2024 Volume 49 Issue 2 Pages 135-153
The cellular slime mold Dictyostelium discoideum, a member of the Amoebozoa, has been extensively studied in cell and developmental biology. D. discoideum is unique in that they are genetically tractable, with a wealth of data accumulated over half a century of research. Fluorescence live-cell imaging of D. discoideum has greatly facilitated studies on fundamental topics, including cytokinesis, phagocytosis, and cell migration. Additionally, its unique life cycle places Dictyostelium at the forefront of understanding aggregative multicellularity, a recurring evolutionary trait found across the Opisthokonta and Amoebozoa clades. The use of multiple fluorescent proteins (FP) and labels with separable spectral properties is critical for tracking cells in aggregates and identifying co-occurring biomolecular events and factors that underlie the dynamics of the cytoskeleton, membrane lipids, second messengers, and gene expression. However, in D. discoideum, the number of frequently used FP species is limited to two or three. In this study, we explored the use of new-generation FP for practical 4- to 5-color fluorescence imaging of D. discoideum. We showed that the yellow fluorescent protein Achilles and the red fluorescent protein mScarlet-I both yield high signals and allow sensitive detection of rapid gene induction. The color palette was further expanded to include blue (mTagBFP2 and mTurquosie2), large Stoke-shift LSSmGFP, and near-infrared (miRFP670nano3) FPs, in addition to the HaloTag ligand SaraFluor 650T. Thus, we demonstrated the feasibility of deploying 4- and 5- color imaging of D. discoideum using conventional confocal microscopy.
Key words: fluorescence imaging, organelle, cytoskeleton, small GTPase, Dictyostelium
Live-cell fluorescence imaging is indispensable for studying the functions of various biomolecules and organelles in cells and tissues. Genetically encoded fluorescent proteins (FP) expressed under specific promoters serve as surrogates to monitor gene expression or track intracellular localization as protein tags. Furthermore, FPs can be used as donor-acceptor pairs for Fluorescence Resonance Energy Transfer (FRET) to measure molecular interactions. Model systems in cell and developmental biology have incorporated new-generation FPs with superior stability and high quantum yields (Rodriguez et al., 2017; M. Wang et al., 2023) to encompass a wide range of visible light spectra from blue to near-infrared (NIR). In budding yeast, the nuclei, cell membranes, kinetochores, tubulins, and F-actin were simultaneously imaged using five FPs (Sakai et al., 2021): mTagBFP2 (Subach et al., 2011), mTurquoise2 (Goedhart et al., 2012), mNeonGreen (Shaner et al., 2013), mCherry (Shaner et al., 2004), and iRFP (Filonov et al., 2011). In Neurospora crassa, four FPs, mTagBFP2, mNeonGreen, mApple (Shaner et al., 2008), and iRFP670 (Shcherbakova and Verkhusha, 2013) were employed to visualize the nucleus, nuclear pore complex, microtubules, and cell polarity markers (Z. Wang et al., 2023). In Caenorhabditis elegans, four FPs, mTagBFP2, TagRFP-T (Shaner et al., 2008), mNeptune2.5 (Chu et al., 2014), and CyOFP1 (Chu et al., 2016) with a genetically encoded Ca2+ indicator GCaMP6s (Chen et al., 2013) were expressed simultaneously to identify individual neurons and monitor their activities (Yemini et al., 2021). These studies are only a small representation of works that show the benefit of utilizing a wide range of fluorescence spectra (i.e., the “color palette”) in imaging-based analysis.
Dictyostelium discoideum is a valuable model system for studying cytokinesis, phagocytosis, cell migration, and chemotaxis, along with other processes such as cell–cell signaling associated with its unique aggregative multicellularity. Although various FP has been utilized in D. discoideum (Table S1), fluorescence imaging using three or more fluorescent markers has not been widely adopted. mTurquoise2, the bright variant of CFP fused to cAMP receptor cAR1, expressed together with RegA-mRFPmars and AcaA-GFP highlighted cell–cell heterogeneity at the early gene expression level (Mukai et al., 2016). The CFP- and YFP-based FRET measurements of cytosolic cAMP combined with the detection of RFP-tagged MS2 coat protein (Masaki et al., 2013) or PH domain fused to RFP to monitor cAMP-induced synthesis of mRNA and phosphatidyl inositol (3,4,5) trisphosphate (PIP3) (Kamino et al., 2017) characterized the coupling between cell–cell signaling, gene expression, and cell migration. The FP repertoire has since expanded to include short wavelength blue fluorescent TagBFP (Kundert et al., 2020) and NIR fluorescent protein iRFP (Ohta et al., 2018); however, these two ends of the spectrum have not been utilized in Dictyostelium except in a case of 4-color imaging (Kundert et al., 2020) with mCerulean (Rizzo and Piston, 2005), Superfolder GFP (Pédelacq et al., 2005), Flamindo2 (Odaka et al., 2014), and mCherry.
In D. discoideum, one of the obstacles to live-cell imaging is its high photosensitivity to blue light (Zhang et al., 2023), which precludes the repeated application of strong excitation light. Earlier generations of FPs have relatively low quantum yields (Matlashov et al., 2020; Seo et al., 2024) making long-duration time-lapse imaging unfeasible. Here, we tested newer-generation FPs, either alone or in a tagged format, to increase the practicality of 4- and 5-color fluorescence imaging in Dictyostelium. We examined mTagBFP2, Achilles (Yoshioka-Kobayashi et al., 2020), mScarlet-I (Bindels et al., 2016) and miRFP670nano3 (Oliinyk et al., 2022), for brightness and intracellular localization in tagged form, and compared them with conventionally used FPs. By employing these FPs along with the large Stoke shift LSSmGFP (Campbell et al., 2022), we demonstrated practical 4- and 5-color fluorescence live-cell confocal imaging in D. discoideum.
Plasmids constructed in this study are summarized in Table S2. For constitutive expression, the majority of vectors employed act15 promoter, except some vectors that employed coaA promoter (Table S2). For details, see Supplementary Methods and Supplementary Data. To introduce extrachromosomal vectors, Ax4 cells were electroporated with 1 μg plasmid DNA following the standard protocol (Nellen et al., 1984). Transformants were selected in HL5 growth medium containing 10 μg/mL G418, 60 μg/mL Hygromycin B and 10 μg/mL Blasticidin S either alone or in combination. For 4- and 5- color imaging, we introduced single vectors (G418 resistance) carrying two tags along with the knock-in of mScarlet-I-tagged PKBR1(N150) to act5 locus (see Supplementary Methods).
