Py3-FITC: a new fluorescent probe for live cell imaging of collagen-rich tissues and ionocytes

Polypyrrole-based polyamides are used as sequence-specific DNA probes. However, their cellular uptake and distribution are affected by several factors and have not been extensively studied in vivo. Here, we generated a series of fluorescence-conjugated polypyrrole compounds and examined their cellular distribution using live zebrafish and cultured human cells. Among the evaluated compounds, Py3-FITC was able to visualize collagen-rich tissues, such as the jaw cartilage, opercle and bulbus arteriosus, in early-stage living zebrafish embryos. Then, we stained cultured human cells with Py3-FITC and found that the staining became more intense as the amount of collagen was increased. In addition, Py3-FITC-stained HR cells, which represent a type of ionocyte on the body surface of living zebrafish embryos. Py3-FITC has low toxicity, and collagen-rich tissues and ionocytes can be visualized when soaked in Py3-FITC solution. Therefore, Py3-FITC may be a useful live imaging tool for detecting changes in collagen-rich tissue and ionocytes, including their mammalian analogues, during both normal development and disease progression.


Introduction
Distamycin is a small antibiotic molecule containing tri-N-methylpyrrole (Py) that binds to the AT-rich DNA minor groove [1]. Poly-Py-based polyamides are used as sequence-specific DNA probes [2]. Therefore, fluorescence-conjugated poly-Py-based polyamides are thought to be useful as nuclear-specific probes. However, several factors influence the nuclear uptake of poly-Pybased polyamides [3]. Best et al. [3] demonstrated that poly-Py-based polyamides with good nuclear uptake affinity have several principal features, such as an eight-ring polyamide DNA-binding domain, one or more positive charges, and a conjugated fluorescein fluorophore. Long-length poly-Pybased polyamides are not always effective. Inversely, short polyamides could bind multiple genomic regions [4]. The addition or deletion of a single charge might change the nuclear uptake effect [5]. A moderate length of fluorophore polyamide has many advantages: increasing stabilization, increasing binding intensity, cell nuclear uptake and protein affinity. Therefore, the localization of fluorescence-conjugated polypyrroles in living animals is unclear. We used three types of FITC-conjugated compounds to assess the effect of pyrrole number on their localization in living animals.
Collagen accounts for a large proportion of body mass and is the most common component of the extracellular matrix [6]. Multiple types of collagen play essential roles in body structure. In animals, triple helices assembled in individual collagen molecules create the sophisticated, hierarchical structure that forms the fibres and networks observed in tissue, bone and basement membranes [7]. Furthermore, collagen is involved in diseases; for example, fibrosis is a disease process involving the destruction of healthy tissue through the deposition of a collagen-rich extracellular matrix [8]. Therefore, visualizing collagen localization or dynamics provides essential information for basic studies or early disease diagnosis. Several methods for visualizing collagen have been developed. Second harmonic generation (SHG) imaging can be used to visualize unstained collagen [9,10]. However, a specific device is needed for SHG imaging. Using collagen-binding peptides or proteins, peptide-or protein-based fluorescence probes have been generated [11,12]. These probes are used for staining living cells and dissected living tissue, and can be intravenously injected to detect fibrotic organs. Nevertheless, these high-molecular-weight probes may not be suitable for the whole-body imaging of living animals. A low-molecular-weight probe, Col-F, has been developed [13]. However, because it is a fluorescein-conjugated physostigmine, an inhibitor of acetylcholinesterase, its side effect potential needs to be determined before it can be applied to living animals. A transgenic strategy has also been used for imaging collagen [14]. However, making transgenic lines takes a long time, and only a specific type of collagen can be labelled this way. Low-molecular-weight, low-toxicityinducing and pan-collagen imaging probes would have advantages for general use in collagen imaging.
In zebrafish (Danio rerio), ionocytes transport ions through respective sets of ion transporters and control body osmotic pressure. In zebrafish embryos, several types of ionocytes, such as H + -ATPase-rich (HR) cells, Na + -K + -ATPase-rich (NaR) cells, Na + -Cl − cotransporter (NCC) cells, solute carrier 26 (SLC26) cells and K + -secreting (KS) cells, are in the skin and gill [15]. HR cells secrete acid, take up Na + and excrete NH 4 + and are similar to mammalian renal proximal tubular cells [15]. Therefore, an easy imaging tool for HR cells may make these cells useful as model systems of mammalian renal proximal tubular cells. Mitochondrial staining reagents, including MitoTracker, can stain HR cells and NaR cells [16]. Sodium green is a Na +dependent fluorescent reagent that can detect HR cells [17]. However, sodium green is used only in living cells and is not applied to fixed samples, such as immunostaining. To our knowledge, there is no specific and useful probe for HR cells.
Here, we show that newly generated MeO-Py3-GABA-FITC (Py 3 -FITC) can be used to detect collagen-rich tissue in living zebrafish embryos. Py 3 -FITC can be used to visualize cartilage, notochord, fin rays, etc. In addition, Py 3 -FITC also stains HR cells. Py 3 -FITC can be used for double staining with Alizarin Red S or antibodies. These data indicate that our newly developed Py 3 -FITC probe is a powerful tool for imaging collagen-rich tissue and HR cells.

