HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 280, Issue 44, 36773-36783, November 4, 2005

This Article Abstract Full Text (PDF) All Versions of this Article: 280/44/36773    most recent M506789200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Qin, S. Articles by Tatulian, S. A. Search for Related Content PubMed PubMed Citation Articles by Qin, S. Articles by Tatulian, S. A. Social Bookmarking                      What's this?

Evidence for the Regulatory Role of the N-terminal Helix of Secretory Phospholipase A₂ from Studies on Native and Chimeric Proteins^*

Shan Qin, Abhay H. Pande, Kathleen N. Nemec, Xiaomei He¶, and Suren A. Tatulian1

From the Biomolecular Science Center, University of Central Florida, Orlando, Florida 32826, the Department of Molecular Biology, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts 02114, and the ^¶ Department of Chemistry, University of Central Florida, Orlando, Florida 32816

Received for publication, June 22, 2005 , and in revised form, August 4, 2005.

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

 
The phospholipase A₂ (PLA₂) enzymes are activated by binding to phospholipid membranes. Although the N-terminal -helix of group I/II PLA₂s plays an important role in the productive mode membrane binding of the enzymes, its role in the structural aspects of membrane-induced activation of PLA₂s is not well understood. In order to elucidate membrane-induced conformational changes in the N-terminal helix and in the rest of the PLA₂, we have created semisynthetic human group IB PLA₂ in which the N-terminal decapeptide is joined with the ¹³C-labeled fragment, as well as a chimeric protein containing the N-terminal decapeptide from human group IIA PLA₂ joined with a ¹³C-labeled fragment of group IB PLA₂. Infrared spectral resolution of the unlabeled and ¹³C-labeled segments suggests that the N-terminal helix of membrane-bound IB PLA₂ has a more rigid structure than the other helices. On the other hand, the overall structure of the chimeric PLA₂ is more rigid than that of the IB PLA₂, but the N-terminal helix is more flexible. A combination of homology modeling and polarized infrared spectroscopy provides the structure of membrane-bound chimeric PLA₂, which demonstrates remarkable similarity but also distinct differences compared with that of IB PLA₂. Correlation is delineated between structural and membrane binding properties of PLA₂s and their N-terminal helices. Altogether, the data provide evidence that the N-terminal helix of group I/II PLA₂s acts as a regulatory domain that mediates interfacial activation of these enzymes.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

 
Enzymes of the phospholipase A₂ (PLA₂)² family hydrolyze the sn-2 ester bond of diacylglycerophospholipids and thus initiate the biosynthesis of lipid-derived mediators, such as eicosanoids and platelet-activating factor, which are potent mediators of inflammation, allergy, and tumorigenesis (1-3). These enzymes are divided into several groups and subgroups based on their amino acid sequences, tissue distribution, and functional features (1, 2).

PLA₂s gain their full activity only when they bind to phospholipid micelles or membranes, an effect known as interfacial activation (4-6). The molecular mechanism of interfacial activation has been the subject of debate over several decades. Initially, similarities of crystal structures of PLA₂s with and without bound substrate mimics argued against conformational changes in the enzymes during interfacial activation (4). Later, the interfacially activated form of group IB PLA₂ was modeled by anion-mediated dimers (7), which revealed significant conformational changes compared with "inactive" monomeric forms, mainly involving the loop that has the functionally important Tyr⁶⁹, and the presence of an assisting water molecule (8). Several spectroscopic studies also identified conformational changes in group IB and IIA PLA₂s upon binding to phospholipid micelles or membranes. Fluorescence studies indicated that the N-terminal helix of porcine group IB PLA₂ becomes rigid upon enzyme-substrate complex formation at the membrane surface (9, 10). This was confirmed by NMR experiments, which indicated that the N-terminal -helix and H-bonded catalytically important residues of porcine group IB PLA₂ were flexible in the aqueous buffer and adopted a fixed conformation in a ternary complex of the enzyme with a substrate analog and phospholipid micelles (11, 12). Stabilization of the structure of membrane-bound human group IB was identified by Fourier transform infrared (FTIR) spectroscopy (13). In contrast, NMR and FTIR data have indicated that group IIA PLA₂s are characterized with stable secondary structure free in solution, including the N-terminal helix, and membrane binding of these enzymes is accompanied with destabilization of -helices (14-17).

Previously, we have shown that deletion of the N-terminal -helix of human IB PLA₂ results in a loss of the productive mode membrane binding capability and a 100-fold inhibition of the enzyme activity (13). Thus, experimental evidence accumulated to date suggests that the N-terminal -helix of group I/II PLA₂s is a crucial structural domain necessary for a productive mode membrane binding and activity of the enzymes (13), conformational changes do occur in these enzymes during interfacial activation involving the N-terminal helix, and these conformational changes are group-specific. In order to understand the significance of differences in membrane-induced conformational changes in group I and II PLA₂s and the role of the N-terminal helix, it is important to examine the structural and functional effects of replacement of the N-terminal helix between PLA₂ isoforms. In this work, we present our findings on the effects of substitution of the N-terminal helix of human group IB PLA₂ (hIBPLA₂) by that of human group IIA PLA₂ (hIIAPLA₂) on the enzyme activity, membrane binding strength, membrane-induced conformational changes, and the precise mode of membrane binding. In order to monitor structural changes in individual helical segments of the protein, we have chemically ligated unlabeled synthetic decapeptides corresponding to the N-terminal helix of either IB or IIA PLA₂s with the uniformly ¹³C-labeled fragment of hIBPLA₂ that lacks the first 10 residues, and we used the semisynthetic, segmentally ¹³C-labeled proteins to assess site-specific conformational effects upon membrane binding. Segmental isotope labeling also allowed determination of the angular orientation of membrane-bound proteins by polarized attenuated total reflection Fourier transform infrared (ATR-FTIR) spectroscopy. Observed conformational differences in the N-terminal and internal helices of hIBPLA₂ and of the chimeric PLA₂ (N10IIA/IB) containing the first 10 residues of hIIAPLA₂ and the rest from hIBPLA₂ indicate an important role for the N-terminal helix as a domain that regulates structural transformations during interfacial activation and facilitates proper membrane binding of group I/II PLA₂s.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

