Pure Appl. Chem., Vol. 81, No. 4, pp. 731–742, 2009.
doi:10.1351/PAC-CON-08-10-20
© 2009 IUPAC
Mapping catalytic promiscuity in the alkaline
phosphatase superfamily*
Stefanie Jonas and Florian Hollfelder‡
Department of Biochemistry, University of Cambridge, 80 Tennis Court Road,
Cambridge, CB2 1GA, UK
Abstract: “Promiscuous” enzymes possess activities in addition to their native ones.
Promiscuous activities could be remnants from an evolutionary ancestor that has been
adapted to fulfil a new function following gene duplication. Alternatively, the observation of
promiscuity could indicate that an enzyme has the potential to evolve into a new catalyst.
Thus, the observation of promiscuity defines functional relationships in enzyme superfamilies. Crosswise promiscuity can provide an additional layer of connectivity between members
of a—usually structurally defined—superfamily to establish a system for tracking the emergence and interconversion of enzymatic function. The systematic analysis of measured
promiscuous rates may serve as a basis for drawing up phylogenetic relationships based on
the potential for catalysis and may be useful for active use in directed evolution, suggesting
evolutionary “short cuts”. We review recent observations of catalytic promiscuity in members of the alkaline phosphatase (AP) superfamily that exhibit reciprocal relationships of
crosswise promiscuity with rate accelerations (kcat/KM)/k2 between 106 and 1018.
Specifically, we focus on the mechanistic features that appear to form the basis of catalytic
promiscuity in this superfamily.
Keywords: catalytic promiscuity; hydrolase; formylglycine; superfamily; metalloenzyme;
phosphate transfer; sulfate transfer; enzymology.
INTRODUCTION
The wealth of information derived from large-scale sequencing efforts and subsequent functional studies of ever more protein catalysts begs the question: Is our fundamental understanding of enzyme catalysis increasing? On one level the answer is no: The engineering of catalytic proteins with novel activities is still a daunting challenge. In favorable cases this challenge can be met by a combination of
rational design, protein engineering, and directed evolution, but such examples are rare [1,2]. More
often, design approaches simply fail to produce large improvements in activity. However, the definition
of protein superfamilies may provide a systematic framework for understanding the evolutionary development of protein structure and function. There are several definitions of superfamilies, using more
or less stringent criteria to group proteins according to function, structure, and conserved residues
which suggest a common evolutionary origin. Classification is usually based solely on structure and sequence criteria [3], without taking experimental data or prediction into account to assign a superfamily.
Examples for this approach are SCOP [4], CATH [5,6], or the SUPERFAMILY database [7]. Catalytic
*Paper based on a presentation at the 19th International Conference on Physical Organic Chemistry (ICPOC-19), 13–18 July
2008, Santiago de Compostela, Spain. Other presentations are published in this issue, pp. 571–776.
‡Corresponding author
731
732
S. JONAS AND F. HOLLFELDER
superfamilies by contrast share a common mechanistic feature [8,9], e.g., a chemical functionality to facilitate a common mechanistic step, and the ability to stabilize specific intermediates or transition states,
requiring a well-defined set of catalytic residues whose structural conservation is the basis of the assignment to a superfamily.
A third way of classifying enzymes into families could be by observation of functional promiscuity. Promiscuity is the ability of some enzymes to perform additional functions other than their native
ones. Despite the standard textbook description of enzymes as highly specific in their substrate recognition properties and in the reaction they perform there are now many examples that constitute exceptions to the rule “one enzyme—one activity”. In this review we focus on promiscuity in the alkaline
phosphatase (AP) superfamily.
Several types of enzyme promiscuity can be distinguished: substrate promiscuity refers to the
ability of enzymes to turn over a range of substrates by making or breaking the same type of bonds,
while catalytic promiscuity involves catalysis of chemically diverse reactions by making and breaking
different bonds in the same active site. An increasing number of enzymes identified as catalytically
promiscuous suggests that promiscuity is not a rare but a rather widespread intrinsic feature of enzymes.
Occasionally these enzymes display surprisingly high efficiencies for the additionally catalyzed reactions [10–13].
Jensen [14] and later O’Brien and Herschlag [11] proposed that these side activities of catalytically promiscuous enzymes might play an important role in the evolution of enzymes with new functions. Although the specific mechanisms involved are under discussion [15], it is widely accepted that
gene duplication is needed in order to relieve one copy of the gene of the selective pressure allowing it
to randomly mutate and evolve a new function [16]. However, the probability that a randomly drifting
gene accumulates unfavorable mutations that eventually are detrimental for the stability and structure
of the protein is high [17]. Thus, acquiring an entirely new function by one or even a few favorable mutations is extremely unlikely. Catalytic promiscuity would provide a head start for the evolution of a new
activity and require a smaller number of mutations to be enhanced up to a selectable threshold than
starting from an enzyme with no such advantage [11]. In directed evolution experiments it has been
shown that a promiscuous function can be substantially improved by only a few single mutations
[18,19] while maintaining the native activity [20–22], thus validating this scenario. This activity increase could be sufficient to confer a selective advantage to the organism [23], resulting in selection for
the duplicated gene and further improvements to the enzyme by evolution. Even rather weak activities
(as low as 0.01 M–1 s–1) have been shown to increase the fitness of the host organism, giving a lower
limit for an evolutionarily relevant promiscuous activity [24,25].
