Improved clearing method contributes to deep imaging of plant organs

Tissue clearing methods are increasingly essential for the microscopic observation of internal tissues of thick biological organs. We previously developed TOMEI, a clearing method for plant tissues; however, it could not entirely remove chlorophylls nor reduce the fluorescent signal of fluorescent proteins. Here, we developed an improved TOMEI method (iTOMEI) to overcome these limitations. First, a caprylyl sulfobetaine was determined to efficiently remove chlorophylls from Arabidopsis thaliana seedlings without GFP quenching. Next, a weak alkaline solution restored GFP fluorescence, which was mainly lost during fixation, and an iohexol solution with a high refractive index increased sample transparency. These procedures were integrated to form iTOMEI. iTOMEI enables the detection of much brighter fluorescence than previous methods in tissues of A. thaliana, Oryza sativa, and Marchantia polymorpha. Moreover, a mouse brain was also efficiently cleared by the iTOMEI-Brain method within 48 h, and strong fluorescent signals were detected in the cleared brain.


Introduction
The development of microscopes, dyes, uorescent proteins (FPs), and sample preparation methods have enabled observation of bright and high-resolution microscopic images, which is the driving force to unravel the mystery of life. In particular, FPs are indispensable tools to visualize cellular components, including organelles, proteins, nucleic acids, and small molecules, and provide information such as temperature, pH, and voltage [1][2][3] . However, it is often di cult to su ciently detect the uorescent signals of FPs emitted from the internal regions of three-dimensionally thick tissues because auto uorescent pigments absorb the light and some cell components, which refract or re ect light, scatter and disturb the signals. Several clearing methods to elute pigments from cells and adjust the refractive index through the specimen to the mounting medium have been developed to observe the tissue deeply embedded in an organ. The clearing methods developed for animal tissues, including ScaleS, CUBIC, PACT, and SeeDB2, can maintain the uorescence of FPs in highly transparent tissues [4][5][6][7][8] . ScaleS uses the hyperhydration effect of urea to make animal tissues transparent, CUBIC and PACT employ effective detergents to remove obstacles, and SeeDB2 uses iohexol solution for transparency. For plant tissues, Scale based method 9 , ClearSee 10,11 , PEA-CLARITY 12 , and TOMEI [13][14][15] were developed as optical clearing methods and preserve the uorescence of FPs. In fact, TOMEI is a powerful technique to analyze the network of vascular tissues in plant organs 16,17 . All of these methods substantially ameliorate the optical properties of specimens; however, each method has drawbacks. Scale-based methods, ClearSee, and PEA-CLARITY are slow to complete clearing. Although TOMEI rapidly clears plant tissues and organs 14,15 , certain auto uorescent pigments remain in plant tissues after clearing and a high concentration of 2,2′thiodiethanol (TDE), used in the TOMEI procedure, reduces the uorescence intensity of FPs 13,18 .
To overcome the problem of auto uorescence from remnant pigments, we modi ed all steps of the TOMEI method, including xation, decolorization, FP reactivation, and mounting, and propose here an improved version of TOMEI, which we designate iTOMEI. We successfully performed deep imaging in plant tissues of Arabidopsis thaliana, Oryza sativa, and Marchantia polymorpha as well as mouse brain tissue using iTOMEI.

Assessment of xation conditions to maintain uorescence of FPs
To improve TOMEI, we rst reconsidered the xation conditions to prevent the reduction in uorescence of FPs. We investigated the effect of ve buffers (PBS, PIPES, HEPES, MOPS, and Tris-HCl) adjusted to pH 7.0 using GFP uorescence, which is unstable under an acidic pH. Arabidopsis thaliana seedlings overexpressing GFP were xed with 2% formaldehyde (FA) in these buffers. The GFP uorescence in the hypocotyl was detected using a uorescence stereomicroscope before and after xation. PBS and PIPES maintained GFP uorescence comparatively better than HEPES, MOPS, and Tris-HCl (Fig. 1a). A 4% FA solution was used for plant tissue xation; however, it dramatically reduced GFP uorescence even in PBS (Fig. 1b). Fixation with 1% FA more effectively maintained uorescence, although the uorescence intensity was reduced to 40% compared with that before xation (Fig. 1b). To examine the effect of 1% FA xation on protein localization and cellular structures, the subnuclear localization of a proliferating cell nuclear antigen (PCNA) fused with EGFP was observed after xation. The PCNA functions as a sliding clamp to tether DNA polymerase to DNA and PCNA-GFP shows nuclear speckles during the late S-phase in living A. thaliana cells (Fig. 1c) 19,20 . Nuclear speckles were not detected after treatment with 0.1% FA because such a low concentration of FA was insu cient for xation. However, treatment with 1%, 2%, and 4% FA maintained the subnuclear and speckle localization of PCNA, suggesting that 1% FA is su cient to x cellular structures and the best xation condition to preserve GFP uorescence in our clearing method ( Fig. 1c).

