ABSTRACT
By taking advantage of MazF, an ACA codon-specific mRNA interferase, Escherichia coli cells can be converted into a bioreactor producing only a single protein of interest by using an ACA-less mRNA for the protein. In this single-protein production (SPP) system, we engineered MazF by replacing two tryptophan residues in positions 14 and 83 with Phe (W14F) and Leu (W83L), respectively. Upon the addition of an inducer (IPTG [isopropyl-β-d-thiogalactopyranoside]), the mutated MazF [MazF(ΔW)] can still be produced even in the absence of tryptophan in the medium by using a Trp auxotroph, while a target protein having Trp residues cannot be produced. However, at 3 h after the addition of IPTG, the addition of tryptophan to the medium exclusively induces production of the target protein at a high level. A similar SPP system was also constructed with the use of a His-less protein [MazF(ΔH)] and a His auxotroph. Using these dual-induction systems, isotopic enrichments of 13C, 15N, and 2H were highly improved by almost complete suppression of the production of the unlabeled target protein. In both systems, isotopic incorporation reached more than 98% labeling efficiency, significantly reducing the background attributable to the unlabeled target protein.
The condensed single-protein production (cSPP) system was developed based on the endoribonuclease activity of an Escherichia coli toxin called MazF, which selectively cleaves cellular mRNAs at the ACA codon sequence (14, 16). Upon induction of MazF, protein synthesis is completely inhibited, and as a result, cell growth is also completely arrested. However, MazF-induced cells are in a quasidormant state, as they are metabolically fully active, producing ATP, amino acids, and nucleotides. Most significantly, in the quasidormant cells, machineries for protein synthesis and mRNA production are also fully functional, and therefore the MazF-induced quasidormant cells are still capable of synthesizing a protein of interest without producing any other cellular proteins, if the mRNA for the protein is engineered to have no ACA sequences (14). This system is thus termed the single-protein production (SPP) system. One of the most remarkable advantages of the SPP system is that the cell culture can be highly condensed without affecting protein yields (7, 12, 13). Using this cSPP system, one can achieve a cost savings of as much as 97.5% by condensing a culture 40-fold. This is particularly valuable when highly expensive isotopes or isotope-labeled compounds, such as amino acids and glucose, are used for the preparation of protein samples for structural study by nuclear magnetic resonance (NMR) spectroscopy. Furthermore, by use of the cSPP system, amino acid analogues or D2O, which is toxic in conventional protein production systems, reducing protein yields, is not toxic, hardly affecting the final protein yields. However, one drawback of the current cSPP system is the use of IPTG (isopropyl-β-d-thiogalactopyranoside) as an inducer for both MazF and a target protein, such that the target protein is also produced at the same time as MazF. Since isotopes or isotope-labeled compounds are added 2 to 3 h after the addition of IPTG to avoid their incorporation into cellular proteins, non-isotope-labeled target protein is also produced during this preincubation period, resulting in a higher background of unlabeled target protein, which may be as high as 20% of the final yield of the target protein produced (10).
The combination of both tetracycline- and IPTG-inducible systems was employed to separate the inductions of target protein and MazF, respectively (10). The main disadvantage of this system is that the expression level of target protein critically depends upon tetracycline concentration. The amounts of tetracycline being added to the cells for induction of target protein significantly affect the level of target protein synthesis, especially in the condensed SPP system. Therefore, a highly precise and accurate optimization of the tetracycline level is required for consistency in the expression of target proteins. To circumvent this problem, we developed a novel dual-induction system using amino acid auxotrophs. It has been shown previously that E. coli cells from a histidine (His) auxotroph can still produce a protein containing no His residues in the absence of histidine in the medium without producing any other cellular proteins (5). Therefore, it is assumed that even if both MazF and a target protein containing His residues or tryptophan (Trp) residues are coinduced by IPTG, only His-less MazF or Trp-less MazF would be produced in the absence of histidine or tryptophan by using a His auxotroph or a Trp auxotroph, respectively. In this fashion, the target protein may be induced by the addition of histidine or tryptophan in the medium a few hours after His-less or Trp-less MazF induction so that background production of the target protein may be avoided. Furthermore, this new SPP system could also be useful in studies involving specific replacement of amino acids with their analogues.
