Spatial mapping of polymicrobial communities reveals a precise biogeography associated with human dental caries

Significance Dental caries remains an unresolved public health problem. The etiology is poorly understood, as the oral cavity harbors diverse communities of microorganisms. Using multiple imaging modalities on human teeth from patients with caries, we discovered a microbial community precisely arranged in a corona-like architecture. Moreover, this organization is mediated by the pathogen Streptococcus mutans through production of an extracellular scaffold that directs positioning of other oral microbes. We developed a methodology to quantify the spatial structure of microbial communities at the micron scale and found a precise spatial patterning of bacteria associated with localized caries onset. These findings are relevant as we approach the post-microbiome era, whereby quantifying the community structural organization may be essential for understanding microbiome function.


SI Figures
. Bacterial FISH probe validation and computational image processing. (A) Differentiation of bacterial group via taxa-specific labelling in planktonic cells. Bacterial groups were differentiated by a specific coding based on the classification of fluorescent labelling. Taxaspecific labelling was capable of separating S. mutans, non-mutans streptococci (NSMU) and non-streptococcal bacteria (NSTR). S. mutans was labelled by EUB, STR and SMU, S. oralis was labelled by EUB and STR, and A. naeslundii was labelled by EUB only. This method was able to exclude the nonspecific labelling and autofluorescence signals of clinical sample. (B) A fluorescence subtraction method for the configuration of polymicrobial community organization in intact plaque biofilm. For cell arrangement configuration (positioning of S. mutans across the microbial consortium) within the polymicrobial biofilm structure, fluorescence subtraction was applied through classification of a set of elements using Image Calculator of ImageJ: SMU, S. mutans alone; Streptococcus (STR) -SMU = non-mutans streptococci (NSMU); All bacteria (EUB) -STR = non-streptococcal bacteria (NSTR). (C) Application of taxa-specific labelling and fluorescence subtraction methods in planktonic cell mixture. To further demonstrate and validate this approach, we prepared planktonic cells of S. mutans (at OD600 1.0; 2×10 9 CFU ml -1 ), S. oralis (at OD600 1.0; 8×10 8 CFU ml -1 ) and A. naeslundii (at OD600 1.5; 8×10 8 CFU ml -1 ), each of which was labelled with taxa-specific probes and acquired image was subjected to fluorescence subtraction method. Top images in the panel C show the taxa-specific labelling (e.g., S. mutans was labelled in SMU, STR, EUB). Bottom images of panel C show how fluorescence subtraction can separate non-mutans streptococci (i.e., S. oralis, depicted in red) and non-streptococcal bacteria (i.e., A. naeslundii, depicted in blue) from planktonic cell mixture via taxa-specific labelling. In the cell mixture at equal proportion of S. mutans, S. oralis and A. naeslundii (1:1:1 ratio for each of the bacterial suspension at ~10 8 CFU ml -1 ), subtraction of SMU from STR is nonmutans streptococci (since both S. mutans and S. oralis were labelled by STR and EUB but S. oralis was not labelled by SMU). Next, subtraction of STR from EUB is non-streptococcal bacteria (A. naeslundii was labelled by EUB only while Streptococcus-genus (i.e., S. mutans and S. oralis) was labelled by STR and EUB). (D) Application of taxa-specific labelling and fluorescence subtraction methods in a mixed-species biofilm. We also prepared a mixed-species in vitro biofilm model to validate the fluorescence subtraction methods. Each of the bacterial suspension was mixed to provide an inoculum with a defined microbial population of S. mutans (10 5 CFU ml -1 ), S. oralis (10 7 CFU ml -1 ) and A. naeslundii (10 6 CFU ml -1 ). The mixed population was inoculated in 2.8 ml of medium containing 1% (w/v) sucrose to form the biofilm community on the apatitic surface. Taxa-specific labelling and fluorescence subtraction were applied as described above. Details of the methodology are provided in SI Methods.  The same enamel blocks were placed under the confocal Raman microscope and the positioning adjusted using the fiduciary marks to match the location of the biofilm structures and the demineralized areas. The distribution and content of phosphate minerals in demineralized lesions were analyzed using Raman spectroscopy as a non-destructive technique. Quantitative data were generated from the dotted-line box areas with 100 measuring points (100×100 µm 2 ). (C) Demineralized lesions associated with the biofilm rotund architecture. Red arrow heads indicate the enamel underneath rotund architecture while white arrow heads indicate the enamel covered by flat communities. Details of the methodology are provided in SI Methods.    Bacterial growth curve of S. mutans wild type (WT) and gtfB mutant strains. Each bacterium was grown in ultrafiltered tryptone-yeast extract broth (UFTYE; 2.5% tryptone and 1.5% yeast extract, pH 7.0) with 1% glucose at 37°C and 5% CO2. (C) Lack of GtfB-EPS in mixed-species biofilm cocultured with S. mutans gtfBΔ resulted in intermixing pattern while segregated cell arrangement was re-established in mixed-species biofilm through GtfB supplementation. (D) Quantitative analysis for bacterial cell arrangement in the mixed biofilm. In S. mutans WT + S. oralis mixed biofilm, the proportional occupancy (PO) of S. oralis increased along with distance when S. mutans was used as a focal point (non-random distribution associated with corona-like arrangement). However, PO did not change across distance in the gtfB mutant + S. oralis mixed biofilm indicating random distribution. In the gtfB mutant + S. oralis mixed biofilm supplemented with GtfB, PO of S. oralis showed a similar pattern to that of mixed biofilm with S. mutans WT, indicating reestablishment of corona-like arrangement.  Data are mean ± SD (n = 4). The data were subjected to analysis of variance (ANOVA) in the Tukey's HSD test for a multiple comparison. Differences between groups were considered statistically significant when P < 0.01. No antibacterial effects were observed with PI at concentrations between 0.125% and 0.5%.

