Matrix Assisted Laser Desorption Ionization Mass Fingerprinting for

Article pubs.acs.org/ac

Matrix Assisted Laser Desorption Ionization Mass Fingerprinting for Identification of Acacia Gum in Microsamples from Works of Art Clara Granzotto*,†,§ and Ken Sutherland‡ †

Northwestern University - Art Institute of Chicago Center for Scientific Studies in the Arts (NU-ACCESS), 2145 Sheridan Road, Tech K111, Evanston, Illinois 60208, United States ‡ The Art Institute of Chicago, 111 South Michigan Avenue, Chicago, Illinois 60603, United States ABSTRACT: This paper reports an improved method for the identification of Acacia gum in cultural heritage samples using matrix assisted laser desorption ionization-time-of-flight mass spectrometry (MALDI-TOF MS) after enzymatic digestion of the polysaccharide component. The analytical strategy was optimized using a reference Acacia gum (gum arabic, sp. A. senegal) and provided an unambiguous MS profile of the gum, characterized by specific and recognized oligosaccharides, from as little as 0.1 μg of material. The enhanced experimental approach with reduced detection limit was successfully applied to the analysis of naturally aged (∼80 year) gum arabic samples, pure and mixed with lead white pigment, and allowed the detection of gum arabic in samples from a late painting (1949/1954) by Georges Braque in the collection of the Art Institute of Chicago. This first application of the technique to characterize microsamples from a painting, in conjunction with analyses by gas chromatography/mass spectrometry (GC/MS), provided important insights into Braque’s unusual mixed paint media that are also helpful to inform appropriate conservation treatments for his works. The robustness of the analytical strategy due to the reproducibility of the gum MS profile, even in the presence of other organic and inorganic components, together with the minimal sample size required, demonstrate the value of this new MALDI-TOF MS method as an analytical tool for the identification of gum arabic in microsamples from museum artifacts.

■

INTRODUCTION Plant Gums: Uses and Composition. Plant gums are natural polysaccharide-based materials obtained as exudates of various trees and shrubs or extracted from some seeds.1 While the major uses of gums are for food formulation, in the biomedical and pharmaceutical industries and for cosmetics and textiles,2,3 the motivation for this work is based on their extensive applications in works of art. Because of their property to form viscous gels or pastes by dissolution or dispersion in water, they have mainly been used as paint binding media and adhesives. Their utility can be traced to as early as the 3rd millennium B.C. in Ancient Egypt.4 Arabic, tragacanth, and fruit tree gums such as apricot and cherry have also been detected variously in mural paintings in Greece of the 4th−3rd centuries B.C.,5 Asian wall paintings of the 2nd−19th centuries,6 and in manuscript illumination of the 16th century.7 A major artistic use of plant gums since the 18th century has been as a medium for watercolor paints.8 Their identification in art and archeological samples can provide important technical information on the artifacts’ manufacture and can be helpful in the development of effective conservation treatment strategies. Plant gum polysaccharides are characterized by a complex, branched structure composed of several neutral sugars and uronic acids (UA) joined by specific glycosidic linkages.9,10 However, besides information on their molecular weight © 2017 American Chemical Society

(MW), which can be as much as a few million Daltons (Da), their exact structure and composition are still poorly understood. The polysaccharide component of gum arabic (obtained from Acacia species) is suggested to comprise a backbone of β1,3-D-galactopyranose (Galp) units with 1,6, 1,4, and 1,2 linked ramified side chains mainly containing β-D-Galp,α-L-Arabinofuranose (Araf), α-L-Rhamnopyranose (Rhap), and β-D-Uronic Acids.11,12 Study has shown that the gum arabic polysaccharide structure resembles that of Type II arabinogalactans (AGs), plant cell branched polysaccharides containing mainly arabinose and galactose.13 However, more and different substituents in the side chains have been proposed in the gum structure, such as disaccharides (e.g., α-D-Galp-(1,3)-αβ-L-Ara), trisaccharides (e.g., α-L-Araf-(1,4)-β-D-Galp-(1,6)-αβ-D-Gal), branched pentasaccharides, and a doubly branched heptasaccharide (Figure 1). Characterization. The complexity of gum chemistry makes their analysis challenging. This is particularly true in the cultural heritage field, since microsamples from works of art pose particular complications as a result of their small size, age, and material complexity. Several analytical techniques have been used: Fourier transform-infrared (FT-IR) and Raman spectrosReceived: December 2, 2016 Accepted: February 2, 2017 Published: February 2, 2017 3059

DOI: 10.1021/acs.analchem.6b04797 Anal. Chem. 2017, 89, 3059−3068

Article

Analytical Chemistry

Figure 1. Illustration of the suggested structure of gum arabic (Acacia senegal) polysaccharide core (based on Nie et al.12). The galactosidic linkages marked in red indicate hypothetical sites of enzyme attack by exo-β-1,3-galactanase.

