HPLC–MS characterization of cyclo-statherin Q-37, a specific cyclization product of human salivary statherin generated by transglutaminase 2

Tiziana Cabras1

Rosanna Inzitari2

Chiara Fanali2

Emanuele Scarano3

Maria Patamia4

Maria T. Sanna1

Elisabetta Pisano5

Bruno Giardina2,4,6

Massimo Castagnola2,4,6

Irene Messana1

1Department of Sciences Appliedto Biosystems, CagliariUniversity, Monserrato Campus,Monserrato, CA, Italy

2Institute of Biochemistry andClinical Biochemistry, CatholicUniversity, Rome, Italy

3Institute ofOtorhinolaryngology, CatholicUniversity, Rome, Italy

4Institute of Chemistry ofMolecular Recognition, NationalResearch Council (CNR), Rome,Italy

5Department of Surgery andOdontostomatological Sciences,Cagliari University, Cagliari,Italy

6International Scientific InstitutePaolo VI for Research on HumanFertility and Infertility, CatholicUniversity, Rome, Italy

Original Paper

HPLC–MS characterization of cyclo-statherin Q-37,a specific cyclization product of human salivarystatherin generated by transglutaminase 2

In the present study the analytical potential of HPLC–MS/MS was utilized for thestructural characterization of a post-translational modification of statherin. Humansalivary statherin (Mav 5380.0 € 0.3 Da) is transformed by the action of transglutami-nase 2 into a cyclic derivative with an average molecular mass of 5363.0 € 0.3 Da.The intra-molecular bridge is generated by the loss of an ammonia moleculebetween the unique lone-pair donating nucleophile Lys-6 and one acceptor amongthe seven glutamine residues of statherin. Digestion of the cyclic derivative withchymotrypsin, proteinase K, and carboxypeptidase Y, monitored by HPLC–electro-spray ionization-ion trap-mass spectrometric analysis, demonstrated that cycliza-tion involved almost specifically Gln-37 (A95%), with the percentage of Gln-39 impli-cated in the cross-linking being less than 5%. The main derivative was named cyclo-statherin Q-37. Guinea pig transglutaminase 2 showed high affinity for statherin invitro (Km = 0.65 € 0.06 lM). Cyclo-statherin was detected in vivo by HPLC-electrosprayionization ion trap-mass spectrometry analysis of whole human saliva and itaccounted for about 1% of total statherin. Detection of cyclo-statherin in wholesaliva is suggestive of a putative role of this molecule in the formation of the “oralprotein pellicle”.

Keywords: Cyclo-statherin / HPLC – ESI IT-MS / Human saliva / Transglutaminase /

Received: June 22, 2006; revised: July 11, 2006; accepted: July 17, 2006

DOI 10.1002/jssc.200600244

1 Introduction

Statherin, a phospho-peptide of 43 amino acid residuesfound in human saliva, exhibits very unusual character-istics [1]. This acidic peptide secreted by different salivaryglands has an anomalously high content of tyrosine, pro-line, and glutamine, and a high degree of structural andcharge asymmetry [2]. An important role of statherin isconnected with its great affinity for calcium phosphateminerals, such as hydroxyapatite (HAP) [3], and with itsability to inhibit precipitation and crystal growth fromsupersaturated solutions of calcium phosphate [4]. It hasalso been suggested that statherin may participate in theexchange and transport of calcium and phosphate dur-

ing the molecular events connected with its secretion inthe salivary glands [2], contributing to the distribution ofthe two ions at the level of the mucosal surface. In viewof its strong affinity for enamel and hydroxyapatite sur-faces, the involvement of statherin in the formation ofprotein layers, i. e. the so-called “pellicles”, has beenhypothesized and then demonstrated in vitro [5].Recently, Li et al. [6, 7] using anti-statherin monoclonalantibodies showed that statherin is an in vivo constituentof the human acquired enamel pellicle. This protein filmis important for the integrity of tooth enamel, since itacts as a boundary lubricant on the enamel surface [8].Moreover, interactions between pellicle proteins andbacterial surfaces are responsible for specificity of thebacterial colonization during the earliest stage of plaqueformation [9]. The in vivo pellicle is thought to be an inso-luble network of proteins generated by post-secretoryprocessing of proteins mainly including cross-linking.Cross-links were demonstrated first by Bradway et al. alsofor the formation of oral mucosal pellicle, a network ofproteins formed by components of saliva adsorbed ontobuccal epithelial cell surfaces that cover the oral mucosalsurface [10, 11]. This protein molecular network could

Correspondence: Professor Irene Messana, Department ofSciences Applied to Biosystems, Cagliari University, MonserratoCampus, 09042 Monserrato, CA, ItalyE-mail: [email protected]: +39-070-6754523

