Transcriptional Activation of the Alpha-1 Acid Glycoprotein Gene by YY1 Is Mediated by Its Functional Interaction with a Negative Transcription Factor

MaryAnnLiebert, Inc.,Publishers Pp. 1029-1036

Transcriptional

Activation

of the

_Alpha-1

Acid

_Glycoprotein

Gene

_by

YY1 Is

Mediated

_by

Its

Functional Interaction with

a

Negative Transcription

Factor

YU-MAY

LEE

and SHENG-CHUNG LEE

ABSTRACT

Regulation

of

alpha-1

acid

_glycoprotein

(AGP)

gene

expression

involves both

positive

and

negative

transcription

factors. We have

_previously

identified two dominant factors:

_positive

and

_negative

transcription factors,

AGP/EBP

and factor

_{B, respectively,}

involved in the

_{transcription}

ofAGP and other

acute-phase

response genes. In this

report,

we

_present

evidence

showing

that the

_{transcription}

of the AGP gene is

positively regulated by

a

transcription

factor,

YY1. The activation of AGP gene

by

YY1 ismediated

by

a

negative

element B in the

AGP

promoter

region.

YY1 can also activate the B motif linked to a

heterologous

promoter. However,

YY1 does not bind

_directly

to the B motifper se.

Rather,

our data

suggest

that the activation of AGPgene

_by

YY1

may be mediated

_by

its functional

interaction

withfactor

B,

which

_recognizes

the B motif.

INTRODUCTION

Alpha-1

ACID glycoprotein

(AGP)

is one of the

most

abundant,

_{acute-phase-responsive, liver-specific}

plasma proteins.

The

_biological

function of AGP is

un-known,

but there are indications that it _{may suppress the}

immune_response(BennettandSchmid,

1980).

Some exper-iments_suggestthat AGP is a

nonspecific

antiinfection_agent

(Friedmann, 1983),

and that it _possesses nerve

growth-promoting activity

(Liu

et al.,

1988).

To

_investigate

the

regulation

of AGP gene

expression during physiological

homeostasis and

_{during perturbed physiological}

conditions (i.e.,acute

inflammation),

it isessentialto

_study

the

_complex

interactionsofvarious

_{fratts-acting}

factorsin these different

physiological

states.

In rats and

mice,

the levels of liver AGP mRNA and

plasma

AGP

_protein

increase10-to 100-fold within24hr of

experimentally

induced inflammation

(Baumann

and Held,

1981;

Baumannet_{al., 1983, Kulkarnieío/., 1985;}Baumann

and

_Maquat

1986;

Darlington

et al., 1986; Gauldi etal.,

1987;

Kleinétal.,

1988;

Prowseand

Baumann,

1988).

The increase in mRNA is

_primarily

attributedto

_{transcriptional}

activationofthe AGP_gene

(Kulkami

etal., 1985;Baumann

and

_Maquat,

1986).

We have

_previously

identified four motifs

_{recognized by}

a

positive transcription

factor,

AGP/

EBP,

and a

negative

eis element

recognized by

a

negative

factor_(i.e.,factor

_B)

_(Change/

al.,

1990;

Leeetal.,

_1993).

During

the

_acute-phase

_response,there is dramatic increase in the levelof

AGP/EBP,

coupled

tothe decrease infactor

B,

resulting

in the induction of AGP gene

transcription

(Leeet al.,

1993).

In additiontoAGP/EBP and factor

B,

glucocor-ticoids

(Klein

et _{al., 1988)} have also been shown to be involved in the induction of AGP gene.

Despite

the substan-tial progress in

understanding

the factors in

_regulating

the activationof AGP_gene

_during

the

_acute-phase

_response,the

regulation

of_{this gene in normal}

_{physiological}

statesremains

tobe

_{investigated.}

Theupstream

regulatory region

of the AGPgenecontains

a_sequence,GA ACATT TT(

—

121to

-113)

that is similarto

amotif

_{recognized by}

a

transcription

_factor,Y Y1 (consensus

sequence,

CGACATTTT) (Shi

et al.,

1991).

YY1 is a

human

_{Kriippel-related protein}

that interacts with DNA elements

(CGACATTTT,

located at -50 to

_-70;

CTC-CATTTT, locatedat the

_{transcription}

initiation

site)

of the adeno-associated virus P5 _promoter. These DNAelements

are

capable

of

repressing transcription

directed

by

heterolo-Institute of

Biological

Chemistry. AcademiaSinica,and Institute of MolecularMedicine,and Institute of Clinical Medicine,

College

of

Medicine,National Taiwan

University, Taipei,

Taiwan.

gouspromoters.El Anot

_only

relieves

_repression

exerted

_by

YY1 but also stimulates

_{transcription through}

the YY1

binding

site

(Shi

et

al.,

_1991).

There is evidencethatYY1 mediates both the

_repression

and activation_responses

(Shi

et

al.,

1991).

The

_ability

ofYY1 to mediate

_opposite

effects

depends

on the intracellular milieu

(e.g.,

the _presence or

absence ofEl

A).

YY1 isrelatedtothe GL1

_{-Krüppel family}

of_genes

_(Ruppert

et

al., 1988).

The

Drosophila Krüppel

protein

canalso repressoractivate

transcription,

_depending

onthecontextofits

_binding

site

(Frasch

andLevine,

1987;

Ruppert

et

_al.,

1988;

Licht et

al., 1990).

Two different

domains in the

_{Krüppel protein}

havebeen

_reported

toexhibit

repression activity

when fused to a

heterologous

DNA-binding

domain. It remains to be determined whether the

Krüppel protein

containsanactivation domain. In additionto

YY1,

several

_{transcription}

factors identical or

homologous

to YY1 have been identified and cloned.

