Adenosine 5′-Phosphosulfate Kinase from Penicillium chrysogenum

The properties of Penicillium chrysogenum adenosine 5′-phosphosulfate (APS) kinase mutated at Ser-107 were examined. Ser-107 is analogous to a serine of the E. coli enzyme that has been shown to serve as an intermediate acceptor in the transfer of a phosphoryl group from ATP to APS. Replacement of Ser-107 with alanine yielded an active enzyme with kinetic characteristics similar to those of wild-type APS kinase. Another mutant form of the enzyme in which Ser-107 was replaced by cysteine was also active. Covalent modification of Cys-107 eliminated catalytic activity, and substrates protected against modification. Mutation of Ser-97, of Ser-99, of Thr-103, of Ser-104 to alanine, or of Tyr-109 to phenylalanine also yielded an active enzyme. The cumulative results indicate that Ser-107 may reside in the substrate binding pocket of fungal APS kinase, but neither it nor any nearby hydroxy amino acid serves as an obligatory phophoryl acceptor in the 3′-phosphoadenylylsulfate synthesis reaction. The results also indicate that the absence of a serine at position 478 in the APS kinase-like C-terminal region of fungal ATP sulfurylase does not account for the lack of APS kinase activity in that enzyme. However, mutating the ATP P-loop residues in APS kinase to those found in the analogous C-terminal region of fungal ATP sulfurylase eliminated enzyme activity.

To date, the amino acid sequences of APS kinases from 15 different sources have been reported, although the actual protein has been purified and characterized from only a few. APS kinases from the filamentous fungus Penicillium chrysogenum (GenBank TM accession number U39393) and from the bacterium Escherichia coli (GenBank TM accession number P23846) are of particular interest because they have quite disparate properties despite having a high degree of homology. The P. chrysogenum enzyme is a homodimer (subunit molecular mass of approximately 24 kDa) that undergoes a temperature-dependent reversible dissociation into inactive monomers (1). The kinetic mechanism of the fungal enzyme is essentially compulsory ordered with MgATP adding before APS and with PAPS dissociating before MgADP. APS is a potent substrate inhibitor. Steady state kinetics (2), protection by reversibly bound ligands against chemical inactivation (3), and direct binding studies (4) all point to the formation of a dead-end E⅐MgADP⅐APS complex as the cause of the substrate inhibition. That is, APS adds to the sulfonucleotide subsite on E⅐MgADP that lies vacant after PAPS dissociates from the ternary complex. (APS has little affinity for the enzyme in the absence of MgADP or MgATP.) Incubation of the P. chrysogenum enzyme with [␥-32 P]MgATP in the absence of APS does not yield detectable levels of labeled enzyme (4). In contrast, the E. coli enzyme follows a hybrid sequential ordered pingpong mechanism, with MgATP adding first (5). At low APS concentrations, MgADP can dissociate before APS adds. Substrate inhibition by APS results from formation of an E⅐APS complex when APS binds to the free enzyme blocking MgATP addition (5,6). Incubation of the bacterial APS kinase with [␥-32 P]MgATP in the absence of APS leads to the formation of a catalytically competent phosphoenzyme. The phosphorylated residue was identified as Ser-109 (5,7), a residue that is conserved in all but one (8) APS kinase that has been sequenced to date. 2 The phosphorylated E. coli enzyme exists primarily as a dimer, whereas dephosphorylation favors a tetrameric form of the enzyme.
