Homogeneous expression of the PBAD promoter in Escherichia coli by constitutive expression of the low-affinity high-capacity AraE transporter

Since the first report of its use as an inducible expression system, the arabinose-inducible araBAD promoter (P_BAD) and regulator (AraC) of Escherichia coli have become popular for controlled gene expression in E. coli and other bacteria (Guzman et al., 1995 ). Whether this system is used for control of a single gene or in conjunction with another inducible promoter for control of several genes, the araCP_BAD system offers regulatable control of gene expression in the presence of inducer and tight control in the absence of inducer. This weak, yet tightly controlled promoterregulator system has enabled simultaneous growth and production of both soluble and insoluble heterologous proteins in high-cell-density E. coli cultures (Lim et al., 2000 ).

Recently, the araCP_BAD system has been introduced into both Gram-positive and Gram-negative bacterial hosts (Ben-Samoun et al., 1999 ; Newman & Fuqua, 1999 ; Sukchawalita et al., 1999 ). In Agrobacterium tumefaciens, the level of control afforded is significant, although less stringent than that observed in E. coli (Newman & Fuqua, 1999 ). In Corynebacterium glutamicum, P_BAD requires both arabinose and araC, indicating that E. coli AraC is capable of interacting with C. glutamicum RNA polymerase to induce transcription, but not the CRP protein, as it does in E. coli (Ben-Samoun et al., 1999 ). Given this high degree of flexibility, this broad-host-range promoter is attractive for genetic and metabolic engineering of several different bacteria, since one could make a single genetic construct and use it in several organisms.

Unfortunately, the araCP_BAD system and the associated high-capacity, low-affinity L-arabinose transporter AraE display autocatalytic behaviour and suffer from all-or-none expression in E. coli (Siegele & Hu, 1997 ). Rather than varying the level of gene expression in individual cells of the culture, the concentration of arabinose in the medium changes the fraction of cells that are fully induced. Recently, we showed that expression of araE from an arabinose-independent (IPTG-inducible) promoter allows regulatable gene expression control from P_BAD in individual cells (Khlebnikov et al., 2000 ). In this paper we show that expression of araE from constitutive promoters of various strengths on medium-copy plasmids or on the chromosome allows homogeneous expression from P_BAD and that the level of araE expression affects the level of expression from P_BAD at a given inducer concentration.

General.
Bacteria and vectors are listed in Table 1. Each of the three pCP vectors (pCP8, pCP13 and pCP18) contains a constitutive promoter of different strength from Lactococcus lactis (Jensen & Hammer, 1998a , b ). All DNA manipulations were performed in E. coli DH10B using established protocols (Sambrook et al., 1989 ) or as indicated below. PCR amplification of DNA was done using the Expand High Fidelity PCR System (Roche Molecular Biochemicals) under the conditions recommended by the manufacturer. Oligonucleotides were synthesized by Genemed Synthesis. The restriction digests and ligation reactions were performed as recommended by the restriction enzyme manufacturer (Roche Molecular Biochemicals). The ligated vectors were transformed into electrocompetent cells (E. coli DH10B or E. coli CW2587) by electroporation (field strength 18 kV cm^-1) using a Bio-Rad E. coli Pulser.

Table 1. E. coli strains and plasmids used in this study

Plasmid-borne constitutive promoters.
To construct the pJAT plasmids, an XbaIXbaI fragment (1981 bp for P_CP8 and P_CP13, 1982 bp for P_CP18) containing the corresponding constitutive promoter region and the erythromycin-resistance gene was subcloned from the pCP plasmids into the XbaI restriction site on pBluescript SK+. The resulting vectors were digested with ClaI (a site for which was located between the XbaI restriction site and the erythromycin-resistance gene on the constitutive promotererythromycin DNA fragment) and EcoRI (which cuts in the multi-cloning site of pBluescript SK+ opposite the constitutive promoter-erythromycin DNA fragment), thereby generating smaller fragments that contained the constitutive promoters and erythromycin-resistance gene (1911 bp for P_CP8 and P_CP13, 1912 bp for P_CP18) and with appropriate restriction sites at either end for cloning into plasmid pJN105. These ClaIEcoRI fragments were ligated to the 4751 bp ClaIEcoRI fragment of the broad-host-range, medium-copy-number plasmid pJN105, containing the gentamicin-resistance gene and the pBBR-1 origin of replication. The resulting plasmids were designated pJAT8, pJAT13 and pJAT18.

