|Home | About | Journals | Submit | Contact Us | Français|
ApolipoproteinC-I (apoC-I) is an important constituent of high-density lipoprotein (HDL) and is involved in the accumulation of cholesterol ester in nascent HDL by inhibiting cholesterol ester transfer protein (CETP) and potentially activating lecithin:cholesterol acyltransferase (LCAT). As the smallest exchangeable apolipoprotein (57 residues), apoC-I transfers between lipoproteins via a lipid-binding motif of two amphipathic α-helices (AαHs), spanning residues 7–29 and 38–52. To understand apoC-I’s behavior at hydrophobic lipoprotein surfaces, oil drop tensiometry was used to compare the binding to triolein/water (TO/W) and palmitoyloleoylphosphatidylcholine/triolein/water (POPC/TO/W) interfaces. When apoC-I binds to either interface, the surface tension (γ) decreases by ~16–18 mN/m. ApoC-I is exchangeable at both interfaces, desorbing upon compression and readsorbing on expansion. The maximum surface pressure at which apoC-I begins to desorb (ΠMAX) was 16.8 and 20.7 mN/m at TO/W and POPC/TO/W interfaces, respectively. This suggests that apoC-I interacts with POPC to increase affinity for the interface. ApoC-I is more elastic on POPC/TO/W than TO/W interfaces, marked by higher elasticity modulus (ε) values on oscillations. At POPC/TO/W interfaces containing an increasing POPC/TO ratio, the pressure at which apoC-I begins to be ejected increases as phospholipid surface concentration increases. The observed increase in apoC-I interface affinity due to higher degrees of apoC-I/POPC interactions may explain how apoC-I can displace larger apolipoproteins, such as apoE, from lipoproteins. These interactions allow apoC-I to remain bound to the interface at higher Πs, offering insight into apoC-I’s rearrangement on triacylglycerol-rich lipoproteins as they undergo Π changes during lipoprotein maturation by plasma factors such as lipoprotein lipase.
Apolipoprotein C-I (apoC-I) is synthesized in the liver and circulates in the plasma bound to the surface of high-density lipoproteins (HDL), very low-density lipoproteins (VLDL), intermediate density lipoproteins (IDL), and chylomicrons (1). As the smallest exchangeable apolipoprotein (57 amino acids), apoC-I transfers between lipoproteins and chylomicrons, altering their metabolic properties (2, 3). On nascent HDL, apoC-I is a potent activator of Lecithin:Cholesterol Acyltransferase (LCAT), which allows for the accumulation of cholesterol ester (CE) and HDL particle maturation. ApoC-I also serves as a secondary activator of LCAT, activating the enzyme with 10–45% efficacy relative to its primary activator, apolipoprotein A-I (4–6). ApoC-I is a potent inhibitor of Cholesterol Ester Transfer Protein (CETP), which promotes exchange of CE and triacylglycerols (TAG) between HDL and VLDL or LDL (7–9). On the surface of HDL, apoC-I displaces hepatic lipase (HL), which hydrolyzes HDL phospholipids and mono- and diglycerides (10–12). In activating LCAT and inhibiting CETP and HL, apoC-I aids in the synthesis and stabilization of mature HDL particles.
Comparatively, apoC-I is known to inhibit activity of lipoprotein lipase (LPL), an enzyme critical for hydrolysis of VLDL triacylglycerols and conversion of VLDL to LDL (13). ApoC-I further inhibits clearance of TAG-rich lipoproteins and their remnants by impairing interactions with the cellular receptors mediating their uptake (14, 15). Impaired uptake of β-VLDL has been linked to apoC-I-induced displacement of apoE, the ligand for the low-density lipoprotein receptor-related protein (LRP) (16–18). Together, these findings suggest that apoC-I lipoprotein surface modifications could displace apoE or HL, while the structure of apoC-I on the lipoprotein surface may allow for a portion of the peptide to interact with and regulate the function of LCAT or CETP. Understanding of the structure of apoC-I is therefore central to understanding its functional effects on the surface of lipoproteins.
ApoC-I, as a lipid-free monomer in solution, is largely unfolded exhibiting an average helical content of 31% (19, 20). Helical content increases to 65–75% in lipid-bound or self-associated states (20–22). NMR and CD studies show that apoC-I contains four 11mer repeats, predicted to produce two lysine-rich class A amphipathic α-helices (AαH), spanning residues 7–29 and 38–52 (23–28). A common lipid surface-binding motif among apolipoproteins and synucleins, the class A AαH is characterized by a large (30–50%) apolar face essential for interactions with apolar lipid moieties and positive residues distributed in the polar-nonpolar interface, which interact with phospholipid head groups and solvent molecules (29–33).
Figure 1 depicts the N-terminal (left) and C-terminal (right) AαH helix wheel diagrams, with a dashed line demarcating their respective hydrophobic faces. The lysine residue (K52) at the end of the C-terminal AαH likely snorkels up out of the hydrophobic face (30, 31). Due to its structural similarity to apoA-I and α-Synuclein, apoC-I’s AαHs are modeled as α11/3 helices (32, 33). Such helices are slightly unwound compared to the standard α-helix, subsuming a rotation of 98.18° as opposed to 100° per residue.
