Navigation – Plan du site

AccueilNuméros51II. Origine des matières première...Proteomic analysis of Renaissance...

II. Origine des matières premières et préparation du stuc

Proteomic analysis of Renaissance stucco: development of the analytical protocol and first results

Analyse protéomique de stucs de la Renaissance : élaboration du protocole analytique et premiers résultats
Amra Aksamija, Stéphanie Devassine, Fabrice Bray, Marc Bormand, Anne Bouquillon, Christian Rolando et Caroline Tokarski
p. 75-83

Résumés

Cet article propose une stratégie fondée sur la protéomique pour l’identification des protéines et de leur origine biologique dans les œuvres en stuc de la Renaissance italienne, dans le cadre du programme de recherche ESPRIT (Étude des stucs polychromés de la Renaissance italienne). Les trois échantillons étudiés proviennent de La Vierge et l’Enfant, type Santa Maria Nuova, d’après Antonio Rossellino (musée du Louvre, Campana 19), une Vierge et l’Enfant d’après Benedetto da Maiano (musée des Beaux-Arts de Strasbourg, MBA 507) et une Vierge et l’Enfant de l’entourage de Luca della Robbia (musée des Beaux-Arts de Strasbourg, MBA 247). À partir d’éprouvettes de plâtre contenant différents liants protéiniques, plusieurs protocoles analytiques permettant l’extraction de très petites quantités de protéines ont été testés. Une fois optimisé, le protocole a pu être appliqué avec succès à des échantillons historiques en débouchant sur l’identification des collagènes et de leur origine biologique.

Haut de page

Texte intégral

Introduction

  • 1 Bankart, 2002; Beard, 1983; Berner, Weber, 1999; Arcolao et al., 1998; Natali, Lorenzini, 2001; Vas (...)
  • 2 Aksamija et al., 2019.

1Stucco has been used since ancient times in architectural decoration, in construction and sculpture. De Architectura (Vitruvius, 30-15 BC) and Naturalis Historia (Pliny the Elder, 77-79 AD) are amongst the first treatises discussing the stucco technique on building. Stucco’s composition predominantly consists of mineral phases (e.g. lime, gypsum) with fillers (e.g. sand, marble dust) and various organic/inorganic additives that were added to modify the physico-chemical properties of final stucco mixture (e.g. to retard the setting to process, to obtain a more suitable mixture for casting / modeling). Its composition, not accurately defined, has evolved through years depending its function and artists’ own recipes1. The stucco technique and composition evolved highly during the Renaissance2. In the context of the project “Esprit – Study of Polychromed Stuccos of the Italian Renaissance” aiming to study serial production techniques from 15th century Florentine sculptors, the study of organic material was investigated.

  • 3 Bankart, 2002; Beard, 1983; Berner, Weber, 1999; Arcolao et al., 1998.
  • 4 Berner, Weber, 1999.
  • 5 Beard, 1983; Arcolao et al., 1998; Natali, Lorenzini, 2001.
  • 6 Beard, 1983; Berner, Weber, 1999; Arcolao et al., 1998.
  • 7 Bankart, 2002; Berner, Weber, 1999.
  • 8 Bankart, 2002; Berner, Weber, 1999; Natali, Lorenzini, 2001.
  • 9 Berner, Weber, 1999; Natali, Lorenzini, 2001.
  • 10 Montana, Ronca, 2002.
  • 11 Caroselli et al., 2020.
  • 12 Montana, Ronca, 2002.
  • 13 Ronca, 1994.
  • 14 Aksamija et al., 2019.
  • 15 Kahn, 1995.
  • 16 Dallongeville et al., 2016; Vinciguerra et al., 2019; Tokarski et al., 2002 and 2003.
  • 17 Tokarski et al., 2006; Leo et al., 2009.
  • 18 Solazzo et al., 2008; Hendy et al., 2018.
  • 19 Krizkova et al., 2014.

2Historical documents are referencing the use of many organic compounds in recipes for stucco preparation among which are found glues3, gum Arabic4, milk and curd milk5, mucilaginous plants and grains6, blood7, egg white8, and various oils (olive, walnut, linseed) and fats (pork9). The study of organic additives in stucco is challenging due to the very low amount of constitutive organic material. Various techniques are proposed to detect proteins, carbohydrates and/or lipids among which colorimetric analysis10, and FT-IR11. Other methods targeting the constitutive moieties of the organic compounds (e.g. amino acids, fatty acids, oligosaccharides) analysis were proposed using thin layer chromatography12, LC13 and GC-MS14. We propose here to adapt a proteomic methodology to identify proteins present in the material. Proteomic designates the study of proteins, including their identification, quantification, and the study of their modifications (post-translational and chemical). It is mainly based on the use of complementary techniques: chromatography, mass spectrometry and bioinformatics15. The mainstream technique called “bottom up proteomics” consists of analyzing the peptide mixture resulting from protein hydrolysis (commonly with an endoprotease like trypsin) either by peptide mass fingerprint (measurements of peptides molecular weights) or peptide sequencing (accurate identification of amino acid chains following peptide fragmentation in the mass spectrometer). Peptidic fingerprints/peptide sequences are then assigned to a protein by means of bioinformatics softwares. The other approach called “top down proteomics” investigates protein analysis without preliminary hydrolysis step. An alternative to top down and bottom up approaches is called middle down that is based on a protein hydrolysis into a few high molecular weight peptides using adapted enzymes or chemicals (such as cyanogen bromide that hydrolyzes proteins at the C-terminus of methionine). Proteomics was introduced in the cultural heritage field in the early 200016, showing its high ability to identify proteins in various types of historic or ancient materials such as artworks17, mineral-based objects or material such as ceramics18, or historical mortars19. Considering the very low amount of organic material in stucco, one of the main difficulties is related to the sample preparation including the protein extraction.

