Regulation of Protein Function by SUMO Modification 1401

Pc2 and SUMOylation

D. Wotton1 and J.C. Merrill
Center for Cell Signaling and Department of Biochemistry and Molecular Genetics, University of Virginia, 800577 HSC, Charlottesville, VA 22908, U.S.A.

Abstract
Polycomb proteins are key regulators of transcription in metazoan organisms. Recent work has shed
light on the nature of the polycomb protein complexes in flies and mammalian cells. Multiple enzymatic
activities have been shown to associate with polycomb complexes, including histone methyltransferase,
histone deacetylase and ubiquitination activities, which are primarily directed towards the modification
of chromatin structure. In addition to these chromatin-based functions, other potential roles for polycomb
www.biochemsoctrans.org

proteins exist. Here, we present a comparison of vertebrate Pc2 (polycomb 2 protein) homologues, and
review the known functions of the mammalian Pc2 focusing on its role as a SUMO (small ubiquitin-related
modifier) E3 ligase. Pc2 is an E3 for several SUMO substrates, but still appears to have a more limited
repertoire than other SUMO E3s, perhaps due to its association with polycomb complexes. One possibility
is that Pc2 represents a relatively specialized polycomb protein, which has additional functions to those
associated with other mammalian Pc (polycomb protein) paralogues.

The Cbx (chromobox) family
The human and mouse genomes each encode at least eight
members of the Cbx family of proteins. This classification is
based on the presence of a single N-terminal chromodomain,
which, in at least some members of the family, can bind to his-
tone proteins via methylated lysine residues. The Cbx family
in humans can be subdivided into two groups: one consists of
Cbx1, Cbx3 and Cbx5, also known as HP1 (heterochromatin
protein 1) β, γ and α respectively. The HP1 proteins consist
of an N-terminal chromodomain and a C-terminal chromo-
shadow domain. The remaining five human Cbx proteins
contain a conserved N-terminal chromodomain, but do not
have a chromo-shadow domain. The only other amino acid
Biochemical Society Transactions
to histone H3 trimethyl Lys-9 [5]. Thus there may be
differential recruitment of Cbx protein-containing complexes
to chromatin, dependent on the type of histone modification
and the specific Cbx protein present within the polycomb
complex. The other conserved domain in this group of
proteins, the C-box, is less well characterized. However,
in Pc2, it has been shown to be required for interaction with
the polycomb protein, Ring1, and is required for localization
of Pc2 to subnuclear foci, termed polycomb bodies [6–8].
PRC (polycomb repressive complex)
complexes
Much of the genetic and functional analysis of polycomb

similarity conserved among all five members of this group
is a small hydrophobic region at the extreme C-termini (the
C-box; Figure 1). Both the chromodomain and the C-box are
also present in the Drosophila Pc (polycomb protein), which
represents the prototypic member of this group. Although
all eight Cbx proteins share a highly conserved N-terminal
chromodomain, there is some distinction between the HP1
proteins and the Pc group. HP1 proteins bind to histone
H3 trimethylated on Lys-9, whereas Drosophila Pc binds
preferentially to trimethylated Lys-27 of histone H3 [1–4].
The situation for mammalian Cbx proteins of the Pc group is
more complex, with two having been shown to bind in vitro
to both trimethylated Lys-9 and Lys-27 of histone H3, and
one [Cbx4/Pc2 (polycomb 2 protein)] binding preferentially
Key words: chromobox 4 (Cbx4), E3 ligase, polycomb 2 protein (Pc2), polycomb repressive
complex (PRC), small ubiquitin-related modifier (SUMO)-binding motif (SBM), SUMOylation.
Abbreviations used: Cbx, chromobox; CtBP, C-terminal binding protein; Dnmt, DNA
methyltransferase; HIPK2, homeodomain-interacting protein kinase 2; HP1, heterochromatin
protein 1; Pc2, polycomb 2 protein; hPc1, human Pc1; PML, promyelocytic leukaemia protein;
PRC, polycomb repressive complex; RanBP2, Ran-binding protein 2; STAT, signal transducer and
activator of transcription; PIAS, protein inhibitor of activated STAT; SUMO, small ubiquitin-related
modifier; SBM, SUMO-binding motif; SUV39H1, suppressor of variegation 3–9 homologue 1; Ubc9,
ubiquitin-conjugating enzyme 9.
1
To whom correspondence should be addressed (email dw2p@virginia.edu).
complexes has been carried out in Drosophila, where
polycomb group genes were discovered [9–11]. Recent bio-
chemical analyses using both Drosophila and mammalian cells
have revealed the existence of two major classes of Polycomb
complexes, termed PRC1 and PRC2 [12–15]. Members of the
Pc group of the Cbx family associate with the PRC1 complex,
the core of which in Drosophila consists of Pc, Polyhomeotic,
Posterior sex combs and dRing [15]. In mammalian cells,
a similar core complex is present, but several variants may
exist, at least in part based on the differential inclusion of
various mammalian Pc homologues [13]. For example, hPc1
(human Pc1; Cbx2), hPc2 (Cbx4), hPc3 (Cbx8) or Cbx7 may
be present in different PRC1 complexes, and may help to
determine the precise function of these different complexes
[16]. Although PRC1 components do not bind directly to
DNA, the inclusion of different Cbx homologues may affect
recruitment to chromatin, and it is also possible that different
Cbx proteins bring with them different activities.
Cbx4 proteins
Alignment of the Cbx4 homologues from several vertebrate
species shows that, in addition to the two domains conserved