Despite the kinetic and structural differences noted above, it seemed reasonable to assume that the chemical mechanism of homologous APS kinases would be the same. That is, if the path of phosphoryl group transfer in the reaction catalyzed by the E. coli enzyme is from MgATP to the enzyme and subsequently to APS, then this mechanism should prevail in all cases. So, whereas the E. coli enzyme forms an isolable stable phosphoenzyme intermediate, the fungal enzyme might form a transient phosphoenzyme, but only when APS is present at the catalytic site. To explore this possibility, we examined the functional properties of P. chrysogenum APS kinase that had been mutated to contain either an alanine or a cysteine residue at position 107, i.e. at the position analogous to Ser-109 of the E. coli enzyme. The effects of substitution at other nearby OHcontaining residues and in the ATP P-loop region were also examined. This research was also motivated by the observations that fungal ATP sulfurylase (a) is allosterically inhibited by PAPS (9, 10) and (b) contains a C-terminal domain that is * The research described in this paper was initiated with support from Grant DMB-91-05143 from the National Science Foundation and was completed with the support of a Bridge Grant from the University of California Davis Office of Research. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ To whom correspondence should be addressed. Tel.: 530-752-3193; Fax: 530-752-3085; E-mail: ihsegel@ucdavis.edu. 1 The abbreviations used are: APS, adenosine 5Ј-phosphosulfate (5Јadenylylsulfate); PAPS, 3Ј-phosphoadenosine-5Ј-phosphosulfate (3Јphosphoadenylylsulfate); DTNB, 5Ј,5Ј-dithiobis(2-nitrobenzoate); PCR, polymerase chain reaction. homologous to APS kinase (11). The C-terminal region (which probably evolved by recombination of the ATP sulfurylase and APS kinase genes) almost certainly serves as the allosteric PAPS binding site. Given the facile monomer-dimer interconversion exhibited by true APS kinase, it is not hard to imagine a similar interaction of C termini on adjacent ATP sulfurylase subunits contributing to the allosteric transition induced by PAPS and substrates. Unlike the bifunctional protein from animal systems (12,13), fungal ATP sulfurylase has no detectable APS kinase activity. The absence of APS kinase activity in fungal ATP sulfurylase could result from several sequence modifications within the C-terminal region, one of which is the substitution of alanine for serine at position 478 (the position analogous to Ser-107 of true APS kinase). The experiments described in this present paper would help determine whether this substitution is sufficient to account for the lack of APS kinase activity in fungal ATP sulfurylase.
Sequenced APS kinase genes were cloned as NdeI-XhoI fragments into the Novagen pET23a(ϩ) plasmids and introduced into the expression host by electroporation. E. coli strain BL21(DE3) was used to express the P. chrysogenum APS kinases described in this report. However, a separate experiment established that active S107A APS kinase was also obtained when the recA Ϫ E. coli strain HMS174(DE3) was used.
Growth and Extraction of E. coli-About 0.2 ml of an 8-h culture was used to inoculate a 3-liter Fernbach flask containing 400 ml of LB Amp medium (15). This culture was grown aerobically at 30°C for 10 -15 h to an A 600 nm of about 6. Although the E. coli strain contains a T7 RNA polymerase that is under control of the lac promoter, we found that induction with isopropyl-1-thio-␤-D-galactopyranoside was not necessary for high enzyme yields. Centrifugation of the culture at 5000 ϫ g for 7 min yielded approximately 3-4 ml of packed cells. These were washed by resuspending the pellets in a total of 400 ml of chilled Buffer A (0.3 M Tris-HCl, pH 7.9, at 5°C containing 5 mM EDTA). After another centrifugation, the pellet was suspended in 20 ml of Buffer A containing 1 mg ϫ ml Ϫ1 lysozyme and incubated at room temperature for 15 min. All subsequent steps were carried out at 5°C. The lysozymetreated cells were centrifuged at 1000 ϫ g for 30 min, and the pellet was frozen by immersing the centrifuge tube in liquid nitrogen. The frozen cells were then thawed by resuspending the pellet in 35 ml of cold Buffer A. After sonication (three 30-s bursts using a Virsonic 300 microprobe on setting 3 with a 1 min ice bath incubation between bursts), the cell debris and unbroken cells were removed by centrifugation at 16,300 ϫ g for 10 min. The supernatant extract was saved, and the pellet was subjected to the above freeze/thaw/sonication/centrifugation sequence two more times. The supernatants were pooled. APS kinase activity of the crude extracts could be measured easily by means of the coupled spectrophotometric assay described below. For example, the cell-free extract of the E. coli expressing the wild-type (Ser-107) enzyme contained approximately 4000 total units of APS kinase activity (measured at 5 mM MgATP plus 3.5 M APS), accounting for about 20% of the total soluble protein as estimated from Coomassie Bluestained SDS gels. No activity could be detected in extracts prepared from cells that did not contain pET23a(ϩ) or from cells containing the plasmid without the APS kinase insert. (The absence of detectable endogenous activity most likely results from repression of the E. coli enzyme by organic sulfur-containing compounds in the rich medium.) Enzyme Purification-The pooled soluble fractions were applied to an Affi-Gel Blue (Bio-Rad) column (2.5 ϫ 10 cm) that had been equilibrated with Buffer B (40 mM Tris-HCl, pH 8.2, at 5°C). The column was washed with Buffer B at a flow rate of 3 ml/min until the effluent had an A 280 nm of 0.005 or less. Then, protein was eluted with a linear gradient of NaCl (0 -1 M) in Buffer B (total volume, 500 ml) at a flow rate of 1.5 ml per min. Seven-ml fractions were collected, and their A 280 nm was measured. Fractions containing enzyme activity (coincident with the major protein peak) were pooled, dialyzed against Buffer B, and then applied to a DEAE-cellulose column (2.5 ϫ 15 cm) equilibrated with the same buffer. After a brief wash, this column was eluted at 1 ml/min with a linear gradient of NaCl (0 -0.4 M) in Buffer B (total volume, 500 ml). Fractions containing APS kinase activity (again coincident with the major protein peak) were collected and pooled. This preparation had an A 280 nm /A 260 nm ratio of 2.0 (indicating good separation from nucleic acid fragments) and was judged to be Ͼ95% pure by SDS gel electrophoresis followed by protein staining with Coomassie Blue. The yield of enzyme at this stage generally exceeded 20 mg of APS kinase per 400 ml of culture.