time-of-flight mass spectrometry (MALDI-TOF MS) analysis.27 A similar strategy was also used for the study of locust bean gum28 and other uncommon galactomannans29 using electrospray ionization tandem mass spectrometry (ESI-MS/MS). These approaches provided important information on structural properties of the gums and gave improved insight into their possible applications, appearing promising for potential use in the analysis of cultural heritage objects. MALDI-TOF MS is routinely used for proteomic studies in the biomedical field30,31 and has already been successfully applied to the identification of proteins in samples from works of art.32−35 The method, called peptide mass fingerprinting (PMF), relies on the fact that proteolytic enzymes such as trypsin cleave proteins at specific amino acid sites producing unique patterns of peptides that are characteristic of the specific protein. Recently a similar strategy was developed for plant gum identification, involving partial enzymatic digestion followed by analysis of the released oligosaccharides by MALDI with tandem mass spectrometry (MALDI-TOF MS/ MS).36−38 The enzymes used are glycoside hydrolases that hydrolyze specific glycosidic linkages in the gum polysaccharides.39 Because of significant differences in the polysaccharide structure of various plant gums in terms of glycosidic linkages, monosaccharide sequence, and presence of more or less branched side chains, the released oligosaccharides are characteristic and specific, thus allowing identification of the polysaccharide and the corresponding plant gum. The method previously reported involves incubation of the gum sample with a mixture of two hydrolases: exo-β-1,3-galactanase and 1,4-βmannanase, which hydrolyze the β-1,3- galactosidic linkages at the nonreducing end and the 1,4-linkages between mannose units, respectively.36−38 The incubation results in the release of oligosaccharides of ∼600 to 2700 Da; the corresponding ions are detected in positive mode after reductive amination of the oligosaccharides. This derivatization step occurs directly on the MALDI plate by effect of 3-aminoquinoline, which is the matrix used for the sample spotting.40 With this method, arabic,

copy allow determination of the presence of a polysaccharide gum, which can be distinguished from other polysaccharides (e.g., starch, cellulose) and organic binding media families (proteinaceous, oil, resinous).14−16 However, the spectral features may be obscured in complex samples and are too poorly defined to provide identification of specific gums. This specificity is important since it may provide information on the formulation and possible sources of artists’ materials. The most widely used approaches for the discrimination of the plant gum source are based on the monosaccharide composition as determined by gas chromatography/mass spectrometry (GC/ MS) after complete acid hydrolysis and chemical derivatization,17−20 or with use of pyrolysis techniques (Py-GC/ MS).21−23 Other analytical methods have been used to avoid derivatization and so reduce the analysis time, such as anion exchange chromatography,5 capillary electrophoresis (CE),24 and enzyme-linked immunosorbent assay (ELISA).25,26 Besides the last method which is still under development for the investigation of plant gums, the most informative approaches based on GC/MS and Py-GC/MS still present difficulties in data interpretation due to chemical similarity of the saccharide monomers and the presence of multiple isomers and derivatives. Considering specific derivatization procedures, uronic acids are not readily detected with pyrolytic methylation,22 and inefficient derivatization of some products was found with pyrolysis-silylation,21 leading to a partial loss of structural information. In addition, the monosaccharide profile of a sample is challenging to interpret when dealing with mixtures of polysaccharide materials and can be altered by the presence of other (organic or inorganic) sample components,20 thus hindering correct identification. Mass Spectrometric Methods: MALDI Mass Fingerprinting. Research on carbohydrate structure for the food industry and other applications has seen in a few cases the employment of mass spectrometric techniques. Structural characteristics of mesquite gum, commonly used for soft drinks formulation, were revealed by partial enzymatic hydrolysis of the gum followed by matrix assisted laser desorption ionization 3060


Article


enzyme amount was 1−100 mU and digestion time between 3 and 24 h (adjustments to these parameters are detailed in the discussion, below). Digestion was quenched by heating the solution at 100 °C for 15 min and the final digested sample was dried in a SpeedVac at 30 °C before analysis. The Gettens samples were solubilized in water, mixed for 5 h, centrifuged (13400 rpm, 20 min, repeated 4−5 times as necessary) until the solution was clear, and the supernatant was digested as described above, using 10 mU enzyme and 24 h digestion. In the case of the samples from the Braque painting, due to their limited amount (tens of μg) and complex composition, enzymatic digestion was performed directly on the samples without any previous gum extraction step. A volume of 30 μL of 50 mM phosphate buffer were added to the sample together with 10 mU of exo-β-1,3-galactanase, and digestion was carried out for 24 h. After centrifugation (13400 rpm, 20 min) the supernatant was recovered, heated at 100 °C for 15 min, and dried in a SpeedVac at 30 °C. Oligosaccharides Purification. Several filter devices, utilizing different purification principles, were tested. The packing buffer of Micro Bio-Spin Chromatography columns (Bio-Rad, 6K MW exclusion limit) was removed according to the manufactures’ instructions. The columns were loaded with the digested solution of 1−1000 μg of gum arabic, and separation of the high MW components were carried out by centrifuging the columns at 1000 rpm for 4 min. In a modification of the recommended procedure, molecules retained on the Micro Bio-Spin column, i.e., those smaller than its size exclusion limit of 6 kDa, were flushed with a total of 500 μL of Tris(hydroxymethyl)aminomethane (Tris) buffer, 50 mM pH 7, for MALDI analysis. Graphitized carbon columns (Hypercarb, Thermo Scientific, 50 mg bed weight) were conditioned with 2.5 columns of the elution solvent (see below) acidified with 0.1% trifluoroacetic acid (TFA), followed by 2.5 columns of water. The digested solution of 1−1000 μg of gum was then loaded and the column washed with a further 2.5 columns of water. Oligosaccharides were eluted with several elution solvents: acetonitrile (ACN) 10%, 20%, 25%, and 40% in water acidified with 0.1% TFA; methanol 25% in water + 0.1% TFA; and ACN/isopropyl alcohol 1/3 in water + 0.1% TFA. Finally, a digest of the same amount of gum (1−1000 μg) was loaded on a cutoff membrane, Amicon Ultra 0.5 mL 3K (Millipore), and the eluted (MW lower than 3 kDa) and retained (MW higher than 3 kDa) fractions were recovered after centrifugation (13000 rpm, 20 min). MALDI-TOF MS. The dried digested sample was resuspended in deionized water and several dilutions were prepared (∼200−10 pmol for the reference gum). Derivatization and matrix preparation were done according to Rohmer et al.40 Briefly, 0.5 μL of sample was spotted on the MALDI plate and mixed with 1 μL of 3-aminoquinoline (3-AQ), prepared 20 mg/mL in ACN/water 1/2 (v/v), pH 5. Analyses were carried out at the Integrated Molecular Structure Education and Research Center (IMSERC), Northwestern University, IL, using a MALDI-TOF Autoflex III (Bruker) equipped with a tripled Nd:YAG laser, 355 nm wavelength. Spectra were obtained with a delayed extraction technology, in reflector positive mode with a grid voltage of 16 kV. A total of 3000 laser shots were accumulated for each spectrum. Data were analyzed using FlexAnalysis 3.0 software. THM-Py-GC/MS. Samples from the Braque painting were placed in Frontier Lab stainless steel sample cups for analysis by THM (thermally assisted hydrolysis and methylation) pyrolysis