C The Authors Journal compilation

C 2007 Biochemical Society
1402 Biochemical Society Transactions (2007) Volume 35, part 6

Figure 1 Alignment of vertebrate Cbx4 homologues

Cbx4 sequences from human, mouse, chicken, Xenopus laevis and Xenopus tropicalis, zebrafish and Tetraodon nigroviridis
were aligned. Shading represents identity (black) and similarity (grey) in at least four of seven. Important sequence features
are indicated: chromodomain and C-box (red), CtBP interaction motif (yellow), SUMOylation site (dark blue), SBM (pale blue)
and phosphorylation sites (green).

among the wider Cbx family, there are several other regions
of high sequence identity (Figure 1). Pc2 has been shown
to interact with the transcriptional co-repressor, CtBP (C-
terminal binding protein) [17], which binds to a short peptide
motif (P-X-D-L-S/R/T; where X is often leucine, isoleucine
or another hydrophobic residue) found within numerous
transcriptional regulators [18–20]. This motif is present in
Pc2 from all vertebrate species, but is not found in other
human Cbx paralogues or in Drosophila Pc. A region of
relatively high sequence identity among Cbx4 orthologues
in the central region of the proteins spans a domain shown to
be important for SUMO (small ubiquitin-related modifier)
E3 activity in human Pc2 [21] and, additionally, contains a
conserved SBM (SUMO-binding motif) consensus (Figure 2;
discussed below). A second potential SBM is also present
in five of the seven species, close to the CtBP interaction
motif. Other notable features of the Pc2 sequences include
the VKPE SUMOylation consensus, which in human Pc2 is
the major SUMOylation site, but is lacking in frog and fish
Cbx4 proteins [7]. Aside from the chromodomain and C-box,
conserved regions and sequence motifs found in vertebrate
Cbx4 homologues are not present in other Cbx proteins, sug-
gesting the possibility of conserved Cbx4-specific functions.
SUMOylation
Pc2 was first shown to be a SUMO E3 for the transcriptional
co-repressor CtBP [7]. Since then, other substrates have
been identified for Pc2, but compared with other E3s, such
as RanBP2 (Ran-binding protein 2) or the PIAS [protein
inhibitor of activated STAT (signal transducer and activator
of transcription)] family, relatively few are known for Pc2.
It is possible that Pc2 has a more limited repertoire of
substrates, perhaps due to its specific subnuclear localization
within polycomb bodies. In addition to Pc2, members of the
PIAS family have been shown to be E3 ligases for CtBP [22].

