The Protein Journal

, Volume 26, Issue 6, pp 371–385

Competitive Inhibitory Effects of Acetazolamide upon Interactions with Bovine Carbonic Anhydrase II

Authors

    • Department of Cell and Molecular Biology, School of BiologyUniversity College of Science, University of Tehran
  • Fatemeh Bagheri
    • Department of Cell and Molecular Biology, School of BiologyUniversity College of Science, University of Tehran
  • Ali Akbar Moosavi-Movahedi
    • Institute of Biochemistry & BiophysicsUniversity of Tehran
  • Massoud Amanlou
    • Faculty of PharmacologyTehran University of Medical Sciences
  • Nader Sheibani
    • Departments of Ophthalmology and Visual Sciences and PharmacologyUniversity of Wisconsin
Article

DOI: 10.1007/s10930-007-9073-4

Cite this article as:
Safarian, S., Bagheri, F., Moosavi-Movahedi, A.A. et al. Protein J (2007) 26: 371. doi:10.1007/s10930-007-9073-4

Abstract

Sulfonamide drugs mediate their main therapeutic effects through modulation of the activity of membrane and cytosolic carbonic anhydrases. How interactions of sulfonamide drugs impact structural properties and activity of carbonic anhydrases requires further study. Here the effect of acetazolamide on the structure and function of bovine carbonic anhydrase II (cytosolic form of the enzyme) was evaluated. The Far-UV CD studies indicated that carbonic anhydrase, for the most part, retains its secondary structure in the presence of acetazolamide. Fluorescence measurements using iodide ions and ANS, along with ASA calculations, revealed that in the presence of acetazolamide minimal conformational changes occurred in the carbonic anhydrase structure. These structural changes, which may involve spatial reorientation of Trp 4 and Trp 190 or some other related aminoacyl residues near the active site, considerably reduced the catalytic activity of the enzyme while its thermal stability was slightly increased. Our binding results indicated that binding of acetazolamide to the protein could occur with a 1:1 ratio, one mole of acetazolamide per one mole of the protein. However, the obtained kinetic results supported the existence of two acetazolamide binding sites on the protein structure. The occupation of each of these binding sites by acetazolamide completely inactivates the enzyme. Advanced analysis of the kinetic results revealed that there are two substrate (p-NPA) binding sites whose simultaneous occupation is required for full enzyme activity. Thus, these studies suggest that the two isoforms of CA II should exist in the medium, each of which contains one substrate binding site (catalytic site) and one acetazolamide binding site. The acetazolamide binding site is equivalent to the catalytic site, thus, inhibiting enzyme activity by a competitive mechanism.

Keywords

Bovine carbonic anhydrase IIacetazolamidebinding studycompetitive inhibitionisoforms

Abbreviations

CA

carbonic anhydrase

BCA

bovine carbonic anhydrase

HCA

human carbonic anhydrase

p-NPA

p-nitrophenylacetate

Trp

tryptophan

His

histidine

ε

extinction coefficient

ANS

anilinonaphtalene sulfonic acid

KI

potassium iodide

CD

circular dichroism

ASA

accessible surface area

mAb/min

milliabsorbance per minute

a.u.

arbitrary unit

Å2

surface unit equal to one square of Angstrom

Introduction

Carbonic anhydrases (CAs, EC 4.2.1.1) are metalloenzymes with one zinc atom per molecule in their active site. These enzymes are encoded by three unrelated gene families including α-CAs (present in the cytoplasm of plants and higher vertebrates including humans), β-CAs, and γ-CAs. In higher vertebrates, there are 14 different carbonic anhydrase isozymes distributed in different tissues and subcellular compartments (Supuran and Scozzafava, 2001). Carbonic anhydrases, including bovine carbonic anhydrase II (BCA II), catalyze the reversible hydration and dehydration of carbon dioxide (CO2 + H2O ↔ HCO3+H+), the fundamental process in the transport of carbon dioxide in animals, plants, and bacteria (Supuran and Scozzafava, 2001; Hakansson et al., 1992). BCA II consists of a single polypeptide chain with 259 amino acid residues corresponding to a molecular mass of about 29 kDa. This enzyme consists of 13 β-strands and seven α-helices surrounding the β-sheets (Hakansson et al., 1992; Ohta et al., 2004). It contains one zinc ion in the active site coordinated with three histidine residues (His 93, His 95, and His 118) and one putative water molecule in a tetrahedral geometry (Saito et al., 2004). The hydrolysis of some esters is also catalyzed by the enzyme, and this activity is utilized by several investigators using p-nitrophenylacetate (p-NPA) and other phenolic esters as substrates (Tashian et al., 1963). These substrates can be handled more easily than carbon dioxide, and the reaction rates can be measured more accurately by simple spectrophotometric methods (Thorslund and Lindskog, 1967).

Acetazolamide (2-acetylamino-1,3,4-thiadiazole-5-sulfonamide), with a molecular formula of C4H6N4O3S2, belongs to the family of sulfonamide drugs (Coleman, 1967; Wingard et al., 1991). All of the unsubstituted members of this drug family inhibit carbonic anhydrase activity (Wingard et al., 1991; Burnham et al., 2001). Pure acetazolamide is available as a white powder with a melting point of 258–259°C and molar extinction coefficient of 7900 M−1cm−1 (Lindskog, 1963; Burnham et al., 2001). The inhibitory effect of sulfonamides on carbonic anhydrase activity is reported by Lindskog and Supuran (Supuran and Scozzafava, 2001; Lindskog and Thorslund, 1968). However, the detailed structural and kinetic effects of acetazolamide on the enzyme have not been determined.