tragacanth, cherry, guar, and locust bean gums could be discriminated according to their unique MS fingerprints.36,37 An important limitation of this initial analytical protocol is that ∼1 mg of plant gum was required to generate MS fingerprints. This is well in excess of the amount of sample typically obtained from a museum artifact, which may be in the order of tens of micrograms and is often complex (i.e., consisting of multiple organic/inorganic materials, in addition to the suspected polysaccharide). This paper reports several approaches that were pursued in order to decrease the amount of sample necessary to generate a reproducible and characteristic MS profile. In particular, since compounds such as salts and buffers can interfere with the ion yield and crystal formation of the MALDI matrix when performing MS analysis of carbohydrates,41 a variety of filter devices for sample purification were tested, along with evaluation of the effect of parameters such as the amounts of enzyme and buffer and the digestion timing. The strategy was developed by working with a single type of gum, gum arabic (sp. Acacia senegal), since it is considered to be historically the most widely used for art and archeology applications. In addition to a modern reference material, the method was applied to the investigation of naturally aged gum arabic samples dating from the 1930s obtained from the historic Gettens materials collection (Straus Center for Conservation, Harvard Art Museums) and of samples from a late painting by Georges Braque, Ajax, in the collection of the Art Institute of Chicago (1949/54, AIC No. 1997.447) that had previously undergone medium analysis using GC/MS.

■

EXPERIMENTAL SECTION Chemicals and Samples. Exo-β-1,3-galactanase (EC 3.2.1.145) was purchased from NZYtech. Standard gum arabic from Acacia senegal in powder form was obtained from Zecchi (Florence, Italy). Gum arabic samples from the Gettens materials collection, unpigmented (no. 53.29) and mixed 2:1 with lead white pigment (no. 82.06), were taken as scrapings from films cast on glass plates in 1933 and 1939, respectively. Two samples from the Braque painting were analyzed, one from an area of matte black paint and the other from a dark green paint area; the samples were not weighed but estimated to be a few tens of micrograms. All other chemicals were purchased from Sigma-Aldrich. Sample Preparation and Enzymatic Digestion. Reference gum samples were solubilized in deionized water at a concentration of 1% w/v, mixed for 2 h at room temperature and then overnight at 4 °C. In the case of turbid solution, due to incomplete solubilization of the gum nodules or the presence of wood or particulate residues, the sample was centrifuged (13 400 rpm, 20 min, repeated 4−5 times as necessary) until the solution was clear and the supernatant recovered. Aliquots of the gum solution, ranging from 1 mg to 100 ng of gum, were collected, dried in a vacuum centrifuge at 30 °C (concentrator 5301, Eppendorf, Hamburg, Germany), and subjected to enzymatic digestion. If samples were not digested immediately, they were dried and kept at −20 °C. Since in this paper attention is focused on a single gum, samples were digested with exo-β-1,3-galactanase, the specific enzyme effective for gum arabic, and not with the previous developed enzyme “cocktail” (mixture of exo-β-1,3-galactanase and 1,4-β-mannanase, which is useful when the type of gum is unknown38). The sample was recovered in 30 μL of 50 mM phosphate buffer (pH 6) and incubated at 50 °C with exo-β-1,3-galactanase; 3061


Article


Figure 2. Mass spectrum from MALDI-TOF MS analysis of a standard sample of gum arabic (Acacia senegal, Zecchi) enzymatically digested with exo-β-1,3-galactanase. Peak numbers refer to Table 1.

