SUMOylation of Blimp-1 is critical for plasma cell ...

sscciieennttiiffiiccrreeppoorrtt

SUMOylation of Blimp-1 is critical for plasma cell
differentiation

Hsia-Yuan Ying1,2, Shin-Tang Su1, Pang-Hung Hsu3,4, Che-Chang Chang5, I-Ying Lin1,2, Yu-Hsuan Tseng1,
Ming-Daw Tsai1,3, Hsiu-Ming Shih5 & Kuo-I Lin1+

1Genomics Research Center, Academia Sinica, 2Institute of Biochemistry and Molecular Biology, National Yang-Ming University,
3Institute of Biological Chemistry, Academia Sinica, Taipei, Taiwan, 4Department of Life Science, National Taiwan Ocean University
and Institute of Bioscience and Biotechnology, National Taiwan Ocean University, Keelung, Taiwan, and 5Institute of Biomedical

Sciences, Academia Sinica, Taipei, Taiwan

Transcriptional repressor B lymphocyte-induced maturation proliferation and in the mature B-cell transcription programme
protein-1 (Blimp-1) is a master regulator of plasma cell [6–9], leading to the specification of plasma cell fate. Blimp-1
differentiation. Here we show that Blimp-1 is covalently modified was recently shown to be modified by SUMO1 [10], but
by SUMO1 at lysine 816, a modification mediated by SUMO E3 the underlying molecular and biological consequences are
ligase PIAS1. Mutation of Blimp-1 lysine 816 reduces transcrip- not known.
tional repression—correlating with a reduced interaction with a
histone deacetylase, HDAC2—and impairs differentiation of SUMOylation, a post-translational modification that involves
antibody-secreting cells. Thus, the SUMO pathway critically the covalent conjugation of the SUMO protein to a target
regulates Blimp-1 function during plasma cell differentiation. substrate, regulates many important molecular and cellular
Keywords: differentiation; plasma cell; SUMO; transcriptional processes, including transcriptional control [11]. The
repression conjugation of SUMO with its substrate involves the E3 ligases,
which enhance substrate selectivity in vivo [12]. The PIAS family,
EMBO reports (2012) 13, 631–637. doi:10.1038/embor.2012.60 one of the three types of mammalian E3 ligases, consists of four
proteins: PIAS1, PIAS2 (PIAS2-a and PIAS2-b), PIAS3 and PIAS4
INTRODUCTION [13]. The interactions of transcription factors with PIAS proteins can
either positively or negatively affect transcriptional activities [14].
B lymphocyte-induced maturation protein-1 (Blimp-1), a
transcriptional repressor, has a central role in the differentiation In this report, we demonstrate the significance of
of B cells into antibody-secreting plasma cells [1] as well as many PIAS1-dependent SUMOylation of Blimp-1 in plasma
other cell lineages of multiple organisms [2]. Although plasma cell cell differentiation.
fate specification can be initiated independently of Blimp-1,
Blimp-1 is required for complete plasma cell differentiation [3]. RESULTS
Blimp-1 is present in virtually all plasma cells induced by T cell- Blimp-1 is modified by SUMO1 at Lys816
dependent or -independent antigen immunization [4], and
gene-targeted deletion of the Blimp-1 gene, Prdm1, in mouse We first verified by co-immunoprecipitation (co-IP) that endogen-
B cells blocks their ability to mount an antibody response [1]. ous Blimp-1 is sumoylated in plasma cells NCI-H929 (H929) that
constitutively express Blimp-1, and in lipopolysaccharide (LPS)-
By interacting with several chromatin-modifying enzymes stimulated mouse splenic B cells that express induced Blimp-1.
including HDAC1/2 [5], Blimp-1 suppresses genes implicated in Indeed, SUMO1 is present in anti-Blimp-1 immunoprecipitates in
both cultured cells, and the slower-migrating form shows a
1Genomics Research Center, Academia Sinica, Taipei 115, Taiwan molecular-weight shift of SUMO1 conjugate, which suggests the
2Institute of Biochemistry and Molecular Biology, National Yang-Ming University, sumoylated Blimp-1 (Fig 1A). Protein SUMOylation occurs at the
Taipei 112, Taiwan lysine (K) residue of the consensus SUMOylation motif, cKXE,
3Institute of Biological Chemistry, Academia Sinica, Taipei 115, Taiwan where c represents hydrophobic residues and X represents any
4Department of Life Science, National Taiwan Ocean University and Institute of amino acid [15]. Using the computational SUMOylation
Bioscience and Biotechnology, National Taiwan Ocean University, Keelung 202, prediction programme SUMOsp [16], we found that there are
Taiwan five putative SUMOylation sites in human Blimp-1, which are
5Institute of Biomedical Sciences, Academia Sinica, Taipei 115, Taiwan located at amino-acid residues K235, K245, K574, K767 and K816
+Corresponding author. Tel: þ 886 2 2787 1253; Fax: þ 886 2 2789 9931; (accession number NP_001189 Fig 1B) and are conserved in
E-mail: [email protected] mouse Blimp-1. Five individual Blimp-1 mutants, K235R, K245R,

