Two-photon imaging with longer wavelength excitation in intact Arabidopsis tissues
In vivo imaging of living organisms is an important tool to investigate biological phenomena. Two-photon excitation microscopy (2PEM) is a laser-scanning microscopy that provides noninvasive, deep imaging in living organisms based on the principle of multiphoton excitation. However, application of 2PEM to plant tissues has not been fully developed, as plant-specific autofluorescence, optically dense tissues, and multiple light-scattering structures diminish the clarity of imaging. In this study, the advantages of 2PEM were identified for deep imaging of living and intact Arabidopsis thaliana tissues. When compared to single-photon imaging, near-infrared 2PEM, especially at 1000 nm, reduced chloroplast autofluorescence; autofluorescence also decreased in leaves, roots, pistils, and pollen grains. For clear and deep imaging, longer excitation wavelengths using the orange fluorescent proteins (FPs) TagRFP and tdTomato gave better results than with other colors. 2PEM at 980 nm also provided multicolor imaging by simultaneous excitation, and the combination of suitable FPs and excitation wavelengths allowed deep imaging of intact cells in root tips and pistils. Our results demonstrated the importance of choosing both suitable FPs and excitation wavelengths for clear two-photon imaging. Further advances in in vivo analysis using 2PEM will facilitate more extensive studies in the plant biological sciences.
KeywordsTwo-photon excitation microscopy (2PEM) Deep imaging Live-cell analysis Autofluorescence Multicolor Arabidopsis thaliana
Recent advances in imaging technique using various fluorescence probes have revealed dynamic events at the subcellular level. Green fluorescent protein (GFP, Shimomura et al. 1962) and various GFP-like fluorescent proteins (FPs) have been widely used in molecular biology studies. In the late 1980s, the confocal microscope (a fluorescent microscope) was developed (Brakenhoff et al. 1985; White et al. 1987) in which a confocal pinhole is used to eliminate the fluorescence from non-focal planes and produce thin (<1 μm) and high-spatial-resolution images (Pawley 2006). Nevertheless, confocal imaging has some limitations, e.g., photobleaching of the fluorophore and phototoxicity to specimens. This problem has occurred via excitation in not only the focal plane, but also above and below the focal planes (Benninger and Piston 2013). In addition, confocal microscopy is less suitable for deep tissue analysis because the signal is degraded due to scattering or absorption from excitation and emission (Centonze and White 1998).
Two-photon excitation microscopy (2PEM) is laser-scanning microscopy using multiphoton excitation processes (Helmchen and Denk 2005; Benninger and Piston 2013). Multiphoton processes offer important advantages for physiological studies in living organisms. For example, out-of-focus photodamage and photobleaching are eliminated, as the multiphoton effect is spatially confined to the femtoliter within the focal plane (König 2000; So et al. 2000). Additionally, 2PEM uses the femtosecond near-infrared pulse laser, which achieves deeper penetration. Thus, 2PEM is a useful tool for performing less-invasive studies (Benninger and Piston 2013) and allows cellular imaging of several hundred to thousand microns inside living organisms (Centonze and White 1998; Helmchen and Denk 2005).
In the animal sciences, 2PEM has been used in a wide range of imaging studies, including those involving the lymphatic organ (Cahalan et al. 2003; Bousso and Robey 2004), kidney (Molitoris and Sandoval 2005), heart (Rubart 2004), skin (Laiho et al. 2005), and brain (Denk and Svoboda 1997; Helmchen and Denk 2002) tissues. Additionally, Kawakami et al. (2013) succeeded in visualizing cerebral cortex layers deeper than 1 mm in a living mouse. Numerous applications of 2PEM in animal tissues have been reported, whereas only a few studies have described its application to plants, although 2PEM technology has been used in Arabidopsis (Tirlapur and König 2001a; Feijó and Moreno 2004; Cheung et al. 2010), waterweed (Tirlapur and König 2001b), and tobacco (Schwille et al. 1999). Cheung et al. (2010) attempted to visualize the pollen tube (PT) growth in a living pistil by using 2PEM with GFP-labeled PTs, but imaging was only successful 120 μm inside the pistils. Plant tissues, such as those in chlorophyll, lignin, cellulose, and phenols (Müller et al. 2013), contain a variety of autofluorescent compounds (Rost 1995) that are causally related to plant autofluorescence and optically dense tissues containing multiple light-scattering structures. Thus, imaging of plant tissue is difficult, even at a depth of just several hundred microns (Feijó and Moreno 2004).