 
Materials—The lipids 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) and 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoglycerol (POPG) were purchased from Avanti%20Polar%20Lipids">Avanti Polar Lipids (Alabaster, AL). N-(Fluorescein-5-thiocarbamoyl)-1,2-dihexadecanoyl-sn-glycero-3-phosphoethanolamine (FPE) was from Invitrogen. The N-terminal decapeptides of hIBPLA₂ and hIIAPLA₂, Ac-Ala-Val-Trp-Gln-Phe-Arg-Lys-Met-Ile-Lys-NH₂ (N10hIB) and Ac-Asn-Leu-Val-Asn-Phe-His-Arg-Met-Ile-Lys-NH₂ (N10hIIA), respectively, were synthesized by Advanced ChemTech (Louisville, KY) and were 99% pure, as confirmed by high pressure liquid chromatography and mass spectrometry. C-terminally thioesterified peptides, containing a -COSCH₂COOH group at the C terminus, were synthesized by SynPep Corp. (Dublin, CA) and were 99% pure. Most of other chemicals were purchased from Sigma.

Production of Recombinant and Semisynthetic Proteins—The semisynthetic, segmentally ¹³C-labeled hIBPLA₂ was produced as described previously (18). Briefly, the plasmid for the PLA₂ fragment N10 that lacks the first 10 residues was constructed by using a human pancreatic cDNA library, and the uniformly ¹³C-labeled N10 was expressed in Escherichia coli BL21(DE3) in an M9 minimal medium using ¹³C₆-D-glucose as a single metabolic source of carbon. The N10 fragment was purified by ion exchange and size exclusion columns, as described (18). The C-terminally thioesterified N-terminal decapeptide N10hIB was ligated with the N-terminal cysteine of the N10 fragment, followed by refolding of the ligated PLA₂ by dialysis against 25 mM Tris-HCl, 5 mM CaCl₂, 5 mM L-cysteine, 0.9 M guanidinium HCl (pH 8.0), and purified by using an ion exchange Mono Q 5/50 column. The chimeric, segmentally ¹³C-labeled N10IIA/IB PLA₂ was produced in a similar manner except that the thioesterified N10hIIA peptide was used for chemical ligation. In both cases, the ligation reaction was allowed to proceed for 6 h at 37 °C. The ligated protein was separated from the unligated peptide and the N10 fragment by an ion exchange Mono Q column. The profile of elution of the chimeric PLA₂ sample from the Mono Q column showed two major protein peaks (Fig. 1A), one of which was identified as the chimeric PLA₂ and the other as the unligated N10 fragment (the peptide was eluted before the ligated PLA₂). Analysis of the elution peaks by SDS-PAGE showed that the chimeric PLA₂ was purified to homogeneity (Fig. 1B). The prokaryotic expression vector with the inserted hIIAPLA₂ gene that incorporates an N1A mutation was kindly provided by Prof. David Wilton (University of Southampton, UK). Expression and purification of hIIAPLA₂ were conducted as described (19), except the purification by a heparin column was followed by additional purification using a size exclusion HiLoad Superdex 75 column, which yielded highly pure and active enzyme.

View larger version (31K): [in this window] [in a new window]   FIGURE 1. Purification of the semisynthetic, segmentally 13C-labeled chimeric PLA2 by ion exchange chromatography. The thioesterified decapeptide Asn-Leu-Val-Asn-Phe-His-Arg-Met-Ile-Lys-NH2-COSCH2COOH was ligated with the 13C-labeled fragment N10 in a buffer containing 6 M guanidinum-HCl, 100 mM sodium phosphate, 5% -mercaptoethanol, 1 mM EDTA, 4% thiophenol, and 4% benzyl mercaptan, pH 7.4. The concentrations of the N10 fragment and the peptide were 16 and 1.6 mg/ml, respectively. The reaction was allowed to proceed for 6 h at 37 °C, followed by refolding of the ligated protein and purification using ion exchange Mono Q 5/50 column chromatography. A, the elution profile of the sample from the Mono Q column (continuous line, left axis). The initial buffer was 20 mM Tris-HCl (pH 9.0). Elution was conducted with a stepwise increase of NaCl concentration in the same buffer (broken line, right axis), and 0.5-ml fractions were collected. The peak at fraction number 25 was the ligated product, i.e. the chimeric PLA2, as confirmed by its molecular mass, high PLA2 activity, and spectroscopic data, whereas the peak at fraction number 32 was the unligated 13C-labeled N10 fragment. The free peptide eluted before the chimeric PLA2. B, Coomassie-stained SDS gel of elution fractions shown in A. The single band corresponding to fraction 25 shows that the ligated chimeric PLA2 was isolated as a pure protein.

Fluorescence and Circular Dichroism Experiments—Fluorescence and CD measurements were conducted using a Jasco 810 spectrofluoropolarimeter (Jasco Corp., Tokyo, Japan), as described previously (13). This particular model is designed for CD experiments and is equipped with an additional photomultiplier tube mounted at 90° for fluorescence measurements, as well as with a Peltier temperature controller. Fluorescence experiments were carried out on samples in 100 mM NaCl, 1 mM NaN₃, and 50 mM Hepes (pH 7.4) contained in a rectangular 0.4-cm optical path length quartz cuvette, at 22 °C, using excitation and emission slits of 4 and 10 nm, respectively. For CD experiments, the proteins were dissolved in a 10 mM sodium phosphate buffer, pH 7.4, and measurements were conducted using a 0.5-mm optical path length quartz cuvette. In both fluorescence and CD experiments, five scans were averaged, and the spectra were corrected by subtracting the spectra of blank buffers.

