Modulating the import of medium-chain alkanes in E. coli through tuned expression of FadL
© Call et al. 2016
Received: 23 February 2016
Accepted: 29 March 2016
Published: 5 April 2016
In recent years, there have been intensive efforts to develop synthetic microbial platforms for the production, biosensing and bio-remediation of fossil fuel constituents such as alkanes. Building predictable engineered systems for these applications will require the ability to tightly control and modulate the rate of import of alkanes into the host cell. The native components responsible for the import of alkanes within these systems have yet to be elucidated. To shed further insights on this, we used the AlkBGT alkane monooxygenase complex from Pseudomonas putida GPo1 as a reporter system for assessing alkane import in Escherichia coli. Two native E. coli transporters, FadL and OmpW, were evaluated for octane import given their proven functionality in the uptake of fatty acids along with their structural similarity to the P. putida GPo1 alkane importer, AlkL.
Octane import was removed with deletion of fadL, but was restored by complementation with a fadL-encoding plasmid. Furthermore, tuned overexpression of FadL increased the rate of alkane import by up to 4.5- fold. A FadL deletion strain displayed a small but significant degree of tolerance toward hexane and octane relative to the wild type, while the responsiveness of the well-known alkane biosensor, AlkS, toward octane and decane was strongly reduced by 2.7- and 2.9-fold, respectively.
We unequivocally show for the first time that FadL serves as the major route for medium-chain alkane import in E. coli. The experimental approaches used within this study, which include an enzyme-based reporter system and a fluorescent alkane biosensor for quantification and real-time monitoring of alkane import, could be employed as part of an engineering toolkit for optimizing biological systems that depend on the uptake of alkanes. Thus, the findings will be particularly useful for biological applications such as bioremediation and biomanufacturing.
KeywordsAlkanes Transport proteins Biosensors Bio-oxidation Import Solvent tolerance
Fossil fuels including oil, coal and gas still serve as our predominant source of energy, meeting over 85 % of our requirements . Environmental, social and ecological issues have spurred on alternative approaches for fulfilling energy demands. In recent years, intensive efforts have been placed on the development of microbial platforms for the production and waste-management of fossil fuel constituents such as alkanes. Numerous studies have shown that microbes can be engineered for the biosynthesis, bioremediation and biosensing of alkanes [2–5]. An important aspect of developing and optimizing such systems is to understand the transport characteristics of the microbial host. However to date, the passage of alkanes across the membranes still remains a poorly understood area.
AlkL, which forms part of the well-characterized alkane degradation pathway in Pseudomonas putida GPo1, was identified as the first known bacterial alkane importer . We showed previously that heterologous expression of alkL in E. coli could improve dodecane (C12) oxidation by up to 2 orders of magnitude . The uptake process was evaluated by using the P. putida alkane monooxygenase (AlkB). This inner membrane complex receives electrons from NADH, via rubredoxin reductase (AlkT) and rubredoxin (AlkG), and transfers one oxygen to the alkane substrate while the other is reduced to H2O . Ultimately, it oxidizes alkanes to fatty alcohols, fatty aldehydes, and fatty acids and has been investigated industrially for the bioconversion of octane to octanol . In P. putida, alkanes oxidized to fatty acids are gradually converted to acyl-CoA by AlkK (FadD in Escherichia coli) for entry into the β-oxidation cycle [7, 9].
Results and discussion
Import of octane in E. coli is mediated by FadL
To determine whether OmpW and FadL transporters might be involved in the native uptake of octane, the ΔfadL and ΔompW strains from the Keio E. coli knockout collection  were transformed with the plasmid pGEc41  encoding the alkane inducible alkBGT alkane monooxygenase complex. With this plasmid, cultivation in the presence of 30 % (v/v) octane leads to the formation of three oxidized products: 1-octanol, 1-octanal and 1-octanoic acid, as reported earlier . The synthesis of 1-octanol and 1-octanal is attributed to AlkB while the synthesis of 1-octanoic acid is catalyzed either by AlkB and/or native enzymes that have yet to be identified.
