Infrared laser–induced gene expression in single cells characterized by quantitative imaging in Physcomitrium patens

Abstract

A spatiotemporal understanding of gene function requires the precise control of gene expression in each cell. Here, we use an infrared laser–evoked gene operator (IR-LEGO) system to induce gene expression at the single-cell level in the moss Physcomitrium patens by heating a living cell with an IR laser and thereby activating the heat shock response. We identify the laser irradiation conditions that provide higher inducibility with lower invasiveness by changing the laser power and irradiation duration. Furthermore, we quantitatively characterize the induction profile of the heat shock response using a heat-induced fluorescence reporter system after the IR laser irradiation of single cells under different conditions. Our data indicate that IR laser irradiation with long duration leads to higher inducibility according to increase in the laser power but not vice versa, and that the higher laser power even without conferring apparent damage to the cells decelerates and/or delayed gene induction. We define the temporal shift in expression as a function of onset and duration according to laser power and irradiation duration. This study contributes to the versatile application of IR-LEGO in plants and improves our understanding of heat shock-induced gene expression.

A study optimizes a system effectively inducing gene expression by infrared laser heating and characterizes the induction profile of single cells in the different conditions through quantitative imaging analysis in the moss Physcomitrium patens.

Introduction

Flexible manipulation of gene expression at the single-cell level enables precise functional analysis of genes and their encoded proteins. Infrared laser–evoked gene operator (IR-LEGO) allows local induction of gene expression in living organisms¹. IR-LEGO consists of an optical microscope and an irradiating system with a 1480 nm infrared (IR) laser, whose energy is well absorbed by water molecules and generates heat. The application of a focused IR laser through microscope optics immediately achieves a peak temperature within a few msec in proportion to the laser output (in milliwatts) in the irradiated cells or local area^1,2. The expression of a gene of interest (GOI) can be made inducible by IR laser irradiation by placing the GOI under the control of the promoter of a HEAT SHOCK PROTEIN (HSP) gene. In eukaryotes, the transcription factor HEAT SHOCK FACTOR 1 (HSF1) is the key regulator of the heat shock response (HSR)^3–5. The trimerization of HSF1 as temperature increases results in the marked induction of heat shock–responsive genes, including HSP genes. HSPs serve as molecular chaperones, playing pivotal roles in repairing heat-denatured proteins to maintain protein homeostasis. Altogether, IR-LEGO can achieve spatiotemporal control of gene expression through the HSR with an IR laser in any organism.

IR-LEGO has been employed in the liverwort Marchantia polymorpha⁶, the flowering plant Arabidopsis (Arabidopsis thaliana)^7–9, fruit fly (Drosophila melanogaster)¹⁰, nematode (Caenorhabditis elegans)¹, medaka fish (Oryzias latipes)^7,11, zebrafish (Danio rerio)^7,12, Iberian ribbed newt (Pleurodeles waltl)¹³, and African clawed frog (Xenopus laevis)^13,14. In C. elegans, the combination of lower power and longer duration in IR laser irradiation yielded a higher probability of gene induction with lower incidence of cell death¹. However, detailed profiles of induced gene expression and cell damage with IR-LEGO remain largely unclear in the previous studies. In particular, the temporal profiles of gene expression will be informative for the application of IR-LEGO to any organism.

Employing IR lasers to manipulate gene expression is especially advantageous in plants, since IR light, unlike visible light, is unlikely to perturb photomorphogenesis or photosynthesis. Here, we tested IR-LEGO in the moss Physcomitrium patens. The availability of the soybean HSP promoter for inducible gene expression was previously demonstrated in the whole plant body of P. patens^15–20. P. patens presents suitable features for the application of IR-LEGO, such as the filamentous structure of its protonemal tissue composed of chloronemal and caulonemal cells^21,22 and the ability to grow along the surface of a glass-bottom dish^23–25. These features minimize IR absorbance by water in out-of-focus areas, leading to more reproducible and efficient heating of living cells. In this study, we interrogated the optimal range of IR laser irradiation conditions to achieve higher efficiency of gene induction with lower invasiveness in single cells. We quantitatively characterized the gene induction profile of single cells subjected to different irradiation conditions using a fluorescent reporter driven by the HSP promoter. Our results allow estimation of the onset time and duration time of gene induction, which was delayed and extended, respectively, in the higher IR laser powers. The data presented here will be instrumental to designing future experiments for transcriptional induction and provide fundamental insights into the HSR.

Results

Gene induction under the promoter of Glycine max HEAT SHOCK PROTEIN 17.3B

We established P. patens lines expressing the yellow fluorescent protein gene mCitrine^26,27 fused with nuclear localization signals (NLS)²⁸ and the β-glucuronidase (GUS)²⁹ (NLS-mCitrine-GUS; hereafter NmYG), under the control of the promoter from soybean (Glycine max) HEAT SHOCK PROTEIN 17.3B (ProGmHSP17.3B)^15–20 (Fig. 1a). This promoter has previously been shown to confer heat-inducible expression in P. patens. The presence of GUS in addition to NLS enhances NmYG localization in nuclei^19,30–33. The nuclear localization of the fusion protein facilitates the identification of each cell affected by IR laser irradiation, as described later, by concentrating the fluorescence signal in the nucleus and by blocking movement through plasmodesmata^24,25. These transgenic lines are referred to as ProGmHSP17.3B:NmYG hereafter. We first tested the relative abundance of NmYG transcripts in response to an increase in temperature from 25 °C to 37 °C. RT-qPCR analysis showed that NmYG transcripts accumulate in a heat-inducible manner in the ProGmHSP17.3B:NmYG transgenic lines #5, #7, and #11 (Fig. 1b and Supplementary Fig. 1a). We also confirmed the heat-inducible detection of mCitrine fluorescence in these lines. mCitrine fluorescence was undetectable at the normal growth temperature of 25 °C even in the line #7 with relatively higher expression level than the lines #5 and #7 (Fig. 1c and Supplementary Fig. 1a). We detected fluorescence in the nuclei of protonemal cells 12 h after 30 min treatment at 37 °C in all the lines (Fig. 1c). As the fluorescence intensity in line #11 appeared higher than that of lines #5 and #7, we used ProGmHSP17.3B:NmYG line #11 for later experiments.

