Crystal structure and silica condensing activities of silicatein a–cathepsinL chimeraswzMichael Fairhead,ya Kenneth A. Johnson,yb Thomas Kowatz,yb Stephen A. McMahon,b

Lester G. Carter,b Muse Oke,b Huanting Liu,b James H. Naismith*b andChristopher F. van der Walle*a

Received (in Cambridge, UK) 28th November 2007, Accepted 25th January 2008

First published as an Advance Article on the web 11th February 2008

DOI: 10.1039/b718264c

Cathepsin L mutants with the ability to condense silica fromsolution have been generated and a 1.5 A crystal structure of oneof these chimeras allows us to rationalise the catalytic mechan-ism of silicic acid condensation.

Silica materials are of interest in biotechnology and drugdelivery1 but controllable silicate formation is very di!cult.Yet a variety of sponge species make highly ordered specificglass structures called spicules.2–5 Harnessing the processesthat underlie this biosynthesis has considerable application.The enzyme silicatein a forms part of the organic filamentfound in spicules which in situ condenses silicate (Fig. 1a),although the exact form of the natural substrate is notknown.5 Both wild type and recombinant silicatein a havebeen shown to catalyse the condensation of siloxanes such astetraethoxysilane4,6 in solution and at surfaces.7 Silicatein ahas also been shown to have the ability to cause the depositionof other compounds at surfaces, including titanium phosphate,titanium oxide, zirconium oxide and l-lactide.8,9

Above 100 ppm, condensation of silicic acid occurs sponta-neously, presumably as the concentration of nucleophilicionised molecules is high enough. There are two possiblemechanisms for enzyme catalysis: (i) stabilise at the active siteone molecule of deprotonated silicic acid (the nucleophile)which will then react with another molecule of silicic acid; or(ii) stabilise a protonated silicic acid (the electrophile) whichwill then react with another molecule of silicic acid. Neitherthe wild type nor recombinant silicatein a is amenable tobiophysical study due to low levels of protein expression andinclusion body formation when recombinantly expressed inE. coli.4,7,8

Sequence comparison identifies a significant degree of simi-larity between silicatein a and the human cysteine proteasecathepsin L4 with an overall 52% identity and 65% similarity.The most notable di"erences between the two sequences are

the presence of a loop of four amino acids in cathepsin L 10

that is absent in silicatein a, a large number of hydroxylcontaining residues (serines, threonines and tyrosines) in sili-catein a that are not present in cathepsin L and the substitu-tion of the catalytic cysteine (C25) in cathepsin L 11 for serine(S25) in silicatein a. This serine residue has been shown bymutagenesis to be essential for the function of silicatein a.12

The other residues in the catalytic triad of cathepsin L (H163and N187) are similarly conserved in silicatein a. A mechanismhas been proposed for polymerisation of Si(OEt)4 in which acovalent protein silicate intermediate is formed; this activeelectrophile is then decomposed by water yielding a moreactive nucleophile.12 Whether this mechanism operates in vivois unresolved. Interestingly, there is also a change in theflanking residues either side of the catalytic serine and histidineand this di"erence is conserved across several di"erent silica-teins (Fig. 1b).We have made a series of cathepsin L mutants that increas-

ingly match the sequence features that are unique to silicateina (Table 1). Mutants were made using a combination of theQuikchangeTM mutagenesis technique and conventional PCR.The mutant proteins were then recombinantly expressed as thepro-protein and purified from Pichia pastoris as describedpreviously.13 Mature cathepsin L–silicatein a chimeras werethen obtained via published procedures.14

The various mutant proteins were assayed for silica con-densation activity (detected by precipitation of silicate) using

Fig. 1 (a) The chemical reaction catalysed by the organic filament of

the sponge spicule. (b) Conservation of residues flanking catalytic

serine and histidine in silicatein-a from various sponge species. Shown

is the amino acid sequence around the catalytic serine/cysteine and

histidine in human cathepsin L (UniProtKB/TrEMBL entry P07711)

and silicatein a from: Tethya aurantia (O76238); Suberites domuncula

(Q2MEV3); Hymeniacidon perleve (Q2HYF6); Lubomirskia baicalensi

(Q2PC18).

a SIPBS, University of Strathclyde, 27 Taylor Street, Glasgow,Scotland, UK. E-mail: chris.walle@strath.ac.uk

b The Centre for Biomolecular Science and The Scottish StructuralProteomics Facility, University of St. Andrews, North Haugh, St.Andrews, Scotland, UK. E-mail: naismith@st-andrews.ac.uk

w Electronic Supplementary Information (ESI) available: Molecularbiology, protein purification and crystallographic details.z The final crystal structure and experimental have been depositedwith the Protein Data Bank, entry 2VHS. Links to PDB visualisationsare included in the online supplementary data.y These authors contributed equally.