Caprylyl sulfobetaine elutes chlorophylls without GFP quenching
The auto uorescence of chlorophylls derived from chloroplasts is an obstacle for deep imaging of plant tissues. To identify a reagent for e cient removal of chlorophylls, we screened 19 detergents. Fixed 2- week-old seedlings of A. thaliana were incubated in PBS buffer supplemented with 10% different detergents for 24 h. The absorbance of chlorophylls, which were eluted from the seedling to the detergent solution, was measured by an absorption spectrophotometer. We determined caprylyl sulfobetaine (#7) and sodium deoxycholate (#3) to be the best and second-best detergents for e ciency in chlorophyll elution (Fig. 1d). To investigate their effect on GFP uorescence, xed wild-type plants and plants overexpressing GFP were incubated with 10% caprylyl sulfobetaine and sodium deoxycholate for 24 h.
Both detergents slightly enhanced GFP uorescence after incubation (Fig. 1e). This tendency of sodium deoxycholate is consistent with a previous report for ClearSee 10 . Our assessment suggested that caprylyl sulfobetaine is the best detergent for chlorophyll elution without reducing GFP uorescence.
To enhance the effect of a zwitterionic detergent, caprylyl sulfobetaine, we tested the combinations of caprylyl sulfobetaine with a nonionic detergent, Triton X-100 (Fig. 1f), an anionic detergent, sodium deoxycholate (Fig. 1g), and urea (Fig. 1h). However, all combinations failed to enhance chlorophylls elution, suggesting that incubation with caprylyl sulfobetaine alone is the best. Subsequently, chlorophyll elution e ciency and GFP uorescence were measured after incubation with PBS supplemented with 10%, 20%, or 30% caprylyl sulfobetaine. The PBS buffer with 20% caprylyl sulfobetaine showed the highest e ciency of chlorophyll elution and maintenance of GFP uorescence, indicating that 20% caprylyl sulfobetaine was optimal for our clearing technology.