Fig. 1. Transgene and NmYG inducibility of the transgenic plants.

a Diagram of the construct used in this study. Red, promoter of GmHSP17.3B (ProHSP); yellow, nuclear localization signals derived from simian virus 40 (NLS); green, mCitrine; sky blue, β-glucuronidase (GUS); gray, terminator of Pisum sativum ribulose 1,5-bisphosphate carboxylase small subunit gene (TerTbcS). b Relative NmYG transcript abundance following incubation at 25 °C or 37 °C for 1 h. The transcript levels were normalized against those of α-tubulin (TUA1) genes³⁰, based on the ΔΔCt method. Data are means ± standard deviation (SD, n = 3 for biological triplicates, with technical triplicates in each sample). The P-values determined by the two-sided Welch’s t-test are shown in the graph. c Bright-field (BF) and fluorescence images 12 h after incubation at 30 min incubation at 25 °C or 37 °C. Scale bars, 100 μm.

Efficiency of single-cell gene induction with IR laser heating

We applied the IR-LEGO system to ProGmHSP17.3B:NmYG line #11 and observed the mCitrine fluorescence 12–24 h after initial IR laser irradiation of chloronemal cells (Fig. 2a). The activity of the HSR is determined by the combination of heating temperature and duration of exposure^34–39. This combination is equivalent to the IR laser power and irradiation duration in the IR-LEGO system¹. To identify optimal conditions for the induction of gene expression in chloronemal cells, we assessed mCitrine fluorescence and cell damage induced after IR laser irradiation at different settings for laser power and irradiation duration. Here, the mCitrine fluorescence was determined by the detectable signals in the nucleus of the irradiated cell (Fig. 2a, b). We also defined cell abnormality as immobilized chloroplasts or expanded vacuoles following irradiation with the IR laser (Fig. 2c, Supplementary Fig. 1a, b, and Supplementary Movie 1). Rupture and shrinkage of cells were scored as cell death (Fig. 2d, Supplementary Fig. 1c, and Supplementary Movie 1). Of note, cell abnormality and death included cells with and without mCitrine fluorescence.

Fig. 2. Efficiency of the induction of the fluorescent reporter gene expression and cell damage in different conditions of IR-LEGO.

BF images and maximum intensity projection (MIP) from seven z-stack fluorescence images in chloronemal cells of ProGmHSP17.3B:NmYG line #11 exhibiting mCitrine fluorescence (a), no fluorescence (b), abnormality (c), or death (d) 12 h after IR laser irradiation. mCitrine fluorescence is shown with the fire lookup table of Fiji. Two fluorescence images without (middle) or with intensity enhancement (bottom) are shown (a–d). Scale bars, 50 μm. Relative distribution of chloronemal cells showing mCitrine fluorescence (red), no fluorescence (blue), or abnormality/death (gray) after IR laser irradiation for 1 s (e) or 60 s (f). The number of IR laser–irradiated cells is shown above the bars (n). Different lowercase letters indicate significant differences (P < 0.05) between groups after Bonferroni test.

We first applied a 1 s irradiation of the IR laser, in accordance with previous reports^{1,6,7,10–14,40}. We detected mCitrine fluorescence in the cells irradiated with the IR laser set to a power level from 12.0 to 20.0 mW without cell abnormality or death (from 2.8% to 59.6% of all cells) (Fig. 2e). Exposure to a 22.0-mW IR laser resulted in the highest percentage (70.7%) of inducible gene expression among 1-s irradiation conditions, although we observed a low frequency of cell abnormalities (1.7%; n = 1) (Fig. 2e). Raising the laser output to 24.0 mW resulted in a lower percentage of cells with detectable mCitrine fluorescence, dropping to 61.1%, while the percentage of cell abnormalities and cell death increased to 8.3%.

Since cells of the land plants generally do not move or grow within minutes, the irradiation duration can be extended to 60 s in plant cells^8,9. According to this extension, we alternatively reduced IR laser power required for induction of NmYG (Fig. 2e, f). Chloronemal cells showed mCitrine fluorescence without cell abnormality or death when irradiated with an IR laser power level from 6.0 to 12.0 mW (26.7% to 93.3% of fluorescent cells) (Fig. 2f). We detected the highest percentage of cells with mCitrine fluorescence at a 12.0 mW output for the IR laser (93.3%), with fewer fluorescent cells at 14.0- and 16.0 mW outputs (69.0% and 42.6%) together with more examples of cell abnormality or cell death (14.3% and 53.7%) (Fig. 2f). We also detected mCitrine fluorescence in the neighboring non-irradiated cell when the target cell was irradiated at an output of more than 14.0 mW for 60 s (Supplementary Fig. 3a), but never at a 12.0 mW output. Unlike 1-s irradiations, a 60 s irradiation achieved the highest efficiency of gene induction and the absence of concomitant cell abnormality or cell death in the irradiated cells with an IR laser output of 12.0 mW. Of the 60 s irradiation conditions tested, a laser output of 10.0 mW was the second most efficient (87.1%) and was more efficient than any 1-s irradiation conditions (≥70.7%). These results indicate that a 60-s irradiation is a more stable method for inducing gene expression with less invasiveness compared to 1 s irradiations.