GC/MS. A volume of 2 μL of a 2.5% solution of tetramethylammonium hydroxide (TMAH) in methanol was added to the sample prior to insertion into a Frontier PY2020iD vertical microfurnace pyrolyzer, with the furnace at 550 °C. The pyrolyzer was attached to a Varian 3800 GC with Restek Rxi-5 ms column (30 m, 0.25 mm i.d., 0.25 μm film), interfaced to a Saturn 2200 MS, transfer line temperature 300 °C. The GC oven was programmed from 40 °C, with a 2 min hold, then increased at 20 °C/min to 300 °C, and held isothermally for 10 min; total run time 25 min. Helium was the carrier gas, with a constant flow of 1 mL/min, split ratio 1:10. The MS was run in scan mode (m/z 40−600) with the ion trap at 210 °C.

extent. The constituent pentoses composing the oligosaccharide ions were attributed to L-Arabinose (Ara), the hexoses could be ascribed to D-Galactose (Gal), the Deoxyhexoses (dHex) to L-Rhamnose (Rha), and the Hexuronic acids (HexA) to DGlucuronic acid (GlcA). All the experiments reported in this paper were evaluated by comparison with the data from this reference sample. Evaluation of Cleanup Methods. It is well established that the presence of salts and detergents, along with high MW compounds such as enzymes and undigested polysaccharide, can significantly interfere with the ionization of carbohydrate in MS analysis, which is already challenging due to their hydrophilic nature. In addition, the buffers remaining in the sample after digestion can prevent matrix crystallization. Purification of saccharides from complex mixtures is therefore usually necessary prior to MS analysis. On the basis of their successful application to sample cleanup in other MS studies, several filters were evaluated in this research following enzymatic digestion of the gum sample, with selected results reported in Table 1. The filters tested employed several separation principles: size exclusion (Micro Bio-Spin Chromatography column), affinity with polar compounds (Hypercarb graphitized carbon column), and MW cutoff membranes (Amicon Ultra centrifugal column). Hypercarb in particular has been established as a useful phase for desalting and fractionation of very polar compounds such as carbohydrates.42,43 Micro Bio-Spin Chromatography columns were tested for their potential to remove interfering high MW species, such as larger oligosaccharides, polysaccharide, and enzyme residues, with a MW higher than the exclusion limit of the resin (6 kDa).

■

RESULTS AND DISCUSSION A typical MS fingerprint obtained from the reference Acacia senegal gum is shown in Figure 2 (data for 1 mg sample, digested using the previously established protocol with 100 mU exo-β-1,3-galactanase for 5 h;38 see section Sample Preparation and Enzymatic Digestion for details) and the complete list of characteristic ion masses and the assigned oligosaccharides are reported in Table 1. The MS profile is characterized by several protonated ions, usually accompanied by the corresponding Na and K cation adducts, covering a mass range from 600 to 1300 Da. Considering the hypothetical structure of gum arabic polysaccharide and the action of the exoglycosidase (Figure 1), and by MALDI-TOF MS/MS performed in a previous work,38 most of the ions have been assigned to specific oligosaccharide. Although the enzyme is supposed to have an exo-action, the observed fragments suggest that it is also able to hydrolyze 1−3 bonds in the inner structure of the polysaccharide to some 3062


Article


Table 1. Experimental Masses and Assigned Oligosaccharides for Ions Observed in MALDI-TOF MS Analysis of Standard Gum Arabic (Acacia senegal)a Hypercarb peak no.

experimental mass [Da]

assignment

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

601.39 623.38 629.38 639.35 651.36 667.35 733.45 755.46 763.46 771.43 785.46 791.46 801.45 813.44 829.42 895.53 917.55 923.52 933.52 945.51 961.49 1055.60 1077.59 1085.62 1093.58 1107.61 1123.60 1187.67

29

1209.67

30 31

1217.70 1225.66

32 33

1239.70 1255.69

PentHex2 [M + H]+ PentHex2 [M + Na]+ HexdHexHexA [M + H]+ PentHex2 [M + K]+ HexdHexHexA [M + Na]+ HexdHexHexA [M + K]+ Pent2Hex2 [M + H]+ Pent2Hex2 [M + Na]+ PentHex3 [M + H]+ Pent2Hex2 [M + K]+ PentHex3 [M + Na]+ Hex2dHexHexA [M + H]+ PentHex3 [M + K]+ Hex2dHexHexA [M + Na]+ Hex2dHexHexA [M + K]+ Pent2Hex3 [M + H]+ Pent2Hex3 [M + Na]+ PentHex2dHexHexA [M + H]+ Pent2Hex3 [M + K]+ PentHex2dHexHexA [M + Na]+ PentHex2dHexHexA [M + K]+ Pent2Hex2dHexHexA [M + H]+ Pent2Hex2dHexHexA [M + Na]+ PentHex3dHexHexA [M + H]+ Pent2Hex2dHexHexA [M + K]+ PentHex3dHexHexA [M + Na]+ PentHex3dHexHexA [M + K]+ Pent3Hex2dHexHexA [M + H]+ Pent2Hex3dHex2 [M + H]+ Pent3Hex2dHexHexA [M + Na]+ Pent2Hex3dHex2 [M + Na]+ Pent2Hex3dHexHexA [M + H]+ Pent3Hex2dHexHexA [M + K]+ Pent2Hex3dHex2 [M + K]+ Pent2Hex3dHexHexA [M + Na]+ Pent2Hex3dHexHexA [M + K]+

MeOH 25%

ACN/IPA

X

X X X X X X X

X

ACN 25%

X X X X

X X X X X X

X

X

X

X

X

X X X X

Amicon 3 kDa X X X X X X X X X X X X X X X X X X X X X X X X X X X X X

X

X X

X X

X X

Results obtained with different filter devices and elution solvents used for the cleanup of a 1 mg sample, with respect to the characteristic oligosaccharides detected, are summarized. All listed oligosaccharides were detected as 3-aminoquinoline derivatives. a