Cell culture, development, and labelingCells were grown in HL5 medium at 22°C either in shake flasks or in petri dishes. The growth media contained 10 μg/mL G418, 60 μg/mL Hygromycin B, and 10 μg/mL Blasticidin S where appropriate. To image the growth stage, the cells were collected, washed twice, and resuspended in Phosphate Buffer (PB: 12 mM KH2PO4, 8 mM Na2HPO4, pH 6.5). For NIR imaging, cells expressing miRFP670nano3 and HaloTag were labeled with biliverdin and NIR HaloTag ligands, respectively (see Supplementary Methods). A drop of cell suspension was plated either directly onto a 24 × 50 mm coverslip (Matsunami, Osaka, Japan) or a φ25 mm round coverslip (Matsunami) mounted on a metal chamber (Attofluor, Invitrogen, Waltham, MA, USA). The cells were allowed to attach to the substrate before image acquisition. To image slug and fruiting bodies, washed cells were suspended at the density of 2 × 107 cells/mL in PB, and 5 μL of the cell suspension was deposited on an agar plate (2% agar (Bacto) in Milli-Q water). For cellulose labeling, the cells were developed on an agar plate containing a cellulose-staining dye (see Supplementary Methods). After 15–24 h, samples were excised together with the agar sheet and placed upside down onto a coverslip with one or two 50 μm height polyethylene terephthalate spacer rings (vinyl patch transparent Ta-3N, Kokuyo, Osaka, Japan). The inner space of the ring was filled with liquid paraffin (Light 26132, Nacalai Tesque, Kyoto, Japan) (Hashimura et al., 2019) before observation.
The following procedure was followed to observe cells dissociated from the slugs. Growing cells were collected and suspended in PB at a density of 107 cells/mL. Then, 1.5 mL of cell suspension was deposited on a 60 mm agar plate (2% agar (Bacto) in Milli-Q water) containing 25 μg/mL biliverdin. Excess water was removed after the cells attached to the substrate. After 20 h of incubation at 22°C, slugs were collected and passed through a 23G syringe needle (NN-2332R, Terumo, Tokyo, Japan) 10 times; approximately 0.4 mL of cell suspension at the density of 1.5 × 105 cells/mL was loaded into a custom-made microfluidic chamber with a 2.5 μm height observation channel (Fujimori et al., 2019). The cells were allowed to attach for 30 min prior to observation.
Image acquisition and analysisTo observe the cells expressing the two FPs, green and red (Fig. 1, 2, 3A, 4, 5, 6, S1 and S2), an inverted microscope (Ti-E, Nikon, Tokyo, Japan) equipped with a laser confocal point-scanning unit (A1R, Nikon) was employed. A 488 and 561 nm laser was used for excitation, together with a multiband 405/488/561 nm dichroic mirror and 525/50 and 595/50 bandpass filters for detection. For 3D time-lapse imaging in Fig. 5, 13 z-section images were taken at 7-μm-intervals with a piezo stage (Nano-Drive, Mad City Labs, Madison, WI, USA). For observations that included blue and/or near-IR (Fig. 7, 9, 10A, 11, S3, and S4), a multibeam confocal scanning unit (CSU-W1, Yokogawa Electric Corporation, Musashino, Japan) and an electron-magnified CCD camera (iXon 888, Andor, Belfast, Northern Ireland, UK) on an inverted microscope (IX83, Olympus/Evident, Tokyo, Japan) were employed. 405, 445, 515, 560, and 638 nm lasers were used for excitation, together with a Di01-T405/488/568/647 dichroic beam splitter (Semrock, Rochester, NY, USA) and appropriate emission filters. To observe the cells expressing LSSmGFP (Fig. 8 and Fig. 10B), a multibeam confocal scanning unit (CSU-W1, Yokogawa) and CMOS cameras (ORCA-Fusion BT, Hamamatsu Photonics, Hamamatsu, Japan) on an inverted microscope (IX83, Olympus) were employed. To check the LSSmGFP fluorescence (Fig. 8), 405 and 488 nm lasers were employed together with a multiband dichroic mirror (T405/488/561/640 nm) and 447/60 and 525/50 bandpass filters. For 5-color imaging (Fig. 10B), 405, 488, 561 and 640 nm lasers were used for excitation, and 447/60, 525/50, 617/73, and 685/40 bandpass filters were used for detection.
Images were analyzed using ImageJ and R statistical packages. To quantify the mean fluorescence intensity of individual cells, cell masks were obtained from brightfield images using Trainable Weka Segmentation (Arganda-Carreras et al., 2017). To quantify fluorescence of slugs, the mean fluorescence intensity of 30 μm2 rectangular region at the anterior or posterior region of the slug was measured. For 5-color fluorescence imaging, bleed-through of LSSmGFP was corrected (see Supplementary Methods). In all statistical analyses, P values were determined using the Wilcoxon rank-sum test with Bonferroni adjustment.
Measurement of maturation and turn-over of EGFP and AchillesFor the induction experiments (Fig. 3A and B), Ax4 cells were transformed with either a doxycycline (Dox)-inducible GFP expression vector, pDM340 (Veltman et al., 2009a) or a doxycycline-inducible Achilles expression vector (Supplementary Methods); 10 μg/mL (final conc.) doxycycline was added to the shaken cultures and incubated for 0, 3, 6, or 22 h. Cells were then collected and washed twice with PB, resuspended in 0.3 mM EDTA containing PB, and the fluorescence intensity was measured using a flow cytometer (LE-SH800AC, Sony, Tokyo, Japan; excitation 488 nm, 525/50 nm bandpass filter). Gating was set based on both forward scatter (FSC) and backward scatter (BSC). Cells with a fluorescence intensity three times the standard deviation above the mean (mean + 3 × SD) of the 0 h cells were counted as positive. The fluorescence signal at each time point was divided by the intensity at 0 h for normalization.
For labile-Achilles fluorescence quantification (Fig. 3C), cells expressing either Achilles or labile-Achilles under V18 promoter were collected at mid-log phase, washed twice by centrifugation, and resuspended in PB. For cytometry of vegetative cells, washed cells were suspended in PB containing 0.3 mM EDTA for loading. For cytometry of developing cells, washed cells (2 × 107 cells/mL in PB) were plated onto a nitrocellulose filter on top of an absorbent pad soaked in PB and incubated in the dark at 22°C. After 3, 5, 7, and 9 h, the cells were harvested by gentle scraping using a syringe needle and suspended in 20 mM EDTA containing PB in 1.5 mL tubes. Cell aggregates were passed through a 23G syringe needle (Terumo) approximately 20 times before flow cytometry. Data processing and statistical analyses were performed using R statistical package.
Green fluorescent proteins utilized in D. discoideum today are mostly spectral variants of the GFP of Aequorea victoria (Chalfie et al., 1994; Kataria et al., 2013; Mukai et al., 2016) selected for reduced crossover of excitation and emission spectra (Goedhart et al., 2012; Nagai et al., 2002). Additionally, mNeonGreen derived from Branchiostoma lanceolatum which is three times brighter than GFP (Shaner et al., 2013) has been adopted more recently (Honda et al., 2021; Nichols et al., 2020; Paschke et al., 2018). To update the FP repertoire in D. discoideum, we first studied the stability and expression of Achilles, a yellow fluorescent protein derived from Venus, for fast maturation (Yoshioka-Kobayashi et al., 2020). We introduced an extrachromosomal vector that drives Achilles or mNeonGreen expression under the control of the strong act15 promoter. Using a confocal microscope (see Materials and Methods), the fluorescence of both Achilles and mNeonGreen appeared uniformly in the cytosol and nucleus of vegetative cells (Fig. 1A). The cells expressing Achilles appeared significantly brighter than those expressing mNeonGreen (Fig. 1B). In the slug stage, Achilles fluorescence remained bright in the majority of cells and appeared uniform in the cytosol (Fig. 1C and D), whereas very few cells showed mNeonGreen fluorescence (Fig. 1C–E).
