.m., CK2-α′ – – – – + – – – – + CK2-α′ – – – – + – – – – + CK2-α′ – – – – + – – – – + n ¼ 6–8. #P o 0.05 versus sham saline. Saline Ang II Anti-His tag MHC DNA Saline Ang II *P o 0.05 versus sham Ang II. **P o 0.05 6 [32P]histone H1 versus Ang II. (d) Visualization of ectopic 5 WB anti–His-CK2-α′ IP CK2-α′ TAT-conjugated p27 protein variants (green) 4 WB endog. CK2-α′ in left ventricular tissue samples of WT mice 3 WB endog. p27 2 transduced with saline or the indicated IP CK2-α′ p27∆Pi WB anti–His1 constructs by indirect immunofluorescence p27∆C 0 WB p27 microscopy with anti-His antibodies. Scale bar, 1 2 3 4 5 6 7 8 9 10 p27∆Pi p27∆Pi – + – – – – + – – – 200 mm. (e) Left ventricular extracts (2–3 mg) WB anti–Hisp27∆C – – + – – – – + – – p27∆C from WT mice transduced with the indicated KD CK2-α′ – – – + – – – – + – Actin CK2-α′ – – – – + – – – – + constructs were subjected to anti–CK2-a¢ 1 2 3 4 5 6 7 8 9 10 p27∆Pi – + – – – – + – – – immunocomplex kinase assays employing p27∆C – – + – – – – + – – histone H1 as substrate (left; top right, row 1). KD CK2-α′ – – – + – – – – + – CK2-α′ – – – – + – – – – + Aliquots of kinase reactions were subjected to immunoblotting using either anti–CK2-a¢ and anti-His (top right; rows 2 and 3) or anti-p27 and anti-His (middle right, rows 4–6). Levels of endogenous p27 or transduced p27 protein variants were analyzed by immunoblotting (60 mg/lane) with anti-p27 or anti-His (bottom right; rows 7–10).

e

BNP/GAPDH ANF/GAPDH

CK2-α′ kinase activity (c.p.m. × 103)

p27∆Pi

Saline

d


Long axis (µm)

Myocyte area (µm2 × 103)

Short axis (µm)

c

p27∆C


b


Ang II

Saline

Heart weight/ body weight (mg/g)

a

cell size was similar to that observed in physiologic hypertrophy42. In contrast, pathological cardiac hypertrophy is characterized by increased ANF, BNP and b-myosin heavy chain (b-MHC) expression and a decrease in a-MHC expression43. We observed no alterations in expression of these hypertrophic mRNAs in p27-knockout mice (Fig. 4c). The hearts of p27-knockout mice showed a marked enhancement in contractility, as indicated by increased fractional shortening compared to WT controls (Supplementary Table 1). Thus, in older mice, p27 loss induces spontaneous hypertrophy in the heart without decompensation.

that p27 is important in preventing pressure overload–induced deterioration of cardiac function. Immunocomplex kinase assays showed that pressure overload activated CK2-a¢ and induced p27 destabilization (Fig. 4g). Because we expected increased cardiomyocyte death associated with cardiac remodeling, we examined the impact of p27 deletion on TAB-induced apoptosis. We observed significantly increased numbers of TUNEL-positive cardiomyocyte nuclei in p27-knockout banded hearts as compared to WT banded hearts (Fig. 4h), suggesting that p27 may have an antiapoptotic function.

p27-knockout mice are hypersensitive to TAB At 2 months of age, heart weight/body weight indices were comparable between p27-knockout and WT mice (Fig. 4a). Next, we analyzed whether p27-knockout mice were more susceptible to hypertrophic stimulation after thoracic aortic banding (TAB), as would be predicted if p27 suppresses hypertrophic signaling. After 3 weeks of TAB, the heart/body weight ratio in p27-knockout mice increased by 63%, as compared to 29% in WT mice (Fig. 4d–f). Fractional shortening was significantly reduced in TAB-treated p27-knockout mice compared with WT mice (Supplementary Table 2 online). p27-knockout hearts showed significantly greater dilatation than WT mice, as determined by increases in left ventricular end diastolic and systolic dimensions (LVESD and LVEDD; Supplementary Table 2). These results indicate

KD CK2-a¢ cannot block hypertrophy in p27-knockout mice To determine whether CK2-a¢–directed p27 turnover plays a part in the development of cardiac hypertrophy in vivo, we chronically infused angiotensin II (ref. 44) and intraperitoneally injected recombinant TAT-conjugated CK2-a¢ and p27 variants into WT and p27knockout mice (Supplementary Fig. 3 online). Indices for cardiac hypertrophy, including heart/body weight ratio, myocyte area and ANF and BNP mRNA levels, were significantly increased in angiotensin II–treated WT mice as compared to WT mice injected with catalytically-inactive KD CK2-a¢ or p27DPi (Fig. 5a–d). Immunocomplex kinase assays showed that angiotensin II–induced CK2-a¢ activity was abolished by transduction of KD CK2-a¢ or p27DPi (Fig. 5e). In angiotensin II–infused WT mice, transduction of KD

