Structural and functional differences between histone Hi sequence variants with differential intranuclear distribution

(1)

Genetics

Structural and functional differences between histone Hi sequence variants with differential intranuclear distribution

(chromatinstructure/DNA-proteinbinding/proteinsequencemotifs/interphasechromosomes/Chironomus)

EKKEHARD SCHULZE*, LOTHARTRIESCHMANN*, BETTINA SCHULZE*, ERWIN R. SCHMIDTt, SABINE PITZEL*,

KASPER ZECHEL4, AND ULRICH GROSSBACH*§

*ThirdDepartment of Zoology-Developmental Biology, University of Gottingen, Berliner Str. 28, D-3400G6ttingen,Federal Republic of Germany;

tDepartmentof Genetics,UniversityofMainz, Saarstr. 21,D-6500Mainz,Federal Republic of Germany; andtMax-Planck-Institut

fOr

Biophysikalische Chemie,Am Fassberg, D-3400Gottingen,FederalRepublic of Germany

Communicated by M. M.Green, September 18, 1992

ABSTRACT The chromatin of most cell types contains several different sequence variants of histone Hi. The functional role of this

heterogeneity

is not known. In the larval tissuesof the midge,Chironomus thummi, there are

Hi

variants of twotypes. Hi 11-1, Hi 11-2,andHirI-1havesimilaramino acidsequences and appear uniformly distributed in polytene interphase chromosomes. The total number of gene copies per genome for this type of

Hi

histones is about 40inC.th. thummi and50-60inC. th. piger. In contrast, histone

Hi

I-1 is encoded by a single copy gene in C.th. thummi and by two to four genes inC. th. piger. It has a divergent structure and is found only in a limited number of condensed chromosome sites. The N-terminal domain of Hi I-1 contains an insertion that is lacking in the otherHi variants and that is part of a variant- specific bipartite sequence

Lys-Ala-Pro-Lys-Ala-Pro-Xaal0-

Lys-Val-Ala in front of the conserved central domain. N-ter- minalpeptides ofHiI-1includingthismotif, in contrast to the homologous peptide from Hi IH-1, competed with the drug Hoechst33258 for binding to the minor groove of the DNA double helix. Repeats of the sequence Lys-Ala-Pro are also present at the same distance from the conserved central domain,ina singleHi variant of a nematode and of agreen alga. Themotifcould interact with linker DNAinintranuclear targetingorpackaginga condensed subtype of chromatin,or both.

Thefolding of the nucleosome chain into the 30-nm fiber of chromatin is mediatedbythe bindingof histone Hi (1).In vitrostudies have shown that

Hi

is a repressor oftranscrip- tion and that its association with nucleosomes is lost or altered duringthe transcriptional process (for reviews, see refs. 2 and 3). On the otherhand, in situdecoration ofgiant chromosomes with monoclonal antibodies and subsequent ultrastructural analysishaverevealed that the chromatin of highly active Balbiani ring genes contains histone Hi (4).

Thesecontradictoryresultsaswellasotherevidence indicate that the role ofHiin the architecture of chromatin may be less uniform than is usually assumed. Within one nucleus, certain

Hi

subtypesmightrepress

transcription

whileothers may not interfere with this process. Several different sequencevariants ofHi have beenfound inmost

eukaryotes

that havebeenanalyzedcarefully (reviewedin refs.5-7),and indirect evidence suggestsafunctional role of this

diversity

(6). Until now,however, direct

experimental

evidence fora distinct intranuclear location of different Hi

subtypes

and for functional differences between them has been

lacking.

Using

subtype-specific

monoclonal

antibodies,

we have shown(8) thatdifferent Hi sequence variants are differen- tiallydistributed withinthe

salivary gland

nuclei ofthe

midge

Chironomus thummi. Several of the Hi variants of this species appear to be present throughout the polytene chromosomes,whereasHi I-1, the variant with the most diverg- ing structure (9), was detected only in a limited number of chromosome bands. Many of these bands replicate late in S-phase(10) and containrepetitive DNA(11, 12). They are morefrequent in C. th. thummi than in C. th. piger. Accord- ingly,Hi I-1 was found to account for about20%oof total Hi inlarvae ofC. th. piger and for about 30% in larvae of C. th.

thummi (8).