Enzyme and Protein Assays-Except where noted, all enzyme assays were performed in 50 mM Tris-Cl, pH 8.1, 30°C (Buffer C) containing 100 mM ammonium sulfate. One unit of activity is defined as that amount of enzyme that promotes the formation of 1 mol of primary product in 1 min. Activity in the forward (PAPS synthesis) direction was measured by the pyruvate kinase ϩ lactate dehydrogenase ϩ nuclease P1-coupled spectrophotometric assay as described previously (2,16), except that the KCl concentration was 7.5 mM, and the pyruvate kinase ϩ lactate dehydrogenase preparation was obtained from Sigma as a 50% glycerol solution instead of as a suspension in ammonium sulfate. The APS kinase levels in the assays were varied between 0.15 and 0.9 g ml Ϫ1 ([E] t approximately 6 -38 nM in terms of sites) in order to achieve a ⌬A 340 nm of about 0.02 per min. The large excess of nuclease P1 continuously regenerates APS from the PAPS formed, thus permitting initial velocities to be measured at micromolar levels of APS. If nuclease P1 was omitted from the assay mixture, the total absorbance change at 340 nm in the presence of excess MgATP was proportional to the APS added, confirming that it was kinase activity that was being measured, not ATPase. The high (NH 4 ) 2 SO 4 decreased the severity of the substrate inhibition by APS and also minimized ionic strength variations encountered when different formulations of the coupling enzymes were used.
The reverse reaction was measured by monitoring the rate of ATP formation from PAPS and MgADP in the presence of excess ATP sulfurylase, hexokinase, and glucose-6-phosphate dehydrogenase (2). APS kinase levels in the assays were 4 g ml Ϫ1 (167 nM in sites). Diadenosine 5Ј-pentaphosphate (50 M) was included in the assay mixture to reduce the background rate of NADPH formation promoted by traces of contaminating adenylate kinase. (NH 4 ) 2 SO 4 (which interfered with the action of diadenosine 5Ј-pentaphosphate) was omitted.
Protein concentrations of the final preparation were estimated from the absorbance of the solution at 260 nm and 280 nm (17). The exact concentration of APS in the stock solution was determined spectrophotometrically from the end point of the reverse ATP sulfurylase reaction coupled to hexokinase plus glucose-6-phosphate dehydrogenase (16).
The covalent modification of S107C by DTNB was monitored directly at 412 nm as described previously for fungal ATP sulfurylase (18). The procedure for measuring enzyme inactivation by DTNB or tetrathionate and protection by ligands was similar to that described by Renosto et al. (3). All modification-inactivation preincubations were performed in Buffer C at 30°C in the presence of 5 mM excess Mg 2ϩ .
Kinetic Data Analyses-The velocity equation describing an ordered bireactant sequence with a dead-end EBQ complex in the absence of products (2) is as follows.
The family of plots do not intersect at a common point but any two lines intersect at the point described by the following equation.
