| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
CIHR Group in Protein Structure and Function, Department of Biochemistry, University of Alberta, Edmonton, Alberta, Canada T6G 2H7
Reprint requests to: Brian D. Sykes, 419 Medical Sciences Building, Department of Biochemistry, University of Alberta, Edmonton, AB, Canada T6G 2H7; e-mail: brian.sykes{at}ualberta.ca; fax: (780) 492- 0886.
(RECEIVED May 20, 2005; FINAL REVISION June 27, 2005; ACCEPTED June 29, 2005)
 |
Abstract |
|---|
 TOP Abstract IntroductionResultsDiscussionMaterials and methodsReferences |
|---|
2 of ~90° in the F104W mutant and ~–100° in the F153W mutant. Using results from 19F-1H heteronuclear NOE experiment, we show the replacement of L-Trp with 5fW at these positions does not change the orientation of the indole ring and the spread of the 5fW side-chain dihedral angles increases moderately for the F104(5fW) mutant and not at all for the F153(5fW) mutant. Based on these structures, we conclude that the substitution of Phe by 5fW at these two positions has minimal effects on the structure of cTnC and that the 5fW indole rings in both mutants have well defined orientation, making the two mutants viable candidates for use in in situ 19F NMR spectroscopy. Keywords: protein stability and mutagenesis; solution NMR; structure and folding
Abbreviations: Tn, troponin • (c, s)TnC, (cardiac, skeletal) troponin C • (c, s)TnI, (cardiac, skeletal) troponin I • 5fW, 5-fluorotryptophan • 4fW, 4-fluorotryptophan • 6fW, 6-fluorotryptophan • F104W, Phe 104 to Trp • F153W, Phe 153 to Trp • F104(5fW), Phe 104 to 5fW • F153(5fW), Phe 153 to 5fW • NOE, nuclear Overhauser enhancement • CSA, chemical shift anisotropy
Article and publication are at http://www.proteinscience.org/cgi/doi/10.1110/ps.051595805.
|
Introduction |
|---|
|
TOPAbstractIntroductionResultsDiscussionMaterials and methodsReferences |
|---|
Fluorine holds several advantages over other nuclei commonly used in NMR. Its inherent absence in most natural organisms reduces its background noise in NMR spectra. The high gyromagnetic ratio of this spin one-half nucleus coupled with 100% natural abundance means the detection of fluorine signal is much easier than those of 13C or 15N atoms. The electronic structure of the fluorine atom gives the atom a wide chemical shift spread, making its resonance frequency far more sensitive to its magnetic environment than most other atoms, a valuable asset for probes of protein orientation and dynamics. However, the wide CSA of the atom also induces faster transverse relaxation in solution, reducing its sensitivity in many situations.
Many methods are used to fluorine label the protein of interest (Sykes and Weiner 1980) and biosynthetic incorporation of fluorinated amino acids are among the most popular schemes used. Despite the widespread application of these fluorinated amino acids, very few systematic studies have been conducted to investigate fluorine’s perturbation on protein structures. In most investigations involving fluorine-labeled proteins, perturbation due to fluorine is considered minimal or inconsequential. However, considering fluorine is known to alter chemical properties of compounds, fluorination cannot be inert in every situation. Since it is impossible to predict how fluorine is going to change a protein’s structure or affect a protein’s biological function, it is essential that we build up semiempirical knowledge on the perturbation to protein structure due to insertion of a fluorine atom into an environment of a specific character, e.g., hydrophobic pocket or charged ligand binding site.
Pratt and Ho (1975) and Browne et al. (1970) have independently examined the relative toxicity of fluorotryptophans on bacteria. Both found that 4-fluorotryptophan (4fW) supported bacterial growth better compared to 5- (5fW) or 6-fluorotryptophan (6fW) (see Fig. 1B
for an illustration of the indole ring carbon numbering system and proton nomenclature). They also examined the specific activity of fluorotryptophan-containing 
-galactosidase and found, once again, 4fW-containing isomers of the enzyme have far higher activity compared to 5fW and 6fW. However, 5fW continues to be widely used in the community (Campos-Olivas et al. 2002; Salopek-Sondi et al. 2003; Abbott et al. 2004; Anderluh et al. 2005). Xiao and coworkers (Xiao et al. 1998) investigated conformational changes induced by global substitution of Tyr by 3-fluorotyrosine in glutathione transferase using X-ray crystallography. While substitution of the active site residue Tyr 6 reduced the catalytic efficiency of the enzyme by 10-fold owing to both the change in the pKa of the aromatic OH group and steric clashes between the fluorine and the substrate, most other mutations produced only minor changes in the protein conformation (Xiao et al. 1998). A number of other studies (both theoretical and experimental) have demonstrated that in some proteins perturbations from 19F are minimal (Lau and Gerig 1997; Campos-Olivas et al. 2002). A notable observation is that a polar 19F atom can sit comfortably inside a hydrophobic pocket without disturbing the hydrophobic core packing (Campos-Olivas et al. 2002).