ACN/IPA and ACN allowed detection of a sufficient number of ions (representing a majority of the diagnostic oligosaccharides) to identify the gum arabic from its MS profile (see Table 1, which summarizes results using 1 mg of gum). However, the column did not perform as well as expected since, comparing the MS profile of 1 mg of digested gum before and after purification, no improvement was obtained in the signal-tonoise (s/n). Finally, Amicon Ultra centrifugal filter with a 3 kDa cutoff was tested for removal of high MW interfering species, since the most characteristic oligosaccharides occur below this MW. Results for sample sizes greater than 10 μg revealed that all oligosaccharides were recovered (see Table 1) and a significant improvement in the s/n was observed. However, when decreasing the amount of gum to lower values, similar to those that might typically be obtained from a museum artifact (10 μg and below), oligosaccharides were retained on the filter, thus preventing the generation of a characteristic MS profile.

Following elution of these larger molecules, the column was flushed with Tris buffer with the aim of eluting the smaller, retained species, including the characteristic oligosaccharides (this is an adaptation of the typical use of the columns, in which it is the high MW components that are retained for analysis). Tris buffer was used as it had proved effective for size exclusion chromatography of plant gums in previous experiments.36 However, none of the characteristic oligosaccharides were detected in subsequent analysis, indicating either that they have been retained in the column or that combined buffers and salts not efficiently retained are interfering with the ionization of any recovered oligosaccharides. The presence of salts in the analyte solution was also suggested by the difficulties in drying of the sample while mixed with 3-AQ on the MALDI plate. Hypercarb graphitized carbon column has proved effective in retaining carbohydrates preferentially so that all other interfering species are removed. Among the elution solvents tested, several (methanol, ACN/IPA, and ACN) resulted in recovery of oligosaccharides for MALDI analysis, but only 3063


Article


Figure 3. Mass spectra from MALDI-TOF MS analysis of gum arabic from Acacia senegal: 10 μg digested for 5 h with (a) 10 mU and (b) 1 mU of exo-β-1,3-galactanase; 0.1 μg digested with 10 mU for (c) 5 and (d) 24 h. Peak numbers refer to Table 1.

In summary, the application of filters in most cases did not result in an improvement of the limit of detection, with the exception of Amicon centrifugal filters that were found to improve oligosaccharide recovery only for larger sample sizes. In particular, the strategy was unsuccessful when working with very small samples, for which most of the characteristic oligosaccharides appear to be retained on the filter media. Improvement of digestion parameters. As an alternative strategy to decrease the amount of gum necessary to generate a reproducible MS fingerprint the effect of parameters relating to the enzyme digestion was tested. Attention was initially focused on the enzyme amount, which is also connected to the quantity of salt in the sample solution since the enzyme is solubilized in a buffer. Since both residual enzyme and salts can potentially interfere with MS analysis of oligosaccharides, tests were made to reduce the amount of enzyme solution in relation to the sample size. While the established digestion procedure using 100 mU of enzyme for 5 h38 allowed to obtain a recognizable MS spectrum from as little as 250 μg gum arabic, improved results were indeed obtained by scaling down the enzyme amount proportionally to the gum quantity. With this strategy a reproducible MS was obtained from just 10 μg of gum digested with 10 mU enzyme (Figure 3a). Characteristic oligosaccharides were also obtained using 1 mU enzyme with the same amount of sample, although the MS profile was incomplete with an high percentage of the protonated [M + H]+ ions and

the Na+ adducts missing, and the result not highly reproducible (Figure 3b). Variation of the digestion timing allowed further improvement in the results. Various sample sizes (100−5 μg) were digested using 1 or 10 mU enzyme for 3, 5, 8, and 24 h. While 1 mU enzyme was not effective in generating a reproducible and complete MS at any of the digestion times (see Figure 3b for 5 h digestion), 10 mU enzyme produced characteristic spectra at 3, 5, and 8 h digestions, with most or all of the characteristic oligosaccharides detected. Reducing the sample size further to 1 or 0.1 μg, digestion with 10 mU enzyme generated only a few characteristic oligosaccharides: Figure 3c shows results for the 5 h digest of 0.1 μg in which low levels of diagnostic ions are observed at 733 [M + H]+ and 1255 Da [M + K]+. A few additional ions were detected in the 8 h digest; however, increasing digestion time to 24 h with 10 mU enzyme produced an almost complete spectrum from 0.1 μg sample (Figure 3d). Referring to Table 1, most of the ions missing (nos. 6, 7, 9, 12, 16, 18, 24, and 28) correspond to the protonated form, which is usually the less abundant due to the low ionization efficiency of oligosaccharides. However, the specific oligosaccharides of the gum are still clearly recognizable from the corresponding Na+ and K+ adducts. This represents a reduction in sample size required for successful MS fingerprinting of several orders of magnitude, and is much closer to the typical amount that might be obtained from a museum artifact. Therefore, the newly developed experimental protocol with optimized digestion 3064