If new family members arise from the additional activities of their ancestors then crosswise
promiscuity within a family could be expected. The AP superfamily is based on common structural features [26], but its members also exhibit efficient promiscuous catalysis of the reactions promoted by
other members [27–33]. In this case, structure and a system of crosswise promiscuity suggest that its
members are evolutionarily related, and that this relationship can be characterized at the level of catalysis and protein fold.
Further mechanistic information and the identification of promiscuous catalytic superfamilies
may provide evolutionary short cuts for the identification and evolution of new functions as well as an
understanding of how enzymatic functionality is interconverted.
ESTABLISHING AN ENZYME AS CATALYTICALLY PROMISCUOUS
Unlike in synthetic chemistry, where spectroscopic techniques such as NMR can report on the identity
and purity of the compounds made, such positive detection is impossible for enzymes. Given that the
rates of primary and promiscuous activity differ by many orders of magnitude (see Table 2), the question of possible minute impurities that could lead to the observed side activities becomes relevant. The
© 2009 IUPAC, Pure and Applied Chemistry 81, 731–742
Multiple catalytic promiscuity
733
following tests have been established to ascertain that the observed second activity is a genuine function of the enzyme studied [27–29,33]:
i.
ii.
iii.
iv.
Copurification of the protein and all catalytic activities over several purification steps suggest that
the same protein is responsible for all activities since a contaminant would not be expected to
have the same physical and chemical properties.
Cross inhibition of the native activity by the substrates for the weaker activities suggests that all
reactions are catalyzed by the same active site.
Coincidence of pH profiles for all activities indicates that the same residues in the active site play
similar roles during catalysis of all reactions. Quantitative agreement suggests that the active site
functionality is used in the same way, qualitative agreement indicates mechanistic differences between native and promiscuous reactions.
Finally, deleterious mutations in the active site should result in a decrease of all activities. No single observation in itself constitutes “proof”, but success with a number of tests collectively indicates a bona fide dual function.
THE ALKALINE PHOSPHATASE SUPERFAMILY
Originally structural features had been used to group a diverse set of metalloenzymes together [26,34]:
hydrolases, isomerases, and transferases which act on phospho- or sulfocarbohydrate substrates. The
name-giving member of this superfamily, alkaline phosphatase (AP), is a non-specific phosphate monoester hydrolase which has been extensively studied [35–37]. Other members include nucleotide pyrophosphatases/phosphodiesterases (NPPs), cofactor-independent phosphoglycerate mutases (iPGMs),
phosphonoacetate hydrolases, phosphonate monoester hydrolases (PMHs), and a large number of sulfatases, such as arylsulfatases (ASs) [26,34,38–40].
Table 1 shows structures of selected members of the AP superfamily including their active sites
and their native and promiscuous activities. The superfamily members share very limited sequence
homology, so alignments are based on broad structural similarities (discussed below). Other characteristics do not immediately suggest a close relationship: AP has three metal ions (of which at least two
are involved in the catalytic mechanism), other members such as the ASs only bear one metal ion.
Specifically, the active site residues are different: besides metal and a nucleophilic group, the residues
lining the pocket diverge considerably. However, the character of the active site residues is conserved:
the second metal ion in AP is substituted by a cationic group in AS A in a very similar position [41].
This functional rather than literal conservation of residues suggests that the common theme might be
the rather similar catalytic tasks carried out by this superfamily: substrates are predominantly negatively
charged with phosphorus and sulfur-containing groups and hydrolytic group-transfer reactions dominate. So perhaps an equally defining common characteristic of the AP superfamily is the similarity of
its functional catalytic features.
© 2009 IUPAC, Pure and Applied Chemistry 81, 731–742
Phosphodiesterase,
phosphonate monoesterase,
sulfatase
[27,28,42]
Promiscuous activities
References
structural features are highlighted in red.
Phosphomonoesterase
Native activity
aConserved
Ca2+
2 Zn2+, Mg2+
Active site metal
[29,30,43]
Phosphomonoesterase,
phosphodiesterase
Sulfatase
Formylglycine (fGly51)
Pseudomonas aeruginosa
Serine (Ser102)
Escherichia coli
Arylsulfatase (AS)
Nucleophile
Active site
Structurea
Organism
Alkaline phosphatase (AP)
Table 1 Structures of AP superfamily members that are catalytically promiscuous.
[33,44]
Phosphomonoesterase,
sulfatase,
sulfonate monoesterase
Phosphonate monoesterase/
phosphodiesterase
Mn2+
Formylglycine (fGly57)
Rhizobium leguminosarum
Phosphonate monoester hydrolase
(PMH)
[31,32]
Phosphomonoesterase,
sulfatase
Phosphodiesterase
2 Zn2+
Threonine (Thr90)
Xanthomonas axonopodis
Nucleotide pyrophosphatase/
phosphodiesterase (NPP)
734
S. JONAS AND F. HOLLFELDER
© 2009 IUPAC, Pure and Applied Chemistry 81, 731–742
Multiple catalytic promiscuity
735
CONSERVED STRUCTURE IN THE ALKALINE PHOSPHATASE SUPERFAMILY
The members of the AP superfamily are largely globular, mixed α/β-proteins with a conserved central
mixed β-sheet of eight strands sandwiched between two layers of α-helices [26,31,34,38,39,44,45].