_Among

them are:

protein

delta,

which bindsto_sequenceelements

(GCNGC-CATC)

downstream of the

_{transcriptional}

start sites in ribosomal

_protein

_{genes and}functionsas anactivator

(Hari-haran et al.,

1991); NF-El,

which binds to the

_IgK

3' enhancer

(CCACCTCCATCTT)

and functionsas a_repressor

(Park

and

Atchison, 1991); UCRBP,

whichbindstotheUCR

core

(CGCCATTTT)

in the

long

terminal_repeatof

Moloney

leukemia virus

(which

also functionsas a

_repressor)

(Flana-ganetal.,

1992);

andF-ACT

1,

which binds

competitively

to

themost

_proximal

serum_responseelement

(CGCCATATT)

of theskeletala-actin_{gene and functions}as a_repressor

(Lee

etal., 1992).

These dataindicate that the

_{binding specificity}

of YY1 orits

homolog

is

unusually

diverse.

Furthermore,

a

negative regulatory

domain in the human

_{papillomavirus}

type 18 _promoter was shown to be the _target site ofYY1

(Bauknecht

etal.,

1992).

Taken

_together,

these data

_suggest

that,

depending

on the _context, YY1 can function as an

activator,

a _repressor, or an initiator of

transcription.

For

thesereasons,wedecidedto

_investigate

the

_possible

involve-mentof the YYl-like motifin the

_regulation

_{of AGP gene}

expression.

In this

_study

we examined the

regulation

of

wild-type

and a series of mutants of AGP _promoter in transient transfection assays. Wefound that YY1 can

acti-vate AGP _promoter

_by

_relieving

the

_negative

action ofB

element,locatedat—40to—66.

However,

YY 1 didnotbind

directly

to this

_negative

element.

Instead,

it relieves the

repression

of AGP _gene

_by

its functional interaction with factor B.

MATERIALS AND METHODS

Preparation of

nuclearextract

Liver nuclearextractwas

prepared according

tothe proce-dures of Gorskietal.

(1986)

from 200-to

_300-gram

Wistar

ratseitheruntreatedortreated with

lipopolysaccharide

(LPS)

(10

p,g/gram

body weight)

for 4 hr. Detailed

_procedures

were asdescribed

_(Gorski

et

al.,

_1986).

Plasmids and

_{oligonucleotides}

The

_{AGP/CAT, AGP/EBP/CAT,}

and CMV-AGP/EBP

were as described

(Chang

etal.,

1990;

Lee etal.,

1993).

CMV-YY1

_(kindly

_{provided by}

Dr. T.

Shenk)

wasdetailed

elsewhere

(Shi

etal.,

1991).

Various CATconstructs

con-taining

the mutated AGP_{promotersequence}were

generated

essentially

asdescribed

(Lee

etal.,

1993).

pCAT-promoter

(CAT

reporter gene controlled

by

SV40

promoter)

and

pSV-ß-Gal (ß-galactosidase

reporter gene controlled

by

SV40

_promoter)

wereobtained from

Promega

(WI).

These

two_reportervectorswereusedfor

constructing

the

recombi-nant

_{plasmids containing}

the

_oligomeric

B motif. All the

constructswereverified

_by

_sequence

_analysis.

The

oligonu-cleotides used were as follows:

B,

TACTGTCCCTGGCT-TCAGTCCCATGCCCT;

U

(UCR),

TAACGCCATTTTG-CAAGGCAT;

mutant UCR

(mt

U);

TAAATACATTTT-GCAAGGCAT.

Analysis of protein—DNA

interaction

Gel

_{mobility-shift}

and

_footprinting

_assays were as

de-scribed (Lee et al.,

_1993).

Briefly,

an end-labeled DNA

fragment

(—180

to +10 from AGP_promoter

_region)

₍₅

_ng)

generated by polymerase

chain reactions

(PCR)

wasaddedto

a

_20-p.l

reaction mixture

_containing

50 mM NaCl, 20 mM Tris-HCl

_pH

7.6,

1 mM

_MgCl2,

0.2 mM

EDTA,

10%

glycerol,

5 mMDTT,and 1 p,g

poly(dl-dC).

Nuclearextract

or

partially purified

fraction

containing

factorBwas

prein-cubated with recombinantYY 1or

partially

purified

YY 1 for

5 minatroom_temperaturefollowed

by adding

the

probe

and

incubated for 90 min on ice. DNase I

(30

p.g/ml, Sigma)

freshly

dilutedin 10mM

_MgCL

and 5 mM

_CaCl2

wasadded

tothe reaction mixture

_(usually

1-3

_pJ

was_used).

Digestion

was

performed

for2-3 minonice and

stopped by

the addition

of 80

_pd

of_stopsolution

_containing

75 u-gof_yeasttRNA/ml,

20 mM EDTA, and 0.5%

_NaDodS04.

The

_samples

were

extracted with

_phenol

chloroform,

precipitated

with ethanol

at -70°C. The

_pellets

were dried and

resuspended

in

se-quencing

solution (95%

formamide,

1%

_{xylene cyanol}

FF,

and 1%

_bromphenol

blue),

heated at 95°C for 3

min,

and loadedontoa

sequencing gel.

For

_{gel mobility-shift}

assays,

protein-DNA complexes

wereformed asdescribed forthe

footprinting

_{assays. After}

incubation for 20 min at room _temperature, 1

_|xl

of 1%

bromphenol

bluewas

added,

and the

sample

wasloadedonto

a 5%

polyacrylamide gel (acrylamide/bisacrylamide

=

30:1)

containing

4%

_glycerol

in

_Tris-Glycin

buffer.

Electro-phoresis

was

performed

at 150 Vatroom _temperaturewith

buffercirculation.

Recombinant YYI and

_antibody

to YY1

Recombinant YYI was

generated

by cloning

the Nco

l-Eco RI

_fragment

of YYI cDNA

(YYI

plasmid

was

kindly

provided by

Dr. T.

Shenk,

Princeton

_University)

into

plas-mid

_pRSET

_(Invitrogen,

San

_Diego,

CA).

