Protein Engineering, Vol. 15, No. 11, 855-859,
November 2002
© 2002 Oxford University Press
19F NMR study of the leucine-specific binding protein of Escherichia coli: mutagenesis and assignment of the 5-fluorotryptophan-labeled residues
Department of Chemistry, Clarkson University, Potsdam, NY 13699, USA
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResults and discussionReferences |
|---|
The Escherichia coli L-leucine receptor is an aqueous protein and the first component in the distinct transport pathway for hydrophobic amino acids. L-Leucine binding induces a conformational change, which enables the receptor to dock to the membrane components. To investigate the ligand-induced conformational change and binding properties of this protein, we used 19F NMR to probe the four tryptophan residues located in the two lobes of the protein. The four tryptophan residues were labeled with 5-fluorotryptophan and assigned by site-directed mutagenesis. The 19F NMR spectra of the partially ligand free proteins show broadened peaks which sharpen when L-leucine is bound, showing that the labeled wild-type protein and mutants are functional. The titration of L-phenylalanine into the 5-fluorotryptophan labeled wild-type protein shows the presence of closed and open conformers. Urea-induced denaturation studies support the NMR results that the wild-type protein binds L-phenylalanine in a different manner to L-leucine. Our studies showed that the tryptophan to phenylalanine mutations on structural units linked to the binding pocket produce subtle changes in the environment of Trp18 located directly in the binding cleft.
Keywords: 5-fluorotryptophan/19F NMR/leucine/mutagenesis/receptor
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionReferences |
|---|
The periplasmic receptors of Escherichia coli are ideal proteins to manipulate in order to study the mode of binding of hydrophobic ligands of interest in the pharmaceutical industry and medical field. Especially interesting are the two periplasmic components of the branched-chain amino acid transport system in E.coli, namely the leucine–isoleucine–valine binding protein (LIV) and the leucine-specific binding protein (LS). These water-soluble receptors localize in the periplasmic space translocate small molecules to cytoplasm of the bacteria. The method of transport relies on the recognition of the binding protein to the membrane-bound components. A change of tertiary structure from the open to closed form is vitally important for this interaction (Luck and Falke, 1991a
; Shilton et al., 1996). The mechanism of cleft closure is a theme throughout nature. Thus, LIV has been used as a model for numerous studies of fastidious human proteins such as the N-terminal domain of the group I metabotropic glutamate receptors and the N-methyl-D-aspartate receptor (Bessis et al., 2000; Paoletti et al., 2000; Lynch and Guttmann, 2001).
LIV and LS are ~80% identical in amino acid content and have virtually the same three-dimensional structure in the unliganded form (Sack et al., 1989b). However, the two periplasmic receptors have different specificities for ligands and the range of ligands for LS is much wider than originally observed in the initial studies of the proteins (Landick and Oxender, 1985; Adams et al., 1991). LS binds leucine and phenylalanine and also the fluorinated analogs of both of these amino acids (Rahmanian et al., 1973; Adams et al., 1991; Luck and Johnson, 2000). LIV recognizes leucine, isoleucine and valine and, to a lesser extent, threonine, serine and alanine, but to date LIV has not been shown to bind any fluorinated analogs of the amino acids. LS has been studied by X-ray crystallography but only in the open form without ligand. LIV has also been examined as an unliganded protein along with a liganded `open’ form where leucine is associated only with the N-terminal lobe of protein (Sack et al., 1989a). No closed structure has been published and the mechanism of ligand binding is unknown.
These periplasmic receptors offer a unique opportunity to explore subtle differences in substrate specificity in the ligand-binding pocket. In addition, we have a unique prospect to explore the engineering of a specific binding pocket to accommodate a ligand of interest. Since LS binds fluorinated substrates, we may be able to expand the range of binding in the cleft to include fluorinated substances such as those used in biological warfare. Thus, our periplasmic binding protein could be used as a basis for building biomolecular scavengers for these agents.
