alternative translation initiation of a haloarchaeal...
TRANSCRIPT
Alternative Translation Initiation of a Haloarchaeal Serine ProteaseTranscript Containing Two In-Frame Start Codons
Wei Tang,a Yufeng Wu,a Moran Li,a Jian Wang,a Sha Mei,a Bing Tang,a,b Xiao-Feng Tanga,b
State Key Laboratory of Virology, College of Life Sciences, Wuhan University, Wuhan, Chinaa; Hubei Provincial Cooperative Innovation Center of Industrial Fermentation,Wuhan, Chinab
ABSTRACT
Recent studies have shown that haloarchaea employ leaderless and Shine-Dalgarno (SD)-less mechanisms for translation initia-tion of leaderless transcripts with a 5= untranslated region (5= UTR) of <10 nucleotides (nt) and leadered transcripts with a5= UTR of >10 nt, respectively. However, whether the two mechanisms can operate on the same naturally occurring haloar-chaeal transcript carrying multiple potential start codons is unknown. In this study, the transcript of the sptA gene (encoding anextracellular serine protease of Natrinema sp. strain J7-2) was experimentally determined and found to contain two potentialin-frame AUG codons (AUG1 and AUG2) located 5 and 29 nt, respectively, downstream of the transcription start site. Mutationalanalysis revealed that both AUGs can function as the translation start codon for production of active SptA, although AUG1 ismore efficient than AUG2 for translation initiation. Insertion of a stable stem-loop structure between the two AUGs completelyabolished initiation at AUG1 but did not affect initiation at AUG2, indicating that AUG2-initiated translation does not involveribosome scanning from the 5= end of the transcript. Furthermore, the efficiency of AUG2-initiated translation was not influ-enced by an upstream SD-like sequence. In addition, both AUG1 and AUG2 contribute to transcript stability, probably by re-cruiting ribosomes to protect the transcript against degradation. These data suggest that depending on which of two in-framestart codons is used, the sptA transcript can act as either a leaderless or a leadered transcript for SptA production in haloarchaea.
IMPORTANCE
In eukaryotes and bacteria, alternative translation start sites contribute to proteome complexity and can be used as a functionalmechanism to increase translation efficiency. However, little is known about alternative translation initiation in archaea. Ourresults demonstrate that leaderless and SD-less mechanisms can be used for translation initiation of the sptA transcript from twoin-frame start codons, raising the possibility that in haloarchaea, alternative translation initiation on one transcript functions toincrease translation efficiency and/or contribute to proteome complexity.
In all three domains of life (i.e., Archaea, Bacteria, and Eukarya),translation of the open reading frames (ORFs) of mRNAs into
proteins is an essential process for gene expression. This processconsists of initiation, elongation, and termination phases, and theinitiation phase largely controls the rate and efficiency of proteinsynthesis (1). In comparison with other phases of translation, theinitiation phase has diverged significantly between the Archaea,Bacteria, and Eukarya. The initiation pathways characterized todate include scanning (5=-end-dependent), internal ribosomalentry site (IRES)-mediated, Shine-Dalgarno (SD)-dependent,SD-less (SD-independent), and leaderless mechanisms (2).
In eukaryotes, the translation start codon is generally identifiedby the scanning mechanism, wherein the 40S ribosomal subunitbinds to the 5= end of the mRNA with the aid of initiation factorsand then scans along the 5= untranslated region (5=UTR) until thefirst AUG codon is reached (3). Some eukaryotic and viral mRNAscontain highly structured IRESs within the 5=UTR, and initiationfactors assist in recruiting the 40S ribosomal subunit to the IRESin order to initiate translation (2). The SD-dependent initiationmechanism, initially identified in bacteria, involves the formationof base pairs between the 3= end of the 16S rRNA and the SD motifin the 5=UTR of the bacterial mRNA (4, 5). Some archaea, such asSulfolobus solfataricus, also utilize SD-dependent translation initi-ation (6). Recent studies on translation initiation in haloarchaearevealed the existence of an SD-less mechanism that operates onmRNAs with a 5=UTR that lacks a SD motif or IRES and does not
involve ribosomal scanning (7–9). In contrast to the aforemen-tioned mechanisms, the leaderless mechanism operates onmRNAs that lack or have very short 5= UTRs and contain nosubstantial ribosomal recruitment signals except the start codon(2). Translation initiation of leaderless mRNAs occurs in all threedomains of life and, because of its simplicity, is hypothesized to bethe ancestral translation initiation mechanism (10, 11). Com-pared with the canonical SD-dependent and scanning mecha-nisms, the mechanistic and physiologic details of SD-less andleaderless initiation are poorly understood (8, 9, 12).
Although translation initiation of eukaryotic mRNAs typicallyoccurs at the first AUG codon from the 5= end via the scanningmechanism, under some rare conditions the first AUG may bebypassed, and initiation can occur at downstream AUGs as well.Bypassing of the first AUG codon is termed “leaky scanning” and
Received 3 March 2016 Accepted 26 April 2016
Accepted manuscript posted online 2 May 2016
Citation Tang W, Wu Y, Li M, Wang J, Mei S, Tang B, Tang X-F. 2016. Alternativetranslation initiation of a haloarchaeal serine protease transcript containing twoin-frame start codons. J Bacteriol 198:1892–1901. doi:10.1128/JB.00202-16.
Editor: W. W. Metcalf, University of Illinois at Urbana-Champaign
Address correspondence to Bing Tang, [email protected], orXiao-Feng Tang, [email protected].
Copyright © 2016, American Society for Microbiology. All Rights Reserved.
crossmark
1892 jb.asm.org July 2016 Volume 198 Number 13Journal of Bacteriology
on July 12, 2018 by guesthttp://jb.asm
.org/D
ownloaded from
enables the production of two independently initiated protein iso-forms from one mRNA (13). Protein isoforms can also be pro-duced by alternative translation initiation on an IRES-containingmRNA (14). Alternative translation initiation, along with differ-ential promoter usage and alternative splicing, contributes to eu-karyotic proteome complexity and under selection pressure canserve as a mechanism for increasing translation efficiency (15).
The SD motif is the most important cis-acting element govern-ing translation initiation in bacterial mRNAs. Jin et al. (16) re-ported that when an SD motif is located between two potentialstart codons, initiation occurs predominantly at the second startsite; the first start site is utilized when the distance between this siteand the downstream SD motif is increased and/or when the sec-ond start site is weakened. In that study, the authors found that theSD motif is systematically avoided in the early coding region ofnatural genes in Escherichia coli, possibly reflecting a strong evo-lutionary constraint for codon usage (16). In rare cases, bacterialgene transcripts contain an internal start codon that is preceded bya strong SD motif (e.g., GGAGG); thus, alternative translationinitiation leads to the generation of two isoforms of the protein, asevidenced by the heat shock proteins ClpB (17) and ClpA (18),initiation factor 2 (IF2) (19), hemolysin (20), and lycopene cyclase(21).
Notably, the majority of haloarchaeal transcripts are leaderless,and most leadered transcripts are devoid of an SD motif (7).Moreover, the haloarchaeon Haloferax volcanii does not use theSD-dependent translation initiation mechanism, despite the pres-ence of an SD motif in the 5=UTR of some transcripts, indicatingthat an SD-less mechanism for translation initiation operates atleast in haloarchaea (8, 9, 22). Experimental studies of alternativetranslation initiation of single haloarchaeal transcripts are limited.Hering et al. (8) reported that the introduction of an additionalin-frame AUG in the 5= end of haloarchaeal transcripts in H. vol-canii results in the simultaneous use of two start sites on the samemessage, raising the possibility that alternative translation initia-tion on one mRNA occurs in haloarchaea. However, whether al-ternative translation initiation occurs extensively on haloarchaealtranscripts with multiple potential start codons is unclear. Theidentification and characterization of haloarchaeal transcriptscontaining multiple potential start codons may provide insightsinto the regulation of translation initiation and its impact on geneexpression in haloarchaea.
