To evaluate the amounts and radial distributions of berberine and palmatine in P. amurense, freeze-fixed blocks from the stems of P. amurense sampled in the fall and summer seasons were cut into serial tangential sections of 100-μm thickness. Every two serial sections (as one sample) were extracted with acetonitrile (ACN) and measured by HPLC ( Figure 1 ).

Berberine and palmatine exhibited similar radial distributions in the stems of P. amurense from summer and fall. They were mainly distributed in the phloem region but also in the cambial zone. In the xylem region, berberine was detected in most sections, while palmatine was only in several sections. Regarding the overall amount of alkaloids, berberine had higher content than palmatine in both seasons, which is consistent with the previous studies (Xian et al., 2014). However, from fall to summer, there was a slight decrease in the content of berberine and a significant increase in the content of palmatine.

It was reported that minor alkaloids, including columbamine, jatrorrhizine, magnoflorine, phellodendrine, 8-oxoepiberberine, and tetrahydropalmatine present in trace amounts in the stem of P. amurense (Xian et al., 2014). To quantify these minor alkaloids, freeze-fixed blocks from the stems of P. amurense in the fall and summer seasons were extracted with a 95% ACN aqueous solution (aq.). Obtained extracts were measured by HPLC at different conditions ( Table 1 ) to achieve optimal separation and quantification of target alkaloids.

Five of the minor alkaloids got separated and quantified at three different conditions, while it was hard to quantify tetrahydropalmatine separately: the signal overlapped with that of columbamine in Condition 1 and was not significantly detected in Conditions 2 and 3 ( Figure S1 ). In this case, the difference between the total area of the overlapping peak and the equivalent peak area of columbamine, which had been quantified in Condition 3, was regarded as the equal peak area of tetrahydropalmatine in Condition 1 and used for its quantification.

Quantification results of six minor alkaloids by HPLC are shown in Table 2 . Compared with the contents in fall samples, five alkaloids, including jatrorrhizine, columbamine, magnoflorine, phellodendrine, and 8-oxoepiberberine had a higher content in summer samples, while tetrahydropalmatine was of slightly lower content.

Alkaloid standards were measured by cryo-TOF-SIMS to determine their specific secondary ions. The acquired standard and the typical spectra obtained from the transverse surface of freeze-fixed P. amurense stems in fall and summer are shown in Figure 2 .

For quaternary ammonium alkaloids, including berberine, columbamine, jatrorrhizine, magnoflorine, phellodendrine, and palmatine, the strongest signals were detected at m/z 336, 338, 342, and 352, respectively, indicating their representative secondary ions to be the molecular ions [M]⁺. However, [M+H]⁺ ion at m/z 356 was detected as the strongest ion for the tertiary ammonium alkaloid tetrahydropalmatine. As for another tertiary ammonium alkaloid, 8-oxoepiberberine, since [M]⁺ ion at m/z 351 and [M+H]⁺ ion at m/z 352 were detected with similar intensities, visualizations of both ions are discussed in 3.4. Fragment ions were also detected as [M−CH₃]⁺, [M−CH₃+H]⁺, or [M−CH₃−CH₃]⁺ ions in the spectra of alkaloids ( Figure 2 ), but their intensity was low. From these results, it was determined to use [M]⁺ and [M+H]⁺ ions to visualize these alkaloids.

Cryo-TOF-SIMS spectra acquired by measuring the surface of actual samples in fall and summer ( Figure 2I, J ) exhibited the signals of characteristic ions of alkaloids as well. For example, berberine, the alkaloid of the highest amount in the stem of P. amurense, was detected with high intensity at m/z 336. Also, overlapping signals of significant ions with the same m/z produced by alkaloids were detected at m/z 338 (columbamine and jatrorrhizine), 342 (magnoflorine and phellodendrine), and 352 (8-oxoepiberberine and palmatine), and tetrahydropalmatine was detected at m/z 356.

Results obtained by cryo-TOF-SIMS/SEM analysis of freeze-fixed stems of P. amurense in fall and summer are shown in Figure 3 , 4 . After cryo-TOF-SIMS measurements, the sample blocks were transferred to cryo-SEM for conducting observation ( Figure 3A , 4A ) of the same measurement area containing phloem, cambial zone and xylem ( Figure 3J , 4J ) after appropriate freeze-etching ( Figure S2 ), which was to enhance the contrast of cryo-SEM images. To further assign detailed tissue structures, sample sections were obtained from P. amurense stems in the fall and summer, stained by toluidine blue, and observed by optical microscopy ( Figure 5 ).

In the results obtained from the transverse surface of the fall P. amurense stem, potassium ( Figure 3C ) was specifically detected in the phloem and cambial zone at higher intensity, as well as specific structures in xylem region at a lower intensity, which tended to be similar to the localization of living cells. However, in the results of summer samples, potassium ions were found to have wide distribution over the area measured, except for some specific positions in the phloem ( Figure 4C ). Therefore, the specificity of potassium distribution in the summer sample was not sufficient to represent the biological activity of cells.