We also compared the above results separately in the first (apical stem cells), third, and fifth cells of protonemata for each IR laser output (Supplementary Fig. 3). In almost all conditions, the percentage of cells possessing detectable mCitrine fluorescence was not significantly different among cells at different positions, indicating no positional effect on the HSR, at least for the first five chloronemal cells. We observed faint mCitrine fluorescence signals more frequently in the fifth cells (83.3%) than in the first or third cells (8.3% and 0.0%, respectively) with a 60 s irradiation at 6.0 mW output (Supplementary Fig. 3b). This difference might reflect a biological fluctuation at the threshold of inducibility. These results demonstrate that IR-LEGO is a viable system for local gene induction in P. patens, as reported in other plants and animals^1,6–14.

Growth after IR laser irradiation

To more precisely assess cell damage caused by IR laser irradiation, we asked whether cell growth occurred in apical stem cells after the heating. Accordingly, we performed time-lapse imaging for 24 h after initial IR laser irradiation with a range of output values resulting in no cell death (Fig. 2e, f). We determined that irradiated apical cells either grew continuously following irradiation, initially stopped before resuming growth, or remained in growth arrest throughout the time-lapse imaging (Fig. 3a-c and Supplementary Movie 2). In the 1 s irradiation conditions, the percentage of growth arrest was strongly influenced by the IR laser power output, with 0.0% of growth arrest at 18.0 mW, 25.0% at 20.0 mW, and 71.4% at 22.0 mW (Fig. 3d). One cell out of the 16 subjected to irradiation at 20.0 mW output displayed abnormal multinucleation (Supplementary Fig. 4a). These data suggest that 1 s irradiation conditions are highly invasive with higher IR laser powers. By contrast, the 60 s irradiation conditions were associated with relatively lower percentages of growth arrest even with higher IR laser powers settings, with 10.0% of growth arrest at 8.0 mW, 0.0% at 10.0 mW, and 27.8% at 12.0 mW. A large fraction (38.9%) of cells also underwent a temporary growth arrest followed by growth recovery at the 12.0 mW output level of the IR laser (Fig. 3d). These data indicate that 60-s irradiation conditions are less invasive than 1 s conditions, allowing cell growth under these highly inducible conditions (Fig. 2c and Fig. 3d). We note that although we showed images with fluorescence for growth assessment (Fig. 3a–c), growth arrest did not always coincide with the detection of mCitrine fluorescence.

Fig. 3. Growth of apical stem cells of protonemata after IR laser irradiation.

BF images and MIP from seven z-stack fluorescence images of apical cells displaying continuous growth (a), recovery after growth arrest (b), or growth arrest (c). The times after IR laser irradiation are indicated. White dotted lines indicate the tip position of the apical stem cells just after IR-laser irradiation (0 min). Scale bars, 50 μm. d Relative distribution of chloronemal cells showing continuous growth (green), recovery after growth arrest (light green), or growth arrest (gray). The number of IR laser–irradiated cells is shown below the bars (n). The lowercase letters indicate significant difference (P < 0.05) between groups after Bonferroni test.

Time-lapse imaging and quantification of the induction level in single cells

To investigate the timing and levels of gene induction in different conditions with the IR-LEGO system, we performed time-lapse imaging and measured mCitrine fluorescence intensity in the fifth cells from the apex of protonemata mainly under conditions yielding no cell death. Because these cells did not show cell elongation or division during the time-lapse imaging in our conditions, their influence was negligible for the measurement of intensity. After IR laser irradiation, we observed an elevation in mCitrine fluorescence intensity, followed by a gradual decline over time (Fig. 4, Supplementary Fig. 5, and Supplementary Movies 3 and 4). For the 1 s irradiations, we detected similar maximum mCitrine fluorescence intensity among the output levels 18.0, 20.0, and 22.0 mW (Fig. 4g–i). However, under the 60 s irradiation conditions, the maximum intensity rose with higher IR laser power (Fig. 4j–l). These data indicate that a 60 s irradiation is advantageous for the higher induction.

Fig. 4. mCitrine intensity after irradiation with the IR laser.

BF images (top) and MIP (bottom five columns) of the fifth cells of protonemal tissues in ProGmHSP17.3B:NmYG exhibiting mCitrine fluorescence after IR laser irradiation with 18.0 mW (a), 20.0 mW (b), or 22.0 mW (c) output for 1 s or with 8.0 mW (d), 10.0 mW (e), or 12.0 mW (f) output for 60 s. BF images show chloronemal cells just after IR laser irradiation. Times after IR laser irradiation in fluorescence images are indicated in minutes. Scale bars, 50 μm. g–l Mean fluorescence intensity of mCitrine in 20 min intervals in chloronemal cells of ProGmHSP17.3B:NmYG after IR laser irradiation with 18.0 mW (g), 20.0 mW (h), or 22.0 mW (i) output for 1 s or with 8.0 mW (j), 10.0 mW (k), or 12.0 mW (l) output for 60 s were shown in black plots. SD is indicated by the shaded area. The p₁, p₂, and p₃ values were determined by fitting the sigmoidal function to the kinetics of mean fluorescence intensity (solid red curves to the peak and dotted red curves after the peak). n values are given for the 1-s irradiation with 18.0 mW (n = 5), 20.0 mW (n = 6), and 22.0 mW (n = 15) and 60 s irradiation with 8.0 mW (n = 18), 10.0 mW (n = 12), and 12.0 mW (n = 14). Residual sums of squares (RSSs) are also indicated.

