Protein J (2009) 28:1–13 DOI 10.1007/s10930-008-9158-8
An Anionic Porphyrin Binds b-Lactoglobulin A at a Superficial Site Rich in Lysine Residues Ivan Silva Æ Samuel Sansone Æ Lorenzo Brancaleon
Published online: 6 January 2009 Ó Springer Science+Business Media, LLC 2008
Abstract Binding of small ligands to globular proteins remains a major research topic in biophysics. We have studied the binding of several photoactive dyes to b-lactoglobulin (BLG), as a model to investigate the photoinduced effects of porphyrins on proteins. A combination of optical spectroscopies (fluorescence, circular dichroism) and molecular docking simulations were used to estimate the pH-dependence of the binding parameters and the docking location for meso-tetrakis(sulfonatophenyl)-porphyrin (TPPS). We have observed that the binding of TPPS is not modulated by the pH-mediated conformational transition of the protein (i.e., Tanford transition). Binding of TPPS appears to occur with some degree of negative cooperativity. Moreover, TPPS remains bound even upon partial denaturation of the protein. These results are consistent with a superficial binding site at a location removed from the aperture of the interior b-barrel. Binding occurs through electrostatic interactions between the negative SO3- groups of TPPS and positively charged Lys and Arg residues. This is the first study that explores the interaction of an anionic porphyrin with BLGA in a pH range that spans across the Tanford transition. Establishing the location of the binding site will enable us to explain the photoinduced conformational effects mediated by TPPS on BLG.
Electronic supplementary material The online version of this article (doi:10.1007/s10930-008-9158-8) contains supplementary material, which is available to authorized users. I. Silva S. Sansone L. Brancaleon (&) Department of Physics and Astronomy, University of Texas at San Antonio, One UTSA Circle, San Antonio, TX 78249, USA e-mail:
[email protected]
Keywords Lactoglobulin Porphyrin Fluorescence spectroscopy Binding Photodynamic therapy (PDT) Abbreviations PPIX Protoporphyrin IX b-lg b-Lactoglobulin GI Gastrointestinal tract DMSO Dimethylsulfoxide KI Potassium iodide S–V Stern–Volmer Trp Tryptophan TPPS4 Meso-tetrakis(sulfonatophenyl)-porphine ANS 1-Anilinonaphthalene-8-sulfonate FRET Fluorescence resonance energy transfer CD Circular dichroism
1 Introduction Because of its small size (162 amino acids) and its well characterized structure [1, 2], b-lactoglobulin (BLG) represents a viable protein model for the investigation of the effects of photoactive ligands (such as porphyrins) on globular proteins. Following our study on the binding of protoporphyrin IX (PPIX) to BLG [3], we have recently shown that irradiation of the porphyrin in the PPIX/BLG complex induces partial unfolding of the protein [4]. Light-induced processes have been extensively applied to study the dynamic structure of proteins [5, 6]. The direct photoinduced, porphyrin-mediated conformational effects [4] could also have consequences for biomedical applications since porphyrins have been used as experimental and/ or clinically useful [7, 8] photosensitizing drugs in photodynamic therapy (PDT). The direct targeting of proteins in
123
2
PDT has been historically overlooked despite the mounting evidence that proteins are indeed directly targeted during cell irradiation in the presence of various photosensitizers [9–11] and that such damage is not confined to membrane proteins [10, 12]. Understanding light-induced, porphyrinmediated conformational changes of model globular proteins will therefore be useful to understand some of the mechanisms that produce protein damage during PDT, but also provide a new approach that could intentionally target specific proteins, thus providing a potential paradigm shift in the PDT protocol. A fundamental element in the study of photoinduced, porphyrin-mediated conformational changes of proteins is establishing the location of the binding site for the porphyrin. Binding of meso-tetrakis(sulfonatophenyl)-porphine (TPPS) to BLG has been investigated near neutral and acidic pH values [13] where the porphyrin is a tetra-anion (TPPS4-) and a di-anion (TPPS2-), respectively. In the acidic regime where BLG is a monomer and TPPS2- tends to form J-dimers, it was found that the protein itself can induce the formation of J-aggregates [13]. Near neutral pH where BLG is dimerized and TPPS4- is monodispersed the signature band of the bound J-aggregate disappeared [13]. The study, however, did not include a wider pH range that could probe the effects of the conformational changes of BLG (the so-called Tanford transition) [14]. We analyzed the binding of TPPS4- to b-lactoglobulin variant A (BLGA) at 6 B pH B 9. In such pH range BLGA is a dimer and each monomer undergoes a conformational transition near pH 7.5 [14, 15]. In this range the presence of TPPS2- or its J-aggregates is negligible [16]. The combined used of emission spectroscopy, fluorescence lifetime measurements, circular dichroism spectroscopy (CD) and computational docking simulations, provided a likely scenario for the binding site of TPPS4- to BLGA and revealed the docking to a superficial protein location stabilized by electrostatic interactions between the negatively charged SO3- groups of TPPS4- and positively charged amino acid residues. This superficial binding site is in agreement with what had been suggested previously [13].
2 Materials and Methods
I. Silva et al.
a mass of 18,278 Da for BLG which is within 1 Da of the NCBI sequence gi|6729725. This sequence differs from the expected BLGA sequence for eight amino acids in the C-terminal region of the protein. 2.1.1 Buffers Aqueous phosphate buffers were prepared using deionized water. One tablet of phosphate buffered saline tablets (Sigma–Aldrich) was dissolved in 200 mL of water to produce a 10 mM buffer at pH 7.4. Buffers were subsequently adjusted to the correct pH by addition of small aliquots of a 0.5 M solution of HCl or a 0.1 M solution of NaOH. In order to investigate the effects of the BLGA conformational transition [14] on the porphyrin binding, the pH of the buffer was adjusted from 6 to 9 at increments of one pH unit. 2.2 Sample Preparation 2.2.1 TPPS UV–Vis Absorption and Fluorescence The porphyrin was dissolved directly in buffer and the concentration adjusted to 0.5 lM using absorption spectroscopy, assuming e413 = 5.10 9 105 M-1 cm-1 in aqueous solution [17]. The fluorescence of TPPS was recorded with excitation at 413 nm and emission in the 580–750 nm range. A stock solution of BLGA at the same pH as the porphyrin solution was prepared at a concentration of 30 lM by dissolving it directly in buffer. The concentration of the protein in solution was always determined by spectrophotometric analysis assuming e280 = 1.76 9 104 M-1 cm-1. The LC/MS data did not detect any aromatic impurity in our sample, therefore we assumed that the content in aromatic amino acids (Tyr ? Trp) was identical to the one expected for other BLG molecules whose molar extinction coefficient, e280, is the one shown above [13, 18]. Increasing aliquots (from 10 to 100 lL) of the protein solution were added directly into 1.5 mL of the aqueous solution of TPPS. The fluorescence spectrum of TPPS was recorded after each addition of the protein. The absorption spectra of the same solutions were also recorded from 250 to 500 nm and used to correct for sample dilution and to normalize the emission spectra [3] but also to survey the possible formation of J- or other aggregates.