Zebrafish
AB (wild type), Tg (fli:dsRed);casper [18,19], and mib ta52bmutant zebrafish lines were used. The casper zebrafish [19] were obtained from the Zon laboratory. Tg (fli:dsRed); casper zebrafish were generated by crossing Tg (fli:dsRed) with casper zebrafish. The mib ta52b mutant was described previously [20]. The zebrafish were raised and maintained under standard conditions [21] with approval by the Institutional Animal Care and Use Committee at Chiba University (nos 1-174, 2-178). Males and females of these strains were mated to generate embryos.

Experimental methods for synthesis
All reactions were performed with dry solvents under argon, and the reagents were purified by the usual methods. Column chromatography purification was performed with silica gel 60 N (spherical, neutral 63-210 mesh). Preparative thin-layer chromatography (PTLC) separations were carried out on 2 mm E. Merck silica gel plates (60F-254). Nuclear magnetic resonance (NMR) spectra were recorded on JEOL-JMN-ECS400 and ECZ400 spectrometers operating at 400 MHz to obtain 1 H-NMR spectra. Data from the 1 H-NMR analysis were reported as follows: chemical shift (δ ppm), multiplicity (s = singlet, br-s = broad singlet, d = doublet, t = triplet and m = multiplet), coupling constants (Hz) and integration. High-resolution mass spectra were measured on a JEOL AccuTOF LC-plus JMS-T100 LP spectrometer (ionization method: electronic supplementary material). MeO-Py 4 -NH2 was obtained [23]. All of the reagents were purchased from Sigma-Aldrich.

Storage and usage of Py 3 -FITC
Py 3 -FITC is stable for over a year in a dry state at -30°C. Stock solution (10 mM in DMSO) in a dark place is stable for several months at −30°C by preventing freezing and thawing and was diluted immediately before use. It would probably cost 500 USD or less to synthesize several hundred milligrams of Py 3 -FITC using our protocol from commercially available materials.

Fluorescent reagent staining
Fluorescein, 5-FITC-labelled benzylamine, Py 3 -FITC and Py 4 -FITC were dissolved in DMSO to generate 10 mM stock solutions, which were stored at −30°C. Fertilized eggs were incubated in E3 solution (5 mM NaCl, 0.17 mM KCl, 0.33 mM CaCl 2 and 0.33 mM MgSO 4 , with 0.0002% methylene blue). For early-stage staining, the chorion was removed by pronase (0.2 mg ml −1 in E3) treatment (28.5°C for about 7 min) at one-cell stage. For later-stage staining, the chorion was manually removed by forceps at 1 day post-fertilization (dpf ). The embryos were treated with 10 µM fluorescein, 5-FITC-labelled benzylamine, Py 3 -FITC or Py 4 -FITC in E3 1 dpf. Then, 0.003% 1-phenyl-2-thiourea (PTU) (w/v) was added to E3 to prevent pigmentation. A fluorescent reagent was added to 24-well dishes and protected from light during the incubation. Immediately before the observation was performed, the fluorescent reagent-treated embryos were washed with E3 medium three times. For imaging at 24 h post-fertilization (hpf ), embryos were treated with each compound for 5 h beginning at 19 hpf. Then, the embryos were anaesthetized by 0.016% (w/v) tricaine (MS-222) in E3 and mounted in 0.7-2% low-melting-temperature agarose in E3 medium with 0.016% tricaine on a glass slide using a silicon rubber ring made in-house. Z-stack images were taken by an SP8 or LSM780 confocal laser scanning microscope (Leica, Germany; Zeiss, Germany). For fixed zebrafish staining, embryos were first fixed with 4% paraformaldehyde (PFA) for 1 h. Embryos were washed and removed into acetone for 1 h (−30°C). Then embryos were stained with 10 µM Py 3 -FITC overnight.