Because the different fixed levels of B were all below 0.5[B] opt measured at near-saturating [A], the circle of ambiguity of the intersection points was small, and the horizontal coordinate of the circle provided a reasonable estimate of Ϫ1/K ia (Fig. IX-64 in Ref. 19). The reverse APS kinase reaction follows rapid equilibrium ordered kinetics (2). In this case, where P ϭ PAPS and Q ϭ MgADP, V max, app at saturating PAPS is the true V max at all fixed [MgADP] (Ref. 19, pp. 320 -322), whereas the K mP, app is described by the following equation.

RESULTS AND CONCLUSIONS
Kinetic Properties of Recombinant Wild-type and S107A APS Kinases- Fig. 1 shows the v versus [MgATP] and v versus [APS] profiles for the expressed wild-type APS kinase and the recombinant mutant enzyme containing alanine in place of the serine at position 107. Although the velocity profiles of the two enzymes differ, both behave qualitatively like native APS kinase purified from P. chrysogenum mycelium (2). The most characteristic feature is the substrate inhibition exerted by APS, a property shared by APS kinases from most sources (5, 20 -24). (APS kinase from B. stearothermophilus has been reported not to be inhibited by high APS levels (25).) Incubation of the S107A enzyme with MgATP and APS in the absence of nuclease P1 yielded a 260 nm-absorbing substance with the same retention volume as authentic PAPS upon high performance anion exchange chromatography (10), further confirming that the observed activity was that of APS kinase.
As shown in Fig. 2, the mutant and wild-type enzymes show similar responses to changing ionic strength, viz., high salt shifts the v versus [APS] curves to the right. Because of the biphasic nature of the velocity curve, high salt "inhibits" the reaction at low APS concentrations, but appears to be an "activator" at high APS. Velocity measurements made at a single APS concentration in the region of the crossover would lead to the conclusion that ionic strength had little or no effect on activity. Kinetically, the shift in the velocity curve results mainly from changes in the two kinetic constants associated with APS. For the wild-type enzyme, the major effect of changing ionic strength is on K IB, app ; for the mutant enzyme, the major effect is on K mB, app (see below and Equation 4). The structural basis of the shift is as yet unknown, but it is noteworthy that high salt and mutation at position 107 have similar effects. Table I summarizes the kinetic properties of the recombinant enzymes. Compared with the wild-type enzyme, the mutant enzyme generally has higher kinetic constants. For example, under standard assay conditions at high ionic strength, the apparent Michaelis constant of the mutant enzyme and the apparent inhibition constant for APS are 12-and 1.7-fold, respectively, greater than those of the wild-type enzyme. MgATP binding (as measured by K ia ) is about the same for both enzymes, although the K mA of the mutant enzyme is 3-fold higher. (K ia is the E⅐MgATP dissociation constant, whereas K mA is a more complex constant that includes the rate constants for catalysis plus the release of products (Ref. 19, pp. 563-564). In the absence of high (NH 4 ) 2 SO 4 , K mB, app and K IB, app of the mutant enzyme are about 2 and 5 times greater, respectively, than the corresponding constants of the wild-type enzyme, whereas the K mP, app for PAPS is 12-fold greater.
Compared with the wild-type enzyme, S107A has a 60% higher theoretical V max, app in the forward direction at high ionic strength and 5 mM MgATP. This higher V max, app does not result from differences in the K mA for MgATP. Nor does it necessarily reflect a higher rate of phosphoryl group transfer from ATP to APS, because an increase in the rate constants for the release of PAPS and MgADP would have the same effect if these steps are rate-limiting in the overall reaction. Certainly, MgADP release must be partially rate-limiting and cannot be very much faster than PAPS release, otherwise there would be no significant steady state level of E⅐MgADP to bind APS and produce the observed substrate inhibition.
