Inducible suppression of global translation by overuse of rare codons.

Inducible Suppression of Global Translation by Overuse of Rare Codons Hideki Kobayashi Japan Agency for Marine-Earth Science and Technology (JAMSTEC), Yokosuka, Japan

R

ecently, artificial gene regulation systems have been developed to control gene expression or cell behavior in synthetic biology (1–6). Deans et al. developed an artificial genetic switch with RNA interference (RNAi) and a repressor protein in mammalian cells (7). In many cases, artificial gene networks have focused on switching gene expression for specific purposes by using the fluorescent protein gene as an output gene (1, 2). The control of cell behavior with artificial gene networks requires an output gene whose biochemical function is clearly defined and a deficient mutant form thereof (4). To extend the application of artificial gene networks, we need a more convenient gene as the output gene in order to control biological activity in various organisms. The present study focused on codon usage bias, and an artificial gene-silencing system was constructed that is different from the repressor protein or interfering RNA found in a natural biological system. Low-usage codons have a genome-wide distribution at a very low frequency, and their tRNAs occur in the cell at lower concentrations than those for normal codons (8–10). The expression of foreign genes is suppressed in Escherichia coli because of differences in codon usage (11, 12). Overexpression of the ␭ integrase mRNA containing 20 rare arginine codons with a strong ribosome binding site reduces the growth rate of the host E. coli (11). Therefore, the pathway of a rare codon in the genome and its corresponding tRNA is one “vulnerability” or “security hole” of biological systems, much like one in a computer system or network (13, 14). In the present study, I constructed an artificial gene containing many low-usage codons to exploit this security hole by monopolizing multiple minor codon tRNAs through its expression, resulting in the suspension of almost all translation in the cell (Fig. 1). This scheme was modeled on a type of computer system or network attack known as a distributed denial-of-service attack. I arranged this artificial gene downstream of the lac promoter of E. coli, the galactose-inducible promoter of yeast, and the doxycycline (Dox)-inducible promoter of mammalian cells. I found repression of the growth of E. coli, yeast, and mammalian cells, as well as downregulation of gene expression with the induction of my artificial gene. The application of this artificial gene provides a novel nonspecific virus defense system in E. coli and

2544

aem.asm.org

human cells. This artificial gfp gene would work as a system device with which to control cell behavior with an artificial gene network with applications in biotechnology. MATERIALS AND METHODS Plasmid construction, genes, cells, phages, and chemicals. (i) E. coli and Saccharomyces cerevisiae. The artificial green fluorescent protein (GFP) genes lgfp and hgfp were designed on the basis of the codon usage database (15) and synthesized by the GenScript Corporation (Piscataway, NJ). The C-terminal deletion mutant genes lgfp⌬1, lgfp⌬2, and lgfp⌬3, lacking 25, 50, and 75% of the length of the C terminus of lgfp, respectively, were obtained by PCR from the lgfp gene (Table 1). The codon adaptation indices (CAIs) of artificial genes were calculated with the formula (16) L log ␻i(l)], where ␻i ⫽ f(i)/max[f(j)], f(i) and f(j) CAI ⫽ exp[(1/L) l⫽1 are frequencies of synonymous codons for amino acids, and L is the number of codons. All plasmids except pRARE for bacterial experiments were constructed from pTAK, pIKE, and pTSMb1 (kind gifts from J. J. Collins, Boston University, Boston, MA) by using standard cloning techniques (2, 17, 18). pHGFP, pLGFP, pLGFP⌬1, pLGFP⌬2, and pLGFP⌬3 were constructed by inserting hgfp, lgfp, lgfp⌬1, lgfp⌬2, and lgfp⌬3, respectively, under the control of the Ptrc promoter of pTAK132 by replacing the cI857gfpmut3 genes. pRARE was isolated from the Rosetta strain (Merck, Darmstadt, Germany). pYEG, pLGFP1, and pLGFP2 were constructed by inserting a yeast-enhanced GFP gene (yEGFP), lgfp, and hgfp under the

兺

Received 11 November 2014 Accepted 21 January 2015 Accepted manuscript posted online 30 January 2015 Citation Kobayashi H. 2015. Inducible suppression of global translation by overuse of rare codons. Appl Environ Microbiol 81:2544 –2553. doi:10.1128/AEM.03708-14. Editor: S.-J. Liu Address correspondence to [email protected]. Supplemental material for this article may be found at http://dx.doi.org/10.1128 /AEM.03708-14. Copyright © 2015, American Society for Microbiology. All Rights Reserved. doi:10.1128/AEM.03708-14 The authors have paid a fee to allow immediate free access to this article.

Applied and Environmental Microbiology

April 2015 Volume 81 Number 7

Downloaded from http://aem.asm.org/ on April 26, 2015 by NYU MEDICAL CENTER LIBRARY

Recently, artificial gene networks have been developed in synthetic biology to control gene expression and make organisms as controllable as robots. Here, I present an artificial posttranslational gene-silencing system based on the codon usage bias and low tRNA content corresponding to minor codons. I engineered the green fluorescent protein (GFP) gene to inhibit translation indirectly with the lowest-usage codons to monopolize various minor tRNAs (lgfp). The expression of lgfp interfered nonspecifically with the growth of Escherichia coli, Saccharomyces cerevisiae, human HeLa cervical cancer cells, MCF7 breast cancer cells, and HEK293 kidney cells, as well as phage and adenovirus expansion. Furthermore, insertion of lgfp downstream of a phage response promoter conferred phage resistance on E. coli. Such engineered gene silencers could act as components of biological networks capable of functioning with suitable promoters in E. coli, S. cerevisiae, and human cells to control gene expression. The results presented here show general suppressor artificial genes for live cells and viruses. This robust system provides a gene expression or cell growth control device for artificially synthesized gene networks.

Artiﬁcial Control Device for Gene Expression

Genome Transcription Translation

tRNAs for minor codons

Deficiency of minor codon tRNAs Cell growth or virus expansion Monopoly of minor codon tRNAs Translation

mRNA Transcription

GFP

Plasmids

FIG 1 Schematic of posttranscriptional inhibition by the rare-codon gene. Rarecodon tRNAs (pink) are monopolized by overexpression of the artificial gfp gene. Translation of the mRNA transcribed from the genome is inhibited because of the shortage of rare-codon tRNAs. All cellular protein synthesis except that of GFP will be inhibited, and cellular activities such as cell growth will be reduced.

control of the PGAL promoter of pYES2 (Invitrogen, Carlsbad, CA). The PGAL-gfp-CYC1TT region was then obtained by PCR and cloned into pAUR112 (GenBank accession no. AB012283; TaKaRa, Kyoto, Japan). pL⌬NG was constructed from pLGFP⌬1 by the insertion of tetR-repressive gfpmut3 obtained from pIKE107. pH⌬NG was constructed from pL⌬NG by TABLE 1 Codons and their fractions in hgfp and lgfp Fractiona in: b

Fraction in:

No. in:

AA

hgfp codon

E. coli

S. cerevisiae

No.

lgfp codon

E. coli

S. cerevisiae

lgfp

lgfp⌬1

lgfp⌬2

lgfp⌬3

Gly Lys Leu Asp Val Glu Thr Ile Asn Phe Tyr Pro Ser His Ala Gln Met Arg Cys Trp Stop

GGC AAA CUG GAA GUG GAU ACC AUU AAC UUU UAU CCG AGC CAU GCG CAG AUG CGC UGC UGG UAA

0.41 0.76 0.50 0.69 0.37 0.63 0.44 0.49 0.55 0.57 0.54 0.53 0.28 0.57 0.36 0.65 1 0.41 0.56 1 0.64

0.19 0.58 0.11 0.70 0.19 0.65 0.22 0.46 0.41 0.59 0.56 0.12 0.11 0.64 0.11 0.31 1 0.06 0.37 1 0.47

