Disruption of the C-terminal helix by single amino acid deletion is directly responsible for impaired cholesterol efflux ability of apolipoprotein A-I Nichinan.

Apolipoprotein A-I (apoA-I) Nichinan, a naturally occurring variant with ΔE235 in the C terminus, is associated with low plasma HDL levels. Here, we investigated the tertiary structure, lipid-binding properties, and ability to induce cellular cholesterol efflux of apoA-I Nichinan and its C-terminal peptide. Thermal and chemical denaturation experiments demonstrated that the ΔE235 mutation decreased the protein stability compared with wild type (WT). ApoA-I Nichinan exhibited capabilities to bind to or solubilize lipid vesicles that are intermediate to that of WT and a L230P/L233P/Y236P variant in which the C-terminal α-helix folding is completely disrupted and forms relatively larger and unstable discoidal complexes, indicating that perturbation of the C-terminal α-helical structure by the ΔE235 mutation leads to reduced lipid binding. Supporting this, apoA-I 209-241/ΔE235 peptide showed significantly decreased ability to form α-helix both in the lipid-free and lipid-bound states, and reduced efficiency to solubilize vesicles. In addition, both apoA-I Nichinan and its C-terminal peptide exhibited reduced activity in ABCA1-mediated cellular cholesterol efflux. Thus, the disruption of the ability of the C-terminal region to form α-helix caused by the E235 deletion appears to be the important determinant of impaired lipid binding and cholesterol efflux ability and, consequently, the low plasma HDL levels of apoA-I Nichinan probands.

the baseline for an appropriate blank sample. The ␣ -helix content was calculated from the molar ellipticity at 222 nm ([ ] 222 ) using the equation: % ␣ -helix = [(-[ ] 222 + 3000)/(36000 + 3000)] × 100 ( 33 ). Thermal unfolding was monitored from the change in [ ] 222 over the temperature range of 20-90°C, as described ( 34 ). Although there was some variability in the initial and fi nal ␣ -helix contents, the apoA-I variants exhibited reversible thermal unfolding. The van't Hoff enthalpy, ⌬ H v , was calculated from the slope of the line fi tted by linear regression to the equation, ln K D = Ϫ ( ⌬ H v / R ) 1/ T + constant, where K D is the equilibrium constant describing the unfolding of apoA-I at each temperature, R is the gas constant, and T is temperature. The kinetics of thermal unfolding of apoA-I in discoidal complexes with 1-palmitoyl-2-oleoyl PC (POPC) prepared using the cholate dispersion method ( 35 ) was monitored by temperature-jump analysis from 25 to 80-90°C ( 36 ).

Chemical denaturation studies
For monitoring chemical denaturation, lipid-free proteins at a concentration of 25-50 µg/ml were incubated overnight at 4°C with GdnHCl or urea at various concentrations. When complexed with POPC, apoA-I was completely denatured by incubating for 72 h ( 37 ). K D at a given denaturant concentration was calculated from the change in either [ ] 222 or wavelength of maximum fl uorescence (WMF) of intrinsic Trp residues ( 15 ). The Trp emission fl uorescence spectrum was recorded from 300 to 420 nm using an excitation wavelength of 295 nm with a Hitachi F-7000 fl uorescence spectrophotometer in Tris buffer (10 mM Tris, 150 mM NaCl, 1 mM EDTA, 0.02% NaN 3 , pH 7.4). The Gibbs free energy of denaturation in the absence of denaturant, ⌬ G D °, the midpoint of denaturation, D 1/2 , and m value, which refl ects the cooperativity of denaturation in the transition region, were determined by the linear equation, ( 19,33 ).

ANS fl uorescence measurements
8-Anilino-1-naphthalenesulfonic acid (ANS) fl uorescence measurements were carried out with a Hitachi F-4500 fl uorescence spectrophotometer. The extent of ANS binding to hydrophobic sites on the apoA-I variants was determined by measuring ANS fl uorescence spectra recorded from 400 to 600 nm at an excitation wavelength of 395 nm, in the absence or presence of 50 g/ml protein and an excess of ANS (250 M) ( 15 ).

