Fluorescence Properties of Glutamine-Binding Protein from Escherichia coli and Its Complex with Glutamine Irina M. Kuznetsova,† Olga V. Stepanenko,† Konstantin K. Turoverov,*,† Maria Staiano,‡ Viviana Scognamiglio,‡ Mose’ Rossi,‡ and Sabato D’Auria*,‡ Institute of Cytology, Russian Academy of Science, 194064 St. Petersburg, Russia, Institute of Protein Biochemistry, C.N.R., 111 80131 Naples, Italy Received October 24, 2004
In this work, the fluorescence of glutamine-binding protein (GlnBP) and its complex with glutamine (GlnBP/Gln) in native and unfolded forms was studied. The experimental data were interpreted on the basis of the results of the analysis of Trp and Tyr microenvironments taking into the account the data for GlnBP mutated forms Trp32Phe(Tyr) and Trp220Phe(Tyr), which have been obtained by Axelsen et al. (Biophys. J. 1991, 60, 650-659). This allowed us to explain the negligible contribution of Tyr residues to the bulk fluorescence of the native protein, the similarity of the fluorescence characteristics of GlnBP and GlnBP/Gln, and the uncommon effect of the excess of the fluorescence intensity at 365 nm (Trp emission) upon excitation at 297 nm respect to the excitation at 280 nm. The last effect is explained by the spectral dependence of the Trp 32 and Trp 220 contributions to the protein absorption. The protein Trp fluorescence dependence on the excitation wavelength must be taken into account for the evaluation of the Tyr residues contribution to the bulk fluorescence of protein, and in principle, it also may be used for the development of an approach for the decomposition of a multicomponent protein fluorescence spectrum. Keywords: Glutamine-binding protein • intrinsic fluorescence • energy transfer • absorption spectrum • tryptophan • tyrosine residues microenvironment
Escherichia coli glutamine-binding protein (GlnBP) is a small (25 kDa), soluble protein belonging to the large family of specific binding proteins that are essential primary receptors in transport various biomolecules, such as sugars, amino acids, peptides, and inorganic ions.1-3 Although these proteins range in size and primary sequence, they have remarkably similar topology.4,5 Their 3D structures are characterized by two globular domains connected by a hinge region. Two domains of GlnBP are referred to as small and large. The small domain consists of three R-helices and four parallel and one antiparallel β-strands connected by a large loop, while the large domain contains two additional R-helices and three β-strands. The domains are linked by two antiparallel β-strands. The deep cleft formed between the domains contains the ligand-binding site (Figure 1). GlnBP is responsible for the first step in the active transport of L-glutamine across the cytoplasmic membrane.5 X-ray crystallographic data5,6 show that the binding of substrate induces large conformational changes in the tertiary structure of the protein (Figure 1). This property makes GlnBP a good candidate for biological recognition element in the develop* To whom correspondence should be addressed: Dr. Konstantin K. Turoverov, Institute of Cytology Russian Academy of Science, Tikhoretsky av., 4, 194064 St.Petersburg, Russia. Tel. 7(812) 247-1957. Fax: 7(812) 2470341. E-mail:
[email protected] and/or Dr. Sabato D’Auria, Institute of Protein Biochemistry, Via P. Castellino, 111 80131 Naples Italy. Phone: +39-0816132250. Fax: +39-0816132277. E-mail:
[email protected]. † Institute of Cytology, Russian Academy of Science, St. Petersburg, Russia. ‡ Institute of Protein Biochemistry, C. N. R., Naples, Italy. 10.1021/pr0498077 CCC: $30.25
2005 American Chemical Society
ment of biosensors aimed at determining Gln concentrations in cell cultures.10,11 In this context, the comparative fluorescence examination of GlnBP in the opened and closed forms is of high interest. Furthermore, this protein is also an interesting model for studying the processes of proteins folding and the role of ligand binding in stabilization of protein structure. Since the majority of protein folding-unfolding studies have been done on small, single domain proteins, such investigations on more complex proteins are of actual.12 In addition, the study of partially folded intermediate states of proteins is also important in view of the assumption of their functional role in living cell.13 In this connection, GlnBP, as well as other two domains ligandbinding proteins, is an interesting object of investigation, as the possible mechanism of protein-ligand complex dissociation consists of a protein transition in a partially folded intermediate state in the vicinity of membrane. A variety of methods are used for investigation of proteins folding and among them intrinsic Trp fluorescence of proteins, which parameters (intensity, spectrum position, anisotropy, excitation lifetime, etc.) are highly sensitive to protein conformational changes.14 This represents an additional reason the study of fluorescence characteristics of GlnBP is actual. Fluorescence characteristics of GlnBP have been already studied by Weiner and Heppel15 and Axelsen et al.16 Nonetheless, for various reasons, we address it once again. There are Journal of Proteome Research 2005, 4, 417-423
417
Published on Web 02/08/2005
research articles
Kuznetsova et al.
