n e w s a n D v i e w s

skp2: caught in the aktKarin Ecker and Ludger Hengst

To control cell proliferation, signal transduction needs to regulate the cell-cycle machinery. Recent findings show that akt — a major kinase that coordinates diverse signalling pathways — phosphorylates skp2, a subunit of the sCF-skp2 ubiquitin ligase that targets key cell-cycle regulators. akt1-dependent phosphorylation activates sCF-skp2 through multiple mechanisms.

Progression through the cell cycle is orches-trated mainly by ubiquitin-mediated pro-teolysis of key regulatory proteins, such as cyclins and cyclin-dependent kinase (CDK) inhibitors. Ubiquitin ligases (E3) function in the last step of a three-enzyme cascade, which leads to covalent attachment of ubiquitin or polyubiquitin chains to Lys residues of sub-strates. Two types of ubiquitin ligases, SCF complexes and the anaphase-promoting com-plex/cyclosome (APC/C), ubiquitylate many cell-cycle regulators and are essential for cell-cycle progression. SCF-type ubiquitin ligases are composed of three invariable subunits, Skp1, Cul1 and Rbx1, and one of about 70 dif-ferent F-box proteins. These proteins bind to Skp1 through their F-box domain and deter-mine target selection by binding to substrates, often in a manner that depends on substrate phosphorylation1. On pages 420 and 397 of this issue, Lin et al.2 and Gao et al.3, respec-tively, show that the F-box protein Skp2 is itself phosphorylated by Akt. Phosphorylation regulates formation and ubiquitin ligase activ-ity of the SCF-Skp2 complex, Skp2 localization and stability, cell migration, cell proliferation and tumorigenesis. This adds an important direct link in the complex regulatory network between phosphatidylinositol-3-kinase (PI(3)K)/Akt signalling and cell-cycle control.

Among the different substrates of SCF-Skp2, the CDK inhibitor p27Kip1 was identified as a key target, and SCF-Skp2 is a major ubiquitin ligase for CDK/cyclin-bound p27 (refs 1, 4). Skp2 is considered a proto-oncogene as its overexpres-sion causes increased proliferation, at least in part through increased p27 proteolysis. Activation of the PI(3)K/Akt pathway by diverse extracel-lular signals triggers a cascade of responses, including cell growth, proliferation, survival

and motility. PI(3)K is antagonized by several lipid phosphatases, including PTEN. The PI(3)K/Akt pathway is known to regulate the Skp2/p27 axis at multiple levels. For example, Akt con-trols p27 transcription, translation, localization, complex formation and stability through direct and indirect mechanisms4. In addition, PTEN inhibits Skp2 expression5 and Akt induces Skp2 transcription6,7. PTEN also inhibits SCF-Skp2 complex formation indirectly by inhibiting Cul1 association with Skp1 or Skp2 (ref. 8).

Both, Lin et al. and Gao et al. observed that Akt1 binds directly to Skp2. Binding was lost on removal of the 90 amino-terminal amino acids of Skp2 (ref. 3), a region dispensable for the assembly of active SCF-Skp2 ligase complexes9. They found that Akt1, but not related kinases such as Akt2, SGK or S6k3, phosphorylates Skp2 on Ser 72 (refs 2, 3). Ser 72 had recently been identified as one of two major Skp2 phosphorylation sites in vivo and, interestingly, Ser 72 phosphorylation was maximal in M phase10. Not only does Ser 72 phosphorylation induce p27 degrada-tion, indicating that Akt stimulates SCF-Skp2 activity, but phosphorylated Skp2 translocates to the cytoplasm2,3.

Lin et al. found that Skp2 phosphoryla-tion at Ser 72 is essential for its ability to pro-mote cell proliferation and tumorigenesis. A phospho-deficient mutant, Skp2S72A mark-edly impaired Skp2-induced cell proliferation in vitro and tumorigenesis in a mouse model2. Therefore Skp2, phospho-Akt and PTEN lev-els were analysed in human prostate and colon tumour microarrays. Skp2 cytosolic localiza-tion correlates strongly with activated Akt, low PTEN levels and lymph node metastasis in colon cancers. This supports a potential role for Akt signalling and Skp2 cytoplasmic localization in tumour metastasis2.

Lin et al. and Gao et al. report different mechanisms that may activate Skp2. Lin et al. found that Ser 72 phosphorylation promotes SCF-Skp2 assembly and enhances ubiquit-ylation of p27 (ref. 2). In contrast, Gao et al. observed that Ser 72-phosphorylated Skp2 is stabilized through its inability to bind to Cdh1, an activator of the APC/C ubiquitin ligase3 that ubiquitylates Skp2 in G1 (refs 11, 12).