23 21 18 18 16 15 15 13 13 12 12 11 9 9 9 8 7 6 2 1 1

GGG AAG CUA GAC GUA GAG ACA AUA AAU UUC UAC CCC UCA CAC GCU CAA AUG AGG UGU UGG UAG

0.15 0.24 0.04 0.31 0.15 0.31 0.13 0.07 0.45 0.43 0.43 0.12 0.12 0.43 0.16 0.35 1 0.02 0.44 1 0.05

0.12 0.42 0.14 0.35 0.21 0.30 0.30 0.27 0.59 0.41 0.44 0.15 0.21 0.36 0.38 0.69 1 0.21 0.63 1 0.23

23 21 18 18 16 15 15 13 13 12 12 11 9 9 9 8 7 6 2 1 1

19 18 11 12 14 12 11 10 10 11 9 7 6 5 5 5 5 5 2 1 1

14 10 8 8 10 9 11 3 2 9 6 6 4 2 4 4 3 3 2 1 1

9 5 6 3 6 5 6 2 1 3 1 4 3 0 1 1 1 0 1 1 1

a b

Fractions are based on E. coli W3110 and S. cerevisiae as described in Materials and Methods. AA, amino acid.



aem.asm.org

2545


tRNAs for major codons

Proteins

replacing lgfp⌬1 with hgfp⌬1, which lacks 25% of the C-terminal region of hgfp. pPAL and pPAH were constructed from pLGFP and pHGFP by replacing the LacI-PL-Ptrc regions with a psp promoter obtained from E. coli AK1 by PCR. The Pyrobest DNA polymerase (TaKaRa) was used for PCR. E. coli Mach1 [F⫺ ␾80(lacZ)⌬M15 ⌬lacX74 hsdR(rK⫺ mK⫹) ⌬recA1398 endA1 tonA; Invitrogen] was used for cloning and GFP expression experiments. E. coli K-12 XL-10 [deoR endA1 gyrA96 hsdR17(rK⫺ mK⫹) recA1 relA1 supE44 thi-1 ⌬(lacZYA-argF)U169 ␾80dlacZ⌬M15 F⫺ ␭⫺ PN25/tetR placIq/lacI; Clontech] was also used for GFP expression experiments as a relA mutant. E. coli JM2.300 [lacI22 LAM e14 rpsL135(Strr) xyl-7 mtl-1 thi-1] was used for phage infection experiments. E. coli AK4 is a ␭-defective mutant derived from E. coli AK1 (4) by UV mutagenesis that was also used for phage infection experiments. Phages T4 (NBRC20004) and T7 (NBRC20007) were purchased from the NITE Biological Resource Center (Kazusa, Japan). Phages ␭ (NCIMB10451), f1 (NCIMB13926), and MS2 (NCIMB10108) were purchased from the National Collection of Industrial, Food, and Marine Bacteria (Aberdeen, United Kingdom). S. cerevisiae YPH499 (MATa his3-⌬200 leu2-⌬1 lys2-801 trp1-⌬1 ade2-101 ura352) was also used for GFP expression (19). (ii) Human cells. The artificial GFP gene hu-lgfp was also designed and synthesized on the basis of Homo sapiens codon usage. pTRE-G1 was constructed by arranging hu-lgfp downstream of the Ptet promoter of the pTRE-Tight vector (Clontech). Plasmid pTRE-Luc carrying a luciferase gene instead of the hu-lgfp gene was used as a control plasmid. HeLa-TetON, MCF7-Tet-ON, and HEK293-Tet-ON cells carrying the tetR gene in their genomes (Clontech) were used for GFP expression and recombinant adenovirus infection experiments. Recombinant adenovirus was purchased from TaKaRa Bio (Shiga, Japan). The DNA sequences of the artificial gfp genes and all of the plasmids used in this study are described in the supplemental material. Growth conditions and chemicals. All E. coli cells were incubated in LB broth (Difco Laboratories, Detroit, MI) containing 100 ␮g/ml of ampicillin (Sigma, St. Louis, MO) at 37°C and 160 rpm. Growth of E. coli was monitored by measuring the optical density at 660 nm (OD660) and counting the CFU. Chloramphenicol was also added to the cultures of E. coli carrying pRARE. The isopropyl-␤-D-thiogalactopyranoside (IPTG;

Kobayashi

2546

aem.asm.org

Infection and counting recombinant adenovirus. HEK293 cells transfected with pTRE-G1 or pTRE-Luc were incubated for 48 h at 37°C. Recombinant adenovirus was added to each culture at a concentration of 1.6 ⫻ 105 infectious units/ml. The surviving cells and recombinant adenovirus particles were counted after 72 h at 37°C. The number of recombinant adenovirus particles was determined with the Adeno-X Rapid Titer kit (Clontech) with HEK293 cells as the host (22). Measurement of hu-lgfp mRNA by qRT-PCR. HeLa cells were incubated for 48 h at 37°C after the transfection of pTRE-G1. mRNA was isolated with RNeasy (Qiagen). The quality of the mRNA was verified with the Agilent RNA 6000 nano kit and an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA). A quantitative real-time PCR (qRT-PCR) assay was carried out with TaqMan one-step RT-PCR master mix reagent on the Applied Biosystems 7500 real-time PCR system (Applied Biosystems, Carlsbad, CA) under the following amplification conditions for the above-mentioned genes: 95°C for 10 min, followed by 60 cycles at 95°C for 15 s and 60°C for 1 min. Forward primer 100609-187F (CGACGCTATC GTATGGTGTACAAT), reverse primer 100669-337R (ACTTCCGCACG CGATTTATAAT), and the TaqMan 100669-271T probe (AAGGTTATA TACAAGAACGTACGATA) were used for measurement of hu-lgfp mRNA. Human ␤-actin mRNA was used as a standard. Nucleotide sequence accession numbers. The DNA sequences of lgfp and hgfp have been deposited in the DDBJ database under accession numbers LC018331 and LC018335, respectively. All of the genes, plasmids, microorganisms, and human cell lines used in this study and the accession numbers of DNA sequences are summarized in Table 2.

RESULTS

Effect of lgfp expression on the growth, salinity response, and gene expression of E. coli. I constructed two artificial genes based on the amino acid sequence of the GFP-encoding gfpmut3 gene (23); one with the lowest- or lower-usage codons (lgfp) and another with the highest-usage codons (hgfp; Table 2 and Fig. 2A). The CAIs of gfpmut3, hgfp, and lgfp were 0.595, 1.00, and 0.394, respectively (16). I constructed pHGFP and pLGFP by arranging the hgfp or lgfp gene downstream of an IPTG-inducible Ptrc-2 promoter (LacI repressed) (2). Both plasmids had the pBR322 ColE1 origin of replication and an ampicillin resistance gene. Expression of lgfp repressed the growth of E. coli without killing the cells; the number of CFU of E. coli neither increased nor decreased for 4 h after induction with IPTG and then increased, whereas induction of hgfp had no effect on cell growth (Fig. 2B). The level of GFP expression from lgfp was about one-third of that from hgfp at 2 h of incubation (Fig. 2C). Next, I checked the response of outer membrane proteins OmpF and OmpC to salinity change under lgfp expression in order to investigate the effect of lgfp induction on the gene networks in E. coli. OmpF was expressed under low-salinity conditions and then repressed, and OmpC was expressed according to the increase in salinity (Fig. 2D, lanes 1 and 3) (24). There was no response to salinity under lgfp expression (Fig. 2D, lane 2). I also examined the effects of the induction of lgfp on ALP activity, which is encoded by one of the housekeeping genes in E. coli cells and is independent of the cell cycle (50, 51). Induction of lgfp expression suppressed ALP production to 2.8% of that in controls at 3 h of incubation (see Fig. S1 in the supplemental material). I observed no growth inhibition upon lgfp induction with ⬍1 mM IPTG. E. coli carrying the pLGFP plasmid containing the p15A or SC101 origin of replication at a medium or low copy number showed normal growth (see Fig. S2 in the supplemental material). The increase in the CFU count after 4 h is dependent on the increases in the levels of




