In vitro excision of adeno-associated virus DNA from recombinant plasmids


Material Information

In vitro excision of adeno-associated virus DNA from recombinant plasmids isolation of an enzyme fraction from HeLa cells that cleaves DNA at polypurine-polypyrimidine sequences rich in GC base pairs
Physical Description:
ixx, 153 leaves : ill. ; 29 cm.
Gottlieb, H. Jonathan
Publication Date:


Subjects / Keywords:
Genetic Engineering   ( mesh )
Dependovirus   ( mesh )
DNA, Recombinant   ( mesh )
Immunology and Medical Microbiology thesis Ph.D   ( mesh )
Dissertations, Academic -- Immunology and Medical Microbiology -- UF   ( mesh )
bibliography   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph.D.)--University of Florida, 1988.
Bibliography: leaves 145-152.
Statement of Responsibility:
by H. Jonathan Gottlieb.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001037379
oclc - 20674908
notis - AFB9806
System ID:

Full Text







To my family

Digitized by the Internet Archive
in 2011 with funding from
University of Florida, George A. Smathers Libraries with support from LYRASIS and the Sloan Foundation


I would like to express my sincerest appreciation to my mentor and

friend, Dr. Nicholas Muzyczka, for the unique opportunity to work with

him during the last several years. In addition, I would like to thank

the members of my graduate committee at the University of Florida and

Drs. Kenneth Berns and William Holloman for their helpful suggestions

and encouragement.

I am grateful to members of the Muzyczka laboratory in New York

and in Florida for their advice, conversation and friendship. Special

thanks are extended to Deirdre Lawe and to my co-defendant, Susan




ACKNOWLEDGMENTS.................................................. iii

LIST OF TABLES ......................... ... ...... ......... ........ vi

LIST OF FIGURES.................................................. vii

KEY TO ABBREVIATIONS............................................. ix

ABSTRACT ......................................................... xi


I. INTRODUCTION............................................. 1

Introduction...................................... 1
AAV DNA Replication ............................... 1
Rescue of Latent AAV Provirus...................... 4
General Aim of the Studies........................ 6

II. MATERIALS AND METHODS .................................... 7

Cell Culture and Viruses........................... 7
Recombinant Clones and Enzyme Substrates........... 7
Purification of Endo R............................ 10
Nuclease Assays................................... 11
DNA Labeling, Sequencing and
Primer Extension.............................. 12
Gel Electrophoresis.............................. 14
Estimation of the Isoelectric Point............... 15
Chromatographic Media, Enzymes
and Materials .......................... ......... 15
Other Methods...................................... 16
Estimate of In Vivo Excision Frequency.............. 16

III.PURIFICATION OF ENDO R................................... 17

Detection of a Site-Specific Endonuclease
Activity in Crude Nuclear Extracts............ 17
Purification of Endo R............................ 20
The Endo R Agarose Gel Assay....................... 25
Molecular Size and Purity of Endo R............... 25


IV. RESCUE OF AAV SEQUENCES .................................. 35

Localization of the Site of Cleavage to
the AAV/Vector Junction...................... 35
Either AAV/Vector Junction Can Be Cleaved
Independently of the Other................... 38
Cleavage of AAV Terminal Mutants.................. 39
Minor Endo R Cleavage Sites........................ 46
AAV Excision In Vivo.............................. 48
Replication of Endo R Products.................... 51

V. ENZYME RECOGNITION........................................ 53

Isolation of the Cleavage Site ................... 53
Sequence at the Site of Cleavage.................. 63
Is Endo R a Site Specific Single-
Stranded Nuclease.............................. 76

VI. PROPERTIES OF ENDO R..................................... 85

Reaction Conditions for Double-Stranded
Cleavage.................................... 85
Alternate Cleavage Secificity in the
Presence of Mn Ions........................ 89
Endo R Specific Cleavage is of a
Double-Stranded Nature....................... 94
Effect of Nucleotides and Polynucleotides
on the Cleavage Activity..................... 100
Are Stoichiometric Amounts of Enzyme
Required for Cleavage........................ 111
Characterization of the Ends of the
Reaction Products............................. 111
Is Endo R a Human Topoisomerase................... 120

VII.DISCUSSION AND CONCLUSIONS............................... 121

Purification of Endo R............................ 121
Excision of AAV Sequences In Vitro................ 122
Is Endo R Responsible for the Rescue of
AAV Sequences In Vivo....................... 123
Enzyme Recognition ................................ 129
Evidence for a Contaminating Nuclease.............. 137
Other Mammalian Endonucleases..................... 138
What Does Endo R Do for the Cell.................. 140

REFERENCES....................................................... 145

BIOGRAPHICAL SKETCH.............................................. 153


Table Page

I. PURIFICATION OF ENDO R.......................... 21

PLASMIDS....................................... 40

III. MINOR CLEAVAGE SITES............................ 47

IV. SUBCLONED CLEAVAGE SITES........................ 58





Figure Page

1-1. Model for AAV DNA Replication.................... 3

3-1. Endo R Agarose Gel Assay.... ... .................. 19

3-2. Enzyme Titration of Fractions I, II
and III....................................... 23

3-3. Enzyme Titration of Fraction III................. 27

3-4. Sedimentation Analysis of Endo R by
Glycerol Gradient Centrifugation............... 30

3-5. Molecular Weight and Purity of Endo R............ 32

4-1. Identification of AAV-Specific and pBR-
Specific Fragments in the Endo R
Reaction................................... 37

4-2. Cleavage of Substrates That Contain Only
One Copy of the AAV Terminal Repeat............ 42

4-3. Cleavage of AAV Plasmids That Contain
Deletions in the Terminal Repeat............... 45

4-4. Cleavage of AAV Plasmids In Vivo................. 50

5-1. Oligomer Clone Cleavage............................ 55

5-2. Oligomer Clone Gel Assay......................... 57

5-3. Endo R Cleavage of pGA38 and pEV136............... 62

5-4. Sequence at the Site of Cleavage-
Dideoxy Method.................................. 67

5-5. Sequence at the Site of Cleavage-
Maxam and Gilbert Method...................... 72

5-6. Sequence at the Site of Endo R Cleavage-
Summary of the Cleavage Sites.................. 74

5-7. Sl and Endo R Nuclease Digestion of pSM620....... 78

Figure Page

5-8. SI and Endo R Activity on Single-Stranded
DNA.......................................... .82

5-9. Endo R Activity on Single-Stranded
Circular DNA.................................... 84

6-1. Sodium and Potassium Inhibition of Endo R........ 88

6-2. Cleavage of Form I and Form III Plasmid
Substrate.................................... 91

6-3. Specificity of Endo R With Mg2+ and Mn2 ......... 93

6-4. Cleavage Reaction Time Course..................... 96

6-5. Single-Stranded Specificity of Endo R............ 99

6-6. ATP Inhibition of Endo R Activity................ 104

6-7. Polynucleotide Competition of Double-
Stranded Cleavage Activity.................... 107

6-8. Stoichiometry of the Endo R Cleavage
Reaction...................................... 110

6-9. Ligation of Endo R Products...................... 116

6-10. 5' End Labeling of Endo R Products............... 118

7-1. Model for the Excision of AAV DNA from
Recombinant Plasmids.......................... 126



ATP adenosine-5'-triphosphate

bp base pairs

BSA bovine serum albumin

CTP cytosine-5'-triphosphate

d daltons

dATP deoxyadenosine-5'-triphosphate

dCTP deoxycytosine-5'-triphosphate

DEAE diethylaminoethyl

dGTP deoxyguanosine-5'-triphosphate

DNA deoxyribonucleic acid

DTT dithiothreitol

dTTP thymidine-5'-triphosphate

E. coli Escherichia coli

EDTA ethylenediamine tetra-acetic acid

form I superhelical plasmid DNA

form II nicked circular DNA

form III linear duplex DNA

G:C double-stranded poly(dG):poly(dC)

GTP guanosine-5'-triphosphate

kb kilobase

kd kilodaltons

PMSF phenylmethylsulfonylfluoride











polydeoxyadenylic acid

polydeoxycytidylic acid

polydeoxyguanylic acid

polydeoxythymidylic acid

ribonucleic acid

revolutions per minute

sodium dodecyl sulfate

E. coli single-stranded binding protein

telomere terminal transferase

left-handed helical double-stranded DNA

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



H. Jonathan Gottlieb

August 1988

Chairman: Nicholas Muzyczka
Major Department: Immunology and Medical Microbiology

When circular recombinant plasmids containing adeno-associated

virus (AAV) DNA sequences are transfected into human cells, the AAV

provirus is rescued. Using these circular AAV plasmids as substrates, a

cellular site-specific endonuclease, endo R, was isolated from HeLa cell

nuclei on the basis of its ability to excise intact AAV sequences in

vitro from the vector DNA and produce linear DNA products. The enzyme

recognizes and cleaves polypurine-polypyrimidine sequences which are at

least 9 residues long and rich in GC base pairs. Such sequences occur

naturally in non-coding regions of higher eucaryotes, are present in AAV

recombinant plasmids as part of the first 15 bp of the AAV terminal

repeat and, in some cases, are present at the AAV/vector junction as the

result of cloning by GC tailing. Plasmid DNA that is transfected into

tissue culture cells is cleaved in vivo to produce a pattern of DNA

fragments similar to that seen with purified enzyme in vitro. The level

of specific cleavage in the standard assay depends on the type and

length of the recognition signal in the substrate. Generally, longer

and more homogeneous stretches of poly(dG):poly(dC) are better

substrates for double-stranded cleavage. The enzyme apparently

recognizes the secondary structure characteristic of G-rich polypurine

sequences in duplex DNA. Sequence analysis of the cut site suggests

that double-stranded cleavage occurs through a series of concerted

single-stranded nicks of the substrate throughout the enzyme recognition


The molecular weight of the active form of endo R was estimated by

gel filtration and sedimentation in glycerol gradients to be

approximately 125,000. The enzyme, which requires magnesium as a

cofactor, does not have significant endonucleolytic activity on single-

stranded DNA, but is equally active on closed circular and linear duplex

DNA substrates. The properties of endo R are consistent with it playing

a role in the rescue of integrated AAV sequences. The enzyme's function

in cellular DNA metabolism is less obvious, but its properties suggest

that it may have a role in cellular DNA replication and recombination.




A major obstacle in the characterization of cellular factors

involved in mammalian DNA metabolism lies in establishing a

biochemically and genetically defined system. For the most part, the

complexity of cellular processes and the lack of genetic analysis in

higher eucaryotes prevent accurate characterization of these processes.

This problem has been partially circumvented with the use of well

characterized mammalian viral systems as probes for cellular functions

(Challberg and Kelly, 1979; Li and Kelly, 1984; Murakami et al., 1986).

In this regard, adeno-associated virus (AAV) is ideal. AAV has been

well characterized genetically (Hermonat et al., 1984), and because of

its small size, is largely dependent on cellular functions for viral

multiplication (Berns et al., 1985).

AAV DNA Replication

The current model for AAV DNA replication is a modified version of

a general model for the replication of eucaryotic DNA molecules first

proposed by Cavalier-Smith (Cavalier-Smith, 1974; Hauswirth and Berns,

1977; Straus et al., 1976). AAV contains a 4.7 kb single-stranded

linear genome consisting of an internal nonrepetitive sequence which is

flanked at each end by inverted terminal repeats that are palindromic

(Lusby et al., 1980). DNA replication proceeds by leading strand

Figure 1-1. Model for AAV DNA Replication.