|
Most EF-hand motifs follow the consensus sequence of h1**h2(x)*(y)*(z)*G(–y)I(–x)**( –z)h3**h4, where x, y, z, –x, –y, and –z represent Ca2+-chelating amino acids, * represents any amino acids; and h represents hydrophobic amino acids. For the C-terminal domain of cTnC, h1 and h2 of the first EF-hand motif, as well as h3 and h4 of the second EF-hand motif are Phe. Our decision was to mutate the h2 of the first EF-hand motif, which is residue 104, and the h3 of the second EF-hand motif, which is residue 153, to Trp (Fig. 1A
). By putting the mutations on two nonparallel helices near two different Ca2+-binding loops, we reduce the chance of protein misfolding induced by having two mutations on the same helix. The skeletal troponin C equivalents of both mutants containing L-Trp have been studied before, and the mutations are found to have minimal effects on the biological function of the protein. A study done on F105W mutant of skeletal TnC (the skeletal analog of the F104W mutant) by Trigo-Gonzalez reasoned, from an observed red shift in the absorbance spectrum of the F105W mutant, that Ca2+-binding caused the tryptophan ring to become more buried (Trigo-Gonzalez et al. 1992). However, no changes in Ca2+-binding affinity of the C-terminal domain was observed by them. Chandra and coworkers (Chandra et al. 1994) did a thorough study on the sTnC analog of the F153W mutant. They observed a twofold decrease in the Ca2+ affinity of the mutant, which they attributed to the destabilization of the hydrophobic pocket packing of sTnC by the bulky side chains of tryptophan. However, physiological assays of the mutant using reconstituted Tn–tropomyosin–actomyosin ATPase complex showed the mutant produced the same ATPase activity as the wild type, both in the presence and absence of Ca2+, indicating the change in the mutant’s Ca2+ affinity is not enough to reduce biological activity (Chandra et al. 1994). In this study, we closely scrutinized the effects of Phe-to-Trp substitutions as well as 5fW incorporation on the structure of cTnC. Using NMR techniques, we determined the high resolution structures of all four isoforms of the protein, putting the emphasis on the structure of the C-terminal domain where the substitutions reside. The orientation of the indole ring in each isoform was also carefully determined using data from homonuclear NOE as well as 19F-1H heteronuclear NOE experiments.
|
Results |
|---|
|
TOPAbstractIntroductionResultsDiscussionMaterials and methodsReferences |
|---|
. However, in the BL21-expressed F104(5fW) mutant the N-terminal domain binds Ca2+ before the C-terminal domain. This is indicated by the migration of the N-terminal domain G70 amide proton signal upon Ca2+ addition before the signal corresponding to amide proton of residue G146 in the Ca2+-saturated C-terminal domain begins to grow in intensity (Fig. 2B). It was also noted that the final intensities of most signals from the Ca2+-saturated C-terminal domain were only half the size of their counterparts from the N-terminal domain. This is in contrast to cTnC isoforms that fold normally. In those cases, signals from the Ca2+-saturated C-terminal domain are the same size as those from the N-terminal domain (Fig. 2A). This indicates that not only is the Ca2+ affinity of the C-terminal domain reduced but a portion of the C-terminal domain is also not folding properly. This hypothesis was further confirmed by 1D 19F spectra taken during the titration, which showed that besides the main peak at –49.1 ppm, multiple signals between –49.6 ppm and –50.0 ppm can be seen at the end of the titration, indicating that many conformational states of the protein exist (Fig. 2C). To ensure the abnormality was not caused by the mutation from Phe to Trp, we performed Ca2+ titrations on mutants containing L-Trp. Both F153W and F104W mutants showed a normal pattern of Ca2+ binding (a representative set of data from the Ca2+ titration of the F104W mutant is shown in Fig. 2A). Even though the {1H, 15N}-HSQC NMR spectra of these Phe-to-Trp mutants in their Ca2+-saturated state are not identical to that of the wild-type cTnC (Fig. 3), the patterns of chemical shift changes upon metal binding are very similar to the wild type and the presence of minor populations or misfolded proteins are not observed. Originally, we thought the observed reduction in the Ca2+ affinity and misfolding we saw in the F104(5fW) mutant’s C-terminal domain was a consequence of 5fW incorporation. However, the same mutant expressed in Escherichia coli Trp auxotrophic strain W3110 Trp A33 exhibited normal Ca2+ binding patterns (data not shown); thus, the phenomenon we observed seems to be restricted to the F104(5fW) mutant produced in BL21 cells.
|
|
Effects of 19F incorporation on backbone amide proton/nitrogen chemical shifts
In many situations, the replacement of a proton by a 19F atom is considered a conservative substitution. Presence of only a single 19F atom often changes little in the way of steric and electrostatic interactions in the region. However 19F does possess a different electronic configuration than a proton. Under the right circumstances, this can induce a change in the local magnetic environment, which can manifest itself as changes in the chemical shifts of nuclei surrounding it. The HSQC spectrum of the F104(5fW) mutant largely resembles that of the F104W mutant (Fig. 3C,E). However, many residues in the F153(5fW) mutant give two signals in the {1H, 15N}- HSQC (Fig. 3E). All of these residues are found in the C-terminal domain of the protein. Each pair of signals consists of a strong peak and a weak peak, with the weak peak having exactly the same chemical shifts as a peak in the F153W mutant {1H, 15N}-HSQC spectrum. Besides residue 153 itself, other residues exhibiting discernable signal doubling included residues 98, 100, 119, 121, 136, 137, 148, and 154 to 158. All the residues are mapped to a region close to the location of the 19F atom (Fig. 4). Analysis of the heteronuclear spectral data also showed that the signals in each pair have the same neighboring residues as well as identical side chain atom chemical shifts and amide proton NOE patterns; thus, they must represent the same residue. Based on these observations, we think these signal doublings are a manifestation of the incomplete 5fW-labeling of the protein. For residues close to the indole ring, the amide nitrogen and amide proton chemical shifts are sensitive enough to detect the difference between a 19F atom and a proton. The fact that signal doubling was not seen with the F104(5fW) mutant means that either the 19F incorporation is more complete for the F104(5fW) mutant, or the side chain of 5fW in the F104(5fW) mutant is not close enough to any residue to produce a significant change in the chemical shifts of amide protons and nitrogens. However, even in the case of the F153(5fW) mutant, not every residue close to the 19F atom exhibited signal doubling. In particular, residues 112 and 117 showed no change in their amide nitrogen/proton chemical shifts. No signal doubling was observed in any atom other than the amide nitrogen and amide proton. This is expected, since amide nitrogens and amide protons are known to be much more sensitive to the environment than the side chain atoms.
