Efficient catalytic promiscuity in an enzyme superfamily: An arylsulfatase shows a rate acceleration of 1013 for phosphate monoester hydrolysis

Article abstract of DOI:10.1021/ja8047943

We report a second catalytic activity of Pseudomonas aeruginosa arylsulfatase (PAS). Besides hydrolyzing sulfate monoesters, this enzyme catalyzes the hydrolysis of phosphate monoesters with multiple turnovers (>90), a k_cat value of 0.023 s

Full text of DOI:10.1021/ja8047943

Published on Web 11/14/2008
Efficient Catalytic Promiscuity in an Enzyme Superfamily: An
Arylsulfatase Shows a Rate Acceleration of 1013 for
Phosphate Monoester Hydrolysis
Luis F. Olguin, Sarah E. Askew, AnnMarie C. O’Donoghue,† and Florian Hollfelder*
Department of Biochemistry, UniVersity of Cambridge, 80 Tennis Court Road,
Cambridge CB2 1GA, United Kingdom
Received June 23, 2008; E-mail: fh111@cam.ac.uk
Abstract: We report a second catalytic activity of Pseudomonas aeruginosa arylsulfatase (PAS). Besides
hydrolyzing sulfate monoesters, this enzyme catalyzes the hydrolysis of phosphate monoesters with multiple
turnovers (>90), a kcat value of 0.023 s-1, a KM value of 29 µM, and a kcat/KM ratio of 790 M-1 s-1 at pH
8.0. This corresponds to a remarkably high rate acceleration of 1013 relative to the nonenzymatic hydrolysis
[(kcat/KM)/kw] and a transition-state binding constant (Ktx) of 3.4 pM. Promiscuous phosphatase and original
sulfatase activities only differ by a factor of 620 (measured by kcat), so the enzyme provides high accelerations
for both reactions. The magnitudes and relative similarity of the kinetic parameters suggest that a functional
switch from sulfatase to phosphatase activities is feasible, either by gene duplication or by direct evolution
via an intermediate enzyme with dual specificity.
duplication,10-14 in which the second gene copy is released from
selective pressure and can accumulate mutations that eventually
Introduction
Most textbooks describe an enzyme as a catalyst with
exquisite specificity for just one preferred substrate that is
accepted out of the large number of alternative molecules in a
cell. However, there is a growing body of evidence suggesting
that the rule “one enzymesone activity” is not absolute. The
versatility of some lipases and esterases to process a variety of
substrates and the enhancement of these additional activities
by directed evolution are industrially important examples in this
respect.1 Even more remarkable than this broad substrate
specificity is the ability of one active site to facilitate the turnover
of several classes of substrates involving the making or breaking
of completely different types of bonds, a phenomenon called
catalytic promiscuity. Recent reviews list about two dozen
examples of such activities,2-7 but it is conceivable that a much
larger number of enzymes with promiscuous activities remain
undetected.7,8
lead to an enzyme with a new function. However, gene
duplication is a slow process (estimated to have a time scale of
millions of years in eukaryotic organisms).11 Therefore, it has
been suggested that the presence of inherent promiscuous
activities allows a more rapid response to an environmental
change.6,7,9,15,16 Directed-evolution and protein-redesign studies
have shown that many promiscuous activities can be enhanced
by one or a few mutations without abandoning the native
original activity, potentially giving an organism dual specificity
even before a process of gene duplication occurs.15,17 Alterna-
tively, it is also possible that evolution occurs directly (i.e.,
without duplication) through an evolutionary intermediate that
has multiple functions at a relevant level (although duplication
could still occur later, to provide a handle for more specific
regulation).14a,15 In view of this background, the number of
experimentally detected promiscuous activities to date may
represent only a small fraction of a more comprehensive system
of phylogenetic relationships for catalysis.
Jensen9 and later O’Brien and Herschlag7 have suggested that
catalytic promiscuity plays an important role in protein evolu-
tion. Evolution is thought to proceed through a process of gene
In this work, we set out to explore the catalytic promiscuity
of Pseudomonas aeruginosa arylsulfatase (PAS). PAS is a
sulfate monoester hydrolase that uses an unusual amino acid
modification to provide an active-site nucleophile. In its active
† Present address: Department of Chemistry, Durham University, South
Road, Durham DH1 3LE, United Kingdom.
(1) Bornscheuer, U. T.; Kazlauskas, R. J. Hydrolases in Organic Synthesis;
Wiley-VCH: Weinheim, Germany, 1999.
(10) Chothia, C.; Gough, J.; Vogel, C.; Teichmann, S. A. Science 2003,
300, 1701–1703.
(2) Bornscheuer, U. T.; Kazlauskas, R. J. Angew. Chem., Int. Ed. 2004,
43, 6032–6040.
(11) Lynch, M. Science 2002, 297, 945–947.
(12) Ohno, S. EVolution by Gene Duplication; Springer: New York, 1970.
(13) Zhang, J. Trends Ecol. EVol. 2003, 18, 292–298.
(14) (a) Bershtein, S.; Tawfik, D. S. Mol. Biol. EVol. 2008, 25, 2311–2318.
(b) Hurles, M. PLoS Biol. 2004, 2 (7), e206.
(3) Copley, S. D. Curr. Opin. Chem. Biol. 2003, 7, 265–272.
(4) Glasner, M. E.; Gerlt, J. A.; Babbitt, P. C. Curr. Opin. Chem. Biol.
2006, 10, 492–497.
(5) Hult, K.; Berglund, P. Trends Biotechnol. 2007, 25, 231–238.
(6) Khersonsky, O.; Roodveldt, C.; Tawfik, D. S. Curr. Opin. Chem. Biol.
2006, 10, 498–508.
(15) Aharoni, A.; Gaidukov, L.; Khersonsky, O.; Gould, S. M.; Roodveldt,
C.; Tawfik, D. S. Nat. Genet. 2005, 37, 73–76.
(16) Glasner, M. E.; Gerlt, J. A.; Babbitt, P. C. AdV. Enzymol. Relat. Areas
Mol. Biol. 2007, 75, 193–239.
(7) O’Brien, P. J.; Herschlag, D. Chem. Biol. 1999, 6, R91–R105.
(8) Yarnell, A. Chem. Eng. News 2003, 81 (49), 33–35.
(9) Jensen, R. A. Annu. ReV. Microbiol. 1976, 30, 409–425.
(17) Toscano, M. D.; Woycechowsky, K. J.; Hilvert, D. Angew. Chem.,
Int. Ed. 2007, 46, 3212–3236.
9
10.1021/ja8047943 CCC: $40.75
2008 American Chemical Society
J. AM. CHEM. SOC. 2008, 130, 16547–16555 16547

A R T I C L E S
Olguin et al.
site, a cysteine is post-translationally oxidized to a CR-
formylglycine (FGly) whose hydrated form coordinates to a
calcium ion. This metal positions and activates the FGly for
nucleophilic attack on the sulfur center of the substrate.18 The
oxidation of cysteine (or serine in other sulfatases) is carried
out by a range of modification systems,19-24 although the
specific modification machinery for PAS is not known. PAS
was found to hydrolyze several aromatic sulfate esters25 and is
overexpressed in response to sulfate starvation conditions;25,26
also, its gene (atsA) is part of an operon coding for sulfate-
starvation-induced proteins (a group of proteins that includes
other sulfatases, sulfate-binding proteins, and sulfate ester
transporters).26 Although its physiological substrate is not
known, these observations suggest that PAS has a genuine
physiological function as a sulfatase involved in sulfate
scavenging.
