The potential of aptamers for the analysis of ceramic bound proteins found within pottery

Archaeological pottery are the most numerous objects found during excavations and reflect the culinary practices of the past. However, their functionality for cooking/storing specific foods or drinks cannot be deduced solely from comparing their shapes and sizes. Analysis of protein residues bound to ceramics can reveal the protein/animal type through their amino acid sequence, thus enabling direct identification of food types. Therefore, the aim of our experimental study was to test sixteen aptamers for the analysis of proteinaceous organic residues found within the porous structure of pottery. Traditionally prepared archaeological ceramic replicas were cooked for 5 days in various food/protein suspensions, were UV aged, buried for a year, excavated, and extensively cleaned. Their shards were analysed using immunofluorescence microscopy with aptamers. Results show that eight aptamers (Clone1 and Kirby for egg residuals; seqU5 and BLG14 for milk residuals; HA for blood residuals; Gli4 for gluten residuals; Par1 for fish residuals; and D1 for collagen residuals) produced a successful/specific immunofluorescence microscopy result when they were hybridised to shards containing target protein residuals. Interestingly, on whole egg control samples, when the egg lysozyme-targeting aptamer Kirby was used, fluorescence intensity was 3.1 times greater compared to that observed with anti-ovalbumin antibodies.


Preparation of fresh control samples
Archaeological ceramic replicas (vessels) were traditionally manufactured and burned in a wood heap, by the Pottery Center Bahor (Institute for the Development and Promotion of Pottery 3,714,195,000, Šaleška cesta 1, 3320 Velenje; Slovenia).Traditional manufacture ensures that the final product is porous facilitating the retention of organic compounds within the ceramic matrix.This is because natural clays and minerals are used, which contain inherent porosity.Moreover, lower burning temperatures within a wood heap to those of modern furnaces contribute to its porosity 27 .Replicas were washed and cleaned three times with Milli-Q water and were broken into ceramic shards (~ 3 × 3 cm pieces) using a chisel and a hammer.Control samples were prepared on ceramic shards, and each shard was impregnated with either one of the two suspensions: (a) 50 g of whole chicken egg mixed with a stick blender; and (b) 5 g of lysozyme from chicken egg white (L6876-5G, Sigma Aldrich) and 50 g of MilliQ water (final lysozyme concentration of 100 g/L).

Preparation of cross-sections of samples in resin
Firstly, a ~ 2 × 2 mm large piece of sample was removed from the concave surface of a control/model sample, using a hammer and a sharp slotted screwdriver.Any piece containing organic residuals, visible under a stereomicroscope, was rejected.The selected piece was embed in copolymer resin Kristal PS (Samson Kamnik LLC, Slovenia), containing a 2% concentration (w/w) of catalysator Kristal PS (Methyl ethyl peroxide (MEK); Samson Kamnik LLC, Slovenia).The resin was polymerised for 5 days at room temperature.Hardened samples were then polished on the table top polishing equipment RotoPol-15 (Struers Westlake, OH), with six successively finer grades of Micro-mesh abrasive MD-Dac pads (320, 600, 800, 1200, 2400 and 4000 mesh, Struers Westlake, OH).Paraffin oil (Agrolit LLC, Slovenia) was used as a lubricant only with the first polishing pad (320 mesh) and for later polishing steps no lubricant was applied.Finally, paraffin oil was removed by shortly wiping (to.to paper sheets, Tosama LLC, Slovenia) the samples with petroleum ether (Sigma Aldrich, Germany), and by dry wiping.Then samples were fixed onto a microscope glass slide, using a Fimo polymer clay (Staedtler Mars GmbH & Co. KG, Germany), so that the cross sections were facing upwards and were completely aligned with the ground surface.

Hybridization of antibodies to sample cross-sections
First, the exposed cross-section was covered with a 50 µl drop of 5% CSA (calf serum albumin; N4637, Sigma-Aldrich), which is used as a blocking solution (see Fig. 1).After a 60 min long incubation period at 23 °C, the blocking solution was removed using a pipette, and 50 µl of the primary anti-chicken egg albumin antibody (anti-ovalbumin Ab produced in rabbit, Sigma-Aldrich, C6534), diluted in blocking solution (1:10), was pipetted onto each cross-section and the samples were incubated overnight at 4 °C.Any unbound primary antibody was then washed off, by applying successive 100 µl drops of PBS at 4 °C to the sample, waiting 5 min, and removing the drop with a pipette.This washing process was repeated three times and great care was taken not to touch the sample itself with the tip of the pipette.After washing, 50 µl of the FITC (fluorescein isothiocyanate)-conjugated secondary anti-rabbit IgG (Ab produced in goat, Sigma-Aldrich, F0382), diluted in blocking solution (1:10), was pipetted onto each cross section and the samples were incubated for 2 h in the dark at room temperature.Unbound secondary antibodies were washed off, as described above, and during these and all of the subsequent steps, we removed all environmental sources of visible bright light.Finally, any remaining liquid around the sample was carefully removed, using a cotton swab (without touching the sample itself) and the samples were left to dry for 1 h in the dark.

