Different regions of a complex satellite DNA vary in size and sequence of the repeating unit

J. .Moz. Biol.

Different

(1979) 135, 483-500

Regions of a Complex Satellite DNA Vary in Size and Sequence of the Repeating Unit MARIAN

CARLSONT~ ASL) DOUGLAS

Depurtment

oj’ Biochemistry

Stanford University Stanford, Calif.

(Ileceived

BRVTLAU

School

of Medicine

94305, U.S.A.

20 June 1979)

The I-ti88 g/cm” satellite DNA of l~rosophila melanogaster is composed primarily of 359 base-pair units repeated in t,andem. Most of these units contain a single rlr!a,vnge sit,e for both Ha+?111 and H&f1 restriction endonucleases; however, some units lack one or both sites. Previously we had shown that the distribution of HaeIlI and HinfI endonuclease sit)rs varies widely between different regions of I.688 g/cm3 satellite DNA; for example, some regions contain Hue111 sites in c’\-cry unit and other regions (> 10,000 base-pairs) contain no HaeIII sites (Carlson & Brutlag, 1977). We have now cloned molecules of I.688 g/cm3 satellite DNA4 wllicll lack HaeIII sites and have shown that the absence of sites is caused by sequence variation rather than haso modification. This result indicates that regions of 1.688 g/cm3 satellite DNA with different, dist,ributions of rest,rict,icnl sites differ in the sequence of their repeating unit,s. We also show that a large fraction of the satellite DNB which is not cleaved by HaeIII endonuclease still contains H&f1 endonuclease sites (and AZuI sites) spaced about 359 base-pairs apart. However, one cloned segment, lacking HaeIII sites was found to contain 33 tandem copies of a novel 254 base-pair unit. Sequence analysis showed that tIlis 254 base-pair unit is homologous to t,he 359 repeat except for a 98 base-pair deletion. These data suggest that both units have evolved from a common ancestor and t,hat each has subsequently becorno amplified into separate t,andem arrays.

1. Introduction ‘I’h(b genome of DrosophiZa melanogaster contains several highly repeated satellite DIVAS, which comprise about 20% of the genome and are primarily located within centromeric heterochromatin. Analysis of the sequence organization of these satellite DNAs is essential for understanding their evolution, and will perhaps provide clues t)o their function. We have studied the 1.688 g/cm3 satellite (1.688 satellite) DNA, which is mainly composed of bandem repeats of a 359 base-pair unit (Hsieh & Brutlag, 1979). Previous studies have shown that although many 359 base-pair units contain a recognition site for both Hue111 a.nd HinfI endonucleases, some units lack one or both sites (Manteuil et al., 1975; Shen et al., 1976). Digestion of I.688 satellite DNA with either enzyme generates monomers, dimers, trimers, and so on, of the 359 basepair unit. as well as very large fragments (> 10 x lo3 base-pairs) resistant to cleavage. In a previous report we presented evidence that 1.688 satellite DNA includes regions t L’rcBsrnt, address: Department’ MtLSS. 021:19, U.S.,\.

of Biology,

Massachuset,ts

.Institut,o of Technology,

Cambridgta>

4x3 00~~~2X36/i9/a40483-18

SOZ.OO/O

;(‘; 1979 Academic

Press Inc. (London)

Ltfl.

484

M. UARLSOX

dNU

I). BlCUTLA(:

with different distributions of HaeIII endonuclease sites. Some regions contain few or no HaeIII endonuclease cleavage sites, while most regions contain sites almost, 1977). Because the absence of a recognition every 359 base-pairs (Carlson & Brutlag, site could result from either sequence alteration or base modification oftht: recognition sequence, it was unclear whether these different regions of 1.688 satellite DNA differ in nucleotide sequence or in degree of base modification. Tn this work we have investigated the absence of HaeIII endonuclease sites in f)ot,h bhe dimer fragments and long HaeIII-resistant fragments produced from 1.688 satellite DNA by HaeIII endonuclease cleava,gc. Using molecular cloning. we have shown that nucleotide sequence, not base modification,‘is responsible for t,he lack of HaeIII

sites.

Hence,

regions

of 1.688

satellite

DNA

with

different,

disbrihutions

of

HaeIII endonuclease sites differ in sequence. We have chara&erizcd 1.688 satellite DNA regions which lack HaeIII sites to determine bheir relationship to the bulk of the 1.688 satellite DNA, which contains bot,h HaeIII and HinfJ endonuclease sites almost every 359 base-pairs. Much of the HaelIL-resistant DNA wa;s shown to 1~ composed of repeated units approximately 359 base-pairs in size and to contain regularly spaced H&f1 endonucleasc sites. Hoxxver, analysis of cloned segments of HaeIII-resistant satellite DNA led to the discovery of a novel 1.688 satellite DNA component with repeating units of 254 base-pairs. Sequence analysis of the 254 basepair unit showed that it is homologous to all but 98 base-pairs of the 359 base-pail unit,. These data suggest tha.t bot’h repeat units evolved from a common ancestor and that

each was subsequently

and 254 base-pair

amplified

to generate

separate

tandem

arrays

of the 359

repeats.

2. Materials and Methods (a) Preparation The

of D,NsLt

l-688

sat,ellite DNA lvas purified from D. ~w,eZanogaster ernbryvuic Illlcltti (Csrlsoll & Brutlag, 1977). Analytical centrifugation showed that 360/, of the recovered DNA had a buoyant density of 1.6885 g/cm3, and the remaining 64O/e formed a broad peak with a density of 1.699 g/cm3. The DNA sample used for the experiment shown in Pig. 1 was further purified by contrifugat,ion in cesium formate and is the same preparation dcscribed before (Carlson & Brutlag, 1977). Dimer fragments and long HaeIII-resistant fragments were purified from l-688 satellite DNA by digestion to completion w&h HaeIII endonuclease. The products were then fractionat,ed by eloctrophoresis in a 1 */0 agarose gel. The dimer fragments and the HaeIII-resistant fragments migrating at, the limit,ing mobility were eluted from the gel, as described below. The HaeIII-resistant fragments comparison were estimated to be larger than 10x IO3 base-pairs by gel electrophoretic wit’h size standards. Procedures for isolation of closed circular plasmid DNAs have been described (Carlson & Brutlag, 1977). (b) Enzywaes and ?.eactio?rs Terminal deoxynuoleotidyl transferase (Chang & Bollum, 1971), HaeIiI endonuclease (Roberts et al., 1975) EcoRI endonuclease (Modrich & Zabel, 1976) and polynucleotide kinase (Richardson, 1965) were gifts of R. Ratliff, D. Charney, P. Modrich and J. Chien, respectively. Escherichia coli polymerase I was purified as described by Jovin et al. (1969) and H&f1 endonuclease was prepared by BioGel A-0.5 M, phosphocellulose and DEAEcellulose chromatography, as described by R. a. Roberts (personal communication). Ah1 and HaeIII endonuoleases were purchased from New England Biolabs. Bacterial alkaline phosphatase was purchased from Worthington and further purified by DEAEcellulose chromatography by M. Goldberg. Reaction conditions for restriction ondonuclease digestion have been described (Carlson & Brutlag, 1977).

