In Vivo Detection of Hetero-association of Glycophorin-A and Its Mutants within the Membrane*

Protein recognition within the membrane is a crucial process for numerous biological activities. Detection of such interaction is limited because of difficulties that arise from the hydrophobic environment of the membrane. We detected direct hetero-oligomerization of the glycophorin-A (GPA) transmembrane segments in vivo through inhibition of ToxR transcription activator dimer formation. We investigated the amino acids important for hetero-oligomerization within the membrane, using peptide analog segments of the transmembrane domain of glycophorin A. The wild type ([WT]GPA) and alanine mutant ([A]GPA) were able to interfere with and inhibit the proper dimerization of the ToxR-GPA transcription factor. Conversely, a second alanine mutant ([A 2 ]GPA), a gly- cine mutant ([G]GPA), and a scrambled analog ([SC]GPA) were virtually inactive. Binding studies reveal similar membrane partitions for [WT]GPA, [G]GPA, and [SC]GPA, whereas membrane partition of [A]GPA and [A 2 ]GPA are lower. Spectral analysis of fluorescent- labeled analogs revealed a significant blue shift, indicat-ing membrane insertion. Our results suggest that the G XXX G motif, found in homo-oligomerization, is not sufficient for hetero-oligomerization in a biological membrane, whereas an extended motif, LI XX G XXX G XXX T, is sufficient. Interfering with hetero-oligomerization within the membrane can be a useful strategy for characterizing such interactions and possibly modulating membrane protein activity. leading the of N-NBD the were thoroughly DMF and methylene peptides were cleaved from the and purified as described previously. of were prepared at a of phosphatidyl- glycerol (PG) (7:3, as previously described Lipids were resus-pended lipid In Vivo Detection of Hetero-association of Proteins within the Mem- brane— The ToxR transcription activator can be used successfully to assess weak protein interactions within the E. coli membrane. A GPA transmembrane-encoding DNA cassette was previously inserted between the maltose-binding protein and the ToxR transcription activator (2). Transcription activation is mediated by expressing the construct, ToxR-GPA, in the indicator strain FHK12. After transforming FHK12 cells, 1-ml cultures (8 repeats) were grown in the presence of chloram-phenicol and 0.1 (cid:2) M isopropyl 1-thio- (cid:3) - D -galactopyranoside for 20 h at 37 °C. (cid:3) -galactosidase activity was quantified in crude cell lysates after the addition of o- nitrophenylgalactoside and monitoring the reaction at 405 nm for 20 min at 30-s intervals at 28 °C. Specific (cid:3) -galactosidase activity was measured in Miller units (2). Hetero-association was de- tected using ToxR-GPA grown in the presence of an exogenous peptide.

Protein recognition within the membrane is crucial for a wide range of processes in all organisms. Examples include intersubunit association of receptors, ion channels, and pumps. Despite major progress, detection of such interactions in vivo is still limited. Thus, most of the current knowledge about heteroassociation of proteins in the membrane comes from in vitro experiments relying on strategies such as competition assays in SDS-polyacrylamide gels (1-3), cross-linking (4), and energy transfer (5)(6)(7).
Many biologically important membrane proteins associate via their transmembrane domains (8). This is demonstrated by reconstitution of a functional bacteriorhodopsin (9, 10) and lactose permease (11) from separate transmembrane segments. Other immunologically related proteins that form hetero-oligomeric complexes are the T cell receptor (12,13) and the MHC II complex (14). Fusion proteins also seem to oligomerize through their transmembrane domains. This includes the viral fusion proteins influenza hemagglutinin and hepatitis E1/E2, and the cellular fusion protein synaptobrevin (3,15,16). Other examples of biologically important membrane proteins include the M13 major coat protein (17), phospholambdan (18), and glycophorin A (GPA) 1 (1). Uncovering the mechanisms that drive these kind of interactions is critical for understanding how these proteins work.
GPA (1), a well documented example of transmembrane homophilic interactions, was previously shown to create ␣-helical right-handed dimers with a specific homo-association motif (1, 2, 5, 19 -21). A part of this motif, GXXXG, was shown to be crucial for the homo-oligomerization process. Furthermore, protein data base studies as well as transmembrane domain bacterial libraries suggest that this motif plays a general role in membrane protein-protein homo-association (19,22).
We chose GPA as a model protein to investigate heteroassociation of different transmembrane segments. GPA has a known homo-oligomerization motif, which was implicated in hetero-oligomerization in vitro (2,19,21,22). Until now, detection of hetero-association between proteins in the membrane was mostly restricted to in vitro methods. In the case of GPA, SDS-polyacrylamide gel electrophoresis experiments suggested hetero-oligomerization of a wild type transmembrane domain linked to staphylococcal nuclease with mutant peptides through the homo-oligomerization motif (19,23). We devised a scheme to detect direct interaction between hetero-proteins in a biological membrane. We then applied this method to wild type GPA ([WT]GPA) and peptide analogs to study their hetero-association.
Our results suggest the importance of an extended motif for hetero-association of GPA within the membrane of Escherichia coli. This study emphasizes the significance of using such displacement strategies in the investigation of membrane complexes. In turn, these strategies may prove useful in modulation of membrane protein activity.