This has interesting implications in our current understanding of the substrate binding of FadL which has long been known to play a physiological role in the import of long chain fatty acids in E. coli [12, 13]. Previously, Black et al.  had shown that deletion of the His110 residue could significantly reduce the rate of fatty acid binding and import, indicating its importance in the interaction of the carboxylic group with FadL. From structural studies, the zwitterionic head group of the detergent molecule, lauryldimethylamine-oxide (LDAO) is positioned close to Arg157 and Lys317 suggesting that these residues are also necessary for carboxylic group interaction . Kinetic studies show an almost 10-fold increase in Km for shorter chain fatty acids (C6-C9) compared to LCFAs (C12-C16) . As octane is an aliphatic hydrocarbon without the negatively-charged carboxylic head group, these results do suggest that the binding of substrates to FadL is not necessarily dependent upon the presence of electrostatic charges.
Optimal import of alkanes requires tuned expression of protein transporters
We noted a complex interplay between IPTG-based membrane transporter overexpression and octane conversion. For example, transforming E. coli HB101 strain with pASKAfadL led to a reduction in the specific activity of octane bioconversion from 6.2 to 4.1 μmol min−1g−1; specific activity further decreased to 2.1 and then to 0.48 μmol min−1g−1 with induction of recombinant FadL at 50 μM and 500 μM IPTG respectively (Additional file 1: Figure S1a). By switching to the E. coli Keio ΔfadL strain complemented with a pASKAfadL, a slight increase in activity from 3.1 to 3.25 μmol min−1g−1 between 0 and 50 μM IPTG induction was initially observed which drastically decreased to 0.71 μmol min−1g−1 at the higher concentration of 500 μM IPTG (Additional file 1: Figure S1b). Taken together, these observations corroborate our previous findings in which octane conversion was found to be inversely correlated with AlkL induction . Even though the mode of inhibition remains to be proven, these findings strongly support the notion that high expression level of membrane importers may not necessarily offer a sound approach for achieving optimal rates of bioconversion and could, in fact, be detrimental to the physiology of the host organism.
Although the tighter rhaBAD promoter may have helped to resolve leaky or poorly controlled expression issues, intrinsic differences in induction, folding and translocation efficiencies may still have impacted optimal inducer concentration. Furthermore, subtle differences in the affinity of AlkL OmpW, and FadL for octane could also account for the variations in optimal inducer concentrations. Nonetheless, it is clear that optimizing import to maximize conversion rates of octane by AlkBGT not only requires choosing the right expression system, but also sensitive tuning of induction levels.
Removal of fadL confers tolerance toward medium chain alkanes
Given that the viability of E. coli in the presence of high alkane concentration is less affected in a ΔfadL genetic background suggests that tolerance could be engineered by reducing or abolishing FadL activity. Such an approach would reduce toxicity by effectively lowering the rate of intracellular influx of alkanes. For the ΔmarR strain, which was found to possess greater growth resilience in the presence of alkanes, the physiological explanation is probably somewhat more complex since multiple stress-related mechanisms are most likely to operate simultaneously . The induction of the multi-drug efflux pump, AcrAB-TolC, is one such response which is known to confer resistance to multiple drugs, in addition to alkanes including octane . It remains possible that, during the course of this study, the native house-keeping activity of the AcrAB-TolC complex may well have partially masked the tolerance benefit imparted by the deletion of FadL. In light of this, AcrAB-TolC mutants may provide a more ideal genetic background for evaluating the significance of FadL in conferring tolerance toward a range of chemicals including alkanes.
Impact of native alkane import on whole-cell alkane biosensing
Controlling the import of signal-eliciting molecules will be important for the development of advanced genetic circuits (including biosensing systems) that respond to external stimuli in a precise, predictable and dose-dependent manner . In this regard, transmembrane importers such as FadL could serve as critical targets for engineering purposes given their importance in influencing the signal response of whole-cell biosensing systems.
In conclusion we show FadL to be the principle route for octane import in E. coli. This study provides a useful toolkit of experimental approaches for evaluating alkane import, and highlights the importance of native outer membrane importers such as FadL for the purpose of developing and optimizing biological applications such as bioremediation and biofuels.
Initial identification of candidate genes was expanded from the literature search using online resources: Ecocyc database, (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC99147/). OmpW was previously identified as a target via BLAST searching, revealing identity with the known alkane importer AlkL . Structural alignment and viewing of FadL, AlkL and OmpW was performed using Pymol. The AlkL models were generated using Swiss Model [29, 30], using the 2f1t x-ray crystal structure as a template. Membrane topology was predicted using the orientations of proteins in membranes database .
