• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of aemPermissionsJournals.ASM.orgJournalAEM ArticleJournal InfoAuthorsReviewers
Appl Environ Microbiol. Apr 2009; 75(7): 2002–2011.
Published online Feb 5, 2009. doi:  10.1128/AEM.02315-08
PMCID: PMC2663224

Random Mutagenesis of the Pm Promoter as a Powerful Strategy for Improvement of Recombinant-Gene Expression[down-pointing small open triangle]


The inducible Pm-xylS promoter system has proven useful for production of recombinant proteins in several gram-negative species and in high-cell-density cultivations of Escherichia coli. In this study we subjected a 24-bp region of Pm (including the −10 element) to random mutagenesis, leading to large mutant libraries in E. coli. Low-frequency-occurring Pm mutants displaying strongly increased promoter activity (up-mutants) could be efficiently identified by using β-lactamase as a reporter. The up-mutants typically carried multiple point mutations positioned throughout the mutagenized region, combined with deletions around the transcription start site. Mutants displaying up to about a 14-fold increase in β-lactamase expression (relative to wild-type Pm) were identified without loss of the inducible phenotype. The mutants also strongly stimulated the expression of two other reporter genes, luc (encoding firefly luciferase) and celB (encoding phosphoglucomutase), and were found to significantly improve (twofold) a previously optimized process for high-level recombinant production of the medically important granulocyte-macrophage colony-stimulating factor in E. coli under high-cell-density conditions. These results demonstrate the potential of using random mutagenesis of promoters to improve protein expression at industrial levels and indicate that targeted modifications of individual functional elements are not sufficient to obtain optimized promoter sequences.

The Pm-xylS promoter system drives the expression of the meta-cleavage operon carried by the Pseudomonas putida TOL plasmid pWW0. The gene products of this operon are involved in the catabolism of alkylbenzoates and are expressed in response to meta-pathway substrates (25). XylS positively regulates Pm when forming an activated complex with effectors like benzoate or its derivatives (13, 26). Transcriptional activation occurs through binding of the activated XylS to two direct imperfect repeats located directly upstream of the −35 region of Pm (12).

Pm-xylS has been shown to be useful for high-level expression of recombinant proteins in a wide range of gram-negative bacterial species (3, 4, 20, 24). The uninduced expression level is low, and the use of different effector compounds at various concentrations can be used to regulate the level of induced expression (35). Many of the inducers are low-cost compounds that enter the cell by passive diffusion. Previously, we reported the use of this system in the construction of broad-host-range expression vectors based on the RK2 minimal replicon (2, 3). One of these vectors, pJB658, has proven useful for tightly regulated recombinant-gene expression in several gram-negative species (2, 3, 5, 31, 35). Industrial levels of production of human proteins under high-cell-density conditions (HCDC) of Escherichia coli has also been demonstrated with the Pm-xylS system coupled to the RK2 replicon (30, 31). By studying the functionality of different secretion signal sequences, the effects of various vector copy numbers, and the conditions of promoter induction, we achieved about 0.8 g/liter of recombinant granulocyte-macrophage colony-stimulating factor (GM-CSF) under HCDC of E. coli (31).

The Pm promoter does not exhibit apparent sequence homology to typical consensus promoter elements. The absence of a −35 element is probably due to the requirement for XylS binding to promote transcription initiation, mediated by two different σ factors: σ32 in the exponential phase and σS in the stationary phase (19). The −10 element was experimentally determined (TAGGCT) by Domínguez et al. (9), and in the same study nucleotide positions important to σ32 and σS recognition were identified.

We have previously reported that the expression level from Pm could be approximately doubled by introducing random mutations in the −10 region (34). These experiments had certain technical limitations, which indicated to us that it might be possible to improve the system further by modifying procedures used previously. Here, we report the screening of a complex library of mutants (in E. coli) generated by cloning a mixture of synthetic oligonucleotides covering the −10 region of Pm. As a reporter gene in the screening we used bla (expressing β-lactamase). This system has the advantage that increased expression from Pm correlates with enhanced ampicillin resistance of the corresponding cells, as described previously (1, 34). Thus, low-frequency mutants could easily be identified by plating large numbers of cells from the libraries on agar medium with various concentrations of ampicillin. Here, we found that about a 14-fold improvement of β-lactamase production could be achieved, and further studies showed that the same mutants also gave rise to enhanced levels of expression of genes other than bla. Finally, one such Pm mutant was found to lead to a doubling (relative to wild-type Pm) of the production of a human protein (GM-CSF) under HCDC in a fermentor.


Bacterial strains, plasmids, and growth medium.

The bacterial strain and plasmids used in this study are described in Table Table1.1. In all experiments, cells were grown in Luria-Bertani (LB) broth (10 g/liter tryptone, 5 g/liter yeast extract, and 5 g/liter NaCl) or on LB agar at 37°C, except in expression studies where 30°C was used. Kanamycin was used for plasmid selection at 50 μg/ml. Ampicillin was used in expression studies, and concentrations are reported with the results. For induction of the Pm-xylS promoter system, m-toluate was used at a concentration of 2 mM.

Bacterial strains and plasmids

DNA manipulations.

