A hydrophobic cluster forms early in the folding of dihydrofolate reductase

The rapid kinetic phase that leads from unfolded species to transient folding intermediates in dihydrofolate reductase from Escherichia coli was examined by site-directed mutagenesis and by physicochemical means. The absence of this fluorescence-detected phase in the refolding of the Trp-74Phe mutant protein strongly implies that this early phase in refolding can be assigned to just one of the five Trp residues in the protein, Trp-74. In addition, water-soluble fluorescence quenching agents, iodide and cesium, have a much less significant effect on this early step in refolding than on the slower phases that lead to native and native-like conformers. These and other data imply that an important early event in the folding of dihydrofolate reductase is the formation of a hydrophobic cluster which protects Trp-74 from solvent.     

Characterization of the folding and unfolding reactions of single-chain monellin: evidence for multiple intermediates and competing pathways

The mechanisms of folding and unfolding of the small plant protein monellin have been delineated in detail. For this study, a single-chain variant of the natively two-chain monellin, MNEI, was used, in which the C terminus of chain B was connected to the N terminus of chain A by a Gly-Phe linker. Equilibrium guanidine hydrochloride (GdnHCl)-induced unfolding experiments failed to detect any partially folded intermediate that is stable enough to be populated at equilibrium to a significant extent. Kinetic experiments in which the refolding of GdnHCl-unfolded protein was monitored by measurement of the change in the intrinsic tryptophan fluorescence of the protein indicated the accumulation of three transient partially structured folding intermediates. The fluorescence change occurred in three kinetic phases: very fast, fast, and slow. It appears that the fast and slow changes in fluorescence occur on competing folding pathways originating from one unfolded form and that the very fast change in fluorescence occurs on a third parallel pathway originating from a second unfolded form of the protein. Kinetic experiments in which the refolding of alkali-unfolded protein was monitored by the change in the fluorescence of the hydrophobic dye 8-anilino-1-naphthalenesulfonic acid (ANS), consequent to the dye binding to the refolding protein, as well as by the change in intrinsic tryptophan fluorescence, not only confirmed the presence of the three kinetic intermediates but also indicated the accumulation of one or more early intermediates at a few milliseconds of refolding. These experiments also exposed a very slow kinetic phase of refolding, which was silent to any change in the intrinsic tryptophan fluorescence of the protein. Hence, the spectroscopic studies indicated that refolding of single-chain monellin occurs in five distinct kinetic phases. Double-jump, interrupted-folding experiments, in which the accumulation of folding intermediates and native protein during the folding process could be determined quantitatively by an unfolding assay, indicated that the fast phase of fluorescence change corresponds to the accumulation of two intermediates of differing stabilities on competing folding pathways. They also indicated that the very slow kinetic phase of refolding, identified by ANS binding, corresponds to the formation of native protein. Kinetic experiments in which the unfolding of native protein in GdnHCl was monitored by the change in intrinsic tryptophan fluorescence indicated that this change occurs in two kinetic phases. Double-jump, interrupted-unfolding experiments, in which the accumulation of unfolding intermediates and native protein during the unfolding process could be determined quantitatively by a refolding assay, indicated that the fast unfolding phase corresponds to the formation of fully unfolded protein via one unfolding pathway and that the slow unfolding phase corresponds to a separate unfolding pathway populated by partially unfolded intermediates. It is shown that the unfolded form produced by the fast unfolding pathway is the one which gives rise to the very fast folding pathway and that the unfolded form produced by the slower unfolding pathway is the one which gives rise to the slow and fast folding pathways.     

Revealing a Concealed Intermediate that Forms after the Rate-limiting Step of Refolding of the SH3 Domain of PI3 Kinase