Unsubstituted sulfonamides (including acetazolamide) in their deprotonated tetrahedral forms (without H2O) bind to the zinc ion through the nitrogen atom of the sulfonamide chemical group. X-ray crystallographic investigations have revealed the binding nature of acetazolamide to the active site of three carbonic abnhydrase isozymes (CA I, CA II, CA IV) (Supuran and Scozzafava, 2001; Abbate et al., 2004). In all of these binding criteria, the deprotonated sulfonamide group (SO2NH) binds through its nitrogen atom to the zinc ion and simultaneously to the oxygen atom of the side chain of Thr199 (Abbate et al., 2004). Furthermore, other van der Waals and hydrophobic interactions between heterocyclic/aromatic moiety of the sulfonamides and the side chains in the active site of the enzyme may contribute to its high-binding affinity for the enzyme active site. The carbonyl oxygen of the amido group, hydrogen bonds with side chain of Gln92 and the methyl group interact with Phe131 (Lindskog, 1997). The binding of acetazolamide to CA II is very dependent on the pH value of the medium in such a way that its binding capacity decreases below pH 6.0 and above pH 8.0 (Coleman, 1967; Lindskog and Thorslund, 1968). It has been reported that the inhibitory effect of acetazolamide on carbonic anhydrase activity at pH 8.45 is exerted by a non-competitive mechanism with the dissociation constant of 2 × 10−7 M (Pocker and Stone, 1967). This non-competitive mechanism of inhibition is in contrast with the X-ray crystallography data, which support the penetration of acetazolamide molecules into the catalytic site of BCA II (Supuran and Scozzafava, 2001; Abbate et al., 2004). Acetazolamide is a weak acid (pKa 7.2), and at pH 8.45 most of its molecular population contained a negative charge resulting in inactivation of the enzyme. Acetazolamide can also partially inactivate the enzyme at acidic pH values (Pocker and Stone, 1967).

Acetazolamide is used for the treatment of glaucoma since sulfonamide-sensitive isozymes of carbonic anhydrase (CA II and CA XII) are responsible for secretion of aqueous humor and bicarbonate in the eye and the rise in intraocular pressure (Wingard et al., 1991; Abbate et al., 2004). Acetazolamide is also used to relieve inflammatory pain by increasing the alkalinity of tissues (Radhakrishnan and Sluka, 2005). Acetazolamide is a good choice for treatment of benign intracranial hypertension, altitude sickness, seizures and some cancers (Supuran and Scozzafava, 2001; Burnham et al., 2001; Abbate et al., 2004; Richter and Hamann, 2004; Agrawal et al., 2004). Therefore, determination of acetazolamide structural effects on carbonic anhydrase, a key enzyme in many physiological processes, has major clinical implications. In the present work, kinetic and structural effects of acetazolamide on carbonic anhydrase were determined.

Materials and Methods

Materials

Bovine carbonic anhydrase from erythrocytes, p-nitrophenylacetate (p-NPA), Trizma base (molecular biology grade), and anilinonaphtalene sulfonic acid (ANS) were obtained from Sigma. Pure Acetazolamide in pharmaceutical grade was provided by Iran Daru pharmaceutical company (Tehran, Iran). Potassium iodide and sodium dithionite were from Merck. All other reagents were of analytical grade. Solutions were prepared in Tris buffer (100 mM Tris, pH 7.5) utilizing double-distilled water with very low conductivity obtained by SANYO Gallencamp PLC (Fistreem Cyclon).

Methods

Kinetic Studies

Carbonic anhydrase assays were performed at 27°C as described before (Armstrong et al., 1966; Pocker and Stone, 1967). The enzyme-catalyzed reaction was followed at 400 nm using a Shimadzu UV-260 spectrophotometer (Japan) in quartz cells with a 1 cm path length. The enzyme concentration was 0.15 μM, which reacted with 0–2 mMp-nitrophenylacetate (p-NPA) as substrate. Recording the effect of the enzyme on higher concentrations of p-NPA (greater than 2 mM) was not possible because above this critical point (about 2–2.5 mMp-NPA) substrate appeared to sediment. The enzyme stock solution was prepared by dissolving approximately 3 mg of enzyme powder in 1 ml of Tris buffer followed by dialysis at 4°C against several changes of the same buffer. The enzyme concentration was subsequently determined after dialysis using the extinction coefficient of \( \varepsilon ^{{1\% }}_{{280}} = 19.0 \) as previously reported (Nyman and Lindskog, 1964). The stock solution of p-NPA (138 mM) was prepared by dissolving 0.05 g of p-NPA in 2 ml of acetonitrile. The esterase activity of the enzyme was monitored at 400 nm by the appearance of p-nitrophenol as the product.

Fluorescence Measurements

Fluorescence emission spectra of the enzyme in four concentrations of acetazolamide (0, 0.15, 0.29, and 0.36 μM) were recorded using a Hitachi spectrofluorimeter model MPF-4. The selected excitation wavelength of 291 nm is specific for excitation of the tryptophanyl residues of the protein. Emission spectra were recorded between the wavelengths of 300–450 nm at the bandwidth of 10 nm. A filter was placed on the pathway of the emitted light to reduce the interfering effects from the scattered photons from wavelengths below 290 nm.

The presence of accessible hydrophobic pockets in the native protein was determined using ANS under similar acetazolamide concentrations. ANS is a good fluorogenic probe, which demonstrates significant changes in the emission spectrum based on the hydrophobic power of the surrounding medium. The emission spectra of the enzyme–ANS complexes at 27°C were recorded between the wavelengths of 400–650 nm at a bandwidth of 10 nm and excitation wavelength of 350 nm. The enzyme and ANS concentrations in all protein samples were 2 and 50 μM in a total volume of 500 μl, respectively. Before recording of the emission spectrum, an incubation time of 1 min was allowed for ANS penetration into the accessible hydrophobic pockets in the protein structure.

The accessibility of the tryptophanyl residues was tested using iodide ion as a fluorescence quencher of the emitted photons leaving the excited tryptophanyl residues. Enzyme (2 μM) samples in medium containing increasing concentrations of KI (0–250 mM obtained from 1 M stock solution of KI containing 2 mM sodium dithionite to prevent oxidation of the iodide ions) were incubated for 1 min under the three acetazolamide concentrations (0, 0.15, and 0.36 μM). This was done until the maximum quenching was obtained. Recording of the emission spectra was performed in the range of 300–450 nm at an excitation wavelength of 291 nm.

Calculation of the Accessible Surface Area

Calculation of the accessible surface area (ASA) was carried out at the amino acid level using the FANTOM program provided by the Sealy Center of Structural Biology at the University of Texas Medical Branch. Cartesian coordinates of the atoms in carbonic anhydrase were obtained using the web service of the Protein Data Bank (PDB).