Achilles as a superior green fluorescent protein alternative in D. discoideum
(A) Vegetative cells expressing Achilles (upper panels) and mNeonGreen (lower panels) under act15 promoter (left: brightfield channel, right: green channel). Scale bar, 10 μm. (B) Violin plot of the single-cell mean fluorescence intensity distribution of Achilles and mNeonGreen expressing vegetative cells (Achilles, n = 297 cells. mNeonGreen, n = 266 cells). The black line is the median. *: P<10–15. (C) A slug harboring Achilles (upper panels) and mNeonGreen (lower panel) act15 promoter expression plasmids (left: brightfield channel, right: green channel). Scale bar, 100 μm. (D) High magnification images of slugs harboring Achilles (upper panels) and mNeonGreen (lower panels) act15 promoter expression plasmids (left: brightfield channel, right: green channel). The anterior-posterior axis of the slug is from left to right. Scale bar, 10 μm. (E) Violin plot of fluorescent intensities measured in the anterior and the posterior region of a slug (Achilles, n = 10 slugs. mNeonGreen, n = 10 slugs). The black line indicates the median value.*: P<10–4. **: P<0.05. NS: not significant (P>0.05).
Next, we compared Achilles and mNeonGreen protein tags. We first examined cells carrying an extrachromosomal expression vector for Achilles or mNeonGreen fused to the PH domain of Akt/PKB (Meili et al., 1999). The strong coaA promoter (Katoh-Kurasawa et al., 2021) was used for more homogeneous expression than in the act15 promoter in the slug (Fig. S1). Using the same microscopy setup as above, in vegetative cells, PHAkt-Achilles and PHAkt-mNeonGreen fluorescence appeared localized in the pinocytic cups (Fig. 2A), as expected from an earlier observation of PHAkt-GFP (Ruchira et al., 2004). Cells expressing PHAkt-Achilles appeared brighter than those expressing PHAkt-mNeonGreen (Fig. 2B). In the slug stage, PHAkt-Achilles fluorescence was bright with small cell–cell variability, showing enrichment in the front cell–cell contact region (Fig. 2C and D). There was no visible effect on the ability of cells to form slugs or fruiting bodies. In contrast, PHAkt-mNeonGreen fluorescence was almost undetectable, except very rare cells that showed localized patterns similar to PHAkt-Achilles (Fig. 2C–E). This indicates that while Achilles behaves well during development, mNeonGreen may be less stable. Similarly, while both the N- and C-terminal fusions of Lifeact-mNeonGreen exhibited the expected localized patterns in vegetative cells (Fig. S2A and B), the fluorescence was markedly diminished in the slug stage of C-terminal fusion (Fig. S2C–E), suggesting that the free C-terminus of mNeonGreen may be subjected to regulated degradation in D. discoideum. These results indicate that, when studying later developmental stages, Achilles is the green/yellow FP of choice for brightness and stability.
Use of Achilles as a yellow fluorescent protein tag in D. discoideum
(A) Cells carrying coaAp:PHAkt-Achilles (upper panels) and coaAp:PHAkt-mNeonGreen (lower panels) (left: brightfield channel, right: green channel). Scale bar, 10 μm. (B) Violin plot of the single-cell mean fluorescence intensity distribution of PHAkt-Achilles and PHAkt-mNeonGreen expressing vegetative cells (PHAkt-Achilles, n = 402 cells. PHAkt-mNeonGreen, n = 599 cells). The black line indicates the median. *: P<10–15. (C) Slugs (PHAkt-Achilles or -mNeonGreen under coaA promoter. Scale bar, 100 μm. (D) High magnification images of a slug (upper panel PHAkt-Achilles, lower panels PHAkt-mNeonGreen). The anterior-posterior axis of the slug is from left to right. Scale bar, 10 μm. (E) Violin plot of the mean fluorescent intensities of the anterior and the posterior region of a slug (PHAkt-Achilles, n = 12 slugs. PHAkt-mNeonGreen, n = 12 slugs). The black line indicates the median value.*: P<10–5. **: P<0.05. ***: P<0.01. NS: not significant (P>0.05).
Next, we studied whether the fast maturation of Achilles (Yoshioka-Kobayashi et al., 2020) could be used in D. discoideum. To measure the rate of increase in fluorescence intensity, strains carrying a dox-on-inducible vector (Veltman et al., 2009b) with either GFP or Achilles under an inducible promoter were employed. Confocal microscopy images showed a clear increase in fluorescence after induction with doxycycline in both the strains (Fig. 3A). Flow cytometry was performed for quantification (see Materials and Methods). While 64% of the Dox-Achilles cells took only 6 h to show detectable fluorescence (Fig. 3B, red), 22 h was required for a similar percentage of Dox-GFP(S65T) cells (68%) to display a comparable level of fluorescence intensity (Fig. 3B, green). These result suggests that despite the low temperature (22°C) of D. discoideum culture, maturation of Achilles was fast and comparable to that demonstrated in mice (Yoshioka-Kobayashi et al., 2020). Rapid maturation makes Achilles ideal for the sensitive detection of early gene expression in D. discoideum. To further extend its applications, we sought to reduce its half-life for a rapid response to gene downregulation (Sunabori et al., 2008). Based on labile-GFP (Deichsel et al., 1999), we constructed “labile-Achilles” where the N-terminus of Achilles was tagged to the first 33 amino acids of S. cerevisiae ubiquitin (Bachmair et al., 1986). We tested its half-life by expressing labile-Achilles under V18 promoter, whose activity diminishes after starvation (Singleton et al., 1989). Fig. 3C shows a plot of the intensity of Achilles and labile-Achilles-expressing cells developed on a nitrocellulose filter (see Materials and Methods). A difference in the reduction of Achilles and labile-Achilles fluorescence was detectable during the first 3 h after starvation. After 5 h, the fluorescence of labile-Achilles decreased by 46% (Fig. 3C, 5 h), whereas it took approximately 7 h for Achilles to reach a comparable level (Fig. 3C, 7 h). These results suggest that labile-Achilles is a fast-responding alternative to labile-GFP for monitoring promoter activity.
Achilles and labile-Achilles as fast responding reporter genes in D. discoideum
(A) Dox-induced GFP(S65T) (upper panel) and Achilles (lower panel) fluorescence in vegetative cells plated from shaken culture. Cells were incubated with 10 μg/mL Dox for indicated time (0, 3, 6, and 22 h). WT indicates parental Ax4 cells without expression of fluorescence protein. (B) Percentages of the cells displaying GFP(S65T) or Achilles fluorescence after doxycycline (Dox) addition. Data obtained through flow cytometry (see Material and Methods). Dots and error bars represent the average of positive cells and their standard deviation (three independent experimental runs; n = 135,000 cells, 45,000 cells per experiment). (C) Flow cytometry of V18p:Achilles and V18p:Labile-Achilles. The violin plots represent three independent experimental runs, totaling 90,000 cells (3,000 cells per experiment). The boxes and whiskers indicate the mean values and the interquartile ranges, respectively. Outliers are represented as dots. A Welch Two Sample t-test was conducted to compare the two groups, and significance levels are indicated (***: P<2.2e-16).
