Photoaffinity labeling of the aglycon binding site of the recombinant human liver UDP-glucuronosyltransferase UGT1A6 with 7-azido-4-methylcoumarin.

7-Azido-4-methylcoumarin (AzMC) is a fluorescent photoactive compound structurally related to 4-methylumbelliferone (4-MU), a marker substrate of the human liver recombinant UDP-glucuronosyltransferase (UGT) 1A6. AzMC was synthesized and utilized to label the substrate binding site of UGT1A6. AzMC exhibits a fluorescence spectrum with maximum excitation and emission wavelengths of 380 and 442 nm, respectively. Upon irradiation, the probe irreversibly inhibited glucuronidation activity measured with para-nitrophenol (pNP) as substrate and interacted with UGT1A6 according to a saturable process indicative of reversible binding before covalent incorporation of the photoaffinity label. This inhibition was both time and concentration dependent and led to the calculation of an inhibition constant, k(2) = 0.113 mM min(-1), and dissociation constant, K(d) = 2.89 mM, for the reaction. Partial photoinactivation of UGT1A6 with AzMC revealed that the probe decreased the apparent V(max) of the pNP glucuronidation reaction, but not the K(m). Moreover, inhibition was partially prevented by 1-naphthol, a surrogate substrate for the enzyme, or by preincubation with an active-site directed inhibitor, 5'-O-[[(2-decanoylamino-3-phenyl-propyloxycarbonyl)amino]-su lfonyl]-2 ',3'-O-isopropylideneuridine. In contrast, UDP-glucuronic acid (UDP-GlcUA) did not have any protective effect against photoinactivation and AzMC did not affect the photoaffinity labeling of UGT1A6 by 5-[beta-(32)P]N(3)UDP-GlcUA, a photoaffinity analog of UDP-GlcUA. Additionally, in the absence of irradiation, AzMC was found to be a competitive inhibitor of 4MU glucuronidation. Collectively, these results strongly indicate that AzMC specifically binds to the UGT1A6 aglycon binding site. Amino acid alignment of phenol-binding proteins revealed a conserved motif, YXXXKXXPXP. It is possible that this motif is involved in phenol binding to UGT1A6 and other phenol-accepting proteins.     

CYP2E1 active site residues in substrate recognition sequence 5 identified by photoaffinity labeling and homology modeling

Despite its biological importance, our knowledge of active site structure and relevance of critical amino acids in CYP2E1 catalytic processes remain limited. In this study, we identified CYP2E1 active site residues using photoaffinity labeling with 7-azido-4-methylcoumarin (AzMC) coupled with a CYP2E1 homology model. In the absence of light, AzMC was an effective competitor against substrate p-nitrophenol oxidation by CYP2E1. Photoactivation of AzMC led to a concentration-dependent loss in CYP2E1 activity and structural integrity resulting from the modification of both heme and protein. The photo-labeling reaction degraded heme and produced a possible heme adduct. Probe incorporation into the protein occurred at multiple sites within substrate recognition sequence 5 (SRS-5). Based on a CYP2E1 homology model, we hypothesize AzMC labels SRS-5 residues, Leu363, Val364, and Leu368, in the active site. In addition, we propose a series of phenylalanines, especially Phe106, mediate contacts with the coumarin.     

3-Azi-1-methoxybutyl D-maltooligosaccharides specifically bind to the maltose/maltooligosaccharide-binding protein of Escherichia coli and can be used as photoaffinity labels

Maltooligosaccharides with two to six (alpha 1-4)-linked glucose residues, carrying at their reducing end a 3-azi-1-methoxybutyl group in either alpha or in beta glycosidic linkage, were synthesized. These maltooligosaccharide analogues inhibit maltose uptake via the maltose-binding-protein-dependent transport system in Escherichia coli. The concentration of half-maximal inhibition of maltose transport, at 15 nM concentration, decreases with increasing chain length of the analogue, levelling off at 40 microM after a chain length of four glucose residues in the alpha series and at 350 microM after a chain length of three glucose residues in the beta series. The inhibition of maltose transport occurs at the level of the periplasmic maltose-binding protein. 3-Azi-1-methoxybutyl alpha-D-[3H]maltotrioside was bound by the maltose-binding protein with a Kd of 0.18 mM. Irradiation at 350 nm of purified maltose-binding protein in the presence of 4 microM of this substrate labeled the protein covalently; labeling was prevented by 1 mM maltose. Using a crude preparation of periplasmic proteins two proteins were labeled, the maltose-binding protein and alpha-amylase. Thus, 3-azi-1-methoxybutyl alpha-D-maltooligosaccharides are potent photoaffinity labels for proteins with maltooligosaccharides-binding sites.     

