protoplasts started leaking certain solute molecules into the
medium altering the osmotic gradient across the plasma membrane.
To take into account the potentially increasing inaccuracy of
data collected from all protoplast including stretched protoplasts,
the contribution of consecutive measurements to the overall model
fit was increasingly down-weighted. In this way most emphasis
was given to the initial readings that are expected to be the most
accurate ones. The down-weighting was hence applied to all the
protoplasts analyzed on the DMF platform. In Fig. 3(b), the two
experimental curves (dashed lines) reflect the typical changes over
time in the major and minor axis occurring in one single isolated
protoplast treated at an osmotic pressure change of −0.99 MPa.
The superimposition of the experimental data (dashed line) and
the modeled theoretical data (solid line) showed good fitting, suggesting
that leaking due to plasma membrane stretching did not
occur in this protoplast. In Fig. 3(c) instead, a proper fit between
the experimental data (dashed line) and the modeled theoretical
data (solid line) could not be obtained, suggesting over stretching
of the plasma membrane of this single isolated protoplast. Thanks
to the down-weighting factor, the model could still be successfully
applied to the overstretched protoplast and a proper value for the
permeability coefficient (P) could be estimated as the information
for estimating P is mainly contained in the first dynamic part of
the curve where the model fits the cell size data well. These results
confirm the validity of the described model and demonstrate the
suitability of the DMF chip for performing this type of analysis.
By means of DMF droplet manipulations, 50 single A. thaliana
protoplasts were successfully analyzed and individual P values
were obtained ranging from 1 m s−1 to 75 m s−1 (Fig. 3(d)).
To compare the distribution of the 50 P values obtained for the
protoplasts treated with two hypo-osmotic conditions, we conducted
Mann–Whitney U test. According to this nonparametric test
(p-value = 0.2548 was obtained), it could be concluded that the Permeability
coefficient values were not statistically differentin the two
hypo-osmotic treatments and could be pooled. Large variations in
P values of populations of cells are commonly observed in these
types of experiments (1.25–540 m s−1) [37], which mostly reflect
the existing variations in aquaporin expression or activity within
plasma membranes of different cells [38]. The P values obtained
are thus in line with the values presented in the literature [37] and
perfectly illustrate the heterogeneity present within a population
of cells when exposed to the same stimuli.
This study confirms that DMF chip is an efficient platform for
generating the steep gradient steps necessary for the implementation
of water potential studies at a single cell level and it further
validates the usefulness of the illustrated immobilization strategy,
which successfully retains small non-adhering cells during
the rapid external solution exchange. Furthermore, this approach
allows a substantial increase in the throughput of the analysis as
an average of seven cells can be monitored and followed simultaneously
within a single field of view.