Proc. Natl. Acad. Sci. USA
`Vol. 83. PP. 4134-4137, June 1986
`Biochemistry
`
`lac repressor blocks transcribing RNA polymerase
`and terminates transcription
`
`(transcription regulation/repressor—operator function)
`
`ULRICH DEUSCHLE*, REINER GENTZ, AND HERMANN BUJARD*l
`Central Research Units, F. Hoffmann-La Roche and Co. AG, CH-4002 Basel, Switzerland
`
`Communicated by Werner Arber, February 5, 1986
`
`Operator sequences are essential elements in
`ABSTRACT
`many negatively controlled operons. By binding repressors,
`they prevent the formation of active complexes between RNA
`polymerase and promoters. Here we show that the Escherichia
`coli lac operator-repressor complex also efficiently interrupts
`ongoing transcription. This observation suggests a mechanism
`of action for operators located distal to promoter sequences.
`
`It is generally believed that repressors of prokaryotic operons
`act exclusively by preventing the onset of transcription. This
`view is supported by a wealth of experimental data, of which
`the most convincing are the structural analyses of regulatory
`regions: operators are located within the DNA sequence
`covered by a promoter-bound RNA polymerase (i.e., be-
`tween positions +20 and -50, where + 1 is the first nucleotide
`transcribed) (1-3) or, as in the case of promoter P1 of the
`Escherichia coli gal operon, within the cAMP—CAP
`(catabolite activator protein) binding sites (4). Thus, by
`occupying an operator, a repressor may either obscure a
`promoter sequence from being recognized by RNA polymer-
`ase or prevent the formation of an active complex between
`the enzyme and the promoter (5-7). However, within the lac
`operon, as well as within the gal operon, additional operator
`sequences were identified well downstream of the regulatory
`region (8, 9), and although an in vivo function for such an
`operator was demonstrated in the gal system (9), its mode of
`action has not been elucidated. The most straightforward
`mechanism, the direct interference of an operator—repressor
`complex with the transcribing enzyme, is generally ruled out
`(7, 10) despite suggestive genetic and biochemical data
`(11-13). Here we present evidence that the lac repres-
`sor-operator complex is indeed an efficient terminator of
`transcription in vivo and in vitro, suggesting an obvious mode
`of action for operator sequences found, for example, within
`structural genes of operons.
`We had observed that, when transformed with plasmid
`pGBU207, E. coli cells showed differences in tetracycline
`resistance depending upon the internal level of lac repressor.
`In pGBU207 (14), a lac operator sequence is located between
`promoter P3207 and the coding sequence of the tet region;
`therefore, we analyzed the effect of an isolated lac operator
`sequence inserted into a transcriptional unit distal to the
`promoter.
`
`MATERIALS AND METHODS
`
`Plasmids and Bacteria. The pDS1 vector system and the
`promoters P525 and PD/E20 have been described (15, 16). The
`lac operator was obtained as a 54-base-pair (bp) Hpa II—Alu
`I fragment from pBU10 (14). Plasmid pDM1.1, which carries
`the lad" gene and the p15A replicon, was a gift of M. Lanzer
`
`The publication costs of this article were defrayed in part by page charge
`payment. This article must therefore be hereby marked “advertisement"
`in accordance with 18 U.S.C. §1734 solely to indicate this fact.
`
`(ZMBH, Univ. of Heidelberg). All plasmids and in vivo
`RNAs were prepared from transformed E. coli DZ 291 (14).
`In Vitro Transcripts. In vitro transcription was carried out
`under standard_conditions (14, 16), whereby a 50-pl assay
`mixture contained 0.2 pmol of template (construct A, carry-
`ing promoter P525 or PD/E20 in plasmid pDS1; Fig. 1), 1 pmol
`of E. coli RNA polymerase, and [a-32P]UTP whenever
`labeling of the transcription products was required. The
`reaction mixtures were incubated at 37°C in the absence or
`presence of lac repressor (gift of M. Lanzer). Repressor was
`inactivated by addition of isopropyl
`[3-D-thiogalactoside
`(IPTG) to a final concentration of 200 ;:.M.
`In general,
`incubation was for 3 min before samples were directly
`prepared for PAGE.