Article

Analytical Chemistry parameters of 24 h with 10 mU enzyme was used for further measurements. Analysis of Aged Reference Materials and Samples from Braque Painting. Further tests were conducted to demonstrate the applicability of the improved MALDI mass fingerprinting strategy to complex, aged materials more typically encountered in works of art. Two naturally aged (∼80 year) gum arabic samples from the Gettens reference collection of artists’ materials, one composed of pure gum and the other mixed with lead white pigment, were analyzed to evaluate the possible influence of aging and pigment interference on the data obtained. Samples of a few hundreds of micrograms of the two aged samples, both taken as scrapings from films cast on glass plates, were prepared as described in the Experimental Section. The obtained mass spectra showed, in both cases, almost all the characteristic ions of gum arabic corresponding to those observed in the Acacia senegal reference. This result confirms that a good pattern for gum arabic polysaccharide can be obtained from aged samples even in the presence of lead white pigment. In addition, as discussed above, no cleanup of the sample is necessary so the possible loss of analyte in multiple and time-consuming steps is avoided. The technique was finally applied to the analysis of samples from a painting by Georges Braque (1882−1963), Ajax, in the collection of the Art Institute of Chicago (Figure 4). Samples from Ajax were analyzed previously by THM-Py-GC/MS with TMAH reagent, a technique that has been applied extensively to characterize a broad range of organic materials in works of art and which revealed the presence of drying oil (indicated by the presence of characteristic fatty and dicarboxylic acids, principally palmitic, stearic, azelaic, and suberic acids) and diterpene resin (Pinaceae, probably pine resin, indicated by abietane compounds such as 7-oxo-dehydroabietic acid), as illustrated in Figure 5. Evidence for a plant gum such as Acacia gum was found in some samples, based on the detection of methylated derivatives of arabinose/xylose (the two monosaccharides are not distinguishable under the experimental conditions) along with 1,2,4-trimethoxybenzene.22,44 The presence of gum had also been indicated by Fourier transform-infrared microspectroscopy (FT-IR) analysis of the samples. However, the data were ambiguous due to the limitations of these methods for discriminating carbohydrates, and the low levels of methylated derivatives detected by THMPy-GC/MS. For this reason the developed MALDI mass fingerprinting strategy was used to further investigate the binding media used by the artist. Two paint samples were selected in which plant gum were indicated by THM-Py-GC/ MS: a matte black paint (containing predominantly carbon black pigment; detail of sampling area shown in Figure 4) and a dark green sample (containing Prussian blue and yellow ochre; pigments used in the painting are described in a separate publication45). The samples were digested without any previous extraction of the polysaccharide in water due to the extremely limited amount (a few tens of micrograms at most). This procedure, described in the Experimental Section, limited the loss of analyte that would likely result from aqueous extraction or filtering steps. The mass spectrum obtained from the black sample is shown in Figure 6. Despite the complexity of the sample, including multiple organic materials in addition to pigment, and its small size, MALDI analysis allowed detection of a majority of the characteristic ions attributable to gum arabic from Acacia

Figure 4. Georges Braque, Ajax, with a detail showing location of black paint sample. 1949/54, oil and mixed media on paper mounted on canvas,179 cm × 71 cm, Bequest of Florene May Schoenborn, Art Institute of Chicago no. 1997.447, Copyright 2016 Artists Rights Society (ARS), New York/ADAGP, Paris.

senegal. Besides the already identified oligosaccharides listed in Table 1, additional ions detected in the reference gum arabic MS profile were also observed in the mass spectrum of the Braque sample, thus providing further support for the identification. Examination of their m/z values indicate that these additional ions are K cation adducts that have not been derivatized by the 3-aminoquinoline matrix. The presence of similar, nonderivatized ions was previously observed in the MS fingerprint of locust bean gum.38 Mass intervals between the ions correspond to pentoses, hexoses, and deoxyhexoses, thus aiding assignment to specific oligosaccharides: for example, the ion at m/z 705.22 can be attributed to Hex4 [M + K]+, m/z 867.30 to Hex5 [M + K]+, and m/z 1131.44 to Pent2Hex5 [M + K]+. For the dark green paint, a different MS pattern was obtained after enzymatic digestion that showed similarity to reference data for another Acacia gum reference, A. seyal, thus suggesting the potential of this method to discriminate different Acacia 3065


Article


The promising result from MALDI-TOF MS reveals that aging and the presence of other organic media and pigments do not appear to interfere significantly with the enzymatic digestion and analysis, thus demonstrating the robustness of the method and the feasibility of applying the newly developed and optimized strategy to samples of significant complexity and age, such as those from works of art and archeological artifacts.

■

CONCLUSIONS This paper reports a method for characterizing Acacia senegal gum in cultural heritage materials using MALDI-TOF MS fingerprinting. The study demonstrates that an optimized protocol using enzymatic digestion for 24 h followed by MALDI analysis results in a characteristic MS profile of the gum polysaccharide, and it is a powerful strategy for its unambiguous identification in aged microsamples from works of art. Several approaches were pursued in order to decrease the amount of sample necessary to generate a reproducible and characteristic MS profile. A variety of filter devices were tested and not found to be effective; however, a significant improvement was obtained by scaling down the amount of enzyme and buffer and modifying the digestion timing. Results showed a remarkable improvement in detection limits, generating for the first time characteristic spectra from as little as 0.1 μg of gum. The analytical approach has been applied successfully to naturally aged gum arabic samples from the historic Gettens materials collection and was validated by the detection of gum arabic in a late painting by Georges Braque, confirming the suspected use of mixed media in this work. These results demonstrate the robustness and reliability of the analytical strategy, since neither the aging nor the presence of other organic (oil and resin) and inorganic (pigment) components resulted in a significant modification of the characteristic MS profile of gum arabic. Another advantage of the MALDI-TOF MS strategy is that the enzymatic digestion could be performed