Alkaline buffers chemically reactivate FP uorescence in plant tissues
Chemical reactivation (CR) enables recovery of reduced GFP uorescence in resin-embedded specimens by alkaline treatment in animal tissues. We tested the effect of CR on GFP uorescence in A. thaliana. Fixed seedlings were treated with trisodium phosphate buffer at pH 10.7 and sodium carbonate buffer at pH 11.3 for 2 h. Both treatments signi cantly enhanced GFP uorescence in xed plants overexpressing GFP after CR (Fig. 2a), demonstrating that CR recovered the reduced uorescence of GFP in plants. In addition, reactivated GFP uorescence after CR was maintained in PBS buffer for 3 h (Fig. 2a). To determine the optimum buffer pH for CR, we modi ed the pH of sodium dihydrogen phosphate buffer, sodium hydrogen phosphate buffer, and trisodium phosphate buffers to prepare buffers ranging in pH from 8.0 to 12.0. Fixed seedlings were incubated in the buffers at pH 7.4 to 12.0 for 2 h. Buffers at pH 8.0 to 11.0 recovered the GFP uorescence to the same degree (Fig. 2b), thus we chose the pH 8 buffer for our clearing method.
Next, aiming to simplify the process and reduce the procedure time, we attempted to combine decolorization with CR. Fixed seedlings were incubated in PBS buffer at pH 7.4 or sodium phosphate buffer at pH 8.0 supplemented with 10% caprylyl sulfobetaine for 24 h. In addition, the seedlings were incubated in PBS buffer at pH 7.4 with 10% caprylyl sulfobetaine for 24 h and sequentially incubated in sodium phosphate buffer at pH 8.0 for 2 h (Fig. 2c). Incubation at pH 8.0 with 10% caprylyl sulfobetaine for 24 h resulted in GFP uorescence of higher intensity than other incubations (Fig. 2c). Thus, decolorization and CR were performed simultaneously with this method.
Iohexol enables clearing of plant organs without FP quenching Finally, we searched for a clearing reagent superior to TDE, which is used for TOMEI. Iohexol dramatically improves the transparency of mouse brain tissue without FP quenching 7 . We investigated whether iohexol improves clarity and does not affect GFP uorescence in plant tissues. Fixed and decolorized seedlings were incubated in PBS buffer, 97% (w/w) TDE, and 70.4% (w/w) iohexol solution for 1 h, and then GFP uorescence intensity was measured. TDE slightly decreased GFP uorescence compared with PBS buffer, whereas iohexol treatment did not affect GFP uorescence (Fig. 2d). Subsequently, the transparency of the cleared leaf was compared by measuring the intensity of grids viewed through the leaf (Fig. 2e). TDE and iohexol decreased the intensity of grids behind the leaf compared with that of PBS buffer (Fig. 2e). These results suggested that iohexol is also a suitable clearing reagent for plant tissues.
The nal protocol of iTOMEI is as follows.
Step 1: a sample is xed in PBS buffer supplemented with 1% FA for 1 h. Step 2: the xed sample is decolorized in sodium phosphate buffer at pH 8.0 supplemented with 20% caprylyl sulfobetaine for 24 h.
Step 3: the sample is incubated in 70.4% iohexol solution for 1 h. iTOMEI is the best clearing technique for plant organs within 26 h The iTOMEI procedure was compared with the clearing methods ClearSee and TOMEI-II. The total processing time for each technique was uni ed to 26 h. iTOMEI-treated seedlings of A. thaliana showed the highest transparency (Fig. 3a). Seedlings overexpressing GFP were xed in PBS buffer supplemented with 1%, 2%, or 4% FA and treated with the three clearing methods. iTOMEI maintained the highest uorescence intensity among the three methods under each xation condition (Fig. 3b). Confocal uorescence images of cotyledons in seedlings expressing histone H2B-GFP were captured after applying the three clearing techniques (Fig. 3c). The uorescent signals were almost invisible at depths of 45 and 75 mm in cotyledons treated with PBS buffer and TOMEI-II, respectively. Although GFP signals were detected at 75 mm depth in cotyledons cleared using ClearSee or iTOMEI, the brightest uorescence at 75 mm depth in the cotyledon was detected after iTOMEI treatment.
iTOMEI reveals the FP expression pattern inside plant tissues and organs Next, we performed iTOMEI to clear organs of rice (Oryza sativa). Three days was required for the leaves to become completely transparent (Fig. 4a). We also observed the expression pattern of the transcription factor OsMADS15, which regulates the transition from the vegetative to the reproductive phase in the shoot apical meristem (SAM). Given that the diameter of the rice reproductive SAM is approximately 150 µm, the FP expression pattern in the rice SAM could not be analyzed without sectioning. After 4% FA xation of the SAM dissected from plants expressing OsMADS15 fused with mOrange 22 , the xed SAM was stained with a uorescent cell-wall-staining optical brightener, SCRI Renaissance 2200 (SR2200), and cleared with TOMEI-II or iTOMEI. Neither uorescent signal was detected in the central area of the SAM in only xed SAMs (Fig. 4b). The SR2200 signal was detected in the central portion of the SAM in TOMEI-IItreated samples, whereas