$$\rightleftharpoonup{xx}$$
$$\longleftharp{xx}$$,
$$\longrightharp{xx}$$,
Membrane proteins (MPs) are challenging targets in protein production studies due to their insolubility in aqueous environments. Conventional MP production platforms comprise cell-based systems such as E. coli, yeast, or eukaryotic cells. The synthesized recombinant MPs are either extracted from cell membranes or refolded from inclusion bodies1. After detergent solubilization, MPs can be transferred into suitable membrane environments by established in vitro reconstitution protocols. Besides vesicles and liposomes, MP reconstitution into planar membranes in the form of nanodiscs2 or salipro3 particles have become routine techniques in recent times. However, all these strategies imply detergent contact with MPs that can result in destabilization, dissociation of oligomers, and even loss of protein structure and activity4. Screening for optimal detergent solubilization and reconstitution conditions can therefore be tedious and time consuming5.
The open nature of cell-free (CF) systems allows the expression reaction to be directly supplied with preformed membranes with a defined lipid composition. In this lipid-based expression mode (L-CF), the synthesized MPs have the opportunity to co-translationally insert into the provided bilayers6,7 (Figure 1). Nanodiscs consisting of a membrane scaffold protein (MSP) surrounding a planar lipid bilayer disc8 appear to be particularly suitable for this strategy9,10. Nanodiscs can routinely be assembled in vitro with a variety of different lipids, they are very stable, and stocks can be concentrated up to 1 mM. However, MSP expression in E. coli and its purification is necessary. As an alternative strategy, MSP can be co-expressed together with the target MP in CF reactions supplied with liposomes11,12,13. DNA templates for both MSP and MP are added into the reaction and MP/nanodiscs can form upon expression. While MSP production is avoided, the co-expression strategy requires careful fine-tuning of the final DNA template concentrations and higher variations in the efficiency of sample production can be expected.
The co-translational insertion of MPs into membranes of preformed nanodiscs is a non-physiological and still largely unknown mechanism independent from translocon machineries and signal sequences13,14,15,16. Major determinants of the insertion efficiency are the type of membrane protein as well as the lipid composition of the provided membrane, with a frequent preference for negatively charged lipids15,17. As the nanodisc membranes are relatively confined in size, a substantial amount of lipids is released upon MP insertion18. Variation of nanodisc size enables insertion and tuning of higher oligomeric MP complexes15,18. Among others, the correct assembly of homooligomeric complexes was shown for the ion channel KcsA, for the ion pump proteorhodopsin (PR) and for the multidrug transporter EmrE15,18. MPs are likely to enter the symmetric nanodisc membrane from both sides at relatively equal frequency. It should therefore be considered that different monomers or oligomers inserted into one nanodisc may have opposite orientations. However, a bias in orientation could be caused by cooperative insertion mechanisms as reported for the formation of PR hexamers and KcsA tetramers18. A further bias in MP orientation might result from orientation switches of inserted MPs probably at the rim of the nanodisc membranes.
The production of CF lysates from E. coli strains is a reliable routine technique and can be performed in almost any biochemical laboratory19,20. It should be considered that besides disulfide bridge formation, most other post-translational modifications are absent if a MP is synthesized using E. coli lysates. While this might generate more homogenous samples for structural studies, it may be necessary to exclude potential effects on the function of individual MP targets. However, the efficient production of high quality samples of G-protein coupled receptors (GPCR)14,21,22, eukaryotic transporters23 or large heteromeric assemblies24 in E. coli CF lysates indicates their suitability for even complex targets. Preparative scale amounts (≈ 1 mg/mL) of a sample can be obtained with the two-compartment continuous exchange cell-free (CECF) configuration, composed of a reaction mixture (RM) and a feeding mixture (FM) compartment. The FM volume exceeds the RM volume 15 to 20-fold and provides a reservoir of low-molecular weight energy compounds and precursors19. The expression reaction is thus extended for several hours and the final yield of the MP target is increased. The RM and FM compartments are separated by a dialysis membrane with a 10-14 kDa cutoff. The two compartments require a special design of the CECF reaction container (Figure 1). Commercial dialysis cassettes as RM containers in combination with tailored plexiglass containers holding the FM are suitable examples. MP yields can further be manipulated by modifying the RM:FM ratios or by exchanging the FM after a certain period of incubation.