`In Vivo Transcripts. E. coli cells transformed with the
`proper plasmid were grown to an OD“ of 0.5 in M9 medium
`containing 10% Luria broth (17). Labeled RNA was obtained
`by adding 500 p.Ci (1 Ci = 37 GBq) of [3H]uridine to 10 ml of
`the logarithmically growing culture. After 1 min at 37°C, cells
`were quickly chilled in liquid nitrogen and RNA was isolated
`according to Glisin et al. (18). High intracellular levels of lac
`repressor were achieved by the simultaneous presence of the
`compatible plasmid pDM1.1. Repressor was inactivated by
`addition of IPTG (200 pg/ml) to the cultures 60 min before
`harvest.
`
`Nuclease Sl Mapping (19). A suitable DNA fragment for the
`characterization of the 3’ ends of in vivo and in vitro
`transcripts was obtained by cleaving construct A (Fig. 1) with
`Acc I and Pvu II. The Ace I cleavage site located 147 bp
`upstream of the operator sequence was filled in with [a-
`32P]dATP, resulting in a 3’-labeled 318-bp fragment covering
`the entire operator sequence. About 0.01 pmol of the labeled
`DNA fragment was denatured and mixed with one-fourth of
`an in vitro transcription assay mixture or with 10 pg of total
`cellular RNA. The nucleic acids were allowed to hybridize
`(volume 30 al, 80% forrnamide/0.4 M NaCl/40 mM Tris/HCI,
`pH 8) for 2 hr before 300 pl of S1 buffer (19) containing 20
`units of nuclease S1 were added. After 2 hr at 14°C, the
`S1-resistant material was analyzed by electrophoresis in 8%
`polyacrylamide/8 M urea gels.
`Quantitation of in Vivo RNA by Hybridization. RNA was
`labeled with [3H]uridine and isolated as described above.
`Dihydrofolate reductase (DHFR)- and chloramphenicol ace-
`tyltransferase (CAT)-specific transcripts were quantified by
`hybridization with an excess of single-stranded M13 DNA
`carrying the proper DHFR and CAT gene sequences, respec-
`tively. The hybridized material was collected by filtration
`
`Abbreviations: bp, base pair(s); DHFR, dihydrofolate reductase;
`CAT, chloramphenicol acetyltransferase; IPTG,
`isopropyl B-D-
`thiogalactoside.
`;
`‘Present address: Zentrum fiir Molekulare Biologic, Heidelberg, Im
`Neuenheimer Feld 282, D-69 Heidelberg, Federal Republic of
`Germany.
`‘To whom all correspondence should be addressed.
`
`Mylan v. Genentech
`Mylan V. Genentech
`IPR2016-00710
`IPR2016-00710
`Genentech Exhibit 2076
`
`4134
`
`

`
`Biochemistry: Deuschle et al.
`
`-_
`P
`
`DHFR
`
`CAT
`
`To
`
`Proc. Natl. Acad. Sci. USA 83 (1986)
`
`4135
`
`FIG. 1. Transcriptional unit used for operator insertion. The
`standard transcription unit of the pDS1 vector system (15) contains
`the coding sequence of the dihydrofolate reductase (DHFR) and the
`E. coli chloramphenicol acetyltransferase (CAT) genes, both of
`which can be brought under control of a single promoter (P).
`Transcripts of defined size are obtained by the function of terminator
`to of phage X, which also prevents extensive read-through into other
`parts of the plasmid (15). Expression of this unit in vivo under the
`control of promoter P51, or Pwm (16) yields exclusively CAT
`protein, since only this sequence carries a functional translational
`start signal ([3). The lac operator (0) sequence (positions -17 to +34,
`ref. 7) was fused to either Hirrdlll or BamHI synthetic linkers and
`inserted into the HindIII (H) or BamHI (B) site,
`resulting in
`constructs A and B, respectively. The distances between the pro-
`moter (P) and sites B, H, and to are about 100, 670, and 1700 bp,
`respectively, depending somewhat upon the position of the promoter
`within the cloned fragment.
`
`through nitrocellulose and its radioactivity was monitored.
`This method has been described previously (20, 21).
`
`RESULTS
`
`lac Repressor—0perator Complex Functions as a
`Regulatable Terminator. The transcriptional unit used in
`these experiments has been described earlier as part of the
`pDS1 vector system (15).
`In these vectors,
`the coding
`sequence of DHFR and CAT genes are under the control of
`a single promoter, giving transcripts of =1700 nucleotides
`due to the terminator to at the end of the CAT gene (Fig. 1).
`The transcription units analyzed here were controlled by
`either one of two promoters of coliphage T5 [P525 or PD/E20
`(16)], and a lac operator sequence was inserted either
`between the DHFR and the CAT sequence (construct A) or
`into the BamHI site near the promoter (construct B in Fig. 1).
