Gene silencing in Escherichia coli using antisense RNAs expressed from doxycycline-inducible vectors


  • N. Nakashima,

    Corresponding author
    1. Biomass Refinery Research Center, National Institute of Advanced Industrial Sciences and Technology (AIST), Sapporo, Japan
    • Bioproduction Research Institute, National Institute of Advanced Industrial Sciences and Technology (AIST), Sapporo, Japan
    Search for more papers by this author
  • T. Tamura

    1. Bioproduction Research Institute, National Institute of Advanced Industrial Sciences and Technology (AIST), Sapporo, Japan
    2. Laboratory of Molecular Environmental Microbiology, Graduate School of Agriculture, Hokkaido University, Sapporo, Japan
    Search for more papers by this author


Nobutaka Nakashima, Bioproduction Research Institute, National Institute of Advanced Industrial Sciences and Technology (AIST), 2-17-2-1 Tsukisamu-Higashi, Toyohira-ku, Sapporo 062-8517, Japan.



Here, we report on the construction of doxycycline (tetracycline analogue)-inducible vectors that express antisense RNAs in Escherichia coli. Using these vectors, the expression of genes of interest can be silenced conditionally. The expression of antisense RNAs from the vectors was more tightly regulated than the previously constructed isopropyl-β-D-galactopyranoside-inducible vectors. Furthermore, expression levels of antisense RNAs were enhanced by combining the doxycycline-inducible promoter with the T7 promoter-T7 RNA polymerase system; the T7 RNA polymerase gene, under control of the doxycycline-inducible promoter, was integrated into the lacZ locus of the genome without leaving any antibiotic marker. These vectors are useful for investigating gene functions or altering cell phenotypes for biotechnological and industrial applications.

Significance and Impact of the Study

A gene silencing method using antisense RNAs in Escherichia coli is described, which facilitates the investigation of bacterial gene function. In particular, the method is suitable for comprehensive analyses or phenotypic analyses of genes essential for growth. Here, we describe expansion of vector variations for expressing antisense RNAs, allowing choice of a vector appropriate for the target genes or experimental purpose.


The genome sequences of many organisms have been completely determined, allowing comprehensive analysis of gene functions through reverse genetic approaches. Gene silencing (knock-down) is one such approach and is much easier and faster than the gene disruption (knock-out) approach. Some methodologies have been developed to silence bacterial genes, and of these, single-stranded antisense RNAs (asRNAs) generated from expression vectors have been frequently used (Yin and Ji 2002; Nakashima et al. 2012). Expressed asRNAs are typically designed to hybridize to the ribosome-binding site (RBS) and the start codon region of target mRNAs (Stefan et al. 2010). Expressed asRNAs are thought to prevent ribosomes from recognizing the RBS and, thus, inhibit translation. However, the precise silencing mechanism, including whether target mRNA levels decreased, is not clear (Pestka et al. 1984; Coleman et al. 1985; Stefan et al. 2010). Yet, in our previous studies, target mRNA level decreased in proportion to decreased protein level (Nakashima et al. 2006).

The expressed antisense approach had a drawback, in that silencing efficacy was low, depending on the targeted genes (Brantl 2002). To improve silencing efficacy, we previously constructed novel expression vectors for Escherichia coli that express asRNAs carrying paired termini (PTasRNAs) (Nakashima et al. 2006; Nakashima and Tamura 2009) (Fig. 1a, pHN1257). The PTasRNAs have flanking 38-bp inverted repeats that create paired double-stranded RNA termini (Fig. 1b). We found that the PTasRNAs had much higher silencing efficacies than ordinary asRNAs, due to their improved RNA stability, which increases the abundance of these RNAs in cells.

Figure 1.

Maps of vectors and structure of the expressed PTasRNA. (a) Arrows indicate open reading frames, promoters or terminators, and boxes indicate the pSC101H replication origin (ori). kanr, kanamycin-resistance gene; tetR, tetracycline repressor gene; Phom, constitutive hom promoter; Ptet, doxycycline-inducible promoter; Pt7, T7 promoter; TrrnB, rrnB terminator; PT5′ and PT3′, 5′- and 3′-regions of paired-termini structure, respectively. Between the NcoI and XhoI sites, unique EcoRI, SnaBI, NotI, SacII, BamHI, HindIII and BglII sites are located. (b) Structure of a PTasRNA is shown. The curved line indicates the antisense sequence (or a multiple cloning site sequence in the case of an empty vector). (c) The vector for integrating the Ptet-T7RNAP and Phom-tetR gene cassettes into the lacZ locus is shown. The region that remains on the lacZ locus after the counter selection step is shown in grey.

