Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

2.1 Introduction

Cell-free (CF) expression has emerged in the last decade as an efficient and fast approach for the production of membrane proteins (MPs) of diverse topologies and origin. Its unique design as an open accessible reaction helps to eliminate several central bottlenecks known from conventional cell-based MP expression systems. In general, problems with cell physiology, expression regulation and cell culture are reduced. On the other hand, the high diversity of CF reaction conditions requests increased time investments in controlling MP quality, fine-tuning of reaction conditions and designing sample evaluation strategies. Poor MP sample quality can be the result if this important requirement is overseen.

CF reactions can basically be operated in two flavors, the single compartment batch configuration and the two-compartment continuous exchange (CECF) configuration (Kigawa and Yokoyama 1991; Spirin et al. 1988). The batch configuration is the method of choice in throughput applications using microplate devices and analytical scale reactions (Kai et al. 2013; Savage et al. 2007; Schwarz et al. 2010). Batch reaction times are limited to few hours with consequently lower yields of protein, although a number of modifications are possible in order to considerably improve efficiencies. Higher protein yields are typically obtained with the CECF configuration containing a reaction mixture (RM) compartment containing all the high molecular weight compounds such as ribosomes, DNA template and enzymes, and a feeding mixture (FM) compartment with a certain amount of precursors such as amino acids and nucleotides. Protocols for batch and CECF configurations are highly variable and among others, expression efficiencies depend on (1) precursor concentrations, (2) energy regeneration systems, (3) RM–FM volume ratios and (4) the implementation of repeated FM exchanges.

An array of new applications, modifications, and strategies for the CF production of MP samples has been developed within the last decade. In particular, the tools for the modulation of MP quality already during translation by CF reaction condition tuning have been widely expanded. We therefore provide a current view on options and perspectives for successful MP production and we summarize diverse strategies based on CF expression technologies.

2.2 Selecting the Background: Different Extract Sources

The origin of the CF extract is the first selection to be made by approaching MP expression. In particular, within the last decade, a considerable number of new extract sources covering eukaryotic as well as prokaryotic origins have been introduced (Table 2.1). Major selection criteria before starting a CF expression approach are usually (1) the required amount of synthesized recombinant protein, (2) to provide the most favorable background for promoting protein folding, (3) to increase the likeability of posttranslational modifications, (4) general handling issues, system availability and costs.

Table 2.1 CF extract sources
Full size table

Expression efficiencies and other characteristics of the various systems still differ significantly and best compromises have to be found. While few micrograms of recombinant protein can usually be obtained in any system, the production of preparative scale levels approaching milligram yields out of 1 ml of RM is currently only routinely possible with extracts of Escherichia coli or wheat germs. Frequent limiting factors for protein production efficiency in cell extracts are high concentrations of endogenous degrading enzymes, poor synchronization of ribosome activity during cell growth, or stability problems of essential enzymes. It should be noted that extracts of cells showing even high expression activities in vivo such as yeasts might not be very efficient in CF expression. However, protocols in particular for the efficient CECF configuration are continuously being optimized and further potential for improved protein synthesis might exist. Most systems have now been adjusted as coupled transcription/translation systems including the efficient T7 promoter for protein production and accepting plasmid or linear DNA templates (Table 2.1). The addition of translation factors or considering specific template modifications might further be necessary depending on the selected system.

A critical issue is the availability of the different CF extracts. Most systems are available as standardized commercial kits, but quality optimization and specific applications often require the set up of individual expression reactions. The preparation protocols for the various cell extracts differ significantly with sometimes even high variations in extract batch quality (Table 2.1). For eukaryotic cell extract preparations, species possible to grow in defined cell cultures might be preferred or commercial sources might be considered. The relatively fast and efficient preparation protocol is a major advantage of using E. coli extracts. In addition, it is best characterized and a large variety of compounds useful for reaction modifications is available. The vast majority of current data on MP production have thus been obtained with E. coli extracts.

Depending on the intended applications, the proper formation of posttranslational modifications can be a key issue for protein sample quality evaluation. Disulfide bridge formation may be triggered independently from extract origins by modulating the reducing conditions, e.g., by adding redox systems into the reaction, by supporting disulfide bridge formation with chaperones or by chemical pretreatment of extracts (Goerke and Swartz 2008, Kim and Swartz 2004, Yin and Swartz 2004). More complex modifications, such as glycosylation, lipidation, or phosphorylation, are so far only described from systems with eukaryotic extracts such as rabbit reticulocytes, insect cells, or wheat germ and at analytical scales (Table 2.1). Many modifications require supplements such as canine pancreas microsomes into the CF reaction. If modifying enzymes are provided, posttranslational modifications such as N-glycosylation appear to be possible even in extracts of E. coli (Guarino and DeLisa 2012). However, it might stay challenging to combine quality and homogeneity of posttranslational modifications with high-level expression purposes.