3This paper describes the proteomics methodology proposed to identify proteins from stucco. First investigation was focused on the development of the best protocol to the study of very low amount of proteins trapped in gypsum-based matrix. To this end, we used mock-up impregnated with model proteins; various proteins were studied including casein and milk, lysozyme, rabbit collagen and bovine gelatin. The method was then successfully applied to the study of Italian Renaissance stucco devotion reliefs identifying the presence of collagen-based material.

Material and methods

Gypsum based mock-up

  • 20 Tokarski et al., 2006.
  • 21 Aksamija et al., 2019.

4Series of reference stucco mock-up samples containing gypsum and one standard protein (α-casein, lysozyme, collagen from rabbit skin) or a protein-based material (milk and bovine gelatin) were prepared. Chemically pure plaster CaSO4·½ H2O (Sigma Aldrich, France) was tempered with aqueous homogeneous solution of standard protein or protein material (~ 1 g of solution with ~ 1.5 g of plaster) and obtained stucco mixture was vigorously mixed, then put in Eppendorf tubes with a spatula and left open to dry during one week, on ambient temperature. After setting, the stucco samples were unmolded and conserved in their Eppendorf tubes before utilization (fig. 1 a-b). One series of blank stucco samples was prepared with water, free of any organic additive. The prepared stucco mock-up before drying was weighted between 2.0 and 2.5 g, including 0.5 % w/w of protein or protein material. Before protein extraction, reference stucco mock-up samples were crushed into homogenous powder before extraction using a pestle with a single-use synthetic resin as described in20. This powder was kept in original Eppendorf during this study and weighted. One stucco sample of each series was kept as control. After drying, a thin “shiny” layer was observed in mock-ups that included proteins as observed in previous studies suggesting a possible protein migration inside the mock-up21.

Fig. 1 a-b. Stucco mock-up made by chemically pure plaster and protein additives; a. In Eppendorf tubes; b. Unmolded after setting. Observed difference of color between mock-ups is not relevant to chemical composition

Fig. 1 a-b. Stucco mock-up made by chemically pure plaster and protein additives; a. In Eppendorf tubes; b. Unmolded after setting. Observed difference of color between mock-ups is not relevant to chemical composition

© A. Aksamija.

Historic samples

  • 22 Aksamija et al., 2019.

5The historic samples were obtained from the Centre de recherche et de restauration des musées de France (C2RMF) and selected in regard to previous GC-MS results22 that allowed positive detection of proteins. The studied Renaissance stucco artworks are presented here (table 1). Samples were taken from the back of the artworks either in powder form (Campana 19 and MBA 507) by micro-drilling with a tungsten carbide mini-drill or as a fragment from the surface (a few mm2, depth 1 to 2 mm; MBA 247) obtained with a micro-burin. The collected samples for organic analysis were kept in aluminum foil to minimize contamination.

Table 1. The samples from historic stucco artworks

Table 1. The samples from historic stucco artworks

Protein extraction

6Several extraction solvents and mixtures were tested to extract proteins from the studied samples. The protocols are detailed in the following sections.

  1. Acid extraction based on H2O, 0.1 % trifluoroacetic acid (TFA): 1.5 mg sample was mixed with 200 µL H2O, 0.1 % TFA and shacked during 24 hours. 10 min centrifugation at 13.4 rpm was performed and the solution was placed into ultra-centrifugal filters with 3 kDa cutoff and centrifuged again at the same speed. The remaining solution was transferred into an Eppendorf tube. Barium acetate (0.02 M) was added to desalt the sample, two times. White precipitate corresponding to BaSO4 was observed. The sample was centrifuged at 13,400 rpm during 10 min and the supernatant was separated and evaporated using speed vacuum. 10 µL of H2O, 0.1 % AF was added and the sample was desalted to discard the remaining acetate. Several desalting procedures were tested including C4 UptiTiTM (Interchim) and micro-spin column with EmporeTM Styrene Divinyl Benzene (SDB-XC, 47 mm Extraction Disks 2240, Supelco, Solid phase Extraction Disk). Best results were obtained with SDB-XC using the following procedure: the protein extract was loaded ten times on desalting disks and elution was performed first with 50 µL of a solution containing 50/50 v/v, ACN/H2O 0.1 % AF; the procedure was repeated two times. Then 50 µL of pure ACN was used; the procedure was repeated two times. The solution containing extracted proteins was placed in vacuum evaporator to eliminate ACN. The extracted and purified protein was recovered in 10 µL of MS-grade water H2O acidified with 0.1 % formic acid (FA).

