0300-9084/$ - see front matterdoi:10.1016/j.biochi.2006.01.00
Codon-specific and general inhibition
of protein synthesis by the tRNA-sequestering minigenes
Luis Delgado-Olivares, Efraín Zamora-Romo, Gabriel Guarneros, Javier Hernandez-Sanchez *
Departamento de Genética y Biología Molecular, Centro de Investigación y de Estudios Avanzados del IPN, Apartado Postal 14-740, México DF 07000, Mexico
Received 11 October 2005; accepted 10 January 2006Available online 03 February 2006
During protein synthesis, it may occur that the elongatingpolypeptide chain terminates abortively by dissociation or“drop off” of the nascent peptidyl-tRNA (pep-tRNA) fromthe ribosome [1]. The pep-tRNA is readily cleaved by pepti-dyl-tRNA hydrolase (Pth) to regenerate aminoacylatable tRNAfor new rounds of polypeptide elongation. The activity of Pthis essential for protein synthesis and viability of bacteria [2]. Amutation in the gene encoding the hydrolase, which confers atemperature-sensitive phenotype, pth(Ts), determines a rapidaccumulation of pep-tRNA upon shift to a restrictive tempera-ture [1,3]. Mutant pth(rap) cells, which have deficient Pth ac-tivity, are unable to maintain the vegetative growth of lambdabacteriophage under conditions that allow exponential cellgrowth [3,4].
Minigenes are DNA segments whose transcripts contain aShine-Dalgarno sequence appropriately spaced for translationfrom short reading frames (from two to six sense codons),which direct the synthesis of oligopeptides [5–7]. The nearlyidentical two-codon barI and barII minigenes, differing only inthe ORF sequence, AUG AUA UAA or AUG AUA UGA,respectively, were originally isolated from lambda bacterioph-age. However, synthetic minigenes of variable length and co-don composition have also been analyzed [6,7]. The expressionof some minigenes can be toxic to Pth-defective cells becausethey sequester tRNA as pep-tRNA reducing the pool of ami-noacylable tRNA essential for protein synthesis [8]. Some tri-plets when located next to the termination codon of the mini-genes generate pep-tRNAs which are not readily hydrolyzed bythe ribosomal peptidyl transferase center [9]. Presumably, theribosome pauses at the termination codon in the minigenemRNA favoring drop off and accumulation of the pep-tRNAspecific for the last sense codon [6,8,9]. Accordingly, tRNAstarvation leads to a general arrest of protein synthesis and celldeath [10] and these effects are reverted by the supplementa-
L. Delgado-Olivares et al. / Biochimie 88 (2006) 793–800794
tion of Pth-defective cells with tRNA or Pth [8,11]. Indeed,there is a strict correlation between the rate of pep-tRNA dropoff and cell toxicity of the minigene [9,12]. The sequestrationof tRNA by minigenes has previously been used to limit thepool of specific tRNAs and promote bypassing by specific nu-cleotide sequences using the lacZ gene as a reporter [13]. Pep-tRNA drop off has also been proposed to explain the seques-tration of tRNA species cognate to the last sense codon of anoverproduced gratuitous protein [14]. However, minigeneshave only been associated with a general inhibition of proteinsynthesis which mostly depended on the rate of pep-tRNAdrop off and the translational signals of the minigene. Thus,in this work we analyzed whether there is a selective inhibitionof target genes containing codons that are present in the lastposition of the tRNA-sequestering minigene. We also analyzedif the inhibition depends on the number of “hungry codons” (aterm coined by Gallant J. [13,15]) in the target gene andwhether the ribosome pauses at hungry codons along the repor-ter messenger. Here, we show that the inhibition of proteinsynthesis can be targeted to genes, which harbor specific co-dons, by the expression of minigenes which carry the samecodon prior to the termination codon. This inhibition dependedon the number of hungry codons of the reporter gene. The ex-pression of an incomplete polypeptide from a lacZ constructinterrupted where AGA codons were substituted at positions352 and 353 after minigene-mediated limitation of tRNAArg4,indicated that the ribosome paused at the AGA codons of themessenger. Finally, a non-specific inhibition of protein synth-
Table 1Strains of E. coli K12 and plasmids
Strains or plasmids GenotypeStrainsP90C pth(rap) araΔ(lac-pro)thi pth(rap)zch::Tn10C600 pth(rap) F- (leu-B6 thi-1 supE44 lac Y1 tonA21) pth(rap)zch::Tn10LDO1 P90C pth(rap) ΔTn10LDO2 P90C pth(rap) ΔTn10 ssrA::KanW3110 ssrA::Kan F-IN[rrnD-rrnE]1 ssrA::KanPlasmids DescriptionpPROEX-1 bla lacIQ pTrcpHZ Derivative of pPROEX-1 with NdeI/StuI chromosomal lacZpHZ-MR Derivative of pHZ with SpeI/HinDIII AUGAGAUAA minipHZ353-MR Derivative of pHZ-MR with CGA353AGA substitution inpHZ352-353-MR Derivative of pHZ-MR with ATT352AGA CGA353AGA spHZS-MR Derivative of pHZ-MR with CGA353AGA GGC354UAApKQV4 bla lacIQ pTrcplacZ pKQV4 containing EcoRI/HinDIII lacZ gene insert from plplexlacZ expression protein vector, containing lacZ gene Apr
placZ2 Derivative of placZwt with ACC2AGA substitution in lacZplacZ2-3 Derivative of placZwt with ACC2AGA, GTA3AGA substipAGA Derivative of pKQV4 with EcoRI/HindIII AUGAGAUAA
of dipeptide MRpCAC Derivative of pKQV4 with EcoRI/HindIII AUGCACUAApAUA Derivative of pKQV4 with EcoRI/HindIII AUGAUAUAApAAA Derivative of pKQV4 with EcoRI/HindIII AUGAAAUAApACG Derivative of pKQV4 with EcoRI/HindIII AUGACGUAApGGG Derivative of pKQV4 with EcoRI/HindIII AUGGGGUAAp101 Derivative of pKQV4 with an inactive EcoRI/HinDIII AUApPNP pCJ11 construct derivative of pUC19 with SalI/BamHI pnppACYC-184 Cmr
pAGG Derivative of pKQV4 with EcoRI/HindIII AUGAGGUAApArg4 Originally pDC952, a derivative of pACYC-184 harboring
esis, irrespective of the codon content of the target gene, wasalso observed in vitro.
