Impedance flow cytometry for rapid quality assessment of protoplast cultures

Abstract

Background

Protoplasts, which are plant cells devoid of cell walls, are valuable tools in plant biotechnology. However, they are highly sensitive to mechanical and osmotic stress during isolation and early culture, often leading to significant loss of viability. Reliable and efficient methods for monitoring protoplast quality are essential for downstream applications.

Results

We applied impedance flow cytometry to assess the viability, cell size, and early division of freshly isolated protoplasts from Arabidopsis thaliana, Brassica napus, and Beta vulgaris. This label-free technique enables fast, objective, and high-throughput assessment of individual protoplasts, allowing reliable monitoring of viability and early division in large populations. Importantly, IFC-derived viability metrics strongly correlated with microcallus formation, demonstrating their predictive value for culture competence.

Conclusions

Impedance flow cytometry provides a robust, efficient and reproducible method for characterizing protoplast cultures. It enables rapid assessment of viability and growth potential, supporting quality control and optimization in plant cell culture workflows.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13007-025-01474-6.

Background

Protoplasts, or plant cells devoid of a cell wall, are versatile platforms in plant biotechnology that are widely used for gene and protein expression analyses, plant cell regeneration studies, somatic hybridization, genetic modification, and the study of cellular processes [1–3]. The regeneration of protoplasts into whole plants has become especially attractive in recent years for gene editing applications, where the regeneration of edited protoplasts into fertile, transgene-free, nonchimeric plants has been demonstrated in major crop species such as rapeseed (Brassica napus) [4], tomato (Solanum lycopersicum) [5], and potato (Solanum tuberosum) [6]. However, the regeneration of protoplasts is highly variable and strongly affected by the genotype, the selected donor plant tissue, and the culture conditions [7, 8].

Protoplast preparation is a multistep procedure that requires empirical optimization [1, 2, 9]. The isolation and culture of protoplasts require careful handling, and multiple rounds are needed to optimize incubation conditions and establish a regeneration protocol. This is a time-consuming and labour-intensive process crucial for developing a robust and effective protocol that often represents the first significant hurdle in protoplast research. A key factor that determines the success of regeneration and other protoplast applications is cell viability [1, 2, 9]. It is therefore critical to obtain protoplast suspensions that are highly viable and have minimal contamination of cell debris and dead cells.

Enzymatic removal of the cell wall imposes mechanical stress on the cells, requiring adjustments to the osmolarity of the incubation solution to compensate for the elimination of turgor pressure. In addition, cell wall digestion separates the cell from its neighbours, disrupting spatial cues that are critical for maintaining cell identity [10, 11]. This cell isolation initiates the dedifferentiation step that precedes regeneration. While in theory, each protoplast is totipotent, in practice, only a fraction of the cells displays the capacity to divide and regenerate. The lack of uniform regeneration competence is attributed to differences in the cell developmental state, stress sensitivity, and metabolic condition of the cell [12]. In general, the regeneration capacity of protoplasts is poor, exacerbating the low efficiency of methods for genetic engineering and underscoring the need to improve cell viability during extraction and culture. Typically, the loss of viable cells occurs primarily during the first days of culture [13]. Additional cell loss occurs when media or incubation conditions are changed, e.g., when polyethylene glycol (PEG) is added for the induction of transfection [14].

Accurate monitoring of protoplast viability during the early stages of culture is therefore critical when optimizing regeneration protocols and improving the success of downstream applications. Usually, cell viability is assessed by microscopic analysis via stains or molecular markers. A commonly used viability stain is fluorescein diacetate (FDA), which is cleaved by nonspecific cytoplasmic esterase, releasing the highly fluorescent compound fluorescein. Green fluorescence accumulates only in live cells with an intact plasma membrane [15]. A major limitation is that many plant cells display autofluorescence in the same spectral range as fluorescein does, interfering with accurate cell viability quantification [16]. In addition, FDA fluorescence levels are dynamic, as the cell continuously synthesizes and degrades fluorescein, leading to a progressive loss of signal from the fluorophore. Furthermore, repeated excitation‒emission cycles increase the risk of photobleaching of fluorescein. All of these imply that a protocol with strict timings and microscopy settings is needed to obtain reproducible results. This, in combination with the requirement of analysing a sufficiently large number of cells per sample, makes this approach very time-consuming. Moreover, the often direct visual interpretation of cell morphology and signal intensity leads to a subjective categorization of cells as alive or dead [17].

Conventional flow cytometric analysis of protoplasts represents a valuable alternative to microscopy-based assays, greatly increasing the speed and number of cells analysed and improving the objectivity of the classification, although at the cost of increased instrument complexity [18, 19]. Like microscopic methods, this technique also depends on robust viability markers such as FDA, which, again, depends on strict time management to obtain quality data.

