| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 279, Issue 18, 18799-18805, April 30, 2004
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
¶**
From the
School of Biomedical Sciences, Institute for Molecular Bioscience, and ||School of Molecular and Microbial Sciences, University of Queensland, Brisbane, Queensland 4072, Australia
Received for publication, November 10, 2003 , and in revised form, January 28, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-estradiol, thyroid hormones, and dehydroepiandrosterone), chemical carcinogens, bile acids, and neurotransmitters. These substrates may undergo sulfonation directly as in phase II metabolism (e.g. minoxidil, paracetamol) or following phase I metabolism after introduction of a functional hydroxyl group (e.g. N-hydroxyl arylamines). The sulfonate donor for these reactions is 3'-phosphoadenosine 5'-phosphosulfate (PAPS; Refs. 1–3). Sulfonation is generally considered to be a detoxification pathway for a broad range of compounds, because conjugation with a sulfonate (SO3-1) renders substrates more hydrophilic and facilitates their excretion. However, evidence has emerged in recent years that this reaction can lead to formation of highly reactive intermediates (e.g. with N-hydroxy heterocyclic and aromatic arylamines) that can bind to DNA resulting in mutagenicity and carcinogenicity (4–6). Therefore, various studies have investigated its role in human cancer and in particular polymorphisms in SULT1A1 have been implicated in the predisposition to lung cancer (7), age of onset in breast cancer (8), and increased risk of breast cancer (9, 10).
Fifty-six cytosolic SULT cDNAs have been cloned and characterized from bacteria to human (11–13). SULT1, SULT2, SULT3, SULT4, and SULT5 gene families have been identified in mammals. Of these, only SULT1, SULT2, and SULT4 have been identified in humans. These include seven members of the SULT1 family (SULT1A1, SULT1A2, SULT1A3, SULT1B1, SULT1C1, SULT1C2, and SULT1E1), two members of the SULT2 family (SULT2A1 and SULT2B), and one member of the SULT4 family (SULT4A1). Although each SULT has distinct substrate preferences, the range of molecules accepted can be broad and may be overlapping. In general, the SULT1 family catalyzes sulfonation of small phenols, monoamine neurotransmitters (e.g. dopamine), thyroid hormones, and estrogens, and the SULT2 family sulfonates steroid hormones. To date, neither an endogenous nor an exogenous substrate has been found for SULT4 A1.
Although there has been an increase in structure function studies as a result of an expansion of SULT x-ray crystal structures published, substrate binding and inhibition mechanisms are still far from clear. The crystal structures of mouse estrogen SULT1E1 (14), human SULT1A3 (15, 16), SULT domain of the human membrane bound heparan sulfate N-deacetylase/N-sulfotransferase 1 (17), human hydroxysteroid SULT2A1 (18, 19), and human estrogen SULT1E1 (20); in addition, site-directed mutagenesis of critical amino acids in these enzymes has provided information on catalytic mechanism of these enzymes. Recently, this laboratory also reported the structure of human SULT1A1, crystallized in the presence of 3'-phosphoadenosine 5'-phosphate (PAP) with two p-nitrophenol (pNP) molecules bound in the active site (21). This structure is similar to the other SULT enzymes as it incorporates a core PAP binding site. However, it was the first structure to show two molecules of the substrate bound in the active site.
Substrate inhibition is characteristic of many SULT enzymes and has been reported previously with human SULTA1 and SULT1A3 with pNP and dopamine, respectively (22–24). The presence of two substrate molecules in the SULT1A1 structure was postulated to explain the slight cooperativity observed at low concentrations of substrate and the inhibition observed at higher concentrations of pNP. In previous work (21), we proposed that inhibition is caused by impeded catalysis when both binding sites are occupied. In SULT1A1, substrate inhibition is evident when the pNP is used as the acceptor molecule but not when dopamine is the substrate. To investigate this further, we modeled dopamine into the substrate-binding pocket of SULT1A1 to show that two dopamine molecules could fit. Dopamine is the preferred substrate of SULT1A3 and our initial resolution of SULT1A3 structure showed large stretches of disorder that account for about 25% of its structure, which may in part be caused by lack of substrate binding (15). Therefore, we have also modeled the substrate-binding pocket of SULT1A3 to investigate dopamine binding and to determine how substrate inhibition takes place in this enzyme. Finally, site-directed mutagenesis was performed on residues interacting with the substrate molecule at the active site of SULT1A1 indicated that the flexible residue Phe-247 is important for substrate inhibition by dopamine.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Site-Directed Mutagenesis—Double-stranded primers encoding a single amino acid mutation in the SULT1A1 gene were used to obtain SULT1A1 mutants. Twenty cycles of PCR were carried out with denaturation at 94 °C for 30 s, annealing at 55 °C for 1 min, and extension at 68 °C for 12 min. The reaction contained 50 ng of DNA template, primers, 200 µM concentrations of each of the four deoxynucleotide triphosphates (dNTPs), 1x PfuTurbo buffer, and 2.5 units of PfuTurbo DNA polymerase. The PCR product was treated with DpnI for 1 h at 37 °C followed by transformation of BL21 (DE3) Escherichia coli cells. The transformed bacteria were used to amplify the plasmid for sequencing as well as for protein expression of the SULT1A1 mutants. The mutated cDNA sequences were confirmed by automated sequencing carried out by the Australian Genome Research Facility (Brisbane, Australia) with the ABI Prism dye terminator cycle sequencing ready reactions (PerkinElmer Life Sciences) using the T7-promoter and terminator universal primers.
