Dry seeds of B. distachyon accession Bd21-3 were imbibed and stratified in a wet paper towel at 6 °C for seven days. Seeds were then sown in 10 cm pots containing potting mix (#2; Conrad Fafard Inc., Agawam, MA, USA). Growth chamber temperature was maintained at 20 °C with 20 h light:4 h dark cycles at a fluence rate of 220 µmol of photons^.m^−2.s⁻¹ and relative humidity of 67 to 69. For further histochemical analysis, the first internode of the tallest stem was removed from plants of three different stages. The first stage, elongation, corresponded to when the first internode above the crown of the tallest stem was elongating. The second stage, inflorescence emergence, was when the first internode was completely elongated and the inflorescence began to emerge from the flag leaf. The third stage, senescence, was when the first internode had senesced and the stem reached its maximum height with all of the leaves showing signs of senescence.

Tiller count (n  =  20) was recorded as the number of stems per plant and height (n  =  25) as the tallest tiller per plant to the tip of the inflorescence of fully senesced plants. Biomass accumulation was quantified as the fresh weight (n  =  25) of the total above ground biomass. Internode length (n  =  13-21) was determined by removing the leaf sheath and imaging the tissue using a stereo dissecting Leica MZ16 F microscope (Leica Microsystems, Buffalo Grove, IL, USA) and the line measurement feature of ImageJ [18].

Approximately 200 µm transverse sections of the first internode were made manually using a No. 11 scalpel blade and a Nikon SMZ445 dissecting microscope (Nikon, Melville, NY, USA). Sections were transferred to 1.5 mL Eppendorf microcentrifuge tubes containing distilled water. Cross-sections were subjected to two different histological stains: toluidine blue or the Wiesner reagent. Following a 30 s treatment with toluidine blue (7.4 µM in H₂O), sections were rinsed twice with distilled water. To stain and visualize lignin, sections were treated with the Wiesner reagent (79 µM phloroglucinol-ethanol in 13.7 mM HCl) for 2 min. Stained cross-sections were mounted on microscope slides and visualized using a Nikon Eclipse E200MV R microscope (Nikon, Melville, NY, USA) attached to a PixeLINK 3 MP camera (PixeLINK, Ottawa, Canada). Images were captured using the associated PixeLINK uSCOPE software (PixeLINK, Ottawa, Canada) and further processed with Adobe Photoshop CS5.5 (Adobe, Waltham, MA, USA). The cellulose-binding module CBM3a (PlantProbes, Leeds, England) was used to detect crystalline cellulose [19]. Stem cross-sections were rinsed twice with phosphate-buffered saline (PBS; 33 mM Na₂HPO₄, 1.8 mM NaH₂PO₄ and 140 mM NaCl, pH 7.2) and suspended for 30 min in blocking buffer (PBG; 0.2% fish gelatin and 2.5% Bovine Serum Albumin in PBS, pH 7.4) prior to incubation with 100 µL of 10 µg/mL CBM3a diluted in PBG for 1 h. The solution was then removed, and two 5 min long washes were followed by two rapid changes of PBS. The cross-sections were then treated with 100 µL of anti-His antibody produced in mouse (Sigma-Aldrich, St. Louis, MO, USA) and diluted in PBG at a 1∶1000 ratio for 1 h. The solution was then removed and two 5 min long washes followed by two rapid changes of PBS. The cross-sections were then incubated with 100 µL of rabbit anti-mouse antibody conjugated to Texas Red fluorophore (Invitrogen, Grand Island, NY, USA) and diluted in PBG at a 1∶100 ratio for 1 h. Afterwards, the solution was removed and two 5 min washes were followed by two changes of PBS. A similar protocol was used to label the sections with LM10, a rat monoclonal antibody that detects xylan [20]. LM10 was diluted in PBG at a 1∶10 ratio and the secondary antibody, a goat anti-rat antibody conjugated to Texas Red fluorochrome, was diluted in PBG at a 1∶100 ratio. No tertiary antibody was required for the detection of the LM10 xylan epitope. All sections were mounted using Fluoromount Aqueous Mounting Medium (Sigma-Aldrich, St. Louis, MO, USA). Fluorescence microscopy was performed using a Leica MZ16 F microscope equipped with a mercury bulb attached to a Leica DFC300FX 1.4 MP camera (Leica Microsystems, Buffalo Grove, IL, USA). The violet filter (425 nm) was used to visualize lignin autofluorescence, and the Texas Red filter (560 nm) was used to visualize crystalline cellulose and xylan. Images were captured using the Image-Pro Plus imaging software (Media Cybernetics, Bethesda, MD, USA) and further processed using Adobe Photoshop CS5.5. Image acquisition was performed so that nominal autofluorescence was observed when analyzing unlabeled sections under Texas Red filter (Fig. S1).

