JOURNAL or BACTERIOLOGY, Oct. 1985, p. 70-77
`0021-9193/85/100070-08$02.00/0
`Copyright © 1985, American Society for Microbiology
`
`Vol. 164, No. 1
`
`Promoters Recognized by Escherichia coli RNA Polymerase
`Selected by Function: Highly Eflicient Promoters from
`Bacteriophage T5
`REINER GENTZT AND HERMANN BUJARD1‘*
`
`Molekulare Genetik der Universitdt Heidelberg, D-6900 Heidelberg, Federal Republic of Germany
`
`Received 6 March 1985/Accepted 24 June 1985
`
`Highly efficient promoters of coliphage T5 were identified by selecting for functional properties. Eleven such
`promoters belonging to all three expression classes of the phage were analyzed. Their average AT content was
`75% and reached 83% in subregions of the sequences. Besides the well-known conserved sequences around
`- 10 and -33, they exhibited homologies outside the region commonly considered to be essential for promoter
`function. Interestingly, the consensus hexamers around -10 (TAT AAT) and -35 (TTG ACA) were never
`found simultaneously within the sequence of highly efficient promoters. Several of these promoters compete
`extremely well for Escherichia coli RNA polymerase and can be used for the efficient in vitro synthesis of defined
`RNA species. In addition, some of these promoters accept 7-mGpppA as the starting dinucleotide, thus
`producing capped mRNA in vitro which can be utilized in various eucaryotic translation systems.
`
`Promoters of the Escherichia coli system start synthesis of
`functional RNAs with vastly dilferent efficiencies. Little is
`known, however, about the rules by which functional pa-
`rameters are implemented within a promoter sequence.
`Despite our knowledge of more than 150 promoter se-
`quences (10) and a wealth of genetic and biochemical data
`(19), we are still unable to make reasonable predictions on
`functional properties of a promoter from structural informa-
`tion alone. Consensus sequences of E. coli promoters de-
`rived from sequence compilations. have elucidated some
`important general features. However, synthesis of consen-
`sus promoters (5) have resulted in signals which are, at most,
`average in function (U. Deuschle and M. Kammerer, per-
`sonal communication). This is not surprising if one considers
`the complexity of the process programmed by a promoter
`sequence as well as the fact
`that
`in the derivation of
`consensus sequences there is usually no value describing
`functional parameters given to individual sequences.
`We approached this problem in a dilferent way. By
`selecting for the most efficient unregulated promoters in the
`E. coli system, we expected to reveal sequences which
`would exhibit pertinent structural features most clearly. The
`selection principles utilized for the identification of eflicient
`promoters were the determination of (i) the rate of complex
`formation between RNA polymerase and promoter in vitro,
`(ii) the relative efliciency of RNA synthesis in vitro under
`competitive conditions, and (iii)
`the relative promoter
`strength in vivo.
`The in vitro analysis of promoter-carrying DNA fragments
`has been described previously (6, 7). For the in vivo study of
`promoters we developed cloning systems which allow the
`stable integration of strong promoters as well as the precise
`determination of their in vivo function (9, 21; U. Deuschle,
`M.S. thesis, University of Heidelberg, 1984). Of about 60
`promoters tested (including those of coliphage T7, fd, and A)
`some of the most efficient signals were found in the genome
`of coliphage T5. Here we describe the application of the
`
`* Corresponding author.
`1‘ Present address: F. Holfmann—La Roche & Co. A.G., ZFE CH
`4002 Basel, Switzerland.
`
`70
`
`pDS1 vector system (21; Fig. 1) for the selective cloning of
`strong promoters, the identification and structural analysis
`of 11 promoters of the phage T5 genome, and some of the
`functional properties of these promoters. As can be seen
`from the results of this and previous studies (9), several
`promoters described here appear especially useful for the
`efficient in vitro synthesis of defined RNA species, and as
`some of the promoters accept 7-mGpppA as the starting
`dinucleotide capped RNAs can be directly obtained in vitro.
`This transcription-coupled capping allows an eflicient and
`selective expression of cloned DNA sequences in vitro
`which has been found to be especially useful in studying the
`translocation of proteins into or through membranes (11, 23).
