molecularbasisforsumoylation-dependentregulationof ... · domains, hsf1 and hsf2 possess a...

Molecular Basis for SUMOylation-dependent Regulation ofDNA Binding Activity of Heat Shock Factor 2*□S

Received for publication, August 19, 2008, and in revised form, November 17, 2008 Published, JBC Papers in Press, November 18, 2008, DOI 10.1074/jbc.M806392200

Yukihiro Tateishi‡§, Mariko Ariyoshi‡1, Ryuji Igarashi‡, Hideyuki Hara¶, Kenji Mizuguchi�, Azusa Seto‡2,Akira Nakai**, Tetsuro Kokubo§, Hidehito Tochio‡, and Masahiro Shirakawa‡3

From the ‡Graduate School of Engineering, Kyoto University, Kyoto 615-8510, the §International Graduate School of Arts andSciences, Yokohama City University, Yokohama, Kanagawa 230-0045, ¶ESR Application, Bruker Biospin K.K., Ibaraki 305-0051,�National Institute of Biomedical Innovation, Ibaraki-City, Osaka 567-0085, and the **Department of Biochemistry and MolecularBiology, Yamaguchi University School of Medicine, Ube 766-8505, Japan

Heat shock factor 2 (HSF2) is a member of a vertebrate tran-scription factor family for genes of heat shock proteins and isinvolved in the regulation of development and cellular differen-tiation. TheDNAbinding property of HSF2 ismodulated by thepost-translational modification of a specific lysine residue in itsDNAbinding domain by small ubiquitin-likemodifier (SUMO),but the consequences of SUMOylation and its underlyingmolecular mechanism remain unclear. Here we show the inhib-itory effect of SUMOylation on the interaction between HSF2and DNA based on biochemical analysis using isolated recom-binant HSF2. NMR study of the SUMOylated DNA bindingdomain ofHSF2 indicates that the SUMOmoiety is flexible withrespect to the DNA binding domain and has neither a noncova-lent interface with nor a structural effect on the domain. Com-bined with data from double electron-electron resonance andparamagnetic NMR relaxation enhancement experiments,these results suggest that SUMO attachment negatively modu-lates the formation of the protein-DNA complex through a ran-domly distributed steric interference.

Theheat shock response is a cellular defense against environ-mental and physiological stresses, such as heat, heavy metals,oxygen radicals, and viral or bacterial infections. Eukaryoticcells maintain homeostasis through elevating the expression ofheat shock proteins (HSPs),4which function asmolecular chap-

erones, in response to various stress stimuli. The heat shockresponse is also evoked during nonstress-induced processes,such as early development and cellular differentiation in mam-mals (1). The HSP genes are up-regulated by heat shock ele-ments (HSEs) comprising multiple, and at least three, invertedrepeats of the pentanucleotide 5�-nGAAn-3�, where n denotesany nucleotide. The proteinswithin the heat shock factor (HSF)family, evolutionarily conserved from yeast to human, bind toHSEs and regulate the transcription level of a set of the HSPgenes. One of the mammalian members, HSF2, is involved indevelopment and differentiation-related processes, whereasHSF1 is responsive to classical stress stimuli (2, 3). Recentgenetic and biological studies have revealed that interplaysbetweenHSF1 andHSF2 gain functional versatility in the stressresponse and developmental signaling pathways. For example,HSF1 and HSF2 form heterocomplexes on the promoters ofchaperone genes, clusterin and hsp70, and are cooperativelyinvolved in the regulation of these genes (4, 5).The HSF members are characterized by a conserved domain

structure that consists of a winged helix-turn-helix motif nearthe N terminus followed by an extended hydrophobic heptadrepeat (HR-A/B). By itself, the helix-turn-helix DNA bindingdomain (DBD) of HSFs is capable of binding to at least twoinverted HSEs, regardless of the repeat directions, that is head-to-head (nGAAnnTTCn) and tail-to-tail (nTTCnnGAAn)repeats (6). However, it has been reported that Drosophila andyeast HSFs preferentially bind to the tail-to-tail elements ratherthan the head-to-head ones (7, 8). The HR-A/B domain facili-tates trimerization through the formation of a coiled-coil bun-dle upon receiving stress signals. The trimer formation enablesHSFs to bind the HSE repeats with high affinity, and hencetranscription of the downstream HSP genes is activated. Nev-ertheless, the isolated DBDs of Kluyveromyces lactis have beenshown to bind to HSEs in vitro (9); thus it is suggested thatisolated DBDs of HSFs retain a binding affinity for HSE repeatsand that the trimer formation of HSF1 and HSF2 has a regula-tive role in DNA binding. In addition to these functional

* This work was supported by grants (to M. S. and H. T.) from the Ministry ofEducation, Culture, Sports, Science and Technology of Japan and in part bythe Global COE Program “International Center for Integrated Research andAdvanced Education in Materials Science” the Ministry of Education, Cul-ture, Sports, Science and Technology of Japan, administrated by the JapanSociety for the Promotion of Science. The costs of publication of this articlewere defrayed in part by the payment of page charges. This article musttherefore be hereby marked “advertisement” in accordance with 18 U.S.C.Section 1734 solely to indicate this fact.

□S The on-line version of this article (available at http://www.jbc.org) containssupplemental Figs. 1– 4.

1 To whom correspondence may be addressed: Graduate School of Engineer-ing, Kyoto University, Nishikyo-ku, Katsura, Kyoto 615-8510, Japan. Fax:81-75-383-2541; E-mail: [email protected].

2 Present address: Dept. of Basic Medical Sciences, Institute of Medical Sci-ence, the University of Tokyo, Tokyo 108-8639, Japan.

3 To whom correspondence may be addressed: Graduate School of Engineer-ing, Kyoto University, Nishikyo-ku, Katsura, Kyoto 615-8510, Japan. Fax:81-75-383-2541; E-mail: [email protected].

4 The abbreviations used are: HSP, heat shock protein; DBD, DNA bindingdomain; SUMO, small ubiquitin-like modifier; HSF, heat shock factor; HSE,heat shock element; EMSA, electrophoretic mobility shift assay; GST, glu-

tathione S-transferase; hHSF2FL, full-length human HSF2�; TDG, thymineDNA glycosylase; SPR, surface plasmon resonance; PDB, Protein Data Bank;MTSL, (1-oxyl-2,2,5,5-tetramethyl-D3-pyrroline-3-methyl) methane thio-sulfonate; SDSL, site-directed spin labeling; HSQC, heteronuclear singlequantum coherence; SIM, SUMO interaction motif; HR, hydrophobicrepeat; DEER, double electron-electron resonance; PRE, paramagneticNMR relaxation enhancement; HtH, head-to-head; TtT, tail-to-tail.

THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 284, NO. 4, pp. 2435–2447, January 23, 2009© 2009 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in the U.S.A.

JANUARY 23, 2009 • VOLUME 284 • NUMBER 4 JOURNAL OF BIOLOGICAL CHEMISTRY 2435

by guest on July 28, 2020http://w

ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

domains, HSF1 and HSF2 possess a C-terminal hydrophobicrepeat (HR-C) that suppresses the trimer formation to main-tain an inactive state in the absence of stimuli (10).Functional properties of mammalian HSFs were found to

be modulated by post-translational modification with smallubiquitin-like modifier (SUMO). Modification with SUMO(SUMOylation) has been recognized as an important regula-tory mechanism in diverse cellular processes such as nucleartrafficking, transcription, and DNA repair (11). SUMO sharesstructural similarity with ubiquitin and can be covalentlyattached to lysine residues within a �KXE consensus sequence(� indicates hydrophobic residue and X is any amino acid) intarget proteins via an isopeptide bond formed through an en-zymatic pathway that is similar to but distinct from that inubiquitination (11–13). Several SUMO paralogues, termedSUMO-1, SUMO-2/3, and SUMO-4, have been identified inmammals to date (14). The majority of SUMOylation sub-strates identified thus far are nuclear factors, including tran-scriptional regulators (11, 15).With some transcription factors,the conjugation of SUMO introduces a new protein-proteininteraction surface without changing the DNA binding proper-ties, leading to the recruitment of distinct multiprotein com-plexes required for transcriptional activation or repression (16,17). In some other cases, SUMOylation regulates gene expres-sion levels by altering the subnuclear localization of transcrip-tional regulators (18, 19). The molecular basis underlying theregulation of protein-DNA interactions by SUMO remains tobe elucidated, but one clear example has been provided by bio-chemical and structural studies on a DNA repair enzyme, thy-mine DNA glycosylase (TDG) (20, 21). The modification withSUMO-1 is suggested to induce a conformational change inTDG that is associated with noncovalent contacts between thecovalently linked SUMO-1 andTDGmoieties, and ultimately itleads to a loss of DNA binding.The HSF isoforms exhibit different SUMOylation patterns