Received 20 October 2011; revised 16 April 2012; accepted 17 April 2012;
published online 4 May 2012

&2012 EUROPEAN MOLECULAR BIOLOGY ORGANIZATION EMBO reports VOL 13 | NO 7 | 2012 6 3 1

scientific report SUMO modulates the function of Blimp-1
H.-Y. Ying et al

A H929 Spl + LPS IP
IP IP
IP

IgG Blimp-1 IgG Blimp-1 Input IgG Blimp-1 IgG Blimp-1 Input

130 130

100 kDa IB: Blimp-1 100 kDa IB: Blimp-1

IB: SUMO1 IB: SUMO1

B EK235NE IK574YE VK816QE D
Proline-rich Zn Acidic
Acidic PR FLAG–Blimp-1
1 PK245RE 825 EGFP-SUMO1 – + + +
EK767EE
WT 170 IP: FLAG
WT 130 IB: GFP
K235R 170
K245R 130 IP: FLAG
K574R IB: FLAG
K767R 100
K816R Input
40 IB: GFP
W130 Input
T IB: FLAG
100
WTkDa
K816R
K818R
C

FLAG–Blimp-1
EGFP-SUMO1 – + + + + + +

170 IP: FLAG F Ab2 Ab3 Ab4 Ab5 Ab6 Ab7
130 IB: GFP
170 90 V LY Q E Q T G G97
130 IP: FLAG Precursor ion: 683.3368 2+ Ay7 Ay6 Ay5 Ay4 Ay3 Ay2 Ay1 Bb2 Bb3
100 IB: FLAG Bb32+
815V L K Q E818
40 Input
130 IB: GFP By2 By2
Input
100 IB: FLAG
kDa

E Bb2 Bb3

Q

GST–Blimp-1 By2 Ay1 Ay2 Ay3 Ay4 Ay5 Ay6 Ay7 By3

–FLAG WT K235R K245R K574R K767R K816R GT Q E Q Y
–+ –+ – + –+ –+
SUMOylation – + Ab2 Ab3 Ab4 Ab5 Ab6Ab7
Q E Q TG

170 *SUMO from

130 * * * * * Blimp-1

110 IB:
FLAG

70

kDa 200 400 600 800 1,000 1,200
m/z

Fig 1 | Blimp-1 is sumoylated at residue 816. (A) Endogenous Blimp-1 is modified by SUMO1. Nuclear extracts harvested from H929 cells or day 3
LPS-stimulated splenic B cells (Spl þ LPS) were subjected to IP using anti-Blimp-1 or control IgG, followed by IB analysis using indicated antibodies.
(B) Potential Blimp-1 SUMOylation sites predicted by SUMOsp (http://sumosp.biocuckoo.org/online.php). (C,D) Lysine 816 is the unique SUMO1
acceptor site. The EGFP-SUMO1 expression plasmid was co-transfected with FLAG–Blimp-1 (WT) or mutant forms of Blimp-1 (K235R, K245R, K574R,
K767R or K816R in C and K816R or E818A in D) into 293T cells. Cell lysates were immunoprecipitated with anti-FLAG. Immunoprecipitates and input
lysates were analysed by IB analysis. Arrowheads in C and D indicate the Blimp-1–EGFP-SUMO1 conjugate. (E) IB analysis with anti-FLAG showing
the in vitro SUMOylation of purified GST- and FLAG-tagged WT Blimp-1 and various Blimp-1 mutants. Asterisks indicate conjugated GST–Blimp-1–
FLAG and SUMO1, and the arrow indicates GST–Blimp-1–FLAG. (F) The MS/MS spectrum confirms the SUMOylation site at K816 of human Blimp-1.
The cross-linked peptide containing chain A (90VYQEQTGG97, from SUMO1_HUMAN) and chain B (815VKQE818, from Blimp-1_HUMAN) was
identified by mass spectrometry. All b and y product ions from the cross-linked peptide are labelled in the spectrum. The amino-acid sequence of the
cross-linked peptide is confirmed by the difference of the b and y product ions. The confirmed amino-acid residues are shown in the figure. Blimp-1,
B lymphocyte-induced maturation protein-1; EGFP, enhanced green fluorescent protein; GST, glutathione S-transferase; IB, immunoblot; IP,
immunoprecipitation; LPS, lipopolysaccharide; MS/MS, tandem mass spectrometry; WT, wild type.

K574R, K767R and K816R, each of which contains a different similar to wild-type (WT) Blimp-1, K235R, K245R, K574R and
mutated SUMO acceptor site, were generated. These FLAG- K767R Blimp-1 were sufficient to conjugate with EGFP-SUMO1,
tagged (FLAG–)Blimp-1 mutants were expressed in the nucleus because a protein species with a higher molecular mass (B150
(supplementary Fig S1A online). Immunoblot (IB) analysis using kDa) corresponding to the addition of EGFP-SUMO1 conjugate
cell lysates from 293T cells transfected with vectors expressing was detected (Fig 1C). In sharp contrast, K816R Blimp-1 failed to
FLAG–Blimp-1 or various FLAG–Blimp-1 mutants and enhanced be sumoylated (Fig 1C). Another Blimp-1 mutant, E818A, which
green fluorescent protein (EGFP)-SUMO1 demonstrated that, carries the disrupted VKQE SUMOylation motif in Blimp-1, was

