5
Biotechnology
The following chapter explains how to practically transform a “-barrel membrane protein (MP) into a nano-channel with desired geometrical and/or functional features, starting from the conceptdesign and design of the respective gene. It will then give an overview on the conventional means of production and purification of bacterial OMPs (outer membrane proteins), stressing on the problems and challenges of over-expressing OMPs into the Gram-negative bacterial outer membrane and of isolating them from the outer membrane. Furthermore the special problems of producing modified “-barrel MPs and ways to overcome these problems by using alternative methods will be named and explained. The different ways of analyzing OMP samples regarding yield, purity and correct folding will be considered briefly and the chapter discusses the distinctive experimental adaptations to scale-up the production of OMPs in general and of modified OMPs especially. As for many industrial applications of OMP derived nano-materials vast amounts of the proteins need to be produced, exceeding the capacities of conventional methods. The chapter will close with a discussion on artificial “-barrel structures to which OMPs are an alternative.
5.1
Outer Membrane Protein Modification
As addressed in Chap. 2 the “-barrel shaped integral OMPs of Gram-negative bacteria are well-suited for the nano-material design. From
10 “-strands on, these proteins form pores and channels that reach to quite substantial dimensions in strand-rich proteins such as the TonB-dependant transporters that harbor 22 strands as for instance the E. coli FhuA or 24 as the so-far largest known natural OMP PapC from E. coli. The robust barrel structure tolerates mutations and facilitates protein-refolding from the fully or partially denatured state. OMPs spontaneously insert into lipid or polymer membranes opening the possibility to design novel hybrid materials for various applications (see Chap. 6), well studied examples are nanocontainer systems, in which a protein nanochannel allows controllable compound release or biosensor systems in which a membrane reconstituted protein channel is used to monitor analytes. Analytes are detected by single channel conductance measurements. The special class of bacterial “-barrel pore-forming toxins is quite interesting in this respect, as they selfassemble from several monomers to form stable pores with stable conductance. Especially the heptameric ’-hemolysin from S. aureus has been successfully employed in the design of nano-pore sensors for the detection of single molecules. Since ’-hemolysin does not contain an intrinsic specificity for a certain molecule or a class of molecules and since specificity is the result of genetic engineering it can be utilized very flexibly [1]. In general and as logic dictates the OMP has to be chosen according to the desired application, in case of nano-sensors for instance
M. Fioroni et al., ß-barrel Channel Proteins as Tools in Nanotechnology, Advances in Experimental Medicine and Biology 794, DOI 10.1007/978-94-007-7429-2__5, © Springer ScienceCBusiness Media Dordrecht 2014
95
96
proteins lacking motile loop-regions that interfere with measurable channel conductance are especially suited. Further characteristics depend on the analyte’s nature, a huge and bulky analyte for example might necessitate a protein pore with larger inner diameter. For a hybrid-catalyst system with an OMP as reaction-specificity inducing scaffold for a metal catalyst instead, a more narrow pore might be of advantage, again depending on the substrate to be transformed [2]. In drug-release systems specificity (apart from size-dependent specificity) is generally not an issue as a nano-container enclosed drug simply has to be released at a target. In this case the ability to reversibly open or close an OMP channel is of greater importance [3, 4] and motile loops that might lead to a temporary channel-closing are again undesirable. All of these applications and further ones such as membrane systems or molecular sieves with certain cut-off might ask for larger than the naturally occurring protein channels. Depending on the lipid- or polymerbased hydrophobic materials used into which the OMPs should be reconstituted, the protein characteristics have to be adapted to the material, if an adaptation of the material is not possible or proves disadvantageous (for instance due to costineffectiveness). As these examples show nanomaterial design often requires the change of a chosen OMP’s natural features or the addition of new features. These changes can be either introduced by chemical modifications or by geneticengineering or a combination of both. Major changes as an altered geometry however will mostly require the introduction of mutations on gene level that are subsequently translated to protein level, while a triggered opening and closing of an OMP channel can be achieved by chemical labeling of adequate amino acid residues within the channel. It might however be necessary to introduce suitable amino acid residues by site-directed mutagenesis. Genetic engineering to transform an OMP into a nanochannel will generally be based on a rational approach. A recent example shows the great potential of “-barrel proteins and their units (i.e. strands and sheets) for the creation of artificial nano-
5 Biotechnology
pores. De Pinto et al. reported the nano-pore design starting from natural “-barrel structures. A multi-alignment of general diffusion porins identified a highly conserved motif coding for two “-strands (obtained from E. coli OmpF) as the basic module of the artificial pores. Hexameric repeats of the respective sequence were obtained through cDNA recombinant technology. The coded protein was expressed, purified and reconstituted in planar bilayer membranes and showed channel-forming ability [5]. The following sections will explain how to modify OMPs for nano-channel design purposes starting out from gene-design considerations.
5.1.1
Gene-Design
The design of completely new genetic information is a very flexible and powerful way to obtain engineered or entirely novel proteins, especially so since the designed genes can nowadays be synthetized commercially at an affordable price of 0.25A C/bp (in 2013) [6]. Furthermore the vast sequence information obtained by the various sequencing projects (as collected on the website of the “International Sequencing Consortium”: http://www.intlgenome.org/viewDatabase.cfm), such as the E. coli genome project [7], allowed the development of computer based gene-design software tools helping the scientist to optimize a newly written DNA sequence. General design parameters and criteria that apply to the design of genes coding for soluble proteins apply just as well to MPs and OMPs. A good way to check whether a planned design is valid is the use of theoretical methods that can supply valid and important information. A comprehensive overview on computation methods and tools that can be applied to “-barrel membrane proteins and nano-systems based on these proteins can be found in Chap. 4. Before designing a gene coding for a modified OMP one generally chooses the expression host (organism and strain). Considerations on how to choose the best suited system will be given in Sect. 5.2. After rationally deciding for positions to be mutated, the desired modification will be
5.1 Outer Membrane Protein Modification
introduced “on paper” using the amino acid sequence of the OMP to be engineered as a template. While point-mutations can be introduced by well-established PCR (polymerase chain reaction) techniques, as will be shown in Sect. 5.1.3, vast mutations that will effect a protein’s overall geometry generally require the redesign of the respective open reading frame (ORF) and the subsequent synthesis of a synthetic gene. In theory the design of a new gene from a given amino acid sequence is rather simple, as the desired amino acid sequence has “just” to be reverse translated to the corresponding DNA sequence. However the practical conversion of such an altered “virtual” amino acid sequence to a DNA sequence that is supposed to code for the desired protein is not so trivial, as there are certain prerequisites to be fulfilled: 1. The resulting protein has to be intact, correctly folded and stable. Sequence elements that are crucial for stability, folding or function should be known as they may not be removed or may be newly introduced. 2. The DNA sequence has to optimally fit the selected expression host in terms of codon usage. Each amino acid can be coded by one to six different codons and different organisms differ in their preference for the various codons [8]. Many companies that offer gene synthesis services will also optimize the gene sequence to the expression host codon bias. 3. For prokaryotic hosts: An important component affecting expression levels is the ribosome binding site (RBS) needed for translation initiation between 5 and 15 bases upstream of the AUG start codon. Sequence changes within the RBS might change expression levels substantially [9]. 4. mRNA secondary structures have to be avoided, as they might occlude the RBS or start codon in prokaryotic expression hosts inhibiting translation [10]. 5. Certain restriction sites might be desired for easy cloning, while others instead might interfere with cloning purposes. 6. The protein has to be produced in sufficient yield. This point can be addressed by using a
97
strong promoter sequence, such as the phage derived T5 or T7 promoters. 7. For purification sequence tags, such as the Hexa-His- [11], glutathione-S-transferase (GST)- [12], streptavidin-binding-peptide (STREP- or STREP II)- [13, 14], FLAG-tag [15] or others can be introduced on DNA sequence level. To enhance solubility the attachment of tags like the maltose binding protein (MBP)- [16] or the N-utilization substance A (NusA) [17] can be useful. A comprehensive review on the different affinity and solubility tags and how they can be combined for best results can be found in [18]. There are various software kits and web platforms available that use information from known DNA sequences to assist the user in optimizing an engineered or newly written DNA sequence. While older tools focused mainly on the organism specific codon bias [19–23], recently developed tools, such as GeMS [24], Gene Composer [25], GeneOptimizer [26], Synthetic Gene Designer [27], GeneDesign [28], Visual Gene Developer [29] or Gene Designer [30] consider a combination of the above mentioned parameters. Apart from these parameters that apply to the synthetic gene design in general certain issues have to be considered when designing a gene coding for a MP or OMP. If the synthetic gene derived OMP should be expressed homologous and into its natural environment (the outer membrane), the N-terminal signal sequence that leads to outer membrane targeting (see Sect. 2.3.2) has to be added on gene level, otherwise the protein won’t be folded correctly and will be expressed into inclusion bodies. Successful heterologous expression of MPs is often difficult [31], as membrane targeting might function poorly or might not occur at all and the protein might be toxic for the expression host cells [31–34]. To avoid or improve these issues lowering the expression level (e.g. by adding a weak promoter sequence) may be considered [34]. Several of the designing parameters listed can be also addressed by choosing the optimal expression plasmid. Especially the larger tags or
98
5 Biotechnology
promoter sequences will be generally located on the plasmid used. In conclusion though the design of truly novel genes is becoming more and more facile due to existing sequence knowledge applied to the software tool development, the knowledge on factors that might lead to unsuccessful protein expression (especially when expressing heterologous and especially when producing OMPs) is not yet sufficient to guaranty success in any case. The experimentalist will still have to test the virtually designed gene in laboratory experiments and considerable effort might still be needed to optimize the protein expression conditions (see Sect. 5.2). When a new protein is then effectively expressed it generally needs to be isolated and purified, which again might prove a rather challenging task, especially when working with MPs (see Sect. 5.2.4). The following Sect. 5.1.2 will explain and give examples on how the OMP geometry can be altered by introducing changes to their “virtual” amino acid sequence in order to derive a synthetic gene from which to express the altered protein. The main example will be the E. coli FhuA protein.
5.1.2
Geometry Modification
Outer membrane “-barrel channel proteins like the E. coli FhuA can be an alternative to artificial chemically synthesized nano-pores (on artificial “-barrels see Sect. 5.4). In order to be useful as channel structures for nano-technological applications channel proteins must be flexible enough to be modified in their geometry, i.e. length and diameter depending on the application and hydrophobic carrier (i.e. lipid or polymer membranes or vesicles). Due to the mentioned robustness of the “-barrel structure that tolerates vast sequence mutations; major changes in channel geometry are rendered possible. As these geometry changes have been quite successfully carried out with the FhuA protein, it will serve as an example on how to mutate a TonB-dependent siderophore transporter protein with wide channel diameter to obtain a passive diffusion
channel, whose diameter can be increased, whose hydrophobic membrane spanning region can be elongated and whose flexible loop regions can be partly removed leading to a more regular channel structure (For details on the FhuA WT protein structure and function see Sect. 2.4).
5.1.2.1 From Ferrichrome Transport to Passive Diffusion The first mutations leading to a modification of the FhuA channel geometry were introduced before the protein was recognized as a nano-channel for biotechnological applications. Partial or total deletions of the proteins N-terminal plug-domain (amino acids 1–159) were carried out to reveal the plug function (see Fig. 5.1). In 1999 Braun et al. created a FhuA deletion variant lacking amino acids 5–160, as they hypothesized that the resulting mutant FhuA5– 160 lacking most of the N-terminal plug should form a stable permanently open channel allowing diffusion of substances smaller than its inner diameter. Active transport of ferrichrome instead was thought to be defective, as the TonB-box (amino acids 7–11) thought to be vital for the energy-providing interaction with TonB, as well as the ferrichrome binding domains within the plug were missing [35]. They found that indeed FhuA5–160 formed stable channels within the E. coli outer membrane, rendering cells sensitive to the large antibiotics erythromycin, rifamycin, bacitracin and vancomycin, and enabling them to grow on maltotetraose and maltopentaose in the absence of the LamB protein that is involved in the transport of maltose and maltodextrins. Furthermore cells lacking the TonB in the inner membrane but expressing FhuA5–160 were able to grow on media containing high concentrations of ferrichrome, while cells lacking TonB and FhuA or lacking TonB and carrying FhuA WT were not. These findings confirmed the passive diffusion channel hypothesis. Interestingly though FhuA5–160 still facilitated active ferrichrome transport (at 40 % of WT level) in cells with active TonB, leading to the conclusion that apart from the TonB-box amino acids there have to be other regions within the FhuA that interact with TonB [35]. The same had been
5.1 Outer Membrane Protein Modification
99
Fig. 5.1 Schematic representation of E. coli FhuA genetic engineering concept leading from the cork-closed WT protein to passive diffusion facilitating variant
FhuA1–159, by deletion of the cork-domain amino acids 1–159 (Proteins are shown in New Cartoon representation made with VMD)
found for plug-lacking variants of the Gramnegative bacteria Salmonella paratyphi B and Salmonella enterica serovar Typhimurium FhuA homologues [36] and for a plug-less E. coli FepA (TonB-dependant ferric enterobactin transporter; see Chap. 2) variant [37]. To further analyze the FhuA5–160 features the protein was purified and reconstituted into black lipid bilayer membranes formed by diphytanoyl PtdCho/n-decane and the membrane current was recorded by high-resolution, singlechannel electrical recordings in 1 M KCl. Results showed that the reconstituted protein led as expected to an increased conductance of 0.5 nS in 1 M KCl, however the recordings showed a high degree of noise, leading to the conclusion that the channels were not permanently open [38], most likely due to the long extracellular, motile loops of the protein that might temporarily close the channel. In a later study by Nallani et al. the fully or partly cork-depleted FhuA was first recognized as a robust protein channel with wide diameter, facilitating passive diffusion and thus a perfect protein-based nano-material. Here the FhuA N-terminal amino acids 1–159 were deleted (Fig. 5.1) leading to variant FhuA1-159 and the deletion of amino acids 1–129 led to
FhuA1-129. Both variants had been successfully reconstituted into ABA triblock copolymer vesicles, where A is poly(dimethylsiloxane) and B is poly(2-methyloxazoline) (PMOXAPDMS-PMOXA). The passive diffusion of single stranded DNA-oligomers through the protein channels was reported and it was shown that mutant FhuA1-159 revealed a higher translocation efficiency [39]. The FhuA1-159 potential as nano-material for biotechnological applications was further demonstrated when it was found that the protein residing in PMOXA-PDMS-PMOXA polymer vesicles (nano-compartments) can be used to selectively recover and trap negatively charged molecules such as sulforhodamine B within the vesicles, by the use of positively charged poly-lysine traps enclosed in the nano-compartment. It was furthermore reported that the same FhuA1159 functionalized nano-compartment system can be used for the enzymatic conversion in nano-compartments, shown on the example of 3,30 ,5,50 -tetramethylbenzidine (TMB) oxidation by vesicle enclosed horseradish peroxidase (HRP) [40]. Sulforhodamine B and TMB are thus able to diffuse through the large FhuA1-159 pore. In 2009 the same FhuA variant was shown to facilitate diffusion of the
100
fluorescein derivative calcein again after being reconstituted into PMOXA-PDMS-PMOXA nano-compartments [3]. Though the FhuA1-159 has not been crystallized up to now, analysis by circular dichroism (CD) spectroscopy showed a clear “-structure (49–65 % contribution of “-sheet content) [3, 4, 41], while for the FhuA WT a CD derived “-sheet content of 51 % had been reported [42]. Furthermore a recent molecular dynamics (MD) simulation study on the FhuA WT and FhuA1–159 performed in a DNPC (1,2-dinervonyl-sn-glycero-3-phosphocholine) lipid bilayer and a water/OES (N-octyl-2hydroxyethyl sulfoxide) detergent solution revealed that the mutant protein shows a remarkable stability in both environments independent from the presence of the cork domain [43]. Though the cork-less FhuA variants usefulness as nano-channel, in nano-compartment systems was further strengthened by the development of a reduction-triggerable [3] or light-triggerable [4] opening mechanism of chemically labeled FhuA1–159 proteins, as will be discussed in more detail in Sect. 5.1.3, the problem of the flexible extracellular loops interrupting channel diffusion properties remained and required for the design of new variants.