Many haloarchaea secrete proteases to degrade proteinsubstrates in the natural environment. Most of these proteasesbelong to the subtilisin-like serine protease (subtilase) superfam-ily, known as halolysins (23, 24). We previously cloned threehalolysin-encoding genes from the genome of the haloarchaeonNatrinema sp. strain J7, designated sptA, sptB (which are arrangedin tandem; GenBank accession no. AY800382), and sptC(GenBank accession no. DQ137266). The enzymatic properties,maturation, and secretion mechanisms of SptA and SptC havebeen characterized (25–28). As the major extracellular protease ofNatrinema sp. strain J7, SptA is secreted primarily during station-ary phase (28, 29). The SptA precursor is composed of a Tat signalpeptide, an N-terminal propeptide, a subtilisin-like catalytic do-main, and a C-terminal extension (CTE). The signal peptide me-diates the translocation of SptA across the cytoplasmic membranevia the Tat pathway, which is used for secretion of folded proteins(30) and then is cleaved by a signal peptidase. The N-terminalpropeptide, which functions as an intramolecular chaperone and
inhibitor of the mature form, is autocatalytically processed toyield the active mature enzyme composed of the catalytic domainand the CTE (26, 28). Notably, it was observed that three in-frameATGs are present at the 5= end of the sptA gene (25). The purposeof this study is to investigate whether the sptA transcript could beinitiated at multiple start codons and to probe the mechanism ofalternative translation initiation in haloarchaea. The transcript ofsptA was experimentally determined, and the potential startcodons and their flanking regions were characterized using muta-tional analysis. Translation efficiencies were calculated based onquantification of transcript and protein levels. The roles of multi-ple start codons in translation initiation and gene expression arediscussed based on our results.
MATERIALS AND METHODSStrains and growth conditions. Natrinema sp. strain J7 (CCTCCAB91141) was isolated from a salt mine in China (31). Natrinema sp.strain J7-2, a subculture of strain J7 lacking plasmid pHH205 (32, 33), wasgrown in liquid modified growth medium (MGM) with 18% total salts asdescribed previously (25). The WFD11 strain of H. volcanii, which hasbeen developed into a model species for studies of biological characteris-tics of haloarchaea (34), was used as the host for expression of the sptAgene and its mutants. H. volcanii WFD11 was grown aerobically at 37°C in18% MGM supplemented with 0.4 �g/ml of novobiocin when necessary,as described previously (25). Escherichia coli DH5� and E. coli JM110 wereused as hosts for plasmid construction and were grown at 37°C in Luria-Bertani medium supplemented with kanamycin (30 �g/ml) or ampicillin(100 �g/ml) as needed.
Determination of the 5= and 3= ends of the sptA transcript. TotalRNA was isolated from late-log-phase Natrinema sp. strain J7-2 grown in18% MGM using RNAiso Plus (TaKaRa, Otsu, Japan) according to themanufacturer’s protocol. RNA circularization, reverse transcription, andPCR amplification were carried out according to the circularized RNAreal-time PCR (RT-PCR) method (7, 35). Briefly, total RNA was self-ligated by incubating 1 to 2 �g of RNA with 20 U of T4 RNA ligase(TaKaRa), 20 U of RNase inhibitor (TaKaRa), 1� T4 RNA ligase buffer,and 0.1% bovine serum albumin (BSA) in a reaction volume of 20 �l at10°C for 16 to 18 h. After the total volume was adjusted to 400 �l, proteinswere removed by phenol-chloroform extraction, and self-ligated RNAwas precipitated with ethanol following standard protocols (36). Afterremoval of genomic DNA, self-ligated RNA was converted into cDNAwith the sptA-specific primer SptART (data not shown). Genomic DNA(gDNA) removal and reverse transcription of self-ligated RNA were car-ried out using a PrimeScript RT reagent kit with gDNA Eraser (PerfectReal-Time; TaKaRa) according to the manufacturer’s instructions. ThecDNA was then amplified with the sptA-specific primer pair P1/P2, fol-lowed by a second PCR with the nested primer pair N1/N2 (data notshown). The second PCR product was inserted into the NcoI-BamHIrestriction site of the vector pET26b (Novagen, Darmstadt, Germany).The resulting plasmid was isolated from E. coli DH5� and used for DNAsequencing to determine the 5= and 3= ends of the sptA transcript.
Plasmid construction and mutagenesis. Genomic DNA of Natrinemasp. strain J7-2 was prepared according to the method of Kamekura et al.(37). The oligonucleotide primers used for plasmid construction areavailable upon request (data not shown). Plasmids pST1 and pST2 wereconstructed by replacing the bop promoter in pSY1 (38) with the sptA andfsptA promoters, respectively. Note that fsptA is a mutant of sptA, andtheir promoters differ from each other at three positions (described be-low). Briefly, the pBluescript SK� fragment was amplified from pSY1 byPCR using the forward primer SK-f and the reverse primer SP-r or FP-r,which contain the complementary sptA or fsptA promoter sequence, re-spectively (data not shown). The PCR products were digested with XbaIand NdeI and then separately inserted into the XbaI-NdeI restriction siteof pSY1, yielding plasmid pST1 or pST2, containing the sptA or fsptA
Alternative Translation Initiation in Haloarchaea
July 2016 Volume 198 Number 13 jb.asm.org 1893Journal of Bacteriology
on July 12, 2018 by guesthttp://jb.asm
.org/D
ownloaded from
promoter (data not shown) (Fig. 1). For the construction of other plas-mids, target DNA fragments were amplified from genomic DNA by PCRusing the appropriate primers (available upon request) and then insertedinto the NdeI-NcoI restriction site of pSY1, pST1, or pST2 (data notshown). Subsequently, the QuikChange site-directed mutagenesismethod (39) was employed to generate mutant constructs (data notshown) using the appropriate primers (available upon request). The se-quences of all recombinant plasmids were confirmed by DNA sequencing.
Determination of transcript levels. Recombinant plasmids were am-plified in E. coli JM110 and then transferred into H. volcanii WFD11 (40).Total RNA was isolated from mid-log-phase (optical density at 600 nm[OD600] of �1.2) H. volcanii WFD11 transformants grown in 18% MGMcontaining 0.4 �g/ml novobiocin using RNAiso Plus (TaKaRa) accordingto the manufacturer’s protocol. Genomic DNA removal and reverse tran-scription were carried out using a PrimeScript RT reagent kit with gDNAEraser (Perfect Real Time; TaKaRa) according to the manufacturer’s in-structions. Quantitative real-time PCR (qRT-PCR) was carried out in a20-�l reaction mixture containing 9.5 �l of diluted cDNA sample, 0.25�M gene-specific primer pair, and 10 �l of 2� iTaq Universal SYBR greenSupermix (Bio-Rad, Hercules, CA). The RT-PCR protocol consisted of a3-min initial denaturation step at 95°C, followed by 44 cycles of 95°C for10 s and 60°C for 30 s. The sptA transcript levels were determined using the
primer pair SptAQRT-f/SptAQRT-r (data not shown). As an internalcontrol, 16S rRNA transcript levels were determined using the primer pair16srRNAQRT-f/16srRNAQRT-r (data not shown). The qRT-PCR resultswere analyzed using the threshold cycle (2���CT) method (41). For quan-tification of sptA transcript levels, 16S rRNA transcript CT values wereused to normalize the CT values of the sptA transcripts. Transcript levels ofthe plasmid-carried novobiocin resistance gene (novR) were also deter-mined by qRT-PCR using the primer pair NovRQRT-f/NovRQRT-r (datanot shown), and the resulting CT values were used to normalize the CT
values of the sptA transcripts for comparison of the stabilities of tran-scripts of different constructs.
For the mRNA decay assay, the mid-log-phase culture of the H. volca-nii WFD11 transformant was supplemented with 100 �g/ml of actinomy-cin D, which has been shown to inhibit transcription in Haloferax medi-terranei (42) and Halobacterium salinarum NRC-1 (43). Culture sampleswere taken immediately after the addition of actinomycin D (time point 0min) and 10, 20, and 30 min later. The samples were directly mixed withequal volume of ice-cold medium in precooled centrifuge tubes and cen-trifuged at 8,000 � g for 5 min at 4°C to collect the cells. Total RNA wasisolated using RNAiso Plus, and mRNA levels were determined by qRT-PCR as described above. 16S rRNA was used as an internal control tonormalize the data.