2.1 Chemicals 2.2.2 BLGA Fluorescence Quenching TPPS and bovine lactoglobulin (genetic variant A, BLGA) were purchased from Sigma–Aldrich (St. Louis, MO) and used without further purification. Electrophoretic analysis of BLGA under denaturing conditions revealed [99% purity. In addition we analyzed the BLG sequence using capillary liquid chromatography/mass spectrometry (LC/ MS) which confirmed the purity of the protein and revealed
123
The solutions for the TPPS-induced quenching of BLGA contained 7 lM of the protein (OD295 & 0.1) [19]. Excitation of protein fluorescence was carried out at 295 nm which selectively excites the Trp residues of BLGA. A 20 lM stock solution of TPPS in buffer was prepared and added at increasing aliquots to a 1.5 mL volume of the
An Anionic Porphyrin Binds b-Lactoglobulin A
3
BLGA solution. The intrinsic fluorescence of the protein was recorded between 300 and 450 nm after each porphyrin addition. Absorption values at the excitation wavelength were recorded for all samples and used to normalize the emission spectra of the protein and correct for the dilution of the sample [3]. 2.2.3 FRET, Fluorescence Anisotropy and Circular Dichroism Spectroscopy All these experiments were carried out at a single BLGA:TPPS molar ratio where the fluorescence spectra indicated the maximum shift of the emission (thus binding saturation). FRET experiments were carried out using steady state and time resolved fluorescence. 2.3 Instrumentation Absorption spectra were recorded on a dual beam spectrophotometer (Evolution 300, Thermo Scientific, Waltham, MA). All spectra were recorded at a rate of 240 nm/min and a spectral resolution of 2 nm. Appropriate baseline and reference cells were used for each scan. Circular Dichroism (CD) experiments were carried out using a J-810 spectropolarimeter (Jasco Inc., Easton, MD). Steady state fluorescence and fluorescence anisotropy were recorded using a double-monochromator fluorimeter (AB2, Thermo Scientific, Waltham, MA). Spectra were recorded at a speed of 1 nm/s and spectral resolution of 4 nm in both emission and excitation. Fluorescence lifetime was carried out using an IBH 5000U instrument (Horiba JobinYvon, Edison, NJ). The fluorescence decay of TPPS was recorded upon excitation with a pulsed diode laser at 405 nm (NanoLED-405L, pulsewidth * 150 ps); the fluorescence decay of BLGA was recorded upon excitation with an LED source at 295 nm (NanoLED-295, pulsewidth & 750 ps). Both sources are manufactured by IBH Ltd. (Glasgow, UK) and are operated at a repetition rate of 1 MHz. 2.4 Methods
The fluorescence spectra of the porphyrin as a function of the addition of the protein were analyzed using Gaussian fitting of the spectra as reported in detail in a previous manuscript [17]. In summary, each spectrum of the tetraanion TPPS4- was analyzed according to.
ðkKf Þ2 Df
þ Ab e
ðkKb Þ2 Db
½TSPPb ðK½BLGAÞn ¼c ½TSPPTot 1 þ ðK½BLGAÞn
ð2Þ
where the parameter n represents the Hill coefficient [24, 25], [TPPS]b and [TPPS]Tot are the concentrations of bound and the maximum concentration of bound TPPS4-, respectively, under the conditions of the experiments, K is the equilibrium binding constant and c is a function of the emission quantum yield for free and bound ligands. The non-linear fitting of Eq. 2 yields the values of K and n, c. 2.4.2 Stern–Volmer (S–V) analysis We assumed that quenching of BLGA fluorescence by TPPS4- is due to the binding between the two molecules; thus static quenching is responsible for the decrease of protein fluorescence upon addition of TPPS4-. Stern– Volmer analysis for static quenching predicts that the ratio F0/F between the fluorescence of BLGA in the absence of TPPS (F0) and the fluorescence of BLGA in the presence of TPPS (F), is, for small amounts of quencher, a linear function of the concentration of the ligand [26] F0 ¼ 1 þ K½TSPP F
ð3Þ
where both F0 and F are corrected for the absorption at the excitation wavelength as well as the potential filter effects with the equation [26]
2.4.1 Fluorescence of TPPS
IðkÞ ¼ Af e
where I(k) is the overall fluorescence spectrum, Af and Ab are proportional to the concentration and the emission quantum yield of the free and bound porphyrin, respectively, and Kf and Kb are the peak position of the free and bound Gaussian, respectively [17], calculated for the Q0 (0,0) peak [20, 21] (Fig. 1). The last summation of Eq. 1 fits the region of the Q0 (0,1) peak [20, 21] (Fig. 1), which is less sensitive to the binding of porphyrins to proteins and is produced by an unknown number of bands. This assumption enabled us to fit the emission spectra with a minimum number of Gaussian components [17]. The quality of the fitting was determined by visual inspection and by using the value of the reduced v2. The amplitude of each Gaussian yields the products Uf[porphyrin]f and Ub[porphyrin]b [17]. In order to determine the binding parameters we assumed the modified Hill model [13] which after the introduction of the Gaussian fitting becomes [17]
þ
X j
Aj e
ðkKj Þ2 Dj
ð1Þ
F ¼ 10
Aex þAem 2
Fraw
ð4Þ
In this case Fraw is the uncorrected emission intensity and Aex and Aem are the optical densities of the sample at the excitation and the maximum of the emission wavelength, respectively. The slope obtained from the linear regression of F0/F vs. [TPPS] provides the static quenching constant, K, which in
123
4
(a)
(b) Q'(0,0)
1
Normalized Fluorescence Intensity
Normalized Fluorescence Intensity
1
0.8
0.6
Q'(0,1)
0.4
0.2
0 600
650
700
0.8
0.6
0.4
0.2
0
750
600
Wavelength (nm)
700
750
(d) 1
Normalized Fluorescence Intensity
1
0.8
0.6
0.4
0.2
0 600
0.8
0.6
0.4
0.2
0 650
700
750