Sigma) and anhydrotetracycline (aTc; Acros Organics, Geel, Belgium) were used to induce the Ptrc and Ptet promoters. S. cerevisiae was cultured in yeast extract-peptone-dextrose (YPD) medium (Difco) containing 0.5 ␮g/ml aureobasidin A (TaKaRa). YPGalactose medium containing 1% (wt/vol) yeast extract (Difco), 2% (wt/vol) Polypeptone (Wako Pure Chemical Industries, Japan), 2% (wt/vol) galactose, 1% raffinose, and 0.5 ␮g/ml of aureobasidin A was used to induce GFP expression. All S. cerevisiae cells were grown at 30°C and 200 rpm. Human cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Invitrogen) containing 10% (vol/vol) fetal bovine serum, penicillin (50 IU/ml)-streptomycin (50 ␮g/ml) (MP Biomedicals), and 200 ␮g/ml G418 (Clontech). All cells were incubated in suitable culture dishes in a 5% CO2 incubator at 37°C. Growth of cells was measured with a OneCell counter by microscopy after cells were removed from the culture dishes with TrypLE express (Invitrogen). All cell cultures were incubated in a CO2 incubator at 37°C. Plasmid transfection of human cells. Plasmid transfection was performed with Xfect transfection reagent (Clontech). Cells were grown in DMEM until half confluent at 37°C. pTRE-G1 or pTRE-Luc was then transfected into cells according to the prescribed protocol. The plasmid was isolated and purified with the EndoFree Plasmid Maxi kit (Qiagen, Germantown, MD). GFP expression and quantification. E. coli cells were grown aerobically in LB medium containing ampicillin overnight at 37°C, diluted 1:500, and regrown in LB medium containing the appropriate inducer at 37°C. Cells were collected by centrifugation at 8,000 rpm for 1 min at 4°C, washed with phosphate-buffered saline (PBS; 75 mM sodium phosphate, 67 mM NaCl, pH 7.4), and suspended in PBS for GFP measurement. S. cerevisiae cells were grown aerobically in YPD medium containing aureobasidin A for 24 h at 30°C. Cells were collected and washed twice with YPGalactose and then suspended in fresh YPGalactose containing raffinose and aureobasidin A, and the OD660 was adjusted to about 0.1. Cells were collected by centrifugation at 1,000 rpm for 5 min at 4°C and suspended in PBS for GFP measurement. All GFP expression data were collected with a FACSCalibur flow cytometer (Becton Dickinson, Franklin Lakes, NJ). All fluorescent measurements of gene expression were obtained from samples of 30,000 cells. Human cells were grown in DMEM for 4 h after transfection. Dox was added to the culture for induction of low GFP at a concentration of 1 ␮g/ml. GFP was observed with a fluorescence microscope. Preparation and electrophoresis of outer membrane protein of E. coli. E. coli cells were harvested (5,000 rpm, 10 min, 4°C), washed with 0.8% (wt/vol) NaCl, and then suspended in 0.5% (wt/vol) N-lauroylsarcosinate. Cells were broken by sonication (20 W, 2 min, 4°C). After centrifugation (2,000 rpm, 10 min, 4°C) to remove unbroken cells, the supernatant was centrifuged (100,000 ⫻ g, 1 h, 4°C) to collect the N-lauroylsarcosinate-insoluble fraction of the outer membrane protein. The outer membrane protein was dissolved in 2% SDS solution, and electrophoresis with a 4 M urea–12% acrylamide gel was carried out by the method of Laemmli (20). ALP activity. Cells were collected from 10-ml cultures by centrifugation at 8,000 rpm for 5 min at 4°C, washed with deionized distilled water (DDW), and then suspended in 300 ␮l of DDW. A 10% (vol/vol) toluene solution was added to each cell suspension to increase the permeability of the substrate. The alkaline phosphatase (ALP) reaction was carried out by incubation at 37°C for 3 h after aliquots of 10-␮l cell suspensions and 100 ␮l of p-nitrophenylphosphate solution (Wako Chemical, Richmond, VA) were mixed. The absorbance at 405 nm (A405) of the solution was measured as ALP activity (21). Phage experiments. E. coli JM2.300 was used as the host for phages T7, T4, and ␭, and E. coli AK4 was used as the host for phages MS2 and f1. Phage solution was prepared from phage-infected E. coli cultures by filtration with a 0.2-␮m-pore-size membrane filter after cells were removed by centrifugation at 8,000 rpm for 10 min at 4°C. The titer of the phage solution in PFU was measured on LB plates covered with LB (T4 and T7), LB plus 0.2% maltose (␭), or 1/4LB soft agar (0.8%; f1 and MS2) containing each host strain.


TABLE 2 Artificial gfp genes, plasmids, host strains, cell lines, phages, and virus used in this study Accession no.

Description

Genes gfp lgfp lgfp⌬1 lgfp⌬2 lgfp⌬3 hgfp hgfp⌬1 hu-lgfp

JC175864 LC018331 LC018332 LC018333 LC018334 LC018335 LC018336 LC018337

The original gfpmut3 gene (2) The gfpmut3 gene with all codons changed to those with the lowest usage in E. coli The modified lgfp gene with 180 bp (25%) deleted from C terminus The modified lgfp gene with 357 bp (50%) deleted from C terminus The modified lgfp gene with 537 bp (75%) deleted from C terminus The gfpmut3 gene with all codons changed to those with the highest usage in E. coli The modified hgfp gene with 180 bp (25%) deleted from C terminus The AcGFP1 gene with all codons changed to those with the lowest usage in H. sapiens

Plasmids pLGFP pHGFP pLGFPm pLGFPl pLGFP⌬1 pLGFP⌬2 pLGFP⌬3 pL⌬NG pH⌬NG pPAL pPAH pRARE pYEG pYLGFP1 pYLGFP2 pTRE-Luc pTRE-G1

LC018338 LC018344 LC018342 LC018342 LC018339 LC018340 LC018341 LC018348 LC018349 LC018350 LC018351 NAa LC018345 LC018346 LC018347 NA LC018352

IPTG-inducible lgfp, ColE1 origin, Ampr IPTG-inducible hgfp, ColE1 origin, Ampr IPTG-inducible lgfp, p15A origin, Ampr IPTG-inducible lgfp, SC101 origin, Ampr IPTG-inducible lgfp⌬1, ColE1 origin, Ampr IPTG-inducible lgfp⌬2, ColE1 origin, Ampr IPTG-inducible lgfp⌬3, ColE1 origin, Ampr IPTG-inducible lgfp⌬1, aTc-inducible gfp, ColE1 origin, Ampr IPTG-inducible hgfp⌬1, aTc-inducible gfp, ColE1 origin, Ampr f1 phage infection-inducible lgfp, ColE1 origin, Ampr f1 phage infection-inducible hgfp, ColE1 origin, Ampr Carries rare-codon tRNAs, p15A origin, chloramphenicol resistance Galactose-inducible yEGFP in S. cerevisiae, aureobasidin A resistance, CEN/ARS Galactose-inducible lgfp in S. cerevisiae, aureobasidin A resistance, CEN/ARS Galactose-inducible hgfp in S. cerevisiae, aureobasidin A resistance, CEN/ARS Dox-inducible luciferase in mammalian cells, G418 resistance Dox-inducible hu-lgfp in mammalian cells, G418 resistance

Host strains E. coli Mach1 XL-10 JM2.300 AK4 S. cerevisiae YPH499

F⫺ ␾80(lacZ)⌬M15 ⌬lacX74 hsdR(rK⫺ mK⫹) ⌬recA1398 endA1 tonA deoR endA1 gyrA96 hsdR17(rK⫺ mK⫹) recA1 relA1 supE44 thi-1 ⌬(lacZYA-argF)U169 ␾80dlacZ⌬M15 F⫺ ␭⫺ PN25/tetR placIq/lacI lacI22 LAM e14 rpsL135 (Strr) xyl-7 mtl-1 thi-1 ␭⫺ lacI::Km mutant of K-12 (CGSC7296) wild-type strain MAT␣ his3-⌬200 leu2-⌬1 lys2-801 trp1-⌬1 ade2-101 ura3-52

