J. Mol. Biol. (t990) 216, 25-37
Replacement of the Escherichia coli trp Operon Attenuation Control Codons Alters Operon Expression Robert Landick~'21-, Charles Yanofsky 3, Karen C h o o ~ and Le P h u n g ~ Departments of 1Biology and 2Biochemistry and Molecular Biophysics Washington University St Louis, MO 63130, U.S.A. and SDepartment of Biological Sciences Stanford University Stanford, CA 94305, U.S.A. (Received 5 February 1990; accepted 8 June 1990) To test features of the current model of transcription attenuation in amino acid biosynthetic operons, alterations were introduced into the trp operon leader region and expression of the mutated operons was examined in miaA and miaA + Escherichia coli strains that lacked the trp repressor. The miaA mutation prevents modification of the adenosine residue immediately 3' of the anticodon of tRNAs that interact with codons beginning with uridine. The undermodified tRNA Trp in miaA strains is thought to increase readthrough at the trp attenuator by slowing ribosome movement over two tandem Trp codons in the 14-codon leader peptide coding region. The rate of translation of these two "control codons" is thought to be the key step in determining the extent of transcription attenuation in the trp leader region. Sequential deletion of trpL DNA specifying the leader peptide initiation region, RNA segment l, RNA segment 2 and RNA segment 3 alternately decreased and increased trp operon expression, a result consistent with previous findings in another bacterium and the generally accepted model for transcription attenuation. Replacement of the tandem Trp control codons by AGG-UGC (Arg-Cys) codons eliminated the miaAdependent increase in transcription readthrough. Replacement of the Trp control codons by AGG-UGA (Arg-stop) codons caused complete reaxtthrough at the trp attenuator as well as abolisi~'mg the miaA effect. Presumably, the ribosome terminating translation at the new UGA c4don mimics the effect of a stalled ribosome at the Trp control codons. This finding suggests that ribosome dissociation at some stop codons is slow relative to the time required for transcription of the trp leader region. Thus, most ribosomes translating the trp leader peptide coding region may remain attached to the natural UGA stop codon until after the attenuation decision is made. This interpretation supports models for trp operon attenuation in which the elevated basal level readthrough is determined by occasional ribosome release prior to synthesis of the 3 : 4 terminator hairpin.
1. I n t r o d u c t i o n
these mechanisms is that for the trp operon of Escherichia coli, where experimental evidence has been provided in support of many of the features of the attenuation model (Landick & Yanofsky, 1987; Roesser & Yanofsky, 1988; Roesser et al., 1989; Landick et al., 1985, 1987; Kolter & Yanofsky, 1984). As currently postulated, the first distinguishable step following transcription initiation is the formation of a paused transcription complex, triggered by formation of the 1:2 RNA hairpin after synthesis of the first 92 nucleotides of the transcript (Farnham & Platt, 1981; Winkler & Yanofsky,
Many amino acid biosynthetic operons in enterobacteria are regulated by transcription attenuation mechanisms that are similar in their key features (Landick & Yanofsky, 1987; Bauer et al., 1983; Kolter & Yanofsky, 1982). The best understood of ¢ Author to whom all correspon.dence should be addressed. :~ Present address: Department of Biochemistry and Molecular Biology, The University of Chicago, Chicago, IL 60637, U.S.A. 0o22-2836/90/200025-13 $03.00/0
25
© 1990 Academic Press Limited
26
R. Landick et al.
1981; Fisher & Yanofsky, 1983; Fisher et al., 1985; Landick & Yanofsky, 1984). This transcription pause presumably provides time for a ribosome to initiate translation of the trp leader peptide coding region. Translation of the 5'-proximal portion of the leader peptide coding region releases the paused transcription complex, perhaps because the moving ribosome pulls RNA segment 1 into its mRNA binding site, dissociating the 1:2hairpin, and allowing re-formation of a normal RNA:DNA heteroduplex transcription substrate (Landick et al., 1985). The translating ribosome then continues translation until it either reaches the UGA stop codon or, if there is a deficiency of charged tRNA Trp, stalls at the Trp control codons. When the latter occurs RNA segment 1 is covered by the ribosome, RNA structure 2:3, the antiterminator, forms and precludes formation of RNA structure 3:4, the terminator. Thus, the consequence of ribosome stalling at the Trp control codons is transcription through the attenuator into the trp operon structural genes. One of the few remaining features of this mechanism requiring experimental clarification is the dynamics of ribosome movement and release relative to RNA polymerase movement (Kolter & Yanofsky, 1984; Roesser & Yanofsky, 1988; Roesser et al., 1989). The absence of leader peptide synthesis in a trpL29t (leader peptide AUG-~AUA) strain decreases trp operon expression by a factor of 2 to 3 relative to the basal level observed in wild-type strains growing in excess Trp (Kolter & Yanofsky, 1984; Zurawski et al., 1978; Roesser et al., 1989). Reduction of basal level readthrough when leader peptide synthesis is prevented has been termed superattenuation (Stroynowski & Yanofsky, 1982; Yanofsky, 1984). It also has been documented in Serratia marcescens, where deletion o f the leader peptide Shine-Dalgarno sequence or AUG codon decreases trp operon expression by a factor of 7 to l0 (Stroynowski//et al., 1982; Stroynowski & Yanofsky, 1982).~Roesser & Yanofsky have shown that wild-type ba'sal level readthrough may be attributed to ribosome release from the leader peptide stop codon, since inhibition of release in release factor mutants decreases trp operon expression (Roesser & Yanofsky, 1988; Roesser et al., 1989). They have argued that the 15°/o basal level readthrough observed in wild-type strains growing in excess Trp can be explained by approximately 24 % of the ribosomes releasing from the leader peptide UGA codon prior to 3 : 4 terminator RNA synthesis,
t trpL is the traditional genetic designation for the trp leader region; that is, the transcribed portion of the trp operon that precedes the trpE AUG codon, trpLep is our designation for the altered trp leader region containing EcoRI and PstI sites (see Figs 1 and 2). Alterations to trpLep are designated by the position of the alteration and the nucleotide present in the derivative, or by A when the nucleotide has been deleted. Deletions of larger portions of trpLep are numbered.
