Sterol carrier protein-2, a new fatty acyl coenzyme A-binding protein.

The ability of sterol carrier protein-2 (SCP-2) to interact with long chain fatty acyl-CoAs was examined. SCP-2 bound fluorescent fatty acyl-CoAs at a single site with high affinity. Kd values for cis- and trans-parinaroyl-CoA were 4.5 and 2.8 nM, respectively. Saturated 10-18-carbon and unsaturated 14-20-carbon fatty acyl-CoAs displaced SCP-2-bound fluorescent ligand. Oleoyl-CoA and oleic acid (but not coenzyme A) significantly altered SCP-2 Trp50 emission and anisotropy decay, thereby increasing SCP-2 rotational correlation time, SCP-2 hydrodynamic radius, and SCP-2 Trp50 remaining anisotropy up to 1.7-, 1.2-, and 1.3-fold, respectively. These changes were not accompanied by significant alterations in protein secondary structure as determined by circular dichroism. Finally, SCP-2 differentially altered the fluorescence emission and anisotropy decays of bound cis- and trans-parinaroyl-CoA. Both fluorescent fatty acyl-CoAs were located within a very ordered (limited cone angle of rotation) environment within SCP-2, as shown by a remaining anisotropy of 0.365 and 0.361 and a wobbling cone angle of 12 and 13°, respectively. These anisotropy values were very close to those of such ligands in a propylene glass. However, the rotational relaxation times exhibited by SCP-2-bound cis- and trans-parinaroyl-CoA, 8.4-8.8 ns, were longer than those for the corresponding free fatty acid, 7.5-6.6 ns. These data show for the first time that SCP-2 is a fatty acyl-CoA-binding protein.

It was recognized over 20 years ago that cytosolic fatty acidbinding proteins, representing 2-6% of cytosolic protein, can bind fatty acyl-CoAs (reviewed in Ref. 1). More than 17 members of this family have thus far been identified, with most of them having low micromolar affinities for fatty acyl-CoAs. More recently a totally different fatty acyl-CoA-binding protein (ACBP), 1 unrelated to the fatty acid-binding protein superfamily, was reported (8). ACBP has higher affinity for fatty acyl-CoAs than the fatty acid-binding proteins (9,10).
In contrast to the ACBP, the cellular sterol carrier protein-2 (SCP-2) has long been associated with the metabolism of lipids, especially cholesterol (reviewed in Ref. 11). Like ACBP, SCP-2 is also a ubiquitous protein, identical in all tissues examined (12). Furthermore, SCP-2 appears localized in subcellular organelles intimately involved with fatty acid oxidation, the peroxisomes, and mitochondria, with lesser amounts in the cytosol and near endoplasmic reticulum (13). The recent discovery that SCP-2 can bind fatty acids with K d values near 0.2 M (14) led us to examine whether SCP-2 could also interact with fatty acyl-CoAs. We show herein for the first time that SCP-2 is a high affinity fatty acyl-CoA-binding protein with K d values near 2-4 nM. The chain length and unsaturation specificity of fatty acyl-CoAs was also examined. Furthermore, we report the effect of fatty acyl-CoAs and fatty acids on the rotational dynamics and structure of SCP-2 as ascertained by phaseresolved fluorescence.
Time-resolved Fluorescence Spectroscopy-Time-resolved fluorescence measurements were performed on a GREG 250 subnanosecond multifrequency cross-correlation and modulation fluorometer equipped with KOALA automatic sample compartment (ISS Instruments) with a Glan-Taylor polarizer in the excitation channel. Excitation was with an Innova-Sabre argon ion laser (Coherent Laser Group, Palo Alto, CA) with automatic wavelength selection of single selectable lines at ϭ 275.4, 300.2, 302.4, and 305.5 nm (power output of 340, 630, 800, and 460 milliwatts, respectively). Laser output was kept stable over several hours in either "Power Track" or the "Light Regulation" mode. In order to avoid scattered emission and Raman scatter artifacts, protein fluorescence emission was observed with a 0.3-cm quartz cuvette (Expotech Inc., Houston, TX) through a model 341 BP15 interference filter (Omega Optical Inc., Brattleboro, VT) with a transmittance maximum at 341 nm (bandwidth of 15 nm). Excitation intensity was optimized to minimize protein photobleaching (typically Յ10%) while maximizing high signal:noise ratio. cis-Parinaric acid and cis-parinaroyl-CoA were excited at ϭ 325 nm by a 4240NB He-Cd laser (Liconix, Sunnyvale, CA), and emission was observed through KV389 low fluorescent cut-off filters (Schott Glass Technologies Inc., Duryea, PA). Sample absorbance at the excitation wavelengths was Յ0.05. All data were obtained in 25 mM phosphate buffer, pH 7.4, at 25°C.