We detected faint mCitrine fluorescence or a signal that was almost indistinguishable from chlorophyll autofluorescence at power levels of 14.0 and 16.0 mW with 1 s irradiation (Supplementary Fig. 5a, b, h, i) and 6.0 and 7.0 mW with 60 s irradiation (Supplementary Fig. 5d, e, n, o), suggestive of detection limits. We thus considered these conditions inappropriate for the induction of gene expression. Relatively invasive conditions of 24.0 mW output for 1-s irradiation and 14.0- or 16.0 mW output for 60 s irradiation were prone to result in mCitrine fluorescence levels that deviated from the population average (Supplementary Fig. 5m, s, t). In these conditions, induction levels in some cases exceeded those seen with lower laser outputs (Supplementary Fig. 5h–t). Induction levels in the other cases were close to the detection limit (Supplementary Fig. 5m, s, t). We also noticed apparent delays in the increase of mCitrine fluorescence intensity (Supplementary Fig. 5m, s, t). Under 60 s irradiation at a laser output of 14.0 or 16.0 mW, we observed several examples of multinucleation or cytoplasmic fluorescence, most likely reflecting the breakdown of the nuclear envelope (Supplementary Fig. 4b-d). Taken together, our results indicated that these invasive conditions are not appropriate for induction of gene expression.

Quantitative analysis of gene induction in single cells

To decipher the temporal profile of induced mCitrine fluorescence from the single cells observed following irradiation with the IR laser under different conditions, we explored the kinetics characteristics of fluorescence by a curve fitting analysis. Elevation of gene expression over time can be defined as a sigmoidal increase^41–44. In individual HCT116 colon cancer cells, the intensity of enhanced green fluorescent protein gene (EGFP) under the control of the human HSP70 promoter also fit a sigmoidal curve⁴⁵. Accordingly, we performed curve fitting with a sigmoidal function to the fluorescence kinetics of each cell (see Methods section) (Fig. 4g–l). We extracted three parameters that indicate the maximum (p₁), gain (p₂), and timing (p₃) of induced NmYG protein accumulation. The p₁ parameter represents the peak of mCitrine fluorescence intensity during the given time window, p₂ refers to the speed shown in reaching the peak, and p₃ denotes the beginning of the increase in mCitrine intensity (Supplementary Fig. 6a-c). We then compared these fitting parameters among all single cells treated with the various induction conditions. In agreement with the above results of normalized mCitrine fluorescence intensity (Fig. 4g–l), the p₁ values obtained here further supported the idea that mCitrine intensity significantly increases according to the increase in the IR laser power following a 60 s irradiation but not a 1 s irradiation (Fig. 5a). p₂ values were higher at lower IR laser power output for both irradiation durations (Fig. 5b), indicating that gentle heating induces NmYG expression, and thus NmYG accumulation more rapidly. p₃ increased with a higher IR laser power output for 1 s irradiations and slightly for 60 s irradiations (Fig. 5c), suggesting that stronger heating might delay the transactivation of ProHSP through the HSR. Integrating these quantitative analyses strongly indicates that a 60 s irradiation with 10.0- to 12.0 mW output enables more efficient induction of genes in the IR-LEGO system than all other conditions tested.

Fig. 5. Quantitative characterization of the elevation in mCitrine intensity after different conditions of IR laser irradiation.

Distribution of p₁ (a), p₂ (b), and p₃ (c) derived from mCitrine intensity in individual irradiated cells in different conditions with box-whisker and violin plots. Violin plots show the density distribution of the data. In the box-whisker plots, gray horizontal lines, boxes, and whiskers indicate median, interquartile range, and data range, respectively. Different lowercase letters indicate significant difference (P < 0.05) between groups after Kruskal-Wallis test. n values are given for each group: 1 s irradiation with 18.0 mW (n = 3), 20.0-mW (n = 4), and 22.0 mW output (n = 10) and 60-s irradiation with 8.0 mW (n = 11), 10.0 mW (n = 12), and 12.0 mW (n = 13) output.

To estimate and compare the actual time of NmYG expression induced by IR laser irradiation, we extracted the onset and duration of the elevation in mCitrine intensity from the obtained fitting parameters. The derivative of the sigmoidal function corresponds to the velocity of the increasing mCitrine intensity (Fig. 6a). In the derivative, the starting point of the increase and peak width thus indicate the onset and duration of NmYG accumulation, respectively. We evaluated the starting point and peak width with the “tangent method” employed in the field of chromatography⁴⁶. The starting point is obtained from the point at the intersection of the left tangent from the derivative and the baseline (y = 0) (Fig. 6a). The peak width is acquired as the length between the two intersection points of the tangents from the derivative and the baseline (Fig. 6a). We note that the starting point and peak width include the time to complete transcription, translation, protein folding, and maturation of the fluorescent protein. The starting point depends on the fitting parameters p₂ and p₃, and the peak width on p₂ (see Methods section). Accordingly, a 60 s irradiation with 8.0 mW output resulted in a significantly earlier initial rise of mCitrine intensity (23 min in the median) compared to all other suitable conditions (Fig. 6b). As the IR laser power increased, the onset was delayed in both the 1- and 60 s irradiation conditions. This finding indicates that more gentle heating is responsible for the earlier onset of mCitrine accumulation. The duration of mCitrine intensity elevation was around 5–7 h (Fig. 6c). This duration was extended by up to 2 h for the conditions with higher IR laser power outputs compared to those with lower outputs (Fig. 6c). The onset and duration appear to change dramatically between 20.0 and 22.0 mW for the 1 s irradiation and between 10.0 and 12.0 mW for the 60 s irradiation. These data suggest that the degree of HSR induced by IR laser heating is dramatically altered between these output ranges. According to the calculations, these less invasive conditions (18.0, 20.0, and 22.0 mW for the 1 s irradiation and 8.0, 10.0, and 12.0 mW for the 60 s irradiation) allow the induction of gene expression to start at around 1 h after initial irradiation and continue for 5–7 h. These values are consistent with the profile of measured values (Fig. 4), suggesting that sigmoidal fitting of each cell is a reasonable evaluation method in this study.