with half of the released RNAs folding into the antitermination configuration (Roesser et al., 1989). To characterize further the fate of a ribosome that translates the trp leader peptide coding region, as well as test general features of the current model for attenuation, we constructed several altered trp leader regions and examined transcription of these variants both in vivo and in vitro. We constructed several deletions of the trp leader region that blocked synthesis of the leader peptide and removed increasing numbers of RNA secondary structurespecifying segments. Examination of these deletions supported the postulated role in attenuation of these elements and also provided controls for a second set of experiments. Here we tested replacements, either with AGG-UGC (Arg-Cys) or AGG-UGA (Arg-stop) codons, of the Trp control codons in the trp leader peptide coding region. Results from these experiments support the idea that a ribosome that reaches a stop codon releases slowly relative to the rate of transcription of the trp operon leader region.
2. Materials and Methods
(a) Bacterial strains, plasmids, media and growth conditions The bacterial strains and plasmids used in this study and their derivations are listed in Table 1. Transfer of trp leader regions from plasmids to the chromosomes of RK2 and RK3 was accomplished by recombination essentially as described by Kolter & Yanofsky (1984). All t~T leader DNAs were transduced from the strain in which recombination occurred into plasmid-free RK2 and RK3, using a generalized P1 transducing phage and selecting for growth in the absence of tryptophan, to ensure that the strains used for assay of trp operon expression were isogenic except for the desired trpL alterations. Difco Tryptic Soy Broth was prepared according to the manufacturer's instructions. Luria-Bertani (LB) broth (Miller, 1972), Terrific Broth (Tartof & Hobbs, 1987) and Vogel-Bonner medium (Vogel & Bonner, 1956) were prepared as described. (b) Chemicals and enzymes High pressure liquid chromatography grade ethyl acetate (Baker) and scintillation grade Triton X-100 (Packard) were used for anthranilate synthase assays. All other chemicals were from Sigma (St Louis, MO). Restriction endonucleases and DNA-modifying enzymes were obtained from New England Biolabs (Beverley, MA), Promega Biotech (Madison, WI) and IBI (New Haven, CT) and used according to the manufacturers' instructions. E. coti RNA polymerase was prepared as described by Burgess & Jendrisak (1975) or was a gift from M. Chamberlin (University of California, Berkeley). NusA protein was purified as described by Schmidt & Chamberlin (1984) from an overproducing E. coli strain (Gribskov & Burgess, 1983). Oligodeoxyribonucleotides were synthesized on an Applied Biosystems model 380B oligonucleotide synthesizer, either at Stanford University or at Washington University, using standard phosphoramidite chemistry.
Replacement of trp Operon Control Codons
27
Table 1 E. coli strains and plasmids Strain or plasmid
Genotype or derivation
Reference
A. 81rains
RL245 RL246 RL289 RL290 I~L291 RL292 RL293 RL294 RL295 RL296 RL297 RL298 RL299 RIA39 RL440 RL441 RIA42 RL443 RL444 RIA45
supE thi A(lac-proAB)/F' traD36 proAB + laclqZAM15 recA1 endA1 gyrA96 thi hsdRl7 (rk-, mk+) ~tpE44 relA1 A(lac-proAB)/F' traD36 proAB + laclqZAM15 W3110 AtrpL-11/180 tnaA2 trpR W3110 AtrpL-ll/180 tnaA2 trpR miaA W31 l0 tnaA2 trpR W31 l0 tnaA2 trpR miaA W3110 trpLep tnaA2 trpR W3110 trpLep tnaA2 trpR miaA W3110 trpLep50A,57A,62C tnaA2 trpR W3110 trpLepSOA,57A,62C tnaA2 trpR miaA W3110 trpLep57A,62A lnaA2 trpR W3110 trpLep57A,62A tnaA2 trpR miaA W3110 trpLep57A,62C tnaA2 trpR W31 l0 trpLep57A,62C tnaA2 trpR miaA W31 l0 trpLep50A,57A62C,63G tnaA2 trpR W3110 trpLepSOA,57A.62C,63G tnaA2 trpR miaA W3110 trpLepASO.57A,62C tnaA2 trpR W3110 trpLepA50,57A,62C tnaA2 trpR miaA W3110 AtrpLeplO1 tnaA2 trpR W3110 AtrpLepl01 tnaA2 trpR miaA W3110 AtrpLepl02 tnaA2 trpR W31 l0 AtrpLepl02 tnaA2 trpR miaA W3110 AtrpLepl03 tnaA2 trpR W31 I0 AtrpLepl03 tnaA2 trpR miaA W3110 AtrpLepl04 tnaA2 trpR W3110 AtrpLepl04 tnaA2 trpR miaA W3110 AtrpLepl05 tnaA2 trpR W3110 AtrpLepl05 tnaA2 trpR miaA W3110 AtrpLep106 tnaA2 lrpR W31 l0 AtrpLepl06 t~mA2 trpR miaA W31 l0 AtrpLepl07 tnaA2 trpR W31 l0 AtrpLepl07 tnaA2 trpR miaA W3110 AtrpLepl08 tnaA2 trpR W3110 AtrpLepl08 tnaA2 trpR miaA W3110 AtrpLeplOltrpLep57A,62A tnaA2 trpR W31 l0 AtrpLep101trpLep57A,62A tnaA2 trpR
R~446 RL447
miaAAtrpLeplO1trpLep57A,62C tnaA2 trpR W3110 W3110 AtrpLep101trpLep57A,62C tnaA2 trpR
RIA48 RL449
W3110 At.rpLep105trpLep57A,62A tnaA2 trpR W3110 AtrpLeplO5trpLep57A,62A tnaA2 trpR
RIA50 RL451
W31 l0 AtrpLeplO5trpLep57A,62C tnaA2 trpR W3110 AtrpLeplO5trpLep57A,62C tnaA2 trpR