EXPERIMENTAL PROCEDURES
Peptide Synthesis and Purification-The peptides were synthesized by a standard solid phase method on PAM-resin as described (24,25). The peptides were cleaved from the resin by HF treatment and purified by reverse phase-HPLC. Purity (ϳ99%) was confirmed by analytical HPLC. The peptide compositions were confirmed by Platform LCZ electrospray mass spectrometry.
Fluorescent Labeling of Peptides-The boc protecting group was removed from the N terminus of the peptides by incubation with trifluoroacetic acid, whereas all the other reactive amine groups of the attached peptides were kept protected. The acidity was neutralized with 5% (v/v) N,N-diisopropylethylamine (DIEA 5%). The resin-bound peptides were then treated with 4-chloro-7-nitrobenz-2-oxa-1,3-diazole * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  fluoride (NBD-F) (2 eq) in dry dimethylformamide (DMF), leading to the formation of resin-bound N-NBD peptides, respectively. After 1 h, the resins were washed thoroughly with DMF and then with methylene chloride. The labeled peptides were cleaved from the resin and purified as described previously.
Preparation of Small Unilamellar Vesicles (SUVs)-Lipid films were prepared at a ratio of phosphatidylethanolamine (PE) phosphatidylglycerol (PG) (7:3, w/w) as previously described (26). Lipids were resuspended by vigorous vortexing. SUVs were obtained by sonication of the lipid suspensions for 2 min in a water bath-type sonicator.
In Vivo Detection of Hetero-association of Proteins within the Membrane-The ToxR transcription activator can be used successfully to assess weak protein interactions within the E. coli membrane. A GPA transmembrane-encoding DNA cassette was previously inserted between the maltose-binding protein and the ToxR transcription activator (2). Transcription activation is mediated by expressing the construct, ToxR-GPA, in the indicator strain FHK12. After transforming FHK12 cells, 1-ml cultures (8 repeats) were grown in the presence of chloramphenicol and 0.1 M isopropyl 1-thio-␤-D-galactopyranoside for 20 h at 37°C. ␤-galactosidase activity was quantified in crude cell lysates after the addition of o-nitrophenylgalactoside and monitoring the reaction at 405 nm for 20 min at 30-s intervals at 28°C. Specific ␤-galactosidase activity was measured in Miller units (2). Hetero-association was detected using ToxR-GPA grown in the presence of an exogenous peptide. Inhibition was calculated according to Equation 1, where I is the inhibitory ability of the peptide, A peptide is the activity of ToxR-GPA in the presence of peptide, A max is the maximal activity of ToxR-GPA without peptide, and A baseline is the baseline activity of the monomer A 16 plasmid (2 where A represents ToxR-GPA transcription factor, B represents the exogenous peptide, I represents inhibition of ␤-galactosidase activity (1-activity) and K is a scaling constant. Peptide Binding to Bilayer Vesicles-The fluorescence of NBD is sensitive to its environment (28). The fluorescence quantum yield is low in solution and high in the membrane-bound state. The degree of peptide association with PE/PG (7:3) SUVs was measured by adding increasing amounts of vesicles to 0.1 M NBD-labeled peptides dissolved in dimethyl sulfoxide (Me 2 SO). The fluorescence intensity was measured as a function of the lipid/peptide molar ratio, with excitation set at 467 nm (8-nm slit) and emission set at 530 nm (8-nm slit). The fluorescence values [f] were corrected by subtracting the corresponding blank (Me 2 SO with the same amount of vesicles) [f-f 0 ], were f 0 represents fluorescence intensity in the absence of SUV. The partition coefficient was determined by nonlinear least-squares (NLLSQ) fitting using Equations 4 and 5, where f bound is the relative fluorescence of the peptide-bound fraction, f max is the fluorescence at full binding, W is the water concentration (55.56 M) and L is the lipid concentration in molar units. NLLSQ analyses and data simulations were performed with the commercial software package Origin 6.0. NBD Fluorescence Measurements-Changes in the fluorescence of NBD-labeled peptides were measured upon their binding to vesicles. NBD-labeled peptide (0.1 M) was added to 0.4 ml of phosphate-buffered saline, containing PE/PG (7:3) SUV (200 M). Emission spectra were recorded (8-nm slit), with excitation set at 467 nm (8-nm slit), and compared with the emission spectra of the NBD-labeled peptide in liposome-free PBS. The peptide/lipid molar ratio was kept at a level such that the majority of the peptides were bound to the vesicles and the contribution of the free peptide to fluorescence could be neglected.