Double staining with Py 3 -FITC and Alizarin Red S
The embryos that were treated with Py 3 -FITC from 1 dpf to 6 dpf, as mentioned above, were transferred into 10 µM Py 3 -FITC and 50 µM Alizarin Red S (Sigma-Aldrich) solution for 24 h and then washed with E3 medium three times. Z-stack images were taken as mentioned above.

Scale staining
Zebrafish (three-month post-fertilization, 3 mpf) were anaesthetized with 0.016% tricaine added to the water. The scales were picked up by forceps and treated with 10 µM Py 3 -FITC (diluted in 1 × PBS) for one hour and then washed with 1 × PBS overnight. Images were taken by a stereoscopic fluorescence microscope and digital camera (Leica M165FC and DFC7000T, Germany).
For the antibody staining of scales, 3 mpf zebrafish were anaesthetized with 0.016% tricaine placed in the water system, and the scales were picked up by forceps. The scales were washed with PBS and treated with 10 µM Py 3 -FITC in PBS for 1 h. The solution was removed, and the scales were washed with PBS twice and incubated overnight. The scales were fixed with 4% PFA for 1 h at 4°C. Then, the scales were incubated overnight with anti-collagen type I (diluted 1 : 100, mouse monoclonal; SP1. D8, DSHB) antibody or anti-collagen type II (diluted 1 : 100, mouse monoclonal; II6B3, DSHB) antibody at 4°C. The scales were washed with PBST (0.05% Triton X-100), incubated overnight with Rhodamine Red-X conjugated anti-mouse IgG (diluted 1 : 400, cat. no. 715-295-151, Jackson ImmunoResearch) at 4°C and washed with PBST (1 × PBS, 0.5% Triton X-100) twice for 15 min each time. Images were taken by a stereoscopic fluorescence microscope with a digital camera (Leica M165FC and DFC7000T, Germany).

Data analysis
Z-stack confocal images were processed and analysed, and z-projection views were generated by ImageJ software royalsocietypublishing.org/journal/rsob Open Biol. 11: 200241 (NIH). Illustrator (Adobe) and C4D (Maxon Computer) were used to construct illustrations of the opercles. The whole-view image of the embryo was composited by using Image Composite editor 2.0. (Microsoft).

Py 3 -FITC can stain specific tissues of the zebrafish embryo
We synthesized 5-FITC-labelled benzylamine and FITCconjugated pyrrole polyamide trimers and tetramers, Py 3 -FITC and Py 4 -FITC, respectively (electronic supplementary material, figure S1a). Specifically, we depict the synthesis of Py 3 -FITC and its 1 H-NMR map in detail (figure 1a; electronic supplementary material, figure S1b,c). We examined Py 3 -FITC staining from the 1-cell stage to the shield stage (6 hpf ). We did not observe any stained cells during these early developmental stages in Py 3 -FITC-treated embryos (electronic supplementary material, figure S2a,b). Therefore, we focused on stained zebrafish embryos at later stage. We treated the zebrafish embryos from 1 dpf to 12 dpf with these compounds at a 10 µM concentration, and no obvious toxic effect was observed. When compared with fluorescein, these compounds selectively stained some body parts In addition, we performed Triton X-100 treatment and 10 µM Py 3 -FITC staining after 4% PFA fixation at 4°C overnight or at room temperature for 1 h. After fixation, any specific tissues of zebrafish could not be stained with Py 3 -FITC (electronic supplementary material, figure S9a,b).