Ser 72 is flanked by two phosphorylation sites, Ser 64 and Ser 75 (ref. 3), all located within a region of Skp2 required for Cdh1-binding11. Within this cluster, Ser 64 is most

Karin Ecker and Ludger Hengst are in the Division of Medical Biochemistry, Biocenter, Innsbruck Medical University, Fritz-Pregl-Str. 3, A-6020 Innsbruck Austria.e-mail: ludger.hengst@i-med.ac.at Published online 8 March 2009; DOI:10.1038/ncb1859

SPPRKRLKSKGSKGS

p27

S S

Cdk2(Cdk1)

Akt1 CK1

40 90

48 57 64 72 75

Putative NLS

D-box

1 100 150 208 390 424

14-3-3Akt1cyclin A

Cul1Skp1

Cdh1

LRRF-box

Figure 1 Schematic representation of the Skp2 protein. Functional domains (D-box, F-box and the Leu-rich repeats, LRR) are indicated. Positions of phosphorylation sites are shown with their assigned kinases. All of these sites cluster within a potential regulatory domain, which contains a putative nuclear localization sequence (NLS, underlined). This domain is involved in binding of Skp2 to various proteins and, in its unphosphorylated form, may serve as an inhibitory domain preventing SCF complex formation. Ser 48 and 57 have only been identified by mass spectrometry analysis16.

nature cell biology volume 11 | number 4 | APrIl 2009 377

© 2009 Macmillan Publishers Limited. All rights reserved.

n e w s a n D v i e w s

highly conserved phylogenetically and found in all Skp2 orthologues, from vertebrates to insects10. Ser 72 is conserved in Skp2 from most mammals, including humans, but altered to Gly in mice and not conserved in other ver-tebrates such as birds, frogs and fish. Ser 75 seems to be the least conserved residue and is found in some mammals, including primates and rodents, but not in dog, cow or boar or non-mammalian vertebrates.

Skp2 is nearly quantitatively phosphorylated in vivo on Ser 64 (ref. 10) by Cdk2 and possibly Cdk1 (refs 10, 13, 14). Others have previously reported that Skp2 phosphomimetic mutants Skp2S64D and, to a lesser extent Skp2S72D, were stable and proposed that Skp2 phosphoryla-tion, mainly on Ser 64, stabilized the protein by weakening its interaction with Cdh1 (ref. 10). Gao et al. confirmed that phosphorylation of Ser 64 stabilizes Skp2, but found that Ser 64 mutants (phosphomimetic S64D or S64A) still interacted well with Cdh1 (ref. 3). These data suggest that Ser 64 phosphorylation may

stabilize Skp2 by a Cdh1-independent mech-anism or that additional modifications may cooperate with Ser 64 phosphorylation to pre-vent Cdh1 binding. Moreover, Gao et al. found that combined phosphorylation of Ser 72 and Ser 75 inhibited Skp2 binding to Cdh1. They observed that Ser 72 phosphorylation primes Skp2 for Ser 75 phosphorylation and phosphorylation of both sites permits Skp2 to escape APCCdh1-mediated ubiquitylation and subsequent degradation3.

Lin et al. identified a different mechanism for Skp2 activation by Akt. They found that Ser 72 phosphorylation is required for effi-cient complex formation and ubiquitin ligase activity of SCF-Skp2. For example, a phos-pho-deficient Skp2S72A mutant bound poorly to Skp1 and Cul1, and inhibition of PI(3)K by LY294002 prevented SCF-Skp2 complex formation. It is of note, however, that in these experiments PI(3)K inhibition also prevented binding of Skp1 to Cul1 (ref. 2), suggesting a broader effect of PI(3)K inhibition on all

Cul1-containing SCF complexes. A possible explanation may be an earlier observation that inhibition of PI(3)K blocked SCF-Skp2 complex assembly by promoting Cul1 seques-tration in CAND1 complexes, which block Cul1 accessibility to Skp1 and Skp2 (ref. 8). Therefore PI(3)K/Akt seems to affect SCF-Skp assembly through more than one pathway.