Therecombinant YYI

_expressed

in Escherichia coli

[BL21 (DE3)]

was

purified

by

Ni column anddetected

_by

Western blot

_using

rabbit

_polyclonal

anti-YYl. Rabbitanti-YYl was

generated

by

immunizing

the

_{antigen expressed}

inE. coli.

_Briefly,

a

cDNA

_{fragment corresponding}

to

_polypeptide

of232 amino acids

_containing

the

_{carboxy-terminal}

_part of YYI was

sequence was verified

_by

_sequence

analysis.

The cDNA

fragment

was cloned into

pRSET

(Invitrogen,

San

Diego,

CA)

and

_expressed

in

BL21(DE3).

Cell

cultures,

DNA

_{transfections,}

_{and CAT assays}

Baby

hamster

_kidney

(BHK)

cells were cultured in

Dul-becco's modified

_Eagle's

medium

_supplemented

with 10% fetal calfserum. DNAtransfectionswere

performed by

the

calcium

_phosphate

_{precipitation technique.}

For each 10-cm

petri

dish,

the calcium

_{phosphate-DNA precipitate}

con-tained 8 p.g of _target

_plasmid,

and 3 _u,g CMV-YYI or

amino-terminal deletedYYI

(CMV-C250),

whichcontains the

_{carboxy-terminal}

250 amino acids.Plasmid

_pGEM4

was

usedas acarrierto make the finalDNA to_{be 15 p-g.} Cells

wereharvested48hr

post-transfection.

Cellextract(30

_u,g)

was usedfor CAT_assay. CAT

_activity

was

quantitated by

densitometric

_scanning

of the

_{autoradiogram}

or

by scanning

ofthe TLC

_plate

with BioRad

_Image

_analysis

_system.

Western blot

_analysis

Rat liver nuclear extracts were

prepared

as described

(Chang

etal.,

1990).

Liver nuclearextracts were obtained

fromcontrol and LPS-treated Wistarrats.

_Varying

amounts

of nuclear extracts were resolved on a 10%

NaPodS04-polyacrylamide gel.

Proteinswereelectroblottedto

nitrocel-lulosemembrane.The

_protein

blotwas

subjected

tostandard Western blot

_development

_using

rabbit anti-YYl and _goat anti-rabbit

_IgG

alkaline

_{phosphatase conjugate.}

< ü -122 GGAACATTTTGCGCAAGA -115 Gl TTC G2 CAC G3 G4 GGG GTA

FIG. 1. 7Va«s-activation of the AGP _gene

_by

YYI. Cotransfection assays ofCMV-YY1

_expression

vectorand

wild-type

and mutated AGP _sequences linked to CAT

re-porter gene. The _upstream

_regulatory

sequence

(—122

to

-155)

of the AGP_{gene and the}

_{corresponding}

mutantsare

shown in the upper

panel.

The results of

CAT

_assay are

shown in the lower

_panel.

CMV-YYI

₍₃

_p,g)

_{and 10 p,g}of reportergenewerecotransfected intoBHK cells.

(—

)Cells

transfected with _{reporter gene without}

CMV-YY1;

(

+

₎

cotransfectedwith CMV-YY1.

RESULTS

Transcriptional

activation

_of

AGP

promoter

by

YYI

To

_investigate

the

_possible

involvementofYYI in AGP gene

regulation,

we initiated functional and biochemical

characterizations of YYI on AGP

regulation.

BHK cells were used for the transfection assays because BHK and

HepG2

cells behaved

_similarly

as demonstrated

by

our

previous

studies

(Lee

et

al.,

1993).

Transfection of

wild-type

and severalmutantsof AGP-CAT

_together

withCMV-YYI showed that YYI could activate AGP _promoter

_(Fig.

1). Mutationsinthe_sequence,

GAACATTTT,

donot_{have any}

effecton YYl'sactivation. These results indicate thatYYI can activate AGP

transcription,

but

_{independently}

of the sequence, GAACATTTT. Further

experiments by gel

mo-bility-shift

or

footprinting

_assay

using

recombinant YYI

derived fromE.coli have failedtoshowthat Y Y1couldbind

tothis_sequence

(data

not

_shown).

_Having

demonstrated that YYI couldrra/w-activate AGP gene, we tested its

possible

interaction with themost

_prominant

activator for AGP_gene, AGP/EBP. Itturnedoutthat activation of AGP_gene

_by

YYI and AGP/EBPis additive

_(Fig.

2). Therefore,

the activation ofAGP gene

by

YYI and AGP/EBPare

independent

event.

To understand the

_implications

of these data

further,

we

generated

a series ofmutants

_spanning

from

—

140 to —94 and a B-element-truncated mutant

(Lee

etal.,

1993).

The

nucleotidesequence and theco-transfectiondataare shown

in

_Fig.

3. YYI could activate

_wild-type

aswell as _every

8

F-< 20

_H

Ü o o + +

AGP/EBP

YY1

FIG. 2. The activation of AGP _promoter function

_by

AGP/EBP and YY-1 is additive.

AGP/CAT,

CMV-YY1,

andCMV-AGP/EBPweredetailed in the

previous

sectionor

elsewhere. BHKcellswereused forthetransfection assays.

AGP/CAT

₍₁₀

_p.g)

and 1 pug of YY1 and/or CMV-AGP/EBPwereusedfor the transfection_assay.

-140 -94 GCATAAAGCTGGCTTGAGGGAAGATTTTGCGCAAGACATTTCCCAAG GIO TAC G8 CTA G7 GTT G6 AGG G5 TCT Gl TTC G2 CAC G3 GGG G4 GTA Dl CC D2 CAC D3 GG D4 AA D5 CC 100 G10 G8 G7 G6 G5 G1 G2 G3 G4 D1 D2 D3 04 D5 dB Mutants

FIG. 3.