This journal is !c The Royal Society of Chemistry 2008 Chem. Commun., 2008, 1765–1767 | 1765

COMMUNICATION www.rsc.org/chemcomm | ChemComm

water glass as a substrate (Fig. 1a). Mutation of the catalyticcysteine to serine (the C25S mutant) did not confer significantlevels of condensing activity. The C25S mutant would havebeen expected to be su!cient from the predicted mechanism ofsilicatein a.6 In order to obtain significant condensing activity,it was found to be necessary to mutate the residues either sideof the catalytic S25 residue (Table 1, Fig. 2) before anysignificant activity was observed. Additional mutations tocathepsin L to further increase its resemblance to silicatein adid not significantly increase the ability to condense silica; infact many of these additional mutations tended to decreasecondensation activity. It is likely, therefore, that these residuesplay other functions in the sponge spicule such as in theassociation of silicatein a with silicatein b and galectin,15,16

or in the templating of silica into specific morphologies, ratherthan underlie catalysis per se. While the various active mutantproteins were found to readily precipitate silica from thenatural sodium silicate substrate, we could not detect anyprecipitate from tetraethoxysilane substrate, which can beutilised by silicatein-a.4,6

In an e"ort to understand the catalytic basis of silicacondensation we have solved the X-ray crystal structure ofthe 4SER mutant (Fig. 3a). Crystals of the 4SER mutant were

obtained by hanging drop vapour di"usions. The structurewas determined to 1.5 A using cathepsin L (PDB 1MHW) as amodel for molecular replacement.19,20 Unsurprisingly, the4SER chimera is identical to cathepsin L with a RMSD ofonly 0.5 A between the two structures. Full experimentaldetails are given in the ESI.wFig. 3b shows the active site architecture of the 4SER

mutant with that of the C25S pro-cathepsin L crystal struc-ture21 (PDB 1CJL). The C25S mutant of cathepsin is itselfinactive as a protease, indicating the nucleophilicity of the OGatom of S25 is low. In 4SER the distance between the OGatom on S25 and ND1 on H163 has increased to 3.6 A(beyond hydrogen bonding distance), further decreasing thenucleophilicity of S25. Mutation of the flanking residuesperturbs a cluster of residues that sit behind S25 (Fig. 3).The mutations remove a hydrogen bond (S24A) and decreasevan der Waal interaction (W26Y). The mutations are both toresidues smaller than the native enzyme. The net result wouldbe to allow the S25 to move away from the H163 and toenlarge the volume at the active site.

Table 1 Mutant constructs made

Cathepsin Lconstructa Mutations Match to silicatein a

C2S C25S Catalytic serineAS2 S24A W26Y Residues flanking

catalytic serineAS2H M161L, D162N,

G164A, V165MResidues flankingcatalytic histidine

LOOP 173ESTESDNN180 toISNNQ

To replicate loop

2SER E159S, D160S Match serines4SER E153S, P154S Match serinesa The mutations are additive from top to bottom, thus 4SER has allthe preceding mutations.

Fig. 2 Silica condensing activity of mutant proteins. To 0.1 mg mL"1

of protein solution in 0.1 M sodium phosphate bu"er pH 7 was added

4.5 mM sodium silicate. Samples were incubated at room temperature

for 24 h; precipitated silica was collected by centrifugation and then

assayed using the Merck spectroquant silicon assay kit as described.17

Shown is the average value of five separate incubations with the error

as shown.

Fig. 3 (a) The structure of the 4SER chimera shown in cartoon form.