Next, we explored the use of Achilles and red FP mScarlet-I as cell-type markers. mScarlet-I is a fast-maturing variant of mScarlet that was developed from synthetic constructs to yield high brightness while avoiding dimerization (Bindels et al., 2016). We constructed strains expressing Achilles or GFP(S65T) under the control of a prestalk-specific ecmAO promoter (Williams et al., 1989). These strains also express either mScarlet-I (Bindels et al., 2016) or mRFPmars under D19 promoter (Aubry and Firtel, 1999) whose activity is prespore-specific and is elevated earlier than the ecmAO promoter (Prabhu et al., 2007). mRFPmars is a monomeric RFP derived from Discosoma DsRed (Fischer et al., 2004) that is commonly used in D. discoideum. At 8 h of development, Achilles fluorescence appeared scattered in the aggregating stream (Fig. 4A) when GFP(S65T) fluorescence was barely detectable, except in a few cells (Fig. 4B). The fluorescence of mScarlet-I and mRFPmars was comparable at this streaming stage (Fig. 4C). Given that both Achilles and GFP(S65T) are expressed from an identical plasmid backbone, and that cells with green and red fluorescence appear mutually exclusive, the results suggest that Achilles allows for the earlier detection of cell-type bifurcation.
Yellow fluorescent protein Achilles allows early detection of developmental gene expression in D. discoideum
(A–C) Streaming-stage cells (8 h after starvation) co-expressing green and red FPs under prestalk and prespore specific-promoter. Panels from left to right: Brightfield images, green fluorescence, red fluorescence and merged image. Scale bars, 100 μm. (A) ecmAOp:Achilles (green) and D19p:mScarlet-I (magenta). (B) ecmAOp:GFP(S65T) (green) and D19p:mScarlet-I (magenta). (C) ecmAOp:Achilles (green) and D19p:mRFPmars (magenta).
We further compared the expression of Achilles, GFP (S65T), mScarlet-I, and mRFPmars at later developmental stages under control of the ecmAO promoter. Three strains each carrying two fluorescent reporters were studied: Achilles/mRFPmars, GFP(S65T)/mRFPmars, and Achilles/mScarlet-I (Fig. 5). In early mound formation, before the appearance of the tip, Achilles and mScarlet-I as well as GFP(S65T) fluorescence became clearly detectable and appeared scattered throughout the mound (11 h, Fig. 5 A middle panel: 12 h, Fig. 5B middle panel: 12 h, Fig. 5C middle and lower panels). In contrast, the fluorescence of mRFPmars was not detected, except in a few cells (11 h, Fig. 5A, lower panel; 12 h, Fig. 5B, lower panel). By the time Achilles and mScarlet-I fluorescence appeared concentrated at the tip where prestalk cells accumulated, mRFPmars fluorescence became visible (12 h, Fig. 5A, lower panel; 13 h, Fig. 5B, lower panel), suggesting that the level of mRFPmars fluorescence was similar to that of other FPs. From the tipped mound to the first finger stage (13–15 h), the fluorescence intensity of Achilles in the anterior region was comparable to that of GFP(S65T) (13–14 h, Fig. 5 A middle panel: 14–15 h, Fig. 5B, middle panel). Similarly, the fluorescence intensities of mScarlet-I and mRFPmars were equally bright (13–14 h, Fig. 5A lower panel: 14–15 h, Fig. 5B lower panel: 14–15 h, Fig. 5C lower panel). These results show that early detection of upregulated genes would benefit from the use of Achilles and mScarlet-I.
Red fluorescent protein mScarlet-I allows early detection of developmental gene expression in D. discoideum
(A–C) Cells co-expressing green and red FPs under prestalk-specific promoter. The numbers indicate hours after nutrient removal. Brightfield images (top panels). Maximum intensity projection of z-stacks (middle and bottom panels). Scale bar, 100 μm. (A) ecmAOp:Achilles (green) and ecmAOp:mRFPmars (magenta). (B) ecmAOp:GFP(S65T) (green) and ecmAOp:mRFPmars (magenta). (C) ecmAOp:Achilles (green) and ecmAOp:mScarlet-I (magenta).
Although mScarlet, a brighter but slow-maturing sibling of mScarlet-I, has been used to label the cytosol and microtubules in D. discoideum (Paschke et al., 2018; Schweigel et al., 2022), we tested the suitability of mScarlet-I as a fusion tag and compared it with mCherry, an earlier-generation monomeric RFP variant with an emission peak at 610 nm (Shaner et al., 2004) also commonly used in D. discoideum. Under a confocal microscope using a 561 nm laser for excitation and bandpass filter at 595 nm, mScarlet-I as a standalone appeared uniform in the cytosol and nucleus of vegetative and slug cells (Fig. 6A). mScarlet-I fused to the Ras-binding domain of human Raf1 (RBDhRaf1), which preferentially binds to the GTP-bound form of Ras (Kae et al., 2004) shows clear localization in pinocytic cups (Fig. 6B). mCherry-RBDhRaf1 exhibited the same pattern, which is consistent with an earlier report (Veltman et al., 2016). mScarlet-I-RBDhRaf1 fluorescence was stronger than that of mCherry-RBDhRaf1 in vegetative cells, possibly because of the intrinsic brightness of mScarlet-I on top of the suboptimal excitation wavelength for mCherry (Fig. 6C). In the slug stage, in addition to the expected localization to the front cell–cell contact site (Fig. 6D) (Fujimori et al., 2019), a considerable number of aggregates were visible for mCherry-RBDhRaf1 but not for mScarlet-I-RBDhRaf1 (Fig. 6D). mScarlet-I-RBDhRaf1 fluorescence was approximately 4–5 times brighter than that of mCherry-I-RBDhRaf1 in both the anterior and posterior regions of the slug (Fig. 6E). Although mCherry has the advantage of having a longer wavelength that peaks at 610 nm, thus reducing the spectral overlap with yellow FP, the absence of aggregates makes mScarlet-I a top candidate red FP in D. discoideum.
mScarlet-I is a first-choice red fluorescent protein for protein tagging in D. discoideum
(A) act15p:mScarlet-I cells. Vegetative cells (left panels) and cells in a slug (right panels). Brightfield (left panels) and fluorescence images (right panels). Scale bar, 10 μm. (B) act15p:mCherry-RBDhRaf1 (upper panels) and act15p:mScarlet-I-RBDhRaf1 (lower panel) expressing vegetative cells. Contrast was adjusted as indicated (top right; color bars). Scale bar, 10 μm. (C) Violin plots of the mean fluorescence intensity of mCherry-RBDhRaf1 and mScarlet-I-RBDhRaf1 expressing vegetative cells (mCherry-RBDhRaf1, n = 298 cells. mScarlet-I-RBDhRaf1, n = 333 cells). Black line indicates the median. *: P<10–15. (D) The slug anterior-region of act15p:mCherry-RBDhRaf1 and act15p:mScarlet-I-RBDhRaf1 cells. Magnified view (right panels) of the white-boxed area (left panels). The anterior-posterior axis of the slug is from left to right. Scale bar, 100 μm (left) and 10 μm (right). (E) Violin plots of fluorescence intensity distribution in the anterior and the posterior region of the slug expressing mCherry-RBDhRaf1 or mScarlet-I-RBDhRaf1 (mCherry-RBDhRaf1, n = 11 slugs. mScarlet-I-RBDhRaf1, n = 12 slugs). The black line indicates the median. *: P<10–5. NS: not significant (P>0.05).
