IMAGE ANALYSIS COMBINED FLUORESCENCE
MICROSCOPY
Examples of ImageJ Software Application in Yeast Studies
Evgeny O. Puchkov
Skryabin Institute of Biochemistry and Physiology of Microorganisms, Institutskaya Str.,5, Pushchino, Russia
Keywords: Computer image analysis, Fluorescence microscopy, Yeast culture viability, Brownian motion, Intracellular
viscosity, Nucleic acid intercalator, Anticancer drug.
Abstract: For Saccharomyces cerevisiae yeast studies, three approaches have been developed. They are based on the
image analysis (ImageJ software) application for the fluorescence microscopy data treatment. The first is a
computer-aided fluorescence microscopy procedure for quantifying of the damaged cells in the ethanol-
producing yeast culture. It was shown to be applicable for the assessment of the culture viability. The
second is a means of characterizing Brownian motion of the insoluble polyphosphate complexes in the
vacuoles. Using this approach, the apparent viscosity in the vacuoles was measured. The third is a method
for locating intracellular sites/targets of the nucleic acid intercalators. This method may be of help in
designing of new DNA-targeted drugs and in preliminary studies of their interaction with eukaryotic cells.
1 INTRODUCTION
The yeast Saccharomyces cerevisiae is an object of a
large research interest for at least two reasons. First,
it is widely used in food industry, including baking
and production of alcoholic beverages, ethanol and
food additives. Second, it serves as a model system
for studies of eukaryotic cells since many of the
basic cellular properties between yeast and humans
are highly conserved. This is also due to the
availability of the DNA sequence of the complete
genome and biochemical data, convenience for
molecular manipulations and the ease of handling.
The yeast cells are shown to be a good model in
many areas of cancer research and for new drug
discovery. Hence, the development of new
techniques and approaches for the studies of the
yeast S. cerevisiae is an important task.
Fluorescence microscopy is a very informative
method for single cell studies. Potentialities of the
method can be significantly improved by combining
it with the digital photography and computer image
analysis.
In this paper, three approaches based on the
ImageJ software (National Institute of Health, USA,
http://rsb.info.nih.gov/ij) application for the
fluorescence microscopy data treatment are
described. The first approach was developed for the
assessment of the yeast culture viability; the second
for to evaluate the apparent viscosity in the vacuoles
of the individual yeast cells; and the third for
locating intracellular sites/targets of the nucleic acid
intercalators.
2 ASSESSMENT OF THE YEAST
CULTURE VIABILITY
This study was undertaken to develop rapid
computer-aided fluorescence microscopy procedure
for the assessment of the ethanol-producing yeast
culture viability.
The fluorescence microscopy of rehydrated S.
cerevisiae cells from a dry commercial Fermiol
preparation (DSM Food Specialties Beverage
Ingredients, The Netherlands) stained by a
combination of ethidium bromide (E) and 4,6-
diamidino-2-phenylindole dilactate (DAPI) revealed
two types of cells. The cells of the first type showed
blue-green fluorescence (DAPI), whereas the cells of
the second type showed bright orange-red
fluorescence (E). In special experiments, it was
shown that the first type of fluorescence appearance
was a characteristic of the intact cells and the second
type was a characteristic of the damaged cells.
258
O. Puchkov E..
IMAGE ANALYSIS COMBINED FLUORESCENCE MICROSCOPY - Examples of ImageJ Software Application in Yeast Studies.
DOI: 10.5220/0003120902580261
In Proceedings of the International Conference on Bioinformatics Models, Methods and Algorithms (BIOINFORMATICS-2011), pages 258-261
ISBN: 978-989-8425-36-2
Copyright
c
2011 SCITEPRESS (Science and Technology Publications, Lda.)
Using ImageJ “RGB Split” option, the initial
fluorescent images of the stained cells were
separated into three images containing red, green
and blue components of the fluorescence. This
procedure enabled to establish that, in the ethidium
stainable cells, along with the red component, there
was well expressed green component of the DAPI
fluorescence, too. This was an indication of the fact
that DAPI stained all the cells. So, RGB split of the
fluorescent image provided a means for assessing a
fraction of the E stainable/damaged or E
unstainable/undamaged cells. To this end, after an
appropriate threshold adjustment, “Analyse
Particles” option could be used for automated
counting of the fluorescing cells (“particles”) in the
“red” and in the “green” parts of the RGB-splitted
image.