`
`MATERIALS AND METHODS
`
`Enzymes and chemicals. Restriction enzymes, T4 DNA
`ligase, calf intestinal alkaline phosphatase, and RNase T1
`were purchased from Bethesda Research Laboratories,
`Gaithersburg, Md.; New England Biolabs, Inc., Beverly,
`Mass. ; or Boehringer Mannheim Biochemicals, Indianapolis,
`Ind.; and T4 DNA kinase was obtained from H. Schaller
`(University of Heidelberg). Reactions were carried out as
`recommended by the supplier. The isolation of bacteriophage
`T5 DNA and E. coli RNA polymerase has been described
`previously (7). Xhol synthetic linkers were obtained from
`Collaborative Research, Inc., (Waltham, Mass.) and were
`present in ligation assays in a 20-fold molar excess relative to
`that of the various DNA fragments. [-y-32P]ATP and [oi-32F]
`UTP were from Amersham & Buchler (Braunschweig,
`Federal Republic of Germany) and 7-mGpppA was obtained
`from P-L Biochemicals, Milwaukee, Wis.
`Plasmids and their nomenclature. The basic pDS1 vector
`system has been described previously, and here we fol-
`low previously proposed nomenclature (21). The identity of
`the promoters and terminators which have been integrated
`can be derived from the designation of the plasmid: pDS1/
`pH207,to1 describes a plasmid-carrying promoter pH2o7 in
`front of the coding sequence (dhfr) for dihydrofolate
`reductase (DHFR) and terminator to from phage lambda at
`site 1 (Fig. 1). Another terminator used was tfd from coli-
`phage fd (9). For the correct in-frame positioning of the
`Mylan v. Genentech
`Mylan V. Genentech
`IPR2016-00710
`Genentech Exhibit 2078
`
`Genentech Exhibit 2078
`
`IPR2016-00710
`
`

`
`VOL. 164, 1985
`
`PROMOTERS FROM BACTERIOPHAGE T5
`
`71
`
`translational start sites within cloned promoter-carrying
`fragments with respect to the dhfr sequence, BamHI linkers
`of 8, 10, and 12 base pairs (bp) in length were inserted into
`pDS1. The length of the BamHI linker relating to the reading
`frame can also be derived from the designation of a plasmid.
`pDS1-8/pH207,t01 denotes the same plasmid as described
`above which carries a BamHI linker of 8 bp.
`Cloning of promoter-carrying T5 DNA fragments. Wild-
`type T5 DNA was digested with Hindlll and separated by
`polyacrylamide gel electrophoresis (PAGE; 6%, 14 h, 5
`V/cm, TBE buffer [90 mM Tris—hydrochloride, 90 mM so-
`dium borate, 3 mM EDTA, pH 8.3]), and the fragments A
`through H, IJ, KL, M, and N (8) were isolated by electroelu-
`tion (1 h, 150 V, buffer [5 mM Tris—hydrochloride, 0.5 mM
`EDTA, pH 7.6]). After the fragments were concentrated by
`ethanol precipitation, they were digested with either Haelll
`or Alul and Rsal and then fused to synthetic Xhol linkers by
`ligation. Upon digestion with Xhol,
`the fragments were
`cloned into pDS1 vectors by standard procedures. Transfor-
`mation of E. coli M15 (9) was carried out as described by
`Morrison (17). Selection of transformants was for a high
`level of resistance to chloramphenicol (CM) (100 to 400
`p.g/ml). Selected colonies were analyzed by isolating their
`plasmid DNA from 1.5-ml cultures (1) and by examining the
`restriction patterns of the DNAs by PAGE; fractions of the
`culture were subjected to sodium dodecyl sulfate (SDS)—
`PAGE, and the presence of chloramphenicol acetyl transfer-
`ase (CAT) was monitored after the protein pattern was
`visualized with Coomassie blue. Plasmids were isolated in
`preparative amounts by the method of Clewell and Helinski
`(3) and Radlofi‘ et al. (18).