sl~:Q~Ii:I1’CE

VARI~~TIONS

IN

SATELT,ITE

4x5

UNrl

(c) Bacteria

h’. coli strains (pSClO1) ((‘ohen

HBlOl (ye&, hsr, ham) (Boyer & Roulland-Dussoix, et al., 1973) were obtained from D. S. Hogness.

1969) and

C6i)O

Hybrid ~nolecul~s were constructed by the poly(dA) . poly(dT) joining methods of isrutlag et wl. (1977). For preparation of pSClOB-dimor Ilybrid molecules, purified 1.688 using t)rrminal deoxyrlucleoticlylsatellite dimer fraprnont,s were extended wit,11 poly(dA),, transferaso. D. Kemp generously provitled UanaHI endonuclease-cleaved pSClO5 DN?I llad been added, and these 2 DNA samples wert) (Cohen et ctl., 1973) to which poly(dT),, allowed to anneal. Similarly, purified HaeIlI-resistant fragments and EcoRI-cleaved respecti\,cly. Transformation was t&C101 \vcre ost,tlnded with poly(dT),, and poly(dA),,, 19ii). Plates containing kanamycin (,fYectcd as described before (Ca,rlson & Brutlag, \vern used to select hactcria transformed with (33 pg/tnl) or tetracycline (15 g/ml) and transformed color&s were tested for ttlck pSC!lO5 or pSCl0 1 DNA, respectively, to a modification of tile colony Ilybridizatiorl fjreserrccs of 1.688 sat,ellite DNA according tc~cllniq~~a of Grunstein & Hogness (1975) (J. Lis and L. PrcLstidge, personal commurricat iota). ‘I’tlr: hybridization probe used irt this procedure \vas prepared as follows. A rcconlt)inallt plasmid corLtaining 1.688 satellite DNA inserted into a vector lacking homology t’o ~)ti(.“lOl or l)SC’lO5 n-a,s constructed by joining poly(tld),,-t,rrmirlatrd 1.688 satellite DNA to pcll,y((ll’)-trrrnillated ColEl DNA (a gift, from (:. Kubin). Selection for transformcxd c,olorlies \vus accomplished by addit,ioll of colicin E 1 (from D. Firmrga~n) to transformc~tl t)actcxriu before plating. The pla,smid ~Dm688~,~~, r9 colltaininp an insnrtecl segment of I.688 +ttellit~t~ DN&l. was recovered. 32P-labe10tl cDm688.52 DNA (5 >: 107 cts/min per pg) was ~~r~parcci t)J. nick-translation with E. cold DNA polpmcrasc~ I, losing [32P]d(:TP (200 C’i/ tnmol : Amr~rstram) as labeled substrat o, according to t)l~o met,hod of Rigby et cd. (I 977). \\‘ork itlvc>lving rc!combinant plasmids \vas czarriad ollt) under t,tlo 1’2 lovcl of pltysical ~~orltuiunlr~nt~ ;I~KL Ehll level of biologic~al containnlctlt, as spccifiecl in the Na.tionwl I test ittltc%s of Hclalt II Guidelines on Recombilrant~ DN;1 R<~sc~nrctl (.~~uII: 1976). (t:) Gel electro~ho~esis

am/

d~itiota

of IlX;.-l

,fro7)h gels

.\garosc gvl c~lc~ctrophoresis has born tlescribod (CarlsoIL & Brrltlag, 1977). Polyacrylz~mitlr gels ~vcro prepared in TBE buffer (89 mlr-Tris. OH, 89 m>r-boric arid, 2.5 rn>I-EDTA, 196 ,v,,v’-methylellopH 8.3; l?()acr)ck & Dingnan, 1968) from a stock of 2096 acrylamido, t)isacrylamidc. Analysis of restriction enzyme cleavage pattcrrts was carried oat on 3”{, 10 i"?' linear polyacrylamide gradient slab gals (14 cm x 11 cm x 0.16 cln) by eloctrolrlr~,r;~~is at 100 V for 3 11. For preparation of DNA fragments, polyacrylamide slab gels (23 cm )’ 14 cm >’ 0.16 cm, or 23 rm x 14 cm x 0.3 cm) \rere used, and electrophorc& \Y~H at 60 V for 12 11. DNA fragments were recovered from grls by r>lectrophoretic elutiorl ill cittlcr THE: huffor or 0.05 X ‘I’RK bufY+r at, 50 V ox-c>rnigllt. DNA Was elect~rophorc~snd illto dialysis t)llbing, and the polarity 1~~s rcvcrsed for 1 min prior to reccn-rry of DNA frown ttlo bag. Residual othidium bromide was somet’imes removed by chromatography on l)owc~x A(:5OM:-X8. The DNA was thon pmcipit,atecl \ritll otllarrol klrltl resuspended in TE: t>nff
‘I’ho pattt,rn of DNA fragments was transferrod from gels to nitrocellulose shr,et,s ac>cording to Sonttlern (1975). Sheets wore pre-inclzbated soveral hours at, 65”C, esccpt whr>re noted, in O.OZo/o bovine serum albumin, 0.02% Ficoll (Pharmacia), 0.02?;, poly\.inyl pyrrolictonc. 5 x SSCP (0.6 >I-NaCI, 0.075 Ilr-sodium citrate. 0.1 ni-sodium phosphate. t)H 5) and ().I’!, sodium dodecyl sulfate (Denlrardt, 1966). Hybridizat,ion with 32P-labcl(~d (%Dm688.52, prepared as above, was carrird out for 12 t,o 48 h under tllo same conditions as t’hn pre-incubation, except t’hat carrier salmon sperm DNA (100 &ml) and poly(A) ( 100 &ml) were also included. The nitrncellulose sheet,s were washed wit,h shaking 3 times in pre-incubation buffer at t,he hybridization temperature ant1 t’hon twice in 2 x FM’ (SW is 0.15 Ir-Na( ‘I, 0.015 M-sodilun cratlml; oactl waslr was for

30 Inill.

486

Jf.