Circular Dichroism Studies

Far-UV CD spectra were recorded on an Aviv CD spectropolarimeter model 215 (Aviv Associates, Inc., USA) using four different protein solutions (0.18 mg/ml prepared in the presence of 0, 0.15, 0.29, and 0.36 μM acetazolamide). The results were expressed as molar ellipticity, [θ] (deg cm2 dmol−1), based on a mean amino acid residue weight (MRW). The molar ellipticity was determined as [θ]λ = (θ × 100 MRW)/(cl), where c is the protein concentration in mg/ml, l is the light path length in cm, and θ is the measured ellipticity in degrees at a given wavelength. The instrument was calibrated with (+)-10-camphorsulfonic acid, assuming [θ]291 = 7820 deg cm2 dmol−1, and with standard non-hygroscopic ammonium (+)−10-camphorsulfonate, assuming [θ]290.5 = 7910 deg cmdmol−1. Noise reduction was accomplished using Aviv software, including the fast Fourier-transform noise reduction routine, which allows the enhancement of most noisy spectra without the distortion of peak shapes. Spectral deconvolution was performed by the method of Bohm et al. employing CDNN version 2.1 (Bohm et al., 1992).

Thermal Scanning

The thermal stability of carbonic anhydrase was investigated by spectrophotometric methods using a Schimadzu UV spectrophotometer (model 3100; Japan) equipped with a Haake D8 water bath temperature controller (Germany) between the range of 27 and 87°C. In this temperature range, drawing of the thermal profiles of the protein in the presence of 0, 0.15, 0.29, and 0.36 μM acetazolamide was performed by changing the temperature with a gap of 5°C followed by deliberation in each achieved temperature point (about 1 min) for complete heat transfer to the protein samples and then writing down the obtained absorbance of each sample at 280 nm. Protein concentration was 2 μM dissolved in Tris-buffer in the presence of 0, 0.15, 0.29, and 0.36 μM acetazolamide.

Spectrophotometric Titration and Binding Theory

Absorbance differences between the sample and reference cells were determined spectrophotometrically at 280 nm using a Shimadzu-3100 double beam spectrophotometer (Japan). In each titration step, determination of the molar concentration for the protein–acetazolamide complexes was performed according to our previous binding theory with a slight modification (Moosavi-Movahedi et al., 2006). In our previous theory, the utilized ligand was a colorogenic molecule with a maximum lambda at 440 nm (visible region), but here we have utilized a colorless ligand molecule without any potency to absorb radiated photons at the visible region. It should be noted that at 280 nm the absorbance of the protein molecules did not interfere with that of the acetazolamide (ligand) molecules in the medium because the utilized acetazolamide concentration was so much lower (nM) than that of the instrumental detection limit. Therefore, attachment of the ligand molecules to the protein structure was monitored through the absorbance alterations recorded at 280 nm (λmax of the protein) after binding of the ligand molecules to the protein structure. The binding isotherm plot of acetazolamide in interaction with the protein could be plotted on the basis of the absorbance differences recorded between the native protein structure (0 μM acetazolamide) and the protein–acetazolamide complex in each titration step. The utilized theoretical equation to analyze the obtained spectrophotometric saturation curve was as follows:
$$ \Delta A = C_{{{\text{p,b}}}} (\varepsilon _{{{\text{p,b}}}} - \varepsilon _{{{\text{p,f}}}} ) + C_{{{\text{p,t}}}} \cdot \varepsilon _{{{\text{p,f}}}}$$
(1)
in which ΔA is the absorbance difference between the sample and reference cells that was recorded in each titration point by the instrument; εp,f and εp,b are molar extinction coefficients of the free and bound protein molecules; while Cp,t and Cp,b are molar concentrations of the total and bound proteins, respectively. In Eq. 1, ΔA (recorded by the instrument), Cp,t and εp,f are known \( (\varepsilon ^{{1\% }}_{{280}} = 19.0{\text{ = 55,100}}\;M^{{ - {\text{1}}}} \;{\text{cm}}^{{ - {\text{1}}}} ), \) and εp,b can be determined at the saturation point of the binding curve. Thus, at the saturation point of the curve the molar concentration of the protein–ligand complex is equal to the total molar concentration of the protein. Evidently, the molar concentration of the bound acetazolamide at this point is n times greater than the molar concentration of the CA II-acetazolamide complex (n is the maximum number of binding sites on the protein). Thus, dividing the absorbance difference at the saturation point (graphically determined by linear extrapolation, see Fig. 12) by the total molar concentration of the protein results in the molar extinction coefficient of the bound protein at 280 nm (εp,b280). Therefore, the molar concentration of the bound protein (Cp,b) can be calculated in each titration point using Eq. 1. Since the molar concentration of the bound acetazolamide is n times greater than the molar concentration of the bound protein, the calculation of the molar concentration of the bound acetazolamide is possible (n is the total number of ligand binding site(s) on the protein). We calculated n by dividing the total acetazolamide molar concentration by the total CA II molar concentration at the inflection point of the saturation curve (= [acetazolamide]inf/[CA II]inf ; the subscript of “inf” means “at the inflection point”). We know that in the binding studies the inflection point of a saturation curve (which mostly, but not always, coincides with the point of the 50% saturation of the whole binding site) is a critical point indicating the maximum binding affinity for a given ligand molecule. In fact, we have seen that many known binding parameters, such as Km or \( P^{{{\text{O}}_{2} }}_{{50}} \), have been commonly reported at the inflection point (half saturation) of the related saturation curves. Thus, at the inflection point in which we encounter with the maximum ligand-binding affinity for the protein, the ratio of the [bound ligand]/[total protein] can be used to determine the total number of existed ligand binding site(s) on the protein. However, the number of total ligand binding site(s) cannot be calculated at the saturation point, because after the inflection point whatever the ligand concentration is increased to the binding affinity of the ligand for the enzyme is being reduced. In other word, reaching to the saturation point will be harder and harder with increasing ligand concentration in the medium. Thus, at the saturation point, the occupation of the enzyme might be obtained with an unnecessary much higher molar concentration of the ligand. Here n was calculated at the inflection point as equal to 1, which coincided well with the results reported by X-ray crystallography indicating the placing of one acetazolamide into the active site of the enzyme (Abbate et al., 2004).