In this light, we are gathering biophysical information concerning these proteins. One of our methods for studying the dynamics of these important proteins is 19F NMR. 19F NMR is a proven technique in the study of protein structure and dynamics because the fluorine nucleus is readily incorporated at specific sites within proteins where it provides a sensitive probe for low-resolution structural information (Luck and Falke, 1991a,b). For proteins too large or unstable for full NMR structural determination, 19F NMR gives valuable information about conformational changes since the nucleus is able to detect changes in the local conformational environment, including van der Waals packing interactions and local electrostatic fields. The biosynthetic incorporation of fluorine into tryptophan residues in receptor proteins has shown that the fluorine has little effect on the structure and function of protein (Danielson and Falke, 1996). Each of the four Trp residues in LS is found in an interesting region of the structure (Figure 1): one in the binding cleft which may interact directly with the ligand, one near the hinge connecting the two domains, one buried in the N-terminal domain and a single Trp located on the C-domain. With these probes it is possible to examine both binding events within the cleft and global conformational change in this protein. In this work, we produced site-directed mutants of all four Trp residues. We labeled LS wild-type and four Trp to Phe mutants with 5-fluorotryptophan (5F-Trp) and assigned 19F NMR resonances to specific Trp residues, which allowed us to study the dynamics and substrate binding.
|
-carbon backbone structure of the receptor (Sack et al., 1989b |
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionReferences |
|---|
Preparation of mutants
Plasmid pKSty containing livK, the gene coding for the leucine-specific binding protein (LS) from E.coli, was obtained from the Oxender Laboratory, Michigan State (Oxender et al., 1980; Adams et al., 1991). LS protein contains four Trp residues at the positions 18, 278, 320 and 336 (Figure 1
). In order to create mutants with Trp replaced by Phe in the primer sequence, a tryptophan codon TGG, at the positions mentioned above, was replaced by a phenylalanine codon TTC. Two primers for each mutant (Table I) were custom synthesized by Integrated DNA Technlogies (Coralville, IA). Site-directed mutagenesis was carried out with the Quick-Change mutagenesis kit (Stratagene, La Jolla, CA). Mutants W18F and W336F were screened by digestion with restriction enzyme TspRI, and mutants W278F and W320F were screened using DpnI and TaqI, respectively. Additionally, all mutants were confirmed by DNA sequencing using an Applied Biosystems 373S Model automated DNA sequencer at the Trudeau Institute (Saranac Lake, NY).
|
Protein production
Wild-type of plasmid pKSty and mutant plasmids were transformed in the BL21(DE3) strain of E.coli. Cells were grown at 37°C, to OD600 = 0.7 in LB medium supplemented with ampicillin (100 mg/ml). Cells were harvested, washed with 100 ml of M9 medium and resuspended in 1 l of M9 medium supplemented with 2 mM MgSO4, 0.1 mM CaCl2, 0.04% glucose, 1% casamino acids, 0.1% thiamine, 1% glycerol, 100 mg/ml ampicillin and 1 g 5F-Trp. The cultures were incubated to exhaust the residual Trp in the medium for 30 min, then isopropyl-ß-D-thiogalactoside (IPTG) was added to a final concentration of 0.5 mM. Cells were pelleted after 4 h and proteins were isolated using a standard osmotic procedure. LS proteins, wild-type and mutants, were further purified by ion-exchange chromatography on a DEAE-Sephacel column using a 0–0.25 M NaCl gradient in 10 mM Tris buffer, pH 7.5. The purified proteins were dialyzed in 10 mM potassium phosphate buffer, pH 6.9, supplemented with 0.02% NaN3. Proteins were screened by SDS–PAGE and visualized with Coomassie Brilliant Blue stain.
Linear combination of spectra
To determine the efficiency of 5F-Trp incorporation into the wild-type and mutant LS, we used a linear combination of UV spectral absorbance (LINCS) (Zemsky et al., 1999; Senear et al., 2002).
