Volun~e 27S, i}mubcf I, 115,11,~

FI~BS (Jp;~18

J;trmm'y

1991

,4 IH)NI~ lXq 4 5+'",+),/IIXx))4,'+

The histone H l-&cZ'/hsion protein produced in Escherichia cob binds to the 5- F T G C i C A n n n T G C C A A . motif on D N A Riitta-Maaria M a n n e r m a a and J o u k o {)ikarinen Colla.,,ei~ Research ('nit. Bim'cnter and l)rtmrtnient q f Medical BtochrmL~trv, Univcrsit r ~q'Ouh¢. Kujaanintic 52/L .'it'. 0 , . ~ Ottlu, Fitlhtnd Received 22 ()ctober 19~Jt]: revised version received 12 Noveml~er 1990 The cnditl~_I region (ff the diickcn histone I t 1.03 gene was cloned to a bacterial expression vector, and tile 291 .llrl~iflo acid LI I ,/!.galacto.~idase I'tl~[on pr,,tein wa.~ i~ol:lted :tfter induction with l lrrG. The li;~ion prt,tcin recognizes the 5':rT()fI{;AnnnTCiC'CAA-.Y molif on DNA, Tile Ill l~lohular domain wa.~ initi;dly ~hown to be re.~poafihlc for the sequcnce.specilic bindin~ by functioual ddetion anal)'~i~, "l'hi~ functioz~ m~y by i/Mi~lren.~iihle I'~r the role of ii I ;ts a detcrminanl of nudcosomc positioning and as a eukaryuti~: rcprc.~oi'. I iistone II1: Nudeo~ome: Chr0matin: DNA binding: l+,eprcssion

1. INTRODUCTION Histone HI has been thought to function as a general represser in cukaryotes, being able to aggregate chromatin in an inactive state [1-41. At least 4 - 5 subtypes of HI exist in chromatin in each species, binding to hinge regions of nucleosomcs [5], formed by two turns o f DNA twined around core histories [6,7]. It has been observed that the 3D structure of the globular domain of histone H5 [8], an erythrocytespecific H I , displays similarity to the DNA binding domain in CRP, the ce.carbon backbones of the two being superimposable, Based on these data and tertiary structure homology of the globular domain with DNAbinding homeodomains, it has been suggested that H1 may be a sequence-specific DNA binding protein [9]. It has recently been demonstrated that H1 binds to a CTF/NF-I recognition sequence in the mouse a'2(I) collagen promoter [10] and to specific sites in the rat albumin promoter [11]. H1 can also be isolated by D N A affinity chromatography using the consensus for CTF/NF-I binding sites, 5'-TTGGCAnnnTGCCAA-3' [12], as the ligand (unpublished results). H1 may therefore be one of the proteins of regulatory importance binding to that motif [13]. The present work was undertaken to prove definitely Correspondence address: J. Oikarinen, Department of Medical Biochemistry, University of Oulu, Kajaanintie 52A, $F-90220 Oulu, Finland

that historic H1 is a .~equence.specific D N A binding protein by stu,lyin$ this function in an HI-iacZ' fusion protein produced in E. colt. 2. EXPERIMENTAL 2.1. Clotling anti production o f histone Hi.03 in bacteria Chicken generate DNA was isolated as described elsewhere [14], The coding region of the H 1,03 gene was anaplified by PCR [15] with I/+g of the generate DNA as the template in the presence of 10~10 DMSO, using a Perkin.Elrner Cetus DNA Thermal Cycler. Oligonucleotides of the sequences 5'-ACGCCAAGCTTCATGGCT GAGACCGCTCCTGTCGCT-3' and 5'-CGAGCTCGGTACCTT TFTCTTAGGGGCCGCCTTCTT-3 ', corresponding to a part of the sense strand at the 5' cad and to a part of the antisense strand at tile 3' end respectively, were used as tile primers. These contained 5' overhangs for the H#zdlll and Kpnl sites respectively. The amplified fragment of 696 bp was purified by PAGE, cut with the two enzymes and ligated to plasmid pUCI9 between the Hindlll and Kptll sites, E, colt JMI09 was transfarnled by the construct, A deletion mutant lacking the sequence coding for the C.terminal tail of FII was generated using all oligonuclcotide of the sequence 5 '.GAGCTCGGTACCTTCTTAGGAGCCTTCTCCTTCAC-3 ' as the 3' primer and plasmid PER51 as the template, Plasmid preparations were isolated, and subjected to DNA sequencing by tile dideoxy method [16l. S~,nthesis of the fusion protein wa~ induced by I0 mM IPTG [17], a 10-ml culture giving a yield of about 10an. In order to isolate the fusion protein, the bacteria ,sere lysed under the conditions described for plasmid isolation [14]. The lysates were loaded onto an anti-~galaetosidase affinity column under the condition~ described for DNA affinity chromatography [10], and the bound material was eluted by raising the pH to 10.8, The eluted material was fractionated and the pH was adjusted to 7.5 as soon as possible.