Thedemonstration within one nucleus of differences be- tweenchromosome sites in theirHi subtypecontent makes C. thummiamodelorganismforstudyingpossible functions of Hi heterogeneity. Therefore, we have determined and compared the sequences of thedifferent Hi histone genes of C.thummiinasearch for structural features that are peculiar toHi I-1 and couldbeinvolved inestablishingaparticular subtype of chromatin. We find thatHiI-1, in contrast to the other Hivariants,contains a novel sequencemotifthat is not knownfrom otherproteins but is also present insingle Hi variants ofother,evolutionarilydistant organisms. An N-terminalpeptide of Hi I-1 that includes this motif also differs from the homologous region in another Hi variant of C.

thummi by competing with the drug Hoechst 33258 for binding to the minor groove of the DNA double helix.

Interaction of the subtype-specific motif with linker DNA may play a role in intranuclear targeting ofHi I-1 or in establishingacondensed type ofchromatin, orin both.

MATERIALS AND METHODS

Clone Isolation and DNA Sequencing. Genomic clones containinghistonegene sequenceswereidentified inaC. th.

thummilibrary (13) byacloned histone generepeatingunit fromDrosophila melanogaster (14), andHindIII restriction blots werescreened forHi sequencesbyanoligonucleotide (15-mer with 32 sequencevariants)deduced from the penta- peptide, Ser-Gly-Ser-Phe-Lys, that represents a sequence highly conserved in Hi proteins. One HindIII fragment containingmostof the codingregionofanHi gene andone AccIfragmentcontainingtheentire gene with its 5' and 3' flanking regions were subcloned in the pUC18 vectorand sequenced. TheHindIIIfragmentwasalso usedas aprobe forscreeningagenomic C. th.piger

library

constructed by L.T.; the DNA of 16 positive clones was restricted

by

HindIII,separated^onagarose

gel,

transferredto

Hybond-N

(Amersham) according ^tothe instructions of the manufac- turer, and hybridized at high

stringency

with a

pUC18

subcloneof theHindIIIfragmentthat contains 400 basepairs (bp) of the Hi gene

coding

for parts of the central and N-terminaldomains.

Appropriate

restriction

fragments

con-

§To whomreprint requestsshouldbe addressed.

Thepublicationcostsof this articleweredefrayedinpartbypagecharge payment.Thisarticlemustthereforebeherebymarked"advertisement"

inaccordance with 18 U.S.C.§1734solelytoindicate this fact.

(2)

taining Hi genes were subcloned in the pUC18 vector and sequenced. All sequencing was performed by the dideoxy method (15).

A gene probe specific for HiI-1of C. th. piger (see Results) was used to screen a genomic C. th. thummi library made by B.S., and a positive clone containing the C. th. thummiHlI-1 gene was subcloned and sequenced.

Isolation andPurificationof Hi Peptides. Histone Hi variants I-1 and II-1 from larvae of C. th. thummi were isolated as described (9), purified by reversed-phase HPLC, and cleaved with BrCN at their single methionine residue. In both proteins, the cleavage yielded a short N-terminal and a long C-terminal peptide. The N-terminal peptide of each protein was purified by reversed-phase HPLC and used in the DNA binding experiments. The N-terminal peptide of Hi variant I-1 from C. th. piger was obtained by expression in anE.coli expression system that uses bacteriophage T7 RNA polymerase (16).

DNA Binding in Competition with Hoechst 33258. Salmon sperm DNA (82.3 nM in phosphate residues) was preincubated with Hoechst 33258 (Hoechst) at a concentration of 10.9 nM. The fluorescence intensity of the solution was measured at 25°C in 10 mM sodium phosphate buffer, pH 7.1/150 mM KCl in a Perkin-Elmer LS 50 spectrofluorimeter at 450 nm (excited at 354 nm). The true fluorescence, F, of the DNA-dye complex was calculated by subtracting the back- ground fluorescence of all components minus dye, and the fluorescence of the dye and all components minus DNA.

Increasing amounts of peptide were then added, and the decrease of F was expressed in percent relative to the true fluorescence of the DNA-dye complex without peptide.