The Keio collection, a knockout library of all non-essential genes in E. coli , and ASKA collection, a clonal plasmid library of all individual E. coli gene open reading frames  were obtained from the NBRP, Nara University of Science and Technology. Both collections are derived from the E. coli K-12 strain and allow systematic studies of gene function based on gene overexpression or complementation. The Keio strains have each gene replaced on the chromosome by a kanamycin resistance gene flanked by FRT recombination sites for removal if necessary. The ASKA strains harbor plasmids with the open reading frame (ORF) for each gene, expressed under control of the LacIa promoter, and a chloramphenicol resistance marker. See Additional file 1: Table S1 for a list of all strains used in this study.
The pGEc41 plasmid is 51.6 kb in size and is constructed from a 21 kb vector pLAFR1 containing a 30 kb fragment with alk operon genes alkBFGHJKL and alkST from Pseudomonas putida with the omission of alkJ (an alcohol dehydrogenase), alkK (an acyl-CoA synthetase) and alkL . Gene induction was placed under the control of the AlkS – pAlkB promoter system. ASKA plasmids containing the ORF for each target gene were isolated from the ASKA collection . They contain ORFs for all E. coli genes cloned between Sfil restriction sites, and under the control of isopropyl-β-D-thiogalactopyranoside (IPTG) inducible lac1q promoter. See Additional file 1: Table S1 for a list of all plasmids used in this study.
Lysogeny Broth (LB) growth medium; composed of yeast extract 10 g/L, tryptone 20 g/L and NaCl 20 g/L; was used for most overnight inoculations, Octane uptake and conversion assays were performed in ‘Wubbolts’ enhanced conversion media, taken from Wubbolts et al. . The composition of Wubbolts medium is as follows: KH2PO4, 4 g/L; K2HPO4 (3H2O), 15.9 g/L; Na2HPO4 (12H2O), 7 g/L; (NH4)2SO4, 1.2 g/L; NH4Cl, 0.2 g/L (all from Sigma Aldrich); yeast extract, 5 g/L; L-leucine, 0.6 g/L; L-proline, 0.6 g/L; thiamine, 5 mg/L. The following components were autoclaved and added separatelyo add post autoclaving: MgSO4 (7H2O) (BDH), 1 g/L (BDH); 1 ml of trace minerals (composition below); 1 ml of 4 % (w/v) CaCl2 (2H2O) (Alfa Aesar) and 10 g/L D-glucose, were added having all been heat sterilized separately. Filter- sterilized antibiotics were added as appropriate. The trace minerals solution contained per litre of 5 M HClL: 40 g FeSO4 (7H2O), 10 g MnSO4 (H2O), 4.75 g CoCl2 (6H2O), 2 g ZnSO4 (7H2O), 2 g MoO4Na2 (2H2O), 1 g CuCl2, (2H2O) and, 0.5 g H3BO3. Antibiotics were used at the following working concentrations: tetracycline (10 mg/L), chloramphenicol (25 mg/L), Zeocin (100 mg/L) and kanamycin (30 mg/L).