2.3 Basic Protocol Development: Improving CF Expression Efficiency

Complexity of MP production in CF systems is mainly reduced to the basic transcription/translation process . Coordination of pathways for trafficking or translocation as well as suppressing toxic effects are usually less relevant issues. Protein expression in most CF systems is controlled by the phage T7-RNA polymerase, and the corresponding regulatory promoter and terminator elements in addition to system specific enhancers have to be provided. However, other promoters could work as well. With E. coli extracts, derivatives of standard Ptac promoters recognized by the endogenous E. coli RNA polymerase could give even relatively high expression levels (Shin and Noireaux 2010). DNA template constructs can be generated by overlap polymerase chain reaction (PCR) strategies and added as linear DNA fragment into the CF reaction (Ahn et al. 2005; Yabuki et al. 2007). Alternatively, plasmid DNA templates based on standard vectors such as, e.g., the pET or pIVEx series can be provided. DNA templates appear to be quite stable in CF reactions and final concentrations in between 2 and 10 ng/μl RM are already saturating (Haberstock et al. 2012) .

Initial problems with low expression efficiency are mainly associated with the translation process. Adjusting the proper Mg2+ ion optimum is mandatory for each new target and suboptimal conditions can have severe impacts on protein production (Rath et al. 2011; Schwarz et al. 2007). Abundance of rare codons could further reduce protein expression and induce mis-incorporation of amino acids or even the premature termination of translation. Low protein yields are even more frequently caused by the formation of unfavorable secondary structures of the mRNA involving the 5-prime end containing the translational initiation site. Modulating the nucleotide sequence of the 5-prime coding sequence can therefore be very efficient in order to improve expression (Ahn et al. 2007; Kralicek et al. 2011). A fast approach is the tag variation screen by analyzing the effects of a small number of short sequence-optimized expression tags (Haberstock et al. 2012). The tag variation constructs are generated by overlap PCR and the resulting products can directly be used as DNA templates in CF expression screens. The construction of large fusion proteins in order to improve expression is therefore usually not necessary. Expression monitoring can initially be performed via immunodetection by using C-terminal purification tags such as a poly(His)10-tag as antigen. In an ideal template design, the coding sequence is therefore modified with a C-terminal purification/detection tag, and, if necessary, with a short N-terminal expression tag (Fig. 2.1). If translation can be addressed properly with the above mentioned procedures, the protein production in CF systems is usually very efficient. In expression screens comprising MP targets of diverse sizes, topologies, and functions, high success rates could be achieved (Schwarz et al. 2010; Savage et al. 2007; Langlais et al. 2007) .

Fig. 2.1
figure 1

Basic steps for the design of CF expression reactions

Full size image

Expression monitoring by taking advantage of C-terminally attached derivatives of green fluorescent protein (GFP) could be useful for CF expression protocol development and fast protein quantification (Kai et al. 2013; Müller-Lucks et al. 2012; Nozawa et al. 2011; Roos et al. 2012). For MP expression, it must be considered that the folding of wild-type or red-shifted variants of GFP is hampered in the presence of most detergents (Roos et al. 2012). More resistant is the superfolder GFP derivative most likely due to its higher tolerance for chemical denaturants and its faster folding kinetics (Roos et al. 2012; Pedelacq et al. 2006). However, the folding of superfolder GFP might not correlate with the productive folding of the N-terminal target protein as it is speculated for other GFP derivatives (Pedelacq et al. 2006). Superfolder GFP might therefore only be considered as general expression monitor while fusions with other GFP derivatives may in addition also give some preliminary evidence of the target protein folding and quality .

2.4 Folded Precipitates: P-CF Expression

Depending on the strategy and choice of supplemented additives, several basic expression modes are possible for the CF production of MPs (Fig. 2.1). The selection of the expression mode may depend on the intended application of the MP sample, but it can also have drastic consequences on the resulting MP quality (Junge et al. 2010; Lyukmanova et al. 2012). An overview on the implementation of the different CF expression modes is given in Tables 2.2–2.4 and representative case studies published during the last decade are listed.