Aptamer selection and reconstruction
Appropriate protein targets were listed according to LC-MS/MS proteomic studies of food residuals found in ancient ceramics [28][29][30][31][32] .The amino acid peptides found (along with histamine, which is not a peptide but is abundant in fish) were sequenced within these studies and thus any of the preserved proteins were identified.Then, within the published scientific articles, we searched for DNA aptamers, which specifically targeted the LC-MS/MS identified proteins (found within ancient ceramics), and for their single strand nucleic acid sequences (Table 1).In literature, these aptamers were developed via chemical synthesis (Systematic Evolution of Ligands by Exponential enrichment or SELEX), which includes reiterative selection cycles, that bind their selected protein www.nature.com/scientificreports/targets with high affinities and specificities, via folding into specific secondary ligand binding structures 26 .The selected protein ligand targets are listed in Table 1 under the "target protein" column.According to the literature obtained DNA aptamer sequences, we used the OligoAnalyzer™ Online Tool (Integrated DNA Technologies, Inc.) with default values (Figs. 2 and 3) to reconstruct the folded (native) secondary structures of the aptamers.Within this software, the ΔG values of the reconstructed aptamers were calculated according to the next equation: where K is the equilibrium constant of the binding reaction; R is the gas constant with a value of 8.314 J K-1 mol-1; T is the temperature of the reaction [K]; and ΔG° is Gibbs free energy change under standard conditions [kcal mol -1 ].A more negative ΔG indicates a stronger binding affinity of aptamers to target proteins (Figs. 2 and 3).

Aptamer synthesis and preparation
The collected aptamer sequences were sent to Sigma-Aldrich (Switzerland), where they were synthesised (synthesis scale of 0.2 µmol), HPLC purified and modified with 5' end biotinylation (Btn).In our laboratory, the delivered aptamers were briefly centrifuged which ensured us that the lyophilized aptamer pellet was at the bottom of the tube.Then they were resuspended in Aptamer Resuspension Buffer (Product # RTW0001; Base Pair Biotechnologies, Inc.) to achieve a 100 µM concentration (using microliter volumes indicated on the Sigma-Aldrich aptamer synthesis report).After that, they were incubated at room temperature for 30 min, vortexed for 20 s and centrifuged at 10,000 × g for 1 min.Aliquots were created and stored at -20 °C.Prior to use, the working aptamer solution was prepared by thawing the aliquots and by diluting them (1:10) in the Aptamer Folding Buffer (Product # RTW0003; Base Pair Biotechnologies, Inc.).Finally, the solution was transferred into 200 µl PCR tubes (strips), heated within the MyCycler Thermal PCR Cycler (BioRad) to 93 °C for 5 min and incubated at room temperature for 25 min.

Hybridization of aptamers to sample cross-sections
First, the exposed cross-section was covered with a 50 µl drop of 5% CSA (N4637, Sigma-Aldrich).After a 60 min long incubation period at 23 °C, the blocking solution was removed using a pipette, and 50 µl of the working aptamer solution, was pipetted onto each cross-section and the samples were incubated overnight at room temperature (see Fig. 4).Any unbound aptamer was then washed off, by applying successive 100 µl drops of aptamer washing buffer (PBS with 5 mM MgCl 2 ) to the sample, waiting 5 min, and removing the drop with a pipette.This washing process was repeated three times.After washing, 50  www.nature.com/scientificreports/isothiocyanate)-conjugated streptavidin (S0966, TCI chemicals), diluted in aptamer washing buffer (1:33), was pipetted onto each cross-section and the samples were incubated for 2 h in the dark at room temperature.Unbound streptavidin was washed off, as described above, and during these and all of the subsequent steps, we removed all environmental sources of visible bright light.Finally, any remaining liquid around the sample was carefully removed, using a cotton swab (without touching the sample itself) and the samples were left to dry for 1 h in the dark.

Fluorescent microscopy
Prior to and after the hybridization, the sample cross-sections were photographed in the widefield fluorescence observation mode of the Axio Imager.Z2m LSM 800 microscope (Carl Zeiss GmbH, Germany), running the software Zen Blue 2.5 (Carl Zeiss GmbH, Germany).The HXP unit (metal halide fluorescence light lamp module; 120 V) coupled with the cube filter set 10 (excitation: 450-490 nm, Beamsplitter: FT 510, emission: 515-565 nm; Carl Zeiss GmbH, Germany, 488,010-9901-000) was turned on and within the Acquisition tab of the Zen Blue 2.5 software, under the »widefield« screen (within the smart setup menu), the FITC channel (519 nm) was selected.The intensity of the lamp module was set to 20.0% and the shift value (under exposure time) was set to 80%.Images were captured at a magnification of 100 × using the Axiocam 503 camera in black and white mode and a green pseudo colour (false colour) was later added.