The model for AAV DNA replication is a modified version of the model for
eukaryotic DNA replication first described by Cavalier-Smith (1974). i)
Single or double-stranded specific cleavage at the AAV/pBR322 junction
is necessary to allow formation of the hairpin primer in the palindromic
region of the termini. Site specific cleavage must occur at the other
AAV/vector junction either before or during DNA synthesis to separate
AAV and vector sequences. ii) DNA replication proceeds from the 3' end
of the primer to the other end of the molecule and includes synthesis of
the distal terminal repeat. iii) A specific nick is required on the
parental strand directly across from the original 3' end of the primer
to allow repair of the parental terminal sequences (iv). Both the
parental and progeny strands are now capable of initiating another DNA
replication cycle.

\\ //
SpBR //


Elongation and
Strand Displacement

iii ................................

SRF Resolution

synthesis from the hairpin primer formed in the palindromic region of

the terminal repeat (Figure 1-1). Strand elongation proceeds from the

3' end of the primer, in a 5' to 3' direction, to the other end of the

molecule. A specific nick is required in the parental strand at a point

directly across from the original 3' terminal base to allow repair of

the parental terminal sequences and to produce a full-length duplex

replicative intermediate. Subsequently, both the parental and progeny

strands are capable of initiating a second round of DNA replication and

displacing a progeny single-stranded DNA molecule.

Rescue of Latent AAV Provirus

AAV is a defective parvovirus and requires the presence of a co-

infecting helper virus for a productive viral infection. Virtually any

member of the adenovirus or herpes virus family can supply the helper

functions (Atchison et al., 1965; Buller et al., 1981; McPherson et al.,

1985). In the absence of a helper virus, AAV readily integrates into

the host DNA (Hoggan et al., 1972; Berns et al., 1975; Handa et al.,

1977), either as a single proviral copy or more often as a tandem head

to tail array of several AAV genomes (Cheung et al., 1980; Laughlin et

al., 1986; McLaughlin et al., 1988). Super-infection of latently

infected cells with helper virus results in the rescue of integrated

provirus and normal AAV replication (Hoggan et al., 1972).

In many proviral cell lines the yield of replicative intermediates

and virus produced by superinfection with helper virus are identical to

those obtained from exogenous AAV infections (Laughlin et al., 1986;

McLaughlin et al., 1988). This suggested that the rescue of AAV

sequences from chromatin is a rapid event initiated by specific cleavage

within the AAV termini. Additional evidence that AAV rescue involves a

site-specific nuclease comes from the study of recombinant AAV plasmids

which, presumably, rescue by a mechanism similar to that used by

proviruses integrated into chromatin (Samulski et al., 1982). When AAV-

pBR322 plasmids are transfected into human cells in the presence of

adenovirus, free linear duplex AAV DNA, the major replicative form

(Straus et al., 1976; Hauswirth and Berns, 1977), is seen within 24

hours (Samulski et al., 1982). However, when viable terminal deletion

mutants are transfected into adenovirus infected cells, the formation of

replicative forms (RF) is delayed (Samulski et al., 1983; Lefebvre et

al., 1984, Samulski et al., 1987), indicating that either gene

conversion of the terminal deletion causes a delay in DNA synthesis, or

that AAV sequences at the termini are required for excision. Studies

with the AAV mutant, psub201+, in which the terminal 13 bp have been

deleted from both ends, suggest that the delay in RF formation is due to

a defect in rescue (Samulski et al., 1987) and that the signal for AAV

excision resides in the terminal sequences. AAV DNA replication in

adenovirus infected cells transfected with psub201+ was eight-fold more

efficient when the AAV sequences were cleaved from the vector sequences

prior to transfection (Samulski et al., 1987). Presumably, cleavage of

the plasmid bypassed the excision step which must normally precede DNA


In principle, the duplex replicative intermediate can be generated

from the covalently closed form I plasmid DNA in one of two ways.

Either the AAV sequences are separated from the plasmid by precise

excision or a single-stranded DNA molecule is generated by AAV-specific

replication. (See Samulski et al., 1983, and Senepathy et al., 1984,

for examples of the second mechanism.) Because the AAV termini are also

the origins for AAV DNA replication (Hauswirth and Berns, 1977; Samulski

et al., 1983; Senepathy et al., 1984), it is difficult to distinguish

between the two mechanisms. Both mechanisms, however, require either a

specific nick or a double-stranded cut at an AAV/vector junction in the

input plasmid DNA (Figure 1-1). Thus, it should be possible to use AAV

recombinant plasmids as substrates for an in vitro excision assay and to

screen cellular extracts for an activity that produces either single-

stranded nicks or double-stranded cuts at the AAV/vector junctions.

General Aim of the Studies

Using the recombinant AAV plasmids as substrates, an enzyme

fraction was isolated from HeLa cell nuclear extracts that rescued

intact AAV DNA from vector DNA in vitro and produced linear DNA

products. The double-stranded cleavage activity has been called endo R

because of its ability to rescue AAV sequences. The activity was of

cellular origin and was stimulated approximately 5 fold by Ad2 infection

in the presence of hydroxyurea.

In the following chapters, the purification and properties of the

enzyme are discussed and the activity observed in vitro is compared with

what is observed in vivo. In addition, the behavior of the enzyme

suggests that it may be involved in cellular DNA metabolism, and these

properties are discussed as well.


Cell Culture and Viruses

HeLa S3 cells were maintained at 370C in suspension culture in

Eagles minimal essential medium supplemented with 5% calf serum, 1%

glutamine, penicillin and streptomycin. Wild type adenovirus 2 (Ad2)

was prepared from a freeze/thaw lysate of HeLa S3 cells as previously

described (Samulski et al., 1983). HeLa monolayer cells were

transfected with 5.0 pg DNA by the DEAE-dextran method (McCutchan and

Pagano, 1968) as described (Muzyczka, 1980), and infected with Ad2 virus

at a multiplicity of infection (moi) of 10. Low molecular DNA was

isolated from the cells by the method of Hirt (1967) as described

(Muzyczka, 1980).

Recombinant Clones and Enzyme Substrates

All recombinant clones were maintained in either the recA host

HB101 (Boyer, 1969) or the recBC, recF, sbcB host JC8111 (Boissy and

Astell, 1985), a gift from Peter Tattersall. This was to prevent

variation in the size of the G:C tails and other repetitive inserts

(Hauswirth et al., 1984) and to prevent deletion of the palindromic AAV

terminal repeat (Samulski et al., 1982).

The AAV wild-type plasmid pSM620 contains a full copy of the AAV

genome inserted by GC tailing into the PstI site of pBR322 (Bolivar et

al., 1979). The plasmid contains 18 bp of poly(dG):poly(dC) (G:C) at

the left pBR322/AAV junction and 28 bp of G:C at the right junction

(Samulski et al., 1983). The AAV terminal mutant clone, pSM703, was

isolated by Samulski et al. (1983) and was sequenced as part of this

study. It contains 100 bp deletions in both AAV terminal sequences, 38

bp of G:C tail at the right AAV/pBR junction, and 36 bp G:C at the left

junction. The plasmids pGM620C and pGM620D contain either the left or

the right terminal PstI fragment, respectively, of the wild-type AAV

plasmid, pSM620 (Samulski et al., 1983), subcloned into the PstI site of

pBR322 (Bolivar et al., 1979). Both clones contain the original

poly(dG)-poly(dC) tails that are present in pSM620. The plasmid pGM1008

was constructed by removing the PstI-BssHII fragment (AAV nucleotides

4254 to 4657) from pGM620D and religating the molecule after producing

blunt ends with the Klenow fragment of DNA polymerase I. The plasmid

contains the parental poly(dG)-poly(dC) tail plus the terminal 21 bp of

the AAV sequence up to the BssHII site at nucleotide 4657.

The plasmids pGM913, pGM1116, pGM1505, pGM1635, pGM1228, pGM1344

and pGM1483 were constructed by inserting chemically synthesized

oligonucleotides (Systec) into a unique restriction site in pBR322. To

construct these clones, 2 pmol of single-stranded oligomer was boiled

with an equal amount of the appropriate complementary strand in 20 1l of

water for 5 minutes. The reaction was then cooled slowly to room

temperature to allow the formation of duplex molecules. Ligation

reaction mixtures containing 2 pmol of annealed oligomer, 0.2 pmol of

PstI linearized pBR322 and 400 units of T4 DNA Ligase (New England

BioLabs) were incubated at 150C for 12-24 hours. The 5' ends of the

single-stranded oligomers were phosphorylated only when multiple tandem

copies of the insert were desired. Dephosphorylation of the vector was

generally not required under these conditions. Positive clones retained

tetracycline resistance, but were sensitive to ampicillin.

The plasmids PGM913 and pGM1116 contain 13 and 9 base pairs of G:C

homopolymer, respectively. PGM1635 contains 20 bp of alternating

copolymer GC, and pGM1505 contains the Dictyostelium telomeric repeat

sequence (C2_6T)5 (Shampay et al., 1984). Two plasmids contain the

sequence of the terminal 21 bp of AAV (ggCCaCTCCCTCTCTgCgCgC), in either

a monomer form (pGM1228) or as an inverted dimer (pGM1344). All of the

above clones were constructed by inserting the oligomers into the PstI

site of pBR322. Additionally, the plasmid pGM1483 contains an insert of

the Tetrahymena telomeric sequence, (C4A2)3 (Blackburn and Szostak,

1984), chemically synthesized and inserted into the EagI site of pBR322.

In each case, both strands of the oligonucleotides were synthesized so

that the restriction enzyme recognition site was maintained after the

two strands of each oligonucleotide were annealed and ligated into

pBR322. The inserts and pBR322 flanking sequences of these plasmids

were confirmed by DNA sequencing.

The plasmid pGA38, a gift from Todd Evans, contains an insert of

the alternating co-polymer (GA)38, cloned into the EcoRI site of pUC9

(Evans and Efstratiatis, 1986). PGAll is a subclone of pGA38 and

contains an insert of (GA)11 cloned into the EcoRI site of pUC9. The

plasmid pEV136, a gift from Eckhardt Wimmer, is an infectious polio

clone and contains an 18 bp G:C insert at the EcoRI site, as well as an

84 bp stretch of poly(dA):poly(dT) (Semler et al., 1984).

Purification of Endo R

All operations were carried out at 0-40C. To minimize the effects

of proteolysis, all buffers contained 0.1 mM phenylmethylsulfonyl-

fluoride (PMSF) and the time between chromatographic steps was kept to a

minimum. Nuclei from eight liters of adenovirus infected HeLa S3 cells,

grown in the presence of 10 mM hydroxyurea (Sigma Chemical Co.), were

isolated as described by Challberg and Kelly (1979). Frozen nuclei were

thawed on ice and 5 M NaC1 was added to a final concentration of 0.2 M.

After incubation on ice for 2 hours, the nuclei were pelleted by

centrifugation at 3000 x g for 15 minutes, the pellet was discarded and

the supernatant was dialyzed against 25 mM Tris-HC1 pH 8.0, 10 mM KC1,

0.1 mM EDTA, 5 mM 2-mercaptoethanol, 0.1 mM PMSF and 20% glycerol

(Buffer A). Any insoluble precipitate in the nuclear extract was

removed by centrifugation at 12000 x g for 20 minutes and the

supernatant was retained as fraction I. Fraction I contained 34.2 mg/ml

protein, as determined by the Bradford protein concentration assay

(Bradford, 1976), in a volume of 25 ml.