320

VOLUME 14

[

NUMBER 3

[

MARCH 2008

NATURE MEDICINE

ARTICLES

4

IP CK2-α′

3 IP CK2-α′

2

WB endog. CK2-α′ WB anti–His-CK2-α′ WB anti–His-p27∆Pi WB anti–His-p27∆Pi

1

Actin 2 +

3 –

4 –

5 +

6 –

KD CK2-α′ –

–

+

–

–

+

1 p27∆Pi – KD CK2-α′ –

1 2 3 4 5 6 p27∆Pi – + – – + – KD CK2-α′ – – + – – +

2 + –

3 – +

4 – –

5 + –

**

BNP/GAPDH

0

6 – +

1 p27∆Pi – KD CK2-α′ –

2 + –

3 – +

4 – –

5 + –

p27 KO

e 5

Saline

4

Ang II # # #

*

*

3 2 1

0 1 p27 KO + – WT – + p27∆Pi – – KD CK2-α′ – –

** 2 +– –+ ++ ––

3 +– –+ –– ++

4 +– –+ –– ––

** 5 +– –+ ++ ––

6 +– –+ –– ++

2

3

4

5

6

lin Sa

6 – + WT

1

e

0 1 p27∆Pi –

6 – +

1

2

3

4

5

6

-α ′

5

5 + –

7 ∆P i

[32P]histone H1

4 – –

p2

Ang II

3 – +

e

Saline

Ang II

2 + –

+

6

Saline

TUNEL nuclei (%)

d CK2-α′ kinase activity (c.p.m. × 103)

KD

KD

Figure 6 Constitutively inactive KD CK2-a¢ does not inhibit cardiac hypertrophy in p27 KO mice. (a) Six-week-old p27 KO mice were Stress signals transduced with the indicated constructs. Heart weight corrected for body Ang II Saline Ang II 10 # weight (left) and Masson stain of myocardial cross sections (right). Data are # 8 means ± s.e.m., n ¼ 6–8. *P o 0.01 versus sham saline. **P o 0.01 CK2-α′ p27 * * CK2-α′ 6 versus Ang II. Scale bar, 2 mm. (b) Cell size analysis of adult cardiac Pi Pi n.s. 4 myocytes isolated from 8-week-old p27 KO mice transduced with the p27 2 ? ? indicated constructs. Data are means ± s.e.m., n ¼ 100. *P o 0.01 0 versus sham saline. **P o 0.01 versus Ang II. (c) Determination of ANF 1 2 3 4 5 6 p27 KO + – + – + – + – + – + – and BNP mRNA levels by quantitative RT-PCR in cardiomyocytes from WT – + – + – + – + – + – + OFF ON p27∆Pi – – + + – – – – + + – – 8-week-old p27 KO mice transduced with the indicated constructs. Data Hypertrophy KD CK2-α′ – – – – + + – – – – + + are means ± s.e.m., n ¼ 6–8. *P o 0.01 versus sham saline. **P o 0.01 versus Ang II. (d) CK2-a¢–dependent kinase activity determined in whole left ventricular tissue samples from 8-week-old p27 KO mice transduced with the indicated constructs. A quantitative analysis of CK2-a¢ kinase activities is shown at left. Left ventricular extracts (2–3 mg) were subjected to anti-CK2-a¢ immunocomplex kinase assays using histone H1 as substrate (right). Anti– CK2-a¢ immunoprecipitates were immunoblotted with anti–CK2-a¢ or anti-His to analyze CK2-a¢–bound p27DPi. Levels of transduced p27DPi were analyzed by immunoblotting (60 mg/lane) with anti-His. (e,f) Ang II induced apoptosis in p27 KO mice. (e) Analysis of apoptosis in left ventricular cardiac myocytes (white arrowheads) by immunocytochemistry and TUNEL assay. Data are means ± s.e.m., n ¼ 200 GATA4-positive nuclei. *P o 0.01 versus sham saline, **P o 0.05 versus sham saline #P o 0.01. (f) Determination of caspase-3 activity in total left ventricular heart tissue samples. Data are means ± s.e.m., n ¼ 4. *P o 0.05 versus sham saline. #P o 0.01. (g) Model for CK2-a¢–dependent regulation of cardiac hypertrophy. Hypertrophic signals stimulate CK2-a¢ kinase, which phosphorylates p27 at Ser83 and Thr187. Phosphorylation of p27 impairs its ability to bind CK2-a¢ and inhibit its activity. Subsequently, p27 is proteasomally degraded, allowing hypertrophic growth to proceed. We hypothesize that p27 restrains cardiac growth by targeting an unknown cytoplasmic protein kinase participating in hypertrophic signal transduction.

f

g

Caspase-3 activity (nmol pNA/min/ mg protein)


1 p27∆Pi – KD CK2-α′ –

2

K2

0

C

50

lin

+

Sa

–

ANF/GAPDH

–

*

*

4

TUNEL

+

6

GATA4

–

n.s.