Kinetic and equilibrium studies of the folding and unfolding of the SH3 domain of the PI3 kinase, have been used to identify a folding intermediate that forms after the rate-limiting step on the folding pathway. Folding and unfolding, in urea as well as in guanidine hydrochloride (GdnHCl), were studied by monitoring changes in the intrinsic fluorescence or in the far-UV circular dichroism (CD) of the protein. The two probes yield non-coincident equilibrium transitions for unfolding in urea, indicating that an intermediate, I, exists in equilibrium with native (N) and unfolded (U) protein, during unfolding. Hence, the equilibrium unfolding data were analyzed according to a three-state N ↔ I ↔ U mechanism. An intermediate is observed also in kinetic unfolding studies, and its presence leads to the unfolding reaction in urea as well as in GdnHCl, occurring in two steps. The fast step is complete within the initial 11 ms of unfolding and manifests itself in a burst phase change in fluorescence. At high concentrations of GdnHCl, the entire change in fluorescence during unfolding occurs during the 11 ms burst phase. CD measurements indicate, however, that I retains N-like secondary structure. An analysis of the kinetic and thermodynamic data, according to a minimal three-state N ↔ I ↔ U mechanism, positions I after the rate-limiting transition state, TS1, of folding, on the reaction coordinate of folding in GdnHCl. Hence, I is not revealed when folding is commenced from U, regardless of the nature of the probe used to follow the folding reaction. Interrupted unfolding experiments, in which the protein is unfolded transiently in GdnHCl for various lengths of time before being refolded, showed that I refolds to N much faster than does U, confirms the analysis of the direct folding and unfolding experiments, that I is formed after the rate-limiting step of refolding in GdnHCl.     

Folding of dihydrofolate reductase from Escherichia coli

The urea-induced equilibrium unfolding transition of dihydrofolate reductase from Escherichia coli was monitored by UV difference, circular dichroism (CD), and fluorescence spectroscopy. Each of these data sets were well described by a two-state unfolding model involving only native and unfolded forms. The free energy of folding in the absence of urea at pH 7.8, 15 degrees C is 6.13 +/- 0.36 kcal mol-1 by difference UV, 5.32 +/- 0.67 kcal mol-1 by CD, and 5.42 +/- 1.04 kcal mol-1 by fluorescence spectroscopy. The midpoints for the difference UV, CD, and fluorescence transitions are 3.12, 3.08, and 3.18 M urea, respectively. The near-coincidence of the unfolding transitions monitored by these three techniques also supports the assignment of a two-state model for the equilibrium results. Kinetic studies of the unfolding and refolding reactions show that the process is complex and therefore that additional species must be present. Unfolding jumps in the absence of potassium chloride revealed two slow phases which account for all of the amplitude predicted by equilibrium experiments. Unfolding in the presence of 400 mM KCl results in the selective loss of the slower phase, implying that there are two native forms present in equilibrium prior to unfolding. Five reactions were observed in refolding: two slow phases designated tau 1 and tau 2 that correspond to the slow phases in unfolding and three faster reactions designated tau 3, tau 4, and tau 5 that were followed by stopped-flow techniques. The kinetics of the recovery of the native form was monitored by following the binding of methotrexate, a tight-binding inhibitor of dihydrofolate reductase, at 380 nm.(ABSTRACT TRUNCATED AT 250 WORDS)     

Effects of the difference in the unfolded-state ensemble on the folding of Escherichia coli dihydrofolate reductase

The unfolded state of a protein is an ensemble of a large number of conformations ranging from fully extended to compact structures. To investigate the effects of the difference in the unfolded-state ensemble on protein folding, we have studied the structure, stability, and folding of "circular" dihydrofolate reductase (DHFR) from Escherichia coli in which the N and C-terminal regions are cross-linked by a disulfide bond, and compared the results with those of disulfide-reduced "linear" DHFR. Equilibrium studies by circular dichroism, difference absorption spectra, solution X-ray scattering, and size-exclusion chromatography show that whereas the native structures of both proteins are essentially the same, the unfolded state of circular DHFR adopts more compact conformations than the unfolded state of the linear form, even with the absence of secondary structure. Circular DHFR is more stable than linear DHFR, which may be due to the decrease in the conformational entropy of the unfolded state as a result of circularization. Kinetic refolding measurements by stopped-flow circular dichroism and fluorescence show that under the native conditions both proteins accumulate a burst-phase intermediate having the same structures and both fold by the same complex folding mechanism with the same folding rates. Thus, the effects of the difference in the unfolded state of circular and linear DHFRs on the refolding reaction are not observed after the formation of the intermediate. This suggests that for the proteins with close termini in the native structure, early compaction of a protein molecule to form a specific folding intermediate with the N and C-terminal regions in close proximity is a crucial event in folding. If there is an enhancement in the folding reflecting the reduction in the breadth of the unfolded-state ensemble for circular DHFR, this acceleration must occur in the sub-millisecond time-range.     