SI Methods
In this section, we provided step-by-step protocols for each of the imaging methods: I) Intact human plaque biofilm imaging, II) Synchronized imaging of biofilm structure, pH and enamel surface, III) Simultaneous analysis of in situ gene expression and pH within biofilm Step-by-Step Protocols

I.
Intact human plaque biofilm imaging Brief description: Imaging of undisturbed spatial structure in biofilm remains challenging as it requires sample collection from clinical sites and further processing that disrupts the original architecture. We have applied an imaging methodology to map the spatial organization of biofilm communities in its native state on extracted teeth from patients affected by severe childhood caries. We used taxa-specific fluorescent probes including S. mutans-specific, Streptococcusand all bacteria-probes with confocal imaging across multiple length scales (from submillimeter to submicron level). Then, we used a fluorescence subtraction method (details in Tutorial Guide 1) to analyze the spatial arrangement and composition across different phylogenetic scale, i.e. S. mutans (SMU), non-mutans streptococci (NSMU), and non-streptococcal bacteria (NSTR). This multi-scale approach allowed us to study both the overall organization and spatial arrangement of the intact biofilm communities (SI methods Fig. 1). Fig. 1. Graphical summary of the multi-scale imaging for intact biofilm on tooth surface.

Procedure:
1. Tooth samples were extracted without perturbing the naturally formed biofilms (see details in Materials and Methods), transferred to a dish (ø35×10 mm) containing sterile, PBS-soaked gauze placed at the bottom to hold the extracted teeth under wet condition (without disturbing the biofilm structure), and immediately transferred to the lab. 2. The biofilm on tooth surface was gently washed twice with PBS (pH 7.4) at room temperature, and fixed with 4% paraformaldehyde (in PBS, pH 7.4) at 4°C for 4 h.
3. After fixation, the specimen was washed twice with PBS, then transferred into 50% EtOH in PBS (pH 7.4), and stored at -20˚C. 4. For fluorescence in situ hybridization (FISH), the specimen was treated with lysis buffer containing 10 mg ml -1 lysozyme (in 20 mM Tris-HCl pH 7.5, 5 mM EDTA) for 14 min at 37°C to enhance cell permeability to the FISH probes. 5. The specimen was incubated in the hybridization solution (25% formamide, 0.9 M NaCl, 0.01% SDS, 20 mM Tris-HCl, pH 7.5) containing FISH oligonucleotide probes (MUT590, 5ʹ-ACTCCAGACTTTCCTGAC-3ʹ with Alexa Fluor 488 for Streptococcus mutans; STR405, 5ʹ-TAGCCGTCCCTTTCTGGT-3ʹ with Cy5 for Streptococcus; EUB338, 5ʹ-GCTGCCTCCCGTAGGAGT-3ʹ with Cy3 for all bacteria at a final concentration of 1 µM) at 46°C for 4 h, and then washed using washing buffer (0.2 M NaCl, 20 mM Tris-HCl pH 7.5, 5 mM EDTA, 0.01% SDS) at 46°C for 15 min. 6. After in situ hybridization, the specimen was immobilized on a small petri dish (ø35×10 mm) using a dental wax (red sprue wax) (which allowed adjustment of the position and the angle of the tooth surface to be in a parallel plane in relation to the confocal microscope objective). 7. Biofilm images were acquired using confocal microscopy (LSM 800 (Zeiss) equipped with a 20× (1.0 numerical aperture (NA)) water immersion objective. The field of view was adjusted using fine focus knob under green fluorescence or refection mode. 8. The biofilms were sequentially scanned using Diode lasers (488, 561 and 640 nm), and the fluorescence emitted was collected with the GaAsP or multialkali PMT detector (490-550 nm for Alexa Fluor 488 (S. mutans), 565-620 nm for Cy3 (all bacteria), and 645-700 nm for Cy5 (Streptococcus). 