Reduction of the amyloidogenicity of a protein by specific binding of ligands to the native conformation

Abstract

It is known that human muscle acylphosphatase (AcP) is able, under appropriate conditions in vitro, to aggregate and form amyloid fibrils of the type associated with human diseases. A number of compounds were tested for their ability to bind specifically to the native conformation of AcP under conditions favoring denaturation and subsequent aggregation and fibril formation. Compounds displaying different binding affinities for AcP were selected and their ability to inhibit protein fibrillization in vitro was evaluated. We found that compounds displaying a relatively high affinity for AcP are able to significantly delay protein fibrillization, mimicking the effect of stabilizing mutations; in addition, the effectiveness of such outcome correlates positively to both ligand concentration and affinity to the native state of AcP. By contrast, the inhibitory effect of ligands on AcP aggregation disappears in a mutant protein in which such binding affinity is lost. These results indicate that the stabilization of the native conformation of amyloidogenic proteins by specific ligand binding can be a strategy of general interest to inhibit amyloid formation in vivo.

The amyloidoses are a group of diseases where proteins or protein fragments change from their native soluble forms into insoluble aggregates or plaques, and accumulate in a variety of organs and tissues (Kelly 1996; Rochet and Lansbury 2000). This phenomenon has devastating consequences for the function of the organs that lead in many cases to death. To date, nearly 20 different amyloidoses have been described, including Alzheimer's disease (Selkoe 1996), Parkinson's disease (Conway et al. 1998; Spillantini et al. 1998), light chain amyloidosis (Buxbaum and Gallo, 1999), Huntington's disease (Perutz 1999), familial amyloid polyneuropathy (Costa et al. 1978; McCutchen et al. 1995), dialysis-related amyloidosis (Drueke 2000), and the spongiform encephalopathies (Jackson and Clarke 2000).

A recent development in this field has been the discovery that proteins other than those associated with diseases are capable of fibril formation under appropriate conditions in vitro (Guijarro et al. 1998; Litvinovich et al. 1998; Chiti et al. 1999; Gross et al. 1999; Konno et al. 1999; Villegas et al. 2000; Yutani et al. 2000). This finding has two major consequences. First, it has implications for understanding the origin of amyloid formation in vivo because amyloid deposition can no longer be assumed to arise from the unique physical properties of a limited number of protein sequences (Dobson 1999; Rochet and Lansbury 2000). In addition, the ability to study the formation of amyloid fibrils by a wide range of proteins has opened up the opportunity to augment studies of the molecular basis of amyloid fibril formation (Jimenez et al. 1999; Chiti et al. 2000; Ramirez-Alvarado et al. 2000; Villegas et al. 2000). Such studies not only allow fundamental aspects of the process of fibrillization to be investigated in greater depth, but also allow possible strategies to retard or eliminate amyloid deposition to be tested.

In this work we use human muscle acylphosphatase (AcP) to investigate the importance of stabilizing the native state of a globular protein as a possible strategy to contrast amyloid formation in vivo. One of the strategies that can be used to stabilize the native state of a protein relative to unfolded conformations is the addition of ligands that bind specifically to the native state (Sancho et al. 1991; Chiti et al. 1998). In the case of transthyretin (TTR), it was found that binding of specific ligands to the native tetrameric state can inhibit the dissociation of the tetramer into a monomeric conformation that is only partially folded, although it retains a globular structure and a topology similar to that of the native state; such binding was found to inhibit formation of amyloid fibrils because the tetramer is known to be less prone to aggregation than the monomer (Baures et al. 1998; Peterson et al. 1998; Baures et al. 1999; Klabunde et al. 2000). Here we explore the effect of specific ligands on aggregation from a slightly different angle. AcP was shown to be capable of aggregation and amyloid fibril formation in vitro under conditions favoring the conversion of the native state into denatured conformations in which the native topology of the protein is fully disrupted (Chiti et al. 1999). By using AcP as a model system, we assess whether the ligand-mediated stabilization of the monomeric native state of a protein, relative to unfolded or partially unfolded states, can have the potential to retard aggregation and subsequent amyloid formation.