Results

Figures 13 show the quenching effect of the iodide ions exerted on the accessible tryptophanyl residues of the carbonic anhydrase in the presence of different concentrations of acetazolamide (0, 0.15, and 0.36 μM). In these figures, the y-axis was scaled using relative units (F/F0). This ratio depicts the recorded fluorescence intensity of the protein solution in the presence of potassium iodide (F) at the emission λmax of the protein (343 nm for 0, 0.15, and 0.36 μM acetazolamide) relative to the fluorescence intensity of the protein in the related acetazolamide concentration when the KI concentration is zero (F0). The increase in iodide ion concentration was accompanied by a decrease in the recorded fluorescence intensity ratio since iodide ions existing near the accessible tryptophanyl residues in the protein are more likely at higher KI concentrations (Figs. 13). The point of maximum quenching can be detected by drawing a double reciprocal curve (F0/F against 1/[KI) and extrapolating the resultant curve (relevant to the last shape of the curve) to the 1/y axis (dashed lines in inset a of Figs. 13). The indicated double reciprocal curve has no theoretical base perhaps only enabling us to determine the maximum fluorescence ratio (F/F0) when the quencher concentration, [KI], tends to infinity.
https://static-content.springer.com/image/art%3A10.1007%2Fs10930-007-9073-4/MediaObjects/10930_2007_9073_Fig1_HTML.gif
Fig. 1

Fluorescence quenching of the tryptophanyl residues of the native structure of carbonic anhydrase. Quenching experiments were carried out in the presence of different concentrations of KI (0–250 mM). The excitation wavelength was 291 nm and the protein concentration was 2 μM prepared in 100 mM Tris buffer, pH 7.5. Inset (a): The point of maximum quenching was determined by extrapolation of the curve (dotted line) to the y axis in the double reciprocal plot. Inset (b): Stern–Volmer plot of native carbonic anhydrase. KS–V is calculated from the slope of the first linear phase and is equal to 4.2 M−1 (the mathematical equation of the dotted line is brought in the figure as a function of y to calculate KS–V). x is a variable equivalent to [KI]

https://static-content.springer.com/image/art%3A10.1007%2Fs10930-007-9073-4/MediaObjects/10930_2007_9073_Fig2_HTML.gif
Fig. 2

Fluorescence quenching of the tryptophanyl residues of the protein in the presence of 0.15 μM acetazolamide. Quenching experiments were carried out under the same conditions as described in the legend to Fig. 1. Inset (a): The point of maximum quenching was determined by extrapolation of the curve (dotted line) to the y-axis in the double reciprocal plot. Inset (b): Stern–Volmer plot for carbonic anhydrase in the presence of 0.15 μM acetazolamide. KS–V is calculated from the slope of the first linear phase and is equal to 3.8 M−1 (the mathematical equation of the dotted line is brought into the figure as a function of y to calculate KS–V). x is a variable equivalent to [KI]

https://static-content.springer.com/image/art%3A10.1007%2Fs10930-007-9073-4/MediaObjects/10930_2007_9073_Fig3_HTML.gif
Fig. 3

Fluorescence quenching of the tryptophanyl residues of the protein in the presence of 0.36 μM acetazolamide. Quenching experiments were carried out under the same conditions as described in the legend to Fig. 1. Inset (a): The point of maximum quenching was determined by extrapolation of the curve (dotted line) to the y-axis in the double reciprocal plot. Inset (b): Stern–Volmer plot for carbonic anhydrase in the presence of 0.36 μM acetazolamide. KS–V is calculated from the slope of the first linear phase and is equal to 3.9 M−1 (the mathematical equation of the dotted line is brought into the figure as a function of y to calculate KS–V). x is variable equivalent to [KI]. Inset (c): ANS checking to identify accessible hydrophobic pockets on the protein structure under four different conditions (0, 0.15, 0.29, and 0.36 μM acetazolamide). The height of the columns 1–4 shows the fluorescence intensity at 520 nm for ANS + 0.36 μM acetazolamide; ANS + Protein; ANS + Protein + 0.15 μM acetazolamide; ANS + Protein + 0.29 μM acetazolamide; ANS + Protein + 0.36 μM acetazolamide, respectively

The number of accessible tryptophanyl residues, with respect to the point of maximum quenching, can be calculated as previously described by us (Safarian et al., 2006). This is achieved by multiplication of the F/F0 ratio (related to the point of the maximum fluorescence quenching) that is subtracted from 1 (percent value of the protein in the absence of KI, F/F0 = 1) to 7 (the total number of the existing tryptophanyl residues in the protein) (Safarian et al., 2006; Freifilder, 1985). The number of accessible tryptophanyl residues calculated under the three cited conditions are: 0.92 (≈1.00), 1.21 (≈1.00), and 2.65 (≈3.00), respectively. These numbers indicate that when the acetazolamide concentration increases from 0 to 0.36 μM in the medium, two additional tryptophanyl residues become exposed to the outer environment of the native structure of the protein.

Figure 4 shows fluorescence spectra of the protein recorded under the three described acetazolamide concentrations. The inset of this figure indicates the fluorescence intensity at the emission λmax of the protein in the absence of KI obtained from each related main spectra. These results indicated that increasing the concentration of acetazolamide in the medium was not accompanied by a significant alteration in the recorded fluorescence intensities. Therefore, the tertiary structure of the protein is not drastically altered in the four selected acetazolamide concentrations (0, 0.15, 0.29, and 0.36 μM). Furthermore, the slope (KS–V) of the first linear phase under each environmental condition was 4.2, 3.8, and 3.9 M−1 for 0, 0.15, and 0.36 μM acetazolamide, respectively, thus further verifying the lack of drastic structural alterations in the protein (insets b in Figs. 13). The slope of the Stern–Volmer plot is equal to the Stern–Volmer constant (KS–V), indicating the degree of exposure of the tryptophanyl residues on the protein (Wimley et al., 1996; Liu and Deber, 1997).
https://static-content.springer.com/image/art%3A10.1007%2Fs10930-007-9073-4/MediaObjects/10930_2007_9073_Fig4_HTML.gif
Fig. 4