CAI%LSON

a\KI)

(g ) Electron

I). UHUTL~Y,U wicroxcopy

DNA \vas prepared for electron microscopy by tile basic protein film teclluiquc! A\-it I c&her 409; or 70 96 formamide (Davis et al., 197 1). Het’eroduplexes \vere formed by tlrcl procedure of Davis et al. (1971), except that rcnat uratiorr \vas carried out if1 rithcr 50”,, formamide at, 25°C or 45Sb formamide at 37 C’.

Por determination of nucleotide seq~wnwx ndjacerlt to -Jlrrl c~utlonuclcast~ clt~ava~:.(~ sites, pDm688.169 (20 pg) wa,s digttsted wit11 *I//( I c~~do~~~lcloitse, tllw prwipitatecl wit11 ct,hanol. The rcsnlting fra.gmrnts wcrc dcpl~osl,l~or~latctl by itlcnbut~ion \vith bacterial alkaline phosphatasc ill 20 rn31-Tris.HCl (pH 8) at. 6O‘(J for 2 II. and tlrcx clrzy~,l~ \v?~s tllrylt removed wit,h 4 phenol ext,ractiolls. After ctlwr c,xt wet ioll atlct pwcipitatiorl wit11 c~tharrol. the DNA was resuspendc~d in 50 ~1 of 10 mar-‘1%~. HCl (pH 8). 50 m~l.KC’l. 0.1 I~\I-EDT=\. 5’- Phoaphorylation w-as theu carried ollt by atlditiorl of 5 ~1 of 0.1 >I-M&Y,, 80 mw dithiothreitol and 5Oo/, glycerol : 200 pm01 of Jy-‘(2P]ATP ( ~1000 propa,rfltl C’i/lumol, according to Maxam & Gilbert (19i7)): nerd 2 ~1 l,o14-“tlcl(,oticlc~ kinaw ( I.1 units/ml). 7’11(~ reaction was incubated at 37°C for 30 min and trnninatrtl by adtlit>ioll of RLI qua1 \-()I. of 3 M-ammonium acetate. The DNA \vas then prccipitatcd -\zittl c%llanol t\viw in tiler prcsvnw of carrier transfer RNA (20 pg). The lab&d fragments WVI‘(~ wparatetl l)?; electropl ~orvsis 011 a 2% agarose gel, and tbo 254 base-pair monomw fragments were rluted from tlw gc>l as described abort. These fragments were next divested lvith Hin.fI c~~rtot~nclrasc. Tllc 254 base-pair fragments containing a Hip/f1 site pa\-<‘ rist’ to 2 clcsavapo protlllcts (88 and 166 fragmcllts by c~lrdrobase-pairs), which \voro scparatrtl frown tllo ~I~Icl~~a\w~ 354 basepair phoresis on a 5:& polyacrylamide gel aud eh~ted from tlrc, g?;(>l. Fragments suitable for analysis of sqttcnws atljacont to H$rrfI t~~~do~laclcilst~ sitw \v<~Y(> at~d propared as folloxvs. pDm688.169 .DNA ( 150 pg) evils rliprstcwl Lvitlr Hinf I cndo~luclcaw, pel electropliorc~sis. ‘I’lrc: tile result,ing fragments were fractio~latetl t)y 5”b polyacr,vla~t~id~ 254 base-pair fragment of satrllitc DNA co-migrates with a fragment of pSC101 DNA. arrtl \VHS thell r~stractcd t\vicc \z.itll bot)li fragments xvere c,ll1ted from the gel. Tlrt- DNA 11-br&anol (to remove thca et~liidiuni brornidt> rlsctl i Ii stai nillh (I‘ DNA4 ill tlic* g’cl), atirl loatletl 0II a small DEAE-cellulose (DE52) columrl (0.2 ml) c~clllilibratc~tl \vitll 0.1 ~~-‘lYris. HCl (pH 7.5). After washing the column with this butiw, the DNA ~vas clntc>tl \vitll 0.7 31.NaCl in 0.1 \I‘Tris.HCl (pH 7.5). Closed circular carrier DNA (12 [Lg.) xvas atldcd. alld tilts DNA \I:W iuid 5’-pliosp1~0rylatio~~ react,ions w(‘r(s precipitated with ctllanol. Dephospllorylat~io~~ nsillg 100 pn101 1y-32P]ATP (2500 Ci/mmol : carried out essentially as described abow, all,1 ,~lcctlYlICN). The labeled fragments were t I,(‘1 L digest WI \vit I> .4///l (~llch,lluClcwC plloresed on a 5% polyacrylamido gc,L. All 4 fragments (2 satrllito DNA fragments aucl -.) \-cdor DNA fraguw Its) \v(~rc’ IW~\~~~IYYI. \vitli chromatographed on DE:AE-cellulose as tlescrilwtl about>, a11,I t~tl~anol precipitated car&r tRNA (25 we). The satellit,c DNA frag>nc,trts (88 a,ncl 166 basal-pairs) \vcI’(’ itlwrtific,tl 011 the basis of size, but all 4 fragments ww’t sc’c~~~r~~wd t,o cwnfirln tlrc, idcrrtificat’ioll. Chenlical rea&ions and clcct,roplloresis 011 scVlue~icillK gels \+ err carried ollt ac:cordill~ gels, a 12”. t,o Maxam & Gilbert (1977). Irt adclitiort to 20’;,‘, alld 25”b polyacrylarnide gel ill 7 \1-1wea atIc 100 InwTris-borate (pH 8.3), 1 mi\~acrylamidr/O~6°/0 bisacrylamidc EDTA was also employed.

