overexpression of cloned rhsa sequences perturbs the ... · translational machinery in escherichia...

JOURNAL OF BACTERIOLOGY, Sept. 2011, p. 4869–4880 Vol. 193, No. 180021-9193/11/$12.00 doi:10.1128/JB.05061-11Copyright © 2011, American Society for Microbiology. All Rights Reserved.

Overexpression of Cloned RhsA Sequences Perturbs the CellularTranslational Machinery in Escherichia coli�

Kunal Aggarwal† and Kelvin H. Lee*School of Chemical and Biomolecular Engineering, Cornell University, Ithaca, New York 14853

Received 11 April 2011/Accepted 2 July 2011

RhsA is a member of the multigene Rhs family and consists of a complex genetic sequence. This sequenceconsists of several distinct components, including a GC-rich core (core open reading frame [ORF]), an AT-richextension (ext-a1) of the core ORF and an AT-rich region following the core extension (dsORF-a1). Thefunctions of RhsA and the different distinct components, which can include open reading frames, are not wellunderstood. Here, we study the effect of overexpression of the ext-a1 sequence and the ext-a1 3� region, whichincludes a partial sequence of dsORF-a1, on Escherichia coli cells. Cells expressing these sequences showreduced cell growth and cell viability. The expression of these sequences dramatically affects different compo-nents of the transcription and translation machinery. Transcriptomic analysis reveals an increase in theexpression of genes involved in transcription, RNA processing, and nucleotide biosynthesis and metabolismand a decrease in the expression of amino acid biosynthesis genes and transfer RNAs. Further, expression ofthe above-mentioned RhsA components increases ribosomal gene expression, as well as rRNA and ribosomeabundance. Proteomic analysis reveals an overall reduction of protein expression at the genome-wide level incells expressing the above-mentioned RhsA components. Based on these observations, we suspect a translationproduct of ext-a1 affects different regulatory mechanisms that control rRNA synthesis.

The Rhs family in Escherichia coli consists of five geneticelements. These elements are found in many, but not all, nat-ural E. coli strains (4, 7, 11, 16). The Rhs (or rearrangement hotspot) elements were originally detected as components of recA-dependent intrachromosomal recombination (8). Subsequentstudies have defined the Rhs loci on the chromosome andelucidated the evolutionary history of these elements (4, 6, 7,19). Members of the Rhs family share one or more discretehomologous structures, one of which, a 3.7-kb GC-rich core, iscommon to all the elements. The Rhs family also containsunique sequences specific to each element, such as dissimilarsequences that flank the core of Rhs elements. Certain com-ponents of the Rhs elements also have a distinct G�C content(approximately 40% or 60%) compared to the rest of the E.coli genome (approximately 50%), suggesting their origin isoutside the E. coli species (4). While the function of Rhselements is not fully understood, the amino acid sequencesuggests that the elements encode hydrophilic proteins withrepetitive sequence elements and divergent C termini (7).

RhsA, a member of the Rhs family, is 8.2 kb long and consistsof some distinct sequence features. The RhsA sequence caninclude five different open reading frames (ORFs). The 3.7-kbGC-rich core includes a long core ORF that extends 417 basesinto a 1.3-kb AT-rich region. This extension is designated ext-a1. The 1.3-kb AT-rich region also contains dsORF-a1, an840-bp ORF. Both ext-a1 and dsORF-a1 contain a high num-ber of infrequently or rarely used codons; however, greater use

is exhibited by ext-a1. Further, core ORF (including ext-a1) isnot expressed in routine laboratory cultivation (4). Vlazny andHill (15) have reported that overexpression of a partial se-quence of ext-a1 (referred to as ORF-ex by the authors) in E.coli imparts a stationary-phase-dependent viability block tocells. This toxic effect is suppressed with overexpression ofdsORF-a1 (15). Here, we study the effect of overexpression ofa fragment of the RhsA genetic element, consisting of thecomplete sequence of the component ext-a1 and a partial se-quence of the component dsORF-a1, on E. coli cells. Wemonitor E. coli cell growth in response to differential expres-sion of these elements. Transcriptomic and proteomic profilesare used to clarify the possible relationship between expressionof the cloned RhsA sequences and translation.

MATERIALS AND METHODS

Strains and plasmids. Table 1 lists the plasmids used in this work. Wild-typeE. coli W3110 cells transformed with plasmids pklr1 and pKQV4 (designatedW3110 pklr1 cells) were used for this study. Plasmid pklr1 derives its identityfrom pKFis4 and consists of RhsA components ext-a1 and a partial sequence ofdsORF-a1, under the control of an IPTG (isopropyl-�-D-thiogalactopyranoside)-inducible tac promoter, and a kanamycin (Km) resistance gene (10). This plas-mid contains the entire sequence of ext-a1 and the first 341 nucleotides ofdsORF-a1, corresponding to nucleotides 4611 to 5446 in GenBank accession no.L19044 (http://www.ncbi.nlm.nih.gov/GenBank/index.html). The dsORF-a1 se-quence is followed by the constitutive transcription terminators present in plas-mid pklr1. Plasmid pKQV4 carries lacIq and encodes ampicillin (Amp) resistance(10). Plasmids pklr1 and pKQV4 were obtained from Wilfred Chen (Universityof California, Riverside, CA).

Variants of plasmid pklr1 (pklr2, pklr3, pklr4, and pklr5) were created afterappropriate restriction enzyme digestion of plasmid pklr1, followed by purifica-tion of the plasmid fragment containing the Km resistance gene and subsequentreligation. Plasmid pklr2 was constructed by digestion of plasmid pklr1 with BtsIto delete the tac promoter, ext-a1, and dsORF-a1 sequences. Plasmid pklr3 wasconstructed by digestion of plasmid pklr1 with EcoRI and MfeI to delete only theext-a1 and dsORF-a1 sequences. Plasmid pklr4 was constructed by digestion ofplasmid pklr1 with BpmI to delete the partial sequence of ext-a1. Plasmid pklr5was constructed by digestion of plasmid pklr1 with EcoNI to delete the partial

* Corresponding author. Present address: Delaware BiotechnologyInstitute and Chemical Engineering, University of Delaware, Newark,DE 19711. Phone: (302) 831-0344. Fax: (302) 831-4841. E-mail: [email protected].

† Present address: Novartis Vaccines and Diagnostics, 45 SidneyStreet, Cambridge, MA 02139.

� Published ahead of print on 15 July 2011.

4869

on March 25, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

sequence of dsORF-a1. Wild type E. coli W3110 cells were then transformedwith plasmids pklr2 and pkQV4, pklr3 and pkQV4, pklr4 and pkQV4, and pklr5and pkQV4 and designated W3110 pklr2, W3110 pklr3, W3110 pklr4, and W3110pklr5 cells, respectively. A codon-optimized variant of plasmid pklr1,pklr1OF1OF2, was created by site-directed mutagenesis (Stratagene, CedarCreek, TX). Specifically, the codons CTA, ATA, GGA, CTA, GGG, ATA, ATA,and ATA at nucleotide positions 4828, 4831, 4846, 4849, 4900, 4903, 4924, and4927 (GenBank accession no. L19044) were replaced with CTG, ATC, GGT,CTG, GGT, ATC, ATC, and ATC, respectively. Wild-type E. coli W3110 cellswere then transformed with plasmids pklr1OF1OF2 and pkQV4 and designatedW3110 pklr1OF1OF2 cells.