AcP is a 98-residue protein with a structure, in its native state, containing two parallel α-helices packed against a five-stranded antiparallel β-sheet (Saudek et al. 1989; Pastore et al. 1992). Kinetic experiments have indicated that AcP folds and unfolds in a two-state manner (Chiti et al. 1998; van Nuland et al. 1998). A simplified mechanism of fibril formation for AcP is the following (Chiti et al. 1999):

where N is the native and globular state of AcP, and D is an ensemble of denatured conformations. Because aggregation occurs from denatured or partially denatured conformations (D), this process is very slow under physiological conditions that favor the conformational stability of the native state N. Denaturants such as trifluoroethanol (TFE) cause the equilibrium between N and D to be shifted toward the latter, thus favoring aggregation and subsequent fibril formation (Chiti et al. 1999). Consistent with this concept, the analysis of a group of variants of AcP having single amino acid substitutions from the wild-type sequence has shown that the process of aggregation becomes more favorable as the destabilization of the native state caused by mutation becomes larger (Chiti et al. 2000). The most destabilized AcP variants can be readily converted into amyloid fibrils even in the absence of any denaturant, under solution conditions substantially more similar to those found in living systems (Chiti et al. 2000).

The negative correlation between the conformational stability of the native state and the propensity to form amyloid fibrils has encouraged us to assess whether simple compounds with a high affinity for the native conformation of AcP are effective as inhibitors of aggregation and amyloid formation. AcP is an enzyme that binds to and hydrolyzes acylphosphates (Stefani et al. 1997). This natural property of AcP has been exploited by studying the effects on aggregation of substrate analogs that bind to the active site of AcP and thereby stabilize the native state.

Results

Various compounds were tested to examine their binding to native AcP, particularly compounds having a phosphate group or a negative charge for which some indication of high affinity for this protein already exists (Stefani et al. 1997). These compounds include phosphate, sulfate, 2,3-bisphosphoglycerate, β-glycerophosphate, phenylphosphate, chloride, and adenosine triphosphate (ATP). Because aggregation can be readily induced by addition of TFE, it is important to determine the affinity of the various ligands for native AcP under such conditions. This objective is not straightforward to accomplish with conventional techniques such as optical spectroscopies or equilibrium dialysis, because the native conformation of the protein in the absence of the ligand is unstable under such denaturing conditions. Nevertheless, the affinity of a ligand for the native state of a protein can be determined, under denaturing conditions, by measuring the decrease of the unfolding rate in the presence of increasing concentrations of ligand, a procedure described in detail by Sancho et al. (1991) and Chiti et al. (1998).

Figure 1a ▶ shows the change of fluorescence of AcP in the presence of 30% (v/v) TFE. The fluorescence trace fits very well to a single exponential function. No further changes of fluorescence have been observed on the long timescale, indicating that unfolding is a monophasic process under these conditions. The change of the unfolding rate constant on addition of progressive concentrations of the phosphate ligand is reported in Figure 1b ▶. The dependence of the unfolding rate on ligand concentration is given by Chiti et al. (1998):

1

where k_u,obs is the observed unfolding rate constant, k_u,E and k_u,EL are the actual rate constants for the free and ligand-bound protein, respectively, K_D is the dissociation constant, and [L] is the ligand concentration. Equation 1 is valid only when the concentration of protein used in the experiments is much lower than that of ligand, which is the case for all of the measurements made here. Furthermore, k_u,EL was found to be several orders of magnitude smaller than k_u,E for AcP (Chiti et al. 1998), causing the equation to be converted into the simpler form

2

The K_D value can be considered as the concentration of ligand at which the unfolding rate decreases to one-half with respect to that in the absence of ligand. The fitting of the experimental data reported in Figure 1b ▶ to equation 2 provides an estimate of the K_D value for the complex formed by native AcP and phosphate. The analysis was repeated for other putative ligands, and the resulting K_D values are reported in Table 1. The various ligands have extremely diverse K_D values, ranging from a value of 0.083 ± 0.008 mM obtained for the sulfate ion to the value of 21 ± 4 mM obtained for the chloride ion.

Fig. 1.

Dependence of the unfolding rate constant of AcP (k_u,obs) on phosphate concentration in the presence of 30% (v/v) TFE. The decrease of k_u,obs is caused by the stabilization of the native conformation of AcP through ligand binding (Chiti et al. 1998). The solid line through the symbols is the best fit of the experimental data to equation 5. The fitting procedure allows the dissociation constant of the complex between native AcP and the ligand to be determined.

Table 1.