Membrane Binding Experiments—In order to measure the binding to phospholipid membranes of the N10IIA/IB chimeric PLA₂, as well as of group IB and IIA PLA₂s and their N-terminal decapeptides, fluorescence spectroscopy was used. The conventional resonance energy transfer technique could not be used because some molecules did not have tryptophan (the N10IIA/IB and hIIAPLA₂) and some did not have any fluorophores at all (the N10hIIA peptide). Instead, the increase in fluorescence emission intensity of FPE in vesicle membranes upon protein or peptide binding was used, as described before (20, 21). FPE was incorporated at 2 mol % in large unilamellar phospholipid vesicles, which were prepared by extrusion through 100-nm pore size polycarbonate membranes, as described (13). The buffer was 10 mM Hepes, 1 mM NaN₃, 1 mM EGTA (pH 7.4), and fluorescence experiments were conducted as described above. Excitation was at 490 nm, and emission spectra were recorded between 502 and 540 nm at 22 °C. After recording an initial spectrum of vesicles in the buffer, proteins or peptides were added to the lipid suspension at increasing concentrations from a stock solution, and consecutive fluorescence emission spectra were recorded following equilibration of the sample for 2 min, with constant stirring. The fluorescence intensity of FPE is sensitive to the ionization state of the carboxyl group of fluorescein; the emission intensity increases with increasing deprotonation. The fluorescein moiety of FPE is located at the interface of the negatively charged membrane, where the local environment is more acidic than the bulk phase because of electrostatic attraction of protons to the membrane. This causes partial protonation of the carboxyl group of fluorescein, resulting in a moderate level of fluorescence intensity. Adsorption of cationic protein or peptide molecules to the membrane reduces the negative surface charge of the membrane, which results in a protein (or peptide) dose-dependent deprotonation of fluorescein and an increase in the fluorescence intensity. This effect was used to measure binding isotherms for PLA₂s and their N-terminal peptides.

Analysis of Membrane Binding Data—Binding of proteins and peptides to membranes was analyzed using a logic described earlier (13). The protein is added to preformed lipid vesicles, resulting in binding of the protein to the external surface of vesicles. If one considers that only a fraction of the total lipid is accessible to the protein, and the binding of each protein molecule makes N lipid molecules inaccessible for other proteins, then the dissociation constant of the binding process can be presented as shown in Equation 1,

(Eq. 1)

where [P] and [L] are the total concentrations of the protein and the lipid, respectively, and [P]_b is the concentration of the membrane-bound protein. When the lipid fluorescence is monitored as the lipid is being titrated with the protein, the relative change in the fluorescence signal is as shown in Equation 2,

(Eq. 2)

where F is the change in fluorescence intensity at a given protein concentration; F_max is the saturating level of F at high protein concentration, and [L]_b is the concentration of protein-bound lipid. We have used F values at 518 nm. Using [P]_b = [L]_b/N = F_rel[L]/N (where F_rel F/F_max), we convert Equation 1 into a quadratic equation with respect to F_rel and solve it as shown in Equations 3 and 4,

(Eq. 3)

where

(Eq. 4)

The plus sign in front of the square root is useless because it results in values of F_rel >1. It is seen from Equation 1 that at half-saturation of protein binding, when F_rel = 0.5 (Equation 5),

(Eq. 5)

where [P] is the protein concentration corresponding to F_rel = 0.5.

Changes in FPE fluorescence intensity, F, were measured at various protein concentrations, [P], and F values were plotted against F/[P]. These linear (Scatchard) plots were used to evaluate F_max and [P] from the extrapolated intercept with the F axis and from the slope, respectively. Experimental binding isotherms were constructed by plotting F/F_max against [P]. The theoretical isotherms were simulated through Equation 3 using K_D and N as fitting parameters. Despite two fitting parameters, determination of both K_D and N by this method is reliable. In Equation 5, [P] can be determined using the experimental data, as described above, and is 0.52 for vesicles 100 nm in diameter, with a membrane thickness of 4 nm (13). Replacement of K_D in Equation 4 by the expression in equation 5 yields an equation with only one unknown, i.e. N. Hence, fitting of the experimental curves can be done by varying N in a physically reasonable range, and once a best fit is achieved, the K_D value can be calculated through Equation 5 by using the best fit value of N. The Gibbs free energies of membrane binding, G_b, were calculated using dissociation constants, as described (13).

Attenuated Total Reflection FTIR Experiments—Polarized ATR-FTIR experiments were carried out using a Vector 22 FTIR spectrometer (Bruker Optics, Billerica, MA), as described previously (13). A model 611 Langmuir-Blodgett trough (Nima, Coventry, UK) was used to deposit a POPC monolayer on a germanium internal reflection plate 0.1 x 2 x 5 cm³ in size and with 45° angles (Spectral Systems, Irvington, NY), followed by injection of sonicated vesicles composed of 80 mol % POPC and 20 mol % POPG and formation of supported phospholipid membranes. The protein or the peptide was dissolved in an H₂O-based buffer consisting of 100 mM NaCl, 1 mM NaN₃, 1 mM EGTA, 50 mM Hepes (pH 7.4) and was injected into the ATR cell that contained the supported membrane. After allowing the protein to bind to the membrane (15 min), the cell was gently flushed with a ²H₂O-based buffer of the same composition, followed by recording a series of spectra at alternating parallel and perpendicular polarizations of the incident infrared light, at 2 cm^-1 resolution. The polarized spectra were used for two purposes. First, the spectra measured at parallel and perpendicular polarization were used to create "corrected" spectra as A_corrected = A_|| + 0.8A , which were used for calculation of the second derivative spectra and evaluation of secondary structures of proteins, as described before (13, 22). Second, the polarized spectra were used to evaluate the linear dichroic ratios and molecular orientation with respect to the membrane normal, as described (13, 23). Simulation of second derivatives was accompanied with 13-point Savitzky-Golay smoothing in order to avoid amplification of high frequency noise. Amide I components were obtained by curve fitting, using the number and positions of the inverted second derivative peaks as input parameters. The results of curve fitting were considered satisfactory when the peak frequencies of all components were within a ±1 cm^-1 range compared with the input frequencies and when the sum of all components (the "curve fit") was practically identical to the measured spectrum.