Fluorescence spectra of four conformations of the native enzyme (prepared in 100 mM Tris buffer pH 7.5) and the protein structures in the presence of 0.15, 0.29, and 0.36 μM acetazolamide. The excitation wavelength was 291 nm and the protein concentration was 2 μM. Please note that when acetazolamide is added to the buffer, there is no appreciable change in the fluorescence intensity at λmax = 343 nm (see the inset)

The protein structural alterations in the presence of acetazolamide can also be evaluated using ANS as a fluorogenic probe. This will allow the detection of exposed and accessible hydrophobic pockets in the protein (inset c of Fig. 3). The differences between each ANS–protein emission intensity at λmax 520 nm relative to the free ANS–acetazolamide emission intensity (at maximum concentration of acetazolamide, 0.36 μM) was recorded as an indication of structural alterations in the enzyme. These findings were consistent with no significant alterations in the intrinsic fluorescence emission of the tryptophanyl residues under the four different acetazolamide concentrations (Fig. 4 and its inset).

Figure 5 shows the thermal stability of the protein structure determined relative to the native structure. The stability of the protein structure in the presence of acetazolamide was slightly increased (about 1.5°C), especially at 0.29 and 0.36 μM acetazolamide (Fig. 5 and insets a and b). This is consistent with no significant alterations in the tertiary structure of the protein as was demonstrated by the fluorescence technique (Fig. 4 and inset c of Fig. 3). The CD spectra of the protein under the four different acetazolamide concentrations were recorded to determine the occurrence of the secondary structural changes (Fig. 6). Our results indicated that the presence of acetazolamide in the medium had a minimal effect on the secondary structure of the protein. This was further confirmed when the molar ellipticity of the protein at 222 nm was determined against the acetazolamide concentration (inset of Fig. 6).
https://static-content.springer.com/image/art%3A10.1007%2Fs10930-007-9073-4/MediaObjects/10930_2007_9073_Fig5_HTML.gif
Fig. 5

Thermal profiles of carbonic anhydrase in the presence of 0, 0.15, 0.29, and 0.36 μM acetazolamide. The value of melting temperature (Tm values) of the protein slightly increased (about 1.5°C) with increasing acetazolamide concentration in the solution. The protein concentration was 2 μM. The bold lines indicate those related sigmoid curves utilized to calculate the Tm values (see the insets showing the inflection points). 0 μM acetazolamide ●; 0.15 μM acetazolamide ◆; 0.29 μM acetazolamide ▲ ; 0.36 μM acetazolamide ■. Inset (a): Calculation of the Tm values, 0 μM acetazolamide, bold line; 0.15 μM acetazolamide, dashed line; 0.29 μM acetazolamide, solid line. dy/dx is first derivative of sigmoidal (ascending) part of the main thermogram. Inset (b): calculation of the Tm values. 0 μM acetazolamide, bold line; 0.36 μM acetazolamide, solid line. dy/dx is first derivative of sigmoidal (ascending) part of the main thermogram

https://static-content.springer.com/image/art%3A10.1007%2Fs10930-007-9073-4/MediaObjects/10930_2007_9073_Fig6_HTML.gif
Fig. 6

CD spectra of carbonic anhydrase in the presence of 0, 0.15, 0.29, and 0.36 μM acetazolamide. The bold, solid, dashed, and shadowed lines indicate the protein structures in 0, 0.15, 0.29, and 0.36 μM acetazolamide, respectively. The inset indicates that relative to the native structure no considerable changes occurred in the secondary structure of the protein in the presence of 0.15, 0.29, and 0.36 μM acetazolamide. The protein concentration was 0.18 mg/ml

The relationship between an increase in acetazolamide concentration and catalytic activity is shown in Figs. 710. The increase in acetazolamide concentration in the medium was followed by a decrease in the enzyme velocity, especially at 0.36 μM acetazolamide (compare the last obtained experimental velocities of the velocity curves; a considerable decrease of the enzyme velocity obtained in parallel with increasing concentration of acetazolamide in the medium, Figs. 710). In these figures the saturation velocity curves are biphasic with a distinct bump around 1 mMp-NPA. Reliability of this bump at the critical point of 1 mMp-NPA can be easily distinguished from the similar pattern repeated in four different experiments performed at different concentrations of acetazolamide (Figs. 710). Moreover, performing of the student t-test with statistical accuracy of 0.95 showed that changing of the enzymatic velocity at this critical point (relative to next point) is significant. The existence of this bump indicates a simple common Michaelis-Menten kinetic model cannot be utilized to describe the kinetic behavior of the enzyme. Therefore, other advanced kinetic equations, like those presented by Bardsley, and utilized by Sergienko and Srivastava, were used to analyze the curves (Bardsley et al., 1980; Sergienko and Srivastava, 1997). They reported that when the plot [L]p/v vs. 1/[L] (L is ligand or substrate) approached a horizontal asymptote at p = 2 and a zero asymptote at p = 3, it suggests that full activity is attained when two molecules of p-NPA bound to the enzyme (insets b and c in Figs. 710). We also analyzed our experimental results by plotting 1/v[Acetazolamide]p vs [Acetazolamide] for 1 mMp-NPA (the critical point in the saturation curves) and the acetazolamide concentration range of 0–0.46 μM (Fig. 11a–c). A horizontal asymptote was obtained when p = 2 and zero asymptote was obtained when p = 3. This suggests that there are two essential binding (regulatory) sites for acetazolamide whose occupation results in complete loss of the bound enzyme’s activity (Sergienko and Srivastava, 1997).
https://static-content.springer.com/image/art%3A10.1007%2Fs10930-007-9073-4/MediaObjects/10930_2007_9073_Fig7_HTML.gif
Fig. 7