Next, we expand the color palette to include blue FPs. Two commonly used blue FPs are mTurquoise2: a blue FP derived from a Venus derivative Super Cyan Fluorescent Protein3A (SCFP3A) for high quantum yield and long fluorescence life-time (Goedhart et al., 2012) and TagBFP which is a monomeric blue FP derived from Entacmaea quadricolor TagRFP (Subach et al., 2008). In D. discoideum, only a few studies have used these FPs alone (Kundert et al., 2020; Paschke et al., 2019) or as a fusion tag (Mukai et al., 2016). Here, we tested further applicability of mTurquoise2 in addition to mTagBFP2 which is a brighter variant of TagBFP (Subach et al., 2011). As a standalone, mTurquoise2 and mTagBFP2 fluorescence appeared uniform in the cytosol and nucleus of vegetative cells (Fig. 7A, upper panels). Although this was also true in slug-stage cells for mTurquoise2 (Fig. 7A, left lower panels), aggregates were noticeable in mTagBFP2 (Fig. 6D; Fig. 7A, right lower panels). The results suggest that use of mTurquoise2 is preferable if aggregates are to be avoided during development.
Blue fluorescent proteins mTurquoise2 and mTagBFP2 for live-cell imaging in D. discoideum
(A–H) The brightfield (left panels) and fluorescence images (right panels) of blue fluorescent protein expressing vegetative cells (upper panels) and cells in a slug (lower panels). The anterior-posterior axis of the slug is from left to right. Scale bar, 10 μm. (A) act15p:mTurquoise2 and (B) act15p:mTagBFP2. (C) act15p:mTurquoise2-Lifeact and (D) act15p:mTagBFP2-Lifeact. (E) coaAp:PHAkt-mTurquoise2 and (F) coaAp:PHAkt-mTagBFP2. (G) act15p:HistoneH1-mTurquoise2 and (H) act15p:HistoneH1-mTagBFP2.
When tagged to Lifeact, we observed strong mTurquoise2 fluorescence throughout the cell cortex of vegetative cells with marked concentration at the pinocytic cups and leading edges (Fig. 7B, left upper panel) in accordance with observations of F-actin binding mCherry-LimEΔcoil (Veltman et al., 2016). In contrast, mTagBFP2-Lifect fluorescence was localized to the rear of the cells (Fig. 7B, right upper panel). This difference was also observed in the slug-stage cells; mTurquoise2-Lifeact appeared uniformly in the cortex, whereas mTagBFP2-Lifeact was more concentrated in the cell rear (Fig. 7B, lower panels). As different actin-binding domains are known to show these localization differences (Lemieux et al., 2014), the results suggest that the fusion of these two FPs likely results in distinctly different affinities for F-actin. Other fusion tags were more forgiving regarding the choice of FPs. The PH domain of Akt/PKB (PHAkt) fused to mTurquoise2 or mTagBFP2 showed a localization pattern similar to that of PHAkt-Achilles for both vegetative and slug-stage cells, with mTurquoise2 exhibiting a higher brightness (Fig. 7C; Fig. 2A, C, and D). They appeared mostly in the pinocytic cups of vegetative cells and at the cell front of the slug. HistoneH1-mTurquoise2 and HistoneH1-mTagBFP2 also showed proper localization to the nucleus in vegetative and slug cells (Fig. 7D). These data demonstrate that mTurquoise2 is well expressed as a standalone fusion protein in the single- and multicell stages of D. discoideum. mTagBFP2 may be less bright and its localization may be more sensitive to the protein tag.
Next, we explored the use of LSSmGFP, a large stokes shift variant of the hyperfolder YFP engineered using a top-down structure-based approach for 405 nm excitation (Campbell et al., 2022). Since green fluorescence emission under 405 nm excitation is not commonly utilized in D. discoideum, we first evaluated autofluorescence in the wild-type Ax4 background. We found no significant autofluorescence except for some small particles whose mean intensity was approximately 5% above the background (Fig. 8A and B). When vegetative cells expressing LSSmGFP under act15 promoter were excited at 405 nm, green fluorescence appeared uniformly in the cytosol, with an average intensity exceeding 150% above the background (Fig. 8C and D). Spectral bleed-through of LSSmGFP fluorescence at 447 nm was undetectable, as determined by comparison with Ax4. When excited at 488 nm, there was minor fluorescence above the background, the intensity of which was approximately 1% of that observed under 405 nm excitation (Fig. 8C and D 488 Ex). The level of cross-excitation was consistent with the reported spectral properties of LSSmGFP (Campbell et al., 2022).
Live-cell imaging using large stokes shift LSSmGFP in D. discoideum
(A, B) act15p:LSSmGFP and (C, D) control Ax4 cells in the vegetative stage. (A, C) From left to right; brightfield, fluorescent images (405 nm excitation, 447/60 nm emission), (405 nm excitation, 525/50 nm emission), (488 nm excitation, 525/50 nm emission). Scale bars, 10 μm. (B, D) Violin plots of the mean fluorescence intensity of Ax4 (B) or LSSmGFP expressing vegetative cells (D) (Ax4, n = 6 cells. LSSmGFP, n = 6 cells). Black line indicates the median. *: P<0.05. **: P<0.01. NS: not significant (P>0.05). (E–G) Fluorescent images of act15p:LSSmGFP (E), act15p:LSSmGFP-Lifeact (F) and act15p:LSSmGFP-HistoneH1 (G). Vegetative cells (top) and the slug cells (bottom). Brightfield (left panels) and fluorescent images (405 nm excitation, 525/50 nm emission; right panels). The anterior-posterior axis of the slug is left to right. Scale bar, 10 μm. (H) 3-color fluorescence images. (Left to right panels) Brightfield, mitochondria marker act15p:GcvH1(N99)-mTagBFP2, nucleus act15p:LSSmGFP-HistoneH1, plasma membrane coaAp:PKBR1(N150)-Achilles and merged fluorescent channels. Scale bar, 10 μm.
