Structure And Function Of The Human Sult1a Subfamily

Overall Crystal Structure

The crystal structure of the major catecholamine SULT, SULT1A3, was the first three-dimensional structure of a human cytosolic SULT solved (Bidwell et al., 1999; Dajani et al., 1999a). Prior to this, Negishi's group had published the crystal structures of the mouse cytosolic estrogen SULT (Kakuta et al., 1997) and the SULT domain of the human membrane-bound heparan sulfate N-deacetylaselN-SULT 1 (Kakuta et al., 1999). Since then, additional crystal structures of human cytosolic SULTs have been published: the hydroxysteroid SULT, SULT2A3 (Pedersen et al., 2000), the dehydroepiandrosterone sulfotransferase, DHEA-ST (Rehse, et al., 2002), the estrogen SULT, SULT1E1 (Pedersen et al., 2002), and the phenol SULT SULT1A1 (Gamage et al., 2003). The Bidwell et al. (1999) structure of SULT1A3 diffracted to 2.4A and was solved with a sulfate ion in the active site, whereas the Dajani et al. (1999a) structure diffracted to 2.5A and was complexed with 3'-phosphoadenosine 5'-phosphate (PAP). Significantly, both structures showed large stretches of disorder that account for approximately 25% of the SULT1A3 structure (Figure 10.4a; residues 64-77, 84-99, 216-261). This is thought to be due to the lack of a bound substrate.

To gain more insight into the structural and functional features of SULT 1A subfamily members we have recently solved the crystal structure of SULT1A1 (Gamage et al., 2003). The SULT1A1 cDNA was expressed as an N-terminal hexa-histidine tagged protein in Escherichia. coli, purified, and crystallized in the presence of the desulfonated cofactor PAP and the model xenobiotic substrate p-nitrophenol. The crystal diffracted to 1.9A, and the structure was fully solved except for seven residues missing from the N terminus. The crystal structure clearly shows binding of the PAP and two molecules of p-nitrophenol in an extended substrate-binding pocket. From both the crystal structures of human SULT1A1 and SULT1A3, we can conclude that they are comprised of a core a/p domain, similar to other SULTs, which forms the backbone of a central five-stranded parallel p sheet surrounded on either side by helices (Figure 10.4; see color photo insert following p. 210).

PAP Binding

From the initial cloning and sequence alignment studies of SULTs from a variety of species, it became apparent that there were at least four highly conserved regions throughout phylogeny (Varin et al., 1992; Weinshilboum et al., 1997). A number of laboratories utilized these data to construct chimeric cDNAs, performed site-directed

SULT1A3:S04 SULT1A1 :PAP:pNP

figure 10.4 Crystal structure of human (a) SULT1A3 complexed with sulfate (pink) and (b) SULT1A1 complexed with PAP (green) and p-nitrophenol (pNP1 orange and pNP2 blue). Disordered regions of SULT1A3 are shown in blue dotted lines (Gamage, N.U., Duggleby, R.G., Barnett, A.C., Tresillian, M., Latham, C.F., Liyou, N.E., McManus, M.E., and Martin, J.L. (2003) J Biol Chem 278:7655-7662. With permission). (See color photo insert following p. 210.)

figure 10.4 Crystal structure of human (a) SULT1A3 complexed with sulfate (pink) and (b) SULT1A1 complexed with PAP (green) and p-nitrophenol (pNP1 orange and pNP2 blue). Disordered regions of SULT1A3 are shown in blue dotted lines (Gamage, N.U., Duggleby, R.G., Barnett, A.C., Tresillian, M., Latham, C.F., Liyou, N.E., McManus, M.E., and Martin, J.L. (2003) J Biol Chem 278:7655-7662. With permission). (See color photo insert following p. 210.)

mutagenesis on these regions, and carried out affinity labeling to gain insight into the substrate and PAPS-binding sites of guinea pig (Komatsu et al., 1994), rat (Tamura et al., 1997), plant (Marsolais and Varin, 1995; Marsolais et al., 1999), and human (Radominska et al., 1996) SULTs. Initially, it was suggested that a motif with the sequence GxxGxxK (GMAGDWK in both SULT1A1 and SULT1A3), which is common to all SULTs, was critical for the interaction between SULT and the cofactor PAPS and considered as the P-loop (Chiba et al., 1995; Driscoll et al., 1995; Komatsu et al., 1994). However, the residues that are in the P-loop, a nucle-otide-binding motif found in ATP- and GTP-binding proteins, actually correspond to residues TYPKSGT (45-51) in mouse SULT1E1 (Kakuta et al., 1997). However, results from Zheng et al. (1994), employing chemical affinity labeling of a rat SULT1A SULT IV by an ATP analog, showed covalent attachment to Lys65 and Cys66 in the N-terminal portion of the protein. Following mutagenesis studies on the plant flavonol 3-SULT, Marsolais and Varin (1995) concluded that both Lys59 and Arg276 are involved in PAPS binding through ionic interactions. These results were in part confirmed by Kakuta et al. (1997) when the crystal structure of the mouse estrogen SULT was published (Kakuta et al., 1997). This structure showed that PAP binding involves residues 257-259 (Arg, Lys, Gly) at the beginning of the GXXGXXK region, a second region before the a-helix 6 loop (Arg130) and on a-helix 6 (Ser138). It also showed that the position of the adenine ring of PAPS is determined by Trp53, Thr227, and Phe229.