View larger version (14K): [in this window] [in a new window]   FIGURE 2. Initial structure-function characterization of the chimeric N10IIA/IB PLA2. A, fluorescence spectrum of the semisynthetic chimeric PLA2 in a buffer containing 100 mM NaCl, 1 mM NaN3, and 50 mM Hepes (pH 7.4), in a 0.4-cm optical path length quartz cuvette. Excitation was at 275 nm, and the emission peak was located at 308 nm, characteristic of tyrosine fluorescence of a folded, tryptophan-less protein. B, circular dichroism spectra of the recombinant hIBPLA2 (dotted line), hIIAPLA2 (dashed line), and of the semisynthetic chimeric PLA2 (continuous line) in a buffer containing 10 mM sodium phosphate buffer, pH 7.4, in a 0.5-mm optical path length quartz cuvette. All three proteins contain 40% -helix, as judged from the ellipticity at 222 nm, but the shape of the spectrum of the chimeric protein suggests more rigid helices compared with those in hIBPLA2 and hIIAPLA2. C, activities of the recombinant hIBPLA2 (dotted line), hIIAPLA2 (dashed line), and the semisynthetic chimeric PLA2 (continuous line) as measured in a buffer containing 100 mM KCl, 10 mM CaCl2, 0.3 mM Triton X-100, 1 mg/ml bovine serum albumin, 25 mM Tris-HCl (pH 7.5), 0.4 mM 5,5'-dithio-bis-(2-nitrobenzoic acid), and 1.5 mM DHTPC as a substrate. PLA2s were added to a final concentration of 1.0 µg/ml, followed by stirring for 10-15 s and recording the absorption at 414 nm. The method is based on PLA2-catalyzed cleavage of the sn-2 thioester bond of DHTPC, exposure of free thiol groups that cause monomerization of 5,5'-dithiobis(2-nitrobenzoic acid) resulting in characteristic absorption at 414 nm. The initial slopes of time dependence of absorption were used to estimate activities of PLA2s. All measurements were conducted at 22 °C. See text for more details.

An ideal way to position precisely the protein at the membrane surface is to determine the coordinates of the protein atoms in a "membrane" coordinate system, _m, with axes x, y, and z, which has its xy plane in the membrane center, between the two lipid leaflets, and the z axis is perpendicular to the membrane surface. This was achieved as follows. The protein molecule was considered at an arbitrary rotation about the z axis, maintaining the angles between the z axis and the three axes of the protein system _p, as determined above. This does not reduce the generality because the protein has an unrestricted rotational freedom about the z axis. This allows determination of all nine angles between the protein and membrane coordinate systems, using the laws of direction cosines, which in turn allows transformation of the protein atom coordinates from the system _p to the system _m.

	RESULTS AND DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

 
Initial Structure-Function Characterization of the Chimeric N10IIA/IB PLA₂—The initial structure-activity characterization of the chimeric N10IIA/IB PLA₂ involved fluorescence and CD experiments and an activity assay. The chimeric protein is devoid of tryptophans but contains nine tyrosines, hence its fluorescence spectrum was measured based on tyrosine emission. As shown in Fig. 2A, the protein had a fluorescence emission maximum around 308 nm when excited at 275 nm, which is close to the tyrosine emission peak at 306 nm in folded proteins (13, 25, 26). Because fluorescence spectra provide little information on the secondary structure of proteins, we used CD spectroscopy to evaluate the secondary structure content in the chimeric N10IIA/IB PLA₂ and to compare it with that of the group IB and IIA PLA₂s. The CD spectrum of the N10IIA/IB PLA₂ demonstrated double minima around 208 and 222 nm, with an absolute value of the mean residue molar ellipticity at 222 nm ([]₂₂₂) comparable with those measured for group IB and IIA PLA₂s (Fig. 2B). This indicates that the chimeric protein is folded into a structure that comprises 40% -helix, which is characteristic of group I/II PLA₂s (4, 5). The shape of the CD spectrum of the chimeric protein was different from those of the group IB and IIA PLA₂s, however. The CD spectrum of the N10IIA/IB PLA₂ exhibited an increased ellipticity ratio []₂₀₈/[]₂₂₂ as compared with both hIBPLA₂ and hIIAPLA₂. Reduced values of the ellipticity ratio []₂₀₈/[]₂₂₂ are associated with deviations of the -helix structure from a standard, rigid helix geometry, e.g. by formation of coiled-coil structures (27, 28) or distorted helices (29). Although two internal -helices of group I/II PLA₂s are nearly antiparallel and disulfide-bonded, they are not likely to form a coiled-coil structure (see below). The observed feature is therefore more likely to indicate a less flexible conformation of the chimeric PLA₂ compared with the hIBPLA₂ and hIIAPLA₂.

To assess the influence of the substitution of the N-terminal helix on PLA₂ activity, the activity of the chimeric PLA₂, as well as of hIBPLA₂ and hIIAPLA₂, was measured by using DHTPC micelles as substrate. The activity of the chimeric enzyme, as evaluated based on the initial slope of the time dependence of product accumulation, was v_o = 9 ± 2 µmol/min/mg (Fig. 2C), which is considerably lower than the activities of hIBPLA₂ and hIIAPLA₂, 24 ± 5 and 52 ± 8 µmol/min/mg, respectively (shown are mean values and standard deviations from three experiments). Although hIIAPLA₂ shows little activity toward zwitterionic phosphatidylcholine membranes (30, 31), it is more active than hIBPLA₂ toward micelles of short chain DHTPC. A possible explanation of this is that the extremely high excess positive charge of hIIAPLA₂ (15 excess cationic charges at pH 7.4, see TABLE ONE) prevents binding to zwitterionic membranes because of strong positive surface charge of membranes that rapidly builds up upon binding of a few PLA₂ molecules. Micelles, on the other hand, can afford binding of only one PLA₂ molecule at a time because of their small size, which precludes electrostatic repulsion effects. These data indicate that in addition to structural differences that are reflected in CD spectra, the chimeric enzyme has a specific activity severalfold lower compared with those of group IB and IIA enzymes.