Subfamilies are formed by adding large insertions or whole domains to this core structure, e.g., an additional C-terminal α/β-domain in the AS family. As it is the case for most α/β-proteins [46], the active site is positioned between the loops coming from sheet 1 and the adjacent sheet 6 (Fig. 1).
Active sites contain one (AS: Ca2+, PMH: Mn2+) [43,44,47], two (NPP: Zn2+, iPGM: Mn2+)
[31,48], or three (AP: 2 Zn2+, 1 Mg2+) [49,50] metal ions. The metal site common to all family members is the one coordinating and activating the nucleophile, the position of the second metal ion present
in NPP, iPGM, and AP is occupied by a basic residue in AS/PMH, taking on a similar role. The metal
binding residues are largely conserved throughout the superfamily. All enzymes contain a nucleophilic
residue (AP and iPGM: Ser, NPP: Thr, AS and PMH: formylglycine, posttranslationally formed from
Ser or Cys [40,51,52]) at the bottom of the active site cavity, which is activated by metal coordination.
Fig. 1 Conserved structural elements in the AP superfamily. Family members are characterized by three core layers
α/β/α, a central mixed β-sheet of eight strands in the order 43516728 (strand 7 is antiparallel to the rest) and a
conserved metal binding site positioned between the loops coming from strands 1 and 6.
It is important to bear in mind that neither structure nor function evolves independently, so the
conceptual difference in functional and structural superfamilies is an artificial distinction: evolution will
take advantage of and select for both.
CONSERVED MECHANISTIC FEATURES IN THE ALKALINE PHOSPHATASE
SUPERFAMILY
Like most metallohydrolases the proteins in this superfamily benefit from the availability of a nucleophile activated (i.e., deprotonated) by general base catalysis and metal coordination, and from the offset of negative charge development on the leaving group of phosphoryl/sulfuryl oxygens by Lewis acid
catalysis [53–56]. This matches the provision of cationic amino acid side chains and metal ions in the
respective active sites.
The unifying mechanistic feature of the group transfer reactions is a double displacement or pingpong mechanism starting with a nucleophilic attack on the central phosphorus or sulfur atom by the
metal activated nucleophilic residue (Fig. 2) [37,38,43]. The transition state is stabilized by coordination to the metal ion(s) and other residues lining the catalytic pocket [57–60]. After departure of the protonated alcohol leaving group, a covalent P/S-nucleophile intermediate is formed which has been
trapped experimentally for some family members [39,61,62]. In AP, NPP, and iPGM this intermediate
is broken down by a second nucleophilic attack on the central P/S by an incoming water or alcohol molecule. In the case of AS and PMH the unusual formylglycine residue allows for an intramolecular hydrolysis of the hemiacetal helped by general base catalysis, avoiding a second direct attack at P/S
© 2009 IUPAC, Pure and Applied Chemistry 81, 731–742
736
S. JONAS AND F. HOLLFELDER
Fig. 2 Schematic of the reaction mechanism of AP (A) and PMH (B) showing the formation of a covalent
intermediate and its breakdown by an inter- (A) or intramolecular (B) step. Metal ions act as Lewis acids to lower
the pKa of the nucleophile, stabilize charges, to bind the substrate and—in AP—to offset charge on the leaving
group. In PMH and AS breakdown of the covalent intermediate is essentially occurring by breaking the same
bonds, regardless of which promiscuous substrates has been turned over: only the leaving groups (in this case, the
phosphonate, for sulfate monoester substrates, the sulfate, etc.) are different.
(Fig. 2) [44,58]. Thus, regardless of the P/S-nucleophile intermediate formed for different promiscuous
substrates, this step remains identical [29,33].
HIGHLY EFFICIENT PROMISCUOUS CATALYSIS IN THE ALKALINE PHOSPHATASE
SUPERFAMILY
O’Brien and Herschlag pioneered the functional exploration of this superfamily and detected additional
promiscuous activities in AP, namely, sulfate monoesterase [27], phosphate diesterase, and phosphonate
monoesterase [28] activities in addition to its original phosphate monoesterase activity. Later, an additional phosphite-dependent hydrogenase activity was found [63], giving AP no less than five activities.
Remarkably, the rate accelerations involved were substantial, with second-order rate enhancements
(kcat/KM)/k2 ranging up to 1011 (Table 2).
© 2009 IUPAC, Pure and Applied Chemistry 81, 731–742
kcat
(s–1)
1.4 × 101
5.5 × 10–1
2.3 × 10–2
n.d.
n.d.
n.d.