that might provide complex physiological effects, whereas Lys-298 in HSF1 is modified by either SUMO-1 or SUMO-2 (22),Lys-82 and Lys-139 in HSF2 are modified by SUMO-1 andSUMO-2/3, respectively (23, 24).Of these sites inHSF2, Lys-82,which is located in a loop region (“wing”) of the DBD, is theprimary SUMOylation site. Lys-139 in the trimerizationdomain has been observed to be much less efficiently sumoy-lated than Lys-82 (23). The importance of SUMOylation atLys-82 of HSF2 has been emphasized by its involvement inbookmarking of the hsp70i gene in mitosis (25). However,the functional consequences of the SUMOylation are stillunclear because the SUMO modification of HSF2 at Lys-82has been reported to both enhance and inhibit its DNA bind-ing (23, 24, 26).To clarify the effect of SUMOylation on theDNA interaction

activity of HSF2, we examined the biochemical properties ofisolated recombinant human HSF2 conjugated with SUMO-1.Furthermore, to understand the detailedmolecularmechanismby which SUMO-1 conjugation alters the DNA binding modeof HSF2, we performed structural analysis on the SUMO-1-conjugated HSF2 DBD using NMR and ESR spectroscopictechniques. Our biochemical and structural data indicate thatSUMOylation of HSF2 interferes with the binding to HSEs

through amolecularmechanismdistinct from that proposed bystructural studies of the SUMOylated TDG. The covalentattachment of SUMO does not cause any notable conforma-tional change in the HSF2 DBD or form a stable noncovalentinteractionwithHSF2.Herewe propose a novelmodel inwhichthe randomly distributed SUMOmoiety tethered to the flexibleloop in HSF2 DBD restricts accessibility of HSF2 to DNAthrough a steric interference in a probabilistic manner, result-ing in less chance for the formation of the encounter complex.

EXPERIMENTAL PROCEDURES

Protein Expression and Purification—The cDNA regionencoding DBD of human HSF2� (hHSF2DBD) was amplifiedby PCR and cloned into the bacterial expression vectorpGEX4T-3 (GE Healthcare) with an N-terminal GST tag. ThehHSF2DBD expression vector was transformed into Esche-richia coli BL21(DE3) cells. Cells were grown in Luria-Bertani(LB)medium at 37 °C to an optical density of 0.5 at 660 nm, andthen induced with 1 mM isopropyl �-D-thiogalactopyranosidefor 6 h at 30 °C. For preparation of 15N-SUMO-1 and 15N-hHSF2DBD,M9minimal media containing 0.5 g/liter 15NH4Clwas used instead of LB media. Cells were lysed by sonication in10mMpotassiumphosphate buffer (pH 7.5) containing 500mMKCl, 1 mM phenylmethylsulfonyl fluoride, and 5 mM EDTA.The clarified lysate was loaded onto glutathione-Sepharose 4Fast Flow beads (GE Healthcare) after the debris was removedby centrifugation. hHSF2DBD was eluted from the beads bycleaving off the GST tag with 60 units/g thrombin protease (GEHealthcare) for 3 h at room temperature. The eluted proteinwas further purified using ion-exchange and affinity columnchromatography with HiTrap SP HP and Hitrap heparin HPcolumns (GE Healthcare), respectively. hHSF2DBD was puri-fied with more than 95% homogeneity.A K139R point mutation was introduced into full-length

human HSF2� (hHSF2FL) by PCR using two complementaryprimers containing mismatched nucleotides at the codon cor-responding to Lys-139. ThePCRmixturewas treatedwithDpnIto remove the template plasmid, followed by direct transforma-tion into E. coli DH5� cells. hHSF2FLK139R was expressed as aGST fusion protein in BL21(DE3)-Codon Plus-RIL cells (Nova-gen) andwas purified byGST affinity column chromatography.The GST tag was cleaved off with PreScission protease (GEHealthcare) on the glutathione beads, resulting in elution ofhHSF2FLK139R from the beads. Further purificationwas carriedout using cation-exchange column chromatography with aHitrap heparin HP column (GE Healthcare) and size-exclusioncolumn chromatographywith aHiLoad 16/60 Superdex 200-pgcolumn (GE Healthcare).Preparation of SUMOylated Protein—For preparation of

SUMOylated hHSF2DBD, a pGEX4T-3 vector encodinghHSF2DBD and a pTS1 vector encoding SUMO-1 andSUMOylation enzymes, Aos1/Uba2 and Ubc9, were co-trans-formed into BL21(DE3) cells (27). The transformed cells werecultivated in LB medium at 37 °C to an optical density of 0.5 at660 nm, and then were induced with 1 mM isopropyl �-D-thio-galactopyranoside at 25 °C for 18 h. About 60% of the expressedhHSF2DBD was SUMOylated. SUMO-hHSF2-DBD was puri-fied in the same manner as the SUMO-free DBD. Unmodified

DNA Binding Inhibition of HSF2 by SUMOylation

2436 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 284 • NUMBER 4 • JANUARY 23, 2009


ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

hHSF2DBD was completely removed during the purificationprocess. 15N-SUMO-1-conjugated nonlabeled hHSF2DBD,nonlabeled SUMO-1-conjugated 15N-hHSF2DBD, andSUMOylated hHSF2FLK139R were prepared using an in vitroSUMOylation method. 15 �M SUMO-1 and 7.5 �M GST-hHSF2 were mixed in 30–50 ml of a reaction mixture contain-ing partially purified His-Aos1/Uba2, 3 �M His-Ubc9, 4 mMMgCl2, 0.05% Triton X-100, and 5 mM ATP. The reaction mix-tures were incubated at room temperature for 4 h. TheseSUMOylated proteins were purified in the same manner asdescribed above.DNA Interaction Assay (EMSA)—Synthesized oligonucleo-

tides were used in EMSA experiments as follows: HtH16, 5�-GGTAGAATATTCTTGG-3�; TtT16, 5�-GGTATTCTAGAA-TTGG-3�; HtH26, 5�-GTTATGGTAGAATATTCTTGGTA-TTG-3�; and TtT26, 5�-GTTATGGTATTCTAGAATTGGT-ATTG-3�. Each duplex oligonucleotide was mixed with theunmodified or SUMOylated protein in a binding reaction mix-ture containing 10mMHEPES-KOH (pH7.5), 50mMKCl, 1mMEDTA, 1 mM dithiothreitol, and 5% glycerol. The reactionmix-ture was incubated at room temperature for 15 min and thenwas subjected to electrophoresis on 6% native polyacrylamidegel in 0.5� TBE. Electrophoresis was performed in 0.5� TBEbuffer at 200 V, 10mA for 30min at 4 °C. The bands containingDNA were visualized by staining with GelredTM (Biotinum)and UV illumination.SPR Measurement—Dissociation constants of hHSF2 for

HSE oligonucleotides were measured by SPR using a BIAcore2000 instrument (GE Healthcare). 5�-Biotinylated HtH andTtT HSE oligonucleotides were each immobilized on astreptavidin-coated sensor chip SA, 200 response units forexperiments with hHSF2DBD, and 100 response units forthose with hHSF2FLK139R. Analytes were diluted with thebinding buffer containing 10 mM HEPES-KOH (pH 7.5), 50mM KCl, 1 mM EDTA and 0.005% Tween 20, and were inject-ed/applied at a flow rate of 20 �l/min (hHSF2DBD) or 30�l/min (hHSF2FLK139R). The sensor chip immobilizing theoligonucleotides was regenerated between runs by flowing 50mM NaOH, 0.5 M NaCl, and 0.001% SDS over the chip surface.SPR sensorgrams were processed and analyzed using the BIA-evaluation 3.0 software package (GE Healthcare) and assumingthe binding occurred at a protein:DNA molar ratio of 2:1.NMR Spectroscopy—All NMR spectra of hHSF2DBD and

SUMOylated hHSF2DBD were acquired with an Avance 700spectrometer equipped with a CryoProbe (Bruker Biospin) at298 K in 20 mM potassium phosphate buffer (pH 6.5), contain-ing 200mMKCl, 5%D2O. Sequential backbone assignments forhHSF2DBD were carried out by HNCA, HNCACB, CBCA-(CO)NH,HN(CO)CA, andHN(CO)CACBexperiments using a15N-13C uniformly labeled sample. NMR data were analyzedusing NMRPipe (28) and SPARKY 3.1 (29). Backbone assign-ments were evaluated using a reverse isotopic labeling tech-nique; the signals derived from Arg and Lys residues were con-firmed by adding natural abundant L-Arg or L-Lys at 100mg/liter in M9 minimal media (30).MTSL Labeling—Cysteine mutants of SUMO-1, C52A/

