repression of a herpes simplex virus immediate-early promoter by

MOLECULAR AND CELLULAR BIOLOGY, Nov. 1994, p. 7633-76420270-7306/94/$04.00+0Copyright X 1994, American Society for Microbiology

Repression of a Herpes Simplex Virus Immediate-Early Promoter bythe Oct-2 Transcription Factor Is Dependent on an Inhibitory

Region at the N Terminus of the ProteinKAREN A. LILLYCROP,1 SALLY J. DAWSON,1 JOHN K. ESTRIDGE,1 THOMAS GERSTER,2

PATRICK MATIHIAS,3 AND DAVID S. LATCHMANl*

Department of Molecular Pathology, University College London Medical School, London WJP 6DB,United Kingdom,1 and Biozentrum der Universitat Basel, CH-4056 Basel,2 and

Friedrich Miescher-Institut, CH-4002 Basel,3 Switzerland

Received 26 January 1994/Returned for modification 12 April 1994/Accepted 5 August 1994

The B-cell form of the Oct-2 transcription factor Oct 2.1 can activate the herpes simplex virus immediate-early gene 3 (IE3) promoter, whereas the neuronally expressed Oct 2.4 and 2.5 forms of the protein, whichcontain a different C terminus, can repress this promoter. Here we show that partial or full deletion of the Cterminus of Oct 2.1 in the presence of an intact N terminus results in a protein which can strongly repress theIE3 promoter. In contrast, deletion of the entire N terminus or a short region within it leaving the C terminusintact results in a very strong activator. Deletion ofboth N and C termini leaving only the isolated POU domaingenerates only a very weak repressor. The N-terminal region defined in this way can repress a heterologouspromoter when linked to the DNA-binding domain of the GAL4 factor, indicating that it can function as an

independent inhibitory domain. These results indicate that a specific region within the N terminus common toOct 2.1, 2.4, and 2.5 plays a critical role in the ability of neuronally expressed forms of Oct-2 to repress the IE3promoter but can do so only when the C-terminal region of Oct 2.1 is altered or deleted.

The Oct-2 transcription factor was originally identified as an

activator of transcription in B lymphocytes (29, 30). Thus, thisprotein can activate promoters such as those of the immuno-globulin genes which contain its DNA-binding site (ATGCAAAT) both in B lymphocytes (25) and following theartificial expression of Oct-2 in cells such as fibroblasts whichdo not normally express this protein (20).

Interestingly however, Oct-2-specific mRNAs have beendetected in the nervous system (10, 11, 32), and an octamer-binding protein with the mobility of Oct-2 has been identifiedin the brain with the use of a DNA mobility shift assay (26).However, comparison of the activities of B-cell and neuronalOct-2 has indicated that in contrast to B-cell Oct-2, neuronalOct-2 can function as a repressor of transcription (5, 32). Thus,neuronal Oct-2 was responsible for the repression of theherpes simplex virus (HSV) immediate-early (IE) genes inneuronal cells (15), which is dependent on the octamer-relatedTAATGARAT motif in the IE promoters (12) and results inthe clinically important phenomenon of latent infection (13).Hence, the Oct-2 protein can function as a repressor oractivator of transcription, depending on the cell type studied.Although Oct-2 is encoded by a single gene in the mamma-

lian genome (4, 25, 29), the protein exists in a number ofdifferent isoforms which are produced by alternative splicing ofthe Oct-2 RNA (35). Moreover, the pattern of splicing inneuronal cells differs from that in B cells; hence, differentforms of the protein predominate in the two cell types (17, 32).Although all of the different forms of the protein can transac-tivate a simple octamer-containing promoter when cotrans-fected into fibroblasts (35), in agreement with earlier results(20), they differ in their effects on the HSV IE promoters insimilar cotransfections (17). Thus, while the predominant

* Corresponding author. Mailing address: Department of MolecularPathology, The Windeyer Building, Cleveland St., London WlP 6DB,United Kingdom. Phone: 44-71-380-9343. Fax: 44-71-387-3310.

B-cell form Oct 2.1 can activate the HSV IE3 promoter whencotransfected into fibroblasts, the predominant neuronal formsOct 2.4 and 2.5 (also called Oct-2B [21]) can repress it (17). Toidentify the reasons for the different activities of the variousforms of Oct-2, we have attempted to characterize the featuresof Oct-2 which are necessary for the repression of IE promoteractivity.

MATERUILS AND METHODS

Plasmid DNAs. The target octamer-containing promotersused in this study were the HSV IE3 gene promoter (from-330 to +33) linked to the chloramphenicol acetyltransferase(CAT) gene (IE-CAT [31]), a simple promoter containing a

consensus octamer motif linked to the TATA box and tran-scriptional start site of the c-fos gene and the CAT gene andlacking any other transcription factor binding sites (Oct-CAT[36]), and the tyrosine hydroxylase promoter from -272 to+27 driving expression of the CAT gene (14). These targetplasmids were transfected with either the cDNA clone of oneof the different alternatively spliced forms of Oct-2 under thecontrol of the constitutive IE promoter of cytomegalovirus(CMV) (35) or a series of Oct-2 mutants lacking either theentire N- or C-terminal region (21) or containing different N-or C-terminal deletions (6, 21). The POU domain plasmid wasprepared by PCR amplification of an Oct-2 cDNA clone withappropriate oligonucleotides and, following addition of anATG translation start codon, was cloned into the pJ7 expres-sion vector.DNA transfection. BHK-21 cells (clone 13 [19]), which lack

endogenous Oct-2 (15, 17), were transfected by the method ofGorman et al. (8). Transfections were carried out with 2 x 106cells on a 90-mm-diameter plate transfected with 10 ,ug of theOct-2 expression vectors.CAT assays. Assays of CAT activity in the transfected cells

were carried out as described by Gorman et al. (8), extracts

7633

Vol. 14, No. 11

on April 15, 2019 by guest

http://mcb.asm

.org/D

ownloaded from

http://mcb.asm.org/

7634 LILLYCROP ET AL.

P L A - U.-_' Ptfl _Homeo _ Leu

16e | _

- k>Z.