The molecular basis for the broad substrate specificity of human sulfotransferase 1A1

Cytosolic sulfotransferases (SULTs) are mammalian enzymes that detoxify a wide variety of chemicals through the addition of a sulfate group. Despite extensive research, the molecular basis for the broad specificity of SULTs is still not understood. Here, structural, protein engineering and kinetic approaches were employed to obtain deep understanding of the molecular basis for the broad specificity, catalytic activity and substrate inhibition of SULT1A1. We have determined five new structures of SULT1A1 in complex with different acceptors, and utilized a directed evolution approach to generate SULT1A1 mutants with enhanced thermostability and increased catalytic activity. We found that active site plasticity enables binding of different acceptors and identified dramatic structural changes in the SULT1A1 active site leading to the binding of a second acceptor molecule in a conserved yet non-productive manner. Our combined approach highlights the dominant role of SULT1A1 structural flexibility in controlling the specificity and activity of this enzyme.     

Active site mutations and substrate inhibition in human sulfotransferase 1A1 and 1A3

Human SULT1A1 is primarily responsible for sulfonation of xenobiotics, including the activation of promutagens, and it has been implicated in several forms of cancer. Human SULT1A3 has been shown to be the major sulfotransferase that sulfonates dopamine. These two enzymes shares 93% amino acid sequence identity and have distinct but overlapping substrate preferences. The resolution of the crystal structures of these two enzymes has enabled us to elucidate the mechanisms controlling their substrate preferences and inhibition. The presence of two p-nitrophenol (pNP) molecules in the crystal structure of SULT1A1 was postulated to explain cooperativity at low and inhibition at high substrate concentrations, respectively. In SULT1A1, substrate inhibition occurs with pNP as the substrate but not with dopamine. For SULT1A3, substrate inhibition is found for dopamine but not with pNP. We investigated how substrate inhibition occurs in these two enzymes using molecular modeling, site-directed mutagenesis, and kinetic analysis. The results show that residue Phe-247 of SULT1A1, which interacts with both p-nitrophenol molecules in the active site, is important for substrate inhibition. Mutation of phenylalanine to leucine at this position in SULT1A1 results in substrate inhibition by dopamine. We also propose, based on modeling and kinetic studies, that substrate inhibition by dopamine in SULT1A3 is caused by binding of two dopamine molecules in the active site.     

Photoaffinity labeling of human retinoid X receptor beta (RXRbeta) with 9-cis-retinoic acid: identification of phytanic acid, docosahexaenoic acid, and lithocholic acid as ligands for RXRbeta

We utilized [20-methyl-(3)H]-9-cis-retinoic acid ([(3)H]9-cis-RA) as a direct photoaffinity probe for the characterization of human recombinant retinoid X receptor beta protein (RXRbeta). The photoaffinity labeling was light- and concentration-dependent, saturable, and protected by unlabeled 9-cis-RA in a concentration-dependent manner, indicating that binding occurred in the RXR retinoid binding site. all-trans-Retinoic acid (atRA) did not affect labeling with the 9-cis derivative, confirming that atRA does not compete for the 9-cis-RA binding site. Several retinoid, fatty acid, and bile acid ligands were evaluated for their ability to recognize the 9-cis-RA binding site. Retinol, atRA glucuronide, 13-cis-RA, dolichol, 5,6-epoxy-RA, and vitamin D(3) did not compete for the 9-cis-RA binding site. However, the saturated diterpenoid phytanic acid (PA) and docosahexaenoic acid, which have been recently shown to activate the nuclear receptor, RXR, competed with 9-cis-RA labeling, showing high affinity for the 9-cis-RA binding site. Oleic acid, arachidonic acid, and butyric acid did not interact. However, the bile acid lithocholic acid competed efficiently with 9-cis-RA for the binding site. These data validated the photoaffinity assay as an excellent system for the identification and evaluation of ligands for RXR.     