In-vivo conversion of octane
Comparative bioconversion assays were carried out in 24 deep-well Teflon plates to reduce the adsorption of product and substrate on the on the walls of individual wells. Strains being tested were inoculated overnight in 5 ml LB in 50 ml Falcon tubes with relevant antibiotics for all strains and plasmids. At least 3 biological repeats were isolated from the original transformation plate for each plasmid and condition tested. 1 ml culture volumes were transferred to each well accordingly, and 300 μl of octane (Sigma-Aldrich) was added at t = 0. The initial assay (Fig. 3) was carried out in LB media in plates sealed with aluminium foil and clamped with a film cover (EnzyScreen) to counter octane evaporation, and without pre-induction of pGEc41. For overexpression of ompW and fadL, appropriate concentrations of IPTG were added prior to overnight culture and octane addition. For later assays (Additional file 1: Figure S2 and Fig. 4), 300 μl of substrate was deemed sufficient to counter the effects of evaporation and ensure that substrate availability was not rate-limiting. Gas transfer was facilitated, via a pinhole in the sandwich cover. For pre-induction of pGEc41 in Additional file 1: Figure S2 and Fig. 4, strains were sub-cultured 1:20 in 1 ml of enhanced conversion media (Wubbolt’s media) for 1 h and then induced with 0.05 % (w/v) Dicyclopropylketone (DCPK) and incubated at 37 °C for a further 4 h. For expression from the pRHA67k-derived plasmids, L-rhamnose was added after 1 h of growth in Wubbolt’s media and 4 h prior to octane addition. Assay plates were clamped shut in the shaker incubator (Kuhner) at 37 °C and shaken at 250 RPM with a throw diameter of 25 mm. Well contents were harvested after 8 h (Fig. 3 and Additional file 1: Figure S1) and 1 h (Fig. 4 and Additional file 1: Figure S2) and transferred to pre-weighed 2 ml Eppendorf tubes, centrifuged at 13,000 RPM for 5 min and the supernatant poured into separate tubes. After addition of 800 μl ethyl acetate (Sigma-Aldrich) to separate tubes containing the supernatant and cell pellet, each was vortexed for 1 min to separate the alkanes along with the alkane-derived oxidized products, and then centrifuged for 1 min to separate the aqueous and organic phases. A 600 μl volume of the top phase (organic phase) was transferred to glass vials for analysis by GC-FID.
Seed cultures of the strains were grown overnight in 96-well DSW plates and sub-cultured (using a 5 % (v/v) inoculum) into fresh LB media at 37 °C, shaking at 250 rpm (25 mm amplitude) with absorbance readings monitored at 600 nm using a BMG Clariostar plate reader after 0 h and 2 h. Half the cultures were spiked with an excess of 10 % v/v n-alkane and incubated for a further 3 h until the alkane had evaporated, after which point a final set of OD readings was taken.
Determination of alkanes and their respective oxidized products in microwells was achieved via sacrificial sampling and extraction of both phases into ethyl acetate directly. 800 μl of ethyl acetate was added to the two-phase supernatant following centrifugation at 13,000 rpm for 5 min in a microfuge. The samples were vortexed for 3 × 20 s, prior to centrifugation for 1 min and removal of 100 μl of the organic phase, and then analysed by GC-FID using a Perkin Elmer Autosystem XL equipped with a SGE BPX5 (30 m long; 0.53 mm internal diameter, 1 μm film) capillary column and helium as the carrier gas under a constant pressure of 4 PSI. The proportion of organic phase was determined by cross-referencing the n-alkane peak size to a calibration curve of known standards. Alkanes along with their oxidized derivatives were separated and quantified using the following GC-FID method. For n-octane oxidation product determination the samples were eluted at 70 °C followed by a linear increase of 5 °C minute−1 to reach a final temperature of 145 °C. All standards were purchased from Alfa Aesar at the highest purity available (>98 %).
Determination of dry cell weight
Cell density measurements were taken by sacrificing the well contents and centrifuging the biphasic samples at 13,000 rpm (19,000 g) for 5 min, marking the aqueous volume on the side of the graduated Eppendorf tube; pellets were washed with Tris–HCl pH7.4 and dried in an 80 °C oven until a constant mass was reached.
Alkane biosensor assay
E. coli BW25113 and the Keio ΔfadL strain were transformed with an alkS-Superfolder GFP biosensor on pSB50C7. The strains were grown up overnight from single colonies in parallel in deep square well plates (1 ml per well) at 37 °C and 250 rpm (25 mm amplitude) in Luria broth containing 5 g/L glucose and then sub cultured (4 % incoculum) with 10 g/L glycerol added, grown for 2 h, and induced using 3 mM-arabinose. After 3 h growth, 1%v/v alkane substrates were added and grown for 3 h, analysing fluorescence every 10 min (excita- tion 485 nm; emission 517 nm) using a BMG Clariostar plate reader.
The authors would like to acknowledge the Engineering and Physical Sciences Research Council (EPSRC) and BBSRC for a Masters training grant BB/H021027/1 for funding. The research was performed at the Advanced Centre for Biochemical Engineering, University College London. The authors would like to acknowledge the NBRP-E.coli at NIG for the KEIO and ASKA collections and Adam Denyer who performed some of the work leading to the toxicity study.