In this study, we analyzed proper excitation wavelengths and FPs for 2PEM in intact Arabidopsis tissues. We demonstrated the advantages of orange FPs with longer wavelength excitation to reduce plant autofluorescence, with 2PEM providing deep and multicolor imaging of Arabidopsis tissues in vivo.
Materials and methods
Plant materials and growing conditions
The Arabidopsis ecotype Columbia (Col-0) was used as the wild-type (WT) plant. We also used transgenic plants carrying LAT52p::FPs (FP: mTFP1, sGFP, Venus, TagRFP, and mRFP), RPS5Ap::tdTomato-LTI6b, GFP-mTalin (Oikawa et al. 2003; Kadota et al. 2009), RPS5Ap::H2B-sGFP (Maruyama et al. 2013), RPS5Ap::H2B-tdTomato (Adachi et al. 2011), and FGR8.0 (Völz et al. 2013). Transformants of LAT52p::FPs and RPS5Ap::tdTomato-LTI6b were prepared as mentioned below. Seeds were germinated on agarose plates at 22 °C under 24-h light. Fourteen-day-old seedlings were transferred to a mixture of Metro-Mix 350 soil (Sun Gro Horticulture, Agawam, MA, USA) and potting compost; seedlings were grown at 22 °C under 24-h light or long-day conditions (16-h light/8-h dark).
The pollen-specific LAT52 promoter (Twell et al. 1989) was used to express FPs in the PT. The complementary DNA (cDNA) of each FP was amplified by polymerase chain reaction (PCR) and cloned into pCR2.1 using the TOPO Cloning Kit (Invitrogen, Carlsbad, CA, USA). The cDNAs were cloned into the spUC vector (Kurihara et al. 2008), which contained the LAT52 promoter, and spUC-LAT52p-FP was cloned into the binary vector pMDC99 (Curtis and Grossniklaus 2003). In RPS5Ap::tdTomato-LTI6b, the RPS5A promoter (Weijers et al. 2001) and tdTomato fused to the start codon of LTI6b (At3g05890) with the (SGGGG)2 linker, which was then cloned into the binary vector pMDC123 (Curtis and Grossniklaus 2003). Transformation of Arabidopsis plants was conducted using the simplified Agrobacterium-mediated transformation method (Narusaka et al. 2010). Then, homozygous and heterozygous transgenic plants were selected, and lines showing strong FP expression were analyzed.
For single- and two-photon imaging, we used a laser-scanning inverted microscope (A1R MP; Nikon, Tokyo, Japan) equipped with a ×25 water immersion objective lens (CFI 75 Apo 25×W MP, WD = 2 mm, NA = 1.10). Two dichroic mirrors, DM495 and DM560, and three emission filters, FF01-479/40, FF02-534/30, and FF01-578/105, were used (Semrock, Rochester, NY, USA). For single-photon excitation, using 488-nm Ar and 561-nm diode lasers, the signal was detected with a photomultiplier tube (PMT) detector. For multiphoton excitation, we used a Ti: sapphire femtosecond oscillator pumped by a pulse laser and a 80-MHz repetition rate Ti:sapphire femtosecond regenerative amplifier (Mai Tai DeepSee 690–1040 nm; Spectra-Physics, Mountain View, CA, USA). In our system, although the Ti:sapphire laser provides 350 nm in the useable tuning range (690–1040 nm, Mai Tai DeepSee; Spectra-Physics), a microscopic system is available for up to 1000 nm. Emitted fluorescence signals were detected using Non-Descanned GaAsP PMT Detectors (Nikon). Laser power delivered to samples and detectors settings were adjusted to avoid saturated pixels and reduce background noise for each experiment. Input and output laser power at each wavelength in this study is shown in Fig. S1. Each image was processed with NIS-Elements AR 4.10 software (Nikon) to create maximum-intensity projection images. Adobe Photoshop CS4 (Adobe Systems, Inc., San Jose, CA, USA) and ImageJ (http://rsbweb.nih.gov/ij/index.html) were used to optimize the images.