the examination of mutated forms of Trp32Phe(Tyr) and Trp220Phe(Tyr), which have been done by Axelsen et al.16
Materials and Methods
Figure 1. Spatial pattern of GlnBP (left) and its complex with Gln(right). A. Cartoon diagram of the protein. Tryptophan residues (red), tyrosine residues (green), Lys 166 (blue) belonging to microenvironment of Trp 220 and Gln are shown as spheres. B. The surface of the protein. Trp32 and Trp220 and Gln are shown in green, red, and blue, respectively. Complex formation results in collapse of protein macromolecule around Gln resulting in cleft close. This figure is constructed based on Protein Data Bank,7 file 1GGG.ent, ref 5 and 1 WDN.ent, ref 6. The drawing was generated by the graphic programs VMD8 and Raster 3D.9
some discrepancies in these works. In particular, in the work by Weiner and Heppel,15 the emission maximum was found at 336 nm (λex ) 280 nm), while in the work by Axelsen et al.,16 it was determined to be at 330 nm (λex ) 295 nm). In addition, there are some self-contradictions in the work of Axelsen et al.16 In particular, it is unclear how Trp residues could have essentially different emission maxima and practically the same accessibility to the solvent. Furthermore, the nature of the unusual absorption spectrum of GlnBP and what determines different absorption spectra of two Trp residues of GlnBP also remain unexplained in the work of Axelsen et al.16 Finally, the Tyr fluorescence of GlnBP was not examined in these two works. The work of Axelsen et al.16 was done when the structure of GlnBP was unknown. In recent years, the structure of GlnBP and its complex with Gln were determined by X-ray analysis.5,6 As a consequence, on the basis of these data, we can analyze the microenvironment of Trp and Tyr residues and peculiarities of their location in GlnBP and GlnBP/Gln. The aim of this work was to reexamine the intrinsic fluorescence of GlnBP and GlnBP/Gln taking into the account the results of our analysis of Trp and Tyr residues microenvironment and the results of 418
Journal of Proteome Research • Vol. 4, No. 2, 2005
Protein Purification. GlnBP from E. coli was prepared and purified according to Dattelbaum and Lakowicz.10 L-Glutamine was obtained from Sigma. All of the other chemicals used were commercial samples of the purest quality. The protein concentration was determined by the method of Bradford17 with bovine serum albumin as standard on a double beam Cary 1E spectrophotometer (Varian, Mulgrade, Victoria, Australia). GdnHCl (Nacalai Tesque, Japan) was used without additional purification. The GdnHCl concentration was determined by refraction index with an Abbe refractometer (LOMO, Russia). Analysis of Protein 3D Structure. The locations of Trp and Tyr residues in GlnBP and in the complex of GlnBP with Gln molecule were analyzed according to the atom coordinates of the crystal structure of GlnBP (files 1GGG.ent; ref 5) and its complex with Gln (1WDN.ent; ref 6). The microenvironment of the Trp or Tyr residue was determined as a set of atoms located at a distance less than r0 from the geometrical center of the indole or phenol ring; r0 was taken as 7 Å.18,19 The nearest atom in the microenvironment to each atom of the indole or phenol ring was specified and the distance between them was determined. For Tyr residues, the neighbors of the OH group were also determined. The packing density of the atoms in a microenvironment was determined as the number of atoms comprising the microenvironment, or as the part of microenvironment volume (V0) occupied by the atoms (d ) ∑Vi/V0). The volume occupied by each atom (Vi) was determined according to its van der Waals radius, and only the part inside the microenvironment was taken into account. The real values of atom volume are slightly smaller, as atoms are incorporated in chemical bonds. Nonetheless, it is found to be not significant for the estimation of microenvironment packing density of Trp and Tyr residues. To estimate the accessibility of Trp residues to the solvent, the radial dependence of atoms packing density about the geometrical center of Trp residue was evaluated as follows
d(r) )
∑ V (r,r + ∆r) i
V0(r,r + ∆r)
(1)
where V0 (r, r + ∆r) is the volume of sphere layer, that is r distant from the geometrical center of the indole ring, ∆r is layer thickness and Vi is the part of the i’s atom volume that is inside this sphere layer. The efficiency of nonradiative energy transferred between any two chromophores was evaluated as follows20 W)
1 2/3 R 1+ 2 k R0
()
6
(2)
Here, R0 is Fo¨rster distance, i.e., the averaged distance between randomly orientated donor and acceptor at which W ) 0.5; R is the distance between the geometrical centers of the indolic (or phenol) rings of a donor and an acceptor; and k2 is the factor of mutual orientation of the donor and the acceptor. k2 ) (cos θ - 3cosθAcosθD)2
(3)
research articles
Intrinsic fluorescence of GlnBP and GlnBP/Gln
where θ is the angle between the directions of the emission oscillator of a donor and the absorption oscillator of an acceptor; θA is the angle between the emission oscillator and the vector connecting the geometrical center of the donor, and θD is the angle between the absorption oscillator and the vector connecting the geometrical center of the acceptor.21 The values of R0 for Trp-Trp, Tyr-Trp, and Tyr-Tyr pairs were taken from Eisenger et al.22 and Steinberg.23 All other parameters were determined according to atoms coordinates.18,19,24 Oscillators were considered to be rigid in all calculations. Fluorescence Measurements. All fluorescence experiments were carried out at 23 °C on a homemade steady-state spectrofluorimeter and pulse spectrofluorimeter for recording fluorescence decay curves.25 Fluorescence spectra were measured with excitation at 297 or 280 nm. The parameter A ) I320/I365 (I320 and I365 are fluorescence intensity at 320 and 365 nm, respectively) was used for characterization of fluorescence spectra position.24 The fluorescence spectra and the values of parameter A were both corrected by the instrument sensitivity. Contribution of Tyr residues to the bulk protein fluorescence was evaluated by the value ∆λ )
( ) ( ) Iλ
I365
-
280
Iλ
I365
Figure 2. Fluorescence spectra of GlnBP and its complex with Gln in native (curves 1, circles and 2, triangles) and unfolded (curves 3, circles and 4, triangles) states. Denaturation was induced by 3.0 M GdnHCl; λex ) 297 nm. All values are reduced to fluorescence intensity of native GlnBP at 365 nm.