Plasmid DNA was prepared by using a WizardPlus SV minipreps DNA purification kit (Promega). Transformations of E. coli were performed by use of heat shock-competent rubidium chloride-treated cells. Enzymatic manipulations (restriction enzyme digestions, ligations, and phosphorylations) were performed as described by the manufacturers. DNA was extracted from agarose gel slabs using a Qiaquick gel extraction kit (Qiagen). PCRs were performed using an Expand High Fidelity PCR system kit (Roche) for cloning purposes. When PCR was used for the generation of templates for DNA sequencing, the polymerase DynazymeII (Finnzymes) was used. PCR templates were treated with the enzyme mixture ExoSapIt (USB) prior to DNA sequencing. Sequencing reactions were carried out using an ABI Prism BigDye sequencing kit (Applied Biosystems) and analyzed using an ABI 3100 Genetic Analyzer (Applied Biosystems). DNA sequences for primers used in PCR amplifications and DNA sequencing are given in Table Table22.

Primers used in genetic modifications

A down-mutation in the mRNA leader sequence of Pm previously identified by Winther-Larsen et al. (34) was introduced into pIB3 using two annealed, complementary oligonucleotides with the mutated Pm leader mRNA sequence and SpeI- and NdeI-compatible ends. Sequences of oligonucleotides were 5′-CTAGTACAATAATAATGAAGTCATGAACA-3′ and 5′-TATGTTCATGACTTCATTATTATTGTA-3′ (introduced mutations are underlined, and restriction sites are shown in italics). The small DNA fragment was cloned into the SpeI/NdeI sites of pIB3, generating pIB4. The AflIII site upstream of the Pm promoter of pIB5 was introduced by site-specific mutagenesis of pIB4 using a QuikChange Site-Directed Mutagenesis Kit (Stratagene) as described by the manufacturer and the primers Pr-AflIII.fwd and Pr-AflIII.rev (sequences are given in Table Table2).2). The BspLU11I site of pIB6 was introduced by site-specific mutagenesis of the SpeI site of pIB5 using the primers Pr-BspLU11I.fwd and Pr-BspLU11I.rev.

Construction of the Pm mutant libraries.

It has previously been established that the ampicillin resistance level of host cells containing a plasmid-encoded β-lactamase gene (bla) is approximately proportional to the copy number of the plasmid (32). The expression increases as a function of the gene dosage, and this property can therefore also be used to estimate changes in the promoter activity. We used the plasmid pIB6 for construction of mutant libraries (Fig. (Fig.1),1), based on the vector pJT19bla previously used in studies of Pm mutants (34). In these constructs β-lactamase is a reporter of Pm activity. Kanamycin resistance (Kmr) allowed for plasmid selection without the involvement of Pm. The expression level of this expression system was lowered by introducing a mutation in the ribosome binding site. This mutation (GGAG → GAAG) has previously been found to reduce expression from Pm (34) and was introduced to avoid potential situations where expression exceeds levels detectable by ampicillin resistance.

FIG. 1.
Map of the plasmid pIB6. bla, gene encoding β-lactamase; Kmr, kanamycin resistance; trfA, gene encoding the essential replication protein TrfA; xylS, gene encoding the activator XylS; oriV, origin of vegetative replication; oriT, origin of transfer; ...

To introduce mutations in the Pm region, we used a strategy involving synthetic oligonucleotides, similar to the protocol described by Winther-Larsen et al. (34). Synthetic oligonucleotides were designed to constitute a double-stranded DNA fragment with the Pm sequence and XbaI- and BspLU11I-compatible ends when annealed for subsequent easy cloning into the pIB6 vector. A schematic illustration of the process is given in Fig. Fig.2.2. For the first mutant library produced (library 1), one of the oligonucleotides corresponded to the wild-type Pm sequence (5′-CATGTTGCATAAAGCCTAAGGGGTAGGCCTTT-3′). The oligonucleotide corresponding to the complementary strand was randomly mutagenized by the use of a oligonucleotide mixture (5′-CTAGAA133224122224413324441432AA-3′, where the numbers in the oligonucleotide indicate the doping percentages of the nucleotides: 1 is 88% A, 4% C, 4% G, and 4% T; 2 is 88% C, 4% A, 4% G, and 4% T; 3 is 88% G, 4% A, 4% C, and 4% T; and 4 is 88% T, 4% A, 4% C, and 4% G).

FIG. 2.
Schematic presentation of the construction of a Pm mutant library.

Wild-type and mutated oligonucleotides were mixed (0.7 mM each) and phosphorylated using polynucleotide kinase. NaCl was added to a final concentration of 200 mM, and the phosphorylated oligonucleotides were annealed by gradual cooling from 95 to 20°C during 20 min in a PCR machine. Dilutions of the resulting DNA fragments were ligated into XbaI-BspLU11I-digested plasmid pIB6, which had been purified by a Qiaquick PCR purification kit and dephosphorylated by calf intestinal phosphatase. The ligated plasmids were transformed into E. coli DH5α using Kmr as a selection marker. The approximately 500,000 transformants obtained were mixed and used as a library of Pm mutants. A second mutant library (library 2) was created by the same strategy, but the sequences of the oligonucleotides were based on the Pm high-level expression mutant ML1-18 (Fig. 3A and B) identified from library 1, using the oligonucleotides 5′-CATGTGCTAAAGTTTAAGGGGTAGGCCTTT-3′ and 5′-CTAGAA1332241222244111244413CA-3′ (where 1,2, 3, and 4 are the doping numbers as described above). The number of transformants in library 2 was estimated to be approximately 300,000.

FIG. 3.
Sequences of the −10 Pm region as determined for high-level-expression mutants identified from library 1 (A) and library 2 (B). (C) Pm sequences for mutants with high levels of uninduced expression, identified by selection on agar medium supplied ...

Screening for Pm high-level-expression mutants and determination of ampicillin resistance.

