Control of Membrane Protein Topology by a Single C-Terminal Residue

See allHide authors and affiliations

Science  25 Jun 2010:
Vol. 328, Issue 5986, pp. 1698-1700
DOI: 10.1126/science.1188950


The mechanism by which multispanning helix-bundle membrane proteins are inserted into their target membrane remains unclear. In both prokaryotic and eukaryotic cells, membrane proteins are inserted cotranslationally into the lipid bilayer. Positively charged residues flanking the transmembrane helices are important topological determinants, but it is not known whether they act strictly locally, affecting only the nearest transmembrane helices, or can act globally, affecting the topology of the entire protein. Here we found that the topology of an Escherichia coli inner membrane protein with four or five transmembrane helices could be controlled by a single positively charged residue placed in different locations throughout the protein, including the very C terminus. This observation points to an unanticipated plasticity in membrane protein insertion mechanisms.

Integral α-helical membrane proteins carry out a wide range of central biological functions. They have two conspicuous structural features: hydrophobic transmembrane α helices and a strong bias in the distribution of positively charged arginine (Arg) and lysine (Lys) residues between cytoplasmic and extracytoplasmic loops, with up to three times the frequency of Arg and Lys found in the cytoplasmic loops (1). Positively charged residues exert local control over the orientation of transmembrane helices in their immediate neighborhood (2, 3), but whether they can also affect the global topology of a protein is unknown. Multispanning membrane proteins insert into their target membrane cotranslationally; therefore, positively charged residues in a more C-terminal region of the protein might be expected not to be able to influence the orientation of distant N-terminal transmembrane helices. However, charged residues can affect the orientation of a protein domain, and individual transmembrane helices do not always insert into the membrane in a strict N- to C-terminal order or can reorient during the insertion process (47), which suggests a certain flexibility in the way multispanning membrane proteins are handled by the insertion machinery.

We asked to what extent global control of membrane protein topology by positively charged residues is possible, and so we looked for a multispanning protein that is poised near the threshold between an Nin topology with the N terminus facing the cytoplasm and an oppositely oriented Nout topology. EmrE is an Escherichia coli inner membrane protein with four transmembrane helices (TMHs) that meets this criterion. It belongs to the Small Multidrug-Resistance family of transporters and imparts resistance to toxic compounds like ethidium bromide (EtBr) and acriflavin (8). The active form of EmrE is a homodimer, and structural (9, 10), biochemical (1113), and phylogenetic (14) data support a so-called dual topology for EmrE, i.e., the monomers have a mixed membrane orientation (Fig. 1A). In the available two- and three-dimensional crystal structures (9, 10), two oppositely oriented monomers form an antiparallel homodimer, although other models have been proposed (13).

Fig. 1

(A) The dual-topology protein EmrE and the EmrE(Nin) and EmrE(Nout) constructs (11). Positively charged Arg and Lys residues are shown as black circles and the functionally important Glu14 residue (19) is shown as a white circle. (B) Growth of serial log10 dilutions of cells expressing the indicated constructs on a pH 7 ampicillin plate supplemented with 45 μg EtBr per ml. The normalized growth of a particular construct is calculated as the ratio of the area under its growth-dilution curve relative to that obtained for wild-type EmrE, after subtraction of the area for the empty vector control. (C) Normalized growth values for EmrE constructs discussed in the text. Error bars indicate ±1 SEM.

The homodimeric, dual-topology EmrE protein can be engineered into a functional heterodimer composed of two oppositely oriented monomers, EmrE(Nin) and EmrE(Nout), by suitable placement of positively charged residues in the two EmrE versions (11) (Fig. 1A). Moreover, the conserved glutamic acid residue Glu14 in TMH1 can be changed to aspartic acid (E14D mutation) in one, but not both, monomers while retaining a functional heterodimer, although the change imparts a lower toxin-resistance level (11). Thus, it is possible to determine the orientation of any EmrE mutant by coexpressing the mutant protein, including the E14D mutation, with either EmrE(Nin) or EmrE(Nout). Any homodimers formed by the mutant protein itself will be inactive because they carry the E14D mutation in both monomers, and coexpression of the mutant protein with either EmrE(Nin) or EmrE(Nout) will result in active protein only when an antiparallel heterodimer is formed. Quantitative activity assays can be carried out either in liquid culture (11) or on toxin-containing plates (15) (Fig. 1B). E. coli cells expressing EmrE(E14D), EmrE(Nin), or EmrE(Nout) by themselves were at best marginally resistant to EtBr (Fig. 1C). Coexpression of EmrE(Nin) with EmrE(Nout) made cells resistant to EtBr. Coexpression of EmrE(E14D) with either EmrE(Nin) or EmrE(Nout) also imparted resistance, as expected if EmrE(E14D) had a dual topology and could form active heterodimers with either EmrE(Nin) or EmrE(Nout).