In the present study, we employed two amino acid auxotrophs, of Trp and His, to construct the dual-induction SPP system. For this purpose, both Trp-less proteins and His-less MazF proteins [MazF(ΔW) and MazF(ΔH), respectively] were generated to create the new SPP system. Using this new system, we tested a number of proteins, such as (i) E. coli EnvZB, which is the ATP-binding domain of the histidine kinase EnvZ (161 residues) (15), (ii) E. coli CspA, which is the major cold shock protein (70 residues) (3), (iii) E. coli YaiZ, which is a plasma membrane protein (80 residues) (7), (iv) the antiapoptotic adenoviral protein E1B19K150 (150 residues) (2), (v) human granulocyte colony-stimulating factor (GCSF; 175 residues) (4), and (vi) human calmodulin (CaM), which is a calcium binding protein (148 residues) (6). It was found that isotope incorporation into these proteins was very tightly regulated so that the background due to the unlabeled target protein was significantly reduced. We also developed a pCold(W) system in which a Trp tag can be added to the N-terminal part of a protein so that the dual-induction expression can still be applied for Trp-less proteins.
MATERIALS AND METHODS
Construction of MazF variants.Two MazF variants, MazF(ΔW) and MazF(ΔH), were constructed by site-directed mutagenesis. For MazF(ΔW), Trp14 and Trp83 in pMazF were replaced with Phe and Leu, respectively. To construct His-less MazF(ΔH), Gly27-His28 was altered to Lys-Arg. Primers used for MazF(ΔW) and MazF(ΔH) were as follows: for W14F, 5′-GATATGGGCGATCTGATTTTCGTTGATTTTGACCCG-3′; for W83L, 5′-GTAAAAAGTATCGCCCTGCGGGCAAGAGGAGCAACG-3′; and for G27K and H28R, 5′-GGTAGCGAGCAAGCTAAACGTCCAGCTGTTGTC-3′. Both sets of mutations were confirmed by sequencing the DNA of the individual plasmids pMazF(ΔW) and pMazF(ΔH).
Construction of Trp and His auxotrophs.In order to construct Trp auxotroph BL21(DE3) ΔtrpC and His auxotroph BL21(DE3) ΔhisB, phage lysates were prepared from the JW1254 and JW2004 strains, respectively, from the Keio collection (1) by using the P1 transduction method (8). The ΔtrpC and ΔhisB lysates were used to transfect BL21(DE3) cells, which were then plated on agar plates with or without tryptophan and histidine, respectively. The colonies which appeared on the plates were confirmed for the presence or absence of kanamycin cassettes by colony PCR technique using specific primers. The BL21(DE3) ΔtrpC and BL21(DE3) ΔhisB auxotrophs were also confirmed by their inability to grow in the absence of tryptophan and histidine, respectively. Competent cells derived thereof were transformed with pMazF(ΔW) and pMazF(ΔH), respectively. The obtained plasmid-bearing transformants were stored at −80°C, after being frozen in liquid nitrogen, until further use.
Toxicity test for pMazF(ΔW) and pMazF(ΔH).To determine the toxicities of the pMazF variants, pMazF, pMazF(ΔW), and pMazF(ΔH) were transformed in BL21(DE3) cells, which were then plated on LB plates containing chloramphenicol. A single colony from each plate was resuspended in 200 μl of LB broth, and 5-μl volumes of each were dropped on LB plates containing 0, 0.05, 0.1, 0.25, 0.5, and 1 mM IPTG. The plates were dried and incubated for at least 24 h at room temperature or overnight at 37°C.
In order to show whether the toxicity of MazF was neutralized by its antitoxin (MazE), pMazF and the pMazF variants were cotransformed with pBADMazE and overnight cultures were transferred and grown in 20 ml LB broth. When the optical density at 600 nm (OD600) reached about 0.4 to 0.5, the three different cultures were equally distributed into four sets of tubes (see Fig. 1c). In the first set, neither IPTG nor arabinose was added; in the second, only IPTG was added. The third set contained only arabinose, and the fourth set contained both IPTG and arabinose. Arabinose was added (0.2%) prior to the addition of IPTG for the production of antitoxin MazE.