PAS has been identified as a member of the alkaline
phosphatase (AP) superfamily27,28 and is classified as a member
of this superfamily in the SCOP database.29 This assignment is
based on high structural homology between PAS and the name-
giving member of this superfamily, Escherichia coli AP. Despite
low sequence homology (only 16% identity over 218 residues),18
the root-mean-square deviation of CR atoms between PAS and
AP is only 3.3 Å over 218 residues.18 Specifically, sulfatases
and AP share a common R/ꢀ-fold core structure, conserved
metal-binding and active-site residues, and a similarly situated
nucleophile.27,28
Figure 1. PAS catalyzes hydrolysis of sulfate and phosphate monoesters.
In solution, these reactions share similar dissociative transition states.
Compound 1 is 4-nitrophenyl phosphate (PNPP), and compound 2 is
4-nitrophenyl sulfate (PNPS).
vastness of sequence space (i.e., the enormous combinatorial
diversity that defies screening or selection approaches without
further information and shortcuts).
The assignment of this class of sulfatases to the AP super-
family on the basis of catalytic as well as structural features
prompts the question of whether PAS may promote other types
of reactions in this superfamily35 despite differences in the
reaction centers of the substrates. Catalytic promiscuity within
this superfamily has been observed previously in E. coli AP,
which exhibits promiscuous sulfatase activity with a substantial
rate acceleration [(kcat/KM)/kw ≈ 109].35 Sulfate and phosphate
monoesters share tetrahedral geometry with comparable bond
angles and lengths.36 However, the amount and distribution of
charge37 is different: while sulfate monoesters bear one negative
charge, phosphate monoesters have two negative charges near
pH 7.0 and above. On the other hand, the uncatalyzed hydrolyses
of sulfate and phosphate monoesters occur with similar rate
constants [kuncat(PNPP)/kuncat(PNPS) ≈ 4]38 and proceed through
dissociative transition states (Figure 1). In each case, there is a
small extent of bond making to the incoming nucleophile and
a large extent of bond breaking to the leaving group.39-43 This
observation suggests that the similarity of the substrate structures
and transition states involved can also be used as a criterion to
predict efficient catalysis and define a mechanism-based
superfamily.35,36,44
Protein superfamilies have been defined by combinations of
criteria, such as sequence and structure30,31 or common mecha-
nistic steps,4,32-34 to provide frameworks for understanding the
evolutionary development of protein structure and function.
There is growing experimental evidence that many enzymes
within a superfamily show promiscuous activities toward the
natural substrates of other members of the same superfamily.6,7,16
Defining these promiscuous relationships may provide an
additional criterion to facilitate more efficient traversal of the
(18) Boltes, I.; Czapinska, H.; Kahnert, A.; von Bulow, R.; Dierks, T.;
Schmidt, B.; von Figura, K.; Kertesz, M. A.; Uson, I. Structure 2001,
9, 483–491.
Previous qualitative evidence suggests that sulfatases can also
hydrolyze phosphate esters. For example, a preparation of human
arylsulfatase B had phosphatase activity, although an impurity
as a possible source of this activity was not ruled out.45 On the
basis of this observation, it has been suggested that sulfatases
might be promiscuous phosphatases.35 More recently, the crystal
structure of human arylsulfatase A showed a covalently bound
(19) Benjdia, A.; Deho, G.; Rabot, S.; Berteau, O. FEBS Lett. 2007, 581,
1009–1014.
(20) Benjdia, A.; Leprince, J.; Guillot, A.; Vaudry, H.; Rabot, S.; Berteau,
O. J. Am. Chem. Soc. 2007, 129, 3462–3463.
(21) Carrico, I. S.; Carlson, B. L.; Bertozzi, C. R. Nat. Chem. Biol. 2007,
3, 321–322.
(22) Dierks, T.; Lecca, M. R.; Schlotterhose, P.; Schmidt, B.; von Figura,
K. EMBO J. 1999, 18, 2084–2091.
(23) Dierks, T.; Schmidt, B.; von Figura, K. Proc. Natl. Acad. Sci. U.S.A.
1997, 94, 11963–11968.
(24) Sardiello, M.; Annunziata, I.; Roma, G.; Ballabio, A. Hum. Mol. Genet.
2005, 14, 3203–3217.
(35) O’Brien, P. J.; Herschlag, D. J. Am. Chem. Soc. 1998, 120, 12369–
12370.
(25) Beil, S.; Kehrli, H.; James, P.; Staudenmann, W.; Cook, A. M.;
Leisinger, T.; Kertesz, M. A. Eur. J. Biochem. 1995, 229, 385–394.
(26) Hummerjohann, J.; Laudenbach, S.; Retey, J.; Leisinger, T.; Kertesz,
M. A. J. Bacteriol. 2000, 182, 2055–2058.
(36) Catrina, I.; O’Brien, P. J.; Purcell, J.; Nikolic-Hughes, I.; Zalatan, J. G.;
Hengge, A. C.; Herschlag, D. J. Am. Chem. Soc. 2007, 129, 5760–
5765.
(27) Galperin, M. Y.; Bairoch, A.; Koonin, E. V. Protein Sci. 1998, 7,
1829–1835.
(37) Nikolic-Hughes, I.; O’Brien, P. J.; Herschlag, D. J. Am. Chem. Soc.
2005, 127, 9314–9315.
(28) Galperin, M. Y.; Jedrzejas, M. J. Proteins 2001, 45, 318–324.
(29) Murzin, A. G.; Brenner, S. E.; Hubbard, T.; Chothia, C. J. Mol. Biol.
1995, 247, 536–540.
(38) See Table 2 and ref 36.
(39) Benkovic, S. J.; Benkovic, P. A. J. Am. Chem. Soc. 1966, 88, 5504–
5511.
(30) Orengo, C. A.; Thornton, J. M. Annu. ReV. Biochem. 2005, 74, 867–
900.
(40) Burlingham, B. T.; Pratt, L. M.; Davidson, E. R.; Shiner, V. J., Jr.;
Fong, J.; Widlanski, T. S. J. Am. Chem. Soc. 2003, 125, 13036–13037.
(41) Hoff, R. H.; Larsen, P.; Hengge, A. C. J. Am. Chem. Soc. 2001, 123,
9338–9344.
(31) Todd, E. A.; Orengo, C. A.; Thornton, J. M. J. Mol. Biol. 2001, 307,
1113–1143.
(32) Brown, S. D.; Gerlt, J. A.; Seffernick, J. L.; Babbitt, P. C. Genome
Biol. 2006, 7, R8.
(42) Kirby, A. J.; Jencks, W. P. J. Am. Chem. Soc. 1965, 87, 3209.
(43) Kirby, A. J.; Varvoglis, A. G. J. Am. Chem. Soc. 1967, 89, 415.
(44) O’Brien, P.; Herschlag, D. Biochemistry 2001, 40, 5691–5699.