5.1.2.2 Smoothing Channel Ends for a Permanently Open Diffusion Channel As it was known from the FhuA WT crystal structure that one of the long extracellular loops, loop 4 (L4, amino acids 318–339; see Fig. 5.2) partially constricts the channel entrance and decreases its diameter to about half the area of the total cross section [44], it was thought that the removal of the cork domain in mutant FhuA5–160 might lead to an increase in loopflexibility, explaining the instabilities in singlechannel potassium conductance the protein variant revealed. Braun et al. therefore deleted L4 amino acids 322–336 in FhuA5–160, ending up with the double mutant FhuA5–160 322– 336. However the variant, as the FhuA5–160
5 Biotechnology
before, increased the conductance of diphytanoyl PtdCho/n-decane lipid bilayer membranes but did not show uniform single-channel conductance. Instead it revealed rapid alternating channel opening and closing [38] and again the active ferrichrome transport mechanism was still intact. Construction of a further variant FhuA5–160 335–355 (deletion of one-third of L4 and half of transmembrane (TM) strand 8) showed uniform single-channel conductance of 2.5 nS in 1 M KCl (2 nS higher than the FhuA5–160 channel conductance), while it did not facilitate active ferrichrome transport [38]. Endriß and Braun consecutively deleted the FhuA extracellular loops in the WT protein and replaced them by the short peptide sequence NSEG(S), to reveal their functions. They found that deletion of L3 or 11 inactivated active ferrichrome transport, deletion of L8 removed receptor activity for colicin M and the phages T1, T5, and ®80, while deletion of L7 removed only colicin M receptor function. Removal of L4 caused resistance against phages T1 and ®80 [45]. While these first FhuA loop-deletion studies were still concerned with the analysis of the protein function, later works were oriented towards the FhuA application in the development of a stochastic single molecule sensing element. Mohammad et al. realized the FhuA’s value for the purpose at hand [46]. Though the suitability of OMPs in general had been recognized before, the so far used proteins, S. aureus poreforming toxin ’-hemolysin [1], E. coli porin OmpG [47] and E. coli porin OmpF [48], hold certain disadvantages as they are multimeric proteins making it difficult to engineer them (For more information on these MPs see Sect. 2.3). It was reported that the heptameric character of ’-hemolysin leads to many combinations of engineered and WT monomers, complicating the separation of a desired modified single sub-unit [49]. Though recently a mutation study led to the development of a monomeric OmpF variant [50]. Furthermore the crystal structure derived inner channel diameters of all three proteins are rather small at the point of highest constriction; one finds for ’-hemolysin: 15 Å [51], for OmpG
5.1 Outer Membrane Protein Modification
101
Fig. 5.2 Schematic representation of further genetic engineering concept starting with FhuA1–159 theoretically leading to regular channel variant (FhuA Reg) lacking 10 of the 11 flexible external loops that tend to close
the channel. The first ever deleted loop (L4) is marked by a black circle (Proteins are shown in New Cartoon representation made with VMD)
13 Å [52] and for OmpF: 6 Å [53, 54], allowing the passage of small, non-bulky molecules with a maximum molecular weight of 700 Da. The FhuA protein combines the general “-barrel advantages with further inborn advantages, such as its monomeric character and much wider elliptic inner channel diameter of 39 46 Å [44, 55]. Mohammad et al. therefore deleted apart from the plug amino acids 1–160 four major extracellular loops (L3, L4, L5, and L11) including L4 that were thought to interrupt channel conductance of FhuA1–160. L3 (amino acids 243–274) and L5 (amino acids 394–419) are large flexible loops folding back into the plugless channel interior, L11 (amino acids 482–704) is another long loop that reaches into the pore interior [56]. Loops were deleted and replaced by short turns of the sequence NSEGS. The resulting mutant FhuAC/4L revealed a high unitary channel conductance in 1 M KCl of 3.9–4.9 nS [46, 57]. Results show that the deleted loop regions actually do interfere with channel conductance, as their removal led to a 9.5-times increase in channel conductance as compared to FhuA1–160. Furthermore conductance occurred uninterrupted over a long time span [57] and as the plug-domain deletion seems already a rather huge protein modification, in the mutant
FhuAC/4L almost one third of all FhuA WT amino acid residues has been removed, indicating to what extend the FhuA structure tolerates protein engineering measures. In a later study the same group showed that the FhuAC/4L structure and conductance remained stable (in contrast to the ’-hemolysin) under harsh conditions, such as acidic pH, low ion concentration or temperatures up to 65 ı C. As a proof of concept that the new FhuA variant may be used as a molecular sensor, it had been reconstituted into planar lipid membranes and was used to monitor the pepsin digest of immunoglobulin G (igG) (fragments of digested igG led to short current interruptions) or to monitor the interaction between the retroviral gag nucleocapsid (NCp7) protein with a DNAaptamer. While free NCp7 blocks the protein channel, interaction with the DNA-aptamer leads to a channel opening [57]. In an unpublished study an attempt toward an even more drastically modified loop-deletion variant of the cork-less FhuA was made [58]. The engineering concept is shown in Fig. 5.2. The FhuA1–159 sequence was used as a template to design a synthetic gene in which apart from the very short loop L1 all 11 extracellular loop-coding regions were cut leading to variant FhuA1–159 Reg (for regular channel structure),
102
5 Biotechnology
Fig. 5.3 Microscopic images at a 400-times magnification of (a) non-induced E. coli BE strain BL 21 (DE3) omp8 with plasmid pET22b C FhuA Reg, showing characteristic rod-shape; (b) E. coli BE strain BL 21
(DE3) omp8 with plasmid pET22b C FhuA Reg after induction, showing altered spherical morphology and formation of cell aggregates (black arrows)
or short FhuA Reg. Loop amino acid residues were deleted apart from terminal non-“-strand loop residues (3–5 amino acids), these residues were kept as “-strand connections (compare FhuA topology in [44]). Including the removed cork-domain 321 amino acid residues had been deleted, leading to a protein with 413 amino acids and an expected molecular weight of 46 kDa. The synthetic gene had been obtained from GeneArt (Regensburg, Germany) and cloned into E. coli expression vector pET22bC (Novagen, Merck, Darmstadt, Germany). Expression of the protein was attempted as previously described [40] using E. coli strain BL 21 (DE3) omp8 (FhsdSB (rB- mB-) gal ompT dcm (DE3) lamB ompF::Tn5 ompA ompC). The general OMP expression methods will be discussed in greater detail in Sect. 5.3. However no over-expressed protein with the expected size could be isolated from the outer membrane. It was observed that expressing E. coli cells showed an altered spherical cell morphology (Fig. 5.3a), while non-expressing cells showed the typical rod-shape (Fig. 5.3a). In liquid culture expressing cells tended to form large macroscopic aggregates. These changes in E. coli cell morphology are known for cells that encounter defects in their cell
division mechanism due to mutations, resulting in cells forming spheres that do not grow or propagate further [59] and also E. coli cells lacking the outer membrane turn elliptical or spherical [60]. The so-called L-form E. coli are spherical, osmo-sensitive cells due to a partial or total loss of their cell envelope occurring either spontaneous or after treatment with “-lactam antibiotics that inhibit murein synthesis [61]. In the case at hand the morphological changes were most likely due to the FhuA Reg over-expression and they might be connected to osmotic imbalances upon the insertion of the regular and expected to be permanently open channel protein into the outer membrane or to other toxic effects of the protein. A similar effect has been reported for the Rhodobacter blasticus porin; its expression into the outer membrane of E. coli resulted in cell lysis before the protein could be produced in sufficient amount [62]. The same phenotype as for FhuA Reg had been observed when expressing an FhuA1–159 variant with increased channel diameter, as will be explained in the following. In that case the problem could be overcome by the optimization of culture conditions including a change of growth media. A possible approach to produce FhuA variant FhuA Reg might be its expression into inclusion bodies with subsequent
5.1 Outer Membrane Protein Modification
103
Fig. 5.4 Schematic representation of further genetic engineering concept again starting from FhuA1–159 leading to a channel variant with expanded channel diameter (FhuA Exp), in which the last two N-terminal “-strands
have been duplicated (black circle) (Proteins are shown in New Cartoon representation made with VMD; the FhuA Exp image has been edited using GNU image modification program GIMP 2.6 to add the duplicated sheet)
refolding to native state or a cell-free expression approach in the future (see also Sect. 5.2.1). A similar loop deletion study had been carried out recently on the OmpF porin of Yersinia pseudotuberculosis. The deletion of three major loops L1, L6 and L8 however did not lead to a change in channel conductivity, but rather had an effect on the antigenic structure of the mutant porins, as revealed by immune-blotting and ELISA. Protein variants were expressed heterologous in E. coli and obtained from inclusion bodies [63].
to an OMP should be possible [64, 65]. A so derived increase in inner channel diameter would further widen the spectrum of substances that can be translocated through the protein channel. The gene was designed considering the following concept: As a proof of concept only one “-sheet (two strands) was to be added, without introducing completely new genetic information, leading to a barrel with 24 strands as it can be found also in nature [66]. Therefore the amino acid sequence of one existing sheet was to be copied from the template protein FhuA1–159 and pasted into the sequence of the new protein FhuA1–159 Exp (for expanded channel diameter), or short FhuA Exp. For this purpose the first two N-terminal strands (30 amino acids) were chosen for duplication as they are connected by the short loop L1, in this way the protein N-terminal sequence is conserved and N- and C-terminus are still expected to close by hydrogen bonding to form the intact barrel [67]. The engineering concept is shown in Fig. 5.4. The resulting protein has 628 amino acids and an expected molecular weight of 66.3 kDa. When assuming a simple regular polygonal geometry (hendecagon for FhuA1-159 and dodecagon for FhuA1-159 Exp), with constrained side length given by the “-sheets connecting hydrogen bonds, the expected diameter increase can
5.1.2.3 Increasing the Number of “-Strands for a 0.4 nm Wider Channel Diameter As the FhuA and some of its variants had been found useful as nano-channel biosensors or to functionalize artificial lipid or polymer vesicle systems and as its tolerance towards extensive sequence mutations (especially deletions) had been shown in several studies, the attempt to further increase the channel diameter of variant FhuA1– 159 by inserting “-strand-forming amino acids seemed not too far-fetched. This was further backed up by a study in which the three-stranded “-sheet of Borrelia burgdorferi protein OspA was successfully extended by two further strands by duplicating a beta-hairpin that formed a new sheet, indicating that the addition of new sheets
104
5 Biotechnology
Fig. 5.5 Microscopic image of FhuA Exp pre-crystals
be estimated: The crystal structure derived FhuA WT inner channel diameter is 4.2 nm [56]. Based on the FhuA WT apothem, the expected FhuA1-159 Exp inner channel cross section is 4.6 nm as calculated from the apothem ratio, resulting in a 16 % increase in channel surface area [67]. The on-paper reverse-translation derived synthetic gene was expressed in E. coli BL21 (DE3) omp8 using standard expression conditions as used for FhuA1–159 [40]. However for the FhuA Reg variant no protein of the expected size could be isolated and cells again showed the described, spherical morphology (see Fig. 5.3). Obtained optical densities were poor (unpublished results, [68]). However after optimizing media and growth conditions (see Sect. 5.2.1 for more details) the protein was over-expressed without affecting the expression host cells and could be isolated from the outer membrane. Channel functionality was verified after protein reconstitution into lipid vesicles, by measuring TMB-conversion enzyme kinetics of vesicle entrapped horseradish peroxidase (HRP), as described in [40]. The channel proved to be functional and TMB conversion occurred 17 % faster than with reconstituted FhuA1–159 cor-
relating with the 16 % increase in inner pore surface area. The FhuA Exp structural integrity had been verified by CD spectroscopy, revealing 63 % “-sheet contribution [67]. In a first attempt to crystallize FhuA Ext to analyze its three dimensional structure by X-Ray diffraction, pre-crystals were obtained (see Fig. 5.5) after an intensive screening of more than 600 different crystallization conditions carried out by Prof. E. Mizohata (Division of Applied Chemistry, Graduate School of Engineering, Osaka University) (unpublished results; [68]). In conclusion the study demonstrated that the FhuA structure tolerated larger insertion mutations as well and that the “-sheet secondary structure information is fully contained in the existing FhuA amino acid sequence, therefore the simple “copy-paste” strategy was successful and led to a protein with increased inner channel cross-section and it seems likely that insertion of further strands should be possible. To the author’s knowledge no similar approach of increasing the inner pore diameter of an OMP has been yet followed in other works. Purified variant FhuA Exp was as mentioned successfully reconstituted into lipid vesicles (of E. coli lipid extract) [67] and it was
5.1 Outer Membrane Protein Modification
attempted to reconstitute the protein into BAB triblock copolymer vesicles, formed by the commercially available, cost effective polymer PIB1000 -PEG6000 -PIB1000 (PIB D Polyisobutylene; PEG D Polyethylene glycol), forming impermeable membranes with a hydrophobic thickness of 5 nm [69]. However FhuA Exp did not insert into PIB1000 PEG6000 -PIB1000 membranes, due to their thick hydrophobic region and the resulting hydrophobic mismatch, the same had been reported for FhuA1–159 [69]. However the use of impermeable, commercially available and biocompatible polymers (e.g. PIB1000 -PEG6000 PIB1000 ) can be of great advantage for the design of functionalized nano-compartments, therefore OMP variants that can be reconstituted into such polymer membranes are desirable. A FhuA1– 159-based variant was especially engineered for this purpose (see below).
5.1.2.4 Overcoming the Hydrophobic Mismatch Polymer vesicles or polymersomes can as lipid vesicles (liposomes) be utilized as nano-sized encapsulation devices for applications such as delivery systems, bio-mimetic membranes, biomedical imaging tools, as protection devices for labile substances or as nano-reactors for sealed-in chemical or enzymatic reactions [70]. In contrast to liposomes, polymersomes formed by self-assembling synthetic amphiphilic block copolymers have been reported to possess superior biomaterial properties, such as better chemical and physical stability and impermeability [71]. The polymersome (or polymer membranes in general) functionalization by inserted MPs is therefore a potent way to design new hybrid materials, such as proteinpolymer nano-compartment systems. As mentioned before the FhuA1–159 had been successfully reconstituted into PMOXAPDMS-PMOXA polymer vesicles [3, 40]. The PMOXA-PDMS-PMOXA polymer had been synthetized to be used for the formation of hollow sphere (polymersome) structures with biomimetic membranes [72, 73]. It had been found suitable for the functional reconstitution
105
of E. coli OmpF [74] before it was used in combination with FhuA1–159. Here the standard strategy for the functional reconstitution of MPs into polymeric membranes was followed. This strategy includes the specific design of a polymer membrane’s characteristics to fit the protein to be reconstituted. The membrane should be as thin and as fluid as possible, to minimize the energetic penalty when exposing a nonpolar/polar interface. However membranes formed by block copolymers can be rather thick, as they vary in thickness from 5 to 22 nm, while “natural” phospholipid membranes are only 3–4 nm thick. The polymer membrane greater thickness may cause problems in protein insertion, due to the hydrophobic mismatch between MP transmembrane regions and hydrophobic membrane parts [75]. Since block copolymers that assemble to membrane systems with thick membranes or hydrophobic portion might still have otherwise desirable features (such as mechanical stability, biocompatibility, cost effectiveness, or commercial availability), a protein nano-channel that can be tailored to fit the polymer is of superior advantage. For a first proof of concept the FhuA protein was chosen for its robust structure and the possibility to introduce vast mutations, and a BAB triblock copolymer PIB1000 -PEG6000 -PIB1000 was chosen for its commercial availability and selfassembling ability. The PIB1000 -PEG6000 -PIB1000 polymer membrane hydrophobic part is 5 nm thick, while the FhuA1-159 hydrophobic portion is only 3 nm thick [69]. Therefore based on FhuA1-159 as a template, a new variant sequence was planned by “copy-pasting” the last five amino acids of each “-strand, leading to a total insertion of 110 amino acids thus increasing the expected hydrophobic length by 1 nm to reduce the hydrophobic mismatch of FhuA insertion into PIB1000 -PEG6000 -PIB1000 polymersomes (design concept shown in Fig. 5.6). The resulting protein FhuA1-159 Ext (for extended channel length), or short FhuA Ext, consisted of 665 amino acids with an expected molecular weight of 74 kDa. The protein was obtained from the outer membrane after cloning
106
5 Biotechnology
Fig. 5.6 Schematic representation of genetic engineering concept starting from FhuA1–159 leading to channel variant with extended channel length (FhuA Ext), by
duplicating the last five amino acids of each “-strand on the periplasmic protein side (black circle) (Proteins are shown in New Cartoon representation made with VMD)
the synthetic gene and expressing the protein in E. coli BL21 (DE3) omp8. The standard FhuA1-159 membrane extraction and protein solubilization protocol [40] had to be modified due to the proteins increased hydrophobicity [69] (for details see Sect. 5.2.1). Again structural integrity could be verified by determining secondary structure by CD spectroscopy and the determined “-sheet content was 75 %. In contrast to FhuA1-159 and FhuA Exp variant FhuA Ext was successfully inserted into PIB1000 -PEG6000 PIB1000 polymersome membranes, as determined by the already mentioned TMB conversion kinetics measurements after TMB translocation through the polymer embedded protein channel [69]. Figure 5.7 shows schematically how the elongated FhuA Ext is expected to reside in membranes of the PIB-PEG-PIB type, with the protein TM region thickness overlapping better with the thickness of the hydrophobic membrane portion (top) than it is the case for the shorter FhuA1–159 [69]. In order to determine the PIB1000 -PEG6000 PIB1000 vesicle size and shape with and without reconstituted FhuA Ext protein, dynamic light scattering (DLS) measurements were carried out after polymersome purification
(details on the DLS method can be found in Sect. 6.1.2), showing that plain polymer vesicles were spherical and had an average diameter of 242 nm (when in presence of detergent 2-Hydroxyethyloctylsulfoxide – OES, used to solubilize FhuA Ext), while vesicles with reconstituted protein (in 2Hydroxyethyloctylsulfoxide) had a larger diameter of 279 nm on average and were also spherically shaped (unpublished data; [76]). The results are shown in Fig. 5.8. Diameters of polymersomes with reconstituted FhuA Ext were 37 nm larger than diameters of plain polymersomes. This finding was rather interesting, as from it might be obtained a first clue on the number of proteins interacting with a polymer vesicle, in case the membrane thickness (and polymer volume) remains constant in presence of or without the protein. As such information cannot be obtained by simple DLS measurements; in the future multi-angle light scattering measurements instead might provide information on the membrane thickness and shape. Other analytical methods such as Cryo TEM may be equally suited to compare the polymersome membrane thickness with and without channel protein (For an overview on the
5.1 Outer Membrane Protein Modification
107
Fig. 5.7 Schematic representation of FhuA Ext (top) and FhuA1-159 (bottom) within triblock copolymer PIB1000 -PEG1500 -PIB1000 membranes. The hydrophobic TM regions of FhuA1-159 Ext (4 nm) and FhuA1-
159 (3 nm) are indicated by lines; the duplicated part of FhuA1-159 Ext is indicated by a broken line (Proteins are shown in New Cartoon representation made with VMD) [69]
state of the art analytical methods to characterize protein functionalized membrane systems see Chap. 6, a focus on DLS methods is given in Sect. 6.1.2). None of the described FhuA variants has been crystalized and analyzed by X-ray diffraction analysis, yet. Nevertheless the functional analysis by indirectly monitoring substrate/product diffusion through liposome or polymersome reconstituted channel proteins, using a vesicle enclosed enzyme or by more accurate single channel conductance measurements or CD spectroscopy derived information on protein secondary structure can give good clues on the correct folding of the new variants. Table 5.1 gives a summary of secondary structure information obtained by CD spectroscopy for FhuA WT and all variants with changed channel
geometry. Each of the variants shows a rather high amount of “-structure indicating that the barrel structure is retained. A definite proof on the accurate folding and tertiary structure of a newly expressed protein variant however can be obtained by X-ray diffraction analysis of protein crystals, protein NMR or Cryo-EM (see also Chap. 3).
5.1.3
Modifications for Chemical/Physical Triggering and Specificity
Especially for drug-release or more general compound-release systems composed of a lipid or polymer membrane or vesicles and an OMP channel, a triggerable (irreversible) or
108
5 Biotechnology
Fig. 5.8 DLS measurement results for PIB1000 -PEG6000 PIB1000 vesicles harboring FhuA Ext (solubilized in OES detergent) and plain polymersomes in presence of OES
(grey crosses). Z-average hydrodynamic diameters (Dh) are given in nm for both samples
Table 5.1 CD spectroscopy derived secondary structure amounts of FhuA WT and variants with changed channel geometry FhuA variant FhuA WT FhuA 1–159 FhuA C/ 4L FhuA Exp FhuA Ext
% “-structure 51 49–65 60.3 63 75
% ’-helix 2 3–13 3.7 7 20
switchable (reversible) channel opening/closing mechanism is desirable. Drugs or chemical substances contained in a nano-compartment can be transported to a particular target, at which a triggered channel opening leads to compound release, without the nano-compartment being destroyed (see also Sect. 6.2.1). As proteins are composed of amino acids of which some expose reactive side-chains the most obvious way to introduce an opening/closing mechanism to a channel protein is to selectively and chemically modify certain amino acid side chains with bulky substances that are able to efficiently seal the protein pore. The bond has to be cleavable upon an external stimulus, so to release the bulky label resulting in channel opening. Proteins as all biological molecules restrict chemical labeling
% random coil 47 23–37 37.2 30 5
Reference [42] [3, 4, 41] [46] [67] [69]
reactions to certain conditions that ensure the maintenance of biomolecule integrity. Chemical reaction to modify proteins generally have to be performed in aqueous environment (OMPs often tolerate low concentrations of organic solvents), at ambient temperature and preferably at neutral pH. For OMPs with a narrower inner channel diameter unlabeled amino acids can be used to introduce channel opening/closing switches, as amino acids, based on their pKa values, answer with a charge change upon changes in pH values. This purpose however generally requires for protein modification by site-directed mutagenesis, e.g. the introduction of a certain amino acid at multiple inside facing positions, so to form a ring at the channel constriction site [77, 78].
5.1 Outer Membrane Protein Modification
Such amino acid substitutions by site-directed mutagenesis are not only useful in compoundrelease applications but furthermore are of importance for OMP based biosensor systems, too. The addition of charged rings to an OMP pore inside for instance can be used to detect charged molecules [78]. Single substitutions might change the specificity of OMPs that specifically transport one substance or substance class [79].