Laser scanning confocal fluorescence microscopy
Hybridised sample cross sections were laser scanned by using the laser-scanning confocal fluorescent observation mode, within the Zeiss Axio Imager.Z2m LSM 800 microscope.Firstly, the LSM 800 laser module LM URGB (contains fibre-coupled, pigtailed and collimated lasers) was turned on and the anti-vibration table Vision IsoStation (Newport, USA), on which the microscope system is positioned, was filled with 1 bar of pressure using compressed air.The manual tube slider (for switching between VIS/Camera) was placed into the camera position (100% camera) and within the Acquisition tab of the Zen Blue 2.5 software, under the »LSM« screen (within the smart setup menu), the FITC channel (519 nm) was selected.The intensity of the 488 nm argon diode laser (10 mW, class 3B), used for the excitation of FITC, was set to 2%.Emission detection was via the LSM 800 MAT Confocal MA-detection module (Carl Zeiss GmbH, Germany), which consists of a main beam splitter (MBS), a variable pinhole with automatic alignment, two variable secondary dichroics (VSD) placed at 10 degree angle to incident beam for most effective excitation light suppression; and an emission filter in front of each of the two multi-alkali (MA) PMT confocal channel detectors.The pinhole was set to 1 Airy unit (AU) with an opening diameter of 36 µm.The digital gain of the PMT detectors was set to 0.9 (digital offset of 0), however the master

Surface topography scans
The topography of hybridised cross section was laser scanned, by using the 405 nm laser scanning mode, within the above mentioned Zeiss microscope system.Therefore, the same physical units and microscope system settings were employed as in Sect.3.8.3.Within the applications tab of the Zen Blue 2.5 software, under the »topography«, the image acquisition option was selected.The intensity of the 405 nm violet diode laser (5 mW, class 3B), used for scanning, was set to 10%.The pinhole was set to 1 Airy unit (AU) with an opening diameter of 25 µm.Even though the master gain settings varied from sample to sample, its values were always close to around 250 ± 10 V.
In order to engulf the entire surface topography of a cross-section, the scanning boarders (set first/last) were set to acquire a 700 µm long interval of Z-stack images (each image was 0.41 µm apart; 100 × magnification) and these were saved as a CZI file format.This format was opened within the ImageJ distribution Fiji, and the images were reduced to 8-bits.Finally, by selecting the 3D Viewer option, within the »Plugins« menu, a 3D surface topography model was generated.

Aptamer testing
Aptamers were tested using the ELISA procedure, according to the protocol published by Heginbotham et al. 48ith some modifications.Firstly, protein suspensions (prepared in section "Immunofluporecence microscopy of control samples") were 50 fold diluted in sodium bicarbonate buffer (SBB: 100 mM of NaHCO3, pH 9.6), were separately applied onto Nunc-immuno microtiter plate (96 MicroWell Plates, MaxiSorp Sigma Aldrich; 80 μL/ well of sample suspension/MilliQ water for negative control/bovine serum albumin A7906 (BSA; Sigma Aldrich; 10 μg/ml prepared in SBB) for standard), and were incubated overnight (2-8 °C).Afterwards, 300 μl of 5% CSA (N4637, Sigma-Aldrich) blocking solution was added, to prevent unspecific binding of antibodies in the further steps of the procedure.For the indirect ELISA, the working aptamer solution (prepared in Sect.3.6) was diluted in Aptamer Folding Buffer, at a ratio of 1:10, and the streptavidin − alkaline phosphatase (AP) (S2890; Sigma Aldrich) was diluted in CSA, at a ratio of 1:100.For both aptamers and streptavidin − AP, 100 μl were applied to each well.To obtain enzyme-substrate reaction, 100 μl of colourless p-NPP (P7998, Sigma Aldrich) was added to the wells and incubated for 30 min at room temperature.The indicator was, in the case of a positive reaction, converted into a coloured product, by the reporting enzyme.Finally, the results were read using the Multiskan Sky (Thermo Scientific) spectrophotometer (optical density at 405 nm).Between each step, the wells were washed twice, using 300 μl of 0.02% Tween-20 in PBS, and twice with PBS only.

Aptamer testing
Results of testing aptamers using the ELISA assay are presented in Supplementary Table 1.After the performed assay, aptamers for lysozyme (from chicken egg), when applied onto the lysozyme suspension wells, had OD 405nm values of 0.463 for Clone1, 0.380 for Kirby and 0.037 for Apt1L.Aptamers for gliadin (gluten), when applied onto the wheat flour suspension wells, had OD 405nm values of 0.577 for Gli1 and 0.519 for Gli4.Aptamers for milk proteins, when applied onto the raw milk wells, had OD 405nm values of 0.510 for seqU5 (casein) and 0.394 for BLG14 (β-lactoglobulin).Aptamers for haemoglobin, when applied onto the haemoglobin suspension wells, had OD 405nm values of 0.388 for Hb, 0.471 for HA and 0.015 for G15 T1.The aptamer Mb1 (for myoglobin), when applied onto the raw beef suspension wells, had an OD 405nm value of 0.488.Aptamers for fish related compounds, when applied onto the fish traut suspension wells, had OD 405nm values of 0.412 for Par1 (for parvalbumin), 0.087 for H2 (for histamin) and 0.031 for H47 (for histamin).Finally, aptamers for collagen, when applied onto the gelatin suspension wells, had OD 405nm values of 0.400 for D1 (collagen type XI) and 0.615 for CTx 2R-2 h www.nature.com/scientificreports/(collagen type I).For BSA standard and for negative control wells (MilliQ water), the signal was low (OD 405nm ˂ 0.1) for all tested aptamers.