NMR measurements
19F NMR spectra were obtained at 470 MHz on a Varian Model 500 spectrometer with a 5 mm 19F/1H probe. Samples containing 0.3–0.7 mM of 5F-Trp-labeled proteins in 10 mM phosphate buffer, pH 6.9, were prepared by adding 10% (v/v) D2O as the lock solvent in the final dialysis buffer. 19F resonance was referenced to 3-fluorophenylalanine at –38 p.p.m. as an external standard. Measurement parameters included a temperature of 30°C, a relaxation delay of 2 s, a sweep width of 9000 Hz and spectral processing with 25 Hz line broadening.
Fluorescence measurements
The experimental solutions for the binding determination and conformational stability experiments were prepared from concentrated stock solutions of 20 mM Tris, pH 7.1, and 10 M urea, each containing 0.5 µM of protein. For studies with ligands, proteins were incubated with a concentration of 50 mM to ensure that the proteins were saturated. The solutions were equilibrated for 12 h at room temperature at 25°C before recording the fluorescence spectra. All fluorescence measurements were carried out at ambient temperature using a Perkin-Elmer (Norwalk, CT) LS 50B luminescence spectrophotometer. The excitation and emission bandwidths were set to 5 nm. The samples were irradiated at 278 nm and the emission was monitored from 300 to 390 nm. The fluorescence intensity at 323 nm for all samples was plotted as a function of urea concentration to create an unfolding curve. We found the unfolding transitions by the Origin program from MicroCal (Northampton, MA). Non-linear regression fitting of a two- or three-state model was carried out using the same program.
|
Results and discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionReferences |
|---|
Protein production
Most laboratories have used auxotroph strains of E.coli to produce 5F-Trp-labeled proteins (Danielson and Falke, 1996). We biosynthetically incorporated the fluorine label into LS proteins, wild-type and mutants, by expression of the livK gene in E.coli, BL21(DE3). The tryptophan auxotroph cell line W3110 did not produce high levels of LS. In addition, we observed protein degradation after induction with IPTG. We also observed lower efficiency of labeling in this auxotroph cell line.
The substitution of Phe for Trp in the mutant proteins did not affect protein expression or purification and the resulting yields were 50–80 mg of 5F-Trp-labeled protein per liter. Based on the LINCS examinations, 5F-Trp incorporation rates in LS and mutant proteins were between 90 and 100% (Table II). These results are much higher than those observed in other proteins expressed by a T7 promoter (Ross et al., 2000).
|
The effect of fluorine labeling on the overall structural stability of LS and mutants was investigated by using intrinsic tryptophan fluorescence to monitor urea denaturation. The free energy of unfolding for both of the unlabeled and 5F-Trp labeled LS wild-type without ligand gave a
G°UH2O = 38.0 ± 4 kJ/mol (Skeels, 2001). The results indicate that fluorine incorporation does not disrupt the overall structural integrity of the protein. 19F NMR assignments for the 5F-Trp residues in LS
The spectrum of wild-type 5F-Trp labeled LS with leucine in the binding pocket was compared with the individual spectra of the 5F-Trp-labeled single replacement mutants also with leucine in the binding pocket (Figure 2). Observing which peak in the wild-type spectrum was eliminated in the spectrum of each mutant led to unequivocal assignment of 5F-Trp resonances in the 19F NMR spectrum. In the wild-type spectrum all four 5F-Trp resonances can be clearly identified. There are two well-resolved peaks at –47.8 and –50.3 p.p.m. corresponding to 5F-Trp18 and 5F-Trp336, respectively. The resonances for 5F-Trp320 and 5F-Trp278 overlap slightly, indicating that they reside in a similar chemical environment. The small resonance at –49.6 p.p.m. in the wild-type spectrum represents a small fraction of unfolded or degraded protein, which is at the same frequency as observed for solvent-exposed 5F-Trp (Luck and Falk, 1991a).
|
Effects of Trp to Phe mutations on 19F NMR chemical shifts
The 19F NMR spectra of the mutants are similar to that of the native 5F-Trp-labeled receptor except that a single resonance associated with the missing Trp is deleted. In addition to the deleted resonances, Trp substitution in the 278 position caused slight structural perturbations at position 18 in the binding pocket.