S50C, C52A/Q53C, C52A/V87C, and C52A/E89C were gener-ated by site-directed mutagenesis as described above, and were

expressed and purified as reported previously (27). E19C andK51C mutants of hHSF2DBD were also generated by site-di-rectedmutagenesis and prepared as described above. ForDEERmeasurements, both hHSF2DBDE19C and hHSF2DBDK51Cwere SUMOylated in vitro with SUMO-1C52A/Q53C and puri-fied as mentioned above. For PRE measurements, 15N-hHSF2DBD was conjugated with each SUMO-1 cysteinemutant using the in vitro SUMOylation system, and was subse-quently purified using anion-exchange chromatography.MTSL, dissolved in DMSO, was added to each SUMOylatedhHSF2DBD mutant in a 4 M excess of cysteine residues, andthen the reaction mixture was incubated at room temperaturefor 12 h. The MTSL-labeled protein was further purified byaffinity chromatography using a Heparin-Sepharose 6 FastFlow column (GE Healthcare).DEER Spectra Measurement—DEER spectra were measured

on 150 �l of each MTSL-labeled protein (100 �M), which wereflash-frozen in liquid nitrogen, as described previously (31, 32).Pulsed DEER data were acquired at 80 K on an ELEXSYS E580X-band FT/CW spectrometer (Bruker Biospin) equipped witha dielectric resonator (ER4118X-MD5-W1) and helium gasflow system (CF935, Oxford Instruments). The obtained DEERspectra were analyzed by using the DEER Fit and DEER Trafoprograms (available on line). For DEERmeasurement, a 4-pulseconstant time DEER sequence was employed (31, 32). Thepump pulse was set to themaximum of the nitroxide ESR spec-trum (�B � 9.58 GHz). The observer pulse was set to 60 MHzhigher (�A � 9.64 GHz), which corresponds to about a 20-Gfield separation. The �/2 pulse width was 16 ns. The � pulsesand the pump pulse were 32 ns.Comparative Modeling—The amino acid sequences of

human and Drosophila HSF DBD were aligned using FUGUE(33). A series of models of hHSF2DBD was generated withMODELLER (34) using each of the 28 NMR-derived models ofDrosophila HSF DBD (35) deposited in the Protein Data Bank(PDB entry 1HKT) as a template. Ten of these hHSF2DBDmodelswere superimposed onto one of the subunits in the crys-tal structure of the K. lactis HSF DBD dimer bound to the TtTrepeat (chain B of PDB entry 3HTS) using the programs SSM(36) and MOLMOL (37). The superimposed structures weredisplayed using PyMOL (38).Calculation of Structural Model—The modeled structure of

the SUMO-1-conjugated hHSF2DBD was calculated withHADDOCK 2.0 (39). An ensemble of the 10 structures of thehHSF2DBD comparative models and 1 SUMO-1 structuretaken from an NMR ensemble (PDB entry 1A5R) were used asthe starting structures. TheC-terminal region of SUMO-1 (res-idues 94–97) and the flexible loop of hHSF2DBD (residues75–87) were defined as “fully flexible.” Two distance con-straints derived from themost populated distances obtained byDEER experiments were used in the calculation: 36–46 and35–45Å for the distances betweenGln-53C-� of SUMO-1 andGlu-19 C-� of hHSF2DBD, and between Gln-53 C-� ofSUMO-1 and Lys-51 C-� of hHSF2DBD, respectively. Lowerlimits were set based on the fact that difference in distancesbetween two C-� atoms (R��) of spin-labeled cysteines andbetween the two nitroxide moieties (RNONO) of a doublyMTSL-labeled protein has been shown to range from 0 to 10 Å




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

HtH TtT

Ladd

erFree SUMO-1 +-

+ +- +

free

bound

-- - -

--

1 2 3 4 5 6

hHSF2-DBD

1

FreeSUMO-1 La

dder

+-+ +

- +

HtH TtT

-- - -

--

+s +s- -

free

bound

8765432

hHSF2βK139R

- 50

0

50

100

150

200

250

300

350

- 200 0 200 400 600 800

Res

pons

eU

n its

Time (s)

- 100

- 50

0

50

100

150

200

250

300

350

400

- 200 0 200 400 600 800

Res

pons

eU

nits

Time (s)

HtH2626 with hHSF2DBD TtT26 with hHSF2DBD

- 50

0

50

100

150

200

250

300

350

- 200 0 200 400 600 800 1000

Res

pons

eU

nits

Time (s)

HtH26 with hHSF2FL HtH26 with SUMO-hHSF2FL

0

50

100

150

200

250

300

0 2 4 6 8 10 12

Res

pons

eU

nits

Conc (µM)

0

50

100

150

200

250

0 2 4 6 8 10 12

Res

pons

eU

nits

Conc (µM)

0

20

40

60

80

100

120

140

160

180

200

0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09

Res

pons

eU

nits

0

50

100

150

200

250

300

350

0.00 0.20 0.40 0.60 0.80 1.00 1.20

Res

pons

eU

nits

- 50

0

50

100

150

200

250

- 200 0 200 400 600 800 1000

Res

pons

eU

nits

Time (s)

0.8 µM0.6 µM0.4 µM0.2 µM0.1 µM

0.8 µM0.6 µM0.4 µM0.2 µM

1.0 µM

10 µM8 µM6 µM4 µM2 µM

10 µM8 µM6 µM4 µM2 µM

9.25 ± 0.54 µM20.96 ± 2.12 µM

12.15 ± 0.27 nM 192.53 ± 2.89 nM

HtH26 TtT26 5’- GTTATGGTAGAATATTCTTGGTATTG-3’ 5’- GTTATGGTATTCTAGAATTGGTATTG-3’

FIGURE 1. SUMOylation inhibits the binding of hHSF2 to HSEs. A, DNA sequences of 26-mer double-stranded oligonucleotides containing two inverted HSErepeats, HtH26 and TtT26. HSE consensus sequences in each oligonucleotides are shown in red. B, EMSA performed with hHSF2DBD and either HtH26 (lanes 1–3) or TtT26(lanes 4 – 6). The 1:2 complex mixtures of oligonucleotides and either non-SUMOylated (lanes 2 and 5) or SUMOylated hHSF2DBD (lanes 3 and 6) were resolved on a 6%native polyacrylamide gel. C, SPR analysis of binding of the hHSF2DBD to two inverted HSE repeats. Upper panels, overlay plots of the sensorgrams obtained for 2–10�M of hHSF2DBD and the immobilized HtH26 (left) or TtT26 (right) oligonucleotides. Lower panels, plots of the amplitude of the hHSF2DBD binding as a function of theprotein concentration. D, EMSA performed with hHSF2FLK139R and HtH26 (lanes 1– 4) or TtT26 (lanes 5– 8). The 1:2 complex mixtures of oligonucleotides and eithernon-SUMOylated (lanes 2 and 6) or SUMOylated hHSF2K139R (lanes 3 and 7) were resolved on a 6% native polyacrylamide gel. The complex mixtures treated withSUMO-specific SENP protease were loaded on lanes 4 and 8. E, SPR analysis of binding of unmodified or SUMOylated hHSF2FLK139R to two inverted HSE repeats, HtH26.Upper panels, SPR sensorgrams obtained with immobilized HtH26 oligonucleotides by increasing the concentration of unmodified (left) or SUMOylated (right)hHSF2FLK139R from 0.1 to 1.0 �M. Lower panels, plots of the amplitude of hHSF2FL K139R binding as a function of the protein concentration.




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

(40). To mimic the isopeptide bond between hHSF2DBD andSUMO, a 2.0 Å distance constraint was introduced between thecarboxyl carbon of Gly-97 of SUMO-1 and the N-� nitrogen ofLys-82 of the hHSFDBDwith a force constant of 500 kcalmol�1

Å�2. A total of 1000 structures was initially generated. The top200 structures were subjected to simulated annealing calcula-tions, and the 20 with the lowest intermolecular energy wereanalyzed and presented.