5.1 × 10–10i
1.1 × 10–11e
2.8 × 10–9c
1.1 × 10–11e
2.8 × 10–9c
5.1 × 10–10i
Sulfatasej (4)
Phosphodiesterased (2)
Phosphomonoesteraseb (1)
Phosphodiesterased (2)
Phosphomonoesteraseb (1)
Sulfataseh (4)
Arylsulfatase (AS)
Nucleotide pyrophosphatase/
phosphodiesterase (NPP)
2.3 × 103
1.1
2 × 10–5
4.9 × 107
2.5 × 105
7.9 × 102
9.2 × 103
1.5 × 104
4.9 × 101
2.2 × 101
5.6 × 10–1
3.3 × 107
5 × 10–2
3 × 10–2
1 × 10–2
kcat/KM
(M–1 s–1)
n.d.
n.d.
n.d.
3 × 1010
5 × 1010
8 × 106
2 × 1013
2 × 1011
2 × 106
2 × 105
4 × 107
1 × 1010
n.d.
n.d.
n.d.
kcat/kuncat
5 × 10–15
3 × 10–9
3 × 10–5
1 × 10–17
4 × 10–17
4 × 10–12
3 × 10–17
1 × 10–15
1 × 10–10
2 × 10–10
2 × 10–9
8 × 10–17
2 × 10–10
1 × 10–9
5 × 10–8
Ktx (M)
1 × 1016
2 × 1010
2 × 106
5 × 1018
1 × 1018
2 × 1013
2 × 1018
5 × 1016
5 × 1011
3 × 1011
3 × 1010
7 × 1017
3 × 1011
4 × 1010
1 × 109
(kcat/KM)/k2a
[31,76]
[31,65]
[27,32]
[27,29]
[30,76]
[27,29,65]
[33,78]
[33]
[30,33]
[33,65,79]
[33,73,80]
[27,65,68,75]
[28,76]
[28,33]
[27,77]
Ref.
ck
H2O
phosphate (pNPP) at 25 °C, pH 8.0.
for hydrolysis on pNPP corrected to 25 °C using the reported temperature dependence, pH 8.3, obtained directly and by extrapolating rates in the presence of amines to zero amine
concentration.
dHydrolysis of bis-p-nitrophenyl phosphate (BpNPP) at 25 °C, pH 8.0.
ek
H2O for hydrolysis of BpNPP corrected to 25 °C using the reported temperature dependence, pH 7.0.
fHydrolysis of p-nitrophenyl phenyl phosphonate (pNPPPn) at 25 °C, pH 8.0.
gk
uncat for hydrolysis of pNPPPn, estimated from the value at 30 °C, obtained by extrapolating kOH to pH 8.0.
hHydrolysis of p-nitrophenyl sulfate (pNPS) at 25 °C, pH 8.0.
ik
H2O for hydrolysis of pNPS corrected to 25 °C, pH 8.0 obtained directly and by extrapolating rates in the presence of amines to zero amine concentration.
jHydrolysis of ethyl-p-nitrophenyl phosphate (EtpNPP), pNPPPn, p-nitrophenyl phenyl sulfonate (pNPPSn), pNPP, and pNPS at 30 °C, pH 7.5.
kExtrapolated to 30 °C using the published temperature dependence for k
OH and extrapolation to pH 7.5.
lObtained by extrapolating k
OH at 30 °C to pH 7.5.
mDerived from k
OH and its temperature dependence extrapolated to 30 °C, pH 7.5.
nDerived from the rate constants for the mono- and dianionic states of pNPP, assuming the pK of the equilibrium between mono- and dianion of pNPP to be 5. The values for k
a
monanion
and kdianion were calculated from the published ∆H‡ and ∆S‡ values for both species.
oDerived assuming that pNPS hydrolysis is mostly pH-independent at pH 4–12.
ak = k
2
uncat/55 M = kw.
bHydrolysis of p-nitrophenyl
5.8
2.7
1.2 × 10–2
7.7 × 10–3
4 × 10–2
2.6 × 10–13k
1.7 × 10–11l
5.5 × 10–9m
4.3 × 10–9n
1.1 × 10–9o
Phosphodiesterase (2)
Phosphonatemonoesterase (3)
Sulfonatemonoesterase (5)
Phosphomonoesterase (1)
Sulfatase (4)
3.6 × 101
n.d.
n.d.
n.d.
kuncat
(s–1)
Phosphomonoesteraseb (1) 2.8 × 10–9c
1.1 × 10–11e
Phosphodiesterased (2)
Phosphonatemonoesterasef (3) 4 × 10–11g
5.1 × 10–10i
Sulfataseh (4)
Activity
Phosphonate monoester hydrolase/
phosphodiesterase (PMH)j
Alkaline phosphatase (AP)
Enzyme
Table 2 Catalytically promiscuous enzymes of the AP superfamily with their respective first- and second-order rate accelerations.
Multiple catalytic promiscuity
© 2009 IUPAC, Pure and Applied Chemistry 81, 731–742
737
738
S. JONAS AND F. HOLLFELDER
The additional reactions mirrored the functions of other members of this superfamily, so we explored whether a structurally homologous family member, the AS from Pseudomonas aeruginosa
(PAS), would also exhibit promiscuity. This is indeed the case: PAS accelerates the hydrolysis of phosphate monoesters (Fig. 3, 1) by 1013 [29]. Even more remarkable, the hydrolysis of phosphate diesters
(2) is accelerated by factors between 1015 and 1018, depending on the structure of the diester substrate
used.