`In vitro transcription. Standard in vitro transcriptions
`were carried out in a volume of 100 pl containing 20 mM
`Tris—hydrochloride (pH 8), 10 mM MgCl2, 0.1 mM EDTA,
`120 mM KCl, 5% glycerol, 2 mM dithioerythritol, 400 p.M
`ATP, 400 |.LM GTP, 200 p.M CTP, 50 or 100 pM UTP, and 0.5
`|LM [oi-32P]UTP (3,000 Ci/mmol). For the synthesis of 5’-
`labeled RNA the concentration of ATP was lowered to 12
`
`jLM, i.e., 9 p.M unlabeled and 3 j.tM labeled ATP [-y-32P]ATP;
`23,000 Ci/mmol). The concentration of UTP in these assays
`was 150 p.M. The reactions were initiated by the addition of
`approximately 0.5 U of RNA polymerase per pmol of DNA
`(corresponding to five enzyme molecules per plasmid) and
`terminated by the addition of EDTA (final concentration, 40
`mM) and phenol. Subsequent
`to phenol extraction,
`the
`nonincorporated nucleotides were separated by chromatog-
`raphy on Sephadex G57 (2-ml column); and the RNA-
`containing fractions were collected, precipitated by ethanol
`suspended in TE buffer, and, after the addition of three
`volumes of 90% formamide, heated to 90°C for 30 s and
`analyzed by PAGE in 7 M urea (8% 20 V/cm). Transcription-
`coupled capping was achieved by lowering the concentration
`of ATP in the standard assay to 5 p.M and by including 250
`p.M 7-mGpppA. To demonstrate the incorporation of the
`dinucleotide, short transcripts were produced by replacing
`CTP with methyl-CTP (200 p.M final concentration). After
`the assays were incubated for 5 min at 37°C, the concentra-
`tion of ATP was raised to 1 mM and the reaction was allowed
`to proceed for another 10 min. The controls contained 400
`p.M ATP and no 7-mGpppA.
`Analysis of in vitro transcri ts with RNase T1. In vitro
`transcripts end labeled with ['y- 2P]ATP were eluted from 8%
`polyacrylamide gels containing 7 M urea. After precipitation
`with ethanol,
`the transcripts were suspended in a buffer
`containing 25 mM sodium citrate (pH 5.0), 7 M urea, and 1
`mM EDTA (final volume, 5 id). Upon the addition of 1 pg of
`
`Xhol E‘°
`
`R] Barnl-ll
`
` 100 hp
`\RsuI
`
`
`
`bla
`
`lird Ill
`
`Site 1
`
`Xbui
`
`Site 2
`
`FIG. 1. pDS1 cloning system. The principle of this vector system
`has been described previously (21). The plasmid contains the ColE1
`replicon (ori; arrows indicate the direction of transcription of RNA
`I and RNA 11; 12) and, as a selectable marker, ampicillin resistance
`(conferred by the bla gene). Two indicator functions, encoded by
`the sequence for dhfr and cat, are fused in such a way that they can
`be brought under the control of one promoter (P) but can also be
`separated by inserting a terminator (T) at site 1. A second tenninator
`at site 2 can prevent transcriptional readthrough into the replication
`region. Whereas the cat gene contains its genuine ribosomal binding
`site, the dhfr sequence is not preceeded by such a signal. The
`BamHI site was created by inserting synthetic linkers of 8, 10, and
`12 bp. The resulting plasmids allow the in-frame positioning of the
`dhfr sequence to any upstream translational initiation site. The first
`Rsal cleavage site within the dhfr sequence which was utilized for
`the production of runofl" transcripts is indicated. The nomericlature
`used with this vector system is described in the text.
`
`tRNA per assay the transcripts were incubated with 1 U of
`T1 RNase for 15 min at 55°C. The reactions were stopped by
`freezing the samples in dry ice and then storing at —20°C.
`Immediately before the gel was loaded, the samples were
`heated to 90°C for 30 s and chilled on ice. For standardiza-
`tion of the electrophoretic pattern, fractions of the various
`transcripts were subjected to limited alkaline hydrolysis. A
`typical assay contained the transcript together with 5 pg of
`tRNA in 50 mM sodium bicarbonate-carbonate (pH 9.2;
`volume, 10 iii). The samples were incubated at 90°C for 7
`min and then chilled on ice. Just prior to loading on a
`sequencing gel, an equal volume (10 pl) of 10 M urea—1.5
`mM EDTA containing 0.05% each of xylene cyanol and
`bromphenol blue was added, and the samples were heated
`for 30 s at 90°C and chilled on ice.
`DNA sequencing. DNA was sequenced by the method of
`Maxam and Gilbert (14, 15). Fragments were dephosphoryl-
`ated (calf intestinal alkaline phosphatase), labeled at the 5’
`termini with T4 polynucleotide kinase and [-y-32P]ATP, and
`then subjected to secondary restriction endonuclease diges-
`tion to generate fragments labeled at one end only. Separa-
`tion of the fragment by PAGE and elution of the radioactive
`material from the gels resulted in end-labeled DNA, the
`
`

`
`72
`
`GENTZ AND BUJARD
`
`J. BACTERIOL.