Spectrum imaging of mature pollen grains and root tips
Newly opened flowers were taken from LAT52p::FPs, which expresses mTFP1, sGFP, Venus, TagRFP, and mRFP. Pollen grains were spread on the surface of a coverslip, and a mixture of pollen grains expressing the FPs was excited at 800-1000 nm. Roots of 2-week-old seedlings from WT and RPS5Ap::tdTomato-LTI6b were also observed. The infrared laser power was adjusted as follows to compare images: 800 nm (12.0 %), 850 nm (10.0 %), 900 nm (10.0 %), 920 nm (15.0), 950 nm (20.0 %), 980 nm (25.0 %), and 1000 nm (50.0 %). Input and output laser power at each wavelength is shown in Fig. S1. Emitted fluorescence signals were detected using a 32-channel PMT array detector ranging from 463.9 to 649.2 nm at 6.0-nm intervals. Spectrum analysis and adding color were also processed by NIS-Elements (Nikon).
Deep imaging inside an intact root and pistil
Deep imaging was performed using the same microscope system described above. Roots of 2-week-old seedlings from RPS5Ap::H2B-sGFP and RPS5Ap::tdTomato-LTI6b were placed on a coverslip with water. Unpollinated and pollinated (1 day after pollination (1DAP)) pistils from the FGR8.0 plant were placed on a coverslip with culture media (Mizuta et al. 2014). Emitted fluorescence signals were detected using the Non-Descanned GaAsP PMT detector (Nikon). Z-stack images were taken for projection images and optical sections. Images were processed with the NIS-Elements (Nikon) and ImageJ to create maximum-intensity projection images.
Near-infrared two-photon excitation reduced chloroplast autofluorescence
The near-infrared wavelength, 1000 nm, reduced autofluorescence compounds in both pollen and roots
From these results, shorter excitation wavelengths (e.g., 800 and 900 nm) were not better for observing FPs in Arabidopsis tissues. For clear observation without green plant autofluorescence, a longer excitation wavelength (e.g., 1000 nm) was optimal (Fig. 2). Choosing both suitable FPs and excitation wavelengths is important, giving consideration to the targeted tissues and their developmental stages.
Simultaneous two-photon excitation of FPs for multicolor imaging
High-resolution imaging inside roots and pistils
The nucleus of cells in the meristematic zone was clear in two-photon images (Fig. 6b), and similar results were obtained from the root tip of RPS5Ap::tdTomato-LTI6b (Fig. 6c, d). In the two-photon excitation image, the X–Y image contrast at the center of the root tip was higher than that from single-photon excitation, and the cell shape was clearer. In the reconstructed X–Z images, cell shapes at the center of both transition and meristematic zones were clear in the two-photon image (Fig. 6d). Although axial resolution decreased as depth increased, two-photon excitation provided higher resolution than single-photon excitation. These results demonstrated the advantage of 2PEM for internal imaging of intact roots.
Then, 1DAP, the pistil from an FGR8.0 plant was also observed by 2PEM. Autofluorescence from the ovary cell wall was stronger than that for an unpollinated pistil (Fig. 7a) due to silique development. Thus, 990 nm was used for two-photon excitation. In the intact silique, each nucleus in the synergid and egg cells was clearly identified (Fig. 7b). Unexpectedly, a one-cell embryo was evident in some ovules (magnified image in Fig. 7b). These results indicated that 2PEM enables clear discrimination of each nucleus in ovules during fertilization, even deep inside the intact pistil.