Fluorescence Lifetime Data Acquisition and Analysis-Fluorescence decay kinetics were measured with the excitation polarizer oriented at 35°to yield rotation-free results (18). Lifetime data were acquired at 12-15 modulation frequencies (20 -225 MHz) using external lifetime standards for protein (p-terphenyl, ϭ 1.05 ns) and cis-parinaric acid (dimethyl-POPOP, ϭ 1.45 ns) in absolute ethanol. Data were acquired until the limit of standard error (phase, 0.2°; modulation, 0.004) and analyzed by ISS-187 software (ISS Instruments) as a sum of exponentials: The minimized 2 parameter was the criterion for goodness of fit to the model, with 2 Ͻ 3 acceptable (19).
Anisotropy Decay Measurement and Analysis-Time-resolved anisotropy data were obtained at 12-15 modulation frequencies for both I ʈ and I Ќ components over the range of 20 -225 MHz with Glan-Taylor and Glan-Thompson polarizers in the excitation and emission channels, respectively. Standard error limits were 0.2, 0.004, and 0.01°for phase, modulation, and polarization. Anisotropy decay was modeled by a sum of exponentials as: r(t) ϭ r(0)⌺g i exp(Ϫt/ i ), where r(0) is anisotropy of a fluorophore in the absence of rotational diffusion, i is the rotational correlation time, and g i is fractional anisotropy. The goodness of fit to the applied model was determined as described above using ISS-187 software (ISS Instruments). The equivalent hydrodynamic radius of the protein was calculated as follows, where is solvent viscosity. The value of the same parameter can be estimated from the hydrated protein volume, where M is molecular weight, N 0 is Avogadro's number, ␦ 1 is fraction of hydration, V 1 is volume of bound water, and V 2 is protein specific partial volume. ␦ 1 ϭ 0.4 g of H 2 O/g of protein; V 2 ϭ 0.72 cm Ϫ3 g Ϫ1 for an average protein (20); and V 1 ϭ 1 cm 3 g Ϫ1 for water in Equation 3. Circular Dichroism-Spectra of SCP-2 (2-4 M, 25 mM phosphate buffer) were recorded (1-mm circular cuvette, 1-nm resolution) on a J-710 spectropolarimeter (JASCO Inc., Easton, MD) (bandwidth 2 nm, sensitivity 10 millidegrees, scan rate 50 nm/min, time constant 1 s) at 23°C. Ten scans were averaged for structure analysis (SELCON) (21) with software from JASCO Inc.
Fluorescent Fatty Acid and Fatty Acyl-CoA Binding-All K d values and stoichiometries of SCP-2 binding parinarates or parinaroyl-CoAs were determined (22) at 24°C with 2-ml samples of 0.18 M SCP-2 (25 mM phosphate buffer) titrated with small increments of fatty acids (1-2 l) dissolved in 10 mM NaOH. We replaced ethanol with NaOH to avoid interference with ligand binding to SCP-2 (14). Fatty acids were added from a 150 M stock solution in 10 mM NaOH. Alternately, fatty acyl-CoAs were added from a 270 M stock solution in phosphate buffer, pH 7.4. Each sample and blank (without SCP-2) were thoroughly mixed and allowed to equilibrate for 1-2 min to allow for stable measurement of the fluorescence signal. K d and binding stoichiometry (n) were calculated as follows (22,23), where F and F max are corrected measured and maximal fluorescence intensity of the ligand, respectively, C L is the total ligand concentration, and E 0 is the protein concentration. A plot of 1/(1 Ϫ F/F max ) versus C L /(F/F max ) yields a linear function with a slope of (1/K d ) and an ordinate intercept of (nE 0 )/K d . Displacement of SCP-2-bound cis-Parinaric Acid by Fatty Acyl-CoA-SCP-2 (0.18 M) in phosphate buffer (pH 7.4) was incubated with cis-parinaric acid (0.44 M) for 5 min at 24°C to obtain maximal fluorescence. SCP-2-bound cis-parinaric acid was displaced by a series of fatty acyl-CoAs with fatty acyl chain length ranging from 0 to 20. Measurements were corrected for the blank (ligand or protein only) and photobleaching (see above).

