Interaction of Biotin with Streptavidin
THERMOSTABILITY AND CONFORMATIONAL CHANGES UPON BINDING*

(Received for publication, May 20, 1996, and in revised form, December 2, 1996)

Martín González Dagger , Luis A. Bagatolli Dagger , Izaskun Echabe §, Jose L. R. Arrondo §, Carlos E. Argaraña Dagger , Charles R. Cantor and Gerardo D. Fidelio Dagger par

From the Dagger  Departamento de Química Biológica, Centro de Quimica Biologica de Cordoba (CIQUIBIC), Facultad de Ciencias Químicas, Universidad Nacional de Córdoba, 5000 Córdoba, Argentina, the § Departamento de Bioquímica, Universidad del País Vasco, E-48080 Bilbao, Spain, and the  Departments of Biomedical Engineering and Pharmacology, Boston University, Boston, Massachusetts 02215

ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
REFERENCES


ABSTRACT

The effect of biotin binding on streptavidin (STV) structure and stability was studied using differential scanning calorimetry, Fourier transform infrared spectroscopy (FT-IR), and fluorescence spectroscopy. Biotin increases the midpoint temperature Tm, of thermally induced denaturation of STV from 75 °C in unliganded protein to 112 °C at full ligand saturation. The cooperativity of thermally induced unfolding of STV changes substantially in presence of biotin. Unliganded STV monomer has at least one domain that unfolds independently. The dimer bound to biotin undergoes a single coupled denaturation process. Simulations of thermograms of STV denaturation that take into account only the thermodynamic effects of the ligand with a Ka ~1015 reproduce the behavior observed, but the estimated values of Tm are 15-20 °C lower than those experimentally determined. This increased stability is attributed to an enhanced cooperativity of the thermal unfolding of STV. The increment in the cooperativity is as consequence of a stronger intersubunit association and an increased structural order upon binding. FT-IR and fluorescence spectroscopy data reveal that unordered structure found in unliganded STV disappears under fully saturating conditions.

The data provide a rationale for previous suggestions that biotin binding induces an increase in protein tightness (structural cooperativity) leading, in turn, to a higher thermostability.


INTRODUCTION

The biological function of many proteins is triggered and modulated by the binding of ligands. For this reason, an understanding of the mechanism of protein-ligand interactions is essential for a detailed knowledge of protein function at the molecular level. Ligand binding, in most cases, involves the formation of noncovalent bonds at specific interacting surfaces between the protein and the ligand. The binding of a ligand can be accompanied by conformational changes at the protein site that sometimes are propagated throughout the entire protein. It is desirable to have a way to monitor these structural changes to understand any new properties acquired by the complex. The high affinity of the biotin-streptavidin binding not only offers useful bioanalytical advantages (1), but it also makes this system an attractive model for studying protein-ligand interactions (2-5). The biotin·STV1 association constant of about 1015 is the highest known in biochemistry.

In the present work we explore protein thermostability by heating STV in the absence of ligand or under conditions of partial or full ligand saturation. A dramatic increase in the Tm of protein denaturation, from 75 °C in absence of biotin to 112 °C at full ligand saturation, was revealed using differential scanning calorimetry (DSC). An analysis of the cooperativity of the denaturation was made on the basis of a reversible non-two-state model of protein unfolding to gain understanding of the system that unfolds differently if biotin is present.

Conformational changes were characterized by FT-IR and fluorescence spectroscopy. The large changes in the thermostability of STV·biotin complex as compared with the free protein can be explained by two simultaneous processes. One is related to the binding energy; the other is related to structural changes induced by the binding of biotin to STV.


EXPERIMENTAL PROCEDURES

Materials

All reagents used were of analytical grade. STV was from Life Technologies, Inc. Biotin was from Sigma.

Calorimetry

Protein denaturation was monitored with a MicroCal MC-2D scanning calorimeter with digital data acquisition and analyzed by means of the software provided by the manufacturer. The calorimetric data were analyzed assuming a reversible non-two-state model of denaturation (6). A possible justification for the applicability of reversible thermodynamics to apparently irreversible processes have been discussed previously (6-9) for the case where reversible unfolding is followed by a rate-limited irreversible step. This model was used after checking that no endotherms were found after complete denaturation and that similar thermograms, and results were obtained either: (i) at lower scan rates (26.8 degrees·h-1 and 9.2 degrees·h-1) than the one usually employed (55 degrees·h-1) or (ii) at different protein concentrations (from 0.020 to 0.080 mM) or (iii) by fitting the whole curve using only the first 50% of the experimental data of the thermogram. The calorimetric enthalpy (Delta H) is determined only by the area under the transition peak of the thermogram, while the van't Hoff enthalpy (Delta HVH) depends on the shape of the transition peak. The sharper the transition, the larger is the Delta HVH, independently of Delta H. The Delta H refers to heat change/mol, while Delta HVH is the heat change per unfolding unit (7). Thus, the ratio Delta H/Delta HVH can be thought of as the number of cooperative units/mol of STV monomer. To interpret the cooperative unit value for oligomeric proteins as a quantitative evidence for coupling of the denaturation process between the monomers, the calculations should be done on the basis of protein monomer concentration.

