Mercury(II) binds to both of chymotrypsin's histidines, causing inhibition followed by irreversible denaturation/aggregation

Abstract

The toxicity of mercury is often attributed to its tight binding to cysteine thiolate anions in vital enzymes. To test our hypothesis that Hg(II) binding to histidine could be a significant factor in mercury's toxic effects, we studied the enzyme chymotrypsin, which lacks free cysteine thiols; we found that chymotrypsin is not only inhibited, but also denatured by Hg(II). We followed the aggregation of denatured enzyme by the increase in visible absorbance due to light scattering. Hg(II)‐induced chymotrypsin precipitation increased dramatically above pH 6.5, and free imidazole inhibited this precipitation, implicating histidine‐Hg(II) binding in the process of chymotrypsin denaturation/aggregation. Diethylpyrocarbonate (DEPC) blocked chymotrypsin's two histidines (his₄₀ and his₅₇) quickly and completely, with an IC₅₀ of 35 ± 6 µM. DEPC at 350 µM reduced the hydrolytic activity of chymotrypsin by 90%, suggesting that low concentrations of DEPC react with his₅₇ at the active site catalytic triad; furthermore, DEPC below 400 µM enhanced the Hg(II)‐induced precipitation of chymotrypsin. We conclude that his₅₇ reacts readily with DEPC, causing enzyme inhibition and enhancement of Hg(II)‐induced aggregation. Above 500 µM, DEPC inhibited Hg(II)‐induced precipitation, and [DEPC] >2.5 mM completely protected chymotrypsin against precipitation. This suggests that his₄₀ reacts less readily with DEPC, and that chymotrypsin denaturation is caused by Hg(II) binding specifically to the his₄₀ residue. Finally, we show that Hg(II)‐histidine binding may trigger hemoglobin aggregation as well. Because of results with these two enzymes, we suggest that metal‐histidine binding may be key to understanding all heavy metal‐induced protein aggregation.

Introduction

Some mercury compounds are known to be quite toxic,1 thus mercury is considered a major environmental pollutant.2 It has three oxidation states—0, I, and II—and can be found in three major forms: elemental mercury (0), inorganic mercury (I, II), and organic or organo‐mercury.3 Mercury species differ significantly in properties, but all are toxic.4 Diethylmercury and dimethylmercury are considered “supertoxic”—a fatal dose of dimethylmercury requires absorption of less than two drops (100 µL, or 1.3 mmol).5

A key theme of mercury toxicology is the metabolic conversion of elemental and organo‐mercury into inorganic mercury.1 Elemental mercury vapor and organo‐mercury compounds both have high bioavailability, as they easily permeate biological membranes and thus spread throughout the body. Intestinal microflora and cellular enzymes degrade organo‐mercury compounds (e.g., methylmercury) to Hg²⁺, after which the inorganic Hg²⁺ accumulates in the central nervous system and the kidney.3, 6, 7, 8 The strong association of inorganic mercury with the kidney is traced to binding of Hg²⁺ to proteins containing thiols (e.g., cysteine) as well as thioethers (e.g., methionine).9

The toxic effects exerted by mercury compounds are most often attributed to the tight binding of Hg²⁺ to cysteine‐thiol and seleno‐cysteine residues of critical enzymes (Fig. 1).9, 10, 11, 12 Hg²⁺ binding causes a number of dramatic changes to the SH group: Hg²⁺ is much larger than the H that it replaces; the nucleophilicity of the S atom is dramatically reduced; and loss of H makes the group nonacidic. Clearly, if the —SH group is key to enzyme activity, it is easy to see how Hg²⁺ binding might well curtail such activity. What is often not appreciated however is that Hg²⁺ (and other heavy metal cations) can also cause proteins to unfold, aggregate, and precipitate out of solution.13, 14, 15, 16, 17, 18, 19 In fact, the Merck Index notes that aqueous mercuric chloride “coagulates albumin.”20 Thus there are two distinct mechanisms by which Hg²⁺ can inhibit enzymes.

Figure 1.

Hg²⁺ binds to two nearby cysteine‐thiols.

Myshkin et al.14, 15, 21, 22, 23 have found that the metal‐induced aggregation of hemoglobin depends on Hg²⁺ binding to a noncysteine side chain along with cysteine‐β93, and that interfering with Hg²⁺‐non‐cysteine binding protects against aggregation.21, 22, 24 Because Hg²⁺ binding to cysteine thiols is already well‐characterized, we set out to study Hg²⁺ inhibition of enzymes that lack accessible cysteine sites. Chymotrypsin, a digestive enzyme found in the small intestine, hydrolyzes protein peptide bonds,25, 26 and it lacks free cysteine‐thiols because its ten cysteines are oxidized to five disulfide bridges and are thus effectively protected from mercury binding.26 Chymotrypsin is thus a good model for studying Hg²⁺ inhibition due to interaction with non‐cysteine amino acid side chains. Furthermore, its hydrolytic activity is easily assayed using chromophoric peptide substrates.

In the absence of accessible cysteine side chains, histidine is a potential binding candidate for Hg²⁺. Applying hard‐soft acid base theory, mercury, a soft acid, should have a moderate affinity for nitrogen‐based molecules.27 The formation constants (logK _f) for various 1:1 Hg²⁺:ligand complexes are listed in Table 1. Although Hg‐N K _fs are 7‐11 orders of magnitude lower than Hg‐S K _fs, Hg²⁺ still has a moderate affinity for N ligands. Specifically, K _f for the Hg²⁺‐histidine and Hg²⁺‐histidine‐methyl ester complexes are 10^7. 28 and 10^5. 33, respectively. Similarly, results from isothermal titration calorimetry yield K _f = 10^7.85 for Hg²⁺‐histidine (cf., 10^11.76 for Hg²⁺‐cysteine).30 Hence it is reasonable to expect Hg²⁺ to bind to histidine, as well as cysteine, side chains.

Table 1.

Formation Constants (logK _f) for Hg²⁺‐ligand 1:1 Complexes, from Ref.28 and Ref.29

S‐Ligand	Log K f	N‐Ligand	Log K f
S2‐	52.7	NH3	8.8
Mercaptosuccinic acid	16.3	CH3NH2	8.66
Mercaptoacetic acid	15.9	Pyridine	5.1
β‐mercaptoethanol	15.1	Imidazole	3.57
Cysteine	14.4	Histidine	7.28
		Histidine‐Me‐ester	5.33

Chymotrypsin has two histidines: his₅₇ is part of the active site catalytic triad, and his₄₀ lies near the active site. The effects of Hg²⁺ on chymotrypsin can be easily assayed through two means: hydrolytic activity, and denaturation leading to aggregation. Mercury‐denatured chymotrypsin aggregates, causing an increase in turbidity; the increased light scattering is easily measured by monitoring low wavelength absorbance,31 for example at 350 nm, and this effective absorbance has been shown to correlate directly (i.e., linearly) with the amount of aggregated/precipitated protein.32 The hydrolytic activity of chymotrypsin can be assayed spectrophotometrically at 410 nm using the chromophoric substrates p‐nitrophenyl acetate (an ester) or N‐succinyl‐l‐phenylalanine‐p‐nitroanilide (an amide), as depicted in Figure 2. Both the amide and ester substrates are hydrolyzed by chymotrypsin to form the yellow chromophores p‐nitroaniline and p‐nitrophenoxide, respectively (Fig. 2).