Immunofluporecence microscopy of control samples
To compare aptamer hybridisation results with the allready established antibody hybridisation procedure [49][50][51] , we hybiridsed the same whole egg control sample (chicken egg mix applied onto the ceramic shard), with either the lysozyme targeting aptamer Kirby or with the anti-ovalbumin antibody (see Figs. 5A and 6A).After both hybridization processes, fluorescence intensities (under the widefield fluorescence observation mode) in the thick upper whole egg layer increased from 28 to 43 (by 15) for the anti-ovalbumin antibody (Fig. 5B and C) and from 30 to 76 (by 46) for the aptamer Kirby (Fig. 6B and C).Therefore, in comparison with the use of antibodies, the increase in fluorescence intensity was 3.1 times greater for the Kirby aptamer.Moreover, the hybridisation with the aptamer produced a much clearer picture (Fig. 6C), and an even clearer localisation of fluorescence within the whole egg layer was observed in the confocal fluorescence observation mode (Fig. 6D).In contrast, antibody hybridisation produced smaller and larger fluorescent stains, mainly scattered around the polished resin area (Fig. 5C and D).
After both hybridisations (antibodies or aptamers), the upper whole egg layer swelled up and increased in size for about 1.5 times.The elevation of the egg layer, caused by the swelling, above the rest of the polished sample's surface is clearly visible from the 3D topography models (Fig. 5E for the anti-ovalbumin Ab and Fig. 6E for the aptamer Kirby).
From the supplementary Figs. 1 K and 1L (yellow letters), we can observe that the negative control sample, with the thick layer of lysozyme applied to it, did not yield a successful immunofluorescence microscopy result after anti-ovalbumin antibody hybridization.This outcome is logical, as anti-ovalbumin antibodies exclusively target ovalbumin protein and not lysozyme.

