Polar Interactions Trump Hydrophobicity in Stabilizing the Self-Inserting Membrane Protein Mistic

Abstract

Canonical integral membrane proteins are attached to lipid bilayers through hydrophobic transmembrane helices, whose topogenesis requires sophisticated insertion machineries. By contrast, membrane proteins that, for evolutionary or functional reasons, cannot rely on these machineries need to resort to driving forces other than hydrophobicity. A striking example is the self-inserting Bacillus subtilis protein Mistic, which is involved in biofilm formation and has found application as a fusion tag supporting the recombinant production and bilayer insertion of other membrane proteins. Although this unusual protein contains numerous polar and charged residues and lacks characteristic membrane-interaction motifs, it is tightly bound to membranes in vivo and membrane-mimetic systems in vitro. Therefore, we set out to quantify the contributions from polar and nonpolar interactions to the coupled folding and insertion of Mistic. To this end, we defined conditions under which the protein can be unfolded completely and reversibly from various detergent micelles by urea in a two-state equilibrium and where the unfolded state is independent of the detergent used for solubilizing the folded state. This enabled equilibrium unfolding experiments previously used for soluble and β-barrel membrane proteins, revealing that polar interactions with ionic and zwitterionic headgroups and, presumably, the interfacial dipole potential stabilize the protein much more efficiently than nonpolar interactions with the micelle core. These findings unveil the forces that allow a protein to tightly interact with a membrane-mimetic environment without major hydrophobic contributions and rationalize the differential suitability of detergents for the extraction and solubilization of Mistic-tagged membrane proteins.

Structural Thermodynamics of myr-Src(2–19) Binding to Phospholipid Membranes

“…Best-fit parameter values and associated 95% confidence intervals are given in Table 2 interactions between the peptide moiety of myr-Src and lipid headgroups, which are not accounted for by GouyChapman theory. For instance, it has been reported that, in addition to purely Coulombic interactions, polar and charged residues can interact with lipid headgroups by monopoledipole and dipole-dipole or, more generally, multipolar contacts (44,45). In the present case, such a scenario excluding major changes regarding the myristoyl chain of myr-Src is supported by NMR data, because its geometric parameters are not noticeably affected by the introduction of DLPS ( Fig.…”

Mistic's membrane association and its assistance in overexpression of a human GPCR are independent processes

“…The present results show that, although the isolated A3 peptide retains its high affinity for LDAO observed in the context of the full‐length protein, H3 or A3 alone interact with lipid membranes neither in vivo nor in vitro , respectively. In fact, the invariant ∼30% helical content of A3 in the presence of various lipids is virtually identical to what has been found for the H3 segment of urea‐unfolded full‐length Mistic in the absence of any membrane‐mimetic system and only about half the average helicity of the native protein in different detergent micelles …”

“…), whereas values for transmembrane insertion were generally unfavorable. Surprisingly, structural studies have pointed out that most of the residues involved in binding the zwitterionic detergent LDAO are located in H3 and H4, which renders LDAO‐solubilized Mistic extremely stable against denaturant‐induced unfolding . The present results show that, although the isolated A3 peptide retains its high affinity for LDAO observed in the context of the full‐length protein, H3 or A3 alone interact with lipid membranes neither in vivo nor in vitro , respectively.…”

“…Thus, the protein constitutes a biophysical enigma because of its localization in the lipid bilayer despite the high content of polar and charged residues, which is incompatible with a canonical transmembrane topology of its four‐helix bundle . Over the past years, several attempts have been made to shed light onto its biophysical properties …”

“…1 Over the past years, several attempts have been made to shed light onto its biophysical properties. 2,[4][5][6][7][8] Commonly used prediction software fails to recognize Mistic as a membrane protein 7 as, contrary to canonical membrane proteins, the four a-helices of Mistic are predicted to have a positive free energy of insertion. 9,10 Interestingly, molecular dynamics simulations suggest that Mistic might even be soluble and folded in the absence of detergents, 7 in contrast with experimental observations that reveal protein aggregation in the absence of a membranemimetic environment.…”