Isothermal titration calorimetry (ITC) measurements
Heats of apoA-I binding to SUV were measured with a Micro-Cal MCS isothermal titration calorimeter at 25°C ( 32 ). To ensure that the injected protein bound completely to the SUV surface, the PC to protein molar ratio was kept over 10,000. Heat of dilution determined by injecting apoA-I solution into buffer was subtracted from the heat for the corresponding apoA-I-SUV binding experiments. The decay rate constants for the heats of binding were obtained from fi tting the titration curves to a one-phase exponential decay model ( 15 ).

DMPC clearance assay
The kinetics of solubilization of dimyristoyl PC (DMPC) vesicles by the apoA-I variants or peptides were measured by monitoring the time-dependent decrease in turbidity at 24.6°C. For the apoA-I proteins, DMPC large unilamellar vesicles (LUVs) extruded through a 200-nm fi lter at a concentration of 0.25 mg/ml were mixed with apoA-I samples (0.02-0.2 mg/ml), and incubated for 15 min to monitor the light scattering intensity at 325 nm with a Shimadzu UV-2450 spectrophotometer ( 20,38 ). For the apoA-I peptides, much lower concentrations of DMPC multi-plasma HDL levels and increase the incidence of cardiovascular disease ( 4,5 ). The large majority of mutations associated with low plasma HDL levels are clustered in the N-terminal helix bundle domain, whereas very few mutations are in the C-terminal domain ( 5 ). ApoA-I Nichinan, which contains the deletion of residue E235 in the C-terminal helical region, is known to be associated with low plasma HDL cholesterol and apoA-I levels ( 21,22 ). Interestingly, this ⌬ E235 mutant is associated with normal plasma LCAT activity but a decrease in the ratio of HDL 2 to HDL 3 particles. Because a number of studies have demonstrated the importance of the C-terminal region of apoA-I in the ABCA1-mediated lipid effl ux from cells and HDL biogenesis (23)(24)(25)(26)(27)(28)(29), it is possible that impaired lipidbinding ability caused by the deletion of E235 in apoA-I Nichinan could underlie the decreased plasma HDL concentrations seen in carriers of the mutation. However, the molecular basis for these effects, especially the structurefunction relationship of apoA-I Nichinan, has not been established yet.
In the present study, we investigated the effects of the E235 deletion in apoA-I and its C-terminal peptide on the structure, lipid-binding properties, and ability to induce ABCA1-mediated cellular cholesterol effl ux. The results suggest that the disruption of the helix-forming ability of the C-terminal region caused by the E235 deletion is the critical determinant of impaired cholesterol effl ux ability and, consequently, low plasma HDL levels in apoA-I Nichinan probands.

Preparation of apoA-I proteins and peptides
The mutation in human apoA-I to delete residue E235 was made using the QuikChange site-directed mutagenesis kit (Stratagene). Human wild type (WT) apoA-I and engineered mutants were expressed and purifi ed as described ( 15,16 ). The apoA-I preparations were at least 95% pure as assessed by SDS-PAGE. The C-terminal apoA-I 209-241 and 209-241/ ⌬ E235 peptides were synthesized using Fmoc chemistry as described ( 30,31 ). The N and C termini were capped with an acetyl group and an amide group, respectively. Peptide purity was verifi ed by analytical HPLC (>97%) and mass spectrometry. In all experiments, apoA-I variants and peptides were freshly dialyzed from 6 M guanidine hydrochloride (GdnHCl) solution into the appropriate buffer before use.