Biotin and STV were dissolved in 100 mM phosphate buffer, pH 7.3. The monomer protein concentration was 0.066 mM in all experiments and was determined at 280 nm using an extinction coefficient epsilon 0.1% = 3.4 (10). The molar ratios of biotin:STV (mole of biotin/mol of STV monomer) used were: 0, 0.25, 0.5, 0.75, and 2. Samples were exhaustively degassed before injection into the calorimeter cell (volume: 1.2388 ml) to prevent air bubble formation at the high temperatures reached (near 120 °C). A nitrogen pressure of 2 atmospheres was applied to both cells. The reference cell was filled with a matching buffer identical to that used with the sample. Samples containing biotin:STV at different molar ratios were mixed and incubated at room temperature for at least 10 min, in a final volume of 1.5 ml, before being degassed and injected into the calorimeter cell. Some representative DSC experiments were performed with the homogeneous pure core STV provided by Apcel Ltd., Berkshire, United Kingdom, with identical results to those found for the heterogeneous STV from Life Technologies, Inc.

Computer Simulation of the Excess Heat Capacity

The thermal denaturation of a protein that binds a ligand at a single binding site is predicted to behave at subsaturating levels of ligand, in an unimodal or bimodal manner, depending on the value of the association constant (11). This behavior arises solely from the coupling of the binding to the denaturing process. The model predicts the thermogram of a monomeric protein that binds to a ligand with no changes in the protein conformation or in the association constant as a function of temperature. Under these conditions the transition temperature Tm (temperature at half-completion), in the presence of ligand, changes according to Ref. 11,
T<SUB><UP>m</UP></SUB>=<FENCE><FR><NU>1</NU><DE>T°<SUB>m</SUB></DE></FR>−<FR><NU>R</NU><DE>&Dgr;H<SUB><UP>m</UP></SUB></DE></FR> <UP>ln</UP> (K<SUB>a</SUB>[<UP>L</UP>]+1)</FENCE><SUP><UP>−</UP>1</SUP> (Eq. 1)
where [L] is the free ligand concentration, Ka is the association constant, Hm is the calorimetric enthalpy of the process that occurs in absence of ligand at Tm°. The equilibrium constant for this process KLeft-right-arrow  D can be expressed in terms of the fractional extent of denaturation = KLeft-right-arrow  D/(KLeft-right-arrow  D + 1); as denaturation takes place, the ratio of the free ligand to the undenatured protein varies. Thus, the free ligand concentration [L] is a function of alpha . Values of Tm are obtained by numerical iteration of these variables as a function of the initial fraction of ligand added. The thermograms were recorded as dalpha /dT to obtain the excess heat capacity as a function of temperature, Cexcess (T) Delta Hm(dalpha /dT) (see Shrake and Ross (11) for a more detailed description of the method). For the calculations we used experimental Delta Hm and Tm° values of 150 Kcal/mol and 75 °C, respectively, determined for STV without biotin. The value of the association constant Ka used was 1015 M-1.

Infrared Spectroscopy

Infrared spectra were recorded in a Nicolet Magna 550 spectrometer equipped with an MCT detector. Samples for IR were made up by dissolving the lyophilized protein in 10 mM Hepes, pH or pD (H2O- or D2O-based buffers) 7.4 and placed in a thermostatted cell equipped with CaF2 windows. A path length of 50 µm was used for D2O measurements and 6 µm for H2O solutions. A total of 1000 scans were accumulated for each spectrum, using a shuttle device. Thermal stability studies were carried out by heating the samples in steps of about 3 °C, in the temperature interval 30-85 °C. After every heating step the sample was left to stabilize for 5 min and the spectrum recorded. Solvent subtraction, deconvolution, band position determination, and curve fitting of the amide I band were performed as reported previously (12). Briefly, band component positions are obtained from deconvolution and derivatization; initial heights are set to 90% in the center and the edges and 70% of the original heights in the other components. Bandwidth estimates are obtained from derivative spectra and the Gaussian fraction is set to 90%. The decomposition method assumes that the absorption coefficient is the same for all the components (12). The results obtained after iterations may not be unique, so restrictions must be applied (13). The most important points are that the band position cannot differ from the initial guesses by more than the distance between data points and that the width of the bands should be less than half of the amide I bandwidth. The combined use of several spectra, recorded at different temperatures below denaturation, makes the solution practically unique (12).

Fluorescence Spectroscopy

Corrected steady-state emission spectra and fluorescence phase lifetime measurements were performed with a phase modulation fluorimeter, SLM-Aminco model 4800C, equipped with a xenon arc lamp and thermostatted cell holder. The excitation wavelength was 295 nm in all fluorescence experiments, using a band-pass filter of 8 nm. Fluorescence intensity was corrected for inner filter effects (14) and dilution. All fluorescence experiments were carried out in triplicate, and the temperature was 25 °C (controlled by a circulating bath). Data were analyzed using the SLM 4800 C software package.