ImageJ was used to analyze images by automatically measuring selected areas of interest after scale calibration [18]. Stained whole stem images were used to observe total vascular bundle count. To measure stem cross-section (n  =  20) and vascular bundle area (n  =  7–10 plants, 1–5 bundles/plant), the ImageJ freehand selection tool was used to trace the appropriate anatomy. For interfascicular region measurements (n  =  7–10 plants, 1–5 bundles/plant), the area in between vascular bundles was traced using the polygon tool. The walls of vessels and four adjacent bundle fibers were measured to estimate vascular bundle wall thickness (n  =  6 plants, 5 bundles/plant, 4 measurements/bundle). For interfascicular region cell wall thickness (n  =  6 plants, 5 bundles/plant, 4 measurements/bundle), lines were drawn across the adjacent cell walls of the first and second row of cells outside the bundle sheath of a vascular bundle. The adjacent walls of the second and third row of cells were also measured. Fluorescence was measured as the average pixel intensity relative to the background (n  =  5 plants, 5 sections/plant, 1–2 measurements/section). Selection tools were used to trace vascular bundles, interfascicular regions, whole stem cross-sections, and background regions. The initial images were further processed and false-colored using Photoshop CS5.5.

Stem tissue was pulverized with 6.35 mm stainless steel beads (BioSpec, Bartlesville, OK) using a Retsch Mixer Mill MM400 for 15 min and dried overnight at 37°C. Three biological replicates of each growth stage were individually placed on a sodium chloride compressed slide (New Era, Cat. No. 603A03). Ten randomly selected areas were scanned for each sample by FTIR-spectroscopy. Infrared spectra were obtained on a Hyperion 3000 microscope with a Vertex 70 Bruker spectrometer and processed using OPUS 7.2 spectroscopy software. Spectra were obtained by averaging 48 scans from 829 to 1829 cm⁻¹ at a resolution of 1 cm⁻¹ and corrected for background absorbance by subtraction of the spectrum of the empty NaCl compressed stage. For each developmental stage, the average line spectra were calculated following post-processing normalization.

For each measurement, 6 to 25 independent plants were sampled. Analysis of variance and Tukey’s contrasts were performed in R v2.15.0.

We selected three developmental stages to characterize internode and vascular development in B. distachyon: elongation, inflorescence emergence, and senescence. For all three stages, the first stem internode above the crown was sampled. During the first developmental stage sampled, the first and only internode was elongating at the base of the tallest tiller (Fig. 1A-B). This was the first appearance of stem tissue, which occurred 18 to 22 days after germination. At the second stage, the inflorescence first emerged from the flag leaf sheath (Fig. 1C-D). This developmental stage occurred 25 to 30 days after germination. These first two stages are roughly equivalent to stages 30 and 51 of the BBCH-scale for cereals [21]. The third stage sampled occurred when the plants had reached their maximum height, and the first internode and basal leaves were completely senesced (Fig. 1E-F). This stage, which occurred 39 to 45 days post germination, is more difficult to equate with the BBCH crop phenology system as it describes grain development characteristics at senescence with developmental stages. The number of stems per plant from the elongation stage to senesce did not change (Fig. 2A). Total stem height significantly increased at each stage by nearly 10 cm, and fresh weight significantly increased from 144 mg to 394 mg (Fig. 2B). The length and width of the first internode increased significantly before inflorescences emerged, but not after (Fig. 2C). These three stages encompass nearly the complete development of a stem internode.

