Molecular Dissection of Rab11 Binding from Coiled-coil Formation in the Rab11-FIP2 C-terminal Domain^†

FIP2 Cloning, Expression, and Purification

Several constructs were designed to investigate the determinants of oligomer formation and Rab11 interaction in the C-terminal region of FIP2. The initial expression construct comprised the C-terminal 70 residues of FIP2 (residues 443-512) subcloned into pGEX-4T-1 (14). New protein fragments were generated by PCR amplification followed by subcloning using BamHI and EcoRI sites into pGEX-4T-1. Clones were verified by sequencing.

Glutathione S-transferase (GST) fusion proteins were expressed in E. coli BL21(DE3) Codon Plus (RIL) cells. One liter of LB medium was inoculated with 50 ml stationary culture and grown until the O.D.₆₀₀ was 0.7 before induction with 0.4 mM IPTG for 3h. Cells were harvested, and the pellet resuspended in BugBuster™ (Novagen) containing Benzonase and a cocktail of protease inhibitors, and subjected to rotary shaking for 30 min at room temperature. The supernatant after centrifugation was mixed with glutathione-Sepharose beads and subjected to rotary shaking for 2 h at 4°C. The beads were collected by centrifugation, washed three times with 20 volumes of 20 mM sodium phosphate buffer (pH 7.2) and three times with 20 volumes of 50 mM NH₄HCO₃. Thrombin digestion was carried out in 50 mM NH₄HCO₃, and the GST-free peptide was released into the supernatant. Proteins were further purified through size exclusion chromatography (Sephadex G75 superfine). The purity (>95%) of the FIP2 coiled-coil domains were verified by SDS-PAGE and mass spectrometry.

Rab11 Cloning, Expression, and Purification

A construct expressing GST- Rab11a in pGEX-5X (14) was modified using site-directed mutagenesis (Stratagene) to be cleavable by thrombin instead of Factor X because of the multiple cleavages observed with Factor X. GST-fusion proteins were expressed in E. coli BL21 (DE3) Codon Plus (RIL) cells in LB medium as described for FIP2. The cells were harvested and the pellets resuspended in a lysis buffer containing 10 mM phosphate, 100 mM NaCl, 1% Triton X-100, 2 mM MgCl₂, 3 mM DTT, 100 μM AEBSF and 0.1% (v/v) Protease Inhibitor Cocktail (Sigma). The cells were lysed using an EmulsiFlex-C5 cell homogenizer (Avestin, Inc.). The clarified lysate was mixed with glutathione-Sepharose beads and subjected to rotary shaking for 2 h. The beads were washed with 10 volumes of lysis buffer and 10 volumes of lysis buffer without Triton X-100. The protein was then subjected to a nucleotide exchange protocol to replace GDP with GMP-PNP as previously described (19). The beads were washed with a buffer containing 20 mM HEPES, 150 mM NaCl, 2 mM Mg⁺⁺, pH 8.0, thrombin was added, and the mixture nutated at room temperature for 2 h. The protein was eluted with 5 column volumes of a buffer containing 20 mM HEPES, 150 mM NaCl, 2 mM MgCl₂, 3 mM DTT, 100 μM AEBSF, pH 7.4, and passed through a Benzamidine Sepharose column (Amersham). Protein was further purified by gel filtration on Sephadex G-50 and assayed for concentration using absorption at 280 nm (20).

Circular Dichroism (CD) Spectroscopy and Denaturation Measurements

Far-UV CD spectra were obtained by averaging 2 scans with a step size of 0.5 nm on a JASCO model 810 spectropolarimeter. All measurements were performed in cuvettes with path lengths of 0.1 cm. Spectra were baseline-corrected by buffer subtraction and data were smoothed before curve fitting. Peptide concentrations were determined by UV absorbance at 280 nm. For thermal denaturation, ellipticity at 222 nm was measured as a function of temperature. The data were fit using a two-state model (folded dimer and unfolded monomer) as described (21-23):

where Θ_N and Θ_D represent the pre-transition (N) and post-transition (D) ellipticity values and f_D is the fraction of dissociation of the dimer at each observed ellipticity (Θ). The equilibrium constant (K) was determined at each temperature using the fitted fd values:

where C_t represents the total concentration of the monomer.

Using the relationship ΔG = -RT ln K where ΔG is the Gibb-Helmholtz free energy of unfolding, the T_m was determined for the midpoint of the transition curve where f_D = 0.5 and

where ΔC_p is the apparent heat capacity of unfolding at constant pressure and ΔH(T_m) is the apparent enthalpy of unfolding at T_m.

