Suppression of amber codons in Caulobacter crescentus by the orthogonal Escherichia coli histidyl-tRNA synthetase/tRNAHis pair.

While translational read-through of stop codons by suppressor tRNAs is common in many bacteria, archaea and eukaryotes, this phenomenon has not yet been observed in the α-proteobacterium Caulobacter crescentus. Based on a previous report that C. crescentus and Escherichia coli tRNA(His) have distinctive identity elements, we constructed E. coli tRNA(His) CUA, a UAG suppressor tRNA for C. crescentus. By examining the expression of three UAG codon- containing reporter genes (encoding a β-lactamase, the fluorescent mCherry protein, or the C. crescentus xylonate dehydratase), we demonstrated that the E. coli histidyl-tRNA synthetase/tRNA(His) CUA pair enables in vivo UAG suppression in C. crescentus. E. coli histidyl-tRNA synthetase (HisRS) or tRNA(His) CUA alone did not achieve suppression; this indicates that the E. coli HisRS/tRNA(His) CUA pair is orthogonal in C. crescentus. These results illustrate that UAG suppression can be achieved in C. crescentus with an orthogonal aminoacyl-tRNA synthetase/suppressor tRNA pair.

Introduction

Nonsense suppression (read-through of the stop codons UAG, UAA, or UGA) has been observed in many biological systems. Since nonsense mutations cause premature termination of translation, suppression of nonsense codons has been extensively applied in prokaryotes and lower eukaryotes for identifying gene functions by conditional expression of such genes [1], [2]. Nonsense mutations in essential genes could be lethal. Moreover, eukaryotic messenger RNAs that contain premature termination codons are degraded through nonsense-mediated decay [3], and may be associated with many genetic disorders [4]. Therefore, nonsense suppression would rescue cells from deleterious effect of nonsense mutations. Genetic studies showed that nonsense suppression is often mediated by suppressor tRNAs with altered anticodons that now recognize nonsense codons and insert an amino acid in response. Such suppressors have been found in many organisms including E. coli [5], yeast [2], C. elegans [6], [7], and human [8].

C. crescentus is an α-proteobacterium common in fresh water ecosystems. It has been widely used as a model organism for cell cycle and cellular differentiation studies because of its unique characteristics during cell division. Recently it was found that C. crescentus tRNAHis lacks a universal feature, G-1, the major identity element recognized by HisRS in acylating tRNAHis. Because of differences in the nature of the discriminator base and of position 72 of tRNA, and the presence or absence of G-1, the HisRS/tRNAHis pairs of C. crescentus and E. coli are orthogonal [9]–[11].

Therefore we reasoned that an E. coli tRNAHis with a CUA anticodon (tRNAHis CUA) should act as a suppressor tRNA for UAG codons in C. crescentus in the presence of the E. coli HisRS (Fig. 1). To test our hypothesis, we decided to express three C. crescentus UAG-containing reporter genes encoding a β-lactamase, the fluorescent mCherry protein, and the C. crescentus xylonate dehydratase.

Figure 1

Suppression of amber codons in C. crescentus by E. coli HisRS/tRNAHis CUA.

The E. coli HisRS/tRNAHis CUA pair with the CUA anticodon is orthogonal in C. crescentus. The E. coli pair suppresses an in-frame amber codon in the reporter gene, which allows the expression of the gene while C. crescentus aminoacyl-tRNA synthetase (aaRS)/tRNA pairs are not able to suppress the amber mutation. E. coli HisRS is shown in blue and C. crescentus aminoacyl-tRNA synthetases are shown in green.

Materials and Methods

Strains and Plasmids

The strains, plasmids and primers in the study can be found in Tables S1 and S2. CB15ΔxylD was a gift of Craig M. Stephens, Santa Clara University. A lac promoter was prepared from pUC19 by PCR and cloned into pRVCHYN-5 [12], leading to pRV-lac2-mCherry. Ala184 of an ampicillin resistance gene, His22 of mCherry gene, His283 or His290 of xylD, and the codon encoding Phe3 of groES were mutated to amber codon by QuikChange (Stratagene).