A comparison of the viability rates of cultures,
determined by the plate count method, and the
relative numbers of intact cells, determined in the
same cultures by the developed procedure, showed a
good correlation between these parameters, thereby
indicating a possibility of using this procedure for
the assessment of the viability of rehydrated Fermiol
preparations. It was noted that fluorescence
microscopy underestimates the viability of yeast
populations by 10–15%, compared to cultural
methods. This disagreement can probably be
explained by the fact that some damaged cells may
recover during the subsequent cultivation of
rehydrated cells.
The main advantage of the proposed approach is
that it allows more rapid assessment of cell viability
as compared to not only the cultural methods, but to
the “manual” microscopy as well. Also, this
approach opens a way to the automation of the
analysis. For more details see (Puchkov, 2006).
3 MEASUREMENT
OF VISCOSITY IN VACUOLES
In the S. cerevisiae cells, at some cultivation
conditions, appear vividly moving particles, <1 µm
in size, known as ‘dancing bodies’. They were
shown to be insoluble polyphosphate complexes
(IPCs) localized in the vacuoles. Upon staining of
the cells by DAPI, IPCs acquire a bright yellow
fluorescent colour, while the nuclei and
mitochondria fluoresce blue.
The aim of this study was to quantitatively
characterize, by fluorescence microscopy combined
with computer image analysis, the movement of
IPCs in the vacuoles of S. cerevisiae VKM Y-2549
cells and to evaluate the apparent viscosity in the
vacuoles, using the obtained data.
The immobilized cells were photographed in a
Speed Burst regime of the Sony DSC-V3 digital
camera. It gave a series of eight frames at intervals
of 0.43 s and an interval between series of 2–3 s.
Using ImageJ, in a frame of a series, a fluorescing
particle was selected as a region of interest (ROI) by
“Oval Selection” option. “ROI Manager” option was
switched on to get the same ROI area in other
frames, although position of the ROI was changed
according to the position of the particle in a new
frame. The locations of the IPCs in the two
dimensional space (X and Y locations) were
evaluated using “Center of Mass” option.
The results of this analysis indicated that IPC
movements were chaotic or, as it is often referred to,
were random walks, or Brownian motion.
The Brownian motion of particles obeys the
Einstein–Smoluchowski equation, which for two-
dimensional (2D) movement is as follows:
<s
2
> = 4κTt/3πηD, (1)
where <s
2
> is the average of the square of
displacement, κ is the Boltzmann constant, T is the
thermodynamic temperature, t is the elapsed time, η
is the viscosity and D is the diameter of the particle.
To evaluate the apparent viscosity in the yeast
vacuoles via Brownian motion of the IPCs, the
average displacement in 2D space and the diameters
of the moving particles need to be estimated
[equation (1)]. As IPCs move in three dimensions, a
criterion for selecting the 2D displacements in the
photorecords, which are in the focusing plane or at
least close to it, must be found. To learn how this
could be done, experiments were performed on
suspensions of fluoresceinisothyocyanate-labeled
latex microspheres of 2.1 μm and 3.1μm diameter in
water. In this model, two parameters were known,
the diameters of the microspheres and the viscosity
of distilled water in which they were moving.
Using photorecords similar to those of IPC, two
parameters of Brownian motion of fluorescing
microspheres were estimated for each series of the
eight speed regime shots – the consecutive two-
dimensional locations and the mean fluorescence
intensities (“Mean Gray Value” option) at these
locations. The normalized fluorescence intensity
served as a quantitative measure of the microsphere
shift from the two-dimensional motion in each eight
frames series of the speed regime shots. It was tested
whether the fluorescence decrease of no more than
15% may be used as a criterion for taking
IMAGE ANALYSIS COMBINED FLUORESCENCE MICROSCOPY - Examples of ImageJ Software Application in
Yeast Studies
259
displacements of the microspheres close enough to
the two-dimensional space for viscosity assessment
by formula (1). Upon selecting “appropriate”
displacements by this criterion from photorecords of
the series of speed regime shots, computation of the
apparent water viscosity using formula (1) has been
made. The viscosity values of water estimated by
these measurements agreed with 30% accuracy with
the value (0,89 сР at 25 °C, the temperature of
measurements) obtained by other methods. This
result indicated that the developed approach may be
used for viscosity assessment.