The vector pHN1257 harbours the trc promoter (Ptrc) and the lactose repressor gene (lacIq) to drive conditional expression of PTasRNAs in the presence of isopropyl-β-D-galactopyranoside (IPTG) (Anthony et al. 2004). Ptrc was chosen because it is a very strong promoter and is insensitive to carbon catabolite repression. However, basal level expression from Ptrc in the absence of IPTG was always observed (Anthony et al. 2004) and was problematic when strict regulation of gene silencing was required (Nakashima and Tamura 2009).

To overcome this disadvantage, we have constructed PTasRNA expression vectors that harbour the doxycycline/tetracycline-inducible promoter (Ptet) (Lutz and Bujard 1997). Expression from Ptet is known to be tightly regulated, and as expected, tight regulation of PTasRNA expression was observed. Furthermore, doxycycline-inducible expression of PTasRNAs was enhanced by combining Ptet with the T7 promoter (Pt7)-T7 RNA polymerase (T7RNAP) (Studier and Moffatt 1986). Because not all vector constructs for expressing antisense RNAs are effective, the expansion of vector varieties shown here is useful.

Results and discussion

Construction of doxycycline-inducible PTasRNA expression vectors

To achieve more tightly regulated silencing compared to the Ptrc-PTasRNA vector (Fig. 1a, pHN1257), a Ptet-PTasRNA vector was constructed (Fig. 1a, pHN1869). Note that pSC101H is a modified version of pSC101, such that the plasmid copy number increases approximately fourfold (Peterson and Phillips 2008). This vector harbours tetR, which encodes a doxycycline-dependent transcriptional repressor for Ptet, and transcription of tetR is driven by the constitutive promoter, Phom (Pátek et al. 2003). Moreover, a set of vectors was constructed (Fig. 1a,c); the Pt7-PTasRNA vector (pHN1868) expresses PTasRNAs from Pt7, and pHN1882 expresses T7RNAP from Ptet and is designed to be integrated into the lacZ locus of the genome. pHN1882 harbours a temperature-sensitive replication origin (pSC101ts), chloramphenicol-resistance gene (chlr), sacB for counter selection after genome integration and the Ptet-T7RNAP and Phom-tetR gene cassettes between the 5′- and 3′-flanking regions of lacZ. At high temperatures, the whole vector is integrated into the lacZ locus via homologous recombination and, except for the Ptet-T7RNAP and Phom-tetR gene cassettes, is removed from the genome on sucrose-containing media, because the sacB gene product is toxic to cells in the presence of sucrose. The resulting strain harbouring the Ptet-T7RNAP and Phom-tetR gene cassettes within the genome was termed MG1655-T7P. When Pt7-PTasRNA vectors are introduced into the MG1655-T7P strain, PTasRNAs are expressed upon addition of doxycycline. Expression from the Pt7 is generally very high and requires a faint amount of T7RNAP.

We further constructed vectors that are equivalent to pHN1869 (Ptet-PTasRNA vector) and pHN1868 (Pt7-PTasRNA vector), but which harbour different replication origins (pBR322-, pACYC- and RK2-replication origins) and antibiotic selection markers (ampicillin-, chloramphenicol- and apramycin-resistance; Fig. S1). These vectors, including pHN1869 and pHN1868, are co-transformable in any combination and can be used for multi-gene silencing. Note that expression from these vectors is not affected by glucose.

Silencing of pepN by the doxycycline-inducible PTasRNA expression vector

To evaluate the vectors, pepN, which encodes aminopeptidase N (Chandu and Nandi 2003), was chosen as a target for gene silencing. The pepN antisense sequence of 102 nucleotides (Fig. 2a) was chosen based on our previous finding that RBS-containing sequences of around 100 nucleotides are the best for silencing (Nakashima et al. 2006). The Ptrc-pepN PTasRNA and Ptet-pepN PTasRNA vectors (pHN1890 and pHN1880) were introduced into the MG1655 strain in parallel with the control empty vectors (pHN1257 and pHN1869), and the Pt7-pepN PTasRNA vector (pHN1878) was introduced into the MG1655-T7P strain in parallel with the control empty vector (pHN1868). After culturing these transformants up to the mid-logarithmic phase, PepN enzyme levels were measured.