CD spectroscopy
Near-and far-ultraviolet (UV) circular dichroism (CD) spectra were recorded with a Jasco J-810 or an Aviv 62DS spectropolarimeter. After dialysis from 6 M GdnHCl solution, the apoA-I protein or peptide solutions of 25-50 g/ml in 10 mM sodium phosphate buffer (pH 7.4) were subjected to near-UV measurements by scanning from 185 to 260 nm in a 1-2 mm cuvette, and solutions of 0.3 mg/ml apoA-I concentration were used for near-UV measurements (270-320 nm) in a 1 cm cuvette. For the mixture with small unilamellar vesicles (SUVs), the apoA-I protein or peptide was incubated for 1 h prior to the measurement with egg phosphatidylcholine (PC) SUVs prepared by sonication as described before ( 32 ). The results were corrected by subtracting RESULTS Secondary structure, thermal unfolding, and near-UV CD spectra of apoA-I C-terminal mutants The secondary structure and thermal unfolding of apoA-I Nichinan were analyzed by far-UV CD spectroscopy, in comparison with the apoA-I L230P/L233P/Y236P mutant in which the C-terminal ␣ -helix is completely disrupted ( 19,25 ). The ␣ -helix contents of the lipid-free apoA-I Nichinan and L230P/L233P/Y236P variants were no different from WT ( Table 1 ), indicating that the deletion of E235 and the proline insertions either 1 ) do not perturb the local ␣ -helical structure signifi cantly or 2 ) the C-terminal region is not helical in the lipid-free state. Figure 1A shows the thermal unfolding curves for WT and the Nichinan variant. As summarized in Table 1 , both apoA-I Nichinan and the L230P/L233P/Y236P variant displayed a lower midpoint ( T m ) and higher van't Hoff enthalpy ( ⌬ H v ) compared with WT, indicating that the E235 deletion and the proline insertions decrease the protein stability but increase the unfolding cooperativity of apoA-I ( 19 ). In contrast, near-UV CD spectra of WT and the Nichinan variant were very similar, exhibiting the negative trough with a minimum at 292 nm (Fig. 1B). This suggests that the orientational ordering of the Trp residues located in the N-terminal domain is not affected by the Nichinan mutation ( 16,43 ).

Chemical denaturation of apoA-I C-terminal mutants
Next, we monitored chemical denaturation profi les of apoA-I variants by the molar ellipticity at 222 nm as well as by the change in WMF of Trp fl uorescence spectra. Molar ellipticity measures the unfolding of all ␣ -helices in the protein molecule whereas Trp fl uorescence measures the stability of the N-terminal helix bundle because human apoA-I contains four Trp residues, all of which are located in the N-terminal domain. In Fig. 2A , the denaturation curve of apoA-I Nichinan monitored by the molar ellipticity was shifted to lower concentrations of urea compared with WT. As a result, the Nichinan variant exhibited signifi cant decreases in the Gibbs free energy and midpoint of denaturation compared with WT apoA-I, whereas the L230P/L233P/Y236P mutation had little effect on these parameters ( Table 1 ). Figure 2B compares the free energies of denaturation with GdnHCl and urea for apoA-I variants as monitored by Trp fl uorescence. It is thought lamellar vesicles (MLVs) and peptides (typically 34 g/ml DMPC and 2-20 g/ml peptides) were used to avoid the aggregation of DMPC/peptide complexes during experiments. The decrease in turbidity was monitored by the right-angle light scattering intensity using a Hitachi F-4500 spectrophotometer with both excitation and emission wavelengths set at 600 nm ( 31,39 ).

Monolayer exclusion pressure measurements
The relative affi nities of apoA-I peptides for the lipid-water interface were determined with a surface balance technique, as described previously ( 40,41 ). Linear extrapolation of the initial surface pressure of egg PC monolayer in the absence of peptides ( i ) versus the change in surface pressure in the presence of peptides ( ⌬ i ) curve to the point at which ⌬ i = 0 gave the monolayer exclusion pressure ( e ). The e value indicates the surface pressure at which the peptides are no longer able to penetrate into the egg PC monolayer.

