J. Cell Sci. 31, 303-313 (1976) 303

Printed in Great Britain

THE ACTIONS OF RESTRICTION

ENDONUCLEASES ON LAMPBRUSH

CHROMOSOMES

D. C. GOULDIBM Thomas J. Watson Research Center, Yorktottm Heights, New York 10598, U.S.A.

H. G. CALLANZoology Department, University of St Andrews, Scotland

AND C. A. THOMAS, JRDepartment of Biological Chemistry, Harvard Medical School, Boston, Massachusetts02115, U.S.A.

SUMMARYLampbrush chromosomes from oocytes of Notophthalmus viridescens were dispersed in

media containing restriction endonucleases isolated from Haemophilus and E. coli. Theseendonucleases cleave duplex DNAs at specific palindromic sequences of nucleotides, and severalsensitive sites occur per micron of DNA. The overwhelming majority of the lateral loops oflampbrush chromosomes are extensively fragmented by these endonucleases, but an occasionalpair of loops is refractory. A notable example of loops showing this refractory property are thegiant loops on chromosome II in the presence of Hae. These loops, whose DNA-containingaxes are several hundred micra long, are sensitive to other nucleases such as EcoB, endo-nuclease I and pancreatic DNase I; their refractory behaviour towards Hae therefore indicatesthat the sequence sensitive to this particular endonuclease is systematically absent. Thisanomalous property can be comprehended if it be assumed that the axial DNA of the giantloops consists of tandem repeats of a sequence which happens not to include the sensitive site.

INTRODUCTION

During the past few years a number of sequence-specific endonucleases have beendiscovered in bacteria. These endonucleases recognize particular sequences of 4-8nucleotide pairs and introduce a double-chain break in this site. Among the bestcharacterized of these endonucleases are those found in various strains of Haemo-philus. For example, one isolated from H. aegyptius (Hae) cleaves duplex DNA at thesequence 5' GGCC (K. Murray, personal communication).

The number average length (in base pairs, bp) of the collection of segments pro-duced when E. coli DNA and Drosophila melanogaster DNA are digested with Hae is350 + 70 (Hamer & Thomas, 1975). In accord with statistical expectation, this endo-nuclease introduces some 30-40 breaks in T7 DNA (42000 bp), about 9 in SV40 DNA(5200 bp), and 13 in <^Xi74RF DNA (5200 bp). Thus we can assume that any non-repeating DNA will be broken into segments some 200-500 bp in length, i.e. some6-12 breaks per micron of double helix.

The lateral loops of lampbrush chromosomes are known each to be composed ofa single DNA double helix, the loop axis (Gall, 1963), with associated ribonucleo-

304 D. C. Gould, H. G. Callan andC. A. Thomas, Jr

protein matrix resulting from transcription. A study of the morphology and develop-ment of these loops has led to the suggestion that loop axes are composed of tandemly-repeating DNA sequences (Callan & Lloyd, i960). If this be the case, then an entireloop, whatever its length, should not be broken by a given restriction endonucleaseunless the unit sequence contains one or more sensitive sites; if it does containsensitive sites then the loop should be broken many times.

Lampbrush chromosomes are in the germ line and hence not exceptional in regardto their genetic organization. The visibility of their lateral loops in the light micro-scope depends on the retention of the associated ribonucleoprotein matrix, though thismaterial is generally not effective in protecting the axial DNA molecule from the actionof nucleases (Callan & Lloyd, i960; Macgregor & Callan, 1962). Thus we can ask: doall restriction endonucleases break all loops, which is to be expected if loops containnon-repeating sequences; or do certain restriction endonucleases leave certain loopsunbroken, which is to be expected if loops contain many tandemly-repeating unitsequences. The results reported here demonstrate that at least one restriction endo-nuclease acts in accord with the second alternative.

MATERIALS AND METHODS

Isolation and characterization of restriction endonucleases

The restriction endonucleases were assayed during isolation by adding aliquots of appropriatefractions to a viscometer that contained a solution of salmon sperm DNA (Sigma); decrease inviscosity was recorded as a function of time (Smith & Wilcox, 1970). The results were plottedin a reciprocal manner (Thomas & Doty, 1956) to obtain total units of activity (1 unit = amountof enzyme giving a 25 % decrease in viscosity in 1 min (Smith & Wilcox, 1970)).