`Since the DHFR sequence is not in-frame with any transla-
`tional start site, the only protein expected from this expres-
`sion unit is CAT.
`With construct A (in pDS1), no CAT synthesis is observed
`in E. coli cells containing high levels of [ac repressor (Fig. 2).
`However, CAT production is rapidly induced to a high level
`by IPTG, as expected with‘ these promoters (16). This
`experiment shows that the operator-bound lac repressor can
`efliciently interfere with ongoing transcription. It raises the
`question whether repressor merely blocks the transcribing
`enzyme or causes a true termination event. Analysis of in
`vitro and in vivo transcripts shows that the lac repres-
`sor—operator complex acts as a transcription terminator. In
`the absence of repressor, or in the presence of repressor and
`IPTG, transcripts of around 1700 nucleotides are the major
`products in vitro (Fig. 3A, lanes 1, 4, and 5). In contrast, when
`active repressor is included in the transcription assay, the
`vast majority of transcripts are terminated at three distinct
`positions (a, b, and c in Fig. 3A), yielding RNAs about 750
`nucleotides long. Of these, only the smallest species can be
`converted into larger products (lanes 6 and 7). The others are
`not affected by prolonged incubation with unlabeled
`nucleoside triphosphates. When IPTG is added together with
`unlabeled nucleoside triphosphates, no increase in radioac-
`tivity is found in the 1700-nucleotide RNA species (data not
`shown). These results show that the repressor does not
`simply induce transcribing RNA polymerase to pause but
`rather triggers an active process of termination. The results
`
`
`
`21-
`
`M
`
`3
`2
`1
`0.5
`0
`Time of induction, hr
`
`Frc. 2. Effect of operator insertion on CAT synthesis in vivo. A
`pDS1 plasmid carrying construct A with PD/E20 as promoter was used
`to transform into E. coli cells carrying the compatible plasmid
`pDM1.1. The latter plasmid contains the lac!“ gene and profides high
`intracellular levels of lac repressor. Cultures of the transformed cells
`were grown to OD“, 0.7 before IPTG (200 pg/ml) was added.
`Aliquots of the culture were removed at times indicated and the
`pattern of the total cellular protein was monitored by NaDo_dSO./
`PAGE. The Coomassie blue-stained gel shows that the CAT protein,
`not visible at the time of IPTG addition,
`is the most prominent
`product after only 30 min. The size markers (lane M) are given in kDa
`at left.
`
`of equivalent experiments carried out in vivo are shown in
`Fig. 3B. Again, in the absence of operator or in the presence
`of IPTG, the major plasmid-specified RNA is about 1700
`nucleotides long (lanes 1 and 3). In cells containing high levels
`of repressor, however, two short transcripts of about 750
`nucleotides are synthesized (lane 2). Lanes 3 and 4 of Fig. 3B
`show an additional RNA species of about 820 nucleotides
`(labeled x). This transcript is only observed in the presence
`of the operator-carrying fragment and when transcription is
`allowed to proceed past the operator either by addition of
`IPTG (lane 3) or by limiting amounts of intracellular repressor
`(lane 4), suggesting that an additional sequence acting as a
`terminator in viva must be located downstream of the 60-bp
`operator fragment.
`'
`By quantifying DHFR- and CAT-specific RNA (refs. 20
`and 21; unpublished work) we find as much as 90% termi-
`nation in vivo (Table 1). This termination can be completely
`reversed by IPTG. Our data also indicate that the repressor-
`independent termination at position x (Fig. 3B) is *--17%
`eficient (data not shown).
`'
`Topography of the lac Repressor—Operator Termination
`Signal. Where does an operator-bound repressor force the
`transcribing RNA polymerase to stop and to release the
`nascent transcript? To answer this question, we used con-
`struct A (Fig. 1), containing promoter P525,
`to produce
`transcripts in the presence or absence‘ of [ac repressor, and
`the 3’ ends of the RNAs terminated around the operator
`sequence were characterized by nuclease S1—mapping with
`3’-labeled DNA fragments (Fig 3C). When repressor is
`bound to the operator, transcription is terminated in vivo and
`in vitro at two sites upstream of the operator sequence (Fig.