Figure 2.

Chymotrypsin chromophoric substrates: Left, the amide N‐succinyl‐l‐phenylalanine‐p‐nitroanilide, and right, the ester p‐nitrophenyl acetate. Hydrolytic cleavage sites are indicated by red ∼, and the chromophoric products are boxed (blue).

Results

Hg²⁺‐induced precipitation of chymotrypsin is moderately inhibited by acetate

In the presence of Hg²⁺, chymotrypsin denatures and aggregates; the resulting particles scatter light, causing an increase in turbidity which can be assayed by the increase in absorbance at low wavelengths, for example, 350 nm (Fig. 3). The increase in A ₃₅₀ with time could almost always be fit to double first‐order kinetics [Eq. (5)]. For the control experiment in Figure 3(A) (without added acetate, black curve), k _fast = 13.28 ± 0.07 min⁻¹, ΔA _350,fast = 0.3656 ± 0.0013, and A _final = 1.0045 ± 0.0002.

Figure 3.

Acetate moderately inhibits Hg²⁺‐induced aggregation of chymotrypsin. [chymotrypsin] = 20 µM, 50 mM MOPS at pH 7.23, [Hg²⁺] = 1.025 mM. (A) A₃₅₀ transients are fit to double‐first order kinetics, that is, a fast first order phase followed by a slower first order phase (Eq. (5), dashed lines). (B) Data were fit to Eq. (3), but modified to decline to a non‐zero value (k _fast,final) at infinite [acetate]. Parameters for the fitted curve are: k _fast,0 = 14 ± 4 µM min⁻¹, k _fast,final = 6.9 ± 2.5 µM min⁻¹, and IC₅₀ = 30 ± 40 µM, R ² = 0.71. Uncertainties in the first order fitted values of k are smaller than the point symbols.

Because chymotrypsin has no cysteine thiols available to bind with Hg²⁺ (all ten of chymotrypsin's cysteines are oxidized to disulfide bridges), the most likely side chains to interact with Hg²⁺ are histidine imidazoles and glutamate/aspartate carboxylates. To determine whether Hg²⁺ can interact with these groups, we examined whether the addition of free acetate or imidazole could inhibit the Hg²⁺‐induced chymotrypsin precipitation. In Figure 3(A) we see that the addition of 100 µM acetate decreased k _fast by 45% (to 7.3 min⁻¹), and increased A _350,max by 27% (to 1.28). In other words, added acetate slows down the rate of the fast phase of the precipitation reaction, but increases the overall amount of precipitation. The acetate‐dependence of k _fast plotted in Figure 3(B) shows a maximum 50% decline in k _fast, with IC₅₀ ≈ 30 µM. Above 200 µM acetate, chymotrypsin aggregation slowed to the point that the absorbance transients were sigmoidal (data not shown). Thus, high concentrations of acetate cause a moderate decrease in the rate of Hg²⁺‐induced chymotrypsin precipitation.

Utilizing EDC and NHS to block carboxylates from binding to Hg²⁺

To further explore the nature of the interaction between Hg²⁺ and chymotrypsin carboxylate side chains, we attempted to block these sites using the carbodiimide EDC, with and without N‐hydroxysuccinimide (NHS, which forms a stable intermediate that enhances the EDC‐carboxylate reaction).33, 34, 35 EDC at 50 mM slowed the Hg²⁺‐induced chymotrypsin precipitation such that the absorbance transients became sigmoidal, as opposed to first order (data not shown). Furthermore, 5 mM EDC with 12.5 mM NHS also reduced the rate of precipitation to give sigmoidal absorbance transients (data not shown).

Although the results above suggested that EDC blocking of chymotrypsin carboxylates inhibited Hg²⁺‐induced precipitation, control experiments [Supporting Information Fig. S4, Eq. (S1)] showed that this effect was in fact due to direct interaction between EDC and Hg²⁺, thus keeping the Hg²⁺ from binding to chymotrypsin. It seems likely that EDC serves as a divalent chelator of Hg²⁺, and thus is not an effective carboxylate labeling reagent in the presence of Hg²⁺.

Hg²⁺‐induced precipitation of chymotrypsin is completely inhibited by imidazole

Imidazole dramatically inhibits Hg²⁺‐induced chymotrypsin precipitation (Fig. 4). Compared to the fast rate constant for chymotrypsin precipitation in the absence of imidazole (13.28 ± 0.07 s⁻¹), k _fast is 94% lower in 500 µM imidazole (0.902 ± 0.003 s⁻¹); the fitted curve in Figure 4 shows IC ₅₀ for imidazole = 78 µM. Unlike acetate, high concentrations of free imidazole bind Hg²⁺ sufficiently tightly to protect completely against chymotrypsin precipitation. Interestingly, we found that at ≥1 mM imidazole, Hg²⁺ caused precipitation in the absence of chymotrypsin; at ≥4 mM imidazole the precipitation reaction was so fast that the rate could not be measured. Hence at imidazole:Hg²⁺ stoichiometric ratios ≥ 1:1, an insoluble precipitate is formed.

Figure 4.

Imidazole inhibits the rate of Hg²⁺‐induced chymotrypsin aggregation. Data were fit to Eq. (3). Parameters for the fitted curve are: k _fast,0 = 13.1 ± 0.8 µM min⁻¹, and IC ₅₀ = 78 ± 21 µM; R ² = 0.94. [chymotrypsin] = 20 µM, 50 mM MOPS at pH 7.23, [Hg²⁺] = 1.025 mM. Uncertainties in the first order fitted values of k _fast are smaller than the point symbols.

Hg²⁺‐induced precipitation of chymotrypsin is pH dependent

Comparing Figure 4 with Figure 3, we learn that imidazole interacts much more effectively with Hg²⁺ than acetate does, protecting against chymotrypsin precipitation. This is not surprising, considering the 14 order of magnitude difference for mercury binding to these two ligands: logK _f for Hg(OAc)₂(s) is 2.5 (calculated from its solubility of 0.09 M at 25°C), compared to 16.7 for the 1:2 mercury(II)‐imidazole complex.36 From these results we concluded that Hg²⁺ is more likely to bind strongly to chymotrypsin's two histidines than to its aspartates or glutamates. We therefore studied the pH‐dependence of the Hg²⁺‐induced chymotrypsin precipitation reaction (Fig. 5) to see if the reaction was enhanced above pH 6, the nominal aqueous pK _a of histidine's imidazole side chain. Sure enough, although turbidity as measured by A₃₅₀ was undetectable below pH 6.2 (green curve in Fig. 5), it was readily observed above pH 6.5; between pH 6.5–7.0 (blue curve), A₃₅₀ increased sigmoidally with time. The lag time characteristic of sigmoidal kinetics is common for aggregation processes that proceed via a rate‐determining nucleation step.37, 38, 39 From pH 7.1 to 8.9 (black curve), the increase in A ₃₅₀ was very fast and first order. Above pH 9 (red curve), the absorbance transients indicated a moderately fast first order increase in A ₃₅₀, followed by a slower first order decrease. The slow decrease in light scattering above pH 9 was due to the sedimentation of highly aggregated particles and their removal from the path length of the cuvette (data not shown). The A ₃₅₀ transients in Figure 5 demonstrate that the rate of Hg²⁺‐induced chymotrypsin denaturation/aggregation increases dramatically in the pH 6.5–7.1 range, commensurate with the nominal aqueous pK _a of histidine's imidazole side chain.