We characterized numerous anatomical features of the stem at all three developmental stages. Brachypodium distachyon possesses an atactostele arrangement of vascular bundles with two circles at the periphery of the stem (Fig. 3A). The innermost vascular bundles are considerably larger than the outer ring of bundles. The cell types within the vascular bundles exhibit a pattern typical of other grasses with the phloem at the outside, and the xylem oriented towards the center of the stem (Fig. 3C). Each vascular bundle is surrounded by bundle sheath cells that also separate the xylem and phloem. The xylem is comprised of large vessels located at the interior end and both sides of the bundle with tracheids located in between the vessels. While not very pronounced, some bundles have a lacuna, which is a region of variable size and shape where protoxylem may have existed. The areas in between the vascular bundles, also known as the interfascicular region, are mostly comprised of sclerenchyma fibers (Fig. 3B). These fibers, along with chlorenchyma, are also located in the layer of cells between the epidermis and outer vascular bundle, a region referred to as the cortex. The center of the stem, a region known as the pith, is populated with parenchyma cells that are larger than surrounding cell types (Fig. 3B). A few parenchyma cells are located in the interfascicular region since the outermost exterior of the pith can be located in between inner vascular bundles (Fig. 3B). Interestingly, all of the vascular bundles were formed at or before the point of elongation, as the number did not significantly change (Fig. 2A). Similarly, the size of the vascular bundles did not change over time with an average size of 3100 µm² for outer bundles and 7800 µm² for inner bundles (Fig. 2D). The area between the vascular bundles significantly increased by 5300 µm² between elongation and inflorescence emergence (Fig. 2D). Therefore, an increase in interfascicular region size and not a change in vascular bundle size accounted for the increase stem internode area before flowering.

First internode cross-sections were treated with a polychromatic dye, toluidine-blue, that stains polysaccharides violet and lignin turquoise. At elongation, the vascular bundle cells appeared darker and had thicker cell walls than other cell types (Fig. 4A-B). At inflorescence emergence, the vascular bundles and interfascicular region appeared slightly darker and had noticeably thickened cell walls (Fig. 4C-D). At senescence, these same cell types appeared darker, and their cell walls were substantially thicker (Fig. 4E-F). For both inner and outer vascular bundles, the total cell wall size of xylem vessels and adjacent bundle sheath significantly increased between elongation and senescence (Fig. 2E). This growth was quite dramatic with cell wall thickness increasing from 2.7 to 4.6 µm for inner vascular bundles and from 2.0 to 3.3 µm for outer vascular bundles. For the interfascicular region, wall thickness of neighboring sclerenchyma fibers also increased significantly between elongation and senescence (Fig. 2E). The sclerenchyma fibers closest to the bundle sheath thickened from 1.7 to 5.0 µm while the sclerenchyma fibers second nearest to the bundle thickened from 1.7 to 4.1 µm. Thus, cell wall deposition appears to increase following elongation and inflorescence emergence in both the vascular bundle and interfascicular region.

Stem cross-sections were stained with the Wiesner reagent to observe changes in lignin. The Wiesner reagent stains low concentrations of lignin yellow and becomes increasingly red at higher concentrations. At elongation, the vascular bundles stained red while the interfascicular region appeared yellow (Fig. 4G-H). At inflorescence emergence, the vascular bundles stained a more distinctive red and the interfascicular region a darker yellow (Fig. 4I-J). At senescence, the vascular bundles and interfascicular region both stained a dark red (Fig. 4K-L).