Figure 2.

Silencing of pepN and measurement of PepN activity. (a) Schematic maps of the pepN mRNA and pepN antisense sequences are shown. The numerals indicate nucleotide number. The pepN antisense sequence covers the RBS and AUG start codon regions within pepN mRNA. (b) Vertical axes indicate relative PepN activity (%) in cells harbouring the Ptrc-pepN PTasRNA vector (pHN1890) compared to the empty Ptrc-PTasRNA vector (pHN1257) at the indicated IPTG concentrations. Horizontal axes are shown in log-scale. (c) Same as for panel b, but Ptet-pepN PTasRNA (pHN1880) and empty Ptet-PTasRNA (pHN1869) vectors and doxycycline were used. (d) Same as for panel c, but Pt7-pepN PTasRNA (pHN1878) and empty Pt7-PTasRNA (pHN1868) vectors were used.

As inducers for Ptet, tetracycline, doxycycline and anhydrotetracycline are known. Doxycycline is chosen here, because anhydrotetracycline is 3000 times as expensive as doxycycline and a manufacturer recommends using doxycycline (, accessed 14 Feb 2013).

The results indicated that, when the Ptrc-pepN PTasRNA vector was used, pepN silencing occurred even in the absence of IPTG (Fig. 2b). Meanwhile, silencing with Ptet was tightly regulated, as expected (Fig. 2c). Tight regulation of silencing was also confirmed when the Pt7-pepN PTasRNA vector was used (Fig. 2d); this was observed at low concentrations of doxycycline.

The amount of pepN mRNA remaining after silencing was quantitated by real-time RT-PCR (Fig. 3). When silenced with the Ptrc-pepN PTasRNA and Ptet-pepN PTasRNA vectors, the pepN mRNA amount decreased by 80 and 60%, respectively, but when silenced with the Pt7-pepN PTasRNA vector, mRNA was decreased by only 40%. The results indicated that the remaining protein levels do not always correlate with the mRNA levels.

Figure 3.

Quantification of pepN mRNA remaining after silencing. (a) Vertical axes indicate the amount of pepN mRNA (%) in cells harbouring the Ptrc-pepN PTasRNA vector (pHN1890), compared to cells harbouring the empty Ptrc-PTasRNA vector (pHN1257), at the indicated IPTG concentrations. Horizontal axes are shown in log-scale. (b) Same as for panel a, but Ptet-pepN PTasRNA (pHN1880) and empty Ptet-PTasRNA (pHN1869) vectors and doxycycline were used. (c) Same as for panel b, but Pt7-pepN PTasRNA (pHN1878) and empty Pt7-PTasRNA (pHN1868) vectors were used.

Quantification of the pepN PTasRNA

Expression levels of the pepN PTasRNA were determined in the same transformants as described above. When the expression level of PTasRNA from the Ptrc-pepN PTasRNA vector (pHN1890) in the presence of 1 mmol l−1 IPTG was set to 1, the relative PTasRNA expression levels from the Ptet-pepN PTasRNA vector (pHN1880) and Pt7-pepN PTasRNA vector (pHN1878) peaked at 0·37 and 1·4, respectively (Table 1). These results indicated that expression enhancement by the Pt7-T7RNAP system was successful. Leaky expression from the Ptrc-pepN PTasRNA vector (pHN1890) was implied by the results shown in Figs 2b and 3a, and this was confirmed by the results described above. Moreover, tightly regulated expression from the Ptet-pepN PTasRNA vector (pHN1880) and Pt7-pepN PTasRNA vector (pHN1878) was confirmed.

Table 1. Quantification of pepN PTasRNA levels
Host strainPlasmidInducer conc.Relative pepN PTasRNA amounta
  1. a

    The amount of pepN PTasRNA in MG1655 cells harbouring pHN1890 in the presence of 1 mmol l−1 IPTG was set to 1, and the relative amount in the different strains is shown. Data are given as mean ± SD from triplicate experiments.