The restriction endonuclease Hind, which consists of a mixture of Hind II and Hind III , wasisolated from Haemophilus influenza Rd, obtained from H. Smith, Johns Hopkins University.It was grown in liquid medium, harvested by centrifugation and ruptured by sonicarion(Smith & Wilcox, 1970) or with a French pressure cell (8-3 x io4 kN m~' (12000 lb in."1)) afterresuspension in 005 M Tris, pH 74, 8 mM glutathione. To the supernatant from a 30-min,10000-g centrifugation of the lysate was added dextran, polyethylene glycol (PEG), and NaClto 3'5M (Klein & Capecchi, 1971). The lower dextran phase which contained the nucleic acidswas discarded; the upper phase was dialysed against 0-05 M Tris, pH 7-4, 1 mM /?-mercapto-ethanol, 5 % glycerol, for 2 h and then (NH4)jSO4 was added to 33 % saturation. The upperphase, which contained PEG, was discarded. (NH4)SO4 was added to the lower phase to give75 % saturation and the resulting precipitate was collected by centrifugation.

The pellet was washed 3 times in 75 % saturated (NH4)8SO4, then dissolved in a volume of005M NaCl, 002M Tris, pH 74, 1 mM /?-mercaptoethanol to give a protein concentration of4 mg/ml; 20 g of bacteria gave 400 mg of protein at this point.

The precipitates from (NH4)8SO4 cuts of 0-40, 40-50, 50-60, and 60-70 % saturation wereeach dissolved in 3 ml of 0-05 M NaCl, 0-02 M Tris, pH 74 , 1 mM /?-mercaptoethanol. Thefraction from the (NH4),SO4 cut of 50-60 % saturation contained approximately 60 % of thetotal activity. 1'his fraction was run through a G-15 Sephadex column, 2 cm x 75 cm, equili-brated with o-oi M KjHPO4, pH 7-2, 1 mM /?-mercaptoethanol. Five-millilitre fractions werecollected. The first 2 fractions emerging at the void volume were pooled and pumped on to a2-ml bed volume phosphocellulose (P-cell) column at 5 ml/h (Whatman P-11 previously equili-brated by running through 50 ml of 0-5 M KC1, o-oi M K,HPO4, pH 7-2 and then 200 ml ofO-O2M K,HPO4, pH 72). The column was washed with O-OIM K2HPO4> pH 72 , 1 mM /?-mercaptoethanol, until the eluate had an A^D < 005 . Protein was then eluted stepwise with3-ml aliquots of KC1, in steps of 01M KC1 beginning with 0 1 M KC1, in O-OIM K.HP0 4 ,pH 72 , 1 mM /?-mercaptoethanol; i-5-ml fractions were collected.

Actions of restriction endonucleases 305

The total endonuclease activity in the fractions eluting at 0 3 M KC1 was 5 0 % higher thanthat eluting at 0 2 M KC1; the specific activity of the 0-3M KC1 fractions was 2-fold higher(3000 U/g protein). Making the 0-3 M KC1 fractions 7 0 % saturated in (NH4)jSO4 (Smith &Wilcox, 1970) caused substantial loss of total activity and little or no increase in specificactivity. For this reason the endonuclease preparation used was most frequently the 0-3 M KC1fraction. Gel electrophoresis of this fraction at pH 104 revealed at least 4 and possibly 7 bandswhen stained with naphthol blue. As is now known (Roy & Smith, 1973), this preparationcontains more than one restriction endonuclease and therefore cleaves duplex DNA at sitesother than that originally defined by Kelly & Smith (1970). The sites at which cleavage occursare specific sequences of nucleotides, however; they are not random.

The growth of Haemophilus aegyptius (American Type Culture Collection) and the isolationof the restriction endonuclease from this bacterium was the same as for H. influenza Kd up tothe step of elution from the P-cell column. The P-cell was equilibrated and protein applied asdescribed above. The column was washed with 001 M K2HPO4, pH 72 , 1 mM /?-mercapto-ethanol and then developed with 3-ml aliquots of KC1 in steps of o-iM KC1 beginning with0*1 M KC1, in 0 1 M KtHPO4, pH 72 , 1 mM /?-mercaptoethanol. Fractions of 15 ml werecollected. Approximately 7 0 % of the total endonuclease activity eluted at O-6M KC1. Theremainder was in the 05M KC1 fractions.

