Triplex formation inhibits HER 2/neu transcription in vitro

Full text

(1)

Triplex formation inhibits HER-2/neu

transcription in vitro.

S W Ebbinghaus, … , G Sanders, D M Miller

J Clin Invest.

1993;

92(5)

:2433-2439.

https://doi.org/10.1172/JCI116850

.

Triplex-forming oligonucleotides (TFOs) have been shown to bind to target DNA sequences

in several human gene promoters such as the c-myc oncogene, the epidermal growth factor

receptor, and the dihydrofolate reductase genes. TFOs have been shown to inhibit

transcription in vitro and gene expression in cell culture of the c-myc and other genes. The

HER-2/neu oncogene, which is overexpressed in breast cancer and other human

malignancies, contains a purine-rich sequence in its promoter, which is favorable for

purine:purine:pyrimidine (R:R:Y) triplex formation. Although its function in the HER-2/neu

promoter is unknown, this purine-rich site is homologous to a protein-binding sequence in

the promoter of the epidermal growth factor receptor that is necessary for efficient

transcription of this gene. We have shown that this sequence is a site for nuclear protein

binding by incubation with a crude nuclear extract. We describe the formation of an

interstrand triplex using a purine-rich oligonucleotide antiparallel to this purine-rich target

sequence of the HER-2/neu promoter. Triplex formation by the oligonucleotide prevents

protein binding to the target site in the HER-2/neu promoter in vitro. We have shown that this

oligonucleotide is a potent and specific inhibitor of HER-2/neu transcription in an in vitro

assay. The triplex target site contains a single pyrimidine base that does not conform to the

R:R:Y triplex motif. In an attempt to abrogate […]

Research Article

Find the latest version:

(2)

Triplex Formation Inhibits HER-2/neu Transcription

In

Vitro

Scot W. Ebbinghaus,* JayE. Gee,* Brad

Rodu,I

CharlesA. Mayfield,*George Sanders,* and Donald M.Miller**il

Departments of *Medicine, tBiochemistry, and Oral Pathology, UniversityofAlabamaatBirmingham, andI"Birmingham Veterans

Affairs MedicalCenter, Birmingham, Alabama 35294

Abstract

Triplex-forming oligonucleotides (TFOs) have been shownto

bindtotargetDNAsequencesinseveral humangenepromoters

suchasthec-myconcogene,theepidermal growth factor

recep-tor, and the dihydrofolate reductase genes. TFOs have been showntoinhibittranscription in vitro andgeneexpression in cell culture of the c-myc and other genes. The HER-2

/

neu oncogene, which isoverexpressed in breast cancerand other

human malignancies, containsapurine-richsequenceinits

pro-moter,which is favorable for purine:purine:pyrimidine (R:R:Y) triplex formation. Although its function in the HER-2/neu

promoteris unknown, this purine-rich site is homologoustoa

protein-binding sequence in the promoter of the epidermal growth factorreceptorthat isnecessaryfor efficient

transcrip-tion of thisgene.We have shown that thissequenceisasite for

nuclear protein binding by incubation withacrude nuclear

ex-tract.We describe the formation ofaninterstrandtriplexusing a purine-rich oligonucleotide antiparallel to this purine-rich

target sequence of the HER-2/neu promoter. Triplex forma-tionby the oligonucleotidepreventsprotein bindingtothe tar-getsiteintheHER-2/neu promoter invitro. We have shown that thisoligonucleotide is a potent and specific inhibitor of HER-2/neu transcription inaninvitroassay.Thetriplex tar-getsite containsasingle pyrimidine base that doesnotconform

tothe R:R:Ytriplex motif. Inanattempt toabrogate the

poten-tially destabilizing effects of this pyrimidine base on triplex

formation,wehave substitutedanabasic linker for the

pyrimi-dine residueinthetriplex forming oligonucleotide. Triplex for-mation with the modified oligonucleotideappearstooccurwith

approximately equivalent binding affinity. Triplex formation in theHER-2

/neu

oncogene promoter prevents transcription in vitro andmayrepresentafuture modality for specific inhibition

ofthisgeneinvivo.(J. Clin. Invest. 1993.92:2433-2439.) Key

words: footprint * gel shift*sequenceDNA*oncogene*gene

expression

Introduction

Theneu oncogenewasoriginally described inNIH3T3

trans-fectantsfrom ethylnitrosourea-inducedtumorsinrats( 1, 2).

The human counterpart, c-erbB-2 orHER-2/neu, encodesa

Address correspondencetoDonald M. Miller, M.D., Ph.D., 1824 6th AvenueSouth, Wallace Tumor Institute, Room 520, Birmingham, AL 35294. Thecurrentaddress for Jay E. Gee is Center for Biotechnology, Baylor College of Medicine, The Woodlands, TX 77380.

Receivedfor publication 15 June 1992 and in revisedform11May 1993.

185-kDprotein that is homologous to the EGF receptor

(EGF-R)'

(3, 4). A point mutation changing a single amino acid in the membrane spanning domain leads to deregulation of

tyro-sine kinase activity and transformation of cells

bearing

this

mutantprotein (5-7). A stimulatoryligand for the HER-2/ neu gene product hasrecently been described (8). HER-2/neu isamplified and/or overexpressed in - 25% of human breast

cancerand correlates with advanceddisease and poor outcome

(9, 10); however, the point mutation described for in vitro

activation of theHER-2/neuoncogene is absent in the ampli-fied HER-2/ neu oncogene inhumantumorspecimens( 11 ).