The araE gene was amplified from genomic DNA of E. coli W3110 using PCR and the primers for the 5'-end of the gene (5'-CGTGAATTCGTCTTACTCTCTGTCGGCAG-3') and the 3'-end of the gene (5'-CTACGATCGAACGGCCAAGTGCCCAATCT-3'), and then digested with EcoRI and PvuI. The medium-copy number vectors pJAT8, pJAT13 and pJAT18 were digested with EcoRI and PvuI, and ligated with the EcoRIPvuI PCR fragment, resulting in plasmids pJAT8araE, pJAT13araE and pJAT18araE.

Construction of strains that express araE constitutively from the chromosome.
E. coli encodes both a high-affinity arabinose transporter (encoded by the araFGH operon) and a low-affinity arabinose transporter (encoded by araE) whose synthesis is inducible by arabinose. In order to uncouple the expression of these transporters from this autocatalytic behaviour, new E. coli strains were constructed in which the araFGH genes are deleted and araE is constitutively expressed. Both of these modifications were facilitated by using Red technology (Datsenko & Wanner, 2000 ) and appropriate PCR products. The DE(araFGH) mutation was made by synthesizing a PCR product on pKD83 as template with the primers 5'-TGCACGTTCTCACTGTAATTCTGCGATGTGATA-TTG/CACGTCTTGAGCGATTGTGT-3' and 5'-GAAAAAACGCTAAATTGTTGCAGAAAAAAGCATCAG/ATTCCGGGGATCCGTCGACC-3', in which bases preceding a slash correspond to homology extensions (H1 or H2) for corresponding priming sites (P1 or P4) as shown in Fig. 1. pKD83 is similar to pKD13 (Datsenko & Wanner, 2000 ), except pKD83 has the kanamycin resistance gene (kan903) from Tn903 (K. A. Datsenko & B. L. Wanner, unpublished results). The DE(araFGH) mutation was recombined into the chromosome and verified as described in Fig. 1. Strains carrying this deletion with or without the kan903 gene are described in Table 1.

(15K): Fig. 1. Construction of the DE(araFGH) mutation. An 1163 bp PCR product containing 36 bp homology extensions (H1 and H2) to the chromosomal araFGH region (Blattner et al., 1997 ) was introduced into BW25113 carrying pKD46 as described elsewhere (Datsenko & Wanner, 2000 ). pKD46 encodes the Red recombinase from phage λ which promotes highly efficient homologous recombination. In these crosses, recombination occurs between the homologous sequences on the ends of the PCR products and the bacterial chromosome (Datsenko & Wanner, 2000 ). KmR transformants were PCR-verified by testing for presence of novel junction fragments of expected sizes when using priming (p) sites that flank the homology regions (BW27269). The kan903 gene was then eliminated by site-specific recombination by using the Flp plasmid pCP20 (Cherepanov & Wackernagel, 1995 ), as described previously (Datsenko & Wanner, 2000 ). The resultant KmS recombinant (BW27378) was similarly verified by PCR. kan903 primers: k4 (5'-ACCAGGATCTTGCCATCCTA-3') and k3 (5'-TAATCGCGGCCTCGAGCAAG-3'). araF upstream primer: 5'-GCATTCCAGCGGTTGCCATA-3'. araH downstream primer: 5'-CGGCAGATGAACGATATGTG-3'. Priming sites 1 and 4 are as described elsewhere (Datsenko & Wanner, 2000 ).