Estimation of Gene Copy Numbers. Histone Higeneprobes were hybridized on Southern blots to equivalent molar amounts of genomic DNA and plasmidDNA. Forcalibration, a series of plasmid DNA concentrations equivalent to increasing numbers of gene copies per genomewasused,based on genome sizes of 2.8 x 108 bp (C. th. thummi) and 2.18 x 108 bp (C. th. piger) as determined by H. G. Keyl(personal communication). DNA of plasmid clones containing the complete coding sequence of Hi genes I-1, II-1, and II-2, respectively, and different stretches of flanking sequences were digested withrestriction enzymes to yield thegenomic insert and were separated on agarose gel together with genomic DNA. Southern blotting onto Hybond-N was performed as described above. Gene probes were as follows:

probe 1, a 319-bp Pst I-BamHI fragment encoding the C-terminal half of HiI-1 of C. th. piger [specificforHi genes of type I-1 in both subspecies (L.T. and B.S., unpublished data)]; probe 2, a 4-kbpHindIII fragment encodingtheentire HiI-1gene of C. th. thummi withflanking sequences; probe 3, a 400-bp Hph I fragment encoding thecentral domainand part of theN-terminalandC-terminal domains of Hl II-1 of C. th. thummi. At high stringency, probe 3 hybridized with the Hi II-1, Hi 11-2, and Hi III-1genes but not withHi I-1 genes (L.T., unpublisheddata). Theprobeswerelabeledwith digoxigenin-11-dUTP, and hybridization wasvisualized bya chemiluminescence reaction with the DIG luminescent de- tection kit (Boehringer Mannheim). In another series of experiments, the probes were labeledwith[32P]dNTPs by the random primer method(17), andhybridizationwasvisualized by autoradiography. Labeled spots corresponding to fragment(s) ofgenomic DNA and plasmidinserts werecutfrom the membrane, the radioactivity was counted in a liquid scintillation counter, and the results wereused to construct a calibration curve to determine the number ofgene copies per genome.

RESULTS AND DISCUSSION

To address the question offunctional differences within a nucleus between sequence variants ofhistone Hi, we have

earliercharacterized the Hi subtypes in larvae of C. thummi.

Seven Hi fractions obtained by two-dimensional electrophoresis were analyzed by comparing the peptide patterns after chemical and enzymatic cleavage and by isolation of their folded central domains. From these results (9) it appeared that most and probably all fractions were sequence variants ofHi. However, HPLC of the Hi histones and their peptides and microsequencing of selected peptides have recently shown that there are altogether four different Hi histones in C. thummi larvae (E.S. and J. R. Wisniewski, unpublished results).

This analysis became possible when the Hi genes of C.

thummi had been sequenced (see below). Peptides that contained a putative amino acid sequence specific for the gene products of the individual Hi genes were isolated and sequenced. The results revealed the identityand thedegree of homogeneity of the individual electrophoretic Hi fractions. Three of them (Hi I-1, Hl II-1, and Hi 111-1) were found to represent single polypeptide chains, two (11-2 and 111-2) were a mixture of a fourth polypeptide (which we denominate 11-2) and a degradation productofII-1,while the remaining fractions represented degraded Hi polypeptides (E.S. and J. R. Wisniewski, unpublished results).

The four Hi histones were regularly detectedinindividual larvae from inbred strains by small-scale two-dimensional electrophoresis (H. Leufgenand U.G., unpublished results).

Therefore, their appearanceisprobably not due to anallelic polymorphism.

According to the number of Hi variants, fourtypes ofHi genes werefoundin C. thummi (see below).AsHivariant I-1 exhibiteda divergent protein structure (9) and wasdetected only in a minority of chromosome bands (8), we were especially interestedinstructural propertiesthat arepeculiar to Hi I-1 and that may be involved in specific functions.

Histone Hi 1-1 Contains an Inserted

Sequence

Motif. We have compared the primary structure of Hi I-1 to the structures of the other Hi variants. For this purpose, the sequences of genes encodingthe Hi histones I-1 and II-1 of both subspecies and II-2 and III-1 of C. th. piger were determined. Clones encodingsingleaminoacid substitutions of variant II-1 were also sequenced inpart.The individual Hi genes were identifiedbycomparing theirdeduced amino acid sequences to sequences ofsuitable peptides of the Hi proteins. The deduced amino acid sequences of the N-terminal domains and partofthe centraldomains have beenalignedin Fig. 1 and show aconspicuous feature ofhistoneHi variant I-1. In both subspecies,theN-terminal domain of thisprotein differs from thehomologousdomains of the other Hi histones by a sequence thatconsistsoftwo parts: (i) the insertion of arepeat of thesequenceLys-Ala-Pro, whichis reiterated five times (three direct and two modified repeats) in the C. th.