|
, HN) were first established using data from the HNCACB (Wittekind and Mueller 1993) and CBCACONNH (Grzesiek and Bax 1992) experiments. Chemical shifts of the side chain 13C and 1H were then assigned using data from the HCCHTOCSY (Sattler et al. 1995), HCCONH (Lyons and Montelione 1993), and CCONH (Grzesiek et al. 1993) experiments. Once the proton chemical shifts were known unambiguously, data from 3D aliphatic and aromatic 13C-edited NOESY as well as 3D 15N-edited NOESY experiments were automatically analyzed using CYANA. Distance and dihedral restraints derived from the data were then used to generate structures using both CYANA and Xplor-NIH. For more details on this procedure, please consult the experimental section. Of all the structures solved in this study, none had a significantly different N-terminal domain structure compared to the wild-type chicken cTnC. This is not surprising, since the N-terminal domain sequence of the mutants (residues 1–90) shares 100% identity with the N-terminal domain sequence of wild-type chicken cTnC. Together with the fact that the two domains of cTnC do not interact in solution, substitutions in the C-terminal domain should have no effect on the N-terminal domain. Therefore, the following analysis on the structures of the mutants will focus entirely on the C-terminal domain, with the emphasis on the orientations of the indole ring in each structure.
Structures of F104W and F104(5fW) mutants
The Phe-to-Trp mutation at position 104 resulted in a change of ~2.5 ppm and 2.6 ppm in the residue’s amide nitrogen and carbon chemical shifts, respectively (1.5 ppm and 2.2 ppm after correcting for the change in residue type and assuming random structure [Wishart et al. 1991]) compared to the wild-type cTnC. The replacement of L-Trp with 5fW in the mutant did not induce additional perturbation judging by the small amide proton/nitrogen chemical shift differences between the F104W and the F104(5fW) mutants. The only significant amide nitrogen chemical shift changes induced by the incorporation of 5fW occurred to a few residues on the F helix (residues 119, 120, 121, and 124). This is not surprising, since the side chains of these residues are very close to atom H
3 of the indole ring. However, no significant chemical shift changes were seen with any other atoms in these residues.
In this study CYANA was used to automatically assign and calibrate the signals in all 3D NOESY spectra. For the F104W mutant, CYANA was able to find a total of 1545 unambiguous distance restraints with 183 of them long-range restraints. Structure calculation carried out using these restraints and the dihedral angle restraints derived using TALOS revealed that the structure of the F104W mutant is very similar to that of the wild-type cTnC (Fig. 5A). The 20 lowest energy structures from a total of 50 structures calculated superimposed with a backbone RMSD of 0.80 ± 0.14 Å while the backbone of the well-defined region superimposed with an RMSD of 0.62 ± 0.11 Å (see Table 1 for definition of well-defined region). A total of 89.8% of all backbone dihedral angles are in the most favored region of the Ramachandran map (Table 1). The major difference between the wild-type cTnC and the F104W mutant lies in the inter-helical angles between helices E and F, as well as helices G and H, with the F104W mutant being slightly more "open" compared to the wild type (Table 2; Fig. 6A). The situation for the F104(5fW) mutant is similar. CYANA found 1272 unambiguous distance restraints for this protein, the least of all mutants. Out of all the restraints, 122 are long-range restraints. Despite the smaller number of restraints, the backbone RMSD of 20 lowest energy structures superimposed is 0.94 ± 0.15 Å, with the well-defined region having an RMSD of 0.61 ± 0.12 Å, almost as good as the F104W structure. In fact, the structure of the F104(5fW) mutant has slightly better agreement with the wild-type cTnC than does the F104W mutant (Fig. 6B).
|
|
|
|
2 angle for the Trp residue. From the 3D aromatic 13C-edited NOESY data, a number of unambiguous distance restraints involving indole ring protons were obtained (Fig. 7A). Structure calculations showed all restraints are consistent with each other and a single, well-defined orientation of the indole ring, with a side-chain 2 dihedral angle of ~96° (Table 3), was able to satisfy all the restraints. The orientation of the ring is similar but not identical to that of the wild-type F104 side chain. The H3 atom is pointed down toward the F helix while the H
1 points toward the side chains of residues 97 and 100. Interestingly, even if no distance restraints are used for any indole ring protons but atoms H3 and H1, the ring orientation is still well defined and agrees with the ring orientation obtained with a full set of restraints. This fact turned out to be relevant in defining the 5fW indole orientation in the F104(5fW) mutant. Owing to the lack of 13C-labeling on the ring, we can only obtain NOE data for the fluorine and the H1 proton on the indole ring. Even though signal intensities in the FHOESY experiment are lower compared to the 3D 13C-edited NOESY experiment, a combination of the directly detected 1H dimension and lack of spectral overlap in the 19F dimension means the FHOESY data has higher resolution and is easier to analyze and less ambiguous compared to the 13C-edited NOESY data. The mixing time used in the FHOESY experiment is also kept short to minimize 1H-1H spin diffusion artifacts in the experiment. Comparing the NOE contact pattern made by the proton H3 on the indole ring with the 19F atom’s heteronuclear NOE contact pattern, we see little, if any, difference (Fig. 7B). This is only possible if the orientation of the 5-fluoroindole ring in the F104 (5fW)mutant is similar to that of the indole ring in the F104W mutant. Just as in the F104W case, specifying only distance restraints involving the 19F and H1 is enough to produce a well-defined ring orientation that is very similar to that of the F104W mutant. In fact, although we were able to obtain the chemical shift assignments for protons H
2 and H3 using the FHOESY experiment, adding distance restraints from these two atoms did not improve the definition of the 5-fluoroindole ring orientation.