The HER-2/neu promoter contains typical CCAAT and TATAelements thatareseparated bya45-basesegment that is

78%purineon thecodingstrand(Fig. 1) ( 12).Aportionofthis

region is homologousto a segmentofthe EGF-Rpromoter that

binds atranscription factor named TC factor (TCF) for the

pyrimidine-rich natureof itsrecognition sequence( 13).This sequence in the EGF-R promoter appearstobe necessary for

efficient transcription of EGF-R ( 13),but its function in the

HER-2/neupromoterisnotknown.

Oligonucleotides havenowbeen showntoform interstrand

DNAtriple helices inanumberof naturally occurring purine rich sequences. The promoters

of

the humanc-myc,EGF-R, dihydrofolate reductase (DHFR), andIL-2 receptor(IL-2-R)

genes have been

successfully

targeted for

site-specific

triplex formation (14-18). In addition, sequences in the HIV long terminalrepeat(LTR) (19) and

coding

sequence(20)are tar-gets for triplex formation. Triplex formation in DNA target sequences caneffectively inhibit protein binding andgene

ex-pression (21). Thetranscription factorSp I hasbinding sites in

the DHFR promoter;triplex formationatthesesitesprevents

Spl

binding

(17).In

addition,

triplex

formation

prevents

nu-clear factor

kappa

B

(NFkB)

binding

tothe a-regulatory

se-quenceof the IL-2-Rpromoter(18).

Triplex-forming

oligonu-cleotides(TFOs) have been showntoinhibit in vitro transcrip-tionofc-mycandHIV-LTR( 15, 19). MessengerRNAlevels ofc-mychavebeenmeasuredinHeLa cellsinculture withand

without TFO, and these oligonucleotides apparently inhibit cellularc-myc

transcription

(22). CellularIL-2-R

transcription

is inhibited byaTFO inachloramphenicol acetyl transferase

assay(

18),

andaTFO has been shownto inhibitHIV

tran-scription

andviral

proliferation

in

virally

infected cells in

cul-ture

(19).

Interstrand DNAtriplex formationtypically occurs at

poly-purine/polypyrimidine

tracts (14-23). An exogenous

poly-purine

oligonucleotide formsthe third strand by occupying the

majorgrooveofthe native duplex and forming Hoogsteen

hy-drogen bonds withthepurinebasesofthe duplex (23-25). The

1.Abbreviations usedinthispaper:CMV, cytomegalovirus; DHFR,

dihydrofolate reductase; EGF-R,EGFreceptor;IL-2-R,IL-2receptor; LTR,long terminalrepeat;TCF,atranscription factor namedTF

fac-tor;TFOs,triplex-forming oligonucleotides.

J.Clin. Invest.

©The American Society for Clinical Investigation, Inc.

0021-9738/93/11/2433/07 $2.00

(3)

The HER-2/neu Promoter

5,

-200 Spl

Triplex Target Sequence

(.69) AGGAGAAGGAGGAGGTGGAGGAGGAGGG (-42) TCCTCTTCCTCCTCCACCTCCTCCTCCC

Spi .100 CCAAT TATA

10bp

Figure1. Mapof the HER-2/neu promotershowingthetriplextarget sequenceand relative positions of control elements includingtwo

Spl bindingsitesat -192to-183 and -136to-127,theCCAAT boxat-76to-73,and the TATAboxat-28to-25 relativetothe major transcription start site(+ 1 ).

basetriplets in thispurine:purine:pyrimidine (R:R:Y) triplex

include G:G:C andA:A:T(third strand basesinbold), and the

Hoogsteenbonding makes the triplexsequence specific(24, 25). While purine-rich sequences occur disproportionately in the promoters ofeukaryoticgenes(23),there are often

pyrimi-dines interspersed in the polypurine tract that make triplex

formation less favorable. Nonetheless, triplex formation has

been shown in severalnaturally occurringsequences that are not strictly homopurine:homopyrimidine (14-20).

Pyrimi-dineinterruptions of the polypurine tract potentially disrupt

theability of thesesequences to formtriplexbecause oftheloss ofHoogsteen hydrogen bonds (25, 26).It would be desirable to

bridge these interruptions in a manner that minimizes steric

interactions that could be destabilizingto triplex formation. Anabasic molecule that maintainsthe properinternucleotide spacing (27) would be ideally suitedassuchabridge.Wehave examined theeffect ofsubstitutinganabasicbridgefor a pyrim-idineresidue inapurine-richtriplex forming oligonucleotide.

Wedetermined the ability ofapolypurine oligonucleotide

toformaninterstrandtriplexwith thepurine-richsegment of the HER-2/neu promoter from -42 to -69 relativeto the

major

transcription

startsite. This sitewasshowntobind

nu-clear

protein

from a crude nuclear extract, and the effect of

triplex

formation on nuclear

protein binding

tothe

HER-2/

neupromoterandtothetarget sequence werealsoexamined. The

effect

of

triplex

formationon acell-freeinvitro transcrip-tionassay wasexamined. Thistargetsequencecontainsa

single

internal pyrimidine

(thymidine

at

-54)

and 27

purines.

We

evaluatedan

oligonucleotide

withasinglesubstitution ofthe

thymidine

residuebyanabasic linker molecule.

pH=8.0) at370C for 15min,thenat roomtemperature for 15min,

thenat4VC for 15 min.Glycerol (final concentration 5%) with brom-phenolblue and xylenecyanolwasusedtoload the reactiononto a20% nativepolyacrylamide gelthatwas run at80 V(8 V/cm)for 14 hat

4VC. Bandswere visualized byautoradiography. AHeLanuclear ex-tract wasprepared (28)foruseintheprotein bindingshifts.

Oligonucle-otideswere incubated withalabeledThaI digested fragment of the HER-2/neu promoter or with a labeled synthetic fragment in the

bufferused in othergelshiftsat roomtemperature for 30 min. Then 3

,ggof poly(dIdC) and 5

jLl

of theextract wereadded andincubatedfor 30 minat roomtemperature.Glycerol withbromphenolblue and xy-lenecyanolwasadded, and the reactionwasloadedonto a5% native

polyacrylamide gelandrunat 150Vfor 2-3 h followedby

autoradiog-raphy.