C The Authors Journal compilation

C 2007 Biochemical Society

Figure 2 Schematic representation of human Pc2
Important motifs and domains are as indicated. Colour scheme is as in
Figure 1. The grey bars indicate regions of high conservation among
vertebrate Cbx4 proteins. Brackets indicate minimal binding domains for
Ring1 and SUV39H1. The regions required for E3 activity towards CtBP
are indicated by arrows below. The boxed sequences are the known
SBM consensus sequences. See the text for references.
It is possible that Pc2 promotes SUMOylation of a specific
pool of protein, which is localized to polycomb bodies,
whereas other E3s may regulate SUMOylation of the same
substrates at other cellular locations.
Recent work has identified at least three other SUMO
substrates for Pc2: (i) the transcriptional regulator, SIP1, (ii)
HIPK2 (homeodomain-interacting protein kinase 2) and (iii)
the Dnmt (DNA methyltransferase), Dnmt3a [23–25]. Unlike
CtBP, there is no evidence that PIAS family proteins are
involved in the SUMOylation of these other Pc2 substrates,
although SIP1 has been shown to interact with members of
the PIAS family. Co-localization of HIPK2, SIP1 and CtBP
with Pc2 at subnuclear foci has been observed, suggesting that
for these three proteins at least, the modified pool of protein
may have a restricted localization [7,24,25]. Pc2 has been
shown to recruit Ubc9 (ubiquitin-conjugating enzyme 9)
to the same subnuclear foci, at least when overexpressed,
suggesting that part of the E3 activity observed in vivo is due
to co-recruitment of E2 and substrate [7]. However, other
additional activities of Pc2 may also contribute to E3 activity.
In vivo analysis with transfected proteins suggests that,
for CtBP modification, simple co-recruitment of E2 and sub-
strate by Pc2 is not enough [21]. N-terminal deletions of Pc2
resulted in diminished E3 activity, and pointed to a require-
ment for amino acids 168–288 of Pc2 for full E3 activity. Inter-
estingly, this region contains a small block of highly conserved
amino acids, which includes a consensus SBM. In addition,
the N-terminal half of Pc2, encompassing this region, had E3
activity in vitro. In vivo, this region of Pc2 was not sufficient
to promote CtBP SUMOylation, unless a CtBP recruitment
motif was added. Our recent results suggest that the SBM
within this domain contributes to E3 activity (J.C. Merrill and
D. Wotton, unpublished work). A second potential, but un-
Regulation of Protein Function by SUMO Modification 1403
tested, SBM is present close to the PIDLR CtBP recruitment
motif in Pc2 (Figure 1). This region of Pc2 was implicated
in Ubc9 recruitment, and its deletion reduced in vivo E3
activity towards CtBP while still retaining interaction with
CtBP. However, as with any deletion analysis, more global
effects on protein structure cannot be ruled out. Thus two
domains of Pc2 appear to be required for CtBP modification
(Figure 2); the C-terminal region, which binds both CtBP and
Ubc9, and a region within the N-terminal half of the protein,
which can also interact with Ubc9 and has in vitro E3 activity.
Three non-covalent SUMO interaction motifs or SBMs
have been identified (Figure 2), and recent evidence suggests
that one of these can also function in reverse orientation [26–
29]. The two potential SBMs in Pc2 match the V/I-X-V/I-V/I
consensus, and its reversed orientation [27,28]. Although
both are conserved (one in all vertebrate species, and one in 5
of the 7 shown in Figure 1), and are present in regions of Pc2
that are required for E3 activity, their functional significance
remains to be determined. A V/I-X-V/I-V/I SBM at the
end of the IR1–IR2 domain of RanBP2 has been shown to
contribute to full in vitro E3 activity, although its relevance
in vivo is unknown [27]. Members of the PIAS family also
contain consensus SBMs, which have been shown to bind
SUMO in vitro, but their functional importance is less clear.
Interestingly, SBMs contribute to the structural integrity
of PML (promyelocytic leukaemia protein) bodies, which
are known to require SUMOylation of PML and other
constituent proteins for their formation [30,31]. This suggests
that SBMs in different contexts may act as stable protein
interaction domains, or as part of the enzymatic machinery.
Neither possibility has been fully explored for Pc2.
For all four known substrates of Pc2, in vitro E3 ligase
activity of purified Pc2 has been demonstrated, although it
appears to be a relatively weak E3 in vitro, despite robust
activity in vivo. One possibility is that other proteins within
the polycomb complex play a role in E3 activity in vivo.
Deletion of the C-box from Pc2 does not affect E3 activity
towards CtBP, suggesting that localization to polycomb
bodies is not essential for E3 activity. Pc2 does not have
any significant homology to other known SUMO E3s, and
there is no evidence for the involvement of a RING (really
interesting new gene) domain protein in its E3 activity.
Indeed, the C-box deletion prevents interaction of Pc2 with
Ring1, without affecting its E3 activity towards CtBP [7].
However, a role for other Pc2-interacting proteins in E3
activity cannot be completely ruled out.
Pc2 and the cell cycle
Recent evidence has demonstrated a critical role for several
polycomb proteins in regulating cell cycle progression.
Increased expression of both Cbx7 and Cbx8 (Pc3) allows
primary cells to bypass senescence, and Cbx8 can co-
operate with active Ras, or c-Myc, to increase the growth rate
of primary cells [16,32]. Cbx7 expression is seen to decrease as
primary cells enter replicative senescence, and specific knock-
down of Cbx7 can induce premature senescence in culture.