  2. Basic extraction based on NH4OH, 2.5 M following a similar procedure to the acidic extraction based on H2O, 0.1 % TFA, except the starting extraction that was performed with 200 µL NH4OH, 2.5 M.

  3. Basic extraction with trifluoroethanol (TFE) following a similar procedure to the acidic extraction based on H2O, 0.1 % TFA, except the starting extraction that was performed with NH4OH 2.5 M, 50 % TFE.

  4. Acid extraction with anionic detergent and urea following a similar procedure to the acidic extraction based on H2O, 0.1 % TFA, except the starting extraction that was performed with 4 % sodium dodecyl sulfate (SDS), 6 M urea in H2O, 0.1 % TFA.

  5. Basic extraction with cationic detergent following a similar procedure to the acidic extraction based on H2O, 0.1 % TFA, except the starting extraction that was performed with 4 % dodecyltrimethylammonium chloride (DTAC) or 4 % cetyltrimethylammonium chloride (CTAC), 6 M urea in H2O.

  6. Acid extraction with urea: 1.5 mg sample was mixed with 200 µL H2O 0.1 % TFA, 6 M urea and shacked overnight. 10 min centrifugation at 13.4 rpm was performed and the solution was placed into ultra-centrifugal filters with 3 kDa cutoff and centrifuged again at the same speed. Barium acetate desalting was performed as described earlier (point 1). A cleaning procedure based on Amicon was added with the following solutions: 6 M urea (2 x 200 µL), 3 M urea (2 x 200 µL), 1.5 M urea (2 x 200 µL) and H2O 0.1 % AF with 10 min centrifuge at 13,400 rpm for each step. Solid extraction was then performed with SDB-XC resin as described for acidic extraction based on H2O, 0.1 % TFA. The solution containing extracted proteins was placed in vacuum evaporator to eliminate ACN. The extracted and purified protein was recovered in 10 µL of MS-grade water H2O acidified with 0.1 % formic acid (FA).

Protein hydrolysis

7CNBr hydrolysis. The digestion protocol was optimized using mock-up prepared with 0.5 % w/w of collagen from rabbit skin. The samples were subjected to extraction procedure as described in the protein extraction section up to desalting step. The final extract containing extracted protein was recovered as a remaining volume in a ultra-centrifugal filter (about 30 µL) and transferred into new tube. 50 µL of denaturation buffer composed by 50 mM Tris HCl, 4 mM dithioethanol (DTT), 60 % TFE was added to the sample and the sample was heated at 70 °C during 2 hours. Then 50 mM of iodoacetamide was added for alkylation and the sample was kept in dark during 1 hour. The alkylated mixture was further concentrated to dry residue in vacuum evaporator, then subjected to chemical digestion by adding 100 µL of freshly prepared CNBr (50 mg/mL in acidified MS-grade water, 10 % AF). The digestion mixture was kept in dark at 45 °C during 24 hours. The resulting peptide mixture was concentrated to dryness using vacuum evaporator and desalted with EmporeTM Styrene Divinyl Benzene (SDB-XC) as described in the previous section. It was concentrated to dryness and redissolved in 10 µL of MS-grade water H2O acidified with 0.1 % AF.

Trypsin digestion based on an enhanced filter aided procedure (eFASP)

8Prior to their use, 10 kDa Amicons® filters (EMD Millipore, Darmstadt, Germany) were washed in 5 % Tween-20 overnight and then rinsed three times with 200 μL water during 20 min each. 25 µL of lysis buffer composed by 4 % sodium dodecyl sulfate (SDS), 0.2 % deoxycholic acid (DCA), 50 mM DTT, 200 mM ammonium bicarbonate was added to each protein extract and let overnight at 4 °C. 200 µL of exchange buffer (8 M urea, 0.2 % DCA, 100 mM ammonium bicarbonate, pH 8.8 (ABC) was added. The samples were placed in 10 kDa Amicons® filters and centrifuged at 13,000 g at 20 °C for 15 min. The samples were washed with exchange buffer (3 × 200 µL), centrifuged at 13,000 g at 20 °C for 15 min after each addition of buffer and the filtrate was discarded. The reduced proteins were alkylated (8 M urea, 50 mM iodoacetamide, and 100 mM ABC, pH 8.8) within the membrane filter during 1 hour at 37 °C in the dark by adding of 100 µL of alkylation buffer. The samples were washed with exchange buffer (3 × 200 µL), then with digestion buffer (50 mM ABC, 0.2 % DCA pH 8.8) (3 × 200 µL), followed with centrifugation (13,000 g at 20 °C for 15 min). After the last washing, the holding tube of membrane filters were changed, then a volume 100 µL of digestion buffer was added. 40 μl of trypsin/LysC (Promega, Madison, USA) was added and the samples were let overnight at 37 °C in the dark. The next morning, the digested samples were centrifuged (13,000 g at 20 °C for 15 min) then followed with washing/centrifugation (13,000 g at 20 °C for 15 min) step using peptide recovery buffer (50 mM ABC pH 8.8; 2 × 50 µL) to collect the peptide mixture filtrate in the holding tube. The peptide mixture filtrates were transferred to a new Eppendorf tube, then 200 µL of ethyl acetate and 2.5 µL of TFA were added, all together vigorously shaken and centrifuged (13,000 g at 20 °C for 10 min). The obtained peptide mixture in the aqueous phase was washed with ethyl acetate (3 × 800 µL) followed with centrifugation (13,000 g at 20 °C for 10 min). The open tubes containing aqueous phase were placed at 60 °C, in a fume hood during 5 min, to eliminate the residual ethyl acetate. The peptide mixtures were then concentrated in vacuum evaporator to dryness, then a volume of 100 µL of 50/50 (H2O/MeOH) was added and the evaporation procedure were repeated once. The dry residues were dissolved in 10 µL in 5 % acetonitrile, 0.1 % formic acid solution.