9. Initially, low magnification (0.5×zoom) (submillimeter-scale: 0.624×0.624 (x,y) mm 2 ) images were acquired. Then, high magnification (1×zoom, 312×312 µm 2 ; 2×zoom, 156×156 µm 2 ; 4×zoom, 78×78 µm 2 ) images were further acquired for biofilm architecture (micron-scale). 10. For quantitative analysis, we selected each of the architectures in a field of view at 312x312 µm 2 , and optical sectioning with a z-stack size of 0.92 µm. 11. The confocal images were analyzed using COMSTAT (available as free download at http://www.imageanalysis.dk), written as scripts for MATLAB software to calculate the biovolume of each architecture (details in refs (1, 2)). 12. Total biovolume was measured from EUB, STR and SMU-labelled cells. The biovolume of EUB and SMU-labelled cells were further used for measuring a relative ratio of SMU to EUB. 13. Following overall biovolume quantification, we employed a fluorescence subtraction method to analyze the spatial arrangement and composition across different phylogenetic scale, i.e. S. mutans (SMU), non-mutans streptococci (NSMU), and non-streptococcal bacteria (NSTR) (see Tutorial Guide 1). 14. After confocal imaging, the same specimen was used to acquire electron micrographs via scanning electron microscopy (SEM). 15. For SEM imaging, the biofilm sample was treated with 2% paraformaldeyhyde/2% glutaraldehyde for 18 h. The sample was gently submerged into the fixative solution, and subsequently washed with PBS. 16. Then, the specimen was serially treated with different concentrations of EtOH (in serial dehydration steps; 50→70→80→90→100%). The sample was kept at each concentration of EtOH for 10 min. 17. Next, the specimen was dried using hexamethyldisilazane (HMDS) (serially treated with EtOH:HMDS at 1:1 ratio, then at 1:4 ratio, and finally exposed to 100% HMDS). The sample was further dried in the fume-hood for 10 min. 18. The dried specimen was mounted on a SEM aluminum-holder with super-glue and painted with silver-liquid at the interface between teeth and holder surface. 19. The specimen was subjected to sputter coating (Au/Pd). 20. The image was acquired using a high-resolution scanning electron microscope (SEM; Quanta 250 FEG eSEM, FEI).

II. Synchronized imaging of biofilm structure, pH and enamel surface Brief description:
We developed a sequential multi-step method for synchronized biofilm and enamel surface imaging and realignment for structural and functional assessment related to spatial localization of caries. To perform this method, we used our in vitro mixed-species biofilm model in which S. mutans (pathogen) and S. oralis (commensal) were inoculated on natural human tooth-enamel and allowed to form a biofilm in the presence of sucrose. Following biofilm formation, we employed a hybrid confocal-stereoscope system using tiled image acquisition to encompass the entire biofilm and enamel surface. This approach allowed optical alignment of the biofilm structure formed on the entire enamel block surface with 10 µm length-scale precision. The structural organization was assessed via multi-labelling approaches using species-specific fluorescent probes and EPS glucan matrix labelling. The pH at the biofilm-enamel interface was visualized and measured using a fluorescent pH mapping method and ratiometric analysis as detailed previously (2); the link for the step-by-step protocol can be found here (https://doi.org/10.1371/journal.ppat.1002623.s006). The enamel surface was assessed for demineralized areas with optical and fluorescence imaging complemented by quantitative transverse microradiography (TMR). This procedure is summarized in SI Methods Figure 2A, and used to match the biofilm architectural features with pH mapping and enamel demineralization. Fig. 2. Synchronized biofilm-surface analysis using in situ human enamel model and multi-step imaging.

