The transcriptional start sitefor a human U6 small ... - BioMedSearch

1 downloads 0 Views 2MB Size Report
Three, 5, or 10 bp were deleted or 3, 5, or 9 bp were inserted between the TATA element and the +1 initiation site (the delTA and inTA mutants). Single spacingĀ ...
Nucleic Acids Research, Vol. 20, No. 18 4903-4912

(D) 1992 Oxford University Press

The transcriptional start site for a human U6 small nuclear RNA gene is dictated by a compound promoter element consisting of the PSE and the TATA box Randal S.Goomer and Gary R.Kunkel* Department of Biochemistry and Biophysics, Texas A & M University, College Station, TX 77843-2128, USA Received May 14, 1992; Revised and Accepted August 1, 1992

ABSTRACT Transcription of vertebrate U6 snRNA genes by RNA polymerase Ill requires two sequence elements in the proximal promoter region: the PSE (proximal sequence element, found in snRNA promoters transcribed by RNA polymerase 11) and the TATA element (found in many mRNA promoters). The locations of the PSE and the TATA box are important determinants for transcriptional start site selection in their respective RNA polymerase 11 promoters. In vertebrate U6 genes the PSE and the TATA elements are located in approximately the same positions as in the polymerase 11 transcribed genes, but their respective roles in initiation site selection are unknown. We have analyzed the effects of spacing changes between the PSE and the TATA element, and between the two elements and the normal U6 start site on human U6 gene transcription. The spacing requirement between the two elements is highly stringent, implying a possible interaction between the factors that bind them. Our results discount the possibility that the location of either the PSE or the TATA element, by itself, dictates efficient selection of a transcriptional start site. Instead, we suggest that the two elements form a compound promoter element whose location dictates the start site of transcription from the human U6 gene promoter. INTRODUCTION Classification of eukaryotic gene promoters is broadly determined according to whether transcription is carried out by either RNA polymerase I, II or HI. Historically, these three classes have been differentiated by the sequence elements present in each promoter. The promoters of many protein-coding genes transcribed by RNA polymerase H (pol H) contain a TATA box, the location of which selects the start site of transcription (reviewed in (1,2)). In other vertebrate pol H genes, such as those that encode most of the abundant spliceosomal snRNAs (Ul, U2, U4 and U5), transcription by RNA polymerase II does not require the presence *

To whom correspondence should be addressed

of a TATA box. Instead the promoters of these genes contain a so-called proximal sequence element (PSE) that is necessary for transcription and controls the site of initiation (reviewed in (3)). In contrast, the earliest characterized RNA polymerase III (pol I1)-transcribed genes contain an intragenic control region (ICR) with two separable sequence elements that are essential for transcription (Class IIIICR genes) (reviewed in (4)). Vertebrate U6 snRNA genes, also transcribed by pol HI (5-7), are examples of a class of pol HI-transcribed genes that do not contain an ICR but require only the presence of 5' flanking sequences for transcription (class IHEXT); reviewed in (8-10)). Vertebrate U6 snRNA genes are unique not only because they lack an ICR but also because promoter elements in the 5' flanking region are very similar to those of pol II-transcribed genes. The proximal promoter of vertebrate U6 genes consists of a TATA box sequence (-29 TATATA -24 in a human gene) and a PSE sequence (-66 to -47 in a human gene). Interestingly, the U6 TATA element is located in approximately the same position as the TATA box of many pol II-transcribed genes. The TATAlike sequence is conserved among all U6 genes thus far isolated from different organisms. The U6 TATA element is required for transcription by RNA polymerase HI, and its elimination can cause the gene to be transcribed by pol 11 (11 - 13). Furthermore in the plant, Arabidopsis, U2 and U6 genes contain both a TATA box and a nearby upstream sequence element (USE). The choice of pol II (U2 gene) or pol III (U6 gene) depends only on the spacing between the USE and the TATA box and between the TATA box and the start site (14). It has been shown recently that recombinant TATA-binding protein (TBP), the pol II basal transcription factor, can promote U6 transcription in vitro (15-17). The PSE is required for efficient transcription of the human U6 gene (12,18). It is about 70% similar in sequence with U1/U2 PSEs, located in approximately the same position and functionally interchangeable (12,19). In the case of Ul and U2 genes, the site of transcription initiation is fixed by the relative location of the PSE (20,21). A partially purified protein has been isolated that binds in the PSE regions of vertebrate U6 snRNA genes (22,23).