piger and fourtimes(twodirect andtwomodifiedrepeats)in the C. th.thummi protein; and(ii) thesequenceLys-Val-Ala that lies 10aminoacid residues apartfrom the Lys-Ala-Pro repeat andreplacesasingleLys-Ala-Pro that is characteristic of the otherHi proteins ofC. thummi atthe same position (Fig. 1). An additionalHi geneofC. th. thummi(13), which we have identified as Hi III-1, also lacks the Lys-Ala-Pro repeat and contains a single Lys-Ala-ProattheLys-Val-Ala site (13). Thebipartite Lys-Ala-Pro-Lys-Ala-Pro-Xaa1o-Lys- Val-Ala motif is thus specific for variant Hi I-1 and is not present in the other Hi proteins of C. thummi.

While the sequence Lys-Ala-Pro is rarein histones anda data bank search has not revealed other known proteins containingthesequence Lys-Ala-Pro-Lys-Ala-Pro, we have identified thebipartite

Lys-Ala-Pro-Lys-Ala-Pro-Xaa1o-Lys-

Val-Ala motif in one of the three published Hi histone sequencesofthenematode Caenorhabditis

elegans

(18-20).

Here the motif lies at

exactly

the same distance as in ChironomusHifromtheconserved sequences in the central

(3)

Hi I-1(p) I-1(t) il-i(t) II-1(p) II-1(p) II-1(p) II-1(p) II-2(p) III-1(p)

clone Li 234 T31 TL4 Li45 L91 L92 L95 Li254 Li413

SDPAPEIEAPVEAAPVASPPKGKKEKAPKAPKAPKSPKAEKPKSDKPKKPKVAPTHPPVSEMVVNAVTTL SDPAPEIEAPVEAAPVASPPKGKKE KAPKAPKSPKAEKPKSDKPKKPKVAPTHPPVSEMVVNAVTTL SDPAVEV APTT PVASPAKAKKEK KPKTDKVKKPKAPRTHPPVSEMVVNAVKTL SDPAVEV APTT PVASPAKAKKEK KPKTDKPKKPKAPRTHPPVSEMVVNAVKTL SDPAVEV ARTT PVASPAKAKKEK KPKTDKPKKPKAPRTHPPVSEMVVNAVKTL

...VAPTT PVASPAKAKKEK KPKTDKVKKPKAPRTHPPVSEMVVNAVKTL SDPAVEV APTT PVASPAKAKKEK KPKTDKPKKPKAPRTHPPVSEMVVNAVKSL SDSAVDV APTT PVASPAKAKKEK KPKTDKPTKPKAPKTHPPVSEMVVNAVKTL SDPAIEV APV PVASPAKAKKEK KPKSDKPKKPKAPRTHPPVSDMIVNAIKTL

FIG. 1. The histoneHi variantsof C. thummi larvae. The derived amino acid sequences of the N-terminaldomainsand adjacentportions ofthe globular domains are shown in alignment. Variants of the two subspecies C. th. thummi and C. th. piger are marked by (t) and (p), respectively. The inserted Lys-Ala-Pro (KAP) repeats, the sequence KAP, the sequence Lys-Val-Ala (KVA) that is thought to form part of a bipartite motifKAPKAPX,oKVA, and themethionine residues areunderlined.Dots indicate an unknown sequence, and the arrowindicates the border between the N-terminal and globular domains.

domain of the molecule (Fig. 2). Its two parts are also separated by 10amino acid residues, but these show a very low degree of homology between Chironomus and Cae- norhabditis. Again, the Lys-Val-Ala part of the motif is replaced by Lys-Ala-Pro in the otherHi sequences of this organism (Fig. 2). Interestingly, the Lys-Ala-Pro repeat is also found inserted, at almost the same distance from a Lys-Ala-Ala sequence and from the conserved region in the centraldomain,in the deduced amino acid sequence of one of the two known Hi genes of the algaV. carteri (21). Here it has the sequence Lys-Ala-Pro-Lys-Gln-Pro-Lys-Ala-Pro- Lys-Ala-Pro and is flanked by additional Lys-Ala-Pro ele- ments(Fig. 2). Again, the secondHi gene in V. carteri that has beensequenced does notcontain a repeatof Lys-Ala-Pro (21). We havethus identified aprotein motif that has been conserved in evolution in very distant organisms and that appears in only one of their Hi histones. As the N-terminal domains ofHi histones exhibitahigh degreeofinterspecific variability (22), the presence of the Lys-Ala-Pro-Lys-Ala- Pro-Xaa1o-Lys-Val-Ala motif at aspecific position in one of theHi histones ofamidge, anematode,andagreenalga is nottrivial and may suggest a functional role of this sequence.