To elucidate the interaction of apoC-I AαHs with phospholipid on the lipoprotein surface, several studies monitored the effect of point mutations on apoC-I binding to and stabilization of phospholipid vesicles (20, 28, 34–36). Thermal denaturation and renaturation of apoC-I:DMPC complexes, monitored via CD, revealed the greatest reduction in apoC-I α-helical content and ability to reconstitute apoC-I:DMPC complexes for mutations interrupting the apolar face of either AαH (20, 28). These studies indicate an essential cooperativity in binding of the α-helices to the phospholipid surface (20). Additional studies, utilizing DMPC-clearance assays and electron microscopy, showed that the C-terminal AαH contains aromatic residues essential for stabilizing the structure of the entire peptide and mediating phospholipid interactions (36).
Apolipoprotein monolayers spread at an air/water (A/W) interface compared with apolipoprotein deposition at phospholipid/air/water (PL/A/W) interfaces provide additional insight into the interaction between apolipoproteins and phospholipid (37–43). The surface pressure-area (Π-A) isotherms generated for compression of apoC-I monolayers at an A/W interface reveal a phase transition at Π ~ 37 mN/m and a collapse at Π ~ 47 mN/m (44, 45). Likewise, compression of apoC-I deposited on DPPC monolayers spread at an A/W interface revealed a phase transition beginning at Π ~ 24–27 mN/m and a collapse at Π ~ 49 mN/m (46–48). Brewster Angle Microscopy revealed that each phase transition was attributable to different phases of apoC-I (44–48). These results suggest a two-step desorption process for apoC-I from A/W and DPPC/A/W interfaces, wherein one AαH desorbs at lower pressures (Π = 37 mN/m and 24–27 mN/m, respectively) before complete peptide desorption from the interface at higher pressures (Π = 47 mN/m and 49 mN/m, respectively).
Together, previous work with apoC-I and DMPC vesicles and apoC-I at a DPPC/TO/W interface highlight the structural changes in apoC-I on lipid-binding and the conformational changes of the peptide as Π changes, as may happen in transition of nascent HDL to mature, spherical HDL or of VLDL to LDL. Such studies are limited, however, by lack of a truly physiologic interface, ability to monitor quantitative, real-time surface modifications induced by apolipoprotein binding, and a full analysis of phospholipid effects on apolipoprotein binding.
To better ascertain how apolipoproteins change conformations or modify lipoprotein surfaces upon binding, the oil/water surface provides a model for peptide interactions with a TAG core, while a phospholipid/oil/water interface more closely resembles the lipoprotein surface. Both interfaces can be further refined by adding other molecules known to be in lipoprotein surfaces (e.g. cholesterol, fatty acids, etc.). To date, only a few studies (37, 49–55) have examined the surface behavior of apolipoproteins or their consensus peptides at oil/water interfaces. Even fewer focused on the surface behavior of apolipoproteins at phospholipid/oil/water interfaces (56, 57).
In this study, we examined the interfacial properties of native apoC-I at TO/W and the more physiological palmitoyloleoylphosphatidylcholine/triolein/water (POPC/TO/W) interfaces. POPC better represents the phospholipid content at a lipoprotein surface than DMPC or DPPC, as POPC is found in abundance in TAG-rich lipoproteins, while DMPC and DPPC do not appear significantly on the surface of lipoproteins (57). Our goal was to understand how apoC-I, via its amphipathic α-helices, binds triacylglycerols (represented here by triolein) and/or phospholipids (POPC) and how surface properties vary based on constituents on the modeled TO/PC surface.
To monitor the modifications induced by apoC-I at the lipoprotein surface, we used an oil-drop tensiometer (53, 58, 59) to measure and compare the adsorption isotherms, desorption behaviors, and elasticity properties of apoC-I at both interfaces. From this, we were able to see how apoC-I AαH interactions with POPC alter apoC-I’s binding to and affinity for the interface. Based on the protocol of Mitsche, et. al. (57), we varied POPC surface concentration and quantitatively determine the effects on apoC-I surface retention as Π changes. From these results and comparison with other peptides we’ve studied, notably the AαH-rich apoA-I (52, 55, 57), we propose a model whereby apoC-I binds and increases Π to such a degree that molecules, such as apoE, HL, or CETP apolipoprotein activators, would be displaced from the surface.
Triolein (>99% pure) was purchased from NU-CHEK PREP, Inc. (Elysian, MN), and its interfacial tension (γ) was 32 mN/m.
POPC (1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine) was purchased from Avanti Polar Lipids (Alabaster, Alabama) dissolved in chloroform to a concentration of 25.0 mg/mL and stored at −20°C. The POPC was dried under Nitrogen and re-suspended in 2mM, pH 7.4 phosphate buffered saline (PBS) at a concentration of 2.5 or 5.0 mg/ml. Using a probe sonicator, the POPC in PBS was constituted into SUVs. POPC SUVs, using the method of Mitsche and Small (57), were injected at varied concentrations into the aqueous phase and given sufficient time to adsorb to the preformed TO drop, such that interfacial tension fell to 23–25 mN/m. At these γ values, roughly 42–37% of the TO drop was covered with POPC, based on calculations by Mitsche and Small (57).
Full-size human apoC-I (57aa) was obtained via solid-state synthesis and purified by HPLC to 95–98% purity as described (20, 28). The peptide N- and C- termini were not blocked. SDS gels, Circular Dichroism, and Electron Microscopy revealed the structure and stability of WT apoC-I prepared by this method closely resemble those of the plasma protein (data not shown). A stock solution of peptide was made by adding 2mM, pH 7.4 phosphate buffer (PB) to a final peptide concentration of 1.0 mg/mL. Varied amounts of apoC-I stocks could be added to the aqueous phase to obtain different protein concentrations (from 7.5 × 10−8 M to 1 × 10−6 M at a TO/W interface and from 1 × 10−7 M to 5 × 10−7 M at a POPC/TO/W interface).