Abbreviations: CE, cell envelope; GPL TG2, guinea pig livertransglutaminase 2; PDA, photodiode array; PMSF, phenyl-methylsulfonyl fluoride; TIC, total ion current; XIC, extractedion current

i 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com

2600 T. Cabras et al. J. Sep. Sci. 2006, 29, 2600 –2608

J. Sep. Sci. 2006, 29, 2600 –2608 LC–MS/MS characterization of salivary cyclo-statherin 2601

interact with the oral epithelial-cell plasma membraneand its associate cytoskeleton and might contribute tothe mucosal epithelial flexibility and turnover. It wasdemonstrated that acidic-proline-rich proteins,statherin, and the major histatins are substrates of oraltransglutaminase 2 and they participate in cross-linkingreactions [12] as putative pellicle precursor proteins. Inparticular, transglutaminase 2 is able to cross-link acidicPRP-1 and statherin in vitro [13]. Therefore, statherinseems to take part in the formation of two different net-works, one involved in the dynamics of acquired enamelformation, the other involved in the turnover of the oralepithelia. Whatever the structure of these protein net-works may be, oral transglutaminases (mainly type 2transglutaminase) are the pivotal enzymes for pellicleformation [14]. TG2 is the ubiquitous tissue enzymeexpressed in oral epithelial cells, which catalyzes differ-ent biological processes and is involved in the formationof cross-links between the c-carboxy group of glutamineand the e-amine of lysine residues through amide bonds[15].

The purpose of the present investigation was to evaluateby different HPLC–ESI-IT-MS approaches the reactivity ofstatherin in the presence of type 2 transglutaminase andto characterize the structure of the reaction products.

2 Materials and methods

2.1 Materials and instrumentation

All general chemicals and reagents were of analyticalgrade and were purchased from Farmitalia-Carlo Erba(Milan, Italy), Merck (Darmstadt, Germany), and Sigma–Aldrich (St. Louis, MO, USA). Guinea pig liver transgluta-minase 2 (GPL TG2) was from Sigma Aldrich.

Purifications and analyses were performed either with aBeckman System Gold 125S HPLC system (Beckman, PaloAlto, CA, USA) equipped with a diode-array detector, orwith the Surveyor HPLC–ESI-MS apparatus from Thermo-Finnigan (San Jose, CA, USA). The Surveyor HPLC systemwas equipped with a photodiode array (PDA) detectorand connected with the mass spectrometer Xcalibur LCQDeca XP Plus ThermoFinnigan (San Jose, CA, USA) by anelectrospray ion source.

The Beckman apparatus was used to perform preparativeseparations with an Hyperprep PEP (Alltech, KY, USA)C18 column (250610 mm, 8 lm particles) using as elu-ent: (eluent A) 0.2% aqueous trifluoroacetic acid (TFA)and (eluent B) 0.2% TFA in acetonitrile–water 80:20 (v/v).A linear gradient from 0 to 88% of B in 40 min, and from88% to 100% of B in 3 min at a flow rate of 2.8 mL/minwas used.

HPLC–ESI-IT-MS experiments were performed with aVydac (Hesperia, CA, USA) C8 column with 5-lm particle

diameter (column dimensions 15062.1 mm) using thefollowing solutions: (eluent A) 0.056% aqueous TFA and(eluent B) 0.05% TFA in acetonitrile–water 80:20 (v/v). Theproteins were eluted using a linear gradient from 0 to55% in 40 min, at a flow rate of 0.30 mL/min. The massspectra, in the positive ion mode, were collected every3 ms. MS spray voltage was 4.50 kV and the capillary tem-perature was 2208C.

Peptide sequence was determined with a Procise 610AProtein Sequencer (Applera, Foster City, CA, USA)coupled on-line with a 140C microgradient system and aSeries 200 UV/vis detector (Perkin Elmer, Boston, MA,USA).

2.2 Preparation of transglutaminase 2 from humanwhole saliva

Fresh human whole saliva (100 mL) was collected fromfive healthy adult volunteers with plastic pipettes at thebase of the tongue and maintained at 48C to avoid proteo-lysis. Saliva samples were pooled, centrifuged at27 0006g for 40 min at 48C and pellet and supernatantseparated. The supernatant was used to isolate statherin,as reported below, whereas the epithelial-cell-rich sedi-ment was used to prepare oral transglutaminase, accord-ing to a modification of a previously reported method[12]. Briefly, the pellet was washed twice with 2.6 mMsodium phosphate buffer, 6.3 mM sodium chloride,0.6 mM calcium chloride, pH 7.5, and after centrifuga-tion re-suspended in 2 mL of the extraction buffer con-sisting of 50 mM Tris-HCl buffer, pH 7.5, 1 mM dithio-threitol, and 10 mg of Mini-Complete EDTA free proteaseinhibitor (Roche Applied Science, Mannheim, Germany),and incubated for 18 h at 48C. The suspension was centri-fuged at 27 0006g for 15 min at 48C and the supernatantcontaining crude TG2 was stored at –208C in 200-lL ali-quots. On the basis of the formation rate of cyclo-statherin at 378C (see Section 2.4) we estimated thatabout 0.02610–3 EC units of enzyme were present inone mL of preparation.