Mass spectrometry analysis

9The protein spectra were obtained using MALDI-TOF mass spectrometer (MALDI-TOF/TOF 4800+ mass spectrometer, Applied Biosystems, Sciex). All spectra were acquired in positive linear mode in the mass range 1 kDa-50 kDa. The matrix used was α-cyano-4-hydroxycinnamic acid (CHCA) matrix at 10 mg/ml in 50/50 ACN/H2O, the sample/matrix ratio was 1/1. The mixture was spotted on a 384-well AB Sciex MALDI target. A total of 1500 laser shots were accumulated for each protein spectrum.

10LC-MS/MS analyses were performed on an Orbitrap Q ExactivePlus mass spectrometer hyphenated to a U3000 RSLC Microfluidic HPLC System (ThermoFisher Scientific, Waltham, Massachusetts, USA). 1 μl of the peptide mixture at a concentration of 1 µg/µL was injected with solvent A (5 % acetonitrile and 0.1 % formic acid v/v) for 3 min at a flow rate of 10 μl.min-1 on an Acclaim PepMap100 C18 pre-column (5 μm, 300 μm ID × 5 mm) from ThermoFisher Scientific. The peptides were then separated on a C18 Acclaim PepMap100 C18 reversed phase column (3 μm, 75 mm i.d. × 500 mm), using a linear gradient (5-40 %) of solution B (75 % acetonitrile and 0.1 % formic acid) at a rate of 250 nL.min-1 in 160 min and then 100 % of solution B in 5 min. The column was washed for 5 min with buffer B and then re-equilibrated with buffer A. The column and the pre-column were placed in an oven at a temperature of 45 °C. The total duration of the analysis was 180 min. The proteins were separated on a C4 pre-column and column (ThermoFisher Scientific). The LC runs were acquired in positive ion mode with MS scans from m/z 350 to 1,500 in the Orbitrap mass analyser at resolution 70,000 at m/z 400. The automatic gain control was set at 1e106. Sequentially MS/MS scans were acquired in the high-energy collision dissociation cell for the 15 most-intense ions detected in the full MS survey scan. Automatic gain control was set at 5e105, and the normalized collision energy was set to 28 eV. Dynamic exclusion was set at 90 s and ions with 1 and more than 8 charges were excluded. The raw files were analyzed with Proteome Discoverer software using UniProt database. The precursor mass and fragment mass were identified with an initial mass tolerance of 10 ppm. The search included variable modifications of methionine and proline oxidation, asparagine and glutamine deamidation, tyrosine, serine and threonine phosphorylation and N-terminal acetylation and glutamine to pyroglutamate conversion, and fixed modifications of carbamidomethyl cysteine.

Results

11The aim of this work was to identify the protein-based material and its biological origin present in historical reliefs in stucco. One of main difficulties was related to the very low concentration of proteins in the sample that required an optimization of the analytical workflow. The proposed line was based on the preliminary preparation of mock-up impregnated with various protein-based materials to improve the protein extraction from the inorganic matrix and the full analytical workflow. The first part of this paragraph deals with the analysis of reference material; this part includes the evaluation of various protein extraction condition using complementary analytical workflow (bottom up/middle down proteomics). On the basis of the optimized method, the second part presents the results obtained from the analysis of three Renaissance stuccoes from the Louvre Museum in Paris and the Museum of Fine Arts in Strasbourg, revealing the identification of collagen proteins and their subfamily classification.