Fatty Acyl-CoA and Fatty Acid Binding Properties of SCP-2
Fatty Acyl-CoA Binding to SCP-2-The relative affinity of SCP-2 for kinked chain fatty acyl-CoA versus straight chain fatty acyl-CoA was examined with fluorescent parinaroyl-CoA derivatives. The parinaric acids are naturally occurring fatty acids oriented to form either an 18-carbon, kinked chain, fluorescent analogue (cis-parinaric acid) of oleic acid or an 18carbon, straight chain, fluorescent analogue (trans-parinaric acid) of stearic acid. Parinaroyl-CoAs fluoresce poorly in water, possibly due to strong electrostatic interactions of these fluorophores with the molecules of the solvent. Upon titration of SCP-2 with cis-or trans-parinaroyl-CoA, a progressive increase in fluorescence intensity with the maximum around 420 nm was observed (not shown), consistent with localization of this ligand in a less hydrophilic environment, i.e. binding to SCP-2. cis-Parinaroyl-or trans-parinaroyl-CoA intensities with SCP-2 measured at 420 nm, corrected for the background (no SCP-2), and plotted as a function of total ligand concentration yield pure saturation curves for both cis-and trans-parinaroyl-CoA as shown in Fig. 1, A and B, respectively. The F max parameters were obtained by fitting the saturation curves to a rectangular hyperbola. Subsequent analysis of the data in 1/(1 Ϫ F/F max ) and C L /(F/F max ) coordinates yielded a strong linear dependence for each parinaroyl-CoA (r 2 ϳ0.9) (Fig. 1, A, inset, and B, inset). The K d and n calculated thereby for cis-parinaroyl-CoA were 4.57 Ϯ 0.75 nM and 0.85 Ϯ 0.01, respectively (Fig. 1A, inset), and for trans-parinaroyl-CoA they were 2.76 Ϯ 0.10 nM and 0.71 Ϯ 0.03, respectively (Fig. 1B, inset). These data are consistent with the presence of one cis-or trans-parinaroyl-CoA binding site per SCP-2 molecule. Furthermore, the relative affinity of SCP-2 for the trans-parinaroyl-CoA was nearly 2-fold greater than for the cis-parinaroyl-CoA.
Fatty Acid Binding to SCP-2-Free fatty acid binding was examined to determine if SCP-2 fatty acyl-CoA structural specificities (straight versus kinked acyl chain) were based on the SCP-2 specificity for the fatty acyl chain structure. When SCP-2 was titrated with cis-or trans-parinaric acid, fluorescence intensity at the emission maximum (420 nm) increased without shifting the emission maximum. Binding curves were constructed as described above, and both cis-and trans-parinaric acid showed clear saturation curves (Fig. 2, A and B). The K d and n for cis-parinaric acid were 0.18 Ϯ 0.01 M and 1.13 Ϯ 0.04, respectively ( Fig. 2A, inset), similar to those reported earlier (14). In contrast, the K d and n of SCP-2 for transparinaric acid were 0.56 Ϯ 0.04 M and 0.85 Ϯ 0.02, respectively (Fig. 2B, inset). Thus, SCP-2 binds the cis-parinaric acid with a 3-fold higher affinity than trans-parinaric acid. Furthermore, SCP-2 bound cis-and trans-parinaric acids with 39-and 203-fold lower affinity than the corresponding cis-and trans-parinaroyl-CoAs.

Fluorescence Properties of Tryptophan in SCP-2⅐Fatty
Acyl-CoA Complex SCP-2 Tryptophan Intrinsic Fluorescence-SCP-2 has a single tryptophan (Trp 50 ) residue and no tyrosine (24,25). The fluorescence emission spectrum of SCP-2 upon excitation at 280 nm showed a sole maximum at 332 nm with a bandwidth (full width at half-height) of 48 nm (data not shown). Both the position of the SCP-2 Trp 50 emission peak (ϳ330 nm) and its bandwidth (ϳ50 nm) were consistent with the Trp 50 being located in a hydrophobic locus in SCP-2 (26). The addition of up to 5 molar equivalents of oleoyl-CoA to SCP-2 did not significantly alter either the fluorescence emission intensity of SCP-2 Trp 50 excited at 280 nm or the shape of the emission spectrum. The same result was obtained when SCP-2 was titrated with oleic acid or CoASH (data not shown). It should also be noted that no detectable fluorescence quenching was observed when an equivalent amount of a free tryptophan in solution (as compared with SCP-2 concentration) was titrated with the same ligands over the concentration range that was used for the titration of SCP-2 (data not shown).