Cell culture and cholesterol effl ux
J774 murine macrophages were grown and maintained in RPMI 1640 supplemented with 10% FBS and 0.5% gentamycin. For effl ux experiments, these cells were seeded in 12-well plates, grown to 80-90% confl uence, and then labeled by incubating the cells for 24 h in RPMI medium supplemented with 1% FBS, 2 µg/ml CP-113,818 ACAT inhibitor, and 3 µCi/ml [ 3 H]cholesterol ( 25 ). After labeling, the cells were washed with MEM-Hepes and incubated with RPMI medium containing 0.2% BSA, 2 µg/ ml CP-113,818 ACAT inhibitor, and 0.3 mM 8-(4-chlorophenylthio)-cAMP for 12 h, to upregulate the expression of ABCA1. Cells were then washed with MEM-Hepes and incubated with or without apoA-I proteins or peptides at the indicated concentrations for 4 h. To determine the free unesterifi ed cholesterol effl ux, aliquots were removed from the incubation medium at specifi c time points, fi ltered, and radioactivity was determined by liquid scintillation counting. The percent cholesterol effl ux was calculated after subtracting the background cholesterol effl ux (without apoA-I) as follows: (counts/min in medium at 4 h/cpm in cells at t = 0) × 100.
For effl ux experiments with baby hamster kidney (BHK) cells, BHK cells transfected with human ABCA1 (a generous gift from Dr. John Oram) were grown and maintained in DMEM containing 10% FBS and 0.5% gentamycin ( 42 ). These cells were seeded in 12-well plates and then labeled by incubating the cells for 24 h in DMEM medium supplemented with 2.5% FBS, 2 µg/ml CP-113,818 ACAT inhibitor, and 3 µCi/ml [ 3 H]cholesterol. ABCA1 was induced by incubating the labeled cells for 18 h in DMEM containing 0.2% BSA and 10 nM mifepristone. Cells were then washed with MEM-Hepes and incubated with or without apoA-I proteins or peptides under the indicated conditions of concentration for 4 h and the percentage cholesterol effl ux was calculated as mentioned above. of the C-terminal domain and partially disrupts the hydrophobic sites.

Lipid interactions of apoA-I C-terminal mutants
To assess the lipid-binding properties of the apoA-I variants, we measured the heats of binding of apoA-I to egg PC SUVs using ITC ( 16,32 ). As listed in Table 2 , WT apoA-I exhibited a large exothermic heat of about Ϫ 93 kcal/mol upon binding to SUVs; this exothermic enthalpy of binding is the major component of the favorable free energy of apoA-I binding to SUVs, and the increase in ␣ -helix content associated with such binding is largely responsible for this enthalpy ( 32,47 ). Thus, the large reduction in the binding enthalpy and small increase in ␣ -helical amino acids upon binding to SUVs for the L230P/L233P/Y236P variant compared with WT appear to result from impaired ability to form ␣ -helix caused by the proline insertions ( 19 ). The intermediate effects of the Nichinan variant on the binding enthalpy and increase in ␣ -helical residues upon binding to SUVs seen in Table 2 suggest that the E235 deletion partially disrupts the ability of the C-terminal region to form ␣ -helical structure, resulting in impaired lipid-binding capability for apoA-I Nichinan.
To further characterize the interactions of the apoA-I variants with lipids, we used a DMPC clearance assay for comparing their abilities to solubilize DMPC vesicles ( 16,20,38 ). The typical timecourses of clearance in Fig. 4A that the electrostatic interactions between amino acids in proteins are shielded by GdnHCl whereas uncharged urea has no effect on the electrostatic interactions ( 44 ). With both denaturants, signifi cant decreases in the free energy of denaturation for the Nichinan variant were observed compared with WT, indicating that the E235 deletion destabilizes the N-terminal helix bundle of apoA-I via both electrostatic and nonelectrostatic interactions. In addition, large differences in ⌬ G D ° value monitored by molar ellipticity and Trp fl uorescence such as 3.3 versus 5.3 kcal/ mol for the Nichinan variant ( Fig. 2B and Table 1 ) appear to come from the fact that molar ellipticity measures only unfolding of ␣ -helical structure whereas Trp fl uorescence measures unfolding of tertiary structure as well so that disruption of helix-helix contacts is included in the free energy change ( 20 ).