(45) Bond, C. S.; Clements, P. R.; Ashby, S. J.; Collyer, C. A.; Harrop,
S. J.; Hopwood, J. J.; Guss, J. M. Structure 1997, 5, 277–289.
(33) Gerlt, J. A.; Babbitt, P. C. Curr. Opin. Chem. Biol. 1998, 2, 607–
612.
(34) Gerlt, J. A.; Babbitt, P. C. Annu. ReV. Biochem. 2001, 70, 209–246.
9
16548 J. AM. CHEM. SOC. VOL. 130, NO. 49, 2008

Efficient Catalytic Promiscuity in an Enzyme Superfamily
A R T I C L E S
Table 1. Summary of Kinetic Data for the Native and Promiscuous Activities of Wild-Type and Mutant PAS at pH 8.0a
native sulfatase activityb
kcat (s-1
promiscuous phosphatase activityc
kcat (s-1 cat/KM (M-1 s-1
k )
KM (µM)
)
k
cat/KM (M-1 s-1
)
KM (µM)
)
wild-type
C51S
C51A
0.29 ( 0.03
0.25 ( 0.06
0.40 ( 0.05
14.2 ( 0.6
(4.9 ( 0.8) × 107
(2.1 ( 0.5) × 104
(1.5 ( 0.4) × 102
29.1 ( 2.0
75.2 ( 5.8
0.023 ( 0.0006
790 ( 58
(5.4 ( 0.2) × 10-3
(9.5 ( 0.3) × 10-6
0.13 ( 0.01
(6.1 ( 1.3) × 10-5
-
-
-
d
d
d
a Conditions: 100 mM Tris, pH 8.0, 25 °C. b For the reaction with PNPS. c For the reaction with PNPP. d The detection limit for kcat/KM was a factor
of ∼2 × 105 less than the wild-type activity, providing an upper limit of 4 × 10-4 M-1 s-1 for kcat/KM for this mutant. Background activity and loss of
enzyme activity over longer time periods (>2 days) precluded reliable measurements.
phosphate attached to the active site FGly after incubation with
a phosphate monoester, pointing to at least a single turnover.46
We now provide further, quantitative evidence for the
postulated evolutionary relationship among the proteins of the
AP superfamily35 and show that PAS is an efficient phosphatase
that produces a rate acceleration of 1013. The relative magnitudes
of the promiscuous and native activities are remarkable. While
the kcat/KM values for the native and promiscuous activities of
AP differ by 9 orders of magnitude, they are in a much narrower
range for PAS: the kcat/KM values are within a factor of 104,
and the kcat values are 620-fold different, facilitating the
acquisition of a new function by provision of a head-start
activity.
Results and Discussion
PAS Is a Proficient Promiscuous Phosphatase. We observe
a strong promiscuous phosphatase activity in a highly purified
sample of PAS. Figure 2 shows a time course of the promiscuous
activity, which exhibits more than 90 turnovers per active site,
and a comparison with the uncatalyzed background reaction.
The promiscuous phosphatase reaction of PAS follows saturation
kinetics described by a KM value of 29.1 µM (100-fold larger
than that for the sulfate substrate) and a kcat value of 0.023 s-1
(620-fold smaller than that for the sulfate substrate) at pH 8.0.
A number of observations (outlined in the following para-
graphs) collectively suggest that the catalysis is not due to a
contaminant but is a genuine activity of PAS. These observations
include copurification of native and promiscuous activities on
three columns, coincidence of the two activities on a native gel,
decrease of native and promiscuous activities in active-site
mutants, and competitive inhibition of the native sulfatase
activity by the promiscuous phosphate substrate.
Figure 2. PAS is a promiscuous phosphatase. (A) Time course for the
reaction of PAS with PNPP (O) under conditions ([PNPP] ) 140 µM, [PAS]
) 1.5 µM) at which multiple turnovers of substrate are observed (>90 per
active site) compared with the uncatalyzed background reaction (0). (B)
Michaelis-Menten representation of initial-rate data, demonstrating satura-
tion kinetics ([PAS] ) 1.15 nM). Kinetic data are shown in Table 1.
Conditions: [Tris-HCl] ) 100 mM, pH 8.0, T ) 25 °C.
1. Copurification of Sulfatase and Phosphatase Activities. We
purified PAS using three different types of columns: anion-
exchange, hydrophobic-interaction, and size-exclusion chroma-
tography. The fractions of the respective elutions were assayed
for phosphatase and sulfatase activities. For each column
purification, the two activities were found to elute simulta-
neously, and the ratio of the two activities remained constant
(see Figure 3 and Figures S1 and S2 in the Supporting
Information). If a protein other than PAS were responsible for
the promiscuous activity, it would be expected not to coelute
with PAS every time, because it would be unlikely to interact
in equal measure with three column materials that employ
different types of interactions with proteins.47 If PAS had
attached a contaminant through specific interactions, it would
not be expected to yield a single band on an SDS-PAGE gel.
2. Analysis by Native Gel. In common with the SDS-PAGE
gels described above, a native gel of highly purified PAS shows
only one band, indicating high purity (see Figure 4A). Figure
4B,C shows incubations of native gels of PAS with phosphatase
and sulfatase substrates (indoxyl phosphate and sulfate, respec-
tively) that produce an insoluble, colored product, allowing a
qualitative activity assay. Only one colored band is visible in
each case, and the two spots appear at the same running distance,
showing that the phosphatase and sulfatase activities are caused
by the same protein. The running distance is also identical to
(46) Chruszcz, M.; Laidler, P.; Monkiewicz, M.; Ortlund, E.; Lebioda, L.;
Lewinski, K. J. Inorg. Biochem. 2003, 96, 386–392.
(47) E. coli is not known to have clearly identified sulfatases but has putative
sulfatase genes identified by homology.49,97,98 In the absence of the
plasmid pME4322, no sulfatase activity was observed in lysates of
the E. coli strain BL21(DE3) used for expression. We also observed
that crude lysates of E. coli BL21(DE3) overexpressing PAS showed
a 15-fold increase in their phosphatase activity compared with cells
treated the same way but harboring an “empty” pET vector lacking
the PAS gene. The observed increase suggests that overexpression of
PAS is responsible for the phosphatase activity.
9
J. AM. CHEM. SOC. VOL. 130, NO. 49, 2008 16549

A R T I C L E S
Olguin et al.
Figure 3. Copurification of native sulfatase and promiscuous phosphatase
activities of PAS. The chromatogram of the purification of PAS on a gel-
filtration column (Superdex 200 HR 10/30) shows that the protein eluted
as a single symmetrical peak and that sulfatase (O) and phosphatase (2)
activities coeluted in the same fractions. The inset shows an SDS-PAGE
gel of the protein recovered from the gel-filtration column, with a band
corresponding to the expected molecular mass of PAS (60 kDa).25 The two
other column purifications showed similar patterns of coelution and are
displayed in Figures S1 and S2 in the Supporting Information.