2. Materials and methods
2.1. Strains and growth conditions
Escherchia coli strains and plasmids used in this study arelisted in Table 1. P90C pth(rap), a Pth-defective strain, con-tains 10 times less Pth activity than wild type cells [16]. Thestrain LDO1, a ΔTn10 derivative of P90C pth(rap) obtained bythe method described by Bochner et al. [17], was used to con-struct LDO2 by cotransduction of ssrA::Kan from strainW3110 ΔssrA [18] using P1vir. Unless different growth condi-tions are indicated, bacterial cultures were grown at 32 °C inLuria–Bertani medium [19]. When required, the medium wassupplemented with ampicillin (Amp) 100 μg/ml or chloram-phenicol (Cm) 34 μg/ml.
2.2. Plasmid constructs
The minigenes used in this work were previously describedin [9]. placZ plasmid containing lacZ gene was constructed byPCR amplifying lacZ gene from pLEX/lacZ vector (Invitro-gen). Primers 5-LXR (5′-GCTAGAATTCATGACCGTACCCGTCGTTTTAC-3′) and 3-LXR (5′-CCGGAAGCTTATTTTTGACACC-3′), which contain EcoRI and HinDIII sites (shown
Source
Lab collection[35]This workThis work[18]
[36]gene insert
gene insert This worklacZ This workubstitutions in lacZ This worksubstitutions in lacZ This work
[37]exlacz This work
InvitrogenThis work
tutions in lacZ This workmini-ORF insert which directs the synthesis Lab collection
mini-ORF insert which encodes MH Lab collectionmini-ORF insert which encodes MI Lab collectionmini-ORF insert which encodes MK Lab collectionmini-ORF insert which encodes MT Lab collectionmini-ORF insert which encodes MG Lab collectionAUAUAA insert Lab collection [8]gene insert [38]
New England BioLabsmini-ORF insert Lab collectiontRNAArg4 gene [39]
L. Delgado-Olivares et al. / Biochimie 88 (2006) 793–800 795
underlined) were used. The amplified product was digestedwith EcoRI and HindIII. As the lacZ ORF has an EcoRI siteclose to the 3′ end, a partial digestion with this enzyme wasdone to obtain the whole gene. The digested products werecloned in the same sites of pKQV4 and the candidates wereselected by β-galactosidase (β-gal) activity, plasmid size andfinally corroborated by sequencing.
We constructed placZ2 and placZ2-3 variants containingone or two AGA codons at the second and third positions, re-spectively, by site-directed mutagenesis (Stratagene) usingeither 5′ primer 5-LXR1 (5′-CAGAATTCATGAGAGTACCCGTCGTTTTACAACG-3′) or 5-LXR1-2 (5′CAGAATTCATGAGAAGACCCGTCGTTTTACAACG-3′) and the corre-sponding complementary 3′ primers, 3-LXR1 (5′-CGTTGTAAAACGACGGGTACTCTCATGAATTCTG-3′) or 3-LXR1-2(5′-CGTTGTAAAACGACGGGTCTTCTCATGAATTCTG-3′).
For pHZ plasmid construction, the E. coli lacZ chromosomalgene was amplified by PCR, using primers 5-LZ (5′-AGGGCGCCCATATGATGACCATGATTACGGATTCAC-3′)and 3-LZ (5′-GCCCGAGGCCTGTTAATTAATTATTTTTGACAC-3′) containing NdeI and StuI restriction sites, respectively(shown underlined). The amplified fragment was restrictedwith NdeI and StuI and cloned at the same sites of pPROEX-1(Life Technologies, Inc., USA). The correct construction wasselected by β-gal expression, resistance to ampicillin, and con-firmed by sequencing. For the construction of the pHZ-MRplasmid containing the AGA minigene (AUGAGAUAA), pri-mers 5-MR (5′-CTAGTTTCACACAGGAAACAGAATTCATGAGATAAA-3′) and 3-MR (5′-AGCTTTTATCTCATGAATTCTGTTTCCTGTGTGAAA-3′), were hybridized in buf-fer MT (50 mM MgCl2 and 50 mM Tris–HCl pH 7.0) andheated at 65 °C for 10 min. The hybrid was restricted with SpeIand HinDIII (shown underlined) and cloned at the same sites ofpHZ, downstream lacZ gene.
