Báo cáo khóa học: How calcium inhibits the magnesium-dependent enzyme human phosphoserine phosphatase potx

Leuven, Leuven, Belgium;2Laboratory of Physiological Chemistry, Christian de Duve Institute of Cellular Pathology, Universite´ Catholique de Louvain, Brussels, Belgium The structure of t

Laboratory for Analytical Chemistry and Medicinal Physicochemistry, Faculty of Pharmaceutical Sciences, K.U Leuven, Leuven, Belgium;2Laboratory of Physiological Chemistry, Christian de Duve Institute of Cellular Pathology, Universite´ Catholique de Louvain, Brussels, Belgium

The structure of the Mg2+-dependent enzyme human

phosphoserine phosphatase (HPSP) was exploited to

examine the structural and functional role of the divalent

cation in the active site of phosphatases Most interesting is

the biochemical observation that a Ca2+ion inhibits the

activity of HPSP, even in the presence of added Mg2+ The

sixfold coordinated Mg2+ion present in the active site of

HPSP under normal physiological conditions, was replaced

by a Ca2+ion by using a crystallization condition with high

concentration of CaCl2(0.7M) The resulting HPSP

struc-ture now shows a sevenfold coordinated Ca2+ion in the

active site that might explain the inhibitory effect of Ca2+on the enzyme Indeed, the Ca2+ion in the active site captures both side-chain oxygen atoms of the catalytic Asp20 as a ligand, while a Mg2+ion ligates only one oxygen atom of this Asp residue The bidentate character of Asp20 towards

Ca2+hampers the nucleophilic attack of one of the Asp20 side chain oxygen atoms on the phosphorus atom of the substrate phosphoserine

Keywords: calcium; HAD superfamily; magnesium-depend-ent enzymes; phosphoserine phosphatase;L-serine

Human phosphoserine phosphatase (HPSP) catalyses the

last and irreversible step of the de novo biosynthesis of

L-serine, i.e the hydrolysis of phosphoserine leading to the

formation ofL-serine and inorganic phosphate (Pi) HPSP is

a member of the haloacid dehalogenase (HAD) superfamily

of which the members are characterized by three short

conserved sequence motifs (Fig 1) The residues of these

motifs cluster together to form the active site All enzymes of

the HAD superfamily use the aspartate residue of the first

conserved DXXX(T/V) motif as a nucleophilic residue for

catalysis [1] The second motif contains a conserved serine or

threonine residue, and the third motif contains a strictly

conserved lysine residue followed, at some distance, by less

conserved residues and a strictly conserved aspartate

Mutagenesis studies on these conserved residues show that

all three motifs play an important role in the catalytic

process [2–4]

Despite the low overall sequence homology among the

enzymes of the HAD superfamily, all known structures of

enzymes of this superfamily display a conserved fold [5]

Indeed, 2-haloacid dehalogenase from Pseudomonas sp YL

and Xanthobacter autotrophicus [6,7],

phosphonoacetalde-hyde hydrolase from Bacillus cereus [8], soluble epoxide

hydrolase [9], the Ca2+-P-type ATPase [10],

b-phospho-glucomutase from Lactococcus lactis [11], phosphoserine phosphatase (PSP) from Methanococcus jannaschii (MJ PSP) [12,13] and HPSP [14,15] all have a core a/b domain resembling the NAD(P)-binding Rossmann fold [5] This fold is characterized by a central six-stranded b-sheet flanked on both sides by two or three a-helices The similar topology and common fold of the central domain, strongly suggest that the members of the HAD superfamily evolved from a primordial, generic domain

Metals are found in a broad variety of proteins where they display important functional or structural roles A bound Mg2+ion is an essential active site component of numerous metalloproteins including nucleases, kinases and phosphatases Such proteins use Mg2+ for phospho-substrate binding, catalysis, or both The HAD superfamily members, except for 2-haloacid dehalogenases [6,7], utilize

Mg2+as a cofactor during catalysis The effects of various metal cations on the activity of PSP were described [16], but key features of their metal binding characteristics remained undetermined Maximum activity of the enzyme, measured

by the rate of Pi release from phosphoserine, is obtained with Mg2+ In the absence of added divalent cations, the activity of PSP is only 9–15% of the maximal activity observed in the presence of Mg2+ Of particular interest was our observation that the replacement of Mg2+by Ca2+in

an activity test caused complete loss of activity of PSP Furthermore Ca2+inhibited the activity measured in the presence of Mg2+ Two interesting questions arise from these observations: is there structural evidence for the fact that Mg2+in the active site cannot be replaced by another divalent cation without loss of activity, and how does an enzyme manage to select a specific cation from the surrounding fluids that contain a broad variety of cations?