MG1655pHN1890 (Ptrc-pepN PTasRNA)IPTG 1 mmol l−11
IPTG 0 mmol l−10·19 ± 0·04
pHN1880 (Ptet-pepN PTasRNA)Doxycycline 320 μg l−10·37 ± 0·04
Doxycycline 100 μg l−10·20 ± 0·04
Doxycycline 20 μg l−15·6 × 10−4 ± 7·5 × 10−5
Doxycycline 0 μg l−14·2 × 10−4 ± 4·5 × 10−6
MG1655-T7PpHN1878 (Pt7-pepN PTasRNA)Doxycycline 8 μg l−11·4 ± 0·33
Doxycycline 2 μg l−19·7 × 10−3 ± 1·5 × 10−3
Doxycycline 0·5 μg l−18·7 × 10−3 ± 5·1 × 10−4
Doxycycline 0 μg l−19·8 × 10−3 ± 1·9 × 10−3

The Pt7-PTasRNA vector (pHN1868) showed tightly regulated and high-level expression of PTasRNA and only required a small amount of doxycycline (Fig. 3). This feature is important, because doxycycline impairs growth of wild-type MG1655 cells at levels exceeding 300–500 μg l−1. However, when MG1655-T7P cells carrying the Pt7-PTasRNA vectors are subjected to more than 10–20 μg l−1 of doxycycline, cell growth was delayed. One possible explanation for this delay is that extremely high-level expression of PTasRNA leads to nonspecific cytotoxicity. Alternatively, high-level expression of T7RNAP may be toxic (Temme et al. 2012).

Interestingly, in the MG1655-T7P strain harbouring the Pt7-pepN PTasRNA vector (pHN1878), only a miniscule amount of pepN PTasRNA was detected after addition of 2 μg l−1 of doxycycline (Table 1), but significant silencing was observed at protein level (Fig. 2d). Furthermore, the maximum decrease in the level of pepN mRNA after silencing was lower for the Pt7-pepN PTasRNA vector (pHN1878) than for the Ptrc-pepN (pHN1890) and Ptet-pepN PTasRNA (pHN1880) vectors (Fig. 3). The reason for these results is unclear, but they possibly reflect differences in transcription speed between native RNAP and T7RNAP; transcription with T7RNAP is eightfold faster (Makarova et al. 1995). In the case of RNAII, which is a natural RNA species encoded by ColEl-type plasmids and which acts as a primer for plasmid replication, it is known that a proper secondary structure cannot form and that it is nonfunctional when transcribed by T7RNAP (Chao et al. 1995). Similarly, secondary structure, stability or mode of action of pepN PTasRNA may be different, depending on the polymerase used.

Silencing of accA, a gene essential for growth

The utility of the inducible silencing system is its applicability to a gene essential for growth. To demonstrate this utility, a growth essential gene, accA, which encodes the acetyl-CoA carboxylase carboxyltransferase α subunit, was silenced, and growth kinetics were monitored. Clear growth inhibition was observed when the Ptrc-accA PTasRNA vector was used, but no growth inhibition was observed with the Ptet-accA PTasRNA vector (Fig. 4); rather, growth inhibition at a high doxycycline concentration (note that MG1655 is sensitive to doxycycline) was observed. The Pt7-PTasRNA vector showed growth inhibition, but inhibition was less clear than with the Ptrc-PTasRNA vector (Fig. 4). This result indicated that using a strong promoter does not always lead to effective silencing.

Figure 4.

Silencing of a growth essential accA gene. (a) Transformants harbouring the empty Ptrc-PTasRNA (left) and Ptrc-accA PTasRNA (right) vectors were cultured in the absence (○) or presence (▵) of IPTG. (b) Transformants harbouring the empty Ptet-PTasRNA (left) and Ptet-accA PTasRNA (right) vectors were cultured in the presence of different concentrations of doxycycline; 0 μg l−1 (○), 100 μg l−1 (▵), 320 μg l−1 (×). (c) Transformants harbouring the empty Pt7-PTasRNA (left) and Pt7-accA PTasRNA (right) vectors were cultured similarly in the presence of different concentrations of doxycycline; 0 μg l−1 (○), 2 μg l−1 (▵), 12 μg l−1 (×).