Human cell lines HeLa-Tet-ON MCF7-Tet-ON HEK293-Tet-ON

HeLa cells optimized for Tet-ON gene expression system MCF7 cells optimized for Tet-ON gene expression system HEK293 cells optimized for Tet-ON gene expression system, host of recombinant adenovirus

Phages T4 T7 ␭ f1 MS2

Enterobacterial phage T4, linear double-stranded DNA phage Bacteriophage T7, linear double-stranded DNA phage Enterobacterial phage l, linear double-stranded DNA temperate phage Enterobacterial phage f1, a filamentous single-stranded DNA phage Bacteriophage MS2, an icosahedral linear single-stranded RNA⫹ phage

Recombinant adenovirus

A recombinant virus vector for gene transfer, no gene was coded in this expt

a

NA, not available.

rare-codon tRNAs because the effects of growth repression by pLGFP were abrogated by cotransformation of pRARE carrying tRNA genes of all types with rare codons (see Fig. S3 in the supplemental material) (27). The cessation of peptide synthesis because of the shortage of rare-codon tRNAs may cause a pseudosignal of amino acid starvation in E. coli and activate a ppGpp cascade, which in turn inhibits growth (28). In fact, Brinkmann et al. reported that the


expression of certain genes rich in rare arginine codons in E. coli activated the ppGpp cascade and slowed the cellular growth rate (29). However, lgfp expression inhibited the growth of E. coli K-12 XL-10, which lacks the relA gene required for ppGpp synthesis (see Fig. S4 in the supplemental material). These observations indicated that the effects of my system were independent of natural gene regulation of the ppGpp network. Taken together, these observations indicate that lgfp gene expression silenced the ex-


aem.asm.org

2547


Gene, plasmid, host strain, cell line, phage or virus

1 0.8 0.6 0.4 0.2 0

hgfp

10 10

lgfp

CFU (/ml)

Fraction (-)

Kobayashi

10 9

hgfp hgfp + IPTG lgfp

10 8

lgfp + IPTG

10 7 10 6

1

51

101

151

201

10 5

0

1

2

3

4

5

6

7

8

Time (h) pHGFP

(kDa) M 1

0 0 0 10 102 104 100

0 102 104 100

102 104

Fluorescence signal (arbitrary unit)

58.5

256 Counts

Counts

256

163.7

90.5 256

31.2

Fluorescence signal (arbitrary unit)

4h Incubation time

3

45.0

0 0 0 10 10 2 10 4 0 10 0 10 2 10 4 10 0 10 2 104

2h

2

97.4

95.2

Counts

pLGFP

136.6 256 Counts

204.2 256 Counts

Counts

256

OmpC OmpF OmpA

21.5

6h

FIG 2 Construction and effects of expression of engineered hgfp and lgfp. (A) Fractions of all of the codons used in hgfp (blue) and lgfp (red). The fraction of codon usage was calculated from the codon usage database (http://www.kazusa.or.jp/codon/) for E. coli strain W3110. (B) Effects of hgfp (blue circles) and lgfp (red triangles) expression on E. coli growth. Cultures of E. coli carrying pHGFP and pLGFP were incubated as described in Materials and Methods. IPTG was added to one of two cultures at a concentration of 10 mM to induce GFP expression from the artificial gfp gene (solid symbols). No IPTG was added to controls (open symbols). Error bars show standard deviations (n ⫽ 3). (C) GFP expression from pHGFP and pLGFP at 2 to 6 h after induction. The mean fluorescent signal values (arbitrary units) are indicated in the graph. GFP was measured as described in Materials and Methods. (D) Effects of induction of lgfp on the expression of E. coli OmpC and OmpF. Outer membrane proteins were prepared from cell samples as described in Materials and Methods. E. coli carrying pLGFP was grown in modified LB not containing NaCl (lane 1), and cells were harvested and incubated in LB medium (1% NaCl) with 2 mM IPTG (lane 2) or without IPTG (lane 3) for 2 h.

pression of various genes, including those required for stringent responses to amino acids. Next, I investigated the number of codons needed for repression of E. coli growth. I constructed three deletion mutant genes, lgfp⌬1, lgfp⌬2, and lgfp⌬3 lacking 25, 50, and 75% of the length of the C terminus of lgfp, respectively. The lgfp, lgfp⌬1, lgfp⌬2, and lgfp⌬3 genes had 111, 83, 59, and 36 rare codons (fractions ⱕ0.15), respectively (Table 1). pLGFP⌬1, pLGFP⌬2, and pLGFP⌬3 were constructed from pLGFP by replacing lgfp with one of the three mutant genes, respectively. The products of the three deletion mutant genes showed no fluorescence. Induction of lgfp and lgfp⌬1 repressed the growth of E. coli, whereas no such effect was observed with lgfp⌬2 or lgfp⌬3 (Fig. 3A). In the expression of the gene containing many rare codons, those close to the N terminus contributed to the stringent response to amino acids (11, 17). In my system, the total number of low-usage codons per gene was important for growth suppression. To confirm posttranscriptional gene silencing by expression of a gene rich in low-usage codons, I constructed the pL⌬NG plasmid with lgfp⌬1 and the reporter gene gfpmut3 (Fig. 3B); both gfpmut3 and lgfp⌬1 were arranged downstream of Ptet and Ptrc, which are independently inducible with aTc and IPTG, respectively. I also constructed control plasmid pH⌬NG containing hgfp⌬1 (75% of hgfp from the N terminus) instead of lgfp⌬1 (Fig. 3B). Figure 3C shows the effect of lgfp⌬1 or hgfp⌬1 on the

2548

aem.asm.org

expression of gfpmut3. With lgfp⌬1 induction, GFP expression from gfpmut3 was repressed to one-quarter of that in the absence of induction after 4 h of incubation (Fig. 3C). The induction of hgfp⌬1 kept GFP expression at 93.5% of that in the absence of induction, confirming that my rare-codon gene can repress the expression of other genes as designed. The extent of repression of GFP expression by lgfp⌬1 was less than that of ALP expression (Fig. 2C). In this experiment, gfpmut3 was placed under the control of the strong Ptet promoter and cloned into a high-copy vector. As a result, high levels of gfpmut3 mRNA were transcribed from the plasmids and would therefore have more chances to bind to rare-codon tRNAs in the cell. Effect of lgfp expression on phage expansion. My artificial gene silencing system provides a novel control system for biological activities. One potential application of this system is as a nonspecific virus defense because all viruses require the translation step of protein synthesis in the host to facilitate expansion. I examined the effects of lgfp expression on outbreaks of various types of phage, i.e., double-stranded DNA phages T4 and T7, temperate phage ␭, single-stranded DNA filamentous phage f1, and singlestranded RNA phage MS2, in E. coli cells (30). The overexpression of lgfp suppressed the expansion of highly virulent phages T7 and T4 to as little as a factor of 20, as well as cell lysis, whereas neither expression of hgfp nor suppression of gfp gene expression prevented phage expansion or cell lysis (Fig. 4A). Both phages T4 and




Number of codons


10 10

( ( ( (

108

7 10

6 10

0

1

2

3

4

5

6

8

7

Time (h) aTc

IPTG

P tet PL

gfpmut3

tetR

P L P trc

lacI

hgfp∆1 or lgfp∆1

pL∆NG or pH∆NG GFP lgfp∆1 Counts

200

0

2.1

2.2

0

10

2

10

4

0

10 10

2

10

21.2

84.4

4

0

10 10

2

10

4

10 100

4

2

10

10

Counts

hgfp∆1 Fluorescence signal (arbitrary unit) 200

0

2.4

0

10

2

10

145.9

2.6

4

0

10 10

2

10

4

0

10 10

2

10

134.2

4

0

10 10

2

10

4

10

Fluorescence signal (arbitrary unit) No addtion 10 mM IPTG 100 ng/ml aTc FIG 3 Effects of lgfp deletion genes on growth and gene expression. (A) Effects of lgfp deletion mutant constructs on E. coli growth. Cultures of E. coli carrying pLGFP (circles), pLGFP⌬1 (triangles), pLGFP⌬2 (squares), or pLGFP⌬3 (lozenges) were incubated at 37°C. IPTG was added to one of two cultures at a concentration of 10 mM (solid symbols). No IPTG was added to controls (open symbols). Error bars show standard deviations (n ⫽ 3). (B) Schematic representation of the gene repression system of pL⌬NG. pH⌬NG was also constructed as a control. (C) Expression of GFP from pL⌬NG and pH⌬NG. GFP expression was induced with aTc, and LGFP⌬1 or HGFP⌬1 expression was induced with 10 mM IPTG. There was no fluorescence in LGFP⌬1 or HGFP⌬1. Mean fluorescent signal values (arbitrary units) are indicated in the graphs.