Analysis of stucco mock-up

12The optimization of the analytical workflow was investigated on stucco mock up made with chemically pure CaSO4·½ H2O plaster and containing various protein-based material. More precisely, five standards were used: standard proteins (lysozyme, caseins, collagens), whole milk and glue. One of the most challenging steps was to formulate the most adapted extraction solution. Six extraction solutions were evaluated among which acidic extractions from simplest composition (trifluoroacetic acid only) to more complex composition including a denaturant (urea) or denaturant and anionic detergent (respectively urea and sodium dodecyl sulfate). Basic extraction conditions were also evaluated using a base agent (ammonium hydroxide) but also using solution with denaturant (trifluoroethanol) and cationic detergents (dodecyltrimethylammonium chloride and cetyltrimethylammonium chloride). The efficiency of the extraction was firstly roughly evaluated using intact protein analysis by MALDI-TOF mass spectrometry. As shown in the figure 2 (fig. 2 a), monocharged ions of α-casein, ß-lactoglobulin and α-lactalbumin from model stucco containing milk were successfully detected respectively at 23,939 Da, 18,353 Da and 14,170 Da using the acidic extraction containing urea. An expected decrease of the signal to noise ratio was observed compared to milk model sample (fig. 2 d); different relative intensities of the main patterns were also observed. Weaker signal to noise ratios of the 3 targeted proteins were observed for the other extraction conditions; e.g. the basic conditions such as DTAC/urea extraction (fig. 2 b) and ammonium hydroxide extraction (fig. 2 c). The efficiency of the acidic extraction with urea was confirmed by LC-MS analysis (data not shown).

Fig. 2 a-d. Normalized MALDI-TOF spectra of extracted proteins from mock-up using: a. TFA and urea; b. DTAC and urea; c. Ammonium hydroxide. The spectrum d. shows the proteins extracted from milk without impregnation in mock up (10 pmol)

Fig. 2 a-d. Normalized MALDI-TOF spectra of extracted proteins from mock-up using: a. TFA and urea; b. DTAC and urea; c. Ammonium hydroxide. The spectrum d. shows the proteins extracted from milk without impregnation in mock up (10 pmol)
  • 23 Chowdhury et al., 1990.

13An adapted desalting procedure had to be investigated due to the numerous adduct ions resulting from the attachment of sulfuric acid molecules (from plaster) to the peptide or protein ions as observed by a mass increment of 98 uma. The figure 3 a shows the 98 uma mass increment considering the lysozyme pattern (9+ ions, m/z 1,590.2) resulting from the extraction from the stucco models (fig. 3 a). Barium acetate was used to desalt the sample in order to produce BaSO4 (that was observed by a white precipitate during the desalting procedure23). The figure 3 b shows the resulting spectrum following the successful desalting procedure (fig. 3 b).

Fig. 3 a-b. NanoLC nanoESI-MS spectrum showing intact lysozyme resulting from the extraction of stucco mock-up; a. Without barium acetate; b. With barium acetate desalting

Fig. 3 a-b. NanoLC nanoESI-MS spectrum showing intact lysozyme resulting from the extraction of stucco mock-up; a. Without barium acetate; b. With barium acetate desalting

14Best results from collagen and glue-based model stucco were also obtained from acidic extraction with urea (0.1 % TFA, 6 M urea) and a barium acetate desalting procedure using a middle down analysis (data not shown). Consequently, the same sample preparation procedure was applied to historic samples.

Analysis of historic stucco

15The developed methodology was applied to three historic stuccoes (table 1): the Virgin and Child, type Santa Maria Nuova (after Antonio Rossellino); the Virgin and Child (after Benedetto da Maiano) and the Virgin and Child (Circle of Luca della Robbia), respectively noted Campana 19, MBA 507 and MBA 247. The whole composition of the artworks is pre-sented by G. Gariani, F. Beaugnon et al. in this volume. The analytical workflow was based on a bottom up approach using trypsin as detailed in the experimental section. The results pointed out a successful identification of collagens in the three studied samples (table 2). Collagen alpha-1(I), alpha-2(I) and alpha-1(III) chains were identified using respectively 14 %, 7 % and 2 % sequence coverages in the Campana 19 sample (sequence coverage is the ratio of experimentally identified amino acids over total number of amino acids from theoretical sequence, result expressed in percentage). For example, the figure 4 is showing a MS/MS spectrum identifying the peptide GSAGPPGATGFPGAAGR of the collagen alpha-1(I) chain in the Campana 19 sample (underlined proline P means that proline is hydroxylated). The spectrum is pointing out the good signal to noise ratio allowing the identification of b and y fragment ions characteristic of the identified sequence. Collagen alpha-1(I) and alpha-2(I) chains were also identified in MBA 507 and 247 samples with sequence coverages respectively of 9 % and 12 % for collagen alpha-1(I) and 16% and 10% for collagen alpha-2(I) chain. On the basis of 72 collagens peptides identified in the 3 samples, the sequence alignments of the identified peptides have resulted in the identification of two peptides belonging to the Bos genus/Bubalus genus (Bos Taurus, Bos mutus, Bos indicus, Bubalus bubalis) in the MBA 507 sample (2e05 E-values for Bos genus and 2e04 E-value for Bubalus genus); e.g. the figure 5 is showing the peptide SGETGASGPPGFVGEK of the collagen alpha-2(I) chain (fig. 5).