Culture conditions. Frozen stocks of W3110 pklr1 cells were streaked onto LBmedium plates containing 80 �g/ml Amp and 40 �g/ml Km and incubatedovernight at 37°C to form single colonies. A flask (1,000 ml) containing 200 mlof YT medium (8 g tryptone, 5 g yeast extract, 5 g NaCl in 1,000 ml water, pH7.2) supplemented with 80 �g/ml Amp and 40 �g/ml Km was inoculated with asingle colony of W3110 pklr1 cells and shaken at 250 rpm and 37°C. The culturewas split at an optical density at 600 nm (OD600) of 0.1 to generate three parallelcultures of W3110 pklr1 cells (50 ml each in 250-ml flasks). The cultures weregrown to an OD600 of 0.4 and induced with three different concentrations ofIPTG (0, 0.1, and 1 mM). The time at which IPTG was added to the cultures wasdesignated 0 h. Cells were harvested for further analysis at appropriate timepoints by centrifugation at 4,000 � g and 4°C for 10 min. A similar cultureprotocol was followed for W3110 pklr2, W3110 pklr3, W3110 pklr4, W3110pklr5, and W3110 pklr1OF1OF2 cells.

Bacterial-viability determination. Equal volumes of W3110 pklr1 cultures (0,0.1, and 1 mM IPTG induced) were harvested 3.5 h after IPTG induction andstained using the Live/Dead BacLight Bacterial Viability Kit L7012 (MolecularProbes, Eugene, OR) according to the manufacturer’s protocol. The stained cellswere placed under a coverslip on a glass slide and observed under a NikonStereoscopic Zoom microscope SMZ1500 (Nikon, Rochester, NY) at �100magnification. Focused images were captured from three different areas of thecoverslip per sample using a Spot Insight charge-coupled-device camera (Diag-nostic Instruments, Sterling Heights, MI) attached to the microscope. Imageanalysis was performed using ImageMaster 2D Platinum software v5.0 (GEAmersham Biosciences, Piscataway, NJ) to count the cells labeled fluorescentgreen (viable) or red (nonviable). The average number of red cells was dividedby the average sum of red and green cells to estimate the fraction of nonviablecells in the population.

mRNA analysis. Cells corresponding to a 10-ml culture volume were harvestedfrom biological replicate W3110 pklr1 cultures at 0 h (0 mM IPTG) and at 1.5and 3.5 h (0, 0.1, and 1 mM IPTG) after IPTG induction using centrifugation.The cell pellets were immediately resuspended in RNAlater (Ambion, Austin,TX) and stored at �20°C until further processing. For RNA extraction, each ofthe 14 samples was washed in cold phosphate-buffered saline to remove theRNAlater. Samples were processed using MasterPure RNA purification kits(Epicentre, Madison, WI) according to the manufacturer’s protocol. RNA wasquantified using A260 measurements, and prior to storage at �20°C, purity wasassessed using the A260/A280 ratio.

RNA samples were amplified and labeled according to the Affymetrix pro-karyotic sample and array-processing protocol (Affymetrix, Santa Clara, CA) andNuGen’s Ovation Biotin RNA Amplification and Labeling System (catalog num-ber 2300; NuGen, San Carlos, CA). Briefly, cDNA fragments were generatedfrom RNA samples via reverse transcription. The resulting cDNA was thenfluorescently labeled and hybridized to E. coli antisense version 2 GeneChipprobe arrays (Affymetrix, Santa Clara, CA). After washing, the hybridized arrayswere scanned to detect signal from different probes. The probe intensity datafrom a total of 14 � 2 scanned arrays (corresponding to 14 different RNAsamples and 2 different labeling protocols) was analyzed using Genetraffic soft-

ware (Iobion Informatics LLC, La Jolla, CA). The probe intensity data werenormalized using a GC-RMA algorithm to estimate gene expression values (17).The mRNA expression values for genes were compared across two biologicalstates to calculate the fold change in gene expression. The genes with statisticallysignificant fold change in expression were identified using Significance Analysisof Microarrays software (14).

Protein analysis. For protein analysis, cells harvested from W3110 pklr1 cellcultures (10 ml each) at 0 h (0 mM IPTG) and 3.5 h (0, 0.1, and 1 mM IPTG)after IPTG induction were used. Multiplexed isobaric tagging technology(iTRAQ) followed by mass spectrometric analysis was employed for proteinanalysis, as described in detail elsewhere (1). Briefly, pelleted cells were washedin a low salt solution, resuspended in lysis buffer, and lysed by sonication. A100-�g protein aliquot from each sample (determined using a Lowry proteinconcentration assay) was reduced, alkylated, digested with trypsin, and labeled withiTRAQ isobaric reagents (Applied Biosystems, Foster City, CA). Labeled sampleswere pooled and subjected to strong cation-exchange fractionation, followed byreversed-phase high-performance liquid chromatography separation prior to massspectrometric (MS) analysis. MS and tandem mass spectrometric (MS-MS) analyseswere performed using a 4700 Proteomics Analyzer (Applied Biosystems). Searchesof the MS and MS-MS data against the E. coli genome were performed usingGPS Explorer version 2.0 (Applied Biosystems). Peptide matches where P was �0.05were used for protein identification. The proteins were quantified (in a relativemanner) using the peak areas of the isobaric tags measured during MS-MS analysis.The set of proteomics results is available at the Proteome Commons TrancheRepository using the following hash (unique identifier): �6p6sCEndPRXcDFbxV/NN9FdMlAFHz8TuHGYCicjQgROYnq7tvdh996AdtwxVj0UvZ2ynv6D6iY5M3wmQ1OSHYfDq9AAAAAAAAABuA��.

Ribosome analysis. Ribosomes were quantified per cell optical density unitusing the procedure of Dong et al. (3) with the following modifications. The cellpellet was washed in the polymix buffer, resuspended in polymix buffer contain-ing 100 �g/ml lysozyme, and incubated for 15 min at room temperature. The cellswere sonicated for 10 s on ice using a 550 Sonic Dismembrator (Fisher Scientific,Pittsburgh, PA) equipped with a cup horn. The cell lysate was mixed with 1 U/mlRNase-free DNase (New England BioLabs, Ipswich, MA) and incubated at roomtemperature for 10 min. Cell debris was pelleted by centrifugation at 17,900 � gand 4°C for 10 min. For each sample, 600 �l of the supernatant was layered atop10.8 ml of a 15 to 35% sucrose density gradient in 14- by 89-mm polyallomertubes (Beckman Instruments, Palo Alto, CA). Gradients were centrifuged in aBeckman SW41Ti swinging-bucket rotor at 288,000 � g at 4°C for 4 h. Fractions(approximately 100 �l in volume) were collected from the bottom of the poly-allomer tube into a 96-well UV-transparent microplate (catalog no. 3635; Corn-ing Inc., Corning, NY). The absorbances of the fractions were measured at 254nm using a Versamax microplate reader (Molecular Devices, Sunnyvale, CA).The fractions were stored frozen at �80°C until further use.