The fractions from the P-cell column that contain restriction endonuclease from H. influenzaeRd also contain substantial but variable amounts of an exonuclease-like activity as determinedby the production of trichloroacetic acid (TCA)-soluble material from 3H-T7 DNA. The pre-paration of the restriction endonuclease from H. aegyptius did not contain this activity, for theactivity producing acid-soluble material eluted in the 0 2 M KC1 and 0-3 M KC1 fractions fromthe P-cell column.

The majority of our preparations of endonuclease from H. influenzae Rd also contain anRNase activity as determined by the production of the acid-soluble material from polyA,polyU, and/or tRNA as substrate and by the action of these fractions on lampbrush chromo-somes (Macgregor & Callan, 1962).

To concentrate the Hind endonuclease, to reduce the activities giving acid-soluble materialfrom 3H-T7 DNA, and to avoid stripping RNP matrix from the lampbrush chromosomes, theO2M KC1 and 0-3M KC1 fractions from the P-cell column were combined, run through aSephadex G-25 column ( i c m x i s c m , equilibrated with 001 M K S H P O 4 , P H 7 2 , 1 mMy?-mercaptoethanol) for desalting, and pumped on to another P-cell column (0-2 ml bed volume,equilibrated as given previously) at 2 ml/h. The P-cell column was washed with 5 ml of O-OIMKaHPO4 ) pH 72 , 1 mM /?-mercaptoethanol and then developed with 2-ml aliquots of 0 2 MKC1, 0-25 M KC1, 03M KC1, 04M KC1 in O-OIM K J H P O 4 , pH.7-4, 1 mM /?-mercaptoethanol.Ten-drop fractions were collected. Approximately 90 % of the activities producing acid-solublematerial from 3H-labelled DNA and RNA eluted from the P-cell column in the 0-2 M and 0-25 MKC1 fractions. Approximately 40 % of the endonuclease activity was recovered in the first 2fractions eluting from the P-cell column with 0 3 M KC1 in K,HPO4, pH 7-2, 1 mM /?-mer-captoethanol. The activities (U/ml) of these fractions were concentrated 2 to 3 times over theactivity of the pooled fractions.

An RNase activity in the endonuclease preparation from H. aegyptius ( 06M KC1 fractionfrom the P-cell column) was effective in producing acid-soluble material from tRNA but didnot cause detectable stripping of ribonucleoprotein (RNP) matrix from the loops of the lamp-brush chromosomes.

In order to inhibit the low level of RNase activity remaining in our preparations of Hind,cytidine-5'-monophosphate (Sigma) was added to a final concentration =S 10 mM in the ob-servation chamber.

The restriction endonuclease from E. coli B was isolated as described previously (Roulland-Dussoix & Boyer, 1969). E. coli B was supplied by Stuart Lin, University of California, SanFrancisco (SL 113). The activity of Eco B was assayed by viscometry and by action on lamp-brush chromosomes. Absence of iS-adenosyl methionine in the reaction mixture resulting in lossof activity was used as control.

The restriction endonuclease from E. coli K, Eco K, was obtained from Bob Yuan, HarvardUniversity. It was purified by Janet Haywood by the procedure of Meselson & Yuan (1968)and contained 130 fig protein/ml. This enzyme showed an RNase-like activity on lampbrushchromosomes which could be suppressed by the addition of 1 mM cytidine-s'-monophosphate.

306 D. C. Gould, H. G. Callan and C. A. Thomas, Jr

Endonuclease I was obtained from Charles Richardson, Harvard Medical School. It con-tained no glycerol and was stored at — 20 °C. Pancreatic DNase I was purchased from Sigma.

The action of the endonucleases Hind and Hae on isolated DNA was determined by the sizedistribution of the scission products from unlabelled and 3H-labelled DNA from T7. Diges-tions were carried out under conditions described previously (Middleton, Edgell & Hutchison,1972). Extent of reaction was determined by sedimentation of the reaction mixture in a 5—20 % linear sucrose gradient. The reaction was allowed to proceed until further addition ofendonuclease or increase in the time of reaction caused no further lowering of a minimal S-valueof the DNA. These terminal digests were then electrophoresed in 1 % agarose-acrylamide gelsunder conditions given by Middleton et al. (1972).