.. .. :::0.-:f::.:.:If...

El-I

377 - 425

377 - 463

426 - 463

1 - 186

353 - 463

POU

I--t

M- -w I* * I- I -

I.- I

* *EIm m I l

I - 55

1 - 98

1 - 160

118-160 Z Il_ I

142 -160 I I I_

142 -181 1

FIG. 1. Schematic representation of the various deletions of Oct-2 used in this study compared with the naturally occurring forms of Oct-2 (Oct2.1, 2.2, 2.4, and 2.5). The region rich in proline, leucine, and glutamine residues (PLQ), the POU-specific (POU) and homeobox (Homeo) regionswhich together form the DNA-binding POU domain, and a potential leucine zipper region (Leu) are indicated. The position of a 16-amino-acidinsert (17) found in Oct 2.2 but not in Oct 2.1 (35) is indicated.

having been previously equalized for protein content as de-scribed by Bradford (3). The percentage conversion of chlor-amphenicol to the acetylated form was determined by scintil-lation counting of the regions of the thin-layer chromatographyplate corresponding to the acetylated and nonacetylated formsas detected by autoradiography. In all cases, the values ob-tained were equalized for differences in plasmid uptake be-tween samples based on the results of dot blot hybridization ofan aliquot of the transfected cell extract with a CAT DNAprobe (12).

RESULTSIn our previous experiments (17), Oct 2.1, 2.2, and 2.3 (also

collectively called Oct-2A [20]), which differ only at the Nterminus of the protein (35), were all able to activate the IE3gene promoter upon cotransfection into BHK cells, whereasOct 2.4 and 2.5 (Oct-2B), which differ from the other forms atthe C terminus but have the same N terminus as Oct 2.1, wereable to repress it. We therefore tested a series of C-terminal

deletions based upon Oct 2.1 (6) (Fig. 1) for the ability tomodulate the activity of the IE3 promoter contained in theplasmid IE-CAT (31). As a control, these plasmids were alsotested for the ability to activate a simple octamer-containingpromoter (Oct-CAT) which can be activated by all the Oct-2forms upon cotransfection of BHK cells (17, 36). In all cases,these mutant forms of Oct-2 had been shown to be expressedat similar levels in transfected HeLa cells (6) or BHK cells(data not shown), indicating that any differences in their effectson cotransfected promoters would be dependent on realdifferences in activity rather than on differences in the levels towhich they accumulate in the transfected cells.

In these experiments (Fig. 2a), any interference with theC-terminal region of Oct 2.1 eliminated its ability to activatethe IE3 promoter and resulted in a factor which was able torepress the IE3 promoter. Although all of the C-terminaldeletions were able to repress the IE promoter, the mostsevere repression was observed with the various mutants whichlack the region from amino acids 377 to 399 either alone or

Oct 2.1

Oct 2.21

Oct 2.4

Oct 2.5

377 - 3991

446 - 463 1

m 2 - -*--I v

I-F-T I

I 0

MOL. CELL. BIOL.

-m

1-,- IrT

J_- .-


http://mcb.asm

.org/D

ownloaded from

http://mcb.asm.org/

AN INHIBITORY DOMAIN IN Oct-2 7635

-

00

.t

c

.20

0

104_

vl l~~~~~~~~~~~~~~~~~~~~~~~~c

150-

140

120

100 -

600

50.

40-

20-

Oct-2 377-399 377-425 377-463 426-46 446-463

700&

600

500

400

300

200

100-

011~~~~~~~~~,Ifw1,J' Ad&1 < 1 1 2

f AA% AOcCt-2 377-3W 377-425 377 426463 64893

FIG. 2. Effect of intact Oct 2.1 or various C-terminal deletion mutants lacking the indicated amino acids on the activity of the IE-CAT (a) orOct-CAT (b) plasmid. Each value is expressed as a percentage of that obtained upon cotransfection of the promoter with the CMV expressionvector not containing any inserted sequence (100%). Values are averages of three replicate experiments whose ranges are shown by the bars.

with additional losses C terminal to this region. Interestingly,the region from amino acids 377 to 399 contains a putativeleucine zipper potentially capable of mediating the ho-modimerization of Oct 2.1 or its heterodimerization with otherproteins (4, 20, 25).

In contrast to the repressive effect of the C-terminal dele-tions on the IE3 promoter, all of the deletions retained someability to activate the Oct-CAT construct containing only anoctamer motif and TATA box (36), although this ability wasimpaired in all cases relative to the wild-type construct (Fig.2b). This ability parallels the activation of this promoterobserved with all of the different forms of Oct-2, including Oct2.4 and 2.5 (17). Interestingly the deletion mutant lacking onlyamino acids 377 to 399 was the mutant least impaired in

transactivation of the Oct-CAT reporter (Fig. 2b) even thoughit was a strong repressor of the IE promoter (Fig. 2a). Inagreement with our previous experiments using Oct 2.1, 2.4,and 2.5 (17), none of the Oct-2 deletion mutants had anypositive or negative effect on the activity of a control Roussarcoma virus promoter, indicating that these effects arespecific to promoters containing octamer or TAATGARATmotifs (data not shown).These experiments indicate that any interference with the