HR cells were stained with Py 3 -FITC
The distribution pattern of Py 3 -FITC-positive cells on the yolk sac and the trunk region is similar to that of ionocytes (figure 1d,e) [16]. Previously, ionocytes were identified in the skin of zebrafish embryos as early as 24 h after fertilization, and they consist of several types of ionocytes: HR cells, NaR cells, NCC cells, slc26 cells and KS cells [15,16,24]. NR cells are widely present in both the ventral and dorsal regions of the trunk and tail, whereas HR cells are more restricted to ventral regions [16]. Since Py 3 -FITC-positive cells were observed mostly in ventral regions, we performed double staining with V-type H+ ATPase antibody and Py 3 -FITC. We found that Py 3 -FITC-positive cells merged with the cells detected by the V-type H+ ATPase antibody, suggesting that Py 3 -FITC can selectively stain HR cells ( figure 2a-e).
In addition, Py 3 -FITC staining was present neither on the membrane of HR cells nor in nucleus. An intracellular dot-like staining pattern was observed at high magnification of Py 3 -FITC and V-ATPase double staining (figure 2f-k). These data suggest that HR cells might actively be stained by Py 3 -FITC.
Notch signalling is involved in the differentiation of HR cells, and the number of HR cells is increased in the Notch signalling-deficient mutant mind bomb ta52b (mib ta52b ) [25]. Therefore, we examined the numbers of Py 3 -FITC-positive cells and found that they were significantly increased in mib ta52b mutants at 24 hpf (figure 2l,m).
Together, these results suggest that Py 3 -FITC-stained cells on the surface of the yolk sac and trunk region are HR cells.
Taking advantage of live staining with Py 3 -FITC, compared to Alcian blue, which requires tissue fixation, we conducted double live staining using Py 3 -FITC and Alizarin Red S, which stains mineralized bone [30]. We observed distinct patterns of staining with Py 3 -FITC and Alizarin Red S, and as expected, they stained cartilage and bone, respectively (figure 3c-j ). Double staining revealed the morphology of early-developing dermal bone, the opercle, which forms a

The heart chamber, the bulbus arteriosus, is stained with Py 3 -FITC
We found that Py 3 -FITC stained the oval ball-like region of the heart after 5 dpf through 12 dpf (electronic supplementary material, figure S4a-l). In zebrafish embryos, the ventricle and atrium are formed from the heart tube in the early developmental stage. Subsequently, the b.a. is formed in the anterior region of the ventricle after 2 dpf [32]. To investigate the FITCpositive region of the heart in detail, we performed Py 3 -FITC staining of Tg(fli:dsRed);casper, which expresses DsRed in endothelial cells [18,19]. Time-lapse video imaging showed rhythmic signals of Py 3 -FITC and DsRed (electronic supplementary material, movie S1). Furthermore, we found that Py 3 -FITC stained outside of the DsRed area in the b.a. (figure 4a-j ). The b.a. is a unique elastic structure in fish that functions to dampen the pressure pulse generated by the ventricle as a form of circulatory adjustment [33]. The external layer of the b.a. is known to be formed by numerous cell types embedded in a collagenous matrix [33]. Considering that Py 3 -FITC stained the cartilage, notochord and fin rays, which are also collagen-rich tissues, Py 3 -FITC may have a high affinity for the collagen-rich matrix.