Supporting Information Available
Including bioconversion assay data at varying inducer concentrations and western blot data. This material is available free of charge via the Internet at http://pubs.acs.org.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
- Wilcox J. Grand challenges in advanced fossil fuel technologies. Front Energy Res. 2014;2(47):1–3.Google Scholar
- Zhu L, Zhu Y, Zhang Y, Li Y. Engineering the robustness of industrial microbes through synthetic biology. Trends Microbiol. 2012;20(2):94–101.MathSciNetView ArticleGoogle Scholar
- Moldes AB, Paradelo R, Rubinos D, Devesa-Rey R, Cruz JM, Barral MT. Ex situ treatment of hydrocarbon-contaminated soil using biosurfactants from lactobacillus pentosus. J Agric Food Chem. 2011;59(17):9443–7.View ArticleGoogle Scholar
- Grant C, Deszcz D, Wei Y-C, Martínez-Torres RJ, Morris P, Folliard T, Sreenivasan R, Ward J, Dalby P, Woodley JM, Baganz F. Identification and use of an alkane transporter plug-in for applications in biocatalysis and whole-cell biosensing of alkanes. Sci Rep. 2014;4:5844.View ArticleGoogle Scholar
- Choi YJ, Lee SY. Microbial production of short-chain alkanes. Nature. 2013;502(7472):571–4.View ArticleGoogle Scholar
- Julsing MK, Schrewe M, Cornelissen S, Hermann I, Schmid A, Bühler B. Outer membrane protein AlkL boosts biocatalytic oxyfunctionalization of hydrophobic substrates in Escherichia coli. Appl Environ Microbiol. 2012;78(16):5724–33.View ArticleGoogle Scholar
- Rojo F. Specificity at the end of the tunnel: understanding substrate length discrimination by the AlkB Alkane hydroxylase. J Bacteriol. 2005;187(1):19–22.MathSciNetView ArticleGoogle Scholar
- Favre-Bulle O, Weenik E, Vos T, Preusting H, Witholt B. Continuous bioconversion of n-octane to octanoic acid by recombinant Escherichia coli (alk(+)) growing in a two-liquid-phase Chemostat. Biotechnol Bioeng. 1993;41(2):263–72.View ArticleGoogle Scholar
- Ji Y, Mao G, Wang Y, Bartlam M. Structural insights into diversity and n-alkane biodegradation mechanisms of alkane hydroxylases. Front Microbiol. 2013;4:58.View ArticleGoogle Scholar
- Eggink G, Lageveen RG, Altenburg B, Witholt B. Controlled and functional expression of the Pseudomonas oleovorans alkane utilizing system in Pseudomonas putida and Escherichia coli. J Biol Chem. 1987;262(36):17712–8.Google Scholar
- Favre-Bulle O, Schouten T, Kingma J, Witholt B. Bioconversion of n-octane to octanoic acid by a recombinant Escherichia coli cultured in a two-liquid phase bioreactor. Nat Biotechnol. 1991;9:367–71.View ArticleGoogle Scholar
- Black PN. Characterization of FadL-specific fatty acid binding in E. coli. Biochim Biophys Acta. 1990;1046(1):97–105.View ArticleGoogle Scholar
- Nunn WD, Simons RW. Transport of long-chain fatty acids by Escherichia coli: mapping and characterization of mutants in the fadL gene. Proc Natl Acad Sci U S A. 1978;75(7):3377–81.View ArticleGoogle Scholar
- van den Berg B. Going forward laterally: transmembrane passage of hydrophobic molecules through protein channel walls. Chembiochem. 2010;11(10):1339–43.View ArticleGoogle Scholar
- Hong H, Patel DR, Tamm LK, van den Berg B. The outer membrane protein OmpW forms an eight-stranded beta-barrel with a hydrophobic channel. J Biol Chem. 2006;281(11):7568–77.View ArticleGoogle Scholar
- Lepore BW, Indic M, Pham H, Hearn EM, Patel DR, van den Berg B. Ligand-gated diffusion across the bacterial outer membrane. Proc Natl Acad Sci U S A. 2011;108(25):10121–6.View ArticleGoogle Scholar
- Van den Berg B. The FadL family: unusual transporters for unusual substrates. Curr Opin Struct Biol. 2005;15(4):401–7.View ArticleGoogle Scholar