The interfacial tension (γ) of TO/W and POPC/TO/W interfaces in the presence of varied concentrations of apoC-I in the aqueous phase was measured with an I. T. Concept (Longessaigne, France) Tracker oil-drop tensiometer (53, 58, 59), with the goal of observing and quantifying the surface activity of apoC-I at either interface. A 16-µL triolein drop was formed in 6.0 mL of gently stirred 2 mM PB, pH 7.4, containing no or varied concentrations of POPC. Varied amounts of apoC-I were added into the aqueous phase when TO/W interfaces had reached an equilibrium γ (γe) of 32 mN/m or when POPC/TO/W interfaces had reached a γ between 23 and 25 mN/m. After POPC had adsorbed to the surface, the bulk buffer with POPC SUVs was exchanged with 250 mL of 2 mM, POPC-free PB, pH 7.4, prior to addition of apoC-I into the aqueous phase. This buffer exchange was done using the protocol of Mitsche and Small (57) and resulted in the removal of >99.9% of the original buffer. The γ of apoC-I at either interface was recorded continuously until it approached a γe. Surface pressure (Π) was calculated from the surface tension of the TO/W interface (γTO = 32 mN/m) minus the surface tension of the interface with POPC and apoC-I (γ), such that Π = 32 mN/m − γ. The temperature was 25.0 ± 0.1 °C for all experiments.
Following adsorption of apoC-I to either a TO/W or POPC/TO/W interface, the drop underwent a series of compressions and re-expansions with the goal of determining if apoC-I completely or only partially desorbed from the respective interface. Once γ approached a γe, the TO drop (16 µL) was compressed by rapidly decreasing the volume by different ratios: 6.25% (1 µL), 12.5% (2 µL), 25% (4 µL), 37.5% (6 µL), 50% (8 µL), or, when possible, 62.5% (10 µL). This sudden decrease in volume induced a decrease in drop surface area, resulting in a sudden compression and abrupt decrease in γ. The oil drop was held at this reduced volume for 5–10 min, with γ recorded continuously. If apoC-I readily desorbed, γ rose back towards a γe, observed as a desorption curve. After 5–10 min, the interface was expanded by increasing the volume of the drop by equivalent ratios back to its initial volume (16 µL). As the surface area increased upon expansion, γ abruptly increased. If apoC-I adsorbed from the bulk phase and adhered to the newly formed extra surface, γ dropped back towards the initial γe, observed as a readsorption curve. This process of stress compression and re-expansion was repeated after the bulk buffer was exchanged with 150 mL of 2 mM PB, pH 7.4, devoid of apoC-I.
The desorption and readsorption protocol not only provided information on the nature of peptide ejection from either interface, but the Π at which such ejection occurs. ΠMAX is the maximum pressure (Π) that a peptide can withstand before all or part of the molecule is ejected from the surface. Upon reaching a γe following apoC-I adsorption to TO/W or POPC/TO/W interfaces, the series of experiments highlighted above were carried out in which the drop volume was decreased abruptly, thereby decreasing γ and increasing Π to a given value, Πo. The change in tension (Δγ) over the following 5–10 minutes as peptide desorbed from the surface was plotted against Πo. Regression of a linear fit to the plot reveals ΠMAX as the point where Δγ = 0, such that not even part of the peptide desorbs from the surface upon compression (51, 52).
Oscillating TO/W and POPC/TO/W interfaces with apoC-I bound allowed for comparison of visco-elasticity. After γ approached γe with apoC-I bound to either interface, the volume of the TO drop was oscillated sinusoidally at different periods ranging from 8 to 128 s and different amplitudes ranging from ± 2 µL (~13%) to ± 8 µL (50%). As the volume was oscillated, interfacial area (A) and surface tension (γ) were recorded continuously. From these γ and A measurements, pressure-area (Π-A) isotherms were derived and plotted, as γ = Π + 32.0 mN/m. These γ and A measurements also yielded the interfacial elasticity modulus, defined as ε = dγ/d(lnA), for each sinusoidal oscillation (60). Via Fourier Transform analysis, the real (ε′ =|ε| cos()) and imaginary (ε″ =|ε| sin()) components of ε were determined, allowing for derivation of the phase angle () between compression and expansion. Increasing degrees of surface visco-elasticity are marked by decreasing ε and increasing values.
The goal of slowly expanding and sequentially compressing TO drops with apoC-I bound, but differing in POPC surface concentration (ΓPOPC), was to quantitatively compare the point on compression at which apoC-I begins to be expelled from the POPC/TO/W interface. After POPC adsorption to a TO/W interface lowered γ to 23–25.0 mN/m, the bulk buffer was exchanged with 250 mL of 2 mM, POPC-free PB, pH 7.4, and γ was brought to 24.5–25.0 mN/m. The volume of the drop was then increased or decreased slowly (±0.02 µL/s), causing increases or decreases in γ. This resulted in corresponding decreases or increases in Π, respectively. The pressure of the interface (ΠI) at which apoC-I was added could be correlated to ΓPOPC, and the equivalent percentage of POPC drop coverage, based on the work of Mitsche, et. al. (57). Following apoC-I adsorption to the POPC/TO/W interface lowering γ to a γe, the bulk buffer was exchanged with 150 mL of 2 mM PB, pH 7.4, devoid of apoC-I. The volume of the TO drop was then linearly increased at 0.02 µL/s until the volume reached 30 µL. After allowing γ to equilibrate for 200 s or more, the volume was linearly decreased at −0.02 µL/s with the lower surface area limit varying, based on the condition that γ remain above 5.0 mN/m. This protocol was repeated for several ΠIs and pressure-area (Π-A) isotherms were calculated for each compression and expansion.