TUNEL

6 –

n.s.

**

GATA4

5 +

*

*

2

Saline

4 –

**

*

4

Ang II

3 –

*

100

-α ′

2 +

150

Ang II

Saline

6

0 8

K2

1 p27∆Pi – KD CK2-α′ –

Long axis (µm)

0

0 6 5 4 3 2 1 0

C

6

1


0

**

2

7 ∆P i

5

10

*

*

3

p2

4

2

*

TUNEL

4

20

**

8

GATA4

2

*

30

Ang II

4

TUNEL

1

KD CK2-α′ 3

Saline Myocyte area (µm2 × 103)

* **

6

p27∆Pi

Saline

*

8

Saline

c

5 Ang II

Saline 40


Ang II

Saline

Ang II

Heart weight/ body weight (mg/g)

10

50

GATA4

b Short axis (µm)

a

CK2-a¢ or p27DPi did not affect LVEDD and LVESD measurements (Supplementary Table 3 online). Injection of p27DC, a mutant protein deficient in binding to CK2-a¢ (p27 amino acids 1–86, Fig. 1c), did not have an antihypertrophic effect; this construct thus serves as a negative control. Notably, transduction of recombinant active CK2-a¢ induced cardiac hypertrophy in the absence of angiotensin II (Fig. 5a–c), and had similar effects as angiotensin II infusion on CK2-a¢ activity and p27 destabilization (Fig. 5e). Collectively, these results suggest that downregulation of p27 is necessary for the development of cardiac hypertrophy. Next, we examined the effect of the transduced proteins on cardiac hypertrophy in p27-knockout mice. Elimination of p27 gene function in these mice completely abrogated the ability of KD CK2-a¢ to block cardiac hypertrophy (Fig. 6a–d). In contrast, hypertrophic responses to angiotensin II in p27-knockout mice were abolished by p27DPi transduction (Fig. 6a–d). Thus, reconstitution of p27 protein abundance in the myocardium of p27-knockout mice was accompanied by

NATURE MEDICINE

VOLUME 14

[

NUMBER 3

[

MARCH 2008

inhibition of angiotensin II action. Moreover, we observed a marked increase in the number of TUNEL-positive cardiomyocyte nuclei and in caspase-3 activity in angiotensin II–infused p27-knockout mice as compared to WT controls (Fig. 6e,f). Notably, p27DPi transduction inhibited cardiomyocyte apoptosis in these mice. This was in direct contrast to the effects seen with ectopic KD CK2-a¢ expression (Fig. 6e,f), which is consistent with our previous work showing that CK2 is cardioprotective45,46. DISCUSSION This study demonstrates the existence of a previously undescribed p27–CK2-a¢ regulatory feedback loop that is analogous to, but distinct from, p27–cyclin-Cdk2 complexes. CK2-a¢–dependent regulation of p27 represents a new link between extracellular growth signaling and surveillance of p27 activity in cardiac myocytes. Our in vivo data describe a role for p27 in nonproliferating tissue, where it exerts growth-suppressive functions (Fig. 6g).