The sulfate at the active site is shown as space filling spheres. The two

‘catalytic’ residues H163 and S25 are shown as sticks. Carbon atoms

are coloured yellow, oxygen red, sulfur green and nitrogen blue. (b)

Close up view of the active site of the 4SER chimera showing the close

contacts (o3.5 A) as dashed lines. The sulfate molecule is shown as

sticks; the colour scheme is the same as in part (a). The extensive

interactions mediated by water would allow extensive proton

shuttling. Figures produced with PYMOL.18

1766 | Chem. Commun., 2008, 1765–1767 This journal is !c The Royal Society of Chemistry 2008

An understanding of the mechanism of silica condensationby the chimera is aided by the presence of a sulfate ion at theactive site. (A second sulfate is at a crystal contact 17 A distantfrom S25). The crystal structure reveals it would be di!cult tofit the bulky Si(OEt)4 at the active site without fairly signifi-cant conformational adjustment. This would seem to explainthe lack of activity of the chimera against Si(OEt)4 and wouldsuggest that silicatein a must have a larger active site than4SER which would allow it to accommodate the bulkySi(OEt)4. We believe the tetrahedral sulfate ion at the activesite is a good mimic of the tetrahedral silicic acid. The sulfate ishydrogen bonded to S25. This is consistent with S25 making anucleophilic attack on Si(OH)4 to generate a covalent enzymeintermediate. However, such a high energy intermediate wouldseem chemically unlikely and prone to immediate hydrolysisback to Si(OH)4.The sulfate is hydrogen bonded to H163, the residue that

activates the protein nucleophile in proteases. We favour amechanism in which H163 binds and stabilises the deproto-nated form of Si(OH)4 at the active site. The extensive networkof water molecules and hydrogen bonds (Fig. 3b) wouldpermit significant proton shuttling, such that the negativecharge may reside on a di"erent oxygen. This deprotonatedenzyme bound species is a su!cient nucleophile that it willattack Si(OH)4 solution species initiating polymerisation(Fig. 4). Loss of a Si(OH)4 hydroxyl group carried out byprotonation and elimination of water could also promote thereaction but we have no evidence for an acid. It is also unlikelythat at pH 7 Si(OH)4 will be a su!cient base to first depro-tonate H163 (whose pKa is likely to be lower than 7 due to thepresence of D187) and then for the hydroxyl group of thesecond Si(OH)4 to perform a nucleophilic attack.In our proposed mechanism, the roles of C25S and flanking

mutations are simply to create a su!ciently sized pocket thatwill allow recognition of the tetrahedral Si(OH)4 molecule insuch a way that H163 can deprotonate it. The deprotonatedSi(OH)4 protein complex can be thought of as a template forcondensation reaction. In this proposal there is no need to

involve a high energy covalent intermediate. The presence of aspecific Si(OH)4 transporter in sponge22 suggests the truesubstrate in vivo is indeed silicic acid, not high energy siliconalkoxides. This being so, the simple acid base activationmechanism we propose seems a good model for the biologicalprocess.This work was supported by the Leverhulme Trust, grant

F/00 273/H. The Scottish Structural Proteomics Facility isfunded by the Scottish Funding Council and the Biotechnol-ogy and Biological Sciences Research Council.

Notes and references

1 S. Giri, B. G. Trewyn and V. S. Lin, Nanomedicine, 2007, 2,99–111.

2 W. E. Muller, S. I. Belikov, W. Tremel, C. C. Perry, W. W.Gieskes, A. Boreiko and H. C. Schroder, Micron, 2006, 37,107–120.

3 H. C. Schroder, O. Boreiko, A. Krasko, A. Reiber, H. Schwertnerand W. E. Muller, J. Biomed. Mater. Res., Part B, 2005, 75,387–392.

4 K. Shimizu, J. Cha, G. D. Stucky and D. E. Morse, Proc. Natl.Acad. Sci. U. S. A., 1998, 95, 6234–6238.

5 J. C. Weaver and D. E. Morse, Microsc. Res. Tech., 2003, 62,356–367.

6 J. N. Cha, K. Shimizu, Y. Zhou, S. C. Christiansen, B. F.Chmelka, G. D. Stucky and D. E. Morse, Proc. Natl. Acad. Sci.U. S. A., 1999, 96, 361–365.