A biological sample is composed of heterogeneous tissues in spatial variation, which causes scattering of excitation light, with longer wavelengths being scattered less than shorter wavelengths (Benninger and Piston 2013). In this study, we found that 2PEM of orange FPs at 1000 nm excitation allowed us to clearly see the cellular components in deep plant tissues by eliminating autofluorescence. In plant studies, GFP is generally used for in vivo imaging of cells and tissues to avoid chlorophyll autofluorescence, but some red FPs, such as mRFP, are not appropriate for in vivo imaging in plant tissues. Berg (2004) and Berg and Beachy (2008) reported emission spectra of Arabidopsis autofluorescent compounds such as sporopollenin, cutin, suberin, lignin, and chlorophyll using two-photon microscopy. Most of these emissions were in the blue to green region (∼440 to ∼540 nm), potentially overlapping with cyan, green, and yellow FPs, which suggested that emissions from autofluorescence were also added to these fluorescent spectra. In contrast, some red FPs are known to show lower quantum yield. Especially in two-photon excitation, some red FPs (e.g., mCherry or mStrawberry) show extremely lower quantum yields than single-photon excitation (Drobizhev et al. 2011). Thus, one must identify the FP that is best for 2PEM in plant cells. Orange FPs were useful for clear two-photon imaging in plant tissues, as TagRFP had a higher value than other orange–red FPs under the tuning range (700–1000 nm) of a Ti:sapphire laser (Drobizhev et al. 2009). TagRFP matures quickly and can be expressed in many different types of organisms; it also works well as a fusion protein (Merzlyak et al. 2007). Drobizhev et al. (2009) suggested that tdTomato has the largest value at 1000–1100 nm excitation, which is a three- to fourfold improvement in two-photon brightness compared to enhanced GFPs. Taking these findings into account, orange FPs such as TagRFP and tdTomato seem to be the best for two-photon imaging in Arabidopsis in eliminating plant autofluorescence. Orange fluorophores, with two-photon excitation and longer wavelengths, will enable us to achieve greater imaging depth in plant tissues.
In recent decades, longer wavelengths (i.e., >1000 nm) have been used in two-photon microscopy. The imaging depth at 1280 nm excitation was deeper than at a lower wavelength (Kobat et al. 2009), and lowest light attenuation occurred at 1200–1300 nm excitation (Chen et al. 2002; Tsai et al. 2006). Some reports showed an improved imaging depth using infrared lasers. The far-red FP was excited with a Ti:sapphire laser at 1120–1230 nm excitation from an optical parametric oscillator (Tsai et al. 2006) and a Cr:forsterite laser at −1300 nm excitation (Kobat et al. 2009). Recent reports demonstrated high resolution of subcortical structures within an intact mouse brain using noninvasive three-photon fluorescence microscopy at 1700 nm excitation (Horton et al. 2013). Another study demonstrated that two-photon excitation digital-scanned light-sheet microscopy (2p-DSLM) achieves high-speed imaging and deep penetration in fruit fly and zebrafish (Truong et al. 2011). The higher-order nonlinear excitation with longer excitation wavelengths will facilitate biological investigations at greater depths in plant tissues.
Longer wavelength excitation also provides advantages for simultaneous imaging. FPs having different spectral characteristics are useful for simultaneous imaging of multiple proteins as intracellular biosensors (Verkhusha and Lukyanov 2004). However, the combination of FPs that enables simultaneous imaging is limited. Some improved FPs with a large Stokes shift have been reported (Kogure et al. 2008). These FPs allow us to do multicolor imaging by simultaneous two-photon excitation. Several FPs have an interesting feature in that the two-photon excitation peaks are blueshifted with respect to the single-photon absorption peaks at twice the wavelength (Albota et al. 1998; Drobizhev et al. 2011). This blueshift feature is remarkable in some FPs such as DsRed2 and enhanced GFP (Drobizhev et al. 2011).