2.3 Purification of statherin

The supernatant obtained from centrifugation of humanwhole saliva was diluted 1:1 with 0.2% aqueous TFA, cen-trifuged at 80006g for 15 min at 48C, and the pellet dis-charged. Statherin was partially purified by precipitationaccording to a modification of the method of Flora et al.[16]. Briefly, the acidic solution was diluted 1:1 with0.5 mM zinc chloride, its pH value brought up to 9.0, andthe solution stored on ice for 20 min. The suspensionwas centrifuged at 12 5006g for 20 min at 48C and theprecipitate, washed with distilled water, was dissolved in1 M HCl. The solution was dialyzed overnight in 30 mMacetate buffer, EDTA 10 mM, pH 5.7, and then lyophi-lized.



Lyophilized samples were dissolved in 3 mL of 0.2% aque-ous TFA and submitted to preparative RP-HPLC to isolatepure statherin. The progress of separation was followedby UV at 214 and 276 nm. Statherin, eluting at 27 min,was collected, lyophilized, and stored at –208C. The pur-ity of statherin was checked by HPLC–ESI-MS using Ther-moFinnigan instrumentation.

2.4 Preparation of cyclo-statherin

Different quantities of statherin (usually 7 lmol/L) wereincubated in 50 mM Tris-HCl, 10 mM CaCl2, 1 mM DTT,and 1 mM EDTA buffer, pH 7.5, with 0.02 EC units/mL ofguinea pig liver TG2 at 378C. One EC unit correspondedto the formation of 1.0 lmole of hydroxamate per min-ute from Na-CBZ-glutaminylglycine and hydroxylamineat pH 6.0 at 378C. Reaction was stopped by addition of0.1 M EDTA (final concentration 33 mM), usually after30 min of incubation. Cyclo-statherin was purified by RP-HPLC. Experiments with human TG2 were performedunder the same experimental conditions using 400 lL ofcrude preparation (0.02610–3 EC units/mL).

2.5 Kinetic studies of cyclo-statherin formation

Kinetic experiments were performed at 378C in the samebuffer used to prepare cyclo-statherin at four differentstatherin concentrations (0.15, 0.2, 0.3, and 0.4 lM) andusing 0.005 EC units/mL of GPL TG2. At established times(30 s, 1, 2, 3, and 4 min) part of the reaction mixture(70 lL) was admixed with 0.1 M EDTA (final concentra-tion 33 mM) and analyzed by RP-HPLC, as reported above.Statherin and cyclo-statherin concentrations were calcu-lated on the basis of the chromatographic peak arearevealed at 276 nm. Each data point was determined induplicate. Kinetic parameters (Km and kcat) were deter-mined by fitting experimental results with the Michae-lis-Menten equation.

2.6 Enzymatic digestion of statherin and cyclo-statherin with chymotrypsin

About 4 nanomoles of cyclo-statherin (or statherin) wasdissolved in 250 lL of 100 mM Tris-HCl, 10 mM CaCl2,pH 7.8, admixed with 0.7 lL of 0.5 mg/mL a-chymotryp-sin (Sigma–Aldrich) and incubated at 308C with stirringfor 3 h. Reaction was stopped by adding phenylmethyl-sulfonyl fluoride (PMSF) at a final concentration of 5 mM.Digestion product was submitted to RP-HPLC–ESI-MS/MSanalysis to determine the molecular mass value and thesequence of the peptides formed by digestion.

2.7 Enzymatic digestion of statherin and cyclo-statherin with proteinase K

About 5 nanomoles of cyclo-statherin (or statherin) wasdissolved in 180 lL of 50 mM Tris-HCl, 5 mM CaCl2,

pH 7.5, admixed with 20 lL of 0.5 mg/mL proteinase K(Pierce Biotechnology, Rockford, IL, USA) and incubatedat 378C with stirring for 3 h. Reaction was stopped byadding PMSF at a final concentration of 5 mM. The diges-tion product was submitted to RP-HPLC–ESI-MS/MS anal-ysis to determine the molecular mass value and thesequence of the peptides originated by digestion. Tan-dem MS experiments were performed on the ionsdetected with a peak width of 2 m/z values and using 40%of the maximum activation amplitude.