7 M. N. Tahir, P. Theato, W. E. Muller, H. C. Schroder, A. Jansho",J. Zhang, J. Huth and W. Tremel, Chem. Commun., 2004,2848–2849.

8 P. Curnow, D. Kisailus and D. E. Morse, Angew. Chem., Int. Ed.,2006, 45, 613–616.

9 M. N. Tahir, P. Theato, W. E. Muller, H. C. Schroder, A. Borejko,S. Faiss, A. Jansho", J. Huth and W. Tremel, Chem. Commun.,2005, 5533–5535.

10 S. M. Smith and M. M. Gottesman, J. Biol. Chem., 1989, 264,20487–20495.

11 R. W. Mason, G. D. Green and A. J. Barrett, Biochem. J., 1985,226, 233–241.

12 Y. Zhou, K. Shimizu, J. N. Cha, G. D. Stucky and D. E. Morse,Angew. Chem., Int. Ed., 1999, 38, 780–782.

13 M. Fairhead and C. F. van der Walle, Protein Pept. Lett., 2008, 15,47–53.

14 R. Menard, E. Carmona, S. Takebe, E. Dufour, C. Plou"e, P.Mason and J. S. Mort, J. Biol. Chem., 1998, 273, 4478–4484.

15 P. Curnow, P. H. Bessette, D. Kisailus, M. M. Murr, P. S.Daugherty and D. E. Morse, J. Am. Chem. Soc., 2005, 127,15749–15755.

16 H. C. Schroder, A. Boreiko, M. Korzhev, M. N. Tahir, W. Tremel,C. Eckert, H. Ushijima, I. M. Muller and W. E. Muller, J. Biol.Chem., 2006, 281, 12001–12009.

17 W. E. Muller, A. Krasko, G. Le Pennec and H. C. Schroder,Microsc. Res. Tech., 2003, 62, 368–377.

18 W. L. DeLano, DeLano Scientific, San Carlos, CA, USA, 2002.19 A. J. McCoy, R. W. Grosse-Kunstleve, L. C. Storoni and R. J.

Read, Acta Crystallogr., Sect. D: Biol. Crystallogr., 2005, 61,458–464.

20 L. C. Storoni, A. J. McCoy and R. J. Read, Acta Crystallogr., Sect.D: Biol. Crystallogr., 2004, 60, 432–438.

21 R. Coulombe, P. Grochulski, J. Sivaraman, R. Menard, J. S. Mortand M. Cygler, EMBO J., 1996, 15, 5492–5503.

22 H. C. Schroder, S. Perovic-Ottstadt, M. Rothenberger, M. Wiens,H. Schwertner, R. Batel, M. Korzhev, I. M. Muller and W. E. G.Muller, Biochem. J., 2004, 381, 665–673.

Fig. 4 The chemical mechanism proposed to operate for the poly-

merisation of silicic acid. We propose H163 deprotonates Si(OH)4; the

extensive water network could permit proton shuttling. We have

illustrated one possibility.

This journal is !c The Royal Society of Chemistry 2008 Chem. Commun., 2008, 1765–1767 | 1767

Supplementary Material (ESI) for Chemical Communications This journal is © The Royal Society of Chemistry 2008

Crystal structure and silica condensing activities of silicatein !/cathepsin L

chimeras.

Michael Fairheada, Kenneth A. Johnsonb, Stephen A. McMahonb, Thomas Kowatzb, Lester G. Carterb,

Muse Okeb, Huanting Liub, James H. Naismith*b and Christopher F. van der Walle*a

a SIPBS, University of Strathclyde, 27 Taylor Street, Glasgow, Scotland. E-mail:

chris.walle@strath.ac.uk b The Centre for Biomolecular Science, University of St. Andrews, North Haugh, St. Andrews,

Scotland. E-mail: naismith@st-andrews.ac.uk

Additional supporting material Molecular biology

The procathepsin-L gene (amino acids 18 to 333) was amplified from clone 6712564 (LGC

Promochem) harbouring preprocathepsin-L (UniProtKB/TrEMBL entry P07711) by PCR using

primers PCLFP01 (5´-end) and PCLRP01 (3´-end) and ligated into Pichia pastoris vector pPICZ-!-B

(Invitrogen) which contains a gene for ZeocineTM resistance as well as an ! factor secretion signal

which leads to secretion of recombinant protein into the growth media. Mutations to the loop region

(amino acids 285-294) were made using primers PCLFP01 and LOOPRP01 in a standard PCR

protocol, digesting the product with PstI and SspI and ligating into pPICZ-!-procathepsin-L, similarly

digested. All other mutations were made using the Quikchange™ method (Stratagene) and the

sequences confirmed by the University of Dundee Sequencing Service.