OsMADS15-mOrange signal was not detected at the depth of 80 µm. After clearing with iTOMEI, both signals were detected in the central region of the SAM at 80 µm depth. OsMADS15-mOrange was strongly expressed in the basal region of a hairy bract and in the outer two cell layers of an incipient primary branch meristem (Fig. 4b), suggesting that iTOMEI can reveal the expression pattern of FPs inside the larger reproductive SAM. We also attempted to detect auxin signaling in the root tips of plants expressing pDR5rev::NLS-3xVenus. iTOMEI visualized pDR5rev::NLS-3xVenus in the root tip, whereas PBS alone did not (Fig. 4c). The pDR5rev::NLS-3xVenus construct was strongly expressed in the central and surrounding metaxylem cells in the stele, quiescent center, and columella cells. In an optical cross-section of the root, ve FP foci surrounding the signal in the center of the root were detected, corresponding to the central and surrounding metaxylem (Fig. 4d).
Next, we attempted to clear organs of the liverwort Marchantia polymorpha. An apical portion of a mature thallus of M. polymorpha was cleared within 26 h as for A. thaliana seedlings (Fig. 5a). Thalli of a 2week-old gemmaling expressing histone H2B-tdTomato were xed and treated with PBS, TOMEI-II, and iTOMEI. tdTomato signals were observed from the dorsal to the ventral surface in the thalli by confocal microscopy. Fluorescent signals were detected at the depth of 200 µm in the iTOMEI-treated gemmaling, but not in PBS-and TOMEI-II-treated gemmalings (Fig. 5b). Calco uor White M2R and H2B-tdTomato signals were observed from the ventral to the dorsal surface of the thallus in a 3-day-old gemmaling ( Supplementary Fig. 1). In the apical region of the thallus, we detected an apical cell or subapical cells that appeared fan-shaped when viewed in the Y-Z optical section (Fig. 5c) 23 . These observations demonstrated that iTOMEI is suitable for deep imaging of M. polymorpha thalli.
RSL class I genes encode basic helix-loop-helix transcription factors conserved among land plants. These transcription factors positively regulate root hair formation in O. sativa and A. thaliana, and rhizoid development in M. polymorpha and Physcomitrium patens [24][25][26][27] . RSL class I genes are expressed in a precursor cell of the root hair and rhizoid in O. sativa, A. thaliana, and P. patens. However, the expression pattern of an ortholog of M. polymorpha, MpRSL1, has not been investigated. The expression pattern of nuclear tdTomato under control of the MpRSL1 promoter in the gemma showed that the promoter was activated in rhizoid precursor cells as expected and, unexpectedly, in the apical regions (Fig. 5d). Thus, MpRSL1 may function in the apical region of gemmae. Next, the gemma cup, a cup-shaped receptacle on the mature thallus, was cleared using iTOMEI. Under observation with a two-photon excitation microscope, the gemmae were observed within the transparent gemma cup through the wall of the gemma cup (Fig. 5e). Many immature gemmae are present in a gemma cup and do not develop synchronously, which enables observation of gemmae at various developmental stages within a gemma cup (Fig. 5f). MpRSL1 was strongly expressed in a preapical region of the immature gemmae at different developmental stages (Fig. 5f, arrowheads), and MpRSL1 expression was detected in rhizoid precursor cells in large immature gemmae (Fig. 5f, arrows) but not in smaller gemmae. These results demonstrate that iTOMEI enables in situ observation of gene expression patterns during gemma development through the wall tissues of the gemma cup.
iTOMEI contributes to deep imaging of mouse brain Finally, we tested whether the range of application of iTOMEI could be expanded to animal tissues. We modi ed the iTOMEI protocol for the mouse brain with reference to the SeeDB2 protocol and designated the procedure iTOMEI-Brain (iTOMEI-B). Fixed mouse brain was sliced at 2 mm and treated with PBS, SeeDB2, and iTOMEI-B. Although 24-hour treatment of iTOMEI-B was insu cient to clear the brain, 48hour treatment achieved high transparency ( Supplementary Fig. 2a). iTOMEI-B inhibited the size change of the slice as much as SeeDB2 (Supplementary Fig. 2a). Mouse brain expressing EGFP was also treated with PBS, SeeDB2, and iTOMEI-B, and observed from the cortical surface of the brain using a two-photon excitation microscope. iTOMEI-B allowed acquisition of uorescence images to 3 mm depth in transparent brain expressing GFP ( Supplementary Fig. 2b). The GFP uorescence was clearly detected in the soma, axon, and dendrite of the hippocampus and cerebral cortex ( Supplementary Fig. 2b-d).
Surprisingly, the axonal projection was observed in the thalamus at the depth of 3 mm from the cortical surface. Deep imaging of the transparent brain using iTOMEI-B enabled identi cation of the layer structure of the hippocampus and cortex. The characteristic dendritic morphology of CA1 pyramidal neurons was maintained (Supplementary Fig. 2d). Our imaging data suggested that iTOMEI can be adapted to not only plant tissues but also the mouse brain.