Competitionof theN-Terminal Domainwith Hoechst33258 for DNA Binding. In Chironomus the Hi variant with the Lys-Ala-Pro-Lys-Ala-Pro-Xaa1o-Lys-Val-Ala motif is tar- getedto aminorityofinterphasechromosome bands. Many of the bands thatcontaintheHi variant I-1 haveproperties ofhighlycondensed chromatin(8).Apossiblefunction of the motifthuscould betobind variantI-1tospecificchromatin sites or to contribute topackaging a condensed subtypeof chromatin, or both. We have asked whether the amino terminus of thisHi proteininteractswith DNA inaway that isspecificforits structure and therefore havecomparedthe bindingtoDNAof theN-terminal

peptides

ofHi I-1 and Hi II-1incompetitionwith Hoechst 33258. Thisdrugbinds in the minor groove ofthe DNA double helix in a selective and well-characterizedway(23,24).Itcompetesfor DNA

binding

with an N-terminalpeptideofseaurchin Hi (25).HiI-1 and

24 VAIV

Hi II-1 from larvae of C. th. thummi were cleaved with BrCN, and their N-terminal peptides (amino acid residues 1-59 and 1-45, respectively; cf. Fig. 1) were used in the experiment. The N-terminal peptide ofHi I-1 from C. th.

piger (amino acid residues 1-62) was obtained in larger amountby expression in an E. coli expression system that usesbacteriophage T7 RNA polymerase (16). The peptides were added to DNA that had been preincubated with Hoechst 33258 in a concentration range where the dye binds specifi- cally to the minor groove (25, 26). BothHi I-1peptidesthat contain the

Lys-Ala-Pro-Lys-Ala-Pro-Xaa1o-Lys-Val-Ala

motif were found to compete with Hoechst 33258forbinding toDNA(Fig. 3). In contrast, the homologous BrCNpeptide of histone Hi II-1 that does notcontain the bipartite Lys- Ala-Pro motif but otherwise has a very similarsequence (cf.

Fig. 1) did not compete with Hoechst33258for DNAbinding under theseconditions (Fig. 3). The synthetic peptide Lys- Ala-Pro-Lys-Ala-Pro-Lys-Ala-Pro-Lys-Ser-Pro-Lys-Ala also showed little competition with Hoechst 33258 (not shown). We presume that the N-terminal domain ofHi I-1 binds to the minor groove of the DNA double helixinvivo, whereas the homologousdomains of the otherHivariants do not, and that this binding is mediated by the Lys-Ala-Pro- Lys-Ala-Pro-Xaa1o-Lys-Val-Ala motif. The Lys-Ala-Pro- Lys-Ala-Propeptide,ontheotherhand,maynotbesuffi'cient but mayfunction inaspecificneighborhoodthat mayrequire the Lys-Val-Ala sequence at a defined distance and con- servedstructure(s) in the central domain of theHimolecule.

The

Lys-Ala-Pro-Lys-Ala-Pro-Xaa1o-Lys-Val-Ala

motif in theN-terminal domain is located in immediateneighborhood to thecentralglobular domain of theHimolecule(Fig. 2).As theglobulardomain bindstothe nucleosome(27),the motif mayinteract withastretchof linkerDNAclosetowhere it enters the histone octamer. This interaction could help to establishacondensed subtype ofchromatin. Weare in the courseofinvestigatingthe site and mode of interaction with DNA invitroandbymicroinjection into polytene nuclei.