Figure 5. Proposed mechanism for the promiscuous phosphatase reaction
of PAS (modified from ref 18), which involves (A) attack of the hydrate of
a metal-coordinated formylglycine as the nucleophile followed by (B)
breakdown of the intermediate by hemiacetal cleavage. It should be noted
that the step (B) of the reaction would involve essentially identical Cꢀ-O
bond-breaking events (albeit with different leaving groups) for the native
and promiscuous reactions.
hydrated form of the formylglycine residue is believed to act
as the catalytic nucleophile in the arylsulfatase reaction, forming
a reaction intermediate.18,51 Formation of the intermediate has
been proposed to be followed by cleavage of the Cꢀ-O bond
of the hemiacetal (rather than the S-O or P-O bond) (Figure
5).18
We mutated the active-site nucleophile Cys51 in PAS to
serine and alanine. Neither mutant can be oxidized to FGly.49
Analyzing these mutants for sulfatase catalysis, we observed
similar KM values in the low µM range but reductions in kcat/
KM by factors of 2,330 in C51S and 327,000 in C51A (Table
1). The reduction in catalysis in the alanine mutant is comparable
to an analogous mutation in AP in which deletion of the
nucleophile in the S102A mutant led to a 105-fold reduction in
kcat/KM.52 For the promiscuous phosphatase catalysis by C51S,
the KM value is double that of the wild-type enzyme but is still
in the µM range. The kcat/KM value is reduced 6080-fold in
C51S. The C51A mutation reduced the activity below the
sensitivity of our assay. C51S exhibits more than 40 turnovers
for the native sulfatase activity but shows less than a single
turnover as a promiscuous phosphatase. The latter experiments
took days and were accompanied by activity loss of the enzyme,
limiting the dynamic range of our experiments.
The observation that both the native and promiscuous
reactions are slowed down by similar factors rules out the
possibility that a contaminant would have copurified with PAS.
A hypothetical contaminant would not be affected by a mutation
in PAS and thus would not give rise to such similarly large
decreases in the two activities. This observation is also consistent
with a crucial role of the formylglycine nucleophile (generated
from Cys51) in both reactions.
Figure 4. Native and promiscuous activities coincide in a native PAGE
gel (pH 8.5). The native PAGE gel was cut into three pieces and stained
with (A) Coomassie Blue stain, (B) a sulfatase substrate, and (C) with a
phosphatase substrate. The assay reactions in (B) and (C) rely on the
formation of an insoluble indigo blue dye (5) formed after hydrolysis of
indoxyl sulfate (4) or indoxyl phosphate (3).48
that of the spot in Figure 4A, indicating that the PAS protein
that brought about these spots, catalyzes both reactions.
3. Active-Site Mutants. PAS has a unique post-translational
modification in which the active-site Cys51 is oxidized to FGly.
E. coli can efficiently mature this residue to give complete
oxidation to formylglycine.49,50 We tested the presence of
unmodified cysteines by mixing a denatured sample of PAS
with 5,5′-dithiobis(2-nitrobenzoic acid) (DTNB). Only 4.3% of
the unique Cys residues (Cys51) in a PAS sample were found
to be unmodified (see Figure S4 in the Supporting Information),
which is within experimental error of full modification. The
4. Inhibition. Since the kcat values for the native and
promiscuous substrates differ by a factor of 620, the promiscu-
(48) Holt, S. J. In General Cytochemical Methods; Danielli, J. F., Ed.;
Academic Press Inc.: New York, 1958; Vol. 1, p 375-398.
(49) Dierks, T.; Miech, C.; Hummerjohann, J.; Schmidt, B.; Kertesz, M. A.;
von Figura, K. J. Biol. Chem. 1998, 273, 25560–25564.
(51) Recksiek, M.; Selmer, T.; Dierks, T.; Schmidt, B.; von Figura, K.
J. Biol. Chem. 1998, 273, 6096–6103.
(50) Szameit, C.; Miech, C.; Balleininger, M.; Schmidt, B.; von Figura,
K.; Dierks, T. J. Biol. Chem. 1999, 274, 15375–15381.
(52) Stec, B.; Hehir, M. J.; Brennan, C.; Nolte, M.; Kantrowitz, E. R. J.
Mol. Biol. 1998, 277, 647–662.
9
16550 J. AM. CHEM. SOC. VOL. 130, NO. 49, 2008

Efficient Catalytic Promiscuity in an Enzyme Superfamily
A R T I C L E S
Figure 7. pH dependence of kcat/KM for native sulfatase activity toward
PNPS (red b) and promiscuous phosphatase activity toward PNPP (blue
9). The shapes of the two activity profiles imply that negatively charged
substrates preferentially bind to protonated residues in the active site. The
larger slope in the profile of the phosphatase activity could represent the
involvement of two residues or the protonation of the PNPP substrate. Both
kcat/KM pH profiles were fit to the following equation:56 y ) (kcat/KM,max)/
Figure 6. Catalysis of PNPS hydrolysis by PAS is competitively inhibited
by PNPP. The fraction of activity is plotted as a function of PNPP
concentration at different fixed concentrations of PNPS. A KI value of 29.2
( 1.3 µM was obtained by simultaneously fitting all of the data to a
competitive inhibition model. This value is in good agreement with the KM
value of 29.1 ( 2.0 µM measured for PNPP hydrolysis (Figure 2).
Conditions: [Tris/HCl] ) 100 mM, pH 8.0, 25 °C. Different fixed amounts
of PNPS [(O) 0.13, (9) 0.32, (4) 0.64, (b) 1.27, (0) 3.18, and (2) 6.36
µM] were hydrolyzed in the presence of increasing amounts of PNPP. The
solid lines represent the best fit of the experimental data to a competitive
inhibition model.
[1 + 10pH-pK ].
a
toward further kinetic complexity that affects catalysis of sulfate
and phosphate hydrolysis differently. This difference could
represent the protonation of different residues in the free enzyme
(pKE) involved in the binding of a substrate carrying one (sulfate
ester) or two (phosphate ester) negative charges. In the case of
the phosphatase activity, the slope of log(kcat/KM) for the pH
profile is greater than 1, which could represent ionization of
two enzymatic groups or ionization of one enzymatic group and
protonation of the dianion of PNPP (pKa ) 4.79).55 However,
these scenarios are merely plausible and not conclusive.
ous substrate should not react on the time scale of the sulfatase
reaction. We were therefore able to probe the ability of a
phosphate monoester (i.e., the promiscuous substrate) to act as
a competitive inhibitor of the native reaction.53 Inhibition was
indeed observed, as shown in Figure 6. The data were fitted by
nonlinear regression analysis to a competitive inhibition model.54
The value of KI obtained for PNPP was 29.2 ( 1.3 µM, which
is in close agreement with the KM value measured for PNPP
hydrolysis (29.1 ( 2.0 µM). The competitive inhibition pattern
with almost identical KM and KI values for the phosphate
monoester is consistent with catalysis of both the native and
promiscuous activities occurring in the same active site.53
5. pH Rate Profiles. Identical pH dependences of native and
promiscuous activities has previously been used as a criterion
indicating that both reactions are carried out by the same active
site. Figure 7 shows the pH profiles of kcat/KM for the sulfatase
and phosphatase activities of PAS between pH 7.2 and 10, which
are in qualitative but not quantitative agreement. Over the
measured pH range, the kcat/KM values for the sulfatase and
phosphatase activities have their maximum values at pH 7.2
and then decline as the pH increases (Figure 7). Reliable data
above pH 10 could not be collected because the enzyme lost
activity in less time than that required to conduct an assay.