The DNA sequence of E. coli HisRS was codon-optimized for C. crescentus by using JCat [13]. It was synthesized (GenScript) and inserted into pBXMCS-2 [12] or pBX-lac2. A precursor of E. coli tRNAHis contains the sequence from –40 to +123 of tRNAHis (G-1 is the first nucleotide of tRNAHis). Since E. coli tRNAHis should be produced by correct processing of the precursor transcripts in C. crescentus, we constructed another transcription unit, tRNAHis2 CUA, by replacing C. crescentus mature tRNAHis from the precursor transcript of tRNAHis with E. coli mature tRNAHis. The 5′ and 3′ flanking regions of tRNAHis2 remained those of C. crescentus sequence. The sequences of the precursor tRNAHis are shown in Figure S1. The template for mature tRNAHis was prepared by PCR.

The promoter sequence of C. crescentus ffs was amplified from the genomic DNA and a transcriptional terminator, rrnC, was obtained from pTECH [14]. For the expression of HisRS, tRNAHis or tRNAHis CUA, the appropriate fragments were combined as described in Fig. 2A. C. crescentus groES was inserted into SacII site of pBX-lac2-HisRS-tRNAHis CUA by In-Fusion (Clontech).

Figure 2

Plasmids for E. coli HisRS/tRNAHis and confirmation of E. coli tRNAHis expression in C. crescentus.

A, A schematic representation of expression units for E. coli HisRS, and tRNAHis. The gene of E. coli HisRS was cloned to NdeI and EcoRI sites of pBXMCS-2 or pBX-lac2. The tRNAHis units were inserted into EcoRI and SmaI sites. E. coli tRNAHis is transcribed as a precursor from the ffs promoter and then processed to generate the mature tRNAHis (in green). The flanking sequence is shown in light green. B, Northern blot analysis of E.coli tRNAHis from C. crescentus cultures. 0.5 fmol of in vitro transcripts of E. coli tRNAHis was used as a control. Eco tRNAHis CUA and Eco tRNAHis2 CUA stand for total RNAs from the cells containing pBX-HisRS-tRNAHis CUA and pBX-HisRS-tRNAHis2 CUA, respectively. Cells were grown in presence or absence of 0.2% xylose.

Northern Blot Analysis

Ten microgram of total RNAs from C. crescentus were separated on a 1% agarose gel and transferred into Hybond-N+ membrane (GE Healthcare Life Sciences) by the semi-dry blotter (Bio-Rad). After 2 h hybridization in Rapid-hyb (GE Healthcare Life Sciences) with 0.2 pmol of 5′ end-labeled HisR1 at 42°C, the membrane was washed twice with Solution I (2xSSC and 0.1% SDS) and then twice with Solution II (0.1xSSC and 0.1% SDS) at room temperature. The membrane was analyzed by autoradiography. The control E. coli tRNAHis was prepared by in vitro transcription of pUCT7/tRNAHis.

Viability Assays

All incubations were performed at 30°C. For ampicillin resistance, CB15NΔbla6 was transformed with pRV-lac2-AmpRTAG, which was constructed by Ala184TAG mutation on the ampicillin resistance gene in pRV-lac2-mCherry, by electroporation. Subsequently, pBX-derived plasmids in E. coli S17-1 were introduced into the strain by mating. The cells were cultured in PYE medium (peptone-yeast extract) containing 5 µg/ml of kanamycin (Kan) and 1 µg/ml of oxytetracycline (OxyTet) overnight. Then 3 µl of cultures were spotted onto the PYE plates containing 0.2% xylose, 20 µg/ml of Kan, 2 µg/ml of OxyTet and 50 µg/ml of ampicillin. For the growth curves, 200 µl of the diluted cultures (A660 nm = 0.01) were further incubated in the same medium in presence or absence of 50 µg/ml of ampicillin. The absorbances at 660 nm were measured at every 30 min by using Synergy HT plate reader (BioTek). The assays were repeated at least three times.

For the xylose assay, CB15ΔxylD was transformed with either pRV-lac2-xylDTAG847 or pRV-lac2-xylDTAG868, and obtained the pBX-lac2-derived plasmids by mating as described above. The cells were cultured in PYE media with Kan and OxyTet overnight and then diluted to A660 = 0.2 with the same media. The diluted cultures were further incubated for another 6 h and collected by centrifugation. The cells were washed twice and then resuspended with M2 minimal media. The cells were spotted onto M2 plates containing either 10 mM glucose (M2G) or 10 mM xylose (M2X). The plates were incubated at 30°C until cell growth was able to be observed. The assays were repeated twice.