Fig. 3 Substrates whose hydrolytic breakdown is catalyzed by enzymes in the AP superfamily.
A further member of the AP superfamily, a PMH, was crystallized by our group [33,44] and subjected to activity tests with substrates similar to those used elsewhere in the superfamily [33]. PMH
shows activity toward the originally assigned cognate substrate phosphonate monoester (3) [47] as well
as phosphate mono- (1) and diester (2), sulfates (4), and sulfonates (5) (Fig. 3). The respective rate accelerations ranging from 1010 to as high as 1018 are summarized in Table 2.
Despite the common net hydrolysis reaction in the AP superfamily, this collection encompasses
a range of diverse substrates: their charges range between 0 and –2, the reactions involve transition
states of a different nature (in terms of bond-making and -breaking) [64–66], attack at two different reaction centers (P and S), and diverse intrinsic reactivities (with half lives between 20 and 85 000 years
under near neutral conditions, see Table 2). The mechanistic question of whether a promiscuous enzyme—in this case AP—changes the intrinsic bond-making and -breaking parameters (Brønsted values) or whether it can accept transition states of different nature has been addressed [67]. A comparative transition-state analysis of the phosphate diesterase and monoesterase activities of AP suggests that
the transition states of both reactions are not unified for the enzymatic reaction. Even though AP is evolutionarily optimized for the dissociative transition state of phosphate monoester hydrolysis [64,68,69],
it can still confer substantial catalysis for the associative transition state of the diester reaction [28].
The rate accelerations observed for AP, PAS, PMH, and NPP surpass the accelerations of moderately efficient natural enzymes [70–72]. They are remarkably versatile chemical machines that manage
to catalyze difficult chemical reactions and do not seem to forfeit catalytic efficiency for the very broad
observed specificity. This would render them useful as scavenging enzymes able to hydrolyze a wide
range of compounds.
There are also evolutionary implications: For PAS, the kcat/KM values are within a range of 105
and the kcat values within 103 for all three activities. When an enzyme evolves a new activity it must
confer a selective advantage to the host. Although it is difficult to predict which level of activity would
endow an organism with a selective advantage, the transition to a new activity will be facilitated if the
promiscuous activity used as the starting point for evolution is high or even similar to the original activity. This would render PAS an advantageous target for evolution.
The observation of efficient catalytic promiscuity also raises the question of the structural and
mechanistic basis for enzyme specificity. In AP, several specificity-determining features have been
identified: the dinuclear Zn2+-center, the substrate binding arginine, and a Mg2+-coordinated water molecule hydrogen-bonding to the substrate. The efficiency of AP for the different reactions correlates with
© 2009 IUPAC, Pure and Applied Chemistry 81, 731–742
Multiple catalytic promiscuity
739
the charge sequestered on the non-bridging oxygen between the two Zn-ions [73,74]. Furthermore, bifurcate substrate binding by Arg166 and the charge stabilization by the hydrogen-bonding water molecule provided by Mg2+ both favor doubly charged substrates [31,60]. For PAS and PMH, no such correlations have been observed yet, but experiments are underway.
ORIGINS OF PROMISCUITY
The comparison of the catalytically promiscuous members of this superfamily suggests a list of possibly general criteria:
i.
ii.
iii.
iv.
Native and promiscuous reactions share some key features. All reactions are hydrolytic, involve
a trigonal–bipyramidal geometry at the reaction center and benefit from the possibility of nucleophilic and general acid/base catalysis. At the same time, some rather more subtle mechanistic features influence the degree of promiscuity very little: Both AP and PAS achieve a higher rate acceleration for phosphate diester hydrolysis (with an associative transition state) than for their
promiscuous sulfate or phosphate monoesterase reaction (with a dissociative transition state resembling the transition state of their native reactions) [29,30,67]. So the nature of the transition
state appears to be relatively unimportant. Also, electrostatic interactions between substrate and
the metal ions that have been found to differentiate between phosphate and sulfate esters in AP
[73,74] are not important to the same extent in PMH or PAS. This means that complete coincidence of the catalytic factors is not important—a subset of similarity seems sufficient.
A reactive active site. In no small part, the observed catalysis must be due to the intrinsic reactivity of the metal ion bound in the active site. The catalytic motif of two metal ions has been
shown in model systems to accelerate hydrolytic reactions by lowering the pKa of the nucleophile
and thus increasing the concentration of reactive, deprotonated nucleophile available for the reaction [81–84]. Notably, several highly promiscuous enzymes (such as AP [11,27,28,68], serum
paraoxonase [20,85,86], phosphotriesterase [87,88], and carbonic anhydrase [21,89]) contain
metal ions in their active sites. The availability of a nucleophile activated by a metal ion will invariably accelerate 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).
All members of the AP superfamily are metalloenzymes, although the number and identity
of metal ions differs. The metal may interact strongly with the negatively charged substrates and
position the catalytic nucleophile—serine, threonine, or formylglycine—for attack. The microenvironment created by three metal ions in AP has been shown to make catalysis sensitive to
charge sequestered between the metal ions [73], but the single-metal sites of PAS and PMH show
no such effect [33].