|
|
carbon chemical shifts. Since there is no sign of secondary structure change at either position, the difference in mutation-induced chemical shift changes between the two mutation sites could be purely due to differences in dynamics or exposure to solvents at the two sites. The side chain of the F153 residue is a more integrated part of the hydrophobic core than that of F104. Replacing the Phe with the bulky Trp, and subsequently the bulkier 5fW, could be potentially disruptive to the overall packing of the core. However, the physiological study done by Chandra et al. (1994) showed that the mutant has the same physiological activity as the wild type, while the Ca2+ affinity is decreased only by half. Inspection of the {1H, 15N}-HSQC of either the F153W mutant or the F153(5fW) mutant showed that both have well-folded C-terminal domains.
The 3D aliphatic 13C-edited NOESY spectrum of the F153W mutant was collected on the INOVA 800 MHz spectrometer. The sensitivity and high resolution of the spectrum are reflected in the large number of distance restraints CYANA was able to derive from the data. A total of 2111 unambiguous restraints were found, with 273 of them being long-range restraints, the most of all mutants. The 20 lowest energy structures, of a total of 50 calculated, superimposed with a backbone RMSD of 0.82 ± 0.10 Å , 0.62 ± 0.12 Å for the well-defined regions. Almost 92% of all backbone dihedral angles are found in the most favored region of the Ramachandran map. The inter helical angles between helices E and F, and helices G and H of the F153W mutant also agreed well with the wild-type C-terminal domain structure. The backbone RMSD between the F153W mutant and the wild-type C-terminal domain structure is only 0.94Å for the well-defined regions. For the F153(5fW) mutant, a total of 1709 unambiguous distance restraints were found, of which 216 were long-range restraints. The RMSD between the 20 lowest energy structures (out of 50 calculated) after superimposition is 0.91 ± 0.22 Å . The well-defined regions in these structures superimposed with an RMSD of 0.60 ± 0.17Å . Just like the other mutants, almost 90% of all backbone dihedral angles are found in the most favored region of the Ramachandran map. There is a slight difference in the interhelical angle of helices E and F between the F153(5fW) mutant and the wild-type cTnC (Table 2), with the F153(5fW) mutant being slightly more "open" than the wild-type cTnC. As a result, the two structures superimposed with a slightly higher backbone RMSD for well-defined regions.
The hydrophobic packing of the F153W and the F153(5W) mutants are little disturbed by the replacement of the phenyl ring with the bulkier indole ring. Super-imposition of the C-terminal domain of the F153W mutant with the C-terminal domain of the wild-type protein revealed that the conformations of all the hydrophobic side chains in the core are in good agreement between the two structures. For the F153W mutant, the 3D aromatic 13C NOESY allowed the extraction of close to 30 unambiguous NOEs for the six protons on the indole ring. Qualitative examination of each indole proton’s NOE contact patterns alone was enough to define the orientation of the ring unequivocally (Fig. 8A). The orientation of the indole ring is such that the amide proton on the ring points away from the hydrophobic pocket and toward the E helix, making several contacts with the side chains of Leu 100 and Leu 97. The H3 on the indole ring points deep into the hydrophobic pocket, contacting side chains of the Ile residues in the middle of the short anti-parallel -sheet that forms the back of the hydrophobic pocket. The 2 angle for the residue is ~95° (Table 3). The normal of the indole ring plane is roughly parallel to the helical axis of helix G.
|
3 proton in the F153W mutant (Fig. 8B). Subsequent structure refinements done using the distance restraints from the heteronuclear experiment revealed an orientation of the fluoroindole ring that is almost identical to the orientation of the indole ring in the F153W mutant.
Owing to the larger number of restraints that were found for the 19F atom, the fluoroindole ring orientation of the F153(5fW) mutant is much better defined compared to the F104(5fW) mutant. The average value of the 2 angle for the 5fW is ~106° with a standard deviation of only 6° (Table 3).