'

A

P

ADV

CPVA1 Wt 9 A----8---

S 21 28

KAm^AmAmarKJAAA ATAl---M-

K K KAP M

KAPKAKAPKQPKAPKAPKEPKAPKE-KAA -M

A T2

---K

M-

1S KAPKAP

¹⁰ KVA ^M^-

S 15 KAKAAK- 10 KAP

T1

FIG. 2. AlignmentoftheKAPKAPX,oKVAmotifs insingleHi variantsof C. thummi(row Al),Volvox carteri(row B3), and C.elegans

(row C5).The structurescommonto the other Hi variants of C. thummi(row A2)and Caenorhabditis(rowC6) and the structure of the other known Hi variant of Volvox(row B4)areshown forcomparison.Theadjacentportionsof theglobulardomains are boxed.Numbering refers tothe aminoacid residues betweentheN-terminalserineandalanine,respectively,theKAPKAPX,OKVAmotif, and conserved residues in theglobulardomain.The Caenorhabditisdata is taken from refs. 18-20,andthe Volvoxdata,from ref. 21.

A

2

B

4 s

C

6

(4)

600 800

FIG. 3. Competition ofHi peptides containing the Lys-Ala-Pro- Lys-Ala-Pro-Xaalo-Lys-Val-Ala (KAPKAPX1oKVA) motif with Hoechst 33258 forDNAbinding.Increasingamountsof N-terminal peptides ofHi I-1 (residues 1-59) fromC. th. thummi larvae (v), bacterially expressedHi I-1(residues1-62)fromC. th.piger(0),and Hi II-1 (residues 1-45) from C. th. thummi (s)wereaddedtoDNA preincubated with the drug. Thetruefluorescence, F,oftheDNA- dye complex wascalculated from the measured fluorescence. Its decrease after theaddition of peptide is expressedinpercentof Fof thecomplexwithout peptide.

AnN-terminalpeptide of Hi fromseaurchinspermthat is 48 amino acid residues long and contains repeats of the protein motifSer-Pro-Lys-Lys, has been reportedtocompete withHoechst33258 for DNA binding (25). It is interestingto notethatthe directSer-Pro-Lys-LysrepeatsinthespermHi histonesfrom three differentseaurchinspecies (22, 25) lieat exactly the same distance from the conserved methionine residue^as dotheLys-Ala-Prorepeats inChironomus, Cae- norhabditis, and Volvox. Therefore, it istemptingto specu-

late that both types ofrepeats are part ofa structure that

A

kbp

1 2 3 4 5 6 7

4.0 -

2.7 4do,0

1.8

brings about a condensed subtype of chromatin, such as thosetypical for total sperm chromatin and for those chro- mosomesitesinChironomus where Hi I-1 islocated.

Numbers of Genomic Copies of the Different Hi Genes.

Probes specific for Hi I-1 genes and for the group of the other Hi genes,respectively, were used to estimate the numbers of copiespresent per genome. Surprisingly,the numbers were found to be dramatically different. The results for C. th.

thummi are shown in Fig. 4 A and B. The Hi I-i-specific probe 2hybridizedtotheHi I-1genein plasmid and genomic DNA(Fig. 4A, lanes 1-4) but not to other Hi genes (Fig. 4A, lanes 4-7). The hybridization signal of genomicDNA (lane 4) appeared about half as strong as that of a plasmid insert equivalent to two gene copies per genome (lane 3). On the otherhand, probe 3 that is specific for the other Hi genes hybridizedtothe Hi II-1 genein genomic and plasmidDNA (Fig. 4B,lanes 4-8) butnot toHi I-1 (Fig.4B,lanes 1-3).The hybridization signal yielded by genomic DNA (lane4) appeared to be similar to that obtained with an equivalent to 40 genecopies(lane7).Weconclude that thegenome ofC. th.

thummiprobably containsone copyof the Hi I-1 gene and about 40copies of the other Hi genes.

Withgenomic DNA of C. th.piger,probe1thatis specific forthe Hi I-1 gene hybridized to twofragmentsof 2.5 and 1.0 kbp (Fig. 4C, lanes 1 and 2). Inbothof them, the hybridiza- tionsignalsappearedtobe similartothoseequivalentto one and two genecopies (Fig. 4C, lanes 3 and 4). Aquantitative determination of the amounts of probes hybridized to the blots wasmadebyidentifying the spots onautoradiograms, cuttingthem from themembrane, andcountingthe radioactivity hybridized to genomic and to standard amounts of plasmid inserts in ascintillation counter. The 2.5-kbpfrag- mentsfromlanes3-7of theautoradiogram shown inFig.4C yieldedalinearcalibrationcurve(notshown) for calculation of the amountsofHi I-1 genecopies in genomicDNA(Fig.