Immunofluorecence microscopy of model samples using aptamers
Model samples cooked in a suspension of egg white were hybridized with lysozyme targeting aptamers Clone1, Kirby and Apt1L; those cooked in lysozyme (extracted from egg white) or ovalbumin suspension were hybridized with Clone1 and Apt1L; those cooked in raw milk were hybridized with seqU5 (targeting casein) and BLG14 (targeting β-lactoglobulin); those cooked in a suspension of black pudding or haemoglobin were hybridized with haemoglobin targeting aptamers G15 T1, HA and Hb; those cooked in a suspension of wheat flour were hybridized with gliadin (gluten) targeting aptamers Gli4 and Gli1; those cooked in a suspension of fish traut were hybridized with Par1 (targeting parvalbumin), H2 (targeting histamin) and H47 (targeting histamin); those cooked in a suspension of gelatin (bovine skin) were hybridized with collagen targeting aptamers D1 and CTx 2R-2 h; and finally those cooked in a suspension of beef were hybridized with a myoglobin targeting aptamer Mb1.
After analysis, immunofluorescence microscopy was not successful in highlighting target proteins in gelatin model sample hybridised with CTx 2R-2 h; in egg white and lysozyme model samples hybridised with Apt1L; in ovalbumin model samples hybridised with Clone1 and Apt1L (Supplementary Fig. 1); in wheat flour model sample hybridised with Gli1; in traut fish model samples hybridised with H2 and H47; in beef model sample hybridised with Mb1; and in black pudding and haemoglobin model samples hybridised with G15 T1 and Hb (Supplementary Fig. 2).The negative results, which were obtained when ovalbumin model samples (made from pure ovalbumin standard) were hybridised with Clone1 and Apt1L, were expected, because these two aptamers only target the lysozyme protein and not ovalbumin.
Immunofluorescence microscopy was successful in the next cases: egg white model samples hybridised with Clone1 and Kirby; lysozyme model sample hybridised with Clone1; raw milk model samples hybridised with seqU5 and BLG14; black pudding and haemoglobin model samples hybridised with HA; wheat flour model sample hybridised with Gli4; fish traut model sample hybridised with Par1; and finally gelatin model sample hybridised with D1.In the below text, these successful cases are presented in detail.
Prior to hybridisation, the organic residuals, present within the thin most upper layer of ceramic, which was directly exposed to cooking in a suspension of egg white or lysozyme (Fig. 7A, E and I), exhibited a very weak autofluorescence (Fig. 7B, F and J).After the hybridisation of egg white or lysozyme samples with aptamers (Kirby or Clone1 on egg white; Clone1 on lysozyme), this layer (defined by white triangles in Fig. 7C, G and K) was highlighted with additional fluorescence, which enhanced its brightness under the widefield fluorescence www.nature.com/scientificreports/observation mode (this applies to egg white as well as to lysozyme model samples).These enhancements were stronger and even clearer in the confocal fluorescence observation mode, which excluded autofluorescence from microscopically small stones within the ceramic (Fig. 7H) as well as from the resin's surface reflection (Fig. 7D, H  and L).
Prior to hybridisation, the organic residuals, present within the thin most upper layer of ceramic, which was directly exposed to cooking in raw milk (Fig. 8A and E), exhibited almost no autofluorescence (Fig. 8B and F).After hybridisation with either BLG14 or seqU5 aptamers, this layer (defined by white triangles in Fig. 8C andG) was highlighted with fluorescence, which enhanced its brightness under the widefield fluorescence observation mode.These enhancements were even stronger and clearer in the confocal fluorescence observation mode, in which no material autofluorescence was observed (from resin's surface reflection and from small stones) (Fig. 8D  and H).
Prior to hybridisation, the organic residuals, present within the thin most upper layer of ceramic, which was directly exposed to cooking in a suspension of black pudding or haemoglobin (Fig. 9A and E), exhibited almost no autofluorescence (Fig. 9B and F).After the hybridisation of black pudding or haemoglobin samples with the aptamer HA, this layer (defined by white triangles in Fig. 9C and G) was highlighted with fluorescence, which enhanced its brightness under the widefield fluorescence observation mode (this applies to haemoglobin as well as to black pudding model samples, although the highlighting of the black pudding sample was very slight).These enhancements were stronger and clearer in the confocal fluorescence observation mode (Fig. 9H), although some unspecific fluorescence within this mode could still be observed (for example on the lower halve of ceramic surface of the black pudding model sample; Fig. 9D).
Prior to hybridisation, the organic residuals, present within the thin most upper layer of ceramic, which was directly exposed to cooking in a suspension of wheat flour or fish traut (Fig. 10A and E), exhibited a weak   Cross-section images of model samples, which were cooked in a suspension of black pudding or hemoglobin.They were photographed, prior to hybridisation with the hemoglobin (blood) targeting aptamer HA, using the reflected-light brightfield observation (frames A and E) and the widefield fluorescence observation (frames B and F) modes.After aptamer hybridisation, samples were again captured in the widefield fluorescence observation mode under the same illumination conditions (frames C and G) and were laser scanned using confocal fluorescence microscopy (frames D and H).Upper layers of ceramic, which were directly exposed to cooking, are marked with white triangles.autofluorescence (Fig. 10B and F); and those cooked in a suspension of gelatin (Fig. 10I) had absolutely no autofluorescence (Fig. 10J).
After Gli4 aptamer hybridisation, a thin layer of ceramic exposed to cooking in a suspension of wheat flour and a whitish non-ceramic layer above it were highlighted with additional fluorescence, which enhanced their brightness under the widefield fluorescence observation mode (Fig. 10C).After Par1 aptamer hybridisation, a thin layer of ceramic exposed to cooking in a suspension of fish trout, was highlighted with fluorescence (Fig. 10G).Interestingly, in the case of applying the D1 aptamer, a relatively thick layer of ceramic exposed to cooking in gelatin, was highlighted (Fig. 10K).
For all of the above presented model samples, this highlighting was clearer in the confocal fluorescence observation mode.Moreover, in the case of wheat flour, the nonspecific fluorescence of the upper whitish non-ceramic layer, which was observed under the widefield mode, was absent in the confocal mode (Fig. 10D).Nonetheless, unspecific fluorescence within the confocal mode could still be observed on the lower halves of ceramic surfaces of wheat flour and fish traut model samples (Figs.10H and L).
Immunofluorecence microscopy protocol consists of two washing steps and in most cases, this was sufficeint enough in removing unbound FITC conjugated streptavidin or FITC conjugated secondary antibodies from the polished cross section surfaces.Nevertheless, for two lysozyme model sampels, which were hybridised with eighter Apt1L aptamer (Supplementary Fig. 1F) or with anti-ovalbumin Ab (Supplementary Fig. 1L), these two washing steps were insufficient.Consequently, smaller and larger fluorescent stains were scattered around the entirety of these two cross-section images.