Fluorescence Lifetime of Apo-and Holo-SCP-2-
The fluorescence Trp 50 in apo-SCP-2 exhibited biexponential decay (data not shown). Biexponential decay is common for proteins with a single tryptophan residue (18,27,28). Upon excitation of SCP-2 at 275 nm (emission at 341 nm) the lifetimes of the major and minor components were 2.06 (fraction 0.55) and 0.69 ns (fraction ϳ0.45). The apparent (mean) lifetime was calculated to be 1.51 ns. The addition of 5 molar equivalents of oleoyl-CoA to SCP-2 significantly increased the long lifetime component, 1 (Table II). The magnitude of this correlation time reflects the rotational diffusion of the entire protein globule (18,29). The measured rotational correlation time, 7.8 ns, was consistent with that expected for the molecular mass, 13.2 kDa, of SCP-2 (20). The SCP-2 hydrodynamic radius calculated from the data in Table  II was 20.5 Å. The same parameter calculated on the basis of the theoretical Equation 2 was 18.8 Å. The small difference between the radius determined from anisotropy data (Equation 1) and from physicochemical characteristics of a globular, spherical protein (Equation 2) is consistent with SCP-2 not being shaped as an ideal sphere but instead having a slightly elongated structure. The addition of 5 molar equivalent of oleoyl-CoA significantly increased the SCP-2 overall rotational rate by 35% and increased the hydrodynamic radius from 20.5 to 22.4 Å (Table II). Under the same conditions oleic acid also increased both the SCP-2 Trp 50 rotational correlation time from 7.8 to 13.3 ns, thereby increasing the hydrodynamic radius of SCP-2 by more than 20% from 20.5 to 24.5 Å. Coenzyme A was without effect on the rotational dynamics of SCP-2 (Table II). In general, the overall rotational dynamics of the SCP-2 protein were sensitive to the interaction with fatty acyl-CoA and fatty acid but not CoASH. The absence of a second, much faster, rotational component in the subnanosecond time frame suggests that Trp 50 in apo-SCP-2 may not undergo rapid segmental rotational motions. This is not the case, however. The fact that the measured remaining anisotropy, r(0)g 1 ϭ 0.167, is much less than that of 0.315 for the tryptophan in a propylene glycol glassy matrix at Ϫ60°C (18) indicates the possible existence of a much faster Trp 50 rotation, i.e. on a picoand/or femtosecond time scale, than measured herein. The SCP-2 Trp 50 wobbling cone angle (amplitude of Trp torsional motions) was 34°. The addition of oleic acid increased the SCP-2 Trp 50 remaining anisotropy from 0.167 to 0.215 and correspondingly decreased the wobbling cone angle from 34 to 27° (Table II). In contrast, oleoyl-CoA and CoASH were without effect on the segmental motion of SCP-2 Trp 50 as indicated by the lack of change in the remaining anisotropy and wobbling cone angle (Table II).

Fluorescence Dynamics of cis-Parinaroyl-CoA and of cis-Parinaric Acid in the SCP-2 Binding Site
Fluorescence Emission Decay-Biexponential decay was observed for SCP-2-bound cis-parinaroyl-CoA, with 1 ϭ 4.3 ns (fraction 0.42) and a mean lifetime of 3.07 ns (Table III). In contrast to cis-parinaroyl-CoA, SCP-2-bound cis-parinaric acid displayed heterogenous emission with three components where the major component 2 was 2.36 ns (fraction 0.78, Table III). The mean lifetime of the bound cis-parinaric acid, 8.07 ns, was 2.6-fold greater than that of bound cis-parinaroyl-CoA. SCP-2 bound trans-parinaroyl-CoA to SCP-2 best fit three-exponential decay (Table III), resulting in a mean lifetime of 4.45 ns. SCP-2-bound trans-parinaric acid also showed three-exponential decay (Table III), resulting in a mean lifetime of 6.64 ns. This mean lifetime of bound trans-parinaric acid was nearly 50% longer than that of bound trans-parinaroyl-CoA. Also, the The anisotropy of SCP-2 was analyzed by a sum of exponentials as r(t) ϭ r(0) ⌺g i exp(Ϫt/ i ). All measurements were done in phosphate-buffered saline (pH 7.4) at 25°C. Values represent mean Ϯ S.E. (n ϭ 3-5). a Hydrodynamic radius was calculated from 1 as described under "Experimental Procedures." b r(0) was allowed to vary. c Wobbling cone angle was calculated in degrees (deg) using cos 2 ϭ ((r/r o ) ϫ 2 ϩ 1)/3, where r is the remaining anisotropy, r ϭ r(0)g 1 . r o is the tryptophan fundamental anisotropy, r o ϭ 0.315 (18).  mean fluorescence lifetime of SCP-2-bound trans-parinaric acid was 1.2-fold less than that of SCP-2-bound cis-parinaric acid.