4C, lanes 1 and 2). Linear calibration curves (not shown) were also obtained in three independent experiments with geneprobe 3atlow stringency in which the total number of B ¹ ² 3 ⁴ 5 6 7 8

C1

2 3 4 5 6 7

_ ^la a*0 -2.5

a6Y

a a .-:

-1.0

FIG. 4. Copynumbersper genomeof thetwotypesof Hi genes.Genomic DNAof C. th.thummi(AandB,lanes4)andC. th.piger (C, lanes1and2)^wascleaved withrestrictionenzymesAccIandHindIII, respectively,andseparatedonagarosegelinparallelwithdigested plasmid DNAcontainingacompleteHi I-1gene(AandB,lanes1-3; C,lanes3-7)oracompleteHi II-1gene(Aand B, lanes5-8).Theamountsof plasmid DNAwerechosentobe equivalenttoincreasingnumbers ofgenecopies per genome.ThegelswereblottedontoHybond-Nand hybridizedwithgeneprobes 1and 2 for Hi I-1(AandC)orwithgeneprobe3 that isspecificfor the other Hi variants(B). (A)Probe 2hybridized toa4-kbp genomic fragment (lane 4)andtoaplasmidHi I-1insert ofcorrespondingsize(lanes 1-3). (B)Probe 3hybridizedtoa1.8-kbp genomic fragment (lane4)andtoaplasmidHi II-1insert ofcorrespondingsize(lanes 5-8).(C)Probe 1hybridizedtotwogenomicDNAfragmentsof 2.5 and 1.0kbp (lanes1and2)andtoa2.5-kbp plasmidHi I-1insert(lanes 3-7).FivemicrogramsofgenomicDNAwasappliedin A andB, and2.5ug,inC.In A andB,theplasmidDNAamountswereequivalentto10,4,and 2genecopiesper genomein lanes 1-3,respectively,and to 10, 20, 40,and 80genecopiesin lanes 5-8. InC,theplasmidDNAamountsinlanes 3-7correspondedto1, 2,5, 10,and 15gene copies,

respectively.

100-

80 -

60 -

UL:

40 -

20 -

0 _.I

o 200 400

Peptide, nM

0

(5)

Table 1. Numbers ofHigenecopies pergenome in C. th. piger

AllHitypes* HiI-lt

blot cpm genes lane cpm genes

1 536 50 1 45 0.9

2 3272 58 2 50 1.7

3 3292 61

Radioactivity of the gene probe hybridized to genomic DNA fragment of appropriate size was measured. Gene copy numbers weredetermined by calibration curves derived from standards on the sameblots.

*Independently labeled probes.

tSamplesfromoneblot.

Hi genes was determined, with plasmid DNAcontaining a copyof the Hl II-2gene as astandard.Theresults areshown in Table 1. The genome of C. th. piger thus contains about 50-60 copies ofthe one group of Hi genes and very few copies of the Hi I-1 gene. The numbers obtained from 2.5-kbp fragments of genomic DNA, 0.9 and 1.7 (Table 1), mustprobably bedoubledbecauseanother1.0-kbpfragment ofgenomic DNA showedhybridizationwiththe HiI-1 probe of corresponding intensity (Fig. 4C, lanes 1 and 2). The genomic fragments of 2.5and 1.0kbp may both contain anHi I-1 gene with a specific genomicorganization.Therefore, we assumethatC.th.piger contains two to four copies of theHi I-1 gene. In situ hybridization has revealed that in both subspecies ofC. thummi, the Hi genes of the I-1 type are located in another chromosomethan are the other 40-60Hi gene copies in the genome (L.T. and B.S., unpublished results).

HistoneHi variants containing theLys-Ala-Pro-Lys-Ala- Pro-Xaa1o-Lys-Val-Ala motifmay be present also in other organisms.The fact that C. th. thummi and C. th.pigerhave onecopyandafewcopies, respectively,oftheHi I-1gene indicates the necessitytosearchfor rarecopiesof variantHi genes.Preliminaryexperimentswith DNA from a numberof plant and animalspeciesinwhicholigonucleotidescoding for the Lys-Ala-Prorepeat were used asprobeshavenotbeen successful.Itshould bepossible, however,toelicitantibod- iesagainstthe motifand to use themto detect Hi histonesof thetype foundinChironomus, Caenorhabditis,andVolvox.