Ribosomal-protein analysis. The sucrose density gradient fractions containingribosomal peaks labeled 50S, 70S, 100S, and polysomes (multiple ribosomestranslating a single mRNA) (see Fig. 6) were thawed to room temperature. A 2�volume of chilled ethanol was added to each of the thawed fractions. Eachmixture was vortexed and kept overnight at �20°C to precipitate protein. Proteinpellets were obtained after centrifugation at 10,000 � g and 4°C for 10 min. Theprotein pellets were directly resuspended in iTRAQ dissolution buffer. A 10-�gprotein aliquot from each sample (determined using a Lowry protein concen-tration assay) was reduced, alkylated, digested with trypsin, and labeled withiTRAQ isobaric reagents (Applied Biosystems, Foster City, CA). The labeledsamples were pooled and separated into two fractions using strong cation-ex-change (SCX) chromatography.

Strong cation exchange fractionation was performed using an Applied Biosys-tems cation-exchange cartridge system. The pooled iTRAQ-labeled tryptic pep-tides (�400 �l) were evaporated down to 50 �l with a centrifugal vacuumconcentrator and reconstituted with 1 ml of loading buffer (10 mM potassiumphosphate, pH 3.0, 25% acetonitrile [ACN]). The pH of the sample was adjustedto 3.0 with phosphoric acid prior to cartridge separation. After conditioning ofthe SCX cartridge with loading buffer, the samples (�40 �g) were loaded andwashed with 2 ml of loading buffer. The peptides were eluted in two steps using1 ml of loading buffer containing 75 mM KCl and 500 mM KCl. The two SCXfractions were partially evaporated to remove ACN and desalted by solid-phaseextraction.

Reverse-phase separation and MS-MS analysis were performed using a pro-cedure described previously (9) with the following modifications. The peptideswere eluted from PepMap C18 RP nanocolumn using a 90-min gradient of 5% to40% acetonitrile in 0.1% formic acid at 250 nl/min. MS and MS-MS analyseswere performed on a hybrid triple-quadrupole linear ion trap mass spectrometer,4000 Q Trap (Applied Biosystems), equipped with a Micro Ion Spray Head ion

TABLE 1. Plasmids used in this work

Plasmid Element/gene Resistance

pklr1 tac� ext-a1� dsORF-a1� Kanamycinpklr2 tac ext-a1� dsORF-a1� Kanamycinpklr3 tac� ext-a1� dsORF-a1� Kanamycinpklr4 tac� ext-a1� dsORF-a1� Kanamycinpklr5 tac� ext-a1� dsORF-a1� Kanamycinpklr1OF1OF2 tac� modified ext-a1� dsORF-a1� KanamycinpKQV4 lacIq Ampicillin

4870 AGGARWAL AND LEE J. BACTERIOL.


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

source. MS data acquisition was performed using Analyst 1.4.1 software (AppliedBiosystems) in positive ion mode for information-dependent acquisition (IDA)analysis. MS-MS spectra generated from nano-liquid chromatography (LC)/electrospray ionization (ESI)-based IDA analyses were interrogated using Pro-teinPilot software 1.0 (Applied Biosystems) for database searches against the E.coli database by the Paragon method. The setting for trypsin with methyl meth-ane thiosulfonate (MMTS) modification of cysteine and methionine oxidationwas used for quantitative processing and identification. Peptide matches where Pwas �0.05 were used for protein identification. The peptides were quantified (ina relative manner) using the peak areas of the isobaric tags measured duringMS-MS analysis.

Total RNA analysis. For total RNA analysis, in accordance with the manu-facturer’s protocol, 1 ml of RNAprotect Bacteria Reagent (Qiagen Inc., Valen-cia, CA) was added to 0.5 ml of each cell culture of interest to stabilize the cells.Total RNA was extracted from the stabilized cell pellet using the RNeasy MiniKit (Qiagen Inc.) following the manufacturer’s protocol. The extracted RNA wasquantified using A260 measurements and was stored at �20°C until further use.The extracted RNA was analyzed using formaldehyde-based denaturing agarosegel electrophoresis as described elsewhere (12). The gels were stained usingSYBR green I Nucleic Acid Stain (Invitrogen, Carlsbad, CA). The stained gelswere imaged on an FLA-3000 (Fujifilm Medical Systems, Stamford, CT). Imageanalysis was performed using ImageMaster 2D Platinum software v5.0 (GEAmersham Biosciences) to measure the intensity of the bands in the gel images.

rRNA sequencing. Bands corresponding to 23S and 16Sp RNA (RNA desig-nations are discussed in “RhsA component overexpression alters rRNA expres-sion” below; see Fig. 8 and 9) were excised from agarose gels. RNA was extractedfrom the excised gel bands using a QIAquick gel extraction kit (Qiagen Inc.)following the manufacturer’s protocol. The RNA was then ligated to differentoligonucleotide sequences at the 3 end using T4 RNA ligase (Ambion Inc.). Theoligonucleotide sequences were designed to serve as templates for primer bind-ing during reverse transcription and were commercially synthesized at IntegratedDNA Technologies (Coralville, IA). A cDNA template for PCR amplificationwas generated from the RNA-oligonucleotide hybrid using the Senscript RT kit(Qiagen Inc.). The PCR-amplified product was purified using agarose gel elec-trophoresis followed by gel extraction and sequenced at the Biotechnology Re-source Center, Cornell University (Ithaca, NY). Sanger sequencing was used topartially sequence the 16Sp and 23S RNA to establish the identities of the RNAbands.

rRNA analysis. The sucrose density gradient fractions containing ribosomeswere thawed to room temperature. RNA was extracted using the RNeasy MiniKit (Qiagen Inc.) following the manufacturer’s protocol except for the followingmodification. The lysozyme lysis step was not performed, and 200 �l of bufferRLT was directly added to the sucrose density gradient fraction. The extractedRNA was analyzed using denaturing agarose gel electrophoresis as mentionedabove.

Reinoculation studies. Equal numbers of cells from 0 and 1 mM IPTG-induced W3110 pklr1 cultures at 3.5 h after IPTG induction were used toinoculate fresh YT medium supplemented with Km and Amp. The cultureinoculated with cells overexpressing RhsA elements was split at an OD600 of 0.1,and one of the cultures was induced with IPTG (final concentration, 1 mM).Culture growth was monitored using OD600 measurements. Total RNA wasextracted from the IPTG-induced cells initially used to inoculate fresh medium.Total RNA was also extracted from reinoculated cell cultures at the point ofIPTG induction and at 3 h after IPTG induction (see Fig. 10). Total RNA wasanalyzed using denaturing agarose gel electrophoresis as mentioned above.

Microarray data accession number. The entire set of supporting microarraydata has been deposited in the Gene Expression Omnibus (GEO) database(http://www.ncbi.nlm.nih.gov/geo/) and can be accessed under GEO accessionnumber GSE12411.

RESULTS

Expression of RhsA components reduces measured OD. Inthis work, we studied the effects of expression of RhsA com-ponents ext-a1 and a partial sequence of dsORF-a1 at differentinduction levels on E. coli W3110 cells. These elements arecarried on a medium-copy-number plasmid under the controlof an IPTG-inducible tac promoter. The OD of W3110 pklr1cells after induction with 0, 0.1, and 1 mM IPTG was moni-tored. As seen in Fig. 1a, W3110 pklr1 cells show a reduction

in final OD after induction with IPTG. This reduction is en-hanced when larger amounts of IPTG are added to the growthmedium. W3110 pklr2 cells, which contain plasmid pKQV4and a variant of plasmid pklr1 lacking the cloned RhsA se-quences and the tac promoter, did not show any reduction infinal OD upon addition of IPTG to the growth medium(Fig. 1b).