4904 Nucleic Acids Research, Vol. 20, No. 18 The proximal region of vertebrate U6 promoters contains two required elements, each of which individually determines the transcriptional start site in their respective pol II promoters. When together, these elements are essential components of a pol IIIspecific U6 promoter. We have analyzed the effect of altered spacing of these elements on start site selection and efficiency of transcription. Our results discount the possibility that either the PSE or the TATA box itself determines the start site. Instead the two elements together form a compound promoter unit that dictates the transcriptional start site. The efficiency of expression is strongly affected by the spacing between the TATA box and the PSE, implying a possible interaction, either direct or indirect, between the factors that bind these two elements.

MATERIALS AND METHODS Construction of Promoter Mutant Plasmids The parent template used for construction of mutants in this study was the U6 maxigene contained in an 800 bp human DNA fragment (18) inserted in the SmaI site of the M13mpl8 vector. Spacing mutants were prepared by oligonucleotide-directed mutagenesis using single-stranded recombinant M 13 DNA that was partially substituted with uracil (24). The oligodeoxynucleotide primers used for mutagenesis are listed below (written as 5' to 3', from left to right): inPI: inP2: inP3: inP5: inPIO: delPI: delP2: delP3: delP5: delPlO: delP15: inTA3: inTA5: inTA9: delTA3: delTA5: delTAIO: inP3/TA-:

TATTTCGATTTCTTGgGCTTTATATATCTTG TATTTCGATTTCTTGgtGCTTTATATATCTTG TATTTCGATTTCTTGgtaGCTTTATATATCTTG TATTTCGATTTCTTGgtaccGCTTTATATATCTTG TATTTCGATTTCTTGgtaccctcgaGCTTTATATATCTTG GTATTTCGATTTCTTGCTTTATATATCTTG AGTATTTCGATTTCTGCTTTATATATCTTG AAGTATTTCGATTTCGCTTTATATATCTTG GAAAGTATTTCGATTGCTTTATATATCTTG AACTTGAAAGTATTTGCTTTATATATCTTG CGTAACTTGAAAGTATTTATATATCTTGTG TCTTGTGGAAAGGACgtcGAAACACCGTGCTCG TCTTGTGGAAAGGACcggtcGAAACACCGTGCTCG TCTTGTGGAAAGGACggatccatcGAAACACCGTGCTCG TATCTTGTGGAAAGGAAACACCGTGCTCGC TATCTTGTGGAAAGGACACCGTGCTCGCTT TATATCTTGTGGAAACCGTGCTCGCTTCGG TCGATTTCTTGgtaGgaggtaccgaCTTGTGGAAAGGACG

Compensatory spacing mutants were constructed in a stepwise fashion using the appropriate primers with single-stranded DNA from the inP3, inPO0 and delPI0 constructs. The mutations were verified by dideoxy sequence analysis. The mutated 800 bp U6 maxigene inserts from phage RF DNAs were subcloned into the polylinker of the pGEM3Zf(-) vector (Promega, Madison, WI). Transfection and RNA isolation Subconfluent monolayer 293 cells in 100 mm tissue culture dishes were transfected with 8 jig of the test plasmid by a calcium phosphate coprecipitation method as described previously (18). A pGEMI plasmid carrying the chicken fl-tubulin gene (cf3; (25)) (5 Ag) was cotransfected as a control for transfection efficiency and RNA recovery. Total RNA was isolated 60 hours post-transfection following a rapid procedure for RNA isolation (26). Briefly, transfected monolayer cells were scraped off the plates, pelleted at 500 x g and washed once with phosphatebuffered saline solution. The cell pellet was suspended in 1.1 ml of solution A (8M guanidine hydrochloride, 0.3M sodium

acetate (pH 7), 1 % Sarkosyl and 0. 14M ,B-mercaptoethanol). The lysate was passed through a 22 gauge needle 10 times, and centrifuged at 12,000g for 30 min. Nucleic acids in the supernatant (1 ml) were precipitated with 0.5 volume 95% ethanol, collected by centrifugation and resuspended in 0.5 ml of solution B (4M guanidine thiocyanate and 25mM sodium citrate, pH 7). After reprecipitation with 0.5 volume 95% ethanol, the nucleic acid pellet was washed 3 times, each, with 95 % and 70% ethanol and dried. Specific RNAs were analyzed by primer extension as described below.