Mutants pHZ353 and pHZ353-MR containing an AGA co-don at position 353 and mutants pHZ352-353 and pHZ352-353-MR containing two AGA codons at positions 352 and353, were derived from pHZ and pHZ-MR plasmids by site-directed mutagenesis, using primers 5-ZR353 (5′-GCCGTTGCTGATTAGAGGCGTTAACCGTCACG-3′), 3-ZR353 (5′-CGTGACGGTTAACGCCTCTAATCAGCAACGGC-3′) and5-ZR352-353 (5′-GGCAAGCCGTTGCTGAGAAGAGGCGTTAACCGTCACG-3′), 3-ZR352-353 (5′-CGTGACGGTTAACGCCTCTTCTCAGCAACGGCTTGCC-3′), respectively. Theprimers 5-ZRS (5′-CAAGCCGTTGCTGATTAGATAAGTTAACCGTCACGAGC-3′) and 3-ZRS (5′-GCTCGTGACGGTTAACTTATCTAATCAGCAACGGCTTG-3′) were used forthe construction of the pHZS and pHZS-MR mutants, whichcontain a stop codon at position 354. All the substitutions indi-cated here, correspond to E. coli chromosomal lacZ gene se-quence positions.
All plasmid constructs were transformed in the indicatedE. coli strains by RbCl2 method and the expression of proteinswas under the control of the IPTG-inducible Ptcr promoter.
2.3. Analysis of the β-galactosidase incomplete peptide
LDO2 cells transformed with plasmids pHZ, pHZ-MR andderivatives were grown in 100 ml of supplemented M9 med-ium to 0.5 DO600 and then induced with 1 mM IPTG for40 min. The cells were harvested by centrifugation (3000 × g)and resuspended in 1 ml lysis solution (50 mM NaH2PO4,300 mM NaCl and 10 mM imidazole, pH 8.0) supplementedwith a protease inhibitor cocktail (Sigma-Aldrich ChemicalCo., Missouri, USA). Cells were lysed by sonication and thecellular debris was removed by centrifugation. Forty microli-ters of supernatant were resolved through an 8.0% SDS-PAGE[20]. An immunodetection assay was carried out as previouslydescribed by Cruz-Vera et al. [16] using rabbit anti-His antibo-dies (Sigma-Aldrich Chemical Co.). The signal was amplifiedwith a second anti-rabbit IgA antibody coupled to horseradishperoxidase and finally revealed using the fluorescent reagent(ECL Western blot, of Amersham Pharmacia Biotech). For β-lactamase (β-lac) detection, C600 rap cells, transformed withpKQV4 or derivatives (pAGA or pAGG), were induced with1 mM IPTG for 40 min. The cells were processed and β-lacwas detected with specific rabbit polyclonal antibodies as de-scribed above.
2.4. Cell-free transcription–translation
Fifty microliter reactions were prepared with 2 mg of plas-mid DNA in TE (10 mM HCl, pH 8.0, 1 mM EDTA), 20 ml ofpremix [87.5 mM Tris-Ac pH 8.0, 476 mM potassium gluta-mate, 75 mM NH4(Ac), 5 mM dithiothreitol, 20 mM Mg(Ac)2,1.25 mM each of 20 amino acids, 5 mM ATP, 1.25 mM eachof CTP, UTP, GTP, 50 mM phosphoenol pyruvate, 2.5 mg/mlE. coli tRNA, 87.5 mg/ml polyethylene glycol (8000 Mr),2.5 mM cAMP, 50 mg/ml folinic acid] and 15 μl of S30 pre-pared as previously described [8,21]. Radiolabeled proteinswere synthesized using a premix lacking methionine. 10 μCiof [35S]-methionine (1 170 Ci/mmol) were then added to a finalconcentration of 2 × 10−11 M. Protein was precipitated from15 μl reactions by addition of 60 μl of acetone followed bycentrifugation. The protein pellets were dried and re-suspendedin SDS sample buffer prior to gradient polyacrylamide gelelectrophoresis [20]. Immediately following electrophoresis,wet gels were dried and then visualized by autoradiography.When indicated, dried gels were analyzed in a Typhon radio-activity scanner (Molecular Dynamics). The electronic densityvalues (pixels), which correspond to radioactivity levels, wereused to determine the percentages of protein synthesis inhibi-tion.
3. Results
3.1. Minigene-mediated inhibition of protein synthesis elicitedby starvation for specific tRNA
Toxic minigenes, which inhibit cell growth and proteinsynthesis, release pep-tRNA due to an abortive translation ter-
L. Delgado-Olivares et al. / Biochimie 88 (2006) 793–800796
mination [7–9]. The expression of toxic two-codon minigenes,under limiting Pth activity, leads to starvation for aminoacyl-able tRNA specific for the second codon and to a general in-hibition of protein synthesis. In addition, the degree of celltoxicity correlates with the relative rate of pep-tRNA accumu-lation [9,12]. Therefore, we analyzed whether minigenes pre-ferentially inhibit protein synthesis of those genes containingthe codons that are present in the last sense codon of the inhi-biting minigene. To this aim, we first ranked the degree of celltoxicity of various two-codon minigenes by transforming theappropriate constructs into Pth-defective cells. We then se-lected some of these minigenes to examine in an in vitro tran-scription–translation system the likelihood that bacterial genescontaining codons read by the minigene-depleted tRNAswould be preferentially inhibited over those lacking these tri-plets. A Pth-defective synthesizing S30 cell-free extract wasused to allow the minigene-mediated sequestration of tRNAas pep-tRNA. The synthesis was carried out in the presenceof [35S]-methionine, and the synthesized proteins were re-solved by SDS-PAGE and visualized by autoradiography. Asto the first point, the cells transformed with the constructs har-boring synthetic two-codon minigenes, were induced for mini-gene expression. A 200-min cell growth culture (Fig. 1)showed the order of minigene inhibition: AUG AAAUAA > AUG ACG UAA > AUG CAC UAA > AUG AGAUAA > AUG AGG UAA > AUG GGG UAA. The strengthof the inhibition correlated with the rate of peptidyl-tRNA ac-cumulation in the cases for which it had been previously deter-mined in [9]. The AAA minigene promotes the buildup of pep-tRNALys proportionally faster than the AGA and CAC mini-genes mediate the accumulation of pep-tRNAArg4 and pep-tRNAHis, respectively. The GGG minigene does not accumu-late the cognate pep-tRNA and the levels of pep-tRNA accu-mulated by minigene AUA have not been determined. We then
Fig. 1. The expression of tRNA-sequestering minigenes inhibits Pth-defectivecell growth. pth(rap) cells were transformed with the indicated plasmidscontaining tRNA-sequestering minigenes described in Table 1. Cells werecultured at 38 °C on LB-Ap medium and minigene expression was inducedwith 2 mM IPTG. At the indicated time intervals, samples were drawn tomeasure OD at 600 nm. Cell growth control, without minigene expression, isshown for cells harboring the tRNAIle2-sequestering minigene (pAAA).