Reduction in OD due to biological activity of RhsA elements.The ext-a1 sequence is rich in rare codons, with a Sharp and Licodon adaptation index for ext-a1 of 0.157 (4). It has beensuggested that overexpression of an mRNA dependent on rarecodons can inhibit cell growth (18). To gain insight intowhether the observed reduction in the final OD during RhsAcomponent expression is due to the biological activity of thecloned RhsA sequences or to the overexpression of a sequencerich in rare codons (ext-a1), we expressed a codon-optimizedvariant of the sequence ext-a1 using plasmid pklr1OF1OF2.Four consecutive rare codon pairs (a total of eight codons) inthe cloned ext-a1 sequence were replaced with synonymouscodons known to be used frequently in E. coli. The replace-ment of the consecutive rare codons in ext-a1 with synonymousfrequently used codons reduces the number of rare codons inthe ext-a1 mRNA and is expected to increase the translationrate of the mRNA with synonymous codon replacements com-pared to the wild-type mRNA. A greater negative effect on thefinal OD is observed in W3110 pklr1OF1OF2 cells upon IPTGaddition than in the W3110 pklr1 cells (Fig. 1c). This observa-tion is consistent with the reduction in OD resulting from somebiological activity of the cloned RhsA sequences.

ext-a1 expression alone reduces cellular growth. To betterunderstand the role of the two cloned sequences (ext-a1 anddsORF-a1) individually, in terms of impact on the final OD, wecreated three variants of plasmid pklr1, namely, pklr3, pklr4,and pklr5. The construction of these plasmids is described inMaterials and Methods. As summarized in Fig. 2, upon IPTGaddition to culture media, a reduction was observed only inW3110 pklr5 cells. Further, the phenotype was lost upon de-letion of a region in the vicinity of the 3 end of ext-a1 (plasmidpklr4). This result is consistent with the work of Vlazny andHill (15), where expression of the last 25 codons of the coreORF (and ORF-ex) was observed to be essential in impartingcellular growth toxicity (15). Thus, our study confirmed thatthe expression of the cloned ext-a1 sequence is toxic to the cellsand is responsible for the observed growth reduction, as mea-sured by OD, in induced W3110 pklr1 cells.

Expression of RhsA components reduces cell viability. Theeffect of overexpression of cloned RhsA sequences on E. coliW3110 pklr1 cell viability was observed by differential fluores-cence staining of the viable and nonviable cells from the cul-tures (induced with 0, 0.1, and 1 mM IPTG and then harvestedat 3.5 h). The viable and nonviable cells appear green and red,respectively, when observed under a fluorescence microscope(Fig. 3). As expected from OD600 measurements shown in Fig.1a, the total number of cells (corresponding to green and redspots) in 1 mM IPTG-induced culture is the lowest among thethree cell cultures. However, the fraction of nonviable cells(red spots) relative to the total number of cells is highest in the1 mM IPTG-induced culture (approximately 36% of the cellsare dead). The fraction of nonviable cells in the 0 mM IPTG-induced culture is 8%, and in the 0.1 mM IPTG-induced cul-

VOL. 193, 2011 RhsA SEQUENCES PERTURB TRANSLATION MACHINERY 4871


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

FIG. 1. (a) Phenotypic response of W3110 pklr1 cells to IPTG addition to the growth medium. (b) Phenotypic response of W3110 pklr2 cellsto IPTG addition to the growth medium. The growth of uninduced W3110 pklr1 cells is also compared to that of uninduced W3110 pklr2 cells.(c) Comparison of ODs in W3110 pklr1OF1OF2 and W3110 pklr1 cells upon addition of IPTG to the growth media. The points of IPTG additionto the cultures are indicated by arrows.



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

ture it is 25%. A decrease in cell viability with increased ex-pression of the cloned RhsA sequences confirms that ext-a1expression is toxic to the cells and is consistent with the pre-vious observation that ORF-ex overexpression is toxic to thecells (15).

RhsA component expression increases mRNA expression oftranslational genes and ribosomal proteins. Changes inmRNA expression due to overexpression of cloned RhsA se-quences were measured from IPTG-induced W3110 pklr1 cellsharvested 1.5 and 3.5 h after IPTG induction. The fold changein gene expression in the induced W3110 pklr1 cells at thesetime points was calculated with respect to the uninduced cellsharvested at the same time points. Only the genes showing astatistically significant change in mRNA expression were con-sidered for further analysis. At least a 4-fold increase in RhsAexpression is observed in IPTG-induced W3110 pklr1 cells.This increase is expected because RhsA components are over-expressed in these cells. A greater-than-2-fold increase in theexpression of lac operon genes (lacA, lacY, and lacZ) is alsoobserved in IPTG-induced W3110 pklr1 cells, as expected.Figure 4 (left panel) shows the numbers of genes in differentfunctional categories that show an increase or decrease inexpression upon overexpression of RhsA elements. A largenumber of genes involved in translation and posttranslationalmodification show an increase in expression in IPTG-induced

W3110 pklr1 cells. Many genes coding for 30S and 50S ribo-somal subunit proteins show at least a 2-fold increase in ex-pression in IPTG-induced W3110 pklr1 cells. Figure 5 showsthe fold change in expression of some genes coding for 30Sribosomal subunit proteins. Many genes involved in transcrip-tion, RNA processing and degradation, nucleotide biosynthesisand metabolism, and DNA replication and repair also show anincrease in expression in IPTG-induced W3110 pklr1 cells. Thefactor for inversion stimulation, fis, which is involved in regu-lation of rRNA synthesis (13), shows at least an 11-fold in-crease in expression in IPTG-induced W3110 pklr1 cells. Asubset of the genes belonging to the above-mentioned func-tional categories is listed in Table 2 with corresponding foldchanges in mRNA expression upon induction with 0.1 and 1mM IPTG.

A greater number of genes involved in amino acid biosyn-thesis and metabolism and in central intermediary metabolismshow a decrease in mRNA expression in IPTG-induced W3110pklr1 cells. To serve as a control for the above-mentionedanalysis, the mRNA expression of 1 mM IPTG-induced W3110pklr2 cells (lacking RhsA elements) was compared to that ofuninduced W3110 pklr2 cells. Apart from the increase in theexpression of lac operon genes (lacA, lacY, and lacZ), no sta-tistically significant changes in gene expression were detected.No significant changes were observed between 1 mM IPTG-induced W3110 pklr2 cells and uninduced pklr1 cells for geneslisted in Table 2 (data not shown).

Based on GeneChip probe array measurements, manytRNAs show a decrease in expression in the IPTG-inducedW3110 pklr1 cells compared to the uninduced cells. This de-crease in expression is observed in the case of tRNAs that arenot present on rRNA (rrn) operons.

Translation and posttranslational modification genes showincreased protein expression upon RhsA component expres-sion. To measure changes in protein expression, we used anisobaric-tagging-based shotgun proteomics approach. In thisapproach, four isobaric reagents were used to label peptidesfrom four different protein samples of interest, thus enablingrelative quantitation by tandem mass spectrometry of proteinsin four different biological samples in parallel (1). Cell lysatesfrom W3110 pklr1 cultures harvested at 0 h (0 mM IPTG) and3.5 h (0, 0.1, and 1 mM IPTG) were processed independentlyto generate peptides for labeling and further analysis. Weidentified and quantified 780 unique proteins in each of the

FIG. 2. Different E. coli cell types and plasmid constructs used inthis study. The dashed lines correspond to the pklr1 sequence missingin the respective plasmids. The filled black box in pklr1OF1OF2 cor-responds to the region where eight rare codons were replaced byfrequently used codons. The phenotypic responses of the cell types toIPTG induction are summarized. Growth reduction refers to a lowerfinal OD.