During the first 90 min of HX, the high frequency wing of the amide I band of membrane-bound, segmentally ¹³C-labeled hIBPLA₂ shifts toward lower frequencies by 20 cm^-1 because of HX, whereas the low frequency wing undergoes a more moderate downshift (Fig. 3A). The second derivative spectra, which were used as a resolution enhancement tool, reveal at least 10 components between 1690 and 1570 cm^-1 (Fig. 3B). A well defined amide I component at 1658 cm^-1 is readily assigned to the unlabeled N-terminal -helix of hIBPLA₂, whereas the components at 1619 and 1609 cm^-1 are assigned to the unexchanged and exchanged ¹³C-labeled helices of the protein, respectively (23, 32-38). The component at 1658 cm^-1 undergoes slow HX, because it is only at 15 min following exposure to ²H₂O that spectral downshift of this component clearly manifests itself (Fig. 3B). On the other hand, the signal assigned to the ¹³C-labeled helices 2 and 3 is split into components at 1619 and 1609 cm^-1, already at 1 min of exposure to ²H₂O, and undergoes little spectral change thereafter. As seen from the unprocessed amide I spectra (Fig. 3A), during the course of HX absorption intensity migrates from the high to the low frequency region, which may partially compensate the decrease in the 1619 cm^-1 component during HX. These spectral features thus may indicate an increased stability of the N-terminal -helix of the hIBPLA₂ compared with the two internal helices.

The amide I bands of the membrane-bound segmentally ¹³C-labeled chimeric N10IIA/IB PLA₂ are different from those of the hIBPLA₂ in terms of both the spectral line shape and the dynamics. The high frequency wing of the amide I band undergoes only an 8 cm^-1 downshift in 90 min, and despite this, the peak of the amide I band is located at 1608-1604 cm^-1 (Fig. 3C), which is 12 cm^-1 lower than the amide I peak of the hIBPLA₂ (Fig. 3A). These characteristics reflect the fact that the component corresponding to the N-terminal helix of the chimeric protein is still located at 1657 cm^-1, but the component corresponding to the ¹³C-labeled helices 2 and 3 is shifted farther down to 1608-1604 cm^-1, as seen from the second derivative spectra of Fig. 3D. Moderate spectral shifts for a protein in ²H₂O indicates an overall more rigid structure. The more rigid conformation of the chimeric PLA₂ agrees with the above conclusion deduced from CD data (Fig. 2B). Although both unlabeled and ¹³C-labeled helices undergo gradual HX, resulting in time-dependent diminution of corresponding signals in the second derivative spectra (Fig. 3D), at 60-90 min following exposure to ²H₂O there still is a major component around 1604 cm^-1, whereas the component around 1657 cm^-1 nearly vanishes. This may result from faster HX of the N-terminal helix compared with the ¹³C-labeled helices, which, in contrast to hIBPLA₂, is indicative of more stable internal helices of the chimeric PLA₂ compared with the N-terminal helix.

A possible interpretation of lower amide I frequency of the ¹³C-labeled -helices of the chimeric PLA₂ compared with hIBPLA₂ might be the formation of a double-stranded coiled-coil by the two internal helices of the chimeric PLA₂ and sequestered, rigid helices in the hIBPLA₂, because two-stranded coiled-coils are characterized by significantly lower amide I frequencies compared with standard -helices (39, 40). However, this is not likely to be the case because the internal helices of group I/II PLA₂s are not long enough and lack heptad repeat features, which facilitate helical coiled-coil formation (41). A more substantiated interpretation is based on consideration of the relative amide I intensity and the spectral location of the ¹³C-helices. The ratios of apparent extinction coefficient of ¹³C-labeled and unlabeled helices of both the hIBPLA₂ and the chimeric PLA₂ were estimated as _13C/_12C = I_13Cn_12C/I_12Cn_13C, where I and n are the integrated amide I intensities and the numbers of amino acid residues for corresponding helices, respectively. The corrected, polarization-independent ATR-FTIR spectra were used for such analysis, which indicated that _13C/_12C = 0.54 for hIBPLA₂ and _13C/_12C = 0.67 for the chimeric protein. Increased absorptivity of the amide I mode of unlabeled versus ¹³C-labeled helices was detected before for synthetic 32-mer peptides containing stretches of three consecutive ¹³C-labeled residues (38). Our data indicate that the amide I absorptivity of the ¹³C-labeled helices relative to the unlabeled helix is higher in the chimeric PLA₂ as compared with the hIBPLA₂. On the other hand, the amide I peak of ¹³C-labeled -helices of the chimeric protein occurs at lower frequencies compared with the hIBPLA₂. Because several studies have indicated that temperature-induced destabilization of segmentally ¹³C-labeled -helical peptides was accompanied with a decrease in intensity and an increase in the frequency of both unlabeled and ¹³C-labeled segments (38, 42, 43), these spectral features may be interpreted in terms of a more rigid structure for the internal helices of the chimeric protein compared with the IB PLA₂. In support of this conclusion, it is worth mentioning that _II-helices, which are characterized by much weaker helical H-bonding and are more flexible than standard _I-helices (44), generate amide I modes at 10-12 cm^-1 higher frequencies compared with regular -helices (45-47).

View larger version (58K): [in this window] [in a new window]   FIGURE 3. Structural characterization of the semisynthetic, segmentally 13C-labeled hIBPLA2 and the chimeric N10IIA/IB PLA2 by ATR-FTIR spectroscopy. A, ATR-FTIR amide I bands of the semisynthetic, segmentally 13C-labeled hIBPLA2 bound to a phospholipid bilayer membrane supported on a germanium internal reflection plate. The membrane was composed of POPC in the lower leaflet (facing the plate) and POPC/POPG (4:1, mol/mol) in the upper leaflet. Initially the protein was dissolved in a buffer containing 100 mM NaCl, 1 mM NaN3, 1 mM EGTA, 50 mM Hepes in H2O (pH 7.4) and was injected to the ATR cell with the supported membrane. After binding of the protein to the membrane, the cell was flushed with a 2H2O-based buffer of the same composition, and spectra were recorded at parallel and perpendicular orientations of the infrared light. Presented here are the corrected and normalized spectra. They were obtained as Acorrected = A|| + 0.8A , which are "polarization-independent" and therefore reflect the true secondary structural features of the protein, followed by normalization of the total integrated area to 1 absorbance unit/cm. The stacked spectra correspond to exposure of the protein to 2H2O for 1, 3, 7.5, 26, 44, 66, and 90 min (from top to bottom), which applies to all four panels. B, the second derivative amide I spectra derived from the spectra of A, accompanied with 13-point Savitzky-Golay smoothing. Here, as well as in D, the wave numbers of prominent amide I components are marked by vertically oriented numbers. C, ATR-FTIR spectra of the semisynthetic, segmentally 13C-labeled chimeric N10IIA/IB PLA2 bound to a supported membrane under identical conditions as in A. D, the second derivative spectra derived from spectra of C, accompanied with 13-point Savitzky-Golay smoothing.