JM101 JM 109 RK2 RK3 RL224 RL225 RL230 RL231 RL233 RL234 RL236 RL237 RL239 RL240 RL242 RL243
Yaniseh-Perron eta/. (1985) Yanisch-Perron et al. (1985) Kolter & Yanofsky (1984) Kolter & Yanofsky (1984) This workt This work: This workt This work: This workt This work:[: This workt This work~ This workt This work:~ This workt This works This workt This work:~ This workt This works This workf This workS This workt This works This workt This works This workt This works This work¢ This work: This workt This work:[: This workt This workS This workt This work:]: This workt This work:
miaA This workt This work:[:
miaA This workt This work~
miaA B. Plasmids pRL243 pRL410 pRL415
pRIA16 pRL419 pRIA20 pRIA21 pRIA22 pRIA23
Wild-type trp 650 HpaII fragment cloned in pUCll AccI site trpLep 490 Sau3A fragment in BamHI site of pUCll9 with trp and lac promoters in the same orientation trpLep SmaI-SalI fragment from pRIA10 inserted between the EcoRI and HindIII sites of pUC119 using DNA polymerase Klenow fragment to fill-in 5' overhangs, trp and/ac promoters in opposite orientations Same as pRIA15 with trp and lac promoters in the same orientation pRIAI6 containing trpLep50A,57A,62C pRL416,contabfing trpLep57A,62A pRIAI5 containing trpLep57A,62C pRIA15 containing trpLep50A,57A,62C,63G pRIA15 containing trpLepA50,57A,62C
This work This work This work
This work This work This work This work This work This work
R. Landick et al.
28
Table 1 (continued) Strain or plasmid )RL441 )RL442 )RL443 )RL444 )RIA45 )RIA46 )RL447 )RL448 )RL449 )RL450 )RIA51 )RIA52
Genotype or derivation
Reference
p R L 4 1 5containing AtrpLepl01 pgL415 containing AtrpLep102 pRIA15 containing AtrpLepl03 pRIA15 containing AtrpLepl04 pRL415containing AtrpLeplO5 pRIAl5 containing AtrpLepl06 pRIA15 containing AtrpLepl07 pRIAl5 containing AtrpLepl08 pRL415containing AtrpLep101trpLep57A,62A pRIAl5 containing AtrpLep101trpLep57A.62C pRIA15 containing AtrpLep105trpLep57A.62A pRIAI5 containing AtrpLeplOStrpLep57A,62C
This work This work This work This work This work This work This work This work This work This work This work This work
t Constructed by reconibination of RK2 with the appropriate plasmid listed and selection for Trp + followed by Plvir-mediated transduetion of the mutant trp allele into RK2. Constructed by recombination of RK2 with the appropriate plasmid listed and selection for Trp + followed by Plvir-mediated transduction of the mutant trp allele into RK3.
(c) D N A isolation, preparation and analysis Plasmid DNAs were prepared on a small scale using the alkaline lysis procedure (Maniatis et al., 1982) and on a large scale using the Triton X-100 lysis procedure (Silhavy el al., 1984) followed by CsCI/ethidium bromide buoyant density gradient centrifugation (Maniatis el al., 1982). DNA fragments for in vitro transcription experiments were isolated by non-denaturing polyaczTlamide gel electrophoresis of appropriate plasmids that had been digested with XbaI and BamHI, or subjected to the polymerase chain reaction (Higuchi el al., 1988) with appropriate oligonucleotide primers, followed by elution with X buffer (Maxam & Gilbert, L980). Unless otherwise stated, all DNA manipulations were performed following standard published protocols (Ausubel el al., 1989; Maniatis et al., 1982). DNA sequences were determined by the dideoxynucleotide sequencing method (Ausubel el al., 1989; Sanger el al., 1977) using phage T7 DNA polymerase (Sequenase ~, US Biochemieals), [a-asS]dATP (Amersham) and either unive/rsal or trp-specific oligonucleotide primers. (d) Construction of pRL415 containing the trpLep
reaction was terminated by thorough mixing with 8 pl of 250mM-Na2EDTA, (pH 8"0) and 25pl of equilibrated phenol. After recovery and extraction of the aqueous layer with CHCI3, the partially double-stranded, linearized DNA was electrophoresed through an 0"80/0 (w/v) low-melting agarose (IBI)/TBE gel and isolated by phenol extraction of a gel slice. After recovery by precipitation with ethanol, the DNA was ligated to a phosphorylated, synthetic duplex DNA fragment [d(ACTGCAGAAAGGTTGGTGGCGCACTTCCTGAATTCGGGC)/d(CCCG AATTCAGGAAGTGCGCCACCAACCTTTCTGCAGT)] and transformed into strain JMl01. A recombinant phage containing the desired sequence was identified by restriction endonuclease site analysis and DNA sequencing. The trpLep leader DNA then was excised from the recombinant Ml3 phage with S~mI and SalI, made blunt-ended by treatment with DNA polymerase I Klenow fl'agment and ligated to pUCll9 (Vieira & Messing, 1987), which had been cleaved with EcoRI and HindIII and similarly made blunt-ended, so that the EcoRI and PslI sites in the trpLep DNA were unique in the resulting plasmid (pRIAl5).