Py 3 -FITC stains collagen-rich tissues
We next compared regions stained by Py 3 -FITC with those stained by anti-collagen type I or anti-collagen type II antibody. Most of the tissues stained by anti-collagen type I and II antibodies were also stained positively by Py 3 -FITC ( figure 5a,b 00 ). The notochord was stained with Py 3 -FITC, which was consistent with both anti-collagen type I and II antibodies positive region (electronic supplementary material, figure S11a-d 00 ). Collagen type I-stained areas overlapping with Py 3 -FITC-stained areas were mainly located in the ear, eye, jaw skeletal elements (m, ch and p), fin elements (sco and ed), b.a., and other areas, as shown in figure 5c,d 00 . On the other hand, the collagen type II-stained areas overlapping with Py 3 -FITC-stained areas were mainly observed in jaw skeletal elements and the tail fin ray region ( figure 5b-b 00 , e-f 00 ). However, not all collagen I-and II-positive areas overlapped with the Py 3 -FITC-positive areas (figure 5a-f 00 ). Additionally, we examined whether Py 3 -FITC-stained adult scales since the zebrafish scale contains collagen type I [34]. We found that most of the collagen type I-positive areas were also Py 3 -FITC positive, and collagen type II was not detected in the scales (electronic supplementary material, figure S11e-f 00 ). Furthermore, we examined the relationship between Py 3 -FITC and collagen using cultured human cells. In HFL-1 cells, transforming growth factor-β1 (TGF-β1) stimulates collagen type I expression [35]. Thus, we performed double staining with anti-collagen type I antibody and Py 3 -FITC in HFL-1 cells treated with or without TGF-β1. We observed that Py 3 -FITC staining was increased in the cytoplasm, as collagen type I accumulated in the cytoplasm after TGF-β1 stimulation (figure 6a-f ). We also performed Py 3 -FITC staining in living HFL-1 cells. Nuclei of living HFL-1 cells could not be stained by Py 3 -FITC clearly (electronic supplementary material, figure S12a, b), and the signal intensity was weaker compared to that in fixed samples. Therefore, the permeabilization step after fixation may be important for Py 3 -FITC staining in cultured cells. Moreover, we did not observe obvious staining with 5-FITC-labelled benzylamine, while Py 4 -FITC showed strong signals in the cell nucleus for both living cells or fixed cells (electronic supplementary material, figure S13a-h). However, neither 5-FITC-labelled benzylamine nor Py 4 -FITC showed an increase in their cytosolic accumulation in response to TGF-β1 stimulation, as was observed for Py 3 -FITC (electronic supplementary material, figure S13a-h).
Taken together, these in vivo and in vitro results suggest that Py 3 -FITC can be used to detect collagen-rich fibrous tissues.
4. Discussion 4.1. Py3-FITC is a new low-molecular-weight fluorescent probe for collagen-rich tissues and others Several fluorescent probes for collagen detection have been developed thus far [11][12][13]. However, to our knowledge, previous probes have not yet been used for detecting collagen in living animals. Transgenic lines in which fluorescence-fused collagen protein is expressed have also been reported [14]. However, in these transgenic lines, the collagen subtypes detectable are limited.
Our study revealed that Py 3 -FITC, a low-molecularweight compound consisting of FITC-conjugated pyrrole trimers, might stain several types of collagen. Consistently, in HFL-1 cells, the Py 3 -FITC signal was only partially merged with that of collagen type I. Collagen type III and type V are also induced by TGF-β1 stimulation in HFL-1 cells [36,37]. Therefore, an increase in other collagens, such as type III and type V, also contributes to an increase in Py 3 -FITC-positive components in TGF-β1-stimulated HFL-1 cells.
Collagen forms a right-handed triple helix [7]. Py 3 -FITC might recognize and interact with this collagen three-dimensional structure. The number of pyrrole repeats may be critical for the interactions with collagen because Py 4 -FITC, which has only one additional pyrrole, rarely stained collagen-rich tissues. For wider applications, we created other colour fluorophore-conjugated Py 3 -based compounds, such as Py 3 -Cy3 and Py 3 -rhodamine. However, these red fluorophore-conjugated Py 3 compounds did not stain the same tissue stained by Py 3 -FITC. These results indicate that the chemical properties of fluorescein might also contribute to the selectivity of Py 3 -FITC staining. In addition, poly-Pybased polyamides are known to be able to interact with the minor groove of the DNA duplex. Fluorophore-conjugated poly-Py-based polyamides, which are expected to localize to nuclei, have been reported [38]. By contrast, some dye conjugates localize mainly in the cytoplasm, not the nucleus, suggesting that localization in the cell depends on the cellular uptake rate and cell type [3,39].
We showed that Py 3 -FITC can localize to nuclei in HFL-1 lung fibroblast cells but rarely in living zebrafish embryos. Moreover, fixed embryo nuclei were not stained by Py 3 -FITC. This indicates that fixation may be necessary but not sufficient for nuclear staining in zebrafish embryos with Py 3 -FITC. Further studies are needed to clarify these discrepancies.