_{7Vww-activationofAGPgeneby}

YYIismediated

by

aB motif. Cotransfection _assaysofCMV-YY1

expres-sionvectorandmutated_upstream

_regulatory

sequence

(span-ning

from -140to

-94)

of the AGP_gene linkedto CAT

reportergene.Themutated sequencesareshown inthe_upper

panel.

dB is thedeletedmutantofthe B motif linkedtoCAT,

as

reported

inourearlier

publication

(Lee

etal.,

1993).

The

amountsof

_plasmid

and thecells usedforthe transfectionare

thesameasin

Fig.

1.

(-

and+)Transfectioninthe absence

andinthepresenceofCMV-YY1.

mutated AGP/CAT_exceptthe B-truncated mutant

_(Fig.

3,

lower

_panel,

_dB).

These data indicatethat YYI is

_unlikely

dependent

ona

single

motif

spanning

from -140to -94of AGP_promoterforthe activation.

However,

whenthe B-mo-tif-deletedmutant wasco-transfected withCMV-YY 1, the

activation was abolished

(Fig.

3,

dB).

As detailed in our

earlier

_publication

(Lee

et al.,

1993),

the B motif is a

negative

element that is

_{recognized by}

factorB. The CAT

activity

ofB-motif-deletedmutantwas

substantially

elevated as

compared

with the

wild-type

orother mutated

AGP/CAT,

consistent with the

_negative

effect ofthe B motif.

_Having

observed that YYI could activate the AGP_promoter but failedtodosowhentheBelementwasdeleted, weinitiated

studiesonthe

possible recognition

ofthe B motif

_by

YYI.

Extensive

_{experiments using}

a number of

oligonucleotide

probes

andrecombinant YYI orYYI

purified

from

HepG2

failedto show that YYI can bind to the B motif

(data

not

shown).

These data_suggestthat the observedactivationof YYI on the AGP _promoter is

unlikely

to be mediated

_by

direct interaction of YYI and the B

element; rather,

a

functional interaction_maybe the

_{likely explanation.}

100

Wt -71 OligoB pCAT-P dB

Reporter

FIG. 4. Transfection _{assays demonstrate that} the activa-tion of AGP

_promoter

function

_by

YY1 is

_through

the B motif. CAT _reporter constructs

_containing

the

_wild-type

AGP

_promoter,

deletion mutants of B motif

_(dB),

and

upstreamof-71

_(-71),

_oligomeric

B motic

_(BIO)

linkedto

SV40

_promoter-CAT

_(oligo

B),

and SV40

_promoter-CAT

(pCAT-P)

wereused fortransfection

(-)

orcotransfection

withCMV-YY1

(+)

into BHK cells.

_{Reporter plasmid}

(10

p-g)

and3 p.gof CMV-YY1 wereused.

YYI can activate theB

motif

linked to

heterologous

promoter

The above

_experiments

showedthatthe activationofYY 1

onthe AGP_promoterismediated

by

theBelement,butnot

by

direct

_protein-DNA

interactions. To address further the issueofhowtheB motif

_might

mediatetheactivationof the AGP _promoter

_by

YYI,

we constructed a

heterologous

promoter

(SV40)

containing

oligomeric

B_sequencesaswell

as -71/AGP

_(upstream

of -71 of AGP

_promoter

was

deleted)

promoterlinkedtoCAT

(Lee

etal.,

1993).

Wild-type and -71/AGP _promoter CAT could be activated

_by

CMV-YY1

_(Fig.

4).

YYI could also activate

_oligomeric

B-linked

_heterologous

_promoters. In

_striking

contrast to

these

data,

YYI had no effect on _promoters that do not

contain the Belement. Consistent with theconclusionthat B isa

negative

element from

previous

studies,

heterologous

promoters

containing oligomeric

B

_(B10)

element showed reduced

_activity.

To rule out the

_possibility

that there is a

difference in transfection

_efficiency

between

_experiments,

we used

B,0-linked-pSV-ß-Gal

for the transfection and co-transfection_assays.The

_B10-linked

_pSV-ß-Gal

wasmuch

weaker thanthe

_pSV-ß-Gal

when

_staining

withX-Gal

_(data

not

_shown).

However,

whenYYI wasco-transfected with

B,0-pSV-ß-Gal,

the

_{ß-Gal activity}

was elevated

(data

not

shown).

Thisincrease in

_activity

isnotduetothe increase in transfection

_efficiency

when YYIwas_present,asevident

by

the _apparentnumberofcells

_positive

inX-Gal

_staining

are

relatively

constant

_(data

not

_shown).

These data_suggestthat YY 1 isinvolvedin

_modulating

thefactorB

_activity

either

_by

direct

_{protein-protein}

orfunctional interactionwith Bfactor or

by repressing

the

expression

oftheBfactor_gene.

To rule out further_any

_possible

artefacts

_resulting

from these cotransfection

_experiments,

we constructed a YYI

expression

vector

(CMV-C250)

in which

_{only protein}

con-taining

the

_{carboxy-terminal}

250 amino acids could be

Zn _Finger

I i CMV-C250

1 25 2.S 1 23 1.1 1 29 2.1 1 25 1.2

FIG. 5. The /ra/js-activation of the AGP gene

by

YYI

requires

the

_full-length

YYI molecule. The _upper

_panel

shows the construction of YYI and the truncated YYI

expression

vectors,CMV-YY1 andCMV-C250. Thelower

panel

shows the trans- activation of AGP/CAT

_by

CMV-YY1 but not

_by

CMV-C250. Cotransfection

_experiments

weredetailed in the

legends

of

Figs.

1and 2.The numberson

topof the lower

_panel

indicate the fold stimulationrelativeto

the basal CAT

_activity

of each CAT _reporter construct

(normalized

to

1).

(—)

Transfection with the CAT_reporter

plasmid

alone;

(YY)

cotransfection with

CMV-YY1; (C)

cotransfection with CMV-C250.