`
`M 1
`
`23456789101112
`
`925oo~ --ma-+::;;1arf.::::g
`2-
`66200 '-
`
`ASOOO -
`
`31000 *
`
`21500 *
`
`14400 —
`
`
`
`FIG. 2. Analysis of total protein from E. coli M15 cells harboring T5 promoter-carrying pDS1 plasmids. DNA fragments identified in the
`first cloning experiments were reintegrated into plasmids pDSl- (8,10,12)/102 (Fig. 1). Colonies highly resistant to CM were grown in LB
`medium, and fractions of the culture were analyzed by SDS-PAGE as described previously (21). The positions of CAT, TU (elongation factor
`of translation), and a fusion product of DHFR are indicated. Molecular weights are shown on the left. Lanes M and 1 show a size marker and
`the pattern of plasmid-free M15 cells, respectively, whereas lanes 2 through 11 exhibit the proteins from clones canying the following T5 DNA
`fragments or promoters, respectively: F20, F33, PHZQ7, pms, F81 (carrying pN2(,), F30, F41, F22, F5, F25*. Lane 12 shows the effect of
`positioning the dhfr sequence in-frame with a translational stan signal located in fragment F25*. Excessive amounts of a DHFR fusion protein
`were produced. Although in this particular case the fusion product is obscuring the CAT protein, the presence of both proteins has been
`demonstrated (see text).
`
`sequence of which was determined by standard procedures.
`In all cases both strands were sequenced.
`Analysis of cloned DNA fragments by Southern hybridiza-
`tion. Wild-type T5 DNA, digested with Hindlll, was sepa-
`rated on a 0.5% agarose gel (14 h, 3 V/cm, 1 X Loening buffer
`[36 mM Tris-hydrochloride, 30 mM NaH2PO4, 1 mM EDTA,
`pH 7.6]). After being stained with ethidium bromide, the gel
`was photographed together with a measuring tape. To ensure
`good transfer of large fragments from the gel, it was soaked
`for 1 min in 100 mM HCl, neutralized (two times for 10 min
`each in 0.5 M Tris-hydrochloride, pH 7.5), and then dena-
`tured in 0.5 M NaOH for 20 min. Upon neutralization (see
`above), the DNA was transferred from the gel to nitrocellu-
`lose filter paper (0.45 pm; Schleicher & Schuell, Inc.) by the
`method of Southern (20). The filters were marked to indicate
`the positions of the fragments, rinsed in 2X SSC (1>< SSC is
`0.15 M NaCl plus 15 mM sodium citrate), and baked in vacuo
`(2 h at 80°C). For hybridization, filters were pretreated for 1
`h at 68°C in hybridization bulfer (5>< SSC, 50% formamide,
`1>< Denhardt reagent [4]). Hybridizations were carried out in
`12-ml plastic vials containing 1 ml of hybridization buffer,
`approximately 105 cpm (Cerenkov) of the denatured DNA
`probe (1 min at 95°C in 0.5 x TE buffer), and the filter strip.
`Hybridizations were allowed to proceed at 42°C for 8 to 14 h
`on a rotary shaker. For some fragments (e.g., pm; and ppm)
`the conditions were optimized to reduce cross-hybridiza-
`tions (hybridization temperature, 42, 44, 46, or 48°C). Sub-
`sequently, the filters were washed three times at 42°C in 5X
`SSC buffer—50% formamide and finally twice in 2x SSC at
`room temperature. Filters were then air dried, and autora-
`diographs were prepared with Kodak XR-5 films.
`
`RESULTS
`
`Cloning of promoter-carrying fragments. For the cloning of
`the promoters of the different expression classes of
`
`bacteriophage T5, the DNA fragments of various digests of
`the 120-kilobase phage genome were fused to synthetic Xhol
`linkers and cloned into the Xhol site of pDS1-10/tf,,1 (Fig. 1;
`see above for nomenclature). Selection of transformed E.
`coli M15 was carried out on agar plates containing increasing
`amounts of CM, and colonies resistant to 100 pg/ml or more
`were grown and analyzed. All isolates showed at least one
`cloned fragment in the size range between 130 and 1000 bp.