ANS binding
To compare the exposure of hydrophobic regions in the apoA-I C-terminal variants, ANS binding experiments were performed. Compared with free ANS, a great increase in ANS fl uorescence spectrum was observed for WT apoA-I ( Fig. 3 ), indicating the presence of ANS accessible hydrophobic surface ( 15,45,46 ). Because the deletion or disruption of the C-terminal helical region such as the L230P/L233P/Y236P mutation greatly decrease the hydrophobic sites created by the C-terminal domain ( 15,19 ), the fi nding that the ⌬ E235 mutation displayed the intermediate effect on ANS fl uorescence between WT and L230P/L233P/Y236P apoA-I (Fig. 3) suggests that the deletion of E235 alters the conformation  In addition, comparison of the timecourse of the protein unfolding induced by a rapid increase in temperature (Fig. 5B) showed that the unfolding of apoA-I Nichinan is faster than that of WT apoA-I in POPC discs. These results indicate that apoA-I Nichinan forms less stable and larger discoidal complexes with lipids compared with WT.

Effects of E235 deletion on structure and lipid interaction of apoA-I C-terminal peptides
The effects of the ⌬ E235 mutation on the C-terminal ␣ -helix properties of apoA-I were also examined using synthetic apoA-I C-terminal peptides. Like WT apoA-I, the 33-mer peptide, apoA-I 209-241, can interact with lipids ( 39,41 ) and promote ABCA1-mediated cellular cholesterol effl ux ( 25,49 ). Table 3 summarizes the monolayer exclusion pressure, e , and ␣ -helix contents of apoA-I 209-241 and 209-241/ ⌬ E235 peptides. Although both peptides displayed similar lipid affi nity as assessed by e , the deletion of E235 greatly decreased the ␣ -helix content of apoA-I 209-241 peptide both in the absence and presence of SUVs, consistent with the previous result for a shorter C-terminal peptide, apoA-I 220-241 ( 31 ). This directly demonstrates that the E235 deletion reduces the ability of the C-terminal region of apoA-I to form ␣ -helix upon lipid binding. As shown in Fig. 6A and 6B, such a weak propensity of apoA-I 209-241/ ⌬ E235 to form ␣ -helix resulted in a reduced ability to solubilize DMPC vesicles compared with apoA-I 209-241 ( 50 ); this effect is similar to that seen for the apoA-I C-terminal variants (Fig. 4). and the dependence of clearance effi ciency on protein concentrations in Fig. 4B demonstrate that the solubilizing effi ciency of the apoA-I variants decreases in the order, WT > Nichinan > L230P/L233P/Y236P, which is similar to the results for the SUV binding ( Table 2 ). Because solubilization of DMPC vesicles by apolipoprotein to form discoidal complexes involves a structural reorganization of the protein molecule ( 48 ), the self-association state of protein can affect solubilization kinetics ( 38 ). Cross-linking experiments demonstrated that apoA-I Nichinan at relatively high protein concentration (1 mg/ml) tends to form more oligomers compared with WT apoA-I (data not shown). However, the decreased DMPC clearance effi ciency of the Nichinan mutant was seen even at a much lower concentration range of 25-200 g/ml (Fig. 4B), indicating that the reduced solubilizing effi ciency of apoA-I Nichinan is not mainly due to the self-association state of this protein.
We also compared the stabilities of WT and the Nichinan mutant in apoA-I-POPC discoidal complexes. Electron microscopy and nondenaturating polyacrylamide gradient gel analyses demonstrated that WT apoA-I-POPC discs have a major diameter of 11 nm whereas, under the same conditions, the Nichinan mutant forms much larger discs with a major diameter of 15 nm (data not shown). In Fig. 5A , the denaturation curve of apoA-I Nichinan in a POPC discoidal complex monitored by the molar ellipticity was shifted to lower concentrations of GdnHCl compared with WT, and the resultant Gibbs free energy and midpoint of denaturation for the Nichinan variant were signifi cantly decreased compared with WT apoA-I (2.2 ± 0.2 vs. 3.0 ± 0.1 kcal/mol  a Increase in ␣ -helical residues was estimated from the difference in ␣ -helix contents of apoA-I in the lipid-free state ( Table 1 ) and bound to SUV. the ratio V max / K m for the Nichinan variant was about half of that of WT apoA-I (2.5 vs. 4.7 for J774 cells and 1.6 vs. 3.6 for BHK cells, respectively). This indicates that the deletion of E235 greatly reduces the effi ciency of cholesterol effl ux to apoA-I in both types of cells ( 21,22 ). Similar to previous results ( 25,49 ), apoA-I 209-241 peptide retained cholesterol effl ux ability, but a linear relationship between cholesterol effl ux and peptide concentration in J774 cells ( Fig. 8 ) indicates that ABCA1mediated cholesterol effl ux to this peptide is a low affi nity process. Strikingly, apoA-I 209-241/ ⌬ E235 peptide at up to 20 g/ml failed to induce a cholesterol effl ux, indicating that the deletion of E235 abolishes the ability of the C-terminal helical region of apoA-I to induce ABCA1mediated cholesterol effl ux. This fi nding is consistent with previous studies with amphipathic ␣ -helical peptides demonstrating that lipid effl ux via ABCA1 is sensitive to the lipid-binding affi nity of the helix ( 41 ) as well as the distribution of acidic residues ( 49 ). In addition, neither the apoA-I 209-241 nor the 209-241/ ⌬ E235 peptide gave any cholesterol effl ux from BHK cells in the concentration range studied (data not shown).