Construction of a synthetic gene for an R-plasmid-encoded dihydrofolate reductase and studies on the role of the N-terminus in the protein

R67 dihydrofolate reductase (DHFR) is a novel protein that provides clinical resistance to the antibacterial drug trimethoprim. The crystal structure of a dimeric form of R67 DHFR indicates the first 16 amino acids are disordered [Matthews et al. (1986) Biochemistry 25, 4194-4204]. To investigate whether these amino acids are necessary for protein function, the first 16 N-terminal residues have been cleaved off by chymotrypsin. The truncated protein is fully active with kcat = 1.3 s-1, Km(NADPH) = 3.0 microM, and Km(dihydrofolate) = 5.8 microM. This result suggests the functional core of the protein resides in the beta-barrel structure defined by residues 27-78. To study this protein further, synthetic genes coding for full-length and truncated R67 DHFRs were constructed. Surprisingly, the gene coding for truncated R67 DHFR does not produce protein in vivo or confer trimethoprim resistance upon Escherichia coli. Therefore, the relative stabilities of native and truncated R67 DHFR were investigated by equilibrium unfolding studies. Unfolding of dimeric native R67 DHFR is protein concentration dependent and can be described by a two-state model involving native dimer and unfolded monomer. Using absorbance, fluorescence, and circular dichroism techniques, an average delta GH2O of 13.9 kcal mol-1 is found for native R67 DHFR. In contrast, an average delta GH2O of 11.3 kcal mol-1 is observed for truncated R67 DHFR. These results indicate native R67 DHFR is 2.6 kcal mol-1 more stable than truncated protein. This stability difference may be part of the reason why protein from the truncated gene is not found in vivo in E. coli.     

Effect of single amino acid replacements on the folding and stability of dihydrofolate reductase from Escherichia coli

The role of the secondary structure in the folding mechanism of dihydrofolate reductase from Escherichia coli was probed by studying the effects of amino acid replacements in two alpha helices and two strands of the central beta sheet on the folding and stability. The effects on stability could be qualitatively understood in terms of the X-ray structure for the wild-type protein by invoking electrostatic, hydrophobic, or hydrogen-bonding interactions. Kinetic studies focused on the two slow reactions that are thought to reflect the unfolding/refolding of two stable native conformers to/from their respective folding intermediates [Touchette, N. A., Perry, K. M., & Matthews, C. R. (1986) Biochemistry 25, 5445-5452]. Replacements at three different positions in helix alpha B selectively alter the relaxation time for unfolding while a single replacement in helix alpha C selectively alters the relaxation time for refolding. This behavior is characteristic of mutations that change the stability of the protein but do not affect the rate-limiting step. In striking contrast, replacements in strands beta F and beta G can affect both unfolding and refolding relaxation times. This behavior shows that these mutations alter the rate-limiting step in these native-to-intermediate folding reactions. It is proposed that the intermediates have an incorrectly formed beta sheet whose maturation to the structure found in the native conformation is one of the slow steps in folding.     

Test for cooperativity in the early kinetic intermediate in lysozyme folding

During the folding of many proteins, collapsed globular states are formed prior to the native structure. The role of these states for the folding process has been widely discussed. Comparison with properties of synthetic homo and heteropolymers had suggested that the initial collapse represented a shift of the ensemble of unfolded conformations to more compact states without major energy barriers. We investigated the folding/unfolding transition of a collapsed state, which transiently populates early in lysozyme folding. This state forms within the dead-time of stopped-flow mixing and it has been shown to be significantly more compact and globular than the denaturant-induced unfolded state. We used the GdmCl-dependence of the dead-time signal change to characterize the unfolding transition of the burst phase intermediate. Fluorescence and far-UV CD give identical unfolding curves, arguing for a cooperative two-state folding/unfolding transition between unfolded and collapsed lysozyme. These results show that collapse leads to a distinct state in the folding process, which is separated from the ensemble of unfolded molecules by a significant energy barrier. NMR, fluorescence and small angle X-ray scattering data further show that some local interactions in unfolded lysozyme exist at denaturant concentrations above the coil-collapse transition. These interactions might play a crucial role in the kinetic partitioning between fast and slow folding pathways.     