FIG 1 Analysis of the sptA promoter and transcript. (A) Nucleotide sequence of the 5= end of sptA. The three in-frame ATGs (ATG0, ATG1, and ATG2) are boxed.The cis-acting elements (BRE motif, TATA box, WW motif, and the pyrimidine at position �1) of haloarchaeal promoters are underlined. Nucleotides arenumbered from the transcriptional start site (�1). The N-terminal methionine residues (M1 and M2) of the two SptA precursor isoforms and the Tat motif(DRRSLL [underlined]) in the signal peptide are shown. (B) Sequence of the PCR product of the circularized sptA transcript. The start and stop codons of sptAare boxed, and the ligation point of the 5= and 3= ends of the transcript is indicated by an arrow. (C) Restriction map of plasmid pSY1. The promoter sequencesof plasmids pST1 and pST2 are shown. In the fsptA promoter sequence, the three nucleotides that differ from the sptA promoter sequence are marked in red. (Dand E) Promoter activity assay. Culture supernatants of different H. volcanii transformants were collected at different time points and subjected to -galactosi-dase activity (D) and azocaseinolytic activity (E) assays. Values are expressed as means standard deviations (error bars) from three independent experiments.(F) Detection of recombinant enzymes. Recombinant SptA proteins in the supernatants of mid-log-phase cultures were detected by anti-His tag immunoblotanalysis; the band of mature SptA is indicated by an arrowhead (upper panel). Halo formation around transformant colonies grown on skim milk agar plateindicates SptA production (lower panel). M, molecular mass markers.
Tang et al.
1894 jb.asm.org July 2016 Volume 198 Number 13Journal of Bacteriology
on July 12, 2018 by guesthttp://jb.asm
.org/D
ownloaded from
Prediction of RNA secondary structure and free energy. The Mfoldprogram (44) was used for in silico prediction of RNA secondary structureand free energy. (The Mfold program is available at http://mfold.rna.albany.edu/?q�mfold/RNA-Folding-Form.)
Enzymatic activity assay and analysis of translation efficiency. H.volcanii WFD11 transformants were grown in 18% MGM containing 0.4�g/ml novobiocin, and samples were withdrawn from cultures at varioustime points. Cells and culture supernatants were separated by centrifuga-tion at 10,000 � g for 10 min at 4°C. Cells of transformants harboringrecombinant plasmids carrying the bgaH gene, which encodes a halophilic-galactosidase, were washed three times with buffer A (50 mM Tris-HCl,2.5 M NaCl, pH 7.2) and then sonicated on ice in buffer B (50 mM Tris-HCl, 2.5 M NaCl, 10 �M MnCl2, 0.1% -mercaptoethanol, pH 7.2). Sub-sequently, the cell extract was separated from cell debris by centrifugationat 13,400 � g for 10 min at 4°C and used for determination of -galacto-sidase activity according to the method by Holmes et al. (45). The culturesupernatants of transformants harboring recombinant plasmids carryingthe sptA gene were assayed for azocaseinolytic activity as described previ-ously (28).
The skim milk plate assay was also used for detection of recombinantSptA produced by the transformants. H. volcanii WFD11 transformantswere grown on 18% MGM agar plate containing 1% skim milk and 0.4�g/ml novobiocin at 37°C for 7 days, SptA-producing capacities of thetransformants were detected by halo formation around the colonies.
The proteolytic activity of mid-log-phase culture supernatants wasassayed using N-succinyl-Ala-Ala-Pro-Phe-p-nitroanilide (suc-AAPF-pNA) (Sigma, St. Louis, MO) as the substrate, as described previously(28). Briefly, enzymatic hydrolysis of suc-AAPF-pNA was carried out at40°C in buffer C (50 mM Tris-HCl, 3 M NaCl, 10 mM CaCl2, pH 8.0)containing the substrate (0.2 mM). The initial velocity of suc-AAPF-pNAhydrolysis was monitored at 410 nm using a thermostat-controlled spec-trophotometer (Cintra 10e; GBC, Victoria, Australia), and the level ofactivity was calculated based on a pNA extinction coefficient of 8,480 M�1
cm�1 at 410 nm. One unit of enzyme activity was defined as the amount ofenzyme needed to produce 1 �mol of pNa per min under the assay con-ditions used. Translation efficiency was calculated by dividing the mea-sured proteolytic activity value (units per milliliter/OD600) by the tran-script level.
SDS-PAGE and immunoblot analysis. Sodium dodecyl sulfate-poly-acrylamide gel electrophoresis (SDS-PAGE) was performed using theTris-glycine system (46). To prevent self-degradation of the protease dur-ing sample preparation, proteins were precipitated with 20% (wt/vol)trichloroacetic acid, washed with acetone, solubilized in loading buffercontaining 8 M urea, and then subjected to SDS-PAGE without prior heattreatment. An anti-His tag monoclonal antibody (Novagen) was used forimmunoblot analysis as described previously (47).
RESULTSDetermination of the promoter and transcript of the sptA gene.There are three in-frame ATGs (designated ATG0, ATG1, andATG2) at the 5= end of the sptA gene (Fig. 1A). We previouslyfound that H. volcanii transformants harboring a pSY1-derivedplasmid that contains the nucleotide sequence of sptA startingfrom ATG0 rather than from ATG1 efficiently produce recombi-nant SptA (25), suggesting that the region between ATG0 andATG1 plays an important role in sptA expression. Sequence anal-ysis of this region showed that it contains a transcription factor Brecognition element (BRE) motif, a TATA box, a WW motif, anda pyrimidine at position �1 (Fig. 1A), which are highly conservedcis-acting elements of haloarchaeal promoters (7, 9). Therefore, itis very likely that the sequence between ATG0 and ATG1 acts as apromoter and that the transcription of sptA is initiated down-stream of ATG0. To test this possibility, cellular mRNAs were iso-lated from Natrinema sp. strain J7-2 and circularized, and the
ligation point of the 5= and 3= ends of the sptA transcript wasdetermined by DNA sequencing (Fig. 1B). The results showed thatthe transcriptional start site of sptA was located 5 bp upstream ofATG1 (Fig. 1A).
Subsequently, plasmid pST1 (which carries bgaH as a reportergene) was constructed by replacing the bop promoter in pSY1 withthe 48-bp DNA fragment of sptA starting from ATG0 to the nu-cleotide immediately upstream of ATG1, with a mutation of ATG0
to ATC (Fig. 1C). Cell extracts of H. volcanii transformants har-boring pST1 exhibited higher -galactosidase activity than trans-formants harboring pSY1 (Fig. 1D). We also constructed plasmidspSY1-SPTA and pST1-SPTA by replacing the reporter gene bgaHin pSY1 and pST1 with the sptA ORF starting from ATG1. Culturesupernatants of transformants harboring pST1-SPTA exhibitedhigher proteolytic activity than transformants harboring pSY1-SPTA (Fig. 1E). Furthermore, immunoblot analyses and skimmilk plate assays showed that transformants harboring pST1-SPTA produced higher levels of SptA than transformants harbor-ing pSY1-SPTA (Fig. 1F). These results demonstrate that the se-quence between ATG0 and ATG1 functions as the sptA promoter,with higher activity than the bop promoter, thereby allowing tran-scription of sptA to be initiated at the site 5 bp upstream of ATG1.
We also found a halolysin-encoding gene (designated fsptA;GenBank accession no. DQ137267) previously cloned from theNatrinema sp. J7-2 mutant strain F7, which exhibits higher extra-cellular proteolytic activity than strain J7-2 (unpublished data).The fsptA promoter carries three point mutations in comparisonwith the sptA promoter (Fig. 1C). The sptA promoter in pST1 andpST1-SPTA was replaced with the fsptA promoter to generateplasmids pST2 and pST2-SPTA, respectively. Compared with H.volcanii transformants harboring pST1 or pST1-SPTA, transfor-mants harboring pST2 or pST2-SPTA produced higher levels oftarget proteins, as evidenced by enzymatic activity assays (Fig. 1Dand E), immunoblot analyses, and skim milk plate assays (Fig. 1F).These data indicate that the fsptA promoter is stronger than thesptA promoter; therefore, the fsptA promoter was used for theexpression of sptA and its mutants in this study.