Time-resolved Anisotropy-SCP-2-bound cis-parinaroyl-CoA anisotropy decay exhibited single-exponential decay (Table IV). The maximal anisotropy for cis-parinaroyl-CoA bound to SCP-2, 0.365, along with its wobbling cone angle, 12°, indicated that the amplitude of cis-parinaroyl-CoA motions in the SCP-2 binding site was significantly restricted and resulted in a rotational correlation time of 8.43 ns (Table IV). The latter was similar to the overall rotational dynamics of the SCP-2 globule as observed by the close agreement with that obtained for SCP-2 Trp 50 (7.8 Ϯ 0.6 (Table II)). Likewise, anisotropy decay kinetics of cis-parinaric acid bound to SCP-2 were monoexponential (Table IV). The measured maximal anisotropy r(0) ϭ 0.335 was close to the static anisotropy of cis-parinaric acid in a propylene glycol glass, r o ϭ 0.39 (30). This fact along with the small wobbling cone angle, 18°, of the bound cis-parinaric acid is consistent with the cis-parinaric acid also being significantly immobilized in the protein matrix, resulting in a measured rotational correlation time of SCP-2-bound cis-parinaric acid of 7.51 Ϯ 0.12 ns (Table IV) that also reflected overall rotational dynamics of the SCP-2 globule. In summary, cis-parinaroyl-CoA bound to SCP-2 exhibited a significantly longer rotational correlation time (8.43 ns) than that of SCP-2-bound cis-parinaric acid (7.51 ns). The analysis of anisotropy decay for transparinaroyl-CoA and trans-parinaric acid bound to SCP-2 displayed an almost identical pattern as compared with the respective cis-isomeric counterparts (Table IV). Bound transparinaroyl-CoA showed a 8.82-ns rotational correlation time, 1.33-fold higher than that of trans-parinaric acid but almost undistinguishable from its cis-parinaroyl-CoA analog. The maximal anisotropy for trans-parinaroyl-CoA bound to SCP-2, 0.361, along with its wobbling cone angle, 13°, indicated that the amplitude of trans-parinaroyl-CoA motions in the SCP-2 binding site was significantly less than that of trans-parinaric acid in the SCP-2 binding site (Table IV). The single rotational correlation time for SCP-2-bound trans-parinaric acid was 6.63 ns, significantly lower (faster) than that of cis-parinaric acid. This finding is consistent with the straight-chain trans-parinaric acid rotating faster than its cis isomer in a highly ordered binding pocket. Nevertheless, the high remaining anisotropy of SCP-2-bound trans-parinaric acid, 0.334, and wobbling cone angle, 18°, were nearly identical to those for cis-parinaric acid and indicated a very ordered ligand binding environment with restricted amplitude of rotation (Table IV).