Changes in interfacial tension (γ) upon injection of apoC-I into the aqueous phase reveal the peptide’s surface activity at both TO/W and POPC/TO/W interfaces. Figure 2A shows a typical set of interfacial tension curves of the TO/W interface with varying concentrations of apoC-I in the aqueous phase. The initial γ of the TO/W interface was about 32 mN/m. When varying concentrations of apoC-I were added into the aqueous phase (7.5 × 10−8M to 1.0 × 10−6M), the peptide adsorbed onto the TO/W interface, lowering γ and approaching a γe. At the lowest concentration (7.5 × 10−8M), a lag time of about 800 s was observed before a sharp decrease in γ tapered into a gradual decrease approaching a γe between 1700 and 4000 s. As apoC-I concentration increased, the lag period shortened and the steepness of the rapid decrease in γ increased. Higher peptide concentrations also exhibited adsorption to lower γe values. The net changes in γ ranged from 16.5 mN/m ([apoC-I] = 7.5 × 10−8 M) to 18.0 mN/m ([apoC-I] = 1.0 × 10−6 M).
Figure 2B shows the set of interfacial tension curves of the POPC/TO/W interface with varying concentrations of apoC-I in the aqueous phase. The initial γ of the POPC/TO/W interface was ~24–25 mN/m. Addition of apoC-I to the aqueous phase was chosen as relative time 0 s. The adsorption curves of apoC-I at this interface also showed three regions: a lag period where γ fell slowly with time, a second period of faster decrease in γ, and a final period characterized by a gradual decrease in γ upon approaching a γe. Similar to apoC-I at a TO/W interface, as apoC-I concentration increased at a POPC/TO/W interface, the lag period shortened and the steepness of the rapid decrease in γ increased. Higher peptide concentrations did not, however, exhibit adsorption to lower γe values. Net changes in γ ranged from 15.5 mN/m ([apoC-I] = 1.0 × 10−7 M) to 16 mN/m ([apoC-I] = 5.0 × 10−7 M). These values are comparable to those for apoC-I at a TO/W interface. Consequently, apoC-I is surface active at both TO/W and POPC/TO/W interfaces, lowering interfacial tension in a similar manner and amount.
Following adsorption of apoC-I to TO/W or POPC/TO/W interfaces, changes in γ induced by instantaneous compressions and subsequent expansions of the TO drop allow for determination of the nature of peptide desorption and readsorption. Figure 3A shows γ and area changes during the sudden compression and re-expansion of the TO/W interface for apoC-I before and after buffer exchange. After apoC-I adsorption lowered γ to a γe of ~15.0 mN/m, the 16 µL TO drop was compressed by decreasing its volume in increments varying from 1–10 µL. This corresponded to area changes of 2–55%. Instantaneous decreases in γ increased with the size of compression. With a compression of 4 µL, γ decreased by ~2 mN/m, while a compression of 10 µL decreased γ by ~4 mN/m.
When the compressed volume was held for several minutes, γ rose towards a γe slightly lower than the initial γe. This rise in γ indicates that peptide desorbs from the surface after instant compression. Since the new γe values are lower than 15.0 mN/m, particularly for larger compressions (Figure 3A), this indicates that bound apoC-I is compressed and potentially undergoes conformational rearrangement at the surface as Π increases. When the drop was re-expanded back to 16 µL, γ increased above the initial γe. Within a relatively short time (< 60s), however, peptide readsorbed and γ fell back to ~ 15.0 mN/m, the initial γe.
150 mL of buffer was exchanged (indicated by the bar in Figure 3A) and [apoC-I] was reduced to virtually zero. γ rose during the buffer exchange from ~15.0 mN/m to ~17.0 mN/m. This indicates that some bound apoC-I may desorb from the surface. The instant compression and expansion procedure was repeated with the TO drop (Figure 3A, right). A notable difference was that γ remained at or near heightened γ values following each re-expansion. If one AαH desorbed on compression and readily re-adsorbed on re-expansion, γ would decrease to γe of ~17 mN/m. γ would only remain high following each re-expansion if the entire peptide desorbed from the surface, with virtually no peptide molecules in the aqueous phase available to readsorb and lower γ.
Figure 3B shows γ and area changes during the sudden compression and expansion of apoC-I at the POPC/TO/W interface before and after buffer exchange. Upon POPC absorption, γ fell to 23.2 ± 0.1 mN/m. A 250 mL buffer exchange caused γ to rise to 25.0 ± 0.3 mN/m, due to POPC exchange with bulk solution. ApoC-I was added in the aqueous phase and γ approached a γe of 9.5 ± 0.1 mN/m. The TO drop was compressed via decreasing its volume in increments ranging from 1–6 µL, corresponding to area changes of 2–22%.