pKD85 (P_CP8araE'), pKD86 (P_CP18araE') and pKD118 (P_CP13araE') were used as templates with the primers 5'-TTTATCTGCTGTAAAATTAGGTGGTTAATAATAA-TCGTGTAGGCTGGAGCTGCTTC and 5'-ATATTCATACGCCGCGTATC-3' to recombine these fusions onto the chromosome. The latter primer corresponds to araE sequences in common with these template plasmids and the chromosome (Fig. 2). These plasmids were constructed by synthesizing approximately 850 bp PCR products on the respective pJAT plasmid with the primers 5'-TCAACTGCCTGGCACAAT-3' and 5'-TTCCGCCTCAATATGACG-3'. These PCR products were digested with XhoI and BclI to release internal promoter-containing fragments of approximately 600 bp, which were then cloned into XhoI- and BamHI-digested pKD12. The latter plasmid is similar to pKD13 (Datsenko & Wanner, 2000 ; K. A. Datsenko & B. L. Wanner, unpublished results). The P_CP8, P_CP13 and P_CP18araE' fusions were recombined into the chromosome of BW25113 by using the Red plasmid pKD46 and selecting Km^R transformants as described elsewhere (Datsenko & Wanner, 2000 ). The resultant recombinants (BW27270, BW27535 and BW27271) were verified as shown in Fig. 2. Derivatives of the E. coli K-12 strain BW25113 carrying both the DE(araFGH) mutation and a P_CP8, P_CP13 or P_CP18araE fusion were constructed by transduction with P1kc (Wanner, 1994 ), resulting in strains BW27749, BW27752 and BW27750. All transductants were similarly verified by PCR before and after Flp-mediated elimination of the kanamycin-resistance genes (BW27783, BW27786 and BW27784). DNA sequence analysis revealed that the CP18 promoter upstream region has two adjacent BamHI sites, which were apparently introduced during its original construction (Jensen & Hammer, 1998a ). Otherwise, the sequences of the P_CP8, P_CP13 and P_CP18araE fusions after recombination onto the chromosome and elimination of the resistance marker were as predicted (Fig. 3).

(16K): Fig. 2. Construction of chromosomal PCP8, PCP13 and PCP18araE fusions. The construction of the chromosomal PCP8araE fusion strain is illustrated. The PCP13 and PCP18araE fusions on pKD86 and pKD118 were similarly recombined onto the chromosome. The resulting strains are BW27270, BW27535 and BW27271. See Methods for details. These KmR recombinants and the corresponding KmS recombinants (BW27379, BW27536 and BW27380) were verified by testing for the presence of novel junction fragments of expected sizes before and after elimination of the kanamycin-resistance gene as described elsewhere. araE upstream primer (araE up): 5'-CATGGCGACCAACAATACTC-3'. araE downstream primer: 5'-TTCCGCCTCAATATGACG-3'. Primers k1 and k2 and priming site 1 are as described previously (Datsenko & Wanner, 2000 ).

(54K): Fig. 3. Sequences of the PCP8, PCP13 and PCP18araE fusions. The respective promoter regions were PCR-amplified from the chromosomes of BW27379, BW27380 and BW27536 and analysed by automated DNA sequencing with the flanking primers (araE up, Fig. 2; and 5'-TATTCATACGCCGCGTATCC-3') in the Microbiology and Molecular Genetics Core Facility at Harvard Medical School. The entire region recombined onto the chromosome in the original PCR (Fig. 2), along with approximately 200 bp of flanking DNA, was analysed. Only the relevant segments are shown. No mutation was detected in regions not shown. Priming site 1 (Datsenko & Wanner, 2000 ) sequences are in bold and overlined; FRT sites are in italics; the CP8, CP13 and CP18 promoter sequences are underlined. The first 20 and last three nucleotides correspond to the kduDaraE intergenic region and the start codon of araE, respectively.