Fraction I was loaded onto a 15 ml DEAE cellulose column

previously equilibrated with buffer A, pH 8.0. The column was then

washed with 3 column volumes of buffer A and eluted in the same buffer

with a 150 ml linear gradient containing 0.01 to 0.5 M KC1. Fractions

emerging between 0.1 and 0.25 M KC1 were pooled to form fraction II and

contained 2.7 mg/ml protein a total volume of 60 mls. Fraction II was

dialyzed against two 1-liter changes of buffer B (10 mM sodium

phosphate, pH 6.5, 5 mM 2-mercaptoethanol, 10% glycerol and 0.1 mM PMSF)

containing 0.01 M NaC1 and loaded onto a 7 ml phosphocellulose column

pre-washed with 21 mls of the same buffer. The column was eluted with a

70 ml linear gradient from 0.01 to 1 M NaCI in buffer B. Active

fractions emerged about halfway through the gradient and were pooled to

form fraction III. Fraction III contained 0.16 mg/ml protein in 45 ml

and had a specific endo R activity of approximately 3000 units/mg.

Fraction III was split into 2 aliquots and each aliquot was concentrated

about 55 fold using an Amicon Centriflo 25 spin concentrator, yielding a

final protein concentration of 18 mg/ml in 2.0 ml. The two 1.0 ml

aliquots of concentrated fraction III were passed separately through a

36 ml (1.5 x 45 cm) column of Sephadex G200 equilibrated with buffer B,

pH 6.5, 250 mM NaC1. Fractions containing endo R activity were pooled

(fraction IV), dialyzed against buffer B containing 0.01 M NaC1 and

loaded onto a 3 ml poly(dG) agarose column equilibrated in buffer B, 10

mM NaC1. The column was washed with 3 bed volumes of the same buffer

and eluted with a 30 ml linear gradient from 0.01-1.0 M NaC1. Active

fractions were pooled (fraction V) and dialyzed into 25 mM Tris-HCl, pH

7.5, 50% glycerol, 0.1 mM DTT. Fraction V was stable for at least 2

months at -200C without measurable loss of activity.

Nuclease Assays

The standard endo R reaction mixtures of 25 pl contained 20 mM

Tris-HCl (pH 7.5), 5 mM MgCl2, 1 mM dithiothreitol, 0.2 pmol form I

plasmid substrate and 0.1-5.0 units of endo R (see below). After

incubation at 370C, the reaction products were phenol extracted, ethanol

precipitated and digested with either BstEII at 60C for 1 hour or with

SphI at 37C for 3 hours. The reaction products were then fractionated

by electrophoresis on 1.4% agarose gels. One unit of endo R activity is

defined as that amount of protein that is required to cleave 50% of

pGM620D substrate under standard conditions. During early stages of the

purification, it was necessary to treat the reaction products with 0.25

mg/ml proteinase K (Sigma) for 1 hour at 370C prior to phenol extraction

and gel electrophoresis. When indicated, endo R cleavage reactions were

terminated with the addition of an equal volume of a 2x proteinase K

stop solution, containing 0.5 mg/ml Proteinase K, 1% SDS and 20 mM EDTA.

After addition of the stop solution, incubation was continued at 370C

for at least 1 hour.

Single-stranded nuclease assays contained 5.0 Ag of heat-denatured

3H-labeled E. coli chromosomal DNA (1 x 105 cpm/pg) in 250 Al. For SI

nuclease, the mixtures contained 50 mM sodium acetate (pH 4.5), 0.3 M

NaCI, 10 mM ZnCl2, and 1.25 units of Sl, while the endo R reaction was

under standard reaction conditions. Twenty-five microliter portions

were removed at the indicated times and added to 2ml of 10%

trichloroacetic acid (TCA) containing 200 pg/ml BSA. The precipitates

were collected on nitrocellulose filters (Schleicher and Schuell type

BA85), washed with TCA and ethanol and then counted. E.coli DNA was

labeled with 3H-thymidine (6.7 Ci/mmole, ICN) and isolated as described

(Holloman et al., 1981).

DNA Labeling. Sequencing and Primer Extension

Nucleotide sequences were determined by the method of Maxam and

Gilbert (1977) and by the dideoxy method (Sanger et al., 1977). For

Maxam and Gilbert sequencing of the pGM clones, 10 pg of DNA was

linearized with Scal, dephosphorylated with calf intestine alkaline

phosphatase (CIAP) and labeled at the 5' end with polynucleotide kinase

and 732P-ATP or labeled at the 3' end with the Klenow fragment, a32p

dCTP and 80 /M each of dGTP, dATP and TTP. Five micrograms of the

labeled DNA was further digested with BamHI and the 5' or 3' labeled

ScaI-BamHI fragments were isolated from 1.2% low melting agarose prior

to sequencing. Endo R cleavage fragments were prepared by incubating

the remaining 5 pg of 5' or 3' labeled DNA with 10 units of fraction V

endo R under standard reaction conditions and isolating the labeled endo

R/ScaI fragments from a 4% non-denaturing acrylamide gel as described

(Maniatis et al., 1982). The sequence at the site of cleavage was

determined directly from the sequencing gels by comparing the mobility

of the endo R fragments and the sequence markers.

For primer extension of endo R products, 2 pg of plasmid DNA was

incubated with endo R in a standard reaction mixture of 50 pl. The form

III product was then isolated from 1.2% low melting agarose and

dissolved in 10 pl of water and 4 pl of 5x reaction buffer (0.3 M Tris-

HC1, pH 8.3; 0.375 M NaCl; 37.5 mM MgC12 and 2.5 mM DTT). The reaction

mixtures were divided into two 7 pl portions and 1 pl of either upper or

lower strand pBR322/PstI primers (2 pmol/pl) was added to each portion.

The solutions were heated to 1000C for 5 minutes under parafin oil and

immediately frozen in a dry ice/ethanol bath. After thawing on ice, 1

1l each of stock solutions containing 0.5 mM dGTP, 0.5 mM dCTP, 0.5 mM

TTP, 0.02 mM dATP; 10 pCi/l a 32P-dATP (>3000 Ci/mM) and 10 units/pl AMV

reverse transcriptase (IBI) were added to each portion to produce a

final reaction volume of 10 p1. After incubation at 420C for 10

minutes, 2 p1 of chase solution (0.25 mM each of dGTP, dATP, dCTP and

TTP) was added and incubation was continued at 420C for an additional 10

minutes. The reaction was terminated with the addition of 7 ~l of stop

buffer containing 1.6 pl 0.25 M EDTA; 2 1l 3 M sodium acetate; 0.2 p1 2

mg/ml tRNA and 3.2 pl water. After ethanol precipitation with 3 volumes

of 95% ethanol, the primer extension products were redissolved in 5 p1

of sequence gel running buffer (Maxam and Gilbert, 1977) and heated to

1000C for 3 minutes prior to gel electrophoresis. The sequence at the

site of endo R cleavage was determined by comparing the mobility of the

primer extension products with dideoxy sequencing ladders produced using

identical single-stranded primers.

Gel Electrophoresis

Electrophoresis in the presence of sodium dodecyl sulfate was done

according to Laemmli (1970) in slab gels of 7.5% acrylamide, 0.2% N,N'-

methylene-bisacrylamide. The gels were fixed in 50% methanol overnight

and silver stained as described (Switzer et al., 1979).

Neutral polyacrylamide and agarose gel electrophoresis of DNA was

as described (Muzyczka, 1980). DNA fractionated on agarose gels were

transferred to nitrocellulose filters (Southern, 1975) and hybridized to

32P-nick translated DNA probe (>2x108 cpm/pg) as described (Muzyczka,


Denaturing polyacrylamide-urea gels were as described (Maxam and

Gilbert, 1977). All sequencing gels were run at 55-600C to prevent

compression of palindromic sequences (Lusby et al., 1980).

For denaturing agarose gels, 1.5% horizontal agarose gels were

poured in 50mM NaCI and allowed to equilibrate in excess running buffer

(30 mM NaOH, 1 mM EDTA) for at least 1 hour. Samples were incubated for

5 minutes in gel running buffer (Maniatis et al., 1982) containing 0.2 N

NaOH prior to electrophoresis at 30 volts for at least 24 hours.

Estimation of the Isoelectric Point

Estimation of the pi of endo R was performed by a modified version

of the method by Yang and Langer (1987). Eight 100 pC aliquots of

packed DEAE-cellulose (Whatman DE32) was equilibrated in 0.1 M Tris-HCl

at pH 7.0, 7.5, 8.0 or 8.5 or in 0.1 M 2, 2'-Bis(hydroxymethyl)-

2,2',2"-nitrilo triethanol (Bis-Tris) at pH 5.5, 6.0, 6.5 or 7.0. This

required several changes of buffer until no change in the pH of the

supernatant was observed. After equilibration, the cellulose was packed

by centrifugation at 1000 x g for 1 minute and the supernatant was

decanted. Mixtures of 20 &l of fraction III endo R and 80 p4 of 0.1 M

Tris or 0.1 M Bis-Tris buffer, at the appropriate pH, were added to the

packed resin and the samples were incubated on ice for 30 minutes with

frequent mixing to allow the protein to adsorb to the resin. After

pelleting at 1000 x g for 1 minute, the supernatant was removed and

assayed for endo R double-stranded cleavage activity as described above.

The pi was determined to be equal to the lowest pH at which all the

activity had been adsorbed from the supernatant.

Chromatographic Media. Enzymes and Materials

DEAE-cellulose (DE32) and phosphocellulose (P-ll) were purchased

from Whatman. Sephadex G200, poly(dG) agarose, and calf intestine

alkaline phosphatase (CIAP) were from Sigma and Bio-Gel A1.5M was

purchased from BioRad. All restriction endonucleases, T4 DNA ligase and

polynucleotide kinase were purchased from New England Biolabs. The

labeled nucleotides, a32P-dNTPs and -32P-ATP (>3000 Ci/mmol) were from

New England Nuclear and 3H-thymidine (6.7 Ci/mmol) was from ICN.

Other Methods

Protein concentrations were determined by the procedure of

Bradford (1976). The Biogel A1.5M (45.5 x 0.7 cm) and the Sephadex G200

(45.5 x 1.0 cm) columns were calibrated in 10 mM NaPO4, pH6.5; 0.25 M

NaCI; 10% glycerol and 5 mM 2-mercaptoethanol. The excluded volume was

determined with the use of Blue Dextran 2000 (molecular weight >2

million, Sigma) and the column was calibrated with the use of B-amylase

(200 kd), alcohol dehydrogenase (150 kd), BSA (66 kd), carbonic

anhydrase (29 kd) and cytochrome C (12.4 kd) as molecular weight

standards (Sigma Chem. Co.).

Estimate of In Vivo Excision Frequency

A typical transfection uses 0.1-10 pg of plasmid DNA per 10 cm

dish or approximately 3x109-3x1011 molecules per dish. Assuming 106

competent cells per dish, each competent cell will have 3x103-3x105

molecules of plasmid DNA. This agrees reasonably well with the amount

of plasmid DNA recovered from transfected cells as measured by filter

hybridization. If the in vitro frequency of cleavage by endo R at an

outboard site of AAV is 1%, cleavage at both outboard sites to produce

an intact replicative form would be 0.01%. This would produce 0.3 to 30

molecules of excised AAV per transfected cell.