Table 2. Identified proteins in the historic samples

Proteins names, species

Accession numbers

Molecular weights [kDa]

Sequence coverage (%), identified peptides, unique peptides

Campana 19

MBA 507

MBA 247

Collagen alpha-1(I) chain, Bos taurus

P02453|CO1A1

138

14 %, 15, 1

9 %, 10, 4

16 %, 16, 1

Collagen alpha-2(I) chain, Bos taurus

P02465|CO1A2

129

7 %, 7, 3

12 %, 12, 5

10 %, 10, 2

Collagen alpha-1(III) chain, Bos taurus

Q08E14|Q08E14

138

2 %, 2, 0

-

-

Names of identified proteins, accession numbers, molecular weights as well as sequence coverages (%) and number of identified and unique peptides are described (unique peptide defines a peptide that exists only in one protein of the proteome of interest).

Fig. 4. MS/MS spectrum allowing the identification of the peptide GSAGPPGATGFPGAAGR of collagen alpha-1(I) chain in the sample Campana 19. The precursor ion is doubly charged at m/z 730.34967 Da (Δm 0.25 ppm) at 60.05 min retention time. The sequence includes oxidation at P6 and P12 (underlined in the sequence above)

Fig. 4. MS/MS spectrum allowing the identification of the peptide GSAGPPGATGFPGAAGR of collagen alpha-1(I) chain in the sample Campana 19. The precursor ion is doubly charged at m/z 730.34967 Da (Δm 0.25 ppm) at 60.05 min retention time. The sequence includes oxidation at P6 and P12 (underlined in the sequence above)

Fig. 5. MS/MS spectrum allowing the identification of the peptide SGETGASGPPGFVGEK of collagen alpha-2(I) chain in the sample MBA 507. The precursor ion is doubly charged at m/z 747.3481 Da (Δm 0.20 ppm) at 53.75 min retention time. The sequence includes oxidation at P10 (underlined in the sequence above)

Fig. 5. MS/MS spectrum allowing the identification of the peptide SGETGASGPPGFVGEK of collagen alpha-2(I) chain in the sample MBA 507. The precursor ion is doubly charged at m/z 747.3481 Da (Δm 0.20 ppm) at 53.75 min retention time. The sequence includes oxidation at P10 (underlined in the sequence above)

16Despite the very low abundance of organic material in the studied historic samples, the proposed proteomic approach allowed to confirm the presence of collagen proteins in the three samples and it revealed the Bovinae subfamily of the collagen-based material in the sample Virgin and Child after Benedetto da Maiano.

Conclusion

17This work shows the efficiency of the optimized proteomic approach based on nanoLC, nanoESI high resolution MS, and MS/MS regarding the identification of organic material in historic stucco and their biological origin. The methodology developed on standard mock up built with commercially available standard proteins or binders, included the evaluation of six extraction conditions. It was demonstrated that the methodology is well-adapted to the study of samples containing a low amount of proteins entrapped in a mineral matrix such as stucco. Finally, the methodology was successfully applied to historic stucco from the Louvre Museum and the Museum of Fine Arts in Strasbourg. Collagens were identified in the three studied samples and the Bovinae subfamily of the collagen-based material was identified in one Virgin and Child after Benedetto da Maiano. The method developed could be used to study the migration of the organic material within the stuccoes to drive sampling protocols (surface / in depth sampling).

The authors acknowledge the project Esprit (Étude des stucs polychromés de la Renaissance italienne) and the JPI JHEP CH Pilot Call LeadART for support. The authors also thank the European Union (FEDER), the CNRS, the University of Lille, the Hauts-de-France Region, the Institut universitaire de France, and the IBISA network (Infrastructures en Biologie Santé et Agronomie).

Haut de page

Bibliographie

Aksamija A., Nowik W., Lehuédé P., Le Hô A.-S., Bormand M., Bouquillon A., 2019, “Investigation of organic additives in Italian Renaissance devotion stucco reliefs from French collections”, Journal of Cultural Heritage, 39, p. 66-81.

Arcolao C. et al., 1998, Le ricette del restauro – malte, intonaci, stucchi dal XV al XIX secolo, Marsilio Editori, Venezia.

Bankart G., 2002, The Art of the Plasterer (first ed. 1908), CRC Press/Routledge, London.

Beard G., 1983, Stucco and decorative plasterwork in Europe, Thames and Hudson, London.

Berner M., Weber J., 1999, “Stucco marble: notes on its preparation according to literature”, in C. Wittenburg (ed.), Baroque Artificial marble: Environmental Impacts, Degradation and protection, ENVIART, Protection and Conservation of European Cultural Heritage, Research Report n° 9, p. 11-20.

Caroselli M., Zumbühl S., Cavallo G., Radelet T., 2020, “Composition and techniques of the Ticinese stucco decorations from the 16th to the 17th century: results from the analysis of the materials”, Heritage Science, 8:102 [https://heritagesciencejournal.springeropen.com/articles/10.1186/s40494-020-00446-4].