Translation of SptA can be initiated at two in-frame startcodons. Haloarchaeal transcripts can be categorized into twogroups: leadered transcripts with a 5= UTR of �10 nt and leader-less transcripts with a 5=UTR of �10 nt (7). As transcription of thesptA gene in Natrinema sp. strain J7-2 was initiated at sites 5 and29 bp upstream of ATG1 and ATG2, respectively (Fig. 1A), it ispossible that the sptA transcript functions as either a leaderless ora leadered transcript based upon whether AUG1 or AUG2, respec-tively, serves as the translational start codon. To test this possibil-ity, three mutant constructs were generated from the SPTA con-struct either by changing one of the two start codons to AUC(M1AUC and M2AUC) or by deleting the sequence starting fromAUG1 to the nucleotide immediately upstream of AUG2 (M2)(Fig. 2A). The constructs were cloned into the NdeI-NcoI site ofplasmid pST2. Three additional nucleotides (CAU) were presentupstream of AUG1 due to the introduction of the NdeI site (Fig.2A). The resulting plasmids were introduced into H. volcanii,which was then grown in 18% MGM until mid-log phase. MatureSptA was detected upon immunoblotting of culture supernatantsof the transformants for the constructs SPTA, M1AUC, M2AUC,and M2 (Fig. 2B), but mature SptA was barely detectable in anal-yses of the cellular fractions (data not shown). Additionally, si-multaneous mutation of the two start codons to AUC (M12AUC)
Alternative Translation Initiation in Haloarchaea
July 2016 Volume 198 Number 13 jb.asm.org 1895Journal of Bacteriology
on July 12, 2018 by guesthttp://jb.asm
.org/D
ownloaded from
completely inhibited the production of SptA (Fig. 2B). These re-sults indicate that both AUG1 and AUG2 can function indepen-dently as the translation start codon for SptA and that the recom-binant enzymes are efficiently secreted from H. volcanii cells andprocessed into mature forms in the culture medium. The con-structs M1AUC and M2AUC contain 5=UTRs of 32 and 8 nt (Fig.2A) and thus function as leadered and leaderless transcripts, re-spectively, for SptA.
SptA protein levels were quantified by assaying the proteolytic
activity of the culture supernatant, and transcript levels werequantified by qRT-PCR analysis of cellular mRNA. The resultingdata were used to calculate the translation efficiency. Comparedwith SPTA and M2AUC, significant (P � 0.01) decreases wereobserved in the extracellular proteolytic activity (Fig. 2C) andtranslation efficiency (Fig. 2D) of construct M1AUC (in whichAUG1 was mutated to AUC), suggesting that translation initiationfrom AUG1 is more efficient than that from AUG2. Additionally,although constructs M2AUC and M2 had the same 8-nt 5= UTR(Fig. 2A), the translation efficiency of M2 was significantly (P �0.01) lower than that of M2AUC (Fig. 2D). These data suggest thatthe nucleotides downstream of the start codon affect the transla-tion efficiency of the sptA transcript.
We next investigated the start codon requirements for the twotranslation initiation sites of sptA transcripts. Four constructswere generated by changing the AUG start codon in M1AUC andM2AUC to GUG or UUG (Fig. 3A). Immunoblot analysis re-vealed that both GUG and UUG could drive translation initiationof leaderless (1GUG and 1UUG) and leadered (2GUG and 2UUG)transcripts in H. volcanii to produce mature SptA protein (Fig. 3B,inset). However, the translation efficiency of GUG- or UUG-con-taining transcripts (1GUG, 1UUG, 2GUG, or 2UUG) was muchlower than that of AUG-containing transcripts such as SPTA,M1AUC, or M2AUC (all P � 0.01; Fig. 2D and 3B). In the case ofthe leaderless transcripts, the efficiencies of GUG- and UUG-ini-tiated translation were only �1/340 and �1/560, respectively, ofthat of AUG1-initiated translation (compare 1GUG and 1UUG inFig. 3B with M2AUC in Fig. 2D). These results demonstrate thatboth GUG and UUG can serve as the start codon to initiate the
FIG 2 Translation initiation capacities of AUG1 and AUG2 in the sptA tran-script. (A) Sequences of the 5= ends of transcripts of different constructs ex-amined. Start codons are shown in boldface italics, and mutated nucleotidesare underlined. (B) Anti-His tag immunoblot analysis of mid-log-phase cul-ture supernatants of H. volcanii transformants harboring plasmids for differ-ent constructs. M, molecular mass markers. (C) Proteolytic activity assay usingsuc-AAPF-pNa as the substrate and qRT-PCR analysis of the levels of targetmRNAs using 16S rRNA as an internal control. Proteolytic activities andmRNA levels of mutant constructs are relative to the wild-type SPTA con-struct. (D) Translation efficiencies of the constructs after normalization toSPTA. Values are expressed as means standard deviations (error bars) fromthree independent experiments (**, P � 0.01).
FIG 3 Start codon selectivity of the sptA transcript. (A) Sequences of the 5=ends of transcripts of different constructs examined. Start codons are shown inboldface italics, and mutated nucleotides are underlined. (B) Effect of non-AUG start codons on translation efficiency. Translation efficiencies were nor-malized to the wild-type SPTA construct. Values are expressed as means standard deviations (error bars) from three independent experiments (**, P �0.01). The inset depicts anti-His tag immunoblot analysis of mature SptA inthe culture supernatants. M, molecular mass markers.
Tang et al.
1896 jb.asm.org July 2016 Volume 198 Number 13Journal of Bacteriology
on July 12, 2018 by guesthttp://jb.asm
.org/D
ownloaded from
translation of sptA transcripts at two start sites, but their transla-tion initiation efficiencies are considerably lower than thatof AUG.
Translation initiation of the sptA transcript at the secondAUG does not involve leaky scanning. To investigate whether thetranslation initiation of the sptA transcript at AUG2 occurs in aeukaryotic-like leaky scanning manner, the sequence betweenAUG1 and AUG2 was substituted for with a sequence known toform a stable stem-loop structure (�G � �18.4 kcal/mol) thattotally inhibits translation initiation at an AUG codon upstreamof the stem-loop structure in haloarchaeal transcripts (8), produc-ing construct SL (Fig. 4A). Significantly lower SptA protein pro-duction (Fig. 4B and C) and translation efficiency (P � 0.01) (Fig.4D) were observed with construct SL. Replacing AUG1 in SL withAUC (construct SLM1AUC) had minimal effect on SptA produc-tion and translation efficiency (Fig. 4B to D). However, the trans-lation of SptA was totally abolished when AUG2 in SL was mutatedto AUC (construct SLM2AUC), as neither SptA protein nor pro-teolytic activity was detected in the culture supernatant of H. vol-canii SLM2AUC transformants (Fig. 4B and C). These results in-dicate that the inserted stem-loop structure blocks translationinitiation at AUG1 but does not affect the initiation at AUG2.Therefore, initiation of SptA translation at AUG2 does not involveribosome scanning from the 5= end of the transcript.
SptA translation efficiency is not influenced by the presenceof an SD-like sequence between the two start codons. The sptAtranscript is predicted to contain an SD-like sequence (GGUGA)between AUG1 and AUG2 (Fig. 5A). The five consecutive nucleo-tides of this sequence are complementary to the 3= end of the 16SrRNAs of Natrinema sp. strain J7-2 (32) and H. volcanii (7). Toinvestigate whether this SD-like sequence plays a role in the trans-lation of SptA, mutant constructs were generated by reducing thebase-pairing capability of this sequence with the 3= end of the 16S
rRNA (Fig. 5A). SptA protein and transcript levels were then de-termined for each construct (Fig. 5B to D), and the translationefficiencies were calculated (Fig. 5E and F). No significant differ-ence in translation efficiency was observed between M1AUC andM1AUCm2, which contains two mutations in the SD-like se-quence (P 0.05) (Fig. 5E).