C-terminal region of Oct 2.1 eliminates its ability to activatethe IE promoter and converts it into a repressor. The simplestmechanism by which this effect could operate would be for anOct-2 molecule lacking the ability to activate the IE promoterto continue to bind to the TAATGARAT sequence and

VOL. 14, 1994


http://mcb.asm

.org/D

ownloaded from

http://mcb.asm.org/


1 zu- a

100-

80-

60-

20

POU Oct 2.4 Oct 2.5 377-463

POU Oct 2.4 Oct 2.5 377-463FIG. 3. Effect of the isolated POU domain (POU) of Oct-2 on the

activity of the IE-CAT (a) or Oct-CAT (b) plasmid compared with thatobserved with Oct 2.4 or 2.5 or a C-terminal deletion mutant. Eachvalue is expressed as a percentage of the promoter activity observedwith the CMV expression vector alone (100%). Values are averages ofat least three replicate experiments whose ranges are shown by thebars.

passively prevent the activation of the IE promoter by prevent-ing the binding of other stimulatory octamer-binding proteinsto this sequence (7). To test whether this was the case, we

prepared an expression construct containing only the isolatedPOU domain of Oct-2 and tested whether this region whichcontains the DNA-binding domain was able to repress the IEpromoter. In these experiments (Fig. 3a), the isolated POUdomain was able to moderately reduce the activity of the IEpromoter when cotransfected into BHK cells. However, thedegree of reduction in activity observed was much smaller thanthat seen with either the naturally occurring C-terminal variantOct 2.4 or 2.5 or any of the artificial deletions of the C-terminalregion (Fig. 3a), although the levels of expression observedwith the isolated POU domain were higher than with the otherconstructs (data not shown). As expected, the isolated POUdomain was also able to moderately reduce the activity of theOct-CAT construct, presumably by preventing binding ofstimulatory octamer-binding proteins. In this case, however, allof the other forms of Oct-2 transactivated the promoter as

before (Fig. 3b).These findings suggest that in the case of the IE promoter,

a v c N

.*:

b

31

0

0

C-)

0 2C

0-0

(1-186) (353-463)

FIG. 4. (a) CAT assay showing the effects of deletion mutants ofOct-2 either containing the C terminus and the POU domain andlacking the N terminus (C) or containing the N terminus and POUdomain and lacking the C terminus (N) on the activity of the IE-CATplasmid compared with effects observed with wild-type Oct-2 (0) orthe expression vector alone (V). (b) Quantitative results of this assay

expressed as percentages of the promoter activity observed with theexpression vector alone (100%). Values are averages of three replicateexperiments whose ranges are shown by the bars. The amino acidsmissing in each construct are indicated in each case.

simple inhibition of DNA binding by a nonactivating form ofOct-2 does not account for the ability of the different forms ofOct-2 to repress the IE promoter. Moreover, the fact thatstrong repression can be achieved by a form lacking virtuallythe entire C terminus of the protein (amino acids 377 to 463)but not by the isolated POU domain suggests that the N-terminal region of the protein may be involved in this effect. Totest this possibility, we examined the effects on the IE pro-moter of cotransfecting mutant forms of Oct-2 either contain-ing the intact N terminus with the POU domain but lacking theentire C terminus or containing an intact C terminus linked tothe POU domain but lacking the entire N terminus (21).

In these experiments (Fig. 4), the C-terminal deletion func-tioned as a very strong repressor of the IE promoter compa-rable in strength to that observed with the other extensiveC-terminal deletions and with a much stronger effect than thatobserved with the isolated POU domain. This strong repres-sion is observed even though this mutant is expressed at a

similar level to wild-type Oct-2 in transfected cells (21). Incontrast, the N-terminal deletion functioned as a strong acti-vator of the IE promoter, and the degree of stimulationobserved was much stronger than that seen with the intactOct-2 protein even though the two proteins are known to beexpressed at similar levels in transfected cells (21). Hence,

2C00

0

0-

2c

C)0-

0O0-0ZD.2t5

-F

,OX

Oct-2 2 N terminal dieletion C terminal deletion

MOL. CELL. BIOL.


http://mcb.asm

.org/D

ownloaded from

http://mcb.asm.org/


N terminal deletion

(1-186)

C terminal deletion

(353-463)

b

100-

20 XN terminal deletion

(2-181)

FIG. 5. Effects of the N-terminal and C-terminal deletion mutants of Oct-2 lacking the indicated amino acids on the activity of the Oct-CATplasmid (a) or the tyrosine hydroxylase promoter driving the CAT gene (b). The results are expressed as percentages of the promoter activityobserved with the expression vector alone (100%). Values are averages of two determinations whose ranges are shown by the bars.

these results suggest that an inhibitory domain at the Nterminus of the protein is responsible for the ability of Oct 2.4,2.5, and the C-terminal deletion mutants to repress the IE3promoter. Moreover, the N-terminal domain also appears tolimit the stimulation of the IE promoter which is observed evenwith forms of Oct-2 such as Oct 2.1 which have an intact Cterminus and which are therefore capable of stimulating thepromoter.