5.1.3.1 Chemical Modification (CM) Several amino acid side chains are suitable to be chemically modified due to their reactive nature. The primary amine moiety of the amino acid lysine for instance can act as a nucleophile and forms amides with N-hydroxysuccinimide (NHS) esters, releasing the NHS. As the reaction occurs spontaneously at pH 8–9 and as no reagents such as bases are necessary it is well suited to label Lys residues in a protein. Sulfonated NHS ester derivatives reveal enhanced water solubility and are therefore preferable. Many protein labeling reagents based on NHS esters or NHS ester derivatives are commercially available. Isothiocyanides likewise react with primary amines under the formation of thiourea, at basic pH (9–9.5). Furthermore the Lys amino can be used for the reductive amination of aldehydes under reversible imine formation. The imine can then be reduced irreversibly to a secondary amine using hydrides such as NaBH3 CN. Furthermore Lys is able to form peptide-bonds with halogenated carboxylic acids by a nucleophilic substitution reaction. Arginine residues can amongst others be modified by methyl glyoxal under formation of a pyrimidine derivative. Carboxylic groups of glutamate or aspartate can be labeled by carbodiimide reagents under peptide-bond formation. The cysteine thiol side residue is a stronger nucleophile than the amino group, therefore cysteine generally reacts faster than Lys. But since Cys residues in proteins form disulfide bonds, proteins have to be treated with reducing agents prior to performance of a labeling reaction. Free thiols can then be modified by a variety of different reactions including the reaction with
109
maleimides, alkylation reactions or disulfide bond formation, to name a few. Arg, His or Tyr are other targets for the introduction of chemical modifications [80]. The interested reader may find detailed information on the bioconjugate chemistry including step-wise description of reactions, commercially available reagents, and practical applications of labeled bio molecules in the book “Bioconjugate Techniques” [81]. In case of chemical amino acid modification to introduce an opening/closing trigger or switch to the inside of a protein channel, the decision for a chemical labeling agent and labeling position depends on several points: 1. The amino acid side chains to be modified have to be exposed within the channel interior. Therefore information on the target OMP structure is a pre-requisite for the selection of the type of amino acid to be labeled. It might be necessary to introduce one or several amino acid labeling targets to the inside of an OMP channel by site-specific mutagenesis. 2. It has to be ensured that labeling of certain positions will not destabilize a protein structure. Theoretical considerations prior to labeling experiments are advisable. 3. The labeling agent has to be bulky enough to block the OMP channel and is preferably commercially available. 4. It has to be possible to cleave off the bulky label to open the channel by an outside stimulus (e.g. reduction agents, light, pH or temperature changes). A collection of literature examples for the OMP amino acid chemical modification to introduce trigger mechanisms or to change the conductance of nano-pore sensing elements is given below: CM: Reduction Trigger In case of the FhuA1–159 the amino acid lysine has been successfully biotinylated using reagent 2-[biotinamido]ethylamido-3,30 dithiodipropionic acid N-hydroxysuccinimide ester [3] or pyridylated using 3-(2-pyridyldithio) propionic-acid [3, 82] (see Fig. 5.9). Both labels contain a disulfide bond that allows removal by
110
5 Biotechnology
Fig. 5.9 Schematic representation of pyridylation of FhuA1-159 (or variants) Lys residues. Labeling leads to effective channel blocking as determined by calcein release assay or indirect determination of TMB diffusion through the channel by measuring HRP TMB conversion
kinetics. The label can be cleaved off by providing reducing conditions (e.g. addition of DTT), cleavage results in channel opening (Protein New Cartoon representations made with VMD)
addition of reductive agents, such as dithiothreitol (DTT) permitting reduction-triggered channel opening. Channel blocking properties of both agents were analyzed utilizing calcein release kinetics measurements or the TMB-HRP assay system upon reconstituting labeled proteins into PMOXA-PDMS-PMOXA polymersomes [3] or liposomes [82]. FhuA1–159 contains 29 lysines of which 19 are located on the OMP surface, 6 are buried within the channel and 4 sit on the barrel rims [3]. As the surface exposed lysines are located in the proteins TM region they are covered by detergents after membrane extraction and therefore unlikely to be modifiable [82]. A mutational study on the six channel buried Lys came to the conclusion that pyridylation of Lys at amino acid position 556 is sufficient to sterically hinder compound flux. This finding has been in accordance with MD simulation based B-factor analysis identifying Lys556 as the most rigid of the investigated lysine residues [82].
The same bioconjugation chemistry had been used to biotinylate Lys residues in FhuA Ext and FhuA Exp. FhuA Ext contains 29 lysine residues of which 22 are facing the outside, four are buried within the channel and three are localized at both channel rims. Labeling of PIB1000 -PEG6000 PIB1000 polymersome inserted FhuA Ext lysines led to a decrease in TMB conversion speed and thus TMB influx (about five-times more slow than for unblocked FhuA Ext), but did not lead to a complete channel blocking as it was found for liposome reconstituted FhuA1-159 [69]. In FhuA Ext one finds 31 Lys residues of which 21 are located on the OMP surface, 6 face the channel inside and 4 are on both barrel ends. Biotinylation in this case again led to a decrease in TMB conversion speed as a measure for compound flux through the liposome inserted protein channel. However TMB conversion occurred only 3-times more slow through the blocked FhuA Exp as compared to the open barrel, most likely due
5.1 Outer Membrane Protein Modification
to the increased channel diameter [67]. These findings show that the label choice depends also on the newly designed protein features and an increase in length and especially in diameter necessitates the use of bulkier agents to guarantee efficient channel sealing. CM: Light Trigger Modification of lysine residues in FhuA furthermore allowed the introduction of a light triggerable channel opening system. For this purpose Lys amino acids were labeled using the well-known photo-cleavable compound 6-nitroveratryloxycarbonyl chloride (NVOCCl) [83]. NVOC-Cl reacts with the Lys primary amine by nucleophilic substitution. The LysNVOC complex is cleaved upon irradiation with light of a wavelength of 366 nm releasing the easily detectable yellow compound onitrosobenzaldehyde and CO2 [4]. Channel blocking was monitored using the TMB/HRP assay system and Lys556 labeling was again found to be sufficient to completely block TMB diffusion through the protein pore [4]. CM: Further Examples Cysteine amino acid labeling of the FhuA WT with biotin-maleimide [N-biotinoyl-N0 (6-maleimidohexanoyl)-hydrazide] had been carried out in order to obtain structural and functional information on cysteine residues in the surface exposed loop regions L4 and L11. As labeling was possible only with Cys in L4 after reduction, while C-terminal cysteines reacted only after replacement of one Cys and subsequent denaturation it was deduced that all four cysteine residues form disulfide bridges [84]. As these naturally occurring Cys in the FhuA protein do not easily react with thiol marking compounds single Cys may be introduced to the protein as labeling targets. In fact two Cys were newly added to L4 of which one proved to be more reactive than Cys in the WT [85]. However newly added Cys might lead to incorrect folding by formation of unwanted disulfide bridges and proteins should be expressed under reducing conditions.
111
The ’-hemolysine heptamer had been modified by covalently attaching one 5,000 Da (or 3,000 Da) poly(ethylene glycol) (PEG) molecule to an introduced cysteine in the protein pore (by mixing modified and unmodified subunits in suitable ratios heptamers with a single PEG chain were obtained). As the PEG was attached via disulfide bond formation, it could be cleaved off by adding DTT. Single channel recordings showed that the modification reduced the pore conductance by 18 % [86]. This approach broadens the design possibilities for new biosensors by attaching polymer chains that respond to analytes. The attachment of responsive polymers to the interior of an OMP channel is furthermore useful to design new switches and triggers to open or close the channel. Non-covalent modification of ’-hemolysine by “-cyclodextrin as adapters for organic molecule analytes was reported. “-cyclodextrin blocked the WT channel by 64 %, it is thought to be retained in the channel by one of two restriction rings. As it is known that adamantane derivatives bind to cyclodextrines two adamantine based model molecules (2-adamantamine and 1-adamantanecarboxylic acid) were tested on their ability to further block the channel. Both molecules led to a further decrease in channel conductance, proving the adapter qualities of the channel retained cyclodextrin [87].
5.1.3.2 Site-Specific Mutagenesis (SSM) Since easy to use site-directed mutagenesis kits for amino acid substitutions, point mutations and small deletions/insertions are available from various suppliers, mutagenesis became a lot more practical and is a useful tool to change a protein’s functional features, provided that a protein’s primary sequence is known and one has some information on its tertiary structure. As mentioned, site-specific mutagenesis to substitute amino acids can be used to introduce amino acid based switches. Furthermore it allows the introduction of protease recognition sites that permit channel opening upon cleavage of an introduced loop structure [88]. Some representative
112
examples of OMP site-specific mutagenesis resulting in property changes (e.g. substrate specificity) or the introduction of opening/closing switches are given in the following: SSM: OmpF pH Switch The E. coli porin OmpF has been used for the development of a pH based channel release switch. OmpF was chosen as it had previously been successfully reconstituted into PMOXA-PDMSPMOX block copolymer vesicles [74]. As histidine has a pKa value of 6, the introduction of a ring of histidines at a channel constriction site was thought to permit the release control of positively charged molecules by shifting the pH from 5 to 7. As target position was chosen a constriction site consisting of two amino acid half rings with three amino acids each (Arg42, Arg82 and Arg132; positively charged and Asp113, Glu117 and Asp121; negatively charged). All six amino acids were substituted by His. The pH dependent release was demonstrated by monitoring the acridine orange translocation by Fluorescence Correlation Spectroscopy at different pH values [77]. SSM: Zinc Binding ’-Hemolysin A Zn(II)-binding ’-hemolysin subunit was designed by substituting amino acids Asn123, Thr125, Gly133 and Leu135 with histidine, while Thr292 was substituted by cysteine. The four histidine imidazole sidechains act as ligands to Zn(II), while the single cysteine was modified by 4-acetamido-40 -[(iodoacetyl)amino]stilbene2,20 -disulfonate, leading to a change in SDSPAGE (sodium dodecyl sulfate polyacrylamide gel electrophoresis) electrophoretic mobility allowing heteromer separation [89]. SSM: Nano-pore DNA-Sequencing MspA Nano-pore DNA sequencing, using a membrane inserted protein pore has received much interest [90]. The method is based on single channel current measurements, exploiting that current changes when single-stranded DNA passes through the channel depending on the DNA sequence. The homo-octameric Mycobacterium smegmatis porin A (MspA) is one of the most stable proteins so far known [91]. Due to its
5 Biotechnology
geometrical features, i.e. pore constriction of 1 nm length and 1 nm width surrounded by sections with much larger diameter [92], MspA is rather suited as a nano-pore DNA sequencing device. The MspA pore has been engineered to allow electrophoretic passage of DNA by replacing negatively charged by neutral amino acids in the constriction site. Furthermore the addition of 24 positively charged residues at the channel entrance of another mutant led to an increase of DNA translocation through the protein pore of 20-times as compared to WT [93]. A combination with the latter mutant and the phage DNA polymerase phi29, allowed the control of the DNA translocation rate [94]. A combination of site-directed mutagenesis and subsequent chemical modification of the newly added amino acid(s) can be used to limit labeling to a defined position, adding a further element of control. The introduction of a single cysteine to the cysteine free ’hemolysine WT made possible the development of a light-activation of pore-formation by cysteine modification with 2-Bromo-2-(2nitrophenyl)acetic acid (BNPA). The BNPA modified protein lost pore-forming abilities on rabbit erythrocyte membranes, while the cleavage of the label by irradiation with near UV light restored pore-forming properties [95]. All variants of “-barrel MPs have to be isolated and purified prior to characterization and to application. Apart from the multimeric pore forming toxins whose subunits are expressed in soluble form, OMPs are generally expressed into the bacterial outer membrane and their overexpression and retrieval can be rather challenging due to their localization and hydrophobic character as will be discussed in the following sections.
5.2
Outer Membrane Protein Production: Challenges and Solutions
When it comes to their over-expression, isolation and purification OMPs are, as all integral MPs, classified as challenging proteins. However when
5.2 Outer Membrane Protein Production: Challenges and Solutions
working with bacterial MPs one has access to a number of rather developed standard procedures from which to start from, when a protocol for a new OMP variant has to be set-up. Protocols for eukaryotic MPs are less established, however certain general rules and guidelines apply to both pro- and eukaryotic expression and in some cases the heterologous expression of a eukaryotic OMP in bacteria such as E. coli might be possible. This section and its sub-sections will deal mainly with the production of bacterial OMPs, as they are the main targets for the developments of new nano-materials. In principle MPs can be obtained from their natural source the inner or outer membrane subsequent to their over-expression. This works however only, if the target protein is highly abundant in the respective membrane, as it is the case with the ’-helical MP bacteriorhodopsin that can be obtained from the purple membrane of Halobacterium halobium [96] or for an outer membrane derived porin of Rhodobacter capsulatus [97]. In most cases though the low abundance of a MP in the corresponding membrane prohibits it’s recovery from the natural source. Therefore most often the ORF coding for the target outer membrane protein will be cloned into a suitable over-expression vector allowing homologous or heterologous over-expression leading to higher expression and membrane insertion levels. Three of the first OMPs expressed from plasmidDNA are E. coli malotoporin LamB [98], the Salmonella typhimurium sucrose porin ScrY produced in E. coli [99] and the E. coli FhuA [100]. The outer membrane of the expression host bacterium, rich in the desired OMP, will then be isolated and the protein has to be solubilized and purified prior to its structural or functional characterization or its use as nano-material. As the membrane offers only a limited space, it restricts the number of correctly folded and inserted OMPs, lowering the obtainable yield. Therefore if the protein shows the ability to spontaneously refold from a fully or partially unfolded state, over-expression into inclusion bodies should be considered, as much higher yields can be reached. This method can be especially useful for engineered channel forming
113
OMPs with changed geometry or with a higher than average hydrophobicity (when in folded state). No matter if an OMP has been obtained from the outer membrane or from inclusion body material, the folded/refolded OMP has to be solubilized in detergent-, polymer-, or organic solvent solutions and extensive screening for the right solubilization agent (depending on protein and purpose) may become necessary. Furthermore the subsequent purification steps that are crucial for crystallization purposes and many applications are often selected following a trial-and-error approach, starting out from standard protocols, as each new variant might require the development of a new purification procedure. But since “many roads lead to Rome”, good knowledge of standard OMP production protocols as well as more unconventional methods can be of advantage and sometimes, as it is so often in science, one has to be bold and try something entirely new. Here a summary of the well-established OMP over-expression, solubilization and purification methods, as well as of newly developed/less conventional protocols, with a focus on adaptations suitable for the production of novel developed OMP variants, will be given. Of further consideration will be the various means of expression quality, protein functionality and yield control.
5.2.1
Conventional OMP Isolation from the Outer Membrane
OMPs are either derived from the outer membrane of an expression host or expressed into inclusion bodies from where they can be isolated and refolded [101]. Recently cell-free expression techniques have been recognized as promising tools for the production of MPs in general [102]. Both expression into inclusion bodies and by cell-free expression systems avoid the limited membrane surface bottle-neck. Apart from the rare case in which a target MP can be obtained from its natural source without preceding over-expression, OMPs have to be over-expressed to reach sufficient concentrations for characterization purposes or for the
114
application as nano-sized building blocks. When working with a new OMP one will generally start by choosing a fitting expression host and in case of bacterial OMPs it is often possible to over-express the protein in the natural host (homologous) [101], while for eukaryotic, higher organism mitochondrial or plastidic OMPs heterologous expression into the outer membranes of Gram-negative bacteria or yeast mitochondrial outer membranes can be advantageous as unicellular organisms are more easy to handle and have much shorter generation times. However when expressed in heterologous systems, OMPs as all MPs can be toxic for the host or misfolded and expression levels are often low due to the limited space offered by the membrane [103]. Furthermore OMPs fold and insert solely into the Gram-negative bacterial or mitochondrial/chloroplasts outer membrane and can therefore not be membrane-expressed into Gram-positive bacteria or for instance the yeast plasma membrane. The most widely used host for the heterologous over-expression of Gramnegative bacterial OMPs is E. coli, as it is a Gramnegative organism itself and as it is a well-studied model organism that is widely used for heterologous protein expression. Numerous bacterial OMPs such as the Chlamydia psittaci major outer membrane protein MOMP [104], Haemophilus ducreyi OMP D15 [105], Chlamydia trachomatis MOMP [106], Pasteurella multocida OmpH [107], Neisseria meningitidis PorA [108] to name but a few, are examples for OMPs that have been over-expressed in E. coli. E. coli however is often not suitable for the recombinant expression of eukaryotic OMPs. When a new variant of an already known OMP is to be expressed then often a good starting point is to choose the same expression host, conditions and protocols as used for the parent OMP. If necessary, procedures can then be adapted and optimized for the newly engineered protein. The general work-flow for the conventional OMP production in bacterial host cells is shown in Fig. 5.10. As step one in Fig. 5.10 shows after selecting the host organism, the target OMP coding ORF is cloned into a suitable over-expression vector. These vectors are usually the same as used
5 Biotechnology
for the expression of soluble proteins therefore the same selection criteria are applied. Further considerations are that affinity tags useful for purification purposes, such as the Hexa-His-tag should be fused to the protein C-terminal end, so to not interfere with the N-terminal signal peptide sequence leading the protein to the outer membrane. The elongation of the His-tag from 6 to 8 or 10 histidines can increase purification success, while it might result in lower expression levels [109]. In case of OMPs however internal His-tags can be the better solution, as C- and N-terminus close the barrel by hydrogen bond formation and an affinity tag may inhibit the correct closing [100] (see also Sect. 5.2.4.1). A standard combination of expression host and expression plasmid (not only for MPs) is E. coli BL21 (DE3) and the pET vector system. Both are readily available and generally present in any molecular biology laboratory. The E. coli BL21 (DE3) [F– ompT gal dcm lon hsdSB (rB mB ) œ(DE3 [lacI lacUV5-T7 gene 1 ind1 sam7 nin5])] is an E. coli B strain carrying a œ prophage carrying the T7 RNA polymerase gene and lacIq , usable with plasmids containing the T7 promotor. Expression of the T7 RNA-polymerase is under the control of an isopropyl “-D-1-thiogalactopyranoside (IPTG) inducible lac promoter [110]. The pET expression vector system is one of the most widely used systems for the expression of recombinant proteins in E. coli since the T7 RNA polymerase is highly specific for the T7 promoter on the plasmid (host genes will not be expressed) and since it is highly active leading to high expression levels. The pET vectors have been derived from plasmid pBR322. Target cloned ORFs are under control of strong bacteriophage T7 transcription and translation signals. Upon induction with IPTG the T7 RNA polymerase is expressed in E. coli host cells, resulting in plasmid protein expression [110]. This element of control avoids “leaky” expression of the plasmid coded ORF (i.e. low expression in non-induced cells), this is of great importance since OMP overexpression can lead to toxic effects or reduced growth rates [111].