T7 can lyse E. coli expressing hgfp and showed expansion by factors of 104 and 106, respectively, and GFP expression allowed phage expansion. In addition, the titers of phages ␭, f1, and MS2 failed to increase in cultures of E. coli overexpressing the lgfp gene (Fig. 4B); the titers of f1 and MS2 decreased especially drastically, to about 2 to 3% of the initial values after 4 h of incubation. These two phages lack the nucleases that highly virulent phages possess to cleave and/or use the host’s DNA and cannot attack plasmids in the cell (25). The halting of peptide synthesis by overexpression of lgfp



aem.asm.org

2549


CFU (/ml)

9 10

) pLGFP (+IPTG) ) pLGFP∆1 (+IPTG) ) pLGFP∆2 (+IPTG) ) pLGFP∆3 (+IPTG)

would be effective against virus protein synthesis and stop the expansion of these two phages. Construction of phage-resistant E. coli with a phage-inducible promoter and lgfp. To utilize my artificial GFP gene as a practical antivirus system, I focused on the promoter of the phage shock protein (psp) operon, which is induced by f1 infection in E. coli, and constructed pPAH and pPAL to induce hgfp and lgfp expression on demand, respectively (Fig. 4C) (26). Artificial gfp genes were inserted downstream of the psp operon promoter with a strong ribosome binding site (AGGAGGTTTTTT) (2). Infection with f1 induced GFP expression from the hgfp gene at a level that was one-sixth of that from hgfp induced with 10 mM IPTG at 2 h (Fig. 4D). The mean value of GFP expression from the f1 phageinducible hgfp gene increased to 50.8 (arbitrary units) at 3 h of incubation. GFP expression was also observed in cells transformed with f1 phage-inducible lgfp upon phage f1 infection, although its level was very low. After infection, the growth of E. coli cultures carrying pPAL was similar to that of uninfected controls (Fig. 4E). In addition, the reproductive rate of f1 in E. coli cultures carrying f1 phage-inducible lgfp decreased to 10⫺4 to 10⫺6 of that in cultures carrying f1 phage-inducible hgfp. Suppression of cell growth because of lgfp overexpression was not observed in this system because of the low level of GFP expression, but it was sufficient to interrupt f1 expansion. These results suggest that an artificial gene rich in low-usage codons and a suitable gene expression system could provide a novel genetic system for virus defense in the cell. My system can interrupt the expansion of unknown or mutated viruses that use the host’s protein synthesis systems. Effect of lgfp and hgfp expression on S. cerevisiae. To examine whether my system is effective in eukaryotes, I examined the effects of lgfp and hgfp on the growth of S. cerevisiae, a model eukaryotic microorganism for which a number of host-vector systems are available (19). In S. cerevisiae, lgfp and hgfp showed similar CAIs of 0.594 and 0.621, respectively. However, hgfp had more rare-codon types than lgfp in this organism, which was the opposite of the case in E. coli (Table 1); the lgfp gene has only two types of rare codon, 23 glycine (GGG) and 18 leucine (CUA), the fractions of which were ⬍0.15, while hgfp has five types of rare codon. Six rare arginine codons (CGC) are present from the middle to the end of the hgfp gene (Fig. 5A, arrows). Auxilien et al. suggested that tRNAArgCGC is a product of enzymatic modification of the AGC codon tRNA (31). I constructed pYLGFP1 and pYLGFP2 containing lgfp and hgfp, respectively, arranged downstream of the galactose-inducible promoter PGAL of S. cerevisiae, and an aureobasidin A resistance gene. Control plasmid pYEG had a yeast enhanced GFP gene (yEGFP) instead of the artificial gfp gene in the same vector (32). To investigate the effects of the expression of each artificial gfp, S. cerevisiae carrying each plasmid was cultured in YPGalactose medium containing aureobasidin A (1 ␮g/ml). S. cerevisiae growth repression was observed only with induction of hgfp expression for 24 h, which was longer than that in E. coli induced lgfp expression (Fig. 5B). The copy number of the yeast plasmid in the cell was about 10 (19), which is equivalent to the medium-copy-number plasmid containing the p15A origin of replication in E. coli. There was no growth repression associated with induction of lgfp expressed from the medium-copy-number plasmid in E. coli (see Fig. S1 in the supplemental material). This difference in growth repression between S. cerevisiae and E. coli will depend on the half-life of the mRNA; the average half-life of mRNA is a few minutes in E. coli compared with 22 min in yeast

LGFP+T4 LGFP+T7 LGFP+IPTG+T4 LGFP+IPTG+T7

10 10 10 10 10

9 8 7 6 5

LGFP+IPTG

pHGFP

8

10

pHGFP+IPTG pLGFP

7

10

pLGFP+IPTG

6

10

5

10

4

10

3

10

2

10

1

10 0.1

0

10

λ

-1

10

f1

MS2

-2

10 0.01 2.0

3.0

4.0

Time (h)

E phage expansion

phage protein synthesis

f1 phage infection Ppsp

artificial RBS

lgfp (or hgfp)

pPAL (or pPAH) Amp r

ColE1

pPAH Counts

256

256

2.8

0 0 10

10

2

Reproductive rate (-)

1.0

10

4

0 0 10

30.1

10

2

4

10

Fluorescence signal (arbitrary unit) no phage phage f1

Survival cells (%)

0

10

8

10

6

10

4

10

2

10

0

pPAH pPAL

120 100 80 60 40 pPAH pPAL

20 0 10

-2

10

-1

10

0

10

1

10

2

10

3

FIG 4 Phage resistance caused by induction of the lgfp gene. (A) Effects of lgfp induction on T4 and T7 phage infections. Cultures of E. coli carrying pHGFP (blue) or pLGFP (red) were incubated at 37°C for 30 min with 10 mM IPTG (solid symbols). No IPTG was added to controls (open symbols). The OD660 of each culture was adjusted to about 0.4, and phage T4 (circles) or T7 (triangles) was added. No phage was added to controls (squares). The multiplicities of infection of T4 and T7 were 0.088 and 0.0012, respectively. E. coli carrying pHGFP (⫾IPTG) or pLGFP (⫺IPTG) was grown to stationary phase at 2 h. The PFU count (upper graph) and OD660 (lower graph) of each culture were measured periodically. Error bars show standard deviations (n ⫽ 3). (B) Effects of lgfp expression on the rates of expansion of various phages. E. coli JM2.300 carrying pHGFP or pLGFP was used as a host for phage ␭, and E. coli AK4 carrying pHGFP or pLGFP was used as a host for phages f1 and MS2. Culture conditions and IPTG induction were as described for panel A. Phages were added to the cultures at a multiplicity of infection of 0.01. The rate of expansion of each phage was calculated from the phage titer after 4 h of incubation at 37°C. Error bars show standard deviations (n ⫽ 3). (C) Construction of a phage response genetic system. pPAL contains the lgfp gene downstream of the modified promoter of the psp operon, the ampicillin resistance gene, and ColE1. pPAH has hgfp instead of lgfp of pPAL. RBS, ribosome binding site. (D) GFP expression from pPAH by phage f1 infection. E. coli AK4 carrying pPAH was incubated with phage f1 (multiplicity of infection, 1) at 37°C for 2 h. No phage was added to controls. GFP expression was then measured as described in Materials and Methods. Mean fluorescent signal values (arbitrary units) are indicated in the graphs. (E) Phage resistance of E. coli AK4 carrying pPAL. Cultures of E. coli AK4 carrying pPAH (blue) or pPAL (red) were incubated with various phage concentrations at 37°C for 4 h. Growth is indicated as the ratio of the OD660 to that of a culture of AK4 without the addition of phage f1 (circles, lower graph). All cultures without phage addition were grown to an OD660 of 0.65 to 0.90. The titers of phage f1 in the cultures were measured as described in Materials and Methods. The reproductive rates were calculated from the titers (squares, upper graph). Error bars show standard deviations (n ⫽ 3).