SI Methods
1. Human enamel blocks were prepared to have a uniform size (4x4 mm 2 ) with distinctive edges as fiduciary marks (SI Methods Fig. 2B) to facilitate optical alignment. 2. Prior to bacterial inoculation for biofilm growth, wide-field image of the entire enamel surface was acquired using a stereomicroscope (Axis Zoom V16, Zeiss) equipped with a 1× objective (PlanNeoFluar Z, Zeiss) under bright field and green fluorescence (see Tutorial Guide 2). 3. Then, a mixed-species biofilm was prepared using saliva-coated human enamel blocks (SI Methods Fig. 2A) using the cariogenic pathogen S. mutans and the commensal streptococci S. oralis. 4. For inoculum preparation, each bacterium was grown in ultrafiltered tryptone-yeast extract broth (UFTYE; 2.5% tryptone and 1.5% yeast extract, pH 7.0) with 1% glucose at 37°C and 5% CO2 to OD600 1.0 (exponential growth phase). 5. Each of the bacterial suspension was mixed to provide an inoculum with a defined microbial population of S. mutans (10 5 CFU ml -1 ) and S. oralis (10 7 CFU ml -1 ). 6. The mixed population was inoculated in 1 ml of UFTYE containing 0.1% (w/v) sucrose and incubated for 19 h to form an initial biofilm community on the enamel surface. 7. Then, the biofilms formed on enamel blocks were transferred to UFTYE containing 1% sucrose to induce environmental changes to simulate a cariogenic challenge at 19 h. The culture medium was changed twice daily until the end of the experimental period (91 h). 8. For biofilm structure, EPS was labelled with 1 μM dextran-conjugated Alexa Fluor 488 and bacterial cells were labelled with 10 μM Syto60 (total) or species-specific FISH probes (MUT590, 5ʹ-ACTCCAGACTTTCCTGAC-3ʹ with Cy5 for S. mutans; MIT588, 5ʹ-ACAGCCTTTAACTTCAGACTTATCTAA-3ʹ with Cy3 for S. oralis each probe at a final concentration of 1 µM) as described in Materials & Methods and in step-by-step protocol I. For pH mapping, 1 μM of Lysosensor yellow/blue dextran conjugate was used as detailed previously (2). 9. After biofilm formation, the wide-field images of entire biofilm formed the enamel block (4x4 mm 2 ) were acquired using stereomicroscope (Axis Zoom V16, Zeiss) for optical alignment of biofilm-surface (Tutorial Guide 2). Then, high resolution images were acquired using confocal microscopy (LSM 800 (Zeiss) or TCS SP8 (Leica)) equipped with a 20× (1.0 numerical aperture (NA)) water immersion objective. 10. For biofilm structure, the biofilms were sequentially scanned using Diode lasers (488, 561 and 640 nm), and the fluorescence emitted was collected with the GaAsP or multialkali PMT detector (490-550 nm for Alexa Fluor 488 (EPS), 565-620 nm for Cy3 (S. oralis), and 645-700 nm for Cy5 (S. mutans). 11. Tile acquisition mode (entire surface 4.396×4.396 mm 2 ) was applied for entire biofilm imaging. Then, image stacks of the area of interest were acquired with optical zoom (1×zoom, 312x312 µm 2 ; 2×zoom, 156x156 µm 2 ; 4×zoom, 78x78 µm 2 ). 12. For pH mapping, the biofilm was scanned using two-photon confocal microscopy for dual emission (450 and 520 nm).The pH values within intact biofilms were measured based on fluorescence intensity ratios of the dual-wavelength Lysosensor fluorophore using the titration curves of ratios versus pH (ranging from 4.0 to 7.0) as described in a published protocol (https://doi.org/10.1371/journal.ppat.1002623.s006). 13. Zen Blue (Zeiss) and Amira 5.4.1 (Visage Imaging) software were used to create 3D renderings to visualize the overall architecture of the biofilms. 14. After biofilm imaging using the hybrid confocal-stereomicroscopy system, biomass was removed with an enzymatic treatment (mixture of 8.75 units of dextranase and 1.75 units of mutanase) at 37°C for 2 h followed by water-bath sonication (for 4 min). 15. The cleaned enamel surface was used for optical and fluorescence imaging via stereomicroscope (Zeiss). 16. Optically defined edges of each enamel blocks (which have distinctive edges as fiduciary marks) were used for image alignment.
17. All acquired images (before and after biofilm removal) were realigned using the edge of enamel blocks. To match community structures (e.g., rotund) and demineralized enamel lesions underneath the biofilm, acquired images were aligned using a grid box (See Tutorial Guide 2). 18. After image realignment, lesion depth of the demineralized enamel regions was determined using transversal microradiography (TMR). 19. For TMR, the enamel blocks were cross-sectioned using a hard tissue microtome (Scientific Fabrications Laboratories). 20. Prior to sectioning, the region of interest was selected based on the realigned biofilmsurface images (SI Methods Fig. 2C). For example, the entire cross-section of the enamel block encompassing both demineralized and non-demineralized areas was selected (see Tutorial Guide 2, SI Methods Fig. 14) 21. The enamel block was mounted on the microtome holder and sectioned transversally across the entire length of the enamel block to obtain a 100-µm thick section and section with a hard tissue microtome. 22. The 100-µm section was mounted on X-ray sensitive plates (Microchrome Technology) along with an aluminum calibration step wedge. The plates were developed according to the manufacturer's instructions. 23. Section was subjected to Ni-filtered Cu-Kα radiation (X-ray; Philips Electronic Instruments) at 30 mA and 20 kV for 65 min. 24. Microradiographic images were analyzed with dedicated software (TMR 2000, Inspektor) with sound enamel defined at 87% mineral volume to obtain mean lesion depth (μm).