(4)
297
Analysis of Fluorescence Decay. The decay curves were analyzed in multiexponential approach: I(t) )
∑R exp(-t/τ ) i
(5)
i
i
where Ri and τi are the amplitude and the lifetime of component i, ∑ Ri ) 1. The values of Ri and τi were determined from convolution of I(t) with lamp impulse or reference decay curve. The fitting routine was based on the nonlinear least-squares method. Minimization was performed according to Marquardt.26 pTerphenyl in ethanol and N-acetyl-tryptophanamide in water were used as reference compounds.27 The contribution of component i to the total emission Si was calculated as follows
∫
Ri Si )
∞
exp(-t/τi)dt
0
∑ ∫ Ri
∞
0
) exp(-t/τi)dt
Riτi
∑ Rτ
(6)
i i
The root-mean square value of fluorescent lifetimes, 〈τ〉, for biexponential decay is determined as 〈τ〉 )
R1τ12 + R2τ22 R1τ1 + R2τ2
)
Figure 3. Decay curves of tryptophan fluorescence of native (A) and unfolded (B) GlnBP. Figures represent excitation lamp profile (curves 1), experimental decay curve (curves 2), best fitted calculated fluorescence decay curve (curves 3) and deviation between the experimental and calculated decay curve (weighted residuals; curves 4). The excitation wavelength was 297 nm, the registration wavelength was 340 nm. The fluorescence decay curves show the best fit with a three-exponential decay model. The values of τi, Si, χ2, and average lifetime 〈τ〉 are shown.
∑S τ i
i
(7)
Results and Discussion Trp Fluorescence of GlnBP and Its Complex with Gln in Native and Completely Unfolded States. Figure 2 shows the Trp emission spectra of GlnBP and GlnBP/Gln in the native and unfolded (in 3.0-6.0 M GdnHCl) states upon excitation at 297 nm. It is accepted that Tyr residues absorption at the red edge of spectrum is negligible and thus emission excited at 297 nm is completely determined by Trp residues.14 The emission spectra appear to be smooth and featureless. The fluorescence maximum is about 332 nm for the native GlnBP. It is slightly (1-2 nm) blue shifted upon Gln binding. These data differ from
those obtained earlier by Weiner and Heppel,15 who recorded the emission maximum at 336 nm with excitation at 280 nm, but are comparable to that of Axelsen et al.16 who recorded emission maximum at 330 nm by excitation at 295 nm. As it would be expected, protein unfolding induced by 3.0-6.0 M GdnHCl results in a significant red shift of the emission spectrum (Figure 2). Emission spectra of GlnBP and its complex with Gln in the presence of 3.0-6.0 M GdnHCl are the same, and the emission maximum is 353 nm. The fluorescence intensity of the unfolded protein is essentially lower than that of the native one that is not surprising as fluorescence quantum yield of Trp residues accessible to the solvent is usually lower than that located in the inner parts of the molecule. The fluorescence decay curves show the best fit with a three-exponential decay model with an average lifetime of the native protein of 5.77 ns and 3.09 ns for unfolded protein (Figure 3). The GdnHCl induced simultaneous decrease of both the fluorescence lifetime (Figure 3) and the Trp fluorescence intensity (Figure 2) suggests an increase of the Journal of Proteome Research • Vol. 4, No. 2, 2005 419
research articles fluorescence dynamic quenching on the protein unfolding. According to Axelsen et al.16 fluorescence decay curves show the best fit with a two-exponential decay model with an average lifetime of 5.93 ns. Since the main component consists more than 90% both in the work of Axelsen et al.16 and in our data, it is possible to conclude that there is a good agreement between the data of Axelsen et al.16 and our. The fluorescence lifetime of the unfolded protein (3.09 ns) is similar to that of other denatured proteins.28 Microenvironment of Trp Residues of GlnBP. To evaluate the contribution of individual Trp residue to the bulk protein emission the analysis of the 3D structure of the GlnBP (file 1GGG.ent; Hsiao et al.)5 and its complex with Gln (file1WND.ent; Sun et al.)6 was done. The analysis showed that both Trp residues of this protein (Trp 32 and Trp 220) are located in the large domain (Figure 1) far from ligand-binding site, Trp32 is in the first R-helix Phe 27-Glu 38, and Trp 220 is in the final one Thr 212-Phe 221. Trp residues are located far from each other (R ) 16 Å) and the efficiency of energy transfer between