`4). The termination site observed ‘in ‘vivo when repressor is
`limiting or inactive has been mapped outside of but adjacent
`to the cloned operator fragment. In Fig. 4, the different sites
`are indicated by hatched columns. It appears most likely to
`us that the repressor terminates transcription at precise
`positions and that the regions of 3-5 nucleotides derived from
`S1-mapping experiments primarily reflect a heterogeneity of
`the S1 digest. The homogeneoustranscript obtained in vitro
`
`

`
`4136
`
`Biochemistry: Deuschlc er al.
`
`Proc. Natl. Acad. Sci. USA 83 (1986)
`
`in vivo
`
`in vilro
`
`D
`
`In VITFO
`R
`R
`
` ‘l23LS678
`
`A
`
`FIG. 3. Analysis of RNA synthesized in the presence or absence of lac repressor. (A) In vitro transcripts obtained from construct A in the
`absence (lane 1) or presence of 2 pg (lanes 2, 4, and 6) or 5 pg (lanes 3, 5, and 7) of purified lac repressor (R) per assay were analyzed by PAGE.
`Assays in which the repressor was inactivated by IPTG are indicated (1). The transcripts seen around position 1700 are terminated at to (see
`Fig. 1). In the presence of functional repressor, three shorter species of RNA are identified (a, b, and c), of which c can apparently be “chased”
`upon addition of an excess of unlabeled UTP and 10 min further incubation (lanes 6 and 7). Addition of IPTG to the transcription assay completely
`abolishes tennination (lanes 4 and 5). Markers (sizes in nucleotides at left) are a digest of pDS1' (15) with BamHI, Psi I, and Xba I. (B) In vivo
`RNA specified by our transcription unit can be visualized directly due to the high efficiency of the promoters utilized. Lane 2 shows the RNA
`pattern from cells containing high levels of active repressor. The two RNA species visible resemble in size the in vitro transcripts terminated
`at sites a and b. Both species disappear if IPTG (I) is present in the culture (lane 3) or if lac repressor-producing plasmid pDM1.l is absent (lane
`4). The majority of transcripts synthesized under these latter conditions are 1700 nucleotides long and comigrate in these gel systems with rRNA.
`A new class of RNA (x) is visible in lanes 3 and 4. This transcript, which is 820 nucleotides long, is not present when the operator is deleted
`at the HindIII site (lane 1). Its termination is repressor-independent but requires the presence of the operator-carrying fragment. Markers are
`as in A but mixed with Hae III-cleaved pBR322. (C) Nuclease S1-mapping (19) of the 3' end of in vivo and in vitro transcripts. Lanes 3 and 4
`show the S1-resistant material obtained with in vivo RNA in the absence (3) or presence (4) of repressor (R). Lane 5 shows the efi‘ect of IPTG
`(1). Similarly, lanes 6 and 7 contain probes of in vitro RNA synthesized in the absence and presence of repressor, respectively. The positions
`of the 3’ termini of the various RNAs (a, b, and x) were determined by inference with size markers (M): a labeled Hae III digest of pBR322
`(lane 1) and the G+A sequencing pattern of the 318-bp Acc I-Pvu II fragment labeled at the 3’ end. (D) Precision of lac repressor-induced
`termination. Construct B (Fig. 1) with promoter Puma was used to produce short transcripts (=140 nucleotides) in the presence of repressor.
`Lanes 2 and 3 show these transcripts, whereas lane 1 contains the repressor-free control. Comparing the width of the bands in lanes 3 and 4
`with those of the markers (M, Hae III digest of pBR322) suggests a precise termination (within 1-2 nucleotides). All gels contained 8 M urea
`and were 4% (A and B) or 8% (C and D) polyacrylainide. Size markers are denoted with M and given in nucleotides.
`
`when construct B is used as template is in support of this (Fig.
`3D).
`
`Table 1. Efficiency of transcriptional termination by the lac
`operator—repressor complex
`
`Promoter
`
`Repressor
`
`IPTG
`
`P025
`
`PD/E20
`
`+
`
`+
`
`+
`+
`
`"
`
`+
`
`_
`'0'
`
`Labeled RNA,
`cpm
`T‘ *
`DHFR
`CAT
`E’ %
`14,764
`1,652
`89
`14,361
`1,628
`89
`21,593
`16,931
`22
`20,748
`16,622
`20
`20.618
`2,094
`90
`20,097
`2,055
`90
`27,356
`23,199
`16
`26,750
`25,140
`10
`
`E_ C0,; cells hubon-ng P1351 (carrying construct A) and PDMLI
`(for repressor production) were grown to an OD“, of =0.4 before the
`cultures were divided and IPTG given to one of them. After further
`incubation at 37°C for 30 min, RNA of both cultures was labeled with
`PI-Iluridine, extracted, and quantified as described in Materials and
`Methods. For both promoters, P575 and Pmm, <10% of transcripts
`are CAT-specific when IPTG is absent. Upon induction, the CAT-
`coding region is expressed, though not with the same efiiciency as the
`DHFR sequence. This difference of ==17% is due to the termination
`signal identified at position x (Fig. 4).