Figure 5. THM-Py-GC/MS data (total ion chromatogram) for black paint sample from Georges Braque’s Ajax. Su = suberic acid, Az = azelaic acid, P = palmitic acid, S = stearic acid, 7oxoDHA = 7-oxodehydroabietic acid, M1 and M2 = monosaccharide markers for arabinose/xylose (see Riedo et al.44); all compounds detected as methylated derivatives.

species. This topic is a focus of ongoing research, using a variety of botanical reference samples of known provenance, that will be discussed in a separate paper.46 The finding of Acacia gum, in addition to drying oil and Pinaceae resin as demonstrated by THM-Py-GC/MS analysis, emphasizes the complementary nature of the analytical approaches, and supports visual observations suggesting that Braque used mixed media for painting Ajax. This technique, combining oil-based and water-based media, has significant implications for the condition and conservation treatment options for the painting.45

Figure 6. Mass spectrum from MALDI-TOF MS analysis of black paint sample from Georges Braque’s Ajax, enzymatically digested with exo-β-1,3galactanase. Peak numbers refer to Table 1; asterisks refer to additional ions observed in the standard gum arabic from Acacia senegal but not previously assigned; see text for discussion. 3066


Article


(5) Colombini, M. P.; Ceccarini, A.; Carmignani, A. Journal of Chromatography A 2002, 968, 79−88. (6) Birstein, V. J. Stud. Conserv. 1975, 20, 8−19. (7) Mills, J. S.; White, R. The Organic Chemistry of Museum Objects; Butterworth-Heinemann: New York, 1994. (8) Ormsby, B. A.; Townsend, J. H.; Singer, B. W.; Dean, J. R. Stud. Conserv. 2005, 50, 45−66. (9) Dey, P. M.; Harborne, J. B. Methods in Plant Biochemistry, Vol. 2, Carbohydrates; Academic Press: London, 1989; pp 483−522. (10) Anderson, D. M. W.; Howlett, J. F.; McNab, C. G. A. Phytochemistry 1986, 26, 309−311. (11) Tischer, C. A.; Gorin, P. A. J.; Iacomini, M. Carbohydr. Polym. 2002, 47, 151−158. (12) Nie, S.-P.; Wang, C.; Cui, S. W.; Wang, Q.; Xie, M.-Y.; Phillips, G. O. Food Hydrocolloids 2013, 31, 42−48. (13) Clarke, A. E.; Anderson, R. L.; Stone, B. A. Phytochemistry 1979, 18, 521−540. (14) Daher, C.; Paris, C.; Le Hô, A.-S.; Bellot-Gurlet, L.; Échard, J.-P. J. Raman Spectrosc. 2010, 41, 1494−1499. (15) Vandenabeele, P.; Wehling, B.; Moens, L.; Edwards, H.; De Reu, M.; Van Hooydonk, G. Anal. Chim. Acta 2000, 407, 261−274. (16) Morris, W. W.; Salmon, W. Anal. Chem. 1963, 35, 600−600. (17) Bleton, J.; Mejanelle, P.; Sansoulet, J.; Goursaud, S.; Tchapla, A. Journal of Chromatography A 1996, 720, 27−49. (18) Pitthard, V.; Finch, P. Chromatographia 2001, 53, S317−S321. (19) Bonaduce, I.; Brecoulaki, H.; Colombini, M. P.; Lluveras, A.; Restivo, V.; Ribechini, E. Journal of Chromatography A 2007, 1175, 275−282. (20) Lluveras-Tenorio, A.; Mazurek, J.; Restivo, A.; Colombini, M. P.; Bonaduce, I. PLoS One 2012, 7, E49383. (21) Chiantore, O.; Riedo, C.; Scalarone, D. Int. J. Mass Spectrom. 2009, 284, 35−41. (22) Riedo, C.; Scalarone, D.; Chiantore, O. Anal. Bioanal. Chem. 2010, 396, 1559−1569. (23) Scalarone, D.; Chiantore, O.; Riedo, C. J. Anal. Appl. Pyrolysis 2008, 83, 157−164. (24) Größl, M.; Harrison, S.; Kaml, I.; Kenndler, E. Journal of Chromatography A 2005, 1077, 80−89. (25) Arslanoglu, J.; Schultz, J.; Loike, J.; Peterson, K. J. Biosci. 2010, 35, 3−10. (26) Lee, H. Y.; Atlasevich, N.; Granzotto, C.; Schultz, J.; Loike, J.; Arslanoglu, J. Anal. Methods 2015, 7, 187−196. (27) López-Franco, Y. L.; de la Barca, A. M. C.; Valdez, M. A.; Peter, M. G.; Rinaudo, M.; Chambat, G.; Goycoolea, F. M. Macromol. Biosci. 2008, 8, 749−757. (28) Simoes, J.; Nunes, F. M.; Rosario Domingues, M.; Coimbra, M. A. Carbohydr. Polym. 2011, 86, 1476−1483. (29) Cerqueira, M. A.; Souza, B. W. S.; Simões, J.; Teixeira, J. A.; Domingues, M. R. M.; Coimbra, M. A.; Vicente, A. A. Carbohydr. Polym. 2011, 83, 179−185. (30) Anderson, N. L.; Anderson, N. G. Electrophoresis 1998, 19, 1853−1861. (31) Kahn, P. Science 1995, 270, 369−370. (32) Tokarski, C.; Martin, E.; Rolando, C.; Cren-Olivé, C. Anal. Chem. 2006, 78, 1494−1502. (33) Hynek, R.; Kuckova, S.; Hradilova, J.; Kodicek, M. Rapid Commun. Mass Spectrom. 2004, 18, 1896−1900. (34) Kuckova, S.; Hynek, R.; Kodicek, M. Anal. Bioanal. Chem. 2007, 388, 201−206. (35) Kirby, D. P.; Buckley, M.; Promise, E.; Trauger, S. A.; Holdcraft, T. R. Analyst 2013, 138, 4849−4858. (36) Granzotto, C. Methodological developments based on mass spectrometry for the analysis of glycoproteins and polysaccharides of plant gums: an application to cultural heritage samples. Ph.D. Thesis, Ca’ Foscari University of Venice, Venice, Italy, 2014. (37) Granzotto, C.; Arslanoglu, J.; Tokarski, C. In MaSC Users’ Group for Mass Spectrometry and Chromatography Meeting Program, Chicago, IL, May 21−22, 2015; p 15 (abstract).