C The Authors Journal compilation

C 2007 Biochemical Society
1404 Biochemical Society Transactions (2007) Volume 35, part 6
However, Pc2 appears to function somewhat differently,
in that forced overexpression in CV-1 cells resulted in a
cell cycle block in G2 [33]. In contrast, overexpression of
a truncation mutant, which lacks the C-box was able to
transform Rat-1A cells in vitro, as judged by growth in
soft agar [8]. This Pc2 mutant is unable to interact with
Ring1 and is delocalized from polycomb bodies, perhaps
suggesting that deregulation of a normal function of Pc2
may be responsible for cellular transformation. Although the
apparent differences between Pc2 and Cbx7 and Cbx8 may
be due to different experimental approaches, it seems likely
that Pc2 functions differently from other Cbx relatives. Since
there is evidence for different varieties of the PRC1 complex,
with different Cbx proteins, it is tempting to speculate that
the balance of Cbx proteins within the cell helps dictate
overall PRC1 function, in part by recruiting specific activities.
Other activities for Pc2
In addition to the core components of the PRC1 complex, Pc2
has also been shown to interact, via its C-terminus, with the
histone methyltransferase SUV39H1 (suppressor of variega-
tion 3–9 homologue 1) [34]. This C-terminal domain of Pc2
appears to be a major site for the recruitment of other pro-
teins, including CtBP, SUV39H1, Ubc9 and Ring1 (Figure 2).
Thus it appears that Pc2 interacts with proteins involved in
histone methylation and ubiquitination, protein SUMOyla-
tion and transcriptional repression. In addition, both HIPK2
and Dnmt3a, which are SUMO substrates for Pc2, are also
themselves enzymes. HIPK2 is not only a substrate for
SUMOylation, but it in turn phosphorylates Pc2 [25]. Phos-
phorylation of human Pc2 by HIPK2 in response to DNA
damage has been suggested to regulate SUMOylation of Pc2
itself and its E3 activity towards HIPK2. This raises the pos-
sibility of a regulatory loop involving mutual phosphoryla-
tion and SUMOylation. One caveat to this is that the primary
phosphorylation site in Pc2 is a threonine residue directly
following the VKPE SUMOylation site (Figure 1), which is
not conserved in other vertebrates. However, other sites in
Pc2 were phosphorylated by HIPK2, which in other species
may be more important. A conserved Akt (also called protein
kinase B) site is present in all but one of the Pc2 sequences
shown in Figure 1, and we have preliminary evidence that Pc2
interacts with, and is phosphorylated by, Akt (J.C. Merrill,
M.H. Kagey and D. Wotton, unpublished work). Thus it is
possible that multiple regulatory inputs affect Pc2 activity. It
will be of interest to see which of the numerous enzymatic
activities with which Pc2 interacts is important for its function
and how these inputs impinge on PRC function. In conclu-
sion, we suggest that Pc2 may act as a platform to recruit
multiple enzymatic activities, and that it may co-ordinate the
interaction of these protein modification pathways.