Similar to γ responses at the TO/W interface, each instant compression caused an instantaneous decrease in γ, but γ rose towards a new γe when the compressed volume was held for several minutes. The new γe values were lower than the initial γe of the interface. This can be most noticeably seen in the γe reached (8.1 ± 0.2 mN/m) after the two 6 µL compressions immediately preceding the buffer exchange in Figure 3B. This again indicates that apoC-I could undergo conformational rearrangement as its AαHs are compressed as Π increases. After each re-expansion apoC-I readsorption lowered γ to a γe similar to the initial γe (9.7 ± 0.2 mN/m in this case). This is preliminary evidence that apoC-I does not remove POPC when it desorbs from the POPC/TO/W interface. If apoC-I removed POPC as it desorbed on compression, γe following re-expansion and readsorption of apoC-I, would have risen to ~15.0 mN/m, that of a POPC-free surface.
ApoC-I was removed from the aqueous phase by buffer exchange (indicated by the bar in Figure 3B) and γe rose to 11.9 ± 0.1 mN/m, indicating that some apoC-I desorbs. Instant compression and expansion protocol similar to that before buffer exchange was repeated (Figure 3B, right). The POPC/TO/W interface, like the TO/W interface, did not relax back to γe levels after re-expansion, but remained at or near the heightened γ. Consequently, at both interfaces, the entire apoC-I molecule is expelled upon compression and readsorbs when present in the bulk solution on re-expansion. Entire peptide ejection on compression is characteristic of exchangeable apolipoproteins (52, 53, 55, 57). The N-terminal 44 residues ([1–44] apoA-I) and C-terminal 46 residues ([198–243] apoA-I) of apoA-I are comprised of one and three AαHs, respectively, and completely desorb from and readsorb to TO/W (55) and POPC/TO/W (57) interfaces.
Instant compression and re-expansion measurements were taken at different apoC-I concentrations to estimate the pressure (ΠMAX) at which the entire peptide, as shown in Figure 3, is ejected from each interface. Differences in ΠMAX between the TO/W and POPC/TO/W interface would elucidate the role of apoC-I/POPC interactions in peptide retention on the lipoprotein surface. Since the initial pressure (ΠI) at the POPC/TO/W interface prior to apoC-I adsorption was always ~7.0 mN/m (γI = 25.0 mN/m) for these measurements, the ΠMAX value corresponds to a surface with roughly 38% POPC coverage. The data points shown in Figure 4 represent compression and re-expansion measurements prior to buffer exchange at both interfaces. Figure 4 shows that ΠMAX of apoC-I at a TO/W interface is 16.8 mN/m (solid circles) and ΠMAX of apoC-I at a POPC/TO/W interface is 20.7 mN/m (open circles). The greater ΠMAX of apoC-I at a POPC/TO/W interface suggests that apoC-I binds more strongly to a POPC/TO/W than to a TO/W interface. ApoC-I AαH interactions with POPC likely increase its affinity for the interface.
The viscoelasticity of apoC-I at a TO/W or POPC/TO/W interface can be determined by oscillating either interface. Viscoelasticity represents a surface’s ability to reduce deviations of γ from its equilibrium value by allowing for relaxation processes (60). Typical relaxation processes consist of periodic adsorption and desorption of surface-active material (here, apoC-I) from expanded and compressed surface elements. This causes a change in the surface being compressed versus that being expanded in an oscillation shown as a hysteresis between the respective pressure-area (Π-A) isotherms.
Derived from the γ and A measurements during oscillations, the elasticity modulus (ε) is defined as the increase in γ corresponding to a small increase in the area of a surface element; ε = dγ/d(ln(A)). The phase angle () represents the difference between the real and imaginary components of ε. A higher ε and lower , therefore, indicates a surface more resistant to deformations or changes. If apoC-I interactions with POPC retain the peptide to higher Πs, as indicated by a higher ΠMAX (Figure 4), apoC-I should be more elastic at a POPC/TO/W than at a TO/W interface. This would be marked by lower hysteresis in oscillation Π-A isotherms, greater ε values, and lower values.
Figure 5 shows Π-A isotherms derived from equilibrium oscillations of apoC-I at TO/W (A) and POPC/TO/W (B) interfaces at different amplitudes and periods prior to buffer exchange. Two sets of Π-A isotherms for apoC-I at a TO/W interface, each representing two or three periods for the given amplitude, are shown in Figure 5A. The Π-A isotherms for apoC-I at a TO/W interface show significant hysteresis between compression and expansion when oscillated at larger amplitudes (± 8 µL) and periods (32 and 128 s, represented by cyan and magenta lines, respectively). This indicates that sufficiently large Π changes and time must be provided for observable desorption of apoC-I on compression.
The top half of Table 1 shows the ε and calculated (as described in Materials and Methods) from the Π-A isotherms depicted in Figure 5A. Such analysis reveals ε and values at the TO/W interface ranging from 20.4 to 29.6 mN/m and 10.6 to 23.5° upon oscillations at amplitudes of ± 2 and ± 8 µL. With increasing periods at a small amplitude of ± 2 µL, ε decreases by 9.2 mN/m and increases by 12.9°. These significant changes reveal more apoC-I desorbs from and readsorbs to the surface with a greater oscillatory period, such that the surface is more viscoelastic.