5.2 Outer Membrane Protein Production: Challenges and Solutions
115
Fig. 5.10 Schematic representation of conventional membrane expressed OMP production process flow. 1 – ORF cloning into expression vector (grey circle), transformation of host cells (screening for best expressing cells and conditions not shown) and cell growth C protein expression in liquid culture; 2 – cell harvest (e.g. by centrifugation); 3 – cell disruption and outer membrane
isolation (by differential centrifugation); 4 – protein solubilization (includes solubilization agent screening); 5 – expression yield and protein purity control (e.g. by SDSPAGE); 6 – characterization and application of purified OMP variant (Protein New Cartoon representations made with VMD)
For the over-expression of OMPs E. coli BL21 (DE3) derivatives lacking genes for the major OMPs, such as E. coli BL21 (DE3) omp8, are available. E. coli BL21 (DE3) omp8 expresses only a small subset of naturally occurring E. coli porins assisting the purification of over-expressed barrel proteins [112]. Cloning, transformation, clone screening, small-scale expression and expression follow standard protocols [113] and will not be further discussed. Generally it can be said though, that decelerated expression rate and bacterial growth rate facilitates the functional OMP expression into the outer membrane of E. coli as it avoids unwanted inclusion body formation and toxic effects of OMP over-expression (but lower expression levels can lead to difficulties in OMP isolation and purification). Lower growth
and expression rates can be achieved by lowering growth temperature after induction (e.g. to 30 ı C), by low inducer concentration or by utilizing a vector system with a weaker promoter. Toxicity can be lowered furthermore by shortening the time of induction [99]. In case of OMPs that might lead to host-cell osmotic imbalances (i.e. caused by OMPs that have high conductivity) carefully considering the expression media can lead to better results. As described in Sect. 5.1.2.3. FhuA variant FhuA Exp, an OMP with a widened channel diameter [67], led to poor cell-densities and protein expression-rates when expressed under conventional conditions (i.e. TY or LB media and a growth temperature of 37 ı C [68]). Lowering the expression temperature to 30 ı C, combined with using the hypertonic NaPy
116
5 Biotechnology
Table 5.2 Fatty acid composition of E. coli outer cell membranes applying different cultivation temperatures [115] Fatty acid Myristic acid Palmitic acid Palmitoleic acid Oleic acid Hydroxy-myristic acid Ratio unsaturated/saturated
% of overall fatty acid content 10 ı C 20 ı C 4 4 18 25 26 24 38 34 13 10 2:9 2:0
medium, such as it is used for L-form E. coli [114] greatly enhanced the obtained cell density and allowed the new OMP variant to be isolated. Another important effect of the cultivation temperature is that it influences the lipid composition of the bacterial outer membrane leading to differences in membrane fluidity [115]. Though E. coli can regulate to some extend the ratio of unsaturated/saturated membrane lipid fatty acids at different temperatures due to the FabF enzyme [116]. Nevertheless the E. coli outer membrane composition is temperature dependent and Table 5.2 lists the fatty acid composition of the E. coli outer membrane at different cultivation temperatures. It shows that at lower temperature the membrane contains more unsaturated fatty acids rendering the membrane less fluid (see also Chap. 2). These differences in membrane fluidity may affect the membrane isolation and OMP solubilization process. A comprehensive review on the bacterial membrane lipid homeostasis can be found in [117]. After expressing the target OMP in a liquid culture, protein expressing cells will be harvested (Fig. 5.10, step 2). In case of bacterial expression hosts harvest will generally be accomplished by low speed centrifugation at 4 ı C and cell pellets are resuspended in suitable buffers (e.g. phosphate buffers pH 7.0–8.0) or unsuspended pellets can be stored at 20ı prior to further processing. In order to be able to isolate the protein carrying outer membrane, harvested cells have to be disrupted and non-outer membrane cell components have to be separated and discarded. Many of the methods used to disrupt cells before the isolation of cytoplasmic proteins can be applied also prior to isolation of outer membrane proteins. Cell disruption procedures that cause
30 ı C 4 29 23 30 10 1:6
40 ı C 8 48 9 12 8 0:38
destabilization or dissolving of the outer membrane by addition of high concentrations of detergents such as SDS, Sarkosyl or Triton X-100 are not advisable because they may lead to a premature OMP solubilization. A rather complete review on cell disruption methods is given in [118] and Table 5.3 summarizes the most common methods leading to outer membrane fragments in vesicle shape that can be used prior to outer membrane isolation. The disruption method of choice depends often on the culture volume and cell mass, as some methods like the disruption by ultrasonic sound are only applicable for small volumes. Furthermore it depends on the target protein, as temperature sensitive proteins or proteins that do not tolerate strong shearing forces forbid the use of methods that lead to sample heating or that disrupt cells by shearing such as the high pressure homogenisator (HPH), French press or ultrasonic cavitations. The much milder enzymatic lysis of bacterial cells using lysozyme might be more suitable, shows however low disruption efficiency and is quite time consuming. Gram-negative cells have to be treated with lysozyme and ethylenediaminetetra-acetic acid (EDTA), as EDTA chelates divalent cations (e.g. Mg2C ) that stabilize the outer membrane. The outer membrane destabilization allows lysozyme to reach the inner membrane murein layer [118]. In all cases the addition of DNase is advisable to decrease the solutions high viscosity caused by released DNA. The addition of commercially available protease inhibitor cocktails is generally useful also, as cells contain proteases that are liberated upon disruption. Disruption of cells is followed by the isolation of the outer membrane (Fig. 5.10, step 3) by differential centrifugation. In case of E. coli cell
5.2 Outer Membrane Protein Production: Challenges and Solutions
117
Table 5.3 Classification of cell disruption methods applicable to disrupt Gram-negative bacterial cells prior to outer membrane isolation Type Mechanical
Method HPH
French Press
Nonmechanical (physical)
Freeze/Thaw
Osmotic shock
Ultrasonic cavitations Nonmechanical (biological)
Enzymatic
Concept Pressure of a cell suspension is raised to 1,000 bar. Then high-velocity release of disrupted material through a valve leads to disruption. Pressure of up to 1,400 bar applied to cell suspension. Abrupt pressure reduction by dropwise release of disrupted material. Repeated freezing/thawing of cell paste; disruption through ice crystal formation. Often combined with grinding or enzymatic lysis. Rapid change of solution of high osmotic pressure to low osmotic pressure. Ultrasonic sound leads to pressure changes in cell solution resulting in cavitations. Lysozyme cleaves “-1,4 glycosidic bonds of peptidoglycan polysaccharide chains.
lysates can be cleared by low speed centrifugation to remove cell debris [125] (this step is not crucial though), cleared lysates are then treated with low concentrations of non-ionic detergents like Triton X-100 or N-lauroylsarcosine to solubilize the inner membrane [126, 127]. Centrifugation at speeds around 20,000–40,000 g allows the removal of solubilized inner membrane and other residual cell components with the resulting supernatant [40]. In a pre-solubilization step proteins loosely bound to the outer membrane can be removed by adding low concentration of solubilization detergent and subsequent ultracentrifugation [82]. If the fusion of a purification tag to the target OMP is possible, the outer membrane isolation can be skipped. To the uncleared lysate instead is added the solubilization detergent and the solution is loaded onto a suitable chromatographic column. This is of great advantage, as outer membrane preparation is a time consuming procedure due to lengthy centrifugation steps and incubation in pre-solubilization buffers. Furthermore the outer membrane pellets are very tough and sticky
Comments Suitable for large volumes, but leads to sample heating (stringent cooling required).
Reference [119]
Same as HPH.
[120]
Simple, inexpensive; but sensitive proteins can be damaged and efficiency is low.
[121]
Simple, inexpensive; usable in combination with enzymatic lysis; but not for stationary phase cells. Simple, inexpensive; but only for small volumes. Heats sample and can destroy sensitive proteins. Mild, but slow and with low disruption efficiency. Has to be combined with EDTA to destabilize the outer membrane.
[122]
[123]
[124]
and resuspension in newly added buffer requires mechanical processing with homogenizer or by ultrasonication. The most crucial step, the solubilization of the target protein from the outer membrane (see step 4 in Fig. 5.10), is achieved by the addition of a suitable detergent or other amphiphilic molecules (though detergents are often rather expensive due to a lack of effective alternatives they are still first choice for the MP solubilization) and a last high speed centrifugation step leading to separation of membrane lipids (pellet) and solubilized proteins (supernatant). The solubilization step is mainly dependent on the careful selection of the right detergent. In case of newly developed variants of an already known protein, one often can resort to protocols developed for the WT protein. While in case of new OMPs or variants with novel features a detergent screening is necessary. Detergents first disintegrate the lipid bilayer, dissociating lipid-protein interactions then their hydrophobic tail-regions interact with hydrophobic surface areas of the released OMPs. Hydrophilic detergent and protein portions are
118
in contact with the aqueous environment (see Fig. 5.10, step 4) [128]. It is assumed that a solubilized membrane protein is properly folded when in contact with the detergent [129]. Smallangle x-ray scattering showed that this is true for “-barrel shaped proteins such as the E. coli OmpF, while detergent-solubilized helical MPs such as bacteriorhodopsin and Ste2p G-protein coupled receptor from S. cerevisiae were not properly folded [130]. Solubilized MPs are always complexes of protein, detergent and membrane lipids, the detergent and lipid content of these complexes lies between 10–50 %, depending on the utilized buffer. Detergents are amphiphilic molecules that can be classified into four distinct groups [131]: 1. Ionic detergents (e.g. SDS) 2. Bile acid salts (e.g. Sodium cholate) 3. Non-ionic detergents (e.g. Triton X-100) 4. Zwitterionic detergents (e.g. 3-[(3-Cholamidopropyl)dimethylammonio]-1-propanesulfonate. – CHAPS) Ionic detergents consist of a charged head group and a hydrophobic hydrocarbon chain tail. They effectively solubilize proteins from the outer membrane, but disadvantageously many ionic detergents have more or less strong denaturing effects [131]. They are useful to dissociate protein-protein interactions. Some proteins can be refolded from their SDSsolubilizes state into a lipid environment by renaturing detergents [131] or amphipathic diol solvents such as 2-methyl-2,4-pentanediol (MPD) [132]. SDS can be removed for instance by dialysis. In an unpublished study the FhuA variant FhuA1–159 has been solubilized from the E. coli outer membrane upon over-expression, using SDS [133]. Solubilization efficiencies were much higher than with the conventionally used non-ionic detergent n-octylpolyoxyethylene (octyl-POE), where most of the protein remains within the membrane fragments. (FhuA1– 159 expressed into the membrane and extracted using octyl-POE allowed to obtain 1.5 mg of correctly folded protein from 1L of culture [41]). Protein solubilized by SDS was present in an unfolded state and tended to precipitate. The
5 Biotechnology
Fig. 5.11 SDS-PAGE result of outer membrane derived FhuA1–159 solubilized in SDS and refolded by dialysis into PE-PEG containing buffer
protein was therefore refolded by SDS-removal via dialysis or during immobilized metal-affinity chromatography (IMAC), where the unfolded protein was bound via an internal His-tag to the column resin and the SDS was diluted by washing steps. In both cases the commercially available and relatively cheap diblock copolymer Polyethylene-Poly(ethyleneglycol) (PE-PEG) had been used as a refolding agent and only with the dialysis method could be obtained reasonable yields of 30 mg/L of culture (see Fig. 5.11), which is about 20-times higher than yields obtained by solubilization with octyl-POE and about 1.5-times more than when isolating and refolding the protein from inclusion bodies (where refolding is again facilitated by PE-PEG) (see also Sect. 5.2.2). Structural integrity of the so obtained protein was verified by CDspectroscopy, revealing a “-structure content of 48 %. Obtained purities however were only up
5.2 Outer Membrane Protein Production: Challenges and Solutions
to 85 %. However in this study only PE-PEG had been used to refold the protein without further optimization or screening for better refolding agents, and no further purification steps were employed. Therefore there might be noteworthy potential to improve yields and especially purity by further optimizing the protocol and by using other solubilization agents. Bile acid salts (or saponin detergents) are mild ionic detergents with a rigid steroidal backbone. In contrast to linear chain detergents they do not form conventional micelles [128, 131]. Nonionic detergents have polyoxyethylene or glycosidic, hydrophilic head groups and they are mild and often not denaturing. They effectively break lipid-lipid and lipid-protein interactions, therefore they are commonly used to solubilize MPs in general. Disadvantages are however that they often lead to low yields [128] and as they do not break protein-protein interactions they do not prevent proteins from forming aggregates that may precipitate. Zwitterionic detergents combine ionic and nonionic detergents properties. Though they are more denaturing than non-ionic detergents they are widely used, as they proved to be of advantage for protein structural study purposes. Examples are the BtuB protein that has been crystallized in complex with TonB using detergent lauryldimethyl amine oxide (LDAO) [134] while FhuA and OmpF were crystallized in dimethyldecylamine-N-oxide (DDAO) [56, 135]. Furthermore many NMR-based structural studies have been carried out using zwitterionic detergents such as dodecylphosphocholine [136, 137]. In order to find the best detergent for a certain outer membrane protein (it should solubilize the protein without denaturing it) several different detergents and solubilization conditions have to be screened. The construction of a new OMP variant from an existing OMP often necessitates to newly screen for optimal solubilization conditions, especially if the new variant shows changes in hydrophobicity. One example is FhuA variant FhuA Ext with extended hydrophobic, TM region (see Sect. 5.1.2.4). In contrast to the
119
parent FhuA1–159 that can be solubilized from the membrane using detergent octylPOE, FhuA Ext could be obtained from the membrane only by a serial extraction with organic solvent mixtures of chloroform:methanol and TFE (2,2,2-trifluoroethanol):Chloroform prior to solubilization in detergent OES [69]. Solubilization yield can be monitored for instance by SDS-PAGE of supernatant and pellet after ultra-centrifugation (Fig. 5.10, step 5). The target protein should be mainly in the supernatant. As a rule of thumb a membrane protein is properly solubilized if it stays in solution after 1 h centrifugation at 100,000 g. A more accurate method to determine whether a MP is solubilized is the gelfiltration of protein in detergent solution. When using Sepharose 6B for instance the protein elutes in the columns inclusion volume in case it is solubilized. Recently an ultracentrifugation dispersity sedimentation assay (combining small volume ultracentrifugation and subsequent SDS-PAGE) has been developed that allows determining rapidly whether a MP is monodispersedly solubilized, thus speeding up screening processes [138]. Nowadays commercially obtainable detergent screening kits containing sets of commonly used detergents may speed up the selection of a detergent for solubilization or crystallization. Above a certain detergent concentration, the critical micellar concentration (cmc), in an aqueous environment, the detergent molecules associate and form multimolecular complexes, so called micelles. Micelles are aggregates with hydrophobic interior and hydrophilic exterior surfaces. The cmc depends on the detergent type, the solution pH, temperature and ionic strength [139]. In general the cmc decreases with alkyl chain length and increases with the introduction of double bonds. Additives that break up the water structure (e.g. urea) increase the cmc in all detergent types, while increased concentrations of counter ions result in cmcreduction for ionic detergents. When using detergents for protein solubilization purposes, buffers and solutions should have a detergent concentration above the cmc, especially when a detergent might have to be removed/exchanged by dialysis [131].
120
Arachea et al. extensively studied the extraction profiles of different detergent types for membranes isolated from bacteria and yeast, on a set of recombinant target proteins. Some general trends were found [140]: 1. The extraction efficiencies of the analyzed detergents increased at higher concentrations. At concentrations below a detergent’s cmc extraction efficiency dropped significantly. The optimal concentration is detergent-dependent. 2. SDS, two alkyl sugar detergents, octyl-b-Dglucoside (OG) and 5-cyclohexyl-1-pentyl-bD-maltoside (Cymal-5), and a zwitter-ionic detergent, N-decylphosphocholine (Foscholine-10), were effective in the extraction of a broad range of MPs. 3. In case of E. coli, SDS was the most efficient at extracting proteins from the inner membrane as well as from the outer membrane, while OG was among the most effective nonionic detergents. However membrane protein extraction efficiencies of detergents vary between different E. coli strains. 4. Fos-choline is not very effective for the extraction of outer membrane proteins. 5. Protein extraction from yeast membranes was reported to be in general more difficult. These findings underline the necessity for thorough detergent screening in order to obtain the highest possible concentration of a correctly folded and functional target MP. Some proteins can be solubilized in detergents but need a lipid environment to be functional, here the use of lipid/detergent systems for solubilization can be an alternative [131]. As detergents that allow functional OMP solubilization are often rather expensive, commercially available, amphiphilic block copolymers [41] or if the protein target is stable enough organic solvents [141] may be considered. A recent approach uses nanoscale phospholipid bilayers stabilized by an encircling membrane scaffold lipo-protein, so called nano-discs [142]. Nano-discs are getting more and more attention as they are now commercially available and have been found to be especially useful for protein biophysical characterization [143].
5 Biotechnology
5.2.1.1 Yield, Purity and Quality Control (OMP Structure and Functionality) Supernatants obtained from centrifugation after solubilization have to be analyzed on target OMP presence, amount and purity (see step 5, Fig. 5.10). Most often this will be accomplished by carrying out SDS-PAGE following the “Laemmli” method [144]. However the sample boiling with SDS and reducing agent containing buffer to fully denature the protein can lead to protein aggregate formation, when working with membrane proteins in general. If this is the case, samples can be incubated at 60 ı C for 30 min or at 37 ı C for 60 min instead. The experimentalist has to be aware of the fact that MPs often do not move according to their molecular weight in SDS-PAGE. They may move faster and will thus appear smaller, which might be due to differences in bound SDS amounts as compared to soluble proteins [145]. For instance for the 37.2 kDa OmpF it was reported that its band position on stained SDS gels varied significantly depending on the detergent used for solubilization. While in SDS it migrated as a protein with lower molecular mass in OG it migrated more slowly, showing a false higher molecular mass [140]. The commonly observed “smearing or tailing” of MPs on SDS gels is due to lipids sticking to the protein or to protein precipitation resulting from too low detergent concentrations (as the protein is concentrated during the gel run), often a higher SDS concentration in sample and running buffer avoids “smearing”. After electrophoresis proteins can be visualized by staining the gel with Coomassie brilliant blue or by a silver staining (for lower protein concentrations). In order to check the target OMP purity using the SDS-PAGE method it is suggestible to further analyze the stained gel by image processing programs. A good freeware program to analyze SDS-gels is the Java based “ImageJ” (available at http://rsb.info.nih.gov/ij; developed by Wayne Rasband, National Institutes of Health, Bethesda, MD). Western-blot analysis, using target protein (or purification tag) specific antibodies can be used
5.2 Outer Membrane Protein Production: Challenges and Solutions
to detect poorly expressing proteins after protein transfer from an SDS gel to a blotting membrane. Due to the OMPs hydrophobicity however their transfer to a membrane can be difficult [146]. The protein concentration can be determined using standard assays, such as the bicinchoninic acid (BCA) [147] or Lowry assay [148], while the commonly used Bradford method [149] is incompatible with detergents. An outer membrane protein’s structural integrity and overall folding can be verified by secondary structure analysis via CD spectroscopy or X-ray diffraction measurements of crystallized OMPs and NMR studies on the tertiary structure (see Chap. 3). OMP-functionality is generally studied after reconstituting/inserting the protein into lipid or polymer membranes and analyzing transport activity or conductance either directly by patchclamp methods or indirectly by enzymatic assays that allow the catalytic transformation of a substrate that passes the OMP to reach the enzyme (unable to pass the OMP channel) and by subsequently detecting the product that again passes the OMP (see Chap. 6, Sect. 6.1.3). While the purity of solubilized OMPs is generally sufficient for functional tests, it often does not suffice for structural analysis. Therefore further purification steps might become necessary. A general discussion on the purification of outer membrane proteins with and without purification tag will be given in Sect. 5.2.4.