In studies on SULT1A1 and SULT1A3, we have shown that the inactive cofactor PAP binds in a similar manner to that found by Kakuta et al. (1997) for mouse SULT1E1. The PSB loop (residues 45-51) interacts with the 5'-phosphate of the PAP molecule (Figure 10.5). The amino acid sequence (45-TYPKSGT-51), which

figure 10.5 Human SULT1A1 interactions with PAP Residues from the PSB loop (K48, S49, G50, T51, T52) form H-bonds to the 5'-phosphate of PAP. Residues R130, S138, R257, K258, and G259 form interactions with the 3'PB site. Residues that are H-bonded to the N6 of the adenine ring are T227 and Y193.

forms the classical strand-loop-helix motif found in all other SULT structures (Kakuta et al., 1997; Yoshinari et al., 2001), is also conserved in the human SULT1A subfamily (Figure 10.3). The backbone amides as well as the side chains of residues 48 to 51 are hydrogen bonded to the 5'-phosphate of PAP (Figure 10.5). In particular Lys48, which hydrogen bonds with the leaving oxygen of the 5'-phosphate group, has been shown by Yoshinari et al. (2001) to act as a general acid in catalysis. This residue is conserved in both SULT1A1 and SULT1A3 and nearly all other SULTs (Kakuta et al., 1998). Therefore, this binding site is considered to play a role in the recognition of PAPS and orientates the sulfate group for sulfonate transfer.

The 3'-phosphate-binding site (3'PB; residues 130, 138, and 257-259) of SULT1A1 is well ordered, showing a conserved strand and helix motif, a structural feature found in all other SULT structures (Kakuta et al., 1997; Pedersen et al., 2000). However, in the human SULT1A3 structure this site is disordered (Bidwell et al., 1999; Dajani et al., 1999a). In the SULT1A1 crystal structure, the two residues of the conserved strand-helix motif (Arg130 at the end of strand 5 and Ser138 from helix 8) interact directly with the 3'-phosphate group of PAP (Figure 10.5). The side-chain interaction of Ser138 with 3'-phosphate of PAP is observed in all known SULT structures. The human SULT1E1 structure, which was solved recently by Pedersen et al. (2002) in the presence of PAPS, showed that Ser138 prevents PAPS undergoing hydrolysis in the absence of substrate. Negishi et al. (2001) have postulated that the introduction of the 3'-phosphate group on the sulfonate donor could be an important force that has evolved in all SULTs. In the SULT1A1 structure, residues 257-RKG-259 at the beginning of the conserved GxxGxxK motif (Figure 10.3) are also found within hydrogen-bonding distance of the oxygen atoms of 3'-phosphate (Figure 10.5). Similar to the mouse SULT1E1 structure, the conserved residues Trp53 and Phe229 form a parallel ring stacking arrangement with the adenine ring of PAP molecule and are stabilized by hydrogen-bond interactions of

Thr227 (strand 12) and Thr193 (strand 6). Thus, SULT1A1 binds PAP in a proper orientation for catalysis similar to other known SULT structures. This conserved catalytic core of the SULT1A1 structure provides further evidence that both cytosolic and membrane-bound SULTs (the SULT domain of human N-deacetylasel N-SULT 1) have similar structural motifs that are responsible for PAP binding. Based on this structural similarity, Yoshinari et al. (2001) have concluded that both cytosolic and membrane-bound SULTs probably evolved from a common ancestral gene.