Py 3 -FITC is a useful tool for live imaging
Py 3 -FITC induces low toxicity and therefore is useful for applications in the fields using live in vivo imaging. Three type tissues were stained in living organisms in this study: cartilage, notochord and scale; heart; and ionocytes.  Cartilage, notochord and scale. Alcian blue and Alizarin Red are two colour stains that are often used to estimate bone formation. However, double staining requires cell fixation and involves multiple staining steps [40]. Py 3 -FITC and Alizarin Red S double staining is a new and easy live imaging method for detecting bone formation in zebrafish embryos. A caveat for using this staining method is that the Py 3 -FITC staining pattern is different from the Alcian blue staining pattern to some extent. For instance, ceratobranchial cartilage is clearly stained by Alcian blue staining but not by Py 3 -FITC. By contrast, fin rays are stained by Py 3 -FITC but not by Alcian blue [40]. These differences might be due to the difference in their targets, namely, Alcian blue stains acidic mucopolysaccharides [41] and Py 3 -FITC may interact with collagen-rich tissues.
Dupret et al. showed that craniofacial cartilage is Alcian blue-negative at 2 dpf [42], but it could be stained with Alcian blue at 3 dpf [43]. These previous results indicate that the earliest emergence of cartilage corresponds to the first detection of Py 3 -FITC signal. In addition, the notochord could be stained with anti-collagen II antibody at 24 hpf [44]. Py 3 -FITC can detect the notochord structure at 24 hpf or even   Heart. Py 3 -FITC can visualize the beating of the b.a. The b.a. is a collagen-rich compartment of the fish heart and functions as a modulator to provide steady blood to gills [45]. It has been reported that WGA lectin also stains b.a. mesenchyme at 72 hpf. However, WGA lectin is a protein and not suitable for use as a live imaging tool. By taking advantage of Py 3 -FITC, cardiac beating and rhythm in live zebrafish can be recorded.
Ionocytes. By contrast to the above two tissues, HR cells are not collagen-rich. Py 3 -FITC stains only HR cells and can be used in live samples. Py 3 -FITC did not stain the surface of the HR cells. Confocal analysis showed many intracellular organelles stained with Py 3 -FITC. However, the cell types with Py 3 -FITC-positive intracellular organelles and the extent to which Py 3 -FITC uptake is linked to HR cells function are currently unknown. Ionocyte analogues are also present in mammals, such as renal proximal tubular cells [15], and Py 3 -FITC may be used to help detect and study them. These issues remain to be addressed in future research.

Conclusion
We generated a new fluorophore-conjugated polypyrrole, Py 3 -FITC. In living zebrafish embryos, Py 3 -FITC can detect collagen-rich tissues such as cartilage, b.a., notochord, fin ray, etc. Py 3 -FITC also detects HR cells, which are one type of ionocytes. It has low toxicity, and the staining method is rapid and easy to use. Py 3 -FITC is a useful live imaging tool for detecting collagen-rich tissue morphogenesis and HR cells distribution. Collagen is related to wound healing and fibrosis. Therefore, with modification, Py 3 -FITC might be used as a new diagnostic or therapeutic agent.
Ethics. All animal procedures were approved by the Chiba University Institutional Animal Care and Use Committee (nos 1-174, 2-178).
Data accessibility. This article has no additional accessible data. Figure 6. Py 3 -FITC can be used to detect TGF-β1-induced collagen type I. HFL-1 cells were treated with TGF-β1 for 0 h (a-c) or 48 h (d-f ) and then stained with Py 3 -FITC (a,d ) and collagen type I antibody (b,e), and their merged images are shown (c,f ). Collagen type I expression was increased by 48 h of TGF-β1 treatment. The cytoplasmic signal of Py 3 -FITC was also increased by TGF-β1 and was partially merged with that of collagen type I (c,f ). Scale bar in a, 50 μm.