Discussion
The iTOMEI procedure can preserve and recover bright FP uorescence of improved intensity in transparent organs and tissues compared with previous methods. This advantage enables threedimensional detection of weak uorescent signals from FP in internal tissues of thick specimens. Chemical reactivation enhances the uorescence intensity of GFP and YFP in resin-embedded biological tissues 21 . Alkaline solutions in the range from pH 9.0 to 12.0 are generally used for CR. Given that biological tissues are morphologically damaged by a solution at such a high pH, adaptation of CR for sample preparation without embedding in resin seemed di cult. However, we showed that a weak alkaline solution of pH 8.0 harbors the uorescence recovery capability (Fig. 2b). In addition, PBS buffer at pH 7.4 slightly reactivated uorescence. These results suggested that sample incubation in solutions at pH 7.0 to 8.0 can reactivate FPs quenched by xation.
Fixation greatly reduces the uorescence intensity of GFP in A. thaliana. The uorescence intensity of a sample xed with 1% FA was approximately double and four-times higher than that of samples xed with 2% and 4% FA, respectively (Fig. 1b). The nal uorescence intensity was less affected by FA concentration because the uorescent intensity of iTOMEI-treated samples xed with 1% FA was not more than double that of iTOMEI-treated samples xed with 4% FA (Fig. 3b). ClearSee and TOMEI-II showed a similar tendency to iTOMEI, suggesting that the GFP inactivated by high FA concentration was easily recovered during each procedure because incubation in near-neutral solutions can recover GFP uorescence.
Clearing of the rice SAM by iTOMEI provided spatial expression-pro ling of the key transcription factor OsMADS15 at three-dimensional single-cell resolution (Fig. 4). OsMADS15, which is a homolog of APETALA1/FRUITFUL in A. thaliana, is essential for bract and oral organ development in rice 28 . Consistent with the reported function, OsMADS15-mOrange signal was observed at the incipient primary branch meristem, which provides owers at an advanced stage, and in the basal region of a hairy bract, which is a suppressed leaf produced by the reproductive SAM 29 . OsMADS15 expression is regulated by the origen Hd3a in the rice SAM 22,30 , thus our results suggested the site of action of rice origen at single-cell resolution. In the Mprsl1-1 mutant of M. polymorpha, rhizoids do not develop from the ventral surface of the thallus, gemmae and mucilage papillae did not develop in gemma cups, and slime papillae did not develop near the apical region 26 . However, it has not been known whether the Mprsl1-1 mutant shows any defects in the apical region. We revealed that MpRSL1 promoter activity was detected in rhizoid precursor cells and the apical region in mature gemmae, and was restricted in the apical region during gemma development in the gemma cup (Fig. 5d-f). This imaging data suggest that MpRSL1 is a useful marker to detect meristematic cells in liverworts.
Although iTOMEI was developed for plant tissues and organs, it can be applied to the mouse brain ( Supplementary Fig. 2). Because each clearing method has its own characteristics, the method should be used properly depending on what is important in the imaging analyses. For example, the transparency in the brain treated with SeeDB2 was lower compared with iTOMEI-B but the clearing brain by SeeDB2 will highly preserve the cell morphology and the uorescence of FPs and minimizes the spherical aberration in super-resolution 3D imaging 7 . The high transparency by iTOMEI-B potentially contributes to observation of the weak FP expression in the cells located deep in the brain.
Recent development in transparent techniques enables us to analyze the expression pattern with FPs in the whole tissues and organs. Our developed iTOMEI is a powerful technique to transparentize thick organs with almost no attenuation of the uorescence from FPs, which were expressed in the profound depth of organs. It will be also a helpful technique to construct the image platform of organ morphology at single-cell resolution.