the case of ground state quenching reduces to the equilibrium binding constant. However, since the BLGA dimer is a multi-Trp complex (two residues per BLGA monomer), it is necessary to establish the fraction of fluorescence quenching induced by binding of TPPS4- in order to correct the values of F0 and F in Eq. 3. Static quenching theory assumes that the formation of a chromophorequencher ground state complex eliminates the fluorescence contribution of the complex from the fluorescence signal [26]. If the fluorescent molecule is a multi-Trp protein, binding of a ligand may quench some but not all of the Trp residues in the protein. Thus a residual fluorescence is left from the protein-quencher complex that, if unaccounted, would lead to an underestimate of the binding constant [27]. The fraction of BLGA fluorescence quenched by binding of TPPS4- was retrieved by applying and extending the classic method introduced by [28] since, formally, the linearity of Eq. 3 is true for both static and dynamic quenching. In this case we have considered m independent Trp residues each contributing F0i, to the overall emission intensity, proportional to the quantum yield U0i. Therefore, the fluorescence quantum yield in the absence of quencher (TPPS4-) can be written as
600
650
700
750
Wavelength (nm)
Wavelength (nm)
123
650
Wavelength (nm)
(c) Normalized Fluorescence Intensity
Fig. 1 Fluorescence of TPPS4and Gaussian fitting of the spectra upon addition of BLGA. Solid black corrected emission spectrum; dashed red fitting from Eq. 1; dotted black Gaussian of free TPPS4-; solid red Gaussian of bound TPPS4-; dotted green, dotted orange, dotted blue Gaussians of the last term of Eq. 1. a pH 6, [BLGA] = 0.13 lM; b pH 6, [BLGA] = 2.30 lM; c pH 9, [BLGA] = 0.13 lM; a pH 9, [BLGA] = 2.30 lM. The plot shows that at increased concentration of BLGA the Gaussian associated with bound TPPS4- increases relative to the one associated with the free porphyrin. The increase is similar at pH 6 and 9 thus, it is not affected by the conformational transition of the protein. In (a) the two emission peaks are labeled according to Refs. [22, 23]
I. Silva et al.
U0 ¼
m 1X U0i m i¼1
ð5Þ
Upon addition of the quencher (Q), the fluorescence detected is [28] m 1X U0i U¼ ð6Þ m i¼1 1 þ Ki ðQÞ where Ki is the static quenching constant for each Trp residue. From Eqs. 5 and 6 m 1X U0i Ki ðQÞ ð7Þ U0 U ¼ DU ¼ m i¼1 1 þ Ki ðQÞ and m X U0 1 þ Ki ðQÞ ¼ U0i Ki ðQÞ DU i¼1
ð8Þ
So if we assume that the BLGA–TPPS4- complex quenches only a fraction of accessible Trp residues (i.e., Trp residues in close proximity to the porphyrin), then for those Trp residues that are not quenched Ki = 0 and Eq. 8 can be simplified to [28]
An Anionic Porphyrin Binds b-Lactoglobulin A
U0 1 1 ¼ þ DU aKQ ðQÞ a
5
ð9Þ
where a is the fraction of Trp residues proximal (quenched) to the bound TPPS4- and KQ is the quenching constant assumed to be identical for all Trp residues. In multi-Trp proteins this assumption may be often considered an oversimplification; however, in BLG this condition may be realistic. Of the two Trp residues in each monomer, it has been shown that Trp61 does not contribute to the fluorescence of the protein, likely because of its proximity to the Cys66–Cys160 disulfide bond [19, 29]. Therefore, in each monomer only Trp19 contributes to the emission which implies that only two Trp resides contribute to the fluorescence in each dimer. Since U and the corrected fluorescence of Eq. 4 are related by an instrumental multiplication constant which cancels out in the left-hand side ratio of Eq. 9, this equation can be employed to obtain the corrected fluorescence of Eq. 3 as F0 ¼ 1 þ K½TSPP aF
ð10Þ
where a operates as a correction factor due to the fact that the BLGA/TPPS4- complex may retain residual fluorescence. As a result, without the correction by a the fluorescence F would be overestimated and the value of K would be underestimated. 2.4.3 Fluorescence Resonance Energy Transfer The occurrence of fluorescence resonance energy transfer between Trp and TPPS4- was probed qualitatively, using fluorescence excitation and emission spectroscopy [3], as well as quantitatively, using fluorescence lifetime to establish FRET efficiency [30]. Because of the artifact introduced by static quenching in the emission and excitation spectra [3], fluorescence lifetime is more reliable in establishing FRET efficiency. 2.4.4 Fluorescence Anisotropy Steady-state fluorescence anisotropy was recorded by selecting the excitation light and the emitted fluorescence with manually rotating polarizers and recording the emission spectra of TPPS and the BLGA/TPPS complex at various temperatures. The anisotropy was calculated according to [31] r¼
IVV GIVH IVH 2GIVH
ð11Þ
where I is the intensity of the emission (integral of the spectrum) and the subscripts indicate, in spatial order, the orientation of the excitation and emission polarizers
(V, vertical, H, horizontal). G is the factor that accounts for the differential sensitivity of the instrument to light polarized parallel and perpendicular, and is calculated as G¼
IHV IHH
ð12Þ
Each value of intensity was calculated as the average of three spectra. By varying the temperature of the sample (using a circulating water bath connected to the waterjacketed sample holder of the fluorimeter) we retrieved data for the Perrin equation 1 1 sRT ¼ þ r r0 gVr0
ð13Þ
which yields the value of the limiting anisotropy (r0) and the hydrodynamic volume of the rotating molecule (V), when plotting 1/r vs. T/g ratio between the temperature of the solution and the viscosity of the medium [26]. R is the gas constant. The Perrin equation was used to retrieve V and establish, whether BLGA formed aggregates (larger than the dimer) either alone or induced by the binding of TPPS4-. 2.4.5 Effects of Urea The effects of TPPS on the stability of BLGA was investigated using the dissociation coupled unfolding model (DCU) proposed for BLG [32]. According to the model the free energy of denaturation is calculated as DGDCU = - RT ln KDCU, where KDCU is given by KDCU ¼ 4½BLGA
FD2 1 FD
ð14Þ
and FD is the fraction of denaturated protein calculated as FDEN ¼
Y Ymin Ymax Ymin
ð15Þ
where Y is an observed property of the protein [32] which in this case is the position of the emission maximum, as a function of urea. The standard free energy of denaturation extrapolated at zero urea concentration is derived from the equation DGDCU ¼ DG0DCU m½urea]
ð16Þ
where m is the denaturant index [33]. 2.4.6 Docking Simulations Docking simulations were carried out using the software package Arguslab 4.0 (Planaria Software Inc., Seattle, WA). The TPPS4- structure, generated and optimized with Chimera (University of California San Francisco) [34], was docked to one of the available structures of BLG dimers