Figure 5.

pH dependence of Hg²⁺‐induced precipitation of chymotrypsin. Buffers are: MES (green and blue); CAPS (red), and MOPS (black). Final concentrations: 20 µM chymotrypsin, 1.025 mM Hg²⁺, and all buffers are 50 mM.

Control experiments were performed to demonstrate that the turbidity that we observed was caused by the interaction between Hg²⁺ and chymotrypsin. No change in A ₃₅₀ was observed at any pH in the absence of Hg²⁺ or chymotrypsin, and increases in the concentration of either Hg²⁺ or chymotrypsin led to increases in both the rate and extent of A ₃₅₀ rise (Supporting Information Fig. S5). Precipitation was generally independent of buffer type, except for Tris and glycine. Tris buffer inhibited chymotrypsin precipitation (relative to other buffers at the same pH); such inhibition was expected, due to the ability of Tris (an amine compound) to bind tightly to Hg²⁺.21, 40 On the other hand, glycine buffer enhanced the precipitation to a degree that the rate was too fast to be measured. These buffers were avoided in future experiments due to their specific interactions with chymotrypsin and/or Hg²⁺.

We also performed some Hg²⁺‐chymotrypsin precipitation experiments in the absence of buffer (Fig. 6). Below pH 6.4, the precipitation first order rate constant was low and roughly pH‐insensitive (k = 0.41 ± 0.19 min⁻¹), but above pH 6.4, k increased dramatically, reaching 2.75 min⁻¹ (almost seven‐fold higher) at pH 6.95. These results, like those in Figure 5, suggested that Hg²⁺‐induced aggregation depends on the deprotonation of an acidic group with a pK _a in the 6.5–7.0 range, again implicating histidine.

Figure 6.

pH dependence of Hg²⁺‐induced precipitation of chymotrypsin, in the absence of buffer. [Chymotrypsin] = 20 µM, and [Hg²⁺] = 1.0 mM.

Activity of chymotrypsin remaining in solution after Hg²⁺‐induced precipitation

We have demonstrated that Hg²⁺ caused chymotrypsin denaturation and aggregation, but by centrifuging our samples and probing the supernatant, we were also able to show that Hg²⁺ inhibited the enzyme that remained in solution. To compare these two effects, we plotted both the percent of chymotrypsin remaining in solution and the relative activity of chymotrypsin remaining in solution, as a function of the concentration of Hg²⁺ (Fig. 7: N‐succinyl‐l‐phenylalanine p‐nitroanilide, an amide substrate, was used to assay the activity of chymotrypsin). It can be seen in Figure 7 that at 0.2 mM Hg²⁺, when more than 90% of chymotrypsin still remained in solution, the enzyme's hydrolytic activity was reduced by over 75%. This trend is very clear at the lower concentrations of Hg²⁺ (0.05–0.3 mM); at 0.5 mM Hg²⁺ and above, very little chymotrypsin remained in solution and what little did remain was completely inhibited (0% activity). The same trends were observed with the ester substrate pNPA (Supporting Information Fig. S7).

Figure 7.

Hg²⁺ decreased how much chymotrypsin remained in solution, and chymotrypsin's activity. Each point is the average of three measurements, and y axis error bars = standard deviation; the activity of each set was normalized to the 0 mM concentration for that day. [chymotrypsin] calculated from A₂₈₁ and ε₂₈₁. Prior to Hg²⁺ additions, solutions contained 50 mM MOPS buffer, pH 7.14 and initial [chymo] = 20 µM. Final [succinyl phenylalanine] = 3 mM.

Although Hg²⁺ forms a coordinate (or dative) covalent bond with imidazole, we found that it behaved similar to a reversible noncompetitive inhibitor. Figure 8 shows the results of Michaelis–Menten experiments in the presence of 0–0.3 mM Hg²⁺, after removal of the precipitated enzyme; fitted Michaelis–Menten parameters are presented in Table 2. K _m values ranged from 1 to 3 mM, and the differences between them were not statistically significant. On the other hand, k _cat declined significantly: Compared to the control value, k _cat was two‐fold lower in 0.1 mM Hg²⁺, and six‐fold lower in 0.2 mM Hg²⁺. (Note that at these two Hg²⁺ concentrations, 95% and 88% of the enzyme remained in solution. Even at 0.3 mM Hg²⁺, 72% remained in solution.) A decline in k _cat with unaltered K _m resembles the effect of a reversible noncompetitive inhibitor, one that leaves substrate binding unaltered while inhibiting the subsequent catalytic step. The experiment in the presence of 0.3 mM Hg²⁺ was anomalous, seeming to show an inordinately high K _m (black Xs curve). Fitting to the Michaelis–Menten equation was unsuccessful, with P > 1 for V _max and K _m, indicating statistical insignificance; in other words, the points in this curve do not fit hyperbolic saturation.

Figure 8.

Michaelis–Menten plot of chymotrypsin in the presence of Hg²⁺, in 50 mM MOPS, pH 7.14. Substrate was the amide succ‐phe, data were fit to Eq. (2). Black circles = control (without Hg²⁺), blue squares = 0.1 mM Hg²⁺, green diamonds = 0.2 mM Hg²⁺, black Xs = 0.3 mM Hg²⁺. Final [chymotrypsin] remaining in solution, and all fitted Michaelis–Menten parameters are listed in Table II. For the 0.3 mM Hg²⁺ curve, V _max and K _m returned P values ≫ 1, indicating that these points do not fit hyperbolic saturation.

Table 2.

Michaelis–Menten Parameters for Chymotrypsin‐catalyzed Hydrolysis of Amide Substrate in the Presence of Hg²⁺

[Hg2±], mM	[chymo] in solution, µM	V max, µM/min	k cat, min−1	K m, mM	R 2
0	40.5 ± 0.7	10.5 ± 1.1	0.259 ± 0.028	1.5 ± 0.4	0.95
0.103	38.4 ± 1.2	5.2 ± 1.8	0.13 ± 0.05	2.9 ± 1.8	0.86
0.205	35.7 ± 1.8	1.56 ± 0.27	0.044 ± 0.008	1.0 ± 0.5	0.77
0.308	29.1 ± 2.7	NSSa	NSS	NSS	0.88

^aNSS = not statistically significant.

Histidine blocking with diethylpyrocarbonate

To further characterize the importance of chymotrypsin's histidines in Hg²⁺‐induced precipitation, we employed DEPC as a histidine‐blocking reagent. In kinetic studies we determined that after low‐concentration DEPC incubations of 5–10 min, chymotrypsin activity declined by 68% to a constant low rate (Supporting Information Fig. S6). This histidine blocking reaction followed first order kinetics, with k = 0.43 min⁻¹ (Supporting Information Fig. S6).