Detection of a Site-Specific Endonuclease Activity
in Crude Nuclear Extracts

The results from studies of AAV replication in vivo from latently

infected cells and transfected AAV plasmids suggest that rescue from

chromatin and from recombinant plasmids occurs by a similar mechanism

(Samulski et al., 1982) and that the sequences at or near the AAV

termini are essential for this excision activity (Samulski et al.,

1987). Taking these observations into account, the wild type AAV

plasmid, pSM620, was used as a substrate to assay for an activity that

would specifically excise AAV sequences from the plasmid. The plasmid

pSM620 contains the entire AAV genome cloned into the PstI site of

pBR322 by GC tailing (Samulski et al., 1982). To detect site-specific

cleavage at the AAV/vector junction, form I supercoiled DNA was treated

with cell-free extract and digested with a one-cut restriction enzyme

that cleaves either in the AAV sequences (BstEII) or in vector DNA

(SphI). The products of the assay were then fractionated by agarose gel

electrophoresis. Figure 3-1 illustrates the fragments that would be

produced if cleavage occurs at either the left or right AAV/pBR

junction. For example, cleavage with cellular extract at the right

junction would generate a linear plasmid DNA molecule which, after

BstEII digestion, would produce a 3.0 kb fragment which consists

exclusively of AAV DNA and a reciprocal 6.1 kb band that contains both

AAV and pBR sequences.

Figure 3-1. Endo R Agarose Gel Assay.

The figure illustrates the products that would result from partial
digestion by endo R and complete digestion by BstEII (left pathway) or
SDhI (right pathway) when the wild type plasmid, pSM620, is used as
substrate. The thick line represents AAV DNA; the thin line represents
pBR322 DNA; vertical lines represent cleavage sites of BstEII (B), SphI
(S), and endo R (R). The filled squares represent the AAV terminal
sequences. AAV cleavage occurs at the junction between the AAV termini
and the vector DNA.




\ +Sph

1.3 S

6.0 S 3.1
m --.1----

1.3 S 3.1


1.7 3.0



11~ 11~.J__

. 7.4'
L..... .- -....

Using the agarose gel assay, crude extracts prepared from Ad2

infected, Ad2/AAV co-infected and uninfected HeLa cell nuclei were

screened for the ability to cut near the AAV/vector junctions and the

products were visualized by transfer to nitrocellulose filters and

hybridization to AAV specific probe (Southern, 1975). Specific double-

stranded cleavage activity was observed in uninfected and Ad/AAV co-

infected cells at comparable levels. However, a 5-fold stimulation was

observed in extracts from Ad2 infected cells treated with hydroxyurea.

Whole cell extracts and extracts prepared from a 0.2 M NaCI wash of

freeze-fractured nuclei (Challberg and Kelly, 1979) contained

approximately equal amounts of activity, while nuclear extracts usually

contained fewer contaminating nucleases. For these reasons, nuclear

extracts from Ad2-infected HeLa cells treated with hydroxyurea were used

as the primary source of activity.

Purification of Endo R

In the procedure outlined in Table I, crude nuclear extracts

(fraction I) were prepared from Ad2-infected HeLa S3 cells, grown in the

presence of 10mM hydroxyurea, and harvested at 21 hours post-infection

(Challberg and Kelly, 1979). Due to the high level of contaminating

endonucleases and cellular DNA (Figure 3-2), specific double-stranded

cleavage activity could be detected at this stage only by Southern

hybridization (Southern, 1975). Fraction I was applied to a column of

DEAE cellulose and the activity that was eluted from the column was

essentially free of nucleic acid and separated from at least one

nonspecific cellular nuclease which did not bind to the column.

Fraction II endo R activity (and the activity in subsequent fractions)



Fraction Volume Activity Proteina Specific
ml units mg units/mg

I. Nuclear Extract 25.0 60000 855.0 70
II. DEAE-cellulose 60.0 60000 159.6 376
III. Phosphocellulose 47.5 48000 38.5 1247
IV. Sephadex G200 19.8 10000 9.1 1099
V. Poly(dG) agarose 3.6 9000 2.9 3103

protein concentrations were determined by the method of Bradford (1976).

Figure 3-2. Enzyme Titration of Fractions I, II and III.

Standard reaction mixtures of 25 pl contained 0.1 pmol of pSM620 form I
plasmid and the indicated amounts (pl) of nuclear extract (NE, Fraction
I), DEAE-cellulose fractions (DEAE, Fraction II) or phosphocellulose
fractions (Pcell, Fraction III). Reactions were terminated at 1 hour
with the addition of proteinase K stop solution (see Chapter II), phenol
extracted, ethanol precipitated and digested with BstEII. The samples
were fractionated on a 1.4% agarose gel, transferred to a nitrocellulose
filter and hybridized to AAV-specific probe.

-- NE --I--*-DEAE---- I Pcell --
1.0 2.5 5.0 7.5 0.5 1.0 2.5 5.0 7.5 0.3 0.5 1.0 2.5 5.0 7.5 10


Sii -3.0

I II -


could be detected directly by ethidium bromide staining of the gel assay

(Figure 3-2). Active fractions from DEAE cellulose chromatography

usually contained an overlapping peak of nonspecific nucleolytic

activity. Further fractionation on a column of phosphocellulose

produced an activity (fraction III) that exhibited substantially more

specific cleavage with little evidence of contaminating nucleolytic

activity at low enzyme concentrations (Figure 3-2). Nevertheless, high

levels of fraction III enzyme resulted in non-specific degradation of

the DNA substrate (data not shown). In order to separate the remaining

contaminating nucleolytic activities from endo R, fraction III was

further purified by chromatography on Sephadex G200 (fraction IV) and

poly(dG) agarose (fraction V). Endo R activity emerged from the

Sephadex G200 column just behind the void volume. This step resulted in

a substantial loss of total endo R activity, but has been retained as

part of the purification protocol for qualitative purposes. Endo R

emerges from the final column, poly(dG) agarose, as a single symmetrical

peak, when eluted with a linear increasing salt gradient, with little

additional loss of activity.

Fraction IV endo R is stable for at least 2 months at

-200C without measurable loss of activity. The stability of the enzyme

is greatly reduced when the concentration of protein falls below 200

pg/ml. For this reason, fraction V enzyme is dialyzed into buffer

containing 50% glycerol and supplemented with 200 pg/ml enzyme grade BSA

for periods of storage of up to 2 months at -200C.

The Endo R Agarose Gel Assay

For the purposes of the gel assay, endo R can be thought of as a

restriction enzyme with two sites in pSM620 (Figure 3-1). The products

of endo R digestion can be seen as the products of a partial digestion

of the substrate at either of the AAV/vector junctions. Further

digestion of endo R products with the restriction enzyme BstEII, which

cuts once asymmetrically in the AAV sequences, produces 6.1 and 3.0 kb

fragments from endo R cleavage at the right AAV/pBR322 junction and 7.4

and 1.7 kb fragments from cleavage at the left junction. The formation

of endo R products, using the plasmid pSM620 as substrate, is

illustrated in figure 3-3 with the titration of partially purified

enzyme (fraction III). As the reaction nears completion, most of the

AAV/vector junctions are cleaved and produce the fragments expected from

cleavage by BstEII and endo R. Linear fragments (form III) are the

result of BstEII digestion of unreacted substrate. In further

discussion, the production of the 3.0 and 1.7 kb fragments, which are

easily monitored, is used to quantitate the amount of cleavage at the

right and left junctions, respectively.

Molecular Size and Purity of Endo R

Estimates of the molecular weight of the native protein were made

by chromatography of endo R on gel filtration columns and by

centrifugation in glycerol gradients. The peak of endo R activity

emerges from a Bio-Gel A1.5M column (see Chapter II: Materials and

Methods) behind the 150 kdal alcohol dehydrogenase molecular weight

marker with an apparent molecular weight of 120-125 kdal (data not

shown). This is in good agreement with the observed behavior in a

Figure 3-3. Enzyme Titration of Fraction III.

The numbers at the top refer to the number of units of endo R used in a
each reaction (see Chapter II: Materials and Methods). The sizes of the
major cleavage fragments are indicated on the right. Standard endo R
reaction mixtures (25 Ip) contained 0.5 pg of form I pSM620 plasmid DNA,
5 mM MgC12, 20 mM Tris-HCl (pH 7.5), 1 mM DTT, and the indicated amount
of fraction III enzyme. After incubation for 1 hour at 370C, the
reaction products were treated with phenol, precipitated with ethanol,
and digested with BstEII. The reaction mixture was then fractionated on
a 1.4% agarose gel, transferred to a nitrocellulose filter and
hybridized to nick-translated P-AAV DNA.

.13 .25 .50 1.0 2.0 3.0

*1 -IP



ii B

-0 g -1.7

glycerol gradient (Figure 3-4A), in which the peak of double-stranded

cleavage activity sedimented slightly more slowly than the 150 kdal

marker. The sedimentation results correspond to an empirically

determined sedimentation coefficient (S20,W) of 6.8 (Figure 3-4B) and a

molecular weight of 115 kd, assuming a spherical structure of the


Pooled fractions from the poly(dG) agarose column, fraction V

(Figure 3-5A, fractions 15-21), contain one major and 4 minor protein

bands when analyzed by SDS polyacrylamide gel electrophoresis and silver

staining. However, when individual fractions across the peak of endo R

activity were analyzed on SDS gels (Figure 3-5B) and compared with the

activity profile in the agarose gel assay (Figure 3-5A), the pattern of

the major 100,000 molecular weight band was most consistent with the

profile of specific cleavage activity. In addition, two other proteins

of minor intensity, with molecular weights of 56 and 89 kd, appeared in

a subset of the active fractions in a pattern not inconsistent with the

activity profile (Figure 3-50). As a result, it is possible that these

peptides are responsible for the cleavage activity or contribute either

as minor subunits or accessory peptides. It is also possible that the

56 and 89 kd peptides are proteolytic breakdown products of the major

100 kd protein. Attempts to purify endo R further or to isolate an

active form from a non-denaturing acrylamide gel have been unsuccessful,

probably due to the instability of the protein at low concentrations or

to the loss of subunits and/or accessory proteins.

Figure 3-4. Sedimentation Analysis of Endo R by Glycerol Gradient

A. Glycerol gradients (20-40%), in 0.1 M Tris-HC1, pH 7.5, 0.1 mM DTT
were prepared in 5 ml polyallomer ultracentrifuge tubes. One milligram
of fraction III endo R was combined with 0.25 mg of alcohol
dehydrogenase, in 0.2 ml 0.1 M Tris-HCl, pH 7.5, 0.1 mM DTT. The
mixture was layered on top of a glycerol gradient and the tube was
centrifuged for 30 hours at 45,000 rpm and 40C in a Beckman SW 50.1
rotor. Three other gradients containing endo R fraction III alone, a
mixture of the molecular weight standards B-amylase (200 kd) and BSA (66
kd), or alcohol dehydrogenase (150 kd) were run in parallel. Fractions
were collected from the bottom of the tube and the absorbance at 280nm
was determined (open circles). Fractions from gradients containing endo
R were also assayed for double-stranded cleavage activity using the
standard excision assay (filled squares). Sedimentation is from right
to left. B. The sedimentation coefficient (S20,W) of endo R was
determined from a plot of the mobility of each protein, represented as a
fraction of the total gradient length, against the S20 W literature
values (CRC Handbook of Biochem., 1979) of the following protein
standards: B-amylase (sweet potato, 8.9), alcohol dehydrogenase (yeast,
7.61), bovine serum albumin (4.41).

j I 0 1 I i I- a -a
0 5 10 15 21

Fraction Number

0.5 0.6 0.7 0.8 0.9










200 150 Endo R
1, 1 1









Endo R

3 4-

Figure 3-5. Molecular Weight and Purity of Endo R.