2.8 Enzymatic digestion of statherin and cyclo-statherin with carboxypeptidase Y

About 5 nanomoles of purified cyclo-statherin (orstatherin) was suspended in 280 lL of a 70 mM sodiumacetate buffer pH 5.5, EDTA 10 mM, and digested with0.016103 EC units of carboxypeptidase Y (Pierce Biotech-nology, Rockford, IL, USA) dissolved in 20 lL of the samebuffer at 378C. One EC unit of carboxypeptidase Y hydro-lyzes one lmol of CBZ-phenylalanyl-alanine per minuteat 258C, pH 6.75. At 5, 10, 20, 40, and 60 min, 60 lL ali-quots of the digestion mixture were mixed with an equalvolume of 0.2% aqueous TFA and analyzed by RP-HPLC–ESI-MS analysis according to the procedure previouslydescribed.

Extracted ion-current (XIC) strategies were applied toanalyze the mixture of the proteolytic products search-ing the peptides originated by the different possibledigestion products of cyclo-statherin.

2.9 Peptide sequencing

One nmol of cyclo-statherin was dissolved in 0.1% aque-ous TFA, and the acidic solution was applied to a pre-cycled biobrene-treated glass fiber filter. PTH amino acidswere separated using a 22062.1 mm C18 reversed-phasecolumn. The solutions utilized for the chromatographywere: (eluent A) 3.5% aqueous tetrahydrofuran mixedwith the Applera premix buffer in 25:960 ratio (v/v) and(eluent B) acetonitrile/2-propanol (12:88, v/v).

2.10 Determination of cyclo-statherin in wholehuman saliva

Whole saliva was collected from 12 healthy adult volun-teers (4 males and 8 females, non smokers, age from 27to 36 years), under resting conditions according to a stan-dard procedure between 2 and 4 p.m. [17]. Acidic solu-tion (0.2% aqueous TFA) was immediately added to sali-vary samples in 1:1 v/v ratio at 48C and then centrifugedat 80006g for 5 min. After centrifugation the superna-tant was separated from the precipitate and immediatelyanalyzed by HPLC-ESI-IT-MS. XIC strategy was employedto search different m/z values related to statherin andcyclo-statherin.



2.11 Data analysis

Deconvolution of averaged ESI-MS spectra was automati-cally performed either by using the Bioworks Browsersoftware provided with the Deca XP instrument or byMagTran 1.0 software [18]. The mass values obtained forstatherin, cyclo-statherin, and their derivatives werecompared with average theoretical mass values usingPeptideMass and FindPept programs available at theSwiss-Prot Data Bank (expasy.org/tools), where statherinis coded as P02808. Experimental tandem MS spectrawere compared with theoretical tandem MS spectra gen-erated by utilizing the MS-product program, available atthe Protein Prospector site (prospector.ucsf.edu/).

3 Results

3.1 Reaction of statherin with types 2transglutaminase

Two transglutaminase preparations were used to per-form the reaction in vitro, commercial guinea pig livertransglutaminase 2 and crude oral transglutaminaseprepared from desquamated oral cells separated fromwhole human saliva. The reaction was followed by RP-HPLC–ESI-MS and comparable results were obtained withboth enzyme preparations. Fig. 1a shows the chromato-graphic profile of the reaction mixture (7 lmol/Lstatherin) after 1 min of incubation with GPL TG2(0.02 EC units/mL). The chromatographic profile shows


Figure 1. Detection of cyclo-statherin by RP-HPLC–ESI-MS. RP-HPLC UV (276 nm) and MS (TIC) profiles of human salivarystatherin (7 lmol/L) incubated with GPL transglutaminase 2 (0.02 EC units/mL) for 1 min at 378C (a). ESI-MS spectra of the chro-matographic peaks recorded in the ranges 27.95–28.80 min (b), 29.14–29.92 min (c) and corresponding average mass valuesobtained by deconvolution of the ESI spectra. They were attributed to statherin (5380.1 Da) and to a cyclic derivative of statherin(5363.1 Da).


the presence of a peak eluting at 29.57 min togetherwith the peak of unreacted statherin. Even after pro-longed incubation times and complete transformationof statherin, no other derivative was detected in the chro-matographic profile. The average mass of statherin(Fig. 1b) obtained by deconvolution of the ESI mass spec-trum registered between 27.95 and 28.80 min corre-sponded to 5380.0 € 0.3 Da (mean of five measurements).The identity of statherin was confirmed by automatedEdman partial N-terminal sequencing of the purifiedpeptide (data not reported). The average mass of thederivative obtained by deconvolution of the ESI massspectrum registered between 29.14 and 29.92 min(Fig. 1c) was 5363.0 € 0.3 Da (mean of five measure-ments). The difference between the mass value ofstatherin and that one of the derivative (Dmass = 17.0 Da),corresponding to the mass of an ammonia molecule,indicated that an intra-chain cross-linking occurredbetween the lateral chains of lysine and glutamine resi-

dues belonging to the same statherin molecule. This cyc-lic derivative was named cyclo-statherin.