them is negligible. This result contradicts to one of the conclusions of the work of Axelsen et al.16 about the existence of Trp-Trp interactions in intact protein. The microenvironment density of both Trp residues is not very high (there are 63 atoms and 66 atoms in the microenvironments of Trp 32 and Trp 220, respectively). According to Axelsen et al.,16 both Trp residues are partially accessible to solvent. However the analysis of protein 3D structure shows that Trp 32 is much more buried in comparison with Trp 220 (Figure 1B and 4). The polarity of their microenvironment differs significantly. There are many hydrophobic groups (Leu 5, Val 7, Ile 35, Ala 36, Leu 39, Leu 41, ring of Tyr 43, Leu 45, Leu 64, Leu 66, Ile 187, and Val 200) and only one polar group of the amino acid side chain (OH group of Tyr 43) in the microenvironment of Trp 32. In the microenvironment of Trp 220 there are six polar groups of the side chains of amino acids (Asp 30, Tyr 163, Lys 166, Lys 219) and only seven nonpolar groups (Pro 15, Phe 18, Val 25, Phe 27, Tyr 163, Ile 216, Phe 221) that is essentially less than in the vicinity of Trp 32. So it is likely that the fluorescence spectrum of Trp 220 is more red shifted in comparison with that of Trp 32, that agrees with the conclusions of Axelsen et al.16 However, the results of the analysis of Trp 32 microenvironment gave no reason to assume that this residue has so blue fluorescence spectrum position and so low quantum yield as it is given in the work of Axelsen et al.16 for mutant forms Trp220Tyr (λmax ) 316 nm; quantum yield 0.024) and Trp220Phe (λmax ) 326 nm and quantum yield 0.043). The authors themselves noticed that the fluorescence spectra of Trp220Tyr and Trp220Phe differ significantly and pointed that side-specific mutation of Trp residues to yield single-Trp forms of multiTrp containing protein is subjected to pitfalls because there is no structurally or photophysically “conservative” substitution for Trp.16 Another self-contradiction of this work is that despite such difference in emission spectrum position and quantum yield, bimolecular quenching constants (kq) for Trp 32 and Trp 220 are practically the same. This is also in conflict with the results of 3D structure analysis (Figure 1B and 4). Fluorescence Emission and Microenvironment of Trp Residues of GlnBP Complex with Gln. As it has been mentioned above, the complex formation of GlnBP with Gln induces only a slight (1-2 nm) blue shift of fluorescence spectra (Figure 420
Journal of Proteome Research • Vol. 4, No. 2, 2005
Kuznetsova et al.
2). To explain the reason for such small changes, we have performed a comparative analysis of Trp residues microenvironment of GlnBP and its complex with Gln. The binding of L-glutamine causes cleft closing and a significant structural change with the formation of the so-called “closed form” structure (Figure 1, file1WND.ent; Sun et al.6) and significant change of interdomain dynamics.29 Nonetheless, Gln binding induces insignificant changes in Trp 32 and Trp 220 microenvironments. In GlnBP/Gln there are 65 atoms in the microenvironments of both Trp residues. As it has been mentioned above, the only polar group in the vicinity of Trp 32 is OH group of Tyr 43. The position of this group changes only slightly with the formation of GlnBP/Gln. Although the number of atoms in the microenvironment of Trp 220 residue is practically the same for ligand-free and ligand-bound GlnBP, its composition changes. In the Trp 220 microenvironment of GlnBP there are Pro 15, Phe 18, Phe 27, Phe 221, Tyr 163, while in the GlnBP/ Gln only Pro 15, Phe 18, and Phe 27 remain in the microenvironment. Phe 221 and Tyr 163 belonging to the other domain leave the microenvironment of Trp 220, when this domain turns at the complex formation. There are more polar groups in the microenvironment of Trp 220 in the absence of the ligand (six) than in the complex with Gln (four). At the complex formation, the OH group of Tyr 163 and N2 Lys 166 are removed from the microenvironment, while the position of other groups are changed only slightly (Figure 1). All of this explains why the ligand-binding does not significantly change the Trp fluorescence spectrum of GlnBP (Figure 2). Before continuing the analysis of the contribution of separate Trp to the bulk protein fluorescence, it is