FIG. 3. Effect of ext-a1 expression on cell viability. W3110 pklr1 cells induced with 0, 0.1, and 1 mM IPTG were harvested 3.5 h after IPTGinduction. The cells were stained using the Live/Dead BacLight cell staining kit and observed under a fluorescence microscope (at �100magnification) as described in Materials and Methods. The live cells fluoresce green, and dead cells fluoresce red.



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

four samples using the above approach. As expected, LacZexpression is observed to increase 5-fold in W3110 pklr1 cellsinduced with 0.1 and 1 mM IPTG compared to the uninducedcells. The proteins quantified in this study were broadly clas-sified into different categories based on their functions. Figure4 shows the number of proteins in each functional categoryexhibiting either an increase or a decrease in expression in 1mM IPTG-induced W3110 pklr1 cells with respect to the un-induced cells harvested at 3.5 h after IPTG addition to themedium. Many proteins involved in translation and posttrans-lational modification show an increase in expression in IPTG-induced W3110 pklr1 cells. However, a greater number ofproteins in many functional categories, including amino acidbiosynthesis, cell processes, metabolism, and transport, show adecrease in expression upon IPTG-induced overexpression ofRhsA elements. There also appears to be an overall reductionof protein expression at the genome-wide level in IPTG-in-duced W3110 pklr1 cells. Of a pool of 780 genes, 22% show agreater-than-50% decrease in protein expression in 1 mMIPTG-induced cells compared to the uninduced cells (Table 3).Furthermore, the number of proteins showing reduced expres-sion increases with an increase in the amount of IPTG addedto the medium. For example, in 1 mM IPTG-induced cells,

approximately 11% of proteins show a greater-than-2-fold de-crease in expression compared to 6% of proteins in 0.1 mMIPTG-induced cells.

RhsA component overexpression increases ribosome abun-dance. The mRNA and protein expression data suggest thatthe genes involved in translation, including the genes codingfor 30S and 50S ribosomal subunit proteins, show an increasein expression in IPTG-induced W3110 pklr1 cells. To investi-gate whether an increase in gene expression of ribosomal sub-unit proteins results in a higher intracellular ribosome concen-tration, ribosomes from W3110 pklr1 cells harvested 1.5 and3.5 h after IPTG induction (0, 0.1, and 1 mM) were purifiedusing sucrose density gradients. Figure 6 shows the absorbanceprofiles of the sucrose density gradient fractions for samplesharvested at 3.5 h. In the case of uninduced cells, three distinctpeaks are observed, and in the case of IPTG-induced cells,more than three peaks are observed in the absorbance profiles.These peaks are labeled peaks 1 to 3 in Fig. 6a and peaks 1 to5 in Fig. 6b and c. To establish the identities of these peaks, weperformed a proteomic analysis on the density gradient frac-tions containing the peaks. Proteomic analysis revealed thepresence of 30S subunit proteins in peak 1 and 50S subunitproteins in peak 2; thus peaks 1 and 2 were labeled 30S and

FIG. 4. Changes in mRNA and protein expression in genes belonging to different functional categories due to overexpression of ext-a1. Thedata reported in the figure correspond to 1 mM IPTG-induced W3110 pklr1 cells collected 3.5 h after IPTG induction. Changes in mRNA andprotein expression in 1 mM IPTG-induced cells were calculated with respect to the uninduced cells harvested at 3.5 h.



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

50S, respectively. Both 30S and 50S subunit proteins wereobserved in peaks 3, 4, and 5, suggesting the presence of 70Ssubunit monomers or multimers in these peaks. We labeledthese peaks 70S, 100S, and polysome fractions, and this assign-ment is discussed below.

As seen in Fig. 6, there is an increase in the peak areas of30S, 50S, and 100S subunits per cell optical-density unit in theIPTG-induced W3110 pklr1 cells compared to the uninducedW3110 pklr1 cells. An increase in the peak areas is consistentwith an increase in the concentrations of 30S, 50S, and 100Ssubunits in the induced cells. Similar density gradient absor-bance profiles were obtained for induced W3110 pklr1 samplesharvested at 1.5 h (data not shown). As a control, ribosomesfrom W3110 pklr2 cells (lacking cloned RhsA sequences andthe tac promoter) harvested at 3.5 h (from 0 and 1 mM IPTG-induced cultures) were also quantified. As seen in Fig. 7, thedensity gradient absorbance profiles of the uninduced and in-duced W3110 pklr2 cells are very similar to each other and tothe absorbance profile of uninduced W3110 pklr1 cells. Theseobservations are consistent with an increase in the concentra-tions of different ribosome subunits and polysomes in theribosome density gradient absorbance profiles of inducedW3110 pklr1 cells upon overexpression of the cloned RhsAsequences.

RhsA component overexpression alters rRNA expression. Toinvestigate the increase in the concentrations of 30S, 50S, and100S subunits, we analyzed total RNA from W3110 pklr1 cellsharvested at 1.5 and 3.5 h after IPTG induction (0, 0.1, and 1mM) as described in Materials and Methods. Approximately15 to 20% greater amounts of RNA were extracted consistentlyfrom IPTG-induced W3110 pklr1 cells than from equal num-bers (based on optical density measurements) of uninducedcells. Assuming equal extraction efficiencies, an increase intotal RNA indicates an increase in the concentration of rRNAin IPTG-induced cells, which is consistent with the observedincrease in the abundance of ribosomal subunits in these cells.The RNA extracted from these cells was further analyzed using

denaturing agarose gel electrophoresis (Fig. 8). Two distinctbands corresponding to the 16S and 23S rRNAs are observedon the gel in the lanes containing RNA extracted from unin-duced cells (at 1.5 and 3.5 h). However, in the lanes containingRNA extracted from IPTG-induced W3110 pklr1 cells, threemore bands, apart from those corresponding to 16S and 23S,are observed. Based on the sizes and locations of these bandson the gel, we have designated them 23S, 16S precursor(16Sp), and 5S precursor (5Sp). Sequencing results for partialregions of 23S and 16Sp (see Materials and Methods fordetails) suggest that the RNAs in these bands have very highsequence homology (99% in the sequenced region) to 23Sand 16S rRNA, respectively. 5Sp could not be sequenced dueto technical limitations. An image analysis of the gel revealsthat the intensities of the 23S, 16Sp, and 5Sp bands increasewith an increase in the expression of the cloned RhsA se-quences. The intensities of these bands are higher in the 3.5-hsamples than in the 1.5-h samples. Also, at any given timepoint, the intensities of these bands increase with the amountof IPTG added to the growth medium.

The RNA from sucrose density gradient fractions corre-sponding to 30S, 50S, 70S, 100S, and polysome peaks wasextracted as described in Materials and Methods and analyzedusing denaturing agarose gel electrophoresis. As seen in the gelimages in Fig. 9, the bands corresponding to 16S and 23S arepredominantly observed in the 30S and 50S peaks, respectively,as expected. The presence of a 16S band in the 50S peak anda 23S band in the 30S peak suggests incomplete separation ofthe 30S, 50S, and 70S ribosomal peaks during sucrose densitygradient fractionation. A band corresponding to 16Sp RNA ispresent only in the 30S peak fraction, while a 23S RNA bandis present in the 50S, 70S, and 100S peak fractions in IPTG-induced W3110 pklr1 cells. Some 23S RNA is also present inthe polysome peak fractions (data not shown). Further, theintensity of the 16Sp band increases with the amount of IPTGadded to induce the expression of the cloned RhsA sequences.