RESULTS

SUMO-1ModificationAbrogates theDNABindingActivity ofhHSF2 DBD—Combined with biochemical data, the crystalstructure of the K. lactis HSF DBD in complex with DNA pre-viously demonstrated a stable dimer formation of its DBDuponbinding to two inverted HSEs arranged in the tail-to-tail orien-tation (9). Each of the HSE repeats is recognized by an individ-ual DBD of K. lactis HSF, and the specific protein-DNA inter-action is stabilized by the intersubunit contacts of the proteindimer. In agreement with this previous observation, our resultsfrom the EMSA indicate that the isolatedDBD of humanHSF2,including amino acid residues 6–110 (hereafter designated ashHSF2DBD), binds to two or three inverted HSE consensusrepeats (Fig. 1,A andB) but not to a single one (data not shown).An analysis using size-exclusion column chromatographyshowed that hHSF2DBD exists as a monomer without DNA insolution (data not shown), supporting the dimerization of DBDupon specific binding to the repeated HSEs.To examine whether hHSF2DBD exhibits a preference for

the direction of the HSE repeats, we performed EMSA using a26-bp-long oligonucleotide containing two inverted repeats ineither the HtH or TtT direction (Fig. 1A). The HSF2DBD-HtH26 and the HSF2DBD-TtT26 complexes each showed aclear band shift (Fig. 1B, lanes 2 and 5). The affinity ofhHSF2DBD for each HSE oligonucleotide was quantified usingSPR. Assuming the binding of protein:DNA in a 2:1molar ratio,the dissociation constants (Kd) for theHtH26 and TtT26 repeatswere estimated to be 9.25 � 0.54 and 20.96 � 2.12 �M, respec-tively (Fig. 1C and Table 1). Thus hHSF2DBD binds to twoinverted HSE arrays without strong preference for either theHtH or TtT arrangement. Next, EMSA was performed withrecombinant SUMOylated hHSF2DBD to examine the influ-ence of SUMO-1 conjugation at Lys-82 on the DNA bindingactivity. SUMO-1-conjugated hHSF2DBD (SUMO-hHSF2DBD)was produced in E. coli using a binary vector system in whichSUMO-1 and an acceptor protein are simultaneously expressed

together with SUMOylation enzymes, SUMO-E1 andSUMO-E2 (27) (supplemental Fig. 1A). In the EMSA using theoligonucleotides described above, interaction of SUMO-hHSF2DBDwith neitherHtH26 nor TtT26 repeats was detected(Fig. 1B, lanes 3 and 6). An accurate dissociation constant ofSUMO-HSF2DBD for both the HtH26 and TtT26 DNA couldnot be determined by SPR measurements, because the SPRresponseswereweak and did not reach a state of equilibrium. Inaddition, SUMO-hHSF2DBD was not able to bind to a longerDNA probe containing three-inverted HSE arrays (data notshown). These data demonstrate that SUMO-1 conjugationinhibits the DNA binding activity of hHSF2DBD regardless ofthe direction of the HSE repeats. These DNA binding assayswere performed in a buffer containing 50 mM KCl. The sameeffect of SUMOylation on the DNA binding affinity ofhHSF2DBDwas also observed in EMSA performed in the pres-ence of 200 mM KCl (supplemental Fig. 1C).SUMO Modification Decreases the DNA Binding Activity of

Full-length hHSF2—The effect of SUMO modification on theDNA binding activity of hHSF2FL was also investigated usingEMSA and SPRmeasurements with recombinant SUMOylatedhHSF2FL (SUMO-hHSF2FL). The SUMO-hHSF2FL samplewas prepared by an in vitro SUMOylation reaction (supplemen-tal Fig. 1B). To assess the consequence of the SUMO conjuga-tion to Lys-82 solely, another possible SUMOylation site, Lys-139, lying in the HR-A/B domain, was mutated to arginine.In the EMSA experiment shown in Fig. 1D, band shifts of the

complex with unmodified hHSF2FLK139R were observed forboth HtH26 and TtT26 oligonucleotides (lanes 2 and 6). On theother hand, SUMO-hHSF2FLK139R bound less efficiently to theHSE repeats than unmodified hHSF2FLK139R, but full DNAbinding activity was recovered by releasing the SUMO moietywith SUMO-specific protease (SENP) treatment (Fig. 1D, lanes3 and 4 and 7 and 8). Based on SPRmeasurements, theKd valuesof hHSF2FLK139R for HtH26 and TtT26 repeats were estimatedto be 12.15 � 0.27 and 19.67 � 0.74 nM, respectively, whereasthe Kd values of SUMO-hSHF2FLK139R were estimated to be192.53 � 2.89 nM for HtH26 and 355.98 � 17.01 nM for TtT26(Fig. 1E and Table 1). The modification with SUMO-1 was alsoshown to decrease the affinity of hHSF2FL for five repeatedHSEs existing in the naturalhsp70proximal promoter region byan EMSA experiment (data not shown). Taken together, theseobservations indicate that hHSF2FL binds to the two invertedHSEs in both theHtH andTtT directionswith similar affinities,and that its DNA binding affinity is reduced by more than10-fold by SUMOylation at Lys-82, regardless of theHSE repeatdirection.hHSFDBD Does Not Form a Noncovalent Interface with the

Covalently Attached SUMO-1—We next addressed whetherthere were stable noncovalent interactions between SUMO-1and hHSF2DBD when they are covalently linked through anisopeptide bond. The molecular interaction and conforma-tional changes induced by the covalent attachment were ana-lyzed using two-dimensional 1H-15N correlation spectra.SUMO-hHSF2DBD was prepared by an in vitro SUMOylationreaction, in which either the SUMO-1 or the hHSFDBDmoietywas uniformly labeled with 15N to selectively observe NMR sig-nals from the labeled moiety. Chemical shift perturbation

TABLE 1Dissociation constants between HSE DNA and hHSF2 or SUMOylatedhHSF2 as determined by SPR measurements

hHSF2DBD-(6–110) SUMO1-hHSF2DBDHtH 26 mer 9.25 � 0.54 �M NDa

TtT 26 mer 20.96 � 2.12 �M NDa

HtH 16 mer 14.37 � 0.26 �M 112.89 � 2.99 �MTtT 16 mer 60.91 � 8.22 �M NDa

hHSF2FLK139R SUMO1-hHSF2FLK139RHtH 26 mer 12.15 � 0.27 nM 192.53 � 2.89 nMTtT 26 mer 19.67 � 0.74 nM 355.98 � 17.01 nMHtH 16 mer 29.99 � 2.36 nM 357.55 � 47.91 nMTtT 16 mer 124.34 � 7.55 nM 3.16 � 0.17 �M

a An accurate dissociation constant could not be determined, due to small SPRresponses.




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/cgi/content/full/M806392200/DC1




http://www.jbc.org/

experiments were carried out in a buffer containing 200 mMKCl. The two-dimensional 1H-15N correlation spectrum of15N-SUMO-1 conjugated to nonlabeled hHSF2DBD overlapswell with that of unconjugated 15N-SUMO-1 (Fig. 2A). A sub-stantial chemical shift difference was observed only for the res-

onance attributed to the main chainamide group of Gly-97 at the C-ter-minal end of SUMO-1. As this resi-due is involved in the formation ofthe isopeptide bond with the targetLys residue of hHSF2DBD, thechemical shift change is most likelycaused by a change in the electro-static properties of the terminal car-boxylic acid group or in the localconformation of Gly-97 associatedwith the isopeptide bond formation.The absence of chemical shift differ-ences for all other signals suggeststhat the structure of SUMO-1 isunchanged upon the covalent link-ing to hHSF2DBD, and thatSUMO-1 has no stable noncovalentcontacts with hHSF2DBD exceptfor a region near the conjugating Cterminus. To complement thisexperiment, the two-dimensional1H-15N correlation spectrum of15N-hHSF2DBD conjugated withnonlabeled SUMO-1 was comparedwith that of unmodified 15N-hHSF2DBD. The chemical shiftchanges because of the formation ofthe covalent link to SUMO-1 arelimited to the conjugation site, Lys-82, and its surrounding residues,Ile-76, Ser-78, Val-81, and Arg-85,of hHSF2DBD. These residues dis-play chemical shift differences lessthan 0.1 and 1.0 ppm in the 1H and15N dimensions, respectively (Fig.2B). These results suggest thathHSF2DBD does not have a welldefined noncovalent interface withthe attached SUMO-1 and excludethe possibility that a major confor-mational change in the DBD is asso-ciated with the SUMO conjugation.Interference between Conjugated

SUMO-1 and the DNA RegionsFlanking HSEs—To examine themechanism of the negative effect ofSUMOylation on the DNA bindingactivity of hHSF2, we modeled thestructure of hHSF2DBD in complexwith HSE repeats based on a crystalstructure of K. lactis HSF DBD in acomplex with DNA (9) and NMR

structures ofDrosophilaHSFDBD (35) (see “Experimental Pro-cedures”). Except for a loop region containing the SUMOyla-tion site Lys-82, hHSF2 DBD shows significant sequence simi-larities to those of yeast and Drosophila homologues, whereasthe sequences of the loop region are less conserved (supple-

hHSF2

Blue Red

15N-SUMO-1

15N-SUMO-1

SUMO-1

15N-hHSF215N-hHSF2

Blue Red

A

B

FIGURE 2. No stable interaction between SUMO-1 and hHSF2DBD in the conjugated complex. A, compar-ison of SUMO-1 and SUMO-1 conjugated to hHSF2DBD. Overlay of 1H-15N HSQC spectra of 15N-labeled SUMO-1(blue) and 15N-labeled SUMO-1 conjugated to hHSF2DBD (red). The SUMO-1 residues that exhibited a signifi-cant chemical shift difference are indicated in the spectra. B, comparison of hHSF2DBD and SUMOylatedhHSF2DBD. Overlay of 1H-15N HSQC-spectra of 15N-labeled hHSF2DBD (blue) and SUMOylated 15N-labeledhHSF2DBD (red) is shown. The hHSF2 residues that showed a significant chemical shift difference are indicatedin the spectra.