leader region
(e) Construction of trp leader deletions'
The trpLep DNA was constructed using a wild-type lrpL 490 bp~ ,.~au3A DNA fragment inserted into the BamHI site of M13 phage mp9 (Messing & Vieira, 1982).
pRL415 that had been cleaved with PstI was treated with Bat31 nuclease (New England Biolabs) for 30 s at 37°C, repaired with DNA polymerase I Klenow fragment and ligated to EcoRI linkers [d(CGGAATTCCG)]. After removal of excess linkers by digestion with EcoRI and spermine precipitation, the DNA was recircularized with DNA ligase and transformed into strain JM109. The extent of deletion in plasmids recovered was determined by restriction endonuclease site mapping and subsequent DNA sequencing. The deletions shown in Fig. 2 were then assembled by religation of the appropriate DNA fragments excised from low-melting agarose gels after generation, by cleavage, of different deletion plasmids and pRIA15 with EcoRI and BamHI (AlrpLepl02, AtrpLepl03, AlrpLepl04, AlrpLepl06, AtrpLepl07 and AlrpLepl08) or by cleavage with PslI or EcoRI followed by generation of blunt-ends with phage T4 DNA polymerase and DNA polymerase I Klenow fragment, respectively, and then BamHI (AtrpLepl01 and AlrpLepl05).
Approx. 1/~g of single-stranded DNA from this phage was hybridized to a 10-fold molar excess of double-stranded 490 bp Sau3A DNA fragment at 100°C for 90 s followed by 60 min at 68°C in 92/~l of l0 mM-Tris" HCI, (pH 7"4), l0 mM-MgCl2, 6 mM-KCI, 100 ~g of BSA/ml. After addition of fl-mercaptoethanol to 6 mM, the DNA was digested with 20 units of HphI (New England Biolabs) for 20 min at 37°C, combined with 4/d of 5mM-Trp, 5~1 of 1 M-NaCI, l ]~g of trp repressor (to prevent cleavage at the RsaI site in the trp operator), and digested with l0 units of RsaI (New England Biolabs) for 15 min at 37°C. The
Abbreviations used: bp, base-pair(s); BSA, bovine serum albumin; DTT, dithiothreitol.
Replacement of trp Operon Control Codons (f) Construction of trp leader regions with altered
control codons Approx. l ttg of pRIA15 was cleaved with EcoRI and PstI and the large DNA fragment was recovered in a slice of low-melting agarose after electrophoresis. Approx. 1/50 of the gel slice was melted and combined with approx. 5 ng of the phosphorylated, sinffle-stranded oligonucleotide [d(AATTCAGGAAGTGCGTCACCTACCTTTCTGCA)] (where the duplicated letters indicate incorporation of 50~/o of each base at that position), ligated and transformed into JM109. The altered trp leader DNAs shown in Fig. 2 were identified by screening of plasmids from the transformants for loss of the HhaI site and acquisition of an NlaIV site followed by DNA sequencing. (g) In vitro transcriplion assays For synchronized, single-round transcription reactions to measure paused transcription complex half-lives, 2"5 p,nol of RNA polymerase and 1 pmol of DNA template were combined and incubated at 37 °C for l0 min in 100/~l of buffer containing 20 mM-Tris-acetate, (pH 8-0), 130 mM-KC1, 4 mM-MgCI2, 0-1 mM-Na2EDTA, 0"1 mM-DTT, 150pM-ATP, 20gM-[~-32p]GTP, 4% (v/v) glycerol. 20/~g of acetylated BSA/ml. Transcription was initiated by addition of (.'TP and UTP (150 #M each, final concentration) and l0 pg of rifampicin/ml. Samples (10 pl) were removed at appropriate time intervals and mixed with l0 pl of 2 x TBE, 0"1% (w/v) bromphenol blue, 0"1% (w/v) xylene cyanol saturated with urea. RNA samples were analyzed by electrophoresis through a 10% (w/v) polyacrylamide/7 M-urea-TBE gel. The radioactive pause RNA bands were excised from the gel and their C,erenkov radiation was determined in a scintillation counter. After subtraction of appropriate background values, these data were plotted versus time on semi-log paper and the pseudo first-order rate of pause RNA disappearance was determined from the slopes of the semi-log plots. Steady-state transcription reactions to determine percentage readthrough at the trp attenuator were performed under the same conditions, but with DNA templates prepared from pl~hmids and forward and reverse pUCll9 sequencing primers by the polymerase chain reaction (Higuchi et al., 1988). After formation of open complexes for l0 min at 37°C, heparin was added to 10#g/ml, and CTP and UTP to 150#M, and the reactions were incubated for 10min at .37°C. After electrophoresis as described above, the R,NAs were quantitated in the gels using an AMBIS v Radioanalytic Imaging System. Percentage readthrough was calculated as: mol ~o run-off RNA/(mol % leader RNA +mol ~o run-off RNA). {h) Secondary structure prediction rCNA secondary structures and their free energies of stabilization were predicted using the RNAFLD program of Zuker & Steigler {1981) as implemented by the University of Wisconsin Genetics Computer Group on a VAX 11/750 in the Department of Biology, Washington University. (i) Anthranilate synthase assays Anthranilate synthase assays
Replacement of the Escherichia coli trp Operon Attenuation Control Codons Alters Operon Expression Robert Landick~'21-, Charles Yanofsky 3, Karen C h o o ~ and Le P h u n g ~ Departments of 1Biology and 2Biochemistry and Molecular Biophysics Washington University St Louis, MO 63130, U.S.A. and SDepartment of Biological Sciences Stanford University Stanford, CA 94305, U.S.A. (Received 5 February 1990; accepted 8 June 1990) To test features of the current model of transcription attenuation in amino acid biosynthetic operons, alterations were introduced into the trp operon leader region and expression of the mutated operons was examined in miaA and miaA + Escherichia coli strains that lacked the trp repressor. The miaA mutation prevents modification of the adenosine residue immediately 3' of the anticodon of tRNAs that interact with codons beginning with uridine. The undermodified