Circular Dichroism of Apo-and Holo-SCP-2-The above data suggest that fatty acyl-CoA and fatty acid binding to SCP-2 altered the secondary structure and/or conformation of SCP-2. The circular dichroism spectrum of the apo-SCP-2 ex-hibited two minima at 208 and 225 nm and one maximum at 193 nm, consistent with the presence of ␣-helical and ␤-sheet fractions in the secondary structure (Fig. 3). Spectrum analysis using the self-consistent method of Sreerama and Woody (21) revealed 32% ␣-helix, 21% ␤-sheet, 25% ␤-turns, and 22% unordered structure in apo-SCP-2. It should be noted that neither the shape of the protein spectrum nor the fitted parameters were affected when the concentration of the free SCP-2 was varied in the 1-6 M range (data not shown). The addition of 5 equivalents of oleic acid to SCP-2 did not significantly alter the SCP-2 secondary structure: 34% ␣-helix, 20% ␤-sheet, 25% ␤-turns, and 21% random coil. Neither did the addition of 5-fold molar excess of oleoyl-CoA alter the SCP-2 secondary structure: 34% ␣-helix, 19% ␤-sheet, 26% ␤-turns, and 21% random coil. In contrast, the addition of CoASH resulted in a small alteration of the circular dichroism spectrum in the holo-SCP-2 ( Fig. 3) and ordered the SCP-2 secondary structure, reducing the random coil from 22 to 18% and ␣-helix content from 32 to 28% while concomitantly increasing the ␤-sheet from 21 to 27%. These observations are consistent with fatty acid and fatty acyl-CoA not significantly altering the SCP-2 secondary structure. Thus, the SCP-2 altered rotational dynamics upon the addition of fatty acid or fatty acyl-CoA were more likely due to SCP-2 conformational changes upon ligand binding. DISCUSSION With rare exceptions, almost nothing is known of the physiological function of SCP-2 in intact cells (31,32). However, considerable insights have been obtained by in vitro studies of ligand binding or stimulation of specific enzyme activities. For example, SCP-2 binds phospholipids, sterols, and long chain isoprenoids (reviewed in Refs. 11 and 33). The fact that SCP-2 also enhances transfer of cholesterol (reviewed in Refs. 11 and 33-35) and phospholipids (reviewed in Refs. 11 and 33) further suggests roles of SCP-2 in the intracellular trafficking and/or metabolism of these ligands. Indeed, studies with intact transfected cells confirmed that SCP-2 enhances intracellular cholesterol trafficking (31,32). Three sets of data prompted an investigation of the fatty acyl-CoA binding properties of SCP-2. First, a role for SCP-2 in fatty acid as well as cholesterol metabolism was suggested by its effect on the microsomal esterification of cholesterol. Both in vitro studies (25,36) and reports on intact transfected cells (11,31,32) demonstrated that SCP-2 enhanced microsomal esterification of cholesterol. Whether SCP-2-mediated stimulation of microsomal acyl-CoA cholesteryl acyltransferase was due to enhanced delivery of cholesterol or of fatty acyl-CoA to the acyl-CoA cholesteryl  acyltransferase enzyme is unclear. Based on an early report indicating that SCP-2 did not specifically bind fatty acids (37), most investigators ascribed SCP-2's ability to stimulate ACAT to its propensity to enhance sterol transfer. However, it was recently shown that, in the absence of organic solvents used as carrier in the earlier study (37), SCP-2 does bind fatty acids with affinity similar to or higher than the fatty acid-binding proteins (14).
Second, the intracellular localization of SCP-2 to organelles intimately involved with fatty acid oxidation also suggests a potential role in fatty acid and/or fatty acyl-CoA metabolism. SCP-2 appears highly enriched in and/or on peroxisomes and mitochondria, and lesser amounts are localized near the endoplasmic reticulum (13,38,39). The observation that SCP-2 binds fatty acids (14) suggests that SCP-2 may also bind the fatty acyl-CoA esters. If this is so, SCP-2 may enhance delivery of fatty acids to and/or translocation of fatty acids across membranes of these organelles.
Third, homologous proteins to SCP-2 in other eukaryotic tissues are induced by and/or interact with fatty acids/fatty acyl-CoAs. For example, Candida tropicalis contains an oleic acid-inducible SCP-2 homologue with 30% identity to SCP-2 (40). The nonspecific lipid transfer protein from carrot is an SCP-2 homologue that binds phospholipids, fatty acids, and oleoyl-CoA (41). The terms nonspecific lipid transfer protein and SCP-2 are synonymous in mammalian systems. Finally, mammalian tissues contain a 58-kDa protein, SCP-x, sharing the same C-terminal 123 amino acid sequence (13.2 kDa) as found in SCP-2 (42). The SCP-x is localized exclusively in peroxisomes (13,38,39) and, furthermore, displays fatty acyl-CoA thiolase activity, an enzyme involved in ␤-oxidation of fatty acids (42).
The present investigation shows for the first time that SCP-2 binds long chain fatty acyl-CoAs with even higher affinity than members of the fatty acid-binding protein family. For example, SCP-2 bound cis-and trans-parinaroyl-CoAs with K d values in the range 2-4 nM. These affinities were 40 -200-fold higher than for the corresponding fluorescent fatty acids (22). More important, when compared with the binding of cis-parinaroyl-CoA to liver fatty acid-binding protein, L-FABP (10,15), or to acyl-CoA-binding protein, ACBP (10), SCP-2 has 3000-and 4-fold higher affinity, respectively, for cis-parinaroyl-CoA.