5.2.2
OMP Isolation from Inclusion Bodies for Improved Yields and for the Expression of Toxic OMPs and Their Variants
The described OMP over-expression and subsequent targeting to the outer membrane often has toxic or lethal effects on the expression-host (see also Sects. 5.1.2.2 and 5.1.2.3) limiting the reachable expression-levels and range of obtainable proteins [101, 150]. A valid alternative to obtain OMPs is their non-functional expression into inclusion bodies with subsequent solubilization of inclusion body material in denaturing
121
buffer (e.g. urea) and refolding by exchanging denaturing agents with amphiphilic solubilization agents (e.g. detergents) (see Fig. 5.12). In E. coli over-expressed proteins that accumulate in the cell are deposited in the form of inclusion bodies. Inclusion bodies are insoluble aggregates of misfolded and inactive protein. These aggregates can be observed by phase contrast microscopy as dark intracellular particles (refractile bodies) with a size of 0.2–0.6 m [151]. As proteins from inclusion bodies can be refolded efficiently by well-established protocols, the expression of proteins in inclusion bodies is a suitable method for the high-level production of recombinant proteins [152]. In general the protein production in inclusion bodies holds several advantages over their functional expression: 1. The over-expression of proteins into inclusion bodies is less stressful for host cells. 2. Proteins that are toxic to the expression host when expressed in their functional state are non-toxic in inclusion bodies. 3. Higher expression levels can be reached. 4. Proteins in inclusion bodies are less prone to proteases. 5. The target protein in inclusion bodies is often already relatively pure. A disadvantage is however that proteins obtained from inclusion bodies have to be refolded back to their native states, which can be quite difficult, especially if the target protein contains disulfide bridges. Generally one has to screen for the optimal conditions to refold a certain protein. The inclusion body expression is a promising way to express bacterial OMPs and their variants that when expressed functionally have toxic effects on the host cells [31–34, 62, 101, 150] or that can be expressed in low concentration only. Especially so since the OMP expression into inclusion bodies followed by solubilization and refolding works rather well in contrast to the production of ’-helical plasma membrane proteins by inclusion body expression. Furthermore the isolation of inclusion body material is rather less time consuming the isolation of outer membrane vesicles (however protein refolding steps may take considerable time, especially when carried out by dialysis). One of the first OMPs ob-
122
5 Biotechnology
Fig. 5.12 Schematic representation of conventional inclusion body expressed OMP production process flow. 1 – ORF cloning into expression vector (grey circle), transformation of host cells and cell growth C protein expression in liquid culture; 2 – cell harvest (e.g. by centrifugation); 3 – cell disruption and isolation of inclusion body material; 4 – solubilization of inclusion
body material using high concentrations of e.g. urea; 5 – urea removal and addition of OMP solubilization agent, for instance by step-wise dialyzing the solubilized inclusion body material; 6 – yield and protein purity control (e.g. by SDS-PAGE); 7 – characterization and application of obtained OMP variant (Protein New Cartoon representations made with VMD)
tained in this way and successfully crystalized afterwards, was the E. coli outer membrane phospholipase A (OmpLA) [153]. Inclusion bodies are formed when the transcription and translation rates are high. Generally it can be said that if expression of a plasmid-coded, over-expressed protein is higher than 2 % of cellular protein expression, inclusion bodies are likely to form [154]. Therefore in order to express a target protein into inclusion bodies a plasmid system with a strong promoter, such as the T7 promoter in combination with growth temperatures that facilitate fast growth are suggestible. Furthermore the use of an E. coli strain optimized to better endure stress caused by the formation of inclusion bodies should be considered. Examples for such strains are the E. coli BL21(DE3) derivatives C41(DE3) and C43(DE3) [103]. Moreover the outer membrane protein N-terminal signal sequence that leads to the transport to the Gram-negative bacteria outer
membrane should be deleted, as over-expressed OMPs that lack their signal sequence accumulate in inclusion bodies [101, 155]. The production of an OMP in its unfolded state in inclusion bodies takes course similar to the inclusion body production of soluble proteins and starts after deleting the N-terminal signal sequence of the outer membrane protein ORF. The ORF is then cloned into an adequate vector (with a strong promoter). The vector is then inserted into the host strain by transformation, cells are allowed to grow and induction is carried out at temperatures that facilitate high growth rates (for E. coli 37 ı C) (see Fig. 5.12, step 1). Cells are harvested and disrupted (see Fig. 5.12, step 2), following the same considerations as when isolating OMPs from the outer membrane. But cells have to be disrupted as completely as possible to avoid contaminations with other cellular components. Best results can be obtained by a combination of ultra-sonication or high pressure
5.2 Outer Membrane Protein Production: Challenges and Solutions
homogenization in presence of EDTA (to destabilize the outer membrane) and detergents, such as Triton X-100 followed by lysozyme and DNase treatment. Here it is important to add high concentrations of Mg in order to chelate and inactivate EDTA, as DNase requires Mg as a cofactor [156]. Since inclusion bodies have a rather high specific density, they can be obtained after cell disruption by centrifugation at moderate speeds [152, 156, 157] (Fig. 5.12, step 3). Lysozyme and DNase treatment can be repeated for enhanced purity. Further washing of inclusion body material with Triton X-100 containing buffer removes residual membrane fractions and purities of the target protein of 95 % can be reached [158]. After washing, the purified inclusion bodies have to be solubilized (Fig. 5.12, step 4). Solubilization is generally achieved by the addition of highly concentrated denaturants (i.e. 6 M guanidinium chloride – GdmCl or 6–8 M urea.) and reducing agents (to keep disulfide bridges from forming). Since GdmCl is a strong chaotroph it allows the solubilization of inclusion bodies that are resistant to solubilization by urea [156, 159] but in certain proteins it might inhibit ionic interactions necessary for correct refolding [160]. Alternatives include the use of SDS [161], sodium N-lauroyl sarcosine [162] or N-cethyl trimethyl ammonium chloride [163]. Protein contained in solubilized inclusion bodies has to be folded back to its native state (see step 5 in Fig. 5.12). Prior to refolding the target protein can be purified further, in case it carries a His-tag, purification can be carried out by IMAC in presence of a denaturant allowing on-column refolding (see Sect. 5.2.4) [164]. Refolding is generally initiated by dilution of the denaturing agent into a suitable buffer. In case of membrane proteins this buffer has to contain lipids, detergent micelles, mixed lipid-detergent micelles or other amphiphilic substances. In case of FhuA variant FhuA1–159 expressed into inclusion bodies after signal sequence deletion, the protein could be refolded into a solution containing the commercially available diblock copolymer PEPEG as an alternative to costly detergents. By this
123
Fig. 5.13 SDS-PAGE result of inclusion body derived FhuA1–159 after IMAC on-column refolding, using PEPEG diblock copolymer
method 19.5 mg of protein (purity of 92 %) per L of culture could be obtained. Furthermore FhuA1–159 was successfully reconstituted into liposomes from the PE-PEG solubilized state [41] showing that the diblock copolymer is suitable for protein reconstitution purposes. Dilution of the denaturing agent can be carried out either rapidly by drop-wise addition of solubilized inclusion bodies to a refolding buffer, stirring the solution vigorously or slowly and step-wise by dialysis, or by on-column refolding during chromatography, decreasing the denaturant concentration with each new dialysis or chromatography washing step, in presence of the refolding buffer [152]. As slow folding might lead to the formation of folding-intermediates that precipitate and as the rapid dilution might equally lead to protein precipitation, both depending on the respective protein, the method of choice has to be selected by testing. Also a screening for the best refolding agent is necessary. The on-column refolding method can facilitate protein refolding especially when combined with the use of an ionic detergent, as it inhibits protein aggregation [101, 165]. In an unpublished study the diblock copolymer PE-PEG has been used to on-column refold FhuA1–159 (that includes an internal His-tag; see Sect. 5.2.4) during IMAC. As the SDS-PAGE result depicted in Fig. 5.13 shows, the highly pure protein (92 %) that was successfully bound and eluted from the column, though a relatively high
124
5 Biotechnology
Table 5.4 Examples for bacterial OMPs expressed in E. coli inclusion bodies, listing protein origin and applied refolding method OMP OprM Major porin Omp21 Opc OmpC OmpA, OmpX OmpT FepA
Origin Pseudomonas aeruginosa Rhodopseudomonas blastica Comamonas acidovorans Neisseria meningitides Salmonella typhi E. coli E. coli E. coli
OmpG FhuAC/4L FhuA1–159
E. coli Variant of E. coli FhuA Variant of E. coli FhuA
FhuA1–159
Variant of E. coli FhuA
amount of protein did not bind to the affinity resin and yields were rather poor and in the same range as when isolating the protein from the membrane using octyl-POE for protein solubilization (See Sect. 5.2.1 and see below). Structural integrity of the refolded protein had been verified by CD spectroscopy, with an ascertained “-structure content of 48 % [166]. Refolding is influenced by a set of parameters, such as protein concentration, ionic strength (high ionic strength might increase the yield), residual low concentrations of denaturing agents (in some cases the retention of denaturing agents in low concentrations can lead to yield-increase) and temperature (often room temperature is preferable to lower temperature, such as 4 ı C) [160]. The refolding methods employed for the refolding of several E. coli inclusion body derived bacterial OMPs and OMP variants are listed in Table 5.4. The yields of correctly folded protein are often higher than when the protein is isolated from the outer membrane. In case of the FhuA1–159 19.5 mg of refolded protein were obtained from 1L of culture, while the same culture volume led to only 1.5 mg of outer membrane derived protein [41]. However as the aggregation of unfolded protein and the precipitation of folding-intermediates occur faster than the correct refolding, the yield
Refolding method Dilution with detergent solution Dilution with detergent solution On column refolding using detergents Dilution with detergent solution Dilution with detergent solution Dialysis into detergent solution Dilution with detergent solution Dialysis into detergent solution including SDS Dilution with detergent solution On column refolding using detergents Dialysis into PE-PEG diblock copolymer solution On column refolding using PE-PEG diblock copolymer
Reference [167] [62] [168] [169] [170] [171] [172] [173] [174] [46] [41] Unpublished results [166]
of correctly folded protein is often limited to 2–5 % (in rare cases up to 20 %) of the total expressed protein. Recent developments in OMP expression using cell-free expression systems, with attempts to directly express and fold the OMP into a suitable folding-buffer containing detergents might lead to further increased yields in the future and will be discussed below. Yields are again checked by SDS-PAGE and protein concentration measurements (see Sect. 5.2.1). The refolding itself can be monitored by CD spectroscopy or by measuring protein tryptophane fluorescence, providing the target protein contains tryptophane residues [175]. Spectra of folded and unfolded proteins as well as folding intermediates differ from each other, as a blue-shift of the excitation wavelength is observed upon folding to the native state [175]. As tryptophane fluorescence measurements can also give clues on the stability of membrane proteins in different environments, such as when reconstituted into liposomes or polymersomes [176], this method will be discussed further in Sect. 6.1.4. Useful information on the inclusion body expression and protein refolding methods for many different proteins can be found in the online database REFOLD (http://refold.med. monash.edu.au/) that currently (2013) holds 1,165 entries, of which about 40 concern OMPs [177].
5.2 Outer Membrane Protein Production: Challenges and Solutions
A recent attempt to develop a universally applicable refolding method is based on the finding that the anionic detergent SDS that is very effective to solubilize proteins (also from the membrane; see Sect. 5.2.1), but has denaturing effects, when in presence of the amphipathic diol solvent (2-methyl-2,4pentanediol) is transformed into a non-denaturing solubilization agent. Using this solvent mix the inclusion body derived eight-stranded bacterial OMP PagP could be refolded after inclusion body solubilization in SDS by the addition of 2methyl-2,4-pentanediol [178]. The same solvent system has been used to refold the Omp2a from Brucella melitensis heterologous expressed into E. coli inclusion bodies [132].
5.2.3
A New Alternative: OMP Production Using Cell-Free Expression Systems
Though first attempts in cell-free protein expression have been made as early as in the 1960s [179] and were carried out chiefly to understand the involved cellular processes, the developments made during the last decade, render cell-free expression systems a powerful, rapid and efficient alternative for the production of proteins in general (allowing product yields higher than the mg/ml scale) and are used since around 2004 also for integral membrane proteins [180]. Especially for ’-helical MPs the cell-free expression approach is a valid alternative to the membrane expression, as their refolding from inclusion body material works only in rare cases. However also in case of OMPs it might be a further way to overcome the limitations implied by the expression into and isolation from the outer membrane, namely low yields due to limited (membrane) space for correctly folded protein, toxic effects of OMP over-expression (especially when working with wide and open passive diffusion channel variants that lead to osmotic imbalances; see Sect. 5.1.2.2) or negative effects on the expression host cells caused by blocked OMP trafficking. Cell-free expression systems contain either the coupled transcription/translation mech-
125
anism or translation mechanism only of bacteria (generally of E. coli [181]) or of eukaryotic organisms such as yeast [182], insect cells [183], wheat germ cells [184] or rabbit reticulocytes [185]. Coupled systems allow to start from a DNA template, while translation mechanism systems require less facile mRNA templates [186]. A cell-free expression system can either be just a crude cell extract supplemented with essential amino acids, nucleotides, salts and ATP or guanosine triphosphate (GTP) replenishing factors, such as creatine kinase or creatine phosphatase [187, 188] or can consist of purified components (with the same supplements). The E. coli translational machinery (consisting of more than 100 proteins) for instance had been successfully purified and functionally reconstituted in 2001 [189] and has been further advanced into the commercially available PURE system 2010 [190]. While crude extracts are much easier to obtain and thus much less expensive, they are useful for the expression of proteins harboring a purification tag only, as the extracts contain a multitude of other proteins. Purified protein cell-free systems instead are difficult to obtain, as each protein has to be purified individually, they have the advantage though that the reaction conditions are better controllable and that target proteins without purification tags can be expressed and isolated. Various systems can be commercially obtained or prepared by standard protocols. As the expression of bacterial OMPs does not necessitate the introduction of post-translational modifications, as is the case for many eukaryotic proteins, E. coli based cell-free systems should generally be suitable and the following discussion will therefore focus on such systems. Comprehensible, recent reviews on cell-free protein expression and their use for the biotechnological protein production can be found in [191–194]. E. coli cell-free expression systems are often based on the optimized coupled transcription/translation E. coli extracts [195]. All coupled cell-free systems contain the following major components: 1. For transcription: recombinant T7 RNA polymerase, nucleotides;
126
2. For translation: tRNAs, initiation, elongation, release factors, ribosomes, amino acyl-tRNA, essential amino acids synthetase; 3. General: ATP, co-factors, adenosine triphosphate (ATP)-replenishing factors. Template DNA can be either plasmid DNA or polymerase chain reaction (PCR) derived DNA carrying a T7 promoter (most commonly used) as well a Shine-Dalgarno box as translation initiation signal. The simplest cell-free protein expression format is the batch, where all components and template-DNA are mixed in a closed reaction container, without further substance addition during the reaction. Though easy to handle the batch format limits obtainable yields significantly, as it has rather short lifetimes below 1 h, due to the fast energy-delivering phosphate pool consumption and due to accumulating free phosphates that complex essential enzyme cofactors such as Mg [196]. This problem can be solved by either employing formats such as the complex continuous-flow cell-free expression with a continuous supply of substrates and energy [197], the simpler continuous-exchange cell-free expression with passive substrate and by-product exchange [198] and the efficient bilayer diffusion system [199] or by finding better alternatives to traditional ATP and/or GTP replenishing systems, while using the user-friendly batch format, such as the “Cytomim” system that uses the energy source pyruvate to obtain high yields of expressed protein [200]. As already mentioned cell-free OMP expression can be a good alternative especially to the expression into the outer membrane as it decouples protein production from cell growth and viability, as well as it avoids protein insertion into the outer membrane, thus avoiding problems caused by toxic effects an over-expressed OMP or OMP variant might cause and allowing higher yields due to the omitted membrane-caused space limitation. However until recently cell-free expression had been utilized to produce ’-helical MPs exclusively, due to the lacking possibility to refold these proteins from inclusion body material. In 2005 the first OMP (E. coli nucleoside
5 Biotechnology
transporter Tsx) had been expressed using an E. coli derived cell-free expression system [201]. One can distinguish between two different modes of cell-free MP expression: 1. The expression mix does not contain detergents or other solubilization agents resulting in the precipitation of unsolubilized protein (somewhat similar to the expression into inclusion bodies, though mild detergents are sufficient for solubilization). Proteins are subsequently solubilized by the addition of detergents or other amphiphilic substances [180] (see Fig. 5.14a, step 1 and 2). Unsolubilized precipitate can be removed by centrifugation. The detergent 1-myristoyl-2-hydroxy-snglycero-3-[phosphorac-(1-glycerol)] (LMPG) has been proven to efficiently solubilize obtained protein precipitates [201]. In general the best suitable detergent has again to be decided on after carrying out a screening. Apart from lipids also lipid-protein nanodiscs have been used to solubilize MPs from precipitates obtained by cell-free expression [202, 203]. The use of nano-discs often assists NMR-studies of the reconstituted protein (see also Chap. 3). 2. The expression mix is provided with detergents or other solubilization agents, resulting in the production of already solubilized protein [201, 204] (see Fig. 5.14b, step 1). The detergent class of polyoxyethylene-alkylethers, with higher polymerization number of the polyoxyethylene moieties, has been shown to effectively solubilize membrane proteins during cell-free expression [201]. Several ’helical MPs have furthermore been successfully cell-free expressed in the presence of liposomes, leading to protein insertion into the lipid membrane [203, 205–207]. Recently the ’-helical MP claudin-2, has been expressed in a wheat germ extract based cell-free system containing block copolymer vesicles of the polymer polybutadiene-polyethyleneoxide (PBD-PEO) [208]. Protein derived from both of the described cell-free expression modes has then to be analyzed by SDS-PAGE, to obtain information on
5.2 Outer Membrane Protein Production: Challenges and Solutions
127
Fig. 5.14 Schematic representation of cell-free OMP expression. a – expression mix lacking detergents (or other solubilization agents), leading to protein precipitates that can be refolded by detergent addition; b – expression mix provided with detergents (or other solubilization agents);
1 – addition of template-DNA and cell-free expression; 2 – purification of solubilized OMP; 3 – yield and protein purity control (e.g. by SDS-PAGE); 4 – characterization and application of obtained OMP variant (Protein New Cartoon representations made with VMD)
yield and purity, before further characterization studies can be carried out or the protein can be used for the desired application (Fig. 5.14, step 3 and 4). Up to now only few examples of cell-free expressed “-barrel shaped MPs including E. coli Tsx [201] and the mitochondrial, voltagedependent anion channel (VDAC1) [209] and its chloroplast homologue OEP24 [210] are reported. Nevertheless together with the inclusion body expression/refolding approach the cell-free expression might become even more important to produce OMP variants for nano-technological applications, as both systems allow the expression of variants with “extreme” geometry, thus going “beyond natural limits”. One thinkable example would be a “-barrel protein with hydrophobic region extended above the limit the biological lipidbilayer can accommodate. In order to obtain such a protein it has to either be produced as insoluble precipitate (i.e. inclusion body or cellfree expression) with subsequent reconstitution using detergents or thick enough lipid/polymer membranes or solubly expressed by use of a cell-free system containing suitable detergents,
lipids or polymers. In conclusion both introduced alternatives to the conventional OMP membrane expression allow higher protein yields, avoid toxic effects on an expression host and open the possibility to produce OMP variants with novel characteristics or altered geometries that cannot be found in nature, allowing to adjust the protein nano-channel’s features to the desired application. Furthermore both inclusion body expression and cell-free expression can in the future probably be coupled with the OMP channel insertion into lipid or polymer membranes, which is a step towards functional nano-systems for various applications.
5.2.4
OMP Purification and Concentration Methods
Solubilized outer membrane proteins often have to be further purified to allow functional and structural characterization, especially for crystallization purposes or to perform CD spectroscopy samples have to be exceptionally pure. This is the case particularly for membrane-derived and cell-free expressed OMPs, but can be necessary
128
also for OMPs refolded from inclusion bodies (though inclusion body derived proteins are often already rather pure). Purification by non-specific chromatographic (e.g. gel filtration) methods often leads to a dilution of the purified target, several concentration techniques may be used to obtain highly concentrated samples, however special considerations have to be made when concentrating membrane proteins (see below). In general the same purification techniques as used for water-soluble proteins can be used to purify MPs also. MPs however require the presence of detergents (or other amphiphiles) to remain in solution. Therefore they are usually purified in complex with detergents or detergents and lipids. Detergent concentrations during purification steps should be above cmc, but may be lower than when used for solubilization. In the following will be given an overview on the common purification and protein concentration methods applied to bacterial OMPs. A complete work on protein purification methods at large can be found in [211] while [212] focusses on the purification of membrane proteins.