Microscopy and Data Analysis

C. crescentus cells were grown at 30°C in M2G+ media (M2G supplemented with 1% PYE) to exponential phase (A660 = 0.3–0.4) supplemented with 5 µg/ml of Kan and 1 µg/ml of OxyTet, and then induced by adding 0.03% xylose for 2–3 h before visualization. The PYE-supplemented M2G medium was used in order to increase growth rate without interfering the microscopic analysis.

Cells were immobilized on 1% agarose pads with M2G+ supplemented with Kan, OxyTet and 0.03% xylose. Cell imaging was performed at room temperature (∼22°C) using a Nikon 80i microscope equipped with a 100X NA 1.4 phase contrast objective and a Hamamatsu Orca II-ER camera controlled by MetaMorph software or a Nikon Ti-U inverted microscope with a 100X NA1.4 phase contrast objective and a Hamamatsu Orca ER camera. Data analysis was performed using the MATLAB-based software MicrobeTracker [15]. The experiments were repeated at least twice.

Purification of GroES and Mass Spectrometric Analysis

For the expression of C. crescentus groES, CB15N with pBX-lac2-HisRS-tRNAHis CUA-groESTAG was grown and then the culture was diluted 1∶200 with 1 L of PYE containing 5 µg/ml of Kan and then incubated at 30°C overnight. Then the temperature was raised to 40°C for 2 h. The protein was purified by Ni-NTA column (Qiagen). The protein was sent to the W.M. Keck Biotechnology Resource Laboratory at Yale University for liquid column chromatography-tandem mass spectrometry (LC-MS/MS) analysis.

Results

Expression of the E. coli HisRS/tRNAHis Pair in C. crescentus

The E. coli HisRS and tRNAHis genes were introduced into C. crescentus with the constructed plasmids. For expression, the HisRS and tRNAHis genes were inserted consecutively into the pBXMCS-2 plasmid with a xylose-inducible promoter (P) or the pBX-lac2 plasmid with a lac promoter (P) (Fig. 2A). The ffs promoter, which is derived from the constitutive promoter of the 4.5S RNA gene in C. crescentus [16], was inserted before the tRNAHis genes. To make an amber suppressor, the anticodon of tRNAHis was mutated to CUA. Because G-1 is a major identity element for E. coli HisRS [17], this nucleotide was included in all tRNAHis constructs. The production of mature E. coli tRNAHis in C. crescentus was confirmed by Northern blot analysis (Fig. 2B).

UAG Suppression by the E. coli HisRS/tRNAHis CUA Pair

The orthogonality of the E. coli HisRS/tRNAHis CUA pair and its efficiency for UAG (amber) suppression in C. crescentus were investigated using three reporter genes with an in-frame UAG codon. Since natural suppressor tRNAs are absent in C. crescentus [18], the productions of full-length proteins indicate that successful suppression resulted from the E. coli HisRS/tRNAHis CUA pair.

First, a β-lactamase gene (bla) with a permissive amber mutation (Ala184TAG) [19], [20] was introduced into the ampicillin-sensitive C. crescentus strain CB15NΔbla6, which does not survive ampicillin concentrations at or greater than 50 µg/ml [21]. When E. coli HisRS and tRNAHis CUA were expressed together, the cells survived 50 µg/ml of ampicillin in both solid (Fig. 3A) and liquid media (Fig. 3C), although the presence of ampicillin in the medium delayed the growth of the strain at lag phase (Figs. 3B and 3C). The other precursor construct of the suppressor, tRNAHis2 CUA, showed similar resistance, while the growth of the strain was slower than in the HisRS-tRNAHis CUA-containing strain (Figs. 3A and 3C). All the strains grew at similar rates in the medium lacking ampicillin (Fig. 3B). Those results indicate that both constructs were able to produce active E. coli tRNAHis in C. crescentus. The increase of ampicillin resistance suggests production of the active, full length β-lactamase resulted from amber suppression. On the other hand, neither E. coli HisRS nor tRNAHis CUA alone increased ampicillin resistance (Figs. 3A and 3C). This indicates that the E. coli HisRS and tRNAHis CUA did not cross-react with the host tRNAs and aminoacyl-tRNA synthetases, respectively.