II.
Simultaneous in situ gene expression and pH mapping Brief description: We developed a method for simultaneous pH mapping and in situ atpB expression (a key gene associated with acid tolerance and fitness) across the biofilm structure. To achieve this, we used an atpB-green fluorescent promoter (GFP) S. mutans construct combined with fluorescence pH mapping approach as detailed previously (3). The atpB promoter activity was measured within intact (control, buffer-treated) and disrupted (dextranase treated) corona structure (SI Methods Figure 3). Fig. 3. Simultaneous in situ gene expression and pH mapping of intact corona and corona-disrupted biofilm.

SI Methods
1. An atpB promoter-GFP construct (PatpB::gfp) of S. mutans UA159 was used for in situ gene expression (details of the construct can be found in (3)). 2. The mixed-species biofilms were formed on saliva-coated hydroxyapatite discs (surface area, 2.7 ± 0.2 cm 2 ; Clarkson Chromatography Inc.) as described in the step-by-step protocol II. 3. Each disc was inoculated with S. mutans (10 5 CFU ml -1 ) and S. oralis (10 7 CFU ml -1 ), and then incubated with Lysosensor yellow/blue dextran conjugate for pH mapping. The pH values within biofilms were measured based on fluorescence intensity ratios of the dualwavelength Lysosensor fluorophore (see below). 4. The atpB promoter activity of S. mutans across biofilm architecture was measured at specific locations (i.e. in close proximity of S. oralis layer) within intact or disrupted corona structure. Disruption of the S. oralis corona can be achieved with dextranase treatment without disturbing the structural integrity of the S. mutans inner core and its cell viability (SI Methods Fig. 3A; details in procedure #9), Hence, we compared the atpB expression in intact (enzyme buffer-treated control) vs. disrupted corona (dextranasetreated) 5. The biofilm formed on sHA with intact corona-like cell arrangement (buffer-treated control) was incubated in Na2HPO4-citric acid buffer (pH 5.0) without dextranase at 37˚C for 60 min. After buffer treatment (at pH 5.0), biofilms were gently dip-washed three times and incubated in the acidic pH buffer, and the confocal images acquired after 10, 30 and 60 min of incubation. Then, the biofilm was incubated in the neutral pH buffer (pH 7.0) and the confocal images acquired after 10, 30 and 60 min of incubation. 6. The confocal images of Lysosensor-incorporated biofilms were acquired using a multiphoton laser scanning microscope (TCS SP8, Leica) equipped with a 20× (1.0 NA) water immersion lens. The biofilms were excited at 700 nm, and emission was detected in two channels: one at 450 nm (using non-descanned detector NDD1; −495 nm) and the other at 521 nm (using non-descanned detector NDD2; 495-560 nm). Bacterial cells were stained with SYTO60 and scanned using 633 nm He-Ne laser. 7. For pH measurement, the ratios of fluorescence intensity of selected areas within each biofilm image were converted to pH value. The fluorophore exhibits a dual-emission spectral peak (fluorescence emission maxima 452 nm and 521 nm), and the ratio between the fluorescence intensity of these two spectral peak is pH-dependent within biofilms as described in a published protocol (https://doi.org/10.1371/journal.ppat.1002623.s006). For visualization of pH distribution in the biofilm, fluorescence intensity ratios (corresponding to the pH values between 7.0 and 4.0) of all confocal images were reconstructed using Image J, then Amira. The fluorescence intensity was converted into grayscale using the Amira tool-box to correlate with the pH range from 7.0 (white) to 4.0 (black). 