Next, we tested the applicability of LSSmGFP to live-cell imaging in D. discoideum. Vegetative cells expressing standalone LSSmGFP showed uniform fluorescence, while slug-stage cells showed somewhat brighter fluorescence in the nucleus compared to the cytosol (Fig. 8E). LSSmGFP-Lifeact was localized to the cell cortex, pinocytic cups in vegetative cells, and leading edges in slug-stage cells (Fig. 8F), similar to mTurquoise2-Lifeact (Fig. 7A). Fluorescence of LSSmGFP-HistoneH1 was restricted to the nucleus, as expected, in both vegetative and slug-stage cells (Fig. 8G). Next, we tested 3-color imaging by constructing a strain that co-expressed GcvH1-mTagBFP2, LSSmGFP-HistoneH1 and PKBR1(N150)-Achilles (Fig. 8H). GcvH1(N99) is the N-terminal mitochondrial localization sequence of GcvH1 gene (Perry et al., 2020). PKBR1(N150) is the N-terminal sequence of PKBR1 that includes a myristoylation signal and labels the plasma membrane (Meili et al., 2000). While these probes showed the expected localization (Fig. 8H; vegetative cells), we noticed with 405 nm excitation some bleed-through from the fluorescence of GcvH1(N99)-mTagBFP2 to the 525 nm emission range (Fig. 8H, cyan and green). While there is some overlap with autofluorescence, the results demonstrate that LSSmGFP is a viable option for multi-color imaging in D. discoideum.
Near-infrared fluorescence imaging in D. discoideumFor NIR fluorescence imaging, iRFP and mIFP derived from Rhodopseudomonas palustris and Bradyrhizobium sp., respectively have been employed in D. discoideum as a standalone (Kundert et al., 2020; Ohta et al., 2018). However, these have limited applications owing to their low brightness, dimerization, and relatively large size. Monomeric NIR FP derived from bacterial phytochrome photoreceptors (Shcherbakova et al., 2015) may overcome these drawbacks. These FPs consist of Per-ARNT-Sim (PAS) and cGMP-specific phosphodiesterase, adenylyl cyclase, and Fhla (GAF) domains, which contain the chromophore biliverdin (Rumyantsev et al., 2015; Wagner et al., 2007). Notably, the truncated variant, miRFP670nano, consisted of only a single GAF domain. Its size (~17 kDa) is close to half that of GFP (27 kDa) and has high molecular brightness (Oliinyk et al., 2019), making it a prime candidate for use in D. discoideum. Here, we prepared cells expressing miRFP670nano3 (Oliinyk et al., 2022)—a brighter variant of miRFP670nano, under act15 promoter. Cells were incubated in culture medium with biliverdin and developed on an agar plate containing 50 μg/mL biliverdin (Kundert et al., 2020). We observed that in vegetative cells as well as in cells within slugs, miRFP670nano3 fluorescence appeared uniform in the cytosol and nucleus (Fig. 9A). HistoneH1-miRFP670nano3 was localized in the nucleus (Fig. 9B). PHAkt-miRFP670nano3 was localized to the pinocytic cup in vegetative cells (Fig. 9C, upper panel) and cell–cell contact sites in slug-stage cells (Fig. 9C, lower panel). As expected, when miRFP670nano3 was expressed under the prespore-specific promoter, miRFP670nano3 signals were detected in the posterior region of the slugs (Fig. 9D). When imaging a fruiting body, caution must be taken to account for weak autofluorescence in the prestalk and stalk regions, whose intensity is approximately 35% of the background (Fig. S3A–C). This may not be an issue when the signal is an order of magnitude higher than the background fluorescence, as in the case of miRFP670nano3 expressed under the control of the D19 promoter (Fig. S3D–F).
Live-cell imaging of D. discoideum in near-IR spectrum
(A–C) Snapshots of vegetative cells expressing act15p:miRFP670nano3 (A), act15p:HistoneH1-miRFP670nano3 (B) and coaA:PHAkt-miRFP670nano3 (C). Brightfield (left panels) and fluorescence images (right panels). Scale bars, 10 μm. (D) A slug expressing miRFP670nano3 under the prespore-specific D19 promoter. Brightfield (left), near-IR fluorescence image (middle), merged image (right). Scale bar, 100 μm. (E–G) Snapshots of vegetative cells expressing act15p:Halo (E), act15p:HistoneH1-Halo (F) and coaA:PHAkt-Halo (G). Left panels, bright field. Right panels NIR fluorescence of Halo-ligand SaraFluor650T. Scale bars, 10 μm. The anterior-posterior axis of the slug is from left to right (A–G).
Multi-color imaging may also be deployed using the self-labeling protein HaloTag, a modified haloalkane dehalogenase that covalently binds to a chloroalkane linker substrate (Los et al., 2008). In D. discoideum, HaloTag-fused PTEN and GFP-tagged PH domains of Akt/PKB have been used together with the red fluorescent TMR-conjugated Halo-ligand to track protein and lipid dynamics in the plasma membrane (Arai et al., 2010; Matsuoka et al., 2012). Since the spectral properties depend on the fluorophore that labels the cell-permeable ligand, one may take advantage of choosing the type of fluorophore and its dosage to avoid spectral overlap with FPs and overstaining (Grimm et al., 2017). Here, we tested the silicone rhodamine-based label SaraFluor 650T, which was excited at 650 nm to obtain NIR fluorescence (Yanagawa et al., 2018). After labeling (see Supplementary Methods), vegetative cells expressing HaloTags showed uniform signals in the cytosol (Fig. 9E, upper panel). Although the cells were labeled only during growth, fluorescence was still visible in the slugs (Fig. 9E, lower panel). PHAkt-Halo and HistoneH1-Halo showed proper localization patterns in both vegetative and slug-stage cells (Fig. 9F and G). These results demonstrate that HaloTag in combination with the ligand SaraFluor 650T provides an alternative to NIR FPs in D. discoideum.
Four-color live-cell fluorescence imaging in D. discoideumWe envisaged that the use of blue and NIR fluorescence in combination with the green and red FP tested above, owing to their brightness and relatively good spectral separation, would render multi-color imaging practical in D. discoideum. We first imaged four FPs, namely, GcvH1(N99)-mTagBFP2, Dajumin-mScarlet-I, PKBR1(N150)-Achilles, and HistoneH1-miRFP670nano3, which target the mitochondria (Perry et al., 2020), plasma membrane (Meili et al., 2000), contractile vacuole (Gabriel et al., 1999) and nucleus (Nagasaki et al., 2002), respectively. We used three plasmids with three selection markers, one of which carried two FP-encoding genes (Supplementary Methods, Table S2) to circumvent the limited number of selection drugs. In vegetative cells, we observed the expected localization patterns in organelles and plasma membranes, with good separation between the respective channels (Fig. 10A, Movie S1). Live-cell imaging revealed the dynamics and positional relationships between these structures. Mitochondria were dispersed in regions free of contractile vacuoles and nuclei (Fig. 10A). Contractile vacuoles showed sequential deformation (Fig. 10A, white arrows). The nucleus continued to change its position between an area adjacent to the plasma membrane (Fig. 10A, 0–50 s) and a more distal area (Fig. 10A, white arrow, 60–70 s), appearing as though it was being pushed away by a contractile vacuole. We also noticed that Dajumin-mScarlet-I fluorescence appeared weakly in the plasma membrane (Fig. 10A), which is in contrast to an earlier report of Dajumin-GFP showing stricter localization to the contractile network (Gabriel et al., 1999).