The DNA contained in the gels was detected by autoradiography using no-screen X-ray film(Kodak) when 3H-T7 DNA was digested, or by fluorescence after staining with ethidium bro-mide (Sigma) when unlabelled T 7 DNA was digested; the latter method of detection utilizesa fluorescence microscope; details of the procedure are to be published (C. Muir & D. Malcolm).

Preparation and examination of the lampbrush chromosomes

The specimens of American red-spotted newt, Notophthalmus (Triturus) viridescevs, whichwere usedin 1970 and 1971, were supplied by Lewis Babbitt, Petersham, Massachusetts, thoseused after 1972 were supplied by William Lee, Oak Ridge, Tennessee. The newts were main-tained in aquaria at 18 °C. Female newts were anaesthetized in MS222 (Sandoz) prior to dis-section and removal of the ovaries. Single ovaries were stored, in no case for more than 6 h,in solid watch glasses (embryo cups) with Vaseline-sealed lids. The watchglass contained noliquid other than a little coelomic fluid carried over with the ovary, and was kept on crushed ice.

For isolation of oocyte nuclei, a small portion of ovary was placed in unbuffered 100 mMNaCl. Yellow oocytes within a size range of 0-6-0-8 mm diameter were selected, and the nucleiisolated manually with a pair of watchmaker's forceps and a steel needle. Each nucleus waspumped free from yolky cytoplasm using a flame-polished narrow pipette (orifice internaldiameter about 0 5 mm), and was then transferred to an observation chamber containing amedium whose composition was adjusted empirically so as to conform as closely as possible tothe requirements demanded by each enzyme for optimal activity. The medium needed also tosatisfy conditions necessary for the observation of lampbrush chromosomes in the phase-contrast microscope while digestion progressed, namely that the nuclear sap should disperse,though not too rapidly, that the ribonucleoprotein matrix on the lateral loops remain in place,and that the digestibility of the chromosomes should be sustained for as long as possible.

The media chosen for observing the digestion of the chromosomes by the enzymes areshown in Table 1, alongside what are considered to be optimal reaction conditions for the variousenzymes. Certain of these reaction conditions are in conflict with those required for the pro-longed observation of lampbrush chromosomes, and it will be apparent that we have had tocompromise. All modifications of the reaction medium resulted in a decreased activity of theendonuclease as determined by viscometry. The modification did not, however, alter thespecificity of the endonuclease for the pattern of fragments of digested T7 DNA (digestedusing the modified reaction medium) upon electrophoresis was the same as observed underoptimal conditions.

Dispersal of the nuclear sap is required so that the chromosomes can be seen by phasecontrast, because their visibility depends on a diminished refractive index of the medium sur-rounding them; furthermore the dispersal of nuclear sap allows the chromosomes to float free,to manifest Brownian motion, and thus ensure that any discontinuities in chromosome andloop axes become apparent as they occur. Too-rapid sap dispersal cannot be tolerated, however,for successful removal of the nuclear membrane without mechanical damage to the chromo-somes depends on a cushioning action of the sap gel. Other things being equal, the gel com-ponent of the nuclear sap of N. viridescens oocytes of the chosen size disperses slowly or not atall in unbuffered 100 mM NaCl; therefore this is an appropriate medium for isolating nuclei.The sap gel will usually disperse reasonably quickly when a nucleus is transferred to unbuffered70 mM NaCl, though the oocytes from occasional animals are recalcitrant. In such cases CaCljwas added to the medium to c i mM, and this ensured dispersal.

The pH stability in the observation chambers was maintained at 7-2-7-4 by using 6-6 mM

Tab

le I

. A

ssay

and

lam

pbru

sh c

hrom

osom

e (L

bc) d

%es

tion

cond

ition

s

Hind

Hae

E

co B

E

co K

E

nd

o I

P

ancr

eati

c D

Nas

e I

&----&

B

ioch

emi-

B

ioch

emi-

ca

l as

say

Bio

chem

i-

Bio

chem

i-

Bio

chem

i-

cal

assa

y (R

oull

and-

ca

l as

say

cal

assa

y ca

l as

say

(Sm

ith

&

Dus

soix

(M

esel

son

(Sho

rtm

an

(Sha

ck &

W

ilco

x,

Lb

c B

ioch

emi-

L

bc

& B

oyer

, L

bc

& Y

uan,

L

bc

& L

ehm

an,

Lb

c B

ynum

, L

bc

i?