DNaseIfootprint.A267-bpfragment oftheHER-2/neu promoter wasamplified from HL-60 DNA by PCR and cloned into aplasmid

vector (TA1000;Invitrogen,SanDiego, CA).Thecodingstrandwas

labeledbyplasmiddigestionfollowed by the Klenow reaction. For the non-coding strand the downstream PCRprimerwas5'end labeledby

T4kinasereaction;end-labeledDNA wasobtainedbyPCR

amplifica-tionofthe HER-2 /neupromoter-plasmidconstruct.The sequences of thecloned andamplified DNAswereconfirmedbyMaxam-Gilbert

sequencing(29). DNA was incubated with theoligonucleotidein a

bufferconsisting of 90 mmTris-HCl, pH 7.4, 10 mmMgCl2as de-scribed above. The buffer differed from theoneused ingelshiftanalysis

becauseofimproved enzymatic activityof DNaseI.3

tig

of poly (dl-dC)wasadded. TheDNA wassubjectedto a5-mindigestion on ice with0.1-1 Uof DNase I and the reaction stopped by adding form-amidewithbromphenolblue and xylene cyanol and heating to950C

for 5 min and thenquickly cooling.Thesamples were loaded onto an 8%polyacrylamidegel, 8Min urea, and runat40Wfor 3 h followed

by autoradiography. The promoter fragment and triplex target

se-quenceareillustrated inFig. 1.

CalculationofKd. The ratio ofduplex (D)totriplex (T)was deter-mined bydensitometryofautoradiogramsand the Kdwascalculatedby

Triplex Target Site

5'AGGAGAAGGAGGAGGTGGAGGAGGAGGG 3' 3'TCCTCTTCCTCCTCCACCTCCTCCTCCC 3'

Antiparallel Oligonucleotide 5'GGGAGGAGGAGGTGGAGGAGGAAGAGGA 3'

Parallel Oligonucleotide

5'AGGAGAAGGAGGAGGTGGAGGAGGAGGG 3'

Linker Substituted Oligonucleotide

5'GGGAGGAGGAGGXGGAGGAGGAAGAGGA 3'

Methods

Oligonucleotidesynthesis. Oligonucleotides were synthesized by stan-dardphosphoramidite chemistry on a DNA synthesizer (Cyclone Plus;

MilliporeCorp., Bedford, MA), purified by reverse phase chromatogra-phy (Sep-Pak), and yields were checked by optical density at 260 nm. Theintegrity ofthe oligonucleotides was verified by gel electrophoresis.

Oligonucleotideswerelabeled byT4polynucleotide kinase reaction with32P-ATP.Thelinker substituted oligonucleotide was synthesized

bydissolvingthe solid reagent

(N-Fmoc-O'-DMT-03-cyanoethoxydi-isopropylaminophosphinyl-2-aminobutyryl-1,3-propanediol, Unilink Aminomodifier; Clontech, Palo Alto, CA) in anhydrous acetonitrile andplacingit in theauxiliaryportof the synthesizer. The sequences of

oligonucleotidesareillustrated in Fig. 2.

Gelmobility shift. Oligonucleotides were incubated together in a bufferconsistingof 90mmTris, 90mmborate, 10 mmMgCl2(final

Abasic Linker (with two guanine residues as GXG)

G

O-P=o

0*

( +

0 NH,

G o-P=O

(4)

D/T= KdX I/[Pu], where [Pu] is the final concentration of purine

oligonucleotide( 14).

In vitro transcription assay. The plasmid containing the 267-bp HER-2 / neupromoter includes the transcription start site and was used for in vitro transcription. NotI digestion of this plasmid yields a 61 0-bp fragment with 368 bp downstream of the transcription start site and will yield an RNA transcript of approximately this size in run-off tran-scription. The template DNA (1

Aig)

wasincubated with varying con-centrationsof the triplex-forming oligonucleotide or a purine-rich con-trololigonucleotide in a buffer identical to the one used for gel shift

analysis for I hat370C. As another control, a cytomegalovirus pro-moter template (100ng)yielding a 363-base transcript (Promega Bio-tec, Madison, WI) was handled identically to the HER-2/neu tem-plate. Asdiscussed in Results below, we found high concentrations of several oligonucleotides toinhibit transcription inthis assay system nonspecifically. Therefore, after triplex formation, the template DNA was passedthrough a spin column with a nominal molecular weight cut

offof 100 bp(Chromaspin-100 in DEPCH20;Clontech) as described by the manufacturer. This column allows the templateDNAand an-nealed TFO topass through into DEPC H20 while retaining- 90%of excessoligonucleotide and99%of the buffer ions. Afterpurifyingthe template DNA from excess oligonucleotide, transcription reactants

were added toachieve a final volume of 75

til

withfinal concentrations asfollows:MgCl2=2 mM;KCl=44mM;Tris-HCl, pH 7.4, 8.8 mM; Hepes, pH 7.9, 4.4 mM; EDTA = 88 tM;DTT = 220,uM;glycerol =8.8% vol/vol; ATP=400

,M;

UTP=400

,M;

CTP=400

,M;

GTP = 16 oM;

32P-{Ia}

GTP at3,000Ci/tomol= 30,uCi (totalactivity).