In vitro transcription in a 293 cell S100 extract S100 extracts were prepared from 293 cells grown in suspension culture as described previously (5,27), and then dialyzed against 0.1M KCl, 20% glycerol, 20mM HEPES (pH 7.9), 0.2mM EDTA, 0.5mM dithiothreitol. Conditions for the in vitro transcription reaction have been described previously, except that 0.5mM unlabeled GTP was substituted for radiolabeled GTP (5). 500 ng of each plasmid DNA were added per reaction, and most tubes contained a-amanitin at a final concentration of 1 ,^g/ml. Subsequent to the transcription reaction, nucleic acids in the reaction mix were treated with DNaseI (RNase-free), extracted with phenol/CHCl3, and precipitated with ethanol. In order to control for RNA recovery, a constant amount of an unlabeled synthetic transcript was added to each transcription reaction after quenching. This synthetic RNA was prepared by transcription of HindIII-cut pGEM3Zf(-) plasmid DNA with T7 RNA polymerase. Transcripts were analyzed by primer extension as described below.

Primer extension All primers were labeled by phosphorylation with 'y-32P-ATP (NEN, 3000 Ci/mmole) and T4 polynucleotide kinase. Primer extension of U6 maxigene transcripts was performed with a 'maxi' primer (5'-CACGCCTCGAGGGAA-3') that anneals to maxiU6 RNA from positions + 85 to +99. Primer extension of cf33 RNA used a labeled 'BT-68' primer (5'-AGGTGCACGATCTCC-3') that anneals to exon 1 of the c,B3 transcript from positions +53 to +68. The recovery control RNA from the in vitro transcription reactions was detected using the 'SP6 promoter primer' (5 '-GATTTAGGTGACACTATAG-3'; Promega), resulting in an 87 nt product after extension. The primer extension reactions were performed as described previously, products were separated by electrophoresis on 10% polyacrylamide gels containing 8.3M urea, and labeled bands were detected by autoradiography (18). The exact 5' ends of transcripts were determined by electrophoretic comparison of primer extension products with a dideoxy sequencing ladder generated from maxiU6 plasmid DNA using a phosphorylated 'maxi' primer.

Quantitative analysis of primer extension gels For Figures 2C and 3C, dried gels containing primer extension products from 3 independent transfection experiments were analyzed by direct counting on a Betascope 603 Blot Analyzer (Betagen Corporation). Radioactivity (CPM) for each maxiU6and cfl3-derived band was measured and corrected for background. The amount of maxiU6 product in each lane was normalized for transfectional efficiency by comparison to the radioactivity detected from the c,B3 product. In each experiment the normalized value of the maxiU6 band derived from each mutant was compared to the wild-type value (set at 100%).

Nucleic Acids Research, Vol. 20, No. 18 4905

RESULTS The general structures of mutant U6 promoter constructions and the location of the maxiU6-specific primer used in this study are shown in Figure 1. All promoter spacing alterations were prepared by oligonucleotide-directed mutagenesis upstream of a U6 maxigene (human U6 gene containing a 9 bp linker located within the transcribed region (18)). Exactly 1, 2, 3, 5 or 10 basepairs were either inserted or deleted between the PSE and the TATA element (the inP and delP mutants). Three, 5, or 10 bp were deleted or 3, 5, or 9 bp were inserted between the TATA element and the + 1 initiation site (the delTA and inTA mutants). Single spacing changes between the PSE and the TATA element result in two distinguishable spacing changes: a change in the distance between the PSE and the TATA element, and between the PSE and the normal start site. Similarly, a change in spacing between the TATA element and the normal start site also changes the spacing between the PSE and the normal start site. In order to separate these spacing variables, we have compensated for alteration in the spacing between the PSE and the TATA element with corresponding changes between the TATA and the normal start site. Starting with the inP3, inPlO and delPIO constructs, several mutants were analyzed that contained compensatory spacing changes further downstream in the promoter (inP3/delTA3, inPIO/delTAlO and delPlO/inTA9). Transcription of different plasmids was analyzed following transfection of human 293 cells or by in vitro transcription in an SI00 extract prepared from 293 cells. RNAs were analyzed by primer extension using a primer specific for the U6 maxigene transcript (the 'maxi' primer; Fig. iB).

A PSE . .