Amino acid sequence of MazF and toxicity assay of MazF(ΔW) and MazF(ΔH). (a) Amino acid replacements in MazF(ΔW) and MazF(ΔH). In MazF(ΔH), two tryptophan residues at positions 14 and 83 in MazF were replaced with phenylalanine and leucine, respectively. In MazF(ΔH), the G27-H28 sequence was changed to KR. (b) Plate assay to compare the toxicity of pMazF with those of pMazF(ΔW) and pMazF(ΔH). Panels 1, 2, 3, 4, 5, and 6 correspond to 0.0, 0.05, 0.1, 0.25, 0.5, and 1 mM IPTG, respectively. a, b, and c correspond to pMazF(ΔW), pMazF, and pMazF(ΔH), respectively. (c) Coexpression of MazF and MazE to demonstrate that MazF toxicity was neutralized by MazE production in vivo. For induction of MazF and MazE, 0.1 mM IPTG and 0.2% arabinose, respectively, were added.
Primer extension analysis.Primer extension analysis was also carried out to determine the specificity of MazF cleavage sites in vivo (16). Total RNA was extracted from BL21(DE3) cells containing pMazF, pMazF(ΔW), or pMazF(ΔH) after induction with 1 mM IPTG for 10 min at 37°C. Primer extension was carried out at 47°C for 1 h with 10 units of avian myeloblastosis virus (AMV) reverse transcriptase (Roche) by using 15 μg of total RNA and 1 pmol of the appropriate primer labeled with [γ-32P]ATP through the use of T4 polynucleotide kinase (Takara Bio). A specific primer for ompA mRNA, 5′-GTTTTTACCATAAACGTTGG-3′, was used for all the reactions. The reaction was stopped by the addition of 12 μl of sequencing loading buffer (95% formaldehyde, 20 mM EDTA, 0.05% bromophenol blue, 0.05% xylene cyanol), and the reaction mixture was heated at 95°C for 2 min and then placed on ice. The products were analyzed using a 6% polyacrylamide gel containing 8 M urea, with a sequencing ladder made with the same primer.
Trp-inducible SPP system.The plasmid pColdI(SP4) or pColdI(W) bearing ACA-less genes for EnvZB, CspA, E1B19K150, GCSF, and CaM was transformed into BL21(DE3) ΔtrpC competent cells containing pMazF(ΔW). After overnight incubation at 37°C, colonies were inoculated in 50 ml of M9 medium containing tryptophan (20 μg/ml), ampicillin (100 μg/ml), and chloramphenicol (25 μg/ml). The overnight culture from 37°C was washed with M9 medium and transferred to 1 liter of M9 medium containing tryptophan along with the appropriate antibiotics. The cells were allowed to grow until the OD600 reached 0.5 to 0.6 at the same temperature, after which they were washed two or three times with isotope-enriched M9 medium and resuspended in Trp-less and isotope-enriched M9 medium containing [15N]NH4Cl, [13C]glucose (deuterated), or 2H2O. This was followed by a cold shock on ice for 5 min, an acclimatization at 15°C for 45 min, and a 20-times condensation of the culture as described previously (11, 13). The cells were incubated for an additional 45 min at this temperature, and IPTG was added at a final concentration of 0.1 to 0.5 mM to induce MazFΔW expression. After 3 h of incubation at 15°C, Trp (20 μg/ml) was added to induce target protein production for the next 16 to 24 h until the cells were harvested.
His-inducible SPP system.In this system, the BL21(DE3) ΔhisB host cell containing pMazF(ΔH) was used for transformation of pColdI(SP4) bearing ACA-less genes for YaiZ, GCSF, CaM, and EnvZB. Colonies from overnight plates were used for inoculum in 50 ml of M9 medium supplemented with histidine. The cells were washed with M9 medium and finally transferred to 1 liter of M9 medium containing histidine and appropriate antibiotics (ampicillin and chloramphenicol). The cells were allowed to grow at 37°C until the OD600 reached 0.5 to 0.6, after which they were washed three times using isotope-enriched M9 medium. The cells were then resuspended in His-less M9 medium, followed by cold shock and acclimatization as described in the previous section. After 3 h of induction with IPTG, the cells were washed again and finally resuspended in isotope-enriched M9 medium containing histidine (20 μg/ml) to induce target protein production for the next 16 to 24 h until the cells were harvested.