It was also found by analysis of the fluorescence
intensity profile (“Line selection” across “Center of
Mass” and “Plot Profile” options) of the
microspheres that the level of the 85% fluorescence
intensity corresponds to their outer borders. That
gave the means of estimating the size of the
fluorescing IPC particles by their fluorescence
intensity profiles.
Using methodology developed on the latex
microspheres, Brownian motions of the IPCs in four
cells were analyzed. Displacements for each IPC in
the two-dimensional space in the series of the eight
speed regime shots were estimated. Outer borders of
IPC were determined by estimating 85% level of
fluorescence intensity in the IPC fluorescence
profile. That gave the dimensions of IPC. Since the
shape of IPC is not known, this parameter was
assumed to be corresponding to the diameter of the
microspheres.
In four yeast cells, the 2D displacements and
sizes of the IPCs were evaluated. Apparent viscosity
values in the vacuoles of the cells computed by the
Einstein-Smoluchowski equation using the obtained
data, were found to be of 2.16 ± 0.60; 2.52 ± 0.63;
3.32 ± 0.9; 11.3 ± 1.7 cP. The first three viscosity
values correspond to 30 – 40% glycerol solutions.
The viscosity value of 11.3 ± 1.7 сР was supposed to
be an overestimation caused by the peculiarities of
the vacuole structure and/or volume in this particular
cell. This conclusion was supported by the particular
quality of the Brownian motion trajectories set in
this cell as compared to the other three cells. For
more details see (Puchkov, 2010).
4 INTRACELLULAR LOCATION
OF INTERCALATORS
The aim of this study was to test if intact (not fixed)
yeast cells of S. cerevisiae can be used as a model
for locating intracellular sites/targets of the nucleic
acid intercalators (NAI). To this end, intracellular
distributions of three fluorescing NAI – the
anthracycline anticancer drug doxorubicin (DR)
(trade name adriamycin; also known as
hydroxydaunorubicin) along with nucleic acid dyes
ethidium (E) and 4',6-diamidino-2-phenylindole
(DAPI), were investigated using fluorescence
microscopy combined with computer image
analysis.
Upon incubating for at least 20 h, all the cells of
S. cerevisiae VKM Y-2549 appeared to be stainable
by the NAI. In the cells stained by DAPI, there were
clearly visible fluorescing blue spots of the nuclei
and small dots of the mitochondria. As compared to
DAPI, intracellular fluorescence distribution of DR
and E did not have the clear “spot and dot”
appearance. Although nuclear region could be
distinguished, even distribution of the DR and E
fluorescence was observed in the regions where
mitochondria should be located. There was
heterogeneity in the expression of staining by DR
and E. The observed difference in this feature of the
individual cells may be a reflection of a specific
combination of the cytoplasmic membrane diffusion
barrier properties and of the presence/activity of
drug export permeases.
Combined application of DR or E with DAPI
visualized the location of the nuclear and of the
mitochondrial DNA. However, visual analysis could
not answer the question whether DR and E were
located in the sites of the DAPI-stainable DNA.
To investigate the potential location of DR and E
in the nuclei and in the mitochondria marked by
DAPI, distribution of the red, green and blue
components of the fluorescence (“pseudospectral
analysis”) in these regions was assessed
quantitatively using ImageJ “Measure RGB“ option
of the “Analyze“ plugin after appropriate ROI
selection.
The obtained data indicated that, as it was
expected, Red/Green ratio of DR and E were
significantly higher than that of DAPI. In contrast to
E, DR fluorescence had appreciably higher the
Red/Green ratio in the nuclei than in the
mitochondrial region. These data indicated that at
least a fraction of DR molecules bond in the nuclei
had more expressed red component of their
fluorescence spectra than those in the mitochondrial
region. Interaction of DR with nucleic acids is
known to increase the “red” maximum in its
fluorescence spectrum. So, the less expression of the
red fluorescence in the mitochondria may be
explained assuming that, along with binding to
DNA, some molecules of DR are bound to the
BIOINFORMATICS 2011 - International Conference on Bioinformatics Models, Methods and Algorithms
260
membranes. Binding to the membranes may also be
a plausible reason for visible even fluorescence
distribution in the mitochondrial region in the cells
stained by DR as compared to “dot”-like staining of
the mitochondrial DNA by DAPI. There may be at
least two binding sites for DR