E.S. and L.T. havecontributedequallytothis work andarelisted inalphabetical order. Wethank Dr. J. R. Wisniewski forhelpand advice and Dr. B.Schmidtforsequencing peptidesand forsynthe- sizingthe KAPpeptide. We aregrateful to Drs. A. Lindauer, K.

Muller, and R. Schmitt for the sharing of unpublished data and for permittancetocite their results prior topublication. We also thank Ms. I.Streichhan and Mr.Arnd Steuernagelfor helpwith thefigures and Ms. K. Wedekind for expert secretarial assistance. Apredoc- toral fellowship award to L.T. from the Fonds der chemischen Industrieis gratefullyacknowledged. This work was supported by a grantfrom DeutscheForschungsgemeinschaft to U.G. (Gr 376/11-1).

1. Thoma, F., Koller, T. & Klug, A. (1979) J. Cell Biol. 83, 403-427.

2. Patient, R. K. & Allan, J. (1989) Curr. Opin. Cell Biol. 1, 454-459.

3. Grunstein, M. (1990)Trends Genet. 6, 395-400.

4. Ericsson, C., Grossbach, U., Bjorkroth, B. & Daneholt, B.

(1990) Cell 60, 73-83.

5. Cole, R.D. (1984) Anal. Biochem. 136, 24-30.

6. Cole, R. D. (1987) Int. J. Peptide Protein Res. 30, 433-449.

7. Wu, R. S., Panusz, H. T., Hatch, C. L. & Bonner, W. M.

(1986) CRC Crit. Rev. Biochem. 20, 201-263.

8. Mohr, E., Trieschmann, L. &Grossbach, U. (1989) Proc. Natl.

Acad. Sci. USA 86, 9308-9312.

9. Hoyer-Fender, S. & Grossbach, U. (1988)Eur.J. Biochem.

176,139-152.

10. Keyl, H.-G. &Pelling,C. (1963)Chromosoma 14,347-359.

11. Keyl,H.-G.(1965)Chromosoma 17, 139-180.

12. Schmidt, E. R. (1984) J. Mol.Biol. 178, 1-15.

13. Hankeln,T.&Schmidt, E. R. (1991)Chromosoma 101, 25-31.

14. Lifton, R. P., Goldberg, M. L., Karp, R. W. & Hogness, D. S.

(1978)Cold Spring Harbor Symp. Quant. Biol. 42, 1047-1051.

15. Sanger, F., Nicklen, S. & Coulson, A. R. (1977)Proc.Natl.

Acad. Sci. USA 74, 5463-5467.

16. Rosenberg, A. H., Lade, B. N., Chui, D.,Lin,S., Dunn, J. J.

&Studier,F. W.(1987) Gene 56, 125-135.

17. Feinberg, A. P. & Vogelstein, B. (1983)Anal. Biochem. 132, 6-13.

18. Vanfleteren, J. R., Van Bun, S. M. & Van Beeumen, J. J.

(1988)Biochem. J. 255, 647-652.

19. Vanfleteren, J. R., Van Bun, S. M., De Baere, I. & Van Beeumen, J.J. (1990) Biochem. J. 265, 739-746.

20. Sanicola, M.,Ward,S., Childs, G. & Emmons, S. W.(1990)J.

Mol. Biol. 212, 259-268.

21. Lindauer, A.,Muller,K.&Schmitt,R.(1993)Gene, in press.

22. Wells, D. &McBride,C.(1989)Nucleic AcidsRes. 17,Suppl., r311-r346.

23.

Pjura,

P. E., Grzeskowiak, K. & Dickerson, R. E. (1987)J.

Mol. Biol. 197,257-271.

24. Teng, M., Usman, N., Frederick, C. A. & Wang, A. H.-J.

(1988) Nucleic AcidsRes. 16,2671-2690.

25. Suzuki,M.(1989)EMBO J.8,797-804.

26. Bontemps, J., Houssier, C. & Fredericq, E. (1975) Nucleic Acids Res. 2, 971-984.

27. Allan, J., Hartman, P. G.,Crane-Robinson,C.&Aviles,F. X.

(1980)Nature(London)288,675-679.