3.2 Structure determination of cyclo-statherin

The structure of statherin is as follows:

DSSEEKFLRR IGRFGYGYGP YQPVPEQPLY PQPYQPQYQQ YTF

where S stands for phosphorylated serine.

Being the sole lysine residue present in the sequence, Lys-6 must necessarily be involved in cyclo-statherin forma-tion. However, we confirmed the participation of Lys-6 bysubmitting RP-HPLC purified cyclo-statherin to auto-mated-Edman sequencing. In agreement with a molecu-lar modification of the lysine residue, at the sixthsequencing step the PTH-Lys derivative was not observed.Since seven glutamine residues are present in statherinat positions 22, 27, 32, 35, 37, 39, and 40 (see structurereported above), it was a more challenging task to estab-


Figure 2. RP-HPLC–SIM-MS/MS analysis of cyclo-statherin chymotryptic digest. (a) RP-HPLC MS profile (TIC) of the digest,with the identification of three chymotryptic peptides confirmed by MS/MS experiments. (b) SIM profile showing the peak of thebi-charged ion at 779.3 m/z ([M+2H+]2+). (c) MS/MS spectrum of the ion at 779.3 m/z, corresponding to the sequence of the linearfragment 22–34.


lish which glutamine residue/s was involved in cyclo-statherin formation. Cyclo-statherin and statherin weresubmitted to automated Edman degradation. However,due to the large number of tyrosine residues that low-ered the yield of the reactions, the sequence of both pep-tides was determined only up to Pro-20 [1]. Therefore, inorder to determine the Gln residue/s cross-linked it wasnecessary to perform different enzymatic digestions.

Digestion with chymotrypsin resulted in the formationof several peptides that were characterized by MS/MSexperiments (Table 1). The detection of the linear frag-ment 22–34 (Mav. = 1556.6 Da, Fig. 2) allowed to excludethe participation to the cross-linking of glutamines 22,27, and 32, but not that of glutamines 35, 37, 39, and 40.

Incubation of cyclo-statherin with proteinase K produceda mixture of peptides, which were analyzed by RP-HPLC–ESI-MS and MS/MS strategies. The two mass values of4964.8 Da and 2484.0 Da (Fig. 3) were potentially attribu-ted to fragments containing a lysine–glutamine bridge.The peptide with the mass of 4964.8 Da could correspondto the cyclic fragment 1–38, where either Gln-35 or Gln-

37 might be cross-linked. The mass at 2484.0 Da couldcorrespond both to the fragment 1–9/27–36 and 1–9/28–37. Thus, the results of proteinase K digestion allowedexclusion of the involvement of Gln-39 and 40, but didnot allow discrimination between Gln-35 and 37.

Since these two residues are located near to the carboxy-terminus, purified cyclo-statherin was submitted to 1 hdigestion with carboxypeptidase Y and the reaction pro-ducts were analyzed by HPLC–ESI-MS after 5, 10, 20, 40,and 60 min. The experimental design was planned bear-ing in mind that the proteolysis of C-terminal gluta-mines catalyzed by the exo-peptidase can provide twodifferent results: i) a loss of 128.1 Da, if the Gln is notcross-linked; ii) an increase of 18.0 Da, if the Gln is linkedto Lys-6. Therefore, the chromatographic profiles wereanalyzed by XIC strategies. Fig. 4 shows part of the resultsof this XIC HPLC-MS analysis after 20 min of digestion.The inset in the upper-left corner shows the chromato-graphic profile (total ion current, TIC) of the digestionmixture recorded between 26.50 and 31.50 min. The twobig insets show the ESI spectra registered between 28.59and 29.75 min (average of 44 spectra) and 26.88 and27.47 min (average of 23 spectra). The other insets (Cyand from (a) to (l)) show the XIC peaks obtained by search-ing the triple- and tetra-charged ions of all the possiblepeptides resulting from digestion up to Gln-37. The areaof each XIC peak, the average mass value, and the elutiontime are also reported. The Cy inset shows the XIC peakof unreacted cyclo-statherin (Cyclo-s). Crucial for theassignment were both detected and undetected XICpeaks. Detection of the XIC peaks corresponding to4951.2 (c), 4823.1 (e), 4695.0 (g), 4531.8 (h), and4549.8 Da (i) permitted the conclusion that Gln-37 iscross-linked. On the other hand, the absence of the XICpeaks corresponding to 4969.2 (d) and 4841.1 (f) Daallowed the exclusion of any involvement of Gln-40 andGln-39 in the cyclization reaction. Finally, the XIC peakcorresponding to 4403.7 Da (l), expected if the cross-link-ing would involve Gln-35, 32, 27, and 22, was notdetected. On the basis of this last result we could alsoexclude any contribution of Gln-35, 32, 27, and 22 to thecross-linking reaction, in agreement with the results ofchymotryptic digestion. Therefore, we can deduce thatthe cyclization reaction is highly specific for Gln-37.However, during the analysis of the samples submittedto digestion for 5 and 10 min, two ions at 1614.7/1211.3