Mechanism of interferon action. The interferon-induced phosphoprotein P1 possesses a double-stranded RNA-dependent ATP-binding site

Protein P1, the interferon-induced protein phosphorylated in the presence of dsRNA in human amnion U-cells, was covalently labeled with [alpha-32P]ATP following ultraviolet irradiation. The photoaffinity labeling of protein P1 was dependent upon double-stranded RNA. Antibody prepared against phosphorylated protein P1 immunoprecipitated the double-stranded RNA-dependent photoaffinity-labeled product. The extent of photoaffinity labeling was significantly decreased by the addition of unlabeled ATP, GTP, or AMP; adenosine had little effect on the photoaffinity labeling of protein P1. These results suggest that protein P1 possesses a site capable of binding an adenine nucleotide in a double-stranded RNA-dependent manner.     

Photoaffinity labeling of peptide binding sites of prolyl 4-hydroxylase with N-(4-azido-2-nitrophenyl)glycyl-(Pro-Pro-Gly)5

The synthesis is described of the photoaffinity label N-(4-azido-2-nitrophenyl)glycyl-(Pro-Pro-Gly)5 for the peptide binding site of prolyl 4-hydroxylase. The photoaffinity label is a good substrate and is capable of light-induced inactivation of prolyl 4-hydroxylase activity. Inactivation depends on the concentration of photoaffinity label and is prevented by competition with excess (Pro-Pro-Gly)5. Two moles of photoaffinity label per mole of enzyme is needed for 100% inactivation of enzymic activity. Oxidative decarboxylation of 2-oxoglutarate measured in the absence of added peptide substrate is not affected by labeling. We conclude that the covalently bound nitreno derivative of N-(4-azido-2-nitrophenyl)glycyl-(Pro-Pro-Gly)5 acts by preventing the binding of peptide substrate to the catalytic site without interfering with the binding of the other substrates and cofactors 2-oxoglutarate, O2, Fe2+, and ascorbate. Labeling is specific for the alpha subunit of the tetrameric alpha 2 beta 2 enzyme. In addition to two catalytic binding sites that are blocked by the photoaffinity label, the enzyme contains binding subsites for peptide substrates, as judged from the capability of photoinactivated enzyme to bind to a poly(L-proline) affinity column. These binding subsites may account for the rapidly increasing affinity for peptide substrates with increasing chain length.     

Identification by photoaffinity labeling of fatty acid-binding protein as a potential warfarin receptor in rat liver

Two different photoaffinity analogs of 4-hydroxy coumarin, 3-(p-azidobenzyl)-4-hydroxycoumarin (AzBHC) and 3-(4-azido-5-iodosalicylamido)-4-hydroxycoumarin (AzISAHC), are being used in the identification of warfarin-binding proteins present in mammalian tissue (Myszka, D. G., and Swenson, R. P. (1990) Biochem. Biophys. Res. Commun. 172, 415-422; Myszka, D. G., and Swenson, R. P. (1991) J. Biol. Chem. 266, 4789-4797). In this study, [14C]AzBHC, but not [125I]AzISAHC, was observed to specifically label a 15,000-dalton protein present in both the microsomal and cytosolic fractions of rat liver. Pretreatment of the crude protein samples with warfarin or dicoumarol completely protected the 15-kDa protein from modification by [14C]AzBHC, indicating that this photoaffinity reagent is specifically labeling a coumarin-binding protein. 4-Hydroxycoumarin itself and AzISAHC were unable to block the incorporation of this photoaffinity probe. The 15-kDa protein was isolated by two-dimensional electrophoresis and subjected to amino-terminal sequence analysis. The first 20 amino acid residues analyzed were found to be identical with the amino-terminal sequence of rat liver fatty acid-binding protein (L-FABP) (Gordon J. I., Alpers, D. H., Ockner, R. K., and Strauss, A. W. (1983) J. Biol. Chem. 258, 3356-3363). Photoaffinity labeling and protection experiments carried out on purified preparations of L-FABP paralleled the labeling results obtained in the microsomes and cytosol, confirming that L-FABP is capable of specifically binding AzBHC, warfarin, and dicoumarol. Oleic acid, an established ligand for L-FABP, can compete with the binding of the photoaffinity probe; however, it was less effective in protecting the protein than warfarin. The specificity of labeling of crude liver fractions by warfarin photoaffinity analogs reported here as well as the high concentration of FABP in liver tissue together suggest that this protein may represent a major hepatic receptor responsible for the uptake and/or transport of various oral 4-hydroxycoumarin-based anticoagulant drugs.     