Cyclophilin-promoted folding of mouse dihydrofolate reductase does not include the slow conversion of the late-folding intermediate to the active enzyme

Cyclophilins accelerate slow protein folding reactions in vitro by catalyzing the cis/trans isomerization of peptidyl-prolyl bonds. Cyclophilins were reported to be involved in a variety of cellular functions, including the promotion of protein folding by use of the substrate mouse dihydrofolate reductase (DHFR). The interaction of cyclophilin with DHFR has only been studied under limited conditions so far, not taking into account that native DHFR exists in equilibrium with a non-native late-folding intermediate. Here we report a systematic analysis of catalysis of DHFR folding by cyclophilins. The specific ligand methotrexate traps DHFR in its native state, permitting a specific analysis of the action of cyclophilin on both denatured DHFR with non-native prolyl bonds and denatured DHFR with all-native prolyl bonds. Cyclophilins from yeast and Neurospora crassa as well as the related prolyl isomerase b from Escherichia coli promote the folding of different forms of DHFR to the enzymatically active form, demonstrating the generality of cyclophilin-catalyzed folding of DHFR. The slow equilibrium between the late-folding intermediate and native DHFR suggests that prolyl isomerization may be required for this final phase of conversion to native DHFR. However, by reversible trapping of the intermediate, we analyze the slow interconversion between native and late-folding conformations in the backward and forward reactions and show a complete independence of cyclophilin. We conclude that cyclophilin catalyzes folding of DHFR, but surprisingly not in the last slow folding step.     

Osmolytes induce structure in an early intermediate on the folding pathway of barstar

Osmolytes stabilize proteins against denaturation, but little is known about how their stabilizing effect might affect a protein folding pathway. Here, we report the effects of the osmolytes, trimethylamine-N-oxide, and sarcosine on the stability of the native state of barstar as well as on the structural heterogeneity of an early intermediate ensemble, IE, on its folding pathway. Both osmolytes increase the stability of the native protein to a similar extent, with stability increasing linearly with osmolyte concentration. Both osmolytes also increase the stability of IE but to different extents. Such stabilization leads to an acceleration in the folding rate. Both osmolytes also alter the structure of IE but do so differentially; the fluorescence and circular dichroism properties of IE differ in the presence of the different osmolytes. Because these properties also differ from those of the unfolded form in refolding conditions, different burst phase changes in the optical signals are seen for folding in the presence of the different osmolytes. An analysis of the urea dependence of the burst phase changes in fluorescence and circular dichroism demonstrates that the formation of IE is itself a multistep process during folding and that the two osmolytes act by stabilizing, differentially, different structural components present in the IE ensemble. Thus, osmolytes can alter the basic nature of a protein folding pathway by discriminating, through differential stabilization, between different members of an early intermediate ensemble, and in doing so, they thereby appear to channel folding along one route when many routes are available.     

Monovalent ion-mediated folding of the Tetrahymena thermophila ribozyme

The time-course of monovalent cation-induced folding of the L-21 Sca1 Tetrahymena thermophila ribozyme and a selected mutant was quantitatively followed using synchrotron X-ray (.OH) footprinting. Initiating folding by increasing the concentration of either Na+ or K+ to 1.5M from an initial condition of approximately 0.008 M Na+ at 42 degrees C resulted in the complete formation of tertiary contacts within the P5abc subdomain and between the peripheral helices within the dead time of our measurements (k>50 s(-1)). These results contrast with folding rates of 2-0.2 s(-1) previously observed for formation of these contacts in 10mM Mg2+ from the same initial condition. Thus, the initial formation of native tertiary contacts is inhibited by divalent but not monovalent cations. The native contacts within the catalytic core form without a detectable burst phase at rates of 0.4-1.0 s(-1) in a manner reminiscent of the Mg2+-dependent folding behavior, although tenfold faster. The tertiary interactions stabilizing the catalytic core interaction with P4-P6 and P2.1, as well as one of the protections internal for the P4-P6 domain, display progress curves with appreciable burst amplitudes and a phase comparable in rate to that of the catalytic core. That the slow folding of the ribozyme's core is a consequence of the alt-P3 secondary structure is shown by the 100% burst phase amplitudes that are observed for folding of the U273A mutant ribozyme within which the native secondary structure (P3) is strengthened. Thus, formation of a misfolded intermediate(s) resulting from the alt-P3 secondary structure is independent of ion valency while the rate at which the respective intermediates are resolved is sensitive to ion valency. The overall portrait painted by these results is that ion valency differentially affects steps in the folding process and that folding in monovalent ion alone for the U273A mutant Tetrahymena ribozyme is fast and direct.     