A) Agarose gel assay of peak fractions from poly(dG) agarose (Fraction
V). Reaction mixtures of 25 pl containing 0.1 pmols of BstEII digested
pSM620 DNA were incubated with 5 pl of the fraction indicated under
standard reaction conditions. The reactions were stopped after 1 hour
with the addition of 1/10 volume of gel running dye containing 0.2 M
EDTA and 1% SDS and then fractionated on a 1.4% agarose gel. B) SDS
polyacrylamide gel electrophoresis of peak fractions from poly(dG)
agarose. A constant volume (20 Il) of the fractions indicated were
added to an equal volume of 2X sample buffer, heated to 1000C for 3
minutes and separated on an SDS acrylamide gel as described in Chapter
II: Materials and Methods. A parallel slot contained the following
protein standards: B-amylase (205 kd), alcohol dehydrogenase (116 kd),
phosphorylase b (97 kd), bovine serum albumin (66 kd), egg albumin (45
kd), and carbonic anhydrase (29 kd). C) Molecular weights of the
peptides in Fraction V endo R. Poly(dG) agarose fractions 15-23 were
pooled to form fraction V and the molecular weights of the protein
species present were determined from their relative mobilities on an SDS
gel. Open circles represent the molecular weight standards described
for B; filled circles are endo R fraction V proteins.

23 25 2729 31

Fm3 -

4.4 -
3.0 -

1.7 -

11 13 15 17 1

11 13 15 17 19 21 23 25 27 29 31 M

- 205

- 116

-m -- -

- 97

- 66

- 45

Figure 3-5 (continued), part B

100 89

o MW Markers
56 Fraction V

20 1 I-
0.0 0.5 1


Figure 3-5 (continued), part C



Localization of the Site of Cleavage to the AAV/Vector Junction

The excision of AAV sequences by a site specific endonuclease must

be rather precise in order to produce substrate forms that are capable

of replication without further modification. While minor deletions of

the terminal sequences incurred during rescue can be repaired through

gene conversion (Samulski et al., 1983), a major deviation in the site

of cleavage would produce replicative forms that either lacked the

ability to self-prime, or contained extraneous vector DNA that would

have to be further processed prior to DNA replication. Therefore, an

enzyme postulated to be responsible for AAV rescue in vivo must take

these constraints on the site of cleavage into account.

To demonstrate that the site of cleavage was at the AAV/pBR322

junction, the products of the endo R reaction were hybridized to AAV-

and pBR322-specific probes (Figure 4-1). Specific cleavage occurring

precisely at the AAV/vector junction would produce a subset of fragments

that consisted of AAV or pBR322 sequences exclusively, and these could

be identified by probing duplicate Southern blots (Southern, 1975) with

AAV or pBR322 DNA (Figure 4-1, right panel). In this experiment,

reaction conditions were chosen to cleave 90% of the substrate (form I

pSM620) at least once. When the products of the reaction were examined

by ethidium bromide staining without further digestion with a

restriction enzyme, most of the product formed was linear plasmid DNA

Figure 4-1. Identification of AAV-Specific and pBR-Specific Fragments
in the Endo R Reaction.

PSM620 substrate was incubated with fraction IV enzyme in a standard
endo R reaction (R). Where indicated the reaction products were also
digested with BstEII or SphI. The products were fractionated in
duplicate on 1.4% agarose gels, transferred to nitrocellulose filters,
and hybridized to either AAV-specific or pBR-specific probe (right
panel). The endo R plus BstEII and endo R plus Shl fragment sizes are
indicated to the left and right of the right panel, respectively. The
left panel illustrates the reaction products stained with ethidium
bromide. Also indicated in the left panel are the sizes and positions
of the marker bands (M) and pSM620 relaxed circular (II) and linear
(III) plasmid DNA. The two unidentified lines to the right of each
panel indicate the position of linear AAV (upper) and linear pBR (lower)

4- -
I=4~ 4
#T t I


M r a: Q

Bst n

9.1 -





m A
co (1)

*l _-.8

... -3.1

(Figure 4-1, left panel, lane R). In addition, approximately 3% of the

products consisted of two fragments, 4.7 and 4.4 kb in length, which

were the sizes expected for excised duplex AAV and linear pBR322 DNA.

These were produced from double-stranded cleavage at both of the AAV/pBR

junctions. The identity of these fragments was confirmed by Southern

hybridization with selective probes. The slower migrating 4.7 kb

fragment hybridized exclusively to AAV-specific probe and the 4.4 kb

linear fragment hybridized only to pBR-specific probe (Figure 4-1, right

panel, R lanes). The fragments generated by further digestion of the

endo R products with either BstEII or SphI (see Figure 3-1) produced a

set of fragments consistent with endo R cleavage at either one or the

other of the AAV/pBR322 junctions. The 3.0 and 1.7 kb fragments

produced from BstEII digestion of endo R products contained only AAV

sequences and, therefore, hybridized only to AAV-specific probe.

Conversely, the digestion of endo R products with SDhI generated 3.1 and

1.3 kb fragments, which consist entirely of pBR322 DNA, and were

detected only with pBR322-specific DNA probe. As expected, the higher

molecular weight fragments contain both AAV and pBR sequences and,

therefore, hybridize to both probes. These results, limited by the

sensitivity of hybridization, placed the position of the cleavage site

to within 25 bp of either AAV/vector junction.

Either AAV/Vector Junction Can Be Cleaved Independently of the Other

One proposed mechanism for gene conversion of AAV terminal mutants

to the wild type form requires the interaction between the two terminal

repeats (Lusby et al., 1980; Samulski et al., 1983). It was therefore

possible that cleavage by endo R required this type of interaction. To

determine if this was the case, subclones of pSM620, pGM620C and pGM620D

were constructed which contained either the left or right terminal

repeat of AAV, respectively (Figure 4-2C). In addition to the 145 bp

terminal sequence, both plasmids contained approximately 350 bp of

flanking AAV DNA. Both pGM620C and pGM620D were efficiently cleaved by

endo R and produced fragments of the expected size and sequence

composition when the endo R products were digested with SphI (Figure 4-

2, A and B). It was, therefore, concluded that either AAV/vector

junction could be cleaved independently of the other and that the

recognition sequence was contained within the terminal 500 bp of AAV.

Figure 4-2 also demonstrates that vector DNA, pBR322, is a poor

substrate for endo R.

Cleavage of AAV Terminal Mutants

A number of mutant AAV clones with deletions in the terminal

sequences were fortuitously produced during the course of constructing a

viable AAV recombinant clone (Samulski et al., 1983). To identify the

endo R recognition signal more precisely, the sequences at the

AAV/vector junctions for several of the mutant AAV plasmids were

determined and the relative frequencies of cleavage were compared

(Figure 4-3 and Table II). The conclusion from these studies was that

there were three sets of sequences near the AAV/vector junctions that

could serve as endo R recognition sites. The first was a stretch of

poly(dG) residues that was present at many of the junctions and was a

result of plasmid construction by GC tailing. A second sequence, which

was called the AAV recognition signal, consists of nucleotides 3 to 15

of the AAV terminal repeat and contains the sequence CCaCTCCCTCTCT.



Junction Frequency


pSM620(L) Vector- (g)19 CCaCTCCCTCTCTgCgCg -AAV 100
pSM620(R) Vector- (g)29 CCaCTCCCTCTCTgCgCg -AAV 100
pSM621(R) Vector- (g)29 CCaCTCCCTCTCTgCgCg -AAV 100
pSM703(R) Vector- (g)38 CCaCTCCCTClaCTaggg -AAV 150
pSM609(R) Vector- (g)10 |CTCCCTCTCTgCgCgCTC -AAV 50


pSM621(L) AAV- CCaCTCCCTCTCTgCIagC -Vector 1
pSM704(L) AAV- CCaCTCCCTCTCTgCIagC -Vector 1
pSM609(L) AAV- CCaCTCCCTCTCTgCIagC -Vector 1


pSM703(L) Vector- (g)36 IagTggCCaaCTCCaTCaC -AAV 150

aThe table lists the sequences at the AAV/vector junctions
and the frequency of cleavage of the AAV clones used to
determine the recognition site for endo R. The first two
bases of the AAV sequence in pSM620 (gg) are counted as
part of the poly(dG):poly(dC) tail. A vertical line (I)
indicates the position of the deletion in a mutant clone.
pSM704(R) and both pAVl plasmid junctions were not sequenced.
(R) and (L) depict the right and left AAV/vector junctions,
respectively. pSM620 and pSM609 had been previously
sequenced (Samulski, et al., 1983) and the sequence of
pSM609(L), which was found to be incorrect, is changed
here. The remaining plasmids were sequenced as part of
this study.

blndicates the approximate yield of the fragment resulting
from cleavage at the indicated site (Figure 4-3), where 100%
cleavage has been defined as the cleavage seen at pSM620(R)
No attempt was made to distinguish between the level of
cleavage observed at the junctions in each category.

Figure 4-2. Cleavage of Substrates That Contain Only One Copy of the
AAV Terminal Repeat.

C) Restriction maps of pGM620C and pGM620D. pGM620C contains the 513 bp
left terminal PstI fragment of the wild type AAV plasmid pSM620,
including the left AAV terminal repeat and the 19 bp poly(dG) tail.
PGM620D includes the right terminal repeat of AAV and the 29 bp GC tail
in a 445 bp PstI fragment subcloned from pSM620. Thick lines represent
AAV DNA, thin lines are pBR DNA, and the solid boxes are the AAV
terminal repeats. Also indicated are the positions of the endo R (R)
and SphI sites. A and B) 0.5 pg of pGM620C, pGM620D or pBR322
supercoiled plasmid DNA were incubated with fraction III enzyme as
described in the legend for Figure 3-3 and digested with ShI. The
reaction products were fractionated in a 1.4% agarose gel, stained with
ethidium bromide (panel A), transferred to a nitrocellulose filter and
hybridized to nick-translated AAV DNA probe (panel B). The position of
the marker bands (M) and the SahI digested endo R product bands are
indicated to the left and right of panel A, respectively. The 620 lanes
contain forms I, II, and III pSM620 plasmid DNA.

C) a
0 0 0 w-





(D CD Wa .



- 1.8

- 1.3


pGM 620 C

1 0

Because the AAV terminal repeat is a palindrome, this sequence occurs

twice within each wild type terminal repeat, at nucleotides 3 to 15 and

109 to 121 (Lusby et al., 1980). The two copies have been designated the

outboard and inboard recognition sequences, respectively. The presence

of the inboard copy of this sequence accounts for the minor cleavage

band derived from the left end of AAV and seen at the 1.6 kb position in

virtually all of the plasmids shown in Figure 4-3. It is essentially

the 1.7 kb Endo R/BstEII fragment minus the first 100 bp of AAV DNA,

which is the distance between the outboard and inboard copies of the AAV

recognition sequence. In 3 junctions (pSM621(L), PSM704(L), and

pSM609(L); Figure 4-3 and Table II), the AAV cleavage site was the only

recognition sequence present because the deletion in these plasmids has

removed the poly(dG) tail and placed the inboard copy of the sequence

adjacent to vector DNA. Accordingly, the 1.6 kb band was the only

fragment seen in these plasmids as the result of cleavage at the left

end. Endo R cleavage of the left junction of the wild type clone,

pSM620(L), produced both the 1.6 and 1.7 kb bands because both the

inboard and outboard copies of the AAV recognition sequence were

present. Presumably, the inboard site at the right end of AAV was also

cleaved, but the separation in this region of the gel was not sufficient

for resolution of the 2.9 kb fragment (Figure 4-3).