Propensity of AcP to form amyloid fibrils in the presence of various inhibitors

Ligand	KD (mM)a	ΔΔG (kJ mol−1)b	Minimal %TFE (v/v) required for aggregation (thioflavine T data)c	Minimal %TFE (v/v) required for aggregation (CD data)c
No ligand	—	0	18.5 ± 1	19.5 ± 1
Phosphate	0.13 ± 0.01	10.8 ± 0.2	23.5 ± 1	24.5 ± 1
2,3-bisphosphoglycerate	0.75 ± 0.04	6.60 ± 0.15	21.5 ± 1	21 ± 1
Adenosine triphosphate	0.15 ± 0.02	10.45 ± 0.4	23.5 ± 1	—
Phenylphosphate	0.78 ± 0.06	6.5 ± 0.2	19.5 ± 1	—
β-glycerophosphate	6.5 ± 0.6	2.3 ± 0.15	19 ± 1	20 ± 1
Sulfate	0.083 ± 0.008	11.9 ± 0.25	22 ± 1	22 ± 1
Chloride	21 ± 4	0.95 ± 0.2	18 ± 1	19 ± 1

^a Dissociation constant of the complex between native AcP and ligand in the presence of 30% (v/v) TFE. The values were determined fitting plots of unfolding rate constants versus ligand concentration to equation 2. The experimental errors are standard deviations.

^b Change of ΔG of unfolding upon addition of 10 mM ligand (free energy of stabilization). The values were calculated in each case using equation 5. The experimental errors are standard deviation and were propagated from the errors of K_D.

^c Minimal TFE concentration required for aggregation. The values correspond to the percentages of TFE at which the protein produces ca. 50% of the maximum observed increase in CD ellipticity at 215–220 nm and in thioflavine T fluorescence. The values were determined in all cases in the presence of 10 mM ligand. For adenosine triphosphate and phenylphosphate the values could not be determined with CD due to the large noise caused by these compounds in the CD spectra.

The specific binding of a ligand causes the native conformation of a protein to be stabilized. To determine, from our kinetic data, the free energy of stabilization resulting from the addition of a ligand, the following expression for a two-state folding protein can be used:

3

where ΔG is the free energy change of unfolding, whereas k_u,E and k_f,E are the unfolding and refolding rate constants, respectively. The folding rate of AcP does not change with the addition of ligand, as a consequence of the fact that the active site is not yet formed in either the unfolded or the transition state (Chiti et al. 1998). From equation 3 one can therefore derive the change of ΔG of unfolding when the ligand is added, that is, the free energy of stabilization on addition of ligand (ΔΔG):

4

where k_u,obs is the unfolding rate constant in the presence of ligand. Substituting equation 2 in equation 4 we obtain:

5

Equation 5 shows that the free energy of stabilization on addition of a ligand depends on both the affinity and concentration of the ligand. The calculated ΔΔG values on addition of 10 mM concentration of various ligands and of phosphate at various concentrations are reported in Tables 1 and 2, respectively.

Table 2.

Propensity of AcP to form amyloid fibrils in the presence of phosphate

Phosphate concentration	ΔΔG (kJ mol−1)a	Minimal %TFE (v/v) required for aggregation (thioflavine T data)b	Minimal %TFE (v/v) required for aggregation (CD data)b
No ligand	0	18.5 ± 1	19.5 ± 1
0.5 mM	3.9 ± 0.2	21 ± 1	20 ± 1
2 mM	6.9 ± 0.2	22.5 ± 1	23 ± 1
10 mM	10.8 ± 0.2	23.5 ± 1	24.5 ± 1
20 mM	12.5 ± 0.2	24.5 ± 1	24.5 ± 1
50 mM	14.75 ± 0.2	25.5 ± 1	27 ± 1
R23Q mutant, no ligand	0	12.5 ± 1	12.5 ± 1
R23Q mutant, 10 mM	∼0	11.5 ± 1	12 ± 1

^a Change of ΔG of unfolding upon addition of the indicated concentration of phosphate (free energy of stabilization). The values were calculated in each case using equation 5 and a value of 0.13 ± 0.01 mM for K_D. The experimental errors are standard deviation and were propagated from the errors in the K_D value.

^b Minimal TFE concentration required for aggregation. The values correspond to the TFE percentages at which the protein produces ca. 50% of the maximum observed increase in CD ellipticity at 215–220 nm and in thioflavine T fluorescence.