The fluorescence phase shifts for phase lifetime determinations were measured at 30 MHz. Fluorescence was observed using a 310-nm cut-off filter to eliminate scattered radiation from the light source. In the reference cell a solution of p-terphenyl in ethanol (tau  = 1.05 ns) was used to correct for possible color effects (14). The standard deviations of the phase lifetime were below 2%. Quenching experiments performed with acrylamide were done at a final concentration of 4.2 µM STV in 100 mM phosphate buffer, pH 7.3. Acrylamide was added by successive pipetting of small aliquots from a stock solution (4.8 M) into the cuvette.

Polyacrylamide Gel Electrophoresis

SDS-polyacrylamide gel electrophoresis analyses were performed in a Bio-Rad Miniprotean II using a 13% acrylamide gel according to the method described by Schägger and von Jagow (15). The cathodic buffer was 0.1 M Tris, 0.1 M Tricine, 0.1% (w/v) SDS, pH 8.25; the anodic buffer was 0.2 M Tris-HCl, pH 8.45; the sample buffer was 0.125 M Tris-HCl pH 6.8, 20% (v/v) glycerol, 4% (w/v) SDS, 0.01% (v/v) bromphenol blue. Each sample, containing 35 µg of STV or biotin:STV at different molar ratios was heated for 5 min at 95 °C, cooled down, then mixed with the sample buffer and incubated at room temperature for 10 min before application to the gel. For protein staining, the gel was previously washed in 20% methanol, 10% acetic acid for at least 30 min, then incubated for 30 min with gentle shaking in a mechanical shaker with 0.15% Coomassie Brilliant Blue R-250 (Sigma), dissolved in 50% methanol, 10% acetic acid. For destaining, the gel was left overnight in the same solution but without Coomassie Brilliant Blue R-250.


RESULTS

Calorimetric and Thermodynamic Analysis of STV·Biotin Thermal Denaturation

The effect of ligand on the thermally induced denaturation of STV was studied at different biotin:STV molar ratios. The results appears to be a biphasic or a monophasic process depending on the biotin concentration (the data in Fig. 1 and Table I are based on STV monomer concentration). A monophasic pattern is obtained only either in the absence of ligand or under fully saturating conditions. Biphasic behavior is observed at subsaturating biotin:STV 0.25 or 0.5 molar ratios; this is on average one or two molecules of biotin per tetramer, respectively (Fig. 1). The thermogram of free STV shows a single, symmetric peak with a Tm centered at 75.5 °C (Fig. 1 and Table I). Under biotin-saturating conditions, the Tm of STV is considerably increased and centered at 112.2 °C (Fig. 1). Denaturation of ligand-stabilized STV is marked by a higher excess heat capacity peak, reaching a value close to three times the one obtained for STV, and by an increase in the calorimetric enthalpy (Fig. 1 and Table I). Concomitantly an increase in the cooperativity of the denaturation process, given by a cooperative unit lower than 1, is also observed (Table I). Each single thermogram peak of unliganded or fully liganded STV can be deconvoluted into two transition components (not shown). This can be attributed to the two major different species of STV present in the samples studied (16). In Table I we show calorimetric data for the overall denaturation process, which is an average over the properties of these two species.


Fig. 1. DSC thermograms for STV under different ligand conditions. Bimodal patterns are observed only at biotin:STV molar ratios of 0.25 and 0.5. Protein concentration and scan rates were 0.066 mM and 55.6 degrees·h-1, respectively, in all runs. Inset, simulated thermogram for a hypothetical protein with a Ka of 1015 M-1. Y axis legend: excess heat capacity (Kcal·degree-1·mol-1); X axis legend: temperature ( °C). The parameters necessary for the simulation were: the enthalpy of the STV denaturation process (150 kcal·mol-1), the half-completion temperature (Tm) of the denaturation process (75 °C), the protein concentration (100 µM), and the different initial fractions of ligand added: 0.05 (--), 0.25 (- - - -), 0.50 (--- --- --- ---), 0.75 (- - - -), 0.99 (· · · ·), 0.999 (---·---·---).
[View Larger Version of this Image (29K GIF file)]


Table I.

Thermodynamic parameters of thermal denaturation of the STV-biotin complex


Conditiona Calorimetric enthalpy Tm Cooperative unitsb

Kcal/mol °C
STV 150.3 75.5 1.70
Biotin:STV (0.25)
  Low T peak 95.7 78.0 1.10
  High T peak 57.2 108.0 0.36
Biotin:STV (0.50)
  Low T peak 64.4 81.1 0.84
  High T peak 90.0 108.0 0.42
Biotin:STV (0.75) 171.3 110.2 0.43
Biotin:STV (2) 209.8 112.2 0.58

a Molar ratio based in moles of biotin/mol of STV monomer.
b Cooperative units defined as the ratio of the calorimetric and van't Hoff enthalpies (Privalov, 1982), see "Experimental Procedures" for details.