Kinetic saturation curve of the native carbonic anhydrase in the presence of p-NPA (0–2 mM). A distinct bump around 1 mMp-NPA shows the non-Michaelis-Menten kinetic behavior of the enzyme. Regarding the insets, it can be seen that a horizontal as well as zero asymptotes have been achieved at p = 2 and p = 3, respectively. Thus, using the advanced kinetic equations (see the text) enabled us to determine the possibility of the existence of two catalytic sites (one on each BCA II isoforms). The enzyme concentration was 0.15 μM

https://static-content.springer.com/image/art%3A10.1007%2Fs10930-007-9073-4/MediaObjects/10930_2007_9073_Fig8_HTML.gif
Fig. 8

Kinetic saturation curve of carbonic anhydrase in the presence 0.15 μM acetazolamide. A similar shape with a distinct bump around 1 mMp-NPA shows a kinetic behavior similar to that of the native enzyme. Regarding the insets, a horizontal as well as zero asymptotes were achieved at p = 2 and p = 3, respectively. Thus, two catalytic sites (one on each BCA II isoforms) should exist on the protein structure. The enzyme concentration was as in Fig. 7

https://static-content.springer.com/image/art%3A10.1007%2Fs10930-007-9073-4/MediaObjects/10930_2007_9073_Fig9_HTML.gif
Fig. 9

Kinetic saturation curve of carbonic anhydrase in the presence 0.29 μM acetazolamide. A similar shape with a distinct bump around 1 mMp-NPA shows a kinetic behavior similar to that of the native enzyme. Regarding insets, a horizontal as well as zero asymptotes were achieved at p = 2 and p = 3, respectively. Thus, two catalytic sites (one on each BCA II isoforms) should exist on the protein structure. The enzyme concentration was as in Fig. 7

https://static-content.springer.com/image/art%3A10.1007%2Fs10930-007-9073-4/MediaObjects/10930_2007_9073_Fig10_HTML.gif
Fig. 10

Kinetic saturation curve of carbonic anhydrase in the presence 0.36 μM acetazolamide. A similar shape with a distinct bump around 1 mMp-NPA shows a kinetic behavior similar to that of the native enzyme. Regarding the insets, a horizontal as well as zero asymptotes were achieved at p = 2 and p = 3, respectively. Thus, two catalytic sites (one on each BCA II isoform) should exist on the protein structure. The enzyme concentration was as in Fig. 7

https://static-content.springer.com/image/art%3A10.1007%2Fs10930-007-9073-4/MediaObjects/10930_2007_9073_Fig11_HTML.gif
Fig. 11

Inhibition of BCA II by acetazolamide (0.15, 0.20, 0.29,0.36, and 0.46 μM) in the presence of 1 mM substrate concentration (the bump) in 100 mM Tris buffer, pH 7.5. The number of inhibitory sites was determined by plotting the experimental results as 1/(v[acetazolamide]p) as a function of [acetazolamide] at different integral values of p (as indicated on the individual plot). Please note a horizontal asymptote when p = 2, suggesting that there are two inhibitory acetazolamide binding sites in the BCA II structure (one on each BCA II isoforms). The enzyme concentration was as in Fig. 7

Figure 12 shows the profile of the absorbance changes at 280 nm against the total molar concentration of acetazolamide. We expected that the effect of our ligand molecule (acetazolamide) on the protein structure would not be similar to a denaturant. Thus, the recorded absorbance differences between the native structure of the protein and that of protein–acetazolamide complexes would not be rigorous. Therefore, the data were collected here by a powerful spectrophotometer (Shimadzu-3100), which its error range was ±0.001 absorbance unit. According to the theory described above, the value of υ (the number of moles of bound acetazolamide per moles of CA II) in each titration step can be determined and plotted versus the logarithm of the free molar concentration of acetazolamide (log[acetazolamide]f), which is called “the binding isotherm” (inset a of Fig. 13). With respect to our previous binding theory, ΔG˚ can be calculated using the typical Boltzmann’s equation as follows (Moosavi-Movahedi et al., 2006):
$$ N{\text{/}}N_{0} = \exp ( - \Delta G^{{\text{o}}} {\text{/}}RT) $$
or
$$ \Delta G^{{0}} = RT\ln N_{0} - RT\ln \nu$$
(2)
where N and N0 are the molar concentration of occupied binding sites and the total molar concentration of binding sites, respectively. N is equal to ν (determined through the binding theory described above) and N0 can be determined via multiplication of the total molar concentration of the enzyme by one (the total number of acetazolamide binding sites on the protein structure).
https://static-content.springer.com/image/art%3A10.1007%2Fs10930-007-9073-4/MediaObjects/10930_2007_9073_Fig12_HTML.gif
Fig. 12

The absorbance profile of BCA II–acetazolamide complexes at 280 nm vs. total concentration of acetazolamide. The dashed line denotes the saturation point of the protein. Inset: First derivative of the absorbance profile to determine the inflection point. Maximum binding affinity obtained at the maximum point of the curve

https://static-content.springer.com/image/art%3A10.1007%2Fs10930-007-9073-4/MediaObjects/10930_2007_9073_Fig13_HTML.gif
Fig. 13

The standard free energy per each υ \( (\Delta G^{{\text{o}}}_{\upsilon } ) \) versus υ at 27°C. The solid line indicates the related free energies calculated according to our theory. The dashed lines show that two types of physical interactions exist when the acetazolamide binds to its two individual binding sites on the protein. Inset (a): The binding isotherm (υ against the logarithm of free concentration of acetazolamide) for BCA II–acetazolamide interaction. Inset (b): The profile of the standard free binding energy as a function of ν for acetazolamide–BCA II binding process. The intersection of the two lines appeared around ν = 0.1 suggesting that during the first binding phase (10% of the total protein molecules are bound to acetazolamide irrespective of the type of isoform) the physical essence (electrostatic) should be different from the second phase (hydrophobic)

In Fig. 13 ΔGυo is plotted against the calculated υ values. The curve indicates two distinct phases, i.e., two types of interactions are responsible for the binding of acetazolamide to the CA II binding site. These two phases are shown by the two dashed lines drawn for one of the curves which is interrupted around ν = 0.1 (see inset b of Fig. 13).

Discussion

The majority of the therapeutic effects of the sulfonamide drug-family members are mediated through suppression of the catalytic activity of the membrane and cytosolic carbonic anhydrases. Therefore, there has been increasing interest in the determination of carbonic anhydrase structural properties in the presence of the sulfonamide drugs. Here the effects of acetazolamide on the structure and function of bovine carbonic anhydrase II (a cytosolic carbonic anhydrase) were examined using several experimental techniques.