Two sets of Π-A isotherms for apoC-I at a POPC/TO/W interface, each representing three periods for the given amplitude, are shown in Figure 5B. The Π-A isotherms for apoC-I at a POPC/TO/W interface show significantly less hysteresis between compression and expansion than those at a TO/W interface (Figure 5A). The lower half of Table 1, representing analysis of Figure 5B, reveals ε and values of the POPC/TO/W interface ranging from 48.4 to 51.5 mN/m and 4.5 to 12.6° at oscillation amplitudes of ± 2 and ± 4 µL. Compared to a TO/W interface, these higher ε and lower values indicate apoC-I is more elastic at a POPC/TO/W interface. With increasing periods at a small amplitude of ± 2 µL, the change in ε is insignificant (3.1 mN/m) and increases by 6.2°. Compared to a TO/W interface oscillated at the same amplitude, these changes are much lower, indicating apoC-I binds with higher affinity to a POPC/TO/W interface, creating a more elastic surface.
The small and slight hysteresis in oscillations of apoC-I at a POPC/TO/W interface is indicative of a very small fraction of apoC-I, not POPC, desorbing from the interface. A series of control oscillations of the POPC/TO/W interface at varying initial Πs were performed over a wide range of amplitudes and periods (data not shown) and no hysteresis was observed, even for large amplitudes. This suggests that the interface is nearly completely elastic, such that there is no surface relaxation in which POPC desorbs and the lag between area changes and tension responses () is nearly zero (60).
To determine the effect of varying ΓPOPC on surface retention of apoC-I, TO drops with apoC-I bound and increasing ΓPOPC underwent slow (±0.02 µL/s) expansions and subsequent compressions. This allows for quantitative comparison of the point on compression (the envelope point) at which apoC-I begins to be expelled from the POPC/TO/W interface. Figure 6 shows compression Π-A isotherms of the POPC/TO/W interface when apoC-I adsorbs to the POPC/TO/W interface at various initial interfacial tension values (γI). Greater γ I values correspond to a lower percentage of phospholipid coverage on the TO drop. Since Π = 32.0 mN/m – γ, lower ΠIs correspond to lower POPC surface concentration (ΓPOPC) while higher ΠIs correspond to higher ΓPOPC (57). As listed in Table 2, ΠI values prior to apoC-I adsorption vary from 5.0 mN/m (31.8% of the drop is coated with POPC) in line (a) to 13.4 mN/m (51.8% of drop is coated with POPC) in line (d). Line (e) represents the Π-A compression isotherm for a POPC/TO/W interface devoid of apoC-I. As it does not overlap with the other compression isotherms, it shows that apoC-I adsorbed to and was being ejected from the POPC/TO/W interface in (a), (b), (c), and (d).
In Figure 6, the envelope point (*) represents the Π and A at which apoC-I begins to be ejected from the POPC/TO/W interface upon compression (direction indicated by arrow). Visually, the envelope point is the point at which the slope of the compression isotherm dramatically decreases as area decreases. As ΓPOPC on the interface increases, the envelope point Π (ΠE) increases, while the envelope point area (AE) decreases. Listed in Table 2, ΠE for apoC-I at the POPC/TO/W interface increases from 21.2 mN/m at the lowest ΓPOPC by 5.2 mN/m to 26.4 mN/m at the highest ΓPOPC. Conversely, AE decreases by 15.1 mm2 from 37. 3 to 22.2 mm2 over the same change in ΓPOPC. These data suggest that the higher the ΓPOPC on the POPC/TO/W interface, the greater affinity apoC-I has for the interface and the more difficult it is to expel apoC-I from the surface. Consequently, apoC-I is retained up to higher pressures and smaller areas on the lipoprotein surface as ΓPOPC increases.
On the surface of lipoproteins, it is reasonably assumed that apolipoproteins, including apoC-I, interact with surface-located phospholipids (55). It is also possible that apolipoproteins interact with core molecules, such as TAG or CE, notably when the number of surface lipid molecules is too small to adequately cover the core (55). In this study, our goals were twofold. First, we sought to characterize the surface activity and reversible binding of human apoC-I at TO/W and the more physiological POPC/TO/W interfaces. Second, we sought to understand the effects of apoC-I interactions with POPC on peptide desorption from and retention on the lipoprotein surface.
The first objective of this study was to demonstrate the surface activity of apoC-I at TO/W and POPC/TO/W interfaces. Figure 2 demonstrates the surface activity of apoC-I at both TO/W and POPC/TO/W interfaces. An increase in the rate of peptide adsorption was observed as apoC-I concentration in the aqueous phase increased. The decrease in γ (or increase in Π) upon binding to each interface was comparable, equal to 16.5–18.0 mN/m and 15.5–16.0 mN/m, respectively. At a concentration of 1.2 × 10−6 M, similar to that of apoC-I used at a TO/W interface, [1–44] apoA-I and [198–243] apoA-I could increase Π by around 13.0 mN/m, respectively, at a TO/W interface (55). This suggests a higher surface affinity for apoC-I than the C- and N-termini of apoA-I.
The observed surface affinity of apoC-I for either interface is imparted by its structure. Apolipoproteins exhibit strong lipid-binding properties due to their characteristic secondary structures, containing amphipathic α-helix (AαH) or amphipathic β-sheets (AβS) (29, 49–53, 61). The lipid-binding motifs of apoC-I lie in its predicted N- and C-terminal AαHs (25, 28), as shown in Figure 1. The hydrophobic face of each AαH (marked by a dotted line) is essential for interactions with apolar moieties on phospholipids (here, POPC) and neutral lipids (here, TO). According to the scale determined by Wimley and White (62), ΔGb→w of partitioning for the hydrophobic face of the N-terminal AαH is 3.76 kcal/mol, or 0.42 kcal/mol per residue. ΔGb→w for the hydrophobic face of the C-terminal AαH is 4.13 kcal/mol, or 0.69 kcal/mol per residue. Preferential lipid-binding is implicated in the unfavorable, positive ΔGb→w of transfer from a bilayer interface to the aqueous phase for the hydrophobic faces of both AαHs.