Subsaturating levels of a high affinity ligand have effects on thermally induced protein denaturation that arise from the coupling of protein denaturation and ligand binding equilibria. In the model developed by Shrake and Ross (11), at subsaturating ligand levels a biphasic thermogram can be obtained, since the ratio of the free ligand concentration to the native protein is continuously changing as the denaturation proceeds. The midpoint denaturation temperature increases proportionally with the free ligand concentration and the value of the association constant Ka (see Equation 1 under "Experimental Procedures"). The model assumes that the Ka and the calorimetric enthalpy are independent of the temperature (11). The inset in Fig. 1 shows the simulated thermograms for a hypothetical STV-like protein with a calorimetric denaturation enthalpy of 150 Kcal/mol, a Tm of 75 °C, and a Ka = 1015 for an initial fractional saturation of 0.05, 0.25, 0.50, 0.75, 0.99, and 0.999. At intermediate levels of ligand saturation, the bimodality is evident and, at full saturation, the biphasic behavior disappears, and the Tm is increased to near 96 °C.

The thermostability induced by high affinity ligand binding, taking into account only the coupling of protein denaturation process to the binding equilibrium, explains only partially the experimental thermograms observed for STV in presence of biotin (compare the Fig. 1 with its inset). The experimental midpoint denaturation temperature values for the STV·biotin complex are higher than the theoretical calorimetric parameters expected for such an interaction. Also, at 0.75 fractional saturation the theoretical thermogram predicts that a considerable fraction of the protein will denature at a lower temperature (inset in Fig. 1). However, the experimental thermogram at a similar biotin:STV molar ratio shows just a single peak at 110 °C.

The observed changes in the cooperativity of thermally-induced unfolding provides calorimetric evidence that there is an increment in the strength of the intermolecular association of the subunits induced by biotin binding. The ratio of the calorimetric enthalpy (the area under the transition peak) to the van't Hoff enthalpy (calculated from the shape of the transition peak) indicates whether protein denaturation represents a simple or a complex cooperative system (7). Since all the calorimetric data shown in Table I were calculated on a basis of STV monomer concentration, the ratios are referred to the numbers of unfolding units of the system. In the absence of biotin, STV denatures with a cooperative unit value closer to 2 than 1 (Table I), indicating probably more than one domain in the STV monomer that would unfold independently. At 0.50 biotin:STV molar ratio the cooperative unit is about 0.5 in the higher transition component, indicating that the dimer undergoes a single coupled transition. As the molar ratio increases the intermolecular interaction among the subunits is larger, and a cooperative unit value higher than 1 is no longer observed. Even when there is only one molecule of biotin per tetramer (0.25 biotin:STV molar ratio), the unliganded subunits are affected by the biotin bound subunit (see Table I and Fig. 1).

The effect of biotin on the stability of different oligomeric forms of STV after heating was investigated. Samples of STV at different biotin molar ratios were subjected to heating (95 °C, 5 min) and analyzed by SDS-polyacrylamide gel electrophoresis (Fig. 2). Non-heated STV remains as a ~60-kDa tetramer, and in the absence of biotin the treatment produces mainly protein bands of ~14 and ~30 kDa, corresponding to the monomer and dimer forms of the protein, respectively. A progressive increase in the biotin:STV molar ratios increases the proportion of the tetramer, and at a 0.75 biotin:STV molar ratio, most of the protein is in the tetrameric form, in agreement with the DSC results. The electrophoretic observations support strongly the idea that a structural cooperativity is built up upon biotin binding to STV. Sano and Cantor (17) reported previously that the exchange of free and radioactive biotin in biotin-liganded STV is hard to achieve even in the presence of high concentrations of free ligand. This could indicate that STV would not release the biotin prior to denaturation, instead the STV·biotin complex might undergo the unfolding transition. The idea of an increment in the intermolecular communication between the STV subunits upon biotin binding was suggested by previously the electrophoretic behavior of the complex in presence of 6 M urea (18). This idea is further supported by site-directed mutagenesis of tryptophan 120, which indicates that this residue, which contacts bound biotin, is strongly involved in intersubunit contacts (17).


Fig. 2. SDS-polyacrylamide gel electrophoresis of STV and biotin:STV (tetramer) at different molar ratios. Lanes: 1, non-heated STV; 2, STV heated at 95 °C for 5 min; 3, biotin:STV 0.125 molar ratio heated at 95 °C for 5 min; 4, 0.25 molar ratio heated at 95 °C for 5 min; 5, 0.33 molar ratio heated at 95 °C for 5 min; 6, 0.50 molar ratio heated at 95 °C for 5 min; 7, 0.75 molar ratio heated at 95 °C for 5 min; 8, 2 molar ratio heated at 95 °C for 5 min. Molecular mass standards (Std), from the top: 66 kDa, bovine albumin; 45 kDa, albumin egg; 36 kDa, glyceraldehyde-3-P dehydrogenase; 29 kDa, carbonic anhydrase bovine; 14 kDa, alpha -lactalbumin bovine. The protein mass was the same for all lanes, but the staining is, evidently, different for the tetramer than the monomer (compare lanes 1 and 2); this is probably due to the higher packing of STV in the tetramer.
[View Larger Version of this Image (68K GIF file)]