Characterization of a R plasmid-associated, trimethoprim-resistant dihydrofolate reductase and determination of the nucleotide sequence of the reductase gene

The trimethoprim-resistant dihydrofolate reductase associated with the R plasmid R388 was isolated from strains that over-produce the enzyme. It was purified to apparent homogeneity by affinity chromatography and two consecutive gel filtration steps under native and denaturing conditions. The purified enzyme is composed of four identical subunits with molecular weights of 8300. A 1100 bp long DNA segment which confers resistance to trimethoprim was sequenced. The structural gene was identified on the plasmid DNA by comparing the amino acid composition of the deduced proteins with that of the purified enzyme. The gene is 234 bp long and codes for 78 amino acids. No homology can be found between the deduced amino acid sequence of the R388 dihydrofolate reductase and those of other prokaryotic or eukaryotic dihydrofolate reductases. However, it differs in only 17 positions from the enzyme associated with the trimethoprim-resistance plasmid R67.     

Folding kinetics of phage T4 thioredoxin

The folding mechanism for bacteriophage T4 thioredoxin is best described by a four-state box mechanism, N----Uc----Ut----It----N, where N indicates native, Uc the unfolded form with the cis proline isomer, Ut unfolded with the trans proline isomer, and It a compact form with a trans proline isomer. Both manual mixing fluorescence and size-exclusion chromatography indicate that there is a cis-trans proline isomerization that is important to the folding pathway. Furthermore, the data suggest that the cis-trans isomerization can also occur in a compact nativelike state which is referred to as It. The slow phase seen in fluorescence seems to be monitoring the cis-trans isomerization in the compact form, not the isomerization which occurs in the denatured state.     

The coordination of the isomerization of a conserved non-prolyl cis peptide bond with the rate-limiting steps in the folding of dihydrofolate reductase

The propensity for peptide bonds to adopt the trans configuration in native and unfolded proteins, and the relatively slow rates of cis-trans isomerization reactions, imply that the formation of cis peptide bonds in native conformations are likely to limit folding reactions. The role of the conserved cis Gly95-Gly96 peptide bond in dihydrofolate reductase (DHFR) from Escherichia coli was examined by replacing Gly95 with alanine. The introduction of a beta carbon at position 95 is expected to increase the propensity for the trans isomer and perturb the isomerization reaction required to reach the native conformation. Although G95A DHFR is 1.30 kcal mol(-1) less stable than the wild-type protein, it adopts a well-folded structure that can be chemically denatured in a cooperative fashion. The mutant protein also retains the complex refolding kinetic pattern attributed to a parallel-channel mechanism in wild-type DHFR. The spectroscopic response upon refolding monitored by Trp fluorescence and the absence of a Trp/Trp exciton coupling apparent in the far-UV CD spectrum of the wild-type protein, however, indicated that the tertiary structure of the folded state for G95A DHFR is altered. The addition of methotrexate (MTX), a tight-binding inhibitor, to folded G95A DHFR restored the exciton coupling and the fluorescence properties through five slow kinetic events whose relaxation times are independent of the ligand and the denaturant concentrations. The results were interpreted to mean that MTX-binding drives the formation of the cis isomer of the peptide bond between Ala95 and Gly96 in five compact and stable but not wild-type-like conformations that contain the trans isomer. Folding studies in the presence of MTX for both wild-type and G95A DHFR support the notion that the cis peptide bond between Gly95 and Gly96 in the wild-type protein forms during four parallel rate-limiting steps, which are primarily controlled by folding reactions, and lead directly to a set of native, or native-like, conformers. The isomerization of the cis peptide bond is not a source of the parallel channels that characterize the complex folding mechanism for DHFR.     

Intramolecular catalysis of a proline isomerization reaction in the folding of dihydrofolate reductase.