Below pH 7, the KM for PNPS was less than 0.1 µM; thus, the
errors in absorbance at concentrations around KM precluded
accurate measurements, and the limited range at which data
could be obtained necessarily limits the reliability of the
mechanistic implications. However, the apparent pKa values in
the kcat/KM pH profiles for the sulfatase and phosphatase
reactions (8.3 ( 0.1 and 6.3 ( 0.2, respectively) differ by 2
pH units. The dissimilarity of the kcat/KM pH profiles points
Rate Accelerations. A comparison of the kcat value with the
first-order rate constant of the corresponding uncatalyzed
reaction in water, kuncat, gives the first-order rate enhancement,
kcat/kuncat. The value of kcat/kuncat for the promiscuous phosphatase
activity is 8.4 × 106 (Table 2). This is an unusually large value
for a promiscuous activity, second only to the promiscuous
phosphodiesterase activity of a phosphotriesterase (kcat/kuncat
≈
6 × 108).7,57 Compared with the rate enhancement of PAS for
its native sulfate-transfer activity (kcat/kuncat ) 2.3 × 1010), this
value is smaller by only 3 orders of magnitude.
The catalytic proficiency is defined as the ratio of kcat/KM to
kuncat and gives information about the amount of transition-state
stabilization provided by a catalyst, relating to free rather than
bound substrate and enzyme.59 For PAS, the promiscuous
activity achieves a catalytic proficiency of ∼1011 M-1, which
can be compared with ∼1016 for the native activity. The inverse
of the catalytic proficiency is the transition-state binding
constant, Ktx, which we calculate as 3.4 pM for the promiscuous
activity. The catalytic proficiency of PAS is higher than that
obtained for the promiscuous activities of carbonic anhydrase60,61
(3.6 × 109 and 1 × 108 M-1 for its esterase and phosphate
(55) Zalatan, J. G.; Fenn, T. D.; Brunger, A. T.; Herschlag, D. Biochemistry
2006, 45, 9788–9803.
(56) Gibby, S. G.; Younker, J. M.; Hengge, A. C. J. Phys. Org. Chem.
2004, 17, 541–547.
(57) Shim, H.; Hong, S.; Raushel, F. J. Biol. Chem. 1998, 273, 17445–
(53) Cornish-Bowden, A. Fundamentals of Enzyme Kinetics, revised ed.;
Portland Press: London, 1995.
(54) Segel, I. H. Enzyme Kinetics; John Wiley & Sons, Inc.: New York,
1993.
17450.
(58) Wolfenden, R. Chem. ReV. 2006, 106, 3379–3396.
(59) Miller, B. G.; Wolfenden, R. Annu. ReV. Biochem. 2002, 71, 847–
885.
9
J. AM. CHEM. SOC. VOL. 130, NO. 49, 2008 16551

A R T I C L E S
Olguin et al.
Table 2. Values of First-Order Rate Enhancement (kcat/kuncat), Catalytic Proficiency [(kcat/KM)/kuncat], and Second-Order Rate Enhancement
[(kcat/KM)/kw] for the Native and Promiscuous Activities of Wild-Type PAS at pH 8.0a
c
activity
kcat/kuncat
(kcat/KM)/kuncat (M-1 b
)
(kcat/KM)/kw
Ktx (M)d
∆∆Gq (kcal/mol)e
sulfatase
phosphatase
2.3 × 1010 f
7.9 × 1016 f
4.3 × 1018
1.3 × 10-17
7.4
8.4 × 106 g
2.9 × 1011 g
1.6 × 1013 h
3.4 × 10-12
a For the reaction of PNPS (native) and PNPP (promiscuous). b The catalytic proficiency is defined as (kcat/KM)/kuncat
.
58 c For hydrolytic reactions, kw
d
) kuncat/(55 M).35 Ktx, the transition-state binding constant, was calculated as the inverse of the catalytic proficiency.58 e ∆∆Gq is the difference in
transition-state stabilization for sulfate monoester and phosphate monoester by PAS and was calculated using the following equation:35 ∆∆Gq ) -RT
ln{[(kcat/KM)/kw for PNPS]/[(kcat/KM)/kw for PNPP]} ) (0.001986 kcal mol-1 K-1)(298 K) ln[(4.3 × 1018)/(1.6 × 1013)] ) 7.4 kcal/mol. f Using kuncat
)
6.2 × 10-10 s-1 for PNPS at 25 °C.39 This value was obtained by correcting the original rate constant at 35 °C to a value at 25 °C using the following
equations: log(k2/k1) ) (Ea/2.303R)(T2 - T1)/(T1T2) and ∆Hq ) Ea - RT. g Using kuncat ) 2.7 × 10-9 s-1 for PNPP at 25 °C. This value was obtained
by correcting the original rate constant at 39 °C42 to a value at 25 °C using the same equations as for PNPS. h The kcat/KM values were larger at lower
pH, corresponding to a larger rate acceleration. However, the large errors in absorbance ascribed to the small values of KM at lower pH precluded more
accurate measurements.
monoesterase activities, respectively)62 and phosphotriesterase57
(1.6 × 1010 for its phosphodiesterase activity).63 Finally, the
second-order rate acceleration for the promiscuous activity of
PAS relative to the nonenzymatic attack by water for the same
reaction [i.e., the quantity (kcat/KM)/kw, where kw ) kuncat/(55
M)] has a value of 1.6 × 1013 (see Table 2). This benchmark
has been used for the promiscuous sulfatase activity of AP,
which gave a value of 109.35
The second-order rate acceleration exhibited by PAS for its
promiscuous activity is larger than the corresponding values for
many enzymes for their cognate reactions (3 × 1010 to 7.8 ×
1026).58,64
naturally evolved enzyme that had been removed in the
auxotroph.65 The kcat/KM value for the promiscuous phosphatase
reaction of PAS is only ∼4 × 104 smaller than the native
phosphatase activity of, e.g., AP44 (as an example for a relevant
phosphatase), and thus it is conceivable that it can potentially
carry out a useful function for an organism. Similar examples
in this respect are the roles played by the glucokinase activities
of four uncharacterized proteins66,67 and the phosphite-dependent
hydrogenase activity of AP.68
It is clear that PAS is an efficient catalyst for both reactions,
as the catalysis for both the native and promiscuous reactions
is substantial. The rate accelerations of the promiscuous activity
listed in Table 2 fall in the lower range of enzymatic rate
accelerations but are well above those exhibited by enzyme
models.69 The observation that the thermodynamic difference
between the catalysis of the two reactions (∆∆Gq ) 7.4 kcal
mol-1) is small compared to the overall transition-state stabi-
lization (∆Gq ) 25.4 kcal mol-1 for sulfate transfer and 18.0
kcal mol-1 for phosphate transfer by PAS)70 is also relevant
for the question of whether evolutionary specialization is
necessary, as catalysts with higher efficiency are evolved. It has
been assumed that negative trade-offs between different func-
tions are unavoidable as evolutionary specialization proceeds.71,72
Aharoni et al.15 have shown that such specialization is not
absolutely necessary and that coevolution of activities can be
observed to some extent, at least for thermodynamically less
demanding reactions. The example of PAS, an enzyme that also
has to be considered as the product of an evolutionary process,
shows that retention of two chemically distinct activities is also
possible for more difficult reactions that require more substantial
transition-state stabilization.