Secondary Structure of Streptavidin by FT-IR Spectroscopy

The secondary structure of STV has been studied by means of infrared spectroscopy. The amide I band envelope from samples in H2O and D2O buffers was decomposed into its constituents. Fig. 3 shows the spectrum of STV in the absence (A) or presence (B) of biotin in a D2O buffer. Table II presents the band positions and the relative area of the components shown in Fig. 3. Seven bands are seen that arise from conformations of the peptide backbone. These are located at around 1692, 1681, 1671, 1641, 1633, and 1629 cm-1. The assignment of the amide I component bands to conformational structures is not yet straightforward because of the complexity of the interactions involved (19). Still, some of the bands can be unambiguously assigned (13). In STV, a high content of beta -sheet is expected because the maximum of the amide I band is around 1630 cm-1, which is coherent with the previous x-ray studies (5, 20). Also, a high frequency band around 1680 cm-1 is produced by antiparallel beta -sheet. The band at 1641 cm-1 is due to unordered structure, and bands between 1660 and 1690 cm-1 arise from beta -turns. alpha -Helix usually gives rise to a band around 1650 cm-1; however, bands originating from turns, with dihedral angles comparable with alpha -helix, have also been described at this frequency (21, 22). Binding of biotin to STV does not change the beta -sheet structure content; that remains at 40%, but the two bands at 1641 and 1651 cm-1 in STV alone appear as a single band at 1647 cm-1 after biotin binding (Table II).


Fig. 3. Decomposition of the band components of the amide I IR spectrum, obtained for free STV (A) and biotin:STV at a molar ratio of 2 (B) in D2O buffer. For details see "Experimental Procedures."
[View Larger Version of this Image (19K GIF file)]


Table II.

Infrared parameters for the various components of the protein in D2O of the STV amide I band, for the free and biotin-bound forms of the protein


STV
STV saturated with Biotin
Band position Area Band position Area

cm-1 % cm-1 %
1692 1.1 1692 1.1
1681 6.3 1682 5.5
1671 4.7 1672 3.2
1663 11.5 1662 17.2
1651 21.9 1647 33.2
1641 13.9
1629 40.7 1630 39.7

The integrity of STV conformation with temperature was also studied by FT-IR (Fig. 4). In our hands, this methodology permits heating up to 85 °C only. STV in the presence of saturating biotin concentrations does not change its FT-IR profile significantly in the temperature range tested (Fig. 4A), while free STV undergoes thermal denaturation with a half-midpoint temperature of 72-73 °C (Fig. 5). This agrees with the Tm value observed by DSC. The denaturation process followed by FT-IR is revealed by broadening of the amide I region due to the emergence of bands at 1620 and 1680 cm-1, both associated with denatured conformations (23). The increase in regular structures associated with a raise in thermal denaturation temperature has also been observed in concanavalin A, a protein with a high beta -sheet content upon metal binding (24).


Fig. 4. Thermally induced changes of STV in the 1800-1500 cm-1 IR spectrum of streptavidin in the presence (A) and absence (B) of biotin.
[View Larger Version of this Image (57K GIF file)]



Fig. 5. Half-width of the amide I band of the IR spectra, as a function of temperature, for STV in the presence (open circle ) or absence of biotin (bullet ).
[View Larger Version of this Image (18K GIF file)]


Fluorescence Experiments

STV has six tryptophans (16). Four of these tryptophan residues are clustered in the hydrophobic binding site in close contact with biotin. The binding pocket is located in a shared interface contact domain, since one of the tryptophans, residue 120, belongs to the adjacent subunit (5, 17, 20). As expected, biotin binding leads to considerable changes in STV fluorescence. When biotin binds to STV, a progressive blue shift and a decrease in fluorescence intensity are observed. Under saturation conditions a blue shift of about 5 nm (from 330 to 325 nm) together with a 25% decrease of the fluorescence intensity are observed (not shown) in agreement with the data reported by Kurzban et al. (25). Control experiments performed by mixing N-acetyltryptophanamide, used as a model tryptophan-containing protein, and biotin do not reveal appreciable changes in N-acetyltryptophanamide fluorescence. These results indicate that neither dynamic nor static tryptophan quenching are the direct result of the presence of biotin (data not shown). Thus, changes observed in lambda max and fluorescence intensity when biotin binds to STV are likely to arise mainly from conformational changes of the protein and not directly from a quenching effect of the ligand.