6 3 2 EMBO reports VOL 13 | NO 7 | 2012 &2012 EUROPEAN MOLECULAR BIOLOGY ORGANIZATION

SUMO modulates the function of Blimp-1 scientific report
H.-Y. Ying et al

also not modified by SUMO1 (Fig 1D). An in vitro SUMOylation including PIAS1, PIAS2-a, PIAS2-b, PIAS3 and PIAS4, were used
assay further demonstrated that WT, K235R, K245R, K574R and to test their effects on Blimp-1 SUMOylation. Interestingly, PIAS1
K767R Blimp-1 were conjugated to SUMO in vitro, whereas dramatically enhanced Blimp-1 SUMOylation, even though the
K816R Blimp-1 failed to be sumoylated (Fig 1E), showing again expression of PIAS1 was lower than that of the other PIAS proteins
that K816 is the major SUMO acceptor site of Blimp-1. Indeed, (Fig 3A). In contrast, PIAS3 was much less capable of promoting
liquid chromatography–tandem mass spectrometry revealed a Blimp-1 SUMOylation, and PIAS2-a, PIAS2-b and PIAS4 were
SUMO peptide branch at K816 of Blimp-1. Lysates prepared from nearly inert in the co-transfection assay (Fig 3A). The direct effect
293T cells that transiently expressed FLAG–Blimp-1 and EGFP- of PIAS1 on promoting Blimp-1 SUMOylation was confirmed by
SUMO1 and that had been processed for anti-FLAG IP and trypsin an in vitro SUMOylation assay in which bacterially expressed
digestion were used for liquid chromatography–tandem mass PIAS1 was able to further increase the SUMO1-modified Blimp-1
spectrometry identification (Fig 1F). in a dose-dependent manner (Fig 3B). We also detected the
interaction between Blimp-1 and PIAS1 at endogenous levels by
Along this line, we wondered whether SUMO2 or SUMO3 co-IP in lysates from H929 plasma cells (Fig 3C). The RING-finger-
can, like SUMO1, modify Blimp-1. In spite of the similar like zinc-binding domain (RLD) of PIAS1 has the SUMO E3 ligase
expression levels of all three EGFP-SUMOs, EGFP-SUMO2 and enzymatic activity [13]. We found that PIAS1 without the RLD
EGFP-SUMO3 failed to conjugate with Blimp-1 (supplementary (DRLD) had decreased ability to promote SUMO1 conjugation
Fig S2 online). with Blimp-1 (Fig 3D), indicating that the RLD domain of PIAS1 is
required in this context. Similarly, depletion of PIAS1 by a
SUMOylation of Blimp-1 is of functional importance lentiviral vector that expresses the short hairpin RNA against
PIAS1, PIAS1i, led to reduced SUMO1 conjugation as compared
Prdm1-deficient (knockout (KO)) splenic B cells that are stimu- with the effect of the control short hairpin RNA, Ctrli (Fig 3E).
lated in vitro by LPS fail to cease cell cycling and to commit to the
plasma cell fate [1]. To address the biological importance of We mapped the Blimp-1 region involved in the PIAS1
SUMO modification of Blimp-1, we examined the proliferation interaction by using various plasmids encoding deleted forms of
rate of LPS-stimulated KO B cells that were retrovirally FLAG–Blimp-1 (supplementary Fig S4A online). These deleted
reconstituted with WT or K816R Blimp-1. Proliferation rate was forms of FLAG–Blimp-1 were expressed in the nucleus
monitored by dilution of the fluorescent tracking dye PKH26. (supplementary Fig S1B online). Analysis indicated that amino-
Notably, after stimulation, cell proliferation was halted by WT acid residues 527–574 of Blimp-1 were crucial for the PIAS1
Blimp-1, but K816R-expressing cells divided as effectively as cells interaction (supplementary Fig S4B online). As expected, a Blimp-
that had been transduced with control vector (Fig 2A). Consis- 1 deletion lacking this region (D527–574) failed to interact with
tently, transduction of WT or K245R Blimp-1 in stimulated KO B PIAS1 (Fig 3F) and was less able to be modified by EGFP-SUMO1
cells restored the formation of plasmacytic cells that express the (Fig 3G). LPS-stimulated KO B cells expressing the D527–574
CD138 plasma cell surface marker (Fig 2B). In contrast, despite Blimp-1 mutant showed decreased formation of CD138 þ
the similar expression of the reconstituted Blimp-1 proteins plasmacytic cells (Fig 3H) and IgM production (Fig 3I) as
(supplementary Fig S3 online), the frequency of CD138 þ cells compared with cells that expressed WT Blimp-1, even though
generated by K816R reconstitution was reduced (Fig 2B). both proteins were expressed at similar levels (supplementary
Similarly, K816R, unlike WT and K245R Blimp-1, was much less Fig S3 online).
able to compensate for IgM production in KO B cells (Fig 2C). The
suggestion that disruption of SUMOylation of Blimp-1 impairs its As the interaction with PIAS1 appears to be important for the
ability to trigger the plasma cell fate was further supported by effect of Blimp-1, we wondered whether PIAS1 contributes to
another reconstitution assay in which E818A Blimp-1 was plasma cell differentiation. Indeed, knockdown of PIAS1 in LPS-
expressed in LPS-stimulated KO B cells and led to results that triggered differentiating splenic B cells caused reduced production
were similar to K816R expression (Fig 2B,C). of antibody (Fig 3J). However, it remains possible that knocking
down PIAS1 during plasma cell differentiation may also affect the
Re-introduction of WT Blimp-1 in KO B cells resulted in Blimp-1. We thus ectopically expressed Blimp-1 in a human
downregulation of Pax5 and upregulation of Xbp1 (Fig 2D), a lymphoblastoid cell line SKW 6.4 (SKW) possessing the ability to
crucial gene for antibody secretion [17]. The induction of Xbp1 produce IgM [9], and found that knockdown of PIAS1 reduced
is due to the de-repression effect from Blimp-1–mediated Blimp-1-driven IgM production (Fig 3K). This result is consistent
suppression of the gene encoding Pax5, which suppresses Xbp1 with another observation that SKW cells that were transduced with
[8]. The mRNA1 level for Pax5 was repressed less, and the K816R Blimp-1 produced less IgM when compared with WT
induction of Xbp1 mRNA was hampered in cells that expressed Blimp-1-transduced SKW cells (supplementary Fig S5 online).
K816R Blimp-1 (Fig 2D). Similarly, the other two Blimp-1 target
genes, Ciita and Bcl2a1 [7,18], were repressed less by K816R than SUMO affects transcriptional repression by Blimp-1
by WT Blimp-1 (Fig 2D), in agreement with the less-differentiated
phenotype in K816R-expressing cells. Lastly, we examined the molecular mode of action underlying the
consequence of Blimp-1 SUMOylation. Cycloheximide-based
Together, these data show that Blimp-1, in the absence of protein chase experiment showed that K816R Blimp-1 is not
SUMOylation, is not as effective at blocking cell proliferation and markedly different with respect to its protein stability, in
is defective in driving plasma cell formation. comparison with WT Blimp-1 (Fig 4A). We next examined if
SUMO-modified Blimp-1 has changed transcriptional repression
Involvement of PIAS1 in Blimp-1 SUMOylation activity. We previously showed that Blimp-1 fused with the Gal-4
DNA-binding domain (DBD–Blimp-1) effectively suppresses
We next dissected the SUMO pathways that contribute to Blimp-1
SUMOylation. Various haemagglutinin-tagged human E3 ligases,