As shown in Figure 1, the N-domain of LS contains Trp320, Trp278 and Trp18, which is also found in the binding cleft. The C-domain houses the single Trp336. In the 19F NMR spectrum in Figure 2 we observe that the resonance corresponding to 5F-Trp336 is unaffected by the single replacement of the other Trp residues. In addition, the resonances of the 5F-Trp residues at positions 320 and 278 are not affected by the mutation of Trp336. There are, however, slight changes in the chemical shift of the resonance for 5F-Trp18 in mutant W278F compared with the corresponding resonance in the spectrum of the wild-type. This change in the chemical shift of 5F-Trp18 in mutant W278F without any other spectral differences from the wild-type suggests that the Trp to Phe mutations in fact perturb the protein structure in the area of the binding pocket. Without a closed form of the crystal structure with leucine bound, we can only surmise that this change in resonance position is due to shifts in secondary elements in the N-terminal domain. Trp278 is buried within the domain structure and shares structural components with the residues of the binding pocket. Reductions in ring size of tryptophan to phenylalanine may well affect the local interactions at Trp278, which may in turn perturb the local environment of Trp18 in the binding pocket through common elements. This similar perturbations of 5F-Trp resonances in Trp to Phe mutants are observed in the soluble tissue factor (Zemsky et al., 1999) and the glucose and galactose binding protein (Luck and Falke, 1991a).
Structural effects of L-leucine binding
The above assignment enables the four 5F-Trp resonances to be used as probes of structural changes at four known positions within the native protein. The spectrum of the 5F-Trp wild-type protein after extensive dialysis in phosphate buffer is shown in Figure 3a. Even though the protein was subjected to NMR within 15 min of removal from dialysis, there is residual ligand present. There are eight broadened peaks for the four labeled Trps in the protein indicating multiple conformers in this sample. Figure 3b shows the same protein after excess L-leucine was added. The arrows in Figure 3a indicate the positions of the closed, liganded conformer and the lines drawn from Figure 3a to b show the structural effects observed at the 5F-Trp positions upon L-leucine binding.
|
The structural effects observed upon L-leucine binding to the mutants yield an analogous picture. From these data we were able to identify frequency shifts for each of the four labeled Trp residues. Figure 4
shows three mutants, LSW18F, LSW320F and LSW336F, with substoichiometric amounts of ligand (Figure 4a, c and e) and bound with L-leucine (Figure 4b, d and f). When L-leucine is added to the protein, the spectra (Figure 4b, d and f) show significant sharpening of the resonances. This indicates that the ligand is bound and the mutants are functional. In addition to proving functionality of the mutants, the missing resonances in the spectra of mutants with substoichiometric amounts of ligand allow one to assign the resonances in the spectrum of the extensively dialyzed sample of LS (Figure 3a) and identify the chemical shifts of all of the 5F-Trps.