2550

aem.asm.org




1.0

OD 660

9

10

10

Reproductive rate (-)

10

11

pfu (/ml)

10

HGFP+T4 HGFP+T7 HGFP+IPTG+T4 HGFP+IPTG+T7

Fraction (-)


1

hgfp lgfp

0.8 0.6 0.4 0.2 0 51

101

151

201

Number of codons

OD660

4.0

3.0

2.0

1.0

0.0

0 h 24 h

pYEG

0h

24 h

0h

24 h

pYLGFP1 pYLGFP2

Counts

200 58.8 0 100

76.3

10.8

0 2 4 4 2 2 4 10 10 10 100 10 10 10 10 Fluorescence signal (arbitrary unit)

pYEG

pYLGFP1 pYLGFP2

FIG 5 Effects of rare-codon genes on S. cerevisiae. (A) Codon usage fractions of hgfp (blue) and lgfp (red) in S. cerevisiae. Codon usage fractions are based on database information from the website cited in the legend to Fig. 2. Arrows indicate arginine (CGC) codons. (B) Growth of S. cerevisiae carrying pYEG, pYLGFP1, or pYLGFP2. Plasmid-carrying S. cerevisiae was grown in YPGalactose plus 1% raffinose to induce GFP expression at 30°C. The OD660s of all of the transformants were ⬎5.0 in YPD medium at 24 h. The error bars show standard deviations (n ⫽ 3). (C) GFP expression from pYEG, pYLGFP1, and pYLGFP2 at 24 h. Mean fluorescent signal values (arbitrary units) are indicated in the graph. GFP was measured as described in Materials and Methods.

(17), and the content of rare-codon mRNA from the artificial gfp gene in the cell will be maintained at high levels in yeast despite the low copy number of the yeast plasmid. The level of expression of GFP from hgfp was much lower than that from lgfp or yEGFP (Fig. 5C). It was reported previously that the translational efficiency correlated with the CAI (33, 34). The lgfp gene showed the same GFP expression as yEGFP, although the CAI of lgfp was lower (Fig. 5C). The translational efficiency and growth repression by the artificial GFP gene would be correlated with the number of types of rare codon rather than the CAI. This result indicates that growth repression caused by monopolization of tRNAs for lowusage codons was effective in both prokaryotes and eukaryotes. Repression of human cell growth and virus expansion by hulgfp expression. Finally, I synthesized a rare-codon gfp gene (hulgfp, CAI ⫽ 0.485) according to human codon usage (Table 3) and examined whether its expression suppresses the growth of human


TABLE 3 The codons and their fractions in lgfp and hu-lgfp lgfp a

hu-lgfp b

AA

Codon

Fraction

No.

Codon

Fractionb

No.

Gly Lys Leu Asp Val Glu Thr Ile Asn Phe Tyr Pro Ser His Ala Gln Met Arg Cys Trp Stop

GGG AAG CUA GAC GUA GAG ACA AUA AAU UUC UAC CCC UCA CAC GCU CAA AUG AGG UGU UGG UAG

0.15 0.24 0.04 0.31 0.15 0.31 0.13 0.07 0.45 0.43 0.43 0.12 012 0.43 0.16 0.35 1 0.02 0.44 1 0.05

23 21 18 18 16 15 15 13 13 12 12 11 9 9 9 8 7 6 2 1 1

GGU AAA CUA GAU GUA GAA ACA AUA AAU UUU UAU CCG UCG CAU GCG CAA AUG CGU UGU UGG UAG

0.16 0.43 0.07 0.46 0.12 0.42 0.11 0.17 0.47 0.46 0.44 0.11 0.05 0.42 0.11 0.27 1 0.08 0.46 1 0.24

22 18 17 17 14 15 17 15 15 13 12 10 11 9 11 7 7 6 2 1 1

a

AA, amino acid. Fraction based on E. coli W3110 and H. sapiens as described in Materials and Methods. b


aem.asm.org

2551


1

cells. I cloned hu-lgfp downstream of the Dox-inducible promoter of the Tet-ON system and constructed plasmid pTRE-G1 (35). First, I tried making a stable mutant cell line by integrating Doxinducible hu-lgfp into the HeLa cell genome; however, there was no stable mutant in ⬎300 recombinant colonies. I then examined the effect of transient hu-lgfp expression on the growth of human cells. I selected three kinds of human cells, human HeLa cervical cancer cells, MCF7 breast cancer cells, and HEK293 kidney cells, for high transfection efficiency. HEK293 cells are used as host cells for recombinant adenovirus amplification. The pTRE-Luc plasmid was used as a control plasmid carrying a luciferase gene instead of hu-lgfp. Microscopic observation of HeLa cell culture showed induction of hu-lgfp suppressed cell growth in 24 h, whereas a control cell culture became confluent. The number of cells in the culture clearly decreased in 48 h (Fig. 6A). Genome degradation caused by apoptosis did not occur in HeLa cells expressing hu-lgfp. Microscopic observation showed no fluorescence in HeLa cells transfected with pTRE-G1. I then confirmed mRNA of hu-lgfp in cells transfected with pTRE-G1 after 48 h of incubation by qRT-PCR. The amount of mRNA was about 50 times as much as that of actB mRNA, which was used as a reference. The induction of hu-lgfp also inhibited the growth of MCF7 and HEK293 cells. The numbers of HeLa, MCF7, and HEK293 cells decreased to 6, 23.5, and 16.7% of the control by 72 h of incubation after transfection, respectively (Fig. 6B). Next, I also examined the effect of hu-lgfp induction on the expansion of recombinant adenovirus (22). HEK293 cells expressing hu-lgfp prevented the reproduction of recombinant adenovirus to about six times the initial concentration and kept growing, whereas 80% of the control cells were killed by adenovirus and the number of adenovirus particles increased to about 80 times the initial concentration (Fig. 6C). These results show that the rare-codon gfp

ACKNOWLEDGMENTS FIG 6 Effect of hu-lgfp induction on the growth of human cells and adenovi-

rus infection. (A) Microscopic observation (magnification, ⫻40) of the growth of HeLa cells transfected with pTRE-Luc or pTRE-G1. HeLa cells transfected with pTRE-Luc were used as a control. (B) Effect of pTRE-G1 on the growth of HeLa, MCF7, and HEK293 cells. All cells were incubated at 37°C for 72 h. The ratio of the growth of cells transfected with pTRE-G1 to the growth of cells transfected with TRE-Luc was calculated. Error bars show standard deviations (n ⫽ 6). (C) Effect of pTRE-G1 on recombinant adenovirus infection. HEK293 cells grown in DMEM for 48 h after transfection were used as hosts. Recombinant adenovirus (1.6 ⫻ 105) was added to each culture well. Cells and recombinant adenovirus particles were counted after 72 h of cultivation at 37°C. Error bars show standard deviations (n ⫽ 6 for multiplication data; n ⫽ 3 for cell growth data).

gene works as an anticancer gene and a virus repressor in human cells. DISCUSSION

In the present study, I tested a novel strategy for controlling a biological system in both prokaryotes and eukaryotes. My system is completely artificial and does not involve natural gene networks, and it is therefore very difficult to escape control through

2552

aem.asm.org

I thank J. J. Collins for kindly providing plasmids and T. Gardner and M. Kaern for valuable discussion. I designed the experiments, performed all of the experiments, and wrote the paper. I have no conflict of interest to declare.