|
Discussion |
|---|
|
TOPAbstractIntroductionResultsDiscussionMaterials and methodsReferences |
|---|
Potential use of the mutants as probes in in situ NMR
As the determination of high-resolution protein structures becomes easier and information in structural biology begins to accumulate, many scientists have come to the realization that knowing the structure of a single protein is no longer adequate in many situations. Far more insight can be gained by understanding how proteins interact with each other. Thus, researchers are switching their focus to investigating the behavior of proteins in complex with other macromolecules in their in situ environment. This is especially true in the field of muscle regulation research, where muscle regulatory proteins exist only as a part of the thin filament. Owing to the sizes of these complexes, conventional structural biology techniques usually are not applicable. As a result, a new approach is often used to investigate these large complexes: Structures of proteins composing the complex are first solved individually, then by obtaining a few crucial pieces of information regarding their relative distances and orientations to each other in the complex, the structure of the entire complex is constructed (Corrie et al. 1999). This is also our strategy for investigating movements of TnC during muscle activation. Utilizing the orientation dependency of 19F CSA tensor, information on the orientation of TnC on the filament during muscle contraction and relaxation can be elucidated if the orientation of the C–F bond vector correlates well with the orientation of the protein. Our study showed that, in both mutants, the fluoroindole rings have well-defined orientations; therefore they should be acceptable probes for the orientation of TnC. The 5fW in the F104(5fW) mutant possessed an average 2 of 83 ± 22°, while the 5fW in the F153(5fW) mutant had an average 2 of –106 ± 6°. The fact that the 5fW in the F104(5fW) mutant had larger 1 and 2 dihedral angle spreads than the F153(5fW) mutant could be a reflection of its more dynamic indole ring motions. Molecular dynamics simulations conducted for both mutants showed that the indole ring in the F104(5fW) mutant undergoes ring oscillations of considerably larger magnitude than that of the F153(5fW) mutant, which experiences almost no movement (X. Wang and B.D. Sykes, unpubl.). However, the motions experienced by the F104(5fW) indole ring are at a much higher frequency than the NMR timescale and thus will have no bearing on the results of the NMR experiments. Owing to the inherent degeneracy in the process of orientation determination using the CSA tensor, two or more CF bond vectors, all of which must have different orientation in the protein, will be needed to solve the protein’s orientation. The CF bond vectors in the two mutants are at an average angle of 24° with each other. This implies, with the exceptions of a few situations, that the two vectors will reveal nondegenerate information, an important detail in the case where protein orientation was to be determined using axially symmetric chemical shift tensors.
Structures of the F104 and F153 mutants
In all mutants, the secondary structures are well preserved. The major difference between the mutant structures and the wild-type structure lies in the difference in interhelical angles. Determination of interhelical angles using NOE derived distance restraints alone is not always a robust process (Chou et al. 2001). The presence of one or two long-range restraints is enough to change the interhelical angle by 20° or more. For this reason, some ambiguous NOE signals from long-range contacts were manually assigned before applying the CYANA automatic assignment protocol. It is also evident from the {1H, 15N}-HSQC spectra as well as the presence of important long-range NOE signals in the 3D 13C-edited NOESY spectra that all mutants are well-folded and adopt the "open" conformation of the wild-type cTnC C-terminal domain. Given the above information, we believe the differences in the interhelical angles between the F104W mutant and the wild-type structure, as well as that between the F153(5fW) mutant and the wild-type structure, are not significant enough to be an indicator of major structural change in the mutants. Since the amount of solvent-accessible hydrophobic surface area is a reliable measure of the degree of "openness" in cTnC (Sia et al. 1997), we also examined sizes of solvent-accessible hydrophobic surface area on the C-terminal domain of each mutant structure using the program VADAR (Willard et al. 2003). For all mutants, the average amount of solvent-exposed hydrophobic area is not significantly different than that of the wild type, indicating no significant change in the openness of the structure is induced by Trp or 5fW.
Signal doubling in the F153(5fW) mutant
Fluorotryptophan incorporation in a protein is never 100%, even using the most carefully controlled protocols. The exact percentage of incorporation can be estimated using a number of methods (Campos-Olivas et al. 2002; Valencia et al. 2003). Our study shows that the {1H, 15N}-HSQC can be used as a measure of the fluorine incorporation efficiency. In the case of the F153 (5fW) mutant, we were able to calculate the ratio between the intensities of the peak produced by the fluorinated isoform and the peak produced by the nonfluorinated isoform. Of the 12 residues showing the signal-doubling phenomenon, 10 residues had clear and nonoverlapping peaks. For all 10 peaks, the intensity ratio of the strong peaks and the weak peaks was approximately 4:1. This implies that about 80% of protein is 5fW-labeled, a number that agrees well with the estimation done using fluorescence measurements.
|
Materials and methods |
|---|
|
TOPAbstractIntroductionResultsDiscussionMaterials and methodsReferences |
|---|
Expression of the protein and incorporation of fluorotryptophan into F153W and F104W mutants
In this study, a combination of unlabeled and 15N/13C-labeled proteins containing either L-tryptophan or 5fW was used. Proteins containing normal tryptophan were expressed in E. coli strain BL21(DE3)pLysS using the procedure described previously (Sia et al. 1997). Incorporation of 5fW into BL21 expressed proteins was promoted by first inhibiting the aromatic amino acid synthesis using 1 g/L glyphosate, an inhibitor of the 5-enolpyruvylshikimic acid-3-phosphate synthase (Steinrucken and Amrhein 1980; Comai et al. 1983). The needed aromatic amino acids were then supplemented exogenously. When the culture reached an OD600 of 0.8–1.0, 150 mg of D,L-5fW (Sigma, F0896) was added. Half an hour later, the culture was induced using 120 mg/L IPTG. The induction was carried out for 3 h before harvesting the culture. To achieve a high fluorine incorporation efficiency, 5fW-containing mutants were also expressed in E. coli Trp auxotroph strain W3110TrpA33 that was infected with phage DE3 and transformed with plasmid pLysS (Valencia et al. 2003). In this procedure, 1 L of minimal M9 medium containing 100 mg of L-Trp was seeded with 10 mL of overnight culture grown in LB at 37°C and grown at the same temperature. When the cells reached an OD600 of ~1.0, they were spun down at 4°C for 10 min with a 4400g centrifugal force and the pellet was resuspended in fresh M9 media containing no Trp and grown for 30 min at 37°C; 150 mg/L of D,L-5fW was then added at the end of the 30 min. The cells were grown for another 30 min before induction was initiated with 120 mg/L of IPTG. The culture was harvested 3 h after the induction.