C The Authors Journal compilation

C 2007 Biochemical Society
References
1 Fischle, W., Wang, Y., Jacobs, S.A., Kim, Y., Allis, C.D. and
Khorasanizadeh, S. (2003) Genes Dev. 17, 1870–1881
2 Min, J., Zhang, Y. and Xu, R.M. (2003) Genes Dev. 17, 1823–1828
3 Jacobs, S.A. and Khorasanizadeh, S. (2002) Science 295, 2080–2083
4 Nielsen, P.R., Nietlispach, D., Mott, H.R., Callaghan, J., Bannister, A.,
Kouzarides, T., Murzin, A.G., Murzina, N.V. and Laue, E.D. (2002) Nature
416, 103–107
5 Bernstein, E., Duncan, E.M., Masui, O., Gil, J., Heard, E. and Allis, C.D.
(2006) Mol. Cell. Biol. 26, 2560–2569
6 Satijn, D.P. and Otte, A.P. (1999) Mol. Cell. Biol. 19, 57–68
7 Kagey, M.H., Melhuish, T.A. and Wotton, D. (2003) Cell 113, 127–137
8 Satijn, D.P., Olson, D.J., van der Vlag, J., Hamer, K.M., Lambrechts, C.,
Masselink, H., Gunster, M.J., Sewalt, R.G., van Driel, R. and Otte, A.P.
(1997) Mol. Cell. Biol. 17, 6076–6086
9 Pirrotta, V. (1998) Cell 93, 333–336
10 Simon, J.A. and Tamkun, J.W. (2002) Curr. Opin. Genet. Dev. 12,
210–218
11 Schuettengruber, B., Chourrout, D., Vervoort, M., Leblanc, B. and
Cavalli, G. (2007) Cell 128, 735–745
12 Czermin, B., Melfi, R., McCabe, D., Seitz, V., Imhof, A. and Pirrotta, V.
(2002) Cell 111, 185–196
13 Levine, S.S., Weiss, A., Erdjument-Bromage, H., Shao, Z., Tempst, P. and
Kingston, R.E. (2002) Mol. Cell. Biol. 22, 6070–6078
14 Muller, J., Hart, C.M., Francis, N.J., Vargas, M.L., Sengupta, A., Wild, B.,
Miller, E.L., O’Connor, M.B., Kingston, R.E. and Simon, J.A. (2002)
Cell 111, 197–208
15 Shao, Z., Raible, F., Mollaaghababa, R., Guyon, J.R., Wu, C.T., Bender, W.
and Kingston, R.E. (1999) Cell 98, 37–46
16 Dietrich, N., Bracken, A.P., Trinh, E., Schjerling, C.K., Koseki, H.,
Rappsilber, J., Helin, K. and Hansen, K.H. (2007) EMBO J. 26,
1637–1648
17 Sewalt, R.G., Gunster, M.J., van der Vlag, J., Satijn, D.P. and Otte, A.P.
(1999) Mol. Cell. Biol. 19, 777–787
18 Turner, J. and Crossley, M. (1998) EMBO J. 17, 5129–5140
19 Schaeper, U., Subramanian, T., Lim, L., Boyd, J.M. and Chinnadurai, G.
(1998) J. Biol. Chem. 273, 8549–8552
20 Chinnadurai, G. (2002) Mol. Cell 9, 213–224
21 Kagey, M.H., Melhuish, T.A., Powers, S.E. and Wotton, D. (2005) EMBO J.
24, 108–119
22 Lin, X., Sun, B., Liang, M., Liang, Y.Y., Gast, A., Hildebrand, J., Brunicardi,
F.C., Melchior, F. and Feng, X.H. (2003) Mol. Cell 11, 1389–1396
23 Li, B., Zhou, J., Liu, P., Hu, J., Jin, H., Shimono, Y., Takahashi, M. and Xu, G.
(2007) Biochem. J. 405, 369–378
24 Long, J., Zuo, D. and Park, M. (2005) J. Biol. Chem. 280, 35477–35489
25 Roscic, A., Moller, A., Calzado, M.A., Renner, F., Wimmer, V.C., Gresko, E.,
Ludi, K.S. and Schmitz, M.L. (2006) Mol. Cell 24, 77–89
26 Hannich, J.T., Lewis, A., Kroetz, M.B., Li, S.J., Heide, H., Emili, A. and
Hochstrasser, M. (2005) J. Biol. Chem. 280, 4102–4110
27 Song, J., Durrin, L.K., Wilkinson, T.A., Krontiris, T.G. and Chen, Y. (2004)
Proc. Natl. Acad. Sci. U.S.A. 101, 14373–14378
28 Song, J., Zhang, Z., Hu, W. and Chen, Y. (2005) J. Biol. Chem. 280,
40122–40129
29 Minty, A., Dumont, X., Kaghad, M. and Caput, D. (2000) J. Biol. Chem.
275, 36316–36323
30 Muller, S., Matunis, M.J. and Dejean, A. (1998) EMBO J. 17, 61–70
31 Shen, T.H., Lin, H.K., Scaglioni, P.P., Yung, T.M. and Pandolfi, P.P. (2006)
Mol. Cell 24, 331–339
32 Gil, J., Bernard, D., Martinez, D. and Beach, D. (2004) Nat. Cell Biol. 6,
67–72
33 Dahiya, A., Wong, S., Gonzalo, S., Gavin, M. and Dean, D.C. (2001)
Mol. Cell 8, 557–569
34 Sewalt, R.G., Lachner, M., Vargas, M., Hamer, K.M., den Blaauwen, J.L.,
Hendrix, T., Melcher, M., Schweizer, D., Jenuwein, T. and Otte, A.P.
(2002) Mol. Cell. Biol. 22, 5539–5553
Received 5 June 2007
doi:10.1042/BST0351401