Substrate binding

The substrate-binding region of the SULT1A subfamily will mainly be reviewed based on the recently solved SULT1A1 structure. Figure 10.6, a-c (see color photo insert following p. 210), shows that a deep hydrophobic pocket comprises the substrate-binding site for cytosolic SULTs, whereas the presumed substrate-binding site of the human membrane-bound NST1 (SULT domain of human N-deacetylasel N-SULT1), which is responsible for the sulfonation of large molecules such as carbohydrates, glucosaminylglycans, and proteins, has a large open cleft (Figure 10.6d; see color photo insert following p. 210). However, as mentioned above, the resolved structure of SULT1A1 revealed an L-shaped substrate-binding pocket that is larger than that anticipated from the disordered SULT1A3 structure. The residues lining the substrate-binding pocket of SULT1A1 were found to be well ordered and predominantly hydrophobic (Phe81, 142, 24, 84, 76, 247, 255, Ile89, His149, Tyr169, Tyr240, Ile21, Ala146, 86, Met248, Met77, Val243, Pro90, and Val148) and are contributed by helices 1 and 6, strands 2 and 4, and several loop regions (Figure 10.3 and Figure 10.7). Figure 10.3 demonstrates that residues located at positions 76, 77, 84, 86, 89, 146, 148, 149, 247 in the binding pocket are not conserved among SULT1A family members, which is consistent with the varying substrate specificity of these enzymes.

Figure 10.7 (see color photo insert following p. 210) shows that two p-nitrophenol molecules are bound to the substrate-binding pocket of SULT1A1. The more tightly bound and deeply buried p-nitrophenol molecule (pNP1) is positioned near the active site. The two highly conserved phenylalanines at positions 81 and 142 appear to form a substrate access gate for pNP1 (Figure 10.7). Petrotchenko et al. (1999) showed that these two residues play a critical role in maintaining the proper structure of the binding pocket. The highly conserved residues His108 and Lys106 of SULT1A1 form hydrogen bonds with the SO31- accepting the hydroxyl group of pNP1. The nitro group of pNP1 interacts with a water molecule and forms van der Waals (vdW) interactions with Phe247, Met248, and Val148, whereas the weakly bound pNP2 molecule does not interact with the catalytic residues but forms vdW interactions with Ile89 and Phe247. In the SULT1A3 structure, similar structural elements are utilized to form the substrate-binding pocket although the residues in loop region 216-261 are disordered (Figure 10.4a). Another unusual feature of the SULT1A3 structure is that part of the presumed substrate-binding pocket is occupied by residues 86-90 from a symmetry related molecule. In the SULT1A1 structure, these residues line the binding pocket of the pNP2 molecule. The disorder that is observed in the SULT1A3 structure is apparently due to lack of the substrate, which may be necessary for disorder-order transition in the substrate-binding pocket in these enzymes.

figure 10.6 Crystal structures of four SULTs showing the substrate-binding pocket. (a) Human SULT1A1 with two p-nitrophenol molecules and PAP bound (Gamage, N.U., Dug-gleby, R.G., Barnett, A.C., Tresillian, M., Latham, C.F., Liyou, N.E., McManus, M.E., and Martin, J.L. (2003) J Biol Chem 278:7655-7662. With permission). (b) Human SULT1A3 with SO42-bound (Bidwell, L.M., McManus, M.E., Gaedigk, A., Kakuta, Y., Negishi, M., Pedersen, L., and Martin, J.L. (1999) J Mol Biol 293:521-530. With permission). (c) Mouse SULT1E1 with E2 and PAP bound (Kakuta, Y., Pedersen, L.G., Carter, C.W., Negishi, M., and Pedersen, L.C. (1997) Nat Struct Biol 4:904-908. With permision). (d) SULT domain of human N-deacetylase/N-SULT 1 (NST1; Kakuta, Y., Sueyoshi, T., Negishi, M., and Pedersen, L.C. (1999) J Biol Chem 274:10673-10676. With permission). Substrate molecules are shown in orange and PAP molecule is shown in the ball-and-stick model. Arrow indicates the substrate-binding pocket. (See color photo insert following p. 210.)

figure 10.6 Crystal structures of four SULTs showing the substrate-binding pocket. (a) Human SULT1A1 with two p-nitrophenol molecules and PAP bound (Gamage, N.U., Dug-gleby, R.G., Barnett, A.C., Tresillian, M., Latham, C.F., Liyou, N.E., McManus, M.E., and Martin, J.L. (2003) J Biol Chem 278:7655-7662. With permission). (b) Human SULT1A3 with SO42-bound (Bidwell, L.M., McManus, M.E., Gaedigk, A., Kakuta, Y., Negishi, M., Pedersen, L., and Martin, J.L. (1999) J Mol Biol 293:521-530. With permission). (c) Mouse SULT1E1 with E2 and PAP bound (Kakuta, Y., Pedersen, L.G., Carter, C.W., Negishi, M., and Pedersen, L.C. (1997) Nat Struct Biol 4:904-908. With permision). (d) SULT domain of human N-deacetylase/N-SULT 1 (NST1; Kakuta, Y., Sueyoshi, T., Negishi, M., and Pedersen, L.C. (1999) J Biol Chem 274:10673-10676. With permission). Substrate molecules are shown in orange and PAP molecule is shown in the ball-and-stick model. Arrow indicates the substrate-binding pocket. (See color photo insert following p. 210.)