Primer Sequence 5´ to 3´ Purpose

PCLFP01 GCGAAGGCTGCAGGAGCTACTCTAACATTTGAT Introduce PstI site

PCLRP01 GATGGGGTGACACATTGGCGCCTAGAGC Introduce SacII site

CATFPCS a CAGGGTCAGTGTGGTTCTTCTTGGGCTTTTAGTGC C25S mutation

CATFPASY a CAGGGTCAGTGTGGTGCTTCTTACGCTTTTAGTGCT

ACTG

S24A W26Y

CATFPLNHAM a GACTGTAGCAGTGAAGACTTGAATCATGCTATGCTG

GTGGAAGGC

M161L, D162N, G164A, V165M

mutation

LOOPRP01 CACGACCACCAACCGATGCCTAAATAGAGGTTGTT

GGTCTTTATAACCGAC

173ESTESDNN180 to ISNNQ mutation

CATFP2S a TTTATTTTGAGCCAGACTGTAGCAGTTCAAGCTTGA

ATCATGCTATGCTGGTGGTTG

E159S, D160S mutation

CATRP4S a GTCCTTCCTGTTCTATAAAGAAGGCATTTATTTTTCG

TCAGACTGTAGCAGTTCAA

E153S, P154S mutation

a only the forward primer is shown

Supplementary Material (ESI) for Chemical Communications This journal is © The Royal Society of Chemistry 2008

Protein expression and purification

Pichia pastoris strain X-33 was transformed with plasmid using lithium chloride as described in the

Invitrogen EasyselectTM Pichia Expression Kit manual. Transformed cells were plated onto YPG-agar

plates (2% yeast extract, 2% peptone, 2% glycerol, 2% agar) containing 100 !g/mL of zeocin. After 3

days at 31 °C the resulting colonies were screened for protein expression. Cells were grown in 50 mL

of YPG medium in a 250 mL baffled flask (Nalgene, UK) for 48 h at 30 °C. Cells were then

transferred to 1000 mL of YPM (2% yeast extract, 2% peptone, 1% methanol) medium and grown in

a 2 L baffled flask for 96 h at 20 °C, with fresh methanol added every 24 h. Wet cell biomass at the

end of this period was typically 70-80 g/l. The clarified supernatant was concentrated to 40 mL,

(NH4)2SO4 added to 1 M and loaded onto a 20 mL Phenyl Sepharose 6 Fast Flow column (GE

healthcare, UK) pre-equilibrated with buffer A (20 mM Tris.HCl, pH 8.2) + 1 M (NH4)2SO4. After

loading, the column was washed with buffer A + 1 M (NH4)2SO4, the protein eluted isocratically in

buffer A and desalted. The sample was then loaded onto a 5 mL Q Sepharose column and eluted with

a linear gradient from 0 to 1 M NaCl. Protein containing fractions were exchanged into buffer B (20

mM CH3COONa pH 5), loaded onto a 5 mL SP Sepharose column and eluted with a linear gradient

from 0 to 1 M NaCl. Typically, 20-40 mg of pure protein were obtained per litre of culture

supernatant, depending upon the construct being expressed. Processing of inactive procathepsin-L

mutants to the mature form was done by incubating 20 mg of the selected mutant with 0.2 mg of

activated wild type cathepsin-L in 0.1 M CH3COONa pH 5, with 2 mM DTT for 48-96 h at room

temperature. Processing was checked by SDS-PAGE and continued until a single band at 24 kDa

(mature form) was present. Processed enzyme was then exchanged into 10 mM CH3COONa pH 5 and

stored at –80 C.