&2012 EUROPEAN MOLECULAR BIOLOGY ORGANIZATION EMBO reports VOL 13 | NO 7 | 2012 6 3 3

scientific report SUMO modulates the function of Blimp-1
H.-Y. Ying et al

A C 250 P < 0.01 KO + YFP
120 200 P < 0.005 Ctrl + YFP
100 150 KO + Blimp-1
80Count YFP (MFI 372 ± 6) P < 0.001 IgM-secreting cells 100 KO + K245R
60 Blimp-1 (MFI 685 ± 39) (per 500 cells) KO + K816R
40 K816R (MFI 326 ± 22) 50 KO + E818A
20 0
0 101 102 103 104
100 PKH26

B KO + YFP Ctrl + YFP KO + Blimp-1 D Pax5 Ciita YFP
104 104 104 P < 0.01 P < 0.005 WT
Repression fold change Repression fold change Repression fold change K816R
5 6

103 103 103 4 5

102 102 102 3 4
2
101 2.63% 101 42.74% 101 48.95% 3

100 100 100 2
100 101 102 103 104 100 101 102 103 104 100 101 102 103 104
1 1

B220 KO + K245R KO + K816R KO + E818A 0 0 P < 0.01
104 104 104 Bcl2a1 Xbp1
103 103 103 P < 0.05 Induced fold change
102 102 102 12
6
10
5
8
4

101 46.55% 101 15.29% 101 15.5% 3 6

100 100 100 2 4
100 101 102 103 104 100 101 102 103 104 100 101 102 103 104 1
2

CD138 0 0

Fig 2 | SUMO modification of Blimp-1 is required for plasma cell differentiation. (A) Histogram of flow-cytometric analysis of PKH26 intensity.
Prdm1f/f CD19Cre þ (KO) B cells were labelled with PKH26, followed by LPS stimulation and retroviral vector transduction. Flow-cytometric analysis
of PKH26 levels using the YFP þ gate. The numbers indicate the PKH26 MFI±s.d. from three experiments. (B) Flow-cytometric analysis of surface
expression of B220 and CD138 on the stimulated and retrovirally transduced (YFP þ ) splenic B cells from Prdm1f/f CD19Cre À (Ctrl) or KO mice.