Transcription was initiated by adding HeLa nuclear extract (Promega Biotec) 102

Ag

and incubating at 30°C for 1 h. Transcription was stopped by adding 4.5 vol of a solution of0.3 MTris-HCl(pH 7.4), 0.3 Msodium acetate, 0.5% (vol/vol) SDS, 2mMEDTA, 3

tog/ml

transfer RNA. Reaction products were extracted once with

phenol/chloro-form/isoamylalcohol(25:24:1 ) and ethanol precipitated. Transcripts were resuspended in 50% formamide, denatured at 95°C, and sepa-rated on an 8M urea, 5% acrylamidegel. Bands were visualized by

autoradiographyandcompared toalabeled,

denaturedOX

174/ HaeIII digested molecular weight marker.

Results

Triplex formation and its effects on

protein binding

to the

HER-2/ neu promoter weredemonstratedby gel mobilityshift

and DNaseI protection assays. The triplextarget site in the

HER-2/neupromoterisa

purine-rich

sequence from -42 to -69 as illustrated in Fig. 1.Thetriplex-forming

oligonucleo-tides include the

antiparallel

andlinkersubstituted

antiparallel

sequences illustrated in Fig. 2. The parallel oligonucleotide (Fig. 2)was usedas acontrol

oligonucleotide

that would not formtriplexwith the target sequence.

Gel mobility shift analysis using oligonucleotides reveals

concentration dependent

triplex

formation

by

theantiparallel

polypurine

strandtargeted to the

HER-2/neu

promoter se-quence

(Fig. 3).

Gel shiftswere

performed

by

adding

succes-sively

greaterconcentrations of theparallelorantiparallel

pur-ine-rich

oligonucleotides

tothe

triplex

targetsequencelabeled

on the

pyrimidine-rich

strand under the conditions described. Theparallel purine-rich

oligonucleotide

is identicalto the

pur-ine-rich strand ofthe target sequence inboth sequence and

orientation; the

antiparallel

oligonucleotide

is identical in

se-quence, but reversed in orientation with respecttothe purine-richstrandofthe target sequence(Fig. 2). Triplex formation

occurs when the

triplex-forming oligonucleotide

is at 3.3

X l0-7 M concentration (100-fold excess with respect to the

duplex concentration),and theshiftiscompleteat3.3X 10-6

M (

1,000-fold

excess). This is consistent with a dissociation

control DNA

S D

oligonucleotide added(M)

parallel antiparallel

Conc(M) 5x109 5x1095x108 5x0 551065x105 3x1053x10°3x1O 3x10 3X10

-molar ratio - 10 100 7000 10,000 1 10 100 1000 10.000

T

T_

D --*- '

.''2

~",d, ak 70X

S

-*D-Figure 3. Gelmobilityshift analysisshowingconcentration depen-denttriplex formation.Inalllanes, the 28-basepyrimidinerich strand

ofthe target sequence is labeled. The labeled strandwasannealedto

itspurine-richcomplement, and the parallelorantiparallelpurine richoligonucleotides were added as described in Methods. Single and double strand controls wereincluded. The concentration of the target sequence was constant at 5 X 10-9Mfor the control lane and parallel lanes and at 3.3 x 10-9M(numbersinfigure have been truncated forclarity)in theantiparallel lanes. Theoligonucleotideadded,its

concentration, andits molarratiotothe targetDNA areindicated.

S,single strand DNA; D, duplex DNA; T,triplex DNA.

constant (Kd)of 4.5 X

10-'

M. Theantiparallel oligonucleo-tide was found to betriplex forming, but the parallel

oligonucle-otide failedtodemonstratetriplex formationatconcentrations

up to 5X I0-5M( 10,000-fold excess).

The sequence specificity of triplex formation was

con-firmedby DNase Iprotection footprinting(Fig. 4). A 267-bp

fragment of the HER-2/neu promoter was labeled on either

thecodingornoncodingstrand andincubated withincreasing concentrations of the triplex-forming oligonucleotide before

limiteddigestionby DNase I.Protection of thetarget sequence

bythetriplex-forming oligonucleotidewasobserved to be

con-centration dependentin a mannerconsistent withthegel

mobil-ity shift analysis. Protection from digestion isdemonstrated at

10-6Mconcentration of theantiparallel oligonucleotide, the

concentration at which the target duplex was completely shiftedto DNAtriplexinthegel shiftexperiments. The parallel

polypurine oligonucleotide,

however, didnotprotect the

HER-2/neupromoterfragment fromDNase Idigestionat the same

concentration.

The effectof triplex formationon protein bindingto the HER-2/neupromoter wasevaluated.Gelmobility shift analy-sis ofa 120-bp restriction

fragment

ofthe HER-2/neu pro-moter incubated with a HeLa nuclear extract shows several shifted bands

representing

different protein-DNA complexes

(Fig.

5). Protein bindingwasinhibited by

preincubation

of the HER-2/neupromoter

fragment

with the

triplex-forming

anti-parallel oligonucleotide at a concentration of 5 x 10-6 M

(1,000-foldexcess).Preventionof protein bindingwasshown todependontheconcentrationofthe

triplex-forming

oligonu-cleotide. The concentration of

oligonucleotide

that prevents

protein binding is consistent with the concentrationthat in-ducedcomplete

triplex

formationin theexperimentsdescribed above. Theparallel

polypurine oligonucleotide

didnotprevent

protein bindingatthis concentration.

Proteinbinding directlytothe28-bptarget sequencewas

(5)

nu-Coding Strand

antiparaelblollgonucleotid. G

10-8 1O-7 106 Io5 4105

1 10 100 1000 4000

Non-Coding

Strand

pauraols

DNA control

C alone entiparallelolgonal@O@dO ,o11go

Conc(M) - 5x10-9 5(10O5x1075x041.5x161.5x10d

molarratio - 2 20 200 2000 6000 6000

4*

w1:

I

-I

(-69)

A T

G Cl

G

C

A T G C

A T

A T

G C

G

C

A T

G C

G C

A T

G C

G C

T A

G C

G C

A T

G C

G

C

A T

G C

G

C

A T

G C

( c

(-42)

*i.> ..