Table 1. Cyclo-statherin chymotryptic fragments characterized by RP-HPLC–ESI-MS/MS experiments (see also Fig. 3).

Elution time (min) Mass (Da) Assignment Sequence by MS/MS

16.0–17.4 775.9 (F)GYGYGPY(Q) (fr. 15–21) GYGYGPY19.0–20.0 918.3 (F)LRRIGRF(G) (fr. 8–14) LRRIGRF20.0–21.5 1556.6 (Y)QPVPEQPLYPQPY(Q) (fr. 22–34) QPVPEQPLYPQPY

Figure 3. Attribution of some mass values detected by RP-HPLC–ESI-MS analysis of proteinase K digest of cyclo-statherin. The Mav. of 4964.8 Da could correspond to thefragment 1–38 of cyclo-statherin, with Lys-6 linked either toGln-35 or to Gln-37. The Mav. of 2484.0 Da could correspondto the fragment 1–9 of cyclo-statherin, linked either to frag-ment 27–36 or to fragment 28–37. Even though the MS/MSexperiments did not allow solving the ambiguity, theseresults excluded the involvement of Gln-39 and 40.


m/z were detected at very low intensity, indicating that adigestion product with a mass of 4841.1 Da was present,albeit in almost negligible amount. Thus, a small percen-tage of cyclo-statherin could be also cross-linked at thelevel of Gln-39. On the basis of the ion intensity, cyclo-statherin Q39 cannot represent more than 3–5% of thetotal cyclo-statherin.

3.3 Kinetics of GPL TG2 cross-linking of statherin

The results of kinetic experiments, performed using fourdifferent concentrations of statherin (0.15, 0.2, 0.3, and0.4 lM) and 0.005 EC units/mL of GPL TG2 and analyzedby using the Michaelis-Menten equation, provided a Km

value of 0.65 € 0.06 lM and a kcat value of 33 € 5 min–1.

3.4 Detection of cyclo-statherin in whole humansaliva

By adopting an XIC strategy and searching the ions at1788.6 and 1341.7 m/z values, we were able to detect cyclo-statherin in all the samples of adult whole saliva analyzed

(N = 12; four males, eight females, non-smokers, agedfrom 27 to 36 years, resting conditions). The XIC peak wasattributed to cyclo-statherin on the basis of chromato-graphic elution time, on the basis of the mass value(5363.0 € 0.5 Da) obtained after deconvolution, and onthe basis of the ESI spectrum of cyclo-statherin that showsonly the triple- and tetra-charged ions in a specific andcharacteristic ratio (see Fig. 1). Fig. 5 shows the XIC resultsobtained on one of the samples. From an approximate eva-luation based on the area of the ionic current peak ofstatherin and cyclo-statherin (Fig. 5), the mean concentra-tion of cyclo-statherin in whole saliva corresponds toabout 1% of that of statherin in healthy subjects. Becausein healthy subjects the mean statherin concentrationunder resting conditions is ca. 5 lmol/L [19], the salivarycyclo-statherin concentration should be ca. 50 nano-mol/L. Further studies will be necessary in order to estab-lish the real biological significance of this cyclizationreaction. In this context it would be of interest to investi-gate if the cyclo-statherin concentration changes underdifferent physiological and/or pathological conditions.


Figure 4. RP-HPLC–ESI-MS analysis of carboxypeptidase Y digest of cyclo-statherin. The inset in the upper-left corner showsthe chromatographic profile (TIC) of the digestion mixture. The two big insets show the ESI spectra registered between 28.59–29.75 min (average of 44 spectra) and 26.88–27.47 min (average of 23 spectra). The other insets (Cy and from a to l) show theXIC peaks obtained by searching the triple- and tetra-charged ions of any possible peptide originated by the digestion. The areaof each XIC peak, the average mass value and the elution time are also reported. Cy inset shows the peak of unreacted cyclo-statherin (Cyclo-s). Interpretation of the data was performed considering that digestion of C-terminal Gln residues could providethe following results: 1) Gln residue is not cross-linked to lysine. Digestion will yield a new peptide showing a mass decrease of128.1 Da. 2) Gln residue is cross-linked to lysine. Digestion will yield a new peptide showing a mass increase of 18.0 Da, sincethe proteolysis leads to the entry of a water molecule but not to the loss of the cross-linked Gln. The experimental results demon-strated that Gln-37 is the almost specific glutamine residue involved in the formation of cyclo-statherin.