Table 2.2 Case studies of P-CF-expressed MPs
Full size table

In absence of any provided hydrophobic environment, the freshly translated MPs instantly precipitate in the RM. Successful expression in this precipitate forming (P-CF) production mode can thus even be monitored by increased turbidity of the RM during incubation. Folded structures of such P-CF-generated MP precipitates can be detected by solid state nuclear magnetic resonance (NMR) and resolubilized precipitates show significant structural overlaps with corresponding MP samples obtained after conventional in vivo production (Maslennikov et al. 2010). P-CF-expressed MPs can simply be harvested by centrifugation. The MP pellet is usually contaminated with a number of co-precipitated proteins from the extract. Washing with buffer containing mild detergents such as Brij derivatives can help to selectively reduce such contaminations. The MPs are then solubilized in buffer containing specific detergents. Best results are usually obtained with 1-myristoyl-2-hydroxy-sn-glycero-3-[phospho-rac-(1-glycerol)] (LMPG), 1-palmitoyl-2-hydroxy-sn-glycero-3-[phospho-RAC-(1-glycerol)] (LPPG), or sodium dodecyl sulfate (SDS; Table 2.2; Klammt et al. 2004; Klammt et al. 2012; Rath et al. 2011). Milder detergents such as n-dodecyl-phosphocholine (DPC) or n-dodecyl-b-D-maltoside (DDM), detergent cocktails or mixtures of detergents and lipids could further be useful depending on the MP target (Ma et al 2011).

Critical parameters for the resulting MP quality can be (1) detergent concentration and volume of the solubilization buffer , (2) temperature of solubilization, and (3) the subsequent exchange of the primary and usually relatively harsh solubilization detergent against secondary and considerably milder detergents, e.g., upon MP immobilization during affinity chromatography. Stabilization and high recovery of ligand binding active GPCRs could be obtained by this strategy (Junge et al. 2010; Klammt et al. 2011). Solubilization of P-CF pellets is fast and usually complete after gentle shaking for approximately 1 h. It should be noted that pellets of CF-expressed soluble proteins cannot usually be solubilized by that procedure as they are much more unstructured. Consequently, MPs having excessive soluble domains could therefore resist solubilization out of P-CF pellets.

Selecting the P-CF expression mode is the fastest approach and usually routinely employed for the first level of MP expression protocol development in order to tune protein production up to the desired yields (Junge et al. 2011). Even complex MPs such as 12 transmembrane segment containing eukaryotic ion transporters or the 10 transmembrane segment containing MraY translocase have been functionally synthesized in the P-CF mode (Keller et al. 2008; Ma et al. 2011). The P-CF mode is furthermore excellent for screening MP libraries (Langlais et al. 2007; Savage et al. 2007; Schwarz et al. 2010) and for the production of MP samples for structural analysis by NMR (Klammt et al. 2004; Maslennikov et al. 2010; Rath et al. 2011; Sobhanifar et al. 2010).

2.5 Production of Proteomicelles: D-CF Expression

CF expression systems can tolerate a considerable number of supplied hydrophobic compounds, while certain variations in between the different extract sources exist. The CF expression in the presence of detergents above their critical micellar concentration (CMC) can result into the co-translational solubilization of the expressed MPs and into the instant formation of proteomicelles (Fig. 2.1) .

Extensive evaluation of detergent tolerance has been performed with E. coli extracts (Blesneac et al. 2012; Gourdon et al. 2008; Klammt et al. 2005; Lyukmanova et al. 2012) as well as with wheat germ extract systems (Beebe et al. 2011; Genji et al. 2010; Kaiser et al. 2008; Periasamy et al. 2013). As primary compounds of choice, long-chain polyoxyethylene-alkyl-ethers such as Brij35, Brij58, Brij78 or Brij98, and the steroid-derivative digitonin have been determined (Table 2.3). These detergents have been successfully used for the solubilization of different G protein-coupled receptors (GPCRs) as well as of prokaryotic MPs (Table 2.3). Commonly employed detergents for the extraction of MPs out of native membranes such as DPC, the alkyl-glucoside n-dodecyl-β-D-maltoside (DDM), or n-octyl-β-D-glucopyranoside (β-OG) are too harsh or only tolerated at lower concentrations. However, the tolerance can sometimes be increased if critical detergents are provided as mixed micelles together with other detergents, e.g., CHAPS together with Fos-choline derivatives (Genji et al. 2010). It might generally be advantageous to combine the provided detergent micelles with some small amounts of lipids in case the translated MPs require interaction with some lipids for stabilization (Arslan Yildiz et al. 2013; Müller-Lücks et al. 2013; Nozawa et al. 2007). For the expression of mitochondrial carrier proteins, the addition of some cardiolipin together with fluorinated surfactants or Brij35 detergent had significant beneficial effects, whereas cardiolipin had negative effects in combination with the detergent Brij58 (Blesneac et al. 2012). The nature or composition of the selected detergent or detergent mixture can certainly affect the efficiency of solubilization as well as the MP quality in view of folding and stability. Systematic screens for detergent type and concentration are therefore necessary in order to determine optimal conditions for each individually expressed MP (Ishihara et al. 2005; Klammt et al. 2005; Liguori et al. 2008; Martin et al. 2009; Rath et al. 2011) .