A sufficiently spacious active site. AP, PAS, and PMH all possess a relatively accessible binding
site that can accommodate substrates with different steric demands by possibly allowing different
binding modes in a large binding pocket. The binding pockets of AP and PMH have a wider diameter (see Fig. 4) and PMH is also deeper, but even the binding cleft of PAS is still large enough
to readily accommodate the very small differences in size between sulfate and phosphate esters.
A recyclable nucleophile. In the case of PMH and PAS, the rates of native and promiscuous reactions cover a relatively narrow range. This suggests that the reaction mechanism of these enzymes may be set up for particularly efficient promiscuity. 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 cysteine. It may be thermodynamically advantageous
to utilize this nucleophile for sulfate transfer hydrolysis, because a sulfate intermediate would be
harder to hydrolyze via cleavage of the S–O bond. Furthermore, the breakdown of the intermediate involves the same C–O bond, rather than S–O or P–O, for both native and promiscuous reac-
© 2009 IUPAC, Pure and Applied Chemistry 81, 731–742
S. JONAS AND F. HOLLFELDER
740
tions. Chemically, this second step (Fig. 2) then differs only in the leaving group, but the requirements for general acid/base catalysis of hemiacetal cleavage are identical. Thus, the formlyglycine nucleophile unifies the second step of the catalytic cycle for all reactions.
Further work has to be done to rank and quantify the contributions of these criteria or ascertain
their generality, but for the moment they provide a working hypothesis for explaining the unusual
promiscuity of the AP superfamily. More generally, the criteria discussed serve to define a chemical dimension to the evolution of function, based on economical conservation of catalytic features.
Fig. 4 Active sites openings of members of the AP superfamily. AP (A) has a very shallow active site (the surface
of the active site residues are colored black). Cuts through the protein surface of PMH (B) and PAS (C) show that
their active sites are buried deeper inside the protein with PAS possessing a comparably narrow cleft.
ACKNOWLEDGMENTS
S.J. was supported by a doctoral fellowship from the German National Academic Foundation
(Studienstiftung des deutschen Volkes). F.H. is a European Research Council Starting Investigator. This
work was supported by the Biotechnology and Biological Sciences Research Council (BBSRC) and the
Medical Research Council (MRC). We thank Ann Babtie for comments on the manuscript.
REFERENCES
1. L. Jiang, E. A. Althoff, F. R. Clemente, L. Doyle, D. Rothlisberger, A. Zanghellini, J. L. Gallaher,
J. L. Betker, F. Tanaka, C. F. Barbas, 3rd, D. Hilvert, K. N. Houk, B. L. Stoddard, D. Baker.
Science 319, 1387 (2008).
2. D. Rothlisberger, O. Khersonsky, A. M. Wollacott, L. Jiang, J. DeChancie, J. Betker, J. L.
Gallaher, E. A. Althoff, A. Zanghellini, O. Dym, S. Albeck, K. N. Houk, D. S. Tawfik, D. Baker.
Nature 453, 190 (2008).
3. C. A. Orengo, J. M. Thornton. Annu. Rev. Biochem. 74, 867 (2005).
4. A. G. Murzin, S. E. Brenner, T. Hubbard, C. Chothia. J. Mol. Biol. 247, 536 (1995).
5. C. A. Orengo, A. D. Michie, S. Jones, D. T. Jones, M. B. Swindells, J. M. Thornton. Structure 5,
1093 (1997).
6. F. Pearl, A. Todd, I. Sillitoe, M. Dibley, O. Redfern, T. Lewis, C. Bennett, R. Marsden, A. Grant,
D. Lee, A. Akpor, M. Maibaum, A. Harrison, T. Dallman, G. Reeves, I. Diboun, S. Addou,
S. Lise, C. Johnston, A. Sillero, J. Thornton, C. Orengo. Nucleic Acids Res. 33, D247 (2005).
7. D. Wilson, M. Madera, C. Vogel, C. Chothia, J. Gough. Nucleic Acids Res. 35, D308 (2007).
8. S. D. Brown, J. A. Gerlt, J. L. Seffernick, P. C. Babbitt. Genome Biol. 7, R8 (2006).
9. M. E. Glasner, J. A. Gerlt, P. C. Babbitt. Curr. Opin. Chem. Biol. 10, 492 (2006).
10. O. Khersonsky, C. Roodveldt, D. S. Tawfik. Curr. Opin. Chem. Biol. 10, 498 (2006).
11. P. J. O’Brien, D. Herschlag. Chem. Biol. 6, R91 (1999).
© 2009 IUPAC, Pure and Applied Chemistry 81, 731–742
Multiple catalytic promiscuity
741
12. U. T. Bornscheuer, R. J. Kazlauskas. Angew. Chem., Int. Ed. 43, 6032 (2004).
13. S. Jonas, F. Hollfelder. In The Protein Engineering Handbook, U. Bornscheuer, S. Lutz, (Eds.),
pp. 47–79, Wiley-VCH, Weinhein (2008).