The propensity of AcP to aggregate was evaluated in the presence of the different ligands listed in Table 1 and in solutions containing TFE. TFE was used to destabilize the native state of AcP and therefore facilitate the growth of aggregates within a reasonable timescale. It was shown that the size and morphology of the amyloid fibrils formed from wild-type AcP in the presence of TFE and from a destabilized mutant in its absence are virtually identical, validating the use of this solvent for the measurements presented here. First we analyzed the effect on the aggregation of AcP at six different concentrations of phosphate, namely 0, 0.5, 2.0, 10, 20, and 50 mM. For each of these concentrations, 20–25 samples containing different quantities of TFE were prepared, all containing AcP and phosphate at the same concentration. The samples were tested after an incubation period of 5 h for the presence of aggregated protein by using far-UV CD spectroscopy and fluorescence in the presence of thioflavine T. The formation of aggregates by AcP is readily revealed by negative minima at 215–220 nm in the CD spectra, typical of β-sheet structure, and by the increase in the thioflavine T fluorescence intensity, indicative of amyloid-like structure (Chiti et al. 2000). After this time period the aggregates are granular; mature amyloid fibrils develop only after some days. It has been shown, however, that such early aggregates are ordered, that they possess many elements of the ultrastructure of amyloid fibrils, and that their formation correlates well with the subsequent development of well-defined amyloid fibrils (Chiti et al. 1999). This procedure therefore enables the relatively rapid analysis of a large number of samples for the presence of aggregates when these are sufficiently small in size to detect by using spectroscopic techniques.

The ranges of TFE concentrations at which protein aggregates were observed in each case are shown in Figure 2a ▶ and Table 2. The boundaries of these intervals were defined by the TFE concentrations at which the protein produces 50% of the maximum observed increase in CD ellipticity at 215–220 nm and in thioflavine T fluorescence. The results obtained by using CD are in close agreement with those obtained with thioflavine T. Although the upper limit of the range of TFE concentration at which aggregation occurs is similar, within experimental error, at the various ligand concentrations, the lowest TFE concentration that is required for formation of protein aggregates increases with ligand concentration. Figure 2b ▶ shows that there is a remarkable correlation between the minimum concentration of TFE that causes protein aggregation after 5 h and the free energy of ligand-induced stabilization of native AcP at different concentrations of phosphate. A value of 0.99 is calculated for the linear correlation coefficient (p < 0.0001), indicating that the correlation is highly significant. Because the aggregation process of AcP becomes faster as the TFE concentration increases (Chiti et al. 2000), the observed shift in the minimal concentration of TFE required for the aggregates to be present after 5 h on addition of phosphate indicates that this compound retards aggregation and amyloid formation.

Fig. 2.

(a) Ranges of TFE concentration at which protein aggregates are present for wild-type AcP and for the R23Q mutant after incubation of 5 h at the indicated concentrations of phosphate. The samples were incubated at 25°C in 50 mM acetate buffer (pH 5.5) and at protein concentrations of 0.375 mg/mL. Solid and hatched bars indicate results from far-UV CD spectroscopy and thioflavine T fluorescence, respectively. Experimental errors for both the lower and upper TFE concentrations at which aggregates occur are ∼1. 1% (v/v). (b) Correlation between the minimal TFE concentration required for producing aggregates after 5 h (taken as a measure of the amyloidogenic potential) and the change of ΔG of unfolding on ligand binding (taken as an index of the conformational stability of the native protein). The high r value (0.99) and the low p value (<0.0001) indicate that the correlation is highly significant. The equation of best fit has the form %TFE = (19±1) + (0.48±0.06) * ΔΔG. The values of TFE concentration on the Y-axis are an average of those determined with the thioflavine T and the CD data. The experimental errors are those indicated in Table 2 for both ΔΔG and %TFE values.

The effect of 10 mM phosphate was also studied on the aggregation process of a mutant of AcP in which the arginine at position 23 is substituted with a glutamine residue (R23Q mutant). Arginine 23, in conjunction with the loop 15–21, forms the binding site for the phosphate moiety of the substrate and all of its analogs; removal of arginine 23 is sufficient to compromise the function of AcP and cause its affinity for the substrate and analogs to be lost (Taddei et al. 1994; Stefani et al. 1997). The minimal concentration of TFE required for aggregation of this mutant drops down to ∼12.5 ± 1% (v/v), a result expected for a destabilized variant of AcP (Chiti et al. 2000). Of importance to the present study is that this value does not increase with the addition of ligand (Fig. 2a ▶; Table 2), providing evidence that the inhibition of the aggregation process exerted by this ligand arises from its specific binding to the native state of AcP.