The fatty acyl chain specificity of SCP-2 binding fatty acyl-CoAs differs markedly from that for fatty acid binding. SCP-2 bound the trans-parinaroyl (straight chain) CoA with 2-fold higher affinity than the cis-parinaroyl (kinked chain) CoA. This was opposite to the ligand specificity of SCP-2 for trans-versus cis-parinaric acid. It is not known whether L-FABP exhibits a specificity for the type of fatty acyl (straight chain or kinked chain) CoA (15). ACBP does not bind fatty acids (8). SCP-2 bound fatty acyl-CoAs with carbon chain lengths ranging from C10 to C18, with maximal interaction for C12, C14, and C16. This suggests a potential role for SCP-2 in protein acylation reactions, since the latter utilize primarily C14 and C16 acyl-CoAs. Unsaturated fatty acyl-CoAs also interacted well with the SCP-2. Similar types of data for L-FABP specificity are not available for comparison. In contrast, ACBP binds fatty acyl-CoA esters over a wider range than SCP-2, with acyl chain length ranging from 8 -22 carbons and highest affinity for 14 -22-carbon chain length acyl-CoAs (43,44).
Both the circular dichroism and the rotational dynamic data present new findings on the structure of SCP-2. This adds to our knowledge of molecular details on the secondary/tertiary structure of apo-SCP-2 and provides some of the first data on the structure of holo-SCP-2. Circular dichroism showed that at pH 7.4 apo-SCP-2 was composed of 32% ␣-helix, 21% ␤-sheet, 25% ␤-turns, and 22% unordered structure. This observation is consistent with one recent report using NMR analysis of apo-SCP-2 secondary structure at pH 6.0, which showed the presence of three ␣-helices (38%) and five ␤-strands (24%) in the protein (45). The higher ␣-helix content reported in the NMR study may be due to lower pH used in that study. The circular dichroism secondary structure of apo-SCP-2 at pH 6 also showed 38% ␣-helix (data not shown). Analysis of the anisotropy decay of apo-SCP-2 at pH 7.4 (Table II) revealed a single rotational component with a rotational relaxation time of 7.8 ns. This was consistent with the rotational relaxation time for a 13.2-kDa globular protein of hydrodynamic radius of 20.5 Å. The anisotropy decays of cis-and trans-parinaric acid and parinaroyl-CoA allow further examination of SCP-2 rotational dynamics. These tightly bound ligands demonstrated rotational relaxation times in the range of 6.6 -8.8 ns, consistent with both the apo-SCP-2 and the holo-SCP-2 basically being globular/slightly ellipsoidal proteins. NMR analysis of the SCP-2 at pH 6 showed the polypeptide chain to be arranged in a secondary structure composed of three ␣-helixes and five ␤-sheets, resulting in a slightly elongated globular protein (45). In contrast, an earlier study of apo-SCP-2 time-resolved anisotropy decay at pH 7.4 estimated the rotational relaxation time of apo-SCP-2 at 15 ns and concluded that apo-SCP-2 was a strongly ellipsoidal shaped protein with an axial ratio of 2.8 (46). Neither the NMR study (45) nor the present data support this conclusion. We suggest that the reason for the apparent discrepancy may be due to the presence of SCP-2 dimers in the earlier study wherein 10 M apo-SCP-2 at pH 7.4 was examined (46). At 10 M apo-SCP-2, but not at 0.2-8.0 M apo-SCP-2 (data shown herein), apo-SCP-2 can form dimers with a rotational relaxation time of 15.7 ns (data not shown).
Oleoyl-CoA binding to SCP-2 significantly altered the conformation of the protein. Both oleoyl-CoA and oleic acid (but not CoASH) increased the SCP-2 overall rotational relaxation time, concomitantly increasing its hydrodynamic radius from 20.5 to 24.5 Å (Table II). This 9-Å increase in SCP-2 diameter represents a substantial shift in protein conformation in the holoversus apo-SCP-2. To our knowledge, there have been no previous reports of molecular details regarding the secondary/ tertiary structure of holo-SCP-2.