Limited Proteolysis

The purified protein was subjected to limited protease digestion at 25 °C in 20 mM phosphate buffer containing 100 mM NaCl using trypsin at molar ratio of 1:2000 to protein. Proteolysis was stopped by boiling the sample in SDS-containing buffer prior to analysis on SDS-PAGE.

N-terminal Microsequencing of Proteolyzed Fragments

The fragments generated from proteolysis were separated by Tris-tricine SDS-PAGE and blotted onto a polyvinylidene difluoride (PVDF) membrane with electroblotting buffer (10 mM N-Cyclohexyl-3-aminopropanesulfonic acid, pH 11, 10% (v/v) methanol). Protein fragments were visualized by Coomassie brilliant blue R250 staining, excised and subjected to N-terminal sequencing.

Chemical Cross-linking Experiments

BS³ is an N-hydroxysuccinimide ester that reacts in pH 7-9 buffers with primary amino groups to form stable amide bonds. Primary amines are available from the side chains of lysine residues and the N-terminus of each polypeptide. FIP2 (40 μM) was incubated with 4 to 60 mM freshly prepared BS³ (Bis(sulfosuccinimidyl)suberate) for 30 min at room temperature. The cross-linking reaction was stopped by incubating with 0.1 M Tris, pH 7.5, for 10 min. The samples were electrophoresed, and the gels were stained with Coomassie Blue to visualize cross-linked products.

ITC Experiments

Samples were dialyzed against 20 mM HEPES, 150 mM NaCl, 2 mM MgCl₂, 2 mM TCEP overnight (24). ITC experiments were conducted at 25°C on a Microcal VP-ITC calorimeter by injecting aliquots of 100 μM FIP2 peptide into 10 μM Rab11-GMP-PNP. Data were analyzed using Origin to integrate peak volumes and calculate baselines, and baselines were then visually inspected and corrected.

Analytical Ultracentrifugation

Sedimentation equilibrium experiments were performed on a BeckmanCoulter XL-I with interference optics. Ultracentrifugation samples were prepared from a concentrated protein stock solution and dialyzed against 20 mM sodium phosphate (pH 7.2) overnight. Protein was tested at four concentrations (1.4 mg/ml, 0.7 mg/ml, 0.3 mg/ml, and 0.175 mg/ml) with no apparent aggregation and at two different speeds (50,000 rpm and 35,000 rpm). The collected data were edited and analyzed using the program WINNONLIN v.1.06 (25). A global fit for each of the eight curves was performed to obtain a single species average molecular weight, M, from the following relationship:

where σ was the reduced molecular weight and obtained through curve fitting, $\stackrel{‒}{v}$ is the partial specific volume; ρ is the solvent density; ω is the rotor speed (radians/s), R is the gas constant; and T is the temperature (K).

Equilibrium solute distributions were obtained using the equation:

where C_r is the solute concentration at radial position r and C₀ is the concentration at a reference position r₀; H is the constant $(1-\stackrel{‒}{v}\rho ){\omega }_2∕2\mathrm{RT}$ n is the multimer ratio, Ka is the association constant, and E is the optical baseline offset (26).

Determination of the Boundaries of the FIP2 Coiled-coil Domain by Limited Proteolysis

A peptide comprising the last 70 residues of FIP2, RH70 (Figure 1), was incubated with trypsin at 25°C in 10 mM sodium phosphate, pH 7.0. After incubation, a smaller band appeared on a 15% Tris-tricine gel that was found to be 5976 Da by mass spectrometry. N-terminal sequencing showed the N-terminus started from Ser450 indicating the cleavage site after Arg449. The mass was consistent with a second C-terminal trypsin cleavage site after Arg497 (Figure 2).

A new construct containing the trypsin resistant core (RH50, residues 450-497) was cloned, expressed, and purified (Figure 1). N-terminal amino acid sequencing of a degradation product that slowly appeared with time had the same N-terminus as the trypsin resistant core, indicating degradation from the C-terminus. The observed mass (5039 Da) indicated cleavage after Met489, consistent with cleavage by a contaminating enzyme with chymotrypsin-like activity.

Another construct comprising this Rab11-FIP homology region coiled-coil (RHCC, residues 450-489) was cloned, expressed, and purified (Figure 1). This construct represents the boundaries of the core FIP2 coiled-coil domain with respect to a chymotrypsin-like activity in E. coli and to trypsin. This resulting peptide was resistant to limited proteolysis with Glu-C (data not shown).