Results are representative of three independent experiments. (C) ELISPOT analysis of the number of IgM-secreting cells in retrovirally transduced
(YFP þ ) Ctrl or KO B cells stimulated with LPS. (D) RT–QPCR shows the relative mRNA levels of Ciita, Bcl2a1, Pax5 and Xbp-1 regulated by WT
and K816R Blimp-1 in YFP þ cells sorted as described in C. Blimp-1, B lymphocyte-induced maturation protein-1; Ctrl, control; KO, knockout; LPS,
lipopolysaccharide; MFI, mean fluorescence intensity; WT, wild type; YFP þ , yellow fluorescent protein þ .

luciferase reporter activity driven by a thymidine kinase (tk) DISCUSSION
promoter composed of four Gal4-binding sites, (Gal4)4-tk-Luc
[19]. Here the (Gal4)4-tk-driven luciferase activity was suppressed Here we provide evidence that, among SUMO family members,
by DBD–Blimp-1, but was suppressed to a lesser extent by SUMO1 selectively mediated Blimp-1 SUMOylation and that
DBD–K816R Blimp-1 (Fig 4B). Similar results were observed in Blimp-1 is subjected to PIAS1-mediated SUMOylation at lysine
Blimp-1-dependent repression of the activities of the CIITA 816, which has a crucial role in plasma cell differentiation. We here
promoter III (Fig 4C). These data suggest that a SUMO conjuga- show that the expression of Blimp-1 downstream genes, including
tion-defective Blimp-1 is a less potent transcriptional repressor. Pax5, Ciita, Bcl2a1 and Xbp1, was altered in K816R-reconstituted
KO B cells when compared with that in WT Blimp-1-reconstituted
We next determined whether the effect of SUMOylation on the cells. The changes in the expression of these genes in K816R- versus
transcriptional repression ability of Blimp-1 results from an WT Blimp-1-rescued cells are marginal but statistically significant,
alteration in the DNA-binding ability of Blimp-1. WT Blimp-1 which could be due to the heterogeneity of splenic B cells that
and K816R bound similarly with the endogenous target gene leads to unsynchronized differentiation. Additionally, the reason
loci, as demonstrated by the anti-Blimp-1 chromatin IP assay, that SUMOylation-defective forms of Blimp-1 cannot drive plasma
which examined the binding of WT Blimp-1 and K816R with cell fate is not due to their influence on cell viability, because we
endogenous target sites, including Myc, Pax5 and Ciita, in B observed a similar apoptotic rate in cells reconstituted with mutated
cells (Fig 4D). HDAC1/2 is involved in Blimp-1-mediated or WT Blimp-1 (supplementary Fig S6 online).
gene suppression [5]. We found that HDAC2 was
co-immunoprecipitated with WT Blimp-1, as expected (Fig 4E). Although there are five members of the PIAS family and all are
Notably, the level of HDAC2 that interacted with K816R Blimp-1 expressed in mouse splenic B cells (unpublished observations),
was statistically decreased (Fig 4E). PIAS1 functioned most effectively as the E3 SUMO ligase of
Blimp-1. The interaction between Blimp-1 and PIAS1 appears to

6 3 4 EMBO reports VOL 13 | NO 7 | 2012 &2012 EUROPEAN MOLECULAR BIOLOGY ORGANIZATION

SUMO modulates the function of Blimp-1 scientific report
H.-Y. Ying et al

A PlAS1 PlAS3 PlAS4 PlAS2-α PlAS2-β F H KO + YFP

HA-PIAS – + + + + + + + + + + + + ++ + FLAG–Blimp-1 – W 104
EGFP-SUMO1 + + + + + + + + + + + + + ++ + HA-PIAS1 + T 103
FLAG–Blimp-1 + + + + + + + + + + + + + ++ + 100 102
Δ527–574
70 + B220+ 101 3.6%
130
IP: FLAG 100100 101 102 103 104
100 IB: HA
130 IB: 100 KO + Blimp-1 KO + Δ527–574
FLAG IP: FLAG 104 104
100 70 IB: FLAG 103 103
100 IB: 130 102 102
HA Input 101 46.2% 101 15.7%
70 100 IB: HA
IB: kDa 100100 101 102 103 104100100 101 102 103 104
55 GFP Input
IB: FLAG
40
kDa CD138

B PlAS1 C G WT I
ΔW5T27–574 K
SUMOylation – +++ FLAG–Blimp-1 Δ527–574 P<0.001
GST–Blimp-1 IgM-secreting cells 200 KO + YFP
–FLAG 170 (per 500 cells) 150 KO + Blimp-1
++++ IP EGFP-SUMO1 – – + +
130 KO + Δ527–574
Input IgG PIAS1 170 100
110
*SUMO form 130 IB: 130 IB: 50
Blimp-1 100 FLAG
** * Blimp-1 100 0