5.2.4.1 Immobilized Metal Affinity Chromatography In the easiest and most universally applicable case the target outer membrane protein contains a His-tag allowing the specific purification and concentration by IMAC. Proteins containing a sequence of at least six histidine residues (Histag) bind to Ni-ions that are immobilized to a chromatography resin, allowing the specific purification of tagged proteins [213]. IMAC is one of the preferred OMP purification methods, especially since the presence of a His-tag allows skipping the time consuming membrane isolation and OMP extraction, by applying uncleared bacterial lysate to the IMAC column and providing a suitable solubilization agent during chromatography (though in this case soluble metal-binding proteins present in the lysate might contaminate the purified fraction). Furthermore His-tagged OMPs derived from purified inclusion bodies can be loaded to an IMAC column after solubilizing the inclusion body material. The protein binds to the Ni-affinity resin, the column can be washed
5 Biotechnology
and by adding detergents (or other amphiphiles) during elution the protein can be refolded oncolumn (see also Table 5.4). The IMAC method can furthermore be used to change the detergent solubilizing the protein, as detergents that are used to extract a protein from the outer membrane are not always suitable for characterization purposes or applications involving the OMP reconstitution into polymer or lipid membranes. Regarding the tag position one has to consider that the N-terminus is often less suitable for OMPs that are functionally expressed into the membrane, since the N-terminal signal sequence is processed and cleaved by a signal peptidase. Furthermore C- and N-terminus close to form the barrel with which the His-tag might interfere. For OMPs that do not tolerate a C-terminal His-tag the tag can be placed within an external loop that is not covered by detergents. Ferguson et al. for instance introduced an internal hexa-His-tag to the FhuA protein at the surface exposed amino acid position 405 [100]. The presence of detergents on hydrophobic protein portions can weaken protein binding to the column matrix, this problem can be addressed by increasing the tag-length (generally up to 10 His), and by increasing binding time. Binding time can be elongated either by batch-wise binding, i.e. mixing the protein sample and affinity matrix prior to column packing (the mixture can be incubated overnight) or by lowering the chromatography buffer flow rate during protein binding. Though batch-wise binding can further improve the overall yield, the long incubation time increases the probability of protein degradation. Several bacterial OMPs have been provided with a His-tag and purified by IMAC either after being expressed into inclusion bodies, including on-column refolding (Table 5.4) or after functional expression into the bacterial outer membrane. Table 5.5 sums few examples of bacterial OMPs (with different tag locations) that were expressed into the E. coli outer membrane and were purified by IMAC after protein solubilization from the membrane fraction. IMAC derived protein samples are often purified further by gel filtration to remove remaining
5.2 Outer Membrane Protein Production: Challenges and Solutions
129
Table 5.5 Examples for bacterial OMPs expressed functionally into the E. coli outer membrane that were purified by Ni-IMAC; listing protein origin and tag position OMP FhuA FhuAC/4L PhoE Omp85 OprM OprN
Origin E. coli Variant of E. coli FhuA E. coli Neisseria meningitidis Pseudomonas aeruginosa Pseudomonas aeruginosa
aggregates or to exchange the buffer, as many reconstitution experiments or structural and functional studies in general require special buffer conditions and the imidazole used to elute proteins from the IMAC column may interfere with certain analytical methods, such as CD spectroscopy, as imidazole is a chiral molecule. As the His-tag might inhibit protein crystallization the insertion of a tag removal protease site should be considered prior to crystallization experiments. The TEV (tobacco etch virus) protease site is commonly used, however one has to be aware that the TEV protease is partly inactive in several detergents that are often used to solubilize OMPs after purification (e.g. OG, LDAO) [217].
5.2.4.2 Ion Exchange Chromatography Ion exchange chromatography (IEXC) involves the use of a charged column matrix. The separation is based on competition between proteins carrying different surface charges for oppositely charged groups present on the matrix. IEXC is a good method to purify un-tagged OMPs or can be used in combination with IMAC (or other affinity chromatography methods) whenever especially high purity is necessary. Elution can be accomplished either by changing buffer pH and thus changing the protein net surface charge or more commonly by changing the ionic strength of the buffer. In order not to shield the matrix surface charge, ionic detergents have to be avoided and non-ionic detergents (such as octyl-POE) or zwitterionic detergents have to be used. Ion exchange chromatography can be used to separate folded OMPs from unfolded or partially folded proteins as the unfolded counterparts often elute at different ionic strengths than
Tag position Internal Internal N-terminal N-terminal C-terminal C-terminal
Reference [101] [46] [214] [215] [216] [216]
the correctly folded OMPs. E. coli FepA for instance elutes at lower ionic strength when in the unfolded state than when in its native form [101, 173]. Thus IEXC is a feasible way to further condition proteins refolded from inclusion body material. Since several OMPs bind to ion exchange resins when in high concentrations of urea, IEXC can be used to on-column refold inclusion body derived proteins (similar to the on-column refolding by IMAC) [101]. This technique has for instance been successfully deployed to refold the Rhodopseudomonas blasticus porin [62].
5.2.4.3 Gel Filtration (Size Exclusion Chromatography) In gel filtration (GF) or size exclusion chromatography proteins in a sample are fractionated due to their relative size and GF in contrast to other chromatographic techniques does not involve any chemical interaction between protein and matrix. The GF matrix consists of porous beads and while molecules with hydrodynamic diameter above pore size are not able to enter the bead pores and will elute immediately, molecules whose hydrodynamic radii allow pore entering will be fractionated according to their radii (big radius > small radius) [218]. GF is generally the last step of a multistep OMP purification protocol. Especially for crystallization purposes, when a completely homogeneous sample is necessary, gel filtration can be used to remove any remaining folding intermediates or protein aggregates, as they show different hydrodynamic diameters as compared to the hydrodynamic diameter of the correctly folded protein [101] and elute at different chromatography stages. GF allows a
130
buffer exchange and the removal of imidazole used to elute His-tagged proteins from a Ni-column. As automated protein chromatography methods often rely on the detection of aromatic amino acid residues present in the proteins within a sample by UV absorption measurement at 280 nm, it has to be considered that detergents containing aromatic rings, such as Triton X-100 absorb in the UV region and lead to false positive signals or false high signals.
5.2.4.4 Concentration Methods Especially for crystallization purposes but also for many applications that involve the OMP reconstitution into polymer or lipid membranes (e.g. for polymersome or liposome based drugdelivery systems) considerably high protein concentrations are required and further concentration of purified protein samples might be necessary. In case of IEXC and especially IMAC purification and concentration can be achieved at the same time, as the target-protein interacts with the column resin and elutes when a certain buffer is added, by applying an as much as possible small elution buffer volume the protein concentration can be increased. If the employed purification protocol ends however with a gel filtration step the sample will be diluted (often by a factor of 3). IMAC or IEXC can again be applied or the protein sample can be concentrated by for example: 1. Ultrafiltration: Ultrafiltration-based concentration techniques can be carried out either by commercially available stirred ultrafiltration cells (for volumes up to 500 ml) that can be pressurized to allow water and small solutes to leave through a narrow pore membrane or by centrifugation filter units (for smaller volumes) that remove liquid by centrifugation. Though commonly used for membrane proteins these devices co-concentrate the present detergent micelles, resulting in a change in detergent:protein ratio. This ratio change often leads to detergent or detergentprotein complex precipitation. Furthermore ultrafiltration membranes can be blocked by the protein or by detergent micelles
5 Biotechnology
2. Dialysis: By dialysis of an already pure OMP sample against high molecular weight solutions of polyethylene glycol (PEG; most often PEG 20,000) combined with the use of dialysis membranes with very small average pore size the protein concentration can be increased, since PEG will deprive the sample of water [212]. The PEG can be dissolved to form a very high concentrated solution (20 %), or can even be used as a solid. However this method will increase the detergent concentration as well, leading to the mentioned negative effects. 3. Lyophilization: Since bacterial porins like OmpF and OmpC are known to be stable even after lyophilization [219], this technique might be a very good way to concentrate a target OMP providing the protein resists such treatment. The lyophilizate in powder form can be resolubilized in a suitable amount of buffer. In this way a co-concentration of solubilization agents (e.g. detergent) can be avoided.
5.3
OMP Scale-Up Production
Up to now and to the best of the author’s knowledge the only ’-helical membrane protein produced on industrial scale is the aquaporin protein class. Since aquaporins can be defined as nanosized water-treatment filters, as they allow the specific passage of water molecules at rates near to the diffusion limit [220] they have been used for the development of water purifying systems by the Danish company Aquaporin A/S (http:// www.aquaporin.dk/) (see also Sect. 2.2.3). Scaleup procedures however mostly rely still on the conventional expression, extraction and purification methods, though a cell-free expression system in synthetic liposomes allowing the production of aquaporin Z amounts on the milligram scale has been demonstrated [221] and the cellfree expression approach has been suggested to have great potentials as a platform for the industrial scale protein production [222]. As no other membrane proteins and therefore OMPs are used for industrial applications, yet, no
5.3 OMP Scale-Up Production
high-end production procedures have been developed. However some general considerations can be made based on the pre-requisite that protein yield and purity have to be maximized, while time-loss and costs need to be minimized in order to make a production procedure industrially interesting: 1. The “-barrel OMP expression into the outer membrane and subsequent extraction/solubilization by mild detergents cannot feasibly be used to develop high level production procedures as the membrane offers rather limited space and especially channel forming OMPs can have toxic effects for the expression host cells. Furthermore mild detergents often show low extraction efficiency and the necessary membrane isolation is a time consuming affair (see Sect. 5.2.1). Expression systems with large internal membrane systems, such as Rhodobacter that are used to over-produce ’helical MPs [223] are not suitable for OMPs due to the nature of the outer membrane. 2. The “-barrel OMP expression into the outer membrane and subsequent extraction/solubilization using strong detergents such as SDS lead to much higher extraction efficiencies but the protein becomes unfolded due to the denaturing effects of SDS. The protein can however be refolded by procedures used also to refold inclusion body derived proteins resulting in considerably high yields (see Sect. 5.2.1, Figs. 5.11 and 5.15). A recent report of the successful transformation of SDS from a denaturing detergent to a non-denaturing one with high extraction/solubilization efficiency [132, 178] shows that SDS surely is a solubilization agent that should be considered when OMPs need to be produced in high concentrations. Especially as SDS also solubilizes inclusion body material. 3. The “-barrel OMP expression into inclusion bodies followed by inclusion body solubilization and protein refolding avoids the space limit posed by the outer membrane and avoids toxic effects (such as osmotic imbalances caused by the over-expression
131
Fig. 5.15 FhuA1–159 obtainable yields (in mg per 1 L of culture) by membrane extraction with the mild detergent octyl-POE, membrane extraction by SDS followed by refolding using PE-PEG and inclusion body expression and refolding again using PE-PEG
of a passive diffusion channel OMP) on the host cells. Obtained yields are high when compared to membrane expression and extraction by mild detergents (Fig. 5.15). The expression into inclusion bodies is a valid outer membrane protein over production platform especially for engineered OMP variants that affect host cells in a negative way when expressed into the membrane or that cannot be accommodated by the membrane due to newly introduced features (i.e. elongated hydrophobic portion) (see Sect. 5.2.2). A drawback are the often time consuming refolding procedures, however this problem can be avoided by the use of SDS in combination with the amphipathic diol solvent 2-methyl-2,4-pentanediol that upon addition leads to protein folding [132, 178]. 4. Cell-free expression systems equally avoid the membrane space limit as well as OMP overexpression related toxic effects on the expression host, as it even is completely independent from the cellular system of a living organism. As optimized cell-free expression systems already report good overall yields (see Sect. 5.2.3) it is surely a powerful protein over-expression method that once efficient ways of direct solubilization or protein reconstitution are found and optimized can be used as a basis for an OMP production scale-up.
132
5. The use of cost-effective solubilization agents (e.g. PE-PEG diblock copolymer) is preferable over often costly detergents. Figure 5.15 shows obtainable yields (in mg protein per L of initial culture) by expression into the membrane and mild detergent extraction [40], expression into the membrane, solubilization by SDS and subsequent dialysis based refolding using diblock copolymer PE-PEG [133] and expression into inclusion bodies, followed by dialysis based refolding again using diblock copolymer PE-PEG [41] on the example of FhuA variant FhuA1–159. In all cases E. coli had been used as expression host. For better comparison values have been normalized assuming a purity of 100 %. While the conventional membrane expression method shows low yields of 1–2 mg/L of culture rendering it unfeasible for any industrial scale production, both membrane expression and solubilization by SDS and inclusion body expression lead to much higher yield, permitting 10– 20 times more obtained protein from the same culture volume. Though SDS solubilization of proteins from the outer membrane led to highest yields it has to be mentioned that the involved procedure is the most time consuming one, as it includes the isolation of the outer membrane, a protein extraction step and protein refolding by step-wise dialysis. The step-wise dialysis necessary to refold inclusion body derived protein renders the respective protocol almost equally time ineffective, even though the isolation of inclusion bodies and inclusion body solubilization can be achieved rather fast. The combination of inclusion body expression with SDS solubilization and refolding by simple addition of 2-methyl-2,4-pentanediol or similar substances seems therefore rather promising and should be tried in the future. In conclusion however it has to be noted that the nature of membrane proteins in general and the resulting special demands they make in terms of hydrophobicity of their environment and presence of a boundary surface (hydrophobic/hydrophilic) an industrial largescale production in the range of bulk chemicals is very unlikely and scale-up intentions will most
5 Biotechnology
likely reach to the scale of fine chemical or pharmaceutics production rates, as required for nano-channel applications such as drug-delivery from nano-containers or OMP use as stochastic nano-sensors (Details on the nano-technological applications of OMPs can be found in Chap. 6).
5.4
Artificial “-Barrel Structures
In Parallel to the efforts to produce nano-channels based on engineered bacterial “-barrel OMPs a different scientific approach instead considered the development of synthetic nano-channels that are based on artificial “-barrel structures able to insert into the lipid bilayer [224–227]. These artificial and self-assembling “-barrels are based on rigid rod molecules. Rigid rod molecules are extremely rigid, synthetic rod-shaped molecules with great potential in material sciences as they have one huge advantage; they do not fold, avoiding folding problems one might encounter during production of any protein derived materials [224]. The repeating rigid rod unit of an artificial “-barrel can be the p-octiphenyl rod. To each of the rods phenyl rings short peptide sequences that show “-strand shape are attached, leading to the barrel monomer. These monomers spontaneously assemble to form antiparallel “-sheets that are rolled up into a cylindrical structure promoted by the amphiphilicity of the individual sheets. In the final barrel adjacent amino acid residues face alternately the inner and outer barrel surface, similar to the geometry of “-barrel protein (see also Sect. 2.3.1) [224, 228]. By introducing hydrophobic and hydrophilic amino acids of which the first face the barrel outside, while the latter face the barrel inside, resulting barrel structures are soluble in organic solvents and can form pores in bilayer membranes [224, 229]. Triggering of opening and closing upon an external stimulus (e.g. pH, ionic strength, voltage) of artificial barrels with further chemically modified inside amino acid residues had been achieved and other chemically modified inside facing residues allowed the recognition of for example magnesium cations, and phosphate anions,
References
nucleotides, carbohydrates, inositol phosphates, pyrenes, calixarenes, and fullerenes, oligosaccharides, RNA, DNA, peptides, and synthetic polymers [224], rendering such pore structures useable as stochastic nano-sensing elements. Future further development of this technology can be considered as an alternative to biologically derived nano-channel materials especially considering the possibility to tune length and diameter of the pores. However the synthesis of these molecules is not trivial and large scale production might be difficult [225] and as described in Sects. 5.1.2.3 and 5.1.2.4 the E. coli FhuA protein structure has been shown to allow the engineering of protein nano-channels with altered geometrical features (changing length and diameter) reaching a similar tune-ability. In the following Chap. 6 protein nano-channel reconstitution methods will be briefly introduced, the main applications of these reconstituted nanochannels will be explored and methods of system characterization will be explained.
References 1. Howorka S, Siwy Z (2009) Nanopore analytics: sensing of single molecules. Chem Soc Rev 38:2360–2384 2. Onoda A, Fukumoto K, Arlt M, Bocola M, Schwaneberg U, Hayashi T (2012) A rhodium complex-linked “-barrel protein as a hybrid biocatalyst for phenylacetylene polymerization. Chem Commun 48:9756–9758 3. Onaca O, Sarkar P, Roccatano D, Friedrich T, Hauer B, Grzelakowski M, Güven A, Fioroni M, Schwaneberg U (2008) Functionalized nanocompartments (Synthosomes) with a reduction-triggered release system. Angew Chem Int Ed 47:7029–7031 4. Güven A, Dworeck T, Fioroni M, Schwaneberg U (2011) Residue K556-A light triggerable gatekeeper to sterically control translocation in FhuA. Adv Eng Mater 13:B324–B329 5. Lolicato M, Reina S, Messina A, Guarino F, Winterhalter M, Benz R, De Pinto V (2011) Generation of artificial channels by multimerization of “-strands from natural porin. Biol Chem 392:617–624 6. Richardson SM, Wheelan SJ, Yearrington RM, Boeke JD (2006) GeneDesign: rapid, automated design of multikilobase synthetic genes. Genome Res 16:550–556 7. Blattner FR, Plunkett G, Bloch CA, Perna NT, Burland V, Riley M, Collado-Vides J, Glasner JD, Rode CK, Mayhew GF, Gregor J, Davis NW, Kirkpatrick
133
8. 9.
10.
11.
12.
13.
14.
15.
16.
17.
18. 19.
20.
21.
22.