Reaction Mechanism

The precise mechanism of how a sulfonate group is transferred from the cofactor PAPS to the substrate has to a certain degree been in dispute. Initially, Duffel and Jakoby (1981) reported the sulfonation of p-nitrophenol by rat liver aryl transferase as a random Bi Bi mechanism in which substrate and PAPS bind in an independent manner. However, recently Duffel et al. (2001) explained substrate inhibition by p-nitrophenol in terms of a sequential mechanism with PAPS binding first followed by p-nitrophenol. Studies on the catalytic mechanism of recombinant human

F142

K106

figure 10.7 Stereo view of the active site of human SULT1A1 showing the hydrophobic nature of residues surrounding the ligands. p-nitrophenol1 (pNP1) and p-nitrophenol2 (pNP2) are shown in orange and blue, respectively. F142 and F81 form the substrate access gate. H-bonds are shown in black dotted lines (Gamage, N.U., Duggleby, R.G., Barnett, A.C., Tresil-lian, M., Latham, C.F., Liyou, N.E., McManus, M.E., and Martin, J.L. (2003) J Biol Chem 278:7655-7662. With permission). (See color photo insert following p. 210.)

F142

K106

o figure 10.7 Stereo view of the active site of human SULT1A1 showing the hydrophobic nature of residues surrounding the ligands. p-nitrophenol1 (pNP1) and p-nitrophenol2 (pNP2) are shown in orange and blue, respectively. F142 and F81 form the substrate access gate. H-bonds are shown in black dotted lines (Gamage, N.U., Duggleby, R.G., Barnett, A.C., Tresil-lian, M., Latham, C.F., Liyou, N.E., McManus, M.E., and Martin, J.L. (2003) J Biol Chem 278:7655-7662. With permission). (See color photo insert following p. 210.)

SULT1E1 suggested that sulfonate transfer follows a random Bi Bi mechanism with two dead-end complexes (Zhang, 1998). These same authors also reported that the binding of ^-estradiol (E2) to human SULT1E1 resulted in two E2-binding sites/catalytic subunit, suggesting that the enzyme contains an allosteric E2-binding site. In contrast to the random Bi Bi mechanism discussed above, two studies using purified SULTs from human brain (Whittemore et al., 1986) and rhesus monkey liver (Barnes et al., 1986) concluded that the sulfonation reaction proceeds via a sequentially ordered Bi Bi reaction. Varin and Ibrahim (1992) have also reported a similar mechanism for a plant flavonol SULT. When the crystal structure of mouse SULT1E1 complexed with PAP and E2 was solved, it became clear that the core structure resembles the uridine kinase enzyme, indicating a similar mechanism to phosphoryl transfer (i.e., SN2 in-line displacement; Kakuta et al., 1997, 1998). The active site and the transition state mimicked by the mouse SULT1E1-PAP-vanadate ion structure complex provided further evidence for this in-line displacement mechanism for the sulfonate transfer reaction catalyzed by SULTs.

Figure 10.8 shows the sulfonate transfer mechanism based on data from the mouse SULT1E1 (Kakuta et al., 1997; Pedersen et al., 2002) and the human SULT1A1 crystal structure. His108, which is common to all SULTs, acts as a catalytic base and deprotonates the acceptor 3' OH group of p-nitrophenol, which in turn converts it to a strong nucleophile that attacks the sulfur atom of PAPS. This leads to building of negative charge at the bridging oxygen and forces Lys48 to donate its proton to the bridging oxygen, and sulfonate dissociation occurs. The conserved Ser138 appears to prevent PAPS hydrolysis in the absence of substrate (Pedersen et al., 2002). Therefore, based on this structural data, the donor substrate

figure 10.8 Schematic representation of sulfuryl transfer mechanism for human SULT1A1 with the substrate p-nitrophenol (pNP). The reaction takes place via SN2 in-line displacement as described by Kakuta, Y., Pedersen, L.G., Carter, C.W., Negishi, M., and Pedersen, L.C. (1997) Nat Struct Biol 4:904-908. With permision.

figure 10.8 Schematic representation of sulfuryl transfer mechanism for human SULT1A1 with the substrate p-nitrophenol (pNP). The reaction takes place via SN2 in-line displacement as described by Kakuta, Y., Pedersen, L.G., Carter, C.W., Negishi, M., and Pedersen, L.C. (1997) Nat Struct Biol 4:904-908. With permision.

PAPS binds first, followed by the binding of sulfonate acceptor substrate, favoring the in-line displacement mechanism (Negishi et al., 2001).