Both high quantum yield and a high extinction coefficient allow the use of lower levels of excitation light, which is desirable in reducing specimen damage due to phototoxicity. Recently, multicolor- and four-dimensional imagings were successful when using the narrow range of a band-pass filter to reduce background signals (Kitano and Okada 2012). 2PEM with a dual-laser beam system was also reported (Zinselmeyer et al. 2009). The wavelength scans with various FPs and the use of filters may suffice in finding a good absorption band in the absence of plant autofluorescence. The same types of tissues may vary in autofluorescence, depending on the condition of the plant tissues (Roshchina 2012). Selecting proper FPs with the best excitation wavelength is important, while considering autofluorescence for multicolor deep imaging by 2PEM in living plant tissues.
Two-photon excitation penetrates deeper into the specimen than single-photon microscopy. For single-photon excitation, observing deep inside plant organs is difficult because of scattering structures (multiple cell layers) and components (Fig. 6a, c). In contrast, the signal intensity is retained under the two-photon excitation (Fig. 6b, d). Although the signal intensity of the single-photon images decreases with depth, the depth of imaging penetration from 2PEM was approximately twice that of single-photon imaging (Fig. 6). 2PEM has become a powerful tool for deep imaging of plant’s living tissues. In Arabidopsis, 2PEM facilitated studying the behavior of sperm cells at three steps in an ovule during double fertilization (Hamamura et al. 2011, 2012). Kurihara et al. (2013) isolated an Arabidopsis ovule by 2PEM that allowed clear discrimination of each nucleus in embryo cells at the globular stage. 2PEM is also useful for noninvasive imaging, as it reduces cell damage and enables the imaging of live cells. Cheung et al. (2010) attempted to observe the in vivo PT growth in intact pistils using 2PEM and observed the GFP-labeled PTs deep inside the pollinated pistils. After observations, the observed plant could be further cultured in the greenhouse and seed set was confirmed (Cheung et al. 2010). In the future, studies focusing on individual plant organs by sequential imaging using 2PEM will be able to do, such as from before fertilization to seed development on intact plants. Further advances of live-cell analysis in vivo by 2PEM will facilitate more in-depth analyses in the plant biological sciences.
We thank Dr. A. Kadota (Tokyo Metropolitan University) and Dr. Y. Sato (Nagoya University) for kindly providing seeds of GFP-mTalin and Dr. R. Groß-Hardt (University of Bremen) for kindly providing seeds of FGR8.0. We thank Dr. D. Maruyama (Nagoya University) and Dr. Y. Hamamura (Université de Montréal) for providing plasmids. We thank S. Nasu, T. Nishii and T. Shinagawa for plant culture and S. Nagahara for providing plant material. This work was supported by grants from the Japan Science and Technology Agency (ERATO project to T.H.) and the Ministry of Education, Culture, Sports, Science and Technology, Japan (no. 26840104 to Y.M.).
Conflict of interest
The authors declare that they have no conflict of interest.
- Adachi S, Minamisawa K, Okushima Y, Inagaki S, Yoshiyama K, Kondou Y, Kaminuma E, Kawashima M, Toyoda T, Matsui M, Kurihara D, Matsunaga S, Umeda M (2011) Programmed induction of endoreduplication by DNA double-strand breaks in Arabidopsis. Proc Natl Acad Sci U S A 108:10004–10009PubMedCentralCrossRefPubMedGoogle Scholar
- Benninger RK, Piston DW (2013) Two-photon excitation microscopy for the study of living cells and tissues. Curr Protoc Cell Biol Chapter 4:Unit 4.11.1–24Google Scholar
- Horton NG, Wang K, Kobat D, Clark CG, Wise FW, Schaffer CB, Xu C (2013) In vivo three-photon microscopy of subcortical structures within an intact mouse brain. Nat Photonics 7. doi:10.1038/nphoton.2012.336
- Roshchina VV (2012) Vital autofluorescence: application to the study of plant living cells. Int J Spectrosc 124672Google Scholar
- Rost FWD (1995) Fluorescence microscopy. Cambridge University Press, CambridgeGoogle Scholar
- Tirlapur UK, König K (2001a) Femtosecond near-infrared laser pulse induced strand breaks in mammalian cells. Cell Mol Biol (Noisy-le-grand) 47:OL131–OL134Google Scholar