_Wild-type

(wt)

or mutant

AGP-CAT

(G1,

G2,

and

G3)

areshownonthebottomline of

the lower

_panel.

demonstrated thatwhentheanimo-terminal

_portion

ofYYI

was

deleted,

it failed to activate the AGP gene

(Fig.

5).

Taken

_together,

the activation of theAGP_gene

_by

YYI not

only requires

the B motif but alsoneeds theamino-terminal

portion

ofYYI molecule.

YYI doesnot

_interfere

with the

_binding

_of

B

_factor

toits

_cognate

_motif

Having

demonstrated thatYY 1 canactivatetheAGP_gene

through

the B

motif,

we would liketo knowwhether YYI interactedwith the Bfactor

_physically.

Purified YYI

(from

liver)

or recombinant YYI was incubated with rat liver-derived B factor. Purified YYI didnotinterfere with theB factor

_binding

to itsmotifin

_{footprinting analysis}

(data

not shown) or in

_{gel mobility-shift}

_{assay in the presence} or

absence of

_{specific antibody}

to YYI

(data

not

_shown).

Purified AGP/EBP or

protein

derived from other column

fractionsbehaved

_similarly.

Furthermore,

recombinantYYI did not inhibitthe

_{footprinting activity}

of B factor. These data indicated that YYI did not form a

protein-protein

complex

with factor B; rather, it interacted with B factor

functionally

without

_interfering

with the

_DNA-binding

activ-ity

ofBfactor.

The level

_of

YYI does not

_change

_during

the

acute-phase

response

We have shown

_previously

that the levelsofAGP/EBP (themost

_{prominant positive}

factor)

and Bfactor

(a

negative

factor)

are induced and reduced

during

the

acute-phase

reaction

_(Change

etal., 1990;Leeet

al.,

1993).

_{This may} account in _part for the net induction of the

acute-phase-response genes. Becausewehavenowshown that YYI can

activatethe

_{transcription}

of_{AGP gene, it would be}

interest-ing

toexamine the

_change

inthe

_protein

levelaswellasthe

binding activity

of YYI

_during

the

_acute-phase

_response. Liver nuclear extracts were

prepared

from untreated or

LPS-treated rats and usedfor

_gel

retardation_assays

_(using

UCR

_{oligonucleotide,}

which also contains an

overlapping

AGP/EBP-binding

site)

and Western blot

_analysis.

Both the

binding

activity

and the

_protein

level of YYI remains

constant whether _pre- or

post-acute-phase

_response

(Fig.

6).

AGP/EBPwasincreased

during

the

acute-phase

_response,

whichwas anice internalcontrol. Taken

together,

these data

suggest that the

_protein

level of YYI does not

_change;

however, itsfunctional

_activity

for interactionwithBfactor maybe elevated

(e.g.,

through post-translational

modifica-tion)

during

acuteinflammation.

DISCUSSION

Mechanism

_of

activation

_of

AGPgene

by

YYI In this _paper we have demonstrated that YYI activates

transcription

directed

_by

the AGP_promoteror

heterologous

(SV40)

promoter

containing

the

_negative

element B. YYI represses

activity

ofthe SV40

enhancer/promoter containing

the

_YYI-binding

motif

(Shi

etal.,

1991).

Thefundamental difference between theYYI motif-mediated and the B-mo-tif-mediated

_activity

of YYI is the involvement of direct Y Y1 -DNA interaction in the formercaseand

only

functional

interaction of YYI and B-motifin the lattercase. Two lines

of evidence _suggest that the B motif of AGP _promoter mediates YYl's activation ofAGP_gene:

_(i)

the activation of AGP _promoter

_by

YYI

_depends

on the presence of B

sequence;

(ii)

YYI canactivateB-motif-mediated

repression

of SV40 _promoter when the B motif is inserted into this promoter. An extensive search failed to _{reveal any DNA}

elementinthe AGP_promoter

_region

thatcanbe

recognized

by

Y Y1. Howcanthe_{B sequence mediate the action of YY 1}

without direct

_DNA-protein

interaction? Because the B elementwasshowntobe

_recognized

_by

a

negative

transcrip-tion factorB

(Lee

etal.,

_1993),

four

_possible

mechanisms

may exist:

(i)

direct YYI and factor B interaction at the

protein

level resulted inthe reduced_repressor

_activity

of B

factor;

(ii)

YY 1 could repress the

expression

of theB

factor,

resulting

inareduced

protein

level of factor

B; (iii)

there is

only

a functional interaction between YYI and factor B without direct

_{protein-protein}

interaction,

e.g.,

modulating

offactor B's

_{activity by}

_{YYI; (iv)}

YYI interacts with factor B

_functionally

viaan

intermediary.

For the first

possibility,

it

could be a

physical

interaction between YYI and factor B.

For the second

_possibility,

_{YYI may} bind to the

_cognate

motif in the

_{regulatory region}

ofthe factor_{B gene and result} in

_repression

of

_expression

of factorB. For the third

possi-bility,

YY 1 mayaffect factorB's

_{activity by}

modification of factorB

_{post-translationally.}

Forthe fourth

_possibility,

there mayexista_corepressorfor theinteraction of YYI and

fac-torB.

Twolines of evidence indicatethat YYI interacts with B factor

_{functionally,}

butnot

_physically

_(i.e.,

_complex

forma-B LPS N

YY-1_AGP/EBP

L N _LNLNLN

m

70--35

1234561

2345

6

FIG. 6. Determination of the

_protein

level and the

_{DNA-binding activity}

of YYI in thenuclearextracts

_prepared

from normal andLPS-treated liver.A. Westernblot

_{analysis showing}

therelativeconstant

_protein

level of YY l in normalratliver nuclearextract

(N)

andextract

_prepared

from LPS-treated liver

(LPS).

Rabbit

_antibody

toYYI wasusedfortheblot

analysis.