14. R. A. Jensen. Annu. Rev. Microbiol. 30, 409 (1976).
15. S. Bershtein, D. S. Tawfik. Mol. Biol. Evol. 25, 2311 (2008).
16. S. Ohno. Evolution by Gene Duplication, Springer, New York (1970).
17. S. Bershtein, M. Segal, R. Bekerman, N. Tokuriki, D. S. Tawfik. Nature 444, 929 (2006).
18. C. Roodveldt, D. S. Tawfik. Biochemistry 44, 12728 (2005).
19. J. E. Vick, D. M. Schmidt, J. A. Gerlt. Biochemistry 44, 11722 (2005).
20. A. Aharoni, L. Gaidukov, O. Khersonsky, Q. Mc, S. Gould, C. Roodveldt, D. S. Tawfik. Nat.
Genet. 37, 73 (2005).
21. S. M. Gould, D. S. Tawfik. Biochemistry 44, 5444 (2005).
22. G. J. Poelarends, C. P. Whitman. Bioorg. Chem. 32, 376 (2004).
23. D. M. Schmidt, E. C. Mundorff, M. Dojka, E. Bermudez, J. E. Ness, S. Govindarajan, P. C.
Babbitt, J. Minshull, J. A. Gerlt. Biochemistry 42, 8387 (2003).
24. W. M. Patrick, I. Matsumura. J. Mol. Biol. 377, 323 (2008).
25. W. M. Patrick, E. M. Quandt, D. B. Swartzlander, I. Matsumura. Mol. Biol. Evol. 24, 2716 (2007).
26. M. Y. Galperin, A. Bairoch, E. V. Koonin. Protein Sci. 7, 1829 (1998).
27. P. J. O’Brien, D. Herschlag. J. Am. Chem. Soc. 120, 12369 (1998).
28. P. J. O’Brien, D. Herschlag. Biochemistry 40, 5691 (2001).
29. L. F. Olguin, S. E. Askew, A.-M. O’Donoghue, F. Hollfelder. J. Am. Chem. Soc. 130, 16547
(2008).
30. A. C. Babtie, S. Bandyopadhyay, L. F. Olguin, F. Hollfelder. Angew. Chem., Int. Ed. (2009). In
press (DOI: 10.1002/anie.200805843).
31. J. G. Zalatan, T. D. Fenn, A. T. Brunger, D. Herschlag. Biochemistry 45, 9788 (2006).
32. J. K. Lassila, D. Herschlag. Biochemistry 47, 12853 (2008).
33. B. van Loo, S. Jonas, A. C. Babtie, M. Hyvonen, F. Hollfelder. (2008). Submitted for publication.
34. R. Bhadra, N. Srinivasan, S. B. Pandit. In Silico Biol. 5, 379 (2005).
35. T. W. Reid, I. B. Wilson. In The Enzymes, P. D. Boyer (Ed.), pp. 373–415, Academic Press, New
York (1971).
36. J. E. Coleman. Annu. Rev. Biophys. Biomol. Struct. 21, 441 (1992).
37. K. M. Holtz, E. R. Kantrowitz. FEBS Lett. 462, 7 (1999).
38. M. Y. Galperin, M. J. Jedrzejas. Proteins 45, 318 (2001).
39. R. Gijsbers, H. Ceulemans, W. Stalmans, M. Bollen. J. Biol. Chem. 276, 1361 (2001).
40. S. R. Hanson, M. D. Best, C. H. Wong. Angew. Chem., Int. Ed. 43, 5736 (2004).
41. G. Lukatela, N. Krauss, K. Theis, T. Selmer, V. Gieselmann, K. von Figura, W. Saenger.
Biochemistry 37, 3654 (1998).
42. B. Stec, K. M. Holtz, E. R. Kantrowitz. J. Mol. Biol. 299, 1303 (2000).
43. I. Boltes, H. Czapinska, A. Kahnert, R. von Bulow, T. Dierks, B. Schmidt, K. von Figura, M. A.
Kertesz, I. Uson. Structure 9, 483 (2001).
44. S. Jonas, B. van Loo, M. Hyvonen, F. Hollfelder. J. Mol. Biol. 384, 120 (2008).
45. C. S. Bond, P. R. Clements, S. J. Ashby, C. A. Collyer, S. J. Harrop, J. J. Hopwood, J. M. Guss.
Structure 5, 277 (1997).
46. C.-I. Braenden. Q. Rev. Biophys. 13, 317 (1980).
47. S. B. Dotson, C. E. Smith, C. S. Ling, G. F. Barry, G. M. Kishore. J. Biol. Chem. 271, 25754
(1996).
48. M. J. Jedrzejas, M. Chander, P. Setlow, G. Krishnasamy. EMBO J. 19, 1419 (2000).
49. J. E. Coleman, K. Nakamura, J. F. Chlebowski. J. Biol. Chem. 258, 386 (1983).
50. E. E. Kim, H. W. Wyckoff. J. Mol. Biol. 218, 449 (1991).
51. B. Schmidt, T. Selmer, A. Ingendoh, K. von Figura. Cell 82, 271 (1995).
© 2009 IUPAC, Pure and Applied Chemistry 81, 731–742
742
S. JONAS AND F. HOLLFELDER
52. C. Szameit, C. Miech, M. Balleininger, B. Schmidt, K. von Figura, T. Dierks. J. Biol. Chem. 274,
15375 (1999).