Substrate Specificity

While considerable progress has been made in recent times with the publication of five crystal structures of cytosolic SULTs (Bidwell et al., 1999; Dajani et al., 1999a; Gamage et al., 2003; Kakuta et al., 1997; Pedersen et al., 2000, 2002), we are still in the process of fully understanding the principles that underpin the substrate specificity of the individual enzymes. The key to understanding the overlapping but distinct substrate specificities displayed within the SULT1A subfamily most probably lies in the substrate-binding sites of these enzymes. In contrast to the PAPS-binding site, which consists of conserved amino acids, the differing substrate specificities of the closely related SULT1A subfamily members, SULT1A1, SULT1A2, and SULT1A3, suggest that the structure of their substrate-binding pockets has undergone modification during evolution. Therefore, the amino acid residues that contribute to differing substrate specificities are likely to reside in regions that display the most variability across members of this subfamily. The ability to investigate the role of critical amino acids in these variable regions has been aided by the substrate preferences of SULT1A subfamily members: SULT1A1 — high affinity for p-nitrophenol and low affinity for dopamine; SULT1A2 — lower affinity for p-nitrophenol and no activity toward dopamine as a substrate; and SULT1A3 — high affinity for dopamine and low affinity for p-nitrophenol (Dajani et al., 1998; Sakakibara et al., 1998; Veronese et al., 1994; Zhu et al., 1996).

Initial studies aimed at understanding the substrate specificity of SULT were carried out using chimeric constructs of flavonol (Varin et al., 1995) and rat hydroxysteroid SULTs (Tamura et al., 1997). These data suggest that the central regions spanning amino acids 92-194 and 102-164 in the plant and rat enzymes, respectively, determine their substrate specificities. Similarly, Sakakibara et al. (1998) used chimeric constructs to investigate the amino acids determining the substrate specificity of SULT1A1 and SULT1A3 isoforms. A sequence encompassing amino acid residues 84-148 was found to be the substrate-specific domain of both SULT1A1 and SULT 1 A3 and highlighted the importance of the variable Regions I (residues 84-89) and II (residues 143-148) in determining their distinct enzymatic properties. A related study from our laboratory showed that substrate affinities are mainly determined within the N-terminal end of SULT1A1 (HAST1), SULT1A2 (HAST4), and SULT 1 A3 (HAST3) and include two regions of high sequence divergence that were termed Region A (residues 44-107) and B (residues 132-164), respectively (Brix et al., 1999a). In parallel with the work of Sakakibara et al. (1998), it was shown by Dajani et al. (1998) that the change of a single amino acid, Glu146Ala, was sufficient to change the catalytic properties and substrate specificity of SULT1A3 so that it mimicked those of SULT1A1. In two related studies, we used a variety of phenolic substrates to functionally characterize the role of the amino acids at position 146 in both SULT1A1 and SULT1A3 (Brix et al., 1999a, 1999b). First, the mutant Ala146Glu in SULT1A1 yielded a SULT1A3-like protein with respect to the Km for simple phenols such as p-nitrophenol. The mutation Glu146Ala in SULT1A3 yielded a SULT1A1-like protein with respect to the Km for both phenols and monoamine compounds (e.g., dopamine). These data provided strong evidence that residue 146 is crucial in determining the substrate specificity of both SULT1A1 and SULT1A3. Further, a negatively charged glutamic acid (Glu) at position 146 is crucial for the recognition of dopamine (Brix et al., 1999b).

The importance of specific amino acid interactions in determining the high activity of SULT1A3 toward dopamine as a substrate has been further highlighted by molecular modeling studies (Dajani et al., 1999a; Yoshinari et al., 2001). Unfortunately, these studies have relied on the crystal structure of the mouse SULT1E1 because the structure of SULT1A3 with bound substrate is not currently available (Bidwell et al., 1999; Dajani et al., 1999a). Therefore, we constructed a computer model of SULT1A3 using the coordinates of SULT1A1, which has 93% sequence identity to SULT1A3 (Barnett et al., 2004). Our SULT1A3 model superimposes on the SULT1A3 crystal structure with a root mean square deviation of 1.1A. PAPS was modeled into SULT1A3, based on the SULT1E1-PAPS structure (PDB code 1HY3; Pederson et al., 2002), and dopamine was docked. With a small shift in orientation, either the 3-OH or 4-OH of dopamine could be placed for sulfonation. However, 3-O sulfonated dopamine is reported to be the predominant metabolite in human (Dajani et al., 1999a); therefore, dopamine was orientated for sulfonation at the 3-O position. As in the SULT1A1 structure, Phe81 and Phe142 form a substrate access gate. In contrast to SULT1A1, the substrate-binding pocket is comprised of charged residues such as Glu and Asp. However, as in SULT1A1, the substrate-binding pocket is large enough to accommodate two molecules of dopamine