3. Results (a) Sequence variation

rather than base m,odi$cation accounts joy a6setzce of Haelfl endonuclease recoynition sites

Most of the 359 base-pair repeating unibs of’ I.688 satellite Dh’A contain a recognition site for HaeIII endonuclease, but many units lack such a site; consequently, digestion of 1.688 satellite DNA w&h HaelI endonuclease produces not’ only 359 base-pair monomers, but also dimers, trimers, and higher oligomers, including large fragments which migrate with limiting mobility on a gel (Fig. 1). The absence of

HaeiIl

Trim ler

Din 7er

Monon ner 359 bp

HinfI

48X

hl.

CARLBOX

‘lN1)

I).

IILcl!‘I’1,,1(;

HaeIII cleavage sites in the fragments larger than monomer could result from either sequence variation or base modification of HaeTIT recognition sequences. To distinguish between these possibilities, we cloned dimer fragments and fragments migrating with limiting mobility on a gel and then ~~xamincd the cloned DXA for the presence of HaeIII sites. Since DNA propagat,ed in host-specific lnodification-tlHfecti~c, E. coli exhibits no base modification of HaellL sites (Ca,rlson B B&lag, 1977), the presence or absence of Hue111 endonuclease s&es in cloned DNA is d&~rrnined oul> by the nucleotide sequence. Dimer fragments were purified from Hael I 1 (~r~ctonuoleasc~-clr;lvt,tl 1.688 sat)ellitt: DNA and insert,ed i?l vitro into the BarnHI sitti of t,hc kanamycin resistance plasmid vector pSC105 by the poly(dA).poly(dT) J‘oining m&hod, ad described in Materials and Methods. The resulting molecules were used to tt~ansf’orm E. coli HBlOl t’o kanamycin resistance. and transformed colonies uwo then screened for the prcseucc of 1.688 satellite DKA by colony hybridization (Grunsteiu k Hognesw. 1975). Colonies containing 1.688 satellite. DNA carried hybrid plasmids witlr insWet satcllitc din1c.l fragments. called pSClO5-dimer plasmids. Plasmid DNA was prepared from 18 clones c~ont,ainiug I+88 dimtlr units. \\ hich The rxpectf,d s;tructure of \vere recovered from ten independent transformations. these pSC105-dimcr plasmids is shown in Figure 2. To confirm that, t’hc hybrid

I+‘Iu. 2. Expected structure to scale) the expected structure inserted dimer fragment, the the junctions between dimer climer is flanked by poly(clA) the pSC106 segment represent et al., 1976).

of pSC105-dimer pl~~smitls. The diagr;rm sho\vs scht,m;rtioirlly (not, of a pSC1056dimer hybrid plasmid. The single line rrpresmts thr ;tnrl the black regions at, double line represents the pSC105 vector, and pSC106 represent poly(tlA) .poly(tlT). In each DNA strand thr at one junction and pdy(tlT) at the other. The hatched areas along an inverted repeat in the knnamycin rrsistancc elemt=nt (Gayer

plasmids have this structure, four plasmid DNA preparations were cleaved with EcoRI endonuclease, and the resulting fragments were denatured, allowed to anneal in a first-order reaction as described in the legend to Figure 3, and mounted for electron microscopy. The poly(dA) and poly(dT) segments were expected to form a short stem and dimer-sized loop, and the two inverted repeat sequences in t’he kanamycin resistance element (Guyer et al.; 1976) were expect’ed to form a long stem and small loop. Figure 3 shows that such structures were observed in the electron microscope, indicating that the four pSC105-dimer plasmids tested have the structure shown in Figure 2. Each pSClO&dimer DNA sample \vas then digeat)ed with HaeIll endonucleasc to

FIG. 3. The presence of a 1.688 satellite dimer in pSC105-dimer plasmids. Each of 4 pSC105climer plasmid DNAs was cleaved with EcoRI endonuclease to produce 2 fragments. The fragments wert? then denatured with aIkali and neutralized according to the heteroduplex procedure of Davis it nl. (197 1) ; however, renaturation was prevented by chilling the DNB before neutralization and mounting it immediately for electron microscopy by the basic protein film technique, with 40% rormamide (Davis et al., 1971). Single-stranded +X174 DNA was included as a size standard. The ~~lcctron micrograph shows 2 molecules with the structure predicted in the text for an EcoRl fragment in which the poly(dA) and poly(dT) segments flanking the dimer have annealed. The third molecule has the structure predicted for an EcoRI fragment in which the inverted repeat in t hn kannmycin resistance element has annealed.

determine whether the inserted dimer fragment could be cleaved. If the inserted dimer had no Hap111 site, digestion would generate, in addition to fragments derivcbd from the pSC10.5 vector, a single novel fragment containing the inserted dimer flanked by poly(dA)*poly(dT) segments and adjacent vector DNA. On the other hand, if the inserted dimer were cleaved by HaeIII endonuclease, two new fragments would 1~ genera&ed. Each of 18 pSCl(&dimer plasmids gavck rise t.o only one‘ IICM

490

RI. CARLSON

AS11

I). IJRUTLAC:

fragment which was? as expected, slightly larger than a 1%88 dimer fragment isolated from satellite DNA. Figure 4 shows the HueIII digestion products of eight representative plasmids; the size variation of the new fragments containing the inserted dimers is presumably dut: t’o differences in length of the poly(dA)*poly(dT). Similarly, the prrs;rnce of tv,o iie\v fragmenbs, both larger than a dimer fragment, in slot four prol~ably indicaks hctcrogeneit,y in the length of the poly(dA) . poly(dT) segment’s witjhin this preparation of plasmid DNA (Brutlag of ccl.. 1977). I.688 I

2

3

4

5

6

7

8

PSC 105

satellite DNA

Dimer inserts +Dimer

I+ Monomer

Fro. 4. Cloned dirnnr fragments lack HMIII endonuclease sites. Eighteen pSC105-tlimer hybrid plasmids were digested with HrteIII endonucleusc~, antI t,hl* (hgeation products were electrophoresotl in l%‘vO agarost: gels. The gel shown here displays tht. fragments generatetl from 8 representat,ivc pSC105-tlimer plasmids and the vector pSC105. The fragments containing inscvtcxi dimers arc’ indicut,od by a bracket. The fragments gonerated from parti:Jly lmrifirrl I.688 s~ttellitc~ l)NA by HoeI digrst,ion WC also shown, and the monorncr arul
SKQUESCE

VARTATIOSS

rN

S.4TEI,I,I’l-F:

1!) I

J)N:2

111a similar experiment, we also investigated whet,hrr sequence varktion or hascb modification is responsible for the absence of HaplIT recognition sites in the long sat)ellite DNA fragments which migrate ab the limiting mobility on a gel after Hael It digestion. These HueIII-resistant fragments were purified as described in Materi& and Methods and inserted into the EcoRI endonuclease cleavage site of the t’rtracycline resistance plasmid vector pSClO1 by the poly(dA) .poly(dT) joining method. Fourteen clones containing l-688 satellite sequences were recovered from SCVIYI independent transformations, and hybrid plasmid DNA was isolated. When clca~c~l with HaeTII endonuclease, 13 of the 14 plasmids produced one long novel fragment. in addit’ion to the fragments expected from the pSClO1 vector (Fig. 5). One hytktl plasmid (Fig. 5, slot 3) gave rise to two new fragments; the presence of one HnvI 1I DNA fragment from \\.lrich cleavage site indicates either that the 1). rnelanogaster this hybrid plasmid was derived contained a single modified HaelIl site or t,hat thr, initial HaeIII endonuclease digestion was incomplete. Of the 14 hybrid plasmids, six carried inserted segments larger than 3.5 x IO” I)asclpairs (see the legend to Fig. 5). Most of the inserted DNA in the six largest hybrid plasmids was shown to be homologous to 1.688 satellite DNA by its ability to hyhridizc~ to cloned sat’ellite DNA (see Fig. 7). In addition, the inserted DNA in four of t~lrcw plasmids (pDm688.132, pDm688.142, pDm688*169> and pDm688.194) was SIIOWII c+apable of pairing with the cloned satellite DNA in pDm688.23 (Carlson $ Brutlag. 1977) by heteroduplex analysis (data not shown). The HaeIII resistance of these long (:loned satellite DNA segments constitutes evidence that long regions of I.688 satdlitc~ I)Xiz lack HapITT cleavage sites due to sequence variation.

(1)) Homolog~y of HaeIII-re&%rnt

1&!M satellite nNA Hae111 sites

to satellite DNA

confnitt

itug

The long HaeIII-resistant satellite DNA fragments cross-hybridize with clonrd I a688 satellite DNA containing HaeIII, HinfI and two AluI sites every 359 base-pairs (Carlson k Brutlag, 1977), but the sequences of these two components differ at least t,o the extent revealed by the frequency of HueIII sites. As a measure of sequenccx relationship, the distribution of HinfI and Ah1 sites in HueIII-resistant, DNA \I.as oxamined. HueIII-resistant fragments isolated from satellite DNA were digested with HuPIII. HinfI and AluI endonucleases. The digestion products were elect,rophoresed in an agarose gel (Pig. 6(a)), transferred to nitrocellulose by the technique of Sout.hrrn (1975), and hybridized with radioactively labeled cloned I.688 sat’ellite DNA. The autoradiogram of bhe nitrocellulose, shown in Figure 6(b), confirms that, the Hap1 IIresistant satellite DNA fragments are not cleaved by HaelLI endonuclease. Howcvor~ digestion with HinfI or Ah1 endonuclease produced a fragment pattern qualitativclj similar to t,hat generated from total 1.688 satellite DNA by Hue111 endonucleasr digestion. This result indicates t,hat’ many of t’he HueIII-resistant 1.688 sat#ellite DNA fragments contain regularly spaced recognition sites for HitzfI and AluT edon&eases. The dist’ribution of HinfI sites in individual cloned segments of HaeIlI-resistant satellite DNA was also investigated. The gel in Figure 7(a) displays the fragments generated by Hid’1 cleavage of the six largest hybrid plasmids described previously. The fragments were transferred to nitrocellulose (Southern: 1975) and hybridized with a radioact)ively labeled cloned I.688 satellite T)XA proho. Fragments homologous

49”

M. CSRLSON

1

2

3

AND

4

1). BRUTLAO

5

6

7

8

Fro. 5. Cloned HneIII-msist.ant fragments lack HneIII cndonualease sites. Fourtrcn hybrid plasmids cont,aining HnelII-resistant, I.688 satellite DNA were digrstetl with 111uzTlI mdonuclease, and t.he resulting fragments were olectrophornsed in 1.4%) agarosc g<‘ls. The fragment patterns of 8 represent,ativo hybrid plasmids are shown. As mentioned in t,ht: text. the inserted DNA in one hybrid plasmid, shown in slot 3, gives rise t,o 2 fragments. The fragment,* common t,o all the hybrid plasmids are derived from the pSClO1 vect,or. The fragments WFLW t,ransferrcd from this gel to nitroeellulose (Southern, 1975) and hybridized with 321’-labeletl cDm688.52. Autoradiography confirmed that all the non-vector fragment,s of the 14 plaxmidn csxhibit homology to 1.688 satellite DNA (dat,a not shown). The inserted HaeIII-resistant segments in the 6 largest plasmids (4 of which are shown above) range in size from about 3.5 x lo3 base-pairs to 0 x 103 base-pairs as determined from the mobility of the H&II fragments containing these inserted segments on a 0.7% agarose gel (data not shown) relative to EcoRI-digested lambda DNA marker fragments (Thomas & Davis, 1975). In addition to an inserted segment, these H&II fragments also contain poly(dA).poly(dT) and $1tot,al of 0.2 Y 103 hnse-pairs of acljwcent vc‘rtor 1)N.Z (Carlson K: Rrutlng, 1977).

SEQUENCE

Hinf

VARIATIONS

Haem

Control

IN

A.-\TELLITI+:

A/d

Hinf

l)N:2

HaeEl

493

Cor tro.