Mutants with increased expression relative to wild-type Pm were identified as described by Winther-Larsen et al. (34). The ampicillin concentration levels tolerated by the candidates identified in the screen were then determined by the following procedure. Individual colonies were inoculated with 100 μl of LB broth with or without inducer in 96-well microtiter plates (Nunc). The cells were incubated at 30°C overnight and diluted twice by a 96-pin replicator and microtiter plates with 200 μl of LB broth in each well. Subsequently, the 96-pin replicator was used to plate the cells on LB agar with or without inducer at various ampicillin concentrations. The plates were incubated at 30°C for approximately 20 h and then inspected for growth. The DNA sequence of the Pm region was determined for selected candidates using the Pr-Pm1.seq primer. Sequencing templates were generated by PCR using the primers Pr-Pm1.fwd and Pr-Pm1.rev. All Pm mutants reported under Results and Discussion were reproduced by one of the following strategies. For mutants in which the BspLU11I site was intact, oligonucleotides analogous to those used for constructing the libraries were designed. However, the sequences of the oligonucleotides corresponded to the Pm sequences of the mutant to be reproduced. The oligonucleotides were phosphorylated, annealed, and cloned into the XbaI/BspLU11I sites of pIB6 as described under “Construction of the Pm mutant libraries.” Finally, the Pm region of transformants was sequenced as described above. For mutants in which deletions had disrupted the BspLU11I site, the mutated Pm region was PCR amplified using the primers Pr-Pm2.fwd and Pr-Pm2.rev. The resulting products were purified using spin columns (Qiagene), cut by the restriction enzymes AflIII/NdeI, and recloned into the same sites of pIB6. The sequences of the Pm region of transformants were determined as described above, but another sequencing primer was used (Pr-Pm2.seq) to obtain the DNA sequence of the complete PCR-amplified region.

Enzyme assays.

E. coli cells with the relevant plasmids were diluted 100-fold from an overnight culture grown in selective medium at 37°C. At an optical density at 600 nm of 0.1, the cells were induced by m-toluate at a concentration of 2 mM. The cells were further grown at 30°C for 5 h. Aliquots of cells were snap-frozen on dry ice and ethanol and stored at −80°C for subsequent enzyme assays. The β-lactamase assay was performed essentially as described by Winther-Larsen et al. (34). Luciferase activities were measured as described by Blatny et al. (2), using the luciferase assay system from Promega and a TD-20/20 luminometer (Turner Design). Phosphoglucomutase activities were measured as described by Fjærvik et al. (10). For all enzyme assays, measurements were carried out with three recurrences for each sample. All enzyme activity analyses were repeated at least twice with enzyme extracts obtained from independently grown cultures.

RNA isolation, cDNA synthesis, and real-time PCR.

E. coli cells with the relevant plasmids were grown as described for the enzyme assays. For stabilization of the RNA, cell cultures were treated with RNAprotect (Qiagen) before the cell pellets were frozen for subsequent RNA isolation. RNA was isolated using an RNAqueous kit (Ambion) as described by the manufacturer. The concentration and purity of the RNA were examined by determining the absorbance at 260 and 280 nm in a Lambda 35 UV/VIS spectrometer (Perkin Elmer). The RNA preparations were treated with DNase (DNA-free; Ambion) to remove any contaminating DNA. cDNA was produced using a First-Strand cDNA synthesis kit (Amersham Biosciences) with random pd(N)6 primers as described by the suppliers.

Real-time PCR was used for quantification of bla, luc (the luciferase-encoding gene), and celB (the phosphoglucomutase-encoding gene) transcripts. Primers were designed using the primer design program of Clone Manager, version 6.0 (Scientific and Educational Software) to give products of approximately 250 bp. For amplification of the bla fragment, the primer pair Pr-bla-rt.fwd and Pr-bla-rt.rev was used; primers Pr-luc-rt.fwd and Pr-luc-rt.rev were used for luc, and Pr-celB-rt.fwd and Pr-celb-rt.rev were used for celB. The iTaqTM SYBR Green Supermix (Bio-Rad) was used for the real-time PCRs, and the reactions were carried out in a MX3000 instrument (Stratagene). Optimal primer concentrations were determined for each primer set and varied from 160 to 240 nM. Amplification for each sample was carried out in triplicate wells. The PCR cycles were as follows: 10 min at 96°C, followed by 40 cycles consisting of 30 s at 95°C, 60 s at 55°C, and 30 s at 72°C. Relative quantities were determined using the software of MX3000 (Stratagene). cDNA produced from cells carrying plasmids with wild-type Pm was used as a calibrator, with pIB6, pOY9, and pLB10 used for bla, luc, and celB, respectively. A fragment from the kanamycin resistance gene, amplified using the primer pair Pr-km-rt.fwd and Pr-km-rt.rev, was used as a normalizer. All real-time PCR experiments were repeated at least twice with cDNA samples obtained from independently grown cell cultures.

Primer extension.

Primer extension experiments were performed to determine transcription start sites. E. coli cells with the relevant plasmids were diluted 50-fold from an overnight culture grown in selective medium. After growth at 37°C for 90 min, the cultures were induced to a final concentration of 4 mM m-toluate and grown at 30°C for an additional 3.5 h. Cell cultures were treated with RNAprotect (Qiagen), and cell pellets were frozen. RNA was isolated using RNAqueous (Ambion) and subsequently precipitated with ethanol and sodium acetate. To avoid nonenzymatic degradation of RNA in the concentrated RNA preparations, Na2-EDTA was added to a final concentration of 0.05 mM. Approximately 15 μg of total RNA and 2.5 pmol of the bla-specific primer Pr-ext-bla with a fluorescent label (FAM [6-carboxyfluorescein]) at the 5′ end were denatured by heating at 80°C for 1 min and then chilled on ice. The cDNA synthesis was carried out using the primer extension system with avian myeloblastosis virus reverse transcriptase (Promega) as described by the manufacturer. For production of sequencing ladders, a region of bla was amplified using the primers Pr-bla.fwd and Pr-bla.rev. DNA sequencing reactions were then performed using a Sequenase, version 2.0, DNA sequencing kit (USB) and the same FAM-labeled primer as in the primer extension reaction. Prior to gel electrophoresis, the primer extension reaction mixtures were precipitated with ethanol and sodium acetate. Gel electrophoresis was performed using an ABI Prism 377 Sequencer (Applied Biosystems). The scanning tool of the software was used to obtain exact positions on the gel for the primer extension and sequencing signals. The signals could thus be precisely mapped relative to the sequencing ladders.