- Simard JR, Pillai BK, Hamilton JA. Fatty acid flip-flop in a model membrane is faster than desorption into the aqueous phase. Biochemistry. 2008;47(35):9081–9.View ArticleGoogle Scholar
- Baba T, Ara T, Hasegawa M, Takai Y, Okumura Y, Baba M, Mori H. Construction of Escherichia coli K-12 in-frame, single-gene knockout mutants: the Keio collection. Mol Syst Biol. 2006;2:2006.0008.View ArticleGoogle Scholar
- Black PN, Zhang Q. Evidence that His110 of the protein FadL in the outer membrane of Escherichia coli is involved in the binding and uptake of long-chain fatty acids: possible role of this residue in carboxylate binding. The Biochemical Journal. 1995;310(Pt 2):389–94.View ArticleGoogle Scholar
- Aono R. Improvement of organic solvent tolerance level of Escherichia coli by overexpression of stress-responsive genes. Extremophiles. 1998;2(3):239–48.View ArticleGoogle Scholar
- Diaz E. Microbial degradation: Genomics and Molecular Biology. Caister Academic Press; 2008Google Scholar
- Foo JL, Leong SSJ. Directed evolution of an E. coli inner membrane transporter for improved efflux of biofuel molecules. Biotechnol Biofuels. 2013;6(1):81.View ArticleGoogle Scholar
- Reed B, Blazeck J, Alper H. Evolution of an alkane-inducible biosensor for increased responsiveness to short-chain alkanes. J Biotechnol. 2012;158(3):75–9.View ArticleGoogle Scholar
- Collao B, Morales EH, Gil F, Calderón IL, Saavedra CP. OmpW is cooperatively upregulated by MarA and SoxS in response to menadione. Microbiology. 2013;159:715–25.View ArticleGoogle Scholar
- Doukyu N, Ishikawa K, Watanabe R, Ogino H. Improvement in organic solvent tolerance by double disruptions of proV and marR genes in Escherichia coli. J Appl Microbiol. 2012;112(3):464–74.View ArticleGoogle Scholar
- Canosa I, Sánchez-Romero JM, Yuste L, Rojo F. A positive feedback mechanism controls expression of AlkS, the transcriptional regulator of the Pseudomonas oleovorans alkane degradation pathway. Mol Microbiol. 2000;35(4):791–9.View ArticleGoogle Scholar
- Zhang F, Carothers JM, Keasling JD. Design of a dynamic sensor-regulator system for production of chemicals and fuels derived from fatty acids. Nat Biotechnol. 2012;30(4):354–9.View ArticleGoogle Scholar
- Kiefer F, Arnold K, Künzli M, Bordoli L, Schwede T. The SWISS-MODEL Repository and associated resources. Nucleic Acids Res. 2009;37:D387.View ArticleGoogle Scholar
- Kopp J, Schwede T. The SWISS-MODEL Repository of annotated three- dimensional protein structure homology models. Nucleic Acids Res. 2004;32:D230.View ArticleGoogle Scholar
- Lomize MA, Lomize AL, Pogozheva ID, Mosberg HI. OPM:orientations of proteins in membranes database. Bioinforma. 2006;22:623–5.View ArticleGoogle Scholar
- Kitagawa M, Ara T, Arifuzzaman M, Loka-nakamichi T. Complete set of ORF clones of Escherichia coli ASKA library (A Complete Set of E.coli K-12 ORF Archive): Unique Resources for Biological Research. DNA Res. 2006;12:291–9.View ArticleGoogle Scholar
- Austin RN, Chang HK, Zylstra GJ, Groves JT. The non-heme diiron alkane monooxygenase of Pseudomonas oleovorans (AlkB) hydroxylates via a substrate radical intermediate. J Am Chem Soc. 2000;122(47):11747–8.View ArticleGoogle Scholar
- Wubbolts MG, Favre-Bulle O, Witholt B. Biosynthesis of synthons in two-liquid-phase media. Biotechnol Bioeng. 1996;52(2):301–8.View ArticleGoogle Scholar
- Boyer HW, Roulland-Dussoix DJ. Mol Biol. 1969;41:459–72.View ArticleGoogle Scholar
- Giacalone MJ, Gentile AM, Lovitt BT, Berkley NL, Gunderson CW, Surber MW. Toxic protein expression in Escherichia coli using a rhamnose-based tightly regulated and tunable promoter system. BioTechniques. 2006;40:355–63.View ArticleGoogle Scholar