tRNA Trp in miaA strains is thought to increase readthrough at the trp attenuator by slowing ribosome movement over two tandem Trp codons in the 14-codon leader peptide coding region. The rate of translation of these two "control codons" is thought to be the key step in determining the extent of transcription attenuation in the trp leader region. Sequential deletion of trpL DNA specifying the leader peptide initiation region, RNA segment l, RNA segment 2 and RNA segment 3 alternately decreased and increased trp operon expression, a result consistent with previous findings in another bacterium and the generally accepted model for transcription attenuation. Replacement of the tandem Trp control codons by AGG-UGC (Arg-Cys) codons eliminated the miaAdependent increase in transcription readthrough. Replacement of the Trp control codons by AGG-UGA (Arg-stop) codons caused complete reaxtthrough at the trp attenuator as well as abolisi~'mg the miaA effect. Presumably, the ribosome terminating translation at the new UGA c4don mimics the effect of a stalled ribosome at the Trp control codons. This finding suggests that ribosome dissociation at some stop codons is slow relative to the time required for transcription of the trp leader region. Thus, most ribosomes translating the trp leader peptide coding region may remain attached to the natural UGA stop codon until after the attenuation decision is made. This interpretation supports models for trp operon attenuation in which the elevated basal level readthrough is determined by occasional ribosome release prior to synthesis of the 3 : 4 terminator hairpin.
1. I n t r o d u c t i o n
these mechanisms is that for the trp operon of Escherichia coli, where experimental evidence has been provided in support of many of the features of the attenuation model (Landick & Yanofsky, 1987; Roesser & Yanofsky, 1988; Roesser et al., 1989; Landick et al., 1985, 1987; Kolter & Yanofsky, 1984). As currently postulated, the first distinguishable step following transcription initiation is the formation of a paused transcription complex, triggered by formation of the 1:2 RNA hairpin after synthesis of the first 92 nucleotides of the transcript (Farnham & Platt, 1981; Winkler & Yanofsky,
Many amino acid biosynthetic operons in enterobacteria are regulated by transcription attenuation mechanisms that are similar in their key features (Landick & Yanofsky, 1987; Bauer et al., 1983; Kolter & Yanofsky, 1982). The best understood of ¢ Author to whom all correspon.dence should be addressed. :~ Present address: Department of Biochemistry and Molecular Biology, The University of Chicago, Chicago, IL 60637, U.S.A. 0o22-2836/90/200025-13 $03.00/0
25
© 1990 Academic Press Limited
26
R. Landick et al.
1981; Fisher & Yanofsky, 1983; Fisher et al., 1985; Landick & Yanofsky, 1984). This transcription pause presumably provides time for a ribosome to initiate translation of the trp leader peptide coding region. Translation of the 5'-proximal portion of the leader peptide coding region releases the paused transcription complex, perhaps because the moving ribosome pulls RNA segment 1 into its mRNA binding site, dissociating the 1:2hairpin, and allowing re-formation of a normal RNA:DNA heteroduplex transcription substrate (Landick et al., 1985). The translating ribosome then continues translation until it either reaches the UGA stop codon or, if there is a deficiency of charged tRNA Trp, stalls at the Trp control codons. When the latter occurs RNA segment 1 is covered by the ribosome, RNA structure 2:3, the antiterminator, forms and precludes formation of RNA structure 3:4, the terminator. Thus, the consequence of ribosome stalling at the Trp control codons is transcription through the attenuator into the trp operon structural genes. One of the few remaining features of this mechanism requiring experimental clarification is the dynamics of ribosome movement and release relative to RNA polymerase movement (Kolter & Yanofsky, 1984; Roesser & Yanofsky, 1988; Roesser et al., 1989). The absence of leader peptide synthesis in a trpL29t (leader peptide AUG-~AUA) strain decreases trp operon expression by a factor of 2 to 3 relative to the basal level observed in wild-type strains growing in excess Trp (Kolter & Yanofsky, 1984; Zurawski et al., 1978; Roesser et al., 1989). Reduction of basal level readthrough when leader peptide synthesis is prevented has been termed superattenuation (Stroynowski & Yanofsky, 1982; Yanofsky, 1984). It also has been documented in Serratia marcescens, where deletion o f the leader peptide Shine-Dalgarno sequence or AUG codon decreases trp operon expression by a factor of 7 to l0 (Stroynowski//et al., 1982; Stroynowski & Yanofsky, 1982).~Roesser & Yanofsky have shown that wild-type ba'sal level readthrough may be attributed to ribosome release from the leader peptide stop codon, since inhibition of release in release factor mutants decreases trp operon expression (Roesser & Yanofsky, 1988; Roesser et al., 1989). They have argued that the 15°/o basal level readthrough observed in wild-type strains growing in excess Trp can be explained by approximately 24 % of the ribosomes releasing from the leader peptide UGA codon prior to 3 : 4 terminator RNA synthesis,
t trpL is the traditional genetic designation for the trp leader region; that is, the transcribed portion of the trp operon that precedes the trpE AUG codon, trpLep is our designation for the altered trp leader region containing EcoRI and PstI sites (see Figs 1 and 2). Alterations to trpLep are designated by the position of the alteration and the nucleotide present in the derivative, or by A when the nucleotide has been deleted. Deletions of larger portions of trpLep are numbered.