The relative intensities of the 1.7 and 1.6 kb band produced from

pSM620(L) cleavage reflect the presence of the poly(dG)lg sequences

adjacent to the outboard site. In general, the homopolymeric poly(dG)

tail sequence was approximately 20 to 100 fold more likely to be cut

than the heteropolymeric AAV recognition sequence (Table II). In

contrast to pSM620, another wild type AAV clone, pAV1, contains no GC

Figure 4-3. Cleavage of AAV Plasmids That Contain Deletions in the
Terminal Repeats.

The various AAV mutant substrates were incubated with fraction III
enzyme and digested with BstEII as described for Figure 3-3. The
action products were transferred to nitrocellulose and probed with
P-labeled AAV DNA. The molecular weights of the endo R/BstEII
products are indicated at the left of the figure. The sequence at each
AAV/pBR322 junction is listed in Table II. The figure was over-exposed
to visualize minor cleavage bands.



Ik n




Cj 0
(D r-



I -I





~ C9

tail and cleavage was directed exclusively by the AAV terminal

recognition site. As a result, cleavage at the inboard and outboard

sites occurs with approximately equal efficiency to produce two junction

bands from each end of equal intensity (Figure 4-3, pAV1; the fragment

sizes are 3.0, 2.9, 1.7 and 1.6 kb). The fact that poly(dG)n alone was

sufficient to act as a recognition sequence was further demonstrated by

cleavage of pSM703(L). This plasmid contains a deletion which has

removed both the inboard and outboard AAV recognition sites, but retains

a poly(dG) stretch that is 36 bp long. The poly(dG) stretch was

sufficient to promote cleavage and produced a band of approximately 1.6

kb. The slightly slower mobility of the 1.6 kb fragment generated from

pSM703(L) cleavage is attributed to the addition of the sizable poly(dG)


Minor Endo R Cleavage Sites

Cleavage by endo R appeared to require a sequence that was at

least 10 bp long and consisted of a polypurine-polypyrimidine tract that

was relatively rich in GC base pairs. These requirements were confirmed

by the identification of some of the minor cleavage fragments observed

in Figure 4-3, which had been predicted on the basis of this sequence.

It was noticed that a number of minor bands appeared consistently in

digests of all of the AAV variant plasmids. Because the size of the

minor fragments did not vary with the size of the terminal deletion,

these fragments must have been generated from cleavage within AAV

sequences (If the fragments had been generated from cleavage within

pBR322, they would have spanned the pBR/AAV junction and their size

would have varied depending on the size of the terminal deletion).



Computer-Predicted Fragments Observed

Sequence Nucleotide kb %C kb

CCCCTCTCCCCTC 4089 2.39 10/13 2.4

CCaCTCCCTCTCT 4678 3.00 8/13 3.0
CCaCTCCCTCTCT 4562 2.86 8/13 NC
CCaCTCCCTCTCT 3 1.70 8/13 1.7
CCaCTCCCTCTCT 121 1.59 8/13 1.6

CTCCaCCCCTCC 157 1.54 9/12 NC
CTaCagCaCCCCTT 3026 1.33 7/14 1.3
CCCTgCCCaCCT 2947 1.25 8/12 1.2
CagCagCCCCCTCT 2775 1.08 8/14 1.1
CCagaCTCCTCCTC 2661 0.96 8/14 ND
CCCgCCTCCggCgCC 753 0.95 10/15 ND
CCCCTCCTCCCaCC 1461 0.24 11/14 TS

aThe table lists the cleavage fragments predicted by a computer
search of the plasmid pSM620 using poly(dC)13 as the search
sequence and allowing for 60% homology between compared sequences.
The sequence at which cleavage would presumably occur and the number
of C residues in each sequence (%C) are tabulated with the size of
the predicted fragment, which was derived by assuming that endo R
cleavage occurred 3' to the poly(dG) sequence. The predicted
fragment sizes are compared with the sizes of the fragments observed
in the agarose gel assay. All fragment sizes are the result of BstEII
digestion of endo R products. NC Not clear; ND Not detected;
TS Too small to be resolved on a 1.4% agarose gel.

Using a computer and allowing for substitutions within the predicted

consensus sequence, poly(dG)19, or the complementary sequence

poly(dC)19, internal AAV sequences were searched for potential endo R

recognition sites. Table III compares the computer predicted cleavage

sites and fragments with the observed fragments. The results show a

good correspondence between the computer-predicted cleavage pattern and

the fragments seen in endo R digests. In fact, all of the observed

fragments that were believed to be the result of cleavage in AAV DNA

were predicted by the computer search. In particular, the best matches

found were with a sequence at nucleotide 4089, which produced the 2.4 kb

minor cleavage fragment, and with both the inboard and outboard AAV

terminal recognition sequences.

AAV Excision In Vivo

A demonstration that endo R was capable of excising AAV sequences

in vivo in a manner similar to that observed in vitro would add

credibility to the hypothesis that this was the enzyme involved in AAV

DNA replication. When AAV plasmids were transfected into tissue culture

cells, the bulk of the input DNA was found in the form of nicked

circular (form II) or linear (form III) plasmid species (Samulski et

al., 1982, 1983; Hermonat and Muzyczka, 1984). This was true regardless

of whether the input AAV DNA was capable of DNA replication. To

determine if the linear plasmid DNA (form III) produced in vivo was the

result of cleavage at a specific sequence, transfected DNA was isolated

after 21 hours, digested with BstEII and hybridized with AAV probe

(Figure 4-4). The 1.7 and 3.0 kb bands were readily seen in cells

infected with adenovirus and transfected with the wild type plasmid,

Figure 4-4. Cleavage of AAV Plasmids In Vivo.

Five micrograms of PSM620 and pSM703 DNA were transfected into Ad-
infected HeLa cells (Chapter II: Material and Methods). At 21 hours
after transfection, low molecular weight DNA was isolated by the method
of Hirt (1967), digested with BstEII, fractionated by agarose gel
electrophoresis and hybridized to AAV-specific probe as described for
Figure 3-3.

.o "5

O r-

3.0 2.9

S 2.4

1.7 1.6

pSM620. In this case the fragments were presumably derived from

replicating AAV DNA. However, bands of the same size were seen at

reduced levels in cells transfected with pSM620 in the absence of

adenovirus or in the presence of adenovirus and hydroxyurea (data not

shown). In addition, two minor bands were seen in cells transfected

with pSM620 which apparently were not amplified by DNA replication.

These were identical to the minor 2.4 kb and 1.6 kb bands observed in


Cleavage fragments generated from pSM703 transfection were

expected to be approximately 100 bp shorter than those seen with the

wild type plasmid due to the deletions in both termini (compare with

Figure 4-3). These bands were readily seen in cells transfected with

pSM703 DNA (Figure 4-4) and because pSM703 is an ori" mutant (Samulski

et al., 1983), the 1.6 and 2.9 kb fragments must have been produced from

the input plasmid sequences. As expected, pSM703 also generated a minor

2.4 kb band. It was concluded, therefore, that a substantial amount of

input plasmid DNA was cleaved at the AAV/vector junction in vivo and

that AAV DNA replication is not necessary for cleavage to occur.

Moreover, the pattern of cleavage was essentially the same in vivo and

in vitro. The remaining input DNA was apparently cleaved randomly as

judged by the background radioactivity.

Replication of Endo R Products

When the products of an in vitro endo R digestion of pSM620 were

transfected into Ad-infected HeLa cells, DNA replication of the AAV

sequences occurred at levels that were indistinguishable from the

transfection of the form I plasmid (not shown). This was true even at

times early after transfection. Differences in the production of

replicated AAV DNA would presumably occur early in the replication

cycle, but only if the excision of AAV sequences was a rate limiting

step in DNA replication. Therefore, endo R products were substrates for

DNA replication in vivo, but cleavage of AAV plasmids in vitro prior to

transfection did not provide a selective advantage for DNA replication.

Endo R products of pSM620 cleavage were also substrates for DNA

Polymerase I in vitro. The products of endo R digestion were heat

denatured and cooled rapidly to favor the formation of the hairpin

primer over renaturation of duplex molecules. In these experiments, a

significant amount of radioactivity was incorporated when these

substrates were incubated with DNA Pol I (not shown). Thus, the

products of endo R cleavage will support DNA replication both in vitro

and in vivo.


Isolation of the Cleavage Site

In previous experiments, a consensus cleavage site was defined on

the basis of commonly occurring sequences found in AAV recombinant

plasmids. These sequences contained homopurine-homopyrimidine stretches

of at least 10 nucleotides which were rich in G:C base pairs. However,

it remained possible that the endo R recognition signal was at a

separate location from the cleavage site, possibly at an internal AAV

sequence. In this case, enzyme recognition would occur at a distant

site and result in cleavage at the nearest available stretch of

homopurine-homopyrimidine sequences. To rule out this possibility and

to further characterize the recognition sequence, several clones

containing subsets of the sequences present at the AAV/vector junctions

were constructed and their ability to act as substrates for endo R

specific cleavage was evaluated. The level of cleavage in several of

these clones is illustrated in Figures 5-1 and 5-2 and the insert

sequence and frequency of cleavage are listed in Table IV. The first

plasmid, pGM1008, is a subclone of pGM620D and contains sequences from

the right AAV/pBR322 junction, including the first 21 bp of AAV and the

31 bp G:C tail. This plasmid was cleaved with approximately the same

frequency as the parental junction in pSM620D (Figure 4-4). Two other

plasmids, pGM913 and pGM1116, contained poly(dG)-poly(dC) stretches of 9

and 13 bp, respectively, inserted at the PstI site of pBR322.

Figure 5-1. Oligomer Clone Excision

A) Standard reaction mixtures of 25 pl, containing 0.4 pmol of the pGM
clones indicated, were digested with endo R at 37 C for 1 hour. The
reactions were terminated with the addition of proteinase K stop
solution (see Chapter II: Materials and Methods), phenol extracted,
ethanol precipitated and digested with EcoRI. B) EcoRI digestion of
endo R cleaved substrate produces 0.8 and 3.6 kb bands from pGM clones
containing oligomer inserts in the PstI site of pBR322, while 1.0 and
3.4 kb fragments are produced from clones which contain inserts in the
EagI site (see Figure 5-6 for the sequence of the inserts). E EcoRI
site, R = endo R site, filled square oligonucleotide insert (See
Chapter II for a description of the clones).

4.4 -




0.8 E 3.6
-- I


3.4 E 1.0

Figure 5-2. Oligomer Clone Gel Assay.

Standard reactions of 25 pl, containing 0.4 pmol of the substrates
indicated, were digested with endo R at 37C for 1 hour. Reactions were
terminated with proteinase K stop solution (Chapter II), phenol
extracted, ethanol precipitated and digested with EcoRI. The assays were
fractionated on a 1.7% agarose gel, transferred to nitrocellulose
(Southern, 1975) and probed with nick translated pBR322 DNA (>1 x 108
cpm/pg). Light and dark exposures (top and bottom, respectively) of the
0.8 kb fragments produced from endo R and EcoRI digestion are shown to
illustrate the relative frequency of cutting in the different


, 21-* o
a re) s
Cn W

'- CN



B 0- .8 -



Inserted Sequenceb

















aThe construction of the clones is described in detail
in Chapter II: Materials and Methods. The sequences are
listed as they appear in the clone on the clockwise
strand in a 5' to 3' direction. The frequency of cleavage
is defined as the percent of starting substrate cleaved in
a standard assay with 1 unit of endo R. The amount of
cleavage for pGM1008 was arbitrarily set to 100%.

bThe complete sequence of the inserts are listed in Figure 5-6.

cNatural poly(dG) stretches of 6 bp occur in pBR322 at
nucleotides 2550 and 2797.

dThe sequences for pGM1344 are repeated in an inverted













No AAV sequences were present in either plasmid. Specific cleavage was

observed with these substrates as well. However, the level of cleavage

was significantly reduced with both substrates. PGM913 (13 bp of G:C)

was cleaved at 30% and pGM1116 (9 bp of G:C) was cleaved at 20% of the

level observed with clones that contained all of the parental AAV/vector

junction (pGM620D, pGM1008, Figures 5-1 and 5-2, Table IV). These

clones approach the limit of enzyme recognition, since the naturally

occurring stretches of poly(dG)6 in pBR322 (Figures 5-1, 5-2 and Table

IV) and poly(dG)7 in lambda phage DNA (not shown) were not detectably

cleaved. These results, in conjunction with those obtained with the

pSM609(R) junction (Table II), suggested that a minimum of 9 G residues

are sufficient for cleavage, while the frequency of cleavage increases

in direct proportion to the length of the homopolymer chain. Thus, AAV

plasmids that contain GC tails at the AAV/vector junction are ideal

substrates for endo R.