321


ARTICLES Growth factors activate cyclin E–Cdk2, which then phosphorylates p27 on Thr187 (ref. 18), promoting its ubiquitination and proteasomal degradation19,20. It has been difficult to reconcile how Cdk2-bound p27 can be phosphorylated at Thr187, as G0-phase p27Cdk2 complexes are catalytically inactive. Recent studies have identified tyrosine phosphorylation of p27 as a mechanism through which p27-Cdk2 complexes are activated39,40. The non–receptor tyrosine kinase Lyn and constitutively active Bcr-Abl (Breakpoint cluster region homolog-Abelson murine leukemia viral oncogene fusion protein) can phosphorylate p27 on Tyr88 (ref. 39). c-Src kinase–mediated p27 phosphorylation on Tyr74 and Tyr88 also leads to a reduction in its inhibition of Cdk2 (ref. 40). Tyrosine-phosphorylated p27 can still bind Cdk2, but it shows markedly impaired kinase inhibition39,40. Consequently, these tyrosine-phosphorylated intermediate forms of p27 may be primed for Cdk2-dependent Thr187 phosphorylation and Skp2 (S phase kinase–associated protein 2)—driven p27 proteolysis39,40. Here we demonstrate that p27 is eliminated by CK2-a¢–dependent mechanisms in cardiomyocytes. CK2-a¢–mediated phosphorylation of p27 at Ser83 and Thr187 relieves CK2-a¢ from p27 inhibition and converts CK2-a¢ into an active kinse. Our findings suggest that CK2-a¢ functions primarily to tag p27 for destruction. Because hypertrophic signaling involves activation of tyrosine kinases35, it will be of interest to determine whether CK2-a¢–unrelated kinase(s) can also phosphorylate p27, thereby mediating growth factor signaling. The CK2-a¢–specific Ser83 motif in p27 is evolutionarily conserved and is present in human p27, implying that Ser83 phosphorylation may also have a role in initiating cellular growth in response to extracellular signals in humans. CK2-a¢ bound to p27Pi-Ser83 was less active than CK2-a¢ bound to p27Pi-Thr187. Moreover, we observed only small amounts of p27Pi-Thr187–bound CK2-a¢ compared to p27Pi-Ser83, indicating that p27 is sequentially phosphorylated by CK2-a¢ in hypertrophic signaling. We hypothesize that if phosphorylation on a single CK2-a¢ site in p27 were able to cause its degradation, small fluctuations in CK2-a¢ activity might suffice to abrogate the ability of p27 to restrain cardiomyocyte growth, potentially leading to a cardiac phenotype (such as heart failure), as observed in mice subjected to TAB or infused with angiotensin II. However, p27 turnover by CK2-a¢ appears to require phosphorylation at two sites, imposing a check on cardiomyocyte growth under basal conditions. CK2 consists of two catalytic (a,a¢) and two regulatory (b) subunits36,37 and is involved in diverse biological processes such as cell proliferation, differentiation and survival signaling47–50. The mechanisms regulating CK2 activity are not well defined36,37. We postulate that hypertrophic signaling leads to CK2-a¢ activation by an as yet unknown mechanism. Expression levels of p27 are drastically decreased in both acute and end-stage heart failure in humans32, indicating that selective pressure against p27 cytostatic responses is strong during maladaptive cardiac growth. Our analysis of the phenotype of p27-knockout mice demonstrates that p27 acts to suppress both physiological hypertrophy in the adult heart and pathological cardiac growth in response to pressure overload. These findings suggest that cardiac hypertrophic signaling is sensitive to the inhibitory action of p27 toward CK2-a¢. The exaggerated cardiac hypertrophy of p27-knockout mice, the inhibition of cardiomyocyte hypertrophy by protein transduction of p27 and the failure of dominant-negative CK2-a¢ to block cardiac growth in p27knockout mice suggest that the development of cardiac hypertrophy involves CK2-a¢–dependent inactivation of p27. This study supports the hypothesis that p27–CK2-a¢ complexes are an important part of the cytosolic signaling pathway interrupting the hypertrophic

322