Phosphatidylcholine is a major component of plant biological membranes and is considered a good marker for the biological activity of plant cells for its specified detection in living cells. It has been reported that phosphocholine ion ([C₅H₁₅NO₄P]⁺) of m/z 184 detected in TOF-SIMS could perform as the marker for phosphatidylcholine ion mapping (Okumura et al., 2017). In addition, the ion of m/z 58 ([C₃H₈N]⁺) was identified as the characteristic and stronger secondary ion derived from phosphocholine in TOF-SIMS analysis (Fletcher et al., 2007; Gunnarsson et al., 2010) ( Figure S3 ), by which ray tissues were much more precisely visualized than that by the potassium distribution. From the above points, cryo-TOF-SIMS images of the ion at m/z 58 ( Figure 3D , 4D ) are supposed to represent the distribution of the living cells compared to those of m/z 39 and 184 ions.

Since characteristic ions of jatrorrhizine and columbamine, as well as phellodendrine and magnoflorine, have the same values in m/z, which resulted in signal overlap in cryo-TOF-SIMS analysis, here we discuss them two-by-two as a whole. [M]⁺ ion of 8-oxoepiberberine with the characteristic ion at m/z 351 showed the same distribution in cryo-TOF-SIMS images as that of the ion at m/z 352 ( Figure S4 ). Since the signal at m/z 352 was overlapped by [M]⁺ ion of palmatine and [M+H]⁺ ion of 8-oxoepiberberine, we make an overall discussion on palmatine and 8-oxoepiberberine as well. Eight alkaloids were present in the phloem, cambial zone, and xylem, and the distribution of different alkaloids varied in different tissues. For berberine of m/z 336 ( Figure 3E , 4E ), jatrorrhizine and columbamine of m/z 338 ( Figure 3F , 4F ), palmatine and 8-oxoepiberberine of m/z 352 ( Figure 3H , 4H ), similar distributions of alkaloids are shown for samples from both seasons. They were detected at high intensity in phloem parenchyma cells and the cambial zone. Also, they were found at lower content in the xylem region, with decreasing order from vessel-neighboring parenchyma cells, axial parenchyma cells, to ray cells. They were not detected in the outer cortex or other dead tissues, such as phloem fibers. In contrast, phellodendrine and magnoflorine of m/z 342 ( Figure 3G , 4G ) were detected with decreasing order from vessel-neighboring parenchyma and axial parenchyma cells in xylem, ray cells to phloem parenchyma cells and cambial zone, for samples from both seasons. The distribution of phellodendrine and magnoflorine was also found in the inner cortex of the fall sample but not in the summer sample. As for tetrahydropalmatine of m/z 356 ( Figure 3I , 4I ), it mainly localized in the inner cortex but was not detected with enough intensity from phloem to xylem in both fall and summer samples.

Sampling was achieved on 5 November 2019 in the fall and 22 July 2020 in the summer. Each sample disk (thickness of 10 mm) was obtained from P. amurense trees (2.0 m in height, 3-year-old) grown in the Botanical Garden at Nagoya University, Higashiyama campus (Nagoya, Japan) and cut into small blocks (circular sector with a radius of ca. 4.5 mm and central angle of π/16) containing phloem, cambial zone, and xylem. The blocks were quick-frozen with liquid Freon^® 22 (DuPont) at −160°C and stored at −80°C. Detailed procedures are schematically illustrated in Figure S5 .

Alkaloid standards including berberine chloride, palmatine chloride, and magnoflorine (SIGMA-Aldrich, MA, USA), tetrahydropalmatine and jatrorrhizine chloride (Tokyo Chemical Industry Co., Ltd., Tokyo, Japan), phellodendrine (Nacalai Tesque Inc., Kyoto, Japan), 8-oxoepiberberine and columbamine (MedChem Express, NJ, USA) were purchased and used as received. Ammonium formate and formic acid (FA) were purchased from Kishida Chemical Co., Ltd. (Osaka, Japan). ACN and distilled water of HPLC grade were purchased from Kanto Chemical Co., Inc. (Tokyo, Japan).

Frozen sample blocks were cut from bark to xylem into serial tangential sections of 100-µm thickness. Every two serial sections were collected in the same plastic tube (1.5 mL in volume) and extracted with 1 mL ACN at 60°C for one hour. HPLC measurements for berberine and palmatine were achieved with Agilent 1100 series LC system (Agilent Technologies Inc.) equipped with a TSKgel ODS-100S column (4.6 mm i.d. × 25 cm, C₁₈, 5.0 μm, Tosoh Corp., Japan). Berberine and palmatine were separated by a binary buffer system of 0.1% (v/v) FA (buffer A) and 0.1% (v/v) FA in ACN (buffer B) at a flow rate of 1 mL/min. The gradient was 35 min in total and set as follows: from 5 to 20% (v/v) buffer B in 5 min, from 20 to 30% (v/v) buffer B in 15 min, from 30 to 95% (v/v) buffer B in 0.1 min, holding at 95% (v/v) buffer B for 9.9 min, declining to 5% (v/v) buffer B in 0.1 min, and holding at 5% (v/v) buffer B for 4.9 min. All the chromatograms were taken at column temperature 30°C and UV detection wavelength 265 nm. The measurements were carried out in three replicates using three blocks obtained from the same sample disk to evaluate the average amount and standard errors of berberine and palmatine.