HA, Goeden MA, Rose DJ, Mau B, Shao Y (1997) The complete genome sequence of Escherichia coli K-12. Science 277:1453–1474 Hershberg R, Petrov DA (2008) Selection on codon bias. Annu Rev Genet 42:287–299 Vimberg V, Tats A, Remm M, Tenson T (2007) Translation initiation region sequence preferences in Escherichia coli. BMC Mol Biol 8:100–113 Zhang W, Xiao W, Wei H, Zhang J, Tian Z (2006) mRNA secondary structure at start AUG codon is a key limiting factor for human protein expression in Escherichia coli. Biochem Biophys Res Commun 349:69–78 Hochuli E, Bannwarth W, Döbeli H, Gentz R, Stüber D (1988) Genetic approach to facilitate purification of recombinant proteins with a novel metal chelate adsorbent. Nat Biotechnol 6:1321–1325 Smith DB, Johnson K (1988) Single-step purification of polypeptides expressed in Escherichia coli as fusions with glutathione S-transferase. Genetics 67:31–40 Schmidt TG, Skerra A (1993) The random peptide library-assisted engineering of a C-terminal affinity peptide, useful for the detection and purification of a functional Ig Fv fragment. Protein Eng 6:109–122 Voss S, Skerra A (1997) Mutagenesis of a flexible loop in streptavidin leads to higher affinity for the Strep-tag II peptide and improved performance in recombinant protein purification. Protein Eng 10:975–982 Einhauer A, Jungbauer A (2001) The FLAG peptide, a versatile fusion tag for the purification of recombinant proteins. J Biochem Biophys Methods 49:455– 465 Kapust RB, Waugh DS (1999) Escherichia coli maltose-binding protein is uncommonly effective at promoting the solubility of polypeptides to which it is fused. Protein Sci 8:1668–1674 Davis GD, Elisee C, Newham DM, Harrison RG (1999) New fusion protein systems designed to give soluble expression in Escherichia coli. Biotechnol Bioeng 65:382–388 Waugh DS (2005) Making the most of affinity tags. Trends Biotechnol 23:316–320 Raghava G, Sahni G (1994) GMAP: a multi-purpose computer program to aid synthetic gene design, cassette mutagenesis and the introduction of potential restriction sites into DNA sequences. Biotechniques 16:1116–1123 Hoover DM, Lubkowski J (2002) DNAWorks: an automated method for designing oligonucleotides for PCR- based gene synthesis. Nucleic Acids Res 30:e43 Fuglsang A (2003) Codon optimizer: a freeware tool for codon optimization. Protein Expr Purif 31:247–249 Gao W, Rzewski A, Sun H, Robbins P, Gambotto A (2004) UpGene: application of a web-based DNA codon optimization algorithm. Biotechnol Prog 20:443–448
134 23. Grote A, Hiller K, Scheer M, Munch R, Nortemann B, Hempel DC, Jahn D (2005) JCat: a novel tool to adapt codon usage of a target gene to its potential expression host. Nucleic Acids Res 1:W526–W531 24. Jayaraj S, Reid R, Santi DV (2005) GeMS: an advanced software package for designing synthetic genes. Nucleic Acids Res 33:3011–3016 25. Lorimer D, Raymond A, Walchli J, Mixon M, Barrow A, Wallace E, Grice R, Burgin A, Stewart L (2009) Gene Composer: database software for protein construct design, codon engineering, and gene synthesis. BMC Biotechnol 9:36 26. Raab D, Graf M, Notka F, Schödl T, Wagner R (2010) The GeneOptimizer Algorithm: using a sliding window approach to cope with the vast sequence space in multiparameter DNA sequence optimization. Syst Synth Biol 4:215–225 27. Wu G, Bashir-Bello N, Freeland SJ (2006) The synthetic gene designer: a flexible web platform to explore sequence manipulation for heterologous expression. Protein Expr Purif 47:441–445 28. Richardson SM, Nunley PW, Yarrington RM, Boeke JD, Bader JS (2010) GeneDesign 3.0 is an updated synthetic biology toolkit. Nucleic Acids Res 38:2603–2606 29. Jung S, McDonald K (2011) Visual gene developer: a fully programmable bioinformatics software for synthetic gene optimization. BMC Bioinformatics 12:340 30. Villalobos A, Ness JE, Gustafsson C, Minshull J, Govindarajan S (2006) Gene Designer: a synthetic biology tool for constructing artificial DNA segments. BMC Bioinformatics 7:285–293 31. Freigassner M, Pichler H, Glieder A (2009) Tuning microbial hosts for membrane protein production. Microb Cell Fact 8:69 32. Steffensen L, Pedersen PA (2006) Heterologous expression of membrane and soluble proteins derepresses GCN4 mRNA translation in the yeast Saccharomyces cerevisiae. Eukaryot Cell 5:248–261 33. Wagner S, Bader ML, Drew D, de Gier JW (2006) Rationalizing membrane protein overexpression. Trends Biotechnol 24:364–371 34. Wagner S, Klepsch MM, Schlegel S, Appel A, Draheim R, Tarry M, Hogbom M, van Wijk KJ, Slotboom DJ, Persson JO, de Gier JW (2008) Tuning Escherichia coli for membrane protein overexpression. Proc Natl Acad Sci USA 105:14371–14376 35. Braun M, Killmann H, Braun V (1999) The betabarrel domain of FhuADelta5–160 is sufficient for TonB-dependent FhuA activities of Escherichia coli. Mol Microbiol 33:1037–1049 36. Killmann H, Braun M, Herrmann C, Braun V (2001) FhuA barrel-cork hybrids are active transporters and receptors. J Bacteriol 183:3476–3487 37. Scott DC, Cao Z, Qi Z, Bauler M, Igo JD, Newton SM, Klebba PE (2001) Exchangeability of N termini in the ligand-gated porins of Escherichia coli. J Biol Chem 276:13025–13033
5 Biotechnology 38. Braun M, Killmann H, Maier E, Benz R, Braun V (2002) Diffusion through channel derivatives of the Escherichia coli FhuA transport protein. Eur J Biochem 269:4948–4959 39. Nallani M, Onaca O, Gera N, Hildenbrand K, Hoheisel W, Schwaneberg U (2006) A nanophosphor-based method for selective DNA recovery in Synthosomes. Biotechnol J 1:828–834 40. Nallani M, Benito S, Onaca O, Graff A, Lindemann M, Winterhalter M, Meier W, Schwaneberg U (2006) A nanocompartment system (Synthosome) designed for biotechnological applications. J Biotechnol 123:50–59 41. Dworeck T, Petri AK, Muhammad N, Fioroni M, Schwaneberg U (2011) FhuA deletion variant 1– 159 overexpression in inclusion bodies and refolding with Polyethylene-Poly(ethylene glycol) diblock copolymer. Protein Expr Purif 77:75–79 42. Boulanger P, Le Maire M, Bonhivers M, Dubois S, Desmadril M, Letellier L (1996) Purification and structural and functional characterization of FhuA, a transporter of the Escherichia coli outer membrane. Biochemistry 35:14216–14224 43. Rodríguez-Ropero F, Fioroni M (2012) Structural and dynamical analysis of an engineered FhuA channel protein embedded into a lipid bilayer or a detergent belt. J Struct Biol 177:291–301 44. Locher KP, Rees B, Koebnik R, Mitschler A, Moulinier L, Rosenbusch JP, Moras D (1998) Transmembrane signaling across the ligand-gated FhuA receptor: crystal structures of free and ferrichrome-bound states reveal allosteric changes. Cell 95:771–778 45. Endriß F, Braun V (2004) Loop deletions indicate regions important for FhuA transport and receptor functions in Escherichia coli. J Bacteriol 186:4818–4823 46. Mohammad M, Howard KR, Movileanu L (2011) Redesign of a plugged “-barrel membrane protein. J Biol Chem 286:8000–8013 47. Chen M, Khalid S, Sansom MSP, Bayley H (2008) Outer membrane protein G: engineering a quiet pore for biosensing. Proc Natl Acad Sci 105:6272–6277 48. Nestorovich EM, Danelon C, Winterhalter M, Bezrukov SM (2002) Designed to penetrate: timeresolved interaction of single antibiotic molecules with bacterial pores. Proc Natl Acad Sci 99:9789– 9794 49. Jung Y, Bayley H, Movileanu L (2006) Temperature-responsive protein pores. J Am Chem Soc 28:15332–15340 50. Naveed H, Jimenez-Morales D, Tian J, Pasupuleti V, Kenney LJ, Liang J (2012) Engineered oligomerization state of OmpF protein through computational design decouples oligomer dissociation from unfolding. J Mol Biol 419(1–2):89–101 51. Song L, Hobaugh MR, Shustak C, Cheley S, Bayley H, Gouaux JE (1996) Structure of staphylococcal alpha-hemolysin, a heptameric transmembrane pore. Science 274:1859–1866
References 52. Subbarao GV, van den Berg B (2006) Crystal structure of the monomeric porin OmpG. J Mol Biol 360:750–759 53. Cowan SW, Garavito RM, Jansonius JN, Jenkins JA, Karlsson R, Konig N, Pai EF, Pauptit RA, Rizkallah PJ, Rosenbach JP (1995) The structure of OmpF porin in a tetragonal crystal form. Structure 3:1041–1050 54. Tieleman DP, Berendsen HJ (1998) A molecular dynamics study of the pores formed by Escherichia coli OmpF porin in a fully hydrated palmitoyloleoylphosphatidylcholine bilayer. Biophys J 74:2786–2801 55. Schulz GE (2002) The structure of bacterial outer membrane proteins. Biochim Biophys Acta 1565:308–317 56. Ferguson AD, Hofmann E, Coulton JW, Diederichs K, Welte W (1998) Siderophore-mediated iron transport: crystal structure of FhuA with bound lipopolysaccharide. Science 282:2215–2220 57. Mohammad M, Iyer R, Howard KR, McPike M, Borer PN, Movileanu L (2012) Engineering a rigid protein tunnel for biomolecular detection. J Am Chem Soc 134:9521–9531 58. Tenne SJ, Dworeck T, Fioroni M (2010) Unpublished raw data. Department for Biotechnology, RWTH Aachen University, Aachen 59. Bendezú FO, de Boer PAJ (2008) Conditional lethality, division defects, membrane involution, and endocytosis in mre and mrd shape mutants of Escherichia coli. J Bacteriol 190:1792–1811 60. Liu P, Duan W, Wang Q, Li X (2010) The damage of outer membrane of Escherichia coli in the presence of TiO2 combined with UV light. Colloids Surf B 78:171–176 61. Glover WA, Yang Y, Zhang Y (2009) Insights into the molecular basis of L-form formation and survival in Escherichia coli. PLoS One 4:e7316 62. Schmid B, Krömer M, Schulz GE (1996) Expression of porin from Rhodopseudomonas blastica in Escherichia coli inclusion bodies and folding into exact native structure. FEBS Lett 381:111–114 63. Sidorova OV, Isaeva MP, Khomenko VA, Portniagina O, Likhatskaia GN, Kim N, Novikova OD, Chistiulin DK, Solov’eva TF (2012) Yersinia pseudotuberculosis mutant OmpF porins with deletions of the external loops: genetic constructions design, expression, isolation and refolding. Bioorg Khim 38:156–165 64. Koide S, Huang X, Link K, Koide A, Bu Z, Engelman DM (2000) Design of single-layer b-sheets without a hydrophobic core. Nature 403:456–460 65. Schulz GE (2000) “-barrel membrane proteins. Curr Opin Struct Biol 10:443–447 66. Huang Y, Smith BS, Chen LX, Baxter RH, Deisenhofer J (2009) Insights into pilus assembly and secretion from the structure and functional characterization of usher PapC. Proc Natl Acad Sci USA 106:7403–7407
135 67. Krewinkel M, Dworeck T, Fioroni M (2012) Engineering of an E. coli outer membrane protein FhuA with increased channel diameter. J Nanobiotechnol 9:33 68. Krewinkel M, Dworeck T, Fioroni M (2011) Unpublished data. RWTH Aachen University, Aachen 69. Muhammad N, Dworeck T, Fioroni M, Schwaneberg U (2011) Engineering of the E. coli outer membrane protein FhuA to overcome the hydrophobic mismatch in thick polymeric membranes. J Nanobiotechnol 9:8 70. Ahmed F, Pakunlu R, Brannan A, Bates F, Minko T, Discher D (2006) Biodegradable polymersomes loaded with both paclitaxel and doxorubicin permeate and shrink tumors, inducing apoptosis in proportion to accumulated drug. J Control Release 116:150–158 71. Lee J, Bermudez H, Discher B, Sheehan M, Won Y, Bates F, Discher D (2001) Preparation, stability, and in vitro performance of vesicles made with diblock copolymers. Biotechnol Bioeng 73:135–145 72. Nardin C, Hirt T, Leukel J, Meier W (2000) Polymerized ABA triblock copolymer vesicles. Langmuir 16:1035–1041 73. Nardin C, Thoeni S, Widmer J, Winterhalter M, Meier W (2000) Nanoreactors based on (polymerized) ABA-triblock copolymer vesicles. Chem Commun 15:1433–1434 74. Nardin C, Widmer J, Winterhalter M, Meier W (2001) Amphiphilic block copolymer nanocontainers as bioreactors. Eur Phys J E 4:403–410 75. Mouritsen O, Bloom M (1984) Mattress model of lipid-protein interactions in membranes. Biophys J 46:141–153 76. Muhammad N, Dworeck T, Fioroni M (2011) Unpublished data. RWTH Aachen University, Aachen 77. Ihle S, Onaca O, Rigler P, Hauer B, Rodr{guezRopero F, Fioroni M, Schwaneberg U (2011) Nanocompartments with a pH release system based on an engineered OmpF channel protein. Soft Matter 7:532–539 78. Cheley S, Gu LQ, Bayley H (2002) Stochastic sensing of nanomolar inositol 1,4,5-trisphosphate with an engineered pore. Chem Biol 9:829–838 79. Fsihi H, Kottwitz B, Bremer E (1993) Single amino acid substitutions affecting the substrate specificity of the Escherichia coli K-12 nucleoside-specific Tsx channel. J Biol Chem 268:17495–17503 80. Basle E, Joubert N, Pucheaul M (2010) Protein chemical modification on endogenous amino acids. Chem Biol 17:213–227 81. Hermanson GT (2008) Bioconjugate techniques. Academic, Amsterdam 82. Güven A, Fioroni M, Hauer B, Schwaneberg U (2010) Molecular understanding of sterically controlled compound release through an engineered channel protein (FhuA). J Nanobiotechnol 8:14 83. Patchornik A, Amit B, Woodwars RB (1970) Photosensitive protecting groups. J Am Chem Soc 92:6333–6335
136 84. Bös C, Braun V (1997) Specific in vivo thiollabeling of the FhuA outer membrane ferrichrome transport protein of Escherichia coli K-12: evidence for a disulfide bridge in the predicted gating loop. FEMS Microbiol Lett 153:311–319 85. Bös C, Lorenzen D, Braun V (1998) Specific in vivo labeling of cell surface-exposed protein loops: reactive cysteines in the predicted gating loop mark a ferrichrome binding site and a ligand-induced conformational change of the Escherichia coli FhuA protein. J Bacteriol 180:605–613 86. Howorka S, Movileanu L, Lu X, Magnon M, Cheley S, Braha O, Bayley H (2000) A protein pore with a single polymer chain tethered within the lumen. J Am Chem Soc 122:2411–2416 87. Gu LQ, Braha O, Conlan S, Cheley S, Bayley H (1999) Stochastic sensing of organic analytes by a pore-forming protein containing a molecular adapter. Nature 398:686–690 88. Walker BJ, Bayley H (1994) A pore-forming protein with a protease-activated trigger. Protein Eng 7:91–97 89. Braha O, Walker BJ, Cheley S, Kasianowicz JJ, Song L, Gouaux E, Bayley H (1997) Designed protein pores as components for biosensors. Chem Biol 4:497–505 90. Kasianowicz JJ, Brandin E, Branton D, Deamer DW (1996) Characterization of individual polynucleotide molecules using a membrane channel. Proc Natl Acad Sci USA 93:13770–13773 91. Heinz C, Engelhardt H, Niederweis M (2003) The core of the tetrameric mycobacterial porin MspA is an extremely stable “-sheet domain. J Biol Chem 278:8678–8685 92. Faller M, Niederweis M, Schulz GE (2004) The structure of a mycobacterial outer-membrane channel. Science 303:1189–1192 93. Butler T, Pavlenok M, Derrington IM, Niederweis M, Gundlach JH (2008) Single-molecule DNA detection with an engineered MspA protein nanopore. Proc Natl Acad Sci 105:20647–20652 94. Manrao EA, Derrington IM, Laszlo AH, Langford KW, Hopper MK, Gillgren N, Pavlenok M, Niederweis M, Gundlach JH (2012) Reading DNA at single-nucleotide resolution with a mutant MspA nanopore and phi29 DNA polymerase. Nat Biotechnol 30:349–354 95. Chang CY, Niblack B, Walker BJ, Bayley H (1995) A photogenerated pore-forming protein. Chem Biol 2:391–400 96. Miercke L, Ross PE, Strout RM, Dratz EA (1989) Purification of bacteriorhodopsin and characterization of mature and partially processed forms. J Biol Chem 264:7531–7535 97. Nestel U, Wacker T, Woitzik D, Weckesser J, Kreutz W, Welte W (1989) Crystallization and preliminary X-ray analysis of porin from Rhodobacter capsulatus. FEBS Lett 242:405–408 98. Stauffer KA, Page MGP, Hardmeyer A, Keller TA, Pauptit RA (1989) Crystallization and prelim-
5 Biotechnology
99.
100.
101.
102.
103.
104.
105.
106.
107.
108.
109.
110.
111.