The analysis was extended to all of the ligands listed in Table 1 to assess whether the inhibiting effect observed for phosphate on aggregation is specific to this compound or, rather, is a general characteristic of molecules binding to AcP. In this case a concentration of 10 mM was used for each ligand to facilitate comparison between the various substances. The free energies of stabilization obtained with a 10 mM concentration of the various ligands are shown in Table 1. The variation in the minimal TFE concentration that is necessary for aggregation after 5 h in the presence of these compounds is shown in Table 1 and in the two panels of Figure 3 ▶. The data points appear to be more dispersed compared with the analysis performed with the different concentrations of phosphate, probably because of additional non-specific effects exerted by these compounds on the conversion of soluble AcP into aggregates. Despite this, the statistical analysis clearly shows that the correlation between the minimal TFE concentration and free energy of stabilization is highly significant (r = 0.89, p = 0.003). The results show that only ligands with a relatively high binding affinity for AcP (K_D < 1 mM) are effective at this concentration in inhibiting the aggregation process.

Fig. 3.

(a) Ranges of TFE concentrations at which protein aggregates are present for wild-type AcP after incubation of 5 h in the presence of 10 mM of the indicated ligands. Experimental conditions and experimental errors are those described in Figure 2 ▶. Solid and hatched bars indicate results from far-UV CD spectroscopy and thioflavine T fluorescence, respectively. Adenosine triphosphate and phenylphosphate were not analyzed with CD because of their high absorbance in the far-UV region. (b) Correlation between the minimal TFE concentration required for producing aggregates after 5 h and the change of ΔG of unfolding on ligand binding. An r value of 0.89 and a p value of 0.003 indicate that the correlation is highly significant. The equation of best fit has the form %TFE = (18 ± 1) + (0.40 ± 0.08) * ΔΔG. The experimental errors are those indicated in Table 1 for both ΔΔG and %TFE values.

Discussion

In this paper we have shown that addition of ligands binding specifically to the native state of AcP retards the formation of aggregates under conditions in which the native state is significantly populated. Two observations indicate that the inhibitory effect of such ligands arises from the ability of these substances to stabilize the native state of AcP by specific interaction. First, the efficiency of inhibition depends not just on the concentration of a ligand, but also on its binding affinity to the native protein (K_D value). The correlation found between the propensity of AcP to aggregate and the free energy of ligand-induced stabilization of the native conformation is highly suggestive that such ligands act as inhibitors of aggregation by binding and stabilizing the native state of the protein. A second experimental result that is consistent with this view is that the inhibitory effect of ligands on the aggregation of AcP disappears in a mutant protein in which such binding affinity is lost.

In our previous study we have shown that single-point mutations destabilizing the native conformation of AcP facilitate aggregation at a denaturant concentration lower than those required for the wild-type protein (Chiti et al. 2000). The minimal TFE concentration required for aggregation was found to correlate closely with the free energy of destabilization of the protein (ΔΔG) (Chiti et al. 2000). The present investigation is complementary to this previous study. Instead of destabilizing the native conformation of AcP by mutation we have used specific ligands to achieve a global stabilization of the protein. The data obtained with different phosphate concentrations (Fig. 2b ▶) and with various ligands at the same concentration (Fig. 3b ▶), produce best-fitted linear equations similar to each other and to that reported previously from the mutational study (see Figs. 2b, 3b, and 4 ▶ ▶ ▶ legends). This indicates that the dependence of the propensity of AcP to aggregate on conformational stability is unique regardless of whether mutations or ligands are used to modify the stability of the native state. Figure 4 ▶ reports a comprehensive analysis in which the two sets of data obtained with ligands and protein mutants are combined to show the relationship between propensity to aggregate and conformational stability over a wide range of ΔΔG of unfolding, spanning from –22 to 12 kJ mol⁻¹. The correlation is highly significant (r = 0.97, p < 0.0001); the parameters of the linear expression that fits best to these data are given in the Figure 4 ▶ legend.

Fig. 4.

Correlation between the minimal TFE concentration required for producing aggregates and the change of ΔG of unfolding on either mutation (empty circles) or ligand binding (filled circles). The cross represents the data point for wild-type AcP in the absence of ligands. The r and p values indicate that the correlation is highly significant. The equation of best fit has the form %TFE = (18.25±0.5) + (0.43±0.04) * ΔΔG, which is very similar to those obtained by separate fits of the data obtained at different concentrations of phosphate (see Fig. 2b ▶ legend), at a 10 mM concentration of various ligands (see Fig. 3b ▶ legend), or with the mutants (in the latter case, the equation reported previously [Chiti et al. 2000] has the form %TFE = (18.5±1) + (0.46±0.08) * ΔΔG).