Yield and quality of a MP frequently require intense optimization of reaction parameters. An important advantage of CF expression is the possibility to modify and fine tune almost any compound according to the individual requirements of a MP. A straightforward strategy is to focus first on improving the yield of a MP by establishing a basic production protocol and then to optimize MP quality by fine tuning reaction and folding conditions. The absence of physiological processes in CF lysates and their reduced complexity result in high success rates for the efficient production of MPs25. Routine considerations for DNA template design and optimization of Mg2+ ion concentration are in most cases sufficient to obtain satisfactory yields26. Depending on expression mode, optimization of MP quality can become time consuming, as a larger variety of parameters need to be screened14,17,22.
To establish the described CF expression platform, protocols are necessary for the production of E. coli CF lysate (i), T7 RNA polymerase (ii), nanodiscs (iii), and the basic CECF reaction compounds (iv) (Figure 1). The E. coli K12 strain A1927 or BL21 derivatives are frequently used for the preparation of efficient S30 (centrifugation at 30,000 x g) lysates. Besides S30 lysates, corresponding lysates centrifuged at other g-forces (e.g. S12) may be used. The lysates differ in expression efficiency and in proteome complexity. The proteome of the S30 lysate prepared by the described protocol has been studied in detail28. The S30 proteome still contains some residual outer membrane proteins which could give background problems upon expression and analysis of ion channels. For such targets, the use of S80-S100 lysates is recommended. A valuable modification during lysate preparation is the induction of the SOS response by combined heat shock and ethanol supply at mid-log growth phase of the cells. The resulting HS30 lysates are enriched in chaperones and can be used in blends with S30 lysates for improved MP folding22. CF expression in E. coli lysates is operated as a coupled transcription/translation process with DNA templates containing promoters controlled by T7 RNA polymerase (T7RNAP). The enzyme can be efficiently produced in BL21(DE3) Star cells carrying the pAR1219 plasmid29.
Two copies of MSP1E3D1 assemble into one nanodisc with a diameter of 10-12 nm, but the protocol described below may also work for other MSP derivatives. However, nanodiscs larger than those formed with MSP1E3D1 tend to be less stable while smaller nanodiscs formed with MSP derivatives such as MSP1 may not accommodate larger MPs or MP complexes. MSP1E3D1 nanodiscs can be assembled in vitro with a large variety of different lipids or lipid mixtures. Preformed nanodiscs are usually stable for > 1 year at -80 °C, while stability may vary for different lipid components. For the screening of lipid effects on folding and stability of a MP, it is necessary to prepare a comprehensive set of nanodiscs assembled with a representative variety of lipids/lipid mixtures. The following lipids may give a good starting selection: 1,2-Dimyristoyl-sn-glycero-3-phospho-(1'-rac-glycerol) (DMPG), 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC), 1,2 dioleoyl-sn-glycero-3-phospho-rac-(1-glycerol) (DOPG), 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC), 1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-(1'-rac-glycerol) (POPG) and 1-palmitoyl-2-oleoyl-glycero-3-phosphocholine (POPC).
A protocol for the preparation of a 3 mL CECF reaction is described. Further up or down scaling in a 1:1 ratio is possible. For the two-compartment CECF configuration, a RM containing all compounds and a FM containing only the low-molecular weight compounds have to be prepared. Commercial 3 mL dialysis devices with 10-14 kDa MWCO can be used as RM containers, which are then placed into custom made plexiglass containers holding the FM (Figure 1D)30. The ratio of RM:FM is 1:20, so 60 mL of FM have to be prepared for 3 mL RM. Quality, concentration, or type of several components can be critical for the final yield and/or quality of the synthesized MP (Table 1). DNA templates should be prepared according to published guidelines and codon optimization of the reading frame of the target can further significantly improve product yield26. For preparative scale CECF reaction, an established protocol for the production of PR is described. To establish expression protocols for new MPs, it is usually necessary to perform optimization screens of certain compounds (Table 1) to improve yield and quality. For Mg2+ ions, a well-focused concentration optimum does exist that frequently shows significant variation for different DNA templates. Other CF reaction compounds such as new batches of T7RNAP or S30 lysates may further shift the optimal Mg2+ concentration. As an example, a typical Mg2+ screen within the range of 14-24 mM concentration and in steps of 2 mM is described. Each concentration is screened in duplicates and in analytical scale CECF reactions. As CECF reaction container, custom-made Mini-CECF Plexiglas containers30 holding the RM are used in combination with standard 24-well microplates holding the FM (Figure 1B). Alternatively, commercial dialysis cartridges in combination with 96-deep well microplates or other commercial dialyzer devices in appropriate setups may be used (Figure 1C).