To further explore the influence of biotin binding on the fluorescence of tryptophan in STV, we performed quenching experiments with acrylamide in the presence and absence of ligand. We also measured the phase lifetime of STV and the STV·biotin complex. The major contribution to the total fluorescence emission intensity, when STV is excited at 295 nm, is provided by the four tryptophans at the binding site of each subunit. The accessibility of these amino acids to the water-soluble, neutral quencher acrylamide was evaluated from steady-state fluorescence data by using the modified form of the Stern-Volmer equation: Fo/F = eVapp[Q] (1 + KSV(app) [Q]); where F and Fo are, respectively, the fluorescence intensities in the presence and in the absence of quencher at concentration [Q]. KSV(app) and Vapp are the apparent dynamic and static contributions to the total quenching, respectively. These parameters were obtained from the best fit of the experimental data. In the case of a multitryptophan protein, the Stern-Volmer parameters are a crude estimate of the average exposure of the fluorescent residues in the protein (26). In unliganded STV, the quenching takes place without a shift in the lambda max of the fluorescence emission spectrum (not shown). This suggests that all forms of tryptophan have similar exposure to the quencher. In contrast, the conformational changes in STV that accompanies biotin binding make these tryptophans practically inaccessible to acrylamide. Consequently, higher values of KSV(app) and Vapp were found for STV alone compared than for biotin:STV at saturating conditions. The results are summarized in Table III.

Table III.

Fluorescence parameters for free and biotin-saturated STV


Sample Phase lifetime KSV(app) V(app)

ns M-1 M-1
STV 1.37  ± 0.03 3.56 0.53
Plus acrylamide 1.01  ± 0.01
STV:biotin 0.70  ± 0.01 0.89 ~0
Plus acrylamide 0.66  ± 0.01

We also measured the fluorescence phase lifetime of STV and STV:biotin in the absence and in the presence of 0.15 M acrylamide. In the absence of acrylamide, the phase lifetime of STV is reduced by almost 50% when biotin binds, reflecting the conformational change that occurs upon ligand binding. In the presence of acrylamide, a reduction of about 30% in phase lifetime of STV was found, whereas for the STV·biotin complex the phase lifetime is practically unchanged (Table III). These results are in agreement with the values obtained for the Stern-Volmer constants, suggesting that the solvent exposure of the tryptophans is strongly reduced in the presence of biotin.


DISCUSSION

Previous x-ray studies of a truncated form of STV have shown a beta -barrel involving around 70% of the residues of this form (5, 20). This barrel involves the beta -strands and the connecting turns, besides other turns and two flexible loops have been described. Our FT-IR studies on the unliganded protein also show antiparallel beta -sheet as the major structure. In the presence of biotin, no changes in the beta -sheet conformation are produced, confirming the previous x-ray data (20). The most striking feature of the infrared spectrum as compared with the x-ray structure is the presence of a band at 1651 cm-1 that could be related in principle to alpha -helix structure. Note that our samples consist primarily of full-length STV. Additionally, as demonstrated by Krimm and Bandekar (21, 22), bands around 1653 cm-1 can also arise from beta -turns that have dihedral angles similar to a turn of the helix. In other beta -sheet proteins, such as concanavalin A, a band around 1651 cm-1 has also been described, even if no structured alpha -helices have been found in the x-ray structure (24). Thus, this band can arise either from helical segments that would be in the full-length species of our sample (about 60%), or more probably they correspond to turns with a helix-like geometry. The presence of biotin produces a significant change in the infrared spectrum. This is appreciable even in the crude spectrum (Fig. 3), it corresponds to an increase in the intensity at 1647 cm-1 at the expense of the bands at 1651 and 1641 cm-1 (Table II). With the help of the H2O spectrum (data not shown), where this shift is not observed, it can be concluded that the component at 1641 cm-1, associated with unordered structure, has decreased. Thus biotin binding leads to a more structured protein. In fact, the x-ray results showed that two loops that cannot be defined in the unliganded STV are defined in the presence of biotin (20). Also, from the infrared results it can be suggested that the residues not included in the truncated form are disordered, and after biotin binding they become structured. This assumption is sustained by the large increase in denaturation temperature observed after biotin binding. Fluorescence quenching experiments are also in keeping with the higher structural order observed upon biotin binding, since a drastic decrease of the aqueous quencher accessibility to the tryptophan residues is found for liganded as compared with free STV (Table III). A difference of around 40 °C in the denaturation temperature should imply a protein reorganization similar to that observed in concanavalin A after demetallization (24), where no changes in the beta -sheet structure are produced, but variations in the protein flexible segments are seen.