`These fragments were isolated and recloned into pDS1-(8,
`10, and 12)/to] or pDS1-(8, 10, and 12)/2'02. All of the more
`than 20 fragments gave rise to colonies highly resistant to
`CM (i.e., 5100 pg/ml). Cultures prepared from such colonies
`(10 ml of Luria broth, 30 pg of CM per ml; 37°C overnight)
`were analyzed with respect
`to their protein pattern by
`SDS-PAGE. Considerable amounts of CAT protein were
`observed in all cases (Fig. 2, lanes 2 to 11). The effect of
`positioning a translational start signal of a promoter-carrying
`fragment in frame with the dhfr sequence is shown in Fig. 2,
`lane 12. The presence of the DHFR fusion protein obscuring
`the CAT protein (Fig. 2, lane 12) has been confirmed by
`dilution of the sample (thereby generating two separable
`bands), as well as by immunoblotting with DHFR- and
`CAT-specific antibodies (data not shown). In this way we
`not only identified efficient promoters but also interesting
`translational start signals which will be described elsewhere.
`In vitro transcription of the cloned fragments and analysis
`of the transcripts. To verify the presence of promoters within
`the cloned material,
`the various chimeric plasmids were
`digested with XhoI, and the liberated fragments were, after
`separation by PAGE and elution from the gel used as
`templates for in vitro transcription. Alternatively, the plas-
`mids were digested with Rsal, and the mixtures of fragments
`were transcribed, producing run-off transcripts from the
`cloned promoter to the first Rsal site within the dhfr se-
`quence (Fig. 1). At limiting amounts of RNA polymerase,
`
`

`
`VOL. 164, 1985
`
`PROMOTERS FROM BACTERIOPHAGE T5
`
`73
`
`the run-off transcripts were labeled with either ['y-32P]ATP or
`[a-32P]UTP. In all but one case (fragment F 22) transcripts of
`the same length were obtained from a given template inde-
`pendent of the method of labeling. No label was introduced
`into the RNA originating from the fragment F 22 if [y-
`32P]ATP was used. Consequently, this transcript does not
`start with an adenosine. Efficient transcription can be initi-
`ated, however, with the dinucleotide uridylyl-(3’,5’)uridine
`(data not shown); this promoter is therefore most likely an
`U—starter (see below). Two transcripts of different lengths
`were produced in comparable amounts from fragment F 41,
`indicating the presence of two promoters within this tem-
`plate separated by about 100 bp (Fig. 3). In all cases the
`RNA populations obtained originated almost exclusively at
`
`41 25*
`- M 6»
`
`1631
`
`~
`
`-~
`
`Q
`
`517 \_
`506 f
`
`396
`
`341+
`
`298
`
`—
`
`-
`
`'
`
`““
`
`....
`
`221 \_
`220 f
`
`__
`
`FIG. 3. In vitro transcription of chimeric plasmids. Plasmids of
`the type pDS1-8/p,to2 containing different promoter-carrying frag-
`ments were digested with Rsal and transcribed in vitro (molar ratio
`of RNA polymerase to plasmid, 5:1). The labeled transcripts were
`separated by 7 M urea PAGE (8%, 20 V/cm, 3 h) and visualized by
`autoradiography. As indicated,
`the different plasmids contained
`fragments F25 (25), F41 (41), and F25* (25*), respectively; M
`denotes markers whose sizes are given in nucleotides on the left. In
`each case the dominating transcripts (§95% of hybridizable counts)
`originated from the cloned promoters. The size of the transcripts
`(designated on the left) allows an estimation of the position of the
`promoter within the cloned fragment.
`
`ab
`
`c
`
`d
`
`-
`
`:33
`"'
`- 2
`
`532°
`"2:
`¢:.“l5
`C~-
`u.
`b-
`‘-10
`‘~-
`C-~
`§-
`
`u_
`
`0-
`
`I—
`
`'\-
`
`O1
`
`_
`
`-35
`33°
`
`53-20
`:2
`-"-'?-15
`:2
`C--
`’-
`C"-10
`on--
`__
`C—-
`
`_‘-_
`
`-_
`
`’_
`
`“
`
`FIG. 4. Examples for the sequence analysis of transcripts. Tran-
`scripts subjected to either partial RNase T1 digestion or to limited
`alkaline hydrolysis were analyzed by PAGE (20%, 8.3 M urea, 50
`V/cm, 2 h). Here the pattern originating from transcripts initiated by
`pxzga and p525 are shown in lanes a and c, respectively. The
`corresponding alkaline hydrolysates are in lanes b and d. Numbers
`on the right are distances to the 5’ end of the transcripts in
`nucleotides.
`
`the phage promoters (Fig. 3; see below), demonstrating their
`ability to favorably compete with the plasmid promoters for
`RNA polymerase.
`Eleven different promoters were characterized within the
`20 fragments originally cloned. Ten of these initiated their
`transcripts with adenosine and gave rise to relatively short
`RNAs. The lengths of these RNAs allowed us to estimate the
`positions of the promoters within the cloned fragments,
`which facilitated their further analysis.