Purification of the mutants
The harvested cells were French pressed and the supernatant was applied to a DEAE-Sephadex (A25) column equilibrated with a buffer of 50 mM Tris, 0.1 M NaCl (pH 8.0). The column was washed extensively with the equilibration buffer at a flow rate of 0.6 mL/min before the protein was eluted with a 0.1–0.45 M NaCl salt gradient. The fractions containing the expressed proteins were pooled and lyophilized and redissolved in approximately 30 mL of a buffer containing 5 mM CaCl2, 50 mM KCl, 50 mM Tris, and 1 mM MgCl (pH 7.5). The sample solution was loaded onto a phenyl-Sepharose column equilibrated with the same buffer. The column was washed with the same buffer for 5–8 h at a flow rate of 0.6 mL/min before an elution buffer containing 50 mM Tris and 1 mM EDTA (pH 7.5) was applied. The intact TnC was eluted as a single sharp peak.
Sample preparation and calcium titrations
All samples used for data collection contained 1.0–1.5 mM of protein, 100 mM KCl, 10 mM imidazole, 1.5 mM DSS, 10% D2O, and 0.05% NaN3 (pH 6.7). If a Ca2+-saturated sample was required, CaCl2 stock solution was added until Ca2+ concentration had reached 10 mM. To evaluate 19F’s effect on the protein, qualitative Ca2+ titrations of the protein were performed as follows: Either 0.5 µL or 1 µL aliquot of 100 mM CaCl2 was added to a 500-µL sample containing ~1 mM of apo protein (pH 6.7), and 100 mM KCl, 10 mM imidazole, 0.2 mM DSS. The titration was monitored with either {1H, 15N}- HSQC or one-dimensional 19F spectroscopy, or both. The titration was stopped when the chemical shift and the line shape of the signal stopped changing.
Data acquisition and analysis
All data were acquired either on a Varian Inova 500 or an Inova 600 spectrometer at 30°C except the 3D aliphatic 13C-edited NOESY for the F153W, which was collected on an Inova 800 spectrometer equipped with a 5-mm triple resonance probe with triple axis gradients. The 500- and 600-MHz spectrometers are equipped with 5-mm triple resonance probes with Z-axis pulse field gradients. With the exception of the 19F-1H heteronuclear NOE experiment, all pulse sequences were obtained from the BioPack pulse sequence package distributed by Varian Inc. A mixing time of 75 msec was used for all 3D NOESY experiments used to derive distance restraints for structure calculations. All data were processed using the software package NMR Pipe (Delaglio et al. 1995) and analyzed using NMR View (Johnson and Blevins 1994). Backbone assignments were done using data from HNCO, HNCACB (Wittekind and Mueller 1993), and CBCACONNH (Grzesiek and Bax 1992) spectra. The software package Smartnotebook (Slupsky et al. 2003) was used to expedite the process. Since most secondary structures in cTnC are -helices, dNN contacts from the 3D 15N-edited NOESY experiment were also used to establish sequential assignments along the backbone. Side-chain assignments were done using data from conventional HCCHTOCSY (Sattler et al. 1995), HCCONH (Lyons and Montelione 1993), and CCONH (Grzesiek et al. 1993) experiments. The -methylene protons and the methyl groups on Val and Leu in the mutants were not stereospecifically assigned.
Structural determination
All proton distance restraints used in determining structures were derived from 3D aliphatic and aromatic 13C-edited NOESY and 15N edited NOESY automatically using CYANA version 2 (Guntert 2004). NMR View was first used to automatically generate a peak list for each NOESY spectrum. The lists were then manually edited to remove spectral artefacts and noise. Some unambiguous peaks representing long-range contacts were also assigned manually before the lists were supplied to CYANA. Dihedral angle restraints obtained using TALOS (Cornilescu et al. 1999) were also included in the calculation. The automatic NOE assignments in CYANA were carried out using a tolerance of 0.025 ppm for the directly detected dimension, 0.04 ppm for the indirectly detected proton dimension, and 0.5 ppm for the 13C or 15N dimension. No distance restraints were allowed to be shorter than 1.8 Å or longer than 6 Å . Distance calibration in CYANA was done using a simple but robust scheme: Peaks with median intensity were assumed to represent a distance of 3.1 Å . Distance restraints from peaks with other intensities were extrapolated from this standard using the R6 distance relationship commonly used for calibrating NOE distances. Twenty-four crystal structure-derived distance restraints between Ca2+ ions and chelating side chains were also used in the structure calculation. The automatic NOE assignment module in CYANA used an eight-round iterative structure refinement protocol in which 100 random structures were calculated in the first round with no restraints applied and 50 structures were calculated for each successive round using NOE assignments based on previous round’s structures. After eight rounds, 20 minimum energy structures were chosen from the 50 calculated as the final ensemble. Because the two globular domains of cTnC were known to be independent of each other, the default NOE assignment script was modified to eliminate any possible NOE assignments connecting C-terminal domain atoms with N-terminal domain atoms. The automatic backbone dihedral angle estimation feature of CYANA was not used, since experimentally determined backbone dihedral angles were available. The set of CYANA-optimized structures was further refined using a water-refinement protocol (Linge et al. 2003) implemented under the software package Xplor-NIH (Schwieters et al. 2003). For CYANA calculations, the force field parameter and topology for 5fW were defined exactly the same way as those for L-Trp. The force field parameter and topology for 5fW in Xplor-NIH calculations were adapted from those for LTrp by changing atom type for atom HZ3 to fluorine and modifying the charges on HZ3 and CZ3 to –0.25 and 0.25, respectively. Bond angles and bond lengths, as well as improper dihedral angles concerning these two atoms were not changed.