Photoaffinity labeling of Ras converting enzyme 1 (Rce1p) using a benzophenone-containing peptide substrate

Isoprenylation is a post-translational modification that increases protein hydrophobicity and helps target certain proteins to membranes. Ras converting enzyme 1 (Rce1p) is an endoprotease that catalyzes the removal of a three residue fragment from the C-terminus of isoprenylated proteins. To obtain structural information about this membrane protein, photoaffinity labeling agents are being prepared and employed. Here, we describe the synthesis of a benzophenone-containing peptide substrate analogue for Rce1p. Using a continuous spectrofluorometric assay, this peptide was shown to be a substrate for Rce1p. Mass spectrometry was performed to confirm the site of cleavage and structure of the processed probe. Photolysis of the biotinylated compound in the presence of membranes containing Rce1p followed by streptavidin pull-down and Western blot analysis indicated that Rce1p had been labeled by the probe. Photolysis in the presence of both the biotinylated, benzophenone-containing probe and a farnesylated peptide competitor reduced the extent of labeling, suggesting that labeling is occurring in the active site.     

Interaction of flap endonuclease-1 and replication protein A with photoreactive intermediates of DNA repair

A new method for enzymatic synthesis of radioactive DNA flapped structures containing a photoreactive dCMP moiety at a branch point with 4-(4-azido-2,3,5,6-tetrafluorobenzylidene-hydrazinocarbonyl)butylcarbamoyl group attached at exo-N-position of cytosine was developed. The formation of complexes of flap endonuclease-1 (FEN-1) with flapped DNA was shown by photoaffinity modification and gel retardation assays. The substrate properties of the flapped structures with different flap lengths were studied in the reaction of endonuclease cleavage catalyzed by FEN-1. It was demonstrated that inhibition of FEN-1 activity by replication protein A (RPA) depends on the length of the single-stranded part of the flapped substrate. A significant inhibition of cleavage was observed when the flap length was sufficient for effective RPA binding, while for structures with short single-stranded part the efficiency of cleavage was independent of the presence of RPA. FEN-1 and RPA were modified by photoaffinity labeling using flap structures with single-stranded parts consisting of 8 and 21 nucleotides. Products of DNA photoattachment to FEN-1 were observed in both cases, while the covalent adducts with RPA were obtained only with the 21-nucleotide-long flap. Photoaffinity modification demonstrated that FEN-1 and RPA compete for the binding of the flapped substrates with long single-stranded parts.     

Structure of a human carcinogen-converting enzyme,SULT1A1. Structural and kinetic implications of substrate inhibition

Sulfonation catalyzed by sulfotransferase enzymes plays an important role in chemical defense mechanisms against various xenobiotics but also bioactivates carcinogens. A major human sulfotransferase, SULT1A1, metabolizes and/or bioactivates many endogenous compounds and is implicated in a range of cancers because of its ability to modify diverse promutagen and procarcinogen xenobiotics. The crystal structure of human SULT1A1 reported here is the first sulfotransferase structure complexed with a xenobiotic substrate. An unexpected finding is that the enzyme accommodates not one but two molecules of the xenobiotic model substrate p-nitrophenol in the active site. This result is supported by kinetic data for SULT1A1 that show substrate inhibition for this small xenobiotic. The extended active site of SULT1A1 is consistent with binding of diiodothyronine but cannot easily accommodate beta-estradiol, although both are known substrates. This observation, together with evidence for a disorder-order transition in SULT1A1, suggests that the active site is flexible and can adapt its architecture to accept diverse hydrophobic substrates with varying sizes, shapes and flexibility. Thus the crystal structure of SULT1A1 provides the molecular basis for substrate inhibition and reveals the first clues as to how the enzyme sulfonates a wide variety of lipophilic compounds.     