Structural Biology

The protein was screened for crystallization against 4 commercial 96-condition screens: 1)Wizard 1

and 2 (Emerald Biosystems); 2)JCSG+, 3)Classics and 4)Pegs (GE Healthcare). Conditions which

gave crystals were used to design optimization screens based on a simple stochastic method using in-

house software1. The crystal used for data collection was obtained by the hanging drop method using

protein at 10mg/ml mixed 2µl:1µl with the well solution of 450µl containing 17.36% PEG3350, 0.1M

sodium acetate, pH4.5, and 0.1M lithium sulphate. Prior to exposure to X-rays, the crystal was

transferred to a cryoprotecting solution containing 20% PEG3350, 0.1M sodium acetate, pH 4.5, 0.1M

lithium sulphate and 20% PEG400. The crystal was flash frozen in liquid nitrogen and transported in a

dry cryogenic dewar to beamline ID29 at the European Synchrotron Radiation Facility. Data were

collected from a single crystal at 100 K at a wavelength of 1.27 Å. The data were indexed and merged

using Denzo and Scalepack in the integrated package HKL20002.

Supplementary Material (ESI) for Chemical Communications This journal is © The Royal Society of Chemistry 2008

The structure was solved using PHASER3, 4 using a monomer of cathepsin (PDB 1MHW) as a model

finding four monomers in the asymmetric unit. Manual modification of the protein to convert the

sequence to 4SER used COOT5. The structure was refined using REFMAC56, TLS parameters,

isotropic B-thermal factors and NCS restraints were employed throughout. Water molecules were

added using COOT5. Structure quality was checked with PROCHECK7 and MOLPROBITY8. The

final structure and experimental data were deposited with the PDB9 and have accession code 2VHS.

Table 1 X-ray data

Data collection 4SER ! (Å) 1.27

Resolution Last shell (Å)

28 – 1.5 (1.55 - 1.5)

Spacegroup P1 Cell (Å) a = 56.8 b = 58.1 c = 70.2

! = 105.7 " = 105.0 # = 105.1 Unique refl’s 108723

Average redundancy 1.9 (1.7) I/$ 20 (2.1)

Complete (%) 92 (74) Rmerge 0.039 (0.293)

Refinement

R % 18.8 (26.4) Rfree % 22.0 (32.1)

rmsd bonds (Å) /angles (º) 0.006 / 0.985 Ramach’n favoured (%) 98

PDB accession code 2VHS

REFERENCES

1. B. Rupp, B. W. Segelke, H. I. Krupka, T. P. Lekin, J. Schafer, A. Zemla, D. Toppani, G. Snell and T. Earnest, Acta Cryst. D, 2002, 58, 1514-1518.

2. W. Minor, M. Cymborowski and Z. Otwinowski, Acta Physica Polonica A, 2002, 101, 613-619.

3. A. J. McCoy, R. W. Grosse-Kunstleve, L. C. Storoni and R. J. Read, Acta Cryst. D, 2005, 61, 458-464.

4. L. C. Storoni, A. J. McCoy and R. J. Read, Acta Cryst. D, 2004, 60, 432-438. 5. P. Emsley and K. Cowtan, Acta Cryst. D, 2004, 60, 2126-2132. 6. G. N. Murshudov, A. A. Vagin, A. Lebedev, K. S. Wilson and E. J. Dodson, Acta Cryst. D,

1999, 55, 247-255. 7. R. A. Laskowski, M. W. MacArthur, D. S. Moss and J. M. Thornton, J. Appl. Cryst., 1993,

26, 548-558. 8. I. W. Davis, L. W. Murray, J. S. Richardson and D. C. Richardson, Nucleic Acids Res, 2004,

32, W615-W619. 9. H. Boutselakis, D. Dimitropoulos, J. Fillon, A. Golovin, K. Henrick, A. Hussain, J.

Ionides, M. John, P. A. Keller, E. Krissinel, P. McNeil, A. Naim, R. Newman, T. Oldfield, J. Pineda, A. Rachedi, J. Copeland, A. Sitnov, S. Sobhany, A. Suarez-Uruena, J. Swaminathan, M. Tagari, J. Tate, S. Tromm, S. Velankar and W. Vranken, Nucleic Acids Res, 2003, 31, 458-462.