Linked with its surface activity, we sought to determine the nature of apoC-I binding to TO/W and POPC/TO/W interfaces. AαH structures typically show reversible binding to the TO/W surface, desorbing from and readsorbing to the surface upon compression and re-expansion (49, 52, 55). Figure 3 reveals that apoC-I reversibly binds to both interfaces, desorbing when each interface is instantaneously compressed above ΠMAX. ApoC-I’s complete peptide desorption is evident in the γ responses of both interfaces for compression-expansion protocols after buffer exchange (right side of Figure 3A and B). With peptide present in the aqueous phase (Figures 3A and B), apoC-I readsorbs to both interfaces on re-expansion, decreasing γ to its initial equilibrium value (15.0 mN/m and 9.5 mN/m, respectively).
The left half of Figure 3A and B suggest conformational rearrangement on compression of the AαHs of apoC-I prior to complete peptide desorption at both interfaces, as the γe values reached after each compression are lower than the initial γe. The rise in γ at both interfaces due to the buffer exchange depicted in Figure 3A and B further suggests peptide conformational rearrangement at the surface, potentially followed by complete peptide desorption into the aqueous solution. To explain such conformational rearrangement, results of previous studies may suggest a two-step desorption model for apoC-I, with one AαH desorbing prior to complete peptide ejection.
Compression of apoC-I deposited on a DPPC monolayer at an A/W interface revealed two phase transitions (46–48). The first occurs between Π = 24–27 mN/m and the second at Π ~ 49 mN/m. Brewster Angle Microscopy showed that only apoC-I was affected in each phase transition (46–48) This led to the proposal that, at Πs from 24–27 mN/m, one AαH on apoC-I desorbs from the subphase, aligning itself towards the air, in a similar direction to that of the hydrophobic DPPC acyl-chains (46). The entire peptide was ejected from the monolayer at Π ~ 49 mN/m.
Indicative of the order of apoC-I AαH desorption, the amide proton resonance line widths and deuterium exchange rates of the NMR structure of apoC-I:sodium dodecyl sulfate complexes suggest that the N-terminal AαH binds less tightly to the detergent than the C-terminal AαH (25). In agreement with these results, James, et. al., used mass spectrometry to show that the C-terminal region was the portion of apoC-I most highly protected from proteolysis (36). DMPC clearance assays and electron microscopy further revealed that single or double mutations in the apolar face of the C-terminal AαH of F42 and F46 to G or A result in a 2–3.5-fold reduction in apoC-I binding affinity for DMPC vesicles and apoC-I:DMPC complexes with atypical diameters (36). However, CD studies of apoC-I:DMPC complexes show that interruptions in the apolar face of either the N- (G15P) or C-terminal (T41P, T45P) AαH greatly reduce protein-lipid binding (20, 28).Of all explored mutations, these also show the lowest ability to reconstitute apoC-I:DMPC disks on thermal renaturation (20, 28). This indicates an essential cooperativity in binding of the α-helices to the phospholipid surface (20).The thermal denaturation and reconstitution of apoC-I:DMPC complexes, even at low ramp rates and small temperature changes, failed to show multiphase kinetics indicative of a two-step peptide desorption from the lipid/water interface (20, 28, 34, 35).
If apoC-I follows a two-step desorption from the lipoprotein surface, the NMR structure of apoC-I and greater ΔGb->w values for the hydrophobic face of the C-terminal AαH predict that the N-terminal AαH desorbs prior to complete peptide desorption. Consequently, in the seconds following each instantaneous compression in Figure 3, the N-terminal AαH desorbs from the surface. This conversion is followed by a longer period of desorption of the C-terminal AαH of apoC-I to reach a new peptide equilibrium with the bulk solution.
Having characterized the reversible binding of apoC-I to TO/W and POPC/TO/W interfaces, the second goal was to determine the effect of POPC on apoC-I surface binding and retention. To that end, the pressure ΠMAX, at which the entire peptide begins to be ejected from TO/W and POPC/TO/W interfaces was determined. Figure 4 reveals a ΠMAX value of 16.8 mN/m for apoC-I at a TO/W interface. This ΠMAX is consistent with previous findings for AαHs at a TO/W interface, falling within the range of 13–19 mN/m (49, 52, 55). Comparatively, the ΠMAX values for [198–243] apoA-I and [1–44] apoA-I are 16.2 and 13.2 mN/m, respectively (55). ΠMAX for apoC-I at a POPC/TO/W depends on ΓPOPC. Figure 4 reveals that apoC-I has a ΠMAX of 20.7 mN/m when ~38% of the TO drop is covered by POPC. This higher ΠMAX at a POPC/TO/W interface is likely due to AαH interactions with phospholipids and neutral lipids increasing peptide affinity for the surface, keeping it bound at higher Πs.
The increase of affinity for the interface induced by apoC-I/POPC interactions was further explored by sinusoidal oscillations of both interfaces following apoC-I adsorption. Figure 5 reveals the elastic properties of apoC-I at TO/W and POPC/TO/W interfaces. Previous studies have revealed that apoA-I, the major apolipoprotein constituent of HDL and primary activator of LCAT, binds TO/W interfaces via its AαH structural motifs and creates a highly flexible visco-elastic interface (52). Similarly, analysis of large oscillations or small oscillations with a slow period in Figure 5A reveals apoC-I forms a visco-elastic surface at the TO/W interface, marked by a large hysteresis in compression and expansion Π-A isotherms, a low ε, and large phase angle ().