DISCUSSION
It is known that human apoA-I mutations can reduce plasma HDL levels and increase the incidence of cardiovascular disease ( 4,5 ). ApoA-I Nichinan that contains the

Cholesterol effl ux ability of apoA-I Nichinan variant and peptide
To assess the ability of the apoA-I Nichinan variant and its C-terminal peptide to promote ABCA1-mediated cellular cholesterol effl ux, the concentration-dependence of cholesterol effl ux from ABCA1-expressing J774 and BHK cells to apoA-I was determined. Figure 7 shows hyperbolic relationships between cholesterol effl ux and the apoA-I concentrations for WT apoA-I and the Nichinan variant in both types of cells. The greater difference between the cholesterol effl ux to WT apoA-I and apoA-I Nichinan observed with BHK cells (Fig. 7B) compared with J774 cells (Fig. 7A) may be due to the fact that J774 macrophages express mouse ABCA1 whereas BHK cells are transfected with human ABCA1. Apparent V max and K m values derived by fi tting these hyperbolic curves to the Michaelis-Menten equation are listed in Table 4 . In both types of cells, the apparent K m value for apoA-I Nichinan was much higher than that for WT, and the catalytic effi ciency expressed as    ( Table 1 ), and similar tertiary packing of Trp residues in the N-terminal helix bundle (Fig. 1B), in contrast to the situation with a mutation in the N-terminal domain such as apoA-I Milano (R173C), which markedly destabilizes the protein structure ( 51 ). However, when the changes in molar ellipiticity and Trp fl uorescence caused by Gdn-HCl or urea denaturation were monitored, signifi cant reductions in the conformational stability, ⌬ G D °, and midpoint of denaturation for apoA-I Nichinan were observed compared with WT apoA-I ( Table 1 and Fig. 2). This indicates that the deletion of E235 not only affects the conformational stability of the C-terminal domain but also destabilizes the N-terminal helix bundle domain.
We have demonstrated recently that intramolecular interaction between the N-and C-terminal tertiary structure domains in apoA-I modulates the conformational stability of the N-terminal helix bundle, and the C-terminal helical region is responsible for this stabilizing interaction ( 20 ). Because the deletion of E235 signifi cantly destabilizes the apoA-I structure, it is possible that the absence of E235 that removes a negative charge at this position and alters the charge distribution of the C-terminal region may result in alterations in stabilizing electrostatic interactions between the N-and C-terminal domains in apoA-I. The fi nding that the E235A mutation did not destabilize the apoA-I structure (unpublished observations) indicates that the E235 residue itself is not involved in the interaction with the N-terminal domain. Rather, the alteration in the charge distribution of the C-terminal region caused by the E235 deletion may disrupt optimal electrostatic interaction between the N-and C-terminal domains, resulting in the destabilization of the protein structure, as seen in humanmouse hybrid apoA-I molecules where the human C-terminal sequence is replaced with that of mouse apoA-I ( 16,20 ). Alternatively, it may be possible that the disruption of the ability of the C-terminal region to form ␣ -helix caused by the deletion of E235 inhibits the hydrophobic helixhelix interaction between the N-and C-terminal domains in apoA-I Nichinan. The fact that the ␣ -helix content of ‫ف‬ 80% found in the crystal structure ( 11 ) is much higher than that for monomeric apoA-I in solution ( ‫ف‬ 50%) ( 9,32 ) indicates that the true conformation in solution is more fl exible and less organized than the compactly folded, highly ordered structure in the crystal state. Indeed, whereas the C-terminal residues 219-242 are shown to be helical in the crystal structure ( 11 ), our recent hydrogen-deuterium exchange analysis of lipid-free apoA-I in dilute solution demonstrates that the C-terminal domain is entirely nonhelical (unpublished results). Thus, it is unlikely that the deletion of E235 affects the stability of deletion of residue E235 in the C-terminal helical region impairs lipid binding and is associated with low HDL cholesterol and plasma apoA-I levels ( 21,22 ). To understand the structural basis for such dysfunctions of the Nichinan mutation in terms of alterations in apoA-I tertiary structure, we examined the structure, lipid-binding properties, and ability to induce cellular cholesterol effl ux of apoA-I Nichinan and its C-terminal peptide.