program. Kinase-dead CK2-a¢ did not inhibit cardiac hypertrophy in p27-knockout mice. The most plausible explanation for the function of CK2-a¢ is that it tags p27 for destruction. Therefore, we hypothesize that p27 also targets CK2-a¢–unrelated factors. Because p27 normally inhibits protein kinase activity, it will be important to identify kinase(s) in the prohypertrophic signaling cascade that are regulated by p27. Our work raises the possibility that manipulation of p27 protein levels could yield therapeutic benefits in the treatment of heart failure. METHODS Primary rat neonatal ventricular cardiomyocyte culture and p27-knockout mice. We cultured spontaneously beating cardiomyocytes from heart ventricles of neonatal rats and subjected them to treatment with angiotensin II. We obtained angiotensin II (hypertensin II; A9525), DMSO, lactacystin and N-acetyl-L-leucinyl-L-leucinyl-methional (LLM; A6060) from Sigma. We purchased 2-dimethylamino-4,5,6,7-tetrabromo-1H-benzimidazol (DMAT) from Calbiochem. Animal usage was in accordance with approved institutional animal care guidelines (G0202/04, Landesamt fu¨r Gesundheit und Soziales, Berlin, Germany; animal utilization protocol 1381, University Health Network, Animal Care Committee, Toronto, Canada). We used age-matched male C57BL/6 WT and p27-knockout mice in this study. We chronically infused angiotensin II subcutaneously by osmotic pump (model 2000, Alzet) at a dose of 1.4 mg/kg/min for 14 d. We injected TAT proteins (10 mg/kg) intraperitoneally once daily for 14 d. We performed TAB and pressure gradient measurements as described46. Lentiviral shRNA constructs. We designed siRNA using the Internet applications of Ambion and Dharmacon. We selected siRNA sequences and BLASTsearched them against the rat and mouse genome sequences to ensure that only one gene was targeted and that the control (nonsilencing) siRNA used had no overlap with known genes. We used the following rat target sequences (the sense strand is indicated) to generate small hairpin RNA (shRNA) lentiviruses in pLenti6/BLOCK-iT-DEST (Invitrogen): p27 (Genbank D86924): 5¢-ACCG AGCACCCCAAGCCTT-3¢; CK2-a¢ (Genbank L15618): 5¢-AACACCGTGC TTTCCAGTGGT-3¢; control (nontargeted): 5¢-AGACACACGCACTCGT CTC-3¢. We used HEK293FT cells (Invitrogen) for homologous recombination and packaging. We concentrated lentiviral supernatants 100-fold by polyethylene glycol precipitation. We determined the virus titer through indirect immunofluorescence staining of HT1080 human fibrosarcoma cells (CCL121; American Type Culture Collection) with antibody to HIV-1 p24 (Abcam). We sequentially transduced cardiomyocytes with lentiviruses at 100 plaqueforming units/cell in the presence of 4 mg/ml polybrene (Sigma) at 12 h and 24 h after isolation of the cardiomyocytes. In vitro transcription-translation reactions and in vitro protein interaction assays. For generation of recombinant proteins, we lysed Escherichia coli BL21 or BL21(DE3)pLysS cells (Promega) transformed with pGEX-2lT (for GST fusions) or pRSET-C (for His6-tagged proteins) encoding WT CK2-a¢, CK2-b, WT p27 or the various phosphorylation-resistant p27 mutants in 20 mM Tris pH 8.0, 100 mM NaCl, 0.5% Nonidet P-40, 1.0 mg/ml lysozyme (Sigma), 50 U/ml DNase I (Sigma) and protease inhibitor cocktail (Roche). For His-tagged proteins, we subjected the extracts to immobilized Ni2+ metal ion affinity chromatography (Chelating Sepharose Fast Flow, GE Healthcare) with ¨ KTA, GE Healthcare). We subjected eluted His-tagged proteins to an FPLC (A gel filtration (Superdex 200 HR 10/30, GE Healthcare) and eluted them in 20 mM Tris pH 8.0, 100 mM NaCl, 1 mM EDTA, 0.5% Nonidet P-40, 10% glycerol and 1.0 mM dithiothreitol. We purified GST fusion proteins with glutathione Sepharose 4 Fast Flow (GE Healthcare) and gel filtration. We performed in vitro binding assays of CK2-a¢ and p27 by incubating immobilized GST–CK2-a¢ or GST-p27 with in vitro–translated p27 or His-tagged CK2-a¢, respectively, for 3 h at 4 1C. For in vitro transcription-translation reactions, we used the Expressway Cell-Free E. coli Expression System (Invitrogen) with minor modifications.