1970

) di

gest

ion

cal

assa

y di

gest

ion

1969

) di

gest

ion

1968

) di

gest

ion

1964

) di

gest

ion

1964

) di

gest

ion

f In

gred

ient

s %

K

CI,

m~

NaC

l, n-m

Tri

s, nm

Mg

CL

m~

C

aCI,

, rn

M

/9-S

HE

, a+

BS

A, p

g/rn

lC+

S

AM

, m~

"*

A

TP

, m

M

ED

TA

, r

m

5'-C

MP

, m

~

pH

Th

e si

ngle

nuc

leus

con

trib

uted

3 m

g/m

l of

nuc

lear

pro

tein

, an

d 0.

004

pglm

l of

lam

pbru

sh c

hrom

osal

DN

A t

o th

e 30

-pl

cavi

ty.

+ /9

-mer

capt

oeth

anol

. +

+ B

ovin

eser

um a

lbum

in, P

ente

x fr

acti

on V

. +

++

S-a

deno

syl-

L-m

ethi

onin

e w

as p

urif

ied

by e

luti

on f

rom

Dow

ex 5

0 b

y N

HC

I, n

eutr

aliz

ed w

ith

an e

quiv

alen

t vo

lum

e of

N N

aOH

and

thi

s so

urce

of

NaC

l ta

ken

into

acc

ount

in

the

stat

ed N

aCl

conc

entr

atio

n. t

Na+

con

trib

uted

by

Tri

s bu

ffer

; sa

lt p

lus

buff

er g

ave

5 m

M i

onic

str

engt

h. t

t va

riab

le,

reng

e 1.

5-25

mM

.

308 D. C. Gould, H. G. Callan and C. A. Thomas, Jr

Tris buffer in the dispersing medium. Although several of the enzymes require MgClj as aco-factor, and ideally at a higher concentration, we were forced to limit MgCl2 concentration toi mM in the dispersing media because higher concentrations coagulate the nuclear sap.

In earlier studies of the digestion of isolated lampbrush chromosomes by pancreatic DNase(Callan & Lloyd, i960; Macgregor & Callan, 1962) it became apparent that the chromosomesbecome progressively less susceptible to breakage with the passage of time, this being probablyoccasioned by oxidation of the protein components of the chromosomes. Cross-linkage betweenthe proteins may then suffice to maintain chromosome integrity even though DNA chains havebeen cleaved. The involvement of disulphide bonds in this process has recently been docu-mented by Hill, Maundrell & Callan (1974). In order to hold oxidation in check as far as possible,aliquots of dispersing medium just prior to use were de-gassed, equilibrated with nitrogen gas,/?-mercaptoethanol added to 1 mM, and the containers rightly stoppered. We noticed that/?-mercaptoethanol itself is one of the few chemical agents, other than DNase, which breaks thelateral loops of lampbrush chromosomes. In control experiments we found that this reagent at1 mM in dispersal medium produces an occasional break, but at a frequency far too low to be ofmaterial concern. The value of /?-mercaptoethanol in maintaining isolated lampbrush chromo-somes susceptible to breakage by enzymes over periods of 30 min and more was clearly demon-strated in the course of this study.

Bovine serum albumen was added to the dispersing media at 1 fig/m\ in order to obviatepossible heavy metal contamination and consequential inactivation of the enzymes. Its value inthe present experiments is unproven; nuclear sap contributes several orders of magnitude moreprotein to the medium surrounding the chromosomes while digestion is in progress (Table 1).

The chambers in which the observations were made consist of a conventional 3 by 1 in.(7-6x2-5 cm) glass microscope slide with a quarter-inch (64-mm) diameter hole boredthrough the centre, and a half-inch (i-3-cm) square coverslip attached to the slide with lowmelting point paraffin wax, across the hole, to form the floor of the chamber. Bored slides andcoverslips were acid-washed, rinsed extensively with water, and wiped dry from 95 % ethanol,before being assembled as chambers. The volume of medium in each chamber is about 30/tlwhen covered with a top coverslip.