Here, we explored the use of impedance flow cytometry for rapid and robust analysis of protoplast viability. Electrical impedance refers to the opposition of an electrical circuit to an alternating electric current (AC). Two descriptors of impedance are amplitude—the peak response of the circuit—and phase angle—the temporal offset between the voltage and current waveforms. Interestingly, cellular components, such as the membrane and cytoplasm, respond to applied ACs in a frequency-dependent manner [20]. At low frequencies (< 1 MHz), the electric current flows around the cell, owing to the insulating properties of the intact cell membrane. This explains how the impedance measured at low frequencies primarily reflects cell size [21]. At intermediate frequencies (approximately 500 kHz to 10 MHz), the membrane capacitance can be measured, providing information about membrane integrity and electric charge. At higher frequencies (>6 MHz), the current can penetrate the membrane, and impedance reflects changes in cytoplasmic conductivity and permittivity, which depend on the ionic composition and physiological state of the cell [21–23]. Jointly, these features allow for a label-free characterization of cells simply by measuring electrical impedance at specific AC frequencies.

The advent of microfluidics chips has revolutionized the field of flow cytometry, increasing the resolution of systems while simultaneously decreasing their size and cost [22]. These technologies have been readily adapted to impedance flow cytometry (IFC) since the early 2000s [21, 24], and IFC has since emerged as a robust, high-throughput, and user-friendly system for analysing a wide range of cellular properties. For example, IFC was used to distinguish between different subpopulations of leukocytes [25] and to estimate viability and size of yeast cultures and insect cells [26].

In plant research, IFC has been effectively used to analyse pollen and microspore development. For pollen, it facilitates the rapid assessment of viability and germination capacity in a variety of species, such as hazelnut [27], tobacco, cucumber, tomato [28], potato, pepper [29], wheat [30] and hemp [31]. In microspore cultures, IFC enables early prediction of embryogenic potential [28, 29, 32]. These examples illustrate the broad applicability of IFC.

The application of IFC methods to characterize plant protoplasts has focused on plant cell wall regeneration. For example, it was used in a microfluidic system capable of distinguishing between protoplasts from different genotypes of Arabidopsis thaliana, as well as between these and protoplasts from Populus trichocarpa [18]. These studies revealed dynamic changes in the electrical properties of protoplasts during cell wall regeneration. Similarly, a custom-built impedance device was able to monitor cell wall formation and differentiate between wild-type A. thaliana and mutants with modified cell wall compositions [33].

In this study, we demonstrate that impedance flow cytometry provides reliable measurements of protoplast viability in A. thaliana, B. napus, and Beta vulgaris. In B. napus, we additionally assess cell size, estimate division rates, and predict microcalli development. Finally, we used IFC to track viability trends during the early stages of protoplast cultivation in both B. napus and B. vulgaris. Our work highlights IFC as a novel methodology for fast and robust quality assessment of early protoplast cultures.

Methods

Plant material

Brassica napus DH12075 protoplasts were isolated from the leaves and cotyledons of seedlings grown in round glass jars (Ø8.5 cm × 8 cm) on ½ strength Murashige and Skoog medium (Duchefa Biochemie, #M0222) with 0.77% agar and 1.4% sucrose at 25 °C under 80 µmol/m²/s light intensity and a 16 h photoperiod. The material was collected seven days after sowing in the case of cotyledons or after 4 weeks for the first and second true leaves.

Two Beta vulgaris genotypes (GENOT1 and GENOT2, Florimond Desprez Belgium) were cultivated in vitro for 3 weeks on Murashige and Skoog media under a 16-hour light/8-hour dark cycle at 22 °C with a 60 µmol/m²/s light intensity.

Seeds of A. thaliana ecotype Col-0 were sown on a square Petri dish (120 × 15 mm) containing the same medium described for B. napus and stratified in the dark at 4 °C for one week before being placed under 160 µmol/m²/s light intensity and a 16 h photoperiod at 23 °C for another 7 days.

Protoplast isolation

B. napus protoplasts were isolated from tissues cut with a scalpel into 0.5 cm strips using 1% Cellulase-R10 (Duchefa Biochemie, #C8001) and 0.2% Macerozyme R-10 (Duchefa Biochemie, #M8002) in 0.47 M mannitol, 10 mM MES and 10 mM CaCl₂ (MMC) buffer [34]. The material was incubated in the dark for 15 to 16 h in an orbital shaking incubator (ES-20, Biosan) set at 25 °C (± 0.5 °C) and 50 rpm. The protoplast suspension was then sieved through a 150 μm nylon cloth to remove larger debris, followed by a 70 μm cell strainer (VWR). The filtered protoplasts were subsequently centrifuged at 120 × g for 5 minutes, after which they were recovered from the pellet and resuspended in 4 mL of MMC. Two 15 mL Falcon tubes with 6 mL of 0.6 M sucrose for leaves or 0.3 M sucrose for cotyledons [34], were overlaid with 2 mL of the protoplast suspension and centrifuged at 80 × g for 10 min (minimum brake strength). Protoplasts were then recovered from the interphase band, resuspended in 50 mL of MMC and centrifuged at 120 × g for 5 min. Afterwards, the pellet was again resuspended in 50 mL of MMC, and the protoplast concentration was calculated via a hemocytometer (Fast-Read 102, Kova). Next, the protoplasts were centrifuged one final time at 120 × g for 5 min and resuspended in M1 culture medium (Kao and Michayluk basal salts; Duchefa Biochemie, #K0214) supplemented with 1 mg/L nicotinic acid, 1 mg/L thiamine HCL, 1 mg/L pyridoxine HCL, 100 mg/L inositol, 8.1% glucose, 1 mg/L 2,4-D, 0.2 mg/L NAA, and 0.5 mg/L BA) at a final density of 10⁵ protoplasts/mL. These were divided into two 24-well plates (VWR) with 600 µl in each well. One plate was used for the time course analysis, while the other was kept for imaging at the 2-week mark. The cultures were kept in the dark at 25 °C (± 0.5 °C) for the entire duration of the experiments.