For efficient translation, the SD motif must be located at anappropriate distance (i.e., 3 to 10 nt in archaea [48]) upstream ofthe start codon. When the distance between the SD-like sequenceand AUG2 in M1AUC (13 nt) was shortened to 7 nt, the resultingconstruct, M1AUCd6, exhibited translation efficiency similar tothat of M1AUC (Fig. 5E). One or two mutations were also intro-duced into the SD-like sequence in M2AUC, and the translationefficiencies of the resulting constructs, M2AUCm1 andM2AUCm2, were similar to that of M2AUC (Fig. 5F). These re-sults suggest that the predicted SD-like sequence does not affectthe translation of the sptA transcript initiated from AUG1 orAUG2.
The two start codons and their flanking regions are impor-tant for transcript stability. In addition to translation efficiency,the total mRNA level also regulates gene expression. In this study,the fsptA promoter was employed for transcription of the con-structs. Given that constructs using the same promoter exhibitedsimilar transcription efficiencies, the total mRNA level for eachconstruct was dependent on the plasmid copy number and tran-script stability. To control for the possibility that plasmid copynumber variations among the different constructs affect compar-isons of transcript stability, the level of sptA mRNA for each con-struct relative to that of the reference novR gene located on thesame plasmid (Fig. 1C) was calculated based on qRT-PCR data. Incomparison with construct SPTA, the relative mRNA levels of allmutant constructs were significantly (P � 0.01 or P � 0.05) lower(Fig. 6A). These data suggest that the nucleotide sequence around
FIG 4 Effect of inserting a stable stem-loop structure between the two start codons on translation initiation. (A) Sequences of the 5= ends of transcripts of thedifferent constructs examined. Start codons are shown in boldface italics, and mutated nucleotides are underlined. The inserted sequence containing a stablestem-loop structure (indicated by arrows) is shaded. (B) Anti-His tag immunoblot analysis of mid-log-phase culture supernatants of H. volcanii transformantsharboring plasmids for different constructs. M, molecular mass markers. (C) Results of a proteolytic activity assay using suc-AAPF-pNa as the substrate andqRT-PCR determination of levels of target mRNAs in transformant cells using 16S rRNA as an internal control. Proteolytic activity and mRNA level data formutant constructs are relative to those of the wild-type SPTA construct. (D) Translation efficiencies of the constructs are shown after normalization to SPTA.Values are expressed as means standard deviations (error bars) from three independent experiments (**, P � 0.01).
Alternative Translation Initiation in Haloarchaea
July 2016 Volume 198 Number 13 jb.asm.org 1897Journal of Bacteriology
on July 12, 2018 by guesthttp://jb.asm
.org/D
ownloaded from
the two start codons in native sptA transcript provides for optimalmRNA stability, which enhances the efficiency of sptA gene ex-pression.
The free energies (�Gs) of the predicted secondary structuresin the 5= region (nucleotides 1 to 50) of the mRNAs were calcu-lated, but no strict correlation between �G and transcript stabilitywas observed (Fig. 6A). Instead, there is a strong correlation be-tween start codon usage and transcript stability. Notably, muta-tion of AUG1 to AUC led to a marked decrease in transcript sta-bility, as evidenced by a remarkable change in mRNA level(compare SPTA with M1AUC in Fig. 6A). In addition, the stabilityof leaderless transcripts with a strong start codon (AUG1) wassignificantly greater than that of transcripts with a weak startcodon (GUG or UUG) (compare M2AUC with 1GUG or 1UUGin Fig. 6A). Therefore, the presence of AUG1 at the 5= end of thetranscript is important for its stability. In the case of AUG2, itsmutation to AUC (M2AUC) also led to a decrease in transcriptstability, but to a lesser extent compared with M1AUC, in whichAUG1 was mutated to AUC (Fig. 6A). The effects of AUG1 andAUG2 on sptA transcript stability were further assessed by mea-suring the time course of mRNA degradation in H. volcanii cellsafter the inhibition of transcription by actinomycin D. As shownin Fig. 6B, the mutant constructs M1AUC, M2AUC, andM12AUC showed a significantly (P � 0.01) faster decay rate thanSPTA: M12AUC decayed faster than M1AUC, which decayedslightly faster than M2AUC. These results suggest that both AUG1
and AUG2 contribute to transcript stability and that AUG1 plays amore important role in stabilizing the transcript, probably by in-creasing the extent of ribosome binding to protect the mRNAfrom degradation.
DISCUSSION
In this study, the promoter and transcript of the halolysin-encod-ing gene sptA were experimentally determined. Our results showthat depending on which of two in-frame start codons is used fortranslation initiation, the sptA transcript can act as either a lead-erless or a leadered transcript for the production of functionalSptA in haloarchaea.
In haloarchaea, the sptA promoter can initiate the transcrip-tion of not only the sptA gene but also bgaH, as it contains all of thebasic elements of a haloarchaeal promoter (7, 9). The TATA boxand the BRE motif interact with the general transcription factorsTATA box-binding protein and transcription factor B, respec-tively. The WW motif may also interact with an as yet unidentifiedprotein (with transcription factor B and RNA polymerase as likelycandidates) (7). In the present study, target protein productionwas 2-fold greater with translation using the fsptA promoter thanthat with the sptA promoter. The fsptA and sptA promoters differin three positions, none of which are located within the basic ha-loarchaeal promoter elements. Our data thus suggest that the nu-cleotides flanking the basic promoter elements play importantroles in the regulation of sptA expression.
Both of the in-frame AUGs at the 5= end of the sptA transcriptare used as translation start codons for SptA production. AUG2-initiated translation does not involve ribosome scanning from the5= end of the transcript and is not influenced by an upstreamSD-like sequence. Instead, initiation of SptA translation at AUG2
most likely occurs via the SD-less mechanism, whose moleculardetails remain to be uncovered (8). Although haloarchaeal tran-scripts have been categorized as leadered transcripts with a 5=UTR
FIG 5 Effect of an SD-like sequence between the two start codons on translation initiation. (A) Sequences of the 5= ends of transcripts of the different constructsexamined. Start codons are shown in italics, and mutated nucleotides are underlined. The SD-like sequence is shaded. (B) Anti-His tag immunoblot analysis ofmid-log-phase culture supernatants of H. volcanii transformants harboring plasmids for different constructs. M, molecular mass markers. (C and D) Results ofproteolytic activity assay using suc-AAPF-pNa as the substrate and qRT-PCR determination of levels of target mRNAs in transformant cells using 16S rRNA asan internal control. The proteolytic activities and the mRNA levels of mutant constructs are relative to those of constructs M1AUC (C) or M2AUC (D). (E andF) Translation efficiencies of the constructs are shown after normalization to M1AUC (E) or M2AUC (F). Values are expressed as means standard deviations(error bars) from three independent experiments.
Tang et al.
1898 jb.asm.org July 2016 Volume 198 Number 13Journal of Bacteriology
on July 12, 2018 by guesthttp://jb.asm
.org/D
ownloaded from
of �10 nt and leaderless transcripts with a 5= UTR of �10 nt, thelower limit of 5= UTRs to direct a transcript to the leadered path-way seems to be 14 to 20 nt, and the upper limit for a 5= UTR toallow usage of the leaderless pathway seems to be around 5 nt (7,8). In this context, the sptA transcript, in which the two functionalAUGs are located 5 and 29 nt downstream of the transcriptionstart site, has the potential to be translated via the leaderless andSD-less mechanisms independently. However, AUG1-initiatedtranslation is more efficient than AUG2-initiated translation. Thisis consistent with the finding that leaderless transcripts not onlyform the majority of transcripts in haloarchaea, they are also veryefficiently translated (7).