|
The resonance of 5F-Trp18 located in the binding pocket shows the largest frequency change upon binding with L-leucine. Trp18 lies directly in the open cleft of the binding pocket and is proposed to interact directly with ligand (Sack et al., 1989b
). This tryptophan is solvent exposed but, upon L-leucine binding, this residue would be buried within the cleft. The chemical shift of Trp18 in the unliganded state at –49.6 p.p.m. is the same as that of the free 5F-Trp in solvent (Luck and Falke, 1991b). A chemical shift from –49.6 to –47.8 p.p.m. as observed when L-leucine is bound would indicate the Trp18 moved to very different chemical environment. This change in chemical shift implicates cleft closure and the direct structural interaction between the 5F-Trp18 and L-leucine. Structural effects of L-phenylalanine on LS
The titration of L-phenylalanine into the partially empty 5F-Trp-labeled LS is shown in Figure 5. With each addition of ligand the resonances in the 19F NMR spectra sharpen and take on the appearance of the spectra of the L-leucine-bound receptor shown in Figure 2, with the exception of the broad peak at –49.6 p.p.m. This resonance has the chemical shift of a solvent-exposed 5F-Trp and is fairly broad with saturating amounts of ligand present. One could surmise that the protein is exchanging between a closed form and an open form. Perhaps the L-phenylalanine is able to bind to the protein within the cleft but does not have the contacts to hold the protein in a closed form. The spectrum indicates a fluxional area within the protein near Trp18. Thus, we see the chemical shifts for three of the four Trps in the closed positions and Trp18 exchanging between two positions. These resonances are in slow exchange on the NMR time-scale. Addition of stochiometric amounts L-leucine to the sample with 1.025 mM L-phenylalanine produced a spectrum identical to that in Figure 3b, which confirms that L-leucine binds the 5F-Trp receptor more readily that L-phenylalanine.
|
From studies in our laboratory we have shown from fluorescence data that L-phenylalanine binds to the unlabeled LS (Luck and Johnson, 2000
). Our data show Kds of 0.4 and 0.18 mM for L-leucine and L-phenylalanine, respectively. Urea denaturation studies of LS without and with ligands, L-leucine and L-phenylalanine, were performed. Figure 6 shows that it takes a higher amount of urea to denature the proteins when L-leucine and L-phenylalanine are bound in the substrate binding cleft. The unbound protein and L-phenylalanine-bound protein show a sigmoidal curve signifying a two-state denaturation process (Skeels, 2001). However, the L-leucine-bound protein shown as triangles produces a three-state denaturation curve. These data further corroborate that L-phenylalanine binds to the protein but not in the same manner as L-leucine. We are investigating this further by X-ray crystallography and a future goal is to obtain a full structure by NMR.
|
), LS with excess L-leucine (
) and LS with excess L-phenylalanine (•). Curves were generated by a non-linear regression fitting of a three-state model (Eftink, 1995To date we know that each of the Trp to Phe mutants binds leucine but the Kds have not been determined. Calorimetric studies to obtain binding data on both labeled and unlabeled LS and mutants are under way. Subsequent papers will report on the binding properties of the mutants. At present we have unequivocally assigned the resonances of the 19F NMR spectrum of the wild-type LS and have shown that the mutants are indeed functional. We have shown that Trp to Phe mutations far from the binding pocket can affect the structure of the pocket and these must be taken into account when designing new proteins. We have shown by 19F NMR that L-phenylalanine binds to the 5F-Trp receptor, but in a different manner to L-leucine. This has been further corroborated by urea denaturation studies on LS.
|
Notes |
|---|
1 To whom correspondence should be addressed. E-mail: luckla{at}clarkson.edu
|
Acknowledgments |
|---|
The authors thank Pamela Scott Adams of the Trudeau Institute for DNA sequencing and Derrick Swartz and Matthew Skeels for the fluorescence studies. They also thank J.B.Alexander Ross and Elena Rusinova for LINCS measurements on the 5F-Trp LS and mutants. This research was performed in part at the National High Magnetic Field Laboratory supported by the National Science Foundation through Cooperative Agreement DMR-0084173. This work was supported by the Petroleum Research Fund (36825-AC4).
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionReferences |
|---|
Adams,M.D., Maguire,D.J. and Oxender,D.L. (1991) J. Biol. Chem., 266, 6209–6214.