Digestions were observed in the following manner. About 40 /tl of dispersing medium werepipetted into a chamber, producing a convex meniscus. Enzyme solution, generally 1 or 2 fi\,was added with a micro-cap dispenser, and the medium stirred. Under a dissecting binocularan oocyte nucleus was now isolated, pumped free from cytoplasm, and transferred to thechamber by pipette with as little of the isolating medium (01 M NaCl) as possible. Beforeextensive swelling of the nucleus had taken place, its membrane was gently removed withsharp No. 5 watchmaker's forceps and a fine-pointed tungsten wire needle, and the nuclearcontents spilled out on to the floor of the chamber. The preparation was now covered with atop coverslip, and immediately examined with a Zeiss inverted (Plankton) phase-contrastmicroscope. If the manipulations are performed quickly enough, little sap dispersal will haveoccurred prior to observation, and any significant mechanical damage will be apparent from theoutset. Mechanically damaged preparations were discarded.

As the nuclear sap disperses the lateral loops of the lampbrush chromosomes begin to showBrownian motion, and the progress of enzyme-induced cutting of loops and main chromosomeaxes can be directly observed. Photographs of preparations were taken at various time intervalsto record the general progress of digestion.

The morphology of the lampbrush chromosomes of N. viridescens has been described indetail by Gall (1954). Fig. 1 shows a working map of these chromosomes redrawn from Gall'soriginal data. In accord with conventional practice the chromosomes are numbered in order I toXI with chromosomes I the largest. The working map shows various 'landmark' structures, butthe only landmarks which concern us in the present paper are the giant loops close to the cen-tromere in the longer arm of chromosome II. These are to be seen in Fig. 2A (p. 310). Thegiant loops arise from 2 loci; between these loci lie a few pairs of loops of ordinary dimensions,and similar loops are situated to left and right as well. The giant loops are remarkable in severalrespects. They are much longer than the general run of ordinary loops, single examples oftenbeing in excess of 300 fim; their RNP matrix is bulky, though its fluffy texture is like that ofmost loops; the matrix shows a polarized distribution which may run uninterrupted from thinto thick insertion in the axial chromomere, indicating uninterrupted transcription, or there may

Actions of restriction endonucleases 309

be discontinuities in the matrix (as evident in Fig. 2 A) such that there are two or more pro-gressions from thin to thick within the length of a single loop; finally in certain oocytes theremay be more than one pair of giant loops originating from a single locus on chromosome II.

Our aim in the present study has been to investigate whether any lateral loops of the N.viridescens lampbrush chromosomes are refractory to digestion by particular restriction endo-nucleases. Such observations should be made with caution, for as well as the DNA in the loopaxis, ribonucleoprotein is present as well. A few loops of Triturus cristatus remain intact in thepresence of pancreatic DNase when all other loops in the preparation have been destroyed. Thisrefractory behaviour is a consequence of particularly dense matrix texture, and is certainly notdirectly occasioned by any sequence peculiarity in the loop axis DNA. It will be seen in the

I (11) 1

II (10) •

III (7) _Q_fi Jt, Q-

IV (9)

V (6)

VI (5) A

VII (8)

VIII (4)

IX (3)

X (2)

xi p 9u9u

Fig. 1. Working map of the lampbrush chromosomes of Notophthalamus (Triturus)viridescens, redrawn from the data of Gall (1954). Vertical arrows indicate the centro-meres ; no indicates the nucleolar organizer; Roman figures enumerate chromosomes ac-cording to size, I being largest; Gall's original chromosome numbers enclosed inbrackets.

present study that in regard to Hae the giant loops of N. viridescens are refractory; however thisrefractory behaviour is not a consequence of peculiar matrix texture because they are extremelysensitive to other DNases. Indeed it was precisely on these giant loops of chromosome II thatGall (1963) made his kinetic study of loop breakage by pancreatic DNase, giving results whichled him to postulate that there is only a single DNA duplex in the loop axis.