figure 10.9 The substrate-binding pocket of the modeled SULT1A3 structure showing the binding mode of PAPS and two molecules of dopamine. Ligands and residues are represented as ball-and-stick models. Dopamine 1 is shown in dark green and dopamine 2 in light green. The cofactor PAPS is shown as stick model. Residues not shown for clarity are Tyr240, Ile21, Phe24, Pro90, Val243, Met248 (Barnett, A.C., Tsvetanov, S., Gamage, N., Martin, J.L., Duggleby, R.G., and McManus, M.E. (2004) Active site mutations and substrate inhibition in human SULT 1A1 and 1A3. J Biol Chem 279:18799-18805. With permission). (See color photo insert following p. 210.)

figure 10.9 The substrate-binding pocket of the modeled SULT1A3 structure showing the binding mode of PAPS and two molecules of dopamine. Ligands and residues are represented as ball-and-stick models. Dopamine 1 is shown in dark green and dopamine 2 in light green. The cofactor PAPS is shown as stick model. Residues not shown for clarity are Tyr240, Ile21, Phe24, Pro90, Val243, Met248 (Barnett, A.C., Tsvetanov, S., Gamage, N., Martin, J.L., Duggleby, R.G., and McManus, M.E. (2004) Active site mutations and substrate inhibition in human SULT 1A1 and 1A3. J Biol Chem 279:18799-18805. With permission). (See color photo insert following p. 210.)

(Figure 10.9; see color photo insert following p. 210). In this model, the side chain of Glu146 is placed within hydrogen-bonding distance of the amino group of the first dopamine molecule to form a charge interaction, further highlighting the importance of this interaction in dopamine sulfonation (Dajani et al., 1999a; Yoshinari et al., 2001). This interaction has been confirmed by functional studies by Brix et al. (1999b) who have shown that by changing the Glu at position 146 of SULT1A3 to a glutamine (Gln), thereby neutralizing the negative charge at this position, a 360fold decrease occurred in the specificity constant for dopamine. Further, Dajani et al. (1999a) have reported that mutating Glu146 to an alanine (Ala) in SULT1A3 significantly reduces the Km values for a range of 4-substituted phenols, suggesting that this residue plays a central role in limiting substrate binding and access to the active site. In a more recent paper, Liu et al. (2000) have shown that the concerted action of three mutations (Asp86Ala, Glu89Ile, and Glu146Ala) is sufficient for the conversion of the substrate phenotype of SULT1A3 (M-PST) to that of SULT1A1 (P-PST). In our SULT1A3 model, Glu89 interacts with the amino group of the second dopamine molecule and Asp86 is positioned close to the first dopamine molecule, in agreement with the above observations. Recent studies have shown that Asp86 is involved in Mn2+ stimulation of the Dopa/tyrosine activity of SULT1A3 (Pai et al., 2003). However, Mn2+ exerts a smaller stimulatory effect on dopamine sulfonation by binding directly to Asp86 without making a complex with dopamine. This again highlights the importance of these residues, and it is interesting that in the active site of SULT1A1, we have also noted the importance of Ala86 and Ile89. In the crystal structure of SULT1A1 (Figure 10.7), Ile89 has been shown to form a critical interaction with the pNP2 molecule. While it has been possible to convert the SULT1A3 substrate phenotype to a SULT1A1-like functional protein, achieving the reverse has not been possible to date (Brix et al., 1999b; Liu et al., 2000).

Kinetics of Human SULT1A1

Substrate inhibition has been reported previously for SULT1A enzymes (Raftogianis et al., 1999; Reiter et al., 1983) though the published studies have generally assumed a Michaelis-Menten model to explain the kinetics of these enzymes (Brix et al., 1999a; Lewis et al., 1996). The kinetic implications of the presence of two p-nitrophenol molecules in the crystal structure of SULT1A1 were investigated (Figure 10.10, a and b). A slight deviation from Michaelis-Menten kinetics (broken line) is observed at low substrate concentrations (Gamage et al., 2003). This could suggest that some positive cooperativity is present for p-nitrophenol binding. The most pronounced feature is the substrate inhibition occurring at higher substrate concentrations (above 2 ^M). According to the general kinetic model constructed (Figure 10.10c), p-nitrophenol can bind to the enzyme at site 1 or 2, and occupancy of site 1 does not prevent subsequent binding at site 2. There are two catalytically competent species, ESj and ESjS2 that form the EP and EPS2 enzyme-product complexes with rate constants k1 and k2, respectively. EP releases the product directly (dissociation constant Kp), while release from EPS2 requires prior release of p-nitrophenol from site 2 (dissociation constant Kps2). The model fits well to the experimental data as shown by the lines in Figure 10.10, a and b. Therefore, the presence of two molecules in the active site revealed by the structure is a real property of the SULT1A1 enzyme, and substrate inhibition at high concentrations of p-nitrophenol is due to the impeded catalysis when both binding sites are occupied.