Mass-spectrometric analyses for isotopic incorporation.For mass-spectrometric analyses of isotope-labeled target proteins, the isotope-labeled target protein bands were excised from SDS-PAGE gel. The protein band was treated with trypsin and subjected to liquid chromatography-mass spectrometry (LC-MS) for further analysis as described earlier (10, 11).
RESULTS
Removal of Trp and His residues from MazF.The wild-type MazF contains two Trp residues and one His residue (Fig. 1a). To construct Trp-less MazF, Trp14 and Trp83 were replaced with Phe and Leu, respectively, in pMazF to create pMazF(ΔW). To construct His-less MazF, Gly27-His28 was altered to Lys-Arg to create pMazF(ΔH). For both plasmids, the mazF gene is under the control of the IPTG-inducible lac promoter so that gene expression can be induced by the addition of IPTG. The toxicities of these mutated MazF proteins were examined on plates containing various concentrations of IPTG by using E. coli BL21(DE3) cells carrying pMazF, pMazF(ΔW), or pMazF(ΔH). As shown in Fig. 1b, both MazF mutants showed toxicity almost identical to that of wild-type MazF in all IPTG concentrations tested, indicating that the amino acid substitutions did not affect MazF toxicity. As well, in the demonstration that these MazF variants could inhibit cell growth independently of MazF endoribonuclease activity, coexpression with MazE, an inhibitor of MazF, was also found to revert the toxic effect, as shown in Fig. 1c (panel 4). In the absence of MazE, induction of MazF by IPTG resulted in cell growth inhibition (Fig. 1c, panel 2). Notably, the Trp and His residues are not conserved in MazF homologues, also supporting the notion that these residues do not play crucial roles in the enzymatic activity of MazF (9).
Specific enzymatic activities of MazF variants.The toxic effects of the MazF variants alone could not entirely support the fact that the specificity of the cleavage sites still remained the same. To demonstrate that the cleavage sites of MazF variants had not been altered, primer extension analysis was carried out using ompA mRNA as the template. The primer extension analysis of ompA using a specific primer showed distinct bands exhibiting the specific cleavage sites after induction of MazF by IPTG (Fig. 2). These bands were identical in MazF(ΔW) and MazF(ΔH), as well as in the wild-type MazF, where the cleavage occurred before or after the A residue of ACA sequences (Fig. 2, lanes 5, 7, and 9, respectively). This result clearly indicates that both the MazF variants cleaved mRNAs at the specific sequence ACA in vivo.
Primer extension analysis of MazF and MazF variants. Total RNA was extracted from both the induced and uninduced cells bearing pMazF, MazF(ΔW), and MazF(ΔH). Primer extension was carried out using an ompA mRNA-specific primer as described in Materials and Methods. This result shows that identical bands were produced from all the RNAs extracted from induced cultures, suggesting their similar MazF cleavage sites (lanes 5, 7, and 9). Lanes 6, 8, and 10 represent the same from the uninduced cultures (wild type [WT]).
Use of pMazF(ΔW) and pMazF(ΔH) in the SPP system.For construction of a Trp auxotroph, the gene for trpC was deleted by replacement with a kanamycin resistance gene, creating BL21(DE3) Δtrp from BL21(DE3). For construction of a His auxotroph, the gene for hisB was deleted by replacement with the same kanamycin resistance gene, creating BL21(DE3) Δhis. First, these deletion strains were transformed with pMazF(ΔW) and pMazF(ΔH) and used as hosts for the SPP expression of several proteins, such as CspA, EnvZB, OmpX, E1B19K150, YaiZ, CaM, and GCSF. All their genes were engineered to be ACA free without altering the encoded amino acid sequences, and their codon usages were optimized for expression in E. coli.
Note that since all proteins tested contain at least one Trp residue (one Trp residue in CspA, two in GCSF, three in EnvZB, and seven in E1B19K150), their syntheses were also stopped even though their mRNAs lack ACA sequences. After 3 h of incubation, tryptophan was added to the medium to initiate the production of the target protein, and EnvZB was expressed very well after overnight incubation (Fig. 3a, lane 3). With EnvZB, after 3 h of IPTG induction but before the addition of tryptophan, very little EnvZB was produced (Fig. 3a, lane 2). After overnight incubation without the addition of tryptophan, very little production of the protein was detected (Fig. 3a, lane 4), and this was likely due to the generation of Trp by degradation of the endogenous proteins during the incubation. Note that identical amounts of culture were applied in all lanes and no changes in the amounts of all cellular proteins (except for EnvZB) were observed. This indicates that no significant amount of cellular protein was produced during the overnight culture.
