Figure 1. Principle of a CECF expression system in a coupled transcription–translation mode. Protein expression takes place in the so‐called “reaction mix” (RM) that holds all components required for transcription and translation, most of which are derived from a cellular extract as, for example, from E. coli. Additionally, the RM contains the T7 polymerase, template DNA, nucleotides, and amino acids as well as an energy regenerating system and appropriate buffer conditions. The RM is dialyzed against a feeding mix that supplies the reaction with low‐molecular‐weight compounds while at the same time removing inhibitory side‐products.

Figure 2. Cell‐free expression of membrane proteins. The cell‐free system offers three different modes of expression. In the absence of a hydrophobic environment the membrane protein will be expressed insolubly (P‐CF). This membrane protein precipitate differs from the inclusion bodies known from E. coli as it readily resolubilizes in detergents. Adding a hydrophobic environment (detergents, mixtures of detergents and lipids, or alternative surfactants) into the cell‐free reaction yields soluble protein incorporated into a micelle or micelle‐like structure that can be used for purification directly (D‐CF). Alternatively, the addition of lipid membranes allows expression within a lipid bilayer (L‐CF). In addition to synthetic liposomes, inner membrane vesicles or soluble bicelles and nanodisks are potential additives.

Figure 3. Transmembrane segment‐enhanced labeling. The six amino acids A, F, G, I, L, and V cluster predominantly in the transmembrane region of helical membrane proteins (a) accounting for about 60% of all residues in this region. Labeling only those amino acids greatly reduces the resonance overlap as illustrated in the – TROSY experiments on the C‐terminal fragment of presenilin (b). At the same time, this labeling scheme preserves a high degree of information for the backbone assignment as long stretches with these labeled residues occur. According to their C ^α chemical shift, three subsets of amino acid groups can be identified distinguishing between G, A/L/F, and I/V (c). Partial backbone assignment can thus be achieved based on the most basic triple‐resonance experiments with short coherence transfer delays, such as HNCA and HNCOCA, and mapping the assigned connectivity onto the protein sequence.

Figure 4. Combinatorial labeling scheme for the E. coli tellurite transporter Δ‐TehA. Three different samples are described varying in the combination of ‐ and 1‐‐labeled amino acids as indicated on the left. In the – TROSY all ‐labeled amino acids can be observed, while in the two‐dimensional HN(CO) spectra only those with a preceding 1‐‐labeled amino acid are detectable. As an example, the indicated resonance in the TROSY spectra is present in all three samples and can thus be identified as Ala. In the corresponding HN(CO) spectra, the resonance is only observed in samples 2 and 3, but not in sample 1, indicating that the preceding residue must be a Leu. This Leu–Ala pair can thus be localized in the protein sequence and in this example corresponds to L117–A118 of Δ‐TehA.