Histidine blocking by DEPC inhibited chymotrypsin's hydrolytic activity quite effectively, with an IC ₅₀ of 35 µM (Fig. 9). At 700 µM DEPC, enzyme activity was 83% inhibited, from v ₀ (control) = 4.8 µM min⁻¹ down to 0.4 µM min⁻¹ (Fig. 9).

Figure 9.

DEPC inhibitor profile. 40 µM chymotrypsin in 50 mM MOPS at pH 7.14, incubated with DEPC for 10 min, after which, the substrate pNPA was added to a final concentration of 2.9 mM. Data were fit to Eq. (3): v _o (control) = 4.98 ± 0.27 µM min⁻¹; IC ₅₀ = 35 ± 6 µM; R ² = 0.97. Uncertainties in v ₀ values (obtained from ΔA ₄₁₀/Δt slopes) are smaller than the point symbols.

DEPC concentration had a biphasic effect on the Hg²⁺‐induced precipitation of chymotrypsin (Fig. 10). Up to 500 µM, DEPC enhanced precipitation, increasing A _350,max by 30–35%; above 500 µM DEPC inhibited precipitation, and 2500 µM DEPC afforded complete protection against precipitation.

Figure 10.

Effects of DEPC on Hg²⁺ induced chymotrypsin precipitation. Final cuvette concentrations: 40 µM chymotrypsin, 50 mM MOPS, pH 7.14, 0.410 mM HgCl₂. Uncertainties in the first order fitted values of A_350,max are smaller than the point symbols.

Discussion

Mercury and other toxic heavy metals disrupt cellular function by inhibiting or denaturing critical enzymes and proteins. Within these proteins, Hg²⁺ binds very tightly to cysteine or selenocysteine residues9, 10, 11, 12 with K _f ≈ 10¹⁴−10¹⁶ (Table 1). However, Hg²⁺ also binds fairly tightly to imidazoles and amines (K _f ≈ 10⁵−10⁹, Table 1), as well as carboxylates7, 8 (K _f ≈ 10³). The binding of Hg²⁺ to critical protein side chains can cause inhibition by changing the active site structure, or by causing general unfolding.10, 13, 41 Mercury is well‐known for its disruption of the active sites of a number of important anti‐oxidant enzymes and compounds (e.g., methionine synthase, thioredoxin reductase, glutathione reductase, glutathione peroxidase).3, 10, 12 It is less widely recognized that a number of divalent metal cations like Hg²⁺, Cd²⁺, Pb²⁺, Ni²⁺, and Zn²⁺ can bind to proteins, causing them to unfold, aggregate, and precipitate out of solution.13, 15, 16, 17, 18 Studies by the Myshkin group on oxy‐ and met‐hemoglobin showed that the binding of Hg²⁺ to cysteine thiols triggered surprisingly small changes in protein conformation, which nevertheless led to denaturation and aggregation.14, 15, 21 In the presence of substoichiometric concentrations of Hg²⁺, both mercurated and non‐mercurated hemoglobin were present, and though the mercurated protein more readily aggregated, the non‐mercurated form also participated in the formation of the aggregates.14 Finally, Hg²⁺ chelators like EDTA and thiourea protected hemoglobin against denaturation and precipitation.24

We found in this study that chymotrypsin, like hemoglobin, aggregates in the presence of Hg²⁺, and that this Hg²⁺‐induced precipitation can be inhibited by chelators [e.g., acetate, Fig. 3; imidazole, Fig. 4; and the carbodiimide EDC, Supporting Information Fig. S4 and Eq. (S1)]. Unlike hemoglobin, chymotrypsin lacks free cysteine thiols, as all ten of its cysteines are oxidized into disulfide bridges. Although Hg²⁺ can bind to one of the two S atoms in an S—S bond, such binding is generally weak in the absence of additional intramolecular chelation sites42, 43 (see section 9 in the Supporting Information for further discussion of this point). Hg²⁺ binding sites within chymotrypsin are therefore likely to be either aspartate/glutamate carboxylate side chains, or histidine imidazole side chains.

Our competition results showed that although free acetate inhibited Hg²⁺‐induced chymotrypsin precipitation moderately (40% decline in rate constant, IC₅₀ = 30 µM, Fig. 3), inhibition by free imidazole was much more effective (>94% decline in rate constant, IC₅₀ = 78 µM, Fig. 4). Our results (94% aggregation rate decline at a 25:1 imidazole:enzyme mole ratio) matched those of Myshkin and Khromova, who used histidine to inhibit hemoglobin aggregation (93% rate decline at a 30:1 histidine:enzyme mole ratio).22 Finally, at 1 mM imidazole, prior to the addition of chymotrypsin, Hg²⁺ and imidazole formed a precipitate. As the concentration of imidazole increased from 1 to 10 mM, the rate of precipitation increased dramatically. This indicated that when the imidazole:Hg²⁺ molar ratio is 1:1 or higher, a neutral mercury‐imidazole salt quickly precipitated out of solution. Imidazole‐Hg²⁺ precipitation is discussed further in section 10 of the Supporting Information and Figure S10.

Only a few authors have examined the influence of pH on metal‐induced protein aggregation. Mantyh et al.19 studied the aggregation of the β‐amyloid peptide induced by Fe³⁺, Al³⁺, and Zn²⁺, and found enhanced aggregation above pH 3–5, with pK _a values around 4–6 for the aggregation increase. This suggested metal interaction with the following residue side chains within the β‐amyloid peptide: aspartate, glutamate, or histidine with a lowered pK _a. Similar to Mantyh et al., we found that higher pH favored Hg²⁺‐induced chymotrypsin precipitation: Figures 5 and 6 show that the chymotrypsin precipitation rate increased dramatically above pH 6.4, indicating the importance of a deprotonated acidic side chain with a pK _a around 6.5. Given that the nominal pKa of aqueous histidine is 6–7, all of these results suggest that the denaturation/aggregation of chymotrypsin is triggered by Hg²⁺ binding to deprotonated histidine side chain(s).

On the other hand, Myshkin and Khromova found that the rate of Hg²⁺‐induced aggregation of hemoglobin declined with increasing pH over the 7.2–8.1 range.23 They interpreted this as OH⁻ binding to Hg²⁺ and hindering its ability to bind to the protein. However, these same authors showed that Tris buffer binds to Hg²⁺, displacing one of its protein ligands and inhibiting aggregation.21, 22 Because the pK _a of Tris is 8.1, the concentration of its deprotonated conjugate base increases four‐fold over the 7.2 to 8.1 pH range, from 12 to 51% of total buffer. Thus, it seems more likely that the increase of deprotonated Tris and its subsequent increased binding to Hg²⁺ causes the aggregation inhibition. This Tris buffer effect, and the fact that hemoglobin contains two accessible cysteine thiols (β‐93) that are known to bind Hg²⁺, whereas chymotrypsin lacks reduced thiols, undoubtedly explain our difference in pH effects.