Cell growth and induction studies.
Induction studies with arabinose were performed in C medium (Helmstetter, 1968 ) with 3·4 % (v/v) glycerol as carbon source. Antibiotics were added to the following concentrations: ampicillin, 100 µg ml^-1; chloramphenicol, 34 µg ml^-1; erythromycin and gentamicin, 20 µg ml^-1. E. coli CW2587 was grown overnight at 37 °C in an air shaker without arabinose to an optical density at 600 nm (OD₆₀₀) of 0·60·8. The cells were collected by centrifugation (5 min, 15000 g) and resuspended in fresh C medium with antibiotics to an OD₆₀₀ of 0·10·2. Arabinose was added (at time 0 in all plots) to different concentrations, and 1 ml samples were taken at 2 h intervals for analysis.

Culture OD₆₀₀ was measured in a Beckman DU 640 spectrophotometer (Beckman Instruments) and fluorescence was measured in a Versafluor Fluorimeter (Bio-Rad) with 360/40 nm excitation and 510/10 nm emission filters. Flow cytometry was performed on a Beckman-Coulter EPICS XL flow cytometer (Beckman Instruments) equipped with an argon laser (emission at 488 nm/15 mW) and a 525 nm band pass filter. The sampled cells were diluted to an OD₆₀₀ of 0·050·1 and kept on ice prior to analysis. For each sample, 30000 events were collected at a rate between 500 and 1000 events per second.

Previously, we showed that independent expression of araE in arabinose-transport-deficient strains led to homogeneous gfpuv expression from the P_BAD promoter (Khlebnikov et al., 2000 ). For induction of the P_BAD promoter a threshold internal arabinose concentration is necessary, and that intracellular arabinose concentration is related to the extracellular arabinose concentration and the arabinose transport capacity of the cell. In order to examine the influence of arabinose concentration and amount of arabinose permease on expression from P_BAD we constructed a series of plasmids with constitutive promoters of different strengths, allowing us to vary the amount of permease.

Expression of araE from plasmid-borne P_CP promoters
Given the results from the previous experiments (the strength of the promoter controlling araE appears to affect culture homogeneity) flow cytometry experiments were conducted to examine the effect of araE expression from various constitutive promoters on gene expression from the arabinose-dependent P_BAD promoter. These experiments were performed in the arabinose-transport-deficient strain E. coli CW2587 containing the arabinose-transport gene araE on the pJAT vectors (P_CParaE) and gfp under control of P_BAD on the high-copy plasmid pCSAK50 (P_BADgfpuv).

All cultures containing the pJATaraE plasmids were induced homogeneously (Fig. 4). The culture-averaged fluorescence (fluorescence/OD₆₀₀) was highest with P_CP18 but lower and approximately equal for P_CP8 and P_CP13 at all inducer concentrations (Fig. 5a). All cultures except CW2587 harbouring pJAT18araE grew at approximately the same rate. Because CW2587 harbouring pJAT18araE grew much more slowly and reached a lower final density than all other CW2587 cultures the culture-averaged fluorescence was higher for that culture. At the highest arabinose concentration (2%) a decline in the culture-averaged fluorescence was observed, suggesting that the P_BAD promoter was saturated. All control cultures without a functional arabinose transport system displayed a single non-fluorescent population and were not able to grow on arabinose as a carbon source.

(29K): Fig. 4. Fluorescence intensity of E. coli harbouring pCSAK50 (araCPBADgfpuv) and the plasmid-borne (left column) or the chromosomally integrated (right column) constitutive promoters controlling araE. The top plot in the right column is the fluorescence intensity for E. coli BW27378 (araE under control of its native promoter on the chromosome) harbouring pCSAK50. In the second row, the first plot is that for E. coli CW2587 harbouring pCSAK50 and pJAT18araE and the second plot is that for E. coli BW27784 harbouring pCSAK50; row 3 is CW2587 pCSAK50 pJAT13araE and BW27786 pCSAK50; and row 4 is CW2587 pCSAK50 pJAT8araE and BW27783 pCSAK50. The fluorescence intensity of individual cells was measured by flow cytometry 6 h after addition of arabinose at the indicated concentrations: grey-shaded curve, 0%; orange curve, 2x10-5%; green curve, 2x10-4%; red curve, 2x10-3%; blue curve, 2x10-2%; brown curve, 2x10-1%; purple curve, 2%.