The parinaroyl-CoA rotational relaxation times were always significantly longer than those of the corresponding respective unesterified parinaric acid. This could be accounted for by perhaps the acyl chain (the fluorophore) of the acyl-CoA derivative not inserting quite as far into the acyl chain binding site of SCP-2. The only other study addressing this issue showed that the rotational relaxation time of diphenylhexatriene-PC was 7.4 ns (46). The latter observation is entirely consistent with those obtained herein with parinaric acids and parinaroyl-CoAs. Furthermore, it also suggests that binding of the diphenylhexatriene-PC ligand (rotational relaxation time of 7.4 ns) to dimeric apo-SCP-2 (rotational relaxation time near 15 ns) may disrupt SCP-2 dimers (46).
The degree of ordered motion of fatty acid and/or fatty acyl-CoA ligands in the SCP-2 binding site was quite different from that observed in the liver (L-FABP) and intestinal (I-FABP) fatty acid-binding proteins. The acyl chains of all four ligands (cis-and trans-parinaric acid and cis-and trans-parinaroyl-CoA) appeared localized in a very ordered binding site, as indicated by high remaining anisotropies ranging from 0.335 to 0.365 and wobbling cone angles of 12-18° (Table IV). In contrast, cis-and trans-parinaric acid were much less ordered when bound to L-FABP (limiting anisotropy, 0.20 -0.22) and to I-FABP (limiting anisotropy, 0.21-0.22) (47). The ordering of cis-parinaroyl-CoA bound to L-FABP, remaining anisotropy 0.388, 2 was similar to that observed in SCP-2 (Table IV). In short, the fatty acid, but not fatty acyl-CoA, binding sites of SCP-2 and the FABPs differ markedly. This suggests that the orientations of the fatty acyl chain of fatty acid and fatty acyl-CoA bound to SCP-2 are similar. In contrast, the large difference in remaining anisotropies of fatty acid versus fatty acyl-CoA bound to L-FABP (47) suggests that the respective acyl chains may be oriented oppositely. The latter suggestion is also consistent with modeling and other data reported earlier (10).
In summary, the data demonstrate that SCP-2 is a fatty acyl-CoA-binding protein as well as a fatty acid-binding protein and has much higher affinity for fatty acyl-CoAs than fatty acids. In fact, the affinity of SCP-2 for fatty acyl-CoAs may even exceed that of ACBP, another fatty acyl-CoA-binding protein.
As shown herein, SCP-2 binds cis-parinaroyl-CoA and transparinaroyl-CoA with K d values near 4.6 and 2.8 nM, respectively. In contrast, it was recently shown by use of titration microcalorimetry that ACBP bound palmitoyl-CoA with a K d near 0.5 ϫ 10 Ϫ13 M, 8 orders of magnitude lower than that observed for SCP-2 binding parinaroyl-CoA (48). This large difference could be due to differences in ligands and/or technique. To resolve this issue, in a preliminary study native rat liver ACBP was isolated, and its affinity for binding of cisparinaroyl was found to be high, with K d near 0.017 M and a binding stoichiometry of 0.8. 3 Thus, the fluorescence binding assay used herein is consistent with SCP-2 having a 4 -6-fold higher affinity for fatty acyl-CoA than ACBP. Furthermore, this also suggests that the 10 8 -fold higher affinity reported for ACBP by microcalorimetry was most likely due to differences in the techniques used rather than the ligand. Why titration calorimetry reports so much higher affinities for fatty acyl-CoAs than reported by any other method (reviewed in Ref. 1) is not known. However, a difference of several orders of magnitude in K d values has also been reported for ligand binding to cytosolic fatty acid-binding proteins using titration microcalorimetry (49) versus fluorescence (22,50,51) binding assays. In contrast, other reports on fatty acid binding to fatty acidbinding proteins indicate close agreement by fluorescence (22,50,51) and titration microcalorimetry (52,53) techniques. In general, it would appear that it is important to utilize the same technique when comparing the relative affinities of different fatty acyl-CoA and/or fatty acid-binding proteins. Finally, while the affinities of SCP-2 and ACBP measured by cis-parinaroyl-CoA binding differ only 4 -6-fold, the intracellular localization of these proteins differs much more markedly. SCP-2 appears primarily enriched inside and/or bound to the surface of peroxisomes and mitochondria, while ACPB is almost exclusively cytosolic (reviewed in Ref. 1). This might suggest separate important roles for these fatty acyl-CoA-binding proteins in fatty acyl-CoA metabolism and/or regulatory functions. Although the physiological significance of SCP-2's high affinity for fatty acyl-CoAs is unknown, these observation suggest that it may be as important as ACBP but in different intracelullar sites. This points to an exciting and new direction for investigating the function of SCP-2.