We have expanded the work of previous authors13, 14, 15, 17, 18, 21, 24 by studying not only enzyme aggregation induced by Hg²⁺, but also how Hg²⁺ inhibited the enzyme remaining in solution. In Figure 7, the discrepancy between percent chymotrypsin remaining in solution and percent relative activity revealed that at concentrations ≤0.2 mM, Hg²⁺ inhibited chymotrypsin in solution without causing precipitation. Michaelis–Menten plots (Fig. 8) showed that Hg²⁺ binding to aqueous chymotrypsin lowered k _cat while leaving K _m essentially unchanged. We conclude from these results that Hg²⁺‐histidine binding, although irreversible, results in enzyme inhibition similar to that of a reversible noncompetitive inhibitor. In other words, in the presence of Hg²⁺, chymotrypsin‐substrate binding is unaltered, while catalysis at the active site is shut down. This observation is supported by Henderson, who found that methylating chymotrypsin's histidines left substrate affinity (K _m) unaltered.44 Above 0.2 mM, Hg²⁺ caused chymotrypsin to denature, resulting in aggregation and precipitation of the enzyme. At 0.5 mM Hg²⁺ and above, almost no chymotrypsin remained in solution, and hydrolytic activity was nil (Fig. 7).

As previously mentioned, chymotrypsin contains two histidines: his₅₇ is a member of the active site catalytic triad, whereas his₄₀ is located just a few Ångstroms behind the active site, through a crevice (Fig. 11). For example, his₄₀ is close enough to hydrogen‐bond to the active site aspartate₁₉₄ side chain when the N‐terminus is deprotonated at pH > 9, causing enzyme inhibition.45, 46 Within chymotrypsin's optimal pH range of 6–8, his₄₀ does not influence the active site, whereas deprotonated his₅₇ base‐catalyzes the critical ser₁₉₅ nucleophile.45 The fact that Hg²⁺ ≤0.2 mM dramatically inhibits chymotrypsin activity (Figs. 7 and 8) indicates that Hg²⁺ may bind readily to his₅₇, leaving it unable to participate in the catalytic mechanism of chymotrypsin [Fig. 12(A)]. Mirroring our results, Henderson showed that methylation of his₅₇ alone caused a 200,000‐fold decrease in k _cat (with relatively little change in K _m).44 Our modest six‐fold decrease in k _cat in the presence of 0.2 mM Hg²⁺ is undoubtedly due to the moderate affinity and accessibility of his₅₇, leaving most enzyme molecules mercury‐free.

Figure 11.

Structure of α‐chymotrypsin (PDB# 4Q2K): (A) front view; (B) rear view.

Figure 12.

Consequences of Hg²⁺‐histidine binding in chymotrypsin. (A) Low concentrations of Hg²⁺ (≤0.2 mM) preferentially bind to chymotrypsin's his₅₇, causing enzyme inhibition, but not aggregation. (B) High concentrations of Hg²⁺ (>0.2 mM) result in binding to both his₅₇ and his₄₀ of chymotrypsin; binding to his₄₀ causes the enzyme to aggregate. (C) Low concentrations of DEPC (<0.5 mM) block his₅₇ with an R group; this shifts Hg²⁺ binding to his₄₀, causing enhanced precipitation. High concentrations of DEPC block both his₅₇ and his₄₀, protecting against Hg²⁺ binding and precipitation.

At and above 0.5 mM Hg²⁺, chymotrypsin precipitates out of solution almost completely (<8% remaining), and catalytic activity of the remaining chymotrypsin falls to zero (Fig. 7). Because chymotrypsin only has two histidines, it is tempting to hypothesize that Hg²⁺ binds differentially to the two histidine imidazole side chains. We propose that his₅₇ is more reactive to Hg²⁺ than is his₄₀. This could be due to steric accessibility, or local protein environment polarity or electrostatics. According to this hypothesis, at ≤0.2 mM, Hg²⁺ binds to his₅₇, causing enzyme inhibition but no aggregation [Fig. 12(A)]; above 0.2 mM, Hg²⁺ binds to his₄₀, causing denaturation and aggregation of the enzyme [Fig. 12(B)].

Our model depicted in Figure 12 suggests that inhibition of chymotrypsin activity should be traceable to Hg²⁺ binding to two different sites: At the first site, his₅₇, Hg²⁺ binding yields noncompetitive inhibition; binding at the second site, his₄₀, leads to denaturation, aggregation, and complete loss of activity. Such two‐site inhibition is sometimes referred to as “parabolic” inhibition, and can be characterized by a Dixon plot (1/v ₀ vs. [inhibitor]₀) that slopes upward parabolically.47 These data are fit to a second order polynomial (1/v ₀ = a + b[I]₀ + c[I]₀ ²), where the polynomial coefficients a, b, and c are related to the Michaelis–Menten parameters [see Supporting Information section 7, Eq. (S2)]. Converting our data from Figure 7 (black curve) to the 1/v₀ vs. [Hg²⁺] Dixon format (Supporting Information Fig. S8), we see that the data do indeed fit well to a second order polynomial (R ² = 0.985). This adds support to our model connecting chymotrypsin inhibition to the binding of Hg²⁺ to two distinct sites within the enzyme.

Further support for the hypothesis presented in Figure 12 is found in experiments in which we blocked the reactive nitrogen of histidine's imidazole side chain by reacting it with diethylpyrocarbonate. DEPC reacts with histidine to form a covalent adduct (Supporting Information Fig. S9) to which Hg²⁺ cannot bind; the formation of this adduct decreases the hydrolytic activity of the enzyme and also alters its Hg²⁺‐induced precipitation. The first order rate constant (k) for DEPC's reaction with chymotrypsin histidines is 0.43 ± 0.23 min⁻¹, and enzymatic activity declined to its final inhibited value after 5–10 min (Supporting Information Fig. S6). The concentration profile showed that DEPC inhibited chymotrypsin activity with an IC ₅₀ of 35 ± 6 µM; 700 µM inhibited the enzyme by 83% (Fig. 9).

The effect of histidine blocking by DEPC showed an interesting biphasic concentration dependence (Fig. 10). Up to about 500 µM DEPC, the extent of Hg²⁺‐induced precipitation of chymotrypsin was enhanced, with A_350,max rising 30–35%. Above 500 µM DEPC chymotrypsin precipitation declined, with 2500 µM DEPC offering complete protection. These results are reminiscent of what we observed regarding Hg²⁺ binding to chymotrypsin's two histidines (Fig. 12). Here we posit that DEPC reactivity with chymotrypsin mirrors that of Hg²⁺, reacting preferentially with his₅₇ at low DEPC concentrations (below 500 µM), then reacting with his₄₀ at higher concentrations [Fig. 12(C)]. Below 500 µM DEPC, blocking of his₅₇ causes partial inhibition of chymotrypsin activity (Supporting Information Fig. S6), and subsequently added Hg²⁺ is then forced to bind to his₄₀, which enhances chymotrypsin precipitation [Fig. 12(B,C)]. Above 500 µM DEPC, prior blocking of his₄₀ (along with his₅₇) leaves Hg²⁺ without any imidazole side chains to bind to, thus completely protecting the enzyme from denaturation and aggregation [Fig. 12(B,C)]. Finally, we note that for prior incubations above 1.25 mM DEPC, subsequent Hg²⁺‐induced precipitation was very slow: It took from 1.5 to 2.5 h for the maximum A₃₅₀ to be reached. These slower rates were undoubtedly due to the paucity of available histidines left for binding to Hg²⁺.