analyzed the effect AUA and AGG minigenes, which areamong the most toxic, on the concurrent expression of β-lacand polynucleotide phosphorylase (PNPase) in a Pth-defectivecell-free in vitro protein synthesis system. ampC gene, whoseORF encodes β-lac, contains four AUAs and no AGG codons[22] whilst the bacterial pnp ORF contains two AGGs but notAUA codons. The used plasmids carried either the mini-ORFAUG AUA UAA or AUG AGG UAA, which have beenshown to inhibit protein synthesis [8,23] and to be toxic inPth-defective cells [9]. The results (Fig. 2A) show that β-lacprotein was significantly inhibited by the AUA but not byAGG minigenes. The opposite was observed with the AGGminigene: a greater inhibition of PNPase as compared withthe β-lac protein. The control with an inactive minigene(p101) did not affect the synthesis of PNPase, however, a mod-erate reduction of β-lac is observed even without minigene(pKQV4). This inhibition is analyzed in next section. Thatthe codon-targeted inhibition was due to the depletion of a spe-cific tRNA, was supported by an experiment performed invivo. β-lac, which contains AGA but not AGG coded arginine,was inhibited by the AGA minigene but not by the AGG mini-gene and the inhibition was reverted by the supplementation oftRNAArg4 from a plasmid construct (Fig. 2B). As expected,AGA minigene also inhibited β-lac synthesis in vitro (Fig. 2A).These results show that minigene expression precipitates thecodon-targeted inhibition of protein synthesis by starvationfor the specific tRNA.
3.2. Toxic minigenes cause moderate in vitro inhibitionof protein synthesis irrespective of the presence of hungrycodons in the genes
The tRNA-sequestering mechanism, spelled out aboveeither in vivo or in vitro, explains the inhibition of those genescontaining codons whose tRNA is depleted by the minigeneexpression. However, we have previously observed a generalin vivo inhibition of protein synthesis, irrespective of the codoncontent [23]. In an attempt to further analyze the in vitro sce-nario, we examined the expression of a gene lacking the codonwhose cognate tRNA is sequestered by the toxic minigene. Wetested in vitro the effect of the AUA minigene on the expres-sion of the pnp gene. As mentioned above, the reading frameof pnp does not contain AUA codons. The inhibition ofPNPase synthesis in an in vitro Pth-defective system, althoughbarely visible under the coincident addition of the minigeneand pnp harboring constructs (Fig. 2C, compare lanes b andd), it became evident when the tRNAIle2-sequestering minigenewas added 15 min prior to the addition of the pnp reporter(Fig. 2C, compare lanes c and e). The lower yield of PNPasein this last instance was the result of the minigene expressionand not the result of the natural decay of the S30 proteinsynthesizing activity, because a 15 min pre-incubation periodeither in the absence of a minigene construct or in the presenceof one of the least toxic minigenes (pGGG), PNPase level wasnot reduced to the same degree (Fig. 2C, lanes e and f). Similarresults were obtained when a lacZ construct not containing
Fig. 2. Reporter genes containing hungry codons are preferentially inhibited.(A). Codon-targeted inhibition of β-lac and PNPase by the tRNA-sequestering minigenes. In vitro reaction mixtures containing S30 extracts from Pth-defective cellspth(rap) were primed with the indicated plasmids. After 1 h incubation at 37 °C the products were processed as described in Section 2. pAUA, pAGA, pAGG andp101 are pKQV4 derivatives containing the β-lac-encoding ampC gene and the AUA, AGA, AGG or 101 minigenes, respectively. The in vitro reactions were primedwith the minigene construct DNAs alone or in combination with the pPNP construct as indicated. The migration positions of polynucleotide phosphorylase (PNPase)and β-lac are indicated.(B). Specific in vivo inhibition of β-lac, by the AGA minigene. Pth-defective cells, C600 pth(rap), were transformed with the indicated plasmids. pArg4 is a pACYC-184 derivative containing tRNAArg4 gene. The cells were treated with 2 mM IPTG for minigene expression and processed as described in Section 2. β-lac wasdetected by Western blot using specific polyclonal antibodies. The migration position of β-lac is indicated (C). PNPase synthesis is unspecifically inhibited by atRNAIle2-sequestering minigene (pAUA). The in vitro reactions were primed with the indicated plasmids from the beginning, except lanes c, e and f; where pPNPwas added 15 min after the start of the reaction. pAUA and pGGG are pKQV4 derivatives containing AUA and GGG minigenes, respectively. The migrationposition of PNPase is indicated.