As phenolic groups fluoresce, fluorescence microscopy was used as relative estimate of lignin content within stems. In elongating internodes, the vascular bundles were visible at low magnification whereas the interfascicular region could only be detected at higher magnifications (Fig. 5A-B). At inflorescence emergence, the interfascicular region was noticeably more fluorescent (Fig. 5C-D). At senescence, the vascular bundles were bright and the interfascicular region became strikingly more fluorescent (Fig. 5E-F). We then quantified the total corrected lignin autofluorescence of the whole stem, vascular bundles, and interfascicular region. The whole stem significantly increased in total autofluorescence between each stage sampled (Fig. 6A). Both inner and outer vascular bundles increased only slightly in total autofluorescence over time (Fig. 6A). The interfascicular region had the most substantial change in total autofluorescence where it increased by nearly two-fold between elongation and senescence (Fig. 6A). These observations demonstrate that significant lignin deposition occurred following elongation and inflorescence emergence in the interfascicular region, but not the vascular bundles.

Stem cross-sections were immunolabeled with the CBM3a recombinant protein probe in order to observe crystalline cellulose. At elongation, the xylem was more fluorescent than the interfascicular region (Fig. 5G-H). The interfascicular region was more apparent at inflorescence emergence but the vascular bundle fluorescence did not increase (Fig. 5I-J). Interestingly, neither the interfascicular region nor the vascular bundles seemed any more fluorescent at senescence (Fig. 5K-L). We then quantified the total corrected crystalline cellulose fluorescence of the whole stem, vascular bundles, and interfascicular region. The total crystalline cellulose fluorescence significantly increased between elongation and inflorescence emergence whereupon it appeared to plateau (Fig. 6B). This same trend was observed for inner vascular bundles and the interfascicular region plateaued, while both vascular bundle types interestingly decreased slightly (Fig. 6B). For outer vascular bundles, there was no change in fluorescence after internode elongation (Fig. 6B). Thus, crystalline cellulose immunolabeling increased in vascular bundles and the interfascicular region between elongation and inflorescence emergence and then seemingly ceased afterwards. To observe changes in xylan content, stem sections were immunolabelled with a monoclonal antibody against xylan. The total stem and interfascicular region fluorescence was greater at inflorescence emergence than elongation (Fig. 5M-P, 6C). In contrast, fluorescence did not significantly increase in the vascular bundles following elongation (Fig. 5O-R, 6C). Interestingly, total stem florescence significantly decreased between inflorescence emergence and senescence. Unlike the CBM3a indirect immunofluorescence detection of crystalline cellulose, the concentration of immunofluorescence detected xylan was appreciably greater in the xylem cells than the interfascicular fibers while their detection were noticeably difficult to observe in the bundle sheath.

The complex and generally hydrophobic secondary wall matrix can be difficult to penetrate with histo- and immunochemical reagents. In an attempt to better quantify cell wall modifications that occurred during stem development, ground tissue obtained from basal internodes was analyzed by FTIR spectroscopy. Typical spectra were obtained for all samples and the relative absorbance at wavenumbers associated with cell wall components were generally greater as the stems matured (Fig. 7). Peak absorbance at 1440 (O-H in-plane bending; [22]), 1380 (C-H bending; [22], [23], [24]) and 1330 cm⁻¹ (C–H vibration and O–H in-plane bending; [23], [25]) associated collectively with lignin, hemicelluloses, and cellulose consistently increased as the stem developed. Wavenumbers specific to lignin were significantly greater in senesced than elongating stems. These include C-H deformation at 1465 cm⁻¹ and aromatic ring vibrations at 1598 and 1508 cm⁻¹ [23], [26]. Spectra for senesced stem internodes were also significantly greater for cellulose at 1159 and 898 cm⁻¹ [22], [27], [28]. Peaks observed at 1,234 and 1,250 cm⁻¹ (C = O stretching; [25], [27]) may represent crystalline cellulose; however, further investigation is needed to verify this possibility. Consistent with the LM10 immunolabeling, peak absorbance at 1730 cm⁻¹ (C = O stretching; [23], [26], [29], [30]) suggests significant xylan accumulation from elongation to inflorescence emergence. Together, the FTIR absorbances at wavenumbers assigned to cellulose, hemicellulose, and lignin were consistent with the immunohistochemistry results that demonstrate changes in cell wall deposition following stem elongation.