Finally, we wish to offer a new interpretation of Myshkin et al.'s results showing that exogenous histidine,22 and the amines Tris,21 EDTA, and thiourea,24 all inhibited the Hg²⁺‐induced aggregation of hemoglobin. Because of the kinetics of the aggregation reaction in the presence of these nitrogen‐based inhibitors, Myshkin et al. concluded that Hg²⁺ binding to four protein ligands induces aggregation: cysteine‐β93, two molecules of Tris buffer, and a non‐cysteine protein side chain. When a nitrogen‐based ligand (e.g., thiourea, EDTA, exogenous histidine or a third molecule of Tris buffer) displaces the non‐cysteine side chain, the Hg²⁺‐hemoglobin complex cannot form aggregates that are large enough to precipitate out of solution. Myshkin et al. tentatively identify this non‐cysteine Hg²⁺‐binding site as aspartate‐β94, based on a statement made by Garel et al.48

We note here two problems with this imputed role for aspartate‐Hg²⁺ binding: (1) Hg²⁺‐oxide binding is quite weak (e.g., the solubility of mercuric acetate at room temperature is almost 0.1 M); recall how weakly acetate inhibited Hg²⁺‐induced chymotrypsin aggregation [Fig. 3(B)]; and (2) Garel et al. cited Muirhead and Greer49 in support of their identification of a key “salt bridge between aspartate‐β94 and the carboxyl[ate] of the terminal histidine [β146].”48 The problem is that Garel et al. misinterpreted the following sentence: “The carboxyl[ate] group (labeled COO⁻) and the terminal histidine side chain, his‐β146, are also clearly visible; the latter bonding to the γ‐carboxyl[ate] of aspartate‐β94 in the same [polypeptide] chain.”49 Correct interpretation of the word “latter” shows that Muirhead and Greer identify histidine‐β146's imidazole side chain, and not its terminal carboxylate group, as aspartate‐β94's partner in this key salt bridge. In fact, Muirhead and Greer go on to implicate this histidine/imidazole‐aspartate/carboxylate salt bridge as “responsible for the… Bohr effect,” with its pK _a of 7: Protonated histidine‐β146/imidazole forms this extra salt bridge with aspartate‐β94 in the deoxy form, stabilizing it and thus inhibiting oxygen binding. From this corrected reading of Muirhead and Greer, hemoglobin's non‐cysteine Hg²⁺ ligand is likely to be the imidazole side chain of histidine‐β146, rather than the carboxylate group of this residue or aspartate‐β94. All of this leads us to suggest that Myshkin et al.'s results, like ours, point to the key role of Hg²⁺ binding to the histidine imidazole side chain in triggering protein denaturation and aggregation. This in turn suggests that metal‐histidine binding may be key to understanding all metal‐induced protein aggregation.

Conclusion

Most studies on enzyme inhibition by Hg²⁺ focus on its binding to cysteine and selenocysteine side chains; we have shown that in chymotrypsin (and likely in hemoglobin), Hg²⁺ binds to histidine imidazole side chains. We have further shown that at low concentrations, both Hg²⁺ and DEPC bind preferentially to chymotrypsin's his₅₇, causing inhibition of native (aqueous) enzyme; at high concentrations, Hg²⁺ and DEPC bind additionally to his₄₀, and the Hg²⁺‐his₄₀ form of the enzyme unfolds and aggregates. To our knowledge, we have presented the first data that distinguish between the inhibitory effects of blocking each of chymotrypsin's two histidines. The ability of Hg²⁺ to bind to residues other than cysteine or selenocysteine could implicate more enzymes in the toxicity of Hg²⁺. Furthermore, the key role of this non‐cysteine‐Hg²⁺ binding in enzyme aggregation suggests that weak Hg²⁺ ligands modeled after imidazole or simple amines might help with in vivo mercury detoxification.

Materials and Methods

A single aqueous stock solution of HgCl₂ (10.253 mM) was used over a 12‐month period, stored in a glass bottle at room temperature. MOPS (4‐Morpholinepropanesulfonic acid) buffer stock solutions were prepared at pH 7.23 and pH 7.14. Chymotrypsin (Sigma‐Aldrich C4129) stock solutions (1 mM) were prepared daily in deionized H₂O. Stock solutions of p‐nitrophenyl‐acetate (pNPA, Sigma‐Aldrich N8130) and N‐succinyl‐l‐phenylalanine p‐nitroanilide (Sigma‐Aldrich S2628) were also prepared daily in absolute ethanol and DMSO, respectively. Diethyl pyrocarbonate (DEPC, Sigma–Aldrich D5758) was prepared in acetonitrile. N‐(3‐dimethyl‐aminopropyl)‐N′‐ethylcarbodiimide hydrochloride (EDC, Sigma–Aldrich E6383) and N‐hydroxysuccinimide (NHS, Sigma–Aldrich 130672) solutions were prepared in deionized water.

Molar absorptivities of p‐nitrophenoxide, p‐nitroaniline, and chymotrypsin

p‐nitrophenylacetate hydrolyzes to acetate, 2 H⁺, and the visible chromophore p‐nitrophenoxide. The molar absorptivity of p‐nitrophenoxide was determined at its λ_max, 400 nm, and also at 410 nm, where some literature measurements are made; we completely hydrolyzed p‐nitrophenyl‐acetate solutions ranging in concentration from 10 to 120 µM. Averaging six calibration curves gave ɛ ₄₀₀ = 16200 ± 700 M⁻¹ cm⁻¹, and ɛ ₄₁₀ = 15300 ± 600 M⁻¹ cm⁻¹ (± standard error).

N‐succinyl‐l‐phenylalanine‐p‐nitroanilide is hydrolyzed to l‐phenylalanine‐N‐succinate, H⁺, and the visible chromophore p‐nitroaniline. The molar absorptivity of p‐nitroaniline was determined at its λ_max, 410 nm; we completely hydrolyzed N‐succinyl‐l‐phenylalanine‐p‐nitroanilide solutions ranging in concentration from 10 to 240 µM. Averaging the results of six calibration curves gave ɛ₄₁₀ = 6600 ± 600 M⁻¹ cm⁻¹.

To measure the concentration of chymotrypsin in solution, its molar absorptivity was determined at λ_max = 281 nm, using standard solutions ranging from 1 to 50 µM. Averaging the results of three calibration curves gave ɛ₂₈₁ = 44300 ± 500 M⁻¹ cm⁻¹.