Production analyses in HCDC.

HCDC of E. coli RV308 cells harboring pMV1, pMV2, and pMV3, as well as the subsequent product analyses, i.e., Western blotting and enzyme-linked immunosorbent assay, were performed as described in Sletta et al. (31). All strains tested were run in two fermentations. The values for production are average numbers for the parallel fermentations.


Construction of a Pm mutant library and establishment of an efficient screening strategy.

A Pm mutant library (library 1) was constructed in plasmid pIB6 by introducing random mutations in a 24-bp region spanning the −10 element of Pm (Fig. (Fig.1).1). The bla gene was used as a reporter for Pm, and therefore alterations in host ampicillin tolerance reflected changes in the Pm promoter activity. Cells from the library were plated on agar medium containing m-toluate as an inducer and various concentrations of ampicillin. The results demonstrate that a fraction of the cells could grow at elevated concentrations of the antibiotic (Fig. (Fig.4),4), indicating that the library contains Pm mutants that strongly stimulate bla expression. The small fraction of cells with elevated ampicillin tolerance in the wild-type population (Fig. (Fig.4)4) could be due to spontaneous mutations, presumably selected during slow growth on the agar medium. Typically, these colonies had a mucoid phenotype, indicating production of an exopolysaccharide, which may be cholanic acid (27). The production of such exopolysaccharides is generally known to lead to increased antibiotic tolerance. Spontaneous mutations could therefore potentially contribute to enhanced ampicillin resistances in addition to the introduced Pm mutations. To eliminate such effects, all Pm mutants described here were reproduced by resynthesizing the mutant oligonucleotides, followed by cloning in pIB6 prior to phenotypic characterizations.

FIG. 4.
Fraction of cells with increased ampicillin resistance in Pm mutant libraries 1 and 2 and for a culture of cells containing unmutagenized plasmid pIB6. Cell cultures with mutant libraries were grown in LB broth for 1 h at 37°C, induced with m ...

The Pm DNA sequence was determined for about 20 randomly picked clones from the library. As expected, since only one of the DNA strands was mutagenized, around 50% of the cells exhibited the wild-type Pm sequence. Two of the clones were found to have multiple inserts of the mutagenized DNA fragment. For the remaining clones, the average number of mutations was found to be 2.4 (sequence data not shown).

Identification of Pm mutants with up to 10-fold elevated expression of the bla reporter gene.

Mutants with putatively increased Pm activity (up-mutants) were identified by selection on agar medium supplemented with m-toluate and ampicillin at concentrations ranging from 0.5 to 7 mg/ml. Pm sequences were determined for about 100 such candidates. The results showed that some mutant Pm sequences were present more than once. To make sure that the phenotypes were caused by the observed mutations, the Pm region was resynthesized for 18 of the mutants and used to replace the corresponding sequence in pIB6 (wild type). The sequences and phenotypes of all these mutants are shown in Fig. Fig.3A.3A. While cells harboring pIB6 (wild-type Pm) tolerated up to 0.5 mg/ml ampicillin under induced conditions, cells carrying pIB6 with mutated Pm sequences grew up to 3.5 mg/ml of the antibiotic. Compared to randomly picked mutants (see above), the average number of Pm point mutations was high (approximately 5). All these mutants also carried deletions in or close to the Pm transcription initiation region, a feature not found in the randomly picked mutants. Coincidentally, we identified two up-mutants, carrying the Pm sequences ML1-1 and ML1-2 (Fig. (Fig.3A),3A), that were changed relative to the wild-type Pm sequence only by deletions of one or two nucleotides in the transcription initiation region. These particular nucleotide positions were not intentionally mutagenized or deleted, meaning either that the deletions were generated during cloning of the synthetic oligonucleotides or that they may have existed in the original oligonucleotide mixture prior to cloning. Comparison of the ampicillin resistance data for all 18 mutants indicates that the ML1-1 and ML1-2 deletions contribute to somewhat less than half of the maximum increment in Pm activity observed.

The nucleotide substitutions in the Pm up-mutants covered most of the mutagenized region, and a majority was positioned upstream of the −10 region. Generally, with respect to the spacer between the −10 and −35 elements of bacterial promoters, the focus has been on the importance of the length rather than nucleotide sequence (23, 36). It has been shown, however, that the nucleotide sequence of the spacer can influence promoter strength (14, 17, 21), which is consistent with the results obtained here. The −10 element of Pm has experimentally been determined to be the hexamer TAGGCT (9), corresponding to nucleotide positions −6 to −11 in Fig. Fig.3A.3A. In the same study, positions −6, −10, and −11 were identified as crucial for both σ32 and σs recognition. Interestingly, none of these positions has been affected by the mutations reported in Fig. Fig.3A3A.