Implications and Conclusions
Relevance for Evolution of Function. Hydrolysis of mono-
phosphates by AP proceeds with a kcat/KM value of 3 × 107
M-1 s-1 35
In contrast, its promiscuous activities are smaller
.
by factors of 108 for diesterases (5 × 10-2 M-1 s-1 44
)
and 109
for sulfatase (1 × 10-2 M-1 s-1 35
)
with respect to the native
reactions. The comparison with PAS is striking: kcat/KM for the
promiscuous phosphatase activity of PAS is a factor of ∼6 ×
104 less than its native activity, and the difference between the
turnover numbers (kcat) for the native sulfatase and promiscuous
phosphatase activities is a factor of only 620. Both comparisons
must be considered to reflect relatively small differences
between chemically distinct promiscuous reactions catalyzed
with high rate accelerations.
The high level of the promiscuous activity and the relative
similarity of the kinetic parameters of the two reactions could
potentially facilitate an evolutionary transition between sulfate
and phosphate transfer. It is difficult to predict a priori whether
the level of this activity is immediately significant, but activities
with relatively small kcat/KM values (as small as ∼0.01 M-1
s-1) have been shown to provide a significant evolutionary
advantage to an auxotrophic E. coli strain.65 In that case, the
promiscuous activity of the rescuing enzyme was more than
107-fold smaller (as measured by kcat/KM) than that of the
Phosphate and sulfate monoesters have similar bond lengths,
geometries, and transition-state charge changes but differ in
substrate charge and the identity of atoms undergoing bonding
changes. PAS remarkably accepts these substrates with little
discrimination (7.4 kcal/mol), while in AP the same substrate
pair is differentiated by 12 kcal/mol.35 It will be interesting to
(60) Gould, S. M.; Tawfik, D. S. Biochemistry 2005, 44, 5444–5452.
(61) Koester, M. K.; Pullan, L. M.; Noltmann, E. A. Arch. Biochem.
Biophys. 1981, 211, 632–642.
(66) Miller, B. G.; Raines, R. T. Biochemistry 2004, 43, 6387–6392.
(67) Miller, B. G.; Raines, R. T. Biochemistry 2005, 44, 10776–10783.
(68) Yang, K.; Metcalf, W. W. Proc. Natl. Acad. Sci. U.S.A. 2004, 101,
7919–7924.
(62) These values were calculated using the first-order rate constants for
the hydrolysis of 4-nitrophenyl acetate (kuncat ) 5.5 × 10-7 s-1 99
)
and PNPP at pH 8.0 (kuncat ) 2.8 × 10-9 s-1 35
phosphate monoesterase (0.28 M-1 s-1) activities.61
)
and the respective
(69) Kirby, A. J. In Stimulating Concepts in Chemistry, 1st ed.; Vo¨gtle,
F., Stoddart, J. F., Shibasaki, M., Eds.; Wiley-VCH: Weinheim,
Germany, 2000; p 341.
kcat/KM values for the promiscuous esterase (2 × 103 M-1 s-1 60
)
and
(63) This value was calculated using the first-order rate constant for the
(70) The overall transition-state stabilizations were calculated using the
equation: ∆Gq ) 2.303RT log[(kcat/KM)/kw]. The transition-state
stabilizations obtained using ∆Gq ) 2.303RT log[(kcat/KM)/kuncat] were
22 kcal/mol for the sulfatase and 15 kcal/mol for the phosphatase.
(71) Partridge, L.; Barton, N. H. Nature 1993, 362, 305–311.
(72) Sgro, C. M.; Hoffmann, A. A. Heredity 2004, 93, 241–248.
hydrolysis of ethyl-4-nitrophenyl phosphate (kuncat ) 1 × 10-10
s-1 57
and the kcat/KM value for the promiscuous phosphodiesterase
)
activity (1.6 M-1 s-1).57
(64) Radzicka, A.; Wolfenden, R. Science 1995, 267, 90–93.
(65) Patrick, W. M.; Matsumura, I. J. Mol. Biol. 2008, 377, 323–336.
9
16552 J. AM. CHEM. SOC. VOL. 130, NO. 49, 2008

Efficient Catalytic Promiscuity in an Enzyme Superfamily
A R T I C L E S
explore the detailed factors contributing to the different degrees
of selectivity in the AP superfamily.
of a nucleophile activated by a metal ion invariably accelerates
hydrolytic reactions as long as the substrate is brought into
proximity, thus reducing the problem of catalysis to some extent
to a problem of binding (i.e., orientation and positioning of the
substrate with respect to the reactive nucleophile).
Molecules with multiple roles are rare in nature, and the need
for tight control over cellular function is thought to bias against
unselective catalysts.73,74 However, the relative lack of selectiv-
ity in PAS could make it a useful scavenging enzyme, enabling
an organism to be readily adaptable by hydrolyzing a wide range
of compounds.
General Criteria for Catalytic Promiscuity. Our observations
suggest that the promiscuous activity of PAS can be explained
by a number of perhaps more general criteria:
(i) Native and promiscuous reactions share key features. In
this case, the reactions share a trigonal-bipyramidal geometry
at the reaction center. In solution, the two types of reactions
proceed through similar dissociative mechanisms (characterized
by a small degree of bond making and a large amount of bond
breaking).39,42,43 If, to a first approximation, the transition states
of the catalyzed reactions are presumed to be of similar size
and geometry and differ only by one extra negative charge in
phosphate monoester dianion,36,41,75 then the high-energy species
arising during the course of the reaction can be stabilized by
similar interactions.
However, the requirement that the transition states have
similar natures (i.e., dissociative or associative) is not absolute.
AP achieves a higher rate acceleration for phosphate diester
hydrolysis (with an associative transition state) than for sulfate
ester hydrolysis.76 Some DNases are phosphomonoesterases in
addition to having endo/exonuclease activity.77 Likewise, purple
acid phosphomonoesterase acts on phosphate mono- and diester
substrates with similar kcat values.78
Alternatively, electrostatic interactions between the substrate
and the metal ion may differentiate between phosphate and
sulfate esters that carry different total charges, leading to
differences in their rate enhancements, as has been found for
AP.36,37
(ii) The catalytic motifs employed by PAS to promote sulfate
monoester hydrolysis are common to many other enzymes that
catalyze hydrolysis reactions. For example, provision of a
reactive metal ion has been shown in numerous model
systems79-82 to accelerate hydrolytic reactions by making a
highly reactive nucleophile available for the reaction. Notably,
several highly promiscuous enzymes (such as AP,7,35,44,83 serum
paroxonases,15,84,85 phosphotriesterase,57,86 and carbonic anhy-
drase60,87) contain metal ions in their active sites. The availability
The members of the AP superfamily for which promiscuity
has been detected35,44,55 are metalloenzymes that share active-
site features in addition to substantial structural homology.28,45,55,88
In a similar core domain,28 the active site of PAS possesses a
Ca2+ ion coordinated by three conserved aspartic acids and one
asparagine.18 In contrast, the active site of AP contains a Mg2+
and two Zn2+ ions, but the Zn2+-chelating residues Asp51 and
Asp369 in AP are superimposable with Asp13 and Asp317 that
chelate the Ca2+ atom in PAS. In both cases, metal ions interact
strongly with the negatively charged substrates and position the
catalytic nucleophile (Ser102 or FGly51) for attack.18,52,88,89 In
the position of the other two phosphatase metals, the sulfatases
have two conserved lysines, providing alternative Lewis acid
catalysts.88 In both enzymes, the catalysis involves a covalent
adduct: in the case of AP, a serine-phosphate intermediate is
broken down by cleavage of the P-O bond, predicting retention
of configuration (i.e., double inversion); in the case of PAS,
the FGly-sulfate adduct is instead broken down by cleavage
by the Cꢀ-O bond (resulting in single inversion).18,51 The
catalytic mechanism of PAS has been postulated to involve
general acid-base catalysis,18,56 while in AP, no candidate for
a general base has been identified. Despite small differences,
active-site functionality appears to be largely congruent in AP
and PAS.