Table 2.3 Case studies of D-CF-expressed MPs
Full size table

The open nature of CF reactions has initiated searches for further and new hydrophobic compounds with improved properties that could substitute classical detergents in co-translational MP solubilization (Fig. 2.2). Fluorinated surfactants as well as phospholipid-like surfactants have been considered as mild hydrophobic supplements in D-CF reactions (Park et al. 2011; Blesneac et al. 2012). Amphipols and in particular the charged A8-35 derivative are not well tolerated but might be interesting as solubilizing agents for P-CF-generated MP precipitates. The polyfructose-based uncharged polymer NV10 was claimed to be beneficial in supporting the solubilization of several class B GPCRs (Klammt et al. 2011). Similar positive effects on GPCR solubilization as described for Brij detergents have been observed with peptide surfactants as D-CF supplements (Wang et al. 2011; Corin et al. 2011; Table 2.3). However, curdlan synthase was inactive if D-CF expressed in the presence of Brij58, but active if the detergent was replaced by peptide surfactants (Periasamy et al. 2013). Other compounds such as the recently described maltose-neopentyl glycol amphiphiles might be considered in future as well (Chae et al. 2010). Although the general availability of several compounds is still somehow limited, it is evident that the variety for designing artificial hydrophobic environments in D-CF reactions is rapidly increasing .

Fig. 2.2
figure 2

Supplements for the co-translational and posttranslational modification of MP sample quality

Full size image

2.6 Designing Protein/Membrane Complexes: L-CF Expression

CF extracts are almost devoid of membranes, although some residual small vesicles originating from the cell membranes might be present after S30 preparation. More complete removal of membrane fragments can be achieved by S100 (centrifugation at 100,000 × g) extract preparation (Berrier et al. 2011). Instead of complex cell membranes, vesicles containing selected lipid compositions can be prepared in vitro and provided into the CF reactions. Lipids are, in contrast to detergents, much better tolerated by CF systems and mostly high final concentrations are possible (Kalmbach et al. 2007; Hovijitra et al. 2009; Roos et al. 2013; Umakoshi et al. 2009). This L-CF (lipid membrane-based) expression mode can thus facilitate the co-translational association of the expressed MPs with supplied bilayers provided as liposomes, as bicelles in combination with specific detergents or as planar membrane discs (Fig. 2.2). The L-CF mode is an excellent approach in order to evaluate lipid effects on the quality and activity of MPs (Table 2.4). Subsequent characterization of L-CF-generated MP samples can furthermore be performed in the natural context of membranes. The co-translational insertion might direct the synthesized MPs in a unidirectional inside-out orientation into the supplied membranes. This could be shown with connexins as example and thus more uniform samples can be generated if compared with conventional posttranslational reconstitution approaches (Moritani et al. 2010).

Table 2.4 Case studies of L-CF-expressed MPs
Full size table

It should be realized that by selecting the L-CF expression mode, translocation problems of the expressed MP can become an issue again. For many MPs, complex translocation machineries are essential for their proper membrane insertion in vivo (Shao and Hedge 2011). The knowledge of translocation mechanisms of MPs in artificial L-CF systems is still at the very beginning. However, it is already evident that the dependency of MP insertion on translocation systems might not be as strict as in vivo. The membrane insertions of the CF-expressed channel MscL and of the MtlA permease were independent of the YidC insertase and of the SecYEG complex, respectively (Berrier et al. 2011; Nishiyama et al. 2006). Characteristics that could facilitate membrane insertion might be that (1) the provided membranes are empty, (2) the membrane concentration in CF reaction can be high, and (3) lipid compositions of the supplied membranes can specifically be modulated, e.g., increase of anionic lipids can improve MP insertion efficiencies (Roos et al. 2012). It is generally advisable to perform a lipid screening with individual MPs in order to determine the appropriate composition of supplied membranes in view of lipid charge, length, and flexibility. If translocation systems appear to be mandatory, vesicles isolated out of complex cell membranes and containing natural translocation machineries could be considered as supplements as well (Stech et al. 2012).