`Sequence analysis of promoter-carrying fragments and their
`major transcripts. DNA sequence determinations were car-
`ried out by the method of Maxam and Gilbert (14, 15).
`Fragments of 500 bp and less in size were directly end
`labeled, and after cleavage with a second restriction endo-
`nuclease they were separated by electrophoresis and se-
`quenced. Large fragments were first digested with various
`restriction enzymes, and the promoter-carrying segments
`were determined by binding RNA polymerase by the nitro-
`cellulose filter binding technique as described previously
`(data not shown; 7). Fragments defined in this way were end
`labeled, sequenced, and recloned into the XhoI or EcoRI site
`of pDS1/tol (Fig. 1). In all cases the sequences for both
`strands were determined. The functional orientation of the
`various promoter sequences was confirmed by relating
`cleavage sites within the promoter-carrying fragments to the
`
`

`
`74
`
`GENTZ AND BUJARD
`
`J. BACTERIOL.
`
`PE
`
`PD/E 2o
`
`PD/E 33
`
`PH 207
`
`Pu 25
`
`PM 26
`
`5
`
`PF so
`
`PK 28a
`
`PK 28b
`
`PH 22
`
`RNA START
`
`TACCf&TTCTGAGA GATAAC
`
`CTTTCATAAATT TGAGA AAGC
`
`‘TCGATTTAGGCAGT..,........
`
`TTAG‘®CCTAATGGATCGACCTT
`
`
`
`PG 25
`FIG. 5. Nucleotide sequences of T5 promoters. Regions of major sequence homology are boxed, and the eight fully conserved positions
`are indicated by filled circles. The AT-rich regions centered around -43 are underlined, and the starting nucleotide of the RNA is circled.
`Since pm; can initiate transcription with uridylyl-(3’,5’)un'dine (data not shown), the starting nucleotide of its RNA most likely must be
`positioned within the T'I"I‘ sequence. Abbreviations: PE, preearly; E, early; L, late.
`
`‘ TTGT'®'TAAAGAGGAGAAATTAAC
`
`unique BamHI site in the pDS1 plasmid (data not shown).
`The precise start point of transcription within a promoter
`sequence can be determined by comparing the DNA se-
`quence coding for the 5’ terminal region of the transcript
`with the sequence of the corresponding RNA. We have
`therefore analyzed the G-pattem of transcripts labeled in
`vitro with [y-32P]ATP isolated from polyacrylamide gels.
`Examples for such analyses, which were carried out for all
`but two transcripts, are shown in Fig. 4. A compilation of
`promoter sequences characterized in this way, Fig. 5.
`Mapping of the cloned promoters along the T5 genome.
`Forty-one major promoters of the T5 genome previously
`have been mapped with respect to their location within the
`dilferent expression classes (preearly, early, late), as well as
`in relation to various restriction maps (22). Since the distri-
`bution of Hindlll cleavage sites throughout the T5 DNA
`permits a rather good distinction between preearly, early,
`and late regions of the genome (8), we used Hindlll digests
`of T5 wild-type DNA which were separated by PAGE and
`transferred to nitrocellulose for an analysis of various pro-
`moter-carrying fragments by the method of Southern (20).
`The “P-end-labeled fragments were hybridized to the immo-
`bilized T5 DNA under stringent conditions. Two fragments,
`F20 and F33 hybridized to both HindIII-D and —E (Fig. 6).
`Since these two HindIII fragments contain the terminally
`redundant preearly region, the promoters located in F20 and
`F33 must belong to that expression class. With the exception
`of F22,
`the remaining fragments hybridized to just one
`fragment from the HindIII digest of T5 DNA and can
`therefore be attributed rather precisely to their respective
`expression classes. Fragment F22 associated, even under
`the most stringent conditions, to HindIII-B, -D/E, -G, and
`-H. Whereas the strongest signal was clearly obtained with
`HindIII-H, considerable sequence homology must exist in
`the other regions as well. In conclusion, we identified and
`
`sequenced 11 promoters of coliphage T5. Two of these
`(PD/E20 and PD/E33) belong I0 the PTCC3-1'13’, SCVCH ([71:30, PH2o7.
`PN25, PN26. PK2_sa, pk28ba and PH22) b31008 t0_th¢ early, and
`two (p15 and p525) belong to the late expression classes.