inary X-ray characterization of maltoporin from Escherichia coli. J Mol Biol 211:297–299 Forst D, Schülein K, Wacker T, Diederichs K, Kreutz W, Benz R, Welte W (1993) Crystallization and preliminary X-ray diffraction analysis of ScrY, a specific bacterial outer membrane porin. J Mol Biol 229:258–262 Ferguson AD, Breed J, Diederichs K, Welte W, Coulton JW (1998) An internal affinity-tag for purification and crystallization of the siderophore receptor FhuA, integral outer membrane protein from Escherichia coli K-12. Protein Sci 7:1636–1638 Bannwarth M, Schulz GE (2003) The expression of outer membrane proteins for crystallization. Biochim Biophys Acta 1610:37–45 Schwarz D, Klammt C, Koglin A, Löhr F, Schneider B, Dötsch V, Bernhard F (2007) Preparative scale cell-free expression systems: new tools for the large scale preparation of integral membrane proteins for functional and structural studies. Methods 41:355–369 Miroux B, Walker JE (1996) Over-production of proteins in Escherichia coli: mutant hosts that allow synthesis of some membrane proteins and globular proteins at high levels. J Mol Biol 260:289–298 Dascher C, Roll D, Bavoil PM (1993) Expression and translocation of the chlamydial major outer membrane protein in Escherichia coli. Microb Pathog 15:455–467 Thomas KL, Leduc I, Olsen B, Thomas CE, Cameron DW, Elkins C (2001) Cloning, overexpression, purification, and immunobiology of an 85-kilodalton outer membrane protein from Haemophilus ducreyi. Infect Immun 69:4438–4446 Findlay HE, McClafferty H, Ashley RH (2005) Surface expression, single-channel analysis and membrane topology of recombinant Chlamydia trachomatis major outer membrane protein. BMC Microbiol 5:5 Singh R, Gupta PK, Durga V, Rao P (2009) Expression and purification of the major outer membrane protein (OmpH) of Pasteurella multocida P52 from Escherichia coli. Vet Arhiv 79:591–600 Haghi F, Peerayeh SN, Siadat SD, Montajabiniat M (2011) Cloning, expression and purification of outer membrane protein PorA of Neisseria meningitidis serogroup B. J Infect Dev Ctries 5:856–862 Lee J, Kim S-H (2009) High-throughput T7 LIC vector for introducing C-terminus poly-histidine tags with variable lengths without extra sequences. Protein Expr Purif 63:58–61 Studier FW, Moffatt BA (1986) Use of bacteriophage T7 RNA polymerase to direct selective high-level expression of cloned genes. J Mol Biol 189:113–130 Gosh R, Steiert M, Hardmeyer A, Wang YF, Rosenbusch JP (1998) Overexpression of outer membrane porins in E. coli using pBluescript-derived vectors. Gene Expr 7:149–161
References 112. Prilipov A, Phale PS, Van Gelder P, Rosenbusch JP, Koebnik R (1998) Coupling site-directed mutagenesis with high-level expression: large scale production of mutant porins from E. coli. FEMS Microbiol Lett 163:65–72 113. Sambrook J, Fritsch EF, Maniatis T (1989) Molecular cloning. Cold Spring Harbor Laboratory Press, New York 114. Onoda T, Enokizono J, Kaya H, Oshima A, Freestone P, Norris V (2000) Effects of calcium and calcium chelators on growth and morphology of Escherichia coli L-form NC-7. J Bacteriol 182:1419–1422 115. Marr AG, Ingraham JL (1962) Effect of temperature on the composition of fatty acids in Escherichia coli. J Bacteriol 84:1260–1267 116. Garwin JL, Klages AL, Cronan JE (1980) “ketoacyl-acyl carrier protein synthase II of Escherichia coli. evidence for function in the thermal regulation of fatty acid synthesis. J Biol Chem 255:3263–3265 117. Zhang YM, Rock CO (2008) Membrane lipid homeostasis in bacteria. Nat Rev Microbiol 6:222–233 118. Harrison STL (1991) Bacterial cell disruption: a key unit operation in the recovery of intracellular products. Biotechnol Adv 9:217–240 119. Kleinig AR, Middelberg APJ (1998) On the mechanism of microbial cell disruption in high-pressure homogenisation. Chem Eng Sci 53:891–898 120. French CS, Milner HW (1955) Disintegration of bacteria and small particles by high-pressure extrusion, vol 1, Methods in enzymology. Academic, New York 121. Johnson BH, Hecht MH (1994) Recombinant proteins can be isolated from E. coli cells by repeated cycles of freezing and thawing. Biotechnology 12:1357–1360 122. Felix H (1982) Permeabilised cells. Anal Biochem 120:211–234 123. Wase DAJ, Patel YR (1985) Effect of cell volume on disintegration by ultrasonics. J Chem Technol Biotechnol 35B:165–173 124. Andrews BA, Asenjo JA (1987) Enzymatic lysis and disruption of microbial cells. Biotechnol Lett 5:273–277 125. Lutkenhaus JF (1977) Role of a major outer membrane protein in E. coli. J Bacteriol 131:631–637 126. Hantke K (1981) Regulation of ferric iron transport in Escherichia coli K-12: isolation of a constitutive mutant. Mol Gen Genet 191:288–292 127. Beis K, Nesper J, Whitfield C, Naismith JH (2004) Crystallization and preliminary X-ray diffraction analysis of Wza outermembrane lipoprotein from Escherichia coli serotype O9a:K30. Acta Crystallogr D60:558–560 128. Arnold T, Linke D (2008) The use of detergents to purify membrane proteins. Curr Protoc Protein Sci Chapter 4:Unit 4.8.1–4.8.30. doi: 10.1002/0471140864.ps0408s53
137 129. Gohon Y, Popot JL (2003) Membrane protein-surfactant complexes. Curr Opin Colloid Interface Sci 8:15–22 130. Mo Y, Lee B-K, Ankner JF, Becker JM, Heller WT (2008) Detergent-associated solution conformations of helical and beta-barrel membrane proteins. J Phys Chem B 112:13349–13354 131. Seddon AM, Curnow P, Booth PJ (2004) Membrane proteins, lipids and detergents: not just a soap opera. Biochim Biophys Acta 1666:105–117 132. Roussel G, Perpete EA, Matagne A, Tinti E, Michaux C (2013) Towards a universal method for protein refolding: the trimeric beta barrel membrane Omp2a as a test case. Biotechnol Bioeng 110:417–423 133. Petri AK, Hariskos I, Dworeck T (2010) Unpublished data. RWTH Aachen University, Aachen 134. Shultis DD, Purdy MD, Banchs CN, Wiener MC (2006) Outer membrane active transport: structure of the BtuB:TonB complex. Science 312:1396–1399 135. Peybay-Peyroula E, Garavito RM, Rosenbusch JP, Zulauf M, Timmins PA (1995) Detergents structure in tetragonal crystals of OmpF porin. Structure 3:1051–1059 136. Evanics F, Hwang PM, Cheng Y, Kay LE, Prosser RS (2006) Topology of an outer-membrane enzyme: measuring oxygen and water contacts in solution NMR studies of PagP. J Am Chem Soc 128:8256–8264 137. Hwang PM, Choy WY, Lo EI, Chen L, FormanKay JD, Raetz CR, Privé GG, Bishop RE, Kay LE (2002) Solution structure and dynamics of the outer membrane enzyme PagP by NMR. Proc Natl Acad Sci 99:13560–13565 138. Gutmann DAP, Mizohata E, Newstead S, Ferrandon S, Henderson PJF, van Veen HW, Byrne B (2007) A high-throughput method for membrane protein solubility screening: the ultracentrifugation dispersity sedimentation assay. Protein Sci 16:1422–1428 139. le Maire M, Champeil P, Möller JV (2000) Interaction of membrane proteins and lipids with solubilizing detergents. Biochim Biophys Acta 1508:86–111 140. Arachea BT, Sun Z, Potente N, Malik R, Isailovic D, Viola RE (2012) Detergent selection for enhanced extraction of membrane proteins. Protein Expr Purif 86:12–20 141. Winstone T, Duncalf KA, Turner RJ (2002) Optimization of expression and the purification by organic extraction of the integral membrane protein EmrE. Protein Expr Purif 26:111–121 142. Civjan NR, Bayburt TH, Schuler MA, Sligar SG (2003) Direct solubilization of heterologously expressed membrane proteins by incorporation into nanoscale lipid bilayers. Biotechniques 35:556–563 143. Inagaki S, Ghirlando R, Grisshammer R (2013) Biophysical characterization of membrane proteins in nanodiscs. Methods 59:287–300 144. Laemmli UK (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680–685
138 145. Ratha A, Glibowickaa M, Nadeaua VG, Chena G, Debera CM (2008) Detergent binding explains anomalous SDS-PAGE migration of membrane proteins. Proc Natl Acad Sci 106:1760–1765 146. Schlegel S, Klepsch MM, Gialama D, Wickström D, Slotboom DJ, de Gier JW (2010) Revolutionizing membrane protein overexpression in bacteria. Microb Biotechnol 3:403–411 147. Walker J (1994) The bicinchoninic acid (BCA) assay for protein quantitation. Methods Mol Biol 32:5–8 148. Waterborg J, Matthews H (1984) The Lowry method for protein quantitation. Methods Mol Biol 1:1–3 149. Kruger N (1994) The Bradford method for protein quantitation. Methods Mol Biol 32:9–15 150. Laage R, Langosch D (2001) Strategies for prokaryotic expression of eukaryotic membrane proteins. Traffic 2:99–104 151. Middelberg A (2002) Preparative protein refolding. Trends Biotechnol 20:437–443 152. Rudolph R (1995) Successful protein folding on an industrial scale. In: Cleland JL, Craik CS (eds) Protein engineering: principles and practices. Wiley, New York, pp 283–298 153. Dekker N, Merck K, Tommassen J, Verheij HM (1995) In vitro folding of Escherichia coli outer-membrane phospholipase A. Eur J Biochem 232:214–219 154. Mitraki A, Fane B, Haase-Pettingell C, Sturtevant J, King J (1991) Global suppression of protein folding defects and inclusion body formation. Science 253:54–58 155. Tsumoto K, Ejima D, Kumagai I, Arakawa T (2003) Practical considerations in refolding proteins from inclusion bodies. Protein Expr Purif 28:1–8 156. Steinle A, Li P, Morris DL, Groh V, Lanier LL, Strong RK, Spies T (2001) Interactions of human NKG2D with its ligands MICA, MICB, and homologs of the mouse RAE-1 protein family. Immunogenetics 53:279–287 157. O’Callaghan CA, Tormo J, Willcox BE, Blundell CD, Jakobsen BK, Stuart DI, McMichael AJ, Bell JI, Jones EY (1998) Production, crystallization, and preliminary X-ray analysis of the human MHC class Ib molecule HLA-E. Protein Sci 7:1264–1266 158. Khan RH, AppaRao KBC, Eshwari ANS, Totey SM, Panda AK (1998) Solubilization of recombinant ovine growth hormone with retention of native-like secondary structure and its refolding from the inclusion bodies of Escherichia coli. Biotechnol Prog 14:722–728 159. Roudolph R, Lilie H (1996) In vitro folding of inclusion body proteins. FASEB J 10:49–56 160. Buchanan SK (1999) Beta-barrel proteins from bacterial outer membranes: structure, function and refolding. Curr Opin Struct Biol 9:455–461 161. Stockel J, Doring K, Malotka J, Jahnig F, Dornmair K (1997) Pathway of detergent-mediated and peptide ligand-mediated refolding of heterodimeric class II major histocompatibility complex (MHC) molecules. Eur J Biochem 248:684–691
5 Biotechnology 162. Burgess RR (1996) Purification of overproduced Escherichia coli RNA polymerase sigma factor by solubilizing inclusion bodies and refolding from sarkosyl. Methods Enzymol 273:145–149 163. Cardamone M, Puri NK, Brandon MR (1995) Comparing the refolding and reoxidation of recombinant porcine growth hormone from a urea denatured state and from Escherichia coli inclusion bodies. Biochemistry 34:5773–5794 164. Hochuli E, Döbeli H, Schacher A (1987) New metal chelate adsorbent selective for proteins and peptides containing neighboring histidine residues. J Chromatogr 411:177–184 165. De Bernadez-Clark E, Schwarz E, Rudolph R (1999) Inhibition of aggregation side reactions during in vitro protein folding. Methods Enzymol 309:217–236 166. Petri AK, Dworeck T (2010) Unpublished data. RWTH Aachen University, Aachen 167. Charbonnier F, Köhler T, Pechère JC, Ducruix A (2001) Overexpression, refolding, and purification of the histidine-tagged outer membrane efflux protein OprM of Pseudomonas aeruginosa. Protein Expr Purif 23:121–127 168. Baldermann C, Engelhardt H (2000) Expression, two-dimensional crystallization, and threedimensional reconstruction of the beta8 outer membrane protein Omp21 from Comamonas acidovorans. J Struct Biol 131:96–107 169. Prince SM, Achtman M, Derrick JP (2002) Crystal structure of the OpcA integral membrane adhesin from Neisseria meningitidis. Proc Natl Acad Sci USA 99:3417–3421 170. Kumar PD, Krishnaswamy S (2005) Overexpression, refolding, and purification of the major immunodominant outer membrane porin OmpC from Salmonella typhi: characterization of refolded OmpC. Protein Expr Purif 40:126–133 171. Pautsch A, Vogt J, Model K, Siebold C, Schulz GE (1999) Strategy for membrane protein crystallization exemplified with OmpA and OmpF. Proteins 34:167–172 172. Kramer RA, Zandwijken D, Egmond MR, Dekker N (2000) In vitro folding, purification and characterization of Escherichia coli outer membrane protease OmpT. Eur J Biochem 267:885–893 173. Buchanan SK (1999) Overexpression and refolding of an 80-kDa iron transporter from the outer membrane of Escherichia coli. Biochem Soc Trans 27:903–908 174. Anbazhagan V, Qu J, Kleinschmidt JH, Marsh D (2008) Incorporation of outer membrane protein OmpG in lipid membranes: protein-lipid interactions and beta-barrel orientation. Biochemistry 47:6189–6198 175. Royer CA (1995) Fluorescence spectroscopy, Methods in molecular biology. Humana Press, Totowa 176. Moon CP, Fleming KG (2011) Using tryptophan fluorescence to measure the stability of membrane proteins folded in liposomes, vol 492, Methods in enzymology. Elsevier, Baltimore
References 177. Chow MK, Amin AA, Fulton KF, Fernando T, Kamau L, Batty C, Louca M, Ho S, Whisstock JC, Bottomley SP, Buckle AM (2006) The REFOLD database: a tool for the optimization of protein expression and refolding. Nucleic Acids Res 34:D207–D212 178. Michaux C, Pomroy NC, Prive GG (2008) Refolding SDS-denatured proteins by the addition of amphipathic cosolvents. J Mol Biol 375:1477–1488 179. Nirenberg MW, Matthaei JH (1961) The dependence of cell-free protein synthesis in E. coli upon naturally occurring or synthetic polyribonucleotides. Proc Natl Acad Sci USA 47:1588–1602 180. Klammt C, Löhr F, Schäfer B, Haase W, Dötsch V, Rüterjans H, Glaubitz C, Bernhard F (2004) High level cell-free expression and specific labeling of integral membrane proteins. Eur J Biochem 271:568–580 181. Zubay G (1973) In vitro synthesis of protein in microbial systems. Annu Rev Genet 7:267–287 182. Gaisor E, Herrera F, Sadnik I, McLaughlin CS, Moldave K (1979) The preparation and characterization of a cell-free system from Saccharomyces cerevisiae that translates natural messenger ribonucleic acid. J Biol Chem 254:3965–3969 183. Swerdel MR, Fallon AM (1989) Cell-free translation in lysates from Spodoptera frugiperda (Lepidoptera:Noctuidae) cells. Comp Biochem Physiol 93B:803–806 184. Roberts BE, Patterson BM (1973) Efficient translation of tobacco mosaic virus RNA and rabbit globin 9S RNA in a cell-free system from commercial wheat germ. Proc Natl Acad Sci 70:2330–2334 185. Mosca JD, Wu JM, Suhadolnik PJ (1983) Restoration of protein synthesis in lysed rabbit reticulocytes by the enzymatic removal of adenosine 50 -monophosphate with either AMP deaminase or AMP nucleosidase. Biochemistry 22:346–354 186. Jermutus L, Ryabova LA, Plückthun A (1998) Recent advances in producing and selecting functional proteins by using cell-free translation. Curr Opin Biotechnol 9:534–548 187. Jackson AM, Boutell J, Cooley N, He M (2004) Cell-free protein synthesis for proteomics. Brief Funct Genomic Proteomic 2:308–319 188. Kigawa T, Yabuki T, Yoshida Y, Tsitsui M, Ito Y, Shibata T, Yokoyama S (1999) Cell-free production and stable-isotope labeling of milligram quantities of proteins. FEBS Lett 442:15–19 189. Shimizu Y, Inoue A, Tomari Y, Suzuki T, Yokogawa T, Nishikawa K, Ueda T (2001) Cell-free translation reconstituted with purified components. Nat Biotechnol 19:751–755 190. Ohashi H, Kanamori T, Shimizu Y, Ueda T (2010) A highly controllable reconstituted cell-free systema breakthrough in protein synthesis research. Curr Pharm Biotechnol 11:267–271 191. Katzen F, Chang G, Kudlicki W (2005) The past, present and future of cell-free protein synthesis. Trends Biotechnol 23:150–156
139 192. Carlson ED, Gan R, Hodgman CE, Jewett MC (2012) Cell-free protein synthesis: applications come of age. Biotechnol Adv 30:1185–1194 193. He M (2008) Cell-free protein synthesis: applications in proteomics and biotechnology. Nat Biotechnol 25:126–132 194. Gite S, Lim M, Rothschild KJ (2006) Cell-free protein synthesis systems: biotechnological applications. Biotechnol Genet Eng Rev 22:151–169 195. Jewett MC, Calhoun KA, Voloshin A, Wuu JJ, Swartz JR (2008) An integrated cell-free metabolic platform for protein production and synthetic biology. Mol Syst Biol 4:220 196. Kim DM, Swartz JR (1999) Prolonging cell-free protein synthesis with a novel ATP regeneration system. Biotechnol Bioeng 66:180–188 197. Spirin AS, Baranov VI, Ryabova LA, Ovodov SY, Alakhov YB (1988) A continuous cell-free translation system capable of producing polypeptides in high yield. Science 242:1162–1164 198. Kim DM, Choi CY (1996) A semi-continuous prokaryotic coupled transcription/translation system using a dialysis membrane. Biotechnol Prog 12:645–649 199. Sawasaki T, Hasegawa Y, Tsuchimochi M, Kamura N, Ogasawara T, Kuroita T, Endo Y (2002) A bilayer cell-free protein synthesis system for highthroughput screening of gene products. FEBS Lett 514:102–105 200. Jewett MC, Swartz JR (2004) Mimicking the Escherichia coli cytoplasmic environment activates long-lived and efficient cell-free protein synthesis. Biotechnol Bioeng 86:19–26 201. Klammt C, Schwarz D, Fendler K, Haase W, Dötsch V, Bernhard F (2005) Evaluation of detergents for the soluble expression of a-helical and b-barreltype integral membrane proteins by a preparative scale individual cell-free expression system. FEBS J 272:6024–6038 202. Shenkarev ZO, Lyukmanova EN, Butenko IO, Petrovskaya LE, Paramonov AS, Shulepko MA, Nekrasova OV, Kirpichnikov MP, Arseniev AS (2013) Lipid-protein nanodiscs promote in vitro folding of transmembrane domains of multi-helical and multimeric membrane proteins. Biochim Biophys Acta 1828:776–784 203. Periasamy A, Shadiac N, Amalraj A, Garajová S, Nagarajan Y, Waters S, Mertens HD, Hrmova M (2013) Cell-free protein synthesis of membrane (1,3)-“-d-glucan (curdlan) synthase: cotranslational insertion in liposomes and reconstitution in nanodiscs. Biochim Biophys Acta 1828:743–757 204. Ishihara G, Goto M, Saeki M, Ito K, Hori T, Kigawa T, Shirouzu M, Yokoyama S (2005) Expression of G protein coupled receptors in a cell-free translational system using detergents and thioredoxin-fusion vectors. Protein Expr Purif 41:27–37 205. Kalmbach R, Chizhov I, Schumacher MC, Friedrich T, Bamberg E, Engelhard M (2007) Functional
140
206.
207.
208.
209.
210.
211.
212.
213.
214.
215.
216.
5 Biotechnology cell-free synthesis of a seven helix membrane protein: in situ insertion of bacteriorhodopsin into liposomes. J Mol Biol 371:639–648 Goren MA, Fox BA (2008) Wheat germ cell-free translation, purification, and assembly of a functional human stearoyl-CoA desaturase complex. Protein Expr Purif 62:171–178 Jarecki BW, Makino S, Beebe ET, Fox BG, Chanda B (2013) Function of Shaker potassium channels produced by cell-free translation upon injection into Xenopus oocytes. Sci Rep 3:1040 Nallani M, Andreasson-Ochsner M, Tan CW, Sinner EK, Wisantoso Y, Geifman-Shochat S, Hunziker W (2011) Proteopolymersomes: in vitro production of a membrane protein in polymersome membranes. Biointerphases 6:153–157 Nguyen TA, Lieu SS, Chang G (2010) An Escherichia coli-based cell-free system for large-scale production of functional mammalian membrane proteins suitable for X-ray crystallography. J Mol Microbiol Biotechnol 18:85–91 Liguori L, Marques B, Villegas-Méndez A, Rothe R, Lenormand JL (2007) Production of membrane proteins using cell-free expression systems. Expert Rev Proteomics 4:79–90 Burgess RR, Deutscher MP (2009) Guide to protein purification, vol 463, 2nd edn, Methods in enzymology. Academic, Waltham von Jagow G, Schägger H (1994) A practical guide to membrane protein purification, vol 2, Separation, detection, and characterization of biological macromolecules. Academic, San Diego Arnold FH, Haymore BL (1991) Engineering metalbinding proteins: purification to protein folding. Science 252:1796–1797 Van Gelder P, Steiert M, El Khattabi M, Rosenbusch JP, Tommassen J (1996) Structural and functional characterization of a His-tagged PhoE pore protein of Escherichia coli. Biochem Biophys Res Commun 229:869–875 Volokhina EB, Beckers F, Tommassen J, Bos MP (2009) The “-barrel outer membrane protein assembly complex of Neisseria meningitidis. J Bacteriol 191:7074–7085 Broutin I, Benabdelhak H, Moreel X, Lascombe MB, Lerouge D, Ducruix A (2005) Expression, purification, crystallizationand preliminary X-ray
217.
218.
219.
220.
221.
222.
223.
224. 225.
226.
227. 228.
229.
studies of the outer membrane efflux proteins OprM and OprN from Pseudomonas aeruginosa. Acta Crystallogr Sect F Struct Biol Cryst Commun 61:315–318 Mohanty AK, Simmons CR, Wiener MC (2003) Inhibition of tobacco etch virus protease activity by detergents. Protein Expr Purif 27:109–114 Stellwagen E (2009) Gel filtration. In: Burgess RR, Deutscher MP (eds) Guide to protein purification, vol 463, 2nd edn, Methods in enzymology. Academic, Waltham Luckey M (2008) Membrane structural biology – with biochemical and biophysical foundations. Cambridge University Press, New York Savage DF, Stroud RM (2007) Structural basis of aquaporin inhibition by mercury. J Mol Biol 368:607–617 Hovijitra NT, Wuu JJ, Peaker B, Swartz JR (2009) Cell-free synthesis of functional aquaporin Z in synthetic liposomes. Biotechnol Bioeng 104:40–49 Tang CY, Zhao Y, Wang R, Helix-Nielsen C, Fane AG (2013) Desalination by biomimetic aquaporin membranes: review of status and prospects. Desalination 308:34–40 Laible PD, Mielke DL, Hanson DK (2005) Membrane protein production: a bacterial “factory” in Rhodobacter. Screening 2:30–32 Sakai N, Mareda J, Matile S (2008) Artificial “barrels. Acc Chem Res 41:1354–1365 Baumeister B, Matile S (2000) Rigid-rod betabarrels as lipocalin models: probing confined space by carotenoid encapsulation. Chemistry 6:1739–1749 Sakai N, Matile S (2003) Synthetic multifunctional pores: lessons from rigid-rod beta-barrels. Chem Commun 21:2514–2523 Schwab PFH, Levin MD, Michl J (1999) Molecular rods. 1. Simple axial rods. Chem Rev 99:1863–1934 Sakai N, Mareda J, Matile S (2005) Rigid-rod molecules in biomembrane models: from hydrogenbonded chains to synthetic multifunctional pores. Acc Chem Res 38:79–87 Clark TD, Buehler LK, Ghadiri MR (1998) Selfassembling cyclic b-peptide nanotubes as artificial transmembrane ion channels. J Am Chem Soc 120:651–656