L. Delgado-Olivares et al. / Biochimie 88 (2006) 793–800 797
AUA codons or the wild type E. coli essential pth gene, whichdoes not contain AUA codons, were added 15 min after thetRNAIle2-sequestering minigene (data not shown). These re-sults suggest that, if the minigene mRNA is allowed to com-promise the translational machinery ahead of the reporter gene,a modest non-specific inhibition of protein synthesis ensued,which even affected mRNAs lacking hungry codons. It shouldalso be noticed the reduction of β-lac synthesis when co-ex-pressed with PNPase (Fig. 2A, last lane). A similar reductionof protein synthesis has been observed when β-lac is expressedat the same time with cat gene in the absence of a tRNA-se-questering minigene (data not shown). These results show thatthere is an inhibition of protein synthesis, regardless of the co-don content, presumably caused by the competition for the invitro S30 components.
3.3. The inhibition of protein synthesis shows dependencyon the number of hungry codons
The nature of the codons at early positions after the initia-tion codon has a strong effect on gene expression. For instance,the codon following the initiation codon, +2 position, affectsthe expression of modified lacZ genes 15-fold [24,25]. AlsoNGG codons, where N is non-G, are associated with a verylow gene expression if located at positions +2, +3 and +5 butwith higher expression if placed a few codons further down-stream [26]. Lastly, int gene of bacteriophage lambda whichnaturally contains the minor arginine AGA and AGG codonsat positions 3 and 4 of the reading frame is expressed poorly
[27,28]. However, the substitution of the favorable AAA codonfor these codons at the same positions results in the enhancedproduction of the Int protein [28]. We took advantage of theseobservations to examine the effect of the toxic AGA minigeneon the expression of modified variants of the lacZ reportergene. The lacZ variants containing no AGA codons (placZ),one AGA codon at position 2 (placZ2), or two AGA codonsat positions 2 and 3 (placZ2-3), were expressed in vitro in thepresence, or not, of the tRNAArg4-sequestering AGA minigeneharbored in the pAGA construct. Although wild type β-gal ex-pression is slightly inhibited by the minigene, the lacZ con-structs containing one or two AGAs show higher levels of pro-tein synthesis inhibition (60% and 80%, respectively, Fig. 3B).These results suggest that the minigene-mediated inhibition ofprotein synthesis depends on the number of hungry codonscontained in the reporter gene.
3.4. AGA minigene induces a pause at AGA codons duringpolypeptide chain elongation
The above results predict that elongating ribosomes wouldstop at codons specific for the tRNA depleted by the minigeneand, therefore, that an incomplete protein would be generated.To analyze whether this inference is correct, the construct pHZ,a derivative of pPROEX-1 harboring the NdeI/StuI chromoso-mal lacZ gene insert, was engineered to yield pLZ-MR. In thisconstruct, lacZ was tagged with six histidine codons next to the5′-end and the two-codon AGA minigene was inserted beyondlacZ (see Section 2). Therefore, the transcript that initiates at
Fig. 3. The inhibition of protein synthesis by a tRNA-sequestering minigenedepends on the number of hungry codons in the reporter gene.(A) In vitro transcription–translation reactions, carried out as described inSection 2, were directed by: placZ, (a, b); placZ2 (c, d) and placZ2-3 (e, f).AGA minigene (pAGA) was also expressed in lanes b, d, and f or pKQV4 inlanes a, c and e. After 1 h incubation, samples were processed as described inSection 2. The mobility of β-gal is indicated. (B) Percentages of proteinsynthesis inhibition were determined as indicated in Section 2.
L. Delgado-Olivares et al. / Biochimie 88 (2006) 793–800798
the IPTG-inducible promoter ptrc contains both the histidinecodon-tagged lacZ and the AGA minigene messengers. Inaddition, single or double AGA substituted lacZ variants(lacZATT352AGA, CGA353AGA) were derived from pLZ-MR asdescribed in Section 2. These constructs were expressed inΔssrA cells to favor the stability of the elongating ribosomalcomplexes and the detection of the incomplete peptides [18].The nature of the β-gal antigenic material was analyzed by im-munoblot assays with anti-His antibodies. The results showedthat the concurrent expression of the AGA-substituted lacZgenes and the AGA minigene yielded an incomplete β-gal pro-tein, the proportion of which, relative to the complete protein,was larger for the double than for the single AGA substitutedvariants (Fig. 4 lanes b and c). The wild type lacZ gene did notproduce the incomplete protein (lane a, Fig. 4). The size of theincomplete protein corresponded to the size expected from atranslational arrest at the site of the AGA substitutions asshown by the size of a fragment generated by the introduction
Fig. 4. AGA minigene induces a pause in bacterial polypeptide chainelongation at AGA codons of lacZ mRNA. ΔssrA, Pth-defective cells weretransformed with a plasmid containing AGA minigene and the lacZ gene with a5′ His-tag or any of the following variants: wild type lacZ (a); lacZ CGA353AGA
(b); lacZATT352AGA/CGA353AGA (c); and lacZCGA353AGA/GGG354TAA (d). Cellswere induced and processed as described in Section 2. Tagged β-gal and theincomplete peptide were detected by Western blot using specific His-tagmonoclonal antibodies. The migration position of β-gal and the incompletepeptide are indicated.
of a termination codon at position 354 of the lacZ gene (Fig. 4,lane d). No incomplete β-gal peptide was detected when theAGA-substituted lacZ gene was expressed in the single Pthdefective, wild type ssrA gene or when the tRNAArg4-seques-tering AGA minigene was not co-expressed (data not shown).This result shows that the specific elongation inhibition occursat the AGA codon, probably as a result of ribosome pausingdue to tRNAArg4 scarcity. Furthermore, the severity of the in-hibition mediated by the AGA minigene was proportional tothe number of AGA codons.