Chowdhury S. K., Katta V., Beavis R. C., Chait B. T., 1990, “Origin and removal of adducts (molecular mass = 98 u) attached to peptide and protein ions in electrospray ionization mass spectra”, Journal of the American Society for Mass Spectrometry, 1, p. 382-388.

Dallongeville S., Garnier N., Rolando C., Tokarski C., 2016, “Proteins in art, archaeology and paleontology: from detection to identification”, Chemical Reviews, 116, p. 2-79.

Hendy J., Colonese A. C., Franz I., Fernandes R., Fischer R., Orton D., Lucquin A., Spindler L., Anvari J., Stroud E., Biehl P. F., Speller C., Boivin N., Mackie M., Jersie-Christensen R. R., Olsen J. V., Collins M. J., Craig O. E., Rosenstock E., 2018, “Ancient proteins from ceramic vessels at Çatalhöyük West reveal the hidden cuisine of early farmers”, Nature Communications, 9:4064 [https://www.nature.com/articles/s41467-018-06335-6].

Kahn P., 1995, “From genome to proteome: looking at a cell’s proteins”, Science (New York, N.Y.), 270, p. 369-370.

Krizkova M. C., Kuckova S. H., Santrucek J., Hynek R., 2014, “Peptide mass mapping as an effective tool for historical mortar analysis”, Construction and Building Materials, 50, p. 219-225.

Leo G., Cartechini L., Pucci P., Sgamellotti A., Marino G., Birolo L., 2009, “Proteomic strategies for the identification of proteinaceous binders in paintings”, Analytical and bioanalytical chemistry, 395, p. 2269-2280.

Montana G., Ronca F., 2002, “The ‘recipe’ of the stucco sculptures of Giacomo Serpotta”, Journal of Cultural Heritage, 3, p. 133-144.

Natali C., Lorenzini G., 2001, “Le ‘ricette’ degli stucchi in Italia settentrionale dal XV al XX secolo”. Atti del Convegno di Studi Lo Stucco Cult Tecnol Conoscenza Bressaone, p. 219-231.

Ronca F., 1994, “Protein determination in polychromed stone sculptures, stuccoes and gesso grounds”, Studies in Conservation, 39, p. 107-120.

Solazzo C., Fitzhugh W. W., Rolando C., Tokarski C., 2008, “Identification of protein remains in archaeological potsherds by proteomics”, Analytical Chemistry, 80, p. 4590-4597.

Tokarski C., Cren-Olive C., Martin E., Rolando C., 2002, “Proteomics approach for binding media studies in art painting”, in R. Van Grieken, L. Van’t dack, G. Meersman (eds.), Non-destructive testing and microanalysis for the diagnostics and conservation of the cultural and environmental heritage, University of Antwerp, Antwerp, Belgium.

Tokarski C., Martin E., Cren-Olivé C., Rolando C., 2003, “Protein studies in cultural heritage”, in C. Saiz-Jimenez (ed.), Molecular Biology and Cultural Heritage, Routledge, London, p. 119-130.

Tokarski C., Martin E., Rolando C., Cren-Olivé C., 2006, “Identification of proteins in Renaissance paintings by proteomics”, Analytical Chemistry, 78, p. 1494-1502.

Vasari G., [1550], 2008, The Lives of the Artists, trans. by J. C. and P. Bondanella, Oxford University Press, Oxford.

Vinciguerra R., Illiano A., De Chiaro A., Carpentieri A., Lluveras-Tenorio A., Bonaduce I., Marino G., Pucci P., Amoresano A., Birolo L., 2019, “Identification of proteinaceous binders in paintings: a targeted proteomic approach for cultural heritage”, Microchemical Journal, 144, p. 319-328.

Haut de page

Notes

1 Bankart, 2002; Beard, 1983; Berner, Weber, 1999; Arcolao et al., 1998; Natali, Lorenzini, 2001; Vasari, 2008; Aksamija et al., 2019.