2,2. Assays for DNA binding Abbreviations: IPTG, isopropyl-~-D-thiogalactopyranoside; CTF/NF-I, CCAAT transcription factor/nuclear factor-l; PCR, poiymerase chain reaction; DMSO, dimethylsulfoxide; EMSA, eiect r o p h o r e t i c mobility shift assay; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophorcsis; PCA, perchloric acid

Published by Elsevier Science Publishers B. V. (Biomedical Division)

Tile binding of proteins to DNA was assayed by EMSA [18,19], slightly modified [20]. Preparative EMSA was carried out as described [211. Briefly, the isolated protein was incubated with an oligonucleotide labeled at its 5 ' ends. The retarded band was cat off, the proteins eluted and separated by glyceroI-SDS-PAGE, The effect

] 15

Volume 278, number

January

FEBS LETTERS

1

1991

13

.a -300

300

cl

600

900

L,,--L-li_L_.i.

I_&

/

1

f4indlll

HindlIt

chicken H-l.03 aen

Km

la

.-e

chicken

C

2 __

___ genomic DNA

I

H1.03 223

1

_“__

HpN -COOkI Globular

PCR-51

domain

-

Fig. 1. The strategy used for amplification of the coding region of the histone H1.03 gene (Panel A) and cloning in the frame in conjunction with the l&Z gene (Banei B). The locations of both the. sense and antisense strand PCR primers that contained overhangs for the N&tdHi and KpnI restriction sites and the internal sequencing primers are depicted in Panel A.

CTFINF-I recognition sequence were annealed and puri~ed by PAGE. Qligonucleotides were radioactively labeled at their 5 ’ ends using polynucleotide kinase. The fusion protein material for protein sequencing was separated by gIyc~rol-§DS-PAGE and the proteins stained with Coomassie brilliant blue for a short period of time. The proteins were electroeluted and subjected to sequencing by Edman degradation using an Applied Riosystems 477A Protein Sequencer and 12OAAnalyzer. Glycerol-SDS-PAGE was performed as described previously [22]. The gels were subjected to staining with Coomassie brilliant blue or silver 1231.

of Hl ~tibodies on binding of the fusion protein to DNA was demonstrated by EMSA, as described [IO]. 2.3. Additional merltods An Hl fraction capable of binding to the CTFINF-I recognition sequence was isolated from rat liver nuclear extracts by heparinagarosc and two consecutive DNA affinity chromatographias [lo]. Oligonucleotides were synthesized using a Beckman DNA Synthesizer. The primers in the PCR or dideoxy sequencing were used without further purification. while the oligonudeotides used for the synthesis of the double-stranded oligonucleotide containing the

1 LlrrusOi~Eonuc~eotidefor the CTF/NF-I birding sequences. Panel A demonstrates Fig. 2. EMSA using the Hi-l&Z fusion protein and the CO-W** the binding. Lane 1 contains only the labeled double-stranded oligonucleotide, and lanes 2 and’4 100 ng of Hl isolated from rat liver using DNA recognition site affinity chromatography, 100 ng of the Hi .03-f&?,‘.fusion protein produced in E. co&and both of them, respectively, in addition to the oligonucleotide. Lanes S to 9 contain the fusion protein and lanes 6 to 8 a lo-, 50. and 100.fold excess of the oligonucleotide unlabeled, respectively. All the lanes contain 2.g yg of double-stranded poly(dI-dC). The retarded complexes and the free labeled fragment are depicted by’ arrows. Pane] B demonstrates the effect of H1 antibodies on the mobility of the fusion protein-DNA complex in EMSA. Lanes 1 and 2 corttain Hl-/u&Z’ incubated in the presence of an increasing concentration non-immune serum and lanes,3 and 4 that incubated in the presence of the, Hi antiserum. Lane 5 is the labeled DNA fragment alone: The slow mobility band observable in the presence of antibodies to.Hl is denoted by an arrow. 116

“’ ,’ _:

Volume 278, number 3. RESULTS

1

FEBSLETTERS

AND DISCUSSION

The coding region of the chicken H1.03 gene [24] was cloned to the polylinker area of plasmid pUC19 between the NiridIII and KpnI sites (Fig. 1) using PCR. The’primers were designed so that the genomic frag‘ment obtained after digestion with the appropriate enzymes could be cloned in the correct orientation in conjunction with the /acZ’ gene to generate an open reading frame of 873 nucleotides coding for an H1.03-lacZ’ fusion protein of 291 amino acids. A number of positive clones was obtained by restriction site analysis, and one of these, PcR~51, was subjected to sequencing by the dideoxy method. Roth strands were sequenced, and the construct was shown to conPain the intact H1.03 coding region in conjunction with the l&Z’ gene. The protein coded by the fusion gene was produced in E. coli, and purified using chromatography on an anti-&Z’ antibody column. An apparent molecular mass 36 kDa protein was obtained, its mobility being in agreement with that of H1.03, which migrates in the 32 kDa position by globular standards despite consisting of only 223 amino acids. An additional molecular mass 14 kDa protein was observed, but this turned out to be lysozyme used to lyse the bacteria. The 36 kDa band was recognized in Western blots by antibodies to Hl and ,&galactosidase, and, final proof of the identity of the proteins was obtained by protein sequencing, the proteins being separated by glycerolSDS-PAGE, eluted from the gel and subjected to determination of the N-terminal sequence by Edman degradation. The molecular mass 14 kDa protein gave

January

1991

the lysozyme seqjuence and the other that of the fusion protein. The HI-1ucZ’ fusion protein binds to the 5 ’ -TTGGCAnnnTGCCAA-3 ’ motif on, DNA, as demonstrated by EMSA (Fig, 28). A signal of slower mobility than in the case of Hl. isolated from rat liver was obtained, the mobilities of the protein-DNA complexes being in good agreement with the relative molecular masses of the two proteins. The binding is specific, since it was counteracted by the addition of an excess of the unlabeled binding motif, and clear competition was obtained with increasing amounts of the competing oligonucleotide. On the other hand, actually no binding was observed to a synthetic oligonucleotide containing only one half-site of the symmetric CTF/NF-I binding motif (Fig. 3). Two additional facts suggest that the relative mobility 36 kDa Hl-lscZ’ fusion protein causes the retardation. Firstly, only that ‘protein could be detected by silver staining in a glycerol-SDS-polyacrylamide gel after electroelution and subsequent ellectrophoresis of proteins from the retarded band (data not shown), and secondly, incubation in the presence of Hl antiserum resulted in the appearance of a low mobility band in EMSA, as in the case of Hl isolated from rat liver (Fig. 2B). A deletion mutant of H1.03 lacking amino acids 147 to 223 was produced in E. coli using the same strategy as for the production of the wild-type protein. The mutant protein binds in EMSA to the oligonucleotide containing the CTP/NF-I recognition motif with a two- to three-fold lower affinity compared with the wild-type protein (Fig. 4). The binding is specific, however, since

Fig. 3. Specificity of binding of the H1.04-la& fusion protein. Panel A demonstrates binding of Ml-k2’ to synthetic labeled ollgonucleotides containing the CTF/NP-I consensus sequence [If] (lane 2). a putative site from the mouse a~([) collagen promoter with dne intact half-site [XI] (lane 4) and an API site from the rat ttansin gene 142) (lane 6). Lanes 1, 3 and 5 do not contain protein. Panel B d emonstrates comnetition for ‘binding of Nl-luc2’ to the CTFINF-I consensus motif from synthetic unlabeled oligonucleotides. Lanes 2 to 4 contain I, 10 and 100 ng of the as(I) collagen oligonucleotide unlabeled, and lanes 6 to 8 that of the transin oligonucleotide, respectively. Lanes 1, 5 and 9 do not contain any competitor+ Lane 10 does not contain protein.

117

1 2 3

Fig.

4,

Pancl

A

depicts

the

.Watcgy

used

for

lhc

production

of

311

tlclctinn mutant th;~ is lnrking the C-terminal tail. 13ancl I3 demOnstrntcs that the m\ltimI prowin binds to DNA in I? scqucnce~ specific manner. Binding Of the fusion protein IO rhc 5’-labclcd oligxmclcotitlc Wi\S Ossnycd using EMSA. LXW+ 1 and 3 ~~llt;~ill 100 tig Of the nluulllt prokin, llllil inllc 2 a IOO-fold CxcCSs Of tllc C’TF/NF.I oligonr~clco~idc UI~lObclcd, as n competitor. The rc!nrdcd somplsx is dcpictcd by an a-row. I-11 -03

increasing amounts of the oligonuclcotide unlabclcd readily counteract it and no sompetition is observable with non-specific oligonucleorides (see Fig. 3). The EMSA experiments demonstrate that the H l-/slcZ’ fusion protein is sapable of sequence-spccifis binding to DNA, thus confirming the results obtained previously using H 1 purified from rat liver by DNA affinity chromatography [ lQ,SO’Jand by PCA extraction [I 11. The globular domain of WI may be responsible for rhis interaction, a5 demonsrratsd using the dektioaa mutant, although it should be noted that the mutant protein contains approximately 30 extra amino acid residues both in the N- and C-terminal ends of rhc globular domain. The major problem in dernontirraring ssquencespecific interaction of the HI-k%’ fusion protein with DNA results from the high level of non-sequencespecific interaction. Consistent with this, many of the histones have recently been demonstrated to contain a novel DNA binding unit, an SPKK motif, capable of

interaction with the exceptionally narrow DNA minor groove in A+ T-rich regions [2§,%6]. The C-terminal tail of HI contains several copies of this motif, but all these were removed in the deletion mutant. We suggested in our previous report that the N-terminal end of