4

.:

....

clearextract wasaddedto alabeled

synthetic

double stranded 28-mer

corresponding

tothetarget sequenceanda

single

domi-nant

shift

wasobserved.

Preincubation

of thetarget sequence

with the

triplex forming oligonucleotide

but notthe

parallel

polypurine

strand(2.5X 10-6Min both

cases) prevented

this shift.

The effects of triplex formation wereexamined in an in vitro "run off" transcription assay (Fig. 7). In thisassay, a

eukaryoticpromoteris incubated withaHeLanuclearextract inthepresenceof nucleotide

triphosphates;

transcriptsare la-beledby the

incorporation

of 32P-GTP and correspond in size

tothelength ofthetemplateDNA.ANotI

digest

ofthe plasmid

containing

theHER-2/neupromoter

yields

a6

10-bp fragment

with 368 bp downstream of the transcription startsite (thus yieldinga368-base transcript). When this promoter fragment

wasincubated withthetriplexforming oligonucleotide,a

con-centration dependent inhibition of HER-2/neu

transcription

wasobserved

(Fig.

7

A).

Substantialinhibitionof

transcription

* w: Figure 4. DNase I footprints of triplex formation by

the antiparallel purine-rich oligonucleotide in the

HER-2/neu promoter.267-bpfragmentsofHER-2/

neupromoterwerelabeled andincubated withthe

oligonucleotidesasdescribed. LimitedDNAse I

di-gestionwasperformed,then the DNA was resolved on an8% polyacrilamide sequencinggel 8 Min urea

followed by autoradiography. Maxam-Gilbert

se-quencing

reactionwere

performed (G

and C) and were runsimultaneouslyfor orientation; the target sequenceis indicated. Concentrationand molarratio of theoligonucleotides with respect to the HER-2/

neu promoterfragmentareindicated. The parallel

purine-rich oligonucleotidewasincludedas a control.

was seen witholigonucleotide concentrations of2.5 ,uM and complete inhibition oftranscription isseenwhen the oligonu-cleotideconcentration is25

,gM.

Apurine-rich control (non-triplex-forming) oligonucleotide targetted to anadjacent po-tential triplex siteintheHER-2/neupromoter had noeffecton in vitro transcription at 25

IM.

Ithas been shown that DHFR transcription in vitro is enhanced when thetemplateDNAisa digested plasmidcontainingboth the promoterand the plas-midbackbone(30).Weobserved similarresults withHER-2/

neu.Using digested plasmidDNA asthe template, high molec-ularweight bandsareobservedat - 600and 1,500 bases;

pre-sumably these bands represent interaction of the protein extractwith theplasmidbackbone and similar highmolecular

weight bands have beenobserved in DHFR transcription in vitro(30). Incontrast,twolower molecular weight transcripts

of - 250and 150 bases represent premature termination of

transcription. These lower molecular weight transcripts are also specificallyinhibitedbythe

triplex-forming

oligonucleo-A f:

(6)

oligonucleotide added (M)

antiparallel parallel

DNA + I Il

DNA extract -9

-8

7~

6 -6

alone alone 5x10 5x10 5x10 5x10 5x10

cleotide is retained in the column (datanotshown). Template DNApurifiedin thiswayretains fullpromoteractivity inthe absence of triplex.

The triplex target sequence in the HER-2/neu promoter

containsasingle pyrimidine residueat-54(Fig. 1).This

thy-midine residue is unable to Hoogsteen bondwith the

corre-sponding thymidine residue of the antiparallel triplex forming oligonucleotide, interrupting the R:R:Y triplex motifat this site. We sought to bridge thispyrimidine interruption with an abasic molecule that preserves internucleotide spacing, and comparethe relative triplex forming ability of the substituted oligonucleotide with the unsubstituted oligonucleotide.

Triplex formation by the oligonucleotide with the abasic linkersubstitution was demonstrated by gel mobility shift

anal-ysis (Fig. 8). Again, aconcentrationdependent shifttotriple helix formationwas observed. Thelinker substituted oligonu-cleotide inducestriplex formationat aconcentration ofl1-0 M,with acompleteshift totriplexat o0-6M.This is consistent with a Kd of - 5 x 10-7M. Therefore, the binding affinity of

the linkersubstitutedoligonucleotide was approximately equiv-alentto the affinity of the unsubstituted oligonucleotide.

Discussion

We have demonstrated theformationof an interstrand DNA triplex of the R:R:Y type by a polypurine oligonucleotide

tar-oligonucleotide

added

DNA+ II

DNA extract

alone alone

AP

P

Figure5.Gel mobility shift analysis demonstrating protein binding by

anuclear extract to the HER-2 / neu promoter. A 1 9-bp fragment of theHER-2/ neupromoter was labeled and incubated with a HeLa nuclear extract asdescribed. In the "oligonucleotide added" lanes, theDNAwaspreincubated with the antiparallel or parallel purine-richoligonucleotides at the concentrations indicated. Arrows indicate protein-DNA complexes that are prevented by the triplex formation.

tide inaconcentrationdependent manner. Tofurther demon-stratethespecificity oftheoligonucleotide effecton

transcrip-tion,

theCMVpromoter was used as anegativecontrol (Fig. 7 B). A 363-base transcriptis synthesized from this CMV

tem-plateinvitro,and neitherthetriplex-forming oligonucleotide nor the control oligonucleotide had an effect on CMV

tran-scription in thisassay system.