A detailed study of the structure of the active site of HPSP with Ca2+bound may provide an insight into the biological

Correspondence to A Rabijns, Laboratory for Analytical Chemistry

and Medicinal Physicochemistry, Faculty of Pharmaceutical Sciences,

K.U Leuven, E Van Evenstraat 4, B-3000 Leuven, Belgium.

Fax: +32 16 32 34 69, Tel.: +32 16 32 34 21,

E-mail: anja.rabijns@pharm.kuleuven.ac.be

Abbreviations: HAD, haloacid dehalogenase; HPSP, human

phos-phoserine phosphatase; Pi, inorganic phosphate; PSP, phosphos-phoserine

phosphatase.

(Received 19 May 2004, revised 1 July 2004, accepted 7 July 2004)

role of metal ions, especially divalent metal ions, in

biological processes

Materials and methods

Structure determination

The expression, purification and crystallization of HPSP

was carried out using methods described previously [17]

The Ca2+ containing crystal structure elucidated to a

resolution of 1.53 A˚ has been described elsewhere The

structure was deposited in the Protein Data Bank (code

1NNL; http://www.rcsb.org) [15] The final HPSP model

consists of 3203 protein atoms, 390 water molecules, six Cl–

and three Ca2+ions and a summary of the crystallographic

quality indicators for this final model is given in Table 1

HPSP activity assay After purification, HPSP was assayed at 30°C by the release of Pi from unlabeledL-phosphoserine in an assay mixture (250 lL) containing 25 mM Mes (pH 6.5), 5 mM MgCl2, 1 mM dithiothreitol, 5 mM L-phosphoserine,

0.1 mgÆmL )1bovine serum albumin and 1–10 mU HPSP Reactions were stopped by the addition of 250 lL of 10% (v/v) trichloroacetic acid and the amount of Pi was measured using a spectrophotometer [18] One unit of enzyme is the amount that catalyses the conversion of

1 lmol of substrate per minute under these conditions To address the effects of Ca2+on the HPSP activity, HPSP was incubated with different concentrations of Mg2+(0.2, 1, 2,

5 mM) and all these set-ups were assayed in the presence of increasing concentrations of Ca2+(0.00, 0.025, 0.05, 0.10, 0.25 and 1.0 mM)

Results and Discussion

Presence of Ca2+in the HPSP active site During the refinement of the HPSP model it became clear that the electron density peak in the active site could best be explained by a Ca2+ ion The plausible reason for the presence of the Ca2+in the active site, instead of a Mg2+ ion as in previously reported structures of the PSP family, is that we used CaCl2in the HPSP crystallization condition; at all times the presence of Mg2+was avoided The two HPSP molecules in the asymmetric unit (Molecule A and B) were refined independently of each other In both molecules, the atoms surrounding the divalent cation and the metal ion itself, were in the same range of B factor values (around

15 A˚2); replacing the Ca2+ion by another ion (e.g Mg2+) causes the R factor to increase substantially (0.7%) during the refinement procedure, as commented by Peeraer et al [15] The presence of a Ca2+ion is further confirmed by the geometry and the metal-donor atom target distances (Table 2)

Role of the divalent cation in the reaction mechanism

of HPSP

To understand the biological role of Mg2+in the catalytic mechanism of PSP, one should keep in mind that the hydrolysis of phosphoserine by PSP proceeds through a stepwise phosphotransfer mechanism, as demonstrated by

Table 1 Data collection, refinement and model statistics for the HPSP

structure at 1.53 A˚ resolution Values in parentheses indicate data in the

highest resolution shell, i.e 1.56–1.53 A˚.