Expression profiles of different proteins by using the Trp-inducible SPP system. (a) EnvZB and CspA (lanes 1 to 4 and 5 to 8, respectively); (b) E1B19K150 and GCSF (lanes 1 to 4 and 5 to 8, respectively). Lanes 1 and 5, before IPTG induction; lanes 2 and 6, after 3-h IPTG induction; lanes 3 and 7, overnight incubation in the presence of tryptophan (20 μg/ml); lanes 4 and 8, overnight in the absence of tryptophan; lanes M, molecular weight markers.
CspA, E1B19K150, and GCSF were also expressed in the same manner as EnvZB, as shown in Fig. 3a, lanes 5 to 8, and b, lanes 1 to 4 and lanes 5 to 8, respectively. In all cases, the target proteins were produced in high yields (Fig. 3a, lane 7, and b, lanes 3 and lane 7, respectively) with very low background production before the addition of tryptophan (Fig. 3a, lane 6, and b, lanes 2 and 6, respectively). Also, overnight production of the target proteins without the addition of tryptophan was very low in all cases (Fig. 3a, lane 8, and b, lanes 4 and 8, respectively). These results demonstrate that protein production can be very tightly controlled by tryptophan by using a Trp auxotroph and effectively induced by the addition of tryptophan in the SPP system.
We also carried out the same experiments with a His auxotroph expressing YaiZ, GCSF, CaM, and EnvZB as shown in Fig. 4 (lanes 3, 6, 9, and 12, respectively). In all cases, their expression was tightly regulated by the addition of histidine, with very low background production of these proteins before the addition of histidine. Since pCold vectors already contain one or more His residues, this system can be used for expression of His-less proteins by using the normal pCold vectors.
Expression profiles of different target proteins (YaiZ, GCSF, CaM, and EnvZB) by using the His-inducible SPP system. Lanes 1, 4, 7, and 10, before IPTG induction; lanes 2, 5, 8, and 11, overnight without histidine; lanes 3, 6, 9, and 12, overnight in the presence of histidine; lane M, molecular weight markers.
Controlled protein synthesis by a single amino acid for Trp-less proteins.In the experiments described above, Trp and His were chosen, as these residues exist in most proteins. However, some proteins may not contain these residues. For this reason, we created a new vector system which adds a Trp residue at the N-terminal end of the protein of interest. To each of the pCold vectors, pColdI(SP4), pColdII(SP4), and pColdIII(SP4), developed for the SPP system (13), a TGG codon for Trp was added immediately after the translation enhancing element (TEE), as shown in Fig. 5a. These vectors were designated pColdI(W), pColdII(W), and pColdIII(W), respectively. We tested pColdI(W) for the production of human CaM, a protein of 148 residues that does not contain Trp. The gene for CaM was engineered to be ACA free, and its codon usage was optimized for E. coli. It was subsequently cloned into pColdI(W), as was pColdI as a control. It is important to note that neither vector contains ACA sequences in its transcript.
Expression of Trp-less protein for the Trp induction system. (a) pCold vectors used in this study were pColdI(SP4) (structure A), pColdI(W) (structure B), pColdII(W) (structure C), and pColdIII(W) (structure D). (b) Expression of a Trp-less human calmodulin by using the dual-induction SPP system. Lanes 1 and 4, before IPTG induction; lanes 2 and 5, after 3-h IPTG induction; lanes 3 and 6, overnight in the presence of tryptophan (20 μg/ml); lane M, molecular weight markers.
There is a substantial difference in the amounts of target protein produced with pColdI and pColdI(W) after 3 h of IPTG induction but before Trp addition (Fig. 5b, lanes 2 and 5, respectively). The presence of one Trp residue in pColdI(W) suppressed the production of CaM very efficiently when Trp was not added to the medium containing IPTG (Fig. 5b, compare lanes 5 and 2).