Domínguez et al. (9) further established that positions −7 and −8 (G and C, respectively) and the C stretch upstream of the −10 element are critical for σ32 recognition and that the G at position −9 is important for σs recognition. Whereas position −7 is conserved among all the Pm up-mutants identified here (except for mutant ML2-2) (Fig. (Fig.3C),3C), substitutions at position −8, −9, and in the C stretch occur frequently among mutants. Thus, these mutations may have affected σ selectivity, although in such a way that high level expression was still obtained. Further studies using E. coli strains with deficient σ factors might help in elucidating the role of σ selectivity in determination of Pm mutant phenotypes.

For mutants carrying pIB6 with the Pm up-mutant sequences ML1-3, ML1-11, ML1-18, and ML1-16, the expression levels were further quantified by determining the amounts of β-lactamase transcript (relative to expression in cells carrying pIB6 with wild-type Pm) by a reverse transcriptase real-time PCR strategy. The results are shown in Fig. Fig.5A.5A. These experiments corroborated the ampicillin resistance data and showed that a large increase in expression was obtained. The expression levels for these mutants relative to wild-type Pm were further investigated by determining β-lactamase activities. Results for the mutants relative to the wild-type Pm construct are shown in Fig. Fig.5A5A and demonstrate up to a 10-fold stimulation of expression at the enzyme activity level (mutant ML1-18).

FIG. 5.
Relative expression as determined from transcript quantitation and enzyme activity of β-lactamase (A), phosphoglucomutase (B), and luciferase (C) from selected Pm mutants. The values are the averages of two biological recurrences, and the error ...

Construction and screening of a second mutant library yielded further improved Pm mutants.

One advantage of the mutagenesis strategy used in this study appeared to be that it allowed the accumulation of multiple stimulatory mutations in a single step, but we still considered it possible that further improved Pm up-mutants might be achieved by introducing additional mutations to the up-mutants identified in the screening of library 1. Mutant ML1-18 was chosen as a candidate for such a study, and a second Pm mutant library (library 2) was constructed on the basis of the Pm sequence of this mutant. As for library 1, the fraction of cells growing at elevated ampicillin concentrations on agar plates with m-toluate was determined. The results indicated that there exist mutants in library 2 that exhibit considerably higher ampicillin resistances than the best mutants selected from library 1 (Fig. (Fig.4).4). For selected mutants growing at elevated ampicillin concentrations in the presence of m-toluate, Pm sequences were determined. These mutant sequences were resynthesized and used to replace the corresponding region in pIB6. The Pm sequences and ampicillin resistance data for 11 such mutants are shown in Fig. Fig.3B.3B. Ampicillin resistance levels reached up to 4.5 mg/ml, compared to 3.5 mg/ml for library 1 (Pm sequence ML1-18), clearly indicating that a further improvement has been reached. Two of the new Pm up-mutant sequences (ML2-1 and ML2-10) exhibit a TATACT element, corresponding to the −10 hexamer in both the consensus [TG(N)0-2CYATNCT] and the optimized [TGTGCTATA(C/A)T] −10 elements for σs-dependent promoters (11, 15). Also the Pm sequence ML1-4 identified from library 1 exhibits this hexamer. These three Pm mutants are characterized by high basal expression levels (Fig. (Fig.3,3, ampicillin resistance data), implying that the promoter element can be recognized by the RNA polymerase holoenzyme even in the absence of activated XylS. However, these mutants are also stimulated at induction.

For mutants carrying Pm sequences ML2-1, ML2-2, and ML2-5, the expression from Pm was further investigated by determination of relative amounts of β-lactamase transcript and the corresponding enzyme activities (as described above for up-mutants identified from library 1). The results demonstrate that an increase of up to about 14-fold in Pm activity was obtained after two rounds of mutagenesis (Fig. (Fig.5A5A).

Pm mutants displaying very high basal expression levels could be identified.

The basal expression level, determined as the ampicillin resistance level during growth on agar medium in the absence of inducer, varied extensively among Pm up-mutants from 0.005 to 0.8 mg/ml ampicillin (Fig. (Fig.3A).3A). Apparently, the ratio of the induced to uninduced expression level varies drastically among mutants, and this was further substantiated by directly screening library 1 for mutants displaying increased expression levels in the absence of inducer. Four such mutants were further characterized, all displaying a higher uninduced resistance level toward ampicillin than the corresponding induced level of the wild type. One of these mutants turned out to be identical to ML1-4 (Fig. (Fig.3A),3A), while the remaining three (ML1-19, ML1-20, and ML1-21) were new and had in common that a TG had been introduced in the 5′ proximity (upstream) of the −10 element (Fig. (Fig.3C).3C). TG motifs upstream of the −10 elements (the “extended −10 element”) are known to contribute to the activity of a large number of E. coli promoters (22). Further, TG motifs were found in an optimized −10 promoter element for expression mediated by σS [TGTGCTATA(C/A)T], identified by using a SELEX (systematic evolution of ligands by exponential enrichment) strategy (11). Interestingly, the same TG motif was also found in mutant ML1-5 (Fig. (Fig.3A),3A), but the uninduced expression level of this mutant is not as high as for the mutants ML1-19, ML1-20, and ML1-21. This means that mutations other than the TG contribute to the phenotypes of these mutants.

Basal expression from wild-type and mutant Pm sequences is independent of XylS.

To investigate whether the basal expression (in the absence of inducer) from wild-type Pm and the Pm up-mutants is dependent on XylS, we deleted the xylS gene from pIB6 (resulting in pIB17; wild-type Pm) and from the pIB6 constructs with mutated Pm sequences displaying various basal expression levels (ML1-19, ML1-4, ML1-18, ML1-16, and ML2-1). The uninduced expression levels, measured as ampicillin resistance of the corresponding cells on agar medium (ampicillin concentrations varying from 0.001 to 1.2 mg/ml), were then determined. No changes in the basal ampicillin resistance level either for wild-type Pm or for the mutants were observed when xylS was deleted, indicating that the basal expression is independent of XylS under these conditions.