123
6
I. Silva et al.
obtained from the Protein Data Bank (1BEB.pdb). The protein was stripped of the water molecules encoded in the structure. The porphyrin ligand was ‘‘manually’’ brought into the vicinity of a possible binding site and the docking computation was carried out using different initial orientations of TPPS4-. Several possible binding sites on each monomer as well as at the monomer/monomer interface were probed.
3 Results 3.1 Absorption Spectroscopy At all pH values the addition of BLGA to a stock solution of TPPS4- produces a small bathochromic shift of the Soret band (from 413 to 416 nm) and simultaneous hypochromicity (Fig. 2). The same effect occurs at every pH probed in this study. In agreement with what was suggested by previous results [13] there is no evidence for the formation of J-aggregates of TPPS4- (appearance of a band near 490 nm) either before or after the addition of the protein. Moreover, there is no clear shift of the Q-bands induced by the presence of BLGA [16]. 3.2 Fluorescence of TPPS4- bound to BLGA At all pH values the addition of BLGA prompts a shift of the emission maximum to longer wavelengths (Fig. 3a, b). Such shift is larger ([4 nm) and occurs much earlier than
0.2
Optical Density
0.15
0.1
the one observed in absorption (Fig. 2). The red-shift is more pronounced at higher pH values (Fig. 3c). However, Fig. 3a, b also show that the normalized and corrected intensity of the TPPS4- emission, changes only slightly in the presence of BLGA. Such small change would not allow the application of fluorescence intensity methods to establish the amount of bound porphyrin [35, 36]. Therefore, the contribution of free and bound porphyrin to the emission spectrum must be retrieved via Gaussian analysis of the emission spectra with Eq. 1 (Fig. 1). The plots of [TPPS4-]b/[TPPS4-]Tot vs. [BLGA] obtained from the Gaussian fitting show the relative increase of the fraction of bound porphyrin (Fig. 4). The range of protein and TPPS concentration has to be kept relatively low because of the restriction on the optical density (O.D.) of the samples imposed by the fluorescence method; in particular, signal linearity requirements impose that the O.D. at the excitation wavelength be \0.2. Nevertheless, fitting of the isotherms in Fig. 4 with Eq. 2, yields an estimate of the binding constant, K, the coefficient n, and the factor c, which are summarized in Table 1. The data show that the binding constant increases by approximately 20% with increasing pH. The value of n remains \1 at all pH values, indicating a negative cooperativity, and decreases by about 16% as the pH increases. A student’s t-test analysis reveals that differences in the value of K between pH are not statistically significant (p-value C 0.2). Similarly, an analysis of the value of n reveals that, although its decrease with increasing pH is consistent with electrostatic surface binding of TPPS4 [37], the differences in its value at various pH are not statistically significant (p-value C 0.1). Thus we are led to conclude that that approximately one TPPS molecule binds to each BLGA unit and that the binding constant is approximately constant across the range of pH. A large increase with the pH occurs in the value of c (Table 1). This factor depends on the emission quantum yields for bound and free TPPS4-. Individual quantum yields for free and bound porphyrins at various pH values could not be experimentally calculated and are not available from the literature; therefore, our fitting only indicates that the large change in the binding isotherm of Fig. 4 is due to changes of the emission quantum yields of the free and bound porphyrin at the different pH values.
0.05
3.3 Quenching of BLGA by TPPS 0 350
400
450
500
Wavelength (nm)
Fig. 2 Absorption spectra of TPPS4- upon addition of BLGA. Black [BLGA] = 0 lM; Red [BLGA] = 0.73 lM; Blue [BLGA] = 2.48 lM; Green [BLGA] = 4.02 lM; Yellow [BLGA] = 5.5 lM
123
Each BLGA monomer has two Trp residues, thus each BLGA dimer has potentially four Trp residues that can be quenched. The contribution of each residue to BLGA fluorescence is still under debate but some convincing studies suggested that only Trp19 at the bottom of the interior barrel contributes to the fluorescence of the protein
An Anionic Porphyrin Binds b-Lactoglobulin A
(a)
(b) 1
1.2 2
2
1
1
Fluorescence Intensity (a.u.)
3
Fluorescence Intensity (a.u.)