53. N. H. Williams, B. Takasaki, M. Wall, J. Chin. Acc. Chem. Res. 32, 485 (1999).
54. D. E. Wilcox. Chem. Rev. 96, 2435 (1996).
55. J. A. Cowan. Chem. Rev. 98, 1067 (1998).
56. T. A. Steitz, J. A. Steitz. Proc. Natl. Acad. Sci. USA 90, 6498 (1993).
57. A. Waldow, B. Schmidt, T. Dierks, R. von Bulow, K. von Figura. J. Biol. Chem. 274, 12284
(1999).
58. R. von Bulow, B. Schmidt, T. Dierks, K. von Figura, I. Uson. J. Mol. Biol. 305, 269 (2001).
59. P. J. O’Brien, J. K. Lassila, T. D. Fenn, J. G. Zalatan, D. Herschlag. Biochemistry 47, 7663 (2008).
60. J. G. Zalatan, T. D. Fenn, D. Herschlag. J. Mol. Biol. 384, 1174 (2008).
61. M. Recksiek, T. Selmer, T. Dierks, B. Schmidt, K. von Figura. J. Biol. Chem. 273, 6096 (1998).
62. J. Wang, E. R. Kantrowitz. Protein Sci. 15, 2395 (2006).
63. K. Yang, W. W. Metcalf. Proc. Natl. Acad. Sci. USA 101, 7919 (2004).
64. F. Hollfelder, D. Herschlag. Biochemistry 34, 12255 (1995).
65. A. J. Kirby, W. P. Jencks. J. Am. Chem. Soc. 87, 3209 (1965).
66. A. C. Hengge, I. Onyido. Curr. Org. Chem. 9, 61 (2005).
67. J. G. Zalatan, D. Herschlag. J. Am. Chem. Soc. 128, 1293 (2006).
68. P. J. O’Brien, D. Herschlag. Biochemistry 41, 3207 (2002).
69. J. G. Zalatan, I. Catrina, R. Mitchell, P. K. Grzyska, J. O’Brien P, D. Herschlag, A. C. Hengge. J.
Am. Chem. Soc. 129, 9789 (2007).
70. R. Wolfenden. Chem. Rev. 106, 3379 (2006).
71. R. Wolfenden, M. J. Snider. Acc. Chem. Res. 34, 938 (2001).
72. A. Radzicka, R. Wolfenden. Science 267, 90 (1995).
73. I. Nikolic-Hughes, J. O’Brien, P. D. Herschlag. J. Am. Chem. Soc. 127, 9314 (2005).
74. I. Catrina, P. J. O’Brien, J. Purcell, I. Nikolic-Hughes, J. G. Zalatan, A. C. Hengge, D. Herschlag.
J. Am. Chem. Soc. 129, 5760 (2007).
75. L. Sun, D. C. Martin, E. R. Kantrowitz. Biochemistry 38, 2842 (1999).
76. J. Chin, M. Banaszczyk, V. Jubian, X. Zou. J. Am. Chem. Soc. 111, 186 (1989).
77. S. J. Benkovic, P. A. Benkovic. J. Am. Chem. Soc. 88, 5504 (1966).
78. J. Purcell, A. C. Hengge. J. Org. Chem. 70, 8437 (2005).
79. A. J. Kirby, A. G. Vargolis. J. Am. Chem. Soc. 89, 415 (1967).
80. E. J. Fendler, J. H. Fendler. J. Org. Chem. 33, 3852 (1968).
81. G. Feng, D. Natale, R. Prabaharan, J. C. Mareque-Rivas, N. H. Williams. Angew. Chem., Int. Ed.
45, 7056 (2006).
82. J. R. Morrow, O. Iranzo. Curr. Opin. Chem. Biol. 8, 192 (2004).
83. N. H. Williams. Biochim. Biophys. Acta 1697, 279 (2004).
84. J. Kofoed, J.-L. Reymond. Curr. Opin. Chem. Biol. 9, 656 (2005).
85. A. Aharoni, L. Gaidukov, S. Yagur, L. Toker, I. Silman, D. S. Tawfik. Proc. Natl. Acad. Sci. USA
101, 482 (2004).
86. M. Harel, A. Aharoni, L. Gaidukov, B. Brumshtein, O. Khersonsky, R. Meged, H. Dvir, R. B.
Ravelli, A. McCarthy, L. Toker, I. Silman, J. L. Sussman, D. S. Tawfik. Nat. Struct. Mol. Biol. 11,
412 (2004).
87. C. Roodveldt, D. S. Tawfik. Protein Eng. Des. Sel. 18, 51 (2005).
88. H. Shim, S. B. Hong, F. M. Raushel. J. Biol. Chem. 273, 17445 (1998).
89. D. W. Christianson, J. D. Cox. Annu. Rev. Biochem. 68, 33 (1999).
© 2009 IUPAC, Pure and Applied Chemistry 81, 731–742