Direct photoaffinity labeling of cellular retinoic acid-binding protein I (CRABP-I) with all-trans-retinoic acid: identification of amino acids in the ligand binding site

Cellular retinoic acid-binding proteins I and II (CRABP-I and -II, respectively) are transport proteins for all-trans-retinoic acid (RA), an active metabolite of vitamin A (retinol), and have been reported to be directly involved in the metabolism of RA. In this study, direct photoaffinity labeling with [11,12-(3)H]RA was used to identify amino acids comprising the ligand binding site of CRABP-I. Photoaffinity labeling of CRABP-I with [(3)H]RA was light- and concentration-dependent and was protected by unlabeled RA and various retinoids, indicating that the labeling was directed to the RA-binding site. Photolabeled CRABP-I was hydrolyzed with endoproteinase Lys-C to yield radioactive peptides, which were separated by reversed-phase HPLC for analysis by Edman degradation peptide sequencing. This method identified five modified amino acids from five separate HPLC fractions: Trp7, Lys20, Arg29, Lys38, and Trp109. All five amino acids are located within one side of the "barrel" structure in the area indicated by the reported crystal structure as the ligand binding site. This is the first direct identification of specific amino acids in the RA-binding site of CRABPs by photoaffinity labeling. These results provide significant information about the ligand binding site of the CRABP-I molecule in solution.     

Sulfotransferase 4A1

In this review, we highlight the physical and enzymatic properties of the novel human sulfotransferase, SULT4A1. The gene is most highly expressed in selective regions of the brain, although work to date has failed to identify any specific endogenous substrate for the enzyme. SULT4A1 shares low homology with other human sulfotransferases. Nevertheless, it is highly conserved between species. Despite the low homology, it is structurally very similar to other cytosolic sulfotransferases with a conserved substrate binding domain, dimerization site and partial cofactor binding sites. However, the catalytic cavity is much smaller, and it has been suggested that the cofactor may not be accommodated within it. A recent link between variability in the 5'UTR of the SULT4A1 gene and schizophrenia has heightened interest in the endogenous function of the enzyme and its possible role in human disease.     

Crystal structure of human cholesterol sulfotransferase (SULT2B1b) in the presence of pregnenolone and 3'-phosphoadenosine 5'-phosphate. Rationale for specificity differences between prototypical SULT2A1 and the SULT2BG1 isoforms

The gene for human hydroxysteroid sulfotransferase (SULT2B1) encodes two peptides, SULT2B1a and SULT2B1b, that differ only at their amino termini. SULT2B1b has a predilection for cholesterol but is also capable of sulfonating pregnenolone, whereas SULT2B1a preferentially sulfonates pregnenolone and only minimally sulfonates cholesterol. We have determined the crystal structure of SULT2B1a and SULT2B1b bound to the substrate donor product 3'-phosphoadenosine 5'-phosphate at 2.9 and 2.4 A, respectively, as well as SULT2B1b in the presence of the acceptor substrate pregnenolone at 2.3 A. These structures reveal a different catalytic binding orientation for the substrate from a previously determined structure of hydroxysteroid sulfotransferase (SULT2A1) binding dehydroepiandrosterone. In addition, the amino-terminal helix comprising residues Asp19 to Lys26, which determines the specificity difference between the SULT2B1 isoforms, becomes ordered upon pregnenolone binding, covering the substrate binding pocket.     

Synthesis of the photoaffinity probe 3-(p-azidobenzyl)-4-hydroxycoumarin and identification of the dicoumarol binding site in rat liver NAD(P)H:quinone reductase (EC 1.6.99.2).

A photoaffinity analog of 4-hydroxycoumarin containing an azidobenzyl group at the 3-position and, if desired, carbon-14 or tritium radionuclides has been synthesized and characterized. This compound, 3-(p-azidobenzyl)-4-hydroxycoumarin, serves as an effective competitive inhibitor of the dicoumarol-sensitive NAD(P)H:quinone reductase (EC 1.6.99.2; DT-diaphorase) from rat liver, having an apparent inhibition constant of 6.6 x 10(-8) M, a value comparable to that observed for dicoumarol (1.7 x 10(-9) M), significantly lower than for Warfarin (3.5 x 10(-5) M) and well within the range required of an effective photoaffinity reagent. Irradiation of the reductase with ultraviolet light in the presence of the photoprobe resulted in the covalent labeling of up to 10% of the protein. Greater than 99% of the covalent incorporation is precluded by the addition of 15 microM dicoumarol, consistent with the specific labeling of the 4-hydroxycoumarin binding site of this enzyme by this photoaffinity reagent. Further evidence of a high degree of specificity is provided by the isolation and sequence analysis of the peptides covalently modified by this reagent. A single region within the protein was found to be labeled, with threonine 127 and tyrosine 128 being the only amino acid residues that were observed to be modified. These results, for the first time, define a portion of the 4-hydroxycoumarin binding site within a protein that has a well established sensitivity to this type of anticoagulant and, because dicoumarol serves as a competitive inhibitor for pyridine nucleotides in this enzyme, may also define a portion of this unusual pyridine nucleotide binding site. In addition, these results suggest that this reagent may be effective as a highly specific photoaffinity probe in the identification of other proteins that are similarly inhibited by 4-hydroxycoumarin derivatives, such as the microsomal enzymes associated with the vitamin K-dependent carboxylation system.     