Expression, Purification, and Enzyme Activity—Expression of the recombinant proteins was performed as described previously (15, 25, 26). The protein was purified utilizing the hexahistadine tag on the N terminus of the SULT1A1 and SULT1A3 cDNAs by TALON cobalt affinity (Clontech Laboratories) chromatography. Protein purity was estimated by SDS-polyacrylamide gel electrophoresis and only fractions with the highest purity were used for enzyme assay. The purified protein was mixed with glycerol to a final concentration of 20% (v/v) and stored at -70 °C before use. Protein concentration was determined by the BCA method (Sigma).
Enzyme activity assays were performed using the method described by Foldes and Meek (27) as modified by Brix et al. (26). The reaction mixture consisted of 20 µM [35S]PAPS, 0.1 µg/ml SULT1A1, 10 mM potassium phosphate buffer, pH 7.0, and varying concentrations of the substrates in a final volume of 500 µl. Reactions were started by adding the enzyme to the reaction mixture and then incubated for 20 min at 37 °C. The reaction was terminated by adding 0.1 M barium hydroxide, 0.1 M barium acetate, and 0.1 M zinc sulfate, precipitating the enzyme and unreacted [35S]PAPS. Assays were performed in duplicate and the radioactivity of the sulfonated product was determined using a liquid scintillation counter (Tri-Carb 2500; PerkinElmer Life and Analytical Sciences). The incubation time and protein concentration used were chosen to be within the linear range for product formation.
Kinetic Data Analysis—Kinetic experiments were analyzed by nonlinear regression, fitting the appropriate equation to the data. This regression analysis was performed using the Grafit program, version 5 (Erithacus Software Limited, Surrey, UK).
Molecular Modeling—Modeling of substrates into the human SULT1A1/SULT1A3 substrate-binding pocket was performed using the docking program GOLD (Cambridge University) with 10 genetic algorithm runs and default parameters (28). The structure of SULT1A3 was modeled using the crystal structure of human SULT1A1 (Protein Data Bank code 1LS6 [PDB] ) and minimized in Crystallography and NMR System software (29). The quality of the model was assessed with Procheck (30).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
A computer model of SULT1A3 was constructed using the coordinates of SULT1A1 obtained from the Protein Data Bank (code 1LS6
[PDB]
). Both enzymes have 295 residues, and 21 residues differ between these two enzymes. In the SULT1A1 structure, these residues were mutated to the corresponding SULT1A3 residues, the model was energy minimized in Crystallography and NMR System software, and the validity of the model was assessed. The backbone residues in the SULT1A3 model can be superimposed on the SULT1A1 crystal structure (295 C
atoms with root-mean-square deviation 1.6 Å). PAPS was modeled into SULT1A3 based on the SULT1E1-PAPS structure (Protein Data Bank code 1HY3
[PDB]
; Ref. 20) and dopamine was docked using the GOLD program with default parameters. The model of SULT1A3 (Fig. 2c) with dopamine showed that two molecules of substrate are easily accommodated in the substrate-binding pocket; positioning of the second dopamine molecule was easier with a leucine residue at position 247.
Kinetic Analysis—Previously, we analyzed the substrate saturation and substrate inhibition kinetics of SULT1A1 toward pNP using an equation (Equation 1) derived from the observation that the enzyme structure contains two bound substrate molecules (21).
 | (Eq. 1) |
The difficulty with this equation is that it contains three kinetic constants (K1, K2, and K3) that have no obvious physical meaning. Only V
, the rate as [S] 
, relates directly to an observable quantity. We realized that this equation may be rewritten in terms of more normal constants,
 | (Eq. 2) |
, if Km was very small. Although this equation fits the data very well, the nonlinear regression analysis showed that there are extremely high parameter correlations between the pairs Vm and Km, Vm and Ki, and Km and Ki. This means that estimates of each of the parameters are unreliable because the values are highly dependent upon one another.