ApoC-I at a TO/W interface exhibits a marked hysteresis in its Π-A oscillation isotherms only at compressions of larger amplitudes and longer periods. This supports the two-step model of apoC-I desorption from the interface, as ample time must be provided for the N-terminal AαH to desorb prior to complete peptide desorption. Oscillations of amplitude ± 2 µL over a range of increasing periods yield Π-A isotherms for apoC-I at a TO/W interface with increasing by 12.9° from 10.6 to 23.5° and ε decreasing by 9.2 mN/m from 29.6 to 20.4 mN/m.
Comparatively, the Π-A isotherms for apoC-I at a POPC/TO/W interface, oscillated over a range of amplitudes and periods, exhibit a much smaller hysteresis, even at larger amplitudes and longer periods. This is indicative of a more elastic interface. Oscillations of amplitude ± 2 µL over a range of increasing periods yield Π-A isotherms for apoC-I at a POPC/TO/W interface with increasing by only 6.2° from 6.4 to 12.6° and ε decreasing by only 3.1 mN/m from 51.5 to 48.4 mN/m. Compared to similar oscillations at a TO/W interface, these higher ε and lower values are more resistant to change at a POPC/TO/W interface, demonstrative of a more elastic interface. These data indicate that apoC-I AαH interactions with POPC increase its affinity for and retention on the surface, thereby increasing surface elasticity.
To elucidate the role of AαH/phospholipid interactions on peptide retention at the lipoprotein surface, TO drops with apoC-I bound and increasing ΓPOPC underwent slow expansions and subsequent compressions. Figure 6 and Table 2 show the dependency of the expulsion of apoC-I from the POPC/TO/W interface on ΓPOPC of the TO drop. Increases in ΓPOPC of the interface yield increases in ΠE and decreases in AE, that is, the envelope pressure and area, respectively, at which apoC-I begins to be ejected from the interface. When POPC covers 32.8% of the TO drop surface (line (a) in Figure 6), ΠE = 21.2 ± 0.2 mN/m and AE = 37.3 ± 0.2 mm2, while at a POPC coverage of 51.8% (line (d)), ΠE = 26.4 ± 0.3 mN/m and AE = 22.2 ± 0.3 mm2. These results reveal that larger POPC:apoC-I ratios cause retention of apoC-I to higher Πs and more compact areas.
Based on the findings of this study, we propose possible mechanisms by which apoC-I may contribute to steps of lipoprotein remodeling, including CETP inhibition and prevention of hepatic uptake of TAG-rich lipoproteins. In order to inhibit CETP activity, the N-terminus of apoC-I could desorb from the lipoprotein surface, interacting with CETP and thereby interfering with its function (7, 36). With no discovered CETP-binding domain on apoC-I, a more likely mechanism of inhibition (66) is that apoC-I has a higher affinity for the lipoprotein surface than CETP or its activators: apolipoproteins A-I, A-II, A-IV, and E (63–67). ApoA-IV has the weakest lipid affinity of any apolipoprotein (68, 69) and Human C-apolipoproteins have been shown to displace apoA-IV from TAG-rich lipoprotein surfaces (69). The C-terminus of apoA-I is predicted to be the most lipophilic portion of apoA-I, with the N-terminal helix bundle providing stability at the lipoprotein surface (70–72). Yet, as shown, apoC-I causes greater Π changes and has a higher ΠMAX than the C-terminal of apoA-I (55). Under this model of inhibition, apoC-I, at a high enough plasma concentration, could bind TAG-rich lipoproteins or compact HDL particles and drastically increase Π. These Π modifications could displace bound CETP and CETP activators or prevent them from binding (66). Structurally similar to apoA-I, apoE could also be displaced from β-VLDL by apoC-I-induced Π modifications, thereby inhibiting lipoprotein uptake (16–18, 72).
In summary, our interfacial studies reveal that apoC-I is surface active at both TO/W and POPC/TO/W interfaces, binding the interfaces via its two amphipathic α-helices and lowering interfacial tension, γ. ApoC-I exhibits properties of an exchangeable apolipoprotein, desorbing upon compression and readsorbing upon re-expansion of either interface. AαH interactions with POPC increase apoC-I affinity for the interface, demonstrated by a higher ΠMAX when roughly 38% of the interface was covered with POPC. AαH interactions with POPC also increase surface elasticity, demonstrated by a higher elasticity modulus (ε), a lower phase angle (), and higher Πs upon compression during interface oscillation. Finally, apoC-I affinity for the POPC/TO/W interface increases as ΓPOPC increases, evident in higher envelope point Πs at lower envelope areas upon compression. We propose that the high affinity binding apoC-I to lipoprotein surfaces provides a possible model for inhibition of CETP activity by displacing or preventing the binding of its apolipoprotein activators. Similarly, the high Π changes on binding to triacylglycerol-rich lipoproteins could displace or prevent the binding of apoE, so as to inhibit their receptor-mediated uptake.
The authors would like to thank Dr. Olga Gursky for providing the human apolipoprotein C-I used. The authors also thank Dr. Matt Mitsche for providing technical help and advice on experimental protocol.
†This work is supported in part by Grant NIH-NHLB1 2P01 HL26335-21