Data collection statistics

Resolution limit (A˚) 1.53 (1.56–1.53)

Completeness of all data (%) 99.8 (98.7)

Completeness of the data I > 2r (%) 95.2 (80.7)

Refinement statistics

Model statistics

Average atomic B factors (A˚ 2 )

rmsd of the model

B, bonded main chain (A˚ 2 ) 1.194

B, bonded side chain (A˚2) 2.077

Fig 1 Multiple sequence alignment of the members of the HAD superfamily The first column indicates the protein and the species it comes from PSP, phosphoserine phosphatase; PMM, phosphomannomutase; HAD, haloacid dehalogenase; ATP, ATPase (Human, Homo sapiens; Meth, Methanococcus jannaschii; Sacc, Saccharomyces cerevisiae; Coli, Escherichia coli; Pssp, Pseudomonas sp.) Numbers indicate the distances to the ends of each protein and numbers in parentheses indicate the sizes of the gaps between the aligned segments The highlighted amino acids are conserved in the HAD superfamily.

mechanistic studies on MJ PSP [19] Structural comparison

between HPSP (PDB code 1NNL) and MJ PSP (PDB code

1L7O) structures (rmsd 1.64 A˚ for 176 residues

super-imposed and rmsd 0.64 A˚ for 16 active site residues

superimposed), reveals that the reaction mechanism of

HPSP involves subsequent nucleophilic attacks and acid/

base catalysis The conserved residues Arg65 and Glu29

play an essential role in orientating the substrate in an

appropriate manner for hydrolysis The side-chain of Glu29

interacts with the amino group of phosphoserine, while the

side-chain of Arg65 forms a hydrogen bond with the

carboxyl group of the substrate When the substrate is

positioned correctly, the enzyme closes and Asp20 performs

a nucleophilic attack on the scissile phosphate The

substrate phosphoserine is then cleaved, resulting in the

departure of the leaving group serine and the formation of a

covalent phosphoaspartyl (Asp20) intermediate (Fig 2)

Asp22 serves as a general acid (Fig 2, Enz-H) donating a

proton to serine and thereby facilitating the expulsion of the

leaving group A water molecule takes the position in the

just-vacated leaving group site The Asp22 carboxylate

anion (Fig 2, Enz-B) that was formed during the

protona-tion of the leaving serine group, can now serve as a base

catalyst in the dephosphorylation of the phosphoenzyme

intermediate Asp22 extracts a proton from the water

molecule in the active site, thereby activating the water

molecule to perform a nucleophilic attack on the

phospho-aspartyl intermediate Opening of the enzyme and

dissoci-ation of the inorganic phosphate completes the catalytic

cycle

The Mg2+ion in the active site is essential for HPSP to

perform the hydrolysis of phosphoserine First of all Mg2+

plays a catalytic role in the reaction mechanism The Mg2+

ion coordinates both an oxygen atom of the phosphate

moiety of the substrate and an oxygen atom of the attacking

Asp20 residue In this way the Asp20 residue is stabilized in

an optimal position to perform an attack on the phosphorus

atom of phosphoserine In addition, the positive charge of

the divalent cation is essential to facilitate the nucleophilic

attack of Asp20 by extracting negative charge from the

phosphate group The fact that haloacid dehalogenases do

not need a divalent cation for activity, while the

phospho-transferases of the same HAD superfamily do, supports the

idea that a divalent cation in HPSP is needed to shield the

negative charges of the phosphate group while the attacking

nucleophile Asp20 is approaching Of interest is that in

haloacid dehalogenase, the corresponding attacking Asp

residue approaches an electropositive carbon centre of the substrate and thus a cation is not required to promote the nucleophilic attack

Besides its catalytic role, the divalent cation in the HPSP active site also plays a purely structural role In the HPSP active site, three Asp residues (20, 22 and 179) are in close proximity to each other and form a carboxylate cluster, thereby generating an excess of negative charge in the binding pocket The positive charge of the divalent cation is therefore necessary to stabilize the overall architecture of this carboxylate cluster by diminishing the electrostatic repulsion between the negative charges of the Asp side-chains The stabilizing, structural role of Mg2+is further illustrated by the fact that Asp179, which belongs to sequence motif III and which coordinates the divalent cation in the active site, is conserved in all the HAD superfamily members with the exception of the enzymes that are Mg2+-independent for their activity Indeed, in the haloacid dehalogenases, the corresponding residue is a Ser which is not essential for catalytic activity [7] This observation suggests that Asp179 is essential for binding

of the divalent cation in the active site Furthermore, mutagenesis studies on HPSP showed that mutation of Asp179 to an Asn or Glu results in a 10-fold decrease in the affinity for Mg2+[4] The same functions for the Mg2+ion are observed in other Mg2+-dependent members of the HAD superfamily like phosphonoacetaldehyde hydrolase, b-phosphoglucomutase and P-type ATPases [8,11,20]