Frc:. 6. HneIII-resistant, 1~658 satellite DNA contains HinfI and AZuI endonucleasc sitw. (a) Long HaeIIT-resistant DNA fragments, purified as described in Materials and Methods, were~l~pc:rt.c~iwith .4ZuI, H&f1 and HaeIIT ondonucleases. For comparison, total 1.688 satellite DE.4 \VILY Ilrgestotl with HaeIII endonuclease. The digestion products were electrophoresed in a l’x, iqau,se gel, with qua1 amounts of DNB loaded on the 3 slots at the left. Although no discrr+l* ~~lvavage products s,ro visible in the AZuI and HinfI digests of the HaeIII-resistant fragments, in c.:rvh r:w:r there is clearly less DNA migra.ting at the limiting mobi1it.y than in the Hue111 digest,. Slrrcr. t hc HwlII-rxsistant. fragment, preparation included contaminant DNA as well as 1.68X i;ltellitc I)N*k fragments homologous t,r) 1.6X8 satrllit,e squcnce.s WWR dr+cct,rd by hybridization. its ~Lvscrihcd in (b). (b) The fruymcnt pattern was transferred from the gel shown in (a) to nitrocellulose (Southern. 1!)75). ‘l’hc fragments were then hybridized with 3ZP-labeled cDm688.52 (9 x 104 &s/&n), a probch +Iwcifir for I.688 satellite DKA. An nutoradiogram of the nitrocellulose is shown. AZuI en&1~uc1cw.s~~ tlipwtion may produce fewer clistinct bands than does Hi,tfT cndnnurlenw digest irnr IWP~~IIWm:tn,y RS9 has+pnir units ham: 0I 41~1 sitw an(I only 1 Hiufl sitrb.

to the probe were detected by autoradiography (Fig. 7(b)). The fragments from three hybrid plasmids include satellite monomers approximately 359 base-pairs in size. However, many fragments generated from the cloned HaeIII-resist,ant satellite DNA are much larger than monomer, and also many fragments are not integral multiples of 359 base-pairs. Ry comparison of ‘Figure 7(a) and (b), it is apparent t,hat most of the fragmenk tlcrived from the inserted D. melanogaster DNA segments contain sequences homologous t,o l+W3 satellite DNA. However, a few fragments, such as the fragment, of pDm688.142 ment,ioned in the legend to Figure 7, did not hybridize with the probcb uudcr the hybridization conditions described in Materials and Methods, nor under

494

11. (‘AKLSOS

132

194

169

168

.\NI)

142

!I7

If.

HRCTJ>A(:

132

194

169

168

Vet :tor

Vet tor Din ner

M onon ier

(a)

(b)

142

117

SEQUENCE

VARIhTIONS

IS

SATELLITE

49.5

DSd

conditions (t, - 40°C; data not, shown). Such fragments contain n. DNA a,djacent to a block of 1.688 sa’tellite DNB in the chromosomes (Cnrlson & Brutlag, 1978). The same six hybrid plasmids were also digested with &LI cndonuclease. Digestion of total 1.688 sat)ellite DNA with this enzyme produces mainly fragments 174. 185. and 359 base-pairs in size (Fig. 8; Hsieh & Brut,lag, 1979). Five of the plasmids gavtb rise to fragments homologous to 1.688 satellite DNA but slightly smaller than these sizes. as judged by their mobility on a 2% agarose gel; moreover, many large fragments homologous to 1.688 satellite DNA were also generated, as was the case for Hi)zfT digestion of these plasmids (Fig. 7). Unexpectedly, AZuI digestion of the sixth hybrid plasmid, pDm688.169, revealed a repeating unit of 254 base-pairs in the inserted satellite DNA. less stringent melanogaster

(c) A region

of 16S8 satellite

DNA

with a 254 base-pair

repeat&g

unit

The HaeIII-resistant satellite DNA segment of pDm688.169 is cleaved almost ent,irely into 254 base-pair fragments by AZuI endonuclease (Fig. 8). The inserted segment, is 8.4 x lo3 base-pairs in size (see the legend to Fig. 8) and must therefore, contain about 33 tandem 254 base-pair repeats. Some degree of sequence homolog? het,ween t’hesc 254 base-pair repeating units and cloned 359 base-pair units was indicat#ed by both the hybridization observed between pDm688.169 and cDm688.52 (Fig. 8(h)) and the formation of heteroduplex structures by pDm688.169 and pDm688.23, mentioned above (data not shown). Sequence analysis of the 254 bits+ pair repcat \vas undertaken to determine the extent’ of homology with t)he knon-n sequencxc of the 359 base-pair unit (.Hsieh 8: Brutlag, 1979). To prepare radioactively labeled fragment’s suitable for sequence analysis, we took advant’age of several HinfI sites present in pDm688.169; HinfI endonuclease cleaves t’hc inserted satellite DNA at three sites (see Fig. 7(a)), and one of the resulting fragment’s is 254 base-pairs in size. For analysis of the nucleotide sequences adjacent to Alul sites, the 254 base-pair fragments produced by AluI endonuclease cleavage were labeled with 32P at the 5’ ends. Subsequent cleavage with HinfI endonuclease allowed separabion of the two labeled ends of those 254 base-pair fragments containing a Hitrfl site. For analysis of the sequences flanking the HinfI sites, the single 254 has+pair fragment produced by HinfI endonuclease digestion was end-labeled and thrn cleaved with AM endonuclease to separate the two labeled ends. Nucleotidt! sequences were determined by the procedure of Maxam & Gilbert (1977) (Fig. 9). Figure 10 compares the sequence of this 254 base-pair repeat with that of a 359 base-pair repeat unit from aDm688*23-24 (Hsieh & Brutlag, 1979). The two repeat unit’s are remarkably similar except for a sequence of 98 base-pairs present only in the 359 base-pair unit. Apart from this sequence, the two monomers are about 800/b homologous and differ mainly by single base insertions, deletions, or changes. The Ilornology between these units suggests that these two types of 1.688 satellite DNA uith different repeat sizes evolved from a common ancestor. The isolation of separate tandem repeats of t’he 254 base-pair unit, (33 tandem repeats) and the 359 base-pair unit (16 tandem repeats) in hybrid plasmids pDm688.169 and pDm688.23 (Carlson R: Hrutlag, 1977) indicates that these two different repeats compose separate regions of I.688 satellite DNB, and suggests that alteration of the ancestral sequence was followed by amplification of distinctive repeat units, resulting in new homogeneous tandem arrays.