RESULTS

Digestion of DNA from the phages T7 and 0X174 with Hind and Hae, electro-phoresis in 1 % agarose-acrylamide and detection of the DNA by autoradiography or

D. C. Gould, H. G. Callan and C. A. Thomas, Jr

Fig. 2. A, photograph showing the giant loops close to the centromeres of chromosomeII, undigested, B, the giant loops of chromosome II, and other loops, at an earlystage of digestion by pancreatic DNase. C, the giant loops of chromosome II, andother loops, at an early stage of digestion by endonuclease I, the double chain breakingnuclease from E. coli. D, the giant loops of chromosome II, and other loops, at anearly stage of digestion by Eco B, the restriction endonuclease from E. coli B (Table i).

Actions of restriction endonucleases 3"-

Fig. 3. A and B both show the giant loops of chromosome II resistant to digestion byHae at a stage when other loops have been pulverized.

by fluorescence after staining with ethidium bromide, was found to give the patternsthat are now widely published in the general literature. The presence of sharp bands inthe gels indicates that the DNA molecules have been broken into a collection of smallpieces having discrete lengths. This means that these endonucleases break the DNA atspecific sites. The mobility and relative intensities of the bands formed from digestedDNA molecules are different for the different endonucleases. This demonstrates thatthese endonucleases break the DNA molecule at different sites. Two of our prepara-tions of Hae were used by Dr K. Murray to determine the sequence recognized bythis endonuclease.

Pancreatic DNase was found to cut all the lateral loops of the lampbrush chromo-somes of N. viridescens, including the giant loops on chromosome II (see Fig. 2 B).Endo I, the double-chain-breaking non-specific nuclease from E. coli, was similarlyfound to cut all the lateral loops without exception (see Fig. 2 c). Both these enzymeswere also found to cut the main chromosome axes.

Eco B, in the presence of 5-adenosyl-L-methionine (SAM) and ATP, producedextensive loop breakage, including breakage of the giant loops on chromosome II(see Fig. 2D).

Eco K, again in the presence of SAM and ATP, produced extensive loop breakage,but the giant loops on chromosome II appeared to be relatively resistant. In theabsence of SAM or of ATP, no loop breakage was observed.

Hind was found to break most of the loops provided the preparation chambers were

312 D. C. Gould, H. G. Callan and C. A. Thomas, Jr

incubated for a few minutes at 37 CC. At room temperature very little breakage wasobserved. There was considerable variation in the effects of different preparationsof this enzyme on the giant loops of chromosome II, early observations suggestingthat these loops are resistant, though later observations were in conflict. We cannotplace much reliance on the results with Hind, in particular because our preparationsprobably contained more than one endonuclease; preservation of loops would havebeen significant even so. We were, however, able to demonstrate visually the depen-dence of Hind on magnesium, for in the absence of MgCl2 in the incubation mediumno loop cutting occurred.

Hae produced by far the most impressive destruction of lampbrush chromosomes,and at room temperature. Moreover, in the presence of this endonuclease, there wasno question about the resistance to digestion of the giant loops of chromosome II(see Fig. 3 A, B). This differential effect was observed on at least 50 occasions, and asa result of digestion by samples of several different preparations of Hae.

A few lampbrush preparations from another urodele, Triturus cristatus cristatuswere also exposed to digestion by Hae. Apart from a few very refractile loops onchromosomes X and occasionally also on chromosome I, which are known to resistdigestion by pancreatic DNase, Hae shattered all the lampbrush lateral loops of thisspecies, and extensive breakage of main chromosome axes also took place.

DISCUSSION

The finding that the giant loops on lampbrush chromosome II of TV. viridescens arerefractory to digestion by Hae, though not to other nucleases, must mean that thesensitive sequence 5' GGCC is systematically absent from the DNA axes of theseloops.

In a random sequence of bases in a DNA duplex, the chance of occurrence of thesite sensitive to Hae is once in every 256 base pairs, (1/4)4, a sequence occupyinga length of just under o-i /tm. Each giant loop on chromosome II is a few hundredmicrometres long. This being so, we must conclude that the DNA axes of the giantloops consist of a large number or repeats of a unit sequence which is not likely to bemore than 3 or 4 times 256 base pairs.