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

mental Fig. 2). In an ensemble of NMR structures ofDrosophilaDBD, the loop region is less converged than other regions (35).The corresponding loop in yeastHSFDBD is unstructured evenin a DNA-bound form (9). These observations imply that thisloop of hHSF2 DBD might be unstructured, and thus theSUMOylation site Lys-82 appears not to be fixed in a specificarrangement in our comparative model (Fig. 3A). In the TtTarrangement, this loop region is located in the vicinity of theprotein dimer interface (Fig. 3B, lower panel), raising the pos-sibility that SUMO-1 conjugation could perturb the dimer for-mation required for stable DNA binding. However, in the HtHarrangement, the two SUMOylation sites are assumed to be

located outside the DNA-bounddimer and to be distant from theputative subunit interface (Fig. 3B,upper panel). As SUMOylationequally decreases the binding affin-ity of HSF2 to HSEs in the HtHand TtT arrangements, we hypoth-esized that SUMO-1 conjugationinhibits the DNA binding activity ofhHSF2 through a mechanism otherthan the perturbation of the proteindimer interface.Therefore, we addressed whether

interference between the flexibleSUMOmoiety and DNA could per-turb the interaction between hHSF2and HSEs. This possibility wasexamined by EMSA using 16-meroligonucleotides that contained thetwo inverted HSE repeats with only3-bp flanking sequences at bothends, which are supposed to be justlong enough to be covered with theHSF2DBD dimer in the model.Unmodified hHSF2DBDwas able tobind to the 16-mer DNA in both theTtT (TtT16) and HtH (HtH16)arrangements. However, SUMO-hHSF2DBD was not able to bind toTtT16 but could form a complexwith HtH16 (Fig. 3C). Thus, thepresence of the 5-bp DNA regionsflanking the HSE repeats, which arenot supposed to make direct con-tact with hHSF2DBD, is likely tocause the negative effect ofSUMO-1 conjugation on the pro-tein-DNA interaction in the HtHarrangement. On the basis of theseobservations, we deduced thatabrogation of the hHSF2-DNAinteraction by SUMOylationcould be attributed to steric orelectrostatic interference betweenthe SUMO moiety and DNA. Thebinding of SUMO-hHSF2DBD to

HtH16 was also observed in EMSA in a binding buffer con-taining 200 mM KCl (supplemental Fig. 1C).SUMO-hHSF2FLK139R also interacts with HtH16 but not

with TtT16 (Fig. 3D). SPR measurements showed that theSUMOylation of hHSF2 caused smaller reduction in bindingaffinity for the HtH16 repeats than for the TtT16 repeats(Table 1).Spatial Distribution of the Attached SUMO-1 Relative to

hHSF2DBD—To estimate the extent of the spatial distributionof the attached SUMO-1 relative to hHSF2DBD, we carried outa DEER experiment combined with site-directed spin labeling(SDSL) using MTSL. The distance can be measured between

FIGURE 3. Comparative model of hHSF2DBD. A, comparative model of the hHSF2DBD dimer bound to DNA.Model structures derived from the NMR structures of Drosophila HSF-DBD are overlaid. The side chains of theSUMOylation site, Lys-82, are shown in red. B, models of hHSF2DBD bound to 26-mer oligonucleotides con-taining HtH or TtT two inverted HSE repeats. DNA sequences of HtH and TtT oligonucleotides are shown belowthe models. The positions of HSEs are indicated by arrows. C, EMSA performed with hHSF2DBD and 16-mer twoinverted HSE repeats, HtH16 (lanes 1–3) and TtT16 (lanes 4 – 6). The 1:2 complex mixtures of oligonucleotides andeither non-SUMOylated or SUMOylated hHSF2DBD were resolved on a 6% native polyacrylamide gel. The DNAbinding assay was performed in a buffer containing 50 mM KCl. D, EMSA performed with hHSF2FLK139R and16-mer two inverted HSE repeats. The 1:2 complex mixtures of oligonucleotides and either non-SUMOylated(lanes 2 and 6) or SUMOylated hHSF2K139R (lanes 3 and 7) were resolved on a native polyacrylamide gel. Thecomplex mixtures treated with SUMO-specific SENP protease were loaded on lanes 4 and 8.




ww

.jbc.org/D

ownloaded from



http://www.jbc.org/




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

two nitroxide spins that are each conjugated to a cysteine resi-due in a target protein. For attachment of MTSL to SUMO-1,the native Cys-52 of SUMO-1 was mutated to alanine, becausethe global fold may be perturbed by the MTSL label at this sidechain that points to the interior of the protein molecule. Addi-tionally we replaced a solvent-exposed residue Gln-53 witha cysteine (SUMO-1C52A/Q53C); the cysteine mutants ofhHSF2DBD, E19C and K51C, were each conjugated to SUMO-1C52A/Q53C. MTSL labeling of these conjugated proteins intro-duced a pair of spin labels at Q53C in the SUMO-1 moietytogether with either E19C or K51C in the hHSF2 moiety ofSUMO-hHSF2DBD (Fig. 4, A and B).The distributions of interspin distances were obtained from

DEER spectra of the spin-labeled SUMO-hHSF2DBD. The dis-tribution of SUMO-1C52A/Q53C-conjugated hHSF2DBDE19Cgave a peak with the maximum at 46.0 Å and a width of 21.3 Å(Fig. 4C, left panel), and SUMO-1C52A/Q53C-conjugatedhHSF2DBDK51C gave a peak with themaximum at 45.0 Å anda width of 31.0 Å (Fig. 4C, right panel). The peak widths of21.3 and 31.0 Å are more than five times larger than the peakwidth (4.0 Å) obtained for spins of MTSLs attached to resi-dues at positions 20 and 35 of a ubiquitin derivative (31). Theflexibility of each of the linkers between the protein back-bone and the spin labels can increase the distribution widthby �3 Å (40); therefore, if the positions and orientation ofC-� atoms are fixed, the distribution width of interspin dis-tance is no more than �6 Å, which originated from the flex-ibility of the pair of linkers. The much larger peak widthsobserved for SUMO-1C52A/Q53C-conjugated hHSF2DBDE19Cand SUMO-1C52A/Q53C-conjugated hHSF2DBDK51C suggestthat the relative position of SUMO-1 to hHSF2DBD exhibitsa large range of distribution.We next investigated the preference of the relative orienta-

tion of SUMO-1 in the conjugation complex with hHSF2DBDusing paramagnetic NMR relaxation enhancement (PRE) com-bined with SDSL using MTSL. The paramagnetic effect of thenitroxide radical of MTSL causes line broadening of NMR sig-nals of nuclei within 20–30 Å, whose magnitude depends onthe distances between the paramagnetic center and the nuclei(41, 42). In this experiment, we prepared a series of 15N-hHSF2DBDs conjugated to SUMO-1 harboring a single MTSLlabel and examined the PREs on the two-dimensional 1H-15Ncorrelation spectra attributed to the hHSF2DBD moiety. Inaddition to the C52A/Q53Cmutant, we prepared the followingthree SUMO-1 mutants that each contained a cysteine substi-tution of a solvent-exposed residue: C52A/S50C, C52A/V87C,or C52A/E89C. The sites for theMTSL labeling were chosen sothat they cover both molecular surfaces of SUMO-1; S50C andQ53C are located in the �-helix on an electrostatistically neu-

tral surface, and V87C and E89C are on the opposite negativelycharged �-sheet surface (Fig. 4B).Various signal attenuations were observed for many reso-

nances in the 1H-15N HSQC spectrum of SUMO-15N-hHSF2DBD, in which SUMO-1 was labeled with MTSL (Fig.4D). Compared with the HSQC spectra of SUMO-15N-hHSF2DBD without SDSL, some of the signals were nearlycompletely attenuated by theMTSL in each spectrum, presum-ably because of extensive line broadening by PRE. The PRE foreach residue of 15N-hHSF2DBD was quantified using the ratioof the NMR peak intensity of the MTSL-labeled sample to thenonlabeled one. In all four experiments, the hHSF2 residueswhose signals exhibited large PREs were mainly distributed onthe DNA interaction surface, in particular in/near the putativeDNA recognition helix of hHSF2DBD, helix �3 (amino acidresidues 57–68). Drastic and extensive effects of PRE wereobserved especially in the spectrum of the SUMO-1V87C andSUMO-1E89C conjugation forms, in which the MTSL wasattached to the acidic surface of the SUMO-1 moiety. Theseresults suggest that, although the relative position of SUMO-1to hHSF2DBDhas a large degree of freedom, the distribution ofSUMO-1 is biased toward the DNA interaction surface ofhHSF2DBD, i.e. the SOMO-1molecule is unequally distributedand is statistically more localized near the DNA-binding sur-face than the opposite side of hHSF2DBD. Furthermore, theacidic surface of SUMO-1 tends to be close to the basic DNAinteraction surface of hHSF2DBD, presumably because of along range electrostatic attraction between the complementa-rily charged surfaces of SUMO-1 and hHSF2DBD. Comparedwith the SUMO-1V87C conjugation form, the SUMO-1E89Cconjugation formexhibited larger PRE effects in the loop regioncontaining Lys-82. It is presumably because of the location ofGlu-89, which is closer to the C-terminal Gly-97 of SUMO-1and thus to the conjugating residue Lys-82 in the flexible loop ofhHSF2. These DEER and PRE experiments were performed in abuffer containing 200 mM KCl, same as the chemical shift per-turbation experiments.Based on the average interspin distances obtained from