Examination of the sequence of RHCC shows a hydrophobic residue every seventh position. Assigning these hydrophobic residues at position ‘a’ of a helical wheel, results in residues at position‘d’ that are mostly hydrophobic (Figure 2). An alternative registry that places the initial group of hydrophobics at position d, results in negatively-charged Glu and Asp residues at position ‘a’ in the helical wheel, which is not seen in coiled-coil structures (27).

Oligomerization State of the FIP2 Coiled-coil Domain by Cross-linking Experiments

FIP2 and most of its family members have been reported to self-interact both in vivo and in vitro (11). While a gel filtration experiment on RCP indicates a dimer (17), no one has determined the oligomerization state of FIP2. A cross-linking experiment was performed using the reagent BS³ [Bis(sulfosuccinimidyl)suberate] (Figure 3). Non-cross-linked peptide appeared at ∼4 kDa, whereas only one cross-linked product with an apparent molecular weight of ∼8 kDa was observed, which clearly shows that the FIP2 coiled-coil domain primarily exists as dimer at a concentration of about 0.2 mg/ml.

Sedimentation Equilibrium Studies

FIP2 was also assessed by sedimentation equilibrium to assess its oligomeric state over a range of concentrations. The equilibrium solute distributions were analyzed as indicated under “Materials and Methods” and residuals evidenced the high quality of the fit (Figure 4). At all velocities and concentrations, an estimated average molecular mass of 9957.5 Da was obtained using an ideal single species model. The mass is consistent with a dimeric model for FIP2 coiled-coil domain (calculated mass 9986 Da). We saw no monomer in the sedimentation experiments, consistent with a high association constant for RHCC dimer formation. Using an oligomeric model, data fitting indicated a mass of 9957 Da data (the dimer) in equilibrium with a second component having a multimer ratio of 3 corresponding to a mass of 29870 Da (a hexamer). Because the dimer mass of the oligomeric model agrees with that of the single-species model, the percentage of hexamer is assumed to be small (less than 5%). The association constant was not consistent among the eight data sets (lnKa varied from -5.05 to 1.49), indicating that the system was not in chemical equilibrium on the time scale of the experiment. Both chemical cross-linking experiments and analytical centrifugation experiments reflect the existence of a homodimeric FIP2 coiled-coil domain.

Thermal Denaturation Studies

Preliminary NMR analysis (data not shown) indicated that the RHCC peptide monomers were arranged in a parallel rather than antiparallel manner. A new peptide was designed, RHcys, which had cysteine added at the C-terminus (28). If the dimer were parallel, the introduced disulfide bridge would be expected to be stabilizing. Purified RHcys peptide was dissolved in ammonium bicarbonate buffer (50 mM, pH 8.5) and stirred in an open beaker overnight to allow oxidation of cysteines to cystine. The peptide was dimeric when analyzed on a Coomassie-stained SDS-PAGE gel. In the presence of the BS³ cross-linker, the RHCys peptide showed quantitatively the same pattern as the RHCC core peptide, except the extent of monomer was slightly greater. RHCC and RHCys (and RH50) showed bands consistent with about 90% dimer and some small amount (ca. 5%) of higher order aggregates with molecular weights of approximately 12 and 16 kDa, consistent with tetramer and hexamer formation (Figure 1S). In addition, analysis of the RHCys peptide by gel filtration on a FPLC system showed little evidence of tetramer formation with predominant peaks corresponding to dimer and some monomer. In thermal denaturation experiments carried out on peptides at a monomer concentration of 20 μM, cooperative transitions were observed for both RHCC and RHcys (Figure 5). The unfolding data were fitted to a two-state model with simultaneous dissociation and unfolding processes. The fitted value of T_m was 71 °C for RHcys and 23 °C higher than RHCC (T_m= 48 °C) consistent with observations made for the disulfide cross-linking of the coiled-coil domain of GCN4 (28). The higher thermal stability for RHcys evidenced that the disulfide-bridge stabilizes the coiled-coil structure and suggests that the RHCC dimer is parallel.