VOLUME 14

[

NUMBER 3

[

MARCH 2008

NATURE MEDICINE

ARTICLES


Generation of antibodies to CK2-a¢ and phospho-p27. To generate polyclonal antibodies to phospho-p27, we used the following synthetic phosphopeptides as immunogens Ser83-phosphorylated p27, ERGSpLPEFYYR (amino acids 80–90 of mouse p27); Thr187-phosphorylated p27, TVEQT pPKKPGLR (amino acids 183–194 of mouse p27). We applied the antisera from immunized rabbits to a phospho-peptide affinity column for positive selection and then applied the bound fractions to an unphosphorylated peptide for negative selection. We collected flow-through fractions and used them in this study. Polyclonal antiserum to CK2-a¢ was raised in rabbits against the peptide sequence KEQSQPCAENTVLSSG (amino acids 330– 345 of mouse CK2-a¢). For initial experiments, antibodies to CK2-a¢ and CK2-b subunits were a kind gift from D.W. Litchfield (The University of Western Ontario). Antibodies. We used antibodies directed against the following proteins: His6Gly (R940-25, Invitrogen), actin (4986), lamin A and lamin C (2032), b-tubulin (2146) and tropomyosin (4002), rabbit polyclonal antibody to p27 (2552, all Cell Signaling), mouse monoclonal antibody to p27 (610241; BD Transduction), highly cross-adsorbed Alexa Fluor 488–conjugated goat antibody to rabbit IgG (A-11034; Molecular Probes) and Alexa Fluor 750–conjugated goat antibody to mouse IgG (A-11029; Molecular Probes). Immunocomplex kinase assays. For immunocomplex kinase assays, we incubated cell extracts with antibody to CK2-a¢ or antibody to Cdk2 covalently linked to protein A–agarose for 3 h at 4 1C. We preincubated p27 and CK2-a¢ proteins for 10–15 min on ice. The reactions were initiated by addition of kinase buffer containing histone H1 substrate. For in vitro CK2-a¢ kinase assays, we incubated 100 ng of WT or mutant p27 with 30 ng CK2-a¢b for 30 min at 30 1C in 50 ml kinase buffer containing 50 mM Tris-HCl pH 7.5, 150 mM NaCl, 10 mM MgCl2, 10 mM b-glycerophosphate, 30 mM ATP (Roche), 5 mCi [g-32P]ATP (10 mCi/ml; NEN) and 3 mM histone H1 (Roche). We stopped the reactions by adding SDS sample buffer. In vitro phosphorylation of p27 by CK2-a¢. We performed kinase reactions with 1 mg of His-tagged WT p27 or phosphorylation site mutant p27 protein variants and 200 ng recombinant active His-tagged CK2-a¢b for 1 h at 30 1C in 50 ml kinase buffer containing 50 mM Tris-HCl pH 7.5, 150 mM NaCl, 10 mM MgCl2, 10 mM b-glycerophosphate, 30 mM ATP and 10 mCi [g-32P]ATP. We initiated the reaction by adding CK2-a¢b. For in vitro binding assays, we prephosphorylated recombinant purified His-tagged WT p27, p27S83A, p27T187A and p27DPi (2.5 mg) by incubation with 400 ng recombinant active His-tagged CK2-a¢b for 3 h at 30 1C in kinase buffer containing 1 mM ATP. We inactivated CK2-a¢ by incubation at 100 1C for 5 min. We removed CK2-a¢ by two rounds of immunodepletion with antibody to CK2-a¢ covalently linked to protein A–agarose. Determination of p27 half-life. For metabolic labeling, we pulse-labeled cardiomyocytes with 100 mCi/ml [35S]methionine-cysteine (1000 Ci/mmol; EasyTag, NEN) for 2 h in methionine-cysteine–deficient medium (DME, ICN). The cardiomyocytes were then incubated in isotope-free media supplemented with 0.1 mg/ml methionine-cysteine and 100 mg/ml cycloheximide (Calbiochem) to block protein synthesis. We immunoprecipitated total cell extracts with antibody to p27 covalently conjugated to protein A–agarose. We subjected aliquots to western blotting and autoradiography. Statistical analyses. We used factorial design analysis of variance (ANOVA) or t-tests to analyze data as appropriate. Significant ANOVA values were subsequently subjected to simple main effects analyses or post hoc comparisons of individual means using the Tukey method as appropriate. We considered P values of 0.005 as significant for studies with cultured cardiomyocytes or 0.05 as significant for in vivo studies. Note: Supplementary information is available on the Nature Medicine website. ACKNOWLEDGMENTS We thank D.W. Litchfield (The University of Western Ontario) for the gift of antibodies to CK2-a¢ and CK2-b subunits. This research was supported by Canadian Institutes of Health Research (MOP-81194) and grants from the Deutsche Forschungsgemeinschaft (Ha-1777/7-3, Ha-1777/9-1) to R.v.H., the