L-CF expression in the presence of nanodiscs appears to be in particular promising in order to obtain soluble and functionally active MP/membrane complexes (Cappuccio et al. 2009; Lyukmanova et al. 2012; Proverbio et al. 2013; Roos et al. 2012). Specific advantages might be that nanodisc membranes are accessible from both sides and it could be speculated that inhomogeneities of the membrane/membrane scaffold protein interface may provide additional entry sides for MP integration. MP/nanodisc complexes are highly soluble and can be used for a variety of applications such as surface plasmon resonance (SPR) measurements (Proverbio et al. 2013). Nanodiscs provide furthermore a membrane compartment that is stable in size and topology. In contrast, liposomes supplied to CF reactions show excessive fusion and increase in size, resulting into their almost quantitative precipitation during the reaction (Barrier et al. 2011; Roos et al. 2013). An interesting modification by using the CECF configuration is the initial supply of a lipid/detergent mixture in the RM, whereas the FM is devoid of any detergent. Freshly translated MPs can therefore first physically associate with the provided lipomicelles and become increasingly trapped into membranes that slowly formed by the continuously decreased detergent concentration in the RM. This method was successfully applied for bacteriorhodopsin with combinations of steroid detergents and phosphatidylcholine lipids (Shimono et al. 2009). In a similar approach, the plant solute transporter AtPPT1 was functionally reconstituted by using a mixture of Brij35 and soybean asolectin lipids (Nozawa et al. 2007).

Besides hydrophobic environments, a variety of further additives could be beneficial for the production of high quality MP samples (Fig. 2.2). Chemical chaperones such as sugars, alcohols, or polyions are tolerated by CF systems and synergies of several compounds could be determined in correlated concentration screens (Kai et al. 2013). Such additives might be in particular beneficial for supporting the folding of larger soluble domains of MPs.

2.7 Handling the Toolbox of CF Expression: Strategies for Protein Quality Optimization

During the last decade when CF expression as a new platform for MP production was emerging, the three basic modes P-CF, D-CF, and L-CF have been employed in more or less comparable frequencies. The variety of supplements useful for MP quality optimization is rapidly expanding and defining specific conditions for the production of sufficient MP quantities that are homogenous, functionally folded, and stable is generally the key issue in CF expression protocols. The systematic screening of (1) expression modes, (2) type and concentration of hydrophobic compounds, (3) additives such as stabilizer or chaperones and (4) post-expression processing procedures generate an array of MP samples that have to be analyzed (Fig. 2.2). A strategic plan for MP quality control is therefore crucial and the first evaluation of solubility, sample homogeneity, protein integrity, stability, or specific activity should be implemented as soon as possible, at best already in the crude reactions. The above mentioned GFP fusion approach can, e.g., provide already a first short list of compounds and compound combinations giving an efficient solubilization of the MP. Enzymatic reactions or binding of labeled ligands might be considered as well in specific cases (Gourdon et al. 2008; Kalmbach et al. 2007). Solubility and integrity can be assayed with immunoblotting by taking advantage of terminal tags. Rare occurring MP fragmentation generated by premature translational termination or proteolytic degradation could be addressed by using synthetic genes and by screening protease inhibitors. In the case of some GPCRs, ligand-binding assays by SPR measurement can be performed with crude RMs containing the expressed MP (Proverbio et al. 2013). Feedback from first quality evaluations should then be considered for re-optimization, fine-tuning of compound concentrations, and for analyzing cocktails of beneficial compounds for synergistic effects. In many cases, tremendous variations in the quality of MP samples produced at different CF conditions have been observed. A striking example is the MraY translocase, a membrane-embedded enzyme responsible for lipid-I precursor formation in the bacterial cell wall biosynthesis pathway. The Bacillus subtilis MraY can be functionally synthesized in a large variety of CF conditions implementing detergents and lipids. In contrast, the E. coli MraY enzyme was only functional if L-CF synthesized in the presence of nanodiscs containing anionic lipids (Roos et al. 2012; Ma et al. 2011).