Fig. 3 Fluorescence spectra of TPPS4- upon addition of BLGA. 1 [BLGA] = 0 lM; 2 [BLGA] = 1.33 lM; 3 [BLGA] = 2.07 lM. a pH 6; b pH 9. The spectra show the red-shift of the emission upon binding to the protein. The spectra also show little change of the emission intensity. c Position of the emission maximum of TPPS4- as a function of the concentration of BLGA. d pH 6, j pH 7, r pH 8, m pH 9. The red-shift occurs nearly at the same rate at all pH values
7
0.8
0.6
0.4
0.2
0 600
650
0.8
0.6
0.4
0.2
750
700
Wavelength (nm)
(c)
3
1
0 600
650
700
750
Wavelength (nm)
648 647
645
em
644
λ
max
646
643 642 641 640 0
5 10
-7
1 10
-6
1.5 10
-6
2 10
-6
-6
2.5 10
[BLGA] M
0.2
Table 1 Binding parameters obtained from the fluorescence of TPPS4- using Gaussian fitting and Eq. 2 c
4-
[TPPS ]b/[TPPS ]tot
0.15
n
pH 6
0.67 ± 0.12
0.63 ± 0.14
0.85 ± 0.04
pH 7
0.72 ± 0.14
0.75 ± 0.17
0.86 ± 0.06
pH 8
0.85 ± 0.13
0.76 ± 0.16
0.69 ± 0.10
pH 9
0.95 ± 0.17
0.78 ± 0.18
0.71 ± 0.08
4-
0.1
K 9 10-5 (M-1)
0.05
0 0
5 10
-7
1 10
-6
1.5 10
-6
2 10
-6
[BLGA] µM
Fig. 4 Plot and fitting of bound/total TPPS4- according to Eq. 3. d pH 6, j pH 7, r pH 8, m pH 9
this ratio is basically independent of the pH of the solution Table 2. Fluorescence spectra of BLGA as a function of added porphyrins (supplemental material) show that the decrease of protein fluorescence is not accompanied by a shift in its emission maximum. The values of the binding constants retrieved from S–V plots (Fig. 5b; Table 2) are numerically in good agreement with the ones retrieved from TPPS4- fluorescence (Table 1) and they do not appear to be pH-dependent. 3.4 Fluorescence Resonance Energy Transfer
[19, 29]. We, therefore, used Eq. 9 to analyze whether the BLGA/TPPS4- complex retains some residual fluorescence. The results show (Fig. 5a) that the porphyrin quenches 80–100% of the monomer fluorescence and that
Steady state and fluorescence lifetime experiments rule out any significant FRET between the Trp residue and the bound TPPS4-. Excitation of Trp residues does not produce
123
8
I. Silva et al.
emission by the ligands and, likewise, excitation spectra recorded at the maximum of the emission wavelength of the porphyrin do not show any contribution from Trp. Fluorescence decay experiments also reveal that there is no change in BLGA emission lifetime upon addition of TPPS4-. Such result is also in strong support of the occurrence of static quenching due to binding of the porphyrin to the protein. Lack of FRET is consistent with what had been observed in other systems [3, 17] and is a direct consequence of the very small overlap between the emission spectrum of BLGA and the absorption spectrum of TPPS4-.
Table 2 BLGA fluorescence quenching. Fraction of quenched fluorescence and binding constant obtained from Eqs. 9 and 10
pH 6
At all pH values, CD experiments in the region of the Soret band did not show any optical activity of TPPS4- which indicates that binding to BLGA does not distort the structure of the porphyrin ring. At the same time the binding of the porphyrin did not produce any change in the dichroic signal of BLGA either in the region of the aromatic amino acids (260–320 nm) or in the region of the amide (180–250 nm). 3.6 Effects of Urea In order to probe whether the porphyrin binds to sites at the monomer/monomer interface we studied the effects of the addition of urea to solutions containing the BLGA/TPPS4non-covalent complex. It is known that, within a certain range of urea concentrations, which depends on the pH of the solution [32], BLGA dimers dissociate without any appreciable change in the tertiary structure of the individual monomers [32]. Thus if TPPS4- binds to the dimer at the interface between the two monomers, the addition of urea may cause the porphyrin to return in solution upon dissociation of the dimers. This in turn would cause the emission
1.56 ± 0.33
pH 7
1.10 ± 0.05
1.10 ± 0.25
0.91 ± 0.03
1.37 ± 0.29
pH 9
0.96 ± 0.08
1.14 ± 0.18
(b) 10
1.8
8
1.6
F0/F
F0/∆F
0.89 ± 0.09
pH 8
(a)
6
1.4
4
1.2
2
1 5
2 10
5
4 10
5
5
6 10
1/[TPPS] M
123
K 9 10-5 (M-1)
maximum of the porphyrin to blue-shift towards the position of free TPPS4-. Our results, however, show (Fig. 6) that addition of urea up to 8 M (a concentration where urea is known to partially unfold the monomeric form of BLGA [32, 38]) does not cause any shift of fluorescence of BLGAbound TPPS4-. Addition of the same aliquots of urea to solutions containing only TPPS produced only a slight (2 nm) red-shift of the emission of the porphyrin, probably due to the sequestration of water molecules by urea. Since the red-shift resulting from binding to BLG is 7 nm, the lack of blue-shift of BLG-bound TPPS upon addition of urea indicates that TPPS4- remains attached to the protein even when this is partially unfolded. The dissociation and unfolding of BLGA at increasing urea concentration was probed by the red-shift and the increase in intensity of the intrinsic fluorescence of the protein [19] (supplemental material). The comparison of urea-induced denaturation of BLGA alone and BLGA in the protein/porphyrin complex shows that the presence of bound TPPS4- stabilizes BLGA, as the sigmoidal transition to the unfolded state appears to be delayed (Fig. 7a, b). Calculation of DG0DCU and m with Eq. 16 produces the values shown in Table 3. By comparison with the data obtained in the absence of TPPS4- it can be seen that while the denaturant index, m, remains virtually unchanged, DG0DCU of the BLGA/porphyrin complex increases, at all pH values. The increased stability
3.5 Circular Dichroism
Fig. 5 a Plot F0/DF according to Eq. 10 (model of fractional quenching of [32]). b Stern–Volmer plot of BLGA fluorescence corrected according to Eq. 11. d pH 6, j pH 7, r pH 8, m pH 9
a
8 10 -1
6
1 10
0
1 10 -6 2 10 -6 3 10 -6 4 10 -6 5 10 -6 6 10 -6 7 10 -6
[TPPS] M
An Anionic Porphyrin Binds b-Lactoglobulin A
9
against denaturation induced by binding of TPPS4- is larger at smaller pH (Fig. 7c).