NATURE MEDICINE

VOLUME 14

[

NUMBER 3

[

MARCH 2008

Volkswagen-Stiftung (Lichtenberg program) and Bundesministerium fu¨r Bildung und Forschung (Center for Stroke Research Berlin) to M.E. AUTHOR CONTRIBUTIONS J.R. performed the two-hybrid screen and contributed to the in vitro experiments. C.H., K.G. and M.N. performed some of the animal experiments and contributed to the in vivo analysis of the mice. J.A. contributed to the generation of recombinant TAT proteins. L.H., R.D. and R.v.H. carried out experimental design, data analysis and manuscript preparation. R.v.H. supervised L.H. and J.R. Published online at http://www.nature.com/naturemedicine Reprints and permissions information is available online at http://npg.nature.com/ reprintsandpermissions

1. Berthet, C. & Kaldis, P. Cell-specific responses to loss of cyclin-dependent kinases. Oncogene 26, 4469–4477 (2007). 2. Koff, A., Ohtsuki, M., Polyak, K., Roberts, J.M. & Massague´, J. Negative regulation of G1 in mammalian cells: inhibition of cyclin E–dependent kinase by TGF-b. Science 260, 536–539 (1993). 3. Hengst, L., Dulic, V., Slingerland, J.M., Lees, E. & Reed, S.I. A cell cycle–regulated inhibitor of cyclin-dependent kinases. Proc. Natl. Acad. Sci. USA 91, 5291–5295 (1994). 4. Kato, J.Y., Matsuoka, M., Polyak, K., Massague´, J. & Sherr, C.J. Cyclic AMP–induced G1 phase arrest mediated by an inhibitor (p27Kip1) of cyclin-dependent kinase 4 activation. Cell 79, 487–496 (1994). 5. Polyak, K. et al. Cloning of p27Kip1, a cyclin-dependent kinase inhibitor and a potential mediator of extracellular antimitogenic signals. Cell 78, 59–66 (1994). 6. Toyoshima, H. & Hunter, T. p27, a novel inhibitor of G1 cyclin–CDK protein kinase activity, is related to p21. Cell 78, 67–74 (1994). 7. Pagano, M. et al. Role of the ubiquitin-proteasome pathway in regulating abundance of the cyclin-dependent kinase inhibitor p27. Science 269, 682–685 (1995). 8. Hengst, L. & Reed, S.I. Translational control of p27Kip1 accumulation during the cell cycle. Science 271, 1861–1864 (1996). 9. Tomoda, K., Kubota, Y. & Kato, J. Degradation of the cyclin-dependent- kinase inhibitor p27Kip1 is instigated by Jab1. Nature 398, 160–165 (1999). 10. Hara, T. et al. Degradation of p27(Kip1) at the G0-G1 transition mediated by a Skp2-independent ubiquitination pathway. J. Biol. Chem. 276, 48937–48943 (2001). 11. Kamura, T. et al. Cytoplasmic ubiquitin ligase KPC regulates proteolysis of p27Kip1 at G1 phase. Nat. Cell Biol. 6, 1229–1235 (2004). 12. Kotoshiba, S., Kamura, T., Hara, T., Ishida, N. & Nakayama, K.I. Molecular dissection of the interaction between p27 and KPC, the ubiquitin ligase that regulates proteolysis of p27 in G1 phase. J. Biol. Chem. 280, 17694–17700 (2005). 13. Ishida, N. et al. Phosphorylation of p27Kip1 on serine 10 is required for its binding to CRM1 and nuclear export. J. Biol. Chem. 277, 14355–14358 (2002). 14. Kotake, Y., Nakayama, K., Ishida, N. & Nakayama, K.I. Role of serine 10 phosphorylation in p27 stabilization revealed by analysis of p27 knock-in mice harboring a serine 10 mutation. J. Biol. Chem. 280, 1095–1102 (2005). 15. Besson, A. et al. A pathway in quiescent cells that controls p27Kip1 stability, subcellular localization, and tumor suppression. Genes Dev. 20, 47–64 (2006). 16. Rodier, G. et al. p27 cytoplasmic localization is regulated by phosphorylation on Ser10 and is not a prerequisite for its proteolysis. EMBO J. 20, 6672–6682 (2001). 17. LaBaer, J. et al. New functional activities for the p21 family of CDK inhibitors. Genes Dev. 11, 847–862 (1997). 18. Sheaff, R.J. et al. Cyclin E-CDK2 is a regulator of p27Kip1. Genes Dev. 11, 1464–1478 (1997). 19. Carrano, A.C., Eytan, E., Hershko, A. & Pagano, M. SKP2 is required for ubiquitinmediated degradation of the CDK inhibitor p27. Nat. Cell Biol. 1, 193–199 (1999). 20. Montagnoli, A. et al. Ubiquitination of p27 is regulated by Cdk-dependent phosphorylation and trimeric complex formation. Genes Dev. 13, 1181–1189 (1999). 21. Nakayama, K. et al. Skp2-mediated degradation of p27 regulates progression into mitosis. Dev. Cell 6, 661–672 (2004). 22. Aleem, E., Kiyokawa, H. & Kaldis, P. Cdc2–cyclin E complexes regulate the G1/S phase transition. Nat. Cell Biol. 7, 831–836 (2005). 23. Malek, N.P. et al. A mouse knock-in model exposes sequential proteolytic pathways that regulate p27Kip1 in G1 and S phase. Nature 413, 323–327 (2001). 24. Fero, M.L. et al. A syndrome of multiorgan hyperplasia with features of gigantism, tumorigenesis, and female sterility in p27Kip1-deficient mice. Cell 85, 733–744 (1996). 25. Kiyokawa, H. et al. Enhanced growth of mice lacking the cyclin-dependent kinase inhibitor function of p27Kip1. Cell 85, 721–732 (1996). 26. Nakayama, K. et al. Mice lacking p27Kip1 display increased body size, multiple organ hyperplasia, retinal dysplasia, and pituitary tumors. Cell 85, 707–720 (1996). 27. Loda, M. et al. Increased proteasome-dependent degradation of the cyclin-dependent kinase inhibitor p27 in aggressive colorectal carcinomas. Nat. Med. 3, 231–234 (1997).

323


ARTICLES 28. Liang, J. et al. PKB/Akt phosphorylates p27, impairs nuclear import of p27 and opposes p27-mediated G1 arrest. Nat. Med. 8, 1153–1160 (2002). 29. Shin, I. et al. PKB/Akt mediates cell-cycle progression by phosphorylation of p27Kip1 at threonine 157 and modulation of its cellular localization. Nat. Med. 8, 1145–1152 (2002). 30. Viglietto, G. et al. Cytoplasmic relocalization and inhibition of the cyclin-dependent kinase inhibitor p27Kip1 by PKB/Akt-mediated phosphorylation in breast cancer. Nat. Med. 8, 1136–1144 (2002). 31. Pasumarthi, K.B.S. & Field, L.F. Cardiomyocyte cell cycle regulation. Circ. Res. 90, 1044–1054 (2002). 32. Burton, P.B., Yacoub, M.H. & Barton, P.J. Cyclin-dependent kinase inhibitor expression in human heart failure. A comparison with fetal development. Eur. Heart J. 20, 604–611 (1999). 33. Poolman, R.A., Li, J.M., Durand, B. & Brooks, G. Altered expression of cell cycle proteins and prolonged duration of cardiac myocyte hyperplasia in p27Kip1 knockout mice. Circ. Res. 85, 117–127 (1999). 34. Sadoshima, J., Xu, Y., Slayter, H.S. & Izumo, S. Autocrine release of angiotensin II mediates stretch-induced hypertrophy of cardiac myocytes in vitro. Cell 75, 977–984 (1993). 35. Chien, K.R. & Olson, E.N. Converging pathways and principles in heart development and disease. Cell 110, 153–162 (2002). 36. Litchfield, D.W. Protein kinase CK2: structure, regulation and role in cellular decisions of life and death. Biochem. J. 369, 1–15 (2003). 37. Meggio, F. & Pinna, L.A. One-thousand-and-one substrates of protein kinase CK2? FASEB J. 17, 349–368 (2003). 38. Luo, Y., Hurwitz, J. & Massague´, J. Cell-cycle inhibition by independent CDK and PCNA binding domains in p21Cip1. Nature 375, 159–161 (1995).