`The nomenclature used for the cloned T5 promoters is as
`follows. The letter described the Hindlll fragment within
`which the promoter is located and the number defines the
`cloned fragment; e.g., the promoter of fragment F5 is p15
`since F5 hybridizes to HindIII fragment J . Exceptions are
`the promoters within the HindIII fragments N and K, which
`have been identified and named previously (2, 22).
`RNA synthesis in vitro initiated by pms and ppm. Promoters
`of coliphage T5 compete efficiently for RNA polymerase
`binding in the presence of other promoters (7). Some of them
`not only bind the enzyme highly efiiciently, but they also
`outcompete other promoters in directing RNA synthesis in
`vivo and in vitro (6, 7, 9). Such promoters are therefore good
`candidates for in vitro synthesis of defined RNAs in prepar-
`ative amounts.
`In Fig. 7 we show the kinetics of the
`production of runoff transcripts initiated from pm; and PNZ6.
`It can be seen that RNA is synthesized at a high rate for
`more than 30 min and that between 12 and 25 pmol of RNA
`per pmol of template can be obtained during this time,
`depending on the size of the transcript. More than 90% of the
`RNA synthesized consists of the expected runofi‘ transcripts,
`and more than 95% of the total RNA is complementary to the
`coding strand. The larger transcripts seen in the upper insert
`of Fig. 7 were not analyzed in detail, but we suspect that
`they originate from incompletely digested templates.
`Transcription-coupled capping in vitro. Various promoters
`were tested for their ability to initiate in vitro transcription
`with 7-mGpppA. Successful incorporation can be monitored
`either by translating the resulting RNA in an eucaryotic
`translation system (23) or by analyzing short transcripts by
`urea PAGE. Short transcripts were obtained by replacing
`
`

`
`VOL. 164, 1985
`
`PROMOTERS FROM BACTERIOPHAGE T5
`
`75
`
`cpm
`X103
`
`f 200
`
`L
`
`1'50
`
`— 100
`
`so
`
`u ta
`
`c d e f
`
`20
`
`Ua
`
`,3 15
`
`E
`5ZO.
`
`E 1
`
`% 10
`§
`62EL
`
`5
`
`CTP with methyl-CTP, which stops transcription usually at
`the position of the first and second cytosine, resulting in two
`major RNA species. Transcripts obtained in this way from
`promoters pmga, plug, and pms are shown in Fig. 8. The
`length of the major RNA species was between 15 and 41
`nucleotides, and a shift in migration corresponding to about
`2 nucleotides was introduced by the incorporation of 7-
`mGpppA. The three promoters described here were found to
`be among the most efficient ones for the in vitro synthesis of
`capped RNAs.
`
`DISCUSSION
`
`The family of promoters described in this report was
`discovered previously by examining specific in vitro proper-
`ties of a variety of promoters (6, 7). The functions monitored
`were the relative rate of complex formation between E. coli
`RNA polymerase and promoters, and the capacity of RNA
`synthesis under competitive conditions. Using the pDS1
`plasmid system (Fig. 1), strong promoters were also readily
`cloned under conditions which, again, allowed a selection
`
`
`
`20
`
`33 207
`
`Fragment No
`25
`26
`30
`41 225 25*
`
`
`
`IE‘)"nr'ntDF'1UJ3>
`
`r"'?<x...
`
`ZZ
`
`FIG. 6. Mapping of the positions of the cloned promoter frag-
`ments within the T5 genome. The autoradiogram depicts the result
`of Southern hybridizations of various labeled promoter-carrying
`fragments to a Hindlll digest of T5 wild-type DNA separated by
`agarose gel electrophoresis. The ethidium bromide-stained pattern
`of the Hindlll digest is shown in lanes M. The numbers above each
`lane designate the fragments. The correlation of the Hindlll cleav-
`age map with the genetic and functional organization of the T5
`genome has been described previously (8, 22). PE, E, and L
`designate the preearly, early, and late expresion classes of phage T5.
`
`1
`25
`
`‘I’
`S
`
`I
`7.5
`
`l
`10
`
`I
`15
`
`1
`20
`
`I "
`30
`
`Hrnin)
`
`FIG. 7. Time course of in vitro RNA synthesis governed by PN25
`and mm. RNA synthesis was allowed to proceed under standard
`conditions in 150 mM KC] and in presence of [or-"P]UTP. The
`templates used were pDS1-10/pN25,to2, digested with Rxal (A, upper
`insert), and pDS1-8/pN2(,,to2, digested with Xbal (O, lower insert).