The structural model of SULT1A3 showed that the substrate-binding pocket is large enough to accommodate a second molecule of dopamine (Figure 10.9). Similar to other SULT1A enzymes, SULT1A3 exhibits substrate inhibition at high concentrations of dopamine (Ganguly et al., 1995). Based on our kinetic analysis of sulfonation of dopamine by SULT1A3, we suggest that the inhibition we observed with dopamine is also caused by binding of a second substrate molecule in the substrate-binding pocket.

Table 10.1 shows that in addition to sulfonating a range of small molecular weight compounds, SULT1A isoforms are capable of metabolizing larger substrates such as 17^-estradiol (E2), iodothyronines, 1-hydroxymethylpyrene, 7-hydroxy-7,8,9,10-tet-rahydrobenzo(a)pyrene, 2-naphthylamine, N-hydroxy-2-acetylaminoflorene, and 2-hydroxylamino-5-phenylpyridine (Glatt et al., 2001; Hernandez et al., 1991; Llerena et al., 2001; Meinl et al., 2002). Further, Chou et al. (1995) have reported that the mutagens/carcinogens N-hydroxy-2-acetylaminofluore, N-hydroxy-2-aminoflourene, N-hydroxy-4, 4'-methylene-bis (2-chloroaniline), N-hydroxy-2-amino-1-methyl-6-imidazo[4,5-b]pyridine, and N-hydroxy-2-amino-6-methyldipyrido[1,2-a:3',2'-d]imi-dazole are preferentially sulfonated by human SULT1A1. The binding of two molecules of p-nitrophenol in the active site of SULT1A1 highlights the large and very hydrophobic nature of the substrate-binding region of this enzyme. This binding

figure 10.10 Kinetics of pNP sulfonation by human SULT1A1. (a) pNP concentrations up to 1.5 ^M. (b) pNP concentrations up to 20 ^M. Open and closed symbols show the results from two independent experiments. Each data point is the mean of duplicate or triplicate assays; the standard deviation is contained within the dimensions of the symbols. (c) Kinetic model of human SULT1A1. The enzyme (E) can bind pNP at site 1 to give ES1 with a dissociation constant of KS1 or at site 2 to give ES2 with a dissociation constant of KS2. Occupancy of site 2 prevents pNP from binding to site 1, while occupancy of site 1 does not prevent pNP from binding to site 2 to give ESjS2 (Gamage, N.U., Duggleby, R.G., Barnett, A.C., Tresillian, M., Latham, C.F., Liyou, N.E., McManus, M.E., and Martin, J.L. (2003) J Biol Chem 278:7655-7662. With permission). The lines in panels (a) and (b) represent the best fit of this model to these data.

figure 10.10 Kinetics of pNP sulfonation by human SULT1A1. (a) pNP concentrations up to 1.5 ^M. (b) pNP concentrations up to 20 ^M. Open and closed symbols show the results from two independent experiments. Each data point is the mean of duplicate or triplicate assays; the standard deviation is contained within the dimensions of the symbols. (c) Kinetic model of human SULT1A1. The enzyme (E) can bind pNP at site 1 to give ES1 with a dissociation constant of KS1 or at site 2 to give ES2 with a dissociation constant of KS2. Occupancy of site 2 prevents pNP from binding to site 1, while occupancy of site 1 does not prevent pNP from binding to site 2 to give ESjS2 (Gamage, N.U., Duggleby, R.G., Barnett, A.C., Tresillian, M., Latham, C.F., Liyou, N.E., McManus, M.E., and Martin, J.L. (2003) J Biol Chem 278:7655-7662. With permission). The lines in panels (a) and (b) represent the best fit of this model to these data.

site can accept small flat aromatic compounds, larger L-shaped aromatics, and extended planar aromatic or aliphatic ring systems.