The apparent strong affinity of endo R for long stretches of

poly(dG) and the observed cleavage of G-rich sites that did not contain

homopolymers of G:C (Figure 4-3, pAV1, pSM609(L), pSM621(L) and

pSM704(L)) raised some important questions concerning alternative enzyme

recognition sites and, particularly, what effect substitutions in the

poly(dG) chain have on cleavage activity. To determine what type of

substitutions were tolerated by endo R, a number of plasmid substrates

containing a variety of repeating polymeric sequences were tested. The

first of these was the plasmid pEV136, an infectious polio clone which

contains 18 bp of poly(dG) at the 5' end and an 84 bp stretch of

poly(dA) at the distal end. When incubated with endo R, pEV136 produced

fragments which were the result of cleavage exclusively at the poly(dG)

site. No detectable cleavage was observed at the poly(dA) site (Figure

5-3). However, clones that contained an insert of the alternating co-

polymer (GA)38 and (GA)11 (pGA38 and pGAll, respectively) were cleaved

by endo R at a reduced frequency (Table IV and Figure 5-3). Thus,

homopolymers of (dA) were not recognized as substrates for cleavage, but

substitutions of (dA) in long stretches of poly(dG) were tolerated and

resulted in a decrease in activity. This was not the case with

alternating polymers of GC that can potentially form altered secondary

structures of left-handed Z-DNA (Peck and Wang, 1983; Singleton et al.,

1983). Clones containing inserts of (GC)10 and (GC)20 were not

detectably digested by endo R (pGM1635, Table IV).

More complex sequences with less symmetrical substitutions of the

poly(dG) sequence are also substrates for endo R. This is exemplified

by two clones, pGM1483 and pGM1505, that contain generic forms of the

Tetrahymena (C4A2) and Dictyostelium (C2-6T) telomeric sequences,

respectively (Blackburn and Szostak, 1984; Shampay et al., 1984). These

substrates were cleaved at a level approaching the frequency observed

with the homopolymeric clone pGM913 (Figure 5-1 and Table IV).

Integrated AAV provirus exists in two basic forms in chromatin, as

well as in plasmids. The first of these consists of at least two tandem

head to tail copies of AAV, each separated by two copies of terminal

sequences in an inverted orientation (McLaughlin et al., 1988).

Presumably, this would place two copies of the endo R recognition site

in an inverted orientation, adjacent to each other. The plasmid

equivalent of this contained one copy of the AAV terminal repeat

adjacent to a GC tail and was cleaved with high frequency (e.g. pGM1008,

Figure 5-1). The second type of integrated form occurs less frequently

Figure 5-3. Endo R Cleavage of pGA38 and pEV136.

A) One microgram of either pGA38 or pEV136 was incubated with 1 unit of
fraction V endo R under standard reaction conditions in a volume of 25
Al. After phenol extraction and ethanol precipitation, one half of each
reaction was further digested with either Scal (pGA38) or B&g2 (pEV136)
and both the endo R treated (R) and the restriction enzyme digested-endo
R treated (RS) samples were fractionated on 1.4% agarose gels. B)
Restriction maps of pGA38 and pEV136. Filled boxes represent the
poly(GA) insert for pGA38 (GA) and the poly(dG) and poly(dA) inserts for
pEV136 (G and A, respectively). The molecular weights (kb) indicate the
size of the products that would be expected from a restriction enzyme
digest of endo R products cleaved at GA in pGA38, and at G plus A in


Fm2- .
Fm1 k ,.

- 4.4
- 2.7

- 1.8

- .94


S|- Fm3
4.4 e 5.7

2.4 --

1.5- .


1.9 1



in chromatin and contains single or multiple copies of AAV separated by

a single copy of the AAV terminal repeat (McLaughlin et al., 1988).

Plasmids that contain this type of arrangement were cleaved at a low

frequency in vitro (e.g. pAV1, Figure 4-3).

To determine if the terminal 23 nucleotides of AAV contained a

recognition signal for endo R and what effect the number of copies and

orientation of these sequences had on activity, two clones containing

either a single copy or two inverted copies of the AAV terminal

recognition sequence (ggCCaCTCCCTCTCTgCgCgC) were constructed. The

plasmid pGM1228 contains a single copy of this sequence and pGM1344

contains two inverted copies of the same sequence, both inserted into

the PstI site of pBR322. Both of these constructs were cleaved by endo

R (Figure 5-2). However, the level of cleavage observed with pGM1228

was approximately 1% of that observed with pGM1008 and could only be

visualized by Southern hybridization. In contrast, the plasmid that

contained two inverted copies of the recognition sequence was cut at a

level at least 5 times greater than that of the single copy plasmid,

while still only about 5% of that seen with the parental junction

plasmids. Thus, the presence of tandem inverted copies of the AAV

recognition signal improves the yield of specifically cleaved product in

vitro. This arrangement may be representative of the preferred

situation in chromatin.

Sequence at the Site of Cleavage

Previous experiments have already shown that endo R makes double-

stranded specific cuts and that cleavage of the substrate occurs at the

recognition signal. In addition, it was possible that cleavage at the

insert/vector junction would produce long single-stranded overhangs at

the ends of the fragments. Conceivably, the ends of these fragments

could anneal and effectively reduce the observed amount of specifically

cleaved product. However, no increase in the amount of specifically

cleaved product was observed after the reaction products were phenol

extracted and heated, indicating that the cleaved fragments did not have

extensive 5' or 3' protruding ends capable of hybridization (not shown).

Several questions remained, however, concerning the mechanism and

pattern of endo R cleavage. For example, it was not known whether endo

R cleavage occurred at one or several sites within the recognition

sequence or whether there was a difference in the cleavage pattern

between substrates that contained different types of recognition

signals. Additionally, it was possible that the separate strands of the

recognition sequence were cleaved in a dissimilar manner. To address

these questions, the sequence at the site of cleavage was determined for

several substrates containing various lengths of poly(dG):poly(dC),

alternating copolymers of GA, telomeric sequences and the AAV

recognition sequences (Figure 5-6). In these studies, the mobility of

the primer extension products of endo R fragments were compared with the

sequence of the insert and flanking regions of the substrate clone on

denaturing acrylamide gels. The sequence of the substrate was

determined by the dideoxy sequencing method (Sanger et al., 1977) using

the same single-stranded primer used for primer extension (Figures 5-4

A-D, see Chapter II: Materials and Methods). These results were

confirmed independently by comparison of the mobility of 5' and 3'

labeled endo R products with the chemically sequenced plasmid insert

(Figure 5-5).

The conclusion from the analysis of the cleavage sites of clones

containing homopolymers of poly(dG) was that cleavage could occur

throughout the G:C insert (Figure 5-4A, Figure 5-6). A slight

preference for cleavage at the 3' end of both strands was observed in

all the homopolymeric clones. This effect was especially pronounced at

the 3' end of the poly(dG) strand in the plasmid pGM1008, which contains

the AAV recognition signal 5' to the G:C insert (Figure 5-6). Cleavage

of the G-strand was particularly strong at the G:C/AAV junction and

continued into the AAV sequences, while cutting in the C-strand in this

clone appears to be more uniform, with only a slight preference for the

3' end. The site of cleavage in clones that contained only

poly(dG):poly(dC) inserts and no AAV sequences (pGM913, pGM1116, Figure

5-6) was also slightly skewed toward the 3' end of both strands, while

cleavage was confined almost entirely to the G:C insert. In substrates

that contained the alternating copolymer (GA)11 and (GA)38 (Figure 5-

4B), cutting occurred throughout the polymer insert between the G and A

residues on the lower strand and between the C and T residues on the

upper strand. In these clones, cleavage was evenly distributed within

the CT strand, but more pronounced at the 5' and 3' insert/vector

junctions in the GA strand.

The cleavage pattern observed with plasmids containing telomeric

sequences exhibited more periodicity. In the plasmid pGM1505 (Figure 5-

4C), which contains the sequence (C2_6T)5, cleavage of the CT-strand

occurred throughout the insert with a particularly strong cut occurring

at regular intervals immediately after the first C residue of every

repeat. A preference for cleavage at the 3' end of the CT-strand was

observed, corresponding to the region containing 6 continuous G:C base

Figure 5-4. Sequence at the Site of Endo R Cleavage Dideoxy Method.

A) Nucleotide sequence determined by the dideoxy method (Sanger, 1977)
and primer extension products of endo R digestions for substrates
containing: A) poly(dG) inserts; pGM1008 (left) and pGM913 (right), B)
alternating copolymers of GA; pGAll (left) and pGA38 (right), C)
telomeric sequences; pGM1505 (left) and pGM1483 (right), and D) a single
copy (pGM1228, left) or two tandem inverted copies (pGM1344, right) of
the AAV recognition site. The plasmid sequence (GCAT) was determined
for each strand of the substrate using either a clockwise (cw) or
counterclockwise (ccw) oligonucleotide primer and 1 pg of plasmid. The
endo R primer extension products (R) were synthesized from 1 pg of endo
R digested substrate with the primer used to generate the corresponding
sequence markers (see Chapter II).

cw ccw









pGA1 1




cw ccw








dw ^ B *'




Figure 5-4 (continued), part B




em- -

am a-"
go w
ma .-

Figure 5-4 (continued), part C





Giai1111 C A T R^


*4 4


* i
t %,

Figure 5-4 (continued), part D



Figure 5-5. Sequence at the Site of Endo R Cleavage Maxam and Gilbert

Fine mapping of the endo R cleavage products of pGM620D (PstD). The G+A
and C+T sequencing reactions were as described (Maxam and Gilbert,
1977). 1 pg of pGM620D labeled at the Scal site at either the 5' or 3'
end was digested with endo R (R) and fractionated on a 8% acrylamide-
urea gel in parallel with sequence markers labeled at the same site.
The poly(dG)28:poly(dC)28 tail and the direction of the AAV and pBR322
sequences are indicated in the figure. A detailed description of the
sequencing protocol and of PGM620D is included in Chapter II.