Other structural features that contribute to catalysis may be
present in PAS and AP. In the case of AP, the tolerance for a
wide variety of substrates35,44,76,90,91 can be explained by its
relatively accessible binding site, which can accommodate
substrates with different steric demands by possibly allowing
different binding modes in a large binding pocket. The binding
pocket for PAS is not as wide open as that of AP but is still
large enough to readily accommodate different binding modes
for sulfate and phosphate esters. In addition, we have used
relatively reactive substrates, although less reactive phenyl
phosphates are also accepted by PAS.92
(iii) It is possible that the unique reaction mechanism of PAS
may be set up for particularly efficient promiscuity. First, the
enzyme takes advantage of nucleophilic catalysis by the unusual
geminal diol nucleophile instead of the much more common
side chains of serine, threonine, tyrosine, or unmodified cysteine.
It may be thermodynamically advantageous to utilize this
mechanism for sulfate-transfer hydrolysis, because a sulfate
intermediate would be harder to hydrolyze via cleavage of the
S-O bond. Other sulfatases without a formylglycine also prefer
to break the C-O bonds in sulfate esters instead of the stronger
S-O bonds.93-95 If the proposed two-step reaction catalyzed
by PAS is correct, the breakdown of the intermediate involves
(73) Bone, R.; Silen, J. L.; Agard, D. A. Nature 1989, 339, 191–195.
(74) Kondrashov, F. A. Nat. Genet. 2005, 37, 9–10.
(75) Hengge, A. C. Acc. Chem. Res. 2002, 35, 105–112.
(76) Zalatan, J. G.; Herschlag, D. J. Am. Chem. Soc. 2006, 128, 1293–
1303.
(77) O’Brien, P. J. Chem. ReV. 2006, 106, 720–752.
(78) Cox, R. S.; Schenk, G.; Mitic, N.; Gahan, L. R.; Hengge, A. C. J. Am.
Chem. Soc. 2007, 129, 9550–9551.
(79) Kofoed, J.; Reymond, J. L. Curr. Opin. Chem. Biol. 2005, 9, 656–
664.
(80) Morrow, J. R.; Iranzo, O. Curr. Opin. Chem. Biol. 2004, 8, 192–200.
(81) Suh, J. Acc. Chem. Res. 2003, 36, 562–570.
(88) Lukatela, G.; Krauss, N.; Theis, K.; Selmer, T.; Gieselmann, V.; von
Figura, K.; Saenger, W. Biochemistry 1998, 37, 3654–3664.
(89) Coleman, J. E. Annu. ReV. Biophys. Biomol. Struct. 1992, 21, 441–
483.
(82) Williams, N. H. Biochim. Biophys. Acta 2004, 1697, 279–287.
(83) O’Brien, P. J.; Herschlag, D. Biochemistry 2002, 41, 3207–3225.
(84) Aharoni, A.; Gaidukov, L.; Yagur, S.; Toker, L.; Silman, I.; Tawfik,
D. S. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 482–487.
(85) Harel, M.; Aharoni, A.; Gaidukov, L.; Brumshtein, B.; Khersonsky,
O.; Meged, R.; Dvir, H.; Ravelli, R. B.; McCarthy, A.; Toker, L.;
Silman, I.; Sussman, J. L.; Tawfik, D. S. Nat. Struct. Mol. Biol. 2004,
11, 412–419.
(90) Hollfelder, F.; Herschlag, D. Biochemistry 1995, 34, 12255–12264.
(91) Simopoulos, T. T.; Jencks, W. P. Biochemistry 1994, 33, 10375–10380.
(92) Dokphrom, U.; Hollfelder, F. Unpublished results.
(93) Bartholomew, B.; Dodgson, K. S.; Matcham, G. W.; Shaw, D. J.;
White, G. F. Biochem. J. 1977, 165, 575–580.
(94) Hanson, S. R.; Best, M. D.; Wong, C. H. Angew. Chem., Int. Ed. 2004,
43, 5736–5763.
(86) Roodveldt, C.; Tawfik, D. S. Biochemistry 2005, 44, 12728–12736.
(87) Christianson, D. W.; Cox, J. D. Annu. ReV. Biochem. 1999, 68, 33–
57.
(95) Pogorevc, M.; Kroutil, W.; Wallner, S. R.; Faber, K. Angew. Chem.,
Int. Ed. 2002, 41, 4052–4054.
9
J. AM. CHEM. SOC. VOL. 130, NO. 49, 2008 16553

A R T I C L E S
Olguin et al.
280 nm was determined according to a reported procedure,96 which
the same Cꢀ-O bond (rather than S-O or P-O) for both the
native and promiscuous reactions. Chemically, the reaction of
the second step (Figure 5, step B) then differs only in the leaving
group (phosphate or sulfate), but the requirement for general
acid-base catalysis of hemiacetal cleavage is identical. Thus,
the unusual nucleophile unifies the second step of the catalytic
cycle.
gave an ε280 value of 100 400 M-1 cm-1
.
The cysteine mutants C51S and C51A were prepared using the
QuikChange XL Site-Directed Mutagenesis Kit (Stratagene) with
pairs of complementary mutagenic primers containing the desired
mutation (C51Sfwd, 5′-cac cgc ctc gac ctc gtc gcc gac ccg ctc-3′;
C51Srev, 5′-gag cgg gtc ggc gac gag gtc gag gcg gtg -3′;
C51Afwd, 5′-aca ccg cct cga ccg cct cgc cga ccc gct-3′; and
C51Arev, 5′-agc ggg tcg gcg agg cgg tcg agg cgg tgt-3′). The
overexpression conditions for these mutants were modified to 30
°C and 0.5 mM of IPTG. Previous reports had mentioned that the
C51A mutant was degraded by the cells on overexpression,49 but
the reduction in temperature and IPTG allowed us to produce 16
mg of purified protein/L of culture.