Mg2+substituted by a Ca2+: implications for the reaction mechanism

Neuhaus & Byrne [16] reported that HPSP activity depends

on the presence of Mg2+ We confirmed this requirement and we determined that the Kafor Mg2+in the presence of a saturating concentration of substrate was 0.2 mM (not shown) From Fig 3 it can be seen that Ca2+inhibited the enzyme activity, and the lower the Mg2+concentration the more apparent this effect was Indeed, a 50% inhibition was observed at 0.01, 0.025, 0.05 and 0.2 mM Ca2+ in the presence of 0.2, 1, 2 and 5 mMMg2+

Several experiments to also obtain a Mg2+-containing HPSP structure, i.e soaking and cocrystallization experi-ments, failed Therefore, to elucidate the inhibitory effect exerted by Ca2+on the activity of PSP, we compared the active site of HPSP, which contains a Ca2+ion, with the MJ PSP active site, containing Mg2+(PDB codes 1F5S, 1L7P

and 1J97) The Mg2+ion in the MJ PSP active site displays

almost perfect octahedral coordination geometry with six

ligands Four ligands are in a plane with O–Mg2+–O angles

of nearly 90°, while the two other ligands are above and

below this plane, respectively The coordination of the Ca2+

ion in the active site of HPSP is distorted from octahedral

geometry as shown in Fig 4 In open conformation, three water molecules and three O atoms (OD1 of Asp20, the main-chain carbonyl group of Asp22 and OD2 of Asp179) occupy six of the coordination sites of the Ca2+, similar to the Mg2+ion in the MJ PSP active site Nevertheless, it can

be seen that one water molecule (Fig 4B, Wat1) is forced out of the plane This distortion of the octahedral geometry

is due to the fact that the Ca2+ion prefers seven ligands instead of six as Mg2+does Because the coordination of spherical metal ions is optimized by maximum packing of ligand atoms, the preferred coordination number is primar-ily a function of the size of the ion [21] The effective ionic radius of a Mg2+ion (0.72 A˚) is considerably smaller than that of a Ca2+ion (1.06 A˚) [22] The smaller size of Mg2+ determines its preference for a coordination number of six

In contrast, the effective ionic radius of Ca2+is such that seven or eight coordinating ligands can be comfortably accommodated [23] As a result the Ca2+ion in the HPSP active site accepts both side-chain oxygen atoms of Asp20

as a ligand, while a Mg2+ion ligates only one oxygen atom

of this Asp residue

Besides the differences in geometry between Ca2+and

Mg2+ in the active site, the metal–ligand distances are also quite different Comparison of the active sites of HPSP and MJ PSP shows that replacement of a Mg2+by

a Ca2+ ion results in an increase in all metal–ligand

Fig 2 General scheme of the reaction cycle of PSP [22] Open conformation of PSP (A) L -Phosphoserine binds to the active site presenting the phosphate group to Asp20 (B) Transition state with nucleophylic attack of Asp20 (C) Covalent phosphoaspartyl enzyme intermediate (D) Transition state with a nucleophylic attack of a water molecule causing the dephosphorylation of Asp20 (E) Phosphate noncovalently bound in the active site (F) Enz-H indicates the general acid Asp22, which after the protonation of the leaving serine group serves as a base catalyst Enz-B.