directly on the sample without any previous extraction step, thus limiting the risk of analyte loss. The work therefore represents a significant breakthrough for the characterization of gums in complex matrixes, providing an analytical strategy of value not only for the cultural heritage field but for scientists in other areas performing carbohydrate analysis. The protocol and data reported here provide a resource for researchers studying gum arabic, including those using MALDI-TOF but without access to MS/MS capabilities. The promising results of this study will be used as a basis of further investigation of plant gums in works of art, exploiting the potential of MALDI-TOF MS to discriminate gums that may have only minor differences in their underlying monosaccharide profile (as revealed by GC/MS). Future research will include method optimization for different types of plant gum, including mixtures; the possibility to discriminate gums from different Acacia species; and analysis of various combinations of materials and media to investigate possible interferences in the MS profiles.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected] (Clara Granzotto). ORCID

Clara Granzotto: 0000-0002-7839-4316 Present Address §

C.G.: The Metropolitan Museum of Art, Scientific Research Department, 1000 5th Avenue, New York, NY 10035. Phone: + 1 212 3965509.

Author Contributions

The manuscript was written through contributions of all authors. Both authors have given approval to the final version of the manuscript. Both authors contributed equally. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The research was possible thanks to a visiting research scholarship at the Northwestern University/Art Institute of Chicago Center for Scientific Studies in the Arts (NUACCESS), funded through a grant from the Andrew W. Mellon Foundation. Supplemental support was provided by the Materials Research Center, the Office of the Vice President for Research, the McCormick School of Engineering and Applied Science and the Department of Materials Science and Engineering at Northwestern University. MALDI analyses were performed at the IMSERC facility at Northwestern University. Francesca Casadio and Allison Langley at the Art Institute of Chicago, and Marc Walton, S. Habibi Goudarzi, Saman Shafaie, and Paul Thomas at Northwestern University are thanked for their support and contributions. Samples from the Gettens materials collection were kindly supplied by Narayan Khandekar, Harvard Art Museums.

■

REFERENCES

(1) BeMiller, J. N. In Kirk-Othmer Encyclopedia of Chemical Technology; John Wiley & Sons, Inc.: New York, 2000. (2) Saha, D.; Bhattacharya, S. J. Food Sci. Technol. 2010, 47, 587−597. (3) Verbeken, D.; Dierckx, S.; Dewettinck, K. Appl. Microbiol. Biotechnol. 2003, 63, 10−21. (4) Lucas, A. Ancient Egyptian Materials and Industries; Kessinger Publishing, LLC: London, 2003. 3067


Article

Analytical Chemistry (38) Granzotto, C.; Arslanoglu, J.; Rolando, C.; Tokarski, C. Sci. Rep., in press. (39) Davies, G.; Henrissat, B. Structure 1995, 3, 853−859. (40) Rohmer, M.; Meyer, B.; Mank, M.; Stahl, B.; Bahr, U.; Karas, M. Anal. Chem. 2010, 82, 3719−3726. (41) Harvey, D. J. Int. J. Mass Spectrom. 2003, 226, 1−35. (42) Packer, N.; Lawson, M.; Jardine, D.; Redmond, J. Glycoconjugate J. 1998, 15, 737−747. (43) Ruhaak, L. R.; Deelder, A.; Wuhrer, M. Anal. Bioanal. Chem. 2009, 394, 163−174. (44) Riedo, C.; Scalarone, D.; Chiantore, O. Anal. Methods 2013, 5, 4060−4067. (45) Langley, A.; Sutherland, K.; Casadio, F.; Maccagnola, S. AIC Paintings Specialty Group Postprints, 2015, 28, in press. (46) Granzotto, C.; Arslanoglu, J.; Sutherland, K. In TECHNART 2017: Non-destructive and Microanalytical Techniques in Art and Cultural Heritage, Bilbao, Spain, May 2−6, 2017 (abstract).

3068


Matrix Assisted Laser Desorption Ionization Mass Fingerprinting for

Recommend Documents