Itshouldbementionedthat large excesses of several oligo-nucleotides (triplex-forming and non-triplex-forming) had a

nonspecific inhibitory effect on the in vitro transcription of both the HER-2/neu and CMV genes (data not shown). This

nonspecific effect was overcome by purifying the template DNAfromexcessoligonucleotide by molecular-sieve

chroma-tography. Specifically, the template DNA wasincubated with the oligonucleotide and then passed through a spin column with anominal molecular

weight

cut off of 100bp (50 kD). Oligonucleotide boundto the template DNApassed through thecolumn as a stable triplex, while > 90% ofunbound

oligonu-Figure 6. Gelmobility

shiftanalysisofprotein bindingtothetriplex

target sequencebya HeLa nuclearextract.

Thepyrimidine-rich

strand of the target

se-quencewassynthesized,

labeled, and annealed

toits complementas

described,then incu-bated witha HeLa nu-clearextract. Inthe

"1oligonucleotide

added"lanes,the DNA waspreincubated with theantiparallel (AP)or parallel(P)

oligonucle-otidesatconcentrations of 2.5 x 10-6Mto show thattriplex

(7)

B CMV Figure 7. In vitro

tran-Transcription scriptionassay

demon-° strating specific

inhibi-tion ofHER-2/neu

o transcription. (A)

Tran-LZ

5> >scriptionfroma

di-,n0,,¢ gestedplasmid

contain-NZ ON

ing

the

HER-2/neu

promoter-template. (B)

*iw Transcriptionfroma

purified CMVpromoter template.

Oligonucleo-* -603 tideswereaddedtothe templateDNAas

de-scribedatthe

concen-trations indicatedabove

each lane.HER-2/neu transcription yieldsa

*9 ' 363 full-length368-base

RNAtranscript

(arrow),

aswellastwo

smallerprematurely ter-minated transcriptsof

- 250 and 150bases.

These transcriptsare

inhibited by triplex

for-mation with the

anti-parallel28 baseoligonucleotideinaconcentrationdependentfashion.

Apurine-rich oligonucleotide antiparallel toanadjacenttarget site doesnotformtriplexandhasnoeffectonin vitrotranscription (control oligo). Largemolecularweightbands in A at 600 and1,500

basesrepresentinteractionof the HeLa extract with theplasmid

backbone andarenotHER-2/neu transcripts. CMVtranscription

withapurifiedpromotertemplate yieldsasingle363-basetranscript

(arrow, B).Neither the TFOnorthe controloligonucleotidehave aneffectonCMVtranscription. MWM,OX174/HaeIII molecular

weightmarker with the 310and 603 base standards indicated.

gettedtotheHER-2/neupromoter.Triplexformationblocked

proteinbindingtothetarget site andpossiblytoother sitesas

well.Triplex formationwasdependentonthepresenceof

mag-nesium ion, but occurred atphysiologic pHwith nanomolar

concentrations ofoligonucleotide. Triplexformation isasite

specific phenomenon in which theoligonucleotideforms

spe-cific G:G:Cand A:A:Ttripletswith the HER-2/neu promoter

target sequencebyHoogsteen hydrogen bonding. Several

re-cent reports suggest thattriplex formation ispossible under

nearphysiologic conditionsofpH, temperature,and ion

con-centrations, and triplex forming oligonucleotides targettedto

cellulargenes maybe expectedto formtriplexstructuresinthe

living cell ( 14, 22). Triplex formationin thec-mycpromoter of

cultured HeLa cells has recently been reported (22). In this

setting, triplex-forming oligonucleotides may have potential applicationasagents thattargetageneinasequencespecific

mannerandpreventgeneexpression by inhibiting

transcrip-tion. Inhibitionof invitroand in vivotranscriptionof another

biologicallyandclinically importantoncogene, c-myc, has

al-ready been demonstrated with triplex formation ( 14, 22). We haveshown the specific inhibition ofHER-2/neutranscription

invitro through triplex formation.

Thetargetsequence for triplex formationin theHER-2/ neuoncogene reported here is likelytobe important forthe

regulation and efficient transcription of the HER-2/neugene.

We have shown that thissequenceisabindingsite foranuclear

protein. The identity of the protein and its significancearenot

known,but thesequencehomology with the TC factor

recogni-tion site in the EGF-Rpromoter( 13)suggestthat itmaybe of

importance in HER-2/neu expression. Inhibition of HER-2/

neutranscription by triplex formationmay occurbyblocking

thebinding of this protein, whichmaybeatranscriptionfactor necessaryfor efficientHER-2/neu expression. In addition, the

triplextargetsiteis flanked bytwoimportant and well-recog-nized transcription factor recognition sites, the CCAAT

se-quence(recognition site for the protein CTF) and the TATA

box (recognition site for the essential transcription factor TFIID) (Fig. 1). Triplex formation in thetargetregion may

have aneffect on the nearby binding of thesetwofactors by changing the local DNA conformation of the flanking se-quences,thereby preventing formation of the transcription ini-tiation complex whereby RNA polymerase recognizes the

tran-scriptionstartsite(+ 1)tobegin RNA synthesis (31 ). Finally, triplex formationmayaffect the DNA conformationatdistal sites (31); triplex formationatthetargetsite of HER-2/neu

maythereforepreventRNApolymeraseIIitself from binding

tothe DNA template.

Animportant limitationtotheapplication oftriplex forma-tiontonaturally occurring DNAsequencesis the requirement

for polypurine:polypyrimidine tracts. Although these se-quences occurinnature,naturallyoccurring polypurinetracts

arefrequently interrupted byone orseveral pyrimidines. These

pyrimidine residues could disrupt triplex formation inan

oth-erwise favorable triplextargetsiteby loss of triplet hydrogen bonds and unfavorable steric interaction of the pyrimidines (26). Wehave examinedanabasic linker molecule that could

provideabridge overthe pyrimidine interruptions that often are found in naturally occurring purine-rich sequences.