PSC 101

117

142

166

194

132

169

1.688 satellite

359

254

185 174

-

-

bp bp

bp

bp

II7

142

168

194

132

169

1.688 satellite

SEQUENCE

VARIATIOSS

IS

S~~TELLITE

1)SA

4!)i

4. Discussion In these studies we have used molecular cloning to show that nucleotide sequence, rather than base modification, is responsible for the absence of Hue111 endonuclcascl sites in the dimer fragments and long HaeIII-resistant fragments generated from 1.688 satellite D1C’A by Hoe111 endonuclease digestion. This finding indicates that regions of 1.688 satellite DNA with different distributions of restriction endonuclease recognition sites (Carlson & Brutlag, 1977) differ in nucleotide sequence rather thaii degree of base modification. In cont’rast to our results; Gautier et al. (1977) harts demonstrated that base modification masks restriction sites in cow satellite D&A. However, the sequence CpG, which is preferentia,lly modified in cow and other DSAs (Sinsheimrr. 1956). is not present in any of the restriction sites in l-688 sabellite DN=\. \Vc have also characterized the long HaeIII-resistant regions of satellit,e DKA to determine their sequence relationship to region, s of 1.688 satellit#e DKA containing Hat!111 sites in nearly every repeat. We had previously shown that the Hue11 Iresistant portion of the I.688 satellite would cross-hybridize with cloned monomer repeats (Carlson & Brutlag, 1977). H ere we have shown that much of this Hae 1I1 resistant’ DNA still contains Hinfl and Alul sit,es distributed at approximately 35Q 1)ase-pair intervals, or integral multiples thereof. Thus, as judged from the distribution of restriction endonuclease sites: most of the HoeID-resistant satellite DNA appears to differ from the bulk of the 1.688 satellite DKA mainly in its lack of Hue111 rndonuclease sites. However, comparison of the five cloned HaeIII-resistant, DNA segments containing repeat units of about 359 base-pairs with the cloned satellit#e Di\‘A segment’ in hybrid plasmid pDm688.23 (Carlson & Brutlag. 1977) shows that the sequence variation between different stretches of satellite DSA is in some cases eon sidrrable. Jn contrast to the lack of HaeTl I sit’es and some\vhat irregular dist,ribut)ion of Hinfl xnd Blul endonuclease sites in the five cloned Hael ll-resista,nt segments. pDm688.23 contains a HaeIII. Hinfl and t’wo Blur endonuclease sites in ew~ch of 16 tandem 359 base-pair units. An example of variation not only in dist,ribution of restriction sites but a,lso in sizcb of t’he repeat) unit is provided by the Haelll-resistant segment’ cloned in pDm688.16Q. This plasmid contains 33 tandem repeats of a unit only 254 base-pairs in size. Sequrnc:~: anal,vsis showed that this 254 base-pair unit is homologous to all but 98 i)ase-pairs of t*hc 35Q base-pair repeat. Although pDm688.169 offers a particularly striking exampl(~ of variation in size of the repeat unit, the finding that the other five cloned segments have repeat units slightly smaller than 359 base-pairs suggests that, this kind of sequence variation may be common in 1.688 satellite DXA. Although one of six HaeIII-resistant sat’ellite clones analyzed contained 254 I)ase pair units. regions composed of 254 base-pair repeat units constitute only a small fraction of the I.688 satellite DKA. When total 1.688 satellite DPiA was cleaved with (b) Thra 1)S.a fragments were transferwtl from the gel lo nitrocollulosr ant1 hpbritliwcl kvith “*I’-lahrlrvl cDm688.52 DNA (1 x 1V cts/min). ;\n autoradiogram of t,hc nitrocellulow is shown. ‘l’ransf~~r of very small fragments to nitrocellulose is variable in efficiency, and in other exlwrirncwt. MY haw tletrcted in each hybrid l>lnsmitl rxccl>t l)Um688.169 a l)air of labclr~tl fragments, WIW~lwmling to the pair visible on t,he gel, of sizes close to 180 bascx-pairs. I’lasmid pl)m-688, I6!) prnertrt.es a strong band of 254 base-pair fragments homologous to 1.688 satellite DNrZ. The size of the inserted segment, in pDm688.169 was determined t,o be X.4 x lo3 base-llaiw b> co-clectrophort,sis of HoeII-digested pDm688~169 DNA with marker fragment,s in a 0.7% agarww gel (data not shown). HneII endonuclcase digestion ofpDm688.169 generates a fragment containing the inserted I)iYA segment flanked by poly(clA)~poly(tlT) am1 a total of 0.3 x IO3 base-l)airs (11 adjacent vector DXA (Carlson & Brutlag, 1977).

G

A+G

C

C+T

G

A+G

C

C+T

30

40

50

60

(b)

(a) I?‘lr:. 9. Representative autoradiograrn of a sequencing gel. A 168 base-pair fragment end-labeled at the H&f1 endnnuclease site and including nuclootick 199 to 254 and 1 to 112 was prepared as described in Materials and Nethods. The chemical PCIactions for gnanine, strong adenine/aealr guaninc, cytosine, and thymine and cytosine cleavage were curried out (Maxam & Gilbert, 1977). The reaction produe& were nlectrophoresed in bot,h a 200,; polyacrylamide gel (a) and a 12”’ /Cl polyacrylamido gel (b) as described in Materials and Methods. Autoradiographs of the gels are shown, and numbers indicate the l>ositiorl in the sequence, as written in Fig. 10, of the nuclootide to which each band corresponds. The sequence read from these gels is complementary to the seyuence shown in B’iguro 10.

220 TCATAl~TCGCXXAGCTXGTAATTA~~TTTCCAATGAAA~TGTGTT=AA~AATGAAAATTA~ATTTTT=~ TCATATCTCACTGAGCTCGTAATAAAATTTCCAATCAAACTGTGTTCAAAAATGGAAATTAAATTTTTTGG 310 320 330 300 290

Ah I 1

130 140 160 120 CCCTAATTTCGGTCATTAAATAATCAGTTTTTTTGCCACAA~TTTAAAA~TAATTGT~TGAATATG~AATG TTCCAAATTTCGGTCATCAAATAATCATTTATTTTGCCACAACATAAAAAATAATTGTCTGAATATGGAATG 250 260 220 240 230

ATGTGACCATTTTTAGCCAAGTTATAACGAAAATTTCGTTTGTAAATATCCACTTTTTTGCAGAGTCTGTTT 170 160 160 180

70 60 TAAATCCTTCAAAAAGTAA TAAATCCTTCAAAAAGTAATAGGGATCGTTAGCACTGGTAATTAGCTGCTCAAAACAGATATTCGTACATCT 80 90 100 110 130

60

T

230

190

340

270

240

200

170

100 ATAif:AACTTTTTGGCAAAATCTGATT

120

CAXAITT G::AATTT AAT::ACCCC ccTT 3~~~AAAATGC406~AAA~TAA~&AA~AAT~GBdATTTCCC CCACATTTTGCAAATTTTGATGACCCCCCTCCTTACAAAAAATGCGAAAATTGATCCAAAAATTAATTTCCC 10 20 30 40 50