`Samples of the reaction mixture were withdrawn at the times (t)
`indicated and divided into two fractions, of which one was analyzed
`by PAGE (inserts) and the other was used to monitor the acid-
`precipitable (5% trichloroacetic acid) counts (Cerenkov). It can be
`seen that RNA is produced at good rates for at least 30 min. The
`inserts show that the transcripts obtained are rather homogenous.
`More than 90% of the acid-precipitable counts can be hybridized to
`the single-stranded coding regions of the respective transcripts.
`Whereas the runoff RNA obtained with the Rsal-digested plasmid is
`192 nucleotides in length (upper insert), the size of the transcript
`shown in the lower insert is about 1,600 nucleotides (lane g, size
`marker). Lanes a through f in the upper insert show the PAGE
`analysis of the samples taken at times 2.5 through 30 min, and lanes
`h through j of the lower insert depict the corresponding analysis of
`samples taken at 5, 20, and 30 min, respectively, from transcription
`assays with Xbal-digested plasmid used as the template.
`
`for function but this time in vivo. CM resistance as well as
`CAT or CAT and DHFR protein production of the plasmid-
`containing cell were used as markers. The promoters iso-
`lated in this way exhibit remarkable structural and functional
`properties. There are striking commonalties among the 11
`promoter sequences despite the fact that they originate from
`the three different expression classes of phage T5 (Fig. 6). (i)
`The AT content of the relevant sequences from +20 to -—55,
`(the regions in contact with RNA polymerase; data not
`shown) is on the average 75%, with blocks of up to 83% AT
`between -56 and -36 as well as between -12 and +8. (ii)
`There are highly conserved regions around +7, +1, -10,
`-33, and -43. (iii) The distance between the -10 and the
`-33 region is 17 bp for all but one promoter, pl-[22, which is
`only 16 bp. This latter promoter is also the only one which
`most likely initiates RNA synthesis with a uridine instead of
`
`

`
`76
`
`GENTZ AND BUJARD
`
`J. BACTERIOL.
`
`Promoter
`
`M 28a 26
`— + — +
`
`25
`— +
`
`242 -
`
`147 -
`
`:3
`
`;;
`"'
`
`67 —
`
`*‘
`
`.C
`
`.C
`
`-
`O
`
`31». —
`
`25 —
`
`C
`
`an
`
`FIG. 8. Transcription-coupled capping in vitro. Plasmids of the
`type pDS1-10/p,t02 (form I DNA) containing the promoters pxm,
`PN26, pmg, respectively, were transcribed under standard conditions
`with (+) or without (-) 7-mGpppA in the presence of methyl-CTP
`instead of CTP. The labeled transcripts were separated by 7 M urea
`PAGE (20%, 50 V/Cm, 3 h) and visualized by autoradiography.
`Incorporation of 7-mGpppA clearly alters the electrophoretic mo-
`bility, permitting the identification of the capped transcription
`products. The positions of the first and second cytosine in the
`transcripts initiated from pm, p26, and p25 are 33 and 40, 15 and 20,
`30 and 36, respectively. M designates markers (HpaII digest of
`pBR322) whose sizes in nucleotides are given on the left.
`
`an adenosine. (iv) Perfect homology is found at five positions
`(-7, -11, -12, -34, and -43; Fig. 6). This number would
`increase to 8 (-33, -35, and -44) if one would allow a 17-bp
`distance between the -10 and -33 region of pun.
`A closer look at these promoter sequences reveals addi-
`tional interesting features. Promoters of the preearly and
`early expression class contain perfect homologies around +7
`(preearly, ATGAGAG; early, T'I‘GA) which is followed
`downstream by a block of purines (usually 10 of 12 bases).
`This, together with the homology near the RNA start posi-
`tion (T C/T ATA), suggests that in contrast to the common
`definition of a promoter sequence,
`regions around and
`downstream of +1 may be relevant for promoter function.
`Experiments to examine this hypothesis are currently in
`progress. Another striking feature of the sequences shown in
`Fig. 6 is the AT-rich block around -43. A clear selection
`against GC base pairs in this region is also seen in other
`promoters,
`including those of coliphage T7,
`the rRNA
`operons, and lpp promoter (for a review, see reference 10),
`all of which are eflicient RNA initiation signals in vivo. Thus,
`
`as pointed out previously (2), regions outside of the classical
`promoter sequence which spans from +1 to -35 are clearly
`under selective pressure, a conclusion which becomes more
`obvious when highly efiicient promoters are analyzed. It
`should also be not