Figure 3

Increase of ampicillin resistance from suppression of an in-frame UAG codon.

A, Growth test on ampicillin-containing plates. C. crescentus cells were incubated for 1 day (upper panel) or 2 days (lower panel) at 30°C. All cells contained pRV-lac2-AmpRTAG and a plasmid for HisRS and/or tRNAHis. 1. pBXMCS-2; 2: pBX-HisRS; 3: pBX-HisRS-tRNAHis CUA; 4: pBX-tRNAHis CUA; 5: pBX-HisRS-tRNAHis2 CUA; 6: pBX-tRNAHis2 CUA. B, Growth curves of the different strains in liquid media. The medium lacked ampicillin. C, Growth curves of the strains in liquid media in presence of 50 µg/ml of ampicillin. The error bars indicate standard deviations.

Second, the efficiency of amber suppression was determined with the production of the fluorescent protein mCherry [22]. When wild-type mCherry was expressed from the plasmid pRV-lac2-mCherry, fluorescence was clearly detected (Fig. 4A) as expected. Introduction of an amber mutation (His22TAG) into the mCherry gene resulted in no fluorescence detected in cells (Fig. 4B). The co-expression with the E. coli HisRS/tRNAHis CUA pair restored 4±2% (in two independent experiments) of the fluorescence intensities compared to the wild-type mCherry (Fig. 4C and 4D). Although the E. coli HisRS/tRNAHis CUA pair only partially restored the fluorescence, this result indicated effective amber suppression under no growth selection pressure. The growth rates of the strains that contained either pBXMCS-2 (the empty vector) or pBX-HisRS-tRNAHis CUA were similar (Fig. S2), which implies that the production of the HisRS/tRNAHis CUA pair does not affect the protein synthesis significantly in the strain.

Figure 4

Suppression of a UAG codon of the mCherry gene.

A, A fluorescence (left) and a phase contrast (right) image. The cells contained pRV-lac2-mCherry and pBX-HisRS-tRNAHis CUA. The scale bar represents 1 µm. B. Images of cells that contained the mutant mCherry gene and pBXMCS-2, which is the empty vector. C. Images of cells expressing the mutant mCherry gene and the E. coli HisRS/tRNAHis CUA. D, Histogram of the ratio between percentages of cells and fluorescence intensities. CB15N stands for C. crescentus CB15N strain harboring pRV-lac2-mCherryTAG. Eco HisRS+tRNAHis CUA stands for the same strain containing the additional pBX-HisRS-tRNAHis CUA plasmid.

Third, the effect of amber suppression was evaluated by rescuing a C. crescentus xylD nonsense mutation. The C. crescentus xylD encodes a xylonate dehydratase that is essential for xylose metabolism. The C. crescentus xylD deletion mutant strain (CB15ΔxylD) cannot utilize xylose as a sole carbon source [23]. For complementation, two constructs of xylD with amber mutation (His283TAG and His290TAG) were individually introduced into the CB15ΔxylD strain pre-transformed with pBX-lac2-HisRS-tRNAHis CUA for expression of the E. coli HisRS/tRNAHis CUA pair. As we expected, all strains grew on M2G plates supplemented with glucose, but only the complementation strains expressing both the E. coli HisRS and tRNAHis CUA were able to grow on M2X plates with xylose as the sole carbon source (Fig. 5). The strain containing xylDTAG847 (His283TAG) grew faster than that containing xylDTAG868 (His290TAG); this suggests that the suppression efficiency is different depending on the position of the amber mutation. The growth of the strain that contained HisRS alone seemed slower than those of the other strains (lane 2 in the M2G plate), however, in another sets of the experiment, the growth of the strain was similar to those of the other strains in both an M2G plate and PYE liquid medium (data not shown).

Figure 5

Growth rescue by suppression of a UAG codon of xylD.

CB15ΔxylD was incubated at 30°C for 3 days (M2G plate) or 9 days (M2X plate). The cultures were diluted serially (1∶5) and then spotted onto the plates. The cells contained either pRV-lac2-xylDTAG847 (upper panel) or pRV-lac2-xylDTAG868 (lower panel). The cells also contained a plasmid for the E. coli HisRS and/or tRNAHis CUA. 1: pBX-lac2-HisRS-tRNAHis CUA; 2: pBX-lac2-HisRS; 3: pBX-lac2-tRNAHis CUA; 4: pBX-lac2.