(28K): Fig. 5. Culture-averaged fluorescence (fluorescence/OD600) of E. coli harbouring the plasmid-borne (a) or the chromosomally integrated (b) constitutive promoters as a function of the inducer concentration. All cultures harboured pCSAK50 (araCPBADgfpuv). Six hours after addition of arabinose, the culture-averaged fluorescence was determined. Data were corrected for the background fluorescence displayed by control cultures. (a) E. coli CW2587 harbouring pCSAK50 and pJAT8araE (black), pJAT18araE (vertical lines) and pJAT13araE (grey). (b) E. coli BW27783 (PCP8araE) harbouring pCSAK50 (black), BW27784 (PCP18araE) harbouring pCSAK50 (vertical lines), BW27786 (PCP13araE) harbouring pCSAK50 (grey) and BW27378 (PEaraE) harbouring pCSAK50 (white).

Expression of araE from P_CP promoters in single copy on the chromosome
The constitutive promoters contained on the pJAT plasmids (pCP8, pCP13 and pCP18) were recombined onto the chromosome of the araFGH strain BW27378 replacing the arabinose-responsive araE promoter. The resulting strains, BW27783 (P_CP8araE), BW27784 (P_CP18araE) and BW27786 (P_CP13araE), as well as the parental strain were transformed with pCSAK50 and induced with various concentrations of arabinose. The homogeneity of induction was measured using flow cytometry.

All cultures containing the chromosomally integrated P_CParaE were homogeneously induced (Fig. 4, bottom three plots of right column), whereas the parental strain displayed a double population (Fig. 4, top plot of right column). The culture-averaged fluorescence increased with inducer concentration (Fig. 5b), although the difference between the induction at high and low inducer concentrations was less than with the plasmid-borne, constitutively expressed transport genes In contrast to the strains bearing the pJATaraE plasmids, the culture-averaged fluorescence was highest in cells carrying the chromosomal P_CP8, was lower for P_CP18, and was lowest for P_CP13. Since these experiments were carried out using different strains than those above, differences are probably attributable to strain background or increased stability of the chromosomal constructs.

Previously, we have shown that providing arabinose-transport-deficient cells with a plasmid-borne araE gene under control of an IPTG-inducible promoter resulted in a homogeneous population of cells expressing the P_BAD promoter at all arabinose concentrations (Khlebnikov et al., 2000 ). As there are many applications for which one may want more than one inducible promoter to control expression of multiple genes, using P_tac to control expression of araE prevented its use to control expression of another gene. The arabinose-inducible P_BAD and the IPTG-inducible P_lac (or P_tac or P_trc) promoters are convenient because they are readily controllable and well characterized. The expression vectors and hosts described here eliminate the all-or-none induction behaviour of P_BAD while freeing the lac promoter for use with another gene of interest.

Expression of the gene encoding the low-affinity, high-capacity arabinose permease from constitutive promoters eliminated all-or-none induction of P_BAD. In general, the level of induction from the P_BAD promoter varied most with the concentration of inducer in the medium and slightly with the constitutive promoter strength controlling the arabinose transport gene, whether expressed from the medium-copy plasmids or from the single-copy chromosome. A relatively linear response in P_BAD induction was observed over a 1000-fold range of inducer concentration.

These constructs and strains should prove useful for controlled production of regulatory proteins, where a consistent and regulatable response from all cells in a culture is desired, or for expression of genes involved in the synthesis of a secondary metabolite, where under- or overexpression of a given pathway could lead to inefficient production of the desired metabolite.

This research was supported by the ERC Program of the National Science Foundation under award number EEC-9731725 and by Award GM63525 from the National Institutes of Health to J.D.K. and by Award MCB-9730034 from the National Science Foundation to B.L.W.