FIG. 5. Log2 fold change in mRNA expression of genes coding for 30S ribosomal subunit proteins in W3110 pklr1 cells induced with 0.1 and1 mM IPTG and harvested 3.5 h after IPTG induction. Fold changes were calculated with respect to the uninduced W3110 pklr1 cells harvestedat 3.5 h. The error bars indicate standard deviations in the calculated fold change values.



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

Image analysis reveals a similar trend in the intensity of the23S band.

RhsA component overexpression-induced change in rRNAexpression correlates with a decrease in final OD. IPTG-in-duced and uninduced W3110 pklr1 cells were harvested at3.5 h for reinoculation into fresh medium (lacking IPTG).Cultures inoculated with IPTG-induced W3110 pklr1 cells ex-hibited a longer lag phase to reach the exponential growth

phase than uninduced cells (Fig. 10). However, in the absenceof IPTG, the two cultures exhibited similar growth rates in theexponential phase. Upon addition of IPTG to the growth me-dia (at an OD600 of 0.3), cultures inoculated with inducedW3110 pklr1 cells exhibited a reduction in the final OD, asexpected. These observations suggest that the effect of expres-sion of cloned RhsA sequences on cellular growth, as measuredby OD, is not permanent and that these cells can resume awild-type W3110 pklr1 phenotype upon introduction into freshgrowth media. Analysis of RNA extracted from IPTG-inducedW3110 pklr1 cells used for inoculation and from cells har-vested from reinoculated cultures at the point of IPTG induc-tion and at 3 h postinduction (circled in Fig. 10) suggests thatin the absence of IPTG (and expression of the cloned RhsAsequences) the 23S RNA band disappears as cells reach ex-ponential phase and enter early stationary phase. However, inthe presence of IPTG, the intensity of the 23S RNA bandincreases with culture time. These observations suggest astrong correlation between the compromised cell growth stateand the presence of 23S RNA in W3110 pklr1 cells.

DISCUSSION

The sequence of Rhs elements consists of discrete structuralcomponents with different evolutionary backgrounds. Some ofthese structural components are preceded by ribosome bindingsites and start codons, suggesting a potential for the translationand physiological role of Rhs elements inside cells. RhsA caninclude five different ORFs. Although the function of the en-tire RhsA element or each of its potential ORFs is not known,overexpression of core ORF extension (ORF-ex) has beenobserved to impart toxicity to cells (15). In the current work,we studied the effect of overexpressing the RhsA components,ext-a1 and partial sequence of dsORF-a1 (both componentsare parts of ORF-ex) on E. coli cells and confirmed that ex-pression of the cloned ext-a1 sequence is toxic to the cells andreduces cell growth, as measured by OD.

Our studies suggest that overexpression of ext-a1 dramati-cally affects the transcriptional, as well as translational, pro-cesses inside E. coli W3110 cells. Gene expression analysis atthe mRNA and protein levels reveals that many genes involvedin translation and posttranslational processing have a morethan 2-fold increase in expression in cells expressing ext-a1.Further, IPTG-induced cells expressing ext-a1 show an alteredribosome density gradient profile compared to the uninduced

TABLE 2. Fold changes in gene expression in IPTG-inducedW3110 pklr1 cells 3.5 h after IPTG induction

Genea Blattner no.

Log2(fold change) incells induced with

IPTG at:b

0.1 mM 1 mM

DNA replication, recombination,modification, and repair

rnpA b3704 3.7 (0.8) 4.6 (0.7)fis b3261 3.6 (0.8) 3.6 (0.7)recR b0472 1.4 (0.5) 3.5 (0.4)priB b4201 2.9 (1.1) 3.1 (1.0)ruvC b1863 1.1 (0.7) 2.4 (0.9)gidB b3740 1.5 (0.5) 1.8 (0.4)rnc b2567 1.3 (0.7) 1.4 (0.7)

Nucleotide biosynthesis andmetabolism

purM b2499 3.1 (0.5) 4.8 (0.6)purF b2312 4.1 (0.4) 4.8 (0.2)purL b2557 3.1 (0.7) 4.4 (0.5)purT b1849 3.0 (1.5) 3.9 (1.4)purK b0522 2.2 (0.8) 3.8 (0.6)purE b0523 2.2 (1.4) 3.5 (1.6)purR b1658 3.0 (0.4) 3.5 (0.3)purC b2476 2.3 (0.6) 3.2 (0.4)purN b2500 1.6 (1.3) 3.0 (1.7)purH b4006 1.9 (1.5) 2.7 (2.0)cmk b0910 1.5 (0.8) 1.5 (0.8)

Transcription, RNA processing,and degradation

deaD b3162 1.8 (0.7) 2.4 (0.5)rpoB b3987 1.8 (0.2) 2.3 (0.2)rpoA b3295 1.7 (1.6) 2.2 (1.8)nusA b3169 2.0 (0.6) 2.2 (0.5)nusG b3982 1.3 (0.7) 1.8 (0.4)rho b3783 0.4 (0.3) 1.7 (0.5)rpoC b3988 0.9 (0.8) 1.2 (1.0)

Translation and posttranslationalmodification

prfC b4375 2.7 (1.3) 2.7 (1.5)infA b0884 2.7 (0.5) 2.7 (1.0)greA b3181 2.9 (0.6) 2.6 (0.5)tsf b0170 2.3 (1.0) 2.5 (1.1)efp b4147 1.7 (0.9) 2.5 (0.9)prfB b2891 1.5 (0.5) 2.3 (0.4)aspS b1866 1.8 (0.5) 2.3 (0.2)rbfA b3167 1.6 (0.6) 2.1 (0.2)trmD b2607 1.8 (0.8) 1.9 (0.5)infB b3168 1.4 (0.6) 1.8 (0.2)leuS b0642 1.2 (0.8) 1.6 (0.9)

a A subset of the genes belonging to four representative functional groups arepresented. Only genes showing a statistically significant change in expression arelisted.

b Fold changes were calculated with respect to the uninduced W3110 pklr1cells harvested 3.5 h after IPTG addition to the growth medium. The standarddeviations for the calculated fold change values are shown in parentheses.

TABLE 3. Percentages of proteins with a particular fold change inprotein expression in IPTG-induced W3110 pklr1 cells

3.5 h after IPTG induction

Foldchange (%)

% Proteins showing a change in expressiona in cells inducedwith IPTG at:

0.1 mM 1 mM

Upregulated Downregulated Upregulated Downregulated

50 2.3 19.4 6.0 22.6100 1.3 6.2 1.3 10.9300 0.4 0.6 0.4 2.7

a The change in protein expression was calculated relative to uninduced cellsharvested 3.5 h after IPTG addition to the growth medium.