Membrane Binding of Group IB and IIA PLA₂s, Their N-terminal Peptides, and the Chimeric PLA₂—Membrane binding is a crucial determinant of PLA₂ activity. Because in this work we consider the structural and functional consequences of the substitution of the N-terminal helix of hIBPLA₂ by that of hIIAPLA₂, we have determined and compared membrane binding parameters of hIBPLA₂, hIIAPLA₂, their N-terminal peptides, and the chimeric N10IIA/IB PLA₂. Fig. 4A shows representative fluorescence spectra indicating a gradual increase in the emission intensity of FPE (2 mol % in membranes of large unilamellar vesicles) with increasing concentration of hIBPLA₂. The increments of fluorescence change at each protein concentration, F, were corrected by taking into account the decrease in fluorescence intensity due to sample dilution (Fig. 4B) and were used to construct the binding isotherms (Fig. 4C), as described under "Experimental Procedures." The binding isotherms were described by Equations 3 and 4, and the dissociation constants (K_D) and the numbers of lipids per membrane-bound protein (N) were evaluated from the best fit between the experimental and simulated isotherms.

There is strong evidence that membrane binding of group I and II PLA₂s has a strong electrostatic component, resulting in poor binding to zwitterionic PC membranes and tight binding to anionic membranes (48-50). For example, binding of human or snake group IIA PLA₂s to phospholipid membranes could only be detected in the presence of 15 mol % acidic lipid in membranes, such as DPPG or POPG (49, 50). Correspondingly, the activity of hIIAPLA₂ against phosphatidylcholine vesicles is 1000-10,000-fold lower than against phosphatidylglycerol vesicles (30, 49, 51). The dissociation constant of group IIA PLA₂ for zwitterionic membranes is in the millimolar range (48, 50, 52). Mammalian group IB PLA₂s also exhibit poor binding to zwitterionic membranes and low activity compared with membranes containing anionic lipids (30). Involvement of a large electrostatic factor in membrane binding of group I and II PLA₂s is a physiologically intrinsic feature of these enzymes. Our binding experiments indicated that not only group IB and IIA PLA₂s but also their N-terminal peptides did not bind to pure phosphatidylcholine membranes but did bind to membranes containing acidic lipid. Given all these circumstances, and the fact that most biological membranes have surface charge corresponding to 20 ± 5% anionic lipids (53, 54), we used membranes containing 20% anionic lipid in binding experiments. We also used membranes containing 40% anionic lipid in order to determine the sensitivity of membrane binding of all five molecules to the membrane surface charge (binding isotherms are only shown for vesicles with 40% POPG in Fig. 4C). Parameters K_D and N for the binding of all five molecules to membranes containing 20 and 40 mol % POPG are summarized in TABLE ONE. The data indicate that hIIAPLA₂ and the N10hIB peptide exhibit tighter binding to membranes containing 20 mol % POPG (K_D = 2.5-2.8 µM) and 40 mol % POPG (K_D = 0.3-0.6 µM) compared with the other molecules. The membrane binding affinity of hIIAPLA₂ (in terms of K_D) undergoes an 10-fold increase upon increasing the content of negative lipid in membranes from 20 to 40 mol %. The sensitivity of membrane binding affinity to the membrane negative surface charge decreases in the sequence: hIIAPLA₂ > N10IIA/IB hIBPLA₂ N10hIB > N10hIIA. In order to interpret these characteristics, the pI values and the excess charge of the molecules, as shown in TABLE ONE, should be taken into account. There is some, but inconsistent, correlation between the affinities for anionic membranes and pI values of the molecules. The N10hIB peptide, which has three cationic and no anionic residues (the C-terminal carboxyl group is suggested to be compensated by the N-terminal amino group), has the highest pI value and relatively high membrane binding affinity but lower sensitivity to the membrane anionic charge than the hIIAPLA₂ and the chimeric protein. It is very likely that the binding of proteins and peptides to anionic membranes is determined not only by the pI values but, perhaps more importantly, by the actual excess cationic charge of the molecule. For example, the hIIAPLA₂ has much higher excess positive charge than any of the molecules studied, which results in strong binding to anionic membranes and high sensitivity to the membrane surface charge. The chimeric protein, on the other hand, has only 1.2 excess cationic charge at pH 7.4, resulting in weaker membrane binding, especially at 20 mol % of acidic lipid in membranes, and lower sensitivity to the membrane anionic charge.

View larger version (17K): [in this window] [in a new window]   FIGURE 4. Quantitative characterization of binding of hIBPLA2, hIIAPLA2, their N-terminal peptides, and the chimeric N10IIA/IB PLA2 to phospholipid membranes. A, dependence of fluorescence emission spectra of FPE, incorporated at 2 mol % in large unilamellar vesicles (100 nm in diameter) containing 58 mol % POPC and 40 mol % POPG, on the concentration of hIBPLA2. The increasing darkness of lines corresponds to increasing concentrations of hIBPLA2, which are indicated in C. The buffer contained 10 mM Hepes, 1 mM NaN3, 1 mM EGTA (pH 7.4). Excitation was at 490 nm. B, control experiments, in which buffer was added instead of PLA2 to FPE-containing vesicles in order to determine the effect of sample dilution, which was used for correction of spectra in A. The increasing darkness of spectra corresponds to increasing sample dilution. Fluorescence spectra were measured using a 0.4-cm optical path length quartz cuvette. C, membrane binding isotherms for hIBPLA2 (open circles), N10hIB peptide (closed circles), hIIAPLA2, (open squares), N10hIIA peptide (closed squares), and the chimeric N10IIA/IB PLA2 (triangles). Lipid composition of vesicles and the buffer are as in A. The relative increase in peak fluorescence intensity of FPE, F/Fmax, is plotted against protein concentration. The theoretical isotherms are simulated through Equations 3 and 4, using the binding parameters (KD and N) summarized in TABLE ONE; all lines are continuous except for that describing the binding of the chimeric PLA2 (broken line).