REFERENCES 1. Elowitz MB, Leibler S. 2000. A synthetic oscillatory network of transcriptional regulators. Nature 403:335–338. http://dx.doi.org/10.1038/35002125. 2. Gardner TS, Cantor CR, Collins JJ. 2000. Construction of a genetic toggle switch in Escherichia coli. Nature 403:339 –342. http://dx.doi.org/10 .1038/35002131. 3. Hasty J, McMIllen D, Collins JJ. 2002. Engineered gene circuits. Nature 420:224 –230. http://dx.doi.org/10.1038/nature01257. 4. Kobayashi H, Kærn M, Araki M, Chung K, Gardner TS, Cantor CR, Collins JJ. 2004. Programmable cells: interfacing natural and engineered gene networks. Proc Natl Acad Sci U S A 101:8414 – 8419. http://dx.doi .org/10.1073/pnas.0402940101. 5. Isaacs FJ, Dwyer DJ, Ding C, Pervouchine DD, Cantor CR, Collins JJ. 2004. Engineered riboregulators enable post-transcriptional control of gene expression. Nat Biotechnol 22:841– 847. http://dx.doi.org/10.1038 /nbt986. 6. Stricker J, Cookson S, Matthew R, Bennett MR, Mather WH, Tsimring LS, Hasty J. 2008. A fast, robust and tunable synthetic gene oscillator. Nature 456:516 –519. http://dx.doi.org/10.1038/nature07389.




natural gene regulation, enzyme reaction, or signal transduction in the cell. My system showed stronger growth repression than a gene containing a single kind of many rare codons (12). In the case of a single rare codon, genes not containing the critical rare codon would be expressed as usual. The growth repression achieved would then be weaker than that of my system. The lgfp deletion mutant constructs did not show the specific codons needed for growth suppression. The most-reduced rare codon in lgfp⌬1 and lgfp⌬2 was AUA (Ile). On the other hand, glycine is the most common amino acid in E. coli, and the lowest-usage glycine codon, GGG, was the most common one in lgfp (36). A shortage of the major amino acid may be needed for the suppression of cell growth. A quantitative analysis of low-usage codons in additional lgfp mutant constructs would reveal the specific codon or total rare codons needed for the silencing of gene expression or cell growth. My system is easily applicable to artificial gene networks in synthetic biology because almost all of the artificial gene networks in synthetic biology usually use the gfp gene or relative fluorescent proteins as reporters (1–6, 37, 38). Furthermore, specific gene deletion mutant constructs, genetic backgrounds, and compensating genes are needed for the use of artificial gene networks to program cell behavior (4, 15, 39–41). My artificial gfp genes would provide cell growth regulation in artificial gene networks for applications in biotechnology without any mutagenesis of host cells (42–45). In future studies, I will also construct an artificial virus defense system by using my artificial gene and promoters activated by virus infection in mammalian cells (25). Linkage of artificial and natural gene control systems will be useful in the construction of effective and safe systematic gene therapy protocols and transgenic organisms (43). For example, many genes or RNAs have been tested for gene therapy for HIV (46, 47). The gene transfer and expression systems are ready in CD4⫹ T cells (48, 49). My artificial gene system for growth suppression and gene silencing would be able to suppress the expansion of HIV similarly to that of adenovirus (Fig. 6C). Determination of the kind of gene, quantitative analysis for length, and codons of the artificial gene for suppression of HIV would be needed for effective gene therapy.



31. Auxilien S, Crain PF, Trewyn RW, Grosjean HH. 1996. Mechanism, specificity and general properties of the yeast enzyme catalysing the formation of inosine 34 in the anticodon of transfer RNA. J Mol Biol 262: 437– 458. http://dx.doi.org/10.1006/jmbi.1996.0527. 32. Cormack BP, Bertram G, Egerton M, Gow NA, Falkow S, Brown AJ. 1997. Yeast-enhanced green fluorescent protein (yEGFP) a reporter of gene expression in Candida albicans. Microbiology 143:303–311. http://dx .doi.org/10.1099/00221287-143-2-303. 33. Blake WJ, Kaern M, Cantor CR, Collins JJ. 2003. Noise in eukaryotic gene expression. Nature 422:633– 637. http://dx.doi.org/10.1038/nature01546. 34. Kellis M, Birren BW, Lander EC. 2004. Proof and evolutionary analysis of ancient genome duplication in the yeast Saccharomyces cerevisiae. Nature 428:617– 624. http://dx.doi.org/10.1038/nature02424. 35. Urlinger S, Baron U, Thellmann M, Hasan MT, Bujard H, Hillen W. 2000. Exploring the sequence space for tetracycline-dependent transcriptional activators: novel mutations yield expanded range and sensitivity. Proc Natl Acad Sci U S A 97:7963–7968. http://dx.doi.org/10.1073/pnas .130192197. 36. Csonka LN, Epstein W. 1996. Chemical composition of Escherichia coli, p 13–15. In Neidhardt FC, Curtiss R, Ingraham JL, Lin ECC, Low KB, Magasanik B, Reznikoff WS, Riley M, Schaechter M, Umbarger H (ed), Escherichia coli and Salmonella: cellular and molecular biology. ASM Press, Washington, DC. 37. Tigges M, Marquez-Lago TT, Stelling J, Fussenegger M. 2009. A tunable synthetic mammalian oscillator. Nature 457:309 –312. http://dx.doi.org /10.1038/nature07616. 38. Nandagopal N, Elowitz MB. 2011. Synthetic biology: integrated gene circuits. Science 333:1244 –1248. http://dx.doi.org/10.1126/science.1207084. 39. Fung E, Wong WW, Suen JK, Bulter T, Lee S, Liao JC. 2005. A synthetic gene-metabolic oscillator. Nature 235:118 –122. http://dx .doi.org/10.1038/nature03508. 40. Friedland AE, Lu TK, Wang X, Shi D, Church G, Collins JJ. 2009. Synthetic gene networks that count. Science 324:1199 –1202. http://dx.doi .org/10.1126/science.1172005. 41. Cardinale S, Joachimiak MP, Arkin AP. 2013. Effects of genetic variation on the E. coli host-circuit interface. Cell Rep 4:231–237. http://dx.doi.org /10.1016/j.celrep.2013.06.023. 42. Khalil AS, Collins JJ. 2010. Synthetic biology: applications come of age. Nat Rev Genet 11:367–379. http://dx.doi.org/10.1038/nrg2775. 43. Ruder WC, Lu T, Collins JJ. 2011. Synthetic biology moving into the clinic. Science 333:1248 –1252. http://dx.doi.org/10.1126/science.1206843. 44. Cameron DE, Collins JJ. 2014. Tunable protein degradation in bacteria. Nat Biotechnol 32:1276 –1281. http://dx.doi.org/10.1038/nbt.3053. 45. Keung AJ, Bashor CJ, Kiriakov S, Collins JJ, Khalil AS. 2014. Using targeted chromatin regulators to engineer combinatorial and spatial transcriptional regulation. Cell 158:110 –120. http://dx.doi.org/10.1016/j.cell .2014.04.047. 46. Rossi JJ, June CH, Kohn DB. 2007. Genetic therapies against HIV. Nat Biotechnol 25:1444 –1454. http://dx.doi.org/10.1038/nbt1367. 47. Hoxie JA, June CH. 2012. Novel cell and gene therapies for HIV. Cold Spring Harb Perspect Med 2:a007179 http://dx.doi.org/10.1101/cshperspect.a007179. 48. Braun SE, Taube R, Zhu Q, Wong FE, Murakami A, Kamau E, Dwyer M, Qiu G, Daigle J, Carville A, Johnson RP, Marasco WA. 2012. In vivo selection of CD4(⫹) T cells transduced with a gamma-retroviral vector expressing a single-chain intrabody targeting HIV-1 tat. Hum Gene Ther 23:917–931. http://dx.doi.org/10.1089/hum.2011.184. 49. Lisziewicz J, Sun D, Smythe J, Lusso P, Lori F, Louie A, Markham P, Rossi J, Reitz M, Gallo RC. 1993. Inhibition of human immunodeficiency virus type 1 replication by regulated expression of a polymeric Tat activation response RNA decoy as a strategy for gene therapy in AIDS. Proc Natl Acad Sci U S A 90:8000 – 8004. http://dx.doi.org/10.1073/pnas .90.17.8000. 50. Wilson IB, Dayan J, Cyr K. 1964. Some properties of alkaline phosphatase from Escherichia coli. Transphosphorylation. J Biol Chem 239:4182– 4185. 51. Shen BH, Boos W. 1973. Regulation of the ␤-methylgalactoside transport system and the galactose-binding protein by the cell cycle of Escherichia coli. Proc Natl Acad Sci U S A 70:1481–1485. http://dx.doi.org/10.1073 /pnas.70.5.1481.