The AGP/EBPlevel

(indicated

by

a35andarrow)wasusedas a

positive

controltoshow its increase

_during

the

_acute-phase

response.B.Gel

_{mobility-shift}

_assay

_showing

theconstantlevel in

_binding

_activity

of YYI in normal livernuclearextractand

extractfrom liver of the

_acute-phase

_response

(lanes

1-6 contain

10, 8, 7, 6,4,

and2 p.g of nuclearextract,

respectively).

The

arrowindicates the

YYl-DNA

_complex;

thearrowheadindicatesthe AGP/EBP-orAGP/EBP-related factors.

tion): (i)

both

_purified

andrecombinantYYI cannot inhibit the

_{footprinting activity}

of factor B, and

_antibody

to YYI failed to

_pattern

of

_complex

formation inabandshift_assay;

(ii)

although

YYI cannotbindtothe B

motif,

itcanactivate

genesthat containaBmotif.The

following examples

show

that YYI _{could affect gene}

_{expression through}

a third

proteins:

(i)

YYI

_repression

of the P5_promoterisrelieved

_by

adenovirus E1A

_protein

_(Shi

et al.,

1991). (ii)

A fusion

protein

of YYI and the GAL4 DNA

_binding

domain

re-presses

transcription

of a

thymidine

kinase _promoter with

GAL4

_binding

sites

(Seto

etal.,

1991).

(iii)

c-Myc

canform

complex

withYYI and thus modulates the

_activity

ofYYI

(Shrivastava

et al.,

1993).

A

_simple

model to

explain

the

functional

_versatility

of YYI is that its interaction with different cellular

_proteins

_{may alter its}

_{activity. During}

the

acute-phase

response, the

binding activity

of factor B

de-creases while both the

protein

level and the

DNA-binding

activity

of YYI do not

_change.

The

_activity

of YYI canbe

modulated

_during

theacuteinflammation_{in such way}sothat

its functional interactionwith factor B isenhanced,whilethe

DNA-binding activity

of factorB is unaffected. However, this modification of YYl's

_activity,

which

_ultimately

affects the

_activity

of factorB,maynotinvolvein the formationof YY1-B

_{complex. Alternatively,}

factorBis modified

_during

the

_acute-phase

_{response and}results inthe enhanced

func-tional interaction with YYI.

Furthermore,

both of these

possibilities

exist. Future

_experiments

shouldbe

_designed

to

resolve this issue.

A number of_systemsin which

_{protein-protein}

interactions resulted inaltered functions ofa

specific

transcription

factor

have been demonstrated

_(Lee

et

al., 1993;

Nishio et al., 1993;Steinetal.,

1993).

The functions ofYYIaremediated

mainly by

its

_binding

toa

specific

DNA motif

(Hariharan

et

al.,

1991;GualbertOÉ>rer/.,

1992;Montalvoe<a/.,

1991;

Park and

Atchison,

1991;

Seto et al.,

1991;

Shi et

ai,

1991;

Bauknechtetal., 1992;

Flanagan

etal.,

1992;

Lee et

_al.,

1992). Therefore,

YYI _mayalso modulatethe

_expression

of factor B

_{transcriptionally through}

its

_binding

to the

regula-tory

regions

of factorB gene. Definitiveanswers mustwait until the factor B_geneisisolated and cloned. Taken

_together,

our _present data and the

rapid

reaction

during

the acute

inflammation_suggestthatthe functional interaction ofYYI andfactorB,rather than the

_{transcriptional}

mechanismis the

more

likely

mechanismin

modulating

thefactorB

activity.

However,

despite vigorous

tests

_(performed

_by

anti-YYl

antibody depletion

followed

_by

_footprinting

_assayortheuse

of anti-YYl

_antibody

in

_{gel mobility-shift}

_assay),

we have

failed to demonstrate that the YYI and factor B

_complex

exists.

TRANSCRIPTION

OF AGP

GENE

Cytokines

(e.g.,LIF, IL-6,

IL-1)

YY"1 _{Factor B}

Functional Interaction

FIG. 7. Schematic

_{representation}

ofthe

_regulation

ofAGP

gene

by

various

transcription

factors.

Cytokines

stimulate AGP/EBP- orAGP/EBP-related /ra/w-activators andresult

in the induction ofAGP

_during

the

_acute-phase

response. YYI

_participates

in the AGP

_{regulation by repressing}

the functional

_activity

of a

_negative

factor,

B.

Thus,

YYI activatesAGP_gene

_{indirectly by}

functional interaction with factorB.

is its induction

_during

the

_acute-phase

_{response. We have}

previously

shown the level of AGP/EBP increased several fold

_during

the

_acute-phase

response.Wealso documenteda

decreaseinthe

_{binding activity}

of factorB. In this_report,we

presented

evidence forthe activation of AGPgene

by

YYl,

including

that

(i)

this activation is mediated

_by

B

motif;

(ii)

the activation of AGP _gene

_by

YYI

_requires

the amino-terminal

_portion

ofYYI molecule;and

(iii)

YYl'sfunction in

_activating

AGP_{gene is}

_independent

ofAGP/EBP.

Boththe

_protein

level and the

_{binding activity}

ofYYI do

not

_change

inratliver

_during

the

_acute-phase

_response

_(Fig.

6).

Becausethemechanism of reduction of factorB'

s

binding

activity

during

the

_acute-phase

_response remains to be

solved,

it is

_{highly possible}

that some

post-translational

mechanisms are involved in

regulating

factor B's

_activity,

suchas

interacting

withYY 1.Ifthisindeed

happens,

then the

functional

_activity

ofYYI undernormal

_{physiological}

con-ditions must be different from that

_during

the

_acute-phase

response. The

regulatory

roleofYYI onfactorB's levelor

activity

offersanattractive model for_gene

regulation:

Reg-ulation of a

negative

factor

by

another

negative/positive

(Yin-Yang-1)

factor. A schematic mode

_summarizing

the

regulation

ofAGPgene is

presented

in

Fig.