Another interesting observation made in these experiments is that the inducer seems to have an easily detectable effect on ampicillin tolerance levels of the cells, irrespective of Pm-xylS. This effect is not specifically related to the RK2 replicon since we also found that DH5α cells containing plasmid pBR322 tolerated about 2.5 times more ampicillin on agar medium in the presence of the XylS inducer than cells without the plasmid. Therefore, for some Pm up-mutants with a very low ratio of induced to uninduced expression (like ML1-19), most of the increased expression upon addition of inducer is probably caused by this type of effect and not by altered transcription from Pm. Hydrocarbons like toluene and xylenes are known to cause damage to bacterial membranes (29), and Domínguez-Cuevas et al. (8) found that hydrocarbons like toluene, o-xylene, and, to some extent, m-toluate activate genes involved in the heat shock response in P. putida. Furthermore, the acrAB-TolC multidrug efflux pump system in E. coli is known to mediate transport of ß-lactams, and transcription of acrAB has been shown to be increased by general stress conditions (18). The addition of m-toluate to the cell cultures could therefore lead to somewhat increased ampicillin resistance of the cells through a stress response which acts independently of Pm activation. However, for wild-type Pm and most mutants reported here, this effect contributes very little to the total resistance observed upon XylS-mediated induction.

Pm mutants showing increased expression levels typically have altered transcriptional start sites.

All the Pm up-mutants were characterized by deletions around the Pm transcription initiation region (Fig. (Fig.3),3), and it seemed plausible to believe that these deletions could affect the transcriptional start site. For cells with wild-type Pm (pIB6) and for the mutant sequences ML1-13, ML1-4, ML1-18, and ML1-16, the transcriptional start sites were experimentally determined by primer extension analysis (Fig. (Fig.6).6). Transcription from wild-type Pm was found to initiate from a C, while for the mutants with the sequences ML1-13, ML1-16, and ML1-18, the transcriptional start site was changed relative to wild-type Pm (Fig. 6A and B). For ML1-4 the start site was found to be identical to that of the wild type, but the first downstream nucleotides were changed due to deletions (Fig. (Fig.6B).6B). It has previously been reported that the efficiency of the transcription initiation process can be influenced by the type of nucleotide at the transcriptional start site (16, 33). Therefore, it is plausible to assume that the change of start site for the Pm up-mutants could also contribute to the increased expression.

FIG. 6.
Transcription start sites for the β-lactamase transcript from cells with wild-type Pm (pIB6) and Pm mutants in pIB6. (A) Gel image showing the mapping of transcriptional start sites. The DNA sequences (right) are complementary to the sequences ...

Pm mutants also confer improved expression of two alternative reporter genes.

It is well known that the efficiency of recombinant gene expression for a specific vector system varies for different genes. The celB (encodes phosphoglucomutase) (6) and luc (encodes firefly luciferase) (7) genes were inserted as reporters for selected Pm up-mutant sequences (ML1-18, ML1-16, ML2-1, ML2-2, and ML2-5). The levels of recombinant transcript and concomitant enzyme activities were compared to cells with wild-type Pm (pLB10 and pOY9 with celB and luc, respectively) for both reporter genes (Fig. 5B and C). For celB the relative amounts of transcript and enzyme activities were comparable to those found for β-lactamase for the same Pm sequences (5- to 7-fold increase for mutants from library 1 and up to a 10-fold increase for mutants from library 2). For luc the increment in expression was four- to fivefold relative to wild-type Pm for mutants selected from library 2. These results show that the mutant Pm sequences screened out for high-level expression with bla as a reporter also support strongly stimulated expression of other reporter genes. This means that the Pm up-mutants represent generally useful new tools for the Pm-xylS expression system.

Pm up-mutants substantially improve GM-CSF production under HCDC.

We previously showed that the Pm-xylS system can be useful to produce high volumetric yields of several industrially and medically important human-derived proteins in E. coli under HCDC (30, 31). For example, by extensive genetic and process optimizations, up to 0.8 g/liter of recombinant GM-CSF was obtained. To investigate whether Pm up-mutants identified in the present study could be used to further improve the production of GM-CSF, three of the high-level-expression Pm sequences (ML1-18, ML1-16, and ML1-14) were examined. The wild-type Pm sequence in plasmid pTA52 was exchanged with the Pm sequences ML1-18, ML1-16, and ML1-14 to generate pMV1, pMV2, and pMV3, respectively. Plasmid pTA52 is identical to the GM-CSF expression vector pGM29pelB (31) except for two new restriction sites introduced for cloning purposes (see Materials and Methods). The three new GM-CSF expression vectors were transformed into production host E. coli RV308, and the resulting recombinant strains were subjected to production analyses under HCDC together with pGM29pelB, as described previously (30, 31). The results of these experiments showed that all three strains produced GM-CSF at significantly higher levels (up to 1.6 g/liter of GM-CSF) than the cells harboring the parental vector (about 0.6 g/liter) (Table (Table3).3). The increments were predominantly found in the insoluble fractions, suggesting that the translocation machinery has been saturated under these conditions. Interestingly, the ranking of the three Pm up-mutants here was different from that obtained with bla as a reporter gene (Fig. (Fig.3).3). The biological reason for this is unknown but may be related to changes in cell physiology caused by growing the cells under HCDC instead of shake flask cultivations. Nevertheless, these results clearly demonstrated that the Pm up-mutants can be useful to improve the production level of a human protein already expressed at industrial levels under HCDC, as long as maximized induced expression levels of the recombinant proteins are desired.