Fig. 6. Evaluation of NmYG expression profiles.

a Diagram of the derivative of the sigmoidal function (blue) (p₁ = 7; p₂ = 0.015; p₃ = 5.5). The red and orange lines are the tangents at the inflection points (circles). The triangle and square are the defined onset and end time of NmYG expression, respectively, obtained by the intersection point of the tangent and the base line of the derivative. The evaluated duration is shown in gray-shaded area between the onset and end. b, c Distribution of the evaluated values of the onset (b) and duration (c) calculated based on the fitting parameters with box-whisker and violin plots. Violin plots show the density distribution of the data. In the box-whisker plots, gray horizontal lines, boxes, and whiskers indicate median, interquartile range, and data range, respectively. Different lowercase letters indicate significant differences (P < 0.05) between groups after Kruskal-Wallis test. Mean values are given below each violin plot. n values are given for each group: 1 s irradiation with 18.0 mW (n = 3), 20.0 mW (n = 4), and 22.0 mW output (n = 10) and 60-s irradiation with 8.0 mW (n = 11), 10.0 mW (n = 12), and 12.0-mW (n = 13) output (b, c).

Gene induction in gametophores

We applied the IR-LEGO system to cells in buds, phyllids, and rhizoids of gametophores to test its versatility in P. patens. We detected mCitrine fluorescence in these tissues at 10.0 mW output for 60 s (Fig. 7 and Supplementary Movies 5–7). In the buds, we observed fluorescence in multiple cells (Fig. 7a), probably resulting from cell divisions of the target cell after IR laser irradiation. A longer duration of irradiation may generate heat that might unintentionally reach neighboring cells due to the small size of bud cells. In addition, mCitrine intensity appeared weaker in bud cells than in chloronemal cells (Fig. 7a and Fig. 4e). Cell divisions may also dilute the accumulated NmYG to daughter cells; bud cells may also have a different sensitivity to a temperature increase by IR laser irradiation. In gametophore rhizoids and phyllids, we observed mCitrine fluorescence at the single cell as in protonemal tissue (Fig. 7b, c). Taken together, these results demonstrated the applicability of IR-LEGO to gametophore buds, phyllids, and rhizoids in conditions similar to those optimized for protonemal tissues.

Fig. 7. Application of IR-LEGO in gametophores.

BF images and MIP from seven z-stack fluorescence images in cells of a bud (a), rhizoid (b), and phyllid (c) in ProGmHSP17.3B:NmYG exhibiting mCitrine fluorescence after IR laser irradiation with 10.0 mW output for 60 s. mCitrine signal is shown with the fire lookup table of Fiji. The yellow dotted lines indicate the apical area of the rhizoid cell or distal area of the phyllid cell (b, c). Scale bars, 50 μm.

Discussion

The precise temporal and spatial control of gene expression offers means to dissect gene function not typically afforded by broadly expressed promoters. In this study, we demonstrated that IR-LEGO can be used to induce gene expression at arbitrary times in targeted single cells in P. patens, as has been previously shown in other plants and animals^1,6–14. Our results show that a 60 s irradiation with an IR laser is suitable for gene induction with lower invasiveness than 1-s irradiation conditions. Among the 60 s irradiation conditions, laser outputs of 10.0 and 12.0 mW yielded around 90% efficiency for the induction of gene expression without concomitant cell death (Fig. 3). In addition, an output of 10.0 mW did not cause growth arrest. These results suggest that a 60-s irradiation at 10.0- to 12.0 mW output is the optimal range for gene induction in the chloronemal cells of P. patens. Our data also indicate that IR laser power does not fully compensate for the lower incidence of gene induction following a 1 s irradiation compared to a 60 s irradiation (Fig. 2). We attribute this effect to the vacuoles present in the irradiated cells. Indeed, vacuoles occupy a large volume in protonemal cells⁴⁷. We applied an IR laser closer to the center of cells to presumably minimize heat leakage into adjacent cells. mCitrine fluorescence in neighboring cells was thus observed only under unsuitable heating conditions that also led to cell abnormality or death (Supplementary Fig. 2a and Supplementary Movie 1). However, irradiating the center of a cell may mostly heat the vacuole rather than the cytosol or the plasma membrane. Heating of the cytosol contributes to the production of unfolded or misfolded heat-labile proteins, which would lead to the release of cytosolic Class A1 HSF (HSFA1) from HSP70 and HSP90 in both plants and animals^48,49. Similarly, heating of the plasma membrane is likely to invoke calcium influx required for HSFA1 activation through cyclic nucleotide-gated channels (CNGCs) in P. patens and Arabidopsis^17,50–54. We speculate that IR laser–mediated heating of the vacuole in each cell, as in this study, would mitigate the resultant levels of heat-labile proteins in the cytosol or calcium influx via CNGCs at the plasma membrane. Compared to the 1 s irradiation, 60 s irradiation conditions can counteract the vacuole effect via propagation of heat to the cytosol and plasma membrane. This hypothesis would account for the higher probability of gene induction seen with a 60 s irradiation relative to a 1 s irradiation (Fig. 2). We obtained the variable induction levels for mCitrine fluorescence in each condition of IR laser irradiation, possibly resulting from the extent of the vacuolar effect provided by different irradiating positions (Fig. 4 and Supplementary Fig. 5). It remains to be addressed whether a control of IR laser irradiation at the subcellular scale would have a strong influence on gene induction. We recently discovered that cell size affects the possibility of heat shock–mediated CRE/loxP recombination in Arabidopsis⁹. Thus, the HSR triggered by IR laser irradiation may be influenced by thermophysical properties derived from the subcellular and cellular structure of the organism.