Free acetate and free imidazole protect chymotryspin from Hg²⁺ induced precipitation

Concentrations of imidazole or acetate, ranging from 0‐1000 µM, were incubated with 1.025 mM Hg²⁺ in 50 mM MOPS, pH 7.23, for 2 min. Chymotrypsin (20 µM) was then added and turbidity increase was monitored at A ₃₅₀. A ₃₅₀ versus time transients were fit to first order kinetics at each [imidazole].

EDC (and EDC + NHS) blocking of carboxylate side chains

Carbodiimides such as EDC react with carboxylate groups. Chymotrypsin (20 µM) was incubated with EDC (0–20 mM) for 15 min, followed by the addition of MES (pH 7.15, 50 mM final concentration) and HgCl₂ (1.025 mM final concentration). Turbidity increase was monitored at A₃₅₀. Because the reaction between EDC and NHS results in a more stable intermediate that reacts more effectively with carboxylates, the experiment above was also performed including NHS at a concentration 2.5x that of EDC.33, 34, 35

Control experiments were performed to determine whether protection against Hg²⁺‐induced chymotrypsin precipitation was due to a direct reaction between Hg²⁺ and the side chain labeling reagents EDC/EDC + NHS (reacts with aspartate/glutamate) and DEPC (reacts with histidine). Hg²⁺ was incubated with each of the above reagents for 30 min, and then chymotrypsin and MES buffer were added (final concentrations: 1 mM Hg²⁺, 20 µM chymotrypsin, 50 mM MES, pH 7.15) and A₃₅₀ was monitored for 4–40 min. The reagent concentrations used were those that yielded extensive or complete protection against Hg²⁺‐chymotrypsin precipitation: 5 mM DEPC; 50 mM EDC; and 5 mM EDC + 12.5 mM NHS. Labeling reagents were incubated with Hg²⁺, followed by addition of buffer, deionized water, and finally, chymotrypsin, followed immediately by measurement of A ₃₅₀. This protocol allowed time for the labeling reagents to interact with mercury, but insufficient time for them to react covalently with chymotrypsin side chains. Hence, under these conditions, any decline in chymotrypsin precipitation must be due to direct interactions between Hg²⁺ and the labeling reagents, thus preventing the mercury from binding to chymotrypsin side chains. Conversely, unaltered chymotrypsin precipitation would show that Hg²⁺ still binds to chymotrypsin side chains, thus Hg²⁺ does not interact directly with the labeling reagents.

EDC‐Hg²⁺ direct interaction was also followed spectrophotometrically, as EDC absorbance at 243 nm increases with added Hg²⁺. EDC (20 µM) in 50 mM MES, pH 7.15, was titrated with Hg²⁺ (0–2 mM final concentrations). Because Hg²⁺ also absorbs at 243 nm, the molar absorptivity of Hg²⁺ was determined at 243 nm, and absorbance due to Hg²⁺ was subtracted in order to isolate only EDC absorbance changes. The A_{243,corrected} versus [Hg²⁺] data were plotted and fit to the equation for hyperbolic saturation [Eq. (1)].

pH dependence of Hg²⁺‐induced chymotrypsin precipitation

Buffer stock solutions were 0.2 M, and reactions were prepared with a final buffer concentration of 50 mM, 20 µM chymotrypsin, and 1.025 mM Hg²⁺; final reaction volumes were 1 mL. Light scattering of turbid/aggregated chymotrypsin particles was monitored via the absorbance increase at 350 nm and, when appropriate, A₃₅₀ versus time transients were fit to first order kinetics [Eq. (5)]. Buffers were used within their buffer zone: pK _a – 1 < pH < pK _a + 1. Buffers used were (with pK _a): sodium acetate (4.76); MES (6.15); MOPS (7.15); HEPES (7.55); Tris (8.08); TAPS (8.4); CHES (9.3); glycine (9.9); and CAPS (10.4). pH‐dependence was also studied in unbuffered solutions; in these experiments, the target pH was attained by adding aliquots of 10 mM NaOH or HCl.

Chymotrypsin remaining in solution after Hg²⁺‐induced precipitation

To study the extent of chymotrypsin precipitation, concentrations of Hg²⁺ ranging from 0 to 3.6 mM were incubated with chymotrypsin (20 µM) in 50 mM MOPS, pH 7.14, for a minimum of two hr. Deionized H₂O was added to the Eppendorf tubes to bring the total volume of solution to 1 mL. After the incubation was complete, solutions were microfuged for 5 min at 12,000 rpm. The supernatant was collected and the concentration of chymotrypsin remaining in solution was calculated from A ₂₈₁ using ɛ ₂₈₁ = 44,300 M⁻¹ cm⁻¹; the “percent remaining” was calculated relative to the concentration of chymotrypsin in the control sample (without Hg²⁺). Using a final concentration of 3 mM N‐succinyl‐l‐phenylalanine p‐nitroanilide (from 0.25 M stock in DMSO) chymotrypsin hydrolysis activity was measured as an increase in A ₄₁₀. The rate v ₀ was calculated from the ΔA ₄₁₀/Δt slope using ɛ ₄₁₀ = 6600 M⁻¹ cm⁻¹ for the p‐nitroaniline chromophoric product. Percent activity was determined relative to the activity of the control (without Hg²⁺).

Michaelis–Menten kinetics of Hg²⁺‐inhibited chymotrypsin

To examine how Hg²⁺ inhibits chymotrypsin remaining in solution, the enzyme was incubated with three concentrations of HgCl₂ (0.103, 0.205, and 0.308 mM). For each Hg²⁺ concentration, seven Eppendorf tubes were prepared. Hg²⁺ was incubated with chymotrypsin (20 µM) in 50 mM MOPS, pH 7.14, for a minimum of 2 h. Deionized H₂O was added to bring the total volume of solution to 1 mL. After the incubation was complete, all seven Eppendorf tubes were microfuged for 5 min at 12,000 rpm to pellet out the aggregated chymotrypsin; the supernatants were transferred to seven cuvettes. Two stock solutions of the amide substrate N‐succinyl‐l‐phenylalanine p‐nitroanilide (0.28 and 0.2 M in DMSO) were prepared and aliquots were added to the seven supernatants, giving seven different final substrate concentrations varying from 0.05 to 3.5 mM. Chymotrypsin hydrolytic activity was measured as an increase in A ₄₁₀. The rate, v ₀ (µM/min), was calculated from the ΔA ₄₁₀/Δt slope as before. The rate was plotted against substrate concentration (mM) and fit to the Michaelis–Menten equation [Eq. (2)]. A control Michaelis–Menten plot was also prepared with uninhibited chymotrypsin (no Hg).