At first glance the observation that Ser 72 phosphorylation is required for SCF-Skp2 complex formation is surprising. Previous studies have shown that the initial 100 amino acid region of Skp2 is dispensable for SCF complex assembly and ubiquitin ligase activity9, and that the Skp2 F-box is suf-ficient to form a quaternary complex with Cul1, Rbx1 and Skp1 (ref. 15). One attractive model that can reconcile the findings of Lin et al. with previous work is to propose that the unphosphorylated N-terminal region prevents the interaction of Skp2 with Skp1 and Cul1; Ser 72 phosphorylation induces a structural change that permits SCF assembly.

PKB/AktP

Skp2

Skp2

Skp2

CKI

14-3-3

Skp2

Importin α5; α7

Skp2

Skp2Skp2

InactiveSkp2

SC(F)

SC(F)

Active

UbUb

UbUb

APCCdh1

Skp2Ub

UbUb

Ub

Skp2Ub

UbUb

Ub

Skp2

Skp2

P

p27

PS72

PS72 PS72

PS72

PS72

PS72PS72

PS72

Figure 2 Model for Akt1-dependent regulation of Skp2 activity and localization. Akt1 phosphorylates Skp2 at Ser 72, leading to cytoplasmic retention of the protein by promoting binding to 14-3-3 and inhibiting binding to the nuclear import receptors, importins. Ser 72 phosphorylation primes Skp2 for subsequent phosphorylation at Ser 75 by casein kinase 1 (CK1). Phosphorylation on both sites interferes with the Skp2/APCCdh1 interaction and stabilizes Skp2. Phosphorylation at Ser 72 also permits an enhanced interaction with Skp1 and Cul1, and activates the SCF-Skp2 ubiquitin ligase.

378 nature cell biology volume 11 | number 4 | APrIl 2009

© 2009 Macmillan Publishers Limited. All rights reserved.

n e w s a n D v i e w s

Consistent with this model, a Skp2 mutant lacking the N-terminal 90 amino acids formed a SCF complex more efficiently than full-length Skp2 (ref. 2).

In contrast to the findings of Lin et al. that Ser 72 phosphorylation is required for SCF-Skp2 assembly and activity2, others found recently that simultaneous mutation of amino acids 64 and 72 (S64D/S72D and S64A/S72A) did not affect SCF assembly or its activity10. In these experiments, Skp2 and all other SCF sub-units, as well as the cofactor Cks1, were over-expressed10, whereas Lin et al. only expressed Skp2. Several experimental differences may explain the opposite outcome. For example, overexpression of all SCF subunits may favour SCF-Skp2 complex formation and ligase activity even in the absence of Ser 72 phosphorylation.

Adding to this complexity, Gao et al. and Lin et al. both found that Ser 72 phosphor-ylation also translocates the protein to the cytoplasm. Again, two different mechanisms seem to contribute to this. First, Ser 72 is located within a putative nuclear localiza-tion sequence (NLS) and its phosphorylation impairs Skp2 binding to nuclear import recep-tors3. Second, Lin et al. and Gao et al. found that Ser 72 phosphorylation facilitates Skp2 binding to 14-3-3 proteins2,3. Skp2 cytoplas-mic localization required 14-3-3β2.

Overexpressed Skp2S72D shows only partial cytoplasmic localization2,3, and a Skp2S64D/S72D double mutant was mainly nuclear10; however, endogenous Ser 72-phosphorylated Skp2 was predominantly cytosolic, but not nuclear2.

Akt1-phosphorylated Skp2 is bound in 14-3-3 complexes, which anchors the protein in the cytoplasm2. Overexpression of Skp2 may exceed the 14-3-3 pool, permitting partial nuclear localization. The cytoplasmic locali-zation of Skp2 raises a number of interesting questions. For example, can Skp2/14-3-3 inte-grate into active SCF complexes? If so, does this alter target substrate selection and what are central substrates of cytosolic SCF-Skp2? If nuclear, how can Ser 72-phosphorylated Skp2 escape 14-3-3 and how is the protein imported? Of note, a Skp2–NES (nuclear export signal) fusion protein was predominantly cytosolic but unable to form an active SCF or to ubiq-uitylate p27 ref. 2). As Cdh1 is usually nuclear, is inhibition of Cdh1 binding physiologically significant as long as most Ser 72-phosphor-ylated Skp2 resides in the cytoplasm?