Production yields of soluble and insoluble recombinant GM-CSF using the Pm wild type and three selected Pm up-mutants

Concluding remarks.

The results reported in this paper demonstrate that there is a significant potential for use of random mutagenesis of an already efficient promoter system to increase recombinant protein production both under laboratory-scale and industry-simulated conditions. It seems clear that the success of such a strategy depends on at least three parameters that must be simultaneously in place. First, since the mutants occur at low frequencies, the library size is critical, and a strong selection scheme is required. Furthermore, multiple mutations should be present in each oligonucleotide cloned. Since deletions seemed to play an important role, the procedures may be improved even further by also deliberately introducing such changes during oligonucleotide synthesis. The random nature of the mutations introduced in the Pm up-mutants illustrates that the relation between nucleotide sequence and promoter activity is complex and does not involve only clearly defined functional elements. It seems unlikely that mutants with similar properties could be obtained by rational design. The strategy described here would probably also function for most other promoter systems used for high-level expression of recombinant proteins.


[down-pointing small open triangle]Published ahead of print on 5 February 2009.


1. Berg, L., R. Lale, I. Bakke, and S. Valla. The expression of recombinant genes in Escherichia coli can be strongly stimulated at the transcript production level by mutating the DNA-region corresponding to the 5′-untranslated part of mRNA. Microb. Biotechnol, in press. [PMC free article] [PubMed]
2. Blatny, J. M., T. Brautaset, H. C. Winther-Larsen, K. Haugan, and S. Valla. 1997. Construction and use of a versatile set of broad-host-range cloning and expression vectors based on the RK2 replicon. Appl. Environ. Microbiol. 63:370-379. [PMC free article] [PubMed]
3. Blatny, J. M., T. Brautaset, H. C. Winther-Larsen, P. Karunakaran, and S. Valla. 1997. Improved broad-host-range RK2 vectors useful for high and low regulated gene expression levels in gram-negative bacteria. Plasmid 38:35-51. [PubMed]
4. Brautaset, T., R. Lale, and S. Valla. 2009. Positively regulated bacterial expression systems. Rev. Microb. Biotechnol. 2:15-30. [PMC free article] [PubMed]
5. Brautaset, T., S. B. Petersen, and S. Valla. 2000. In vitro determined kinetic properties of mutant phosphoglucomutases and their effect on sugar catabolism in Escherichia coli. Metab. Eng. 2:104-114. [PubMed]
6. Brautaset, T., R. Standal, E. Fjaervik, and S. Valla. 1994. Nucleotide sequence and expression analysis of the Acetobacter xylinum phosphoglucomutase gene. Microbiology 140:1183-1188. [PubMed]
7. de Wet, J. R., K. V. Wood, D. R. Helinski, and M. DeLuca. 1985. Cloning of firefly luciferase cDNA and the expression of active luciferase in Escherichia coli. Proc. Natl. Acad. Sci., USA 82:7870-7873. [PMC free article] [PubMed]
8. Domínguez-Cuevas, P., J. E. González-Pastor, S. Marqués, J. L. Ramos, and V. de Lorenzo. 2006. Transcriptional tradeoff between metabolic and stress-response programs in Pseudomonas putida KT2440 cells exposed to toluene. J. Biol. Chem. 281:11981-11991. [PubMed]
9. Domínguez-Cuevas, P., P. Marín, J. L. Ramos, and S. Marqués. 2005. RNA polymerase holoenzymes can share a single transcription start site for the Pm promoter: critical nucleotides in the −7 to −18 region are needed to select between RNA polymerase with σ38 or σ32. J. Biol. Chem. 280:41315-41323. [PubMed]
10. Fjærvik, E., K. Frydenlund, S. Valla, Y. Huggriat, and M. Benziman. 1991. Complementation of cellulose-negative mutants of Acetobacter xylinum by the cloned structural gene for phosphoglucomutase. FEMS Microbiol. Lett. 77:325-330.
11. Gaal, T., W. Ross, S. T. Estrem, L. H. Nguyen, R. R. Burgess, and R. L. Gourse. 2001. Promoter recognition and discrimination by Eσs RNA polymerase. Mol. Microbiol. 42:939-954. [PubMed]
12. González-Pérez, M. M., S. Marqués, P. Domínguez-Cuevas, and J. L. Ramos. 2002. XylS activator and RNA polymerase binding sites at the Pm promoter overlap. FEBS Lett. 519:117-122. [PubMed]
13. Inouye, S., A. Nakazawa, and T. Nakazawa. 1981. Molecular cloning of gene xylS of the TOL plasmid: evidence for positive regulation of the xylDEGF operon by xylS. J. Bacteriol. 148:413-418. [PMC free article] [PubMed]
14. Jensen, P. R., and K. Hammer. 1998. The sequence of spacers between the consensus sequences modulates the strength of prokaryotic promoters. Appl. Environ. Microbiol. 64:82-87. [PMC free article] [PubMed]
15. Lacour, S., and P. Landini. 2004. σs-Dependent gene expression at the onset of stationary phase in Escherichia coli: function of σs-dependent genes and identification of their promoter sequences. J. Bacteriol. 186:7186-7195. [PMC free article] [PubMed]
