Label-free imaging and analysis of subcellular parts of a living diatom cylindrotheca sp. using optical diffraction tomography

Highlights • Subcellular parts of a living diatom cell was well visualized by digital holographic microscope.• Subcellular parts could be identified as differences of refractive indexes.• The observation was achieved without any pre-treatment of the living cell.


Method details
Quantitative phase imaging (QPI) techniques have been exploited to investigate both morphology and optical properties of living cells, due to its label-free and quantitative imaging capability. [1][2][3][4][5] For example, 3D QPI techniques, such as optical diffraction tomography (ODT) or holotomography (HT), reconstructs three-dimensional structures of a living cell from multiple 2D optical field images with various illumination angles. The reconstructed 3D refractive index (RI) distributions of a cell can provide both the morphological information about the cell and the quantification of the dry mass of the cell and the dry mass concentration. [6][7][8][9][10] Although QPI provides several attractive benefits in bioimaging -without any pre-treatments such as fixation of cells or staining with dyes, its applications to the field of biology and medicine as a standard method has been stymied due to the limitations of instruments, including limited spatial resolution and unfavorable usability in biological laboratories. For the last decade, there have been significant developments in the instrumentations of QPI. [2,7,[11][12][13][14][15][16][17] For example, the spatial and temporal resolution of QPI had been improved and reached the limitations of farfield diffraction resolution and the speed of a high-speed image sensor. [18][19][20][21][22] Recently, many researchers have intensively reported biological applications of QPI, in particular, medical applications.  [4] Stretching cells were one of the unique applications. [26] On the other hand, red blood cells (RBCs) have also been interests for many researchers. [6,[27][28][29] Furthermore, the quantitative imaging capability of QPI and its synergistic applications in combination with artificial intelligence (AI) became one of the hot topics. [30][31][32][33][34][35][36] One of the important but not fully explored applications of QPI is the study of diatoms. Diatoms are major photosynthetic planktons that produce 20% of primary products on the earth. Previously, the shape of diatoms was measured using holographic microscopy [37][38][39]. However, the spatial resolution had been mostly limited. Recently, Zetsche reported improved QPI images of diatoms in details in 2014 with an enhanced spatial resolution using DHM, but it only addressed 2-D information. [1] More recently, due to the advances in ODT and HT techniques, 3-D RI tomograms of diatoms have been measured. Simon et al. measured the 3D RI tomogram of a non-living Coscinusdiscus sp. diatom using ODT. [40] Lee et al. reported the 3D RI tomograms of phytoplanktons, including Naviculaes, Pseudonitzschia, and Thalassiosira . [41] However, the 3D RI tomogram of living diatoms and their quantitative analysis have not been performed.
In this paper, we report the measurements of 3D RI tomograms of living cylindrotheca sp. diatom in seawater using HT. We also perform the quantitative analysis to the measured RI tomograms and systematically investigate the subcellular parts of the diatom, including frustules, protoplasm, vacuole, and chloroplast of a living diatom cell.
Unlabeled live diatom samples were prepared and imaged using HT. The sample preparation is as follows. An aliquot of seawater Katase-Enoshima area was diluted into Guillard f/2 culture medium and incubated for several weeks at 20 • C. Then, an appropriate volume of the incubated seawater was injected in an imaging dish (TomoDish, Tomocube Inc., Daejeon, South Korea), and then, directly observed using an HT instrument (HT-2, Tomocube Inc., Daejeon, South Korea). For the observation, an objective lens was immersed in water, not in a culture medium. The theoretical spatial resolution of the used HT system is 110 nm and 360 nm for lateral and axial direction, respectively. In order to validate the HT system, we measured latex beards (6 μm in diameter, 19102-2, Polysciences, Inc., Warrington, PA) as a standard sample. The representative RI tomograms of the latex beads are presented in Supplementary Information Fig. S1. The averaged RI value of the beads was measured as 1.594 ± 0.001 ( n = 61). The value is well corresponded to the RI values of the beads provided by the manufacturer (1.5983). Figure 1 shows a representative RI tomogram image of a diatom cell, cylindrotheca sp. The unique morphology of the cell was well observed. In addition, because an aliquot of seawater was directly observed, unknown fiber-like objects were also clearly observed.
In order to investigate the internal structures of the diatom cell, the measured RI tomogram was analyzed based on their spatial distributions and RI values. Figure 2 represents four different 3D structures of the cell. From the measured RI tomogram, four different 3D images were retrieved based on specific ranges of the RI values. The black volume in Fig. 2 (a) represent cellular parts that had RI values between 1.352 and 1.357. For visualization purpose, the extracted volumes in Fig. 2 are also presented as video clips (see Supplementary Information Fig. S2). The morphology of the black volume is almost similar to the well-known frustule structures of cylindrotheca sp. [42,43] This means, diatom frustule morphology could be visualized by observing RI distribution.
The green, blue, and red volumes in Figs. 2 (b) −(d) were generated with the RI ranges 1.363 −1.381, 1.388 −1.395, and 1.403 −1.436. The green, blue, and red structures could be understood as protoplasm, vacuole, and chloroplast based on their morphologies. In the case of the red parts, 'volume' and 'volume of color' in the table indicated almost the same values. On the other hand, in blue parts, 'volume' value was much larger than 'volume color' value. If we think red and blue parts revealed vacuole and chloroplast, the difference is reasonable for the following reason. The chloroplast is totally chloroplast, so that RI values might be uniform. That is why 'volume' and 'volume of color' showed similar values. Vacuole membrane and vacuole inside might have different RI values. So, when Frustules, protoplasm, vacuole, and chloroplast of a living diatom cell were well distinguished based on the HT observation without any staining and isolation. The RI values of each part were also estimated from the measured RI tomograms. Our results clearly revealed future potentials of HT method in researches of diatom and other microorganisms.