Further investigation showed that an empirical simplification could eliminate one parameter and these undesirable correlations, while maintaining a good fit. The revised equation is as follows.
 | (Eq. 3) |
has the same definition as in Equations 1 and 2. The third parameter, Ks, is a "substrate constant" with no exact physical meaning but that acts as a scaling factor for the substrate concentration axis. It can be used as an intermediate vehicle to calculate the more meaningful parameters Sp (the concentration of substrate that results in the peak rate) or Vp (the rate at [S] = Sp) using the relationships below.
 | (Eq. 4) |
 | (Eq. 5) |
Although it would be desirable to use Equation 2 for the data analysis, we believe that it is preferable to use an Equation that yields reliable parameter values even though it has an empirical element. Kinetic data that exhibit substrate inhibition were fitted using the combined Equations 3, 4, and 5 to obtain best fit values and standard errors for ks, Vp, and V. Where data showed no substrate inhibition, a reparameterized version of the Michaelis-Menten Equation 6 was used for the analysis to obtain best fit values and standard errors for ks and Vm.
 | (Eq. 6) |
To determine whether substrate inhibition involves the three residues that interact with both pNP molecules (Ile-89, Phe-247, and Val-148; Fig. 2a), in particular residue 247 as suggested by the computer modeling, we employed site-directed mutagenesis. Residues 247 and 148 were exchanged to their counterpart residues in human SULT1A3 (F247L, V148A), whereas residue 89 was converted to the smaller ala-nine residue. His-108 was not mutated because it is reported to function as a catalytic residue for sulfonation (31, 32) and is found in all human sulfotransferases.
SULT1A1-F247L—When modeled into the SULT1A1 structure, the leucine in the mutant of phenylalanine at 247 to leucine (F247L) does not form van der Waals interactions with the nitro groups of pNP1 and pNP2, but there is no change to the overall protein structure. Unlike wild-type human SULT1A1 with dopamine, this mutant showed substrate inhibition when dopamine was used as the acceptor molecule (Fig. 3a) and retained substrate inhibition with pNP (Fig. 3b). When this mutant was analyzed kinetically, it showed an increased affinity for pNP (Table I) and exhibited substrate inhibition at a slightly lower pNP concentration.
|
|
SULT1A1-V148A and SULT1A1-I89A—In the SULT1A1 crystal structure, the mutation of valine at 148 to alanine (V148A) removes van der Waals interaction to the nitro group of pNP1. The mutation of isoleucine at 89 to alanine (I89A) disrupts the binding of pNP2 to the enzyme. These mutants did not show any significant effect on substrate inhibition pattern with dopamine or pNP (Table I).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In our present study, the differing patterns of substrate inhibition relating to the substrates (pNP and dopamine) of SULT1A1 and SULT1A3 suggested to us that the substrate-binding pocket of SULT1A1 may not accommodate two molecules of dopamine or that binding of the second substrate molecule may have been impeded in some way. To investigate this, we modeled dopamine into the substrate-binding pocket of the human SULT1A1 structure. We found that the substrate-binding pocket of SULT1A1 could accommodate two molecules of dopamine, although one enzyme residue (Phe-247) made unfavorable interactions with the second molecule. This led us to hypothesize that this residue may be important in substrate inhibition. Site-directed mutagenesis of Phe-247 to the smaller leucine residue found in SULT1A3 gave rise to substrate inhibition in SULT1A1, suggesting that dopamine can be accommodated in the second binding site without making unfavorable interactions with the enzyme. Therefore, this residue seems to be involved in the phenomenon of substrate inhibition, because mutations in other residues that interact with pNP (Ile-89 and Val-148) exhibited kinetic profiles that are similar to the wild-type SULT1A1.
The size and nature of the substrate binding pocket of SULT1A1 and SULT1A3 suggest that much larger molecules than pNP can be accommodated. Endogenous substrates such as iodothyronines and 17-estradiol (E2) have shown to be catalyzed by SULT1A1 with a clear substrate inhibition (34, 35). SULT1A3 also sulfonates the above substrates with a low affinity (34, 36), and higher concentrations of these substrates needed to be analyzed to investigate substrate inhibition. In fact, 3,3'-diiodothyronine (T2) was recently identified as the endogenous substrate of SULT1A1 (37), and we identified a catalytically feasible binding conformation when T2 modeled into the active site of SULT1A1 (21). In this conformation, the two phenyl rings of T2 occupy each of the pNP binding sites by adopting a bent conformation. We also discussed in the same article that a rigid multi-ring substrates such as E2 cannot accommodate the binding pocket without a conformational change in the enzyme. However, with both T2 and E2, it is unlikely that two molecules will occupy the substrate binding site of these two isozymes. Therefore, substrate inhibition would probably occur by some other mechanism.
