1774 (2007) 1247–1253www.elsevier.com/locate/bbapap

Biochimica et Biophysica Acta

Analysis of the structural consensus of the zinc coordination centers ofmetalloprotein structures

Kirti Patel, Anil Kumar, Susheel Durani ⁎

Department of Chemistry, Indian Institute of Technology Bombay, Mumbai-400076, India

Received 3 January 2007; received in revised form 16 June 2007; accepted 20 July 2007Available online 8 August 2007

Abstract

In a recent sequence-analysis study it was concluded that up to 10% of the human proteome could be comprised of zinc proteins, quite variedin the functional spread. The native structures of only few of the proteins are actually established. The elucidation of rest of the sequences of notjust human but even other actively investigated genomes may benefit from knowledge of the structural consensus of the zinc-binding centers ofthe currently known zinc proteins. Nearly four hundred X-ray and NMR structures in the database of zinc–protein structures available as of April2007 were investigated for geometry and conformation in the zinc-binding centers; separately for the structural and catalytic proteins andindividually in the zinc centers coordinated to three and four amino-acid ligands. Enhanced cysteine involvement in agreement with theobservation in human proteome has been detected in contrast with previous reports. Deviations from ideal coordination geometries are detected,possible underlying reasons are investigated, and correlations of geometry and conformation in zinc-coordination centers with protein function areestablished, providing possible benchmarks for putative zinc-binding patterns of the burgeoning genome data.© 2007 Elsevier B.V. All rights reserved.

Keywords: Metalloprotein; Zinc–protein; Zinc binding pattern; Zinc coordination; Structural zinc-center; Catalytic zinc-center

1. Introduction

The second most abundant transition element in biologicalsystems after iron, zinc is essential for life. With closed d-shell,unlike iron and copper, zinc (II) finds biological relevance as aredox-stable element in metalloprotein structures. Interactingmainly with amino-acid side chains and occasionally with non-protein ligands in mostly tetra, penta, and hexa-coordinatecomplexes, zinc contributes both structural and catalytic roles inmetalloprotein structures [1,2]. Progressive studies of zincproteins have established their involvements in diverse cellular

Abbreviations: H, His, Histidine; C, Cys, Cysteine; D, Asp, Aspartate; E,Glu, Glutamate; Zn, Zinc; PDB, Protein Data Bank; ZBP, Zinc Binding Patterns;C4, tetra coordinated zinc site; C5, penta coordinated zinc site; C6, hexacoordinated zinc site; Eq, Equatorial; Ax, Axial; O–Zn–O′, is angle betweentwo bidentate oxygen atoms bonded to zinc. Structural proteins are those wherezinc plays structural role whereas catalytic proteins are those where zinc is incatalytic role⁎ Corresponding author. Tel.: +91 22 25767164; fax: +91 22 25767152.E-mail address: sdurani@iitb.ac.in (S. Durani).

1570-9639/$ - see front matter © 2007 Elsevier B.V. All rights reserved.doi:10.1016/j.bbapap.2007.07.010

processes, like, transcription, translation, and metabolism [3–8].Based on homology based searches of human genome, Andreiniet al. [9] concluded that∼2800, i.e.,∼10%, of the proteins couldbe zinc proteins, encompassing all the six International Union ofBiochemistry recognized classes of enzymes [10]; in fact zinc isthe only metal so widespread in the distribution. The zinc-fingertype transcription regulators, comprising more than 4% ofhuman proteome [9], call upon zinc as a tetra-coordinatingelement; involvements of zinc as catalytic element, on otherhand, call upon its electrophilic character, in stabilizing anionicintermediates, and dynamic coordination geometry, in facilitat-ing bond rearrangements. In protein engineering and designapplications, zinc could be an aid in control of both con-formation and function [11].