Recording of the CD and fluorescence emission spectra of the protein under different acetazolamide concentrations indicated no significant structural changes in the enzyme (Figs. 14 and 6 and their related insets; comparisons among the calculated amounts of secondary structures are tabulated in Table 1). The CD technique can distinctly determine secondary or tertiary structural alterations in proteins. However, a simple fluorescence experiment will not reveal tertiary structural changes. Therefore, it was necessary to perform fluorescence profile tests using different probes and strategies for the detection of tertiary structural alterations. In the near UV-CD determinations the protein concentration used is around 1 mg/ml. This is an instrumental limitation not allowing the use of ligand/protein molar ratios that are usually used in the kinetic experiments. Thus, the sensitive fluorescence technique was used here for determination of the tertiary structural alterations, which does not require sample solutions with high protein concentration.
Table 1.

Calculated Amounts of the Five Types of Secondary Structure in Bovine Carbonic Anhydrase II

Secondary structure (200–260 nm)

0 μM Acetazolamide

0.15 μM Acetazolamide

0.29 μM Acetazolamide

0.36 μM Acetazolamide

Helix

21.2%

20.0%

19.3%

20.0%

Anti-parallel

17.1%

18.0%

18.2%

18.0%

Parallel

10.3%

10.5%

10.7%

10.5%

β-Turn

17.7%

17.7%

17.5%

17.7%

Random Coil

33.5%

33.6%

34.3%

33.6%

Inset of Fig. 4, drawn using the emission intensity at the emission λmax of the protein (343 nm) under four different acetazolamide concentrations, demonstrated that the tertiary structure of the protein remains unchanged. In addition, the fluorescence experiments using the ANS fluorogenic probe indicated that small structural alterations may occur in the protein such that the exposure of the hydrophobic pockets on the protein surface is minimal (inset c of Fig. 3). The minimal changes observed in the CD and ANS fluorescence spectra for the native and ligand-bound protein structure suggest possible structural alterations induced by acetazolamide. These alterations can occur locally near the active site, the most flexible and instable part of the enzyme (Shoichet et al., 1995), and not systemically on the whole protein structure.

Fluorescence experiments using KI confirmed the existence of minimal conformational alterations induced by the acetazolamide molecules in the protein structure (Figs. 13). These small tertiary structural changes may occur in the form of spatial reorientations of distinct amino acyl residues (e.g., tryptophanyl residues). For example, in 0.36 μM acetazolamide, the specific reorientations of some flexible tryptophanyl residues, like those placed near the active site, was confirmed by the KI experiments indicating the accessibility of two additional tryptophanyl residues (Fig. 3). The lack of systematic structural changes on the protein was also supported by the similar KS–V values when the acetazolamide concentration in the medium was increased; in Figs. 13 the calculated KS–V values are represented as 4.2, 3.8, and 3.9 M−1 for 0–0.36 mM acetazolamide, respectively.

Our theoretical methods for calculating ASA and determining the two additional exposed tryptophanyl residues (described above) are based on the method of Shrake and Rupley (1973). These calculations indicated that among the seven tryptophanyl residues of carbonic anhydrase, one residue (Trp 243) has the highest accessibility to the solvent or other small molecules (e.g., iodide ions). The calculated ASA for Trp 243 was greater than those of other Trp residues (ASATrp243 = 35.96 Å2; ASATrp4 = 23.41 Å2; ASATrp190 = 13.63 Å2; ASATrp122 = 8.86 Å2; ASATrp207 = 2.23 Å2; ASATrp15 = 0.73 Å2; ASATrp96 = 0.00 Å2), thus indicating a greater chance of colliding with small molecules in solution such as the iodide ions. These findings are also consistent with the results obtained in the fluorescence quenching experiments. In these experiments, the exposure of only one tryptophanyl residue (likely Trp 243) was proposed for the native structure of the protein in pure buffer (Fig. 1). According to the crystallographic coordinates Trp 243 and the other two more accessible tryptophanyl residues (Trp 4 and Trp 190, especially Trp 4) are placed near the enzyme active site. The exposure of the two additional tryptophanyl residues calculated from Figs. 2 and 3 suggest that reorientation and exposure of some distinct aminoacyl residues (Trp 4 and Trp 190 by themselves or those residues which are placed near these tryptophanes) are responsible for the decreased enzyme velocity (compare the last experimental velocity values in Figs. 710 showing the decreasing pattern of enzyme velocity). Thus, the small tertiary structural alterations of carbonic anhydrase and the reorientation of some essential residues may explain the considerable drop in the velocity of carbonic anhydrase in the presence of acetazolamide.

The thermal stability of carbonic anhydrase, determined under the four acetazolamide concentrations, showed a slight increase (about 1.5°C) in the melting temperature of the enzyme especially at acetazolamide concentrations greater than 0.29 μM (Fig. 5 and its insets). This is consistent with the lack of significant structural alterations, which are further supported by the fluorescence and CD experiments (Figs. 14 and 6). In addition, the recorded absorbance differences between the thermally unfolded structures of the free enzyme, as well as the enzyme–acetazolamide complexes (near 0.5 absorbance unit), and the absorbance of the free folded structure of the enzyme (near 0.1 absorbance unit) are very high (Fig. 5). These results indicated the existence of a high aggregation velocity for the protein in the presence or absence of the acetazolamide. In fact, when the aggregation velocity of the unfolded enzyme molecules is higher than the applied scan rate of the instrument at each exerted temperature point, higher amounts of aggregated protein molecules may form resulting in more scattering of the radiated photons and in more recorded absorbance of the sample protein solution (Fig. 5). Therefore, the similarities in melting temperature (Tm) and aggregation velocity between the free enzyme and the enzyme–acetazolamide complexes suggest that the binding of the acetazolamide to the enzyme is not accompanied by significant structural changes. This is also further supported by the conserved kinetic pattern of CA II in the presence or absence of acetazolamide with a distinct bump around 1 mMp-NPA (Figs. 710). In addition, the similar number of substrate binding sites (two binding sites, p = 2, calculated for the free as well as acetazolamide-bound protein complexes) also confirms the existence of minimal acetazolamide-induced structural alterations on the protein. Alternatively, the similarities in the shape of the kinetic saturation curves reveal the existence of a similar catalytic mechanism supporting minimal structural alterations of the protein.