IB: 70 IB: 40 IB:
FLAG kDa PIAS1 kDa GFP

70 J P<0.005 P<0.05 P<0.001

110 PIAS1 30,000 Ctrli IgM secretion 100
kDa (ng per 106 cells) 90 Ctrli
IgM secretion 25,000 PIAS1i 80 PIAS1i
(ng per 105 cells) 70
D E Ctrli PIAS1i 20,000 60
50
HA-PIAS1 – – WT ΔRLD EGFP-SUMO1 – + – + 15,000 40
EGFP-SUMO1 – ++ + FLAG–Blimp-1 + + + + 10,000 30
FLAG–Blimp-1 + ++ + 20
170 5,000 10
0
170 130 IB: 0
130 FLAG Experiment: 1 GFP GFP–Blimp-1
IB: 2
100 FLAG 0.54 ± 0.13 Ctrli PIAS1i
100 100 (P<0.01)
IB: Modified/unmodified 1 IB: Ctrli PIAS1i GFP– GFP–
70 HA PIAS1 100 IB:
100 GFP Blimp-1 GFP Blimp-1
IB: IB: 70 PIAS1
GFP 70 GFP 130 IB:
40 55 IB: 100 Blimp-1
kDa 40 kDa tubulin IB:
kDa 70 PIAS1
55
IB:
kDa tubulin

Fig 3 | PIAS1 regulates Blimp-1 SUMOylation. (A) PIAS1 enhances SUMO1 modification of Blimp-1 in a co-transfection experiment in 293T cells.
(B) IB analysis with anti-FLAG to show that in vitro SUMOylation of purified GST- and FLAG-tagged WT Blimp-1 is enhanced by bacterially
expressed PIAS1. The asterisks and arrow are as described for Fig 1E. (C) Endogenous Blimp-1 interacts with endogenous PIAS1 in H929 plasma cells.
(D) The RLD of PIAS1 is required for Blimp-1 SUMOylation as demonstrated by a co-transfection assay in 293T cells. (E) Depletion of PIAS1 reduced
SUMOylation of Blimp-1. 293T cells transduced with vectors expressing shRNA for PIAS1 (PIAS1i) or control shRNA (Ctrli) for 2 days were then
co-transfected with vectors encoding the indicated proteins for 2 days. The relative ratios of the quantification results for Blimp-1–EGFP-SUMO1
(arrowhead; modified) versus Blimp-1 (arrow; unmodified) band intensities are shown. (F,G) Plasmids encoding indicated proteins were co-transfected
into 293T cells, followed by IP and IB analysis in F and IB analysis in G. (H) Flow-cytometric analysis of surface expression of B220 and CD138 on
LPS-stimulated KO splenic B cells transduced with indicated retroviral vectors. (I) The number of IgM-secreting cells was determined by ELISPOT
assay for cells expressing the indicated proteins. Sorted YFP þ cells from H were used here. (J) Knockdown of PIAS1 reduces the formation of
antibody-secreting cells. Stimulated splenic B cells were transduced with lentiviral construct encoding GFP and expressing either PIASi or Ctrli,
followed by ELISA analysis of the amounts of IgM from supernatants of sorted GFP þ cells. Results are from triplicate data from two independent
experiments. (K) PIAS1 knockdown reduced the IgM production in SKW cells ectopically expressing GFP or GFP À Blimp-1 fusion protein by
lentiviral vectors. IB analysis in J and K showed knockdown of PIAS1 in sorted GFP þ PIAS1i-expressing cells. Arrowheads in A,D and G indicate the
SUMO1-conjugated form. Blimp-1, B lymphocyte-induced maturation protein-1; EGFP, enhanced green fluorescent protein; ELISA, enzyme-linked
immunosorbent assay; GFP, green fluorescent protein; HA, haemagglutinin; IB, immunoblot; KO, knockout; LPS, lipopolysaccharide; RLD, RING-
finger-like zinc-binding domain; shRNA, short hairpin RNA; WT, wild type; YFP þ , yellow fluorescent protein þ .

be of biological importance, as a Blimp-1 mutant that lacks the co-repressor for some transcription factors, whereby it prevents
region required for PIAS1 interaction was much less capable of their binding to target genes without the involvement of SUMO
promoting plasma cell differentiation. Also, depletion of PIAS1 [20]. Our result that lack of RLD in PIAS1 shows reduced ability to
attenuated the Blimp-1-mediated production of IgM. In addition promote Blimp-1 SUMOylation suggests that PIAS1 relies on its E3
to serving as a SUMO E3 ligase, PIAS1 is a transcriptional ligase domain to regulate Blimp-1.

&2012 EUROPEAN MOLECULAR BIOLOGY ORGANIZATION EMBO reports VOL 13 | NO 7 | 2012 6 3 5

scientific report SUMO modulates the function of Blimp-1
H.-Y. Ying et al

A WT K816R CE
24 24
CHX (h) 0 P < 0.05
Repression fold change
100 80 8 10

IB: FLAG WT K816R
Blimp-1 8 70 IP:FLAG

6 55 IB: HDAC2
1 0.6 ± 0.02 (P < 0.001)
55 IB: 4 IP / input
tubulin IP: FLAG

kDa 2 100 IB: FLAG
1 0.76 0.38 0.16 1 0.76 0.38 0.14 70
Input
B Proline-rich Zn Acidic 0 55 IB: HDAC2
K816R FLAG WT K816R 100
DBD–Blimp-1 Input
Acidic PR IB: kDa IB: FLAG
100 Blimp-1
Gal4-DBD
55 IB:
DBD–K816R kDa tubulin