The currently known zinc centers could provide the structuralconsensus for characterization of the newer zinc-bindingsequences, as well as serve as the guidelines for proteinengineering and design applications. Reviews of geometry andstereochemistry of zinc-coordination centers have appearedfrom time to time [12–21]. However, the structural database ofzinc proteins has grown at least five times since the last survey

Table 1PDB codes of 228 structural zinc proteins, 102 NMR (Panel a) and 126 X-ray determined (Panel b), and 154 catalytic zinc proteins, 7 NMR (Panel c) and 147 X-raydetermined (Panel d), investigated in this study

Panel a a

1A1T 1FRE 1M36 1PFT 1U2N 1VYX 1XPA 1ZFO 2AYJ 2FNF1AJY 1G25 1M9O 1PXE 1U6P 1WIG 1XRZ 1ZGW 2AZH 2GAT1BBO 1G47 1MM2 1PYC 1U7J 1WII 1XWH 1ZR9 2BL6 2GHF1BOR 1IBI 1N0Z 1R9P 1U85 1WJ0 1Y0J 1ZU1 2C6A 2HGH1DSQ 1IML 1NCS 1RXR 1UW0 1WJ2 1Y8F 1ZW8 2CON 2I5O1DSV 1IYM 1NEE 1SRK 1UW2 1WO3 1YUI 2A20 2CQE 2IDA1DX8 1J2O 1NJ3 1T3K 1VA1 1WPK 1YWS 2A23 2DB6 2IHX1DXW 1JJD 1NYP 1TF3 1VA2 1WWD 1Z60 2A51 2F8B 2JM31E4U 1JN7 1OVX 1TM6 1VA3 1XF7 1ZE9 2AF2 2FEJ 2JNE1EF4 1LPV 1P7M 1TOT 1VD4 1XJH 1ZFD 2AQC 2FK4 2ODX1EXK 1LV3

Panel b b

1A2P 1NW2 1R1H 1U5K 1YW4 2AU3 2CYE 2FU5 2H6L 2OGW1A7W 1OAO 1R2Z 1VJ0 1Z05 2AYD 2DGE 2FYG 2HCV 2OH31BTK 1OJ7 1R5Y 1W4R 1Z83 2AZ4 2DS5 2G84 2HJN 2OIK1DVP 1ONW 1RUT 1WWR 1Z84 2B3J 2ESL 2GC0 2I13 2OMH1ENR 1OQJ 1RYQ 1WY2 1ZED 2B4Y 2EV6 2GFE 2I9W 2OSO1GL4 1P3J 1SW1 1X0T 1ZKP 2B8T 2F5N 2GFO 2IMR 2OU31H3N 1P5D 1T0B 1XC8 1ZSW 2B9D 2F5P 2GLZ 2IQJ 2OWA1INN 1P6O 1T64 1XM8 1ZZM 2BOQ 2FE3 2GNR 2J21 2OX01J30 1PZW 1T9H 1XTM 2A1K 2C1D 2FE8 2GPY 2J6A 2P571JW9 1Q08 1TDZ 1XV2 2AP1 2C2U 2FEA 2GVI 2J9U 2PEB1KWG 1Q1A 1TJL 1XWY 2APO 2CIH 2FPR 2GWG 2O6D 2UVL1L1T 1Q74 1TQX 1YC5 2AQ2 2CJS 2FQP 2GWN 2O8J 4KMB1MA3 1Q9U 1U0A 1YQD 2AS9 2CS7

Panel c c

1DXW 1FFW 1XXE 1XYD 2AF2 2HDP 2JMO

Panel d d

1A4M 1GVF 1KO3 1P5X 1TAZ 1XOV 1Z1L 2BO0 2FR5 2HPO1AMP 1H2B 1KQ9 1Q2O 1TBF 1XP2 1Z5R 2BZ1 2G3F 2HSI1ATL 1HET 1LML 1Q7L 1TQW 1XRT 1ZDE 2C1I 2G64 2I0O1C7K 1HKK 1LR5 1QH5 1U2W 1XRU 1ZKL 2CC0 2G7N 2IW01C8Y 1HP1 1LU0 1QIP 1UDV 1XVX 1ZY7 2CDC 2GFO 2J9A1CA1 1IA9 1M65 1QWR 1UUF 1Y0Y 1ZZ1 2CFU 2GIV 2JG61D8W 1IE0 1M7J 1QWY 1V4P 1Y2K 2A21 2CKI 2GMW 2NX91ED9 1IM5 1MFM 1R55 1VHH 1Y7W 2A7M 2CKL 2GSO 2NXF1EU3 1ITU 1MXG 1R5T 1VYK 1Y93 2A97 2CTB 2GU1 2O1Q1EVL 1J79 1NTO 1S2Z 1W5Q 1Y9A 2AFW 2DDF 2H6F 2OB31F4T 1JD0 1OHT 1S4B 1WCZ 1YB0 2AIO 2DVT 2HBV 2OOT1FRP 1JIW 1OI0 1SG0 1WN5 1YIX 2BC2 2F4M 2HC9 2ORW1FUA 1JJT 1OJR 1SG6 1X8H 1YLK 2BGX 2FOU 2HEK 2OVY1GKP 1K7H 1ONW 1SR9 1XEM 1YM3 2BIB 2FPQ 2HF1 8TLN1GUD 1K9Z 1OYW 1T0A 1XOC 1YT3 2BNM