aem.asm.org

2553


7. Deans TL, Cantor CR, Collins JJ. 2007. A tunable genetic switch based on RNAi and repressor proteins for regulating gene expression in mammalian cells. Cell 130:363–372. http://dx.doi.org/10.1016/j.cell.2007.05.045. 8. Ikemura T. 1981. Correlation between the abundance of Escherichia coli transfer RNAs and the occurrence of the respective codons in its protein genes: a proposal for a synonymous codon choice that is optimal for the E. coli translational system. J Mol Biol 151:389 – 409. http://dx.doi.org/10 .1016/0022-2836(81)90003-6. 9. Ikemura T. 1985. Codon usage and tRNA content in unicellular and multicellular organisms. Mol Biol Evol 2:13–34. 10. Yamao F, Andachi Y, Muto A, Ikemura T, Osawa S. 1991. Levels of tRNAs in bacterial cells as affected by amino acid usage in proteins. Nucleic Acids Res 19:6119 – 6122. http://dx.doi.org/10.1093/nar/19.22.6119. 11. Calderone TL, Stevens RD, Oas TG. 1996. High-level misincorporation of lysine for arginine at AGA codons in a fusion protein expressed in Escherichia coli. J Mol Biol 262:407– 412. http://dx.doi.org/10.1006/jmbi .1996.0524. 12. Zahn K. 1996. Overexpression of an mRNA dependent on rare codons inhibits protein synthesis and cell growth. J Bacteriol 178:2926 –2933. 13. Bosworth S, Kabay ME. 2002. Computer security handbook, 4th ed. John Wiley & Sons, London, United Kingdom. 14. Bishop M. 2003. Computer security: art and science. Addison-Wesley, Boston, MA. 15. Nakamura Y, Gojobori T, Ikemura T. 2000. Codon usage tabulated from international DNA sequence databases: status for the year 2000. Nucleic Acids Res 28:292. http://dx.doi.org/10.1093/nar/28.1.292. 16. Carbone A, Zinovyev A, Kepes F. 2003. Codon adaptation index as a measure of dominating codon bias. Bioinformatics 19:2005–2015. http: //dx.doi.org/10.1093/bioinformatics/btg272. 17. Lodish H, Berk A, Zipursky SL, Matsudaira P, Baltimore D, Darnell J. 2004. Molecular cell biology. W. H. Freeman & Co., New York, NY. 18. Sambrook J, Fritsch E, Maniatis T. 1989. Molecular cloning: a laboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 19. Sikorski RS, Herter PA. 1989. System of shuttle vectors and yeast host strains designed for efficient manipulation of DNA in Saccharomyces cerevisiae. Genetics 122:19 –27. 20. Laemmli UK. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680 – 685. http://dx.doi.org/10 .1038/227680a0. 21. Coleman JE. 1992. Structure and mechanism of alkaline phosphatase. Annu Rev Biophys Biomol Struct 21:441– 483. http://dx.doi.org/10.1146 /annurev.bb.21.060192.002301. 22. Graham FL, Smiley J, Russell WC, Nairn R. 1977. Characteristics of a human cell line transformed by DNA from human adenovirus type 5. J Gen Virol 36:59 –74. http://dx.doi.org/10.1099/0022-1317-36-1-59. 23. Cormack BP, Valdivia RH, Falkow S. 1996. FACS-optimized mutants of the green fluorescent protein (GFP). Gene 173:33–38. http://dx.doi.org /10.1016/0378-1119(95)00685-0. 24. Csonka LN, Epstein W. 1996. Osmoregulation, p 1210 –1220. In Neidhardt FC, Curtiss R, Ingraham JL, Lin ECC, Low KB, Magasanik B, Reznikoff WS, Riley M, Schaechter M, Umbarger H (ed), Escherichia coli and Salmonella: cellular and molecular biology. ASM Press, Washington, DC. 25. Flint SJ, Enquist LW, Racaniello VR, Skalka AM (ed). 2009. Principles of virology, 3rd ed, vol 2, p 52–163. ASM Press, Washington, DC. 26. Model P, Jovanovic G, Dworkin J. 1997. The Escherichia coli phageshock-protein (psp) operon. Mol Microbiol 24:255–261. http://dx.doi.org /10.1046/j.1365-2958.1997.3481712.x. 27. Gursky YG, Beabealashvilli SH. 1994. The increase in gene expression induced by introduction of rare codons into the C terminus of the template. Gene 148:15–21. http://dx.doi.org/10.1016/0378-1119(94)90228-3. 28. Magnusson LU, Farewell A, Nystrom T. 2005. ppGpp: a global regulator in Escherichia coli. Trends Microbiol 13:236 –242. http://dx.doi.org/10 .1016/j.tim.2005.03.008. 29. Brinkmann U, Mattes RE, Buckel P. 1989. High-level expression of recombinant genes in Escherichia coli is dependent on the availability of the dnaY gene product. Gene 85:109 –114. http://dx.doi.org/10.1016/0378 -1119(89)90470-8. 30. Birge EA. 2000. Bacterial and bacteriophage genetics, 4th ed. Springer Verlag, New York, NY.

Regulation of translation by upstream translation initiation codons of surfactant protein A1 splice variants.

Translation of 'Systematization of Codons in the Genetic Code [I]' by Yu. B. Rumer (1966).

Ribosome rescue and translation termination at non-standard stop codons by ICT1 in mammalian mitochondria.

Accurate Recovery of Ribosome Positions Reveals Slow Translation of Wobble-Pairing Codons in Yeast.

Rare codons capacitate Kras-driven de novo tumorigenesis.

Out-of-frame start codons prevent translation of truncated nucleo-cytosolic cathepsin L in vivo.

Emergent rules for codon choice elucidated by editing rare arginine codons in Escherichia coli.

Swedish translation and validation of a web-based questionnaire for registration of overuse problems.

Measurements of translation initiation from all 64 codons in E. coli.

Codon-reading specificities of mitochondrial release factors and translation termination at non-standard stop codons.

Alternative Translation Initiation of a Haloarchaeal Serine Protease Transcript Containing Two In-Frame Start Codons.

Translation Initiation from Conserved Non-AUG Codons Provides Additional Layers of Regulation and Coding Capacity.

Ribosome profiling: new views of translation, from single codons to genome scale.

Alternative translation initiation codons for the plastid maturase MatK: unraveling the pseudogene misconception in the Orchidaceae.

tRNAHis pair.

Structural basis of suppression of host translation termination by Moloney Murine Leukemia Virus.

Global mRNA selection mechanisms for translation initiation.

Reversible suppression of cyclooxygenase 2 (COX-2) expression in vivo by inducible RNA interference.

Suppression of inducible nitric oxide synthase pathway by 7-deacetylgedunin, a limonoid from Xylocarpus sp.

Reduction of gastric cancer proliferation and invasion by miR-15a mediated suppression of Bmi-1 translation.

Absolute in vivo translation rates of individual codons in Escherichia coli. The two glutamic acid codons GAA and GAG are translated with a threefold difference in rate.

The suppression of defective translation by ppGpp and its role in the stringent response.

Adjacent Codons Act in Concert to Modulate Translation Efficiency in Yeast.

Global analysis of translation termination in E. coli.