B. vulgaris protoplasts from the leaves of in vitro-grown plants were isolated as described earlier by Hall et al. [35, 36]. In brief, a protoplast population of mainly stomatal guard cells is obtained by introducing a grinding step to enrich the epidermal cell fraction prior to overnight enzymatic digestion with cellulase RS and macerozyme R10. Next, intact stomatal guard cells are isolated via centrifugation using a sugar gradient.

A. thaliana protoplasts were isolated from entire seedlings according to the protocol described by Jeong et al. [34], with modifications to the enzyme mixture (1% Cellulase-R10, 0.4% Macerozyme R-10 and 0.1% Pectolyase Y-23; Duchefa Biochemie #P8004).

Microscopy

For microscopic viability assessment, aliquots were taken and stained with fluorescein diacetate (FDA; 2 mg/mL) at a ratio of 1:100 (v/v) FDA solution to protoplast suspension for 1 min. The stained protoplasts were counted using an Olympus IX81 inverted epifluorescence microscope equipped with a 10x/0.4 NA objective, a U-MWIBA3 filter cube (Olympus, excitation wavelength of 460–495 nm, emission wavelength of 510–550 nm) and an Olympus XM10 camera. Imaging was performed via Olympus cellSens Dimension software (v1.18), and the cells were counted with a hemocytometer. Protoplasts exhibiting bright green cytoplasmic fluorescence were considered viable.

At 14 days post-isolation, protoplast cultures in 24-well plates were imaged using the multiple image alignment (MIA) function of cellSens to automatically acquire and stitch images. Partial well views were reconstructed from a 3 × 4 grid of images (9.13 × 5.81 mm per image, 53.02 mm² total area) centred on the middle of the well and using a 4x/0.16 NA objective. The rectangular region of interest, covering 27.7% of the well surface area, was consistently selected across all the wells to exclude artefacts caused by their curved edges.

Impedance flow cytometry

The impedance measurements were performed using an Ampha Z40 impedance flow cytometer (Amphasys AG) equipped with a chip featuring a 120 μm × 120 μm sensing channel. Measurement frequencies were set to 0.49 MHz and either 8 MHz for protoplasts from B. napus cotyledons and A. thaliana seedlings, 10 MHz for B. napus leaves or 11 MHz for B. vulgaris leaves. These account for differences in membrane properties and cytoplasmic conductivity across different tissues. We used AmphaSoft Pro (v2024.08.09) for data acquisition, visualization, and analysis.

Specialized templates were developed to optimize the analysis of protoplasts from the different samples used. Among other features (Amplification 5, Modulation 4, Demodulation 1), the template’s trigger level setting (0.1) was aimed at excluding protoplast debris from the analysis. The trigger level was set based on the optimized mix ratio of measurement buffer and protoplast medium which influences the buffer conductivity. So the data recording includes all cell signals and excludes smaller cell debris that cannot be clearly separated from the electronic noise.

All the samples were prepared by mixing 2 mL of AmphaFluid 6 buffer (Amphasys AG) with 0.5 mL of protoplast suspension in M1 culture medium and filtering through either a 40 or a 100 μm nylon sieve (VWR) just before measurement.

A vertical viability gate was applied across all measurements at 300° phase angle. An additional gating strategy for cell division estimation is illustrated in Fig. 3.

Fig. 3.