19F-1H heteronuclear NOE determination
The distance restraints between 19F and 1H were determined using data from a 19F-1H heteronuclear NOESY experiment (FHOESY), whose pulse sequence was adapted from a pulse sequence designed to measure heteronuclear NOEs (Rinaldi 1983; Kover and Batta 1987) for use with 19F. In particular, 19F was the indirectly detected nucleus and 1H the observed nucleus. No 19F decoupling was applied during the acquisition. Mixing times of 150 msec and 75 msec were used for the F104(5fW) and the F153(5fW) mutants, respectively.
|
Acknowledgments |
|---|
|
References |
|---|
|
TOPAbstractIntroductionResultsDiscussionMaterials and methodsReferences |
|---|
Anderluh, G., Razpotnik, A., Podlesek, Z., Macek, P., Separovic, F., and Norton, R.S. 2005. Interaction of the eukaryotic pore-forming cytolysin equinatoxin II with model membranes: 19F NMR studies. J. Mol. Biol. 347: 27–39.[CrossRef][Medline]
Browne, D.R., Kenyon, G.L., and Hegeman, G.D. 1970. Incorporation of monoflurotryptophans into protein during the growth of Escherichia coli. Biochem. Biophys. Res. Commun. 39: 13–19.[CrossRef][Medline]
Campos-Olivas, R., Aziz, R., Helms, G.L., Evans, J.N., and Gronenborn, A.M. 2002. Placement of 19F into the center of GB1: Effects on structure and stability. FEBS Lett. 517: 55–60.[CrossRef][Medline]
Chandra, M., McCubbin, W.D., Oikawa, K., Kay, C.M., and Smillie, L.B. 1994. Ca2+, Mg2+, and troponin I inhibitory peptide binding to a Phe- 154 to Trp mutant of chicken skeletal muscle troponin C. Biochemistry 33: 2961–2969.[CrossRef][Medline]
Chou, J.J., Li, S., Klee, C.B., and Bax, A. 2001. Solution structure of Ca2+-calmodulin reveals flexible hand-like properties of its domains. Nat. Struct. Biol. 8: 990–997.[CrossRef][Medline]
Comai, L., Sen, L.C., and Stalker, D.M. 1983. An altered aroa geneproduct confers resistance to the herbicide glyphosate. Science 221: 370–371.
Cornilescu, G., Delaglio, F., and Bax, A. 1999. Protein backbone angle restraints from searching a database for chemical shift and sequence homology. J. Biomol. NMR 13: 289–302.[CrossRef][Medline]
Corrie, J.E., Brandmeier, B.D., Ferguson, R.E., Trentham, D.R., Kendrick-Jones, J., Hopkins, S.C., van der Heide, U.A., Goldman, Y.E., Sabido-David, C., Dale, R.E., et al. 1999. Dynamic measurement of myosin light-chain-domain tilt and twist in muscle contraction. Nature 400: 425–430.[CrossRef][Medline]
Delaglio, F., Grzesiek, S., Vuister, G.W., Zhu, G., Pfeifer, J., and Bax, A. 1995. NMRPipe: A multi-dimensional spectral processing system based on UNIX pipes. J. Biomol. NMR 6: 277–293.[Medline]
Eccleston, J.F., Molloy, D.P., Hinds, M.G., King, R.W., and Feeney, J. 1993. Conformational differences between complexes of elongation factor Tu studied 19F-NMR spectroscopy. Eur. J. Biochem. 218: 1041–1047.[Medline]
Grage, S.L., Wang, J., Cross, T.A., and Ulrich, A.S. 2002. Solid-state 19FNMR analysis of 19F-labeled tryptophan in gramicidin A in oriented membranes. Biophys. J. 83: 3336–3350.
Grzesiek, S. and Bax, A. 1992. Correlating backbone amide and side chain resonances in larger proteins by multiple relayed triple resonance NMR. J. Am. Chem. Soc. 114: 6291–6293.[CrossRef]
Grzesiek, S., Anglister, J., and Bax, A. 1993. Correlation of backbone amide and aliphatic side-chain resonances in C-13/N-15-enriched proteins by isotropic mixing of C-13 magnetization. J. Magn. Reson. Ser. B 101: 114–119.
Guntert, P. 2004. Automated NMR structure calculation with CYANA. Methods Mol. Biol. 278: 353–378.[Medline]
Ho, S.N., Hunt, H.D., Horton, R.M., Pullen, J.K., and Pease, L.R. 1989. Site-directed mutagenesis by overlap extension using the polymerase chain reaction. Gene 77: 51–59.[CrossRef][Medline]
Hull, W.E. and Sykes, B.D. 1974. Fluorotyrosine alkaline phosphatase. 19F nuclear magnetic resonance relaxation times and molecular motion of the individual fluorotyrosines. Biochemistry 13: 3431–3437.[CrossRef][Medline]
———. 1975. Fluorotyrosine alkaline phosphatase: Internal mobility of individual tyrosines and the role of chemical shift anisotropy as a 19F nuclear spin relaxation mechanism in proteins. J. Mol. Biol. 98: 121–153.[CrossRef][Medline]
———. 1976. Fluorine-19 nuclear magnetic resonance study of fluorotyrosine alkaline phosphatase: The influence of zinc on protein structure and a conformational change induced by phosphate binding. Biochemistry 15: 1535–1546.[CrossRef][Medline]
Johnson, B.A. and Blevins, R.A. 1994. NMRView: A computer program for the visualization and analysis of NMR data. J. Biomol. NMR 4: 603–614.[CrossRef]
Kover, K.E. and Batta, G. 1987. Theoretical and practical aspects of one-dimensional and two-dimensional heteronuclear Overhauser experiments and selective C-13 T1-determinations of heteronuclear distances. Prog. Nucl. Magn. Reson. Spectrosc. 19: 223–266.