Assessment of Structure by Circular Dichroism Spectroscopy

The far-ultraviolet CD spectra of RH50 and C-terminal truncation mutant RHCC showed two ellipticity minima around 208 nm and 222 nm consistent with an α-helical structure (Figure 7). Assuming that the ellipticity at 222 nm is -33400 deg cm² dmol^-1 for 100% α-helix (29), the calculated helical content was 63% for RH50 and 55% for RHCC. However, the difference is not likely to be significant given the probable errors in measuring protein concentrations and small differences in dynamics. The degree of coiled-coil formation was estimated from the Θ₂₂₂/Θ₂₀₈ ratio. The ratio is 0.83±0.03 for non-interacting α-helices and 1.03±0.03 for two-stranded coiled-coils (30). The Θ₂₂₂/Θ₂₀₈ ratio changes because upon formation of the two-stranded coiled-coil, the parallel-polarized amide π-π* electronic transition at 208 nm is reduced, while the n-π* transition at 222 nm remains the same (31). The Θ₂₂₂/Θ₂₀₈ ratio was 1.14 for RHCC and 0.99 for RH50. A reason that the ratio for RHCC was higher than 1.03 may be due to a sizeable contribution by an aromatic side-chain such as one of the tyrosines to the far-UV CD spectrum, an unexpected change in the π-π* or n-π* transitions of the peptide bond, or differences in dynamics (32, 33).

To define the boundaries of the coiled-coil more tightly, two additional constructs were studied. One was truncated by 8 residues at the N-terminus of RH50 (RHNQ) and the other 5 residues at the C-terminus of RHCC (RHCN). In the far-UV CD spectrum of RHNQ (Figure 7), the intensity of the molecular ellipticity at both 222 nm and 208 nm was dramatically reduced, indicating a poorly structured peptide. The non-dimeric nature of RHNQ is supported by BS³ cross-linking experiments in which it did not cross-link (Figure 1S). This suggests the first eight residues in RH50 were crucial for α-helix and coiled-coil dimer formation. RHCN was also not structured in CD measurements. The structural conclusions from the CD data are supported by NMR data in which the one-dimensional NMR spectrum of RH50 showed good resonance dispersion, which indicates a structured protein, while that of RHNQ showed poor resonance dispersion, corresponding to random coil (Figure 2S). Therefore, RHCC is the minimal structural core of the coiled-coil domain thus far defined.

Isothermal Titration Calorimetry Studies

Having defined a structural core of the FIP2 coiled-coil domain, we then investigated the relationship between structure and the ability to bind Rab11. To test interactions with Rab11, various constructs were subjected to ITC experiments (Figure 6). The binding constant (K_d), stoichiometry (n), enthalpy (ΔH), and entropy (ΔS) were obtained by curve-fitting the binding isotherms using a one-site model and calculating FIP2 concentrations for the dimeric form. The data showed that RH50 and RHNQ have similar binding affinity with K_d values of about 840 nM. These values are 20-fold lower than observed with larger constructs including ∼100 residues N-terminal to RH50, suggesting interactions between Rab11 and FIP2 outside the RBD (17). The binding stoichiometry values were 0.37±0.02 for RH50 and 0.36±0.02 for RHNQ. Assuming that the active Rab11 concentration may be over-estimated given that the purity of GMP-PNP used was 92%, these values are consistent with previous studies indicating that the molecular complex of FIP2 and Rab11 is a heterotetramer comprising a dimer of FIP2 and a dimer of Rab11 (17). Although appearing as a dimer in complex with FIP2, the oligomeric state of Rab11 alone is less well defined. While it was monomeric in solution by gel filtration experiments (17), it has been observed to be dimeric crystallographically (34) and diffusion measurements on Rab11 at 13 mg/ml using NMR spectroscopy show that it is a dimer (G.D. Henry & J.D. Baleja, unpublished results). The discrepancy may reflect an equilibrium between dimer and monomer that favors dimer formation at the high concentration found in crystals and NMR but a favored monomer at lower concentrations.

When the RHCC construct lacking C-terminal amino acids relative to RH50 was used as a titrant, there was no detectable exothermic heat response indicating no detectable interaction. Although the CD spectra indicated that unlike RH50, RHNQ was not well structured, it had the same binding affinity as RH50. On the other hand, RHCC was as well structured as RH50, but did not bind Rab11 with appreciable affinity. Therefore, we conclude that a preformed coiled-coil is not required for Rab11 binding.

Association of Rab11 with the FIP2 Coiled-coil Domain Induces a Conformational Change

CD spectra of RH50 alone, RHNQ alone, Rab11 alone, and the mixtures each peptide with Rab11 and the calculated sum of the relevant intensities are shown in Figure 8. The intensity of the mixture of Rab11 and RH50 was 11% higher than the calculated sum of the separated components. For the mixture of Rab11 and RHNQ, the intensity was 20% higher. The higher intensities at 208 and 222 nm indicate additional α-helical structure in the Rab11-RHNQ complex is especially apparent and is likely to mostly derive from coiled-coil formation within FIP2 rather than a conformational change in Rab11. This suggests that the coiled-coil conformation may stabilize the complex, even though it is not required for direct interaction.