1.5
[urea] = 0 M [urea] = 2 M
3.7 Fluorescence Anisotropy
Fluorescence Intensity
[urea] = 4 M
Perrin plots from Eq. 13 are linear and show, as expected, a steeper slope for free TPPS than for the BLGA/TPPS complex at all pH (Fig. 8a, b). From the slope of the plot and upon measurements of free and bound TPPS fluorescence lifetime (9.8 and 11 ns, respectively) we estimated the radius of the free ligand to be \1 nm and the radius of the complex to be 2.5–3 nm. The radius of the BLGA/ TPPS complex is in excellent agreement with the size of the BLGA dimer reported by others [39, 40]. This in turn confirms that in the pH and concentration range investigated, binding of TPPS4- does not induce aggregation of BLGA.
[urea] = 6 M
1
[urea] = 8 M
0.5
0 600
650
700
750
Wavelength (nm)
3.8 Docking Simulations
Fig. 6 Fluorescence of TPPS4- in the protein/porphyrin complex as a function of urea concentration. The spectra are normalized and offset with respect to each other to show that the emission maximum of bound TPPS4- is not affected by the addition of up to 8 M urea
(a)
(b) 1
0.8
0.8
0.6
0.6
FDEN
FDEN
1
0.4
0.4
0.2
0.2
0
0 -2
0
2
4
6
8
10
0
2
4
6
8
10
1.5
DCU
1.4
1.3
1.2
1.1
0
DCU
(BLGA/TPPS4-)/∆G
0
(BLGA)
(c)
-2
[urea] M
[urea] M
∆G
Fig. 7 Denaturation curves. FDEN calculated according to Eq. 18 using the wavelength of the emission maximum as the parameter Y. a pH 7: d BLGA/ TPPS4- complex, s BLGA alone; b pH 9: d BLGA/ TPPS4- complex, s BLGA alone. The data are representative of the trend at the other pH values. c Relative increase of DGDCU induced by the presence of bound TPPS4as a function of pH
A survey of the BLGA dimer structure (PDB file 1BEB.pdb) provides several potential binding sites for TPPS4-. There are several Lys and Arg residues on the surface as well as in
1 5.5
6
6.5
7
7.5
8
8.5
9
9.5
pH
123
10
m BLGA alone
DG0DCU (kJ/mol) BLGA/TPPS4- complex
DG0DCU (kJ/mol) BLGA alone
pH 6
6.38 ± 0.52
4.44 ± 0.91
74.4 ± 4.1
51.1 ± 5.1
pH 7
4.46 ± 0.78
2.86 ± 0.40
53.3 ± 4.5
38.0 ± 4.3
pH 8
3.38 ± 0.42
2.79 ± 0.27
45.3 ± 4.3
35.8 ± 4.5
pH 9
2.53 ± 0.61
2.53 ± 0.22
40.1 ± 4.0
34.8 ± 5.0
(a)
(b)
4.4
4.8 4.6
4.2 4.4 4 4.2
1/r
Fig. 8 Perrin plots of the steady state fluorescence anisotropy (Eq. 16). a pH 6: d BLGA/ TPPS4- complex, j TPPS4alone; b pH 9: d BLGA/ TPPS4- complex, j TPPS4alone. The data are representative of the trend at the other pH values and show a larger slope for TPPS alone than for the porphyrin bound to the proteins
m BLGA/TPPS4complex
1/r
Table 3 Thermodynamic parameters for the urea-induced unfolding of BLGA in the complex with TPPS4- and alone at different pH values
I. Silva et al.
4
3.8
pH 9
3.8 3.6 3.6 3.4 3.4 3.2 4 4 4 4 4 4 4 4 4 2.5 10 3 10 3.5 10 4 10 4.5 10 5 10 5.5 10 6 10 6.5 10
T/η (K/Pa)
the grove formed between the 3-turn helix and the outside of the b-barrel [14, 22, 41]. In addition we investigated the docking at intradimeric sites on the side of the 3-turn helices as well as on the side of the aperture of the barrel [42]. We attempted the docking to all these locations by approaching TPPS4- at different orientations to each site. Our results show that there is no stable binding configuration for either the interior of the b-barrel or the outside grove. In addition, docking at the monomer-monomer interface did not yield stable binding configurations. Our docking computation revealed that the most likely binding site for TPPS4- is on the surface of BLGA at a
Fig. 9 Simulation of the docking of TPPS4- to BLGA. a The most stable docking configuration produces TPPS4(red) laying flat on the surface of BLGA with its four SO3groups interacting with Lys (yellow) and Arg (blue) positively charged residues. b Different view of the same binding site which shows that ´˚ away TPPS4- (red) is \8 A from the indole ring of Trp19 (blue)
123
3.2 4 4 4 4 4 4 4 4 4 2.5 10 3 10 3.5 10 4 10 4.5 10 5 10 5.5 10 6 10 6.5 10
T/η (K/Pa)
location which involves a near interdigitation of three of the four SO3- groups of TPPS4- and the positively charged side chains of Lys14, Lys47, Lys100 and Arg124 (Fig. 9a). ´˚ This location also places the ligand in proximity (*7.2 A ) of the fluorescent Trp19 residue (Fig. 9b) which would explain the static quenching produced by the binding of TPPS4- to BLGA. Thus, the computed binding site involves regions that include the N-terminus, the B-strand, the FG-loop, and the loop between strand I and the 3-turn a-helix [14, 41]. Because of the large number of surface Arg and Lys it is likely that other surface regions may also offer docking sites to the porphyrin.