A detailed study directed to the inhibition of aggregation by design of compounds binding specifically to the native conformation has been performed with TTR, a homotetrameric protein working as a carrier of the thyroid hormone and which is involved in fibril deposition in senile systemic amyloidosis and familial amyloid polyneuropathy (Kelly 1996). The finding that the thyroid hormone can by itself inhibit fibril formation by TTR (Miroy et al. 1996) has encouraged efforts directed to finding suitable drugs that could inhibit aggregation in vivo by specific binding (Baures et al. 1998, 1999; Peterson et al. 1998; Klabunde et al. 2000). A recent report proposes that ortho-trifluormethyl-phenyl anthranilic acid and N-(meta-trifluoromethylphenyl) phenoxazine 4,6-dicarboxylic acid could represent two very promising classes of TTR fibril formation inhibitors, not just for their high specificity and binding affinity to the homotetrameric conformation of TTR, but also for the minimization of side effects associated with their prolonged use (Klabunde et al. 2000). The rationale for the inhibitory effect on aggregation of these compounds is that the native tetrameric state of TTR is stabilized relative to the more amyloidogenic monomeric state (Miroy et al. 1996). Our data on AcP show that disfavoring denaturation of a globular protein by addition of ligands binding specifically to the native state is also effective against aggregation. This reinforces previous views that, at least for some proteins, aggregation occurs from denatured states in which the native topology is fully disrupted (Dobson 1999). Although the ligands investigated here have binding affinities lower than those required for drug design, these data underlie the concept that stabilization of the native conformation of amyloidogenic proteins by specific binding can represent a strategy of general interest for inhibiting aggregation and amyloid formation in vivo.

Materials and methods

Materials

AcP was purified from an Escherichia coli-based expression system, as described previously (Taddei et al. 1996). The protein used in this study has cysteine at position 21 replaced by a serine residue to avoid complexities arising from a free cysteine residue (van Nuland et al. 1998). The R23Q mutant of AcP was produced by using an oligonucleotide-directed mutagenesis kit from Pharmacia Amersham (Taddei et al. 1996). It also presents the C21S replacement. Protein concentration was measured by UV absorption by using an ɛ₂₈₀ value of 1.49 mL mg^–1 cm^–1. All compounds used as specific ligands of AcP were purchased from Sigma-Aldrich.

Stopped-flow kinetics

A Bio-Logic SFM-3 stopped-flow device was used to measure the rate of TFE-induced unfolding of AcP in the presence of various concentrations of ligands. In each experiment, 1 volume of a solution containing native AcP was mixed with 10 volumes of a denaturing solution containing TFE and the ligand at the desired concentration. In all experiments, the final conditions after mixing were 0.02 mg mL^–1 AcP, 30% (v/v) TFE, 50 mM acetate buffer (pH 5.5) at 25°C. For each ligand, the unfolding rate of AcP was measured at eight different concentrations. The various kinetic traces obtained in the presence of ligand were fitted to single exponential functions to determine the unfolding rate constant by using the Kaleidagraph software package.

Circular dichroism

Far-UV CD spectra were acquired at 25°C by using a Jasco J-720 spectropolarimeter and cuvettes of 1-mm path length. In a typical experiment, 20–25 samples of AcP at a concentration of 0.375 mg/mL (34 μM) were prepared at different TFE concentrations ranging from 0% to 45% (v/v) in 50 mM acetate buffer (pH 5.5) at 25°C. The far-UV CD spectrum of each sample was acquired after 5 h of incubation. This analysis was repeated in the presence of each of the seven ligands considered here at a concentration of 10 mM. For phosphate, the analysis was also performed at five different concentrations ranging from 0.5 to 50 mM. All mother solutions containing AcP were centrifuged and their protein concentrations measured immediately before preparation of the samples.

Thioflavine T dye binding

Aliquots of the samples prepared as described in the previous paragraph for the CD measurements were also used to test for thioflavine T binding (LeVine 1995). A sample of 133 μL was mixed with 867 μL of 25 μM thioflavine T in 25 mM phosphate buffer (pH 6.0). The fluorescence was measured immediately after mixing by using excitation and emission wavelengths of 440 and 485 nm, respectively. A Shimadzu RF-5000 spectrofluorometer was used for fluorescence measurements.

Abbreviations

• AcP, acylphosphatase

• CD, circular dichroism

• TFE, 2,2,2-trifluoroethanol

• TTR, transthyretin