necessary to examine the participation of Tyr residues in the emission of this protein. Tyrosine Fluorescence of GlnBP. Usually, the use of protein intrinsic fluorescence for examination of protein foldingunfolding means the use of its Trp fluorescence. This is due to high sensitivity of all Trp fluorescence characteristics to their microenvironment and consequently to all structural changes of protein. However, there is a risk to monitor only local structural changes while using Trp fluorescence. That is why along with Trp fluorescence, the recording of Tyr one could be also useful. Contrary to Trp residues, usually there are many Tyr residues in proteins. In particular, GlnBP has two Trp and ten Tyr residues: four Tyr residues are located in the small domain, and the other six Tyr residues are in the large domain (Figure 1). The position of fluorescence spectrum of Tyr residues only slightly depends on the properties of their microenvironment. In Trp containing proteins the only characteristic of Tyr fluorescence, which can be used for examination proteins structural changes, is the contribution of Tyr residues to the bulk protein fluorescence. The advantage of monitoring Tyr fluorescence is that Tyr residues are located all over protein macromolecule, and there is a chance to find some other transitions in protein unfolding, which might be not detected by Trp fluorescence. Usually, the contribution of Tyr residues to the bulk fluorescence protein is evaluated from the comparison of emission spectra excited at 280 nm and 295-297 nm, as it is believed that Tyr residues have no absorption in the long-wave edge of absorption spectrum. Figure 5 shows the results of such evaluation for GlnBP in native and unfolded states. It is clear that though GlnBP contains ten Tyr residues, their contribution to the bulk fluorescence of native GlnBP is negligible. Analysis of the microenvironments of Tyr residues showed that there
research articles
Intrinsic fluorescence of GlnBP and GlnBP/Gln
Figure 4. Radial dependence of packing density of atoms in macromolecule about geometrical centers of indole rings of tryptophan residues Trp 32 and Trp 220 of GlnBP (solid lines) and GlnBP/Gln(dash-dot lines). This figure is constructed based on Protein Data Bank,7 file 1GGG.ent, ref 5 and 1 WDN.ent, ref 6.
Figure 5. Contribution of tyrosine residues to the bulk fluorescence of the native and the unfolded GlnBP. Fluorescence spectrum of native (curves 1 and 2) and unfolded (curves 3 and 4) GlnBP excited at 280 nm (curves 1 and 3) and 297 nm (curves 2 and 4). Curves 5 and 6 represent the difference between curves 1, 2 and 3, 4, respectively. Table 1. Efficiency of Nonradiative Energy Transfer from Tyrosine to Tryptophan Residues of GlnBP (1GGG.ent) and GlnBP/Gln(1WDN.ent) Trp 32
Trp 220
Tyr\Trp
GlnBP
GlnBP + Gln
GlnBP
GlnBP + Gln
Tyr 24 Tyr 43 Tyr 85 Tyr 86 Tyr 123 Tyr 143 Tyr 163 Tyr 185 Tyr 213 Tyr 217
0.89 0.96 0.95 0.86 0.00 0.02 0.42 0.89 0.40 0.29
0.91 1.00 0.91 0.86 0.00 0.02 0.36 0.90 0.34 0.28
0.83 0.88 0.43 0.94 0.05 0.56 1.00 0.91 0.76 0.99
0.78 0.89 0.31 0.96 0.00 0.73 1.00 0.92 0.70 0.99
are two reasons of their low quantum yield in GlnBP: (1) There are conditions for an effective energy transfer from the majority of Tyr to Trp residues. All Tyr residues except Tyr 123 can effectively transfer their excitation energy to Trp 32 and/or Trp 220 directly (Table 1), or via other Tyr residues (Table 2); (2) Most of the Tyr residues could be quenched not only by energy transfer to Trp residues, but also by quenching groups in their vicinity or by effective energy transfer to Tyr residues which are quenched (Table 3). Tyr fluorescence intensity increases significantly when GlnBP is unfolded in 3.0-6.0 M GdnHCl (Figure 5). Thus, the change of Tyr residues contribution to the bulk fluorescence of protein can be regarded as a parameter for monitoring protein unfolding.
Figure 6. Change of the intensity of Trp fluorescence (λem ) 365 nm) of GlnBP in the process of protein unfolding induced by GdnHCl. Fluorescence was excited at 297 and 280 nm (curves 1 and 2, respectively). In these dependences, all values are reduced to the value of fluorescence intensity of native protein excited at 297 nm, and value of fluorescence intensity of unfolded protein excited at 280 nm was set equal to that excited at 297 nm.