This mechanistic feature together with the commensurate
modification of activity in the C51S and C51A mutants
discussed above suggests that at least the catalytic effects
associated with the nucleophile in the PAS active site promote
both activities to the same extent. This would mean that at least
one catalytic feature is intrinsically promiscuous, and given how
highly cooperative the interactions in enzyme active sites are,
the same may be true for the other catalytic effects in operation
in PAS. In addition, it has been proposed that general
acid-general base catalysis occurs in the sulfuryl transfer from
substrate to enzyme in PAS (His211 and Asp317)18 and in the
Aerobacter aerogenes sulfatase.56 This could potentially enhance
the rate of the first step of the native and promiscuous reactions.
Kinetic Measurements. Reactions were followed by measuring
the absorbance of the 4-nitrophenolate product at 405 nm in a
Spectramax plate reader (Molecular Devices) at 25 °C. All of the
reactions were performed in buffered solutions (100 mM Tris-HCl,
pH 8.0). The experiments were performed in triplicate and carried
out in a final volume of 220 µL in 96-well plates (Nunc). The
extinction coefficient for 4-nitrophenolate at pH 8.0 was determined
with a standard curve of known concentrations using a volume of
220 µL. Initial rates (less than 5-20% of the reaction) were
calculated from the linear part of the time course and then used to
construct Michaelis-Menten plots. Values for Vmax and KM were
obtained by fitting the data to the Michaelis-Menten equation using
a nonlinear least-squares fit (Kaleidagraph, Synergy Software).
Values of kcat were obtained by dividing Vmax by the corresponding
enzyme concentration. Enzyme concentration ranges of 0.1-0.5
and 20-50 nM were used in the measurements of the wild-type
sulfatase and promiscuous phosphatase activities, respectively. The
activities of the mutants C51S and C51A were measured in 1.0 cm
path length quartz cuvettes using the same equipment. The
concentration of enzyme in these assays was increased ∼100 and
1,000-fold for the C51S and C51A mutants, respectively. The
substrate concentration was typically varied from 0.2KM to 10KM,
provided that the limits of detection and solubility permitted it.
Higher substrate concentrations led to substrate inhibition. Dilutions
of the enzyme at concentrations lower than 500 nM led to rapid
loss of activity. Addition of 500 µg/mL bovine serum albumin
(BSA) stabilized the protein at low concentrations (<500 nM). BSA
itself did not accelerate the sulfatase or phosphatase reaction. PAS
copurified with Ca2+ in its active site,18,49 and this metal was not
added to the kinetic assays. Addition of further Ca2+ did not
enhance the enzymatic reaction, suggesting that the enzyme was
saturated. The pH profiles were constructed by measuring the
steady-state kinetic parameters over a pH range of 7.2 to 10 using
the following buffers: Tris (100 mM, pH 7-9) and CHES (100
mM, pH 9-10).
It will be interesting to test other sulfatases following the
same mechanism for their ability to catalyze phosphate transfer
as efficiently as PAS or to trace whether differentiation between
the two activities has taken place. It may be that for a further
increase in efficiency, more pronounced specialization is neces-
sary. However, PAS is an example where two activities are
catalyzed at a relatively high level, suggesting that the trade-
off between efficiency and specificity can be relaxed to create
an efficient yet multispecific enzyme.
An extrapolation from this single example suggests that
efficient promiscuity is most likely to be found when the
geometries of the transition states are similar, the substrates have
similar steric demands, the promiscuous reaction is similarly
or less thermodynamically demanding than the native one, and
(in a multistep reaction) some of the steps of the promiscuous
reaction are identical with steps in the native reaction. Further
work must be done to rank and quantify the contributions of
these criteria or ascertain their generality, but they may provide
a basis for systematic analysis and exploration of the ability of
existing protein structures to catalyze a wider range of reactions
than currently known.
Materials and Methods
Materials. The plasmid pME4322 was a gift from Dr. Michael
Kertesz.49 The sodium salt of 4-nitrophenyl phosphate hexahydrate
1, the potassium salt of 4-nitrophenyl sulfate 2, indoxyl phosphate
3, and indoxyl sulfate 4 were purchased from Sigma.
For the inhibition experiments, the data were fit by nonlinear
regression analysis to a competitive inhibition model54 using
Kaleidagraph. Alternative inhibition models did not converge with
the experimental data.
Overproduction and Purification of PAS and Its Mutants.
Overproduction of the wild-type sulfatase proteins was carried out
in E. coli BL21(DE3) harboring the plasmid pME4322 following
the procedure described in ref 49. The protein was purified in the
same way as described previously,49 except that an additional step
of gel filtration was performed using a Superdex 200 HR 10/30
column. The copurification data described in the Results and
Discussion were identical for virginal and reused columns, ruling
out possible “leaching” of phosphatases from the columns over time
as the source of the observed promiscuous activity. Enzyme-
containing fractions were concentrated and stored at -20 °C.
Typical yields of purified protein were in the range 18-23 mg of
protein/L of culture. The extinction coefficient of the protein at
Native PAGE Gel. A native PAGE gel at pH 8.5 was loaded
with 4 µg of purified PAS on each lane and run at 100 V. Next,
the gel was divided into three portions; one was stained with
Coomassie Blue solution, another incubated in 7 mL of a solution
containing indoxyl sulfate (30 µg/mL) in 50 mM Tris (pH 8.0) for
60 min, and the third incubated with indoxyl phosphate (400 µg/
mL) in 50 mM Tris (pH 8.0) for 6 h.
Acknowledgment. We thank Prof. Michael Kertesz for making
the plasmid PME45322 available to us and acknowledge prelimi-
nary experiments by Sean Wallace. L.F.O. acknowledges support
from the Consejo Nacional de Ciencia y Tecnologia de Mexico
(97) Hoffman, J. A.; Badger, J. L.; Zhang, Y.; Huang, S. H.; Kim, K. S.
Infect. Immun. 2000, 68, 5062–5067.
(98) Myette, J. R.; Shriver, Z.; Claycamp, C.; McLean, M. W.; Venkat-
araman, G.; Sasisekharan, R. J. Biol. Chem. 2003, 278, 12157–12166.
(96) Pace, C. N.; Vajdos, F.; Fee, L.; Grimsley, G.; Gray, T. Protein Sci.
1995, 4, 2411–2423.
9
16554 J. AM. CHEM. SOC. VOL. 130, NO. 49, 2008

Efficient Catalytic Promiscuity in an Enzyme Superfamily
A R T I C L E S
(CONACYT) and an Overseas Research Award. A.C.O. received
an EU Marie-Curie Fellowship. We thank Tony Kirby, Pat O’Brian,
Dan Tawfik, Jesse Zalatan, and members of the Hollfelder labora-
tory for comments on the manuscript. This work was supported by
the BBSRC, the Newton Trust, the MRC, and the EU network
ENDIRPRO. F.H. is an ERC Starting Investigator.
unmodified cysteines in PAS (Figure S3), wavelength scans of
product formation (Figure S4), and pH-KM and pH-kcat profiles
for sulfatase and phosphates activities (Figure S5). This material
is available free of charge via the Internet at http://pubs.acs.org.
JA8047943
Supporting Information Available: Chromatograms of the
purifications of PAS (Figures S1 and S2), determination of
(99) Jencks, W. P.; Carriuolo, J. J. Am. Chem. Soc. 1960, 82, 1778–1786.
9
J. AM. CHEM. SOC. VOL. 130, NO. 49, 2008 16555