16. Lewis, D. E. A., and S. Adhya. 2004. Axiom of determining transcription start points by RNA polymerase in Escherichia coli. Mol. Microbiol. 54:692-701. [PubMed]
17. Liu, M., M. Tolstorukov, V. Zhurkin, S. Garges, and S. Adhya. 2004. A mutant spacer region between −35 and −10 elements makes the Plac promoter hyperactive and cAMP receptor protein-independent. Proc. Natl. Acad. Sci. USA 101:6911-6916. [PMC free article] [PubMed]
18. Ma, D., D. N. Cook, M. Alberti, N. G. Pon, H. Nikaido, and J. E. Hearst. 1995. Genes acrA and acrB encode a stress-induced efflux system of Escherichia coli. Mol. Microbiol. 16:45-55. [PubMed]
19. Marqués, S., M. Manzanera, M. M. González-Pérez, M. T. Gallegos, and J. L. Ramos. 1999. The XylS-dependent Pm promoter is transcribed with σ32 or σ38 depending on the growth phase. Mol. Microbiol. 31:1105-1113. [PubMed]
20. Mermod, N., J. L. Ramos, P. R. Lehrbach, and K. N. Timmis. 1986. Vector for regulated expression of cloned genes in a wide range of gram-negative bacteria. J. Bacteriol. 167:447-454. [PMC free article] [PubMed]
21. Miksch, G., F. Bettenworth, K. Friehs, E. Flaschel, A. Saalbach, T. Twellmann, and T. W. Nattkemper. 2005. Libraries of synthetic stationary-phase and stress promoters as a tool for fine-tuning of expression of recombinant proteins in Escherichia coli. J. Biotechnol. 120:25-37. [PubMed]
22. Mitchell, J. E., D. Zheng, S. J. W. Busby, and S. D. Minchin. 2003. Identification and analysis of “extended −10” promoters in Escherichia coli. Nucleic Acids Res. 31:4689-4695. [PMC free article] [PubMed]
23. Murakami, K. S., S. Masuda, E. A. Campbell, O. Muzzin, and S. A. Darst. 2002. Structural basis of transcription initiation: an RNA polymerase holoenzyme-DNA complex. Science 296:1285-1290. [PubMed]
24. Ramos, J. L., M. Gonzalez-Carrero, and K. Timmis. 1988. Broad-host range expression vectors containing manipulated meta-cleavage pathway regulatory elements of the TOL plasmid. FEBS Lett. 226:241-246. [PubMed]
25. Ramos, J. L., S. Marqués, and K. N. Timmis. 1997. Transcriptional control of the Pseudomonas TOL plasmid catabolic operons is achieved through an interplay of host factors and plasmid-encoded regulators. Annu. Rev. Microbiol. 51:341-372. [PubMed]
26. Ramos, J. L., A. Stolz, W. Reineke, and K. N. Timmis. 1986. Altered effector specificities in regulators of gene expression: TOL plasmid xylS mutants and their use to engineer expansion of the range of aromatics degraded by bacteria. Proc. Natl. Acad. Sci. USA 83:8467-8471. [PMC free article] [PubMed]
27. Sailer, F. C., B. M. Meberg, and K. D. Young. 2003. β-Lactam induction of colanic acid gene expression in Escherichia coli. FEMS Microbiol. Lett. 226:245-249. [PubMed]
28. Santos, P. M., I. D. Bartolo, J. M. Blatny, E. Zennaro, and S. Valla. 2001. New broad-host-range promoter probe vectors based on the plasmid RK2 replicon. FEMS Microbiol. Lett. 195:91-96. [PubMed]
29. Sikkema, J., J. A. de Bont, and B. Poolman. 1995. Mechanisms of membrane toxicity of hydrocarbons. Microbiol. Rev. 59:201-222. [PMC free article] [PubMed]
30. Sletta, H., A. Nedal, T. E. V. Aune, H. Hellebust, S. Hakvåg, R. Aune, T. E. Ellingsen, S. Valla, and T. Brautaset. 2004. Broad-host-range plasmid pJB658 can be used for industrial level production of a secreted host-toxic single-chain antibody fragment in Escherichia coli. Appl. Environ. Microbiol. 70:7033-7039. [PMC free article] [PubMed]
31. Sletta, H., A. Tøndervik, S. Hakvåg, T. E. Vee Aune, A. Nedal, R. Aune, G. Evensen, S. Valla, T. E. Ellingsen, and T. Brautaset. 2007. The presence of N-terminal secretion signal sequences leads to strong stimulation of the total expression levels of three tested medically important proteins during high-cell-density cultivations of Escherichia coli. Appl. Environ. Microbiol. 73:906-912. [PMC free article] [PubMed]
32. Uhlin, B. E., and K. Nordstrom. 1977. R plasmid gene dosage effects in Escherichia coli K-12: copy mutants of the R plasmid R1drd-19. Plasmid 1:1-7. [PubMed]
33. Walker, K. A., and R. Osuna. 2002. Factors affecting start site selection at the Escherichia coli fis promoter. J. Bacteriol. 184:4783-4791. [PMC free article] [PubMed]
34. Winther-Larsen, H. C., J. M. Blatny, B. Valand, T. Brautaset, and S. Valla. 2000. Pm promoter expression mutants and their use in broad-host-range RK2 plasmid vectors. Metab. Eng. 2:92-103. [PubMed]
35. Winther-Larsen, H. C., K. D. Josefsen, T. Brautaset, and S. Valla. 2000. Parameters affecting gene expression from the Pm promoter in gram-negative bacteria. Metab. Eng. 2:79-91. [PubMed]
36. Young, B. A., and T. M. Gruber. 2002. Views on transcription initiation. Cell 109:417-420. [PubMed]

Articles from Applied and Environmental Microbiology are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...