Dependence of Trp Fluorescence of GlnBP from Excitation Wavelength. To test the existence of an effective energy transfer from Tyr to Trp residues we have recorded the dependence of fluorescence intensity at 365 nm (Trp fluorescence) upon concentration of GdnHCl for two excitation wavelengths: 280 and 297 nm (Figure 6). In these dependences, all values are reduced to the value of fluorescence intensity of native protein excited at 297 nm, and the value of fluorescence intensity of the unfolded protein excited at 280 nm was set equal to that excited at 297 nm. It is worth stressing that it is possible to equal these values because protein unfolding induces the loss of the conditions of Tyr f Trp energy transfer, and compensates all peculiarities of the microenvironments of separate Trp and Tyr residues. We can expect that for native protein we will obtain (I280/I297)365 > 1, if at 280 nm the population of the Trp excited states is determined by absorption of Trp residues themselves and energy transfer from Tyr residues (Tyr f Trp). This ration must be equal to unity if there is no energy transfer from Tyr to Trp. Our data show that (I280/I297)365 < 1. This can be due to the change of absorption spectrum induced by protein unfolding, which can induce the decrease of the value (I280/I297)365 if for molecular extinction coefficient the equation (280/297)N < (280/297)U is valid. The change of the shape of absorption spectra induced by protein unfolding must be considerable enough to overbalance the effect of the energy transfer TyrfTrp, which contrary induces the excess of I280 over I297. Our interpretation of the relation (I280/I297)365 < 1 is based on the work Axelsen et al.,16 in which the reliable differences in secondary derivative absorption spectra of wild type of GlnBP and four single-Trp GlnBP mutants have been shown. The second derivative absorption spectra reveal a small shoulder in the absorption spectra (i.e., negative second derivative) at 293 nm for wild type and mutants with Trp 220 changed for Tyr or Phe. This shoulder in the long-wave range of absorption spectrum of GlnBP was reliably recorded in our experiments (Figure 7A). According to Axelsen et al.16 Trp 32 is responsible for this shoulder in the absorption spectrum. According to our data Trp 220 gives the major contribution on the red edge of absorption spectrum of GlnBP. This conclusion is made on the basis of the dependence of parameter A on the excitation wavelength (Figure 7B). As it was shown above, the Tyr residues give negligibly small contribution to the fluorescence spectrum of native GlnBP. As a consequence the dependence of parameter A on the excitation wavelength Journal of Proteome Research • Vol. 4, No. 2, 2005 421
research articles
Kuznetsova et al.
Table 2. Efficiency of Energy Transfer between Tyrosine Residues in GlnBP (file 1GGG.ent)a Tyr 24
Tyr 24 Tyr 43 Tyr 85 Tyr 86 Tyr 123 Tyr 143 Tyr 163 Tyr 185 Tyr 213 Tyr 217
0.59 0 0.01 0 0 0.08 0.06 0.01 0.06
Tyr 43
Tyr 85
Tyr 86
Tyr 123
Tyr 143
Tyr 163
Tyr 185
Tyr 213
Tyr 217
12.8
16.39 14.63
18.13 20.03 7.05
33.04 37.34 23.25 18.16
23.94 32.95 23.61 17.51 17.96
14.85 22.40 16.11 11.58 23.57 11.86
16.83 19.34 6.36 5.15 18.08 18.50 13.78
21.00 18.23 7.95 7.95 23.97 24.34 16.32 11.57
20.58 21.53 10.57 5.63 20.47 18.61 11.63 10.76 6.24
0 0 0 0 0.02 0 0.03 0.06
0.49 0.02 0.02 0 0.97 0.89 0.28
0 0.29 0.78 0.99 0.95 0.99
0.04 0 0.09 0 0
0.72 0.26 0.05 0.15
0.36 0.18 0.58
0.58 0.36
0.01
a
The Table shows the values of the efficiency of nonradiative energy transfer W (lower left part of the table) and the distances between the geometrical centers of the rings of tyrosine residues (upper right part of the table).
Table 3. Characteristics of the Microenvironment of Tyrosine Residues in GlnBP (1GGG.ent)
tyrosine residue
Na
24
(52)
43 85
(46) (83)
86 123 143 163 185 213
(68) (35) (61) (47) (56) (64)
217
(66)
amino acid side chains - potential quenchers of tyrosine fluorescence
11 11 32 28 28 217 127 156 167 88 85 85 86 224
Thr Thr Trp Asp Asp Tyr Asn His Thr Ser Tyr Tyr Tyr Glu
O OG1 NE1 OD1 OD2 OH OD2 ND1 OG1 OG O N OH OXT
oxygen atoms of bounded water ROH (Å)
3.1 5.6 4.2 2.9 3.6 3.8 3.6 3.9 5.6 3.7 2.7 3.5 3.8 4.5
NHOHb
ROHc (Å)
∆Iexp(λ) ) I297(λ) - I280(λ) ) ∆I297(∆Trp220) ∆I280(TyrfTrp220) - ∆I280(TyrfTrp32) (8)
317 225
5.3 3.1
281
2.7
332 231
3.1 4.3
a N is the number of atoms in the microenvironment of tyrosine residue, in brackets the number of the atoms in the microenvironment of oxygen atom of hydroxyl group of tyrosine is given. b NHOH is the number of the molecule of bounded water c ROH is distance from the oxygen atom of hydroxyl group of tyrosine residue.