7.

Additive

_{effect of}

activation

_by

YYI

and

AGP/EBP

on the AGP

promoter

We have shown

_previously

that AGP/EBP couldactivate AGP

_{transcription}

and that AGP/EBP is the

_key

nuclear factor in

_conveying

the

_signals

[i.e.,

cytokines

like interleu-kin-1

_(IL-1),

IL-6,

or leukemia

inhibitory

factor]

to their targetgenes

(e.g.,

AGP)

during

the

_acute-phase

_response.To

determine whether the activation effects of AGP/EBPand YYI onAGPare

independent

_events,we

performed

trans-fection

_{experiments using}

_wild-type

AGP/CAT,

CMV-AGP/ EBP, and CMV-YY1. As shown in

_Fig.

2,

AGP/EBPand YY1could activate theAGP_promoter,and theeffect of these

two

_expression

vectors seems to be additive. These data

suggest that AGP/EBP and YYI donot interact with one

another to activate the AGP _{gene, but rather}act

indepen-dently

in

_activiating

the AGP_gene.These resultsare

consis-tentwith thedata that the

_activity

ofY Y1

_depends

onthe B

motifwhereas AGP/EBP

_depends

on other motifs. Taken

together,

YYI,

AGP/EBP,

C/EBP-a, and factor B are

important trans-acting

factors in

_regulating

AGPgene

tran-scription.

The data _present in this _paper are

_particularly

important

in the

_potential

modulation of YYl's

_activity

during

theacuteinflammation and its

_subsequent

interaction withfactorB. Further

_{understanding along}

these lines _may

clarify

not

_only

factor B's

_regulation

but alsoYY 1'$mode of action and

_regulation.

ACKNOWLEDGMENTS

WethankDr.

_Tung-Tien

Sunforacriticalreview of the

manuscript,

Mr.

_Tso-pang

Yao for

_performing

thePCR of YYI cDNA,Dr.

_Wuh-Liang

Hwufortechnical

_help,

Wen-Hai Tsai fornuclearextracts, and Drs.T. Shenkfor CMV-YY1andK. Ozato forthe

_expression

vectorofUCRBP. This research was

supported by

_grants NSC82-0412-B002-251

fromtheNationalScience Council.

REFERENCES

BAUKNECHT, T., ANGEL, P., ROYER, HD., and ZUR

HAUSEN, H. (1992). Identification ofa

negative

regulatory

domain in the human

papillomavirus

type 18_promoter: interac-tion with the

_{transcriptional}

_{repressor YYI.} EMBO J. 11,

4607-4617.

BAUMANN, H., and _HELD, W.A. _(1981).

_Biosynthesis

_and

hormone-regulated expression

ofsecretory

glycoproteins

in rat

liver and

hepatoma

cells.J. Biol. Chem.256, 10145-10155.

BAUMANN, H., and

_MAQUAT,

L.E. (1986). Localization of DNA_{sequences involved}in

dexamethasone-dependent

expres-sionoftherat

_alpha-1

_acid

_glycoprotein

_{gene. Mol. Cell. Biol.6,}

2551-2561.

BAUMANN, H.,FIRESTONE,G.L.,BURGESS, TL., GROSS, K.W., YAMAMOTO,K.R., and_HELD, W.A. (1983). Dexa-methasone

_regulation

of

alpha-1

acid

glycoprotein

and other

acute

_phase

reactants _in rat_{liver and}

_hepatoma

cells. J. Biol. Chem.258,563-570.

BAUMANN, H., HILL, R.E.,SAUDER, D.N., andJAHREIS,

G.P.(1986).

Regulation

ofmajoracute

_{phase plasma proteins by}

hepatocyte stimulating

factors of human squomas carcinoma cells. J. Cell Biol.102,370-383.

BENNETT, H., andSCHMID,K. _(1980).

_{Immunosuppression}

of

human

plasma alpha-1

acid

_{glycoprotein: Importance}

of the

carbohydrate

moiety.

Proc. Nati. Acad. Sei. USA 77,

6109-6113.

CHANG, C.J., CHEN, T.T., LEI, H.Y., CHEN, D.S., andLEE,

S.C. (1990). Molecular

_cloning

ofa

transcription

factor, AGP/

EBP,that

_belongs

tomembersof the C/EBP

_family.

Mol. Cell. Biol. 10,6642-6653.

DARLINGTON, G.D., WILSON. R.. and LACHMAN, L.B.

(1986).

Monocyte-conditioned

medium, interleukin-1, and

tu-mornecrosis factor stimulate theacute

_phase

_{response in human}

hepatoma

cells in vitro. J. Cell Biol. 103,787-793.

FLANAGAN,J.R.,BECKER, K.G., ENNIST,D.L.,GLEASON, S.L., DRIGGERS, P. H., LEVI, BZ., APPELLA, E., and

OZATO, K. (1992).

Cloning

ofa

negative transcription

factor

that bindstothe_upstreamconserved

region

of

_Moloney

murine leukemia virus. Mol. Cell. Biol. 12,38-44.

FRASCH,M.,andLEVINE, M.(1987).

Complementary

patterns

of

even-skipped

and fushitarazu

_expression

involve their differ-ential

regulation

by a common set of

segmentation

genes in

Drosophila.

Genes & Dev. 1,981-995.

FRIEDMANN, M.L. (1983). Control of malaria virulence by

alpha-1

acid

_glycoprotein

(orosomucoid),anacute

_phase

(inflam-matory)reactant.Proc. Nati. Acad. Sei. USA80,5421-5424.

GAULDI,J.,RICHARDS, C,HARNISH. D.,LANDSORP,P.,

and BAUMANN, H. (1987). Interferon beta-2 shares

identity

with monocytederived

hepatocyte-stimulating

factor and regu-lates the

_major

acute

_{phase protein}

_{response in liver cells. Proc.}

Nati. Acad. Sei. USA84,7251-7255.