Effects of deletion of E235 on apoA-I structure
Lipid-free human apoA-I is organized into two structural domains comprising an N-terminal helix bundle domain and a separate C-terminal domain ( 9,11,15 ). Because the mutation in apoA-I Nichinan occurs in the C-terminal helical region, the effects of the E235 deletion on apoA-I tertiary structure are expected to be primarily in the C-terminal domain. Indeed, there was no or little difference in the ␣ -helix content, thermal denaturation parameters between WT and the Nichinan variant  sible that the loss of E235 residue cause a "frameshift" that could disrupt the optimal charge distribution on the polar and nonpolar faces of the C-terminal helix, leading to the impaired ␣ -helix formation and lipid binding. Supporting this, the E235A mutation in apoA-I protein (unpublished observations) and in the C-terminal peptide ( 31 ) has little effect on the DMPC solubilizing ability.
Effects of deletion of E235 on the cholesterol effl ux ability of apoA-I It is well established that the C-terminal ␣ -helical region of apoA-I is important for effective cellular lipid effl ux (23)(24)(25)55 ) and formation of nascent HDL particles ( 26,56 ) via the ABCA1 pathway. In this apoA-I/ABCA1 reaction, the microsolubilization of exovesiculated domains of cell plasma membrane by apoA-I is a rate-limiting step to create discoidal HDL particles ( 28 ). Consequently, the ability of apoA-I variants to solubilize phospholipid bilayer such as DMPC vesicles is correlated with the effi ciency to effl ux cellular lipids via the ABCA1 pathway ( 28 ). As the solubilization process requires penetration of apoA-I helices into the phospholipid bilayer with the C-terminal ␣ -helix playing a key role ( 15,16,41 ), it follows that the decreased lipid-binding ability of the C-terminal helical region in apoA-I Nichinan should lead to the reduced rate of cholesterol effl ux, as actually seen in the previous ( 22 ) and this study (Fig. 7). In addition, it is likely that the decrease in the hydrophobic binding site in the C-terminal domain of apoA-I Nichinan (Fig. 3) prevents the formation of pre ␤ -HDL particles because this site is important for the acquisition of a few phospholipid molecules for the formation of pre ␤ -HDL ( 56 ). It should be noted that the reduced ability of apoA-I Nichinan to mediate cholesterol effl ux is unlikely to result from impairment of direct interactions between apoA-I and ABCA1 because the C-terminal deletion mutants, ⌬ 190-243 and ⌬ 223-243 apoA-I, which have reduced lipid affi nity, cross-link to ABCA1 as effectively as WT apoA-I ( 27 ).
Such a reduction in the abilities to induce cellular cholesterol effl ux and formation of HDL particles in the apoA-I Nichinan probands could result in a large pool of lipid-free apoA-I that is susceptible to rapid catabolism ( 57 ), consistent with a previous report that used in vivo turnover studies in rabbits to demonstrate that the Nichinan mutant is rapidly cleared (22% faster) compared with WT apoA-I ( 21 ). In addition, the fact that the size distribution of HDL particles is altered in the apoA-I Nichinan probands despite normal plasma LCAT activity ( 21 ) appears to be consistent with the notion that the stability of the N-terminal helix bundle and the lipid-binding affi nity of the C-terminal domain are key factors that modulate the distribution of apoA-I among HDL subclasses ( 16,53 ). In addition, the fi nding that the Nichinan mutant forms less stable and larger discoidal complexes with lipids compared with WT (Fig. 5) may also underlie the higher HDL 2 levels seen in the Nichinan probands.
In summary, the present study demonstrated that the disruption of the helix-forming ability of the C-terminal region by the mutation of E235 deletion contributes not the N-terminal helix bundle through disrupting the C-terminal helical structure in the lipid-free apoA-I. Rather, it is likely that the helix-disrupting effect of the Nichinan mutation in the apoA-I molecule is relevant only to the lipid-bound state. In addition, the higher ␣ -helix content seen for apoA-I 209-241 peptide than that for apoA-I 209-241/ ⌬ E235 in the lipid-free state is likely due to the self-association of apoA-I 209-241 peptide at the concentration studied ( 52 ).