Assessment of dividing protoplasts using microscopy and IFC. A Scatter plots of the impedance amplitude vs. phase angle of B. napus cotyledon protoplasts at 4 d.a.i. Samples were filtered through a 40 μm strainer (purple) or a 100 μm strainer (red). The sectors “Live-Dividing (LD)” on the right and “Dead-Divided (DD)” on the left are highlighted in red. Histograms for the phase angle (top) and amplitude (right) are shown on the sides of the scatter plot. B Microscopy image of FDA-negative protoplasts showing cytokinesis at 2 d.a.i. (arrow). The scale bar is 40 μm. C Time course line plot (days: D4, D5 and D6) of the ratio of FDA-stained and non-FDA-stained dividing protoplasts (green) compared with the ratio of LD to DD gates (blue)

Coulter counter

Coulter analysis of protoplast cultures was performed via a Multisizer 3 (Beckman Coulter) with a 100 μm aperture. Protoplast cultures were measured daily from the day of isolation to 3 days after isolation. For each measurement, 1 mL of protoplast suspension was diluted in 20 mL of Isoton II (Beckman Coulter). The device used a current of 800 µA and a 60-second counting time, with coincidence correction enabled. The particle diameter was analysed in the 7–100 μm range, with a total of 5 × 10⁴ particles counted per run. Two technical replicates were included for each time point. Data were acquired using the Multisizer 3 software (v3.51; Beckman Coulter).

Heat inactivation and time course experiments

Heat inactivation was performed on fresh protoplasts of B. napus and A. thaliana via a heating block (Labnet AccuBlock Mini) to warm 1 mL aliquots of protoplast suspension. Specifically, suspensions were measured after being exposed for 5 min to 50 °C, 60 °C, 70 °C and 80 °C, in the case of A. thaliana, and additionally 90 °C for B. napus. Each aliquot was exposed to a specific temperature, and measurements were performed immediately after the treatment.

A time course analysis of B. napus DH12075 protoplast viability was conducted on a set of 7 experiments, consisting of 3 independent isolations of leaf protoplasts and 4 independent isolations of cotyledon protoplasts. The same experiment was also performed on 2 different genotypes of B. vulgaris, both of which were isolated on the same day. These protoplast suspensions were monitored at three different time points: day 0 (before plating), day 1 and day 2 of culture.

For both the time course and heat inactivation assays, each sample was measured in two technical replicates.

Evaluation of cell division

For the cell division time course, the cells were stained with FDA and placed in a hemocytometer for viability assessment and counting. This was done at days 4, 5 and 6 after culture initiation. Only cells showing clear signs of cell division (e.g., oblong shape, peripheral organelle arrangement, central nuclear localization, membrane constriction along the cell midline, presence of a division plate) were considered. The selected cells were divided according to their viability and measured with the measuring tool in CellSens. A ratio of stained over non-stained cells showing signs of cytokinisis was established for all days to serve as a reference.

For the IFC analysis of dividing protoplasts, aliquots of the same samples used for FDA staining were filtered through either a 40 μm strainer (to exclude dividing cells and generate negative controls) or a 100 μm strainer (to retain dividing cells). Impedance data were acquired at 0.49 MHz, and scatter plots of phase angle versus amplitude were used to define gates representing dead-divided (DD) and live-dividing (LD) cells.

Gates were initially defined at 4 d.a.i. by comparing the impedance profiles of the 100 μm-filtered samples against the negative controls while attempting to approximate the FDA-stained microscopy reference. The DD gate was placed around the population with low amplitudes and left-shifted phase angles, while the LD gate was placed around the population with high amplitudes and right-shifted phase angles. The gating boundaries were set manually based on the data distribution at 4 d.a.i. and then applied unchanged to the datasets from 5 to 6 d.a.i. The ratio of LD: DD events was used as a proxy for the viability of dividing cells, and this was compared to same-day FDA staining results.

Image analysis of microcalli mass

The day 14 images were analysed via ImageJ (v1.54p) within Fiji [37]. Batch processing was done using a custom macro to import images using the Bio-Formats Importer, apply automatic thresholding using the ‘Default’ method for each image/well, convert the image to binary, generate a selection, and measure the selected area. The percentage of this area in relation to the total image size was calculated and saved in a results table for further analysis.

Data analysis and visualization

Data analysis and plotting were performed via R (v4.3.3) in RStudio (v2024.12.1.563; Posit team [38]).

For viability assays, counts from IFC gating were summarized as proportions of total events. Comparisons between IFC- and FDA-derived viability were assessed with boxplots.

For size analysis, protoplast diameters from IFC and Coulter counter measurements were normalized relative to the day 0 median for each method. Relative changes were calculated as , where is the p-th percentile at day d such that corresponds to the day 0 median. Boxplots were formatted with whiskers spanning the 10th–90th percentiles, boxes spanning the 25th–75th percentiles, and a central line marking the median.

For predictive modelling of microcallus mass, the dependent variable was the proportion of the image area covered by protoplast-derived structures at 14 d.a.i. Candidate predictors were IFC-derived features measured on isolation day and 2 d.a.i. Forward stepwise regression was performed with the corrected Akaike information criterion (AICc) for model selection. Model assumptions were checked by plotting residuals. Performance was summarized with adjusted R² and AICc. Independent isolations excluded from training were used for model validation.

Results

IFC for estimating protoplast viability

To evaluate the suitability of IFC for assessing protoplast cultures, we tested whether the impedance phase angle can be used to estimate protoplast viability. Protoplasts were isolated from different species and tissues, including protoplasts from A. thaliana seedlings, B. napus leaves and cotyledons, and B. vulgaris guard cells.