`“Termination efiiciency.
`
`DISCUSSION
`
`The data presented above show that the complex between lac
`repressor and operator can efficiently halt transcribing RNA
`polymerase and cause the release of nascent RNA. The two
`sites where RNA synthesis is intemipted both lie upstream of
`the operator sequence. Of these, the major site utilized in vivo
`and in vitro (site a in Fig. 4) immediately borders the operator
`sequence, indicating that the active center of the transcrip-
`tional elongation complex can move very close to the
`hindering repressor—operator complex. This suggests that, in
`contrast to the promoter-bound enzyme, the transcribing
`RNA polymerase barely extends in front of its catalytic site.
`The second site (b in Fig. 4), where release of RNA occurs,
`is 10 bp upstream of site a. In vivo, both sites are utilized with
`about the same frequency, whereas site a is the preferred one
`in vitro. The intracellular concentration of repressor may
`have an effect on this phenomenon. The weak termination
`signal identified at site x is most likely created by integrating
`the operator sequence into this particular environment, since
`it occurs about 15 bp outside of the inserted fragment and at
`a distance 45 bp from the center of the operator sequence.
`Although several lines of evidence have indicated that an
`operator-bound lac repressor may interfere with ongoing
`transcription (9, 12),
`the view that a transcribing RNA
`polymerase would “peel off’ ’ such DNA-bound proteins was
`generally accepted. This was also suggested by in vitro data
`(12) that showed that, in the presence of lac repressor and
`RNA polymerase, the lac UV5 promoter/operator sequence
`
`

`
`Biochemistry: Deuschle et al.
`
`Proc. Natl. Acad. Sci. USA 83 (1986)
`
`4137
`
`CCGGCCAAGCTTGGCGGCTCGTATGTTGTGTGGAATTGTGAGCGGATAACAATTTCACACAGGAAACAGCCAAGCTTGGCGAGATTTTCAGGAGCTAAGGA
`
`91
`
`b
`
`u
`
`X
`
`FIG. 4. Sequences involved in repressor-induced transcriptional termination. The central region of the operator sufficient to bind repressor
`(1, 22) is boxed, and the inverted repeat of the sequence is delineated by arrows. The G in the center of the operator sequence has been used
`to define position 0. The sites where transcription is terminated are indicated by the hatched columns. The width of the columns reflects the
`heterogeneity of the S1-resistant material and the height of the columns represents the relative frequency of termination at the respective site.
`The columns above and below the sequence describe the in vitro and in vivo results, respectively. Sites a, b, and x correspond to the designations
`used in Fig. 3. In the presence of repressor, transcripts are terminated upstream of the operator sequence. Termination at site x occurs outside
`of the original operator fragment. The HindIII cleavage sites used to insert the 54-bp fragment between the DHFR and the CAT sequence yielding
`construct A (Fig. 1) are underlined.
`
`can only transiently block a transcriptional elongation com-
`plex. By contrast, our data demonstrate that a lac repres-
`sor-operator complex located distal to a promoter sequence
`can directly interfere with gene expression by efficiently
`terminating transcription. This sheds new light on the pos-
`sible role of operators found outside of the primary regulatory
`region. Thus, operator/repressor systems could have func-
`tions in addition to the one commonly considered—namely,
`(i) to prevent readthrough from upstream regions into the
`repressed operon and (ii) to establish a polarity pattern within
`an operon that is dependent on the level of inducer and the
`affinity between a particular operator sequence and a repres-
`sor.
`
`These properties could play a role in the fine tuning of gene
`expression at the transcriptional level and may be considered
`as a type of attenuation. Systems to examine this hypothesis
`could be the gal as well as the lac operon (8, 23, 24). Both
`operons contain a second operator sequence about 50 and 400
`bp downstream of the RNA initiation site, respectively. In
`the gal operon, operator 2, which is located within the
`structural gene galE, may affect transcription from both
`promoters P1 and P2, but primarily from P2 by attenuation,
`whereas operator 1 is the main control element of P1. Finally,
`Sellitti and Steege (25) reported that transcription from the
`lac! promoter is “punctuated” within the lac control region
`and that this punctuation is strongly influenced by the lac
`repressor. These data suggest that the mechanism of action
`for repressor/operator systems proposed here is utilized in
`the E. coli lac system.