It is well known that ligand binding gives an apparent increment in the thermal stability of a protein due only to the coupling of binding with unfolding (11). So, it is clear that not only thermodynamic factors are influencing the experimental STV stability upon thermally induced unfolding (Fig. 1). The increase in structural order seen by previous authors in x-ray crystals (5, 20) and also confirmed by FT-IR in a more dynamic aqueous solution condition is perhaps the main reason for the increment from 75 to 112 °C in the thermal stability of the STV·biotin complex. The calorimetric data also indicate an increase in the cooperativity of the unfolding process. The sharpening and the 3-fold increment in the excess heat capacity changes the calorimetric cooperative unit from 1.7 in unliganded STV to 0.5 at full ligand saturation (Table I). A cooperative unit in calorimetry is a thermodynamic concept that may be related to a structural domain and represents a region of a protein with a thermodynamically stable structure (7). The structural changes in STV upon biotin binding as observed by several techniques include Trp-120 from one subunit in the dimer moving across the surface binding site of the second interacting subunit (20). At full ligand saturation, the increased packing due to the interdigitation of the region that contains Trp-120 apparently leads the dimers to undergo a coupled single transition upon heating. In unliganded STV the interaction of the Trp-120 domain in the binding cleft might not be strong enough, and this would imply the possibility of observing some intermediate state in the transition ascribable to one of the cooperative units (or units of unfolding). In a non-cooperative model of the biotin-STV interaction (see below) and at subsaturating conditions, a considerable portion of the unliganded STV monomer is in close contact with the neighboring liganded monomer. So, it is probable that, at subsaturating conditions, part of free STV monomer melts at a lower temperature, and a second domain (including the loop containing Trp-120) melts cooperatively with the liganded neighboring STV. This may explain the lower calorimetric enthalpy of the low transition component found at subsaturating conditions (Table I). At full saturation, the molecular interdigitation between the STV monomers is at maximum, and the system undergoes denaturation with maximum calorimetric enthalpy (Table I). Replacement of Trp-120 by Phe reduces substantially the affinity constant to approximately 108 M-1, indicating that the contact made by Trp-120 to biotin has a considerable contribution to the tightness of biotin binding to STV (17). So, it can be expected that the Phe-120 mutant would behave with a different cooperativity of unfolding upon heating than the native protein.

The association constant, Ka for the binding of biotin to STV is among the highest for any known non-covalent structure (2, 17). Even with this huge Ka it has been suggested previously that biotin induces binding cooperativity (17, 18, 27). Usually the term cooperativity in binding is used to explain changes in the affinity (with Ka in micromolar-nanomolar range) of an oligomeric protein in the presence of increasing amounts of bound ligand. The typical textbook example is oxygen binding to tetrameric hemoglobin, in which one molecule of oxygen increases the affinity of the second and so forth. It is very difficult to imagine (and experimentally difficult to evaluate) whether the cooperativity that appears to occur with biotin binding to STV works in a similar way than that observed for the interaction of oxygen with hemoglobin, even when the quaternary structural changes seen in STV are consistent with the possibility of cooperativity.

A model in which the conformational communication between dimers in the tetrameric structure seen in the present work and in previous papers, induced by the interaction of a first ligand, influences the affinity of the remaining ligands giving an overall macroscopic Ka of biotin (~1015) is not in agreement with the calorimetric data. Instead, biotin appears to induce structural cooperativity into the subunits without substantial changes in the binding affinity. According to the conventional picture of cooperative binding, at subsaturating conditions of 0.25 biotin:STV monomer, the fully liganded tetrameric and unliganded species should predominate in the population of the species (option A in Fig. 6). If this were the case, and assuming that STV melts with the biotin attached, the calorimetric profile should be biphasic with a main transition centered at 75 °C (STV tetramer without biotin) and a transition centered at 112 °C (fully liganded STV tetramer). However, the actual experiments shown an intermediate calorimetric profile that depends on the occupancy (biotin:STV molar ratios, see Fig. 1), so option B, pictured in Fig. 6, is more consistent with the data. Evidence for non-cooperativity in biotin binding to STV has been recently reported by Jones and Kurzban (28) by using equilibrium binding and separation of the present species by anion exchange chromatography. Binding appears to occur with cooperative structural changes without affecting the binding affinity of the overall process. The STV-biotin system may be thought as an extreme case of Koshland-Nemethy-Filmer sequential model of binding (29).


Fig. 6. Schematic illustration of two extreme models for STV-biotin interaction. Cooperative (A) and non-cooperative (B) binding with changes in the conformation of unliganded protein neighbors propagated as a consequence of the binding of biotin to one STV subunit.
[View Larger Version of this Image (30K GIF file)]



FOOTNOTES

*   The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
par    To whom correspondence should be addressed: Dept. de Química Biológica, CIQUIBIC, Faculdad de Ciencias Químicas, Ciudad Universitaria, Apartado Postal 4-Casilla de Correo 61, 5000-Córdoba Argentina. Tel.: 54-51-334171 or 54-51-334168; Fax: 54-51-334074; E-mail: gfidelio{at}dqbfcq.uncor.edu.
1   The abbreviations used are: STV, streptavidin; DSC, differential scanning calorimetry; FT-IR, Fourier transform infrared spectroscopy; Tricine, N-[2-hydroxy-1,1-bis(hydroxymethyl)ethyl]glycine.