Diethyl pyrocarbonate reacts with histidines in chymotrypsin

The reagent diethyl pyrocarbonate (DEPC) was utilized to block chymotrypsin histidines (see Supporting Information Fig. S9). To study the kinetics of DEPC reacting with histidine, the enzymatic activity of a DEPC‐chymotrypsin solution was assayed at various times. A 9.0 µL aliquot of 100 mM DEPC (in 98.5% vol/vol acetonitrile) was added to 720 µL of 1.0 mM chymotrypsin; this initial solution was thus 988 µM chymotrypsin, 1.2 mM DEPC, and 1.2% vol/vol acetonitrile. Over the course of 1 h, sixteen 40.5 µL aliquots of the chymo/DEPC solution were mixed in a cuvette with buffer, water, and pNPA substrate (0.32 M stock solution in ethanol) to give the following final concentrations: 40 µM chymotrypsin, 49 µM DEPC, 50 mM MOPS, pH 7.14, and 2.9 mM pNPA. Hydrolytic activity of chymotrypsin was measured as an increase in A₄₁₀; the rate v ₀ (µM/min) was calculated from the ΔA ₄₁₀/Δt slope using ɛ ₄₁₀ = 15300 M⁻¹ cm⁻¹ for the p‐nitrophenoxide chromophoric product. The rate v ₀ (µM/min), plotted versus DEPC incubation time (min), showed first order exponential decline. During the course of these experiments we discovered that DEPC is soluble in acetonitrile but nearly insoluble in deionized H₂O. Experiments that followed this discovery utilized DEPC stock solutions prepared in acetonitrile. Low concentrations of acetonitrile did not reduce the activity of chymotrypsin significantly (see Supporting Information Fig. S1).

To test the effect of histidine‐blocking by DEPC on the hydrolytic activity of chymotrypsin, the enzyme (40 µM) in 50 mM MOPS, pH 7.14, was incubated for 15 min with concentrations of DEPC varying from 0 to 700 µM (5 and 50 mM stock solutions in acetonitrile); the final volume of incubating solutions was 1 mL. The substrate pNPA (0.32 M stock in ethanol) at a final concentration of 2.9 mM was then added. The hydrolytic activity of chymotrypsin was measured as an increase in A ₄₁₀, with v ₀ (µM/min) calculated from the ΔA ₄₁₀/Δt slope as above. The v ₀ (µM/min) versus [DEPC] data were plotted and fit to the equation for enzyme inhibition concentration‐dependence [Eq. (3)].

To assay the effects of DEPC on Hg²⁺‐induced chymotrypsin precipitation, 0–2.5 mM DEPC (from 50 mM stock solution in acetonitrile) was incubated for 15 min with chymotrypsin (20 µM) in 50 mM MOPS pH 7.14, in a final volume of 960 µL. After the incubation, 40 µL of 10.253 mM Hg²⁺ was added, giving a final concentration of 0.410 mM Hg²⁺. Chymotrypsin precipitation was monitored at 350 nm, and the max A ₃₅₀ was recorded and plotted against DEPC concentration (µM). In control experiments similar to those described above for EDC, we determined that DEPC does not interact directly with Hg²⁺.

Data analysis

hyperbolic saturation: Hg²⁺‐EDC binding

The absorbance of EDC at 243 nm (corrected for Hg²⁺ absorbance) was plotted against Hg²⁺ concentration, and fit to the general equation for hyperbolic saturation [Eq. (1), a version of the Michaelis–Menten equation] to determine A _max and C ₅₀, the Hg²⁺ concentration giving half‐maximal A ₂₄₃.

$$A=\frac{A_{\text{max}}}{1+\frac{C_{50}}{[H{g}^{2+}]}}=\frac{A_{\text{max}}·[H{g}^{2+}]}{[H{g}^{2+}]+C_{50}}$$ (1)

Hyperbolic saturation: Michaelis–Menten analysis

Chymotrypsin activity (v ₀, in µM/min) was plotted against initial substrate concentration [S]₀, and fit to the Michaelis–Menten equation [Eq. (2)] to determine V _max and K _m.

$$v_0=\frac{V_{\max }}{1+\frac{K_{\mathrm{m}}}{{[S]}_0}}=\frac{V_{\max }\cdot {[S]}_0}{{[S]}_0+K_{\mathrm{m}}}$$ (2)

where [S]₀ is the initial concentration of substrate, V _max (µM/min) corresponds to the maximum reaction rate when the enzyme is saturated with substrate, and K _m corresponds to the substrate concentration at which v ₀ = ½ V _max; K _m is a measure of enzyme‐substrate binding affinity.

Hyperbolic saturation: enzyme inhibition

Chymotrypsin activity (v ₀, in µM/min) was plotted against DEPC concentration, and fit to the general equation for the concentration‐dependent inhibition of enzyme activity [Eq. (3)]:

$$v_0=\frac{v_0(control)}{1+\frac{{[I]}_0}{I{C}_{50}}}$$ (3)

where [I]_o is the initial inhibitor concentration (0–700 µM), v₀ (control) = the reaction rate of the uninhibited enzyme, and IC ₅₀ is the inhibitor concentration that gives v₀ = ½ v₀ (control).

Protein aggregation kinetics

A detailed discussion of protein aggregation kinetics is included in Section 2 of the Supporting Information; here we include a brief synopsis. Protein aggregation is generally modeled as a multi‐step process that features nucleation followed by particle growth.37, 38, 39 If nucleation is rate‐determining, then aggregation kinetics are sigmoidal, featuring an initial lag time, followed by rapid aggregation and eventual saturation.37, 38, 39, 50 If nucleation is fast (either inherently,51, 52, 53 or due to seeding of nuclei,37 pH ≥ 7,19 or addition of heavy metal cations14, 15, 16, 17, 18, 19), then the entire aggregation process follows first order kinetics.37, 38, 39

In our studies of the kinetics of Hg²⁺‐induced chymotrypsin aggregation, between pH 6.5–7.0, turbidity increased sigmoidally with time, and the A₃₅₀ versus t transients were well‐fit using either the Finke‐Watzky38 or Crespo et al.50 exponential equations, the latter of which is given in Eq. (4):

$$A_{\mathrm{t}}=A_0+\Delta A-\frac{\Delta A}{k_{\mathrm{b}}(e^{k_{\mathrm{a}}t}-1)-1}$$ (4)

where A ₀ is the initial absorbance (usually zero), ΔA is the maximum change in absorbance from time zero to infinity, k _a is a scaled first order rate constant for the growth phase, and k _b is a unitless ratio of the nucleation/growth rate constants.50

Above pH 7 we observed no lag time, and our data were best fit with two sequential first order processes [Eq. (5)], with the half‐time of the fast process ≈ 4 s, and τ _slow ≈ 30 s (see Supporting Information Table SI; it is worth noting that the Crespo et al. equation also gave reasonably good fits).

$$A_{\mathrm{t}}=A_{\text{final}}-\Delta A_{\text{fast}}*\text{exp}\left(k_{\text{fast}}*t\right)-\Delta A_{\text{slow}}*\text{exp}\left(k_{\text{slow}}*t\right)$$ (5)

where k _fast and k _slow are first order rate constants for the fast and slow processes, respectively, and ΔA _fast and ΔA _slow are the maximum amplitudes (i.e., the total absorbance change from time zero to infinity) attributed to the fast and slow processes, respectively. ΔA _slow/A _final is the fraction of A _final contributed by the slow process, and 1 − ΔA _slow/A _final is the fraction of A _final contributed by the fast process. Presumably the slower process represents the formation of larger light‐scattering particles, however, as this process usually contributed less than 10% of the A₃₅₀ signal, and never more than 25% (see Supporting Information Table SI), we based our results on the fast process.

Supporting information

Supporting Information