Gal4-DBD

Repression fold change P < 0.005 170 DBD D 0.4 YFP
14 130 DDBBDD––BKli81m6p-R1
12 % of input 0.3 WT
10 100 IB: 0.2 K816R
Blimp-1 0.1
8 55
6 kDa IB: 0
4 tubulin
2
0 Myc Pax5 Ciita Ciita 3′UTR

DBD DBD– DBD–
Blimp-1 K816R

Fig 4 | SUMO-modified Blimp-1 is a more potent transcriptional repressor. (A) Stability of Blimp-1 and K816R. 293T cells transfected with vectors
encoding WT or K816R Blimp-1 were subjected to CHX (100 mg/ml) chase experiment and lysed at the indicated time points for IB analysis. The
relative ratios of protein levels are quantified by normalization of the band intensity of Blimp-1 to that of tubulin for each time point, and the data are
presented as fold changes as compared with the sample at 0 h. Results are representative of two experiments. (B) Representation of the Gal4 DBD
construct. Relative luciferase activity of mouse 18–81 pre-B cells transfected with plasmids encoding indicated proteins together with (Gal4)4-tk-Luc.
(C) Relative luciferase activity of 18–81 cells transfected with plasmids encoding WT or K816R Blimp-1 and a luciferase reporter driven by the CIITA
promoter III. In B and C, total cell lysates were analysed by IB. Tubulin was used as a loading control. (D) A chromatin IP assay with anti–Blimp-1
shows that ectopically expressed WT and K816R Blimp-1 bound as effectively to the indicated endogenous loci in 18–81 cells. Regions encompassing
the Blimp-1 binding site in the indicated loci and a negative control region located in the 30 UTR of Ciita were amplified by QPCR. Input DNA was
used to normalize the data. Results are the mean±s.e.m. from triplicate samples of two independent experiments. (E) K816R is less capable of
interacting with HDAC2 than is WT Blimp-1. 18–81 cells electroporated as described in (D) were used for co-IP using anti-FLAG, followed by IB
analysis. The number indicates the quantitative ratios of band intensities obtained from IP versus its corresponding input. Blimp-1, B lymphocyte-
induced maturation protein-1; CHX, cycloheximide; DBD, DNA-binding domain; IB, immunoblot; IP, immunoprecipitation; tk, thymidine kinase;
UTR, untranslated region; WT, wild type.

In many cases, SUMO modification converts the function of Blimp-1, which may provide new strategies for regulating humoral
transcription factors from positive to negative regulators [21]. immune responses.
It appears that SUMO-modified Blimp-1 is a more potent
transcriptional repressor. In the case of K816R Blimp-1, its METHODS
reduced transcriptional repression may not result from globally
changing subcellular localization, as immunofluorescence Detailed experimental procedures are described in the
staining illustrated that the majority of WT and K816R Blimp-1 supplementary information online.
had a speckled nuclear distribution pattern and that WT and IP and IB analyses. IP and IB analyses were performed as
K816R Blimp-1 did not co-localize with promyelocytic leukaemia described [19]. Antibodies used in this study are described in
protein (supplementary Fig S7 online). It is more likely that the the supplementary information online.
level of HDAC2 that interacted with K816R Blimp-1 was Transfection and luciferase reporter assay. The luciferase repor-
decreased. Although a recent paper suggested that SUMOylation ter fused to the CIITA promoter III was described previously [7].
results in proteasomal degradation of Blimp-1 [10], we did not Renilla luciferase reporter driven by the tk promoter (RL-tk) was
observe that an absence of the SUMOylation site affected Blimp-1 used for normalization. Mouse 18–81 pre-B cells were transfected
expression. It is possible that the fusion of SUMO1 to the amino by electroporation; the amounts of DNA were chosen and the
terminus of Blimp-1, the approach used in that study, caused the procedure performed as described previously [5]. Cells were lysed
conflicting observation. 24 h after transfection and were used for measurement of firefly
luciferase and Renilla luciferase activities using the Dual-
In summary, our study here provides a mechanistic insight into Luciferase Reporters Assay kit (Promega). The luminescence was
the modulation of cell proliferation and plasma cell fate detected by TopCount NXTt (Perkin Elmer). Fold repression was
commitment through a SUMO site in the master regulator calculated as described [19].

6 3 6 EMBO reports VOL 13 | NO 7 | 2012 &2012 EUROPEAN MOLECULAR BIOLOGY ORGANIZATION

SUMO modulates the function of Blimp-1 scientific report
H.-Y. Ying et al

Generation of retrovirus or lentivirus and virus transduction. The 6. Lin Y, Wong K, Calame K (1997) Repression of c-myc transcription
procedure for preparing viral vectors was as described [19]. B cells by Blimp-1, an inducer of terminal B cell differentiation. Science 276:
were transduced with virus, at a multiplicity of infection of 5–10 596–599
for lentiviral transduction or 1–5 for retroviral transduction, in the
presence of 5 mg/ml of polybrene. Transduced B cells were 7. Piskurich JF, Lin KI, Lin Y, Wang Y, Ting JP, Calame K (2000) BLIMP-I
analysed or sorted by flow cytometry 2–3 days after infection. mediates extinction of major histocompatibility class II transactivator
Statistics. Statistical significance was determined by two-tailed expression in plasma cells. Nat Immunol 1: 526–532
Student’s t-tests. Po0.05 was considered significant. Data shown
here are the mean±s.d. from at least three experiments unless 8. Lin KI, Angelin-Duclos C, Kuo TC, Calame K (2002) Blimp-1-dependent
otherwise indicated. repression of Pax-5 is required for differentiation of B cells to
Supplementary information is available at EMBO reports online immunoglobulin M-secreting plasma cells. Mol Cell Biol 22:
(http://www.emboreports.org). 4771–4780