a The dataset of 102 structural zinc proteins (NMR structures).b The dataset of 126 structural zinc proteins (X-ray structures).c The dataset of 7 catalytic zinc proteins (NMR structures).d The dataset of 147 catalytic zinc proteins (X-ray structures).

1 http://dunbrack.fccc.edu/PISCES.php.

1248 K. Patel et al. / Biochimica et Biophysica Acta 1774 (2007) 1247–1253

[18]. Specifically, explosive growth of the NMR characterizedzinc–protein structures has been witnessed in addition to theX-ray characterized structures [22]. The combined resource ofabout 400 structures will possibly better capture the consensusin geometry and conformation in the newly uncovered zinc-binding patterns of diverse genomes.We have examined the zinccenters available in Protein Data Bank [22] as of April 2007 fordetails of structure and conformation and summarize the resultsin this report.

2. Materials and methods

Protein Data Bank (PDB) [22] was queried for keyword “ZN” with asequence identity cut-off ≤50% cross checked on PISCES1 server of Dunbrack[23], furnishing a total of 1321 structures. Restricting to high-resolution X-raystructures (≤2.0 Å resolution, b45 average B-factor, and b0.25 R-value) butincluding all NMR structures, gave 382 PDB entries corresponding to 228structural proteins and 154 catalytic zinc–proteins. Among the structural

Fig. 1. Ideal geometries for 4, 5, and 6 coordinated Zinc. Ax=axial,Eq=equatorial.

1249K. Patel et al. / Biochimica et Biophysica Acta 1774 (2007) 1247–1253

proteins, 102 entries were NMR structures and 126 were X-ray structures, whileamong 154 catalytic proteins, 7 entries were NMR structures and 147 were X-ray structures. Classification of the protein structures as the structural or catalyticstructures was made on basis of the available descriptions in specific PDBentries. The NMR structures were only considered for conformation in the zinc-coordination centers, while the X-ray structures were considered for bothgeometry and conformation. The overall dataset includes apo-enzymes,enzyme–inhibitor complexes, and mutants. The PDB codes of the proteinsanalyzed, their classification, and the technique of structure determination aresummarized in Table 1.

The analysis performed with in-house programs2 involved searching PDBcoordinates for the amino-acid residues within 3.0 Å cut-off of zinc atom,calculating bond lengths and angles of the atoms coordinated with zinc, andevaluating sequence positions of the amino acids coordinated with zinc. Theangle of imidazole planes of coordinating histidines with metal ion was alsodetermined with in-house program. ϕ, ψ Dihedrals of zinc coordinated amino-acid residues were extracted with PROSS program.3

3. Results and discussion

3.1. Primary coordination sphere of Zn-binding sites

Zn sites are classified according to the coordination numberin the primary coordination sphere. Zn readily forms four, five,and six coordinate complexes, as shown in Fig. 1, but tetra-coordinated structures (C4) in tetrahedral geometry are the mostprevalent. Among penta-coordinated structures trigonal-bipy-ramidal and square-pyramidal geometry prevails, while amonghexa-coordinated structures (C6) octahedral geometry prevails.As noted in Fig. 2, among 126 structural proteins (X-raystructures) a majority of the zinc sites (82%) are in C4coordination; only 14% being in C5 and 4% in C6 coordination.Of 147 catalytic proteins (X-ray structures), 58% are in C4

2 The programs are available on request from the authors.3 http://www.roselab.jhu.edu/.

coordination, 31% in C5 coordination, and 11% in C6coordination.

The C4 coordination sites are always tetrahedral. Majority ofthe C5 coordination sites are trigonal bipyramidal, while the C6coordination sites are always octahedral. Presumably electro-static and steric interactions among the ligands define thecoordination geometries. Deviations from ideal geometry occurfor some of the following reasons:

(1) A ligand extraneously H-bonded often in secondarycoordination sphere and sometime with solvent.