the DEER spectra of MTSL-labeled SUMO-1C52A/Q53C-con-jugated hHSF2DBDK51C and SUMO-1C52A/Q53C-conjugatedhHSF2DBDE19C and the comparative models of hHSF2DBD,we calculated the relative positions of the conjugated SUMO-1to hHSF2DBD based on a protein-protein docking approachusing the HADDOCK 2.0 program (39). As no global confor-mational changes upon SUMO conjugation were evident fromthe results of chemical shift perturbation experiments, wechose 10 comparable modeled structures of hHSF2DBD (Fig.3A) and a structure of unliganded SUMO-1 taken from anensemble of NMR structures (PDB entry 1A5R) as start struc-

FIGURE 4. Spatial distribution of the SUMO-1 moiety in a conjugation complex. A, MTSL spin labeling positions in hHSF2DBD used for DEER measurementsare indicated on the comparative model of the HSF2DBD monomer bound to DNA. Representative structure of the comparative models (Fig. 3A) is shown asa ribbon diagram. MTSL was attached to the cysteine residue, which replaced Glu-19 (orange) or Lys-51 (green). Red stick models indicate the side chains of theSUMOylated target residue, Lys-82. The DNA recognition helix �3 is colored in blue. B, MTSL spin-labeling sites in SUMO-1 for DEER and PRE measurements areindicated on the ribbon representation (PDB entry 1WYW). Each cysteine mutant at Ser-50 (orange), Gln-53 (cyan), Val-87 (blue), and Glu-89 (magenta) waslabeled with MTSL. C, DEER measurements of SUMO-1C52A/Q53C-hHSF2DBDE19C (left) and SUMO-1C52A/Q53C-hHSF2DBDK51C (right). Upper and lower panels showDEER time-domain data and inter-spin distance distributions, respectively. D, paramagnetic relaxation enhancements observed in hHSF2DBD conjugated toeach MTSL-labeled SUMO-1 mutant (S50C, Q53C, V87C, or E89C). Using a color scheme corresponding to the NMR signal reduction ratio, the relaxation effectscaused by MTSL attached to the SUMO-1 moiety onto hHSF2DBD residues are mapped on the molecular surface calculated from the model of the hHSF2DBD-DNA complex displayed in A.




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

tures. Two hundred structures were generated for SUMO-hHSF2DBD using restraints for inter-C-� distances betweenGln-53 of SUMO-1 and Lys-51 of hHSF2DBD and betweenGln-53 of SUMO-1 and Glu-19 of hHSF2DBD, which werederived from the DEER experiments (see “Experimental Proce-dures”). Eighteen of 20 structures that showed the lowest inter-molecular energies were clustered into a group (supplementalFig. 3), whereas the remaining structure is inconsistent with theresults of PRE experiments, and thus is omitted from furtherdescription below (Fig. 5A). In these 18 structures, the averageC-� distance between Gln-53 and Glu-19 is 37.01� 0.62 Å andbetween Gln-53 and Lys-51 is 36.65 � 0.59 Å. On the otherhand, the DEER experiments showed the most populatinginterspin distances between Gln-53 and Glu-19 and betweenGln-53 and Lys-51 to be 46 and 45Å, respectively (Fig. 4C). Thedifference in distances between two C-� atoms (R��) of spin-labeled cysteines and between the two nitroxide moieties(RNONO) of a doubly MTSL-labeled protein has been shown torange from 0 to 10 Å, depending on the conformations of thelinker tethering the backbone and the nitroxide (40). Thus, the18 structures generated by the HADDOCK calculation arecompatible with the results of the DEER data. Furthermore,the average distance between Gly-97 C� and Lys-82 N-� ofthe final structures was 2.35 � 0.005 Å, which allows isopep-

tide bond formation. In the groupof HADDOCK structures, but nottwo outside the cluster (Fig. 5A),the acidic surface at the �-sheet ofthe SUMO is positioned in thevicinity of the DNA-binding sur-face of hHSF2DBD, and the struc-tures are consistent with theresults of the PRE experiments.These Haddock models presum-ably assume one of the most pop-ulated orientations of the SUMOmoiety in the conjugated complex.

DISCUSSION

HSF2 has been shown to be abona fide SUMOylation substrate,and its DNA binding activity toHSEis modulated by conjugation ofSUMO to the target lysine residue,Lys-82, in the DBD. Previously, theHSE binding activity of SUMOy-lated mouse HSF2 was investigatedby two independent groups, andopposite effects of SUMOylation,both enhancement and inhibition ofthe HSF2-DNA interaction, havebeen detected in EMSA using invitro-translated proteins (23, 24).Such inconsistent outcomes maybe due to differences of multipleSUMOylation states, interplayswith other post-translational modi-fications, or variations in DNA

probes or conditions used in the binding assays. In fact, aminorSUMOconjugation site, Lys-139, has been identified inHSF2 inaddition to Lys-82 (23), and the polycomb protein, Mel-18 hasbeen recently reported to inhibit the SUMOylation of HSF2(43). In this study, the direct influence of the SUMO conjuga-tion on the HSF2-DNA interaction was assessed using an iso-lated protein in a system that excluded the effects of other pro-tein factors and SUMOylation at Lys-139. Our biochemicaldata, obtained from EMSA and SPR measurements, haveclearly demonstrated that the interaction of full-length hHSF2with the HSE DNA is repressed by the conjugation of SUMO-1at its Lys-82. In addition, this negative effect of SUMOylationwas observed on the DNA binding activity of the hHSF2DBD, aconstruct that does not have the oligomerization domain,which implies that a loss of DNA binding cannot be attributedmainly to the perturbation of the trimer formation. Consis-tently, Anckar et al. (23) have reported that SUMOylation doesnot interfere with HSF2 oligomerization. The model ofhHSF2DBD in complex with DNA suggests that the dimerinterface of the DBDs in the TtT arrangement is seeminglydestabilized by SUMO-1 conjugation, although not for theHtHarrangement (Fig. 3B). However, in our experiments, the DNAinteractions of hHSF2 were equally reduced with both the HtHand TtT repeats. Taken together, we conclude that the modifi-

FIGURE 5. Possible orientations of the SUMO moiety conjugated to hHSF2DBD. A, superpositions of the 20lowest intermolecular energy structure models generated by the HADDOCK calculation using distancerestraints derived from DEER measurements. hHSF2DBD is shown in blue and SUMO-1 in green. Gray showsminor orientations of SUMO-1 (see text). C-� atoms of the spin-labeled residues are indicated by spheres.B, ribbon diagram of a representative model structure (left). Middle and right panels show its complex modelswith the right halves of HtH16 DNA and HtH26 DNA, respectively, each containing a single HSE repeat.




ww

.jbc.org/D

ownloaded from



http://www.jbc.org/

cation with SUMO-1 provides a direct impact on the interfacebetween HSF2 and DNA.To date, multiplemechanisms whereby SUMOylation exerts

functional regulation have been proposed. First, SUMOylationcompetes with other post-translationalmodifications for targetlysine residues. Alternative modifications of specific lysine res-idues with ubiquitin and an acetyl group have been reported inproliferating cell nuclear antigen and a tumor suppressorHIC1,respectively (44, 45). Second, the conjugated SUMO sequestersa binding site of the substrate protein for its downstream effec-tor (46). Third, as proposed based on the crystallographic stud-ies of the SUMO-conjugated TDG, conjugation of SUMOinduces conformational changes in substrates and alters theirmolecular interactions (20). Finally, SUMO that is covalentlylinked to substrates can serve as a binding tag, without a con-formational change in the substrate, which provides an addi-tional molecular interface for association with downstreameffectors, as observed for RanGAP1 (47–49). Recently, a shortpeptide motif, termed the SUMO interaction motif (SIM), wasidentified in SUMO-interacting proteins (50, 51). The SIMgen-erally contributes to establishing a new protein-protein inter-action that is mediated by SUMO.Of these mechanisms, the third one has been shown tomod-

ulate a protein-DNA interaction. The high binding affinity ofTDG for DNA containing an abasic site is greatly reduced bythe SUMOylation of TDG, likely through a conformationalchange mediated via noncovalent interactions between theconjugated SUMO and a SIM-like sequence in TDG (20).Morerecently, another example of the modulation of protein-DNAinteractions by SUMO has been reported. The DNA bindingactivity of a human transcription factor, glial cell missing Dro-sophila homologue a (GCMa), is repressed by the SUMOyla-tion of its DNA binding domain (52).hHSF2 has the primary SUMOylation site, Lys-82, in its