4. Discussion
We have shown that the expression of toxic minigenes in-hibits primarily the synthesis of proteins encoded by genes thatcontain the codon recognized by the tRNA which has beensequestered by the minigene. This specific or codon-targetedinhibition may result from ribosomes pausing at the codonsstarved for aminoacyl-tRNA i.e. hungry codons, as indicatesthe production of an incomplete peptide of the expected sizeby an AGA-substituted lacZ construct. In addition to the co-don-targeted inhibition, a non-specific in vitro inhibition ofprotein synthesis was also recognized.
The minigenes used to assess the codon-targeted inhibitionof protein synthesis were chosen from a group of minigeneswhose cell toxicity was monitored in Pth-defective cells. Thetoxicity correlated with the rate of pep-tRNA accumulation inthe cases for which it had been previously determined in [9].Among these, AAA, AGA, AGG and AUA minigenes are themost toxic [12,9]. AGG and AUA minigenes were selected toanalyze the specificity of the codon-targeted inhibition of pnpand ampC reporter genes. The expression of the pnp gene,which lacks AUA but contains AGG codons, was strongly in-hibited by the AGG minigene but not by the AUA minigene;whilst the expression of the ampC gene, that lacks AGG butincludes AUA codons, was unaffected by the AGG minigeneand inhibited by the AUA minigene. A similar effect was pro-duced by the AGA minigene on the synthesis of β-lac, whichcontains two AGA codons (Fig. 2A). These results demonstrateunambiguously a preferential inhibition of genes containing thecodons that are present in the last position of the minigene.Therefore, this type of inhibition can be targeted to any genecontaining codons similar to the last sense codon of the tRNA-sequestering minigene. The extent of the inhibition alsoshowed dependency with the number of hungry codons. Thus,two contiguous AGA codons at the positions 2 and 3 weremore inhibitory than a single AGA codon at position 2(Fig. 3). In the absence of minigenes, the over-expression ofgenes with a high content of AGA codons may result per se,in a defective synthesis of the corresponding protein in wildtype cells. Particularly, the closer the tandem of two to fiveAGG codons is placed to the translational start (up to the11th codon) the stronger the inhibition of protein synthesis[29]. This effect is averted by the over-expression of tRNA(s)cognate to the rare arginine codon(s), indicating that the inhibi-tion is related to starvation for the tRNA [30–32]. The inhibi-
L. Delgado-Olivares et al. / Biochimie 88 (2006) 793–800 799
tion seems to involve ribosome stalling as it is accompanied bypep-tRNA accumulation [28]. It is assumed that low levels inthe pool of specific aminoacyl-tRNA determine a ribosomalpause at the cognate codons of the mRNA [8] that eventuallyleads to the release of pep-tRNA. In addition, over-expressionof the test gene in Pth-defective cells, increases the above in-hibition effects [11,23]. Our results show that the concurrentexpression of a tRNA-sequestering minigene pushes furtherthe scarcity of the specific tRNA pool, inhibiting more severelythe translation of those genes containing hungry codons.
That the expression of the tRNAArg4-sequestering AGAminigene promoted ribosome pausing at the AGA codons ofan AGA substituted lacZ mRNA, is supported by the accumu-lation of the incomplete β-gal polypeptide of the expected size.The relative levels of the incomplete polypeptide are higherwith two AGA codons as compared with the one-AGA codonvariant. The expression of the AGA substituted construct incells lacking the trans-aminacylation system (ΔssrA) predictsthat once the ribosome pauses at the hungry codons duringtranslation elongation, the incomplete peptide is retained inthe ribosome, probably as pep-tRNA in the ribosomal P site.Preliminary results with anti-His-tag monoclonal antibodies re-veal that the peptide is mostly found in the ribosomal fraction.Furthermore, peptide release from the ribosome upon puromy-cin treatment suggests that it is located in the ribosomal P site(data not shown; Bueno-Martínez J. and León-Avila G).
In addition to the codon-targeted inhibition discussed above,our in vitro results evidenced a non-specific inhibition of pro-tein synthesis, initiated by minigene expression irrespective ofthe cognate tRNA levels (Fig. 2C). The inhibition was moreevident when the minigene was let to compromise with thebiosynthetic machinery ahead of the reporter gene, either inthe absence of minigenes or in the presence of one of the leasttoxic minigenes. Therefore, the inhibition may be caused bythe direct competition of both the toxic minigene and the re-porter gene for the transcriptional and translational machinery.In this regard, a not specific inhibition of β-lac is also observedwhen co-expressed in vitro with PNPase (Fig. 2, lanes 4, 5).Indeed, protein synthesis of many genes, including ampC, isnoticeably reduced when co-expressed in vitro with cat gene(data not shown). We have observed that this inhibition de-pends on the strength of the transcriptional and translationalsignals of the gene tested and is usually lower than the co-don-targeted inhibition (data not shown). This type of inhibi-tion may account partially for the previously reported generalin vivo inhibition of protein synthesis by toxic minigenes [23].However, it is possible that the synthesis of proteins not con-taining hungry codons could be indirectly inhibited in the cellby the tRNA-sequestering mechanism. For example, it is likelythat the AAA minigene is very toxic (Fig. 1) primarily becauseit inhibits the synthesis of a great number of E. coli proteins.The frequency of the lysine codons AAA and AAG, read bythe unique tRNALys, is among the highest in E. coli genes [33].The pth gene, which is essential for E. coli, contains 12 AAAand one AAG codons [3]. Indeed, the synthesis of Pth proteinis inhibited by the AAA minigene (S. Vivanco-Dominguez,
G. Guarneros, unpublished results). The tRNA-recycling func-tion of Pth declines as the levels of Pth protein are reduced dueto the depletion of the lysyl-tRNA pool by the minigene. Thisin turn leads to a reduction of the whole pool of aminoacylabletRNAs and to a global inhibition of protein synthesis in thecell. That the non-specific inhibition observed in vivo also de-pends on the codon-targeted inhibition is supported by the re-versibility of the whole inhibition by the supplementation ofPth or the specific tRNA [8,11,23]. A similar scenario couldarise with the AGA minigene. A computer search of rareAGA/AGG codons in E. coli genes showed that they are pre-ferentially located within the first 25 codons. More than 100genes containing a single AGG or AGA codon within the first25 codons are associated with essential functions [34]. The in-hibition of these genes by the AGA minigene should reason-ably lead to a general inhibition of protein synthesis and cellgrowth, irrespective of the codon content.