REFERENCES

  1. Wilchek, M., and Bayer, E. A. (1988) Anal. Biochem. 171, 1-32 [Medline] [Order article via Infotrieve]
  2. Vajda, S., Weng, Z., Rosenfeld, R., and DeLisi, C. (1994) Biochemistry 33, 13977-13988 [Medline] [Order article via Infotrieve]
  3. Miyamoto, S., and Kollman, P. A. (1993) Proc. Natl. Acad. Sci. U. S. A. 90, 8402-8406 [Abstract/Free Full Text]
  4. Weber, P. C., Pantoliano, M. W., and Thompson, L. D. (1992) Biochemistry 31, 9350-9354 [Medline] [Order article via Infotrieve]
  5. Hendrickson, W. A., Palher, A., Smith, J. L., Satow, Y., Merritt, E. A., and Phizackerley, R. P. (1989) Proc. Natl. Acad. Sci. U. S. A. 86, 2190-2194 [Abstract]
  6. Sturtevant, J. (1987) Annu. Rev. Phys. Chem. 38, 463-488 [CrossRef]
  7. Privalov, P. L. (1982) Adv. Protein Chem. 35, 1-104 [Medline] [Order article via Infotrieve]
  8. Lepock, J. R., Ritchie, K. P., Kolios, M. C., Rodahl, A. M., Heinz, K. A., and Kruuv, J. (1992) Biochemistry 31, 12706-12712 [Medline] [Order article via Infotrieve]
  9. Sanchez-Ruiz, J. M. (1992) Biophys. J. 61, 921-935 [Abstract]
  10. Green, N. M., and Melamed, M. D. (1966) Biochem. J. 100, 614-621 [Medline] [Order article via Infotrieve]
  11. Shrake, A., and Ross, P. D. (1990) J. Biol. Chem. 265, 5055-5059 [Abstract/Free Full Text]
  12. Arrondo, J. L. R, Castresana, J., Valpuesta, J. M., and Goñi, F. M. (1994) Biochemistry 33, 11650-11655 [Medline] [Order article via Infotrieve]
  13. Arrondo, J. L. R., Muga, A., Castresana, J., and Goñi, F. M. (1993) Prog. Biophys. Mol. Biol. 59, 23-56 [CrossRef][Medline] [Order article via Infotrieve]
  14. Lakowicz, J. R. (1983) Principles of Fluorescence Spectroscopy, pp. 39-46 and 86-91, Plenum Press, New York
  15. Schägger, H., and Von Jagow, G. (1987) Anal. Biochem. 166, 368-379 [Medline] [Order article via Infotrieve]
  16. Argaraña, C. E., Kuntz, I. D., Birkin, S., Axel, R., and Cantor, C. R. (1986) Nucleic Acids Res. 14, 1871-1882 [Abstract]
  17. Sano, T., and Cantor, C. R. (1995) Proc. Natl. Acad. Sci. U. S. A. 92, 3180-3184 [Abstract]
  18. Sano, T., and Cantor, C. R. (1990) J. Biol. Chem. 265, 3369-3373 [Abstract/Free Full Text]
  19. Surewicz, W. K, Mantsch, H. H., and Chapman, D. (1993) Biochemistry 32, 389-394 [Medline] [Order article via Infotrieve]
  20. Weber, P. C., Ohlendorf, D. H., Wendolosky, J. J., and Salemme, F. R. (1989) Science 243, 85-88 [Medline] [Order article via Infotrieve]
  21. Krimm, S., and Bandekar, J. (1992) Adv. Protein Chem. 38, 181-364
  22. Bandekar, J. (1992) Biochim. Biophys. Acta 1120, 123-143 [Medline] [Order article via Infotrieve]
  23. Fernandez-Ballester, G., Castresana, J., Arrondo, J. L. R., Ferragut, J. A., and Gonzalez-Ros, J. M. (1992) Biochem. J. 288, 421-426 [Medline] [Order article via Infotrieve]
  24. Arrondo, J. L. R., Young, N. M., and Mantsch, H. H. (1988) Biochim. Biophys. Acta 952, 261-268 [Medline] [Order article via Infotrieve]
  25. Kurzban, G. P., Gitlin, G., Bayer, E. A., Wilchek, M., and Horowitz, P. M. (1990) J. Protein Chem. 9, 673-682 [Medline] [Order article via Infotrieve]
  26. Eftink, M. R., and Ghiron, C. A. (1977) Biochemistry 16, 5546-5551 [Medline] [Order article via Infotrieve]
  27. Sano, T., Pandori, M. W., Smith, C. L., and Cantor, C. R. (1994) in Advances in Biomagnetic Separation (Uhlen, M., Hornes, E., and Olsvik, O., eds), pp. 21-29, Eaton Publishing, Natick, MA
  28. Jones, M. L., and Kurzban, G. P. (1995) Biochemistry 34, 11750-11756 [Medline] [Order article via Infotrieve]
  29. Cantor, C. R., and Schimmel, P. R. (1980) Biophysical Chemistry, Part III: The Behavior of Biological Macromolecules, pp. 945-978, W. H. Freeman & Co., New York

©1997 by The American Society for Biochemistry and Molecular Biology, Inc.