TATA

------

g TTACCGTAACTTGAAGTatttcgatttcttggc,TTTATATA cttgtggaaaggacgaaacaccGtgctcgcttcgg.

inP 1,2,3,5,10

inTA 3,5,9 delT'rA 3,5,10

delP 1,2,3,5,10,15

spacing

sato

+

+

Compensatory

3 inP 3/delTA 1 0/delTA 10 _inP deiP 9 10/inTA

B 0

000-tCD00 + + + +

4---- ---------------'maxf primer -

99 nts

-I

Figure 1. Structures of the human U6 promoter mutants and the primers used in this study. The PSE (from -66 to -47) and the extended TATA element (from -31 to -23) sequences are capitalized. (A) InPx mutant constructs: x bp were inserted between the PSE and the TATA element at the indicated position. DelPx mutant constructs: x bp were deleted between the PSE and the TATA element. InTAx mutant constructs: x bp were inserted between the TATA element and the normal U6 start site at the indicated position. DelTAx mutant constructs: x bp were deleted between the TATA element and the normal U6 start site. Compensatory change of spacing constructs: Spacing changes between the PSE and the TATA element were compensated by spacing changes between the TATA element and the normal start site, thereby placing the PSE at the wild-type position with respect to the normal U6 start site. (B) The U6 maxigene contains a 9 bp XhoI linker (open box) inserted within the transcribed region of the human U6 gene (18). The 15 nt 'maxi' primer (stippled box) is centered over the XhoI linker and anneals specifically to the U6 maxigene transcripts. The extension product from the 'maxi' primer to the normal U6 start site is 99 nucleotides (dashed line).

A Change in Spacing Between the TATA Element and the PSE Primarily Affects the Efficiency of Expression Insertion of 1 or 2 bp between the PSE and the TATA element (inPl, inP2; Fig. 2A) did not have a significant effect on the relative level of expression in transfected cells (Fig. 2B, lanes 3 and 4, and 2C). However, expression levels were greatly diminished for the inP5 and inPIO plasmids (Fig. 2B, lanes 6 and 7, and 2C). Expression from the inP3 mutant was quite high but the initiation site was heterogeneous. A separate transcript that initiates three nucleotides upstream of the normal start site (the -3 transcript) was produced along with a RNA initiating at the normal start site (Fig. 2B, lane 5). These start sites were determined exactly by comparing the mobility of inP3 primer extension products with the dideoxy sequencing ladder obtained after annealing a phosphorylated 'maxi' primer to U6 maxigene plasmid DNA (results not shown). It has been reported that a mutant human U6 promoter lacking the TATA element is transcribed by RNA polymerase II with initiation sites 3 or 4 nucleotides upstream of the normal U6 start site (12). We explored the possibility that the -3 transcript resulting from the inP3 mutant construct may have been transcribed by RNA polymerase II. We inserted the 3 bp that constitute the inP3 mutation between the PSE and a disrupted TATA element (from plasmid U6/dl(-32,-23) described previously (28)) to test whether this altered PSE spacing induced a higher level of snRNA pol II (TATA-independent) transcription from the U6 promoter (inP3/TA- mutant, see sequence in Materials and Methods). However, no transcript initiating at the - 3 position was observed for this mutant (results not shown). Furthermore, using a downstream primer (5' end at position + 191 of the U6 maxigene), we did not detect any transcript initiating at + 1 or -3 from the inP3 mutant, expecting that a pol II transcript might extend beyond the (dT)5 po' III terminator (results not shown). In addition, we observed the -3 start site after in vitro transcription of the inP3 template in an S100 extract in the presence of a concentration of aamanitin that inhibits polymerase II transcription (Fig. 2D, lane 5). Therefore we conclude that at least the majority of transcripts from the inP3 promoter which initiate at -3 are not synthesized by pol II. In vitro transcription of all inP constructs is shown in Figure 2D. InPl and inP2 mutations did not affect transcriptional efficiency significantly (Fig. 2D, lanes 3 and 4), whereas synthesis of maxiU6 RNA was greatly diminished for the inP5 and inPIO mutants (Fig. 2D, lanes 6 and 7). In contrast to the transfection results, a low level of transcription that starts at + 1 was still detectable after in vitro transcription from inP5 and inPlO. Deletion of 1 bp between the PSE and the TATA element (delP1) did not affect the relative level of U6 maxigene expression (Fig. 3B, compare lanes 1 and 2). Expression levels were significantly reduced in the delP2, 3, 5, 10 and 15 constructs (Fig. 3B, lanes 3 through 7). Although poorly expressed relative to the wild-type promoter, the delPIO construct supported a somewhat enhanced level of expression in vivo in comparison to the delP5 template (Fig. 3B, compare lanes 5 and 6, and Fig. 3C). Thus, removal of one-half helical turn between the PSE and TATA elements is somewhat more deleterious than removal of a complete turn of DNA. A change in the helical orientation between the control elements of some eukaryotic promoters has been shown to affect expression levels (29 -32). However, the results from several independent experiments indicated that

4906 Nucleic Acids Research, Vol. 20, No. 18

B

Ii 1-VIV(

t-

w~~~~~~~~~~~~~~~~~~~~~ 0 -a

,,. &

-3lnit. +I init.