U 0 -4
S4-1 0
M 3 .C *4 0
0 a F4cLFa a
4 QJ 04J 4-4
"I4 tB 4.J 0

) ,C 3 3M > n u
30 4-1 0 a *- 04 c
o a C O !-4 4J E Q)

> >-o C Co a
a) Z a)*uU
0) 0 0 1 W (

S* M L o
o 040 0 X
( -I 0 O
41 0 41O w >o-

O *4O 0 t~O
44-4 W 0 W 0l 0) (4-4

0 1 0 .0 cE
*Oa 0 ) E C 4- CO J
,C 4 u = m.
SC U 4B "0 e

cg *M C M ,C a
bo 0

S 4- C bO04l 03 D
i- 1 0 0 0)

0 3 g -* >
> C I 10C 0

cc o 4 4Cu1 r.C9
O) 44 r. tkO 44 UC C
' t00 4 0
C Q) 4J O-0
u .14-4 u w Q]'4
Sa *r 0 0 c e r

,0 tB 0 ,0
4J 01 4 O4C 0-Cu4-l

0 1 4 0 S Ma
.,j 4) Q) 0 C
w L 44 cd to 44
4 C a0 0 4= to 4
41 C a) 0 4-u c)o
S 001 c 4a 4 0
4J W Lt)i 4 C
4- bO Cl I Cu u w
Ba -*r 0 W V) 41
44 4J CBO C o U)
a) ) o0 w 1 r-. Q1 z
r to Q) w :3 0 -H w
3 CO -r a 3 >.o ,0

* a 0- c 'o -
Ur 41 H OB c M-( mC0
Sa W0 r. CL to a)

W c *Q) 1a *0
E w 0 04 c 4

W 0 w a *n C 0
V0 3 0 U 0 3 41 4J)

*0 4 0 0 0 ml
1o w 4- 41 0c
F4 0-i O M 10 0 14 u


< <

< I-
0 0

C 3 oa
00 00

00 00 o
00 00
C.5 0 <00
< <30 00

S0" a, O
do o I- <
00 00

0 i 00 00 1--00
0.00 00 0 0

00 00 0 0
C-0 0

00 00 < 0-
< 0 0 0 0 I- <0 0
0 0. C, 00 -00 0 00
00 ~ ..Q00. 00 00

oo 00 00 00 00 00 0
0 0g0 000 00
003 <. P-00- 0 00 0 00 00 00 0

0 00 oo oo q0'7. 00
b00 C- 00 0 .00 < F

C.0 0- 00 oo 00 000 ) r) 00
0 0 0 0 Q( ) C. a 0 D o o 0 0
L 0 1--- < <0 < o F- <0 o- <
00) 0 00 H< 000 -0 00 <00

0 00 o00 -O O 0 C 0 o 0
00 0.000(30Q
< 000.0 <0. L) 00

00 00 00 00 ...< S :oo* ^ -~ 00 00


00 CL og oo 00 C
00 0o 0o 00 00 00 o

o^ Qj

pairs (Figure 5-6). Curiously, a strong preference for cutting was seen

at the 3' end of the GA strand, even though the longest stretches of G:C

base pairs are at the 5' end of this strand. In contrast, the pattern

of cutting observed with the plasmid pGM1483, which contains the

sequence (C4A2)3, exhibited little preference for either end of the

insert (Figure 5-4C). The cleavage pattern for this clone was

repetitive and symmetrical throughout the oligonucleotide insert (Figure

5-6). Cleavage appeared to occur in directly opposite positions on each

strand immediately before the A residues on the CA strand.

Cleavage of the plasmids pGM1228 and pGM1344, which contain single

and inverted double copies of the AAV terminal sequences, respectively,

was contained within the C-rich region of the AAV recognition sequence

and showed a preference for the 3' end of each strand (Figure 5-4D,

Figure 5-6). This is especially noticeable in pGM1344, where the cut

sites are localized exclusively within the C-rich region of each strand

and would appear to produce sizable 3' overhangs.

The results from the sequence studies indicated that cleavage

occurred throughout the recognition sequence and frequently extended

into the flanking vector sequences. A preference for cleavage at the 3'

end of both strands was observed in most substrates, while cutting in

the C-strand had a tendency to be more uniform. In addition, the

nucleotide order had no affect on the site of cleavage, where cutting

occurred between every base combination. These observations, in

conjunction with the observed frequency of cutting with different

substrates, suggest that cleavage at the nucleotide level is structural

rather than sequence dependent.

Is Endo R a Site Specific Single-Stranded Nuclease

A number of single-stranded endonucleases will cleave polypurine-

polypyrimidine stretches that are present in duplex DNA. The activity

has been demonstrated for Sl, mung bean, and Neurospora endonucleases as

well as for snake venom phosphodiesterase (Evans and Efstradiatis, 1986;

Cantor and Efstradiatis, 1984; Schon et al., 1983; Pulleyblank et al.,

1985). One important difference between these activities and endo R is

that the single-stranded nucleases require supercoiling in substrates

that contain short polypurine stretches. In contrast, endo R readily

cleaves linear substrates which contain 13 bp of poly(dG) (data not

shown). However, there is enough similarity in sequence recognition to

suspect that endo R may have properties common to single-stranded


To determine if Endo R was a single-stranded DNA endonuclease, the

products of Sl and endo R digestions of pSM620 were compared. Sl

produced a cleavage pattern similar to that seen with endo R. However,

unlike endo R, Sl did not seem to discriminate between the inboard and

outboard ends of the AAV terminal palindrome. It apparently cut equally

well at both ends of the terminal repeat, producing a full length pBR322

linear DNA molecule (4.4 kb) and the shorter 1.6 and 2.9 kb AAV

fragments (Figure 5-7). The 100 bp deletion resulting from the loss of

the AAV terminal sequences is especially noticeable when comparing the

1.7 kb endo R products with the 1.6 kb fragments produced by Sl. It is

not known whether the loss of the AAV terminal sequences was due to

selective degradation of the palindrome or to specific cleavage at both

the inboard and outboard recognition sites. Similar experiments with

mung bean endonuclease did not produce discrete bands (not shown).

Figure 5-7. Sl and Endo R Nuclease Digestions of pSM620.

One microgram (0.2 pmol) of pSM620 DNA was titrated with Sl nuclease or
fraction IV endo R in a 50 pl reaction volume under standard conditions
for each enzyme (see Chapter II: Materials and Methods). Numbers at the
top indicate the number of units of each enzyme used.

SI EndoR

.10 .25 .50 M 40 1.0 2.0





Sl nuclease typically releases acid soluble oligo- and mono-

nucleotides as products of the reaction (Vogt, 1973). After normalizing

endo R and Sl activities on duplex substrates, the activities on single-

stranded 3H-labeled E.coli DNA were compared by measuring the release of

acid soluble radioactivity (Figure 5-8). When equivalent amounts of Sl

and endo R (as measured on duplex DNA) were incubated with 3H-labeled

single-stranded E.Coli DNA, Sl converted 70% of the DNA to acid soluble

product within 30 minutes, while endo R produced no acid soluble

products after 2 hours. The results were similar regardless of whether

the reactions were carried out at pH 6.0 for both enzymes, or whether

each enzyme was assayed at its own pH optimum (pH 4.5 for Sl and pH 7.5

for endo R). On the basis of this assay, it was estimated that endo R

was approximately 50-100 fold less active on single-stranded DNA than


The ability of endo R and Sl to degrade single-stranded circular

DNA was also compared. Under standard reaction conditions, single-

stranded circular OX and M13 DNA are degraded to oligonucleotides of

200-400 base pairs in length (Figure 5-9). When assayed by gel

mobility, endo R cut single-stranded circular DNA at a rate 20 fold

lower than Sl (not shown). This was also true in reactions which

contained standard amounts of plasmid substrate in addition to OX DNA

(not shown). In these experiments, 50% of the plasmid substrate was

cleaved specifically, while activity on single-stranded circles was

similar to that observed without plasmid substrate. It was concluded

from these experiments that endo R was clearly not a typical single-

stranded nuclease. However, it was not clear whether endo R activity on


single-stranded circles was an intrinsic property of endo R or the

result of a residual contaminating endonuclease.

Figure 5-8. Sl and Endo R Activity on Single-Stranded DNA.

Five micrograms of single-stranded, uniformly 3H-labeled E. coli DNA was
incubated with Sl nuclease (1.25 units) or endo R (5.0 units) in a 250
pl reaction at 370C under standard reaction conditions for each enzyme
(see Materials and Methods). Equivalent amounts of double-stranded
cleavage activity for both enzymes were used for each reaction (See
Figure 5-7). At the indicated times, 25 pl portions of each reaction
mixture were acid precipitated and counted. The percent (%) acid-
precipitable counts remaining at each time point is plotted. Closed
circles, endo R; open circles, Sl.







Figure 5-9. Endo R Activity on Single-Stranded Circular DNA.

A standard reaction mixture of 125 ip contained 5 pg of OX single-
stranded circular DNA and 2.5 units of endo R. At the indicated times,
25 1l portions were removed, the reaction was terminated with the
addition of 0.1% SDS and 20 mM EDTA, and the reaction products were
fractionated on a 1.4% agarose gel.

0 10 25 40 60

Circles -
Linears -

200-400 [


Reaction Conditions for Double-Stranded Cleavage

Purified endo R requires Mg2+ cations at 5-10 mM for optimal

cleavage activity (Table V). Calcium ions (Ca2+) at 0.1-1.0 mM can

substitute for Mg2+, but only about half of the activity is achieved

under these conditions. No activity is observed with Zn2+ alone at any

concentration, and both Ca2+ and Zn2+ are inhibitory at higher

concentrations in the presence of Mg2. When Mn2 is substituted for

Mg2+, a different specificity of cleavage is observed at concentrations

below 0.1 mM, while greater concentrations result in non-specific

degradation of the substrate.

Endo R is extremely salt sensitive (Figure 6-1). A 50% inhibition

was observed with NaCl or KC1 at concentrations of 10 mM. Because of

this sensitivity, fractions emerging from columns eluted with salt

gradients were dialyzed before being assayed.

Endo R is active over a wide pH range, from pH 6 to pH 9, in Tris-

HC1, sodium acetate or sodium phosphate buffers (not shown). The pi of

endo R was estimated to be between pH 6.5 and 7.0. This was determined

by observing the binding characteristics of the enzyme to DEAE cellulose

at different pHs (see Chapter II: Materials and Methods). No increase

in the rate of the reaction was observed with temperatures above 370C,

while temperatures below 370C resulted in a decreased rate of cleavage.









% Double








aBstEII linearized pSM620 (0.2 pmol) was incubated with 1 unit
of fraction V endo R under standard reaction conditions with
the amount of divalent metal cation indicated in the table.
The reaction products were fractionated on 1.4% agarose gels,
stained with ethidium bromide and photographed. The amount of
double-stranded cleavage was determined from densitometer
tracings of the 3.0 and 1.7 kbp endo R/BstEII products as
described in the legend to figure 6-1.

Concentrations of MnC12 >0.1 mM result in nonspecific degradation
of substrate DNA, while concentrations between 0.01 and 0.10 mM
produce a different cleavage specificity (see Figure 6-3).

Figure 6-1. Sodium and Potassium Inhibition of Endo R.

Standard reactions containing 0.2 pmol of BstEII linearized PSM620 were
incubated with 1 unit of fraction V endo R with increasing
concentrations of NaCI or KC1 for 1 hour at 370C. The products of the
reaction were fractionated on a 1.4% agarose gel, stained with ethidium
bromide and photographed with Polaroid Type 55 film, which produces a
negative and a positive print. The intensity of the 3.0 kbp product was
measured from the negative using the LKB ultrascan XL densitometer and
the LKB gel scan program, version 1.0. The percent activity is plotted
against the concentration of Na and K chloride in the reaction. The
amount of cleavage at 1 mM NaC1 (the salt contribution from the
extracts) is assigned 100% activity.

- 50
< 40

0 10 20 30 40 50
Na, K (mM)