360

280

210

369

264

1

Hinf I

140

70

FIG. 10. Comparison of the 254 and 359 base-pair ropoitts uf 1.688 stltellite DNA. The 254 base-pair sequence shown is derived from those 254 base-pair units in pDm688.169 which contain HinfI endonuclease sites, and the 359 base-pail sequence was determined by Hsieh & Hrutlag (1979) for a single cloned monomer of satellite DNA. The symbol X in the sequence indicates that the nucleotide has not been determined, and underlining indicates that the sequence is uncertain. The sites of HinfI and AZuI endonuclease cleavage of the 254 base-pair repeat me shown. Three instances of sequence heterogeneity among tho 254 base-pair units were detected as discrepancies between the sequence obtained from fragments end-labeled at AZuI endonuclease sites and that from fragmrntjs end-labeled at HinfI endonuclease sit,es. The alternat)e nucleotide is indicated at positions 158, 162, and 179. Other sequence variations may have gone undetected. Hyphens have been omitted for clarity. bp, base-pairs.

254 bp 359 bp

254 bp 359 bp

254 bp 359 bp

254 bp 359 bp

254 bp 359 bp

500

31. CARLSOS

A.NI,

I).

HRUTLAQ

rllul and fractionated hy polyacrglamide gel electrophoresis, DNA fragment-s 254 base-pairs in size were a very minor product’ in comparison wit’11 t#hose derived from the 359 base-pair repeat (data not shown). Like the satellite DNA segment in pDm688.23, the 8.4 x lo3 hase-pair satellite DKil segment in pDm688.169 appears relatively homogeneous by restriction site mapping; it’ conbains an Ah1 site in every repeat. and the major heterogeneity detected is t,hr presence of three H&f1 sites within the segment. Moreover. the only heterogeneity detected among those 254 base-pair units used for sequence analysis were single ba.se changes. Comparison of the regions of satellit’r DNA cloned in pDm688.23 (Hsieh CV Brutlag. 1979) and pDm688.169 shows that the sequence variations within each region consist primarily of occasional base changes hetween neighboring repeats, while t hr variabions between units of the two different regions are both more numerous a.nd more major. These data, together with previous work (Carlson 8: Rrutla’g. 1977). imply regional homogeneity in 1.688 satellite DNA. The similarity between the 254 and 359 base-pair units suggests that, they evolved from a common ancestral repeating sequence. The existence of tandem arrays of each of t,hese units would then imply that amplification followed their divergence with respect to size. Regional homogeneity may he a direct result, of the ampliiicabion mechanism. This IVe thank Michael Goldberg for advice on DNA sequencing techniques. supported by a Basil O’Connor Starter Grant from the National Foundation Dimes and by a grant from the National Institute of General Medical Sciences. 0rw of us (M. C.) was a Ullit.ed States Public Health Service Trainee.

work \vas Xwch r,f

.KEFEREN(!ES Hover, H. W. & Roulland-Dussoix, I). (1969). ./. AVIol. Viol. 41, 459-472. Brutlag, D., Fry, K., Nelson, T. & Hung, P. (1977). Cell, 10, 509-519. Carlson, M. & Brutlag, D. (1977). Cell, 11, 371 381. Carlson, M. & Brutlap, D. (1978). Proc. Nat. Acatl. S’ci., l1.S.d. 75, 5898-5902. (‘hang., L. M. S. & Bollurn, F. J. (1971). ,/. Hiol. Chem. 246, QOQ--916. CoIlen, 6. N., Chang, A. C. T.. Bayer. H. W. $ Helling, R,. B. (1973). I’roc. .\‘at. dcatl. Sci., C.S.A. 70, 3240-3244. Davis, R. W., Simon, M. & Davidsot), N. (lQ71 ). 111Xethods in &zymology ((:rossrnau, I,. & Moldaue, K., eds). x-01. 21D, pp. 413-428, Academic Press, New York. DcsrAardt, D. T. (1966). Biochem. Bioph,ys. Res. (!mmrnur?,. 23, 641--646. Gautiw. F ., Biinernann, H. & Grotjahn, L. (1977). Eur. J. Biochem. 80, li5-183. Grul&oin. M. & Hogwss, D. S. (1975). Proc. Sat. Acarl. SC%., Cr.S.A. 72, 3961--3Q65. Buyer, M. S.. Figurski. D. & Davidson, X. (IQ76). ,1. Bacterial. 127, 988-997. Hsiclr. T. & l3rutlag. D. (1979). .7. ,Vol. Rio/. 135, 46.5 -481. ,Jo\.itl. T. M.. Englund, P. T. & Bertsch, L. L. (1 Q69). .I. Biol. Chem. 244, 2996 --3008. Malttcuil, S., Hamer, D. H-. & Thomas, (.‘. >A., .Jr (lQ75). Cell, 5, 413-422. Masarn. A. M. & Gilbert, W. (1977). Z’TOC. A’at. &ad. A%‘.. I’.S.A. 74, 560 -564. Modrictr. LJ. & Zabel, I>. (lQ76). J. L(iol. Cl&em. 251. 5866G5874. Poacwck, A. C. & Dingman, C. II:. (1968). Biochemistry, 7, 668.-674. Riclrardson. C. C. (1965). Proc. Nat. Acaci. Sci., T’.S.A. 54, 158-165. Rigby. I’. W. J., Dieckmann, M., Rhodes, C. &, Berg, P. (1977).S. illol. Biol. 113, 237-251. Roberts, R. .J., Breitmcyer, .J. B.. Tabnchnik. N. 1”. & Myers, 1’. A. (lQi5). .J. Mol. Biol. 91, 121-123. Slwn. (‘.-K. J., Wiesehahn, G. 8 Hewst, .I. E:. (1’176). Surl. Acids Res. 3, MSl--951. Sinshcirner, R. L. (1955). .7. Hiol. ClLern. 215, 57Q 583. Soutllern, E. M. (1975). J. Mol. Biol. 98, 503-,517. Tl~ornas, M. & Davis, R. W. (1875). .J. ,Ilo/. Kiol. 91, 315k328.