(2) Bidentate coordination of Asp (D) and Glu (E) carbox-ylates via both oxygens; among 400 carboxyl ligands, 67instances of bidentate arrangement were observed.

(3) Simultaneous ligation of backbone N and CO to Zn: thiswas observed in 3 cases.

(4) Multi-Zn sites involving Asp, Glu, or His (H) as bridgingligand between Zn atoms.

The structural Zn sites are generally tetrahedral with all thefour ligands from protein. In the catalytic Zn sites water orinhibitor usually occupies the fourth ligation site. In structuralzinc proteins conservation of length of the spacer between theamino acid residues ligated with zinc is observed. Theconservation is less stringent among the catalytic zinc sites,possibly due to the requirement of conformational flexibility forcatalytic function. Short and long spacers generally link theprotein ligands in both structural and catalytic sites. The shorterspacers, generally 2 to 7 residues, occur between the first andsecond zinc ligand in both the structural and catalytic sites, andbetween the third and fourth zinc ligand in the structuralproteins. A comparatively longer N20 residue spacer occursbetween the second and third zinc ligand in both the structuraland catalytic sites (Tables S1–S4), similar to the observations ofVallee et al. [24].

Fig. 2. Frequency of specific coordination numbers of Zinc sites.

Fig. 3. Amino acid propensities (%) to be Zinc ligands in X-ray and NMRstructures. Panel a: structural zinc proteins. Panel b: catalytic zinc proteins.

1250 K. Patel et al. / Biochimica et Biophysica Acta 1774 (2007) 1247–1253

3.2. Properties of Zn ligands

The statistics of protein residues serving as Zn ligands are inFig. 3. A higher prevalence of Cys (C) than previously known[8,12,14,17,18,21] is detected in conformity with the results ofAndreini et al. [9] for human genome. In particular, increasingnumber of cysteine coordinated structural proteins have beencharacterized in recent years [9,10,25] including zinc-fingerproteins and metallothianines with 3–9 mainly cysteine-coordi-nated zinc ions.

Fig. 4. Tri-coordinated Zinc Binding Patterns (ZBP). ZBPs in structural proteins (paorder of appearance in sequence (from N to C terminal). Other ZBPs correspond to dH for histidine, E for glutamic acid, D for aspartic acid, K for lysine, N for aspargine,backbone N/O atoms). Panel a: structural proteins. Panel b: catalytic proteins.

Zinc sites are classified according to the patterns of thecoordinated amino acids, i.e., the type of amino acid and theposition along sequence. Frequencies of the observed zinc-binding patterns (ZBP) considering only the protein residues areshown in Figs. 4 and 5. The tetra-coordinated ZBP patterns aremore varied among the catalytic than structural zinc proteins;among the structural proteins the top 10 patterns contribute∼75% of the sites, while among the catalytic proteins the top 10patterns contribute only 46% of the sites. Notably, even amongZBPs of identical ligands, e.g., two Cys and two His in thestructural proteins, there is often a stronger preference for agiven order in amino acid sequence, e.g., the pattern CCHH ismuch more prevalent than HHCC as noted in Fig. 5a, but thereis no such prevalence observed in case of three C and one H. Asnoted in Fig. 5a CCHC, CHCC, and HCCC are with similarpropensities. There are specific preferences noted among thecatalytic sites; e.g., HHHD and HHDD are more prevalent thenHDHH and DHDH. Among tri-coordinated ZBPs, the patternsare similar in both the structural and catalytic Zn sites. Asexpected, the tri-coordinated sites are 2.5 times as prevalent incatalytic proteins as in structural proteins. The observedstatistics are similar to those in human proteome [9].

The propensities of zinc coordinated His, Cys, and Aspresidues for specific secondary structure are summarized inTable 2. His ligands reflect a clear preference for the helical-type ϕ, ψs in the structural proteins, but comparable preferences

nel a), and catalytic proteins (panel b). In specific ZBPs, the Zinc ligands are iniverse combinations of the ligands C, H, E, D, K, N and W; with C for cysteine,and W for tryptophan (the amino acids K, N and Ware coordinated through the

Fig. 5. Tetra-coordinated Zinc Binding Patterns (ZBP). ZBPs in structural proteins (panel a), and catalytic proteins (panel b). In specific ZBPs, the Zinc ligands are inorder of appearance in sequence (from N to C terminal). Other ZBPs correspond to diverse combinations of the ligands C, H, E, D, K, N and W; with C for cysteine,H for histidine, E for glutamic acid, D for aspartic acid, K for lysine, N for aspargine, and W for tryptophan (the amino acids K, N and Ware coordinated through thebackbone N/O atoms).