Plant Materials and Growth Conditions
Arabidopsis thaliana ecotype Columbia-0 was used as the wild type. The transgenic lines p35S::GFP, pPCNA1::PCNA1-EGFP, and p35S::H2B-GFP were previously reported 13,20 . Seeds were germinated and grown on half-strength Murashige and Skoog plates supplemented with 1% sucrose and 14-day-old plants were transferred to soil. Plants were grown in a growth chamber (16 h light/8 h dark, 22°C).
Oryza sativa 'Nipponbare' was used for SAM analysis. Plants were grown in the growth chamber under short-day conditions (10 h light at 27°C and 14 h dark at 25°C, light intensity 400-700 nm, 100 µmol m − 2 ·s − 1 ). Vegetative and reproductive SAMs were isolated by hand dissection of the basal region of rice under a microscope 31 . Generation of the transgenic lines OsMADS15-mOragene and DR5rev:NLS-3xVenus was previously reported 22,32 .
A male accession of M. polymorpha Takaragaike-1 (Tak-1) was used as the wild type. Plants were grown on half-strength Gamborg's B5 plates in a growth chamber (16 h light/8 h dark, 22°C).

Plasmid Construction And Transformation
The MpRSL1 (Mp3g17930) genomic region, including a 6964-bp fragment upstream of the 15th Met codon in the rst exon, was ampli ed as the promoter region from Tak-1 genomic DNA by PCR sing a gene-speci c primer pair (proMpRSL1-F1, CACCCCCAAATGCAATTCTATTGTGTATTCAT; proMpRSL1-R, AGCTGGGTCGGCGCGCATGTTGTTCCTCCTGCTCAGTGT) and subcloned into the pENTR/D-TOPO vector (Thermo Fisher Scienti c, USA). A DNA cassette including the promoter region was transferred to a binary vector pMpGWB116 33 using the Gateway LR Clonase II Enzyme mix (Thermo Fisher Scienti c). To create a construct for proMpEF1:H2B-tdTomato, a SalI-NotI fragment including a coding sequence for H2B-tdTomato from the plasmid spUC-RPS5A::H2B-tdTomato 34 was ligated with SalI-and NotI-digested pENTR-1A vector (Thermo Fisher Scienti c), and the resulting vector was used for LR reaction with a binary vector pMpGWB303 33 . The binary vectors were used for Agrobacterium-mediated transformation of Tak-1 thalli and Tak-1 ⋅ Tak-2 F1 sporelings, respectively.

Measurements of uorescence intensity in A. thaliana
The uorescent images of seedlings were captured under the same optical condition using a uorescence stereomicroscope (SMZ18; Nikon, Japan) equipped with a DS-Ri2 digital camera (Nikon) before and after treatment. The uorescence intensity of GFP was measured in a hypocotyl of the seedling using ImageJ software. We modi ed the pH of sodium phosphate buffer to prepare buffers at pH 8.0 by mixing sodium hydrogen phosphate solution and sodium dihydrogen phosphate solution, sodium hydrogen phosphate solution at pH 9.0 by NaOH, and trisodium phosphate solution at pH 10, 11, and 12 by NaOH.

Measurements of chlorophyll absorbance
Fourteen-day-old seedlings were xed with 2% FA in PBS buffer for 1 h. A xed seedling was placed into a plastic tube containing various detergent solutions and was incubated for 24 h. As a negative control, the detergent solutions without the seedling were incubated for 24 h. The absorbance of chlorophylls, eluted from the seedling to the detergent solution, was measured using an absorption spectrophotometer (Nanophotometer Pearl; IMPEL, Germany).

Optical clearing of A. thaliana by iTOMEI
Seedlings and leaves were xed in PBS buffer with 1%, 2%, and 4% FA for 1 h with evacuation for the rst 10 min and then the samples were washed three times in PBS buffer for 5 min. The xed samples were treated with a decolorization solution (100 mM sodium phosphate buffer [a mixture of sodium hydrogen phosphate and sodium dihydrogen phosphate] at pH 8.0 with 20% caprylyl sulfobetaine [TCI, Japan]) for 24 h with gentle shaking. After washing in PBS buffer for 5 min, the samples were incubated in a mounting solution (70.4% iohexol [TCI or Merck, USA] in PBS buffer) for 1 h with gentle shaking. Finally, the samples were mounted on a glass slide in the mounting solution and a coverslip applied. Exposure of the mounting solution to air for several minutes results in water evaporation and a skin forms on the solution surface. To inhibit skin formation, the mounting procedure should be conducted rapidly. As the mounting solution is prone to mold growth, it should be stored at 4°C. All procedures were performed at 25°C.
Optical clearing of A. thaliana by TOMEI-II Seedlings and leaves were xed in PBS buffer with 1%, 2%, and 4% FA for 1 h with evacuation for the rst 10 min and then the samples were washed three times in PBS buffer for 5 min. The xed samples were gradually treated in PBS buffer with a graded series of Tissue-Clearing Reagent TOMEI (TCI) for 10 min each (10%, 30%, 50%, 70%, and 100%). The samples were incubated in 100% Tissue-Clearing Reagent TOMEI for 25 h with gentle shaking. Finally, the samples were mounted between the glass slide and the coverslip using 100% Tissue-Clearing Reagent TOMEI. All procedures were performed at 25°C.