Using the coexpression assay, we determined the topology for a series of EmrE(E14D) mutants with one extra positively charged Arg or Lys residue placed in the N-terminal tail, in one of the loops between the TMHs, or at the C terminus. The orientation of each mutant faithfully followed the positive-inside rule, i.e., the added positively charged residue always promoted the topology where this residue was in the cytoplasm (Fig. 2A), even for the most C-terminal mutation, where an Arg was added as the last residue in the protein. To confirm these results, we made a construct totally devoid of positively charged residues and then added back Arg or Lys residues, one at a time. The E14D mutant of the construct lacking Arg and Lys residues imparted EtBr resistance only when coexpressed with EmrE(Nout) (fig. S1), which indicated that it had adopted the Nin orientation. As expected, adding one Arg or Lys residue to the N-terminal tail, to the loop between TMH2 and TMH3, or to the C-terminal tail did not change the Nin orientation. Positively charged residues added to the TMH1-TMH2 or TMH3-TMH4 loops progressively changed the orientation to Nout. However, compared with the single-charge effects observed for EmrE(E14D), higher numbers of Arg or Lys were required to reach full inversion of the topology, presumably because, in this case, the starting topology was Nin rather than dual.

Fig. 2

(A) Topological effects of a single positively charged residue placed in the indicated positions in EmrE(E14D). Normalized growth values during coexpression with EmrE(Nin) (blue bars) and EmrE(Nout) (red bars). The predominating topology is shown for each construct as a miniature cartoon where black circles indicate positively charged (Arg or Lys) residues in EmrE(E14D), the crossed white circle indicates Asp14, and the red circle indicates the position of the added positively charged residue. C, cytoplasm; P, periplasm. (B) Topological effects of adding C-terminal His tails to EmrE (Glu14 is retained in these constructs). Blue circles show the normalized growth values for the indicated EmrE construct expressed by itself. Blue and red bars indicate coexpression with EmrE(Nin) and EmrE(Nout), respectively. Error bars indicate ±1 SEM.

That the topological effect of a positively charged residue was equally strong regardless of where in the protein it was placed prompted us also to study the effect of His residues. Histidine carries only a partial positive charge at the pH prevailing in the E. coli cytoplasm (16), and if charge is the controlling factor, its effect on topology should be weaker but qualitatively similar to that of Arg. Indeed, EmrE constructs with three or more His added at the C terminus were inactive when expressed alone but yielded a highly active protein when coexpressed with EmrE(Nout) (Fig. 2B). This appeared to be a result of the charge and not the length of the C-terminal tail, because the addition of up to six Gly had little effect on the dual topology of EmrE (fig. S2).

Finally, to examine whether a C-terminal positively charged residue could influence the global topology when moved even farther from the N terminus, we extended EmrE by adding a fifth TMH, composed only of alanines and leucines, to the C terminus. Given its composition, this TMH was not expected to interact in any specific way with TMHs 1 to 4. EmrE-TMH5 had an Nin topology, as it was inactive when expressed alone but imparted EtBr resistance—albeit at a lower level than wild-type EmrE—when coexpressed with EmrE(Nout) (Fig. 3). However, adding a C-terminal Lys to EmrE-TMH5 resulted in a protein [EmrE-TM5(K)] that imparted EtBr resistance only when coexpressed with EmrE(Nin). Thus, the C-terminal Lys can reverse the orientation of as many as five upstream TMHs. Finally, the Nin topology was regained when the C-terminal Lys was complemented with an N-terminal Lys [EmrE(K3)-TMH5(K)].

Fig. 3

Topological effects of adding Lys residues to the N- and C-termini of the EmrE-TMH5 construct (Glu14 is retained in these constructs). TMH5 (green in the miniature cartoons) is a GGPG…GPGG-flanked 19-residue-long segment composed of four Leu and 15 Ala. Normalized growth values during coexpression with EmrE(Nin) (blue bars) and EmrE(Nout) (red bars) are shown. Error bars indicate ±1 SEM.

In summary, the membrane orientation of the 4-TMH, dual-topology protein EmrE and a 5-TMH version of the same protein could be shifted both to Nin and to Nout by adding a single positively charged residue in various locations throughout the protein. In all cases, the shift in orientation was as predicted by the positive-inside rule. A C-terminal Arg or Lys was as effective in this regard as were positively charged residues placed in other locations closer to the N terminus. Apparently, the protein remains “topologically uncommitted” until the last residue has been synthesized. These and other observations of a related kind (17) raise important questions regarding the mechanism of membrane protein insertion and assembly. Specifically, how much protein can the translocon pore accommodate? Are translocon-associated proteins, such as YidC (18), involved in chaperoning membrane proteins to their final topology? Is postinsertion conversion between different topologies, so far seen only under conditions of extreme alterations in membrane lipid composition (17), possible also in wild-type cells?

Supporting Online Material

Materials and Methods

Figs. S1 and S2


References and Notes

  1. Materials and methods are available as supporting material on Science Online.
  2. We thank Ann-Louise Johansson for technical assistance. This work was supported by grants to G.v.H. from the European Research Council (ERC-2008-AdG 232648), the Swedish Cancer Foundation, the Swedish Research Council, and the Swedish Foundation for Strategic Research, to J.S.S. from the International Human Frontier Science Program Organization, and to P.L.G. from the European Communities (TranSys PITN-2008-215524).
View Abstract

Stay Connected to Science

Navigate This Article