Interestingly, deletion of the N-terminal domain also en-

hanced the ability of Oct-2 to stimulate the Oct-CAT constructwhereas deletion of the C-terminal domain reduced the acti-

vation (Fig. 5a). This finding suggests that the N-terminaldomain can also limit the extent of activation on otherpromoters which can be activated via the C-terminal domain.To further investigate this possibility, we tested the effect of

the N- and C-terminal deletions on the tyrosine hydroxylasepromoter, which we have previously shown to be repressed byall the forms of Oct-2 (4a). In these experiments (Fig. Sb), thispromoter was strongly repressed by the C-terminal deletionbut was unaffected by the N-terminal deletion. This findingindicates that the C-terminal domain does not activate thetyrosine hydroxylase promoter whereas the N-terminal domain

300 -

-6200°- _

c000-° _

0-0-

*> 100C.)

a

Oct-2.2

0

0c0

0ol

.5

Oct-2.2 C terminal deletion

(142-181)

VOL. 14, 1994


http://mcb.asm

.org/D

ownloaded from

http://mcb.asm.org/


can repress it, accounting for our initial observation that all theforms of Oct-2 can repress this promoter regardless of theirdifferent C-terminal regions. As in our previous experimentswith these naturally occurring forms of Oct-2, the inhibitoryeffect of the C-terminal deletion was dependent on the pres-ence of a binding site for octamer-binding proteins in thetyrosine hydroxylase promoter, indicating the specific nature ofthis effect (data not shown).

Hence, the N-terminal inhibitory domain is capable ofrepressing several different octamer-containing promoters,with the precise effect observed in each case being dependenton the extent to which they are stimulated by the C-terminalactivation domain. To map the N-terminal inhibitory domain,we used a series of N-terminal deletion mutants of Oct-2, all ofwhich have been shown to be expressed at similar levels intransfected cells (21) (Fig. 6). In these experiments (Fig. 6),progressive N-terminal truncation of the N-terminal region ofOct 2.2 leaving the C terminus intact led to only a smallincrease in the ability to activate the IE3 promoter when up tothe N-terminal 160 amino acids were deleted. However afurther truncation removing the N-terminal 186 amino acidsresulted in a large increase in transactivation, suggesting thatamino acids 161 to 186 are critical for the inhibitory effect thatwe observe on the IE promoter. In agreement with this idea, aninternal deletion of Oct 2.1 which lacked only amino acids 142to 186 (6) showed an enhanced ability to activate the IE3promoter, whereas deletion of amino acids 142 to 160 reducedthe transactivation observed (Fig. 6). Similar results were alsoobtained when these deletions were assayed on the tyrosinehydroxylase promoter with a construct lacking the N-terminal161 amino acids repressing the promoter, whereas one lackingthe N-terminal 186 amino acids did not have such an effect(data not shown).Having mapped the inhibitory domain, we wished to test

whether this region could function when it was linked to theDNA-binding domain of another factor. We therefore clonedeither the entire N terminus of Oct-2 or the region encodingamino acids 142 to 181 downstream of the DNA-bindingdomain of the yeast transcription factor GAL4 in the vectorpSG424 (24). These hybrid constructs were then cotransfectedwith a construct in which four DNA-binding sites for GAL4had been cloned upstream of the thymidine kinase (tk) pro-moter. This promoter was chosen because it has a relativelyhigh basal activity and had previously been used to map theinhibitory domains in other transcription factors (28). In theseexperiments, a clear reduction in promoter activity was ob-served upon cotransfection of the promoter construct with theconstructs containing either the entire N terminus or aminoacids 142 to 181 compared with the level observed uponcotransfection of the pSG424 vector alone. Hence, both theentire N-terminal region and the isolated region from 142 to181 can function to inhibit promoter activity in the absence ofthe POU domain when located C terminal to a heterologousDNA-binding domain (Fig. 7).To further investigate the activity of the isolated inhibitory

domain, we cloned this region in the absence of the DNA-binding domain or other regions of Oct-2 together with anadded ATG initiation codon downstream of the simian virus 40promoter, allowing its expression in isolation. We then inves-tigated the effect of this isolated region which cannot bind toDNA on the repression of the IE3 promoter by Oct 2.5. To dothis, constant amounts of IE-CAT and of the Oct 2.5 expres-sion plasmid were cotransfected into BHK cells in the presenceof an increasing amount of the inhibitory domain expressionplasmid and a correspondingly decreasing amount of theplasmid vector so as to maintain the amount of transfected

a

600

500

0

>' 400.0

C )

300

200

100

0>Oct 2 1-55 1-98 1-160 1-186 118-160142-160142-181

b 1 2 3

FIG. 6. (a) Effects of N-terminal deletion mutants of Oct-2 lackingthe indicated amino acids on the activity of the IE-CAT plasmidcompared with effects observed with the wild type. The results areexpressed as percentages of the promoter activity observed withexpression vector alone (100%). Values are averages of at least tworeplicate experiments whose ranges are shown by the bars. (b) DNAmobility shift assay using a radiolabelled octamer motif and proteinextracts prepared from untransfected BHK cells (track 2) or similarcells transfected with Oct-2 lacking either amino acids 1 to 98 (track 1)or ammno acids 1 to 186 (track 3). The position of the endogenousOct-i protein is indicated by the line labelled 1, and positions of theexogenous truncated Oct-2 proteins are indicated by the arrowheads.The band of intermediate mobility is produced by a non-sequence-specific DNA-binding protein.

MOL. CELL. BIOL.


http://mcb.asm

.org/D

ownloaded from

http://mcb.asm.org/


a b

100 1-

so -

H0N4-,

0u

60 -

0

4-)

-H

4-

40 -

20 -

0

100 -

H0u4-,

0U

0

T

4-,-H-

4-,Ui*

80 -

60 -

40 -

20 -

0

T

V GAL4 GAL4 GAL4

(2-181) (142-181)

V GAL4 GAL4 GAL4

(2-181) (142-181)FIG. 7. Effect of the entire N-terminal domain (amino acids 2 to 181) of Oct-2 or amino acids 142 to 181 linked to the GAL4 DNA-binding

domain on the activity of constructs containing four binding sites for GAL4 upstream of the tk promoter (a) or the tk promoter lacking GAL4binding sites (b). The effect of the isolated GAL4 DNA-binding domain on both promoters is also shown (GAL4). Results are expressed aspercentages of promoter activity observed upon cotransfection of the expression vector lacking any inserted sequence derived from GAL4 or Oct-2(V). Values are averages of two determinations whose ranges are shown by the bars.