Isotopic incorporation with 15N-, 13C-, and 2H-enriched media.The use of media enriched with [15N]NH4Cl, [13C]glucose, and D2O (2H) showed no significant decrease in the expression level of target proteins with 20-times-condensed cultures. LC-MS analyses of the tryptic fragments obtained from EnvZB, E1B19K150, and CaM were conducted. The obtained tryptic fragments are described in Table 1. The LC-MS results for EnvZB are shown in Fig. 6. The data showed that the levels of incorporation were consistent for all the peptides being studied, with maximum labeling efficiencies higher than 90% (Table 1). The efficiency of isotopic incorporation was analyzed based on retention time and tandem-MS information by integrating the areas of peaks corresponding to the isotope-labeled peptides (10, 11). The ratio of labeled peak area to unlabeled peak area was employed for estimation of the amount of isotopic incorporation. The conventional IPTG-inducible SPP system usually resulted in as high as 20% unlabeled peaks (10). Using the present Trp- and His-inducible systems, this unlabeled background was significantly reduced, resulting in a maximum yield of 99% labeled peaks (Table 1).
Graphical representations of LC-MS results showing isotopic enrichment of EnvZB tryptic fragments. (a) 15N enrichment for HLFQPFVR; (b) 15N enrichment for EIETALYPGSIEVK; (c) 13C enrichment for HLFQPFVR; (d) 13C enrichment for EIETALYPGSIEVK; (e) 2H enrichment for AWFQVEDDGPGIAPEQR; (f) 2H enrichment for AVANMVVNAAR.
Isotopic incorporation rates of several peptides generated from proteins produced by the cSPP dual-induction system
DISCUSSION
In this study, we have developed a novel and very tight induction method for protein production in the SPP system by the addition of an amino acid which is required for cell growth. Although in this system both the mazF mRNA and the ACA-less mRNA for the target protein are simultaneously induced, only the mazF mRNA is translated in the absence of tryptophan by Trp-auxotrophic E. coli cells, since two Trp residues in MazF are replaced with other amino acid residues while the target protein contains one or more Trp residues. In this manner, only after cellular mRNAs are completely removed can the translation of the ACA-less mRNA for the target protein be initiated by the addition of Trp to the medium. We have developed a similar dual-induction SPP system with the use of His-less MazF and a His auxotroph. With these dual-induction systems, we are now able to label a target protein with isotopes at a very high efficiency, thereby significantly reducing the background due to the unlabeled target protein.
The lowering of the background is very important for protein NMR studies, since the unlabeled target protein contributes to a reduced yield of the isotope-enriched species and a lower signal-to-noise ratio in such studies. In this study, we achieved a near absence of unlabeled background protein at 3 and 24 h without Trp or His (Fig. 3 and 4). It is important to note that in the dual-induction systems developed, a trace amount of residual Trp or His in the cells could result in translation of target proteins. Therefore, extensive washing of the cells, as well as prior incubation of the cells to deplete residual Trp or His in the cells before induction of the target proteins, is essential. Notably, even proteins which do not contain Trp or His residues can be produced by the present dual-induction system simply by adding one Trp residue by use of pColdI(W) or His residues by use of pColdI(SP4) or pColdII(SP4), which add a His tag at the N-terminal end of the target protein. We are currently applying the dual-induction system developed here to a number of proteins for NMR structural studies.
Furthermore, the present SPP system may be highly useful for incorporation of specific isotope-labeled amino acid(s) as well as toxic (or nontoxic) amino acid analogues useful for functional or structural studies of a protein, since they are incorporated only into a target protein, thus eliminating any cytotoxic effects of amino acid analogues on the cells.
ACKNOWLEDGMENTS
We thank H. Zheng and C. Zhao of the Biological Mass Spectrometry Facility, CABM, for carrying out all the mass spectrometric analysis.
This work was supported by funding from the National Institutes of Health (grant 5R01GM085449, to M.I.), the Northeast Structural Genomics Consortium (to G. Montelione), and the New York Consortium on Membrane Protein Structure (to W. Hendrickson). This work was also supported by a grant from Takara Bio Inc.
FOOTNOTES
- Received 1 April 2010.
- Accepted 12 July 2010.
- Copyright © 2010 American Society for Microbiology