Our binding studies indicated that the binding of acetazolamide to the protein occurred in a 1:1 ratio (νmax = 1). This finding is consistent with the X-ray crystallographic data supporting the existence of one acetazolamide-binding site on the enzyme, which is equivalent to the enzyme active site (Supuran and Scozzafava, 2001; Abbate et al., 2004). However, our kinetic results supported the existence of two acetazolamide binding sites on the protein structure. The occupation of each of these binding sites by acetazolamide completely inactivates the enzyme (Fig. 11). Thus, the resolution of conflicting results observed here (the number of acetazolamide binding site(s) in the binding study and the kinetic investigations) can be achieved if, and only if, it is considered that two isoforms of CA II exist in the medium, each contained one acetazolamide-binding site equivalent to the catalytic site (competitive ligand binding). Interestingly, the evidence for two isoforms of carbonic anhydrase II has been presented using immunoblotting techniques (Riehl and Schlue, 1993). Moreover, the possible existence of two isoforms of BCA II can be revealed from the distinct bumps appeared in the kinetic saturation curves around 1 mM acetazolamide (Figs. 710). In these curves, the bumping points, around 1 mM acetazolamide, separate the related saturation curves into two kinetic phases each of which related to one isoform’s kinetic behavior.

Our advanced analysis of the kinetic results revealed that there are two substrate-binding sites (p-NPA binding sites) whose simultaneous occupation is essential for full activity of the enzyme (Figs. 710). These findings reinforce the notion of equivalency between the substrate and acetazolamide binding sites, which is further supported by X-ray crystallography (Supuran and Scozzafava, 2001; Abbate et al., 2004). If our proposition about binding of acetazolamide into the active site is correct the mechanism of inhibition for acetazolamide is exerted through competitive inhibition. Therefore, the previously suggested non-competitive mechanism of inhibition for acetazolamide upon interaction with BCA II needs to be re-evaluated.

In Fig. 13 there are two thermodynamic phases with related strong and mild slopes, respectively. In fact, during the first thermodynamic phase releasing of heat is much greater than the second phase. The intersection of the two linear dashed lines (inset b; each line related to one phase) appeared around ν = 0.1. This suggests that during the first binding phase, at low concentrations of acetazolamide, about 10% of the total protein molecules (irrespective of the type of isoforms) are bound to acetazolamide via the very high exothermic (high affinity) physical interactions (e.g., the electrostatic interactions). However, at high acetazolamide concentrations the physical essence of the binding interactions (irrespective of the type of isoforms) was changed since low negative free energy values (note the slight slope of the second phase in Fig. 13) were calculated for the second binding phase. Thus, the majority of acetazolamide binding to the BCA II isoforms (90% of the enzyme molecular population) occurred during the second binding phase, perhaps through hydrophobic interactions with lower affinity (lower releasing of heat) than electrostatic interactions (Hinz, 1983). The p-NPA is a hydrophobic molecule, dissolved in acetonitrile, and it should penetrate into the enzyme catalytic site via hydrophobic interactions. Thus, the penetration of acetazolamide into the catalytic (inhibitory) site can occur through the similar hydrophobic interactions. In the complex with human CA II, the thiadiazole ring of acetazolamide is in van der Waals contact (electrostatic interaction produced by the formation of the transient dipoles) with Val 121, Leu 198, and Thr 200, while the carbony1 oxygen of the amido group hydrogen bonds (electrostatic interactions) with the side-chain amide of Gln 92 and the methyl group interacts with Phe 131 (hydrophobic interaction) (see Scheme 1 provided from Vidgren et al., 1990). Two additional structures involving thiadiazole sulfonamides have also been determined, the methazolamide-HCA I and the aminobenzolamide-HCA II complexes in which the same orientation as acetazolamide–HCA II complex exists (Chakravarty and Kannan, 1994; Vidgren et al., 1993). The stronger binding of aminobenzolamide can be rationalized by additional hydrophobic interactions between the phenyl ring and Phe 131 and Leu 198. Interestingly, the crystallographic recommended residues, which are coming into contact with acetazolamide molecule in the active site of the enzyme, are placed near the flexible tryptophanyl residues (e.g., Trp 190, Trp 243 which are determined in our work) reoriented in order to bind acetazolamide into the specific binding site of the enzyme (the active site). This similarity between our results and the crystallographic data also exists for the two types of the proposed bounding interactions between the acetazolamide and its binding sites on the enzyme (electrostatic interactions, e.g., van der Waals interactions or hydrogen bonds and hydrophobic interactions).
https://static-content.springer.com/image/art%3A10.1007%2Fs10930-007-9073-4/MediaObjects/10930_2007_9073_Sch1_HTML.gif
Scheme 1

Schematic representation of the acetazolamide–HCA II complex. This scheme is provided from Vidgren et al.(1990)

In summary, there are two isoforms of BCA II, each of which contains one catalytic site which can be occupied by one substrate or one acetazolamide molecule. This penetration of acetazolamide into the catalytic (regulatory) site allows some concealed or semiconcealed flexible residues (e.g., Trp 4 and Trp 190 by themselves or those critical aminoacyl residues which are placed near the cited tryptophans) to become partially exposed to the solvent. Under these conditions, the tertiary structure of the protein isoforms is minimally altered without any considerable changes in the secondary structure. These types of structural changes finally decrease the catalytic activity of the enzyme since one acetazolamide molecule penetrates into the catalytic site (especially via the hydrophobic interactions) and inhibits the settling of the substrate molecules for product formation.

Acknowledgments

Financial support for this work was provided by the Research Council of the University of Tehran. The authors would like to thank Dr. Christine Sorenson (University of Wisconsin, Madison, Wisconsin) for valuable comments. The authors would like to thank the Sealy Center for Structural Biology at the University of Texas Medical Branch and the RCSB organization for their valuable web service.

Copyright information

© Springer Science+Business Media, LLC 2007