DNA constant. In these experiments (Fig. 8), the inhibitoryeffect of Oct 2.5 could be abolished by the inclusion of highlevels of the inhibitory domain plasmid. Hence, this domainwhen present alone in trans in the absence of the DNA-bindingdomain can relieve the inhibitory effect of intact Oct-2 con-taining the DNA-binding domain as well as the inhibitoryregion. This observation indicates that the N-terminal regionmay function by recruiting a trans-acting inhibitory factor tothe promoter, with this factor being titrated out by the isolatedinhibitory domain expressed in isolation in the absence of theDNA-binding domain.

DISCUSSION

The simplest mechanism to account for the ability of differ-ent forms of Oct-2 to repress the HSV IE3 promoter would bethat these forms can bind to the octamer/TAATGARAT motifin the IE promoter and prevent the binding of activatingoctamer-binding proteins while failing to activate the IEpromoter themselves (see reference 7 for a review of mecha-nisms of gene repression by transcription factors). In agree-ment with this model, Stoykova et al. (32) have shown that anaturally occurring form of Oct-2 (Mini-Oct) consisting of

VOL. 14, 1994


http://mcb.asm

.org/D

ownloaded from

http://mcb.asm.org/


a 1 2 3 4

* 9*4Ii GAII...-

eel'0.b ll

1 0 0

90

80

0

H

0

u

70

60

4-40

do

0\0eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee

42.

1-

42)U

30

20

1 0 -

0 --.

V V V 2.5 2.5 2.5

P Rs Ras P Rs Ras

FIG. 8. CAT assay showing the effect of the isolated inhibitorydomain on the repression of the IE-CAT plasmid by Oct 2.5. TheIE-CAT plasmid (5 ,ug) was transfected with either 20 ,ug of emptyexpression vector (track 1), 5 jig of the Oct 2.5 plasmid and 15 jig ofvector (track 2), 5 jig of Oct 2.5, 10 jig of the plasmid expressing theisolated inhibitory domain, and S jig of vector (track 3), or 5 jig of Oct2.5 and 15 jig of the inhibitory domain plasmid (track 4). (b) Activityof the IE-CAT plasmid when cotransfected with either the emptyCMV expression vector (V) or the Oct-2.5 CMV expression vector(2.5) together with a plasmid expressing the isolated repressor domainin either the sense (Rs) or antisense (Ras) orientation or with theempty simian virus 40 expression vector (P). Results are expressed as

percentages of the activity observed upon cotransfection of expressionvector alone. Values are averages of two determinations whose ranges

are shown by the bars.

virtually only the DNA-binding POU domain can repress theactivation of octamer-containing promoters by endogenousoctamer-containing factors in embryonal carcinoma cells,which contain several octamer-binding proteins not found inother cell types (21). Similarly, we have recently shown (16)

that the isolated POU domain of Oct-2 can prevent theactivation of the TEl and IE3 promoters by the strong trans-activating complex formed between the cellular transcriptionfactor Oct-1 and the HSV virion protein Vmw65, which playsa critical role in activating the IE promoters following HSVinfection of permissive cell types (22, 23).

In contrast, as shown here, in uninfected BHK cells whichcontain only Oct-1, the isolated POU domain has only a weakeffect on the basal activity of the IE3 promoter, and a similarweak effect is also observed when the POU domain and the IEpromoter are cotransfected into the ND7 neuronal cell line(37), which contains only Oct-1 in addition to endogenousOct-2 (34) (data not shown). Such a finding is in agreementwith the relatively weak transactivating ability of Oct-1 (33) inthe absence of Vmw65, which would result in only a smallreduction in the basal activity of the IE promoter even ifbinding of Oct-1 were prevented by the DNA-binding domainof Oct-2.

Hence, the inhibition of the IE promoter by neuronal formsof Oct-2 in transfected BHK cells or naturally in ND7 neuronalcells (15, 17) does not appear to operate solely via a mecha-nism in which DNA binding mediated via the POU domain ofOct-2 blocks the binding of a transactivating factor. Rather, thedata presented here suggest that the N-terminal region ofOct-2 is required for this effect, whereas the C-terminal regionof the protein present in Oct 2.1 or 2.2 antagonizes the effectand stimulates the IE promoter. Thus, deletion of the N- orC-terminal region results in a strong activator or a strongrepressor, respectively, of transcription when tested on the IEpromoter. Similarly, in forms of the protein such as Oct 2.1,which contain intact N and C termini, the activating effect ofthe C terminus predominates and the IE promoter is stimu-lated, whereas in forms of the protein with an altered Cterminus such as Oct 2.4 and 2.5, the inhibitory effect of the Nterminus predominates and the IE promoter is repressed.The stimulation of the IE3 promoter is likely to be mediated

by an activation domain which has been shown by a number ofgroups to be located at the C terminus of the molecule andwhich is responsible for the ability of Oct-2 to stimulate avariety of different octamer-containing promoters not acti-vated by Oct-1 (2, 6, 33). Interestingly, however, previousstudies have suggested that this stimulatory effect is dependenton only the C-terminal 18 amino acids in the case of apromoter containing a proximal octamer motif (6) or theC-terminal 64 amino acids in the case of a downstreamoctamer-containing enhancer (2). In neither case was anydetrimental effect observed when the region from amino acids377 to 399 containing a putative leucine zipper was deleted (2,6).