It is important to note that in the case of SULT1A1, Chen et al. (38) have reported that residues Glu-83, Asp-134, Glu-246, and Asp-263 are involved in the catalytic activity of SULT1A1 based on their SULT1A1 model. However, the crystal structure of SULT1A1 suggests that the impaired activity seen with mutations of these residues is caused by a perturbation of the 4 strand and 6 helix, potentially disrupting the structure of the backbone. In a recent publication, the same authors (39) used their SULT1A1 model (although the crystal structure of SULT1A1 was published at the time) to demonstrate the importance of arginine residues for SULT1A1 catalytic activity. For example, they showed that when arginine at position 78 was mutated to a glutamic acid (R78E), SULT1A1 lost its activity for 2-naphthol, whereas when it was mutated to alanine (R78A), the enzyme retained activity. Based on their SULT1A1 model, they suggested that the positive charge at position 78 is not critical for activity. When it was mutated to a negatively charged residue, there was a significant reduction in activity, which they suggested was caused by an effect on lysine 106 (Lys-106). However, the SULT1A1 crystal structure shows that Lys-106 is 12 Å from Arg-78 and therefore unlikely to be affected by the mutation. A more likely cause is a disruption of a hydrogen bond or charge-charge interaction network involving Asp-59 and Gln-56. These two residues are located on the 2 helix just preceding the phosphate sulfate binding loop, which is critical for PAPS binding (14, 16, 20, 21). The crystal structure suggests that the R78E mutant would probably disrupt this network, thereby destabilizing PAPS binding because of unfavorable charge-charge interactions.
The crystal structure of the major catecholamine sulfotransferase, SULT1A3, was the first three-dimensional structure of a human cytosolic sulfotransferase to be solved. This structure was published simultaneously by both our group (15) and Coughtrie and co-workers (16). The Bidwell et al. (15) structure was solved with a sulfate ion bound at the active site, whereas the Dajani et al. (16) structure was complexed with the product 3'-phosphoadenosine 5'-phosphate. Significantly, both groups reported that about 25% of the SULT1A3 structure consisted of disordered regions, probably because of the lack of a bound substrate. Most of the 68 disordered residues in the apoSULT1A3 structure correspond to the substrate-binding regions in SULT1A1 (21).
When examining the model of SULT1A3, it highlighted the importance of residue 247 (leucine in SULT1A3 and phenylalanine in SULT1A1) in accommodating two dopamine molecules in the substrate binding site. These modeling data convincingly demonstrate that the observed substrate inhibition in SULT1A3 with dopamine is caused by impeded catalysis when two molecules of dopamine are present in the active site. The molecular mechanism of this substrate inhibition can be explained by using the proposed kinetic model for SULT1A1 (21).
In contrast to SULT1A1, the modeled SULT1A3 substrate-binding pocket has more charged residues, such as aspartate and glutamate (Fig. 2c), and these would favor binding of substrates having positively charged groups, such as dopamine. The model not only highlighted the importance of residue 247 but also showed the importance of three other residues: Asp-86, Glu-89 and Glu-146 for substrate interaction. With the aid of molecular modeling (based on the mouse SULT1E1 structure) and site-directed mutagenesis, we and others have previously identified these three residues as responsible for the high dopamine affinity of SULT1A3 (26, 40–42). In our SULT1A3 model, Glu-146 is placed within hydrogen-bonding distance of the amino group of the first dopamine molecule (Fig. 2c). The importance of this interaction in dopamine sulfonation is confirmed by the functional studies of Brix et al. (41). They have shown that by changing Glu-146 in SULT1A3 to glutamine (abolishing the negative charge), there is a 360-fold decrease in the specificity constant for dopamine. Furthermore, the change of a single amino acid, E146A, in SULT1A3 was sufficient to change the catalytic properties and substrate specificity so that it mimicked those of SULT1A1 (40, 41, 43). A related study from our laboratory showed that the mutant A146E in SULT1A1 yielded a SULT1A3-like protein with respect to pNP (41). Taken together, these data provide strong evidence that residue 146 is crucial for recognition of dopamine and determining the substrate specificity in SULT1A1 and SULT1A3.