Binding of a photoaffinity analogue of glutathione to MRP1 (ABCC1) within two cytoplasmic regions (L0 and L1) as well as transmembrane domains 10-11 and 16-17

MRP1 (or ABCC1) is an ABC membrane protein that transports a wide range of natural products as well as glutathione (GSH)-, glucuronate-, and sulfate-conjugated metabolites. In addition, free GSH is required for MRP1 to transport several chemotherapeutic drugs. However, the mechanisms regulating the influence of GSH on MRP1 is poorly understood, and the location of GSH binding site(s) within MRP1 have yet to be determined. To address these issues, we have synthesized a [(125)I] labeled azido-derivative of GSH (IAAGSH) to photoaffinity label MRP1. Our results revealed that IAAGSH labeled MRP1 with high specificity, and binding was inhibited by MRP1 substrates leukotriene C(4) and MK571. Interestingly, verapamil and vincristine enhanced IAAGSH photolabeling of MRP1, in agreement with observations that both drugs enhance GSH transport. We observed GSH to be the best inhibitor of photoaffinity labeling, as compared to oxidized glutathione (GSSG) and four different GSH alkyl derivatives. These observations indicate that IAAGSH interacted with MRP1 in a similar manner as unmodified GSH. Moreover, using eight MRP1-HA variants, each containing hemagglutinin A (HA) epitopes inserted at different sites in MRP1, we mapped the GSH binding sites in MRP1. Our GSH analogue photoaffinity labeled four MRP1 polypeptides that were located within two cytoplasmic domains in linker sequences (L0 and L1) as well as transmembrane domains 10-11 and 16-17. The photoaffinity labeling of polypeptides within L0 and L1 domains is further confirmed using two MRP1-specific monoclonal antibodies (MRPr1 and QCRL1) with epitopes within the linker domains. Taken together, this study provides the most precise information to date on the location of GSH binding sites in MRP1.     

Arginine residues in the active site of human phenol sulfotransferase (SULT1A1)

Cytosolic sulfotransferases (STs) catalyze the sulfation of hydroxyl containing compounds. Human phenol sulfotransferase (SULT1A1) is the major human ST that catalyzes the sulfation of simple phenols. Because of its broad substrate specificity and lack of endogenous substrates, the biological function of SULT1A1 is believed to be an important detoxification enzyme. In this report, amino acid modification, computer structure modeling, and site-directed mutagenesis were used for studies of Arg residues in the active site of SULT1A1. The Arg-specific modification reagent, 2,3-butanedione, inactivated SULT1A1 in an efficient, time- and concentration-dependent manner, suggesting Arg residues play an important role in the catalytic activity of SULT1A1. According to the computer model, Arg78, Arg130, and Arg257 may be important for SULT1A1 catalytic activity. Site-directed mutagenesis results demonstrated that the positive charge on Arg78 is not critical for SULT1A1 because R78A is still active. In contrast, a negative charge at this position, R78E, completely inactivated SULT1A1. Arg78 is in close proximity to the site of sulfuryl group transfer. Arg257 is located very close to the 3'-phosphate in adenosine 3'-phosphate 5'-phosphosulfate (PAPS). Site-directed mutagenesis demonstrated that Arg257 is critical for SULT1A1: both R257A and R257E are inactive. Although Arg130 is also located very close to the 3'-phosphate of PAPS, R130A and R130E are still active, suggesting that Arg130 is not a critical residue for the catalytic activity of SULT1A1. Computer modeling suggests that the ionic interaction between the positive charge on Arg257, and the negative charge on 3'-phosphate is the primary force stabilizing the specific binding of PAPS.