with half of the released RNAs folding into the antitermination configuration (Roesser et al., 1989). To characterize further the fate of a ribosome that translates the trp leader peptide coding region, as well as test general features of the current model for attenuation, we constructed several altered trp leader regions and examined transcription of these variants both in vivo and in vitro. We constructed several deletions of the trp leader region that blocked synthesis of the leader peptide and removed increasing numbers of RNA secondary structurespecifying segments. Examination of these deletions supported the postulated role in attenuation of these elements and also provided controls for a second set of experiments. Here we tested replacements, either with AGG-UGC (Arg-Cys) or AGG-UGA (Arg-stop) codons, of the Trp control codons in the trp leader peptide coding region. Results from these experiments support the idea that a ribosome that reaches a stop codon releases slowly relative to the rate of transcription of the trp operon leader region.
2. Materials and Methods
(a) Bacterial strains, plasmids, media and growth conditions The bacterial strains and plasmids used in this study and their derivations are listed in Table 1. Transfer of trp leader regions from plasmids to the chromosomes of RK2 and RK3 was accomplished by recombination essentially as described by Kolter & Yanofsky (1984). All t~T leader DNAs were transduced from the strain in which recombination occurred into plasmid-free RK2 and RK3, using a generalized P1 transducing phage and selecting for growth in the absence of tryptophan, to ensure that the strains used for assay of trp operon expression were isogenic except for the desired trpL alterations. Difco Tryptic Soy Broth was prepared according to the manufacturer's instructions. Luria-Bertani (LB) broth (Miller, 1972), Terrific Broth (Tartof & Hobbs, 1987) and Vogel-Bonner medium (Vogel & Bonner, 1956) were prepared as described. (b) Chemicals and enzymes High pressure liquid chromatography grade ethyl acetate (Baker) and scintillation grade Triton X-100 (Packard) were used for anthranilate synthase assays. All other chemicals were from Sigma (St Louis, MO). Restriction endonucleases and DNA-modifying enzymes were obtained from New England Biolabs (Beverley, MA), Promega Biotech (Madison, WI) and IBI (New Haven, CT) and used according to the manufacturers' instructions. E. coti RNA polymerase was prepared as described by Burgess & Jendrisak (1975) or was a gift from M. Chamberlin (University of California, Berkeley). NusA protein was purified as described by Schmidt & Chamberlin (1984) from an overproducing E. coli strain (Gribskov & Burgess, 1983). Oligodeoxyribonucleotides were synthesized on an Applied Biosystems model 380B oligonucleotide synthesizer, either at Stanford University or at Washington University, using standard phosphoramidite chemistry.
Replacement of trp Operon Control Codons
27
Table 1 E. coli strains and plasmids Strain or plasmid
Genotype or derivation
Reference
A. 81rains
RL245 RL246 RL289 RL290 I~L291 RL292 RL293 RL294 RL295 RL296 RL297 RL298 RL299 RIA39 RL440 RL441 RIA42 RL443 RL444 RIA45
supE thi A(lac-proAB)/F' traD36 proAB + laclqZAM15 recA1 endA1 gyrA96 thi hsdRl7 (rk-, mk+) ~tpE44 relA1 A(lac-proAB)/F' traD36 proAB + laclqZAM15 W3110 AtrpL-11/180 tnaA2 trpR W3110 AtrpL-ll/180 tnaA2 trpR miaA W31 l0 tnaA2 trpR W31 l0 tnaA2 trpR miaA W3110 trpLep tnaA2 trpR W3110 trpLep tnaA2 trpR miaA W3110 trpLep50A,57A,62C tnaA2 trpR W3110 trpLepSOA,57A,62C tnaA2 trpR miaA W3110 trpLep57A,62A lnaA2 trpR W3110 trpLep57A,62A tnaA2 trpR miaA W3110 trpLep57A,62C tnaA2 trpR W31 l0 trpLep57A,62C tnaA2 trpR miaA W31 l0 trpLep50A,57A62C,63G tnaA2 trpR W3110 trpLepSOA,57A.62C,63G tnaA2 trpR miaA W3110 trpLepASO.57A,62C tnaA2 trpR W3110 trpLepA50,57A,62C tnaA2 trpR miaA W3110 AtrpLeplO1 tnaA2 trpR W3110 AtrpLepl01 tnaA2 trpR miaA W3110 AtrpLepl02 tnaA2 trpR W31 l0 AtrpLepl02 tnaA2 trpR miaA W3110 AtrpLepl03 tnaA2 trpR W31 I0 AtrpLepl03 tnaA2 trpR miaA W3110 AtrpLepl04 tnaA2 trpR W3110 AtrpLepl04 tnaA2 trpR miaA W3110 AtrpLepl05 tnaA2 trpR W3110 AtrpLepl05 tnaA2 trpR miaA W3110 AtrpLep106 tnaA2 lrpR W31 l0 AtrpLepl06 t~mA2 trpR miaA W31 l0 AtrpLepl07 tnaA2 trpR W31 l0 AtrpLepl07 tnaA2 trpR miaA W3110 AtrpLepl08 tnaA2 trpR W3110 AtrpLepl08 tnaA2 trpR miaA W3110 AtrpLeplOltrpLep57A,62A tnaA2 trpR W31 l0 AtrpLep101trpLep57A,62A tnaA2 trpR
R~446 RL447
miaAAtrpLeplO1trpLep57A,62C tnaA2 trpR W3110 W3110 AtrpLep101trpLep57A,62C tnaA2 trpR
RIA48 RL449
W3110 At.rpLep105trpLep57A,62A tnaA2 trpR W3110 AtrpLeplO5trpLep57A,62A tnaA2 trpR
RIA50 RL451
W31 l0 AtrpLeplO5trpLep57A,62C tnaA2 trpR W3110 AtrpLeplO5trpLep57A,62C tnaA2 trpR