Lau, E.Y. and Gerig, J.T. 1997. Effects of fluorine substitution on the structure and dynamics of complexes of dihydrofolate reductase (Escherichia coli). Biophys. J. 73: 1579–1592.
Li, M.X., Saude, E.J., Wang, X., Pearlstone, J.R., Smillie, L.B., and Sykes, B.D. 2002. Kinetic studies of calcium and cardiac troponin I peptide binding to human cardiac troponin C using NMR spectroscopy. Eur. Biophys. J. 31: 245–256.[CrossRef][Medline]
Linge, J.P., Williams, M.A., Spronk, C.A., Bonvin, A.M., and Nilges, M. 2003. Refinement of protein structures in explicit solvent. Proteins 50: 496–506.[CrossRef][Medline]
Lyons, B.A. and Montelione, G.T. 1993. An HCCNH triple-resonance experiment using C-13 isotropic mixing for correlating backbone amide and side-chain aliphatic resonances in isotopically enriched proteins. J. Magn. Reson. Ser. B 101: 206–209.
Pearlstone, J.R., Chandra, M., Sorenson, M.M., and Smillie, L.B. 2000. Biological function and site II Ca2+-induced opening of the regulatory domain of skeletal troponin C are impaired by invariant site I or II Glumutations. J. Biol. Chem. 275: 35106–35115.
Pratt, E.A. and Ho, C. 1975. Incorporation of fluorotryptophans into proteins of proteins of Escherichia coli. Biochemistry 14: 3035–3040.[CrossRef][Medline]
Rinaldi, P.L. 1983. Heteronuclear 2D-NOE spectroscopy. J. Am. Chem. Soc. 105: 5167–5168.[CrossRef]
Salopek-Sondi, B., Vaughan, M.D., Skeels, M.C., Honek, J.F., and Luck, L.A. 2003. 19F NMR studies of the leucine-isoleucine-valine binding protein: Evidence that a closed conformation exists in solution. J. Biomol. Struct. Dyn. 21: 235–246.[Medline]
Sattler, M., Schwendinger, M.G., Schleucher, J., and Griesinger, C. 1995. Novel strategies for sensitivity enhancement in heteronuclear multi-dimensional NMR experiments employing pulsed-field gradients. J. Biomol. NMR 6: 11–22.
Schwieters, C.D., Kuszewski, J.J., Tjandra, N., and Clore, G.M. 2003. The Xplor-NIH NMR molecular structure determination package. J. Magn. Reson. 160: 65–73.[CrossRef][Medline]
Sia, S.K., Li, M.X., Spyracopoulos, L., Gagné, S.M., Liu, W., Putkey, J.A., and Sykes, B.D. 1997. NMR structure of cardiac troponin C reveals an unexpected closed regulatory domain. J. Biol. Chem. 272: 18216–18221.
Sixl, F., King, R.W., Bracken, M., and Feeney, J. 1990. 19F-NMR studies of ligand binding to 5-fluorotryptophan- and 3-fluorotyrosine-containing cyclic AMP receptor protein from Escherichia coli. Biochem. J. 266: 545–552.[Medline]
Slupsky, C.M., Boyko, R.F., Booth, V.K., and Sykes, B.D. 2003. Smartnotebook: A semi-automated approach to protein sequential NMR resonance assignments. J. Biomol. NMR 27: 313–321.[CrossRef][Medline]
Steinrucken, H.C. and Amrhein, N. 1980. The herbicide glyphosate is a potent inhibitor of 5-enolpyruvyl-shikimic-acid 3-phosphate synthase. Biochem. Biophys. Res. Commun. 94: 1207–1212.[CrossRef][Medline]
Sykes, B.D. and Weiner, J.H. 1980. Biosynthesis and 19F characterization of fluoroaminoacid containing proteins. In Magnetic resonance in biology (ed. J.S. Cohen), pp. 171–196. John Wiley & Sons, New York.
Trigo-Gonzalez, G., Racher, K., Burtnick, L., and Borgford, T. 1992. A comparative spectroscopic study of tryptophan probes engineered into high- and low-affinity domains of recombinant chicken troponin C. Biochemistry 31: 7009–7015.[CrossRef][Medline]
Ulrich, A.S. 2005. Solid state F-19 NMR methods for studying biomembranes. Prog. Nucl. Magn. Reson. Spectrosc. 46: 1–21.[CrossRef]
Valencia, F.F., Paulucci, A.A., Quaggio, R.B., Da Silva, A.C., Farah, C.S., and Reinach, F.C. 2003. Parallel measurement of Ca2+ binding and fluorescence emission upon Ca2+ titration of recombinant skeletal muscle troponin C. Measurement of sequential calcium binding to the regulatory sites. J. Biol. Chem. 278: 11007–11014.
Willard, L., Ranjan, A., Zhang, H., Monzavi, H., Boyko, R.F., Sykes, B.D., and Wishart, D.S. 2003. VADAR: A Web server for quantitative evaluation of protein structure quality. Nucleic Acids Res. 31: 3316–3319.
Wishart, D.S., Sykes, B.D., and Richards, F.M. 1991. Relationship between nuclear magnetic resonance chemical shift and protein secondary structure. J. Mol. Biol. 222: 311–333.[CrossRef][Medline]
Wittekind, M. and Mueller, L. 1993. HNCACB, a high-sensitivity 3D NMR experiment to correlate amide–proton and nitrogen resonances with the - and