The gene expression level required for normal growth differs from one essential gene to another (Goh et al. 2009), and, in the case of accA, low gene expression may be sufficient to support growth. It may be possible that accA PTasRNA expression from Ptet could be insufficient for inhibiting growth. The reason why Ptrc produced better results than Pt7 remains unclear, but differences in transcription speed may be involved, as described above.

These facts, when taken together, we conclude that it is important for effective silencing to choose appropriate promoter according to the target genes or experimental purposes. If a target gene is sensitive to antisense silencing like pepN, the Ptet or Pt7 should be used to achieve tightly regulated silencing, while if high-level PTasRNA expression is necessary, the Ptrc or Pt7 should be used. We previously constructed arabinose-inducible PTasRNA vectors (using the bad promoter; Pbad) in an attempt to generate tightly regulated PTasRNA vectors (Nakashima and Tamura 2009); however, these vectors could not be used in the presence of glucose due to the carbon catabolite repression.

Materials and methods

Bacterial strains, plasmids, and general techniques

Escherichia coli MG1655 (wild-type CGSG6300) (kindly provided by Dr. Masaaki Kanemori (Kanazawa University) strain was used as a host for expressing PTasRNAs. If not otherwise stated, E. coli cells were cultured in Luria Broth (LB) in the presence or absence of appropriate antibiotics at 37°C. General techniques are described elsewhere (Nakashima et al. 2006). Plasmid construction procedures and deduced plasmid sequences are described in Data S1, and plasmids used in this study are listed in Table S1. For measurements of growth kinetics, cells were pregrown overnight and diluted 1 : 25 000 with fresh medium and were then cultured in the Safire microplate reader (Tecan Group Ltd., Männedorf, Switzerland) at 37°C with agitation every 12 min in 200-μl volumes.

Construction of pepN and accA PTasRNA vectors

The pepN antisense sequence was PCR amplified from the genome of the MG1655 strain [primers (5′-3′): cgctcgaggaatagactgaacaccagactc and taccatggtaatccggcgcacgataatcg], and the amplified fragment was treated with NcoI and XhoI and cloned into the NcoI-XhoI sites of pHN1257, pHN1869 and pHN1868, yielding pHN1890, pHN1880 and pHN1878, respectively (Fig. 1, Table S1). The accA antisense sequence was similarly amplified using the specific primers [(5′-3′): aactcgagtttgactaatacaggaatacta and ccccatggctaaccgcagtcagagaatcga], and the fragment then cloned into the NcoI-XhoI sites of pHN1009 (Nakashima and Tamura 2009), pHN1477 and pHN1933, yielding pHN1994, pHN1995 and pHN1996, respectively (Table S1).

Integration of the Ptet-T7RNAP and Phom-tetR gene cassettes into the lacZ locus

Integration and counter selection of pHN1882 were carried out according to the method developed by Hamilton et al. (1989). Colony PCR was used to verify integration and counter selection steps.

PepN enzyme assay

To measure the enzymatic activities of PepN, transformants were pregrown overnight; diluted 1 : 400 with fresh medium; and cultured up to the mid-logarithmic phase. Where required, doxycycline and IPTG were added at the beginning of the culture. The PepN assay was performed as described previously using a fluorogenic peptide substrate, suc-LLVY-AMC (Bachem AG, Bubendorf, Switzerland) (Nakashima and Tamura 2009).

Real-time PCR assay

A real-time quantitative RT-PCR assay was performed with the One Step SYBR PrimeScript RT-PCR Kit II (Takara Bio Co., Ohtsu, Japan) and the Mx3000P Real-Time PCR System (Stratagene, La Jolla, CA, USA). As a normalizer gene, 16S rRNA was amplified [primers (5′-3′): ttgctcattgacgttacccg, acgcccagtaattccgatta] in parallel with pepN PTasRNA [primers (5′-3′): aggaatagactgaacaccagactc, gtgacggtatttggcttgtg] or pepN mRNA [primers (5′-3′): attgggtacagtggcaggac, aacgcgtggtaaaggtatcg]. The pepN PTasRNA primers were designed not to amplify the pepN mRNA. Culture conditions were the same as for the PepN enzyme assay.


We are grateful to Mr. Shobu Sakashita and Mr. Kazuhiro Takahashi for technical assistance. This work was supported in part by KAKENHI (23780096). The authors declare no conflict of interest.