In contrast, in our experiments deletion of this regionproduced the strongest inhibitory effect on IE-CAT expressionof all the mutants tested. This may indicate that the boundaryof the C-terminal activation domain required is different fordifferent octamer/TAATGARAT promoters, depending, forexample, on the precise distance of the octamer motif from thepromoter and/or on the nature of the other transcriptionfactors binding in the vicinity. In agreement with this idea, weobserved that deletion of the putative leucine zipper motif didhave some effect on the ability of Oct-2 to stimulate theOct-CAT construct, whereas Tanaka and Herr (33) found thatdeletion of all 122 amino acids of Oct-2 C terminal to the POUdomain was required for the abolition of its ability to stimulateanother octamer-containing promoter. It is therefore possiblethat the entire C-terminal region functions as an activationdomain which directly stimulates the IE3 promoter. Alterna-tively, it remains possible that the C-terminal region of Oct-2

MOL. CELL. BIOL.

4 u


http://mcb.asm

.org/D

ownloaded from

http://mcb.asm.org/


stimulates the IE promoter indirectly by recruiting anotheractivating molecule to the promoter via the leucine zippermotif.Whatever the precise means by which it stimulates the IE

promoter, it is clear that deletion or alteration of the C-terminal region allows a region located at the N terminus ofthe molecule to exert an inhibitory effect on transcription. Thisresults in repression of theIE promoter below the basal levelobserved in the absence of Oct-2. Moreover, this region canalso limit the degree of activation of a promoter that can beactivated even by forms of Oct-2 with a truncated or altered Cterminus. Thus, the degree of activation of the Oct-CATconstruct was also enhanced in our experiments by deletion ofthe N-terminal region. Interestingly, the naturally occurringforms of Oct-2 as well as the C-terminal deletion are allcapable of repressing the tyrosine hydroxylase promoter,whereas the N-terminal deletion does not affect the activity ofthis promoter. This finding suggests that on the tyrosinehydroxylase promoter, the C-terminal activation domain isinactive, allowing the N-terminal domain common to all of thedifferent forms to exert its inhibitory effect. Hence, the preciseeffect of Oct 2.1 or 2.2. on a specific promoter may depend onthe balance between the inhibitory effect of the N-terminaldomain and the stimulatory effect of the C-terminal domain onthat promoter.

Previous studies (21) have shown that deletion of the first 99amino acids at the N terminus of Oct-2 results in an increase inits ability to activate two artificial octamer-containing promot-ers when tested in HeLa cells. In the case of the HSV IE3promoter, however, deletion of this region did not significantlyenhance the transactivating ability of Oct-2. Instead, we iden-tified a region between amino acids 161 and 186 which appearsto be responsible for the inhibitory effect that we observe. Thiselement forms part of a short region (labelled - in Fig. 1)which is rich in both acidic amino acids and proline residues,with seven prolines and six acidic residues being found in the25 amino acids from positions 161 to 186. Interestingly, these25 amino acids are interrupted by a 16-amino-acid insert in Oct2.2 which is absent in the other forms (35). This insert containsa further four prolines and one acidic residue and may enhancethe inhibitory activity of the region, since whereas both Oct 2.1and 2.2 can transactivate the IE3 promoter, we have observedthat Oct 2.2 transactivates the promoter more weakly than Oct2.1 and 2.3, in which the insert is absent (17).Although two different types of activation domain contain a

high content of either acidic or proline residues, respectively,the region of Oct-2 between residues 161 and 186 does notform part of the N-terminal activation domain described inprevious studies (21, 33). Moreover, a high content of prolineresidues has also been shown to be characteristic of severalinhibitory domains in different transcription factors such as theDrosophila homeodomain proteins even-skipped and paired(9). It is therefore unlikely that this region normally functionsas an activation domain and represses the IE promoter by anonspecific squelching effect in which adaptor factors requiredfor its activation are removed from the IE promoter (for areview, see reference 7). In agreement with this observation,both in our previous experiments (17) and those describedhere, the inhibitory effect of Oct-2 was specific for the IE3promoter and some others containing the octamer/TAATGARAT motif and was not observed with the Rous sarcoma viruspromoter (8) or with the HSV tk promoter in the vector pBL2CAT (18). Similarly, cotransfection of a plasmid expressing theisolated inhibitory domain relieved the inhibitory effect of Oct2.5 on the IE3 promoter rather than enhancing it as would be

expected if this region binds positively acting adaptor factors(Fig. 8).

It is possible that the inhibitory effect is achieved by pre-venting the binding of other transactivating factors to theirbinding sites in the IE3 promoter adjacent to the octamermotif. This is unlikely, however, since the inhibitory effects ofOct 2.4 and 2.5 can also be observed in artificial promoters inwhich the TAATGARAT motif is located in a different contextthan in the IE promoter (17). Moreover, in these promotersthe activity in the presence of Oct 2.4 or 2.5 can be reducedbelow that observed with the same promoter in the absence ofthe octamer motif or with a mutant motif which cannot bindoctamer-binding proteins (5).The conclusion from these studies that Oct 2.4 and 2.5 can

inhibit promoter activity by a mechanism which does notinvolve inhibiting the binding of other stimulatory factors issupported by the observation that the N-terminal inhibitorydomain can inhibit a heterologous promoter when linked tothe DNA-binding domain of GAL4. Hence, this domain caninhibit promoter activity via a different DNA-binding site in adifferent promoter when directed to this site via an appropriateDNA-binding domain, which suggests that the N-terminalregion of Oct-2 contains an inhibitory domain capable ofrepressing IE promoter activity as seems to be the case for theDrosophila homeodomain proteins even-skipped and paired(9). Interestingly, however, we observed that the inhibitoryeffect of Oct 2.5 on the IE3 promoter could be relieved byintroducing the isolated inhibitory domain in trans in theabsence of any DNA-binding domain. Hence, rather thanhaving an inherently inhibitory effect, this domain may func-tion by recruiting another inhibitory factor to the promoter,with this factor being titrated out by the isolated inhibitorydomain, hence relieving this effect. Further studies will berequired to define the precise mechanism by which this inhib-itory region exerts its effect.