We quantitatively characterized the single-cell gene expression induced by the HSR by fitting mCitrine intensity to a sigmoidal function. The obtained fitting parameter p₁ revealed higher induction levels under 60 s irradiation conditions with laser output levels of 10.0 and 12.0 mW (Fig. 5). The fitting parameters p₂ and p₃, representing the gain and beginning of the mCitrine intensity elevation, respectively, showed that higher laser power output is associated with a slower and later induction of gene expression. Lower p₂ and higher p₃ values are likely involved in the lower efficiency and extended resumption, respectively, of the transcription and/or translation of NmYG. In Arabidopsis, heat shock treatment induces the formation of stress granules in which mRNAs or translation factors are sequestered^55–58. Thus, we hypothesize that the change from 10.0- to 12.0 mW output may reinforce the stress granule–mediated processes involved in the translation of NmYG. Higher laser output may also give rise to a greater abundance of misfolded translation factors^55,56,58. Together, these changes in cellular physiology may shape when gene expression starts and reaches its peak through the release of mRNAs from stress granules and disaggregated translation factors by HSP101. To the best of our knowledge, gene expression in response to a heat treatment has not been characterized at the single-cell level in multicellular organisms, perhaps because of the difficulty of individually tracking and comparing cells subjected to various patterns of heat treatment. In contrast, our experimental system with IR-LEGO applied to P. patens can easily design and confer various heat treatment to single cells even in only one sample dish, enabling quantitative and mathematical analysis of gene expression in response to heat following laser irradiation. The fitting parameters p₁, p₂, and p₃ can be new indices to analyze mutant phenotypes related to the HSR using our system.

Even with the IR laser power output set lower than 18.0 mW for 1 s or 8.0 mW for 60 s, the change in mCitrine intensity toward its peak seemed to follow a roughly sigmoidal curve, although the data from lower intensities could not be often fitted to a sigmoidal function (Supplementary Fig. 5h, i, n, o). With IR laser power output set higher than the optimal range, the pattern of mCitrine intensity showed two different trends (Supplementary Fig. 5m, s, t). One was a large increase, which can be explained by the activation of the HSR caused by higher laser powers. The other was a small increase similar to that seen with lower IR laser outputs. Both trends were subject to a delay in the mCitrine intensity reaching its peak, possibly due to mRNA storage in stress granules and HSP101-mediated disaggregation of translational factors. However, in the case of small increases of mCitrine fluorescence intensity, mRNA decay in processing bodies (P-bodies) may additionally or preferentially be engaged in the lower levels of NmYG induction, rather than mRNA storage in stress granules^59,60.

Different heat treatment protocols result in different cellular responses⁵⁷. IR laser irradiation is distinct from conventional methods relying on heat incubation of whole tissues or entire organisms. We estimated the actual time of the onset and duration of gene expression (Fig. 6b, c). Using the most suitable conditions of 10.0- to 12.0 mW output for 60 s, we established that gene expression (as estimated by detectable mCitrine fluorescence) starts 39–79 min after initial irradiation and persists for 318–410 min (median values). This calculation is informative for designing experiments with IR-LEGO and analyzing the function of a GOI.

We successfully induced gene expression with IR-LEGO in buds, rhizoids, and younger phyllids (Fig. 7). The intensity of mCitrine was relatively lower in buds and rhizoids than in chloronemal and phyllid cells. One possible explanation for this lower response is that bud and rhizoid cells may be less responsive to heat. In the case of buds, we observed weak fluorescence in multiple cells, suggesting that cell division of the irradiated cell may dilute mCitrine among daughter cells and thus decrease fluorescence intensity in single cells. Another possibility is that multiple cells were heated through the irradiated cell by the IR laser. Although optimal conditions for particular cell types will need to be determined for future work, we demonstrated the versatility of IR-LEGO in different tissues in P. patens. Gene expression under HSP promoters is transient, but expression of GOI can be maintained by repeated heat shock treatment as previously reported¹⁷. Likewise, repeated irradiation of IR laser may help to ensure and sustain the gene induction. Apart from irradiation methods, the combination with the CRE/loxP system in several organisms has broadened the applicability of IR-LEGO^6,8,9,11,13. In this strategy, DNA sequences flanked by loxP sites in the same orientation are removed through expression of a CRE recombinase gene driven by an HSP promoter in a cell subjected to IR laser irradiation. The expression of the GOI downstream from the loxP site is then induced under any purpose-designed promoters conferring constitutive or tissue-specific expression of the GOI. This technical integration may offer a means to improve the low expression levels in bud and rhizoid cells subjected to IR laser irradiation.

Irradiations of 1 s are less suitable than 60 s irradiations for stable gene induction with lower invasiveness. However, 1-s irradiation may be potentially applicable for cell ablation. IR-LEGO was employed in animals for ablation experiments to investigate neural functions and cell lineages^12,61–64. Such ablation experiments need conditioning in plants, as IR laser ablation in the animal experiments was performed under strong laser power and shorter irradiation durations (70- to 80 mW output for several msec). For either gene expression or cell ablation, measurement of the actual temperature in living cells is crucial for future work. We recently developed B-gTEMP (blue light–excitable genetically encoded temperature indicator) as a genetically encoded nanothermometer based on fluorescent proteins⁶⁵. Transgenic lines expressing B-gTEMP will uncover the thermophysical properties of living cells during IR laser irradiation, facilitating our understanding of the mechanism underlying the HSR. As aforementioned, we previously reported the effect of cell size on HSR in A. thaliana⁹. In this study, we showed the relationship between IR laser power and irradiation duration in HSR. These factors should also be closely associated in common with issues of thermal biology in animals. Besides, IR-LEGO is applicable for gene induction in both plants and animals. Optimization procedure in this study can be technically expanded to other species to perform cell labeling or linage tracing. Thus, our study would promote researches in widespread fields.

Methods

Plant materials and growth conditions

The wild-type strain of Physcomitrium patens used in this study was Cove-NIBB⁶⁶. The plants were cultured on BCDAT medium solidified with 0.8% (w/v) agar under continuous light at 25 °C⁶⁶.