Re-cently, aneutral abasiclinkerwasreportedtosubstitute fora

base inapolypyrimidine triplex forming oligonucleotide ina

Y:R:Y triplex motif (32). We have shown that a positively

chargedlinker substitutedoligonucleotide retains the abilityto

formtriplex with itstargetsequenceinanR:R:Y triplex. It is

control Xa

DNA

S D I

M-EA

5x10 10-4

20.000

10. 0.2 Conc(M) 5x10-9

molarratio

T -_

-0*

oNgonuclootideadded(M)

lo-8 o7 1o-6 10o5 104 2 20 200 2000 20,000

Figure8. Gelmobilityshiftanalysisdemonstratingconcentration

de-pendenttriplexformationbythe linkersubstitutedoligonucleotide.

Conditions for theexperimentweresimilartothose describedfor

othergelshifts.Singleand double strandedcontrolswereincluded,

andtriplexformation with the unmodifiedoligonucleotidewas

in-cludedasacontrol. Theconcentrationofoligonucleotideaddedas

wellasthe molar ratiotothe labeled DNAareindicated.S, single

strandDNA; D,double strandDNA;T,triplestrand DNA. A HER-2lneu

Transcription

-5

TFO o

00

2.

603-368 ,:

310- ^

*A5

v"W'",

a'

pop

(8)

notsurprising that asinglebase substitution does not cause a

substantialincrease inaffinity, since the loss oraddition of a single hydrogen or salt bond would not have a large effect on the binding of this already stable triplex. Clearly, complete binding is observed in the substituted oligonucleotide and this linker molecule may be useful in targeting sequences with sev-eral interruptions in thepolypurine motif that will not form triplex with otheroligonucleotides.The positivechargeon the amino group of this abasic linker may be responsible for its ability to retain equalbinding affinity compared to the

unsub-stituted oligonucleotide. A previously reportedabasic linker molecule is neutral and apparently confers a lower binding affinityonoligonucleotides ina Y:R:Ytriplex(32).

The Hoogsteenbonding responsiblefor triplex formation

can be expected to confer gene specificity on a triplex-forming

oligonucleotide. The conditions ofR:R:Y triplex formation

suggestthatitmay bepossibleto target anoligonucleotideto

thepromoterofalivingcell.Triplexformation may represent animportanttool formodulation ofgeneexpressionin a gene

specificmanner. The datapresentedinthisreportsuggest the potential future applicationof this toolinthe specific modula-tion of HER-2/neu expression.

Acknowledgments

Thiswork wassupported bygrantsfrom theAmerican CancerSociety (PRTF- 153) (S. W.Ebbinghaus), Public Health Service(CA 09467-08) (S.W.Ebbinghaus),National Institutes of Health(CA42664and

ROI CA 42337) (D. M.Miller), U. S.Army(DAMD 17-93-J-3018) (D. M. Miller), The ShareFoundation (D. M. Miller), anda National

ResearchService Award(HL07457)(C.A.Mayfield).

References

1.Schubert, D., S.Heinemann, W.Carlisle, H. Tarikas, B.Kimes,J.Patrick,

H. Steinbach, W. Culp, and B. L. Brandt. 1974. Clonal celllines fromthe rat central nervous system. Nature(Lond.).249:224-227.

2.Shih, C., L. C. Padhy, M. Murray, and R. A.Weinberg. 1981.Transforming

genesof carcinomas and neuroblastomas introduced into mousefibroblasts. Na-ture(Lond.).290:261-264.

3.Schechter,A.L., D. F.Stem,L.Vaidyanathan,S. J.Decker,J.A.Drebin,

M. I.Greene, R. A.Weinberg. 1984. The neu oncogene: an erb-B-related gene

encodinga185,000-M,tumorantigen.Nature(Lond.).312:513-516. 4. L.Coussens, T. L. Yang-Feng, Y. L. E.Chen, A. Gray, J. McGarth, P. H. Seeburg, T. A.Libermann,J. Schlessinger, U. Francke,A.Levinson,and A. Ullrich. 1985.Tyrosinekinase receptorwith extensivehomology to EGF receptor shares chromosomal location with neu oncogene. Science (Wash. DC).

230:1132-1139.

5. Bargmann, C. I., and R.A.Weinberg. 1988.Oncogenic activation ofthe neu-encoded receptorprotein bypoint mutation anddeletion.EMBO (Eur. Mol. Biol. Organ.) J. 7:2043-2052.

6.Hung,M-C., D. H. Yan, X. Zhao. 1989.Amplificationof the proto-neu oncogenefacilitatesoncogenicactivationby asingle pointmutation. Proc. Natl.

Acad.Sci. USA. 86:2545-2548.

7.Bargmann, C. I., M. C. Hung, R. A.Weinberg. 1986. The neu oncogene encodes an epidermal growthfactor receptor-related protein. Nature(Lond.). 3 19:226-230.

8.Peles, E., S. S. Bacus, R. A.Koski,H. S.Lu,D.Wen, S. G.Ogden, R.B. Levy, and Y. Yarden. 1992. Isolation of the neu/HER-2 stimulatory ligand: a 22 kdglycoprotein that induces differentiation of mammary tumor cells. Cell.

69:205-216.

9.Slamon,D.J., G. M.Clark, S. G. Wong,W.J.Levin,A.Ullrich,W. L. McGuire. 1987. Human breast cancer: correlation of relapse andsurvival with amplification of the HER-2/neu oncogene. Science(Wash. DC). 25:177-182.

10.Slamon,D. J., W.Godolphin,L.A.Jones, J.A.Holt,S. G. Wong, D. E. Keith, W. J.Levin,S.G.Stuart, J. Udove, A.Ullrich,and M. F. Press.1989.

Studies of the HER-2/neu proto-oncogene in human breast and ovarian cancer. Science (Wash.DC). 244:707-712.