Optical clearing of A. thaliana by ClearSee
Seedlings and leaves were xed in PBS buffer with 1%, 2%, and 4% FA for 1 h with evacuation for the rst 10 min and then the samples were washed three times in PBS buffer for 5 min. The xed samples were treated in ClearSee (Fuji lm, Japan) for 26 h with gentle shaking. Finally, the samples were mounted between the glass slide and the coverslip using ClearSee. All procedures were performed at 25°C.
Optical clearing of O. sativa by iTOMEI A 5-day-old leaf was xed with 4% FA for 1 h with evacuation for the rst 10 min and the samples were washed three times in PBS buffer for 5 min. The xed sample was incubated in the decolorization solution for 3 d with gentle shaking. The solution was exchanged with the mounting solution and the samples were incubated in the mounting solution for 1 h with gentle shaking.
The SAMs of rice plants were isolated by hand dissection of wild-type and transgenic plants. The isolated SAMs were xed in PBS buffer with 4% FA for 1 h with evacuation for the rst 10 min. The samples were stained in 0.1% (v/v) SR2200 (solution from supplier was considered as 100%: Renaissance Chemicals, UK) at this step. Then the samples were washed three times in PBS buffer for 5 min for the rst washing and 10 min for the second and third washing. The xed samples were treated in the decolorization solution for 24 h with gentle shaking. The solution was exchanged with the mounting solution and the samples were incubated in the mounting solution for 1 h without shaking. Finally, the samples were mounted between the glass slide and the coverslip using the mounting solution.
Optical clearing of O. sativa by TOMEI-II The SAMs of rice plants were isolated by hand dissection of wild-type and transgenic plants. The isolated SAMs were xed in PBS buffer with 4% FA for 1 h with evacuation for the rst 10 min. The samples were stained in 0.1 % (v/v) SR2200 (solution from supplier was considered as 100 %: Renaissance Chemicals) at this step. Then the samples were washed three times in PBS buffer for 5 min for the rst washing and 10 min for the second and third washing. The samples ware incubated in 10%, 30%, 50%, and 70% TOMEI-II solution for 10 min sequentially. After clearing, the samples were mounted in 70% TOMEI-II.
Optical clearing of M. polymorpha by iTOMEI Thalli, gemmalings, and gemmae were xed in PBS buffer with 1% FA for 1 h with evacuation for the rst 10 min and then the samples were washed three times in PBS buffer for 5 min. The xed samples were treated in the decolorization solution for 24 h with gentle shaking. After washing in PBS buffer for 5 min, the samples were stained in Calco uor White Stain (Calco uor White M2R 1 g/l, Evans blue 0.5 g/l; Merck) or 3-fold diluted CyStain UV Precise P staining buffer (Sysmex, Japan) containing 4′,6-diamidino-2-phenylindole (DAPI) for 10 min and 15 min, respectively. After washing in PBS buffer for 5 min, the samples were incubated in the mounting solution for 1 h with gentle shaking. Finally, the samples were mounted between the glass slide and the coverslip using the mounting solution. For observation of immature gemmae in gemma cups using a two-photon excitation microscope, the whole gemma cup was mounted and the immature gemmae were observed through the side wall of the gemma cup (Fig. 5e).
Optical clearing of M. polymorpha by TOMEI-II Two-week-old gemmalings were xed in PBS buffer with 4% FA for 1 h with evacuation for the rst 10 min and then the samples were washed three times in PBS buffer for 5 min. The xed samples were gradually treated in PBS buffer with a graded series of Tissue-Clearing Reagent TOMEI (TCI) for 10 min each (10%, 30%, 50%, 70%, and 100%). The samples were incubated in 100% Tissue-Clearing Reagent TOMEI for 25 h with gentle shaking. Finally, the samples were mounted between the glass slide and the coverslip using 100% Tissue-Clearing Reagent TOMEI. All procedures were performed at 25°C.