RESULTS
A series of five peptides corresponding to the transmembrane domain of GPA were synthesized (Table I) using t-boc strategies. To overcome solubility problems and to deliver the hydrophobic transmembrane segments into the bacterial membrane, a host sequence of two consecutive lysines was used (29). These peptides were tested for their ability to inhibit the formation of a ToxR-GPA transcription activator complex in the inner membrane of E. coli as described in the model depicted in Fig. 1 (2). Briefly, the ToxR-GPA is a transcription factor that activates a lacZ reporter gene as a consequence of dimerization. The dimerization is solely through the transmembrane segment, namely the GPA transmembrane segment (13 amino acids long) (2). Hetero-association of an exogenous peptide with the ToxR-GPA transcription factor will result in a decrease in the lacZ signal.  (Table I). A more complex equilibrium is possible, although it is unlikely because the transcription activator homodimer demands a spatial proximity of the two transcription domains for its activity. For example, inhibition of the dimer through complexes such as heterotrimerization with the peptide is probably not sufficient to prevent the cytoplasmic domains from dimerization (Fig. 1).   (Fig. 3A). The free energy of binding and partition coefficient of all the peptides is within the range reported for known membrane binding peptides (25,29,30,32,33) (Table I).

Weak Membrane Binding Is Not the Reason for Peptide Inactivity-To
Membrane Insertion Analysis-Membrane insertion was investigated through the spectral analysis of the NBD moiety. The GPA analogs were labeled with the NBD fluorescent probe. NBD is highly sensitive to the polarity of the environment (26,31), thus a transfer into a hydrophobic environment will result in a blue shift concomitant with an increase in intensity. Blue shift analysis of GPA NBD-labeled fluorescent analogues revealed similar results for all peptides. Furthermore, the 10 -12 nm blue shift observed is consistent with the localization of the N termini of the peptide in the hydrophobic membrane milieu (Fig. 3B) (31, 34 -36). DISCUSSION This study sets a precedent for the detection of direct in vivo-specific hetero-oligomerization. We found that the homooligomerization motif (GXXXG) (19) is not enough for heteroassociation within bacterial membranes, although an extended motif (LIXXGXXXGXXXT) (19,21) is essential for hetero-association. In fact, [A]GPA retains wild type oligomerization capability despite multiple alanine mutations, suggesting that the extended motif alone is enough for hetero-association. Moreover, when comparing the partition coefficients of the [WT]GPA to that of [A]GPA one can see that [A]GPA has about four times weaker binding, and consequently the IC 50 is around three times higher (Table I). This indicates that the specificity of [A]GPA is very similar to that of [WT]GPA. In contrast, the [A 2 ]GPA peptide, containing only a GXXXG motif, had very low affinity toward the wild type transmembrane domain. This implies a crucial role for the extended homo-oligomerization motif. Moreover, the [G]GPA peptide, having mutations in the extended homo-oligomerization motif, completely loses its affinity to the wild-type transmembrane domain without a reduction in its membrane binding affinity. A possible explanation for the loss of hetero-assembly is that the glycine mutations interfere with the helical packing. Statistical analysis has previously shown that ␤-branched amino acids in positions I and I ϩ 4 are favorable for packing two transmembrane domains, especially when next to the GXXXG motif (19,22). Mutating two such pairs may have lowered the packing energy of the oligomer.
It is worth noting that GXXXG is the most common motif found in oligomeric transmembrane domains (19,22). By creating a second GXXXG motif in [G]GPA, situated on a different helical face than the original, we allowed further freedom. However, this motif did not contribute to oligomerization. The comparison of the extended motif on the [A]GPA peptide with the GXXXG motif on the [A 2 ]GPA peptide further confirms the importance of the extended motif for hetero-oligomerization.
Polarity is the major characteristic implicated in determining the specificity of membrane interactions (37). Additionally, hydrophobic amino acids and especially ␤-branched amino acid pairs are also characteristic of transmembrane segments. The ␤-branched amino acid pairs were previously suggested to be structurally important for helical packing (19,22). We have challenged these characteristics by introducing multiple alanine mutations in hydrophobic and ␤-branched positions, leaving the polar amino acids alone. It is logical to assume that these drastic structural changes are bound to interfere with the helical packing. Still, the specificity of [A]GPA remains similar to that of the wild type, suggesting that in this case the polar interactions play a larger role than the structure and helical packing. In contrast, [G]GPA lost its ability to hetero-associate with the wild type transmembrane domain, despite the presence of the polar amino acids, supporting the notion that hydrophobic interactions are sufficient for assembly. Other hints to this end can be found in the literature where all-hydrophobic transmembrane segments were shown to specifically interact with each other in a biological membrane (38).
Protein-protein recognition and cross-talk in the membrane remains a largely uncharted field. Nevertheless, structural data of membrane proteins is scarce (39). The method presented herein can directly detect the interaction between two different transmembrane domains provided that one of the helices can also homo-oligomerize. Thus, this method can advance structural and functional research of a broad range of such membrane proteins. In the case of the T cell receptor, for example, exogenous addition of a peptide corresponding to the transmembrane segment is known to modulate the activity of the receptor (40). Our method can clarify whether this is due to direct displacement of the corresponding helix from the receptor complex. Characterization of the mechanism by which the transmembrane domain affects receptor activity is highly important both in terms of the basic understanding of receptor mechanics and from a therapeutic perspective. Our results indicate that interfering with hetero-oligomerization within the membrane can be a useful strategy to characterize such interactions, as well as to modulate membrane protein activity.