Skp2 overexpression by gene amplification is frequently observed in metastatic tumours5. Lin et al. found that Skp2–/– MEFs showed a profound defect in cell migration, which could be compensated by Skp2S72D but not Skp2S72A (ref. 2), and a predominantly cytosolic Skp2–NES fusion protein rescued migration of null MEFs. These findings suggest that cytoplasmic Skp2 has a potential function in metastasis. Although p27 has a well-established role in cell migration4, regulation of cell motility by cytoplasmic Skp2 seems to be independent of its ability to ubiquitylate p27, as Skp2–NES fails to form a ubiquitin ligase2. Further studies should elucidate mechanisms by which cyto-plasmic Skp2 affects cell motility.

Taken together, these studies provide com-pelling evidence that Skp2 phosphorylation on Ser 72 has a central role in tumorigenesis. Skp2 phosphorylation seems to affect Skp2 localiza-tion and activity by several complementary mechanisms. The cluster of three phosphoryla-tion sites of different phylogenetic conservation located within a region of Skp2 required for Cdh1 binding and adjacent to the F-box sug-gests possible redundant functions that could explain the variable molecular consequences observed in response to phosphorylation. It is interesting that although mouse Skp2 lacks Ser 72, most molecular consequences of Akt phosphorylation are also observed in mice2 sug-gesting that the Akt–Skp2 axis is functionally conserved but may use distinct mechanisms.

1. Frescas, D. & Pagano, M. Nature Rev. Cancer 8, 438–449 (2008).

2. Lin, et al. Nature Cell Biol. 11, 420–432 (2009).3. Gao, et al. Nature Cell Biol. 11, 397–408 (2009).4. Chu, I. M., Hengst, L. & Slingerland, J. M. Nature Rev.

Cancer 8, 253–267 (2008).5. Hershko, D. D. Cancer 112, 1415–1424 (2008).6. Reichert, M., Saur, D., Hamacher, R., Schmid, R. M. &

Schneider G. Cancer Res. 67, 4149–4156 (2007).7. Barré, B. & Perkins, N. D. EMBO J. 26, 4841–4855

(2007).8. Jonason, J. H., Gavrilova, N., Wu, M., Zhang, H. & Sun,

H. Cell Cycle 6, 951–961 (2007).9. Schulman, B. A. et al. Nature 408, 381–386 (2000).10. Rodier, G., Coulombe, P., Tanguay, P. L., Boutonnet, C.

& Meloche, S. EMBO J. 27, 679–691 (2008).11. Bashir, T., Dorrello, N. V., Amador, V., Guardavaccaro,

D. & Pagano, M. Nature 428, 190–193 (2004).12. Wei, W. et al. Nature 428, 194–198 (2004).13. Zhang, H., Kobayashi, R., Galaktionov, K., Beach, D.

Cell 82, 915–925 (1995).14. Yam, C. H., Ng, R. W., Siu, W. Y., Lau, A. W. & Poon,

R. Y. Mol. Cell Biol. 19, 635–645 (1999).15. Zheng, et al. Nature 416, 703–709 (2002).16. Dephoure, N. et al. Proc. Natl Acad. Sci. USA 105,

10762–10767 (2008).

Targeting protein ubiquitylation: DDB1 takes its RinG offSarah Jackson and Yue Xiong

Ubiquitin e3 ligases of the RinG and HeCT families are distinct not only in their catalytic mechanisms but also in targeting substrates. now it seems that one heterodimeric complex can target substrates to both types of e3 ligase.

Protein ubiquitylation has a broad and criti-cal role in regulating a wide range of cellu-lar processes. The addition of Lys 48-linked polyubiquitin chains to specific substrate

proteins regulates timely degradation by the 26S proteasome. In addition, like other covalent modifications, ubiquitylation can modulate the function of a substrate by caus-ing a conformational change. Ubiquitylation begins with the ATP-dependent activation of ubiquitin by the E1 enzyme, and is followed by the subsequent transfer of ubiquitin to one of a small family of E2 ubiquitin-conjugating

enzymes; finally, an E3 ubiquitin ligase is responsible for recognizing a specific substrate and promoting ubiquitin ligation. More than 1,000 distinct E3 ligases are predicted to exist, either as individual proteins or multi-subunit complexes, in mammalian cells.

There are two major families of E3 ligases distinguished by their active domains: the HECT family (‘homologous to the E6-AP

Sarah Jackson and Yue Xiong are in the Department of Biochemistry and Biophysics, Lineberger Comprehensive Cancer Center, University of North Carolina at Chapel Hill, North Carolina 27599, USA.e-mail: yxiong@email.unc.edu

nature cell biology volume 11 | number 4 | APrIl 2009 379 nature cell biology volume 11 | number 4 | APrIl 2009 379

© 2009 Macmillan Publishers Limited. All rights reserved.