Figure 7. Spectral characteristics of GlnBP. A. Absorption spectrum. B. The dependence of the parameter A upon excitation wavelength. C. Fluorescence spectra excited at 297 nm (curve 1) and 280 nm (curve 2). Curve 3 represents the difference ∆Iexp(λ) ) I297(λ) - I280(λ).
can be determined only by the change of relative contribution to the bulk protein fluorescence of Trp 32 and Trp 220. The character of this dependence suggests that the contribution of Trp with more red fluorescence spectrum position increases with the increase of the wavelength. Figure 7C shows fluores422
cence spectra of native GlnBP excited at 297 and 280 nm, which are given taking into account the determined relation (I280/I297)365 ) 0.93 (see Figure 6). This figure gives also the differented spectrum I297(λ) - I280(λ)
Journal of Proteome Research • Vol. 4, No. 2, 2005
The value of ∆Iexp(λ) represents only the that part of Trp 220 emission, which is determined by the excessive absorption of this residue on the red edge of the absorption spectrum ∆I297(∆Trp220). In eq 8, the values ∆I280(TyrfTrp220) and ∆I280(TyrfTrp32) are the parts of Trp220 and Trp32 emission, which are determined by their transition to the excited-state induced by energy transfer from Tyr residues at 280 nm. Fluorescence emission spectrum of Trp residues is much more sensitive to its microenvironment than the absorption spectrum. The differences of the absorption spectra of individual Trp are not large and usually are not taken into consideration in the analysis of fluorescence data. GlnBP is an unusual example of protein in which one of two Trp residues apparently plays the dominant role in absorption at the red edge of spectrum. The analysis of the fluorescence experimental data and characteristics of microenvironment of two Trp residues of GlnBP allowed us to conclude that Trp 220 is the residue which plays a dominant role in the protein fluorescence emission in the case of excitation at the red edge of the absorption spectrum. Later on this conclusion should be proved by the reexamination of the mutant forms Trp 32 Phe and Trp220Phe of GlnBP. However, we believe that the used approach, based on the analysis of experimental fluorescence data, which takes into account Trp residue characteristics of microenvironment and peculiarities of location in protein, is fruitful and has predictive capacity. In fact, this approach already allowed us to suggest that the quenching of the single Trp 48 of azurin by its Cu center has a long-range character and it is induced by electron removal from indole state in the excited state.18 This statement was later proven experimentally.30 In addition, we also showed the existence of “cavities” near Trp 48 and that the unique blue fluorescence spectrum of azurin is determined not by high density and rigidity, but by the exclusively hydrophobic character of its microenvironment.18 This conclusion was also proved later.31 In the case of actin, this approach allowed to predict the contribution of different Trp in the bulk fluorescence of this protein.32 On the basis of the analysis of Trp residues microenvironment and peculiarities of their location in protein macromolecule, we concluded that Trp 79 and Trp 86 are quenched due to the existence of quenching groups in their vicinity (especially SG
research articles
Intrinsic fluorescence of GlnBP and GlnBP/Gln
atom of Cys 10 near NE1 atom of the indole ring of Trp 86) and effective energy transfer between them, while the major contribution to the bulk actin fluorescence is given by two other Trp residues (Trp 340 and Trp 356), which are located in hydrophobic microenvironment.32 Our conclusions were completely confirmed in the work of Doyle et al.,33 in which mutant forms of recombinant actin with Trp changed by Phe were examined. Unfortunately, the reasons for the anomalous absorption spectrum of Trp 220 are not understood. Also, it is unclear how often so different absorption spectra of Trp residues are met in proteins. The existence of differences in the absorption spectra (excitation spectra) and fluorescence spectra of Trp residues may be used for the development of the approach for decomposition of the multicomponent protein fluorescence spectrum. In conclusion, the results of this work show that Trp fluorescence excited at 280 and 297 nm can differ. This must be taken into account while evaluating the contribution of Tyr residues to protein fluorescence excited at 280 nm. This effect, apparently, can be the reason of the differences in the longwave range of some proteins fluorescence spectra (reduced to fluorescence intensity at 365 nm) excited at 280 and 297 nm. In this case, presenting these spectra in comparable units, it is necessary to take into account the value of the ratio(I280/I297)365. To determine this ratio, it is necessary, alongside the spectra of native protein, to record fluorescence spectra of unfolded protein at 280 and 297 nm.