.14

a ..4

C'.-n

a

a .

_ _ _ _ _

>

A *ffA fA2

PSE_ LPSE_l t

IuPSE Lp!E YPsE

rAT

_J

VIf I

;-.

ilI II '2, i11i1 r ii-1 '5( Jliii l(k,),

G

'ATA -, . -

A

JI

a

C..

G

rrATAJ

.M...

C

1 l I.\;i / l..;

120,

M

m;r)l; \p >

m

-!

.-,

2 3 4 5 6

i

",

7

-

in'.

80.

ho

ammamma

D

I ()0

Z.

c! 3

0 ;1i \\7.'1>. ll' [ !t i" vl!' ~v.r ~ ~ ~

jil{

I.11

l

:

:3

40().

3 iiiit. +1 init

20

COntrolĀ£

0 XV

---'''l-- I'''

i

'-

P-

L

4

._ *

a4

a-r.

P-

-__4f mm

2

_m_4__4I_ do -'

1

3

6

7

Figure 2. Effects of insertions between the PSE and the TATA element (the inP constructs). (A) Mutated U6 proximal promoters showing the indicated spacing changes. The inserted nucleotide bases are capitalized. The boxed 'TATA' denotes the entire U6 A/T-rich element (TTTATATAT). The asterisk indicates an additional transcriptional initiation site other than the normal U6 start site. The 'G' in bold type is the normal initiation site. (B) Transfection experiments: 293 cells (100 mm plates) were transfected with the indicated U6 plasmids (8 'g) together with 5 Ag of a plasmid containing the chicken (3-tubulin gene (cjS3) as a control for transfection efficiency. Total RNA was collected 60 hours post-transfection and analyzed by primer extension using the 'maxi' primer and the 'BT-68' primer. The 'BT-68' primer anneals to ci33 RNA and yields major primer-extension products that are 68 and 69 nt (labeled 'c(33'). Bands are labeled that correspond to the U6 maxigene transcripts initiated at the normal U6 transcriptional start site, and at position -3, observed with the inP3 mutant. (C) Quantitation of expression from inP mutant plasmids: Radioactivity in bands on dried gels was measured using a Betascope 603 Blot Analyzer (Betagen Corp.). Band intensities were normalized using the 13-tubulin signal to correct for transfection efficiency and RNA recovery. The results shown are the average of three independent transfection experiments. Error bars represent one standard deviation from the mean value. (D) In vitro transcription: U6 plasmids (500 ng) were incubated in a dialyzed S100 extract prepared from 293 cells. One tg/mI ax-amanitin was included in each reaction. An exogenous RNA, transcribed from the pGEM3Zf(-) vector using T7 RNA polymerase was added to each sample as an internal control for RNA recovery and primer extension. U6 maxigene transcripts were analyzed by primer extension with the 'maxi' primer (bands labeled '-3 init.' and '+1 init.'), and the internal control was analyzed with the 'SP6 promoter primer' (Promega) that anneals to the exogenously added RNA and yields an 87 nt product (band labeled 'control'). The lane labeled 'vector' is a control for transcription from vector sequences and contains products of primer extension after transcription of the pGEM3Zf(-) plasmid (Promega) lacking any human DNA insert.

further contraction in the PSE-TATA spacing (delP15) did not reduce expression to that from the delP5 template, although synthesis from delP15 was slightly lower than that from delPO0 (Fig. 3C). It should be noted that in the delPl5 template the PSE consensus sequence and the A/T-rich segment that constitutes an extended TATA element are almost as close as they can be positioned without overlap (Fig. 3A). In the context of this system we cannot determine conclusively whether a strict helical periodicity between the PSE and TATA element is important. Primer extension products of the delP mutants transcribed in vitro are shown in Figure 3D. The transcriptional levels of all the delP constructs were reduced moderately in comparison to

the wild-type construct. In some experiments the reduction was more severe (results not shown). However, a periodic modulation of transcriptional efficiency, such as in the transfection experiments, was not apparent.