JM101 JM 109 RK2 RK3 RL224 RL225 RL230 RL231 RL233 RL234 RL236 RL237 RL239 RL240 RL242 RL243
Yaniseh-Perron eta/. (1985) Yanisch-Perron et al. (1985) Kolter & Yanofsky (1984) Kolter & Yanofsky (1984) This workt This work: This workt This work: This workt This work:[: This workt This work~ This workt This work:~ This workt This works This workt This work:~ This workt This works This workf This workS This workt This works This workt This works This workt This works This work¢ This work: This workt This work:[: This workt This workS This workt This work:]: This workt This work:
miaA This workt This work:[:
miaA This workt This work~
miaA B. Plasmids pRL243 pRL410 pRL415
pRIA16 pRL419 pRIA20 pRIA21 pRIA22 pRIA23
Wild-type trp 650 HpaII fragment cloned in pUCll AccI site trpLep 490 Sau3A fragment in BamHI site of pUCll9 with trp and lac promoters in the same orientation trpLep SmaI-SalI fragment from pRIA10 inserted between the EcoRI and HindIII sites of pUC119 using DNA polymerase Klenow fragment to fill-in 5' overhangs, trp and/ac promoters in opposite orientations Same as pRIA15 with trp and lac promoters in the same orientation pRIAI6 containing trpLep50A,57A,62C pRL416,contabfing trpLep57A,62A pRIAI5 containing trpLep57A,62C pRIA15 containing trpLep50A,57A,62C,63G pRIA15 containing trpLepA50,57A,62C
This work This work This work
This work This work This work This work This work This work
R. Landick et al.
28
Table 1 (continued) Strain or plasmid )RL441 )RL442 )RL443 )RL444 )RIA45 )RIA46 )RL447 )RL448 )RL449 )RL450 )RIA51 )RIA52
Genotype or derivation
Reference
p R L 4 1 5containing AtrpLepl01 pgL415 containing AtrpLep102 pRIA15 containing AtrpLepl03 pRIA15 containing AtrpLepl04 pRL415containing AtrpLeplO5 pRIAl5 containing AtrpLepl06 pRIA15 containing AtrpLepl07 pRIAl5 containing AtrpLepl08 pRL415containing AtrpLep101trpLep57A,62A pRIAl5 containing AtrpLep101trpLep57A.62C pRIA15 containing AtrpLep105trpLep57A.62A pRIAI5 containing AtrpLeplOStrpLep57A,62C
This work This work This work This work This work This work This work This work This work This work This work This work
t Constructed by reconibination of RK2 with the appropriate plasmid listed and selection for Trp + followed by Plvir-mediated transduetion of the mutant trp allele into RK2. Constructed by recombination of RK2 with the appropriate plasmid listed and selection for Trp + followed by Plvir-mediated transduction of the mutant trp allele into RK3.
(c) D N A isolation, preparation and analysis Plasmid DNAs were prepared on a small scale using the alkaline lysis procedure (Maniatis et al., 1982) and on a large scale using the Triton X-100 lysis procedure (Silhavy el al., 1984) followed by CsCI/ethidium bromide buoyant density gradient centrifugation (Maniatis el al., 1982). DNA fragments for in vitro transcription experiments were isolated by non-denaturing polyaczTlamide gel electrophoresis of appropriate plasmids that had been digested with XbaI and BamHI, or subjected to the polymerase chain reaction (Higuchi el al., 1988) with appropriate oligonucleotide primers, followed by elution with X buffer (Maxam & Gilbert, L980). Unless otherwise stated, all DNA manipulations were performed following standard published protocols (Ausubel el al., 1989; Maniatis et al., 1982). DNA sequences were determined by the dideoxynucleotide sequencing method (Ausubel el al., 1989; Sanger el al., 1977) using phage T7 DNA polymerase (Sequenase ~, US Biochemieals), [a-asS]dATP (Amersham) and either unive/rsal or trp-specific oligonucleotide primers. (d) Construction of pRL415 containing the trpLep
reaction was terminated by thorough mixing with 8 pl of 250mM-Na2EDTA, (pH 8"0) and 25pl of equilibrated phenol. After recovery and extraction of the aqueous layer with CHCI3, the partially double-stranded, linearized DNA was electrophoresed through an 0"80/0 (w/v) low-melting agarose (IBI)/TBE gel and isolated by phenol extraction of a gel slice. After recovery by precipitation with ethanol, the DNA was ligated to a phosphorylated, synthetic duplex DNA fragment [d(ACTGCAGAAAGGTTGGTGGCGCACTTCCTGAATTCGGGC)/d(CCCG AATTCAGGAAGTGCGCCACCAACCTTTCTGCAGT)] and transformed into strain JMl01. A recombinant phage containing the desired sequence was identified by restriction endonuclease site analysis and DNA sequencing. The trpLep leader DNA then was excised from the recombinant Ml3 phage with S~mI and SalI, made blunt-ended by treatment with DNA polymerase I Klenow fl'agment and ligated to pUCll9 (Vieira & Messing, 1987), which had been cleaved with EcoRI and HindIII and similarly made blunt-ended, so that the EcoRI and PslI sites in the trpLep DNA were unique in the resulting plasmid (pRIAl5).