Fig 3 Effect of Ca 2+ on HPSP activity The effect of Ca2+on HPSP

activity was assayed in the presence of 0.2 (j), 1 (m), 2 (.) or 5 (r) m M

Mg 2+ HPSP activity was assayed as in Materials and methods.

distances, with average distances of 2.1 A˚ for Mg2+and

2.4 A˚ for Ca2+ The observed distances match to a large

extent the ideal distances for Ca2+-donor atom

combina-tions and similar distances are observed in various

metalloproteins [24] Combination of the dissimilar

geo-metry and changed metal–ligand distances drastically

affects the reaction mechanism of HPSP when the

Mg2+ ion is substituted by a Ca2+ ion (Fig 5) Upon

substrate binding, one of the three water molecules

coordinating the divalent ion is replaced by an oxygen

of the phosphate moiety of phosphoserine The fact that Asp20 in HPSP acts as a bidentate ligand in the sevenfold coordination of Ca2+(OD1 and OD2 to Ca2+distances

of 2.38 and 2.77 A˚, respectively), while the corresponding Asp11 in MJ PSP is a monodentate ligand in the sixfold

Mg2+ coordination, hampers the nucleophilic attack of OD1 on the substrate The corresponding OD1 of Asp11

in the Mg2+ bound MJ PSP is 3.33 A˚ away from the cation and therefore it is free to perform an attack on the phosphorus atom of the substrate The distance between

Fig 4 Detailed overview of the Ca 2+ + ion in the active site of HPSP (A) The residues are represented in ball and stick form with oxygen, carbon and nitrogen atoms coloured red, light-blue and dark-blue, respectively The Ca2+ion is shown in green Three of the Ca2+ligands are water molecules, shown as red balls The dashed lines represent hydrogen bonds and metal–ligand interactions Asp20, Asp22 and Asp179 directly coordinate the

Ca 2+ ion Asp179 and Gly180 interact with Asp183, thereby stabilizing the loop on which they are located (B) The coordination of the Ca 2+ is distorted from ideal octahedral geometry with six ligands because it forms an extra interaction with one of the oxygen atoms of Asp20 This extra interaction between Ca 2+ and Asp20, shown in green, does not occur with a Mg 2+ ion in the active site.

Fig 5 Active site of MJ PSP with a Mg2++and phosphoserine in the active site (PDB codes 1F5S and 1L7P) (A) and HPSP (PDB code 1NNL) with a

Ca2+ion bound and the modelled substrate in the active site (B) For clarity only four ligands are shown, i.e a ligating water molecule and Asp13/22 (HPSP/MJ PSP) are omitted in this figure In contrast to a Mg2+ion, the Ca2+ion in HPSP ligates both oxygen atoms of Asp20 thereby preventing

it to perform a nucleophilic attack on the phosphorus atom of the substrate In addition, a Ca2+ion displays longer metal–ligand distances than a

Mg 2+ ion As a consequence the partial positive charge on the phosphorus atom of phosphoserine is smaller if a Ca 2+ takes position in the active site In this manner, a Ca 2+ will further hamper the nucleophilic attack of the catalytic Asp residue on the substrate.

the OD1 of the attacking Asp residue and the

phos-phorus atom of the substrate is increased from 2.96 A˚

with a Mg2+ ion to 3.36 A˚ with a Ca2+, further

hampering the nucleophilic attack on the substrate In

MJ PSP the distance between oxygen O2 of the phosphate

part ofL-phosphoserine and the Mg2+ion is 2.41 A˚ [19]

Replacing the Mg2+ ion by a Ca2+ ion results in an

increase of this distance to 3.20 A˚ This will undoubtedly

result in a smaller attraction of negative charge from the

phosphate moiety of the substrate, thereby suppressing the

nucleophilic attack of Asp20 on the phosphorus atom

HPSP selectivity for Mg2+

For the Mg2+binding site of the related CheY enzyme

[25,26], it was proposed that the carboxylate cluster in the

active site provides charge specificity to the Mg2+binding

site by excluding monovalent cations like Na+and K+,

because they do not possess sufficient positive charge to

stabilize the highly negative carboxylate cluster [27]

Ana-logously, HPSP can exploit the negative charge of the

carboxylate cluster, composed of Asp20, Asp22 and

Asp179, to provide the necessary charge specificity by

excluding monovalent cations

On the other hand, it seems that the Mg2+binding site

in HPSP is weakly protected against the binding of other

divalent cations like Ca2+, as Ca2+ displays inhibiting

properties even in the presence of Mg2+ ([16] and this

paper) The weak size-selectivity of HPSP can originate

from the fact that in HPSP in open conformation three of

the ligands to the divalent cation are water molecules [15]