In contrast to other known initiation mechanisms, leaderlesstranslation can be driven by undissociated 70S/80S ribosomes (49,50) and operate on leaderless transcripts that lack any substantial
signals for ribosomal recruitment aside from the start codon (2,10). Previous reports indicated that the translation of leaderlesstranscripts in E. coli requires the use of AUG as the start codon andnot just the codon-anticodon complementarity (51, 52). The re-sults of subsequent cross-linking experiments suggested that spe-cific ribosomal proteins contribute to recognition of the AUGstart codon of leaderless transcripts and facilitate initial transcriptrecognition and binding in the absence of initiator tRNAs andinitiation factors in E. coli (53). Similarly, only AUG serves as astart codon on leaderless dhfr transcript in H. volcanii, and singlepoint mutations to GUG or UUG completely abolish translationinitiation (8). However, several lines of evidence suggest that al-though an AUG start codon is highly preferred, it is not absolutelyrequired for translation of leaderless transcripts in haloarchaea.For instance, a CAU-to-GAC anticodon mutant of the H. volcaniiinitiator tRNA can initiate Bop protein synthesis using GUC as theinitiation codon in Halobacterium salinarum, suggesting that therequirement for AUG as the start codon for translation of leader-less transcripts does not extend to H. salinarum (54). Recently,Chen et al. (55) reported that GUG can drive translation of lead-erless bgaH transcripts in H. volcanii, with a 5- to 37-fold-lowertranslation efficiency compared with the use of AUG as the startcodon. In this study, we found that not only GUG but also UUGcan initiate the translation of leaderless sptA transcript in H. vol-canii. Despite their ability to initiate translation of leaderless sptAtranscripts, GUG and UUG are much weaker start codons thanAUG1, as evidenced by the finding that the efficiencies of GUG-and UUG-initiated translation are only �1/340 and �1/560 thatof AUG1-initiated translation.
Interestingly, we found that the substitution for AUG1 withGUG or UUG results in not only a dramatic decrease in transla-tion efficiency but also a marked decrease in sptA transcript sta-bility. In addition, the mutation of AUG1 to AUC also leads to asignificant decrease in transcript stability. These findings are con-sistent with those of a previous study that demonstrated a strictcorrelation between translation initiation and the presence ofleaderless transcripts in H. salinarum (54). The life span of a bac-terial mRNA is strongly influenced by its association with ribo-somes, and the initiation phase of translation plays a significantrole in determining mRNA stability (56). Because the removal ofjust a few nucleotides from the 5= end of leaderless transcripts thathave very short or no 5= UTRs is sufficient to disrupt the transla-tion initiation site, it is reasonable to assume that the 5= end ofleaderless transcripts should be properly protected against degra-dation by RNases. In this context, the strong preference for AUGas the start codon of leaderless transcripts in haloarchaea not onlyallows for efficient translation initiation, it also has the advantageof improving the affinity of ribosome binding to protect leaderlesstranscripts against degradation through preventing access of a 5=-end-dependent RNase, such as RNase E-like endoribonuclease(57) or RNase J-like 5=-to-3= exoribonuclease (58), to the 5= end ofthe transcript. The AUG2-initiated translation of the sptA tran-script most likely occurs via the SD-less mechanism, which neitherinvolves ribosome scanning from the 5= end of the transcript normakes use of the SD mechanism for translation initiation at 5=UTRs (8, 9). However, bioinformatics analyses of many nativeSD-less haloarchaeal 5= UTRs have failed to identify conservedmotifs responsible for ribosome binding (7, 8). Nevertheless, thereplacement of AUG2 with AUC leads to a significant decrease intranscript stability, implying that AUG2 may also act as a ribosome
FIG 6 Comparison of transcript stabilities. (A) mRNA level relative to SPTA.Levels of target mRNAs were determined by qRT-PCR using the mRNA of thenovR gene on the plasmid as an internal control. The mRNA levels of mutantconstructs are relative to that of wild-type construct SPTA. Free energy (�G)values (kilocalories per mole) of predicted secondary structures in the 5= re-gion (nucleotides 1 to 50) of the mRNAs are shown above the columns. (B)Time course changes in the remaining amount of mRNA after actinomycin Dtreatment. Levels of target mRNAs in transformant cells collected immediatelyafter the addition of actinomycin D (0 min) and 10, 20, and 30 min later weredetermined by qRT-PCR, using 16S rRNA as an internal control. Each datapoint represents a percentage of the zero-time value for each series. Values areexpressed as means standard deviations (error bars) from three independentexperiments (**, P � 0.01; *, P � 0.05).
Alternative Translation Initiation in Haloarchaea
July 2016 Volume 198 Number 13 jb.asm.org 1899Journal of Bacteriology
on July 12, 2018 by guesthttp://jb.asm
.org/D
ownloaded from
recruitment signal to protect the transcript against RNase-medi-ated degradation. Because the mutation of AUG2 to AUC has littleeffect on sptA expression, initiation of translation of the sptA tran-script occurs predominantly at AUG1. However, the expression ofsptA is not prevented either by mutating AUG1 to AUC or bydeleting the sequence starting from AUG1 to the nucleotide im-mediately upstream of AUG2, although with a low translation ef-ficiency. It seems that AUG2 can serve as a remedial initiation sitein case AUG1 is disrupted due to mutation or RNase-mediateddegradation.
Two SptA precursor isoforms with different N termini can begenerated from the sptA transcript, depending on whether AUG1
or AUG2 serves as the translation start codon. Because the twoAUGs are located upstream of the codons for the amino acid res-idues of the Tat motif (DRRSLL) in the signal peptide (Fig. 1A),alternative utilization of the two translation start sites does notaffect the function of the signal peptide and the secretion of SptAin haloarchaea. Moreover, the signal peptide is removed after en-zyme secretion, and the resulting proform is then converted intothe mature enzyme by autoprocessing of the N-terminal propep-tide. Therefore, the sequence of mature SptA is not influenced byalternative translation initiation at either AUG1 or AUG2. Thephysiologic significance of alternative utilization of the two startcodons on the sptA transcript remains to be elucidated. One pos-sible explanation is that different translation initiation mecha-nisms differentially respond to cellular signals and/or stress. Be-cause the ribosome can span about 30 nucleotides on mRNA uponbinding (59), it is unlikely that AUG1 and AUG2 in a single sptAtranscript can be simultaneously used as translation start codons.The AUG2-initiated translation may be inhibited if AUG1 is intactand functional. When AUG1-initiated translation fails undersome conditions, AUG2 could be used as the start codon thatensures proper expression of sptA. More importantly, our resultsconfirm that alternative translation initiation can take place on thesptA transcript carrying two start codons. A genome-wide surveyof codon usage in the 4,302 ORFs of Natrinema sp. strain J7-2showed that 253 ORFs contain at least two in-frame ATGs withinthe first 10 codons; 945 ORFs contain at least two potential in-frame start codons when GTG and TTG within the first 10 codonsare also taken into consideration (data not shown). Given thattranslation of the sptA transcript can be initiated at one of twosites, it is reasonable to anticipate that further transcriptional andtranslational analyses of these ORFs will result in the identificationof additional haloarchaeal transcripts demonstrating alternativetranslation initiation. It would be interesting to determine if alter-native translation initiation plays an important role in increasingthe diversity of gene products in haloarchaea.
ACKNOWLEDGMENTS
We thank W. F. Doolittle (Dalhousie University, Halifax, Canada) for thegenerous gift of H. volcanii WFD11, and we thank Zhisheng Xu (WuhanUniversity, Wuhan, China) for assistance with qRT-PCR analyses.
This work was supported in part by the National Natural ScienceFoundation of China (grant no. 31570062) and the National Infrastruc-ture of Natural Resources for the Science and Technology Program ofChina (grant no. NIMR-2014-8).
REFERENCES1. Tuller T, Carmi A, Vestsigian K, Navon S, Dorfan Y, Zaborske J, Pan
T, Dahan O, Furman I, Pilpel Y. 2010. An evolutionarily conserved
mechanism for controlling the efficiency of protein translation. Cell 141:344 –354. http://dx.doi.org/10.1016/j.cell.2010.03.031.
2. Malys N, McCarthy JE. 2011. Translation initiation: variations in themechanism can be anticipated. Cell Mol Life Sci 68:991–1003. http://dx.doi.org/10.1007/s00018-010-0588-z.