Bessis,A.S., Bertrand,H.O., Gavez,T., DeColle,C., Pin,J.P. and Acher,F. (2000) Protein Sci., 9, 2200–2209.[Web of Science][Medline]
Danielson,M.A. and Falke,J.J. (1996) Annu. Rev. Biophys. Biomol. Struct., 25, 163–195.[CrossRef][Web of Science][Medline]
Eftink,M.R. (1995) Methods Enzymol., 259, 487–512.[Web of Science][Medline]
Landick,R. and Oxender,D.L. (1985) J. Biol. Chem., 260, 8257–8261.
Luck,L.A. and Falke,J.J. (1991a) Biochemistry, 30, 4248–4256.[CrossRef][Medline]
Luck,L.A. and Falke,J.J. (1991b) Biochemistry, 30, 4257–4261.[CrossRef][Medline]
Luck,L.A. and Johnson,C. (2000) Protein Sci., 9, 2573–2576.[Web of Science][Medline]
Lynch,D.R. and Guttmann,R.P. (2001) Curr. Drug Targets, 2, 215–231.[CrossRef][Web of Science][Medline]
Oxender,D.L., Anderson,J.J., Daniels,C.J., Landick,R., Gunsalus,R.P., Zurawski,G. and Yanofsky,C. (1980) Proc. Natl Acad. Sci. USA, 77, 2005–2009.
Paoletti,P., Perin-Dureau,F., Fayyazuddin,A., Le Goff,A. and Neyton,J. (2000) Neuron, 28, 911–925.[CrossRef][Web of Science][Medline]
Rahmanian,M., Claus,D.R. and Oxender,D.L. (1973) Bacteriology, 116, 1258–1266.
Ross,J.B.A., Rusinova,E., Luck,L.A. and Rousslang,K.W. (2000) In Lakowicz,J.R. (ed.), Topics in Fluorescence Spectroscopy, Vol. 6: Protein Fluorescence. Kluwer Academic/Plenum Press, New York, pp. 17–42.
Sack,J.S., Saper,M.A. and Quiocho,F.A. (1989a) J. Mol. Biol., 206, 171–191.[CrossRef][Web of Science][Medline]
Sack,J.S., Trakhanov,S.D., Tsigannik,I.H. and Quiocho,F.A. (1989b) J. Mol. Biol., 206, 193–207.[CrossRef][Web of Science][Medline]
Senear,D.F., Mendelson,R.A., Stone,D.B., Luck,L.A., Rusinova,E. and Ross,A.J.B. (2002) Anal. Biochem., 300, 77–86.[CrossRef][Web of Science][Medline]
Shilton,B.H., Flocco,M.M., Nilsson,M. and Mowbray S.L. (1996) J. Mol. Biol., 264, 350–363.[CrossRef][Web of Science][Medline]
Skeels,M.C. (2001) PhD Thesis, Clarkson University, Potsdam, NY.
Zemsky,J., Rusinova,E., Nemerson,Y., Luck,L.A. and Ross,J.B.A. (1999) Proteins, 37, 709–716.[CrossRef][Web of Science][Medline]
Received February 1, 2002; revised June 17, 2002; accepted August 13, 2002.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  A. H. Ahmed, A. P. Loh, D. E. Jane, and R. E. Oswald Dynamics of the S1S2 Glutamate Binding Domain of GluR2 Measured Using 19F NMR Spectroscopy J. Biol. Chem., April 27, 2007; 282(17): 12773 - 12784. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
U. Magnusson, B. Salopek-Sondi, L. A. Luck, and S. L. Mowbray X-ray Structures of the Leucine-binding Protein Illustrate Conformational Changes and the Basis of Ligand Specificity J. Biol. Chem., March 5, 2004; 279(10): 8747 - 8752. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Koyanagi, T. Katayama, H. Suzuki, and H. Kumagai Identification of the LIV-I/LS System as the Third Phenylalanine Transporter in Escherichia coli K-12 J. Bacteriol., January 15, 2004; 186(2): 343 - 350. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||