As previously described for pollen and microspores [27–29, 32], protoplasts were initially analysed via IFC following heat inactivation to characterize healthy versus dying populations.

The control IFC analysis revealed a bimodal distribution of cells along the phase angle axis, with a first peak below 250° phase angle, spanning a relatively small amplitude range, and a second peak at phase values above 300°, with a higher amplitude maximum (Fig. 1). With stepwise increases of the heat shock temperature, the percentage of high phase datapoints gradually decreased (Fig. S1). Heat shocks at 90 °C for B. napus and 80 °C for A. thaliana resulted in the complete disappearance of the higher phase population. All detected events had a phase angle lower than 300° (Fig. 1A, B).

Fig. 1.

Impedance phase angle discriminates between live and dead protoplasts. A Scatter plots of the phase angle vs. amplitude for fresh and heat-inactivated B. napus (DH12075) cotyledon protoplasts. B Scatter plots of phase angle vs. amplitude for fresh and heat-inactivated A. thaliana seedling protoplasts. The vertical lines represent viability gates, separating viable (right side) from nonviable protoplasts (left side). Mean percentage values are given for the populations classified as viable. Two technical replicates were run per sample, and replicates per treatment are overlaid in the same color. Phase angle frequency distributions are plotted above each plot

Next, the performance of IFC was compared with that of traditional viability estimation methods based on fluorescein diacetate (FDA) staining (Fig. 2A–B). Viability was assessed via both approaches for B. vulgaris guard cell protoplasts and for B. napus protoplasts isolated from leaves and cotyledons, immediately after isolation and again after one and two days of culture. Both approaches detected a significant reduction in cell viability within the first two days of culture [2] (Fig. S2). The estimated percentages of viable cells were highly correlated between both techniques (Fig. S3; R² = 0.608; r = 0.78). The standard deviation between technical replicates for IFC data (± 2.12%; n = 48) was almost half that of the FDA measurements (± 4.10%; n = 80).

Fig. 2.

Comparison of IFC and FDA staining viability assays and coulter analysis for protoplast size estimation. A Microscopy images of FDA-stained B. napus cotyledon protoplasts at different time points during culture (days D0, D1 and D2). The scale bar is 100 μm. B Boxplot of percent viable protoplasts over culture time estimated by counting FDA-positive vs. total cells (green) and IFC analysis (blue). Individual samples were analysed in parallel with both methods (n = 4 biological replicates (1–4) × 2 technical replicates). C Boxplot of protoplast size over culture time, as measured by IFC (blue) and coulter counter analysis (red). The y-axis represents the relative change in size compared with the 50th percentile (median) on day 0, calculated by averaging two technical replicates. The percentiles are normalized as , where is the day 0 median specific to each method. Boxplot whiskers represent the 10th and 90th percentiles, the box spans the 25th to 75th percentiles, and the line indicates the median

IFC for estimating protoplast size

We compared impedance amplitudes at 0.49 MHz for B. napus cotyledon protoplasts at zero to three days after isolation to size distributions obtained using a Beckman Coulter Multisizer 3. The Coulter principle generates discrete diameter measurements (µm), whereas the impedance amplitude is an indirect volumetric measure [25]. To account for this difference in units, we compared percentile values in the Multisizer histograms with those from the IFC amplitude distributions (Fig. 2C). Direct comparison between the two measurement methods was made possible by centring data around the respective population medians on the day of isolation so that all size differences are expressed as relative changes to day 0’s 50th percentile.

Both methods resulted in similar trends with increasing protoplast sizes during the first 48 h of culture (Fig. 2C). This increase in impedance amplitude was also observed in our viability time course analysis (Fig S2), where the average amplitude measured at 0.49 MHz for B. napus cotyledon protoplasts ranged from 0.76 ± 0.06 on the day of isolation to 0.84 ± 0.10 at 2 d.a.i. For the leaf samples, these values changed from 0.39 ± 0.08 to 0.53 ± 0.08. This initial increase was followed by a slight decrease in size at 3 d.a.i. (Fig. 2C). A strong positive linear correlation is observed between the IFC and CC medians (r = 0.77).

To further link impedance amplitude and protoplast size, we filtered two aliquots of B. napus cotyledon protoplasts at 4 d.a.i. through either a 40 μm or a 100 μm strainer. The removal of cells larger than the filter cut-off resulted in a downwards shift in the maximal amplitude. Protoplasts strained through a 100 μm filter had amplitudes of up to 5, whereas those strained through a 40 μm filter had an amplitude maximum of 3 (Fig. 3A). This is also visible in the amplitude histogram, on the right-hand side of the scatterplot in Fig. 3A.