An interesting feature of this coordination structure is that

one hemisphere of the bound ion is coordinated by three

protein oxygens, while the other hemisphere is

coordina-ted by three solvent molecules (Fig 4) These water

molecules can easily accommodate changing metal–ligand

distances if Mg2+is replaced by a larger divalent cation

such as Ca2+ In addition, the larger Ca2+ ion can

employ Asp20 as a bidendate ligand in order to complete

its preferred sevenfold coordination geometry Thus, the

difference in ionic radii of Ca2+ and Mg2+ is not a

sufficient criterion for HPSP to select Mg2+, as the

binding cavity of the enzyme is flexible and able to adjust

easily to different ionic radii and changing coordination

geometry

In view of the facts outlined above it becomes clear that

the metal-binding pocket of HPSP is charge-selective in

order to discriminate between mono- and divalent cations,

but not size-selective enough to single out particular divalent

cations as Mg2+ and Ca2+ Nevertheless, in living cells

HPSP uses Mg2+as a cofactor and not the larger Ca2+

The latter seems logical, as Mg2+ is the most abundant

divalent cation in eukaryotic cells, with concentrations of

free Mg2+ranging from 0.1 to 1.0 mM, while the Ca2+

concentration is 104-fold lower in resting eukaryotic cells

[28] Thus, HPSP has chosen Mg2+as a cofactor during

evolution based mainly on its natural abundance in living

cells In this scenario it is not the protein metal-binding

pocket architecture itself but the cell homeostasis that

controls the process of metal binding by regulating the

appropriate concentrations of Mg2+and other cations in

various biological compartments

Conclusions

The HPSP reaction mechanism involves nucleophilic attack

of Asp20 on the substrate with acid/base catalysis mediated

by Asp22 The Mg2+ion in the active site is essential for normal enzymatic activity, i.e the Mg2+ion promotes the nucleophilic attack of Asp20 by withdrawing negative charge from the phosphorus atom of the substrate In addition, the divalent cation is essential for the correct orientation of the attacking Asp20 residue towards the substrate A Ca2+ ion however, employs Asp20 as a bidentate ligand, thereby inhibiting the nucleophilic attack

of this catalytic residue Furthermore, it seems that the

Mg2+binding site in HPSP is weakly protected against the binding of other divalent cations, as Ca2+displays inhib-iting properties even in the presence of Mg2+ Therefore it is probable that HPSP has chosen Mg2+as a cofactor during evolution based mainly on the natural abundance of Mg2+

in living cells

Acknowledgements

A.R is a Postdoctoral Research Fellow of the Fund for Scientific Research-Flanders (Belgium) and J.-F.C was Charge´ de Recherches of the Belgian FNRS Work in the lab of E.V.S is supported by the Interuniversity Attraction Poles Program-Belgian Science Policy and by the FRSM We thank the beam line scientists at DESY for technical support and the European Union for support of the work at EMBL Hamburg through the Access to Research Infrastructure Action of the improving human potential programme, contact no HPRI-CT-1999-00017.

References

1 Collet, J.F., Stroobant, V., Pirard, M., Delpierre, G & Van Schaftingen, E (1998) A new class of phosphotransferases phos-phorylated on an aspartate residue in an amino-terminal DXDX(T/V) motif J Biol Chem 273, 14107–14112.

2 MacLennan, D.H., Clarke, D.M., Loo, T.W & Skerjanc, I.S (1992) Site-directed mutagenesis of the Ca 2+ ATPase of sarco-plasmic reticulum Acta Physiol Scand Suppl 607, 141–150.

3 Lingrel, J.B & Kuntzweiler, T (1994) Na + ,K + -ATPase J Biol Chem 269, 19659–19662.

4 Collet, J.F., Stroobant, V & Van Schaftingen, E (1999) Mechanistic studies of phosphoserine phosphatase, an enzyme related to P-type ATPases J Biol Chem 274, 33985–33990.

5 Rossmann, M.G., Moras, D & Olsen, K.W (1974) Chemical and biological evolution of a nucleotide-binding protein Nature 250, 194–199.

6 Hisano, T., Hata, Y., Fujii, T., Liu, J.Q., Kurihara, T., Esaki, N.

& Soda, K (1996) Crystal structure of L-2-haloacid dehalogenase from Pseudomonas sp YL J Biol Chem 271, 20322–20330.