Positioning the Chimeric PLA₂ at the Membrane Surface—Previously, we have determined the precise positioning of hIBPLA₂ on phospholipid membranes (18), and we have shown that the N-terminal -helix is a crucial determinant for the productive mode membrane binding of the enzyme (13). In order to understand the effect of the substitution of the N-terminal -helix on the membrane binding mode of PLA₂, here we have studied the angular orientation of the membrane-bound chimeric N10IIA/IB PLA₂. Because our task was to achieve the structure of the membrane-bound protein at atomic details, first we determined the structure of the chimeric PLA₂ by the internet-based server Swiss-Model (24), using the structure of the porcine pancreatic PLA₂ as a template. The obtained structure revealed three -helices. The C- atom coordinates of the three helices were used to determine the interhelical angles, ₁₂ = 35.5°, ₁₃ = 36.6°, and ₂₃ = 6.2°, where _ij is the angle between helices i and j. The fact that the helices 2 and 3 are nearly parallel to each other allowed us to use a common order parameter for these two helices in terms of their spatial orientation.

View larger version (22K): [in this window] [in a new window]   FIGURE 5. Determination of the angular orientation of the membrane-bound segmentally 13C-labeled chimeric N10IIA/IB PLA2 by polarized ATR-FTIR spectroscopy. The composition of the supported phospholipid bilayer and the method of sample preparation are the same as in Fig. 3. The segmentally 13C-labeled N10IIA/IB PLA2 was dissolved in an H2O-based buffer of 100 mM NaCl, 1 mM NaN3, 1 mM EGTA, 50 mM Hepes (pH 7.4) and injected into the cell containing the supported membrane. After allowing the protein to bind to the membrane (15 min), the cell was flushed with a similar buffer prepared in 2H2O, and ATR-FTIR spectra were recorded at parallel (A) and perpendicular (B) polarizations of the infrared light. Amide I components were generated by curve fitting, using the second derivatives for resolution enhancement. Amide I components corresponding to the N-terminal unlabeled -helix (dotted line) and the two 13C-labeled -helices (dashed line) are located at 1657-1654 and at 1608-1604 cm-1, respectively. The sum of all components is shown by dashed-dotted line, which nearly coincides with the measured amide I contour.

The intensities of polarized amide I components for the N-terminal helix and for the ¹³C-labeled helices 2 and 3, which were spectrally resolved because of an 50 cm^-1 downshift of the signal from the ¹³C-labeled helices, were used to evaluate the respective dichroic ratios as R₁ = 1.45 ± 0.08 and R_2,3 = 1.35 ± 0.05, where the subscripts indicate the corresponding -helices (shown are mean values and standard deviations obtained from curve fitting of several polarized amide I spectra). The mean values of dichroic ratios yield the order parameters of S₁ = -0.268 for the helix 1 and S_2,3 = -0.392 for the helices 2 and 3, with ranges of variation from -0.183 to -0.368 and from -0.331 to -0.457, respectively. Interpretation of these uncertainties in terms of angular orientation indicates that the amplitude of wobbling of the helices relative to the membrane normal does not exceed 10°, which is reasonable regarding the fluid nature of the membrane. Thus, the ATR-FTIR measurements provide the helical orientations with respect to the membrane with an acceptable degree of accuracy. The order parameters were used to find the average tilt angles of the respective helix axis relative to the membrane normal, cos₁ = 0.394 and cos_2,3 = 0.268. At this point, we had the orientations of all helices in the protein coordinate system, _p, and their orientations relative to the membrane normal (the z axis). This allowed expression of cos₁ and cos_2,3 through the cosines of angles between helical axes and the axes of the protein system _pand the cosines of angles between the z axis and the axes of the coordinate system _p, cos(zx^*), cos(zy^*), and cos(zz^*), assuming these two coordinate systems are brought to a common origin (see Ref. 18 for details). Combination of these with the law of direction cosines: cos²(zx^*) + cos²(zy^*) + cos²(zz^*) = 1, allowed determination of cos(zx^*) =-0.392, cos(zy^*) =-0.158, and cos(zz^*) = 0.906. These three cosines determine the angular orientation of the protein relative to the membrane normal.

In order to determine the structure of the membrane-bound protein, i.e. the protein atom coordinates in the membrane coordinate system _m, the remaining six angles between the axes of the coordinate systems _m and _p are also required. Because the membrane is only anisotropic in the vertical dimension (the z axis) but is isotropic in the lateral dimension (the x and y axes), the protein has a rotational freedom about the z axis but not the x or the y axes. Therefore, all azimuthal orientations of the protein, i.e. all angles of rotation about the z axis, are equivalent. This allowed us to select arbitrarily one of the six angles, e.g. cos(yx^*) = 0.456, and use the laws of direction cosines to determine the remaining angles. Thus, we obtained all nine angles between the axes of the coordinate systems _m and _p: cos(zx^*) = -0.392, cos(zy^*) = -0.158, and cos(zz^*) = 0.906, cos(yx^*) = 0.456, cos(yy^*) = 0.822, cos(yz^*) = 0.341, cos(xx^*) = -0.799, cos(xy^*) = 0.547, cos(xz^*) = -0.250. These nine cosines were used to transform the protein atom coordinates from the protein system _p to the membrane system _m, assuming the two systems have a common origin.