Frozen sample blocks from fall and summer were extracted with 95% ACN aq. (1 mL for every two blocks) at 60°C for one hour. A flash chromatography system (Pure C-810, Büchi, Switzerland) was used to separate the minor alkaloid group from berberine and palmatine using a Flashpure Select C₁₈ 4g column (particle diameter 20–35 μm, Büchi). Minor alkaloids were separated by a binary buffer system of 0.1% (v/v) FA aq. (buffer A) and 0.1% (v/v) FA in ACN (buffer B) at a flow rate of 15 mL/min. The gradient was 60.7 min in total and set as follows: holding at 5% (v/v) buffer B in 0.5 min, from 5 to 20% (v/v) buffer B in 5 min, from 20 to 40% (v/v) buffer B in 30 min, from 40 to 95% (v/v) buffer B in 0.1 min, holding at 95% (v/v) buffer B for 13 min, declining to 5% (v/v) buffer B in 0.1 min, and holding at 5% (v/v) buffer B for 12 min. Fractions from 0 to 33 min were collected and concentrated, then re-dissolved with 95% ACN aq. for HPLC analysis.

HPLC quantification of minor alkaloids was achieved with the same LC system equipped with an XSelect CSH C₁₈ Column (4.6 mm i.d. × 25 cm, C18, 5.0 μm, Waters Corp., USA). Three conditions ( Table 1 ) were applied to achieve better compound separation. All the chromatograms were taken at column temperature 30°C and UV detection wavelength 265 nm. The measurements were performed in two replicates with blocks from two different sample disks to evaluate the average amount of minor alkaloids. MS1 and MS2 spectra ( Figure S6 ) of HPLC peaks were obtained for alkaloid standards and the plant samples using Esquire 3000 (Bruker Corp., USA) connected to Agilent 1100 series LC system to confirm the HPLC peak assignments.

Details of the manufactured cryo-TOF-SIMS/SEM system were described previously (Kuroda et al., 2013; Masumi et al., 2014). For each sample from fall and summer, a frozen sample block was fixed in a copper holder by ice embedding and cut in the glove box under a dry N₂ environment (sample temperature < −30°C) to achieve a clean and flat surface, then transferred to the cryo-TOF-SIMS system by cryo-vacuum shuttle for analysis. Positive ion images were obtained by cryo-TOF-SIMS (TRIFT-III spectrometer, ULVAC-PHI Inc., Japan). 22 keV A u 1 + at the current of 5−7 nA was used as the primary ion beam, and a low-energy pulsed electron gun (30.0 eV) was used for surface charge compensation. Other conditions were set as follows: raster size at 300 × 300 μm, measurement time of 10 min, pulse width at 13 ns (non-bunched, image) or 1.8 ns (bunched, spectrum), spot size at 1.0 μm (image), the temperature at −120 to −130°C, vacuum level below 1.0 × 10⁻⁷ Pa. Standard chemicals of alkaloids were dissolved at ca. 50 mM in ACN, dropped on the achieved smooth and flat surface of ice tables (made up of 50 mM KCl solution), and dried up, then measured by cryo-TOF-SIMS in the same procedure in bunched mode.

After cryo-TOF-SIMS measurements, the plant sample block was transferred to cryo-SEM by a cryo-vacuum shuttle. To enhance the contrast of SEM images, the frozen and hydrated sample surface was freeze-etched at −90°C, then observed at around −130°C to obtain images of the same region measured by cryo-TOF-SIMS. The acceleration voltage was set at 1.5 kV, and the working distance was 10 mm.

Obtained cryo-TOF-SIMS images were connected using WinCadence 5.1.2.8 (ULVAC-PHI Inc., Japan) and MATLAB R2019b (The MathWorks Inc., USA) with PLS Toolbox 8.8.1 (Eigenvector Research Inc., USA) without any ion count normalization. Color scales of images were adjusted using ImageJ software (The National Institutes of Health, USA) (Schneider et al., 2012). Cryo-SEM images were connected using Photoshop CS5 Extended (Adobe Systems Inc., USA).

Cryo-sections were prepared by Kawamoto’s film method (Kawamoto, 2003; Kawamoto et al., 2021) with some modifications at the submerging steps to obtain better sections ( Figure S7 ). First, frozen blocks of P. amurense stems in fall and summer were immersed in the embedding medium (SCEM, SECTION-LAB Co., Ltd., Japan) at room temperature for 30 min. Then, thawed blocks were embedded in SCEM and cut into sections at 2-μm thickness using a sliding microtome (REM-710, Yamato Kohki Industrial Co., Ltd., Japan). Before staining, sections obtained were pretreated by the following steps: thawed in the air for 20 s, submerged three times in ethanol for 20 s, 10 s, and 1 min, submerged in acetone for 1 min, submerged in 4% paraformaldehyde solution for 5 min, rinsed under running water for 5 min. After rinsing, sections of 2 μm from the fall block and 4 μm from the summer block were stained with toluidine blue and observed using an optical microscope (BX-60, Olympus Corp., Japan).