4 Discussion

This study clearly shows the great potential of differentbiochemical strategies based on the analytical strengthof HPLC–ESI-MS for the detection of a cyclic peptide.Indeed, HPLC–ESI-MS analysis allowed us to establish thekinetics of cyclo-statherin formation and HPLC–ESI-MScoupled with different proteinase digestions and follow-ing different experimental plans permitted determina-tion of the specific structure of the cyclic derivative.

Statherin may act in vitro as a complete substrate for TG2,since Gln and Lys residues of the same peptide arerecruited for the almost specific formation of cyclo-statherin Q-37. It is important to underline that poly-meric forms of statherin produced by inter-chain cross-linking were not detected even after prolonged incuba-tion time (more than 30 min) or changing statherin con-centration in the enzymatic reaction. The rate of thecyclization reaction in vitro is compatible with the reac-tion in vivo. Indeed, the cyclic derivative has beendetected in whole human saliva (12 samples), suggestinga possible physiological role for cyclo-statherin. How-ever, the occurrence in vivo of heterotypic cyclo-statherinand/or statherin cross-linking with other salivary pep-tides (mainly acidic and basic proline-rich peptides) orepithelial proteins or small salivary diamines cannot beexcluded, due to the presence of many glutamine resi-dues in these salivary peptides.

TG-induced cyclization of glutamine-rich proteins hasbeen already observed in the formation of protein net-

works that cover and protect different epithelia. Forinstance, intra-chain cross-linked (cyclic) derivatives ofloricrin, SPR3, and SPR1, the major epidermal cornifiedcell envelope (CE) proteins, have been obtained in reac-tions catalyzed by TG1, TG2, and TG3 [20–22] and it waspostulated that the cyclic derivatives play an essentialrole in folding the envelope proteins into more compactstructures [23]. It is interesting to outline that TG2 showsan affinity toward statherin ca. twenty times higher thanthat reported for loricrin, SPR3, and SPR1 [20, 22].

The detailed MS structural characterization of cyclo-statherin carried out in the present study showed thatthe Gln-37 residue is a specific lone-pair acceptor of theLys-6 residue. Although the consensus sequences recog-nized by TG2 are not known [24], it has been reportedthat the enzyme is much less selective towards the donoramine lysine residue than towards the acceptor gluta-mine residues. Whereas the recognition of specific lysineresidues seems to be governed only by their steric hin-drance, the spacing and structure of neighboring resi-dues seems to be a crucial factor for the TG2 specificitytowards the targeted glutamine residues. In particular,proline residues seem to be relevant for glutamine recog-nition. A glutamine residue is not recognized as a sub-strate if it occurs between two proline residues [25], andthis seems to be the case for the Gln-32 residue ofstatherin. Moreover, whereas the enzyme is able to recog-nize QxP residues, a +1 or +3 flanking proline residueseems to completely abolish TG2 recognition [26]. Thenon-involvement of glutamine residues at position 22,


Figure 5. RP-HPLC–ESI-MS detection of cyclo-statherin in human whole saliva. The upper panel shows the TIC profile regis-tered between 25.78 and 32.70 min. The middle panel shows the extracted ion current peak of statherin, obtained by searchingthe triple- and tetra-charged ions at 1346.3 and 1794.4 m/z. The bottom panel shows the extracted ion current peak of cyclo-statherin, obtained by searching the triple- and tetra-charged ions at 1341.7 and 1788.7 m/z. The area of the peaks and the elu-tion times are also reported. The peak eluting at 28.74 min corresponds to the salivary proline-rich P-B peptide.


27, 32, and 35 in the formation of cyclo-statherins is inagreement with this rule. On the contrary, the almostspecific involvement of Gln-37 could be related to the –1flanking proline residue, as also observed for glutaminerecognition on some model peptides [27]. Two adjacentglutamine residues may act as amine acceptors in a con-secutive reaction [24, 27] and Gln-39 of statherin seemsto be indeed recognized as a minor reaction site, whilethe Gln-40 residue is not involved at all.