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

cells (Fig. 6). Proteomic analysis of the multiple peaks ob-served in these gradient profiles suggests the presence of 70Ssubunit monomers or multimers in peaks 3, 4, and 5 (Fig. 6band c). Based on this observation, there are two plausibleexplanations for the occurrence of peaks 3, 4, and 5. In the firstscenario, peak 3 could contain altered forms of 50S subunits(possibly due to a lack of certain 50S ribosomal proteins) in the70S assembly, resulting in an overall smaller 70S size. Peak 4could contain functional 70S monomers, and peak 5 couldcontain 100S subunits (or nonfunctional 70S dimers). In asecond possible scenario, peaks 3, 4, and 5 could contain func-tional 70S monomers, 100S subunits, and polysomes, respec-tively. Proteomic analysis of peak 3 does not suggest any dis-crepancy in the stoichiometric composition of 50S ribosomalproteins (such as the absence of specific ribosomal proteins),thus ruling out the possibility of formation of alternate/smaller50S subunits (and consequently smaller 70S subunits). Further,if there were indeed alternate/smaller 50S subunits, we wouldexpect to see another peak succeeding the 50S subunit peak inthe absorbance profiles. Thus, based on the observations, thesecond scenario seems more likely. The extra peaks observed

in IPTG-induced cells correspond to 100S and polysomes. Thepeaks in Fig. 6 have been labeled according to the secondscenario discussed above.

Cells expressing the cloned RhsA sequences show an in-crease in the concentrations of 30S, 50S, and 100S ribosomesubunits. An increase in the 30S and 50S ribosome subunits hasbeen previously reported to accompany an increase in rRNAsynthesis in E. coli (5). Although we do not have direct evi-dence of increased rRNA synthesis in cells expressing thecloned RhsA sequences, we observed that mRNA expression offis, an activator of rRNA synthesis, increased 11-fold in thesecells. Further, we consistently extracted more RNA from cellsexpressing the cloned RhsA sequences, suggesting an increasein rRNA abundance (and hence, rRNA synthesis) in thesecells. From transcriptome analysis, the genes coding for ribo-somal proteins also show an increase in expression. Theseobservations are consistent with the observed increase in 30Sand 50S ribosomal subunit concentrations in cells expressingthe cloned RhsA sequences. We suspect the increase in theconcentration of 100S subunits in cells expressing the clonedRhsA sequences to be a result of increased sequestration of 70S

FIG. 6. Effect of ext-a1 overexpression on cellular ribosome content. Ribosomes were fractionated, as described in Materials and Methods,from W3110 pklr1 cells induced with 0 (a), 0.1 (b), and 1 (c) mM IPTG and harvested 3.5 h after IPTG induction. The absorbance profiles (A254)of the sucrose density gradient fractions for the treatments are shown. The peaks observed in the absorbance profiles are labeled peaks 1 to 3 inpanel a and peaks 1 to 5 in panels b and c. Based on proteomic analysis, peaks 1 to 5 were designated 30S, 50S, 70S, 100S, and polysomes.

FIG. 7. Effect of IPTG induction on cellular ribosome content of W3110 pklr2 cells. Ribosomes were fractionated, as described in Materialsand Methods, from W3110 pklr2 cells induced with 0 (a) and 1 (b) mM IPTG and harvested 3.5 h after IPTG induction. The absorbance profiles(A254) of the sucrose density gradient fractions for the treatments are shown.



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

ribosomes (which are probably more abundant as a result of anincrease in 30S and 50S subunit concentrations) into the trans-lationally inactive dimeric form (100S). An increase in thepolysome abundance in cells expressing the cloned RhsA se-quences could be due to a decrease in the polypeptide elon-gation rate during translation. A decrease in the tRNA abun-dance and expression of amino acid biosynthesis genes, asobserved in cells expressing the cloned RhsA sequences, canresult in a decreased abundance/availability of amino-acylatedtRNAs (aa-tRNAs) during translation, thus increasing ribo-some pausing/stalling on the transcript and decreasing thepolypeptide elongation rate.

A total RNA analysis suggests changes not only in RNAsynthesis, but also in RNA maturation and degradation pro-cesses inside cells expressing ext-a1. Besides 16S and 23SRNA, additional RNA sequences with 99% homology to 16Sand 23S (in the sequenced region) were observed in cells ex-pressing RhsA sequences. Based on the sequencing results andthe sizes of these RNA sequences, we have labeled these

RNAs 16S precursor RNA (16Sp) and 23S. The concentrationof 16Sp is observed to increase with culture time, as well aswith an increase in RhsA component expression inside cells. If16Sp is indeed a precursor to 16S RNA, then we can speculateon why its concentration may increase. The concentration ofrRNA precursors depends on rRNA synthesis and rRNA mat-uration (or processing) rates. The rRNA synthesis rate exceedsthe maturation rate during the exponential growth phase; thus,the concentration of rRNA precursors is higher in the expo-nential phase. However, rRNA precursor abundance decreasesas cells enter the stationary growth phase because rRNA syn-thesis slows down and eventually ceases (2). In the cells ex-pressing cloned RhsA sequences, we suspect that rRNA syn-thesis continues in the stationary phase (at 3.5 h) and therRNA synthesis rate exceeds the rRNA maturation rate, re-sulting in an increase in 16Sp over time. We do not fullyunderstand the reason for the presence of 23S in cells express-ing ext-a1. However, it is possible that 23S is a partially de-graded product of 23S. The presence of 23S RNA only in theext-a1-expressing cells with a compromised growth phenotypeindicates a correlation between the ext-a1 biology, the pres-ence of 23S RNA, and the observed phenotype of compro-mised cell growth as measured by OD. This inference is furthersupported by the observation that ext-a1-expressing cells donot enter the exponential growth phase upon introduction intofresh medium until all the 23S RNA has been processed/degraded (Fig. 10).

RNA analysis of the ribosome density gradient peaks sug-gests that 16Sp is present only in the 30S peak fraction and23S is present specifically in the 50S, 70S, 100S, and polysomepeak fractions. The abundances of 23S and 23S RNAs appearto be similar in the 70S (and 100S) subunit (from the intensityof 23S and 23S bands in Fig. 9). It is not known whether 70Sribosomes containing 23S RNA are functional. However, co-precipitation of 70S ribosomes containing 23S RNA withthose containing 23S RNA during sucrose density gradientfractionation suggests that the two ribosome assemblies havesimilar molecular weights and shapes.

Gene expression analysis of cells expressing ext-a1 also sug-gests an overall decrease in mRNA and protein expression in

FIG. 8. Effect of ext-a1 overexpression on total cellular RNA. TotalRNA was extracted and analyzed using denaturing agarose gel elec-trophoresis as described in Materials and Methods. W3110 pklr1 cellsinduced with 0, 0.1, and 1 mM IPTG and harvested 1.5 and 3.5 h afterIPTG induction were used for RNA analysis. An equal amount ofRNA from each sample was loaded on the gel. Bands corresponding to23S, 16Sp, and 5Sp were observed in the lanes containing RNA fromext-a1-expressing cells.

FIG. 9. Effect of ext-a1 overexpression on rRNA. RNA was extracted from sucrose density gradient fractions containing 30S (a), 50S (b), and70S and 100S (c) ribosomal subunits and analyzed using denaturing agarose gel electrophoresis. The density gradient fractions corresponding toW3110 pklr1 cells induced with 0, 0.1, and 1 mM IPTG were used for analysis. In panel c, the lanes marked with asterisks were loaded with RNAextracted from 100S subunit fractions. 16Sp RNA was observed only in 30S subunit fractions, and 23S and 5Sp RNA were observed in 50S, 70S,and 100S ribosomal subunit fractions in samples corresponding to IPTG-induced cells.