Transgenic lines

Transformation was performed based on the polyethylene glycol–mediated method⁶⁶. The transgenic lines expressing NLS-mCitrine-GUS (NmYG)^26–29 under the control of the promoter region of GmHSP17.3B ( − 506 to −1 bp from the start codon of Glyma.08G069000, GlymaFiskIII.08G067000, GlymaLee.08G063600, or GmISU01.08G063100)¹⁸ (ProGmHSP17.3B:NmYG) were generated by introducing the fragments of pTPH-NmYG (accession No. OR672127) digested by KpnI and SacI (TOYOBO) into the wild-type background. Stable transformants with zeocin resistant gene derived from pTPH-NmYG were selected on BCDAT agar medium containing 50 mg/L zeocin (Invitrogen) twice. Transgene insertion was verified by genomic PCR with the primers listed in Supplementary Table 1 (Supplementary Fig. 1b).

Quantitative RT-PCR

Protonemal tissues of ProGmHSP17.3B:NmYG lines were cultured for 7 d at 25 °C under continuous white light on BCDAT agar medium overlaid with a cellophane sheet. The cellophane sheet with protonemal tissues was then transferred onto BCDAT agar medium preheated to 25 °C or 37 °C and cultured for an additional 1 h in a V11S02SUN incubator (MRT) before samples were collected and frozen in liquid nitrogen for investigation of NmYG expression in response to the temperature elevation.

Total RNA was purified from the protonemal tissues with RNeasy Plant Mini Kit (QIAGEN). First-strand cDNA was synthesized with ReverTra Ace (TOYOBO) according to the manufacturer’s instructions. qPCR was performed using a Mx3000P qPCR System (Agilent) with Thunderbird SYBR qPCR Mix (TOYOBO). The data were analyzed by the method for relative quantification of transcript levels and were normalized to those of two α-tubulin genes PpTUA1s (Pp3c16_6220 and Pp3c16_6240)³⁰. The data were collected from three independent biological replicates with technical triplicates and subjected to statical analysis with two-sided Welch’s t-test. The primers used for qPCR are listed in Supplementary Table S1.

Gene induction by heat treatment of whole tissues

A small amount of one-week-old protonemal tissues of ProGmHSP17.3B:NmYG lines #5, #7, and #11 in the PCR tube was incubated for 30 min at 25 °C or 37 °C in a MiniAmp thermal cycler (Applied Biosystems). The protonemal tissues were incubated in the PCR tube for 12 h at room temperature, and then mounted between two 0.13–0.17 mm thick cover glass glasses (Matsunami Glass). NmYG fluorescence images of 1024 × 1024 pixels were acquired under a ZEISS LSM900 confocal microscope (Carl Zeiss) with a Plan-Apochromat 10x/0.45 M2, and 488-nm (10 mW, 1.5% output) laser. The detection wavelength was 491–601 nm to detect the mCitrine signal.

Microscopy and local gene induction with the IR-LEGO system

Three- to five-day-old protonemal tissues were embedded into 500 μL of BCDAT medium containing 0.5% (w/v) glucose (BCDATG medium) solidified with 0.5% (w/v) gellan gum in four-well chamber slides (no. 5222-004; Iwaki) and covered with the addition 250 μL of medium. For the observation of gametophores, BCD medium containing 0.5% (w/v) glucose (BCDG medium) was used instead. The protonemal tissues in the dish were grown for 10 d. To suppress the overlap of protonemal branches in the field of view for microscopy, branching of protonemata was decreased via culture under red light as previously reported^24,25.

The local heat induction of NmYG was performed with an IR-LEGO 1000 system (Sigma-Koki) in ProGmHSP17.3B:NmYG line #11. The cells of protonemal or gametophore tissues were heated by irradiation with an IR laser (1480 nm) through a neutral density (ND) filter (IR-ND-5; Sigma-Koki) with different combinations of irradiation duration and power through a custom-made objective lens (UAPO 40×/0.90 NA, Olympus) on an IX-81 inverted optical microscope (Olympus). The transparency of this objective lens to IR laser is higher than commercially available lenses such as UPlanSApo, making it more useful for IR-LEGO systems. However, we found that commercial objective lenses can also be used by changing the laser oscillator power and ND filters (IR-ND-5, IR-ND-20, and IR-ND-50; Sigma-Koki) (Supplementary Fig. 7). We measured the IR laser power on the microscope stage using a thermal sensor (7Z02637; Ophir) connected to a VEGA display (7Z01560; Ophir).

The mCitrine fluorescence of NmYG was observed under an IX-81 inverted optical microscope (Olympus). To identify cells showing mCitrine fluorescence in the protonemal tissue, 1024 × 1024-pixel images were acquired with an objective lens (UPlanSApo 20×/0.75 NA), a 458- to 482 nm LED light source (SPECTRA X-6-LCR-SA, Lumencor) (approximately 10% output; 196 mW), and a scientific CMOS camera ORCA-Flash 4.0 (Hamamatsu Photonics) using the 2 × 2 binning mode in Metamorph (Molecular Devices). An emission filter set F (F01-520/35-25 520 nm, Semrock) was used to detect mCitrine signals. The exposure time was 150, 200, or 300 ms.

To evaluate the efficiency of NmYG induction and cell death, the chloronemal cells of ProGmHSP17.3B:NmYG lines were observed 12–24 h after IR laser irradiation. For assessment of cellular growth in ProGmHSP17.3B:NmYG lines, time-lapse imaging of the tip cells was performed at 20 min intervals for 24 h, starting within 20 min after IR laser irradiation. Time-lapse imaging to measure mCitrine intensity was carried out within 20 min after IR laser irradiation under an IX-81 inverted optical microscope (Olympus). Seven z-stack images were acquired with 6 μm steps. The exposure time was 150 ms. In the above time-lapse imaging, samples of protonemal tissue placed on the microscope were exposed to white LED light only during the intervals in between imaging to allow cells to grow.