GORSKI, K., CARNEIRO, M., and SCHIBLER. U. (1986).

Tissue-specific

in vitro

transcription

from the mouse albumin

promoter.Cell47,767-776.

GUALBERTO, A., LePAGE, D., PONS, G., MADER, S.L., PARK, K., ATCHISON, M.L.,andWALSH,K.(1992). Func-tional

antagonism

between YY 1 and theserumresponse factor. Mol. Cell. Biol. 12,4209-4214.

HARIHARAN, N., KELLEY, D.E., and PERRY, R.P. (1991). Delta,a

transcription

factor that bindstodownstream elements in several

_polymerase

II_promoters, isafunctionally versatile zinc

finger protein.

Proc. Nati. Acad. Sei. USA88,9799-9803.

KLEIN, E.S., DlLORENZO, D., POSSECKERT. G.. BEATO, M.,andRINGOLD,G. M.(1988).

Sequences

downstream of the

glucocorticoid

regulatoryelement mediate

cycloheximide

inhibi-tion of steroid-induced

expression

from the rat

_alpha-1

acid

glycoprotein

promoter:evidence foralabile

transcription

factor.

MoL Endocrinol.2, 1343-1351.

KULKARNI, A.B., REINKE, R., andFEIGELSON, P. (1985).

Acute

_phase

mediators and

glucocorticoids

elevate

_alpha-1

acid

glycoprotein

gene

transcription.

J. Biol. Chem. 260, 15386-15389.

LEE, J.S., GALVIN, K.M.. and SHI, Y. (1993). Evidence for

physical

interaction between the

zinc-finger transcription

factor YYI

_andSpl.

Proc. Nati. Acad. Sei. USA90,6145-6149.

LEE, T.C., SHI, Y.,andSCHWARTZ,R.J. ₍1992).

Displacement

of BrdUrd-induced YYI by serum response factor activates skeletal

_{alpha-actin transcription}

in_{embryonic myoblasts.}Proc. Nati. Acad. Sei. USA89,9814-9818.

LEE, Y.M., TSAI, W.H., LAI, M.Y.,CHEN, D.S., andLEE,

S.C. (1993). Induction of liver

_alpha-1

acid

glycoprotein

gene

expression

involves both

positive

and

_{negative transcription}

factors. Mol. Cell. Biol.13,432-442.

LICHT,J.D.,GROSSEL,M.J..FIGGE,J.,andHANSEN,U.M.

(1990).

_{Drosopkila Krüppel protein}

isa_{transcriptionalrepressor.}

Nature346,76-79.

LIU,H.TAKAGAKI, M.K., andSCHMID, K. (1988). In vitro

nerve-growth-promoting activity

of human

plasma alpha-1

acid

glycoprotein.

J. Neurosci. Res. _20,64-72.

MONTALVO, E.A., SHI, Y., SHENK, T., andLEVINE, A.J.

(1991).

Negative regulation

of theBZLF1 _promoterof

Epstein-Barr virus. J. Virol.65,3647-3655.

NISHIO, Y„ ISSHIKI, H., KISHIMOTO, T., and AKIRA. S.

(1993). A nuclear factor for interleukin-6

_expression

(NF-IL6)

and the

glucocorticoid

receptor

synergistically

activate

transcrip-tion of therata-1 acid

_glycoprotein

gene via direct

protein-protein

interaction. Mol. Cell. Biol. 13, 1854-1862.

PARK, K.,andATCHISON,M.L.(1991).Isolation ofacandidate

repressor/activator,

NF-E1(YY-1, delta),that bindstothe

immu-noglobulin

k 3' enhancer and the

_{immunoglobulin}

_heavy-chain uEl site. Proc. Nati. Acad. Sei. USA88,9804-9808.

PROWSE,K.R.,and BAUMANN,H.(1988).

Hepatocyte-stimu-lating

factor, beta-2-interferon, and interleukin-1 enhance

ex-pression

of the rat

_alpha-1

acid

_glycoprotein

_{gene via} adistal

upstream

regulatory region.

Moll. Cell. Biol.8,42-51.

RUPPERT, J.M., KINZLER, K.W., WONG, A.J., BIGNER, S.H., KAO, FT.,LAW, F.L., SEUANEZ, H.N., O'BRIAN, S.J.,and_VOGELSTEIN,B.(1988).

TheGLl-Kriippel family

of human genes. Mol. Cell. Biol.8,3104-3113.

SETO, E., LEWIS, B., and SHENK, T. (1993). Interaction between

_{transcription}

factors

Spl

and YYI. Nature _365, 462-464.

SETO, E., SHI, Y., and SHENK,T. (1991). YYI is an_initiator

sequence-binding protein

that directs and activates

transcription

in vitro. Nature_354, 241-245.

SHI, Y., SETO, E., CHANG, L.S., and SHENK, T. (1991).

Transcriptional repression by

YY-1, a human

GLl-Kriippel-related

_protein,

and relief of

_{repression by}

adenovirus El A

protein.

Cell67,377-388.

SHRIVASTAVA, A., SALEQUE, S., KALPANA, G.V., ARTANDI.S., GOFF, S.P.,andCÁLAME,K.(1993). Inhibi-tion of

_{transcriptional}

_regulator

_{Yin-Yang-1 by}

association with

c-Myc.

Science262, 1889-1892.

STEIN, B.. COGSWELL, P.C., and BALDWIN, A.S. (1993).

Functional and

physical

associations between NF-kB and C/EBP

family

members: A Rel domain-bZIP interaction. Mol. Cell. Biol. _{13,3964-3974.}

Address

reprint

requests to: Dr.

_Sheng-Chung

Lee Institute

of

Molecular Medicine

College of

Medicine

NationalTaiwan

University

Taipei,

Taiwan Received for_publicationMarch22, 1994,and in revised form_May24. 1994.