8. For measurement of atpB promotor activity (GFP expression) with intact corona structure, biofilm was sequentially scanned using the 488 nm Argon laser to minimize the crosstalk between GFP (515-545 nm) and Lysosensor (no signal emitted), and the fluorescence emitted was collected with the internal spectral detectors (515-545 nm). For visualization of gene expression level, we applied ImageJ's lookup table (LUT) Fire (Fire LUT) as described previously (3). Fire LUT was used for in situ atpB expression, while green and red were used for S. mutans and S. oralis respectively (SI Methods Fig. 3B and C). 9. The same imaging procedures were employed for corona-disrupted biofilms using dextranase. Briefly, biofilms were treated with 100 units of dextranase in Na2HPO4-citric acid buffer (pH 5.0) at 37˚C for 60 min (SI Methods Fig. 3A). After dextranase treatment (at pH 5.0 of buffer), biofilms were gently dip-washed three times and incubated in the acidic pH buffer, and the confocal images acquired after 10, 30 and 60 min of incubation. Then, the biofilm was incubated in the neutral pH buffer (pH 7.0) and the confocal images acquired after 10, 30 and 60 min of incubation. In situ pH and atpB promoter activity within the biofilms were simultaneously imaged as described above (procedure #7-9). 10. Quantification of signal intensity of selected area of interest was measured using ImageJ as follows; Analyze→Tool→ROI Manager. For comparison between intact corona and corona-disrupted, similar size (diameter and thickness) of the biofilm structures were selected. 11. To visualize the S. oralis corona cell arrangement, fluorescence subtraction method was applied as follows: total bacteria stained by SYTO60 image stack -S. mutans GFP image stack = S. oralis. Fig. 4. Differentiation of bacterial group via taxa-specific labelling in planktonic cells.  Figure 4 shows the taxa-specific labelling of each bacterial cell. S. mutans is labelled by EUB, STR and SMU, S. oralis is labelled by EUB and STR, and A. naeslundii is labelled by EUB only (SI Methods Fig. 4).

1.2.
Fluorescence subtraction 1 Acquire image with similar signal intensity (signal intensity must be checked using confocal microscope software or ImageJ). Fig. 5. Taxa-specific labelling of planktonic cell mixture.
Step 2: Check the signal intensity of each image stack using ImageJ; Analyze→Tool→ROI Manager. *During the confocal imaging acquisition, signal intensity of each channel is monitored by Zen Blue software. 1.2.1.8.

2.2.
Transversal microradiography (TMR): Measurement of demineralized lesion depth 2.2.1. All acquired images (before and after biofilm removal) are realigned using the edge of enamel blocks. To match community structures (e.g., rotund) and demineralized enamel lesions underneath the biofilm, acquired images are aligned using a grid box (described in section 2.1). 2.2.2. After image realignment, lesion depth of the demineralized enamel regions is determined using TMR as described below: 2.2.3.
Step 1: Select the area of interest (after biofilm-surface image realignment) (highlighted in green-colored box, SI Methods Fig. 14).

2.2.4.
Step 2: Map the location (y-axis section) of the area of interest in a grid box (4.7 mm× 4.7 mm; 1 mm major gridline and 0.1 mm minor gridline). 2.2.5. For example, the entire cross-section of the enamel block encompassing both demineralized and non-demineralized areas is selected (a dotted-line box in SI Appendix Fig. 14). 2.2.6. Step 3: Mount specimen on a plastic rod and section with a hard tissue microtome (cut from 200-300 µm from left and right sides of the area of interest).

2.2.7.
Step 4: A 100-µm section (encompassing the area of interest) is mounted on X-ray sensitive plates (Microchrome Technology) along with an aluminum calibration step wedge. The plates are developed according to the manufacturer's instructions. 2.2.8.
Step 5: The section is subjected to Ni-filtered Cu-Kα radiation (X-ray; Philips Electronic Instruments) at 30 mA and 20 kV for 65 min. 2.2.9.