An Anionic Porphyrin Binds b-Lactoglobulin A
4 Discussion Despite its solubility in aqueous solutions, it had been previously established that TPPS is able to bind small globular proteins such as BLG and tubulin [13, 17]. This investigation introduces, however, the analysis of the effects of the pH-dependent conformational changes of BLG [14, 15] on the binding of TPPS4- and establishes a possible location for the docking of the porphyrin. Assuming that the contribution of TPPS2- is negligible in the pH range investigated (the pKa for TPPS is \4.5 [16, 43]), we considered TPPS4- as the only significant ligand. First of all our results show that for 6 B pH B 9 there is no formation of the porphyrin J-aggregate either before binding or induced by BLGA. The binding parameters obtained from the fluorescence of the porphyrin and the quenching of the protein (Tables 1, 2) show that the numerical values of K retrieved from S–V plots (Fig. 5b) is higher than the one obtained from the curves of Fig. 4. If the data obtained with the two methods are compared with a student’s t-test, the analysis reveals that 0.05 B p B 0.09. Although the difference can be considered somewhat statistically significant, one has to consider the very small sampling associated with our measurements which makes the differences between the values of K retrieved with the two methods, difficult to interpret. Therefore, we can assume that the data obtained from TPPS fluorescence and BLGA quenching (with Eqs. 2, 10, respectively) yield the same qualitative results that the affinity of the porphyrin for BLGA does not change with the pH of the solution. Since across the pH range probed with this study, BLGA undergoes a major conformational change that involves the access to the interior protein pocket (b-barrel) [14], our results rule out that the pocket, or its proximity, represents a binding site for the porphyrin. This is somewhat expected since the insertion of one or more SO3- substituents into the hydrophobic pocket would carry a very unfavorable entropic contribution [14, 44]. The unsuitability of this site as a binding location for TPPS4- is also confirmed by the absence of stable binding configurations in our docking simulations inside or in proximity of the interior pocket of the protein. Possible alternative binding sites could be present on both sides of the ‘‘seam’’ at the monomer/monomer interface of the BLGA dimer [42]. However, our experimental and computational data rule this site out. Urea experiments show (Fig. 6) that even upon substantial denaturation of the protein, the porphyrin remains attached to BLGA, thus ruling out that the separation of the two monomers detaches the porphyrin from the protein. In agreement with this, computational docking did not yield any stable binding conformation at the monomer/monomer interface.
11
All our experimental results instead seem to point in the direction of a superficial site. The fact that the pH does not affect binding constant of TPPS4- is in agreement with such model. Because of the small size of BLGA, if the binding site is not inside or in proximity of the aperture of the interior barrel, then the binding possibilities for a relatively large ligand, such as TPPS4- (Fig. 9), is limited. Since our computational docking simulations rule out binding to the grove formed between the outside of the barrel and the 3-turn a-helix, the remaining binding sites are on the surface of the protein. A superficial binding site is also consistent with the lack of dichroic signal from the porphyrin. The lack of CD signal signifies that the ring of the porphyrin remains substantially undistorted as it would if the whole porphyrin laid on the surface of the protein. The urea data (Fig. 6) also support a superficial binding site since even extensive (albeit not total [32, 45]) unfolding of BLGA does not prompt the release of TPPS4- in solution. Such an effect would be unlikely if the porphyrin was bound to a ‘‘fold’’ or a pocket of the protein through nonCoulombic interactions. How would TPPS4- bind to the surface of BLGA? In fact, why should a water soluble porphyrin such as TPPS4bind to a protein in the first place? We have shown that the same porphyrin binds another, larger, globular protein (tubulin) also at the surface [17] and that such interaction is stabilized by the Coulombic interaction between one of the SO3- groups and positively charged amino acids (His and Arg in that case [17]). Thus we suggest that the interaction between TPPS4- and BLGA occurs through Coulombic interactions between the positively charged amino acids on the surface of the protein and the negatively charged groups on the porphyrins. Such interaction would substitute the weaker ion-dipole interaction between the SO3- groups and water molecule (solvation), thus explaining why a water-soluble ligand would bind a protein in aqueous solution. Even a substantial denaturation of the protein would still keep the porphyrin ‘‘sticking’’ to the surface of the protein since urea does not affect the protonation of amino acid residues. This scenario would also explain why the fitting with Eq. 2 yields values of n \ 1. BLGA is not a specific ‘‘receptor’’ for TPPS thus, one would expect a set of nindependent, non-interacting binding sites for the porphyrin (i.e., there is no reason, and no CD evidence that BLGA changes its conformation upon binding of one TPPS molecule to accommodate or inhibit the binding of additional porphyrin molecules). One would expect a value of n = 1 according to the Hill theory of cooperative binding [46]. A value of n \ 1, however, indicates some degree of cooperativity but because there is no evidence to support the presence of interacting sites for TPPS4-, a likely explanation for the values of n is that the binding of one porphyrin
123
12
molecule to the surface of the protein leads to a decrease of the available surface charges and effectively reduces the probability of binding of additional TPPS4- molecules. Finally, the superficial binding is confirmed by our docking simulations (Fig. 9a) which place the negative groups of the porphyrin in close proximity (in fact almost perfectly interdigitated) with positively charged amino acids such as Lys and Arg residues. Although, as explained above at the pH values of our experiments both molecules carry an excess of negative charge, it is very likely that many of the basic amino acids, Lys and Arg in particular, still carry a net positive charge [23] making the interaction with TPPS4- possible. According to our anisotropy data (Fig. 8) the electrostatic ‘‘shielding’’ provided by the binding of TPPS4- is not sufficient to cause aggregation of BLGA as seen instead for other proteins [47]. Finally a superficial binding site such as the one suggested in this work would explain also the stabilization effect that TPPS4- produces on BLGA. The delay in the urea-dependent unfolding and the higher free energy of unfolding (Fig. 7) can be explained by the fact that TPPS4could act like ‘‘glue’’, linking different domains of the protein [e.g., the N-terminus, the B-strand, the FG-loop, and the loop between strand I and the 3-turn a-helix (Fig. 9)]. The Coulombic interactions between the negative porphyrin and the positive amino acid residues would stabilize this region and as a consequence the whole protein against urea denaturation. At lower pH values the anionic porphyrin would be more effective in stabilizing the protein as the surface of the polypeptide contains a larger number of positively charged residues. As shown by the data in Table 3 the increase in DG0 between the BLGA/ TPPS complex and BLGA alone is larger at smaller pH. The difference in the values of DG0 between complex and protein alone is statistically significant (p-value B 0.01) at pH 6 and 7 while it becomes smaller and statistically insignificant at pH 8 and 9 (p-value C 0.05). Thus is further indication of the stabilization effect that tetra-anionic TPPS produces on BLGA at low pH. Acknowledgments The research was supported by the 2006 Faculty Research Award of the University of Texas at San Antonio (to L.B.) and by the AFRL/HE grant # FA8650-07-1-6850 (to L.B.). The author would also like to thank Dr. Markandeswar Panda for the use of the CD spectrometer.