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

the cells. Many genes belonging to functional categories otherthan translation and posttranslational processing show a de-crease in expression. It can be speculated that the cell is di-verting its resources to ext-a1 expression perturbed rRNA andribosomal protein synthesis. A correlation between an increasein rRNA synthesis and a decrease in cellular growth has beenpreviously reported in the literature (10). In this work, ext-a1expression is observed to result in reduced final OD and cellviability. The compromised state of cells expressing ext-a1 isalso evident from the longer “lag” time taken to reach theexponential growth phase upon introduction into fresh culturemedia. Here, the samples used for gene expression and othermeasurements were not treated to screen out dead cells. Wenote that the presence of dead cells in the samples can affectour experimental measurements and confound the observa-tions and inferences. However, a hydrolyzed amino acid anal-ysis of the culture supernatants reveals increased protein abun-dance in the case of ext-a1-expressing cells with respect to theuninduced cells (data not shown). This increased protein abun-dance in the culture medium is likely due to accumulation ofcellular protein from the lysed dead cells in the cultures ex-pressing ext-a1 over time after IPTG addition.

The direct mechanism by which ext-a1 expression perturbsE. coli physiology is not understood. Nevertheless, an increasein ext-a1 expression appears to be directly correlated with adecrease in cell growth characteristics as measured by OD andviability, an increase in rRNA synthesis and processing, and anincrease in ribosomal-protein expression. Ext-a1 expressioncould affect the global regulatory mechanisms that shut downrRNA synthesis upon the approach of stationary phase. Al-

though we do not have experimental evidence, we believe thatthe ext-a1 protein product may directly interact with DNA,metabolites, or other transcription factors and affect differentregulatory mechanisms. Further experiments investigating theinteraction of the ext-a1 translation product with other biomol-ecules may provide better insight into the function and mech-anism of action of the RhsA component ext-a1.

ACKNOWLEDGMENTS

We thank Leila Choe (University of Delaware) and Sheng Zhang(Cornell University) for mass spectrometric analysis and Wilfred Chen(University of California, Riverside, CA) for providing pklr1 andpKQV4 plasmids.

This work was funded in part by USDA Agricultural ResearchService Specific Cooperative Agreement 58-1907-1-146 and NationalScience Foundation BES-0120315.

REFERENCES

1. Aggarwal, K., L. H. Choe, and K. H. Lee. 2005. Quantitative analysis ofprotein expression using amine-specific isobaric tags in Escherichia coli cellsexpressing RhsA elements. Proteomics 5:2297–2308.

2. Cangelosi, G. A., and W. H. Brabant. 1997. Depletion of pre-16S rRNA instarved Escherichia coli cells. J. Bacteriol. 179:4457–4463.

3. Dong, H. J., L. Nilsson, and C. G. Kurland. 1995. Gratuitous overexpressionof genes in Escherichia coli leads to growth inhibition and ribosome destruc-tion. J. Bacteriol. 177:1497–1504.

4. Feulner, G., et al. 1990. Structure of the RhsA locus from Escherichia coliK-12 and comparison of RhsA with other members of the Rhs multigenefamily. J. Bacteriol. 172:446–456.

5. Gourse, R. L., Y. Takebe, R. A. Sharrock, and M. Nomura. 1985. Feedback-regulation of ribosomal-RNA and transfer-RNA synthesis and accumulationof free ribosomes after conditional expression of ribosomal-RNA genes.Proc. Natl. Acad. Sci. U. S. A. 82:1069–1073.

6. Hill, C. W., et al. 1995. Correlation of Rhs elements with Escherichia colipopulation-structure. Genetics 141:15–24.

7. Hill, C. W., C. H. Sandt, and D. A. Vlazny. 1994. Rhs elements of Escherichia

FIG. 10. Reinoculation studies of uninduced and induced W3110 pklr1 cells grown in fresh media. Compared to uninduced W3110 pklr1 cells,ext-a1-expressing W3110 pklr1 cells take a longer time to reach exponential growth phase upon introduction into fresh medium lacking IPTG. Atan OD600 of 0.3, IPTG was added to a culture that was initially inoculated with induced W3110 pklr1 cells. Addition of IPTG restores the originalphenotype of a reduced growth rate in cultures inoculated with cells expressing ext-a1. At different time points (circled), RNA was extracted fromthe cells for further analysis. The gels showing RNA analysis are located close to the respective time points at which cells were harvested for RNAextraction. The intensity of the 23S RNA band initially present in induced W3110 pklr1 cells used to inoculate fresh media decreases as the cellsreach exponential phase and disappears in stationary phase in the absence of IPTG. However, the intensity of the 23S RNA band increases afteraddition of IPTG to growth media.



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

coli: a family of genetic composites each encoding a large mosaic protein.Mol. Microbiol. 12:865–871.

8. Lin, R. J., M. Capage, and C. W. Hill. 1984. A repetitive DNA sequence,Rhs, responsible for duplications within the Escherichia coli K-12 chromo-some. J. Mol. Biol. 177:1–18.

9. Morris, R. M., et al. 2007. Comparative proteomics of Dehalococcoides spp.reveals strain-specific peptides associated with activity. Appl. Environ. Mi-crobiol. 73:320–326.

10. Richins, R., and W. Chen. 2001. Effects of FIS overexpression on cell growth,rRNA synthesis, and ribosome content in Escherichia coli. Biotechnol. Prog.17:252–257.

11. Sadosky, A. B., A. Davidson, R. J. Lin, and C. W. Hill. 1989. Rhs gene familyof Escherichia coli K-12. J. Bacteriol. 171:636–642.

12. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: alaboratory manual, 2nd ed., vol. 1. Cold Spring Harbor Laboratory Press,Cold Spring Harbor, NY.

13. Schneider, D. A., W. Ross, and R. L. Gourse. 2003. Control of rRNAexpression in Escherichia coli. Curr. Opin. Microbiol. 6:151–156.

14. Tusher, V. G., R. Tibshirani, and G. Chu. 2001. Significance analysis ofmicroarrays applied to the ionizing radiation response. Proc. Natl. Acad. Sci.U. S. A. 98:5116–5121.

15. Vlazny, D. A., and C. W. Hill. 1995. A stationary-phase-dependent viabilityblock governed by 2 different polypeptides from the RhsA genetic element ofEscherichia coli K-12. J. Bacteriol. 177:2209–2213.

16. Wang, Y. D., S. Zhao, and C. W. Hill. 1998. Rhs elements comprise threesubfamilies which diverged prior to acquisition by Escherichia coli. J. Bacte-riol. 180:4102–4110.

17. Wu, Z., R. A. Irizarry, R. Gentleman, F. M. Murillo, and F. Spencer. 2004. Amodel based background adjustment for oligonucleotide expression arrays.J. Am. Stat. Assoc. 99:909–917.

18. Zahn, K. 1996. Overexpression of an mRNA dependent on rare codonsinhibits protein synthesis and cell growth. J. Bacteriol. 178:2926–2933.

19. Zhao, S., et al. 1993. Rhs elements of Escherichia coli K-12: complex com-posites of shared and unique components that have different evolutionaryhistories. J. Bacteriol. 175:2799–2808.



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

overexpression of cloned rhsa sequences perturbs the ... · translational machinery in escherichia...

Documents