The mean fluorescence intensity of mCitrine from NmYG was measured manually in the whole regions of each IR laser–irradiated cell, to which the mean fluorescence intensity of the surrounding medium region was subtracted at each time point with Fiji software (ImageJ; imagej.net/Fiji)⁶⁷. The images for measurements and analyses of mCitrine intensity were acquired with a 150-ms time exposure. The kinetics of mCitrine intensity reaching a maximum were fitted by a sigmoidal function as

$$I\left(t\right)=\frac{p_1}{1+{\exp }^{-p_{2}t+p_3}}+I_0$$ 1

where I(t) and I₀ are the fluorescence intensity of mCitrine at time t and onset of time-lapse imaging, respectively, normalized to the mean fluorescence intensity per pixel at onset of time-lapse imaging (I₀ = 1) indicating chloroplast autofluorescence. p₁ is the peak of elevated mCitrine intensity after IR laser irradiation, p₂ is a velocity of the increase in mCitrine intensity, and p₃ is the start of the increase in mCitrine intensity; p₁, p₂, and p₃ are free parameters determined by minimizing the residual sum of squares between the measured data and the fitting function (1) using custom MATLAB scripts⁶⁸.

mCitrine intensity increased after IR laser irradiation in different manners with different irradiation conditions. To evaluate the onset and duration of mCitrine increase in response to heat by IR laser irradiation, we calculated the time when the velocity of mCitrine intensity increase is above zero. The velocity is given by the first derivative of the fitting function (1) as

$$I^{\left(1\right)}\left(t\right)=\frac{p_1{p}_2{\exp }^{-p_{2}t+p_3}}{{\left(1+{\exp }^{-p_{2}t+p_3}\right)}^2}+I_0$$ 2

We assumed the range of the velocity with the tangent method. The onset and end time of the velocity are the intersections of the tangent at the inflection points and the x-axis (I⁽¹⁾(±∞) = 0). The x-coordinates of the left and right inflection points (a₁, b₁) and (a₂, b₂) on the first derivative (2) are described as follows:

$$I^{\left(3\right)}\left(a_{1,2}\right)=0$$ 3

The x-coordinates were obtained by solving for a_1,2:

$$a_1=\frac{p_3+\log \left(2-3^{\frac{1}{2}}\right)}{p_2}$$ 4

$$a_2=\frac{p_3+\log \left(2+3^{\frac{1}{2}}\right)}{p_2}$$ 5

When assigning a₁ and a₂ to the function (2), the y-coordinates of b₁ and b₂ are obtained as:

$$b_{1,2}=\frac{p_1{p}_2}{6}$$ 6

The function f(t) of the tangents on the inflection points is described as:

$$f\left(t\right)-b_{1,2}=I^{\left(2\right)}\left(a_{1,2}\right)\left(t-a_{1,2}\right)$$ 7

Thus, by solving the following equation, the onset (t₁) and end time (t₂) of the velocity are calculated as:

$$t_1=\frac{p_3+\log \left(2-3^{\frac{1}{2}}\right)-3^{\frac{1}{2}}}{p_{2}t}$$ 8

$$t_2=\frac{p_3+\log \left(2+3^{\frac{1}{2}}\right)+3^{\frac{1}{2}}}{p_{2}t}$$ 9

The duration is finally calculated as:

$$t_2-t_1=2\log \left(2+3^{\frac{1}{2}}\right)+\frac{2\times 3^{\frac{1}{2}}}{p_2}$$ 10

Using the above Eqs. (8) and (10), the onset and duration of NmYG expression were defined from the profile change in mCitrine intensity of chloronemal cells of ProGmHSP17.3B:NmYG.

Statistics and reproducibility

We performed statistical analyses using MATLAB (Mathworks). The statistical analysis and data sampling for gene induction by heat treatment of whole tissues are referred to in the above “Quantitative RT-PCR”. Frequencies of chloronemal cells showing mCitrine fluorescence (Fig. 2e, f and Supplementary Fig. 3a, b) and growth arrest (Fig. 3d) were analyzed by Bonferroni test. The fitting parameters of induction profile (Fig. 5a–c) and the derivative indexes (Fig. 6b, c) in different IR laser irradiation conditions were analyzed by Kruskal-Wallis test. In the analysis of frequencies of the first, third and fifth cells showing mCitrine fluorescence (Supplementary Fig. 3a, b), we repeated the experiment of IR laser irradiation to the cells more than twice except for the condition of 1 s with 12.0 mW output. In the analysis of growth arrest (Fig. 3d), we repeated the experiment of IR laser irradiation to the cells more than three times except, whereas the intact condition was tested once. The data of normalized mCitrine intensity (Fig. 4g–i and Supplementary Fig. 5h–t) was obtained through more than independent twice experiment of IR irradiation except for the condition of 1 s with 14.0- and 16.0 mW and 60 s with 16.0 mW (Supplementary Data 1).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

T.T., Y.T., and Y.Kamei designed the research; T.T., Y.Y., S.O., Y.Kabeya, M.H., T.M., T.K., J.S., Y.T. and Y.Kamei performed the research; T.T. analyzed the data; T.T., J.S., Y.T. and Y.Kamei wrote the paper.

Peer review

Peer review information

Communications Biology thanks the anonymous reviewers for their contribution to the peer review of this work. Primary Handling Editor: David Favero. A peer review file is available.

Data availability

All source data generated and/or analyzed during this study are included in this article, Supplementary Data 1, or deposited in Zenodo (see Code availability statement). The data are also available from the corresponding authors on reasonable request. Plasmids have been deposited as pTPH-NmYG (accession No. OR672127).

Code availability

All scripts of MATLAB with the source data have been deposited and publicly available in Zenodo⁶⁸.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s42003-024-07141-1.