Chemical Modification of S107C-Another mutant enzyme, this one containing cysteine in place of serine at position 107, was also produced. S107C was active and displayed the usual biphasic (The wild-type enzyme, which does not contain cysteine, did not react with DTNB.) After complete modification with either DTNB or tetrathionate, S107C had no detectable activity. At 40 M DTNB (pH 8.1, 30°C), t1 ⁄2 for modification of the enzyme (4 M in sites) was 0.14 min in the absence of ligands (Fig. 3). In the presence of MgADP ϩ Mg 2ϩ (5 mM each) or MgATP ϩ Mg 2ϩ (5 mM each), the t1 ⁄2 values increased to 0.8 and 1.2 min, respectively. APS at 20 M had no effect by itself, but it enhanced the protection of Cys-107 provided by MgADP ϩ Mg 2ϩ , increasing t1 ⁄2 to 1.6 min. (This synergism is consistent with the ordered binding of the two nucleotides.) Qualitatively similar protection results were obtained when modification was monitored by the loss of activity at 0.6 M enzyme sites and 6 M DTNB. Although Ser-107 is not essential for activity, the above results suggest that it may be located in the substrate binding pocket. In that case, covalent modification of the analogous Cys-107 (leading to an increase in side chain bulk) could eliminate activity just by sterically blocking the entry of MgATP. Similarly, the binding of MgATP would block the access of the modifier. Of course, without a knowledge of the structure of the enzyme, there is always an alternative explanation for inactivation by modifiers and protection by reversibly bound ligands, viz., that the target residue is not near the substrate binding pocket, but its covalent modification induces a conformational change in the protein that prevents ligands from binding (or prevents catalytic residues from aligning properly). And conversely, ligand binding stabilizes a conformation of the enzyme that renders the target residue less accessible to modifiers.
P. chrysogenum APS Kinases Mutated at Other Potential Phosphoryl-accepting Residues-The catalytic activity of S107A eliminated Ser-107 as an indispensable residue in P. chrysogenum APS kinase. However, the results did not necessarily eliminate a mechanism that includes a phosphoenzyme intermediate. It was possible that the fungal enzyme has a tertiary structure that differs slightly from that of the E. coli enzyme, and as a result, a different side-chain is juxtaposed to the ␥-phosphoryl of bound ATP. The OH groups on Ser-97, Ser-99, Thr-103, Ser-104, and Tyr-109 were possibilities. The last three are predicted to be in the same kinked helical segment as Ser-107. Also, the first four are changed to alanine in the APS kinase-like C-terminal region of fungal ATP sulfurylase. Accordingly, five new recombinant fungal APS kinases were produced, with alanine replacing one of the serine or threonine residues noted above or with phenylalanine replacing Tyr-109. The activity of these mutant enzymes could be measured readily in unfractionated E. coli extracts. (The levels were about the same as in extracts of the strains containing the recombinant wild-type, S107A, and S107C enzymes.) Fig. 4 shows the normalized v versus [APS] profiles of the above mutant enzymes. S97A, S99A, and T103A behave like the wild-type enzyme under the same high ionic strength condition ([APS] opt ϭ 2.5-4 M). S104A at high ionic strength behaves like the wild-type enzyme at low ionic strength ([APS] opt ϭ 1 M) but is still sensitive to ionic strength: when the high (NH 4 ) 2 SO 4 was omitted, [APS] opt of this mutant enzyme decreased to 0.3 M. The velocity curve of the Y109F mutant is shifted far to the right and resembles a normal hyperbola. Nevertheless, substrate inhibition became obvious when the [APS] range was extended to 50 M ([APS] opt was approximately 15 M). Although the kinetic consequences of the mutations cannot yet be correlated with structural changes in the proteins, the results exclude Ser-97, Ser-99, Thr-103, Ser-104, and Tyr-109 (along with Ser-107) as compulsory phosphoryl acceptors in the APS kinase-catalyzed reaction.
P-loop Mutants-The APS kinase-like C-terminal region of fungal ATP sulfurylase (residues 396 -540) has a high degree of homology to the conserved central sequences of true APS kinases (Fig. 5). Yet fungal ATP sulfurylase has no APS kinase activity. If a phosphoserine intermediate is excluded in the reaction catalyzed by P. chrysogenum APS kinase, then the substitution of alanine for Ser-478 in fungal sulfurylase cannot be the reason for the absence of APS kinase activity. Most likely, any one of several changes can account for the inactivity of the C-terminal region. As suggested earlier (10), one such change may be in the region of the ATP P-loop (26 -28), which is GLSASGKST in fungal APS kinase (residues 32-40) 3 but is altered to GYMNSGKDA in fungal ATP sulfurylase (residues 402-410). To check this possibility, two APS kinases were constructed containing alterations in this region. These were mutant P1, which contained YMN in place of LSA at positions 33-35, and mutant P2, which contained DA in place of ST at positions 39 -40. Both mutant enzymes were expressed to the usual high level and in soluble form, but neither purified protein 3 The P-loop is often indicated as GXXXXGK(S/T) because the internal residues differ among different ATP-binding proteins. However, the X residues are highly conserved in APS kinases, which suggested that they may be just as essential for MgATP binding by this protein as the conserved Gly and Lys residues. A P-loop consensus for 14 of the 15 APS kinases whose sequences have been reported is GLS(G/A)(S/A)GK(S/T). This is followed by a T in all cases. (The Archeoglobus enzyme has a Pro in place of the Leu in the second position.)  had detectable APS kinase activity, presumably because MgATP could not bind properly (or at all). The limit of detection with the purified preparations was about 0.01 units ϫ mg of protein Ϫ1 .