From the above results, a two-stage minigene-mediatedmodel of protein synthesis inhibition in vivo seems to emerge.Initially the sequestration of a specific tRNA, which provokesribosome pausing at the cognate triplets in the messengerRNA, leads to the specific inhibition of protein synthesis;thereafter, a more general protein synthesis inhibition ensuesperhaps due to the codon-targeted inhibition of essential andregulatory genes. This scenario applies partially to the in vitroinhibition of protein synthesis by minigenes. In this case, inaddition to the codon-targeted inhibition, an unspecific inhibi-tion was also observed, regardless of the codon content of thereporter gene. This probably resulted from the competition forthe limited components of the S30 system.
Acknowledgments
This work was supported by CONACyT grants 34836-Nand 28401N to J.H.S. and G.G., respectively. L.D.O. and E.Z.R. were recipients of loan-fellowships from CONACyT andCOSNET. We acknowledge J.G. Bueno Martinez for technicalassistance and V. Ramirez Nieves for clerical work.
References
[1] J.R. Menninger, Peptidyl transfer RNA dissociates during protein synth-esis from ribosomes of E. coli, J. Biol. Chem. 251 (1976) 3392–3398.
[2] A.G. Atherly, J.R. Menninger, Mutant, E. coli strain with temperaturesensitive peptidyl-transfer RNA hydrolase, Nat. New Biol. 240 (1972)245–246.
[3] R. García-Villegas, F. De la Vega, J. Galindo, M. Segura, R. Bucking-ham, G. Guarneros, Peptidyl-tRNA hydrolase is involved in λ inhibitionof host protein synthesis, EMBO J. 10 (1991) 3549–3555.
[4] P. Guzman, G. Guarneros, Phage genetic sites involved in growth inhibi-tion by the E. coli rap mutant, Genetics 121 (1989) 401–410.
[5] J. Hernandez, C. Ontiveros, J.G. Valadez, R.H. Buckingham, G. Guar-neros, Regulation of protein synthesis by minigene expression, Biochi-mie 79 (1997) 527–531.
[6] T. Tenson, J. Vega, P. Kloss, G. Guarneros, A. Mankin, Inhibition oftranslation and cell growth by minigene expression, J. Bacteriol. 181(1999) 1617–1622.
L. Delgado-Olivares et al. / Biochimie 88 (2006) 793–800800
V. Dinçbas, V. Heurgue-Hamard, R.H. Buckingham, R. Karimi, M. Eh-renberg, Shutdown in protein synthesis due to the expression of mini-genes in bacteria, J. Mol. Biol. 291 (1999) 745–759.J. Hernández-Sánchez, G. Valadez, J. Vega, C. Ontiveros, G. Guarneros,λ bar minigene-mediated inhibition of protein synthesis involves accu-mulation of peptidyl-tRNA and starvation for tRNA, EMBO J. 17(1998) 3758–3765.L.R. Cruz-Vera, E. Hernández-Ramon, B. Perez-Zamorano, G. Guar-neros, The rate of peptidyl-tRNA dissociation from the ribosome duringminigene expression depends on the nature of the last decoding interac-tion, J. Biol. Chem. 278 (2003) 26065–26070.
] D. Pérez-Morga, G. Guarneros, A short DNA sequence from λ phage in-hibits protein synthesis in E. coli rap, J. Mol. Biol. 216 (1990) 243–250.
] N.A. Oviedo de Anda, L. Kameyama, J.M. Galindo, G. Guarneros, J.Hernandez-Sanchez, Evidence of bar minigene expression and tRNAIle2
sequestration as peptidyl-tRNAIle2 during lambda bacteriophage develop-ment, J. Bacteriol. 186 (2004) 5533–5537.
] J.R. Menninger, The accumulation as peptidyl-transfer RNA of isoac-cepting transfer RNA families in E. coli with temperature-sensitive pep-tidyl-transfer RNA hydrolase, J. Biol. Chem. 253 (1978) 6808–68013.
] D. Lindsley, J. Gallant, G. Guarneros, Ribosome bypassing elicited bytRNA depletion, Mol. Microbiol. 48 (2003) 1267–1274.
] J. Menez, V. Heurgue-Hamard, R.H. Buckingham, Sequestration of spe-cific tRNA species cognate to the last sense codon of an overproducedgratuitous protein, Nucleic Acids Res. 28 (2000) 4725–4732.
] J.A. Gallant, D. Lindsley, Ribosomes can slide over and beyond “hun-gry” codons, resuming protein chain elongation many nucleotides down-stream, Proc. Natl. Acad. Sci. USA 95 (1998) 3771–3776.