Acknowledgment. This project was realized in the frame of CRdC-ATIBB POR UE-Campania Mis 3.16 activities (S.D., M.R). This work was supported by grants from F.I.R.B. (S.D., M.R.), the Italian National Research Council (S.D, M.R), INTAS 2001-2347 (K.T.), INTAS 04-83-3162 (O.S.), RFBR 05-0448588 (O.S.), and from Presidium of Russian Academy of Sciences for the program “Molecular and Cell Biology” (K.T.). References (1) Higgins, C. F. Annu. Rev. Cell Biol. 1992, 8, 67-113. (2) Boos, W.; Lucht, J. M. Periplasmic binding-protein-dependent ABC transports. In E. coli and Salmonella typhimurium: Cellular and molecular biology (Lin, E., Ed), 1995 pp 1175-1209, American Society for Microbiology, Washington, DC. (3) Gestwicki, J. E.; Strong, L. E.; Borchardt, S. L.; Cairo, C. W.; Schnoes, A. M.; Kiessling, L. L. Bioorgan. Med. Chem. 2001, 9, 2387-2393.
(4) Adams, M. D.; Oxender, D. L. J. Biol. Chem. 1989, 264, 1573915742. (5) Hsiao, C.-D.; Sun, Y.-J.; Rose, J.; Wang, B.-C. J. Mol. Biol. 1996, 262, 225-242. (6) Sun, Y. J.; Rose, J.; Wang, B. C.; Hsiao, C. D. J. Mol. Biol. 1998, 278, 219-229. (7) Bernstein, F. C.; Koetzle, T. F.; Williams, G. J. B.; Meyer, E. F., Jr.; Brice, M. D.; Rodgers, J. R.; Kennard, O.; Shimanouchi, T.; Tasumi, M. J. Mol. Biol. 1977, 112, 535-542. (8) Humphrey, W.; Dalke, A.; Schulten, K. J. Mol. Graphics 1996, 14, 33-38. (9) Merritt, E. A.; Bacon, D. J. Methods Enzymol. 1977, 277, 505524. (10) Dattelbaum, J. D.; Lakowicz, J. R. Anal. Biochem. 2001, 291, 8995. (11) D’Auria, S.; Lakowicz, J. R. Curr. Opin. Biotechnol. 2001, 12, 99104. (12) Zhou, R.; Huang, X.; Margulis, C. J.; Berne, B. J. Science 2004, 305, 1605-1609. (13) Bychkova, V. E.; Ptitsyn, O. B. Chemtracts-Biochem. Mol. Biol. 1993, 4, 133-163. (14) Lakowicz, J. R. Principles of Fluorescence Spectroscopy., 2nd ed.: Kluwer Academic/Plenum Publishers: New York, 1999. (15) Weiner, J. H.; Heppel, L. A. J. Biol. Chem. 1971, 246, 6933-6941. (16) Axelsen, P. H.; Bajzer, Z.; Prendergast, F. G.; Cottam, P. F.; Ho, C. Biophys. J. 1991, 60, 650-659. (17) Bradford, M. M. Anal. Biochem. 1976, 72, 248-254. (18) Turoverov, K. K.; Kuznetsova, I. M.; Zaitzev, V. N. Biophys. Chem. 1985, 23, 79-89. (19) Kuznetsova, I. M.; Turoverov, K. K. Tsitologia 1998, 40, 747-762. (20) Forster, Th. Rad. Res. Suppl. 1960, 2, 326-339. (21) Dale, R. E.; Eisinger, J. Biopolymers 1974, 13, 1573-1605. (22) Eisinger, J.; Feuer, B.; Lamola, A. A. Biochemistry 1969, 8, 39083915. (23) Steinberg, I. Z. Annu. Rev. Biochem. 1971, 40, 83-114. (24) Turoverov, K. K.; Kuznetsova, I. M. J. Fluorescence 2003, 13, 4157. (25) Turoverov, K. K.; Biktashev, A. G.; Dorofeyuk, A. S.; Kuznetsova, I. M. Tsitologia 1998, 40, 806-817. (26) Marquardt, D. W. J. Soc. Indust. Apl. Math. 1963, 11, 431-441. (27) Zuker, M.; Szabo, A. G.; Bramall, L.; Krajcarski, D. T.; Selinger, B. Rev. Sci. Instrum. 1985, 56, 14-22. (28) Grinvald, A.; Steinberg, I. Z. Biochim. Biophys. Acta 1976, 427, 663-678. (29) Pang, A.; Arinaminpathy, Y.; Sansom, M. S.; Biggin, P. C. FEBS Lett. 2003, 550, 168-74. (30) Petrich, J. W.; Longworth, J. W.; Fleming, G. R. Biochemistry 1987, 26, 2711-2722. (31) Reshetnyak, Y. K.; Koshevnik, Y.; Burstein, E. A. Biopys. J. 2001, 81, 1735-1758. (32) Kuznetsova, I. M.; Yakusheva, T. A.; Turoverov, K. K. FEBS Lett. 1999, 452, 205-210. (33) Doyle, T. C.; Hansen, J. E.; Reisler, E. Biophys. J. 2001, 80, 427-434.
PR0498077
Journal of Proteome Research • Vol. 4, No. 2, 2005 423