DISCUSSION
As shown in Fig. 5, there is a high degree of sequence identity in APS kinases from various sources. This similarity in sequence implies similar structures and mechanisms. Consequently, it was expected that the reactions catalyzed by the P. chrysogenum and E. coli enzymes would be essentially the same, viz. that the phosphoryl transfer from ATP to APS would proceed via a phosphoenzyme intermediate as reported for the latter (5, 7). The only difference might be that in the reaction catalyzed by the E. coli enzyme, phosphoenzyme formation can occur before APS adds, whereas with the P. chrysogenum enzyme, phosphoryl transfer to the enzyme occurs only after formation of the ternary central complex. However, substitution of the putative phosphoryl-accepting residue of P. chrysogenum APS kinase (Ser-107) with either alanine or cysteine resulted in catalytically active enzymes with kinetic properties similar to those of the native and recombinant wild-type enzyme. Clearly, Ser-107 does not serve as an obligatory phos- gives the sequence of the APS kinase-like C-terminal region of P. chrysogenum ATP sulfurylase (U07353). Shaded areas show residues that are identical in at least 50% of the sequences. A. brasilense and R. meliloti APS kinases are encoded by part of the larger nodQ gene. Consequently, the sequences shown may include some residues upstream from the start of the APS kinase protein. The A. thaliana sequence may include some chloroplast transit peptide residues at the N-terminal end. The sequences of the Urechis and mouse enzymes may include some C terminus residues that link the APS kinase and ATP sulfurylase domains. (Both sulfate-activating enzymes are present in a bifunctional protein). The positions of suggested nucleotide triphosphate (26 -28) and PAPS (7) binding motifs are shown. phoryl acceptor in the reaction catalyzed by the fungal enzyme. Serine, threonine, and tyrosine residues located in the immediate vicinity of Ser-107 are also not essential for the reaction. If the phosphoryl acceptor of the E. coli enzyme was correctly identified and phosphoryl migration (29) did not occur during isolation or amino acid analysis of the [ 32 P]labeled peptide (7), then we must conclude that despite the strong sequence similarities of the fungal and bacterial APS kinases, the reaction mechanisms are different. Another (albeit remote) possibility that is not refuted by the present data is that phosphoryl transfer in APS kinases can proceed with equal facility by either of two simultaneously available routes: directly from ATP to APS or indirectly via a phosphoenzyme intermediate. This would require that Ser-107 be located at the active site, a requirement consistent with the observed inactivation of S107C by SH-targeted reagents and protection against inactivation by substrates.
The second major question-why does the C-terminal region of fungal ATP sulfurylase lack APS kinase activity?-can be answered in part. When P-loop stretch 33-35 (Leu-Ser-Ala) or 39 -40 (Ser-Thr) of P. chrysogenum APS kinase was changed to its ATP sulfurylase counterpart (Tyr-Met-Asn or Asp-Ala, respectively), a soluble but inactive protein was produced. Although either change was sufficient to eliminate catalytic activity in true APS kinase, we cannot conclude that they are solely responsible for the absence of APS kinase activity in ATP sulfurylase. To arrive at that conclusion, it would be necessary to change the corresponding Tyr-Met-Asn and Asp-Ala sequences of ATP sulfurylase to Leu-Ser-Ala and Ser-Thr, respectively, and demonstrate the acquisition of APS kinase activity. As noted earlier, multiple changes may have tailored the Cterminal domain of fungal ATP sulfurylase to serve as a PAPS receptor rather than as an APS kinase. For example, besides the modifications to the P-loop, there are several changes in the PAPS motif that may play a role in altering sulfonucleotide preference (see Fig. 5).
X-ray crystallographic studies on P. chrysogenum APS kinase and ATP sulfurylase are under way and may help clarify some of the structure-function relationships of the two sulfateactivating enzymes.