4- and 5-color fluorescence live-cell imaging in D. discoideum
(A) Vegetative cells expressing 4 FP-tags. From top to bottom: merged channel, act15p:GcvH1(N99)-mTagBFP2 (mitochondria; cyan), coaAp:PKBR1(N150)-Achilles (plasma membrane; yellow), act15p:Dajumin-mScarlet-I (contractile vacuole; magenta), and act15p:HistoneH1-miRFP670nano3 (nuclei; red). White allows: contractile vacuole going through cycles of deformation. Scale bar, 5 μm. (B) Slug cells expressing 5-FP-tags. From top to bottom: merged channel, act15p:GcvH1(N99)-mTagBFP2 (mitochondria; cyan), act15p:LSSmGFP-Lifeact (F-actin; green), act15p:Golvesin-Achilles (Golgi apparatus and ER network; yellow), act5p:PKBR1(N150)-mScarlet-I-2x (plasma membrane; magenta) and act15p:HistoneH1-miRFP670nano3 (nucleus; red). Scale bar, 10 μm.
Four-color imaging of mitochondria, plasma membrane, contractile vacuoles and nucleus in the vegetative cells (related to Fig. 10A)
Fluorescent images of vegetative cells expressing 4-color markers. Fluorescence images of act15p:GcvH1(N99)-mTagBFP2 (cyan), coaAp:PKBR1(N150)-Achilles (yellow), act15p:Dajumin-mScarlet-I (magenta), act15p:HistoneH1-miRFP670nano3 (red) and merged images are shown. Images were acquired at 10 seconds intervals. Scale bar, 5 µm.
Download VideoNext, we imaged five FPs: LSSmGFP-Lifeact, GcvH1(N99)-mTagBFP2, Golvesin-Achilles, PKBR1(N150)-mScarlet-I and HistoneH1-miRFP670nano3. These target the F-actin, mitochondria, Golgi apparatus (Schneider et al., 2000), plasma membrane, and nucleus. To obtain a strain that expressed all five FPs, the PKBR1(N150)-mScarlet-I-2x sequence was first inserted into the act5 locus (Paschke et al., 2018) by homologous recombination, followed by the introduction of other FPs into extrachromosomal vectors (see Supplementary Methods). Fig. 10B shows representative time-lapse images of the cells dissociated from the slugs. Here, LSSmGFP-Lifeact and GcvH1(N99)-mTagBFP2 images were first obtained by 405 nm excitation followed by sequential acquisition of fluorescence images in the yellow, red, and NIR spectra by excitation at 488, 560, and 640 nm, respectively. To isolate LSSmGFP fluorescence in the green channel, weak cyan fluorescence originating from 405 nm-excited GcvH1(N99)-mTagBFP2 was estimated from other channels and subtracted. To isolate Achilles fluorescence in the yellow channel, weak cyan and green fluorescence from 405 nm-excited GcvH1(N99)-mTagBFP2 and 488 nm-excited LSSmGFP-Lifeact were estimated and subtracted (see Supplementary Methods). Although more thorough spectral unmixing should be employed for rigorous signal separation (Leiwe et al., 2024), subtracting these two components results in well-isolated LSSmGFP and Achilles signals. While there was noticeable cell–cell heterogeneity in the fluorescence intensity, for cells that appeared positive in all 5-channels, the FP tags exhibited the expected intracellular localization in the respective emission spectra. Golvesin-Achilles appeared not only in the perinuclear region, as reported for Golvesin-GFP (Schneider et al., 2000) but also in the ER, likely because of the brightness and fast maturation of Achilles. This allowed observation of the movement of the nucleus and mitochondria, which are tightly embedded in the adjacent ER network (Fig. 10B and Movie S2). Four-color time-lapse imaging was performed using PHAkt-mTurquoise2 (PIP3), mCherry-RBDhRaf1 (Ras-GTP), CRIBPakB-Achilles (Rac-GTP), and Halo-Lifeact (F-actin) (Supplementary Text, Fig. S4, Movies S3 and S4), demonstrating the applicability of multi-color imaging to D. discoideum. While the localization patterns of FP-tagged proteins must be further scrutinized by other means, such as immunostaining, these results demonstrate the relative feasibility of multi-color fluorescence imaging for the detailed analysis of collective cell migration in D. discoideum.
Five-color imaging of mitochondria, cytoskeleton, plasma membrane, Golgi apparatus and nucleus in the slug cells (related to Fig. 10B)
Fluorescent images of cells dissociated from slugs expressing the F-actin probe and organelle markers. Fluorescence images of act15p:GcvH1(N99)-mTagBFP2 (cyan, top middle), act15p:LSSmGFP-Lifeact (green, top right), act15p:Golvesin-Achilles (yellow, bottom left), act5p:PKBR1(N150)-mScarlet-I-2x (magenta, bottom middle), act15p:HistoneH1-miRFP670nano3 (red, bottom right), and the merged image (top left) were shown. Images were acquired at 10 seconds intervals. Scale bar, 10 µm.
Download VideoFour-color imaging of PIP3, Rac-GTP, Ras-GTP and F-actin in the vegetative cells (related to Fig. S4A)
Fluorescent images of vegetative cells expressing 4-color markers. Fluorescence images of coaAp:PHAkt-mTurquoise2 (cyan), coaAp:CRIBPakB-Achilles (yellow), act15p:mCherry-RBDhRaf1 (magenta), act15p:Halo-Lifeact (red), and merged images are shown. Images were acquired at 10 seconds intervals. Scale bar, 10 µm.
Download VideoFour-color imaging of PIP3, Rac-GTP, Ras-GTP and F-actin in the slug cells (related to Fig. S4B)
Fluorescent images of vegetative cells expressing 4-color markers. Fluorescence images of coaAp:PHAkt-mTurquoise2 (cyan), coaAp:CRIBPakB-Achilles (yellow), act15p:mCherry-RBDhRaf1 (magenta), act15p:Halo-Lifeact (red), and merged images are shown. Images were acquired at 10 seconds intervals. Scale bar, 10 µm.
Download VideoFluorescent dyes may also be useful in multichannel imaging. In fixed Dictyostelium samples, calcofluor white is conventionally used to label cellulose, which is a major component of the extracellular matrix, referred to as the slime sheath, stalk tube, surface of the mature stalk, and spore cells in the fruiting bodies (Blanton et al., 2000; Yamada and Schaap, 2019). By including calcofluor in the agar substrate, thereby allowing it to diffuse and be absorbed in the sample, we found that it could be utilized for live-cell imaging (Fig. 11A, Movie S5). Furthermore, we observed that Direct Fast Scarlet 4BS (Pontamine Fast Scarlet 4 B), a carbohydrate-binding dye used in plants to visualize cellulose at an excitation wavelength of 561 nm (Anderson et al., 2010) also stained the stalk region of slugs and the fruiting body without affecting development (Fig. 11B). The staining patterns of the two dyes are indistinguishable. In the example shown in Fig. 11C, the stalk was stained with calcofluor white, and two prestalk subtypes, pstAO and pstB, and prespore cells were visualized using FPs in the fruiting body (Fig. 11C).