Spacing changes between the TATA element and the normal start site generate new transcriptional initiation sites Short DNA segments were either added or deleted between the

TATA element and the normal start site by oligonucleotidedirected mutagenesis (the inTA or delTA mutants, Figs. IA and 4A). In these constructions the PSE and the TATA element were moved coordinately without disrupting their normal relative

[PE|,l1r";.^XlTAt{'TAv|.

A~~~~~~~~~1TT &

at

Q

t

t ; a'' g :T:*. _

PSEt t ,i zit

t

PSE lf{t PSE l

t

tPSEL

C

*:

'

c.1

i. a-g a *:c ._ -

IR'.': .Q

:a:: G

...

Nucleic Acids Research, Vol. 20, No. 18 4907

B

I II -ivo

+1 init.

*

W.T.

* t :G.... delPI 1.i i t . v :

G..... + (i>u 5+; iifdelP3 ''4"''o''' l'l l'*2 c113 ,i za.: ; .* .hR i ,*..X-i delP5 ..................... e4

ti

t

*34

d6

delPI0 12 34 5 67 delP1'42:4iii 5 :''G.......................

WCG......

I delP expression in-vivo I

D

In-vitro 1

(0 la

P4

3 _ 8! e) t

.2

'0

'0

0

5

6

'0 '0

+ 1 init. -

-t

a.

-o

-Ce

z

'0

LO

'U.

0 0

C.

10 -4

control -*

W*

S

a.

tu

'U

1

2

3

4

7

8

Figure 3. Effects of deletions between the PSE and the TATA box (delP constructs). (A) U6 proximal promoter regions showing the indicated spacing changes. (B) In vivo expression: 293 cells were transfected with the indicated U6 plasmids as described in Materials and Methods. The primer extension products corresponding to the normal U6 transcriptional start site are labeled '+ 1 init.'. The bands labeled 'c,B3' correspond to c,33 RNA primer-extension products, a control for transfectional efficiency between different plates. (C) Quantitation of expression from delP mutant plasmids: Relative radioactivity from dried gels was measured and normalized to expression from the wild-type U6 promoter as described in Materials and Methods. The results are the average of three independent transfection experiments. Error bars represent one standard deviation from the mean. (D) In vitro transcription: U6 plasmids were transcribed in a 293 cell S100 extract as described in Materials and Methods. U6 maxigene transcripts were analyzed by primer extension with the 'maxi' primer (band labeled '+ 1 init.'), and the internal control RNA was extended with the 'SP6 promoter primer' (band labeled 'control'). The 'vector' lane is a control for transcription from the pGEM3Zf(-) plasmid lacking any insert of human DNA.

spacing. Expression from the inTA or delTA constructs resulted in new transcriptional initiation sites whose locations tended to be a relatively fixed distance downstream of the TATA element. Insertion of 3 bp between the PSE/TATA element and the normal start site (inTA3) resulted in a majority of transcripts that initiated 3 nucleotides upstream of the normal start site and a minority that initiated at the + 1 position (Fig. 4B, lane 2). Insertion of 5 bp (inTA5) resulted in heterogeneous start sites with a major one at the same position as inTA3 (-3), and minor start sites at the same position as inTA9 (-8) and at the +1 position (Fig. 4B, lane 3). Likewise, insertion of 9 bp between the PSE/TATA element and the normal start site (inTA9) resulted in a major initiation site at -8 and a minor one at the + 1 position (Fig. 4B, lane 4).

Just as the inTA mutant templates produced major transcripts starting upstream, deletion of segments downstream of the TATA element (delTA mutants) usually induced downstream start sites. For example, deletion of 5 bp yielded a major transcript that initiated downstream at the +7 position while deletion of 10 bp resulted in a major transcript that initiated further downstream at the + 12 position (Fig. 4B, lanes 6,7; note that mobilities of bands in lane 6 are reduced). In one case, delTA3, the transcriptional start site was unchanged, and the efficiency was reduced only moderately (Fig. 4B, lane 5). Transcription of inTA and delTA spacing mutants in vitro was comparable to the transfection results (compare Figs. 4B and 4C). Only one minor difference was apparent; transcription from the inTA5 construct initiated at the -8 position almost as efficiently

4908 Nucleic Acids Research, Vol. 20, No. 18

B 0-i

A

:.

:*. \1;.xi/1.1.

.:

4 I.1 ..

, !