Figure 4. The extrinsic FRET pair of SecY**EG reports on the distance across the lateral gate.
(A) Structures of SecYEG-SecA from T. maritima (left; part-open; PDB code 3DIN [Zimmer et al., 2008]) and SecYEβ from M. jannaschii (right; closed; PDB code 1RHZ [Van den Berg et al., 2004]). TMs 1–5 of SecY are coloured pink (TM2 highlighted magenta), TMs 6–10 grey (TM7 highlighted black), SecE orange, SecG/β green and SecA pale blue (2HF highlighted in bright blue). Residues equivalent to A103 and V353 in E. coli (K103 and I342 in T. maritima; I94 and I356 in M. jannaschii) are shown as yellow spheres, with the distances between them (Cβ-Cβ) marked out: 43.5 Å in the open complex, and 34.1 Å in the closed complex. (B) FRET efficiencies of 100 nM SecY**EG in PLs (light grey) or DDM-solubilised (dark grey): with 1 µM SecA; with SecA and 1 mM AMPPNP; with SecA and 1 mM ADP; and alone. Data are normalised to SecYEG with SecA but without added nucleotide, and error bars represent the standard error from six repeats.
Figure 4—figure supplement 1. The extrinsic FRET pair of SecY**EG reports on the distance across the lateral gate.
(A) ATPase assay on SecA in the presence of saturating amounts of SecYEG or SecY**EG PLs.Rates are shown both before and after addition of pre-protein (pOA) to 0.7 µM, and demonstrate that SecY**EG is active. (B) Schematic diagram of SecY, with the ten TMs shown, and the two labelling positions (A103 and V353) shown as black circles. The primary and secondary trypsin cleavage sites – between TMs 6–7 and TMs 8–9, respectively – are also marked out. Cleavage with trypsin yields one fragment of ~28 kDa containing the A103 position (F1), and another of ~21 kDa containing the V353 position (F2). F2 is partially cleaved to produce another 10 kDa fragment (F3), which also contains the V353 position. (C) Fluorescence visualisation of SDS-PAGE following trypsin cleavage of the two single mutants, SecYEGA103C (A) and SecYEGV353C (V), and the double mutant (AV) all conducted in DDM. Each pair of lanes shows SecYEG without (-) and with (+) trypsin treatment; in the absence of trypsin, fluorescent protein runs as a full length band (FL), while trypsin cleavage produces fragments F1–F3. Numbers in red show percentages of the total intensity for the bands in the lane. As expected, fluorophores on the A103 (A) position run in band F1 (samples 3 & 7), while fluorophores on V353 (V) are distributed between bands F2 and F3 (samples 4 & 8). Protein labelled with a single dye at both positions (samples 2 and 6) shows both positions labelled equally (samples 2 & 6). The SecY**EG mutant (samples 1 and 5) is labelled equally on both positions, with both dyes. Therefore, we can rule out an unexpected asymmetric dye loading on the two sites that would skew the FRET results. (D) Analysis of SecY**EG PLs by trypsin proteolysis, results shown without (-) and with (+) trypsin. Fluorescence quantification indicates that the band corresponding to full-length SecYEG accounts for 54% of the total signal in the '+' lane, once the lower molecular weight bands in the '-' lane are corrected for. This is equivalent to 46% outward facing SecYs: consistent, within error, with previous reports that show equal amounts of inward- and outward-facing SecYEG (Mao et al., 2013; Schulze et al., 2014). (E) Normalised FRET efficiencies of 50 nM SecY**EG in PLs in the presence of 500 µM ADP (red) or 500 µM AMPPNP (blue), as a function of SecA concentration. Data are fitted to a tight binding equation and show a difference in apparent FRET efficiency at saturation, but no measurable difference in affinity. Due to variability in protein preps, labelling efficiencies and reconstitutions, variation in the total magnitude of the signal was greater than the individual differences between ADP and AMPPNP. To account for this, we calculated the differences in FRET pairwise for each data set (inset). Error bars represent the standard error of mean (SEM) of five replicates, and the dotted blue line represents where the data would cluster if the ADP and AMPPNP titrations gave the same signal.
Figure 4—figure supplement 2. Interpreting ensemble FRET in PLs.
The FRET changes observed in Figure 4B are lower than would be expected from the structures of the open and closed channel (Figure 4A), given the known R0 of the dye pair (6 nm). Three factors are likely to cause this: (i) about half of all SecY**EG molecules will be facing inwards after reconstitution (Mao et al., 2013; Schulze et al., 2014) (Figure 4S1) and will thus be unavailable for contact with SecA; (ii) SecYEG forms a dimer in the membrane (Bessonneau et al., 2002; Breyton et al., 2002) and SecA engages only one channel for translocation (Osborne and Rapoport, 2007; Zimmer et al., 2008; Deville et al., 2011); and (iii) at most half of the SecYs will carry the donor-acceptor pair due to the random labelling – possibly fewer if labelling is incomplete. In total, therefore, only ~12.5% of all SecYEG molecules are expected to respond to the presence of translocation partners. Two of these three problems are obviated by using single molecule FRET: (ii) is solved by using high lipid to protein ratio (see Methods for details) and extruding the PLs to 100 nm, such that only a single copy of SecY is present in each PL (Deville et al., 2011); and (iii) is solved because only particles containing a single donor and a single acceptor are used for analysis (see Methods and Figure 5—figure supplement 1). Because we use intact PLs, problem (i) still remains even at the single molecule level. We account for this in the final reckoning (Figure 5H) by assuming that half the particles do not respond to SecA as they are inward-facing (Figure 4—figure supplement 1D) and behave as SecA with no additional binding partners. This was done simply by removing 50% of the 'alone' populations from each other condition, then multiplying the result by two to give 100%.