Some of the simple sequence satellite DNA of Drosophila melanogaster is refractoryto Eco RI, Hind and Hae (Hamer & Thomas, 1975), and there are other recent findingsof a like nature (Sommerville & Malcolm, personal communication). The DNA of thegiant loops may also perhaps consist of a manifold-repeated satellite sequence, but ifso this satellite differs from all others so far investigated (Walker, Flamm & McLaren,1969) in that, at least in oocytes, it is transcribed into RNA.

RNA transcription on the giant loops of chromosome II of N. viridescens is currentlybeing studied by Hartley & Callan. Preliminary autoradiographic evidence indicatesthat whereas cytidine, adenine and uridine are all incorporated into this particularloop's RNA, incorporation of guanosine is not detectable. If this finding is substan-tiated, and there is no reason to doubt the early observations, there must be a majorimbalance between the bases on the transcribed and the non-transcribed strands, all

Actions of restriction endonucleases 313

the cytosine residues being confined to the latter. In the light of this observation,the resistance of the giant loops on chromosome II of N. viridescens to digestion byHae is not in the least surprising. Whether other lateral loops will prove to be resistantto digestion by other sequence-specific endonucleases is a question for the future.

REFERENCES

CALLAN, H. G. & LLOYD, L. (i960). Lampbrush chromosomes of crested newts Trituruscristatus (Laurenti). Phil. Trans. R. Soc. Ser. B 243, 135-219.

GALL, J. G. (1954). Lampbrush chromosomes from the oocyte nuclei of the newt. J. Morph.94, 283-352.

GALL, J. G. (1963). Kinetics of deoxyribonuclease action on chromosomes. Nature, Land. 198,36-38.

HAMER, D. H. & THOMAS, C. A. JR. (1975). The cleavage of Drosophila melanogaster DNA byrestriction endonucleases. Chromosoma 49, 243-267.

HILL, R. J., MAUNDRELL, K. & CALLAN, H. G. (1974). Nonhistone proteins of the oocytenucleus of the newt. J. Cell Set. 15, 145-161.

KELLY, T. J. JR. & SMITH, H. O. (1970). A restriction enzyme from Hemophilus influenzae.II. Base sequence of the recognition site. J. molec. Biol. 51, 393-409.

KLEIN, H. A. & CAPECCHI, M. R. (1971). Polypeptide chain termination: purification of therelease factors, Rx and Ra, from Escherichia coli. J. biol. Chem. 246, 1055-1061.

MACGREGOR, H. C. & CALLAN, H. G. (1962). The actions of enzymes on lampbrush chromo-somes. Q. Jlmicrosc. Set. 103, 173-203.

MESELSON, M. & YUAN, R. (1968). DNA restriction enzyme from E. coli. Nature, Lond. 217,1110—1114.

MiDDLETON, J. H., EDCELL, M. H. & HUTCHISON, C. A. (1972). Specific fragments of 0X174deoxyribonucleic acid produced by a restriction enzyme from Haemophilus aegyptius, Endo-nuclease Z. J. Virol. 10, 42-50.

ROULLAND-DUSSOIX, D. & BOYKR, H. W. (1969). The Escherichia coli B restriction endonuclease.Biochim. biophys. Ada 195, 219-229.

ROY, P. H. & SMITH, H. O. (1973). DNA methylases of Hemophilus influenzae Rd. I. Purifica-tion and properties. J. molec. Biol. 81, 427-444.

SHACK, J. & BYNUM, B. S. (1964). Interdependence of variables in the activation of deoxy-ribonuclease. I. J. biol. Chem. 239, 3843-3848.

SHORTMAN, K. & LEHMAN, I. R. (1964). The deoxyribonucleases of Escherichia coli.VI. Changes in enzyme levels in response to alterations in physiological state. J. biol. Chem.239, 2964-2974.

SMITH, H. O. & WILCOX, K. W. (1970). A restriction enzyme from Hemophilus influenzae.I. Purification and general properties. J. molec. Biol. 51, 379—391.

THOMAS, C. A. JR. & DOTY, P. (1956). The mild acidic degradation of desoxyribose nucleicacid. J . Am. Chem. Soc. 78, 1854-1860.

WALKER, P. M. B., FLAMM, W. G. & MCLAREN, A. (1969). Highly repetitive DNA in rodents.In Handbook of Molecular Cytology (ed. A. Lima-de-Faria), pp. 53-72. Amsterdam andLondon: North-Holland Publishing.

{Received 9 January 1976)