1 1.Lemoine,N.R., S. Staddon, C.Dickson, D.M.Barnes, andW.J.Gullick. 1990.Absenceof activating transmembrane mutations in the c-erbB-2 proto-on-cogene.Oncogene.5:237-239,1990.

12.Ishii,S., F. Imamoto, Y. Yamanashi, K. Toyoshima, and T. Yamamoto. 1987.Characterizationofthe promoterregionofthe human c-erbB-2 proto-on-cogene. Proc. Natl. Acad. Sci. USA. 84:4374-4378.

13.Johnson,A.C., Y.Jinno,andG. T. Merlino. 1988. Modulation of

epider-malgrowthfactor receptor proto-oncogenetranscriptionbyapromotersite sensi-tive to S1 nuclease. Mol.Cell. Biol. 8:4174-4184.

14.Cooney, M., G. Czemuszewicz,E.H.Postel, S. J.Flint,and M. E.Hogan. 1988.Site-specific oligonucleotide binding repressestranscriptionof the human c-myc gene in vitro. Science(Wash. DC).241:456-459.

15.Durland, R. H., D. J. Kessler, S.Gunnell,M.Duvis,B. M.Pettitt,and M. E.Hogan. 1991.Binding of triple helixforming oligonucleotidestosites in gene promoters.Biochemistry.30:9246-9255.

16.Blume, S. W., J. E. Gee, K. Shrestha, and D. M. Miller. 1992.Triplehelix formation bypurine-rich oligonucleotides targetedtothe humandihydrofolate

reductase promoter.Nucleic AcidsRes.20:1777-1784.

17.Gee, J. E., S. Blume, R. C. Snyder, R. Ray, and D. M. Miller. 1992.Triplex formationpreventsSpIbindingtothedihydrofolatereductase promoter. J. Biol. Chem.267:11, 163-11,167.

18.Grigoriev,M., D. Praseuth, P.Robin,A.Hemar, T.Saison-Behmoaras,A. Dautry-Varsat, N. Thuong,C. Helene, and A. Harel-Bellan. 1992.A

triplex-he-lix-formingoligonucleotide-intercalator conjugateacts asatranscriptional re-pressorviainhibition of NF KB binding tointerleukin-2receptora-regulatory sequence.J.Biol.Chem.267:3389-3395.

19.McShan, W. M., R. D. Rossen, A. H. Laughter, J.Trial,D.J.Kessler, J. G. Zendegui, M. E. Hogan, and F. M. Orson. 1992.Inhibition of Transcription of

HIV-1 inInfected Human Cells byOligodeoxynucleotides DesignedtoForm DNATriple Helices. J. Biol. Chem. 267:5712-5721.

20.Giovannangeli,C., M. Rougee,T.Garestier,N.Thuong, and C. Helene. 1992.Triple-helix formationbyoligonucleotidescontaining the three bases thy-mine,cytosine,andguanine.Proc.Natl.Acad.Sci. USA.89:8631-8635.

21. Maher,L.J.,III,B.Wold, and P. B. Dervan. 1989.InhibitionofDNA

binding proteins by oligonucleotide-directed triple helix formation. Science

(Wash. DC). 245:725-730.

22.Postel,E.H.,S.J.Flint,D.J.Kessler, M. E.Hogan. 1991.Evidencethata

triplexforming oligodeooxyribonucleotide bindstothe c-myc promoter in HeLa

cells,therebyreducingc-myc mRNA levels. Proc.Natl.Acad.Sci. USA 88:8227-31.

23.Wells, R. D., D.A.Collier,J.C. Hanvey, M.Shimizu,and F. Wohlrab.

1988.Thechemistryandbiology ofunusual DNA structuresadopted by oligo-purineoligopyrimidinesequences. FASEB(Fed.Am.Soc.Exp. Biol.)J. 2:2939-2949.

24.Letai,A.G.,M. A.Palladino,E.Fromm, V.Rizzo,andJ. R. Fresco.1988.

Specificityinformation of triple-stranded nucleic acid helicalcomplexes:studies withagarose-linkedpolyribonucleotide affinity columns. Biochemistry. 27:9108-9112.

25.Beal,P. A., andP.B. Dervan.1991. Second structuralmotif for

recogni-tionofDNAby oligonucleotide-directed triple-helix formation.Science(Wash.

DC). 251:1360-1363.

26. Mergny, J. L., J. S. Sun,M.Rougee, T.Montenay-Garestier,F. Barcelo, J.

Chomilier,and C. Helene. 1991.Sequencespecificityintriple-helix formation: experimentalandtheoretical studies oftheeffect ofmismatchesontriplex stabil-ity.Biochemistry. 30:9791-9798.

27.Nelson, P. S., R. Sherman-Gold, and R. Leon. 1989.A newandversatile

reagent forincorporatingmultiplealiphaticamines into synthetic

oligonucleo-tides.NucleicAcidsRes. 17:7179-7186.

28. Dignam,J. D.Preparationof extracts from higher eurkaryotes. 1990.

Methods Enzymol. 182:194-203.

29. Maxam,A.and W.Gilbert. 1980.Sequencing end-labeled DNAwith

base-specificchemicalcleavages. MethodsEnzymol.65:499-560.

30.Farnham,P. J., and R. T. Schimke. 1986. In vitro transcription and delimitationof promoter elements of the murine dihydrofolate reductase gene. Mo!.Cell. Biol. 6:2392-2401.

31.MaherIII,L.J.,P. B.Dervan, and B. Wold.1992.Analysis of

promoter-specific repressionbytriple-helicalDNAcomplexes inaeukaryotic cell-free

tran-scriptionsystem.Biochemistry.31:70-81.

Figure

Updating...