In conclusion, the data presented here indicate that theN-terminal region of Oct-2 is responsible for the inhibitoryeffects of specific Oct-2 isoforms on the IE3 promoter and iscapable of achieving this effect when the stimulatory effect ofthe C-terminal region is abolished by mutation or deletion.The high levels observed in neuronal cells of forms such as Oct2.4 and 2.5 (17) or the Oct-2c form (32), which containalterations or deletions inactivating the C-terminal region,indicate that these forms are likely to play a critical role in theinhibition of many octamer/TAATGARAT-containing pro-moters observed in neuronal cells, which in the case of theHSV IE promoters plays a critical role in producing latentrather than lytic infection following initial exposure to HSV(13, 15).

ACKNOWLEDGMENTS

We are most grateful to D. Chikaraishi for the tyrosine hydroxylasepromoter construct, to M. Parker for the pSG424 vector, to T. Shenkfor the tk plasmid containing GAL4 binding sites, and to I. Wirth forthe gift of the Oct 2.1, 2.2, 2.4, and 2.5 expression vectors and theOct-CAT plasmid.

This work was supported by Action Research and the MedicalResearch Council.

REFERENCES1. Abken, H., and B. Reifenrath. 1992. A procedure to standardizeCAT reporter gene assay. Nucleic Acids Res. 20:3527.

2. Anweiler, A., M. Muller-Immergluck, and T. Wirth. 1992. Oct-2transactivation from a remote enhancer position requires a B-cell-restricted activity. Mol. Cell. Biol. 12:3107-3116.

3. Bradford, M. M. 1976. A rapid and sensitive method for thequantitation of microgram quantities of proteins utilizing the

VOL. 14, 1994


http://mcb.asm

.org/D

ownloaded from

http://mcb.asm.org/


principle of protein-dye binding. Anal. Biochem. 72:248-254.4. Clerc, R. G., L. M. Corcoran, J. H. LeBowitz, D. Baltimore, and

P. A. Sharp. 1988. The B-cell-specific Oct-2 protein contains POUbox and homeo box-type domains. Genes Dev. 2:1570-1581.

4a.Dawson, S. J., S. 0. Yoon, D. M. Chikaraishi, K. A. Lillycrop, andD. S. Latchman. 1994. The Oct-2 transcription actor repressestyrosine hydroxylase expression via the heptamer motif in thepromoter. Nucleic Acids Res. 22:1023-1028.

5. Dent, C. L., K. A. Lillycrop, J. K. Estridge, N. S. B. Thomas, andD. S. Latchman. 1991. The B cell and neuronal forms of theoctamer binding protein Oct-2 differ in DNA binding specificityand functional activity. Mol. Cell. Biol. 11:3925-3930.

6. Gerster, T., C.-G. Balmaceda, and R. G. Roeder. 1990. The celltype-specific octamer transcription factor OTF-2 has two domainsrequired for the activation of transcription. EMBO J. 9:1635-1643.

7. Goodbourn, S. 1990. Negative regulation of transcriptional initia-tion in eukaryotes. Biochim. Biophys. Acta 1032:53-77.

8. Gorman, C. M., G. T. Merlino, M. C. Willingham, I. Pastan, andB. Howard. 1982. The Rous sarcoma virus long terminal repeat isa strong promoter when introduced into a variety of eukaryoticcells by transfection. Proc. Natl. Acad. Sci. USA 79:6777-6781.

9. Han, K., and J. L. Manley. 1993. Transcriptional repression by theDrosophila even-skipped protein: definition of a minimal repres-sion domain. Genes Dev. 7:491-503.

10. Hatzopoulos, A. K., A. S. Stoykova, J. R. Erselius, M. Goulding, T.Neuman, and P. Gruss. 1990. Structure and expression of themouse Oct-2a and Oct-2b, two differentially spliced products ofthe same gene. Development 109:349-362.

11. He, X., M. N. Treacy, D. M. Simmons, H. A. Ingraham, L. S.Swanson, and M. G. Rosenfeld. 1989. Expression of a large familyof POU-domain regulatory genes in mammalian brain develop-ment. Nature (London) 340:35-42.

12. Kemp, L. M., C. L. Dent, and D. S. Latchman. 1990. Transcrip-tional repression of herpes simplex virus immediate-early genesand of octamer containing cellular promoters in neuronal cells ismediated by the octamer DNA element. Neuron 4:215-222.

13. Latchman, D. S. 1990. Molecular biology of herpes simplex viruslatency. J. Exp. Pathol. 71:133-141.

14. Lewis, E. J., C. A. Harrington, and D. M. Chikaraishi. 1987.Transcriptional regulation of the tyrosine hydroxylase gene byglucocorticoid and cyclic AMP. Proc. Natl. Acad. Sci. USA 84:3550-3554.

15. Lillycrop, K. A., C. L. Dent, S. C. Wheatley, M. N. Beech, N. N.Ninkina, J. N. Wood, and D. S. Latchman. 1991. The octamerbinding protein Oct-2 represses HSV immediate-early genes incell lines derived from latently infectable sensory neurons. Neuron7:381-390.