ACKNOWLEDGEMENTS 9. Shaffer AL et al (2002) Blimp-1 orchestrates plasma cell differentiation by
We thank Dr. W-H Lee for discussion and L-W Lo for technical assistance. extinguishing the mature B cell gene expression program. Immunity 17:
This work was supported by Academia Sinica AS-99-CDA-L12, by 51–62
NSC99-3111-B-001-005 from National Science Council and EX-9509NC
from National Health Research Institutes, Taiwan. 10. Shimshon L, Michaeli A, Hadar R, Nutt SL, David Y, Navon A,
Waisman A, Tirosh B (2011) SUMOylation of Blimp-1 promotes its
Author contributions: H.-Y.Y. performed the experiments, analysed the proteasomal degradation. FEBS Lett 585: 2405–2409
data and compiled the figures. S.-T.S., I.-Y.L. and Y.-H.T. performed the
experiments. P.-H.H., C.-C.C., M.-D.T. and H.-M.S. contributed critical 11. Seeler JS, Dejean A (2003) Nuclear and unclear functions of SUMO.
analytical tools. K.-I.L. designed the research and wrote the paper. Nat Rev Mol Cell Biol 4: 690–699

CONFLICT OF INTEREST 12. Kim KI, Baek SH, Chung CH (2002) Versatile protein tag, SUMO:
The authors declare that they have no conflict of interest. its enzymology and biological function. J Cell Physiol 191:
257–268
REFERENCES
1. Shapiro-Shelef M, Lin KI, McHeyzer-Williams LJ, Liao J, 13. Shuai K, Liu B (2005) Regulation of gene-activation pathways by PIAS
proteins in the immune system. Nat Rev Immunol 5: 593–605
McHeyzer-Williams MG, Calame K (2003) Blimp-1 is required for the
formation of immunoglobulin secreting plasma cells and pre-plasma 14. O’Shea JJ, Watford W (2004) A peek at PIAS. Nat Immunol 5:
memory B cells. Immunity 19: 607–620 875–876
2. Bikoff EK, Morgan MA, Robertson EJ (2009) An expanding job description
for Blimp-1/PRDM1. Curr Opin Genet Dev 19: 379–385 15. Rodriguez MS, Dargemont C, Hay RT (2001) SUMO-1 conjugation
3. Kallies A et al (2007) Initiation of plasma-cell differentiation is in vivo requires both a consensus modification motif and nuclear
independent of the transcription factor Blimp-1. Immunity 26: 555–566 targeting. J Biol Chem 276: 12654–12659
4. Kallies A, Hasbold J, Tarlinton DM, Dietrich W, Corcoran LM,
Hodgkin PD, Nutt SL (2004) Plasma cell ontogeny defined by 16. Xue Y, Zhou F, Fu C, Xu Y, Yao X (2006) SUMOsp: a web server for
quantitative changes in blimp-1 expression. J Exp Med 200: 967–977 sumoylation site prediction. Nucleic Acids Res 34: W254–W257
5. Yu J, Angelin-Duclos C, Greenwood J, Liao J, Calame K (2000)
Transcriptional repression by blimp-1 (PRDI-BF1) involves recruitment 17. Reimold AM, Iwakoshi NN, Manis J, Vallabhajosyula P, Szomolanyi-
of histone deacetylase. Mol Cell Biol 20: 2592–2603 Tsuda E, Gravallese EM, Friend D, Grusby MJ, Alt F, Glimcher LH (2001)
Plasma cell differentiation requires the transcription factor XBP-1. Nature
412: 300–307

18. Knodel M, Kuss AW, Lindemann D, Berberich I, Schimpl A (1999)
Reversal of Blimp-1-mediated apoptosis by A1, a member of the Bcl-2
family. Eur J Immunol 29: 2988–2998

19. Su ST, Ying HY, Chiu YK, Lin FR, Chen MY, Lin KI (2009) Involvement of
histone demethylase LSD1 in Blimp-1-mediated gene repression during
plasma cell differentiation. Mol Cell Biol 29: 1421–1431

20. Liu B, Mink S, Wong KA, Stein N, Getman C, Dempsey PW, Wu H,
Shuai K (2004) PIAS1 selectively inhibits interferon-inducible genes and
is important in innate immunity. Nat Immunol 5: 891–898

21. Gill G (2004) SUMO and ubiquitin in the nucleus: different functions,
similar mechanisms? Genes Dev 18: 2046–2059

&2012 EUROPEAN MOLECULAR BIOLOGY ORGANIZATION EMBO reports VOL 13 | NO 7 | 2012 6 3 7