Table 2Amino acid propensities to occur in specific secondary structures

% Helix % Strand % Others

Aspstr 21 31 48Aspcat 19 33 48Cysstr 19 18 63Cyscat 18 28 54Hisstr 34 20 47Hiscat 29 26 45

Asp— Aspartic acid, Cys— Cysteine, His— Histidine. Subscript denotes thetype of Proteins: str —Structural proteins, cat — Catalytic proteins.

1251K. Patel et al. / Biochimica et Biophysica Acta 1774 (2007) 1247–1253

for the helical and strand-type ϕ, ψs in the catalytic proteins.Cys ligands reflect comparable preferences for the helical andstrand-type ϕ, ψs, while Asp ligands reflect a greater preferencefor the strand-type ϕ, ψs.

Among 228 structural proteins, 37 Zn sites are coordinated to awater molecule, while 15 Zn sites are coordinated to an inorganicmoiety. Among 154 catalytic proteins, 106 Zn sites are withmetal-bound water or hydroxide ion and 60 sites are with a Zn-bound inhibitor, with the inhibitor having either displaced wateror increased the coordination number of zinc. Among the catalyticproteins, the Zn sites tend to be solvent accessible, while amongthe structural proteins, they tend to be solvent sequestered withexception of insulin which has one site exposed to solvent.

His side chain can ligate in alternate tautomeric forms, eithervia δ tautomer with δ nitrogen protonated or via ε tautomer withε nitrogen protonated. About 73% of His residues are coor-dinated via ε nitrogen and 27% via δ nitrogen. Being similar incoordination geometry, the tautomers are treated together fromnow on. Imidazole rings display ±10° deviations from the meanplane of metal coordination.

3.3. Geometry of primary coordination sphere of Zn

There are no significant variations in theZn–Xbond lengths (Xbeing any coordinating atom), except withX as sulfur. The data onbond lengths, including standard deviations in the bond lengths,are summarized in Table S3 under Supplementary Materials. Zn–

N and Zn–O bonds are of comparable length, while Zn–S bondsare appreciably longer, due to larger atomic size of sulfur.

The data on angles between Zn and coordinating atoms are inTable 3. Deviations from ideal values occur and could involveelectron pair repulsions, or effects like bidentate coordination ofcarboxyl group, extraneous H-bonding to secondary coordina-tion sphere or solvent, and bridging ligands in multi-Zn sites. Inthe dataset of angles, O–Zn–O angles of protein residues areclassified in two categories, the angle involving oxygen atomsfrom two different residues and the angles involving bothoxygen atoms from the same residue, i.e., bidentate arrange-ment, denoted by O–Zn–O′. The latter angles are smaller withaverage value of ∼53° for Eq–Zn–Ax (Here, Eq=Equitorial,Ax=Axial) and in smaller deviations than other angles. Someimportant observations are as follows.

Table3

Ang

lesof

Zinccoordinatedatom

s

Ang

les

N–Zn–

NN–Zn–

SN–Zn–

ON–Zn–

O(w

)N–Zn–

O(i)

S–Zn–

S!S–Zn–

OS–Zn–

O(w

)S–Zn–

O(i)

O–Zn–

OO–Zn–

O′

O–Zn–

O(w

)O–Zn–

O(i)