References 1. Kuwata K, Hoshino M, Forge V, Era S, Batt CA, Goto Y (1999) Prot Sci 8:2541–2545 2. Oliveira KMG, Valente-Mesquita VL, Botelho MM, Sawyer L, Ferreira ST, Polikarpov I (2001) Eur J Biochem 268:477–484
123
I. Silva et al. 3. Tian F, Johnson K, Lesar AE, Moseley H, Ferguson J, Samuel IDW, Mazzini A, Brancaleon L (2005) Biochim Biophys Acta 1760:38–46 4. Fernandez NF, Sansone S, Mazzini A, Brancaleon L (2008) J Phys Chem B 112:7592–7600 5. Kubelka J, Hofrichter J, Eaton WA (2004) Curr Opin Struct Biol 14:76–88 6. Lee CT Jr, Smith KA, Hatton TA (2005) Biochemistry 44: 524–536 7. Dougherty TJ, Gomer CJ, Henderson BW, Jori G, Kessel D, Korbelik M, Moan J, Peng Q (1998) J Nat Canc Inst 90:889–905 8. Pope AJ, Bown SG (1991) Br J Urol 68:1–9 9. Castelli M, Reiners JJ, Kessel D (2004) Cell Death Differ 11:906–914 10. Tsaytler PA, O’Flaherty MC, Sakharov DV, Krijgsveld J, Egmond MR (2008) J Proteome Res 7:3868–3878 11. Usuda J, Chiu SM, Murphy ES, Lam M, Nieminen AL, Oleinick NL (2003) J Biol Chem 278:2021–2029 12. Kessel D (2002) Photochem Photobiol Sci 1:837–840 13. Andrade SM, Costa SMB (2002) Biophys J 82:1607–1619 14. Qin BY, Bewley MC, Creamer LK, Baker HM, Baker EN, Jameson GB (1998) Biochemistry 37:14014–14023 15. Tanford C, Bunville LG, Nozaki Y (1959) J Am Chem Soc 81:4032–4036 16. Maiti NC, Ravikanth M, Mazumdar S, Periasamy N (1995) J Phys Chem 99:17192–17197 17. Tian F, Johnson EM, Zamarripa M, Sansone S, Brancaleon L (2007) Biomacromol 8:3767–3778 18. D’Alfonso L, Collini M, Baldini G (2002) Biochemistry 41: 326–333 19. Harvey BJ, Bell E, Brancaleon L (2007) J Phys Chem B 111:2610–2620 20. Bilsel O, Buchler JW, Hammerschmitt P, Rodriguez J, Holten D (1991) Chem Phys Lett 182:415–421 21. Lang K, Mosinger J, Wagnerova DM (2004) Coord Chem Rev 248:321–350 22. Folgolari F, Ragona L, Zetta L, Romagnoli S, De Kruif KG, Molinari H (1998) FEBS Lett 436:149–154 23. Fukamizo T, Juffer AH, Voge HJ, Honda Y, Tremblay H, Boucher I, Neugebauer WA, Brzezinski R (2000) J Biol Chem 275:25633–25640 24. Coval ML (1970) J Biol Chem 245:6335–6336 25. Weiss JN (1997) FASEB J 11:835–841 26. Lakowicz JR (2006) Principles of Fluorescence spectroscopy, 3rd edn. Springer, New York 27. Tatikolov AS, Costa SM (2004) Photochem Photobiol 80:250–256 28. Lehrer SS (1971) Biochemistry 10:3254–3263 29. Cho Y, Batt CA, Sawyer L (1994) J Biol Chem 269:11102–11107 30. Wu P, Brand L (1994) Anal Biochem 218:1–13 31. Chuang TJ, Eisenthal KB (1972) J Chem Phys 57:5094–5097 32. Galani D, Apenten RKO (1999) Food Res Internat 32:93–100 33. Eftink MR (1994) Biophys J 66:482–501 34. Pettersen EF, Goddard TD, Huang CC, Couch GS, Greenblatt DM, Meng EC, Ferrin TE (2004) J Comput Chem 25:1605– 1612 35. Collini M, D’Alfonso L, Baldini G (2000) Prot Sci 9:1968–1974 36. Horowitz P, Prasad V, Luduena RF (1984) J Biol Chem 259:14647–14650 37. Tulp A, Verwoerd D, Hard AA (1997) Electrophoresis 18:767–773 38. Hamada D, Dobson CM (2002) Protein Sci 11:2417–2426 39. Capron I, Nicolai T, Durand D (1999) Food Hydrocoll 13:1–5 40. Gottschalk M, Nilsson H, Roos H, Halle B (2003) Protein Sci 12:2404–2411 41. Brownlow S, Morais Cabral JH, Cooper R, Flower DR, Yewdall SJ, Polikarpov I, North ACT, Sawyer L (1997) Structure 5: 481–495
An Anionic Porphyrin Binds b-Lactoglobulin A 42. Fessas D, Iametti S, Schiraldi A, Bonomi F (2001) Eur J Biochem 268:5439–5448 43. Kelbauskas L, Bagdonas S, Dietel W, Rotomskis R (2003) J Luminesc 101:253–262 44. Jameson GB, Adams JJ, Creamer LK (2002) Int Dairy J 12: 319–329
13 45. Busti P, Scarpeci S, Gatti C, DeLorenzi N (2002) Food Res Internat 35:871–877 46. Dahlquist FW (1978) Methods Enzymol 48:270–299 47. Tominaga TT, Yushimanov VE, Borissevitch IE, Imasato H, Tabak M (1997) J Inorg Biochem 65:235–244
123