16. Lillycrop, K. A., J. K. Estridge, and D. S. Latchman. 1993. Theoctamer binding protein Oct-2 inhibits transactivation of theherpes simplex virus immediate-early genes by the virion proteinVmw65. Virology 196:888-891.

17. Lillycrop, K. A., and D. S. Latchman. 1992. Alternative splicing ofthe Oct-2 transcription factor is differentially regulated in B cellsand neuronal cells and results in protein isoforms with oppositeeffects on the activity of octamer/TAATGARAT-containing pro-moters. J. Biol. Chem. 267:24960-24966.

18. Luckow, B., and G. Schutz. 1987. CAT constructions with multipleunique restriction sites for the functional analysis of eukaryoticpromoters and regulatory elements. Nucleic Acids Res. 15:5490.

19. Macpherson, I., and M. Stoker. 1962. Polyoma transformation ofhamster cell clones-an investigation of the genetic factors affect-ing cell competence. Virology 16:147-151.

20. Muller, M. M., S. Rupert, W. Schaffner, and P. Matthias. 1988. Acloned octamer transcription factor stimulates transcription from

lymphoid-specific promoters in non-B cells. Nature (London) 336:544-551.

21. Muller-Immergluck, M. M., W. Schaffner, and P. Matthias. 1990.Transcription factor Oct-2a contains functionally redundant acti-vating domains and works selectively from a remote enhancerposition in non-lymphoid (HeLa) cells. EMBO J. 9:1625-1634.

22. O'Hare, P., and C. R. Goding. 1988. Herpes simplex virus regula-tory elements and the immunoglobulin octamer domain bind acommon factor and are both targets for virion transactivation. Cell52:435-445.

23. Preston, C. M., M. C. Frame, and M. E. M. Campbell. 1988. Acomplex formed between cell components and a herpes simplexvirus structural polypeptide binds to a viral immediate-early generegulatory DNA sequence. Cell 52:425-434.

24. Sadowski, I., and M. Ptashne. 1989. A vector for expressing GAL4(1-147) fusions in mammalian cells. Nucleic Acids Res. 17:7539.

25. Scheidereit, C., J. A. Cromlish, T. Gerster, K. Kawakami, G.-C.Balmaceda, R. A. Currie, and R. G. Roeder. 1988. A humanlymphoid-specific transcriptase factor that activates immunoglob-ulin genes is a homeobox protein. Nature (London) 336:551-557.

26. Scholer, H. R., A. K. Hatzopoulos, R. Balling, N. Suzuki, and P.Gruss. 1989. A family of octamer-specific proteins present duringmouse embryogenesis: evidence for germline-specific expressionof an Oct factor. EMBO J. 8:2543-2550.

27. Scholer, H. R., S. Ruppert, N. Suzuki, K. Chowdhury, and P.Gruss. 1990. New type of POU domain in germline-specificprotein Oct-4. Nature (London) 344:435-439.

28. Shi, Y., E. Seto, L.-S. Chang, and T. Shenk. 1991. Transcriptionalrepression by YYi, a human GL1-kruppel-related protein andrelief of repression by adenovirus EIA protein. Cell 67:377-388.

29. Singh, H., R. Sen, D. Baltimore, and P. A. Sharp. 1986. A nuclearfactor that binds to a conserved sequence motif in transcriptionalcontrol elements of immunoglobulin genes. Nature (London) 319:154-158.

30. Staudt, L. H., H. Singh, R. Sen, T. Wirth, P. A. Sharp, and D.Baltimore. 1986. A lymphoid-specific protein binding to theoctamer motif of immunoglobulin genes. Nature (London) 323:640-643.

31. Stow, N. D., M. D. Murray, and E. C. Stow. 1986. Cis-acting signalsinvolved in the replication and packaging of herpes simplex virustype-1 DNA. Cancer Cells 4:497-507.

32. Stoykova, A. S., S. Steiner, J. R. Erselius, A. K. Hatzopoulos, andP. Gruss. 1992. Mini-Oct and Oct-2c: two novel functionallydiverse murine Oct-2 gene products are differentially expressed inthe CNS. Neuron 8:541-558.

33. Tanaka, M., and W. Herr. 1990. Differential transcriptional acti-vation by Oct-I and Oct-2: interdependent activation domainsinduce Oct-2 phosphorylation. Cell 60:375-386.

34. Wheatley, S. C., C. L. Dent, J. N. Wood, and D. S. Latchman. 1991.A cellular factor binding to the TAATGARAT DNA sequenceprevents the expression of the HSV immediate-early genes follow-ing infection of nonpermissive cell lines derived from dorsal rootganglion neurons. Exp. Cell Res. 194:78-82.

35. Wirth, T., A. Priess, A. Annweiler, S. Zwilling, and B. Oeler. 1991.Multiple Oct-2 isoforms are generated by alternative splicing.Nucleic Acids Res. 19:43-51.

36. Wirth, T., L. Staudt, and D. Baltimore. 1987. An octamer oligo-nucleotide upstream of a TATA motif is sufficient for lymphoidspecific promoter activity. Nature (London) 329:176-178.

37. Wood, J. N., S. J. Bevan, P. Coote, P. M. Darn, P. Hogan, D. S.Latchman, C. Morrison, G. Rougon, M. Theveniau, and S. C.Wheatley. 1990. Novel cell lines display the properties of nocicep-tive sensory neurons. Proc. R. Soc. Lond. Ser. B 241:187-194.

MOL. CELL. BIOL.


http://mcb.asm

.org/D

ownloaded from

http://mcb.asm.org/

repression of a herpes simplex virus immediate-early promoter by

Documents