Statherin cyclization affects the peptide charge near tothe N-terminus, since the positively charged aminogroup of lysine after the reaction is transformed into anamide bond and the positive charge of the lateral chainis lost. The very negative N-terminal segment of statherinis the domain responsible for the absorption onto hydro-xyapatite [2]. Thus, it could be hypothesized that cycliza-tion may provide an increase of adsorption on HP. Thisproperty of cyclo-statherin will be the aim of future stud-ies.

We acknowledge the financial support of Cagliari University,Catholic University of Rome, MIUR, Italian CNR, and Regione Sar-degna, according to their programs of scientific research promo-tion and diffusion.

5 References[1] Schlesinger, D. H., Hay, D. I., J. Biol. Chem. 1977, 252,

1689–1685.

[2] Raj, P. A., Johnson, M., Levine, M. J., Nancollas, G. H., J.Biol. Chem. 1992, 267, 5968 –5976.

[3] Gururaja, T. L., Levine, M. J., Pept. Res. 1996, 6, 283 –289.

[4] Hay, D. I., Smith, J., Schluckebier, S. K., Moreno, E. C., J.Dent. Res. 1984, 63, 857 –863.

[5] Hay, D. I., Arch. Oral Biol. 1973, 18, 1531 –1541.

[6] Li, J., Helmerhorst, E. J., Corley, R. B., Luus, L. E., Troxler,R. F., Oppenheim, F. G., Oral Microbiol. Immunol. 2003, 18,183 –191.

[7] Li, J., Helmerhorst, E. J., Yao, Y., Nunn, M. E., Troxler, R.F., Oppenheim, F. G., Arch. Oral. Biol. 2004, 49, 379 –385.

[8] Douglas, W. H., Reeh, E. S., Ramasubbu, N., Raj, P. A., etal., Biochem. Biophys. Res. Commun. 1991, 180, 91 –97.

[9] Gibbons, R. J., Hay, D. I., Infect. Immun. 1988, 56, 439 –445.

[10] Bradway, S. D., Bergey, E. J., Jones, P. C., Levine, M. J., Bio-chem. J. 1989, 261, 887 –896.

[11] Bradway, S. D., Bergey, E. J., Scannapieco, F. A., Rama-subbu, N., et al., Biochem. J. 1992, 284, 557 –564.

[12] Yao, Y., Lamkin, M. S., Oppenheim, F. G., J. Dent. Res.1999, 78, 1696–1703.

[13] Yao, Y., Lamkin, M. S., Oppenheim, F. G., J. Dent. Res.2000, 79, 930 –938.

[14] Hannig, C., Hannig, M., Attin, T., Eur. J. Oral Sci. 2005,113, 2–13.

[15] Lorand, L., Graham, R. M., Nat. Rev. Mol. Cell Biol. 2003, 4,140 –156.

[16] Flora, B., Gusman, H., Helmerhorst, E. J., Troxler, R. F.,Oppenheim, F. G., Protein Expr. Purif. 2001, 23, 198 –206.

[17] Castagnola, M., Congiu, D., Denotti, G., Di Nunzio, A.,Fadda, M. B., Melis, S., Messana, I., Misiti, F., Murtas, R.,Olianas, A., Piras, V., Pittau, A., Puddu, G., J. Chromatogr.B Biomed. Sci. Appl. 2001, 751, 153 –160.

[18] Zhang, Z., Marshall, A. G., J. Am. Soc. Mass Spectrom. 1998,9, 225 –233.

[19] Contucci, A. M., Inzitari, R., Agostino, S., Vitali, A., et al.,Oral. Dis. 2005, 11, 95 –99.

[20] Candi, E., Melino, G., Mei, G., Tarcsa, E., et al., J. Biol.Chem. 1995, 270, 26382 –26390.

[21] Candi, E., Tarcsa, E., Idler, W. W., Kartasova, T., et al., J.Biol. Chem. 1999, 274, 7226 –7237.

[22] Candi, E., Oddi, S., Terrinoni, A., Paradisi, A., et al., J. Biol.Chem. 2001, 276, 35014 –35023.

[23] Steinert, P. M., Marekov, L. N., J. Biol. Chem. 1995, 270,17702 –17711.

[24] Esposito, C., Caputo, I., FEBS J. 2004, 272, 615 –631.

[25] Pastor, M. T., Diez, A., P�rez-Pay�, E., Abad, C., FEBS Lett.1999, 451, 231 –234.

[26] Piper, J. L., Gray, G. M., Khosla, C., Biochemistry 2002, 41,386 –393.

[27] Parameswaran, K. N., Velasco, P. T., Wilson, J., Lorand,L., Proc. Natl. Acad. Sci. USA 1990, 87, 8472 –8475.


Documents

HPLC–MS characterization of cyclo-statherin Q-37, a specific cyclization product of human salivary statherin generated by transglutaminase 2