leader region
(e) Construction of trp leader deletions'
The trpLep DNA was constructed using a wild-type lrpL 490 bp~ ,.~au3A DNA fragment inserted into the BamHI site of M13 phage mp9 (Messing & Vieira, 1982).
pRL415 that had been cleaved with PstI was treated with Bat31 nuclease (New England Biolabs) for 30 s at 37°C, repaired with DNA polymerase I Klenow fragment and ligated to EcoRI linkers [d(CGGAATTCCG)]. After removal of excess linkers by digestion with EcoRI and spermine precipitation, the DNA was recircularized with DNA ligase and transformed into strain JM109. The extent of deletion in plasmids recovered was determined by restriction endonuclease site mapping and subsequent DNA sequencing. The deletions shown in Fig. 2 were then assembled by religation of the appropriate DNA fragments excised from low-melting agarose gels after generation, by cleavage, of different deletion plasmids and pRIA15 with EcoRI and BamHI (AlrpLepl02, AtrpLepl03, AlrpLepl04, AlrpLepl06, AtrpLepl07 and AlrpLepl08) or by cleavage with PslI or EcoRI followed by generation of blunt-ends with phage T4 DNA polymerase and DNA polymerase I Klenow fragment, respectively, and then BamHI (AtrpLepl01 and AlrpLepl05).
Approx. 1/~g of single-stranded DNA from this phage was hybridized to a 10-fold molar excess of double-stranded 490 bp Sau3A DNA fragment at 100°C for 90 s followed by 60 min at 68°C in 92/~l of l0 mM-Tris" HCI, (pH 7"4), l0 mM-MgCl2, 6 mM-KCI, 100 ~g of BSA/ml. After addition of fl-mercaptoethanol to 6 mM, the DNA was digested with 20 units of HphI (New England Biolabs) for 20 min at 37°C, combined with 4/d of 5mM-Trp, 5~1 of 1 M-NaCI, l ]~g of trp repressor (to prevent cleavage at the RsaI site in the trp operator), and digested with l0 units of RsaI (New England Biolabs) for 15 min at 37°C. The
Abbreviations used: bp, base-pair(s); BSA, bovine serum albumin; DTT, dithiothreitol.
Replacement of trp Operon Control Codons (f) Construction of trp leader regions with altered
control codons Approx. l ttg of pRIA15 was cleaved with EcoRI and PstI and the large DNA fragment was recovered in a slice of low-melting agarose after electrophoresis. Approx. 1/50 of the gel slice was melted and combined with approx. 5 ng of the phosphorylated, sinffle-stranded oligonucleotide [d(AATTCAGGAAGTGCGTCACCTACCTTTCTGCA)] (where the duplicated letters indicate incorporation of 50~/o of each base at that position), ligated and transformed into JM109. The altered trp leader DNAs shown in Fig. 2 were identified by screening of plasmids from the transformants for loss of the HhaI site and acquisition of an NlaIV site followed by DNA sequencing. (g) In vitro transcriplion assays For synchronized, single-round transcription reactions to measure paused transcription complex half-lives, 2"5 p,nol of RNA polymerase and 1 pmol of DNA template were combined and incubated at 37 °C for l0 min in 100/~l of buffer containing 20 mM-Tris-acetate, (pH 8-0), 130 mM-KC1, 4 mM-MgCI2, 0-1 mM-Na2EDTA, 0"1 mM-DTT, 150pM-ATP, 20gM-[~-32p]GTP, 4% (v/v) glycerol. 20/~g of acetylated BSA/ml. Transcription was initiated by addition of (.'TP and UTP (150 #M each, final concentration) and l0 pg of rifampicin/ml. Samples (10 pl) were removed at appropriate time intervals and mixed with l0 pl of 2 x TBE, 0"1% (w/v) bromphenol blue, 0"1% (w/v) xylene cyanol saturated with urea. RNA samples were analyzed by electrophoresis through a 10% (w/v) polyacrylamide/7 M-urea-TBE gel. The radioactive pause RNA bands were excised from the gel and their C,erenkov radiation was determined in a scintillation counter. After subtraction of appropriate background values, these data were plotted versus time on semi-log paper and the pseudo first-order rate of pause RNA disappearance was determined from the slopes of the semi-log plots. Steady-state transcription reactions to determine percentage readthrough at the trp attenuator were performed under the same conditions, but with DNA templates prepared from pl~hmids and forward and reverse pUCll9 sequencing primers by the polymerase chain reaction (Higuchi et al., 1988). After formation of open complexes for l0 min at 37°C, heparin was added to 10#g/ml, and CTP and UTP to 150#M, and the reactions were incubated for 10min at .37°C. After electrophoresis as described above, the R,NAs were quantitated in the gels using an AMBIS v Radioanalytic Imaging System. Percentage readthrough was calculated as: mol ~o run-off RNA/(mol % leader RNA +mol ~o run-off RNA). {h) Secondary structure prediction rCNA secondary structures and their free energies of stabilization were predicted using the RNAFLD program of Zuker & Steigler {1981) as implemented by the University of Wisconsin Genetics Computer Group on a VAX 11/750 in the Department of Biology, Washington University. (i) Anthranilate synthase assays Anthranilate synthase assays