Mn2+ has shown to stimulate dopamine sulfonation by SULT1A3, and this is proposed to occur via interaction between Mn2+ and Asp-86 which is in the mobile loop having residues 86–90 (44, 45). Further, at higher concentration of Mn2+, it appeared that there was no substrate inhibition observed for dopamine. We can therefore use our model to propose how this takes place if substrate inhibition no longer occurs in the presence of Mn2+. This would imply that the second dopamine molecule would be displaced by Mn2+ binding. We investigated the region of Asp-86 and found that the amine of second dopamine molecule interacts with Glu-89, which is only 4–6 Å from Asp-86 (Fig. 2c). Glu-89 represents the only non-hydrophobic interaction for the second dopamine molecule. We therefore propose that Mn2+ binding at the SULT1A3 binding site may involve both Asp-86 and Glu-89 side chains. This interaction would prevent binding of the second dopamine molecule by removing a major protein interaction. This is further supported by the fact that Glu-89 is replaced by isoleucine in SULT1A1, this protein shows very low substrate binding affinity for dopamine, and its activity is unaffected by adding Mn2+ (44, 45).
The mutation of Glu-89 in SULT1A3 to isoleucine results in reduced activity toward dopamine (42, 45). In the SULT1A3 crystal structure (15), residues 86–90 form a mobile loop that intercalates into the active site of a symmetry-related monomer. The residues in this region are poorly ordered and were suggested to have the potential to undergo disorder-order transition upon substrate binding. In our model of SULT1A3, this flexible loop tucks back in to the same subunit to close over the active site, as in the crystal structure of SULT1A1, and Glu-89 is placed within hydrogen-bond distance to the second dopamine molecule. Even though the mechanism for low activity remains unclear, this mutant may have altered structural interactions of the flexible loop with the substrate. Then again, the decreased overall charge of the substrate-binding pocket may make it less able to bind dopamine.
The data from these investigations support the reliability of our SULT1A3 model, as well as the previously proposed hypothesis that inhibition in SULT1A1 and SULT1A3 is caused by binding of a second substrate molecule at the substrate binding site, and suggest a major role for residue 247 in substrate inhibition of human SULT1A1.
|
FOOTNOTES |
|---|
¶ Supported by a fellowship from the Australian Research Council.
** To whom correspondence should be addressed: Faculty of Biological and Chemical Sciences, Computer Science Building (Bldg. 69, Rm. 304), University of Queensland, Queensland 4072, Australia. Tel.: 61-7-3365-1609; Fax: 61-7-3365-1613; E-mail: m.mcmanus{at}uq.edu.au.
1 The abbreviations used are: SULT, sulfotransferase; PAPS, 3'-phosphoadenosine 5'-phosphosulfate; pNP, p-nitrophenol.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  L.-Y. Lu, Y.-C. Hsieh, M.-Y. Liu, Y.-H. Lin, C.-J. Chen, and Y.-S. Yang Identification and Characterization of Two Amino Acids Critical for the Substrate Inhibition of Human Dehydroepiandrosterone Sulfotransferase (SULT2A1) Mol. Pharmacol., March 1, 2008; 73(3): 660 - 668. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Gamage, A. Barnett, N. Hempel, R. G. Duggleby, K. F. Windmill, J. L. Martin, and M. E. McManus Human Sulfotransferases and Their Role in Chemical Metabolism Toxicol. Sci., March 1, 2006; 90(1): 5 - 22. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. U. Gamage, S. Tsvetanov, R. G. Duggleby, M. E. McManus, and J. L. Martin The Structure of Human SULT1A1 Crystallized with Estradiol: AN INSIGHT INTO ACTIVE SITE PLASTICITY AND SUBSTRATE INHIBITION WITH MULTI-RING SUBSTRATES J. Biol. Chem., December 16, 2005; 280(50): 41482 - 41486. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. C. Sacco and M. O. James SULFONATION OF ENVIRONMENTAL CHEMICALS AND THEIR METABOLITES IN THE POLAR BEAR (Ursus maritimus) Drug Metab. Dispos., September 1, 2005; 33(9): 1341 - 1348. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L.-Q. Wang, C. N. Falany, and M. O. James TRICLOSAN AS A SUBSTRATE AND INHIBITOR OF 3'-PHOSPHOADENOSINE 5'-PHOSPHOSULFATE-SULFOTRANSFERASE AND UDP-GLUCURONOSYL TRANSFERASE IN HUMAN LIVER FRACTIONS Drug Metab. Dispos., October 1, 2004; 32(10): 1162 - 1169. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