`
`We thank M. Lanzer for supplying purified lac repressor and
`pDM1.1; W. Kammerer, M. Lanzer, and D. Stueber for helpful
`discussions; and J. Scaife and S. Le Gtice for critical comments and
`reading of the manuscript. The help of Y. Kohlbrenner in preparing
`the manuscript is gratefully acknowledged.
`
`1. Gilbert, W. & Maxam, A. (1973) Proc. Natl. Acad. Sci. USA
`70, 3581-3584.
`2. Gunsalus, R. P. & Yanofsky, C. (1980) Proc. Natl. Acad. Sci.
`USA 77, 7117-7121.
`3. Adhya, S. & Miller, W. (1979) Nature (London) 279, 492-608.
`
`Shanblatt, S. H. & Revzin, A. (1983) Proc. Natl. Acad. Sci.
`USA 80, 1594-1598.
`Reznikofi, W. S. (1972) in RNA Polymerase, eds. Losick, R.
`& Chamberlin, M. (Cold Spring Harbor Laboratory, Cold
`Spring Harbor, NY), pp. 441-454.
`Majors, J. (1975) Proc. Natl. Acad. Sci. USA 72, 4394-4399.
`Reznikofi, W. S. & Abelson, J. M. (1980) in The Operon, eds.
`Miller, J. H. & Reznikoff, W. S. (Cold Spring Harbor Labo-
`ratory, Cold Spring Harbor, NY), pp. 221-243.
`Gilbert, W., Gralla,
`.l., Majors, J. & Maxam, A. (1975) in
`Protein-Ligand Interactions, eds. Sund, H. & Blauer, G. (de
`Gruyter, Berlin), pp. 193-210.
`Irani, M. H., Orosz, L. & Adhya, S. (1983) Cell 32, 783-788.
`Mitchell, D. H., Reznikofl‘, W. S. & Beckwith, J. (1975) J.
`Mol. Biol. 93, 331-350.
`Reznikoff, W. S., Miller, J. H., Scaife, J. G. & Beckwith,
`J. R. (1969) J. Mol. Biol. 43, 201-213.
`Horowitz, H. & Platt, T.
`(1982) Nucleic Acids Res. 10,
`5447-5465.
`Herrin, G. L. & Bennet, G. N. (1984) Gene 32, 349-356.
`Gentz, R., Langner, A., Chang, A. C. Y., Cohen, S. N. &
`Bujard, H. (1981) Proc. Natl. Acad. Sci. USA 78, 4936-4940.
`Stueber, D. & Bujard, H. (1982) EMBO J. 1, 1399-1404.
`Gentz, R. & Bujard, H. (1985) J. Bacteriol. 164, 70-77.
`Miller, J. H. (1972) Experiments in Molecular Genetics (Cold
`Spring Harbor Laboratory, Cold Spring Harbor, NY), pp.
`431-435.
`Glisin, V., Crkvenjakov, R. & Byus, G. (1974) Biochemistry
`13, 2633-2637.
`Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular
`Cloning: A Laboratory Manual (Cold Spring Harbor Labora-
`tory, Cold Spring Harbor, NY).
`Deuschle, U.
`(1986) Dissertation (Univ. of Heidelberg,
`Heidelberg).
`Bujard, H., Deuschle, U., Kammerer, W., Gentz, R., Bann-
`warth, W. & Stueber, D. (1985) in Sequence Specificity in
`Transcription and Translation: UCLA Symposia, eds. Calen-
`dar, R. & Gold, L. (Liss, New York), pp. 21-29.
`von Hippel, P. H. (1979) in Biological Regulation and Devel-
`opment, ed. Goldberger, R. E. (Plenum, New York), Vol. 1,
`pp. 279-347.
`De Crombrugghe, B., Busby, S. & Buc, H. (1984) Science 223,
`831-838.
`Majumdar, A. & Adhya, S. (1984) Proc. Natl. Acad. Sci. USA
`81, 6100-6104.
`Sellitti, M. A. & Steege, D. A. (1985) J. Cell. Biochem. 85, 205.
`
`.".°‘
`
`11.
`
`12.
`
`13.
`14.
`
`15.
`16.
`17.
`
`18.
`
`19.
`
`20.
`
`21.
`
`22.
`
`23 .
`
`24.
`
`25.