Effects of deletion of E235 on apoA-I lipid binding
On the basis of the two-domain tertiary structure of apoA-I, we have proposed a two-step mechanism for binding of apoA-I to a phospholipid surface; the initial binding step occurs through hydrophobic ␣ -helices in the C-terminal domain, followed by a conformational opening of the N-terminal helix bundle to interact with lipids ( 10,15,53 ). Thus, removal ( 15,32 ) or modifi cation ( 19,23 ) of the C-terminal helical region drastically reduces the lipid-binding ability of apoA-I. The moderate effects of the E235 deletion in apoA-I Nichinan on the binding affi nity to SUV ( Table 2 ) and DMPC clearance effi ciency (Fig. 4) are consistent with the notion that the deletion of E235 partially disrupts local ␣ -helix formation of the C-terminal region. Indeed, comparison of the ␣ -helix contents between synthetic C-terminal peptides, apoA-I 209-241 and 209-241/ ⌬ E235 in the lipid-free or SUVbound states ( Table 3 ) clearly demonstrates that the E235 deletion markedly reduces the ability of the C-terminal region of apoA-I to form ␣ -helix both in the absence and presence of lipids. Because the helical propensity of apolipoproteins is critical for their lipid-binding affi nities ( 19,31,50,54 ), such a weak propensity to form ␣ -helix caused by the E235 deletion results in the less effective ability of the Nichinan protein and its C-terminal peptide to solubilize DMPC vesicles (Figs. 4 and 6). In agreement with this, apoA-I Nichinan forms relatively larger and unstable discoidal complexes with lipids compared with WT (Fig. 5). Interestingly, it was reported that the E235K mutation in apoA-I does not affect the lipid-solubilizing effi ciency ( 23 ), indicating that the absence of negative charge of E235 itself does not appear to be involved in the impaired lipid-binding ability of apoA-I Nichinan. Rather, it is pos- only to the decreased lipid-binding affi nity but also to the destabilized protein structure of apoA-I Nichinan. Such alterations in physical properties of the apoA-I Nichinan protein appear to impair the abilities to induce cellular cholesterol effl ux and biogenesis of HDL particles, resulting in lower plasma HDL cholesterol and apoA-I levels, as well as an altered HDL particle size distribution.