3. Hinnebusch AG. 2014. The scanning mechanism of eukaryotic transla-tion initiation. Annu Rev Biochem 83:779 – 812. http://dx.doi.org/10.1146/annurev-biochem-060713-035802.
4. Shine J, Dalgarno L. 1974. The 3=-terminal sequence of Escherichia coli16S ribosomal RNA: complementarity to nonsense triplets and ribosomebinding sites. Proc Natl Acad Sci U S A 71:1342–1346. http://dx.doi.org/10.1073/pnas.71.4.1342.
5. Steitz JA, Jakes K. 1975. How ribosomes select initiator regions in mRNA:base pair formation between the 3= terminus of 16S rRNA and the mRNAduring initiation of protein synthesis in Escherichia coli. Proc Natl Acad SciU S A 72:4734 – 4738. http://dx.doi.org/10.1073/pnas.72.12.4734.
6. Condo I, Ciammaruconi A, Benelli D, Ruggero D, Londei P. 1999.cis-acting signals controlling translational initiation in the thermophilicarchaeon Sulfolobus solfataricus. Mol Microbiol 34:377–384. http://dx.doi.org/10.1046/j.1365-2958.1999.01615.x.
7. Brenneis M, Hering O, Lange C, Soppa J. 2007. Experimental charac-terization of cis-acting elements important for translation and transcrip-tion in halophilic archaea. PLoS Genet 3:e229. http://dx.doi.org/10.1371/journal.pgen.0030229.
8. Hering O, Brenneis M, Beer J, Suess B, Soppa J. 2009. A novel mecha-nism for translation initiation operates in haloarchaea. Mol Microbiol71:1451–1463. http://dx.doi.org/10.1111/j.1365-2958.2009.06615.x.
9. Kramer P, Gabel K, Pfeiffer F, Soppa J. 2014. Haloferax volcanii, aprokaryotic species that does not use the Shine Dalgarno mechanism fortranslation initiation at 5=-UTRs. PLoS One 9:e94979. http://dx.doi.org/10.1371/journal.pone.0094979.
10. Grill S, Gualerzi CO, Londei P, Blasi U. 2000. Selective stimulation oftranslation of leaderless mRNA by initiation factor 2: evolutionary impli-cations for translation. EMBO J 19:4101– 4110. http://dx.doi.org/10.1093/emboj/19.15.4101.
11. Londei P. 2005. Evolution of translational initiation: new insights fromthe archaea. FEMS Microbiol Rev 29:185–200. http://dx.doi.org/10.1016/j.fmrre.2004.10.002.
12. Benelli D, Londei P. 2009. Begin at the beginning: evolution of transla-tional initiation. Res Microbiol 160:493–501. http://dx.doi.org/10.1016/j.resmic.2009.06.003.
13. Kozak M. 2002. Pushing the limits of the scanning mechanism for initia-tion of translation. Gene 299:1–34. http://dx.doi.org/10.1016/S0378-1119(02)01056-9.
14. Bonnal S, Pileur F, Orsini C, Parker F, Pujol F, Prats AC, Vagner S.2005. Heterogeneous nuclear ribonucleoprotein A1 is a novel internalribosome entry site trans-acting factor that modulates alternative initia-tion of translation of the fibroblast growth factor 2 mRNA. J Biol Chem280:4144 – 4153. http://dx.doi.org/10.1074/jbc.M411492200.
15. Kochetov AV. 2008. Alternative translation start sites and hidden codingpotential of eukaryotic mRNAs. Bioessays 30:683– 691. http://dx.doi.org/10.1002/bies.20771.
16. Jin H, Zhao Q, Gonzalez de Valdivia EI, Ardell DH, Stenstrom M,Isaksson LA. 2006. Influences on gene expression in vivo by a Shine-Dalgarno sequence. Mol Microbiol 60:480 – 492. http://dx.doi.org/10.1111/j.1365-2958.2006.05110.x.
17. Squires CL, Pedersen S, Ross BM, Squires C. 1991. ClpB is the Esche-richia coli heat shock protein F84.1. J Bacteriol 173:4254 – 4262.
18. Seol JH, Yoo SJ, Kim KI, Kang MS, Ha DB, Chung CH. 1994. The65-kDa protein derived from the internal translational initiation site of theclpA gene inhibits the ATP-dependent protease Ti in Escherichia coli. J BiolChem 269:29468 –29473.
19. Plumbridge JA, Deville F, Sacerdot C, Petersen HU, Cenatiempo Y,Cozzone A, Grunberg-Manago M, Hershey JW. 1985. Two translationalinitiation sites in the infB gene are used to express initiation factor IF2alpha and IF2 beta in Escherichia coli. EMBO J 4:223–229.
20. Mackman N, Nicaud JM, Gray L, Holland IB. 1985. Identification ofpolypeptides required for the export of haemolysin 2001 from E. coli. MolGen Genet 201:529 –536. http://dx.doi.org/10.1007/BF00331351.
21. Kim SW, Jung WH, Ryu JM, Kim JB, Jang HW, Jo YB, Jung JK, KimJH. 2008. Identification of an alternative translation initiation site for thePantoea ananatis lycopene cyclase (crtY) gene in E. coli and its evolution-
Tang et al.
1900 jb.asm.org July 2016 Volume 198 Number 13Journal of Bacteriology
on July 12, 2018 by guesthttp://jb.asm
.org/D
ownloaded from
ary conservation. Protein Expr Purif 58:23–31. http://dx.doi.org/10.1016/j.pep.2007.11.004.
22. Gabel K, Schmitt J, Schulz S, Nather DJ, Soppa J. 2013. A comprehen-sive analysis of the importance of translation initiation factors for Ha-loferax volcanii applying deletion and conditional depletion mutants.PLoS One 8:e77188. http://dx.doi.org/10.1371/journal.pone.0077188.
23. Siezen RJ, Leunissen JA. 1997. Subtilases: the superfamily of subtilisin-like serine proteases. Protein Sci 6:501–523.
24. De Castro RE, Maupin-Furlow JA, Gimenez MI, Herrera Seitz MK,Sanchez JJ. 2006. Haloarchaeal proteases and proteolytic systems. FEMSMicrobiol Rev 30:17–35. http://dx.doi.org/10.1111/j.1574-6976.2005.00003.x.
25. Shi W, Tang XF, Huang Y, Gan F, Tang B, Shen P. 2006. An extracel-lular halophilic protease SptA from a halophilic archaeon Natrinema sp.J7: gene cloning, expression and characterization. Extremophiles 10:599 –606.
26. Xu Z, Du X, Li T, Gan F, Tang B, Tang XF. 2011. Functional insight intothe C-terminal extension of halolysin SptA from haloarchaeon Natrinemasp. J7. PLoS One 6:e23562. http://dx.doi.org/10.1371/journal.pone.0023562.
27. Zhang Y, Wang M, Du X, Tang W, Zhang L, Li M, Wang J, Tang B,Tang XF. 2014. Chitin accelerates activation of a novel haloarchaeal serineprotease that deproteinizes chitin-containing biomass. Appl Environ Mi-crobiol 80:5698 –5708. http://dx.doi.org/10.1128/AEM.01196-14.
28. Du X, Li M, Tang W, Zhang Y, Zhang L, Wang J, Li T, Tang B, TangXF. 2015. Secretion of Tat-dependent halolysin SptA capable of autocat-alytic activation and its relation to haloarchaeal growth. Mol Microbiol96:548 –565. http://dx.doi.org/10.1111/mmi.12955.
29. Feng J, Wang J, Zhang Y, Du X, Xu Z, Wu Y, Tang W, Li M, Tang B,Tang XF. 2014. Proteomic analysis of the secretome of haloarchaeonNatrinema sp. J7-2. J Proteome Res 13:1248 –1258. http://dx.doi.org/10.1021/pr400728x.
30. Pohlschroder M, Prinz WA, Hartmann E, Beckwith J. 1997. Proteintranslocation in the three domains of life: variations on a theme. Cell91:563–566. http://dx.doi.org/10.1016/S0092-8674(00)80443-2.