Discussion
Aptamers were assessed for their functionality (specificity) using the ELISA assay.Results indicated that out of the aptamers targeting egg lysozyme, two were functional (Clone1 and Kirby), while one was not (Apt1L).For the fish-related material, only aptamers targeting parvalbumin were functional (Par1), whereas those targeting histamine did not function (H2 and H47).Additionally, two aptamers targeting haemoglobin showed functionality (Hb and HA), while one did not (G15 T1).Lastly, all aptamers tested against gluten, milk, beef, and collagen demonstrated functionality (Gli1 and Gli4 for gluten; seqU5 and BLG14 for milk; Mb1 for beef; and D1 and CTx 2R-2 h for collagen).The ELISA aptamer validation tests showed strong correlation with the results obtained from using aptamers for immunofluorescence microscopy performed on ceramic model samples.Nonetheless, there were four discrepancies between the ELISA and immunofluorescence microscopy results.For example, immunofluorescence microscopy was not successful (after aptamer hybridisation the thin uppermost layer of ceramic, directly exposed to cooking in a protein suspension, did not show bright green fluorescence) when the haemoglobin model sample was hybridized with Hb, when the flour model sample was hybridized with Gli1, when the beef model sample was hybridized with Mb1, and when the gelatin model sample was hybridized with CTx 2R-2 h.Therefore, in all of these four cases immunofluorescence microscopy failed even though it was performed using functional aptamers.It is well acknowledged that in ELISA everything is dissolved in liquid and therefore biochemical interactions and reactions work well and just a few molecules of target protein may get detected 48,50,52 .Moreover, the ELISA was performed on freshly prepared protein suspensions, whereas immunofluorescence microscopy was performed on ceramic model samples, which were cooked in an oven for a period of 5 days, UV aged and subsequently buried.Therefore, target proteins may be degraded such that a given epitope may be preferentially lost 48 .Regarding specificity, all ELISA-tested aptamers showed no cross-reactivity with the BSA standard.Additionally, for immunofluorescence microscopy, the hybridization of egg lysozyme targeting Clone1 or Apt1L on ovalbumin model samples, as expected, did not yield a successful immunofluorescence microscopy result.
The functionality of the egg lysozyme targeting aptamer Kirby was compared with that of anti-ovalbumin (egg albumin) antibody, which is usually employed for immunofluorescence microscopy for cultural heritage binder analysis 51,53 .After immunofluorescence microscopy on the cross-sections of whole egg control samples, flourescence intensity was 3.1 times greater for the Kirby aptamer giving a much clearer picture.Typically, the ordered IgG antibodies (molecular weight ~ 150,000 g/mol) are present in a concentration range of 80-120 μM 54 , while our lyophilized aptamer pellet was diluted to 100 μM.Since both antibody and aptamer stocks were diluted by a factor of 10 before their application onto the cross-section, a similar molecular weight (of antibodies and aptamers) was employed for immunofluorescence microscopy.The greater functionality of aptamers observed here lies in their higher specificity, a result of strict selection during their production involving several SELEX selection cycles.Furthermore, aptamers exhibit high stability at room temperature, a long shelf-life with no batch-to-batch variability, and, compared to antibodies, their molecular weight is only ≈8 kDa, giving them excellent sample penetration 25 .
One thing to note is that during the hybridization of aptamer Kirby or anti-ovalbumin antibody onto the cross-sections of whole egg control samples, the thick egg layer significantly swelled up in size, elevating itself from the rest of the surface.This phenomenon typically occurs when freshly prepared, non-heated and unaged samples are analysed, such as our control samples.Proteins from older materials appear to be sufficiently denatured to become insoluble in aqueous solutions 55 .Moreover, the thicker the proteinaceous layer, the more water it absorbs.Lastly, swelling was not observed in our heated and aged model samples, of which only the thin uppermost layer of ceramic, directly exposed to cooking in a protein suspension, contained residual amounts of protein.
Prior to aptamer hybridisation, the thin uppermost layer of ceramic model samples containing residuals of either egg white, lysozyme, wheat flour or fish traut exhibited a weak autofluorescence.For other model samples no or almost no autofluorescence was observed.In general, proteins contain amino acids tryptophan, tyrosine, and phenylalanine and therefore show some degree of autofluorescence 56 .The weak autofluorescence could therefore reflect the residual presence of proteins within the porous structure of the thin upper ceramic layers of model samples.In any case, autofluorescence did not interfere with the immunofluorescence microscopy results.
The wheat flour model sample, prior to hybridization with Gli4 (Fig. 10B), showed autofluorescence from a whitish non-ceramic layer on top of the ceramic material.However, this non-ceramic layer was not wheat flour.After excavation, ceramic shards of all model samples were cleaned and rinsed, removing all visual traces of organic residuals (Supplementary Fig. 3).Therefore, this whitish layer resulted from the procedure of embedding the sample in resin, during which an air bubble became entrapped in the solid resin.Subsequently, during the cross-section polishing procedure, a paste of fine solid particles, primarily composed of resin and ceramic, accumulated within this bubble.Most of the paste was removed by eliminating the bubble from the cross-section with further polishing and the use of a needle; however, small fractions still remained.
In comparison to the widefield fluorescence observation mode, the confocal observation mode achieved a much more specific and clear immunofluorescence microscopy picture.This was because in the confocal observation mode nearly all of the unspecific fluorescence originating from micro sample's autofluorescence (Figs.10C  and D; wheat flour with Gli4); from microscopically small stones within the ceramic (Fig. 7G and H; egg white with Clone1) or from resin's surface reflection was successfully removed.Nevertheless, the unspecific fluorescence, originating from the lower halves of ceramic surfaces of wheat flour and fish traut model samples, could not have been avoided (Figs.10D and H).The, confocal observation mode uses a monochromatic laser source to scan across a defined sample area (in z-axis), while the use of a very small aperture (pinhole) in the optical path allows to detect light emitted within the focal plane at different sample depths and, at the same time, to discard the out-of-focus light 52,57 .In this way it is possible to obtain optical sections of the sample and reconstruct its 3D structure from the detected light independent of the position of the scanning spot 50 .This allows for the confocal observation mode to be unaffected by various reflection angles, which form when a complex surface structure, such as a gutter/crack, is exposed to excitation light.
The most capable method for residual protein analysis within ceramics is high-throughput whole protein sequencing using LC-MS/MS proteomics 4,7,[28][29][30][31]58 . Firtly, proteins need to be extracted from samples requiring at least 10 µg of protein 32 .Regarding the information obtained, immunological methods can hardly compete with proteomics, simply because the obtained amino acid sequences give exact and species-specific results of the food source material for every encountered protein.In immunology, only proteins in question are analysed and not the entire spectrum of extracted proteins (could also be species-specific).Nevertheless, regarding the bottleneck process that is protein extraction (extraction yields of ≤ 0.1% 59 ), ELISA, being more sensitive (in theory, it may detect a single positive antigen-antibody binding 60 ), requires much lower protein yields compared to proteomics.Moreover, in proteomics, proteins need to be isolated from the extraction buffer (using dialysis or methanol/chloroform procedures), because detergents and urea used in the extraction hinder the very sensitive detection of LC-MS/MS 28 .However, in ELISA, such purifications are unneeded, which means that the sample preparation is faster, retaining more protein for the final analysis 19 . Latly, in immunofluorescence microscopy, analysis is conducted directly on the sample's cross-section, meaning that protein extraction/purification is entirely excluded 51,53 .Additionally, valuable stratigraphic localization of food proteins on the cross-section can be achieved.This stratigraphic information is lost in proteomics, because samples need to be crushed into ceramic powder prior to protein extraction and purification.