IFC for estimating protoplast cell division

To assess whether IFC can be used to quantify viable dividing protoplasts, we monitored a culture of B. napus cotyledon protoplasts from 4 to 6 d.a.i. via microscopy and simultaneously determined impedance characteristics using IFC (Fig. 3). The smallest cell that showed signs of cell division under the microscope had a length of 44 μm (out of 274 cells with signs of cytokinesis) We also identified 2 different types of structures in the observed samples, which were comprised of both FDA-stained and non-FDA-stained cells (Fig. 3B). Based on these observations, cells larger than 40 μm were considered putatively dividing, independently of their viability, and 40 μm straining was used to generate negative controls, while 100 μm straining retained dividing structures for IFC analysis.

Compared with higher frequency measurements (Fig. 1), the nonlinear volumetric amplitude response and the influence of cell size on the phase angle at 0.49 MHz led to a reduced spread of the data cloud along individual axes and produced a more diagonal trend in the impedance scatter plot (Fig. 3A). Nonetheless, basing ourselves on the previous results pertaining to viability, higher phase angle populations were still considered as viable.

A population with higher amplitudes and right-shifted phase angles was detected in the 100 μm strained sample but not in the 40 μm controls. This transformed the unimodal left-skewed purple phase histogram into the bimodal red distribution shown in Fig. 3A. In the negative controls, the impedance amplitude values were mostly less than 2 (Fig. 3A), although some data points above 2 were still present.

Using predefined impedance gates corresponding to live- and dead-dividing cells, IFC-based LD: DD ratios aligned closely with viability estimates from FDA-stained microscopy. This agreement was evident already at 4 d.a.i. and became even stronger at 5 and 6 d.a.i. (Fig. 3C).

IFC metrics as early predictors of microcalli formation

The cell cluster area after 14 days of culture was selected as a measure of regenerative potential (Fig. 4A, left). This time point marks the end of the incubation of the culture in the dark. The cluster mass was quantified by imaging a central rectangular region in each well of the protoplast culture plate. The cell clusters were then segmented (Fig. 4A, right), and the total cluster area was expressed as a percentage of the imaged region to estimate the biomass.

Fig. 4.

Exploring early IFC metrics to predict microcalli mass. A Brightfield image of protoplast cultures in a 24-well plate at 14 d.a.i. (4× magnification; left) and segmentation of the protoplast-derived structure area (white; right). The scale bar is 1 mm. B Regression analysis of cell mass measured at 14 d.a.i. (expressed as a proportion of the image area covered by protoplast-derived structures) versus viability measured with IFC on the isolation day. Individual points represent six different protoplast isolations from B. napus cotyledons. Blue round points correspond to training points, from which the fitted line and 95% confidence interval are drawn. Red triangles indicate the two validation isolations with the red bars corresponding to their 95% prediction interval

We evaluated five IFC-derived metrics as possible early predictors of competence, including viability percentages and viable protoplast concentrations on the day of isolation and at 2 d.a.i., as well as the ratio of live-dividing to dead-divided (LD: DD) structures at 2 d.a.i. These were collected from four different isolations of B. napus cotyledon protoplasts.

To maintain model simplicity, linear regression models limited to two predictors were fitted using forward stepwise selection with AICc. This criterion balances model fit with complexity, applying stronger penalties for smaller sample sizes to prevent overfitting [39].

The best-fitting model included day 0 viability and the LD: DD ratio on day 2, with an R² of 0.986 (Fig. 4B). However, this model was not statistically significant (p = 0.12, α = 0.05). In addition, the two-factor model offered only a marginal improvement over the model based only on day 0 viability (Fig. 4B), which had an R² of 0.956 and was statistically significant (p = 0.02, α = 0.05). The final model can be expressed as:

To further assess the robustness of this relation, we tested the final model against two additional independent cultures where only day 0 viability and day 14 cell mass were recorded. The two validation points fell within the 95% prediction interval and resulted in a mean absolute error of 8.24%. When adding the validation dataset to the training data, the R² changes to 0.901 (p = 0.00; Fig. S5).

The LD: DD ratio on day 2 also showed predictive capacity as a single regressor (R² = 0.726, p = 0.15), although it was not statistically significant.

To compare the predictive capacity of IFC with FDA viability staining, we replaced the IFC predictor in the model with FDA viability measured on the day of isolation. The FDA-based model performed worse at predicting culture competence, yielding a lower R² of 0.716 and a non-significant p-value of 0.15 (α = 0.05).

Discussion

Implications of IFC-based viability assessment

The first days of protoplast culture are characterized by a sharp loss in viability, which has been linked primarily to various stresses experienced during isolation [2]. Identifying reliable, early markers of cell viability is therefore crucial for evaluating the quality of isolated protoplasts and for optimizing culture conditions.

In this study, A. thaliana was selected because it is a well-studied model for fundamental research, whereas B. napus, an allotetraploid oil crop, is characterized by substantial size variation (10–40 μm) (Fig. S4B). Guard cell protoplasts from B. vulgaris are typically very sturdy and small (10–15 μm), and they are exploited in a well-established transformation and regeneration protocol [35, 36].