31. Shen P, Chen Y. 1994. Plasmid from Halobacterium halobium and itsrestriction map. Yi Chuan Xue Bao 21:409 – 416.
32. Feng J, Liu B, Zhang Z, Ren Y, Li Y, Gan F, Huang Y, Chen X, Shen P,Wang L, Tang B, Tang XF. 2012. The complete genome sequence ofNatrinema sp. J7-2, a haloarchaeon capable of growth on synthetic mediawithout amino acid supplements. PLoS One 7:e41621. http://dx.doi.org/10.1371/journal.pone.0041621.
33. Ye X, Ou J, Ni L, Shi W, Shen P. 2003. Characterization of a novelplasmid from extremely halophilic archaea: nucleotide sequence andfunction analysis. FEMS Microbiol Lett 221:53–57. http://dx.doi.org/10.1016/S0378-1097(03)00175-7.
34. Soppa J, Baumann A, Brenneis M, Dambeck M, Hering O, Lange C.2008. Genomics and functional genomics with haloarchaea. Arch Micro-biol 190:197–215. http://dx.doi.org/10.1007/s00203-008-0376-4.
35. Kuhn J, Binder S. 2002. RT-PCR analysis of 5= to 3=-end-ligated mRNAsidentifies the extremities of cox2 transcripts in pea mitochondria. NucleicAcids Res 30:439 – 446. http://dx.doi.org/10.1093/nar/30.2.439.
36. Sambrook J, Fritsch EF, Maniatis T. 1989. Molecular cloning: a labora-tory manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold SpringHarbor, NY.
37. Kamekura M, Seno Y, Holmes ML, Dyall-Smith ML. 1992. Molecularcloning and sequencing of the gene for a halophilic alkaline serine protease(halolysin) from an unidentified halophilic archaea strain (172P1) andexpression of the gene in Haloferax volcanii. J Bacteriol 174:736 –742.
38. Yang Y, Huang YP, Shen P. 2003. The 492-bp RM07 DNA fragment fromthe halophilic archaea confers promoter activity in all three domains oflife. Curr Microbiol 47:388 –394.
39. An Y, Lv A, Wu W. 2010. A QuikChange-like method to realize efficientblunt-ended DNA directional cloning and site-directed mutagenesis si-multaneously. Biochem Biophys Res Commun 397:136 –139. http://dx.doi.org/10.1016/j.bbrc.2010.05.042.
40. Cline SW, Lam WL, Charlebois RL, Schalkwyk LC, Doolittle WF. 1989.Transformation methods for halophilic archaebacteria. Can J Microbiol35:148 –152. http://dx.doi.org/10.1139/m89-022.
41. Livak KJ, Schmittgen TD. 2001. Analysis of relative gene expression data
using real-time quantitative PCR and the 2���CT method. Methods 25:402– 408. http://dx.doi.org/10.1006/meth.2001.1262.
42. Jager A, Samorski R, Pfeifer F, Klug G. 2002. Individual gvp transcriptsegments in Haloferax mediterranei exhibit varying half-lives, which aredifferentially affected by salt concentration and growth phase. NucleicAcids Res 30:5436 –5443. http://dx.doi.org/10.1093/nar/gkf699.
43. Hundt S, Zaigler A, Lange C, Soppa J, Klug G. 2007. Global analysis ofmRNA decay in Halobacterium salinarum NRC-1 at single-gene resolu-tion using DNA microarrays. J Bacteriol 189:6936 – 6944. http://dx.doi.org/10.1128/JB.00559-07.
44. Zuker M. 2003. Mfold web server for nucleic acid folding and hybridiza-tion prediction. Nucleic Acids Res 31:3406 –3415. http://dx.doi.org/10.1093/nar/gkg595.
45. Holmes ML, Scopes RK, Moritz RL, Simpson RJ, Englert C, Pfeifer F,Dyall-Smith ML. 1997. Purification and analysis of an extremely halo-philic beta-galactosidase from Haloferax alicantei. Biochim Biophys Acta1337:276 –286. http://dx.doi.org/10.1016/S0167-4838(96)00174-4.
46. King J, Laemmli UK. 1971. Polypeptides of the tail fibres of bacteriophageT4. J Mol Biol 62:465– 477. http://dx.doi.org/10.1016/0022-2836(71)90148-3.
47. Cheng G, Zhao P, Tang XF, Tang B. 2009. Identification and character-ization of a novel spore-associated subtilase from Thermoactinomyces sp.CDF. Microbiology 155:3661–3672. http://dx.doi.org/10.1099/mic.0.031336-0.
48. Dennis PP. 1997. Ancient ciphers: translation in archaea. Cell 89:1007–1010. http://dx.doi.org/10.1016/S0092-8674(00)80288-3.
49. Moll I, Hirokawa G, Kiel MC, Kaji A, Blasi U. 2004. Translationinitiation with 70S ribosomes: an alternative pathway for leaderless mR-NAs. Nucleic Acids Res 32:3354 –3363. http://dx.doi.org/10.1093/nar/gkh663.
50. Andreev DE, Terenin IM, Dunaevsky YE, Dmitriev SE, Shatsky IN.2006. A leaderless mRNA can bind to mammalian 80S ribosomes anddirect polypeptide synthesis in the absence of translation initiation factors.Mol Cell Biol 26:3164 –3169. http://dx.doi.org/10.1128/MCB.26.8.3164-3169.2006.
51. Van Etten WJ, Janssen GR. 1998. An AUG initiation codon, not codon-anticodon complementarity, is required for the translation of unleaderedmRNA in Escherichia coli. Mol Microbiol 27:987–1001. http://dx.doi.org/10.1046/j.1365-2958.1998.00744.x.
52. O’Donnell SM, Janssen GR. 2001. The initiation codon affects ribosomebinding and translational efficiency in Escherichia coli of cI mRNA with orwithout the 5= untranslated leader. J Bacteriol 183:1277–1283. http://dx.doi.org/10.1128/JB.183.4.1277-1283.2001.
53. Brock JE, Pourshahian S, Giliberti J, Limbach PA, Janssen GR. 2008.Ribosomes bind leaderless mRNA in Escherichia coli through recognitionof their 5=-terminal AUG. RNA 14:2159 –2169. http://dx.doi.org/10.1261/rna.1089208.
54. Srinivasan G, Krebs MP, RajBhandary UL. 2006. Translation initiationwith GUC codon in the archaeon Halobacterium salinarum: implicationsfor translation of leaderless mRNA and strict correlation between transla-tion initiation and presence of mRNA. Mol Microbiol 59:1013–1024. http://dx.doi.org/10.1111/j.1365-2958.2005.04992.x.
55. Chen W, Yang G, He Y, Zhang S, Chen H, Shen P, Chen X, Huang YP.2015. Nucleotides flanking the start codon in hsp70 mRNAs with veryshort 5=-UTRs greatly affect gene expression in haloarchaea. PLoS One10:e0138473. http://dx.doi.org/10.1371/journal.pone.0138473.
56. Deana A, Belasco JG. 2005. Lost in translation: the influence of ribosomeson bacterial mRNA decay. Genes Dev 19:2526 –2533. http://dx.doi.org/10.1101/gad.1348805.
57. Franzetti B, Sohlberg B, Zaccai G, von Gabain A. 1997. Biochemical andserological evidence for an RNase E-like activity in halophilic archaea. JBacteriol 179:1180 –1185.
58. Hasenohrl D, Konrat R, Blasi U. 2011. Identification of an RNase Jortholog in Sulfolobus solfataricus: implications for 5=-to-3= directionaldecay and 5=-end protection of mRNA in Crenarchaeota. RNA 17:99 –107.http://dx.doi.org/10.1261/rna.2418211.
59. Yusupova GZ, Yusupov MM, Cate JH, Noller HF. 2001. The path ofmessenger RNA through the ribosome. Cell 106:233–241. http://dx.doi.org/10.1016/S0092-8674(01)00435-4.
Alternative Translation Initiation in Haloarchaea
July 2016 Volume 198 Number 13 jb.asm.org 1901Journal of Bacteriology
on July 12, 2018 by guesthttp://jb.asm
.org/D
ownloaded from