Conclusions
Our experimental study represents the first application of aptamers as detection molecules in cultural heritage science, aimed at analysing protein residues bound to ceramics.We have successfully validated eight out of sixteen tested aptamers on shards of traditionally prepared, aged, buried (for a year), and excavated archaeological ceramic replicas.These aptamers are designed to detect protein targets originating from various sources, including egg residuals (Clone1 and Kirby), cereal residuals (Gli4), milk residuals (seqU5 and BLG14), blood residuals (HA), fish residuals (Par1), and bone broth residuals (D1).Finally, our findings reveal that the specific fluorescence intensities of the aptamer assembly are three times greater than those of standardly employed antibodies, underscoring the superiority of aptamers in this application.Nonetheless, it has to be stated that this is an experimental study conducted exclusively on heated and aged model samples and therefore, for future experimentation, real archaeological case-study samples (at least 100 years old) will have to be included.

42 Fish[ 8 44 H47[ 46 C
°C) myoglobin from Sigma-Aldrich Mb1 [Btn]CCC TCC TTT CCT TCG ACG TAG ATC TGC TGC GTT GTT CCGA 85.5 Boiled (for 30 min) fish parvalbumin extract Par1 [Btn]TTT TTT TTG CCA AAG GAG GCG AGA GAT AAA AGA TTG CGA ATC CAT TCG 83Btn]AGC TCC AGA AGA TAA ATT ACA GGG AAC GTG TTG GTT GCG GTT CTTCC GAT CTG CTG TGT TCT CTA TCT GTG CCA TGC AAC TAG GAT ACT ATG ACC CCGG 92.Btn]GCC TGT TGT GAG CCT CCT AAC ATT TCT ATG CTG CAG CCA ACT TTT CCA T ACT TCC AGC TTA CCA TTT ATC CAT GCT TAT of a human collagen type XI D1 [Btn]TTT TTG GTT GAC GGC AGT CGG CGG TAT GCG CAT ATC GTG TTG GTA 76.1 -telopeptide of human type I collagen CTx 2R [Btn]ATC CGT CAC ACC TGC TCT AGA CGA ATA TTG TAT CCT CAT TAG ATC AAA AAC GGG TGG TGT TGG CTC CCG TAT https://doi.org/10.1038/s41598-024-70048-8www.nature.com/scientificreports/Fluorescence intensity calculation Fluorescence intensities (FI) were quantified from whole egg control samples images obtained using widefield fluorescent microscopy.Firstly, the .tifformat was opened, within the ImageJ distribution Fiji, and the fluorescence intensity of the thick upper organic proteinous layer (defined by white triangles; see Figs.5 and 6) was quantified, by marking the region of this layer using the »Polygon selections tool« and by selecting the »Measure« function (preselect »Area« and »Mean Gray Value« under the »Set measurements« options), within the »Analyse« menu.