Owing to the capacitive properties of the intact membrane, live cells have greater reactance, which leads to more important time shifts between the signal of the applied voltage and the measured current. At appropriate frequencies, this results in a clear phase angle separation between live and dead cells, with the latter exhibiting lower phase angles [26]. Our results show that this is also the case for protoplasts, as had otherwise been demonstrated for pollen [27–31], microspores [28, 29, 32], and animal and yeast cells [26].

At high frequencies, the separation between live and dead peaks was sufficiently distinct (Fig. 1) to allow some flexibility in defining the viability cut-off. We opted for a conservative gate at 300° of phase, with our results suggesting that signals below the 300° mark correspond to dead cells, whereas those above 300° represent living cells, providing a fast (< 2 min/sample) estimate of protoplast viability in a culture.

Overall, IFC is a robust and label-free method for assessing protoplast viability, with lower technical variation and greater speed than traditional labour-intensive microscopy-based assays. This finding suggests that IFC is a promising tool for high-throughput screening and routine monitoring of protoplast cultures.

IFC for monitoring protoplast physiology

Protoplasts display dielectric properties, modulating an electric field in a frequency-dependent manner [40]. At high frequencies, impedance shifts inform on the cytoplasmic conductivity and membrane capacitance of the protoplast, whereas at low frequencies, they depend primarily on cell size [19, 22, 23]. This relationship between low frequencies and cell size has also been reported for pollen and microspores [27–29] and is consistent with our findings.

In B. napus cotyledons, which yield protoplasts with a broad size distribution, impedance-based size estimates strongly agreed with Coulter counter (CC) measurements, a well-established method for protoplast sizing [41].

The similar trends of increasing size observed in both methods during the first 48 h of culture likely reflect metabolic reactivation and initiation of cell division [11]. The slight decrease in size at 3 d.a.i. (Fig. 2C) may be due to the loss of oversized, damaged or otherwise compromised cells [42, 43]. The strong positive linear correlation observed between the IFC and CC medians supports the use of impedance amplitude measurements for assessing cell size as an alternative to Coulter analysis.

In addition to cell size, IFC also provided insights into protoplast division. By combining size-based filtering with impedance gating, we were able to distinguish what we named live-dividing (LD) from dead-dividing (DD) structures and quantify their relative abundance in the form of a ratio (LD: DD). The progressive improvement in the LD: DD ratio between 4 and 6 d.a.i. mirrored the observed FDA-based staining patterns, demonstrating that impedance features can be used to approximate cell division in regenerating protoplast cultures. The development of this ratio was particularly important, as challenges encountered during the analysis prevented more direct assessments of cell division.

Soon (2–4 h) after the removal of the digesting enzymes, protoplasts initiate the regeneration of new cell walls [44], a prerequisite for subsequent cell division and the formation of microcalli [42]. Upon cell wall formation, protoplasts lose their typical spherical form and adopt an elongated or oblong shape, initiating cytokinesis, which further leads to changes in volume and cytoplasm composition (Fig. 3B). Hence, we postulated that these structural changes influence the dielectric properties of protoplasts in a manner detectable via IFC. However, detecting division with impedance is complicated by the fact that cell division occurs in walled protoplasts of various sizes, meaning that size or amplitude alone cannot be used directly as proxies for division.

Furthermore, the development of the gating strategy was challenged by a small proportion of negative control events (strained through a 40 μm sieve) with amplitude values above 2 (Fig. 3A). These likely arose from incomplete size exclusion due to deformation, orientation effects, or mesh flexibility or from smaller cells possessing intracellular features that increase the impedance amplitude.

Microscopy revealed both FDA-stained and nonstained dividing cells (i.e., live, and dead structures), indicating that cell death can still occur after division has begun (Fig. 3B). This is likely because, in protoplasts, successful cytokinesis requires a specific cellular physiology, which involves dispersion of chromatin in the nucleus, reorganization of the microtubule network, and cell wall regeneration. Failure to meet these requirements can disrupt the spatial cues necessary for preprophase band development, ultimately resulting in incomplete or aberrant mitosis [42, 43]. These mitotic defects are frequently associated with aneuploidy in protoplasts due to chromosome loss [45] and may trigger programmed cell death. This may explain the persistence of DD events in our impedance-based profiles, which, in turn, offered us an opportunity to qualify cytokinesis in the culture through a gating strategy that did not depend solely on amplitude or size.

Overall, our results support the utility of IFC for quickly and accurately determining the proportion of viable dividing protoplasts, a key parameter underlying culture quality and development that is difficult to assess via microscopy [11, 46].

IFC metrics as a predictor of regeneration potential

The variability in the regenerative capacity of protoplast cultures means that early predictors of the ability to produce microcalli are highly valuable for screening and protocol optimization. Such metrics enable timely evaluation of successful incubation conditions and media compositions for regeneration and help to prioritize cultures showing great potential or discard poorly responsive ones. To aid this decision-making process, we sought to identify early IFC-based metrics that correlate with later growth and development of the culture.