] L.R. Cruz-Vera, J.M. Galindo, G. Guarneros, Molecular basis for thetemperature sensitivity of E. coli pth(Ts), J. Bacteriol. 182 (2000)1523–1528.
] B.R. Bochner, H. Huang, G.L. Schieven, B.N. Ames, Positive selectionfor loss of tetracycline resistance, J. Bacteriol. 143 (1980) 926–933.
] E. Roche, R. Sauer, SsrA-mediated tagging caused by rare codons andtRNA scarcity, EMBO J. 18 (1999) 4579–4589.
] J. Sambrook, E.F. Fritsch, T. Maniatis, Molecular Cloning: A LaboratoryManual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor NY,1989 (A.1).
] J.K. Laemmli, Cleavage of structural proteins during the assembly of thehead of bacteriophage T4, Nature 227 (1970) 680–683.
] G.A. Mackie, B.C. Donly, P.C. Wong, Coupled transcription–translationof ribosomal proteins, in: G. Spedding (Ed.), Ribosomes and ProteinSynthesis, A Practical Approach, IRL Press, Oxford, 1990, pp. 191–211.
] P. Balbas, X. Soberon, E. Merino, M. Zurita, H. Lomeli, F. Valle, N.Flores, F. Bolivar, Plasmid vector pBR322 and its special-purpose deri-vatives a review, Gene 50 (1986) 3–40.
] J.G. Valadez, J. Hernández-Sanchez, M.A. Magos, C. Ontiveros, G.Guarneros, Increased bar minigene mRNA stability during cell growthinhibition, Mol. Microbiol. 39 (2001) 361–369.
] A.C. Looman, J. Bodlaender, L.J. Comstock, D. Eaton, P. Jhurani, H.A. de Boer, P.H. van Knippenberg, Influence of the codon following
the AUG initiation codon on the expression of a modified lacZ gene inE. coli, EMBO J. 6 (1987) 2489–2492.
] T. Sato, M. Terabe, H. Watanabe, T. Gojobori, C. Hori-Takemoto, MiuraKi, Codon and base biases after the initiation codon of the open readingframes in the E. coli genome and their influence on the translation effi-ciency, J. Biochem. (Tokyo) 129 (2001) 851–860.
] E.I. Gonzalez de Valdivia, L.A. Isaksson, A codon window in mRNAdownstream of the initiation codon where NGG codons give strongly re-duced gene expression in E. coli, Nucleic Acids Res. 32 (2004) 5198–5205.
] K. Zahn, A. Landy, Modulation of lambda integrase synthesis by rarearginine tRNA, Mol. Microbiol. 21 (1996) 69–76.
] J.J. Olivares-Trejo, J.G. Bueno-Martinez, G. Guarneros, J. Hernández-Sanchez, The pair of arginine codons AGA AGG close to the initiationcodon of the lambda int gene inhibits cell growth and protein synthesisby accumulating peptidyl-tRNAArg4, Mol. Microbiol. 49 (2003) 1043–1049.
] G.T. Chen, M. Inouye, Suppression of the negative effect of minor argi-nine codons on gene expression; preferential usage of minor codonswithin the first 25 codons of the E. coli genes, Nucleic Acids Res. 18(1990) 1465–1473.
] K. Chen, T.C. Peters, J.R. Walker, A minor arginine tRNA mutant limitstranslation preferentially of a protein dependent on the cognate codon, J.Bacteriol. 172 (1990) 2504–2510.
] A.H. Rosenberg, E. Goldman, J.J. Dum, F.W. Studier, G. Zubay, Effectsof consecutive AGG codons on translation in E. coli, demonstrated witha versatile codon test system, J. Bacteriol. 175 (1993) 716–722.
] A.G. Zdanovsky, M.V. Zdanovskaia, Simple and efficient method forheterologous expression of clostridial proteins, Appl. Environ. Microbiol.66 (2000) 3166–3173.
] T. Ikemura, Correlation between the abundance of E. coli transfer RNAsand the occurrence of the respective codons in its protein genes, J. Mol.Biol. 146 (1981) 1–21.
] G.T. Chen, M. Inouye, Role of the AGA/AGG codons, the rarest codonsin global gene expression in E. coli, Genes Dev. 8 (1994) 2641–2652.
] G. Guarneros, P. Guzman, E. Garay, Genetic and physical location of theE. coli rap locus, which is essential for growth of bacteriophage lambda,J. Bacteriol. 169 (1987) 5188–5192.
] D. Polayes, A.J. Hughes, Efficient protein expression and simple purifi-cation using the pPROEX-1k system, Focus 16 (1994) 81–84.
] M.A. Strauch, G.B. Spiegelman, M. Perego, W.C. Johnson, D. Burbulys,J.A. Hoch, The transition state transcription regulator abrB of Bacillussubtilis is a DNA binding protein, EMBO J. 8 (1989) 1615–1621.
] J. García-Mena, D. Asis, A. Sanchez-Trujillo, C. Portier, C. Montañez, Anovel mutation in the KH domain of polynucleotide phosphorylase af-fects autoregulation and mRNA decay in E. coli, Mol. Microbiol. 33(1999) 235–248.
] B.J. Jr, J.M. Del Tito, J. Ward, C.J.L. Hodgson, H. Gershater, L.A. Ed-wards, F.A. Wysocki, G. Watson, J.F. Sathe, Kane, Effects of a minorisoleucyl tRNA on heterologous protein translation in E. coli, J. Bacter-iol. 177 (1995) 7086–7091.