DBD. Therefore, we considered the possibility that theSUMOylation of hHSF2 appeared to influence the DNA inter-action activity through a conformational change of hHSF2 in asimilar manner to the SUMOylation of TDG. However, ourNMR spectroscopic analyses on SUMO-hHSF2DBD haveclearly demonstrated that the inhibition of the hHSF2-DNAinteraction by SUMOylation cannot be attributed to either astable interaction between the SUMO and HSF2 moieties or toa conformational change in HSF2. Another postulated mecha-nism was that the SUMO attachment destabilizes the dimerinterface of hHSF2DBD,which is essential for the specific bind-ing to the HSE repeats. This model accounts for the inhibitionof hHSF2DBDbinding to the TtTHSE repeats but not for bind-ing to the HtH repeats. Conversely, the results of the EMSAusing oligonucleotides HtH26 and HtH16 indicate that theSUMO-1 tethered to the hHSF2DBD interferes with the regionof DNA flanking the HSEs.DEER experiments combinedwith SDSL reveal that the posi-

tion of the SUMO-1 conjugated to hHSF2DBD is not fixed butdisplays a substantial spatial distribution around hHSF2DBD.Consistent with this notion, the SUMOylation site, Lys-82, islocated in a flexible loop as shown in the comparativemodels ofthe hHSF2DBD (Fig. 3A). Nevertheless, the results of the PREexperiments indicate that the spatial distribution and orienta-

tion of the conjugated SUMO-1 are biased in such a way thatSUMO-1 is more frequently positioned near the DNA-bindingsurface of hHSF2DBD with its acidic surface facing towardhHSF2DBD. Based on the results of the DEER and PRE exper-iments, we constructedmodel structures for a part of the highlypopulated conformations of SUMO-hHSF2DBD (Fig. 5A). Thepreferential distribution of SUMO-1 in relation to hHSF2DBDis likely caused by an electrostatic interaction between theacidic surface of SUMO-1 and the basic DNA-binding surfaceof hHSF2DBD. Nevertheless, the widths of the distribution ofthe interspin distance derived from the DEER spectrum indi-cate that the SUMO moiety can deviate from the most popu-lated structure, shown in Fig. 5, up to 10 Å (40). In support ofthe notion of an electrostatic interaction between the acidicsurface of SUMO-1 and the basic DNA-binding surface ofhHSF2DBD, chemical shift perturbation experiments ofSUMO-hHSF2DBD in a buffer containing 50 mM KCl showedthat themain chain amide resonances of Ile-34, Thr-42, His-43,Gln-85, Glu-93, and Gln-94 of SUMO-1 and Ser-60, Arg-63,Gln-64, Lys-72, and His-75 of hHSF2DBD, in addition to resi-dues near the conjugation site, exhibited small but significantchemical shift changes upon conjugation (supplemental Fig. 4,A andB). These chemical shift changeswere not observed in thepresence of 200mMNaCl (Fig. 2). Except forHis-43 andThr-42of SUMO-1, these residues are located on the acidic surface ofSUMO-1 or the basic DNA-binding surface of hHSF2DBD(supplemental Fig. 4C). Thus, the observed chemical shift per-turbation at low ionic strengthmay be caused by transient con-tact between the complementarily charged surfaces ofSUMO-1 and hHSF2DBD.The spatial distribution of SUMO-1, as deduced from the

DEER, PRE, and chemical shift perturbation experiments,explains the mechanism for the SUMOylation-induced de-crease of DNA binding affinities of hHSF2DBD. A comparisonbetween the model structures of SUMO-hHSF2DBD (a groupof 18 HADDOCK structures shown in Fig. 5A) and thehHSF2DBD-DNA complex suggests that SUMO-1 distributesin the space so as to make a steric exclusion with the phosphatebackbone of the DNA region flanking HSE in a probabilisticmanner (Fig. 5B). In contrast, the comparable model of theSUMOylated hHSF2DBD-DNA complex shows that SUMO-hHSF2DBD can binds to HtH16 without severe steric collision.The DNA lacks the flanking region that seemingly interfereswith SUMO-1. Therefore, the steric interference betweenSUMO-1 and DNA seems to be a major cause of the observeddecrease of DNA binding activity of hHSF2 upon SUMO-1attachment. Nevertheless, we cannot exclude the possibilitythat the attached SUMO-1 may also exert an electrostaticrepulsive effect that leads to the reduced DNA binding affin-ity of hHSF2DBD; the presumed distribution of SUMO-1suggests that in majority of the possible conformers theacidic �-sheet surface is predicted to be within 20 Å from thephosphate backbone of DNA. Collectively, the attachedSUMO-1 likely exerts steric and electrostatic interferenceswith the DNA backbone, whose magnitudes are dependingon the relative position of SUMO-1 in the distribution, whenthe complex between SUMOylated hHSF2 and DNA isformed.




ww

.jbc.org/D

ownloaded from




http://www.jbc.org/

Our biochemical and structural data demonstrate a novelmolecular mechanism whereby covalent attachment ofSUMO-1 negatively modulates formation of the protein-DNAcomplex. The attached SUMO-1 appears to make no stableinteractions with the DNA-binding surface of hHSF2, but dis-tributes in a certain area proximal to the surface, and stericallyor electrostatically interferes with DNA binding to hHSF2 in aprobabilistic manner. The biased distribution of SUMO-1 islikely attributable to the long range electrostatic attractionbetween the negatively charged �-sheet surface of SUMO-1and the positively charged DNA-binding surface ofhHSF2DBD.AsDNA-binding surfaces of proteins are generallyhighly positively charged, such electrostatic attractions mayoccur in some classes of SUMOylated DNA-binding proteins.Thus, the mechanism proposed for HSF2 here might be able toexpand to some of the nuclear factors whose DNA bindingactivities are suppressed by SUMOylation.HSF2 exhibits a relatively short half-lifetime in cells, which is

controlled by the ubiquitin-proteosome pathway (53); on thecontrary, HSF1 is stably and constitutively expressed in cells,and its activity is regulated by association with chaperon pro-teins (54). Interestingly, co-localization of HSF1 and HSF2 innuclear stress bodies, which is a subnuclear compartmentinduced by various stress signals, was previously observed (55),although HSF1 and HSF2 are supposed to be involved in dis-tinct processes, classical heat shock response and developmen-tal processes, respectively. In addition, HSF2 has been shownrecently tomodulate expressions of stress-inducible genes acti-vated by HSF1 either positively or negatively (5). On the otherhand, HSF2 has been shown to be modified by SUMO-1 in amitotic cell cycle-dependent manner, coupled to hsp70i genebookmarking (25). SUMOylation of HSF2 DBD could possiblyserve a molecular marker to control both its DNA bindingactivity and a timing of its turnover, independently on the ubiq-uitin-proteosome pathway. Further investigations are neces-sary to elucidate how the DNA binding activity of HSF2 sup-pressed by SUMOylation is linked to these HSF2-relatedcellular events.

Acknowledgments—Weare grateful toDr.Hisato Saitoh for providingexpression vectors of SUMOylation enzymes and the in E. coliSUMOylation system. We thank Dr. Akio Ojida and Prof. ItaruHamachi for their help in recording the SPR data.

REFERENCES1. Akerfelt, M., Trouillet, D., Mezger, V., and Sistonen, L. (2007) Ann. N. Y.

Acad. Sci. 1113, 15–272. Loones, M. T., Rallu, M., Mezger, V., and Morange, M. (1997) Cell. Mol.

Life Sci. 53, 179–1903. Sistonen, L., Sarge, K. D., Phillips, B., Abravaya, K., and Morimoto, R. I.

(1992)Mol. Cell. Biol. 12, 4104–41114. Loison, F., Debure, L., Nizard, P., le Goff, P., Michel, D., and le Drean, Y.

(2006) Biochem. J. 395, 223–2315. Ostling, P., Bjork, J. K., Roos-Mattjus, P., Mezger, V., and Sistonen, L.