Combining cellulose staining with FP-tags to image multicellular body of D. discoideum
(A) A migrating slug stained with cellulose-specific dye calcofluor white. Brightfield (top), fluorescence (middle), and merged images (bottom). The slug migrated toward the left side of the images. Scale bar, 100 μm. (B) Time-lapse imaging of culminating D. discoideum stained with cellulose-specific dye Direct Fast Scarlet 4BS (DFS4BS). Brightfield (left), DFS4BS (middle), and merged (right). Slug stage (upper panels) with cellulose-rich trailing slime sheath on agar. A culminant (lower panels) with a stalk tube. Scale bar, 100 μm. (C) Culminating D. discoideum stained with calcofluor white (cyan) expressing Achilles (PstB cell; yellow), mScarlet-I (PstA and PstO cell; magenta), and miRFP670nano3 (prespore cell; red) under ecmB, ecmAO, and D19 promoter, respectively. Scale bar, 100 μm.
Cellulose imaging of extracellular matrix in the slug trail (related to Fig. 11)
Fluorescence images of cellulose in the extracellular matrix labeled with calcofluor white. The left panel of the movie shows a brightfield image. The right panel shows fluorescence images of calcofluor white in agar used for labeling the cellulose in the slime sheath. Images were acquired at 20 seconds intervals. Scale bar, 10µm.
Download VideoHere, we demonstrated that Achilles and mScarlet-I are two-color alternatives for live-cell imaging of D. discoideum both in the vegetative and development stages. Along with brightness, their rapid maturation facilitates the early detection of prestalk-specific gene expression. These characteristics, along with the degradation tag, aids in addressing temporal and spatial patterns of developmental gene expression. This study expands the FP color palette to include the blue and NIR spectra. For blue FP, mTurquoise2 was the protein of choice for D. discoideum because of its high fluorescence and absence of aggregates. While blue FPs have not been well adopted, our results indicate that mTurquoise2 can be considered for practical use, providing access to microscopes with high-quantum-efficiency photomultipliers or sCMOS cameras. Although mTagBFP2 showed some aggregation and interference in the localization of tagged proteins (Lee et al., 2013; Vještica et al., 2020), its use is worth considering when the emission spectrum is more compatible with the DAPI filter set. Regarding NIR imaging, our data indicated that miRFP670nano3 was sufficiently bright, did not aggregate, and exhibited minimal interference with protein localization in D. discoideum. It should be noted that, under our optical setup, the duration of excitation and signal detection had to be extended by approximately five-fold or more compared with those required for other FPs to achieve a comparable S/N ratio. Owing to this limitation, NIR FP should be employed for tagging strongly expressed proteins, reserving brighter FPs for those with weak expression. In the future, this limitation may be partially overcome by introducing the phycocyanobilin synthesis system (SynPCB) (Sakai et al., 2021) as an alternative to exogenous biliverdin.
Using these four emission spectra, we demonstrated the feasibility of four-color fluorescence imaging using confocal microscopy. Using a conventional confocal microscope, image acquisition can be performed serially in tandem or accelerated, for example, by repeating 2-laser excitation 2-emission readout serially, provided that appropriate dichroic mirrors and filter sets are available. Given the bleed-through of LSSmGFP to 525 nm with 405 nm excitation, a combination of FPs for practical application in multi-color confocal microscope imaging without requiring spectral unmixing is mTurquoise2, Achilles, mScarlet-I, and HaloTag with the NIR ligand. Further assessment of spectral unmixing is necessary if five or more emission spectra are required.
Another challenge for deploying multi-color fluorescence imaging in D. discoideum is the choice of vehicles on which genetically encoded FPs and tags are introduced. Here, extrachromosomal vectors that use the origin of replication of naturally occurring plasmids (Veltman et al., 2009b) were employed. Provided that the tags are small enough, one may choose to have a single plasmid carrying two fusion tags (Supplementary Methods), as we did here with GcvH1(N99)-mTagBFP2 and HistoneH1-miRFP670nano3 (Fig. 10), and PHAkt-mTurquoise2 and Halo-Lifeact (Fig. S4). This allowed the expression of four tags with three commonly available selection markers in D. discoideum: G418, Blasticidin and Hygromycin. However, cells expressing all four probes equally at high levels were rare, likely because of heterogeneity in the plasmid copy number. Such heterogeneity was already noticeable in vegetative cells expressing Achilles (Fig. 1A) and miRFP670nano3 (Fig. 9A) alone. This may arise from physiological burden by highly expressed FPs. In addition, in the latter case, variations in the uptake efficiencies of biliverdin and the HaloTag ligand may also underly cell-cell heterogeneity. Cloning a transformant may help improve uniformity in some cases. We should note that maintenance of these plasmids requires the constant selection of multiple drugs, potentially affecting cell physiology. These caveats may be mitigated by knock-in strategies so that all tags are single-copy and the resulting cell lines may be stable in the absence of the selection drug (Paschke et al., 2018). rps30 locus in the duplicated region of chromosome 2 of the Ax3 and Ax4 strains (Eichinger et al., 2005) and act5 locus, which is redundant among 17 actin genes (Joseph et al., 2008) have been shown to be useful as a “safe harbor” (Corrigan and Chubb, 2014; Muramoto et al., 2010). Here, we employed a knock-in at act5 locus to introduce an FP expression cassette and three extrachromosomal vectors to express four FPs for 5-color imaging. Exploring additional safe harbors should facilitate applications where the homogeneous expression of multiple FPs is important. The disadvantage of this approach is the reduced brightness due to the lower expression level and relatively laborious process of generating a knock-in. Other methods include the use of randomly integrated plasmids (Fey et al., 1995) which should also be considered for a high copy number.
This work was supported by JST CREST JPMJCR1923, JSPS KAKENHI JP19H05801, JP23H00384, JP19H05416, HFSP Research Grant RGP0051/2021 to SS. JP20J00751, JP21K15081, and JP23H04304 to HH; JP22K15119 for SK; JP24KJ0051 to HN.
We thank Douwe Veltman for pDM304, pDM323, pDM324, pDM326, pDM340, pDM358 (National BioResource Project Nenkin plasmid; G90008, G90073, G90074, G90067, G90063, G90009), pDM1208, pDM1209 and pDM1489, Chris Thompson for pecmO:labile-GFP, Jonathan Chubb for providing mNeonGreen vector, Robert Kay for pDM1501 (Addgene plasmid; #108997), Atsushi Miyawaki and the RIKEN BRC for providing Achilles through the National BioResource Project of the MEXT, Japan. The plasmids constructed in this study are available from the NBRP-Nenkin stock center (https://nenkin.nbrp.jp/).
cyan fluorescent protein
CRIBCdc42/Rac interactive binding
FPfluorescent protein
GFPgreen fluorescent protein
NIRnear-infrared
PHPleckstrin homology
PIP3phosphatidylinositol (3,4,5) trisphosphate
RBDRas-binding domain
RFPred fluorescent protein
YFPyellow fluorescent protein