The above operations only provide the angular orientation of the protein relative to the membrane. The vertical location of the protein relative to the membrane plane is still required to position the protein on the membrane. This can be determined by membrane depth-dependent fluorescence quenching of a protein-bound fluorophore by lipids that contain quenchers at various positions of hydrocarbon chains. We have determined previously the depth of insertion of hIBPLA₂ into POPC/POPG membranes by this method (18). Although hIBPLA₂ is an appropriate protein for this purpose because it has a single tryptophan at position 3, which is a part of the membrane binding face, this is not the case with the chimeric N10IIA/IB PLA₂ because this molecule is devoid of tryptophans. Although other native fluorophores, such as tyrosines, can also be used for quenching experiments, the abundance of tyrosines in N10IIA/IB PLA₂ prevents any meaningful interpretation of the fluorescence quenching data in terms of the depth of membrane insertion. Therefore, we used the information regarding membrane insertion of hIBPLA₂ in order to locate the chimeric N10IIA/IB PLA₂ along the membrane normal. For the hIBPLA₂, the geometric center of the indole ring of Trp³ was located at 9 Å from the membrane center (18). By assuming insertion of the chimeric N10IIA/IB PLA₂ into the membrane to a similar depth, and by taking into consideration the difference in size between Trp and Val, we positioned the side chain of Val³ of the chimeric protein at 10 Å from the membrane center. This was done by adjusting the z coordinates of all atoms of the protein so the z coordinate of the geometric center of Val³ was 10 Å, whereas the angular orientation of the protein with respect to the membrane was maintained as determined above. These procedures resulted in a model for the structure of the membrane-bound chimeric N10IIA/IB PLA₂, which is shown in Fig. 6A. In Fig. 6, A and B, three planes of atoms are introduced that are perpendicular to the membrane normal and schematically identify the locations of terminal methyl carbons of the acyl chains of glycerophospholipids (membrane center), the sn-1 carbonyl oxygens, and the phosphorus atoms of lipid phosphate groups. The z coordinates of these three planes are z = 0, z = 14.5 Å, and z = 20 Å, respectively, which is based on x-ray diffraction data on POPC bilayers (55).

The structure of the membrane-bound N10IIA/IB PLA₂, which we refer to as the quinary structure of a membrane protein (18), is consistent with previously reported models of membrane anchoring of the group I/II PLA₂s (8, 56-58), with one important difference, i.e. the depth of membrane insertion of the protein. Although penetration of secretory PLA₂s into membranes has not been reported by others, we have documented insertion of hIBPLA₂ into POPC/POPG membrane to a depth of 9 Å from the membrane center (18), which is the basis for positioning the chimeric PLA₂ relative to the membrane surface as shown in Fig. 6. The mode of membrane binding of the chimeric PLA₂ resembles that of the hIBPLA₂ (18). In both cases, the -helices 2 and 3 are tilted by 75° from the membrane normal toward the membrane surface. The N-terminal -helix is oriented obliquely relative to the membrane plane and is by 7° more tangential in the chimeric PLA₂ than in hIBPLA₂. As shown below, the membrane-bound N10hIIA peptide also turned out to be oriented more laterally on the membrane surface compared with the N10hIB peptide, implying that the orientation of the membrane-bound PLA₂ is probably affected by the intrinsic properties of the N-terminal helix. Membrane binding of the protein is stabilized by hydrophobic, ionic, and H-bonding interactions. In Fig. 6A, the cationic side chains of Arg⁷, Lys¹⁰, and Lys¹¹⁶ are located at the level of the hydrophobic/polar interface of the membrane and are very likely to be involved in H-bonding with lipid carbonyl oxygens. Side chains of other three lysines (residues 113, 121, and 122) are at the level of lipid phosphate groups and are likely engaged in ionic contacts with those groups. Side chains of three nonpolar residues, Val³, Phe¹⁹, and Leu²⁰, act as hydrophobic anchors for the PLA₂-membrane interaction. In this particular orientation, a hydrophobic spot of the protein, marked by Leu⁶⁴ and Leu⁶⁵, is in the polar region of the ester carbonyl groups of membrane lipids. Because the model of Fig. 6A presents an enzyme bound to a membrane in the fluid phase rather than a rigid, immobile structure, the protein is likely to have some degree of motional flexibility, including the angular orientation. The above estimates of the helical order parameters, which are the basis for orientation determination, indicated that within the accuracy of the measurements the orientation of PLA₂ can fluctuate within the limits of 10°. Rotation of PLA₂ by 10° about an axis parallel to the x axis and passing through the center of the protein molecule results in a structure presented in Fig. 6B. This seems to be a viable structure because it is still close to the limits of experimental error and maintains the characteristic features of the structure obtained using the average values of helical order parameters (Fig. 6A). In this structure, the side chains of Leu⁶⁴ and Leu⁶⁵ are conveniently embedded in the hydrocarbon chain region of the membrane, which stabilizes the membrane binding of PLA₂. This is in accord with earlier indications that the loop containing these leucines is a part of interfacial binding region of group IB PLA₂s (9, 56) that is involved in conformational changes in the enzyme upon interfacial activation (8). On the other hand, in this model Lys¹²¹ and Lys¹²², which form ionic bonds with the lipid phosphate groups in the structure of Fig. 6A, are moved farther away from the membrane surface. Also, the residues Lys¹¹³ and Lys¹¹⁶ are located in a less energetically favorable manner than in the model of Fig. 6A. These considerations lead to an inference that the actual membrane binding mode of the protein is likely to be between the two models presented in Fig. 6. It is also possible that PLA₂ undergoes transitions between distinct orientations during the catalytic process.

View larger version (22K): [in this window] [in a new window]   FIGURE 6. Model for membrane-bound chimeric N10IIA/IB PLA2 obtained by homology modeling, polarized ATR-FTIR spectroscopy, and analytical geometry algorithms. The structure in A was obtained by using the mean values of dichroic ratios of the unlabeled N-terminal -helix and of the two 13C-labeled -helices. The structure of PLA2 was homology-modeled by Swiss-Model using porcine pancreatic PLA2 structure (Protein Data Bank entry 1P2P) as a template. The protein is presented in a ribbon format, with the -helices 1-3 colored yellow, purple, and light blue, respectively. The side chains of amino acids directly involved in physical interactions with the membrane are shown in a stick format; hydrophobic residues are shown in green and the cationic residues in blue. The catalytic His48 is shown in ball-and-stick format and colored rose. The three layers of nonprotein atoms are introduced to schematically show membrane sections corresponding to the acyl chain terminal methyl carbons (gray), the sn-1 carbonyl oxygens (red), and the phosphorus atoms (orange) of membrane glycerophospholipids. The z-coordinates of these layers are 0, 14.5, and 20 Å, respectively. The whole structure is turned about the z axis by 10° and slightly (1°) about the y axis. Without these rotations, the plane of the picture would correspond to the yz plane. B, the protein has been rotated by 10° about an axis parallel to the x axis, which results in immersion of Leu64 and Leu65 into the hydrophobic part of the membrane.

,||

,

,||

,

,||

,