C4

109.5

107(8),71

109(6),113

106(15),91

110(10),38

113(18),25

110(8),40

210

7(8),30

106(6),18

108(9),24

109(15),35

53(2),6

111(24),19

111(20),11

106(7),

122

111(7),

4810

6(16),14

9112(17),61

112(14),34

111(7),

8510

7(7),

2210

7(7),

18117(10),10

111(22),66

53(4),

12111(22),25

108(12),14

C5

Eq–

Zn–

Ax

9097

(4),13

92(6),34

97(7),6

88(14),7

93(7),19

54(4),8

88(12),14

89(14),7

95(7),

4493

(8),

7792

(7),

2192

(8),

2610

3,1

97(10),5

93(6),

1854

(3),

1890

(12),34

89(10),8

Eq–

Zn–

Eq

120

112(5),6

118(8),20

123(8),15

128,

1118(7),8

119(14),4

122(15),3

116(12),27

109(2),

3118(7),

2512

0(9),

3512

0(9),

1713

3,1

124,

1118(10),13

120(10),13

129(12),3

Ax–Zn–

Ax

180

165,

115

8(13),7

151,

115

6(16),6

137,

115

6(2),2

166(9),

415

6(10),13

162(6),

316

5(8),

515

9(10),11

164(9),

317

0(12),5

C5′

Adjacent

9091

(10),4

91(11),8

98(12),7

56,1

82(17),3

99(6),

2510

0(5),

294

(9),

4295

(10),20

96(7),

2091

,1

97(16),21

56(3),

790

(11),24

92(8),

12Opp

osite

180

169,

114

0,1

141(6),2

167(5),2

156(13),5

164(10),8

157(10),10

147,

116

0(9),

1016

0(10),10

161(7),

2C6

Adjacent

9097

(5),9

93(8),37

94(6),16

87(7),5

107(17),12

57(4),6

91(11),29

85(11),11

98(6),

1892

(8),

8395

(7),

2393

(9),

1799

(17),28

54(4),

1592

(8),

3993

(13),47

Opp

osite

180

146(18),4

168(6),11

165(8),4

165(10),7

169(6),4

160,

115

5(14),8

167(8),

1616

4(6),

1016

3(12),22

156(4),

315

8(10),10

Anglesaregivenin

degrees,standard

deviations

aregivenin

parenthesesfollo

wed

bythenumberof

observations.C4=Tetrahedralgeom

etry,C5=Trigo

nalBipyram

idal,C5′=Squ

arePyram

idal,C6=Octahedral,

Eq-Equatorial,Ax-Axial.V

aluesin

italicstype

areforstructural

zinc

sitesin

proteincrystalstructures

from

thePDBandvalues

inbold

areforzinc

sitesin

crystalstructures

ofenzymes.

1252 K. Patel et al. / Biochimica et Biophysica Acta 1774 (2007) 1247–1253

3.4. Zn in catalytic sites

Among C4 sites, though the average angles are close to ideal,some distortions like O–Zn–ONN–Zn–O∼N–Zn–N are noted.This could be explained on the basis of electron pair repulsionmodel. In C5 coordination sites, the Eq–Zn–Ax angles are nearerthe ideal value but are greater in the range of deviation. The Ax–Zn–Ax angles are farther from the ideal and with greater standarddeviation. The angles in octahedral geometry are generally fartherfrom the ideal value. The observed deviations are similar to thosein C4 and C5 geometries. In contrast of the angle deviations in C5geometry, the diagonal angles in C6 are nearer to the ideal values.

3.5. Zn in structural sites

The trends are similar to those in the catalytic proteins. Theaverage values in C4 Zn sites are close to the ideal tetrahedralgeometry. In C4 coordination sites, the observed trend in anglesinvolving nitrogen and oxygen is N–Zn–N∼N–Zn–ObO–Zn–O, in agreement of simple electron pair repulsion model, butconverse of the reports ofAlbert et al. [18]. A possible explanationis that in the Albert et al. study the monodentate and bidentateangles involving oxygen atom were clubbed, while these areanalyzed separately in this study; as a result of the separation fromthe bidentate O–Zn–O′ case, the average O–Zn–O angle isincreased. The S–Zn–S angles are closer to the ideal tetrahedralvalue, in contrast of the angles in catalytic proteins, whereasS–Zn–N angles are similar to those in the catalytic proteins.

A comparatively greater deviation of C5 angles from theideal geometry is observed. Particularly, Ax–Zn–Ax anglescould be up to 30° deviated from the ideal angle 180°. TheO–Zn–O angles are on the higher side for Eq–Zn–Ax whenthe bidentate arrangements are excluded, in accordance withthe earlier reported observation [18].

The observed deviations in C6 geometry are similar to thosein C4 and C5 geometries. With absence of Cys in C6 systems,there is an increased propensity for oxygen donors. The angle

Table 4Ligand to metal distances in bidentate arrangements in Zinc proteins

Co-ordNo.