111Wcintrnub, 14.(1985)Cell 42, 709-711. PI Wolffc, A.P, (1989)EMBO .I. 8, 527-537.

I31 Sun, J.-M., Scicncc

Windsrkicwicz,

R. and

RuirGrrillo,

A.

(1989)

245, 68-71,

141Zlalanova, J. (1990)Trends

Uiochem. Sci. 15, 273-276, Staynov, D.Z. and Cranc-Robinson, C, (1988) EMBQ J, 7, 3689-3691, T.J., Finch, J.T., Rushton, B., Rhodes, D. and 161 Richmond, Klug, A. (1984) Nature 311, 532-537. I71 Morse, R.M. and Simpson, R.T, (1988) Cell 54, 285-289. AM, Nilges, M,, Sukumnran, PI Clorc, G.M., Groncnborn, D.K. and Zarbock, J, (1989) EMRO J, 6, 1833-1842.

PI

PI Manncrmaa, R.-M. and Oikarinen, J. (1989) Biochem. Uiophys. RSS. Common. 162, 427-434. [lOI Ristinicmi,J. and Qikarinen, J. (1989) J. Biol. Chem. 264, 2164-2174. Ssvall, J.S. (1988) Biochemistry 29, 5038-5044. tt:/ Nitsson, P., Hallberg, B., Thornell, A. and Grundstrtim, T. (1989) Nucleic Acids Res. 19, 4Ml-4075. [131 Santoro, C., Mcrmod, N., Andrcws, PC. and Tjian, R. (19&j Nature 334, 218-224. 1141 Maniatis, T., Fritsch, E.F. and Sambrook, J. (198%) Molecular Cloning. A Laboratory Manuel. Cold Spring t-iarbor Laboratories, Cold Spring Harbor, New York, USA. b4Sl Saiki, R.K., Oelfand, B.H., Stoffet, S.. Scharf, S.J., Miguchi, R., Horn, G.T., Mullis, K.B. and Ertich, H.A. (1988) Science 239, 487+1, WI San&r, F., Nicktcn, §. and Coutson, AR. (1979) Pros. Natl. Acad. Sci. USA 74, 5463-54G7. (171 Kadonaga, J.T., Carrier, K.R., Masiarz, F.R. and Tjian, R. (1987)Ccll 51, 1079-1090.

VOhilll¢ 278, llllitlb¢i' I

I"i~IIS i,,I~:fl'f!R

,hliHl;Ify I')91

Ilsl i1'~t

I~71 '~:'fal',¢,R0bhi~OlhI~',;rod l~il~., ill|, i19~9) PmI~lu l!Ul~,2.

i201 ( )ik;~'hwn. 1,. Ii;ll;Ul~o~'hi, A, ;rod lh~ ('f,:mbwi~'~h~. It, ( I ~)S7) J, Itlot. ('hem, 2t~2~ 11(~.1:. l l071X 1211 llul)p, P,,A,WI ;~II~IF;il~p¢l,A,I!, 119,~?) Nu~l¢i~' A~:id~ I{¢', I~, ~)')'l):' ,)7~t~, I.%~1

[)9] l¢i,d~i'd-Foy. I'i, m~d

I.t~,l I~.tl (.!oI,:~.i,,,~,,R(~l)io~, A,J,. Madl~y, I,,K, ;rod W¢Ik, ,I,R,I",, 12~1 ,~li/~t'ki,NI, (19S91 EMIiO J, S, 79'~,,,Sl~h

ll~i~oh ¢;,I,, (1~)~'/) l i M I l O

,I, h,

(~0] lh~d~'i~u¢~,Cm~q~, A,, ,~himam~.';h A, and W~.'~el, A, (19~9) Jl, M~I, lliol,209, 13~,.-,I,~0, I ,I,~ 2,1 '' I ,t~,'~0. 1,~21 I

The histone H1-lacZ' fusion protein produced in Escherichia coli binds to the 5'-TTGGCAnnnTGCCAA-3' motif on DNA.

The coding region of the chicken histone H1.03 gene was cloned to a bacterial expression vector, and the 291-amino acid H1-beta-galactosidase fusion p...
757KB Sizes 0 Downloads 0 Views