(2007) J. Biol. Chem. 282, 7077–70866. Perisic, O., Xiao, H., and Lis, J. T. (1989) Cell 59, 797–8067. Bonner, J. J., Ballou, C., and Fackenthal, D. L. (1994) Mol. Cell. Biol. 14,

501–5088. Xiao, H., Perisic, O., and Lis, J. T. (1991) Cell 64, 585–593

9. Littlefield, O., and Nelson, H. C. (1999) Nat. Struct. Biol. 6, 464–47010. Rabindran, S. K., Haroun, R. I., Clos, J., Wisniewski, J., and Wu, C. (1993)

Science 259, 230–23411. Seeler, J. S., and Dejean, A. (2003) Nat. Rev. 4, 690–69912. Gill, G. (2004) Genes Dev. 18, 2046–205913. Rodriguez, M. S., Dargemont, C., and Hay, R. T. (2001) J. Biol. Chem. 276,

12654–1265914. Saitoh, H., and Hinchey, J. (2000) J. Biol. Chem. 275, 6252–625815. Gill, G. (2005) Curr. Opin. Genet. Dev. 15, 536–54116. Chang, C. C., Lin, D. Y., Fang, H. I., Chen, R. H., and Shih, H. M. (2005)

J. Biol. Chem. 280, 10164–1017317. Lin, D. Y., Fang, H. I., Ma, A. H., Huang, Y. S., Pu, Y. S., Jenster, G., Kung,

H. J., and Shih, H. M. (2004)Mol. Cell. Biol. 24, 10529–1054118. Dobreva, G., Dambacher, J., and Grosschedl, R. (2003) Genes Dev. 17,

3048–306119. Sachdev, S., Bruhn, L., Sieber, H., Pichler, A.,Melchior, F., andGrosschedl,

R. (2001) Genes Dev. 15, 3088–310320. Baba, D., Maita, N., Jee, J. G., Uchimura, Y., Saitoh, H., Sugasawa, K.,

Hanaoka, F., Tochio, H., Hiroaki, H., and Shirakawa, M. (2005) Nature435, 979–982

21. Hardeland, U., Steinacher, R., Jiricny, J., and Schar, P. (2002) EMBO J. 21,1456–1464

22. Hong, Y., Rogers, R., Matunis, M. J., Mayhew, C. N., Goodson, M. L.,Park-Sarge, O. K., and Sarge, K. D. (2001) J. Biol. Chem. 276, 40263–40267

23. Anckar, J., Hietakangas, V., Denessiouk, K., Thiele, D. J., Johnson, M. S.,and Sistonen, L. (2006)Mol. Cell. Biol. 26, 955–964

24. Goodson,M. L., Hong, Y., Rogers, R.,Matunis,M. J., Park-Sarge,O. K., andSarge, K. D. (2001) J. Biol. Chem. 276, 18513–18518

25. Xing, H., Wilkerson, D. C., Mayhew, C. N., Lubert, E. J., Skaggs, H. S.,Goodson, M. L., Hong, Y., Park-Sarge, O. K., and Sarge, K. D. (2005)Science 307, 421–423

26. Hilgarth, R. S., Murphy, L. A., O’Connor, C. M., Clark, J. A., Park-Sarge,O. K., and Sarge, K. D. (2004) Cell Stress Chaperones 9, 214–220

27. Uchimura, Y., Nakamura, M., Sugasawa, K., Nakao, M., and Saitoh, H.(2004) Anal. Biochem. 331, 204–206

28. Delaglio, F., Grzesiek, S., Vuister, G. W., Zhu, G., Pfeifer, J., and Bax, A.(1995) J. Biomol. NMR 6, 277–293

29. Goddard, T. G., and Kneller, D. G. (2002) Sparky3, University of Califor-nia, San Francisco

30. Vuister, G.W., Kim, S. J.,Wu, C., and Bax, A. (1994) J. Am. Chem. Soc. 116,9206–9210

31. Hara, H., Tenno, T., and Shirakawa,M. (2007) J. Magn. Reson. 184, 78–8432. Pannier, M., Veit, S., Godt, A., Jeschke, G., and Spiess, H. W. (2000) J.

Magn. Reson. 142, 331–34033. Shi, J., Blundell, T. L., and Mizuguchi, K. (2001) J. Mol. Biol. 310,

243–25734. Eswar, N., Eramian, D.,Webb, B., Shen,M. Y., and Sali, A. (2008)Methods

Mol. Biol. 426, 145–15935. Vuister, G. W., Kim, S. J., Orosz, A., Marquardt, J., Wu, C., and Bax, A.

(1994) Nat. Struct. Biol. 1, 605–61436. Krissinel, E., and Henrick, K. (2004) Acta Crystallogr. 60, 2256–226837. Koradi, R., Billeter, M., and Wuthrich, K. (1996) J. Mol. Graphics 14,

29–3238. DeLano, W. L. (2002) The PyMOL Molecular Graphics System, De Lano

Scientific, San Carlos, CA39. Dominguez, C., Boelens, R., and Bonvin, A. M. (2003) J. Am. Chem. Soc.

125, 1731–173740. Borbat, P. P., McHaourab, H. S., and Freed, J. H. (2002) J. Am. Chem. Soc.

124, 5304–531441. Bloembergen, N., and Morgan, L. O. (1961) J. Chem. Phys. 34, 842–85042. Iwahara, J., and Clore, G. M. (2006) Nature 440, 1227–123043. Zhang, J., Goodson, M. L., Hong, Y., and Sarge, K. D. (2008) J. Biol. Chem.

283, 7464–746944. Hoege, C., Pfander, B., Moldovan, G. L., Pyrowolakis, G., and Jentsch, S.

(2002) Nature 419, 135–14145. Stankovic-Valentin, N., Deltour, S., Seeler, J., Pinte, S., Vergoten, G., Guer-

ardel, C., Dejean, A., and Leprince, D. (2007) Mol. Cell. Biol. 27,2661–2675




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

46. Pichler, A., Knipscheer, P., Oberhofer, E., vanDijk,W. J., Korner, R.,Olsen,J. V., Jentsch, S., Melchior, F., and Sixma, T. K. (2005) Nat. Struct. Mol.Biol. 12, 264–269

47. Pfander, B., Moldovan, G. L., Sacher, M., Hoege, C., and Jentsch, S. (2005)Nature 436, 428–433

48. Reverter, D., and Lima, C. D. (2005) Nature 435, 687–69249. Uchimura, Y., Ichimura, T., Uwada, J., Tachibana, T., Sugahara, S., Nakao,

M., and Saitoh, H. (2006) J. Biol. Chem. 281, 23180–2319050. Hecker, C. M., Rabiller, M., Haglund, K., Bayer, P., and Dikic, I. (2006)

J. Biol. Chem. 281, 16117–1612751. Kerscher, O. (2007) EMBO Rep. 8, 550–55552. Chou, C. C., Chang, C., Liu, J. H., Chen, L. F., Hsiao, C. D., and Chen, H.

(2007) J. Biol. Chem. 282, 27239–2724953. Mathew, A., Mathur, S. K., and Morimoto, R. I. (1998)Mol. Cell. Biol. 18,

5091–509854. Morimoto, R. I. (2002) Cell 110, 281–28455. Alastalo, T., Hellesuo, M., Sandqvist, A., Hietakangas, V., and Kallio, M.

(2003) J. Cell Sci. 116, 3557–3570




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

Azusa Seto, Akira Nakai, Tetsuro Kokubo, Hidehito Tochio and Masahiro ShirakawaYukihiro Tateishi, Mariko Ariyoshi, Ryuji Igarashi, Hideyuki Hara, Kenji Mizuguchi,

of Heat Shock Factor 2Molecular Basis for SUMOylation-dependent Regulation of DNA Binding Activity

doi: 10.1074/jbc.M806392200 originally published online November 18, 20082009, 284:2435-2447.J. Biol. Chem.

10.1074/jbc.M806392200Access the most updated version of this article at doi:

Alerts:

When a correction for this article is posted•

When this article is cited•

to choose from all of JBC's e-mail alertsClick here

Supplemental material:

http://www.jbc.org/content/suppl/2008/12/01/M806392200.DC1

http://www.jbc.org/content/284/4/2435.full.html#ref-list-1

This article cites 53 references, 24 of which can be accessed free at


ww

.jbc.org/D

ownloaded from

http://www.jbc.org/lookup/doi/10.1074/jbc.M806392200

http://www.jbc.org/cgi/alerts?alertType=citedby&addAlert=cited_by&cited_by_criteria_resid=jbc;284/4/2435&saveAlert=no&return-type=article&return_url=http://www.jbc.org/content/284/4/2435

http://www.jbc.org/cgi/alerts?alertType=correction&addAlert=correction&correction_criteria_value=284/4/2435&saveAlert=no&return-type=article&return_url=http://www.jbc.org/content/284/4/2435

http://www.jbc.org/cgi/alerts/etoc

http://www.jbc.org/content/suppl/2008/12/01/M806392200.DC1

http://www.jbc.org/content/284/4/2435.full.html#ref-list-1

http://www.jbc.org/

molecularbasisforsumoylation-dependentregulationof ... · domains, hsf1 and hsf2 possess a...

Documents