To flip or not to flip: lipid-protein charge interactions are a determinant of final membrane protein topology

Abstract

The molecular details of how lipids influence final topological organization of membrane proteins are not well understood. Here, we present evidence that final topology is influenced by lipid-protein interactions most likely outside of the translocon. The N-terminal half of Escherichia coli lactose permease (LacY) is inverted with respect to the C-terminal half and the membrane bilayer when assembled in mutants lacking phosphatidylethanolamine and containing only negatively charged phospholipids. We demonstrate that inversion is dependent on interactions between the net charge of the cytoplasmic surface of the N-terminal bundle and the negative charge density of the membrane bilayer surface. A transmembrane domain, acting as a molecular hinge between the two halves of the protein, must also exit from the membrane for inversion to occur. Phosphatidylethanolamine dampens the translocation potential of negative residues in favor of the cytoplasmic retention potential of positive residues, thus explaining the dominance of positive over negative amino acids as co- or post-translational topological determinants.

Figure 1.

PE-induced post-insertion topological reorganization of LacY. (A) Change in topology of LacY assembled in −PE cells (left panel; Bogdanov et al., 2002) after post-assembly synthesis of PE (right panel; this paper) in strain AT2033. Rectangles define the TMs (Abramson et al., 2003) oriented with the cytoplasm above the figure. TMs (roman numerals), extramembrane domains (P for periplasmic and C for cytoplasmic as oriented in +PE cells), N terminus (NT), and C terminus (CT) are indicated. The locations of negatively charged and positively charged residues involved in salt bridges between TMs are indicated. The locations and names of amino acids substituted by cysteine and used for SCAM analysis are indicated near circles (extramembrane) or squares (within TMs). TMVII is indicated in a periplasmic location (left) and a membrane-inserted location (right). (B–D) AT2033 cells without (−) or with (+) sonication were treated with MPB as described in Materials and methods. Labeling was performed on samples either after initial assembly of LacY in −PE cells (−PE, with IPTG induction but before addition of aTc) or after removal of IPTG and induction of PE synthesis (+PE) for 3 h (maximum PE level) during logarithmic growth (OD₆₀₀ increased from 0.4 to 1–1.4). Western blotting was used to detect biotinylation of diagnostic cysteines that were accessible to MPB; domain and/or substitution positions are indicated. Images are horizontal strips of the LacY (33 kD) position. See Materials and methods for details of image acquisition. (B) MBP labeling was performed at pH 7.5 except for P3/I103C*, where labeling was done at pH 10.5. (C and D). Labeling pH is indicated, and the cysteine substitutions in TMII, TMVII, and TMIX were L54C, I230C, and S300C, respectively. The substitutions in P1 and C6 are those noted in (B), and “−C” refers to cysteine-less LacY. Sonication of the TMVII samples from −PE and +PE cells followed by SCAM analysis at pH 10.5 gave the same result as shown in lanes 3 and 4 from the left in D, respectively (not depicted).

Figure 2.

Effect of net positive charge of the N-terminal bundle cytoplasmic domains on TM orientation in −PE cells. (A) TM orientation of LacY in +PE cells (Bogdanov et al., 2002) (see Fig. 1 A for TM orientation in −PE cells and other details). Locations of positively (red) and negatively (green) charged residues are indicated (see Fig. S4 for exact locations). Name and number of residues changed in B–E are indicated, as well as the net charge of each extramembrane domain. LacY contained either a single cysteine replacement (B, C, and E) at H205 (C6) or an additional cysteine replacement (D) at G13 (NT). SCAM analysis with (+) or without (−) sonication (Son) is shown for LacY with a change of net charge from +2 (“wild type”) in each cytoplasmic domain C2, C4, or C6 to +3 for domain C2 (D68N) or C4 (E126Q) or C4 (E139Q) or C6 (E215Q). The combination of D68N and K73N resulted in no change in net charge for C2, and addition of L72K but retaining D68 increased C2 charge by +1. The presence (strain AL95/pDD72) or absence of PE (strain AL95) is indicated in B, and all derivatives in C–E were expressed in −PE cells (AL95). Images are horizontal strips of the LacY (33 kD) position. See Materials and methods for details of image acquisition.

Figure 3.

Effect of net negative charge of the N-terminal bundle cytoplasmic domains on TM orientation in +PE cells. (A) TM orientation of LacY in +PE cells (see Fig. 1 A for other details). Name and number of residues changed are indicated, as well as the net charge and change in net charge of each domain C2, C4, or C6. (B) LacY derivatives contained a single diagnostic cysteine at H205 (C6) (lanes 1–10) or G13 (NT) (lanes 11 and 12) were expressed in +PE cells (strain AL95/pDD72). SCAM analysis with (+) or without (−) sonication (Son) is shown for LacY with a change in net charge from +2 (lanes 1 and 2) to −2 for domain C2 either separately (K69E and K74E, lanes 3 and 4) or in combination with a change in net charge of domain C6 from +2 to −2 (K211E and R218E, lanes 5 and 6). The changes in domains C2 and C6 were combined with a change in domain C4 from +2 to 0 (K131E, lanes 7 and 8) or from +2 to −2 (K131E and R142E lanes 9–12). Images are horizontal strips of the LacY (33 kD) position. See Materials and methods for details of image acquisition.

Figure 4.

Effect of increasing hydrophobicity of TMVII. (A) See Fig. 1 A for additional details. Positions of cysteine replacements are indicated by black and white rectangles. The “X” in TMVII indicates the position of the D240I substitution. (B) LacY with a D240I substitution in TMVII and a cysteine substitution in either C6 (H205C) or TMVII (I230C) or NT (G13C) and C6 (H205C) was expressed in +PE (+, strain AL95/pDD72) or −PE (−, strain AL95) cells. Cells were treated with MPB without (−) or after (+) sonication (Son) at either pH 7.5 (−) or 10.5 (+) and analyzed by SCAM. Additional controls are shown elsewhere as follows: accessibility of C6 (H205C) at pH 7.5 without the D240I substitution in +PE and −PE cells is shown in Fig. 1 B; accessibility of TMVII at pH 10.5 (I230C) without the D240I substitution in +PE and −PE cells is shown in Fig. 1 D. Images are horizontal strips of the LacY (33 kD) position. See Materials and methods for details of image acquisition.

Figure 5.

Effect of TMVII hydrophobicity on topology of LacY mutants in +PE cells. SCAM analysis is shown for C2/C4/C6 (+2/+2/+2) or C2/C4/C6 (−2/−2/−2) LacY (see Fig. 3 B) expressed in +PE cells (AL95/pDD72). The domain containing the indicated diagnostic cysteine replacement is shown on the left. The amino acid at position 240 was either aspartic acid or isoleucine (see Fig. 4 A), as indicated. Biotinylation was performed at pH 7.5 (top panel) or pH 10.5 (bottom panel). Images are horizontal strips of the LacY (33 kD) position. See Materials and methods for details of image acquisition.

Figure 6.

PE and the positive inside rule. In the left panel a cytoplasmic domain is shown containing a mixture of negative and positive amino acids. PE suppresses or neutralizes the presence of negative residues (yellow circles), which increases the effective positive charge potential, thus favoring cytoplasmic retention of the domain. In the absence of PE (right panel) negative residues (green circles) exert their full potential, resulting in translocation of the domain with a lower effective net positive charge. The proton motive force (arrow) positive outward determines domain directional movement depending on the domain effective net charge as influenced by the lipid environment.