Our results demonstrate that early IFC metrics, particularly day 0 viability, can be exploited to predict the potential of a protoplast culture to produce microcalli. These findings are in line with a previously reported linear model for embryogenesis in Triticum aestivum microspore cultures [32], which also highlights the predictive power of early viability metrics obtained with IFC. Specifically, they described a strong predictive relationship (R² = 0.95, p = 0.02) between the viability of T. aestivum microspore cultures at 7 d.a.i. and the number of microspore-derived embryos counted at 30 d.a.i.

To further assess the robustness of our model, we tested it on a small set of independent cultures. The model successfully captured the trends in these additional samples, supporting the predictive value of day 0 viability. Nonetheless, the limited size of both the training and validation datasets, combined with the inherent variability of protoplast cultures, means that these results should be interpreted carefully. Expanding the dataset in future studies will be important to confirm and strengthen the observed predictive relation.

In our study, FDA viability measurements obtained at the time of isolation were less predictive than the contemporaneous IFC metrics, further underscoring the added value of IFC in predicting culture outcomes.

Although the LD: DD ratio at 2 d.a.i. did not reach statistical significance in our dataset, its predictive signal supports its relevance as a descriptor of early culture development and indicates some possible co-dependency or interchangeability between day 0 viability and early cell division dynamics.

Limitations of IFC assessing protoplast culture quality

Microscopic observations of protoplasts reveal substantial amounts of cell debris—presumably, organelles from lysed cells. In microscopy-based viability assays, however, such cellular fragments and organelles are readily disregarded, as they are recognized from the images, regardless of their size. Although debris particles can also be distinguished from live protoplasts by IFC, large numbers of non-protoplast events can confound population statistics and skew viability estimates. To exclude these signals from small and irregular particles, IFC trigger thresholds should be adjusted per sample type. Nonetheless, the inclusion of larger cell debris in IFC analyses may still result in lower viability scores than does FDA staining, a trend corroborated by a comparison of the boxplots in Fig. 2B and regression analysis (Fig. S3), where the slope of IFC versus FDA measurements is approximately 0.64.

A stringent trigger threshold can also artificially inflate viability estimates, as smaller dead cells and ruptured protoplasts can fall below this threshold, consequently being removed from the analysis and changing the measured culture density and proportional viability. As a remedy for this, it is important to consider both viability and the concentration of live events when comparing different treatments or time points.

Due to the complex nature of impedance, cells of the same size do not consistently have the same amplitude values, resulting in a nonlinear relationship between the cell diameter observed by microscopy and the IFC amplitude, which is largely influenced by volume. For this reason, coulter counters present a better choice if the objective is simply to separate populations based on size.

Conclusion

In this study, we used impedance flow cytometry (IFC) to characterize A. thaliana, B. napus and B. vulgaris protoplasts isolated from leaves, cotyledons, or entire seedlings during the early culture stages, from isolation to initial cell division, and to predict microcallus formation competence. Despite challenges related to the inability to visually validate individual data points—a common challenge in non-imaging cytometry—our findings establish IFC as a versatile and objective tool for evaluating protoplast cultures. We show that this method enables rapid, robust, label-free estimation of viability, provides a proxy for monitoring size-related physiological changes, and allows estimation of cell division rates. Moreover, early IFC-based viability measurements were found to be strong predictors of microcallus formation, highlighting its value for the early selection of competent cultures.

Our work positions IFC as a valuable addition to the methodological toolkit for protoplast research, with potential applications in high-throughput screening, early evaluation of viability and division, and optimization of media and culture conditions for regeneration. The application of this technology revealed species- and tissue-specific differences in the heat sensitivity of fresh protoplasts, significant declines in protoplast viability immediately after isolation, and intricate growth and division patterns over the first days of culture. Future work could further explore such responses and extend the IFC approach to protoplasts from other species and tissue types.

Abbreviations

AC

Alternate Current

AICc

Corrected Akaike information criterion

CC

Coulter counter

D.a.i.

Days after isolation

DD

Dead-divided

FDA

Fluorescein diacetate

IFC

Impedance flow cytometry

LD

Live-dividing

MMC

Buffer solution containing Mannitol, MES, and Calcium

Author contributions

T.R., R.L. and D.G. conceived the study. T.R. and R.L. designed the experiments. T.R. performed the experiments. T.R., R.L. and G.S. acquired the data. T.R. analysed the data. T.R., R.L. and D.G. interpreted the results. T.R. drafted the manuscript. All authors reviewed the manuscript.

Funding

This work was funded by a grant from the Research Foundation Flanders (FWO-SBO grant PASCell; S001423N). Amphasys AG supported this work by providing the impedance flow cytometer at a reduced cost as part of their partnership in the project.

Data availability

The datasets generated and/or analysed during the current study are available upon request.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare the following competing interest: Grit Schade, an employee of Amphasys AG. Amphasys AG provided the impedance flow cytometer at a reduced cost and acted as a project partner.