Confirmation of Histidine Incorporation by the E. coli HisRS/tRNAHis CUA Pair

GroES is a chaperonin that forms a complex with GroEL [24]. In C. crescentus, GroES is expressed from groESL operon and the expression of groES is elevated at 40°C [25]. We cloned C. crescentus groES including the intrinsic promoter and mutated the phenylalanine codon at the third position to amber codon. The mutated gene was expressed with E. coli HisRS/tRNAHis CUA pair in C. crescentus. Histidine incorporation at amber codon was confirmed by LC-MS/MS analysis. Only a few ions were detected but they confirmed the fragment peptide, HRPLGDR, which contained histidine at the UAG codon of GroES (Fig. 6). We detected a histidine immonium ion from the N-terminus of the peptide, contributing to the identification of the fragment. We did not observe other amino acids at position 3 from the LC-MS/MS analysis.

Figure 6

Mass spectroscopic confirmation of histidine incorporation by the E. coli HisRS/tRNAHis CUA pair.

Annotated MS/MS spectra and ions for the HRPLGDR peptide from GroES. H stands for an immonium ion of histidine.

Discussion

In this study, we demonstrated that the orthogonal E. coli HisRS/tRNAHis CUA pair suppressed amber codons in C. crescentus. In three reporter systems, the amber suppression led to increased ampicillin resistance, production of the fluorescent protein mCherry, and complementation of xylD mutations. Based upon the fluorescence intensities of cells expressing mCherry, the suppression efficiency was about 4% (Fig. 4D). To account for the suppression efficiency, two factors are to be considered. First, alteration of the tRNAHis anticodon to CUA decreases about 20-fold decrease the in vitro aminoacylation efficiency (k cat/Km values) of E. coli HisRS [26]; thus tRNAHis CUA is a poorer substrate than wild-type tRNAHis. Second, the bases around termination codons have been suggested to affect the efficiencies of both termination [27], [28] and suppression [29], [30]. Indeed, the suppression efficiencies were different depending on the positions of amber mutations in xylD (Fig. 5). Therefore, the suppression efficiencies of mutant mCherry and other systems may also be influenced by context effects.

Orthogonal tRNA synthetase/tRNA pairs have been applied in bacterial and eukaryotic systems for translational incorporation of non-canonical amino acids (reviewed in [31]–[33]). Since the E. coli HisRS/tRNAHis CUA pair is orthogonal in C. crescentus, this pair can serve as a new tool for protein engineering in the important model organism C. crescentus. In this organism protein constructs can be secreted to the cell surface by fusion of the recombinant protein with RsaA, an abundant surface layer protein in C. crescentus. This process facilitates high yield and convenient protein purification by filtration [34], [35] and makes C. crescentus an ideal organism for production of some engineered proteins. Second, C. crescentus has been widely used for studying cell cycle, cell division, and cellular differentiations. These studies could be aided by the application of engineered E. coli HisRS/tRNAHis CUA pairs that would efficiently acylate histidine analogs and lead to incorporation of non-canonical amino acids with desirable properties.

The precursor tRNAHis sequences. A. The sequence of tRNAHis CUA that corresponds to E. coli precursor tRNAHis from −40 to +123. The sequence of mature tRNAHis is shown in uppercase. The anticodon for amber codon is labeled in red. B. The sequence of the tRNAHis2 CUA. It contains the identical E. coli mature tRNAHis sequence. In contrast to tRNAHis CUA, the flanking sequences originated from the C. crescentus sequences.

(PDF)

Click here for additional data file.

rowth curves of CB15N. Both strains carried pRV-lac2-mCherryTAG as well as either empty vector (pBXMCS-2) or the vector containing HisRS and tRNAHis CUA (pBX-HisRS-tRNAHis CUA). The growth curve in blue came from the strain with pBXMCS-2 and the growth curve of the strain that has pBX-HisRS-tRNAHis CUA is shown in red. The cultures grew in PYE medium with 1 µg/ml oxytetracycline, 5 µg/ml kanamycin and 0.2% xylose. The error bars indicate standard deviations.

(PDF)

Click here for additional data file.

Bacterial strains and plasmids used in this study.

(PDF)

Click here for additional data file.

Oligonucleotides used in this study.

(PDF)

Click here for additional data file.