Mice and Surgery
All experimental protocols were evaluated and approved by the Regulation for Animal Research at Tokyo University of Science. All experiments were conducted in accordance with the Regulations for Animal Research at the Tokyo University Science. Adult C57Bl/6J male (4-5 months old) mice were used. Mice were maintained under a 12 h light/12 h dark cycle (light period 07: 30-19:30), and ad libitum feeding and drinking conditions. The plasmid AAV8 CaMKIIa-EGFP was purchased from UNC Vector Core. We used a titer of approximately 1 × 10 12 vg/ml of EGFP viruses in this study. Mice were mounted in a stereotaxic apparatus, anesthetized with pentobarbital (80 mg/kg) and subcutaneously injected carprofen (5 mg/kg) and dexamethasone (0.2 mg/kg). A 2-mm-diameter craniotomy was performed above the hippocampus. A 0.3 µl virus solution was infused using a Hamilton syringe through a glass micropipette at the following coordinates: relative to bregma (mm): anteroposterior axis (AP): −2.0, mediolateral axis (ML): 1.4, and dorsoventral axis (DV): 0.6, 1.2, and 1.7 from dura mater, taken from the mouse brain atlas 28 at a rate of 0.1 µl/min. A glass capillary was left in place for an additional 10 min. A brain sample was harvested 4 weeks after surgery to allow for recovery and su cient expression of genes. Mouse brain sample preparation.
Mice were deeply anesthetized with pentobarbital and transcardially perfused with 4% paraformaldehyde in 0.1 M sodium phosphate buffer (pH 7.4). The brains were excised, sagittally dissected, and post xed with the same xative at 4°C overnight. The xed brains were sectioned at a thickness of 2 mm with a laser blade to make brain slices.
Optical clearing of mouse brain by iTOMEI-B The hemi brain and brain slices were treated with 20% caprylyl sulfobetaine in 0.1 M sodium phosphate buffer (pH 8.0) for 16 h. The samples were next treated with 18.7% (w/w) iohexol and 20% caprylyl sulfobetaine in PBS for 6 h. Subsequently, the samples were treated with 28.1% (w/w) iohexol and 20% caprylyl sulfobetaine in PBS for 6 h and then with 56.2% (w/w) iohexol in PBS for 12 h. Finally, the hemi brains were treated with 56.2% (w/w) iohexol in PBS for 8 h and the brain slices were treated with 70% (w/w) iohexol in PBS for 8 h. The hemi brains were mounted with Omnipaque 350 and observed by a twophoton excitation microscope. The brain slices were mounted with 70% (w/w) iohexol in PBS and observed with a stereomicroscope. All procedures were performed at 25°C.
Optical clearing of mouse brain by SeeDB2 The hemi brain and brain slices were treated with 2% saponin in PBS buffer for 16 h. The samples were next treated with 2% saponin in a 1:2 mixture of Omnipaque 350 (Daiichi-Sankyo, Japan) and water for 6 h. Subsequently, the samples were treated with 2% saponin in a 1:1 mixture of Omnipaque 350 and water for 6 h and then 2% saponin in Omnipaque 350 for 12 h. Finally, the hemi brains were treated with Omnipaque 350 for 8 h and the brain slices were treated with 2% saponin in SeeDB2S for 8 h. The hemi brains were mounted in Omnipaque 350 and observed with a two-photon excitation microscope. The brain slices were mounted with SeeDB2S and observed with a stereomicroscope. All procedures were performed at 25°C.

Declarations
Author contributions:   Table 1) measured using an absorption spectrophotometer. e Fluorescence intensity quanti ed from uorescence images of the wild type and plants expressing GFP before and after incubation in PBS buffer supplemented with sodium deoxycholate (#3) and caprylyl sulfobetaine (#7). j Relative uorescence intensity quanti ed from uorescence images of GFP expressing plants before and after incubation in PBS buffer supplemented with 10%, 20%, or 30% caprylyl sulfobetaine (#7).
Signi cance was determined using the Tukey-Kramer method (a) and unpaired two-sided t-test (b, e-j). Boxplots show the median (middle bar), 25th and 75th percentiles (upper and lower limits of the box), and 1.5 × interquartile range (whiskers). Each data point is represented by an open circle.  percentiles (upper and lower limits of the box), and 1.5 × interquartile range (whiskers). Each data point is represented by an open circle, and the mean is represented by a red cross.

Figure 4
Deep imaging of transparent rice organs using iTOMEI a Leaf of 5-day-old seedling of O. sativa treated with PBS and iTOMEI. b Shoot apical meristem of a plant expressing MADS15-mOrange (yellow) stained with SR2200 (white) and treated with PBS, TOMEI-II, and iTOMEI. Confocal optical sections were captured under the same optical conditions. The arrow indicates strong mOrange signal in the basal region of a hairy bract. The arrowhead indicates mOrange signal at the incipient primary branch meristem. c, d Root tip of a plant expressing DR5rev::NLS-3xVenus (yellow) stained with SR2200 (white) and treated with PBS and iTOMEI. Confocal longitudinal (c) and transverse optical sections (d) of the root are shown. Scale bars = 50 µm. Figure 5