7 Ridder, I.S & Dijkstra, B.W (1999) Identification of the Mg 2+ -binding site in the P-type ATPase and phosphatase members of the HAD (haloacid dehalogenase) superfamily by structural similarity to the response regulator CheY Biochem J 339, 223– 226.

8 Morais, M.C., Zhang, W., Baker, A.S., Zhang, G., Dunaway-Mariano, D & Allen, K.N (2000) The crystal structure of Bacillus cereus phosphonoacetaldehyde hydrolase: insight into catalysis of phosphorus bond cleavage and catalytic diversification within the HAD enzyme superfamily Biochemistry 39, 10385–10396.

9 Argiriadi, M.A., Morisseau, C., Hammock, B.D & Christianson, D.W (1999) Detoxification of environmental mutagens and

12 Cho, H., Wang, W., Kim, R., Yokota, H., Damo, S., Kim, S.H.,

Wemmer, D., Kustu, S & Yan, D (2001) BeF3 acts as a

phosphate analog in proteins phosphorylated on aspartate:

structure of a BeF3 complex with phosphoserine phosphatase.

Proc Natl Acad Sci USA 98, 8525–8530.

13 Wang, W., Kim, R., Jancarik, J., Yokota, H & Kim, S.H (2001)

Crystal structure of phosphoserine phosphatase from

Methano-coccus jannaschii, a hyperthermophile, at 1.8 A˚ resolution

Struc-ture 9, 65–71.

14 Kim, H.Y., Heo, Y.S., Kim, J.H., Park, M.H., Moon, J., Kwan,

D., Yoon, J., Shin, D., Jeong, E., Park, S.Y., Lee, T.G., Jeon,

Y.H., Ro, S., Cho, J.M & Hwang, K.Y (2002) Molecular basis

for the local conformational rearrangement of human

phospho-serine phosphatase J Biol Chem 227, 46651–46658.

15 Peeraer, Y., Rabijns, A., Verboven, C., Collet, J.F., Van

Schaf-tingen, E & De Ranter, C (2003) High resolution structure of

human phosphoserine phosphatase in open conformation Acta

Cryst D59, 971–977.

16 Neuhaus, F.C & Byrne, W.L (1959) Metabolism of

phospho-serine II Purification and properties of O-phosphoserine

phos-phatase J Biol Chem 234, 113–121.

17 Peeraer, Y., Rabijns, A., Verboven, C., Collet, J.F., Van

Schaf-tingen, E & De Ranter, C (2002) Purification, crystallization and

preliminary X-ray analysis of human phosphoserine phosphatase.

Acta Cryst D58, 133–134.

calcium pump Annu Rev Biophys Biomol Struct 32, 445–468.

21 Yang, W., Lee, H.W., Hellinga, H & Yang, J.J (2002) Structural analysis, identification and design of calcium-binding sites in proteins Proteins: Stuct Func Genet 47, 344–356.

22 Shannon, R.D (1976) Revised effective ionic radii and systematic studies of interatomic distances in halides and chalcogenides Acta Cryst A36, 751–767.

23 Falke, J.J., Drake, S.K., Hazard, A.L & Peersen, O.B (1994) Molecular tuning of ion binding to calcium signaling proteins.

Q Rev Biophys 27, 219–290.

24 Harding, M.M (2001) Geometry of metal-ligand interactions in proteins Acta Cryst D57, 401–411.

25 Bellsollel, L., Prieto, J., Serrano, L & Coll, M (1994) Magnesium binding to the bacterial chemotaxis protein CheY results in large conformational changes involving its functional surface J Mol Biol 238, 489–495.

26 Ridder, I.S., Rozeboom, H.J., Kalk, K.H., Janssen, D.B & Dijkstra, B.W (1997) Three-dimensional structure of L -2-haloacid dehalogenase from Xanthobacter autotrophicum GJ10 complexed with the substrate analogue formate J Biol Chem 272, 33015– 33022.

27 Needham, J.V., Chen, T.Y & Falke, J.J (1993) Novel ion speci-ficity of a carboxylate cluster Mg(II) binding site: strong charge selectivity and weak size selectivity Biochemistry 32, 3363–3367.

28 Romani, A & Scarpa, A (1992) Regulation of cell magnesium Arch Biochem Biophys 298, 1–12.