If a short list of few promising reaction conditions has been determined, more time-consuming assays using purified samples such as size exclusion chromatography, multi-angle light scattering, or circular dichroism spectroscopy can be implemented in order to analyze homogeneity, folding, and oligomeric states of the MP. Functionality of MPs is often more difficult to analyze and currently available case studies are compiled in Tables 2.2–2.4. As some examples, ligand binding of GPCRs and transporters were shown by radioassays (Gao et al. 2012; Ishihara et al. 2005; Sansuk et al. 2008; Yang et al. 2011), fluorescence anisotropy measurement (Junge et al. 2010), SPR (Kaiser et al. 2008; Proverbio et al. 2013), or thermophoresis (Corin et al. 2011). Functional samples of channels and transporters after co-translationally or posttranslationally reconstitution into lipid bilayers were obtained from MscL (Berrier et al. 2004, 2011), EmrE (Elbaz et al. 2004), TetA (Wuu and Swartz 2008), PorA and PorH (Rath et al. 2011), eukaryotic organic ion transporters (Keller et al. 2008, 2011), as well as from aquaporins (Hovijitra et al. 2009; Kai et al. 2010; Müller-Lucks et al. 2013).

2.8 Perspectives for Structural Approaches

The possibility to produce pure and concentrated samples of even very difficult MPs in a short time by CF expression is certainly of major interest for structural studies. Efficient incorporation of selenomethionine is of value for X-ray crystallography . In addition, MPs could already co-translationally be stabilized by inhibitors or other ligands. X-ray structures of CF-expressed MPs are still limited to the multidrug transporter EmrE (Chen et al. 2007) and Acetabularia proteorhodopsin (Wada et al. 2011) synthesized in a mixture of detergents and lipids. Crystallization was furthermore successful with the CF-expressed human voltage-gated anion channel VDAC1 (Nguyen et al. 2010; Deniaud et al. 2010). Despite some success, the still relatively low number of X-ray structures derived from CF-expressed MPs indicates that reaction conditions obviously may often not have been optimal in order to obtain crystallization grade MP samples. Systematic case studies are needed in order to identify the key parameters to be considered by choosing CF expression approaches. Sample homogeneity might be optimized by (1) intensive screening of reaction modes and hydrophobic supplements, (2) providing detergents in cocktails with some stabilizing lipids, (3) adjusting redox conditions for proper disulfide bridge formation, and (4) fine-tuning of solubilization conditions of P-CF samples.

The efficient and cost-effective labeling opportunities of CF expression are excellent prerequisites for structural approaches by NMR spectroscopy and a variety of sophisticated labeling tools and schemes have been developed and are already established standards (Klammt et al. 2012; Reckel et al. 2008; Ozawa et al. 2006). Nevertheless, liquid-state NMR of MPs is still a challenging task. MPs in micellar as well as in lipid environment are prone to signal broadening due to the large size and slow rotational tumbling. Furthermore, amino acids in α-helical structures tend to display narrow-range chemical shifts, resulting in severe peak overlaps. Numerous liquid-state NMR structures of P-CF (Klammt et al. 2012; Maslennikov et al. 2010; Sobhanifar et al. 2010) or D-CF (Reckel et al. 2008) expressed MPs have been reported (see also chapter by L. Catoire and D. Warschawski in this volume). For solid-state NMR, P-CF precipitates, samples posttranslationally reconstituted into liposomes, or L-CF samples by additions of liposomes or nanodiscs can be used (see also chapter by L. Catoire and D. Warschawski in this volume). Examples of successfully analyzed MPs are the mechanosensitive channel MscL (Abdine et al. 2010) and the multidrug transporter EmrE (Lehner et al. 2008). In particular, L-CF expression in the presence of nanodiscs can become attractive as shown in initial studies with proteorhodopsin (Mörs et al. 2013).

2.9 Conclusions

Besides exploring new approaches in particular for drug screening or single-molecule approaches, some major current challenges are the scale-up of CF reactions to industrial dimensions, the streamlined determination of expression conditions for crystallization grade MP samples, and the production of larger assemblies and MP complexes. Initial milestones have already been achieved and give promising perspectives. The manufacturing of multigram to kilogram scales in 100 L or even higher CF reaction volumes appears to become feasible (Zawada et al. 2011). Providing complex and more elaborated hydrophobic expression environments might be a direction for producing homogeneous samples suitable for crystallization (Wada et al. 2011). Tuning expression and template design might be a prerequisite for successful synthesis and assembly of MP complexes as recently been shown for the 542-kDa ATP synthase complex (Matthies et al. 2011).