Recent studies have suggested that SULT1A1 is primarily responsible for the sulfonation of 3,3'-diiodothyronine (3,3'-T2) in the human placenta and developing liver (Li et al., 2001; Stanley et al., 2001). While other SULTs such as human SULT1B1 (Fujita et al., 1997; Wang et al., 1998), human SULT1C1 (Li et al., 2001), and human SULT1E1 (Kester et al., 1999) are also capable of sulfonating thyroid hormones, it is apparent from their expression levels that they play a minor role compared to SULT1A1 during the ontogeny of the liver (Richard et al., 2001). 3,3'-T2 has been shown to stimulate mitochondrial respiration in different tissues, which is not mediated by the T3 receptor (Li et al., 2001; Moreno et al., 1997). Therefore, it is possible that T2 and its sulfonated product could play a physiological function in the human fetus. Based on the geometry of the two bound p-nitrophenol molecules in the substrate-binding pocket of the SULT1A1 crystal structure, we docked 3,3'-T2 into the SULT1A1 structure (Figure 10.11a; see color photo insert following p. 210). The outcome was a catalytically competent binding model for 3,3'-T2 when Phe247 was changed to an alternate conformation to prevent steric clash with one of the iodine atoms of the substrate. This structural model supports the metabolic results of Richard et al. (2001) and suggests that iodothyronines are endogenous substrates of SULT1A1 (Gamage et al., 2003).

It is reported that SULT1A1 sulfonates the endogenous substrate 17^-estradiol (E2) with a lower affinity than other substrates (Adjei and Weinshilboum, 2002; Falany and Falany, 1997). Therefore, E2 was modeled into the SULT1A1 active site using the binding mode identified in the SULT1E1:PAP:E2 structure (Figure 10.11b). However, E2 cannot be accommodated in the SULT1A1 structure even with the altered conformation for Phe247, as it makes unfavorable interactions with residues involved in two loops (residues 146-154 between a6 and a7 and 84-90 immediately preceding a4; Figure 10.11c). These two loops close over the active site of SULT1A1 more tightly than they do in the SULT1E1 structure, thus restricting the space available for ligands that bind in an extended conformation (Figure 10.11d). Therefore, we propose that binding of extended fused ring systems such as E2 to SULT1A1 results in an energy cost due to conformational rearrangement. Sulfonation could occur only if there was a conformational change in the SULT1A1-binding site, and thus the SULT1A1 substrate-binding site must be sufficiently plastic to accept flexible and rigid ligands, as shown in Table 10.1.

Through the elucidation of the crystal structures of SULT1A1 and SULT1A3, plus the kinetic analysis of chimeric constructs and mutated enzymes, a degree of insight has been obtained into the critical amino acids that control the substrate specificity of these SULTs. However, considerably more work is required before we fully understand the distinct but overlapping substrate specificities of SULT1A subfamily members. For example, the reasons SULT1A2 has a low affinity for p-nitrophenol and no dopamine activity, although it shares 97% and 94% homology to SULT1A1 and SULT1A3 respectively, is yet to be elucidated. Further, the fact that SULT1A3 is capable of sulfonating a range of substrates from smaller catechols to larger endogenous substrates, such as iodothyronines (Kester et al., 1999), suggests that its substrate-binding pocket may be flexible, similar to that of SULT1A1. In general, the substrate specificity in SULT1A subfamily members appears to be determined by both binding affinity and proper positioning of a substrate in the active site. For example, SULT1A1 has a predominant hydrophobic-binding pocket, which aids its interaction with substrates such as p-nitrophenol. On the other hand, the substrate-binding pocket of SULT1A3 appears to be less hydrophobic, and specific charge interactions between functional groups govern the basis of its substrate specificity. Overall, the above data suggest some commonality between the abc

figure 10.11 Human SULT1A1 active site plasticity. (a) T2 (yellow) docked into the active site of human SULT1A1, solvent accessible surface was calculated with Phe247 in an alternate rotamer conformation (pink). p-nitrophenol1 (orange) and p-nitrophenol2 (blue) are shown to indicate the orientation. PAPS is modeled at the cofactor-binding site. (b) Solvent accessible surface was calculated for E2 with Phe247 having an alternate conformation (pink). E2 makes clashes with Val148, Phe84, and the alternate conformation of Phe247. (c) Superimposition of human SULT1A1 and mouse SULT1E1 showing the striking variation in loop region. (d) The solvent accessible surface for the two pNP molecules as in the human SULT1A1 structure (Gamage, N.U., Duggleby, R.G., Barnett, A.C., Tresillian, M., Latham, C.F., Liyou, N.E., McManus, M.E., and Martin, J.L. (2003) J Biol Chem 278:7655-7662. With permission). (See color photo insert following p. 210.)

SULT1A1 and SULT1A3 active sites, which is probably not unexpected since they metabolize many of the same substrates, albeit with different affinities and activities.

Low Carb Diets Explained

Low Carb Diets Explained

You can burn stored body fat for energy and shed excess weight by reducing the carbohydrate intake in your diet. Learn All About The Real Benefits of Low Carb Diets And Discover What They Can Really Do To Improve The Quality Of Your Life Today.

Get My Free Ebook


Post a comment