Figure 3 .
Figure 3. Folded secondary structures of aptamers, reconstructed within the OligoAnalyzer™ Online Tool, which target hemoglobin (black pudding); parvalbumin and histamine within fish; and collagen (bone broth).

Figure 5 .
Figure 5. Cross-section images of the control sample, for which a chicken egg (whole egg) mix was directly applied onto the ceramic shard.It was photographed prior to anti-ovalbumin antibody (Ab) hybridisation using the reflected-light brightfield observation (frame A) and the widefield fluorescence observation (frame B) modes.After antibody hybridisation, the sample was again captured in the widefield observation mode under the same illumination conditions (frame C) and was laser scanned using confocal fluorescence microscopy (frame D) and surface topography scans (frame E).Fluorescence intensities of sample layers in frames B and C are marked with white, horizontally directed triangles.

Figure 6 .
Figure 6.Cross-section images of the micro sample, for which a chicken egg (whole egg) mix was directly applied onto the ceramic shard.It was photographed, prior to hybridisation with the lysozyme (egg) targeting aptamer Kirby, using the reflected-light brightfield observation (frame A) and the widefield fluorescence observation (frame B) modes.After aptamer hybridisation, the sample was again captured in the widefield observation mode under the same illumination conditions (frame C) and was laser scanned using confocal fluorescence microscopy (frame D) and surface topography scans (frame E).Fluorescence intensities of sample layers in frames B and C are marked with white, horizontally directed triangles.

Figure 7 .
Figure 7. Cross-section images of model samples, which were cooked in a suspension of egg white or lysozyme (extracted from egg white).They were photographed, prior to hybridisation with lysozyme targeting aptamers Kirby (frames C and D) or Clone1 (frames G, H, K and L), using the reflected-light brightfield observation (frames A, E and I) and the widefield fluorescence observation (frames B, F and J) modes.After aptamer hybridisations, samples were again captured in the widefield fluorescence observation mode under the same illumination conditions (frames C, G and K) and were laser scanned using confocal fluorescence microscopy (frames D, H and L).Upper layers of ceramic, which were directly exposed to cooking, are marked with white triangles.

Figure 8 .
Figure 8. Cross-section images of model samples, which were cooked in Raw milk.They were photographed, prior to hybridisation with the β-lactoglobulin (milk) targeting aptamer BLG14 (frames C and D) or the β-casomorphin-7 (milk) targeting aptamer seqU5 (frames G and H), using the reflected-light brightfield observation (frames A and E) and the widefield fluorescence observation (frames B and F) modes.After aptamer hybridisations, samples were again captured in the widefield fluorescence observation mode under the same illumination conditions (frames C and G) and were laser scanned using confocal fluorescence microscopy (frames D and H).Upper layers of ceramic, which were directly exposed to cooking, are marked with white triangles.

Figure 9 .
Figure 9. Cross-section images of model samples, which were cooked in a suspension of black pudding or hemoglobin.They were photographed, prior to hybridisation with the hemoglobin (blood) targeting aptamer HA, using the reflected-light brightfield observation (frames A and E) and the widefield fluorescence observation (frames B and F) modes.After aptamer hybridisation, samples were again captured in the widefield fluorescence observation mode under the same illumination conditions (frames C and G) and were laser scanned using confocal fluorescence microscopy (frames D and H).Upper layers of ceramic, which were directly exposed to cooking, are marked with white triangles.

Figure 10 .
Figure 10.Cross-section images of model samples, which were cooked in a suspension of wheat flour, fish traut or gelatin (bovine skin).They were photographed, prior to hybridisation with the gliadin targeting aptamer Gli4 (frames C and D) or the parvalbumin (fish) targeting aptamer Par1 (frames G and H) or with the gelatin aptamer D1 (frames K and L), using the reflected-light brightfield observation (frames A, E and I) and the widefield fluorescence observation (frames B, F and J) modes.After aptamer hybridisations, samples were again captured in the widefield fluorescence observation mode under the same illumination conditions (frames C, G and K) and were laser scanned using confocal fluorescence microscopy (frames D, H and L).Upper layers of ceramic, which were directly exposed to cooking, are marked with white triangles.

Table 1 .
Sixteen DNA sequences of aptamers employed within our study, for targeting proteins within eggs, cereals, milk, blood, red meat, fish and bone.

Target protein Name 5'-3' end DNA single strand sequence modified with 5' end biotinylation [Btn] T m [°C] Source of aptamer sequence in literature
40 Hb [Btn]GGC AGG AAG ACA AAC ACC AGG TGA GGG AGA CGA CGC GAG TGT T AGA TGG TAG CTG TTG GTC TGT GGT GCTGT 93.8