324

39. Grimmler, M. et al. Cdk inhibitory activity and stability of p27Kip1 are directly regulated by oncogenic tyrosine kinases. Cell 128, 269–280 (2007). 40. Chu, I. et al. p27 phosphorylation by Src regulates inhibition of cyclin E–Cdk2. Cell 128, 281–294 (2007). 41. Nagahara, H. et al. Transduction of full-length TAT fusion proteins into mammalian cells: TAT-p27Kip1 induces cell migration. Nat. Med. 4, 1449–1452 (1998). 42. Hunter, J.J. & Chien, K.R. Signaling pathways for cardiac hypertrophy and failure. N. Engl. J. Med. 341, 1276–1283 (1999). 43. Hirota, H. et al. Loss of a gp130 cardiac muscle cell survival pathway is a critical event in the onset of heart failure during biomechanical stress. Cell 97, 189–198 (1999). 44. Satoh, M. et al. Requirement of Rac1 in the development of cardiac hypertrophy. Proc. Natl. Acad. Sci. USA 103, 7432–7437 (2006). 45. Li, P.F. et al. Phosphorylation by protein kinase CK2: a signaling switch for the caspaseinhibiting protein ARC. Mol. Cell 10, 247–258 (2002). 46. Donath, S. et al. Apoptosis repressor with caspase recruitment domain is required for cardioprotection in response to biomechanical and ischemic stress. Circulation 113, 1203–1212 (2006). 47. Seldin, D.C. & Leder, P. Casein kinase IIa transgene–induced murine lymphoma: relation to theileriosis in cattle. Science 267, 894–897 (1995). 48. Xu, X., Toselli, P.A., Russell, L.D. & Seldin, D.C. Globozoospermia in mice lacking the casein kinase IIa¢ catalytic subunit. Nat. Genet. 23, 118–121 (1999). 49. Kato, T. Jr, Delhase, M., Hoffmann, A. & Karin, M. CK2 is a C-terminal IkB kinase responsible for NF-kB activation during the UV response. Mol. Cell 12, 829–839 (2003). 50. Scaglioni, P.P. et al. A CK2-dependent mechanism for degradation of the PML tumor suppressor. Cell 126, 269–283 (2006).

VOLUME 14

[

NUMBER 3

[

MARCH 2008

NATURE MEDICINE

C O R R I G E N D A / E R R ATA

Corrigendum: Protein kinase CK2 links extracellular growth factor signaling with the control of p27Kip1 stability in the heart Ludger Hauck, Christoph Harms, Junfeng An, Jens Rohne, Karen Gertz, Rainer Dietz, Matthias Endres & Rüdiger von Harsdorf Nat. Med. 14, 315–324 (2008); published online 2 March 2008; corrected after print 21 March 2008


In the version of this article initally published, one author, Junfeng An, was missing from the author list and the Author Contributions section. The errors have been corrected in the HTML and PDF versions of the article.

Corrigendum: Robo4 stabilizes the vascular network by inhibiting pathologic angiogenesis and endothelial hyperpermeability Christopher A Jones, Nyall R London, Haoyu Chen, Kye Won Park, Dominique Sauvaget, Rebecca A Stockton, Joshua D Wythe, Wonhee Suh, Frederic Larrieu-Lahargue, Yoh-suke Mukouyama, Per Lindblom, Pankaj Seth, Antonio Frias, Naoyuki Nishiya, Mark H Ginsberg, Holger Gerhardt, Kang Zhang & Dean Y Li Nat. Med. 14, 448–453 (2008); published online 16 March 2008; corrected after print 18 April 2008 In the version of this article initially published, the affiliation of Rebecca A. Stockton was incorrect. Her correct affiliation is affiliation 5: the Department of Medicine, University of California, San Diego, 9500 Gilman Dr., La Jolla, California 92093-0726, USA. The error has been corrected in the HTML and PDF versions of the article.

Erratum: Detection of colonic dysplasia in vivo using a targeted heptapeptide and confocal microendoscopy Pei-Lin Hsiung, Jonathan Hardy, Shai Friedland, Roy Soetikno, Christine B Du, Amy P Wu, Peyman Sahbaie, James M Crawford, Anson W Lowe, Christopher H Contag & Thomas D Wang Nat. Med.14, 454–458 (2008); published online 16 March 2008; corrected 21 March 2008 In the version of this article initially published online, the name of the first author, Pei-Lin Hsiung, was misspelled as Pei-Lei Hsiung. The error has been corrected in all versions of the article.

NATURE MEDICINE VOLUME 14 | NUMBER 5 | MAY 2008

585