The cis/trans isomerization of the peptide bond preceding proline residues in proteins can limit the rate at which a protein folds to its native conformation. Mutagenic analyses of dihydrofolate reductase (DHFR) from Escherichia coli show that this isomerization reaction can be intramolecularly catalyzed by a side chain from an amino acid which is distant in sequence but adjacent in the native conformation. The guanidinium NH2 nitrogen of Arg 44 forms one hydrogen bond to the imide nitrogen and a second to the carbonyl oxygen of Pro 66 in wild-type DHFR. Replacement of Arg 44 with Leu results in a change of the nature of the two slow steps in refolding from being limited by the acquisition of secondary and/or tertiary structure to being limited by isomerization. The simultaneous replacement of Pro 66 with Ala (i.e., the Leu 44/Ala 66 double mutant) eliminates this isomerization reaction and once again makes protein folding the limiting process. Apparently, one or both of the hydrogen bonds between Arg 44 and Pro 66 accelerate the isomerization of the Gln 65-Pro 66 peptide bond. The replacement of Arg 44 with Leu affects the kinetics of the slow folding reactions in a fashion which indicates that the crucial hydrogen bonds form in the transition states for the rate-limiting steps in folding.     

Microsecond subdomain folding in dihydrofolate reductase

The characterization of microsecond dynamics in the folding of multisubdomain proteins has been a major challenge in understanding their often complex folding mechanisms. Using a continuous-flow mixing device coupled with fluorescence lifetime detection, we report the microsecond folding dynamics of dihydrofolate reductase (DHFR), a two-subdomain α/β/α sandwich protein known to begin folding in this time range. The global dimensions of early intermediates were monitored by Förster resonance energy transfer, and the dynamic properties of the local Trp environments were monitored by fluorescence lifetime detection. We found that substantial collapse occurs in both the locally connected adenosine binding subdomain and the discontinuous loop subdomain within 35 μs of initiation of folding from the urea unfolded state. During the fastest observable ∼ 550 μs phase, the discontinuous loop subdomain further contracts, concomitant with the burial of Trp residue(s), as both subdomains achieve a similar degree of compactness. Taken together with previous studies in the millisecond time range, a hierarchical assembly of DHFR—in which each subdomain independently folds, subsequently docks, and then anneals into the native conformation after an initial heterogeneous global collapse—emerges. The progressive acquisition of structure, beginning with a continuously connected subdomain and spreading to distal regions, shows that chain entropy is a significant organizing principle in the folding of multisubdomain proteins and single-domain proteins. Subdomain folding also provides a rationale for the complex kinetics often observed.     

Probing the cytochrome c' folding landscape

The folding kinetics of R. palustris cytochrome c' (cyt c') have been monitored by heme absorption and native Trp72 fluorescence at pH 5. The Trp72 fluorescence burst signal suggests early compaction of the polypeptide ensemble. Analysis of heme transient absorption spectra reveals deviations from two-state behavior, including a prominent slow phase that is accelerated by the prolyl isomerase cyclophilin. A nonnative proline configuration (Pro21) likely interferes with the formation of the helical bundle surrounding the heme.     

Multiple-probe analysis of folding and unfolding pathways of human serum albumin. Evidence for a framework mechanism of folding.

The changes in the far-UV CD signal, intrinsic tryptophan fluorescence and bilirubin absorbance showed that the guanidine hydrochloride (GdnHCl)-induced unfolding of a multidomain protein, human serum albumin (HSA), followed a two-state process. However, using environment sensitive Nile red fluorescence, the unfolding and folding pathways of HSA were found to follow a three-state process and an intermediate was detected in the range 0.25-1.5 m GdnHCl. The intermediate state displayed 45% higher fluorescence intensity than that of the native state. The increase in the Nile red fluorescence was found to be due to an increase in the quantum yield of the HSA-bound Nile red. Low concentrations of GdnHCl neither altered the binding affinity of Nile red to HSA nor induced the aggregation of HSA. In addition, the secondary structure of HSA was not perturbed during the first unfolding transition (<1.5 m GdnHCl); however, the secondary structure was completely lost during the second transition. The data together showed that the half maximal loss of the tertiary structure occurred at a lower GdnHCl concentration than the loss of the secondary structure. Further kinetic studies of the refolding process of HSA using multiple spectroscopic techniques showed that the folding occurred in two phases, a burst phase followed by a slow phase. An intermediate with native-like secondary structure but only a partial tertiary structure was found to form in the burst phase of refolding. Then, the intermediate slowly folded into the native state. An analysis of the refolding data suggested that the folding of HSA could be best explained by the framework model.