2 Aksamija et al., 2019.

3 Bankart, 2002; Beard, 1983; Berner, Weber, 1999; Arcolao et al., 1998.

4 Berner, Weber, 1999.

5 Beard, 1983; Arcolao et al., 1998; Natali, Lorenzini, 2001.

6 Beard, 1983; Berner, Weber, 1999; Arcolao et al., 1998.

7 Bankart, 2002; Berner, Weber, 1999.

8 Bankart, 2002; Berner, Weber, 1999; Natali, Lorenzini, 2001.

9 Berner, Weber, 1999; Natali, Lorenzini, 2001.

10 Montana, Ronca, 2002.

11 Caroselli et al., 2020.

12 Montana, Ronca, 2002.

13 Ronca, 1994.

14 Aksamija et al., 2019.

15 Kahn, 1995.

16 Dallongeville et al., 2016; Vinciguerra et al., 2019; Tokarski et al., 2002 and 2003.

17 Tokarski et al., 2006; Leo et al., 2009.

18 Solazzo et al., 2008; Hendy et al., 2018.

19 Krizkova et al., 2014.

20 Tokarski et al., 2006.

21 Aksamija et al., 2019.

22 Aksamija et al., 2019.

23 Chowdhury et al., 1990.

Haut de page

Table des illustrations

Titre Fig. 1 a-b. Stucco mock-up made by chemically pure plaster and protein additives; a. In Eppendorf tubes; b. Unmolded after setting. Observed difference of color between mock-ups is not relevant to chemical composition
Crédits © A. Aksamija.
URL http://journals.openedition.org/techne/docannexe/image/8990/img-1.jpg
Fichier image/jpeg, 160k
Titre Table 1. The samples from historic stucco artworks
URL http://journals.openedition.org/techne/docannexe/image/8990/img-2.jpg
Fichier image/jpeg, 336k
Titre Fig. 2 a-d. Normalized MALDI-TOF spectra of extracted proteins from mock-up using: a. TFA and urea; b. DTAC and urea; c. Ammonium hydroxide. The spectrum d. shows the proteins extracted from milk without impregnation in mock up (10 pmol)
URL http://journals.openedition.org/techne/docannexe/image/8990/img-3.png
Fichier image/png, 56k
Titre Fig. 3 a-b. NanoLC nanoESI-MS spectrum showing intact lysozyme resulting from the extraction of stucco mock-up; a. Without barium acetate; b. With barium acetate desalting
URL http://journals.openedition.org/techne/docannexe/image/8990/img-4.png
Fichier image/png, 30k
Titre Fig. 4. MS/MS spectrum allowing the identification of the peptide GSAGPPGATGFPGAAGR of collagen alpha-1(I) chain in the sample Campana 19. The precursor ion is doubly charged at m/z 730.34967 Da (Δm 0.25 ppm) at 60.05 min retention time. The sequence includes oxidation at P6 and P12 (underlined in the sequence above)
URL http://journals.openedition.org/techne/docannexe/image/8990/img-5.png
Fichier image/png, 70k
Titre Fig. 5. MS/MS spectrum allowing the identification of the peptide SGETGASGPPGFVGEK of collagen alpha-2(I) chain in the sample MBA 507. The precursor ion is doubly charged at m/z 747.3481 Da (Δm 0.20 ppm) at 53.75 min retention time. The sequence includes oxidation at P10 (underlined in the sequence above)
URL http://journals.openedition.org/techne/docannexe/image/8990/img-6.png
Fichier image/png, 90k
Haut de page

Pour citer cet article

Référence papier

Amra Aksamija, Stéphanie Devassine, Fabrice Bray, Marc Bormand, Anne Bouquillon, Christian Rolando et Caroline Tokarski, « Proteomic analysis of Renaissance stucco: development of the analytical protocol and first results »Technè, 51 | 2021, 75-83.

Référence électronique

Amra Aksamija, Stéphanie Devassine, Fabrice Bray, Marc Bormand, Anne Bouquillon, Christian Rolando et Caroline Tokarski, « Proteomic analysis of Renaissance stucco: development of the analytical protocol and first results »Technè [En ligne], 51 | 2021, mis en ligne le 15 décembre 2022, consulté le 23 mars 2025. URL : http://journals.openedition.org/techne/8990 ; DOI : https://doi.org/10.4000/techne.8990

Haut de page

Auteurs

Amra Aksamija

Chercheur, Miniaturisation pour la Synthèse, l’Analyse et la Protéomique USR CNRS 3290, université de Lille (amra.aksamija[at]gmail.com).

Stéphanie Devassine

Assistante ingénieur, Miniaturisation pour la Synthèse, l’Analyse et la Protéomique USR CNRS 3290, université de Lille (stephanie.devassine[at]univ-lille.fr).

Fabrice Bray

Ingénieur d’étude, Miniaturisation pour la Synthèse, l’Analyse et la Protéomique USR CNRS 3290, université de Lille (fabrice.bray[at]univ-lille.fr).

Marc Bormand

Conservateur général du patrimoine, département des Sculptures, musée du Louvre, Paris (marc.bormand[at]louvre.fr).

Articles du même auteur

Anne Bouquillon

Ingénieur de recherche, département Recherche, C2RMF, Paris – UMR CNRS 8247, Paris (anne.bouquillon[at]culture.gouv.fr).

Articles du même auteur

Christian Rolando

Directeur de recherche, Miniaturisation pour la Synthèse, l’Analyse et la Protéomique USR CNRS 3290, université de Lille (christian.rolando[at]univ-lille.fr).

Caroline Tokarski

Professeur, Miniaturisation pour la Synthèse, l’Analyse et la Protéomique USR CNRS 3290, université de Lille (caroline.tokarski[at]u-bordeaux.fr).

Haut de page

Droits d’auteur

CC-BY-NC-ND-4.0

Le texte seul est utilisable sous licence CC BY-NC-ND 4.0. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.

Haut de page
Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search