Aspartate Glutamate

Zn–O′ Zn–Oʺ Zn–O′ Zn–Oʺ

C4 1.97 (0.17), 42.02 (0.20), 11

2.69 (0.10), 42.62 (0.25), 11

2.14 (0.02), 22.07 (0.08), 8

2.52 (0.28), 22.60 (0.29), 8

C5′ 2.20 (0.15), 4 2.52 (0.23), 4 2.07 (0.12), 7 2.47 (0.17), 7C5 1.92 (0.0), 2

1.99 (0.06), 42.76 (0.13), 22.58 (0.28), 4

2.05 (0.14), 82.04 (0.08), 6

2.67 (0.19), 82.48 (0.39), 6

C6 2.08,12.08 (0.07), 5

2.31,12.52 (0.31), 5

2.05 (0.0), 21.99 (0.12), 3

2.50 (0.04), 22.64 (0.29), 3

O′ and Oʺ denotes the two carboxylate oxygen atoms bound to zinc in abidentate arrangement. Zn–O′ shows the average distance of the nearer oxygenatom while Zn–Oʺ denotes the average distance between farther oxygen ofcarboxylate and zinc ion. Distances are in Å, standard deviations given in theparentheses followed by the number of observations. Values in italics type arefor structural zinc sites in protein crystal structures from the PDB and values inbold are for catalytic zinc sites in protein crystal structures. Coordinationnumbers 4, 5, 6 are denoted by C4, C5 and C6 respectively. C5′ denotes squarepyramidal geometry while C5 denotes trigonal bipyramidal geometry.

1253K. Patel et al. / Biochimica et Biophysica Acta 1774 (2007) 1247–1253

between diagonally placed atoms is closer to the ideal 180° thanin C5 geometry.

3.6. Bidentate carboxylate groups

Among different kinds of bidentate arrangements possible,the most common are the simple and the bridging. Nearly all-bidentate interactions of Zn are of simple type; both oxygenatoms of Glu or Asp tend to ligate the same Zn. As noted inTable 4, the Zn–O distances in bidentate carboxylates are quitevaried; typically one of the oxygen tends to be nearer to Zn, butthere are also cases of similar distance; the distances are nearlythe same irrespective of the coordination number. The averagedistances for Zn coordination are shorter in the catalytic zincproteins than in the structural zinc proteins, with aspartate as theexceptions. A significant observation is that the bond length ofZn–O′ in bidentate arrangement is shorter than the Zn–Odistances in monodentate arrangement, which may result fromthe smaller O–Zn–O′ angle allowing a greater penetrations ofthe ligand atom into the Zn coordination sphere.

4. Conclusion

Extending previous reports [8,12,14,17–21,26], we haveanalyzed Zn coordination centers in the database of currentlyavailable NMR and X-ray protein structures for specific featuresof geometry and conformation. Structural databases will beinherently biased against the proteins unstructured and theproteins recalcitrant to structure characterization. But for thiscaveat, the present analysis encompassing a larger dataset ofbetter-filtered structures perhaps better captures both thediversity and the consensus for possible structures in thecurrently known ZBPs. Although excellent in the exposition ofstructure and catalysis the currently available reviews of thestructure in zinc centers [8,12–18,21] do not adequately capturethe diversity and consensus for zinc binding and function. Manymore Zn–S4 coordinated structures have been documented in therecent years, establishing Cys as the prominent Zn ligand overHis in structural proteins [9,10]. Better capturing the consensusof structure [9], and the subtleties of metal coordination, thepresent study will possibly better serve the cause of character-ization, engineering, and design of Zn proteins.

Acknowledgements

This work was supported by grant from Board of Research inNuclear Sciences, India (BRNS) and Council of Scientific andIndustrial Research (CSIR), Government of India. KP and AKare recipients of fellowship from CSIR, India.

Appendix A. Supplementary data

Supplementary data associated with this article can be found,in the online version, at doi:10.1016/j.bbapap.2007.07.010.

References

[1] J.M. Berg, Y. Shi, The galvanization of biology: a growing appreciation forthe roles of zinc, Science 271 (1996) 1081–1085.

[2] D.S. Auld, Structural zinc sites, Handbook of Metalloproteins, vol. 3,2004, pp. 403–415.

[3] J.E. Coleman, Zinc proteins: enzymes, storage proteins, transcriptionfactors, and replication proteins, Ann. Rev. Biochem. 61 (1992) 897–946.

[4] H. Cho, S. Ramaswamy, B.V. Plapp, Flexibility of liver alcoholdehydrogenase in stereoselective binding of 3-butylthiolane 1-oxides,Biochemistry 36 (1997) 382–389.

[5] S. Lindskog, A. Liljas, Carbonic anhydrase and the role of orientation incatalysis, Curr. Opin. Struck. Biol. 3 (1993) 915–920.

[6] U. Derewenda, Z. Derewenda, E.J. Dodson, G.G. Dodson, C.D. Reynolds,G.D. Smith, C. Sparks, D. Swenson, Phenol stabilizes more helix in a newsymmetrical zinc insulin hexamer, Nature 338 (1989) 594–596.

[7] L. Banci, I. Bertini, F. Cramaro, C.R. Del, M.S. Viezzoli, Solution structureof Apo Cu,Zn superoxide dismutase: role of metal ions in protein folding,Biochemistry 42 (2003) 9543–9553.

[8] J.S. Hanas, D.J. Hazuda, D.F. Bogenhagen, F.Y. Wu, C.W. Wu, Xenopustranscription factor A requires zinc for binding to the 5 S RNA gene, J. Biol.Chem. 258 (1983) 14120–14125.

[9] C. Andreini, L. Banci, I. Bertini, A. Rosato, Counting the zinc–proteinsencoded in the human genome, J. Proteome Res. 5 (2006) 196–201;C. Andreini, L. Banci, I. Bertini, A. Rosato, Zinc through the three domainsof life, J. Proteome Res. 11 (2006) 3173–3178.

[10] D.S. Auld, Zinc coordination sphere in biochemical zinc sites, BioMetals14 (2001) 271–313.

[11] L. Regan, Protein design: novel metal-binding sites, Trends Biochem. Sci.20 (1995) 280–285.

[12] D.W. Christianson, Structural biology of zinc, Adv. Protein Chem. 42 (1991)281–355.

[13] S. Karlin, Z.Y. Zhu, Classification of mononuclear zinc metal sites inprotein structures, Proc. Natl. Acad. Sci. U. S. A. 94 (1997) 14231–14236.

[14] D.W. Christianson, R.S. Alexander, Another catalytic triad? Nature19 (346) (1990) 225.

[15] P. Chakrabarti, Geometry of interaction of metal ions with histidineresidues in protein structures, Protein Eng. 4 (1990) 57–63.

[16] A. Vedani, D.W. Huhta, A new force field for modeling metalloproteins,J. Am. Chem. Soc. 112 (12) (1990) 4759–4767.

[17] B.L. Vallee, D.S. Auld, Zinc: biological functions and coordination motifs,Acc. Chem. Res. 26 (10) (1993) 543–551.

[18] I.L. Alberts, K. Nadassy, S.J. Wodak, Analysis of zinc binding sites inprotein crystal structures, Protein Sci. 7 (1998) 1700–1716.

[19] M.M. Harding, The geometry of metal–ligand interactions relevant toproteins. II. Angles at the metal atom, additional weak metal–donorinteractions, Acta Crystallogr., D Biol. Crystallogr. 56 (2000) 857–867.

[20] M.M. Harding, Small revisions to predicted distances around metal sites inproteins, Acta Crystallogr., D Biol. Crystallogr. 62 (2006) 678–682.

[21] D.W.Christianson, R.S. Alexander, Carboxylate–histidine–zinc interactions inprotein structure and function, J. Am. Chem. Soc. 111 (16) (1989) 6412–6419.

[22] H.M. Berman, J. Westbrook, Z. Feng, G. Gilliland, T.N. Bhat, H. Weissig,I.N. Shindyalov, P.E. Bourne, The Protein Data Bank, Nucleic Acids Res.28 (2000) 235–242.

[23] G. Wang, R.L. Dunbrack Jr., PISCES: a protein sequence culling server,Bioinformatics 19 (2003) 1589–1591.

[24] B.L. Vallee, D.S. Auld, Short and long spacer sequences and other structuralfeatures of zinc binding sites in zinc enzymes, FEBS Lett. 257 (1) (1989)138–140.

[25] B.L.Vallee, J.E.Coleman,D.S.Auld, Zinc fingers, zinc clusters, and zinc twistsin DNA-binding protein domains, Proc. Natl. Acad. Sci. U. S. A. 88 (3) (1991)999–1003.

[26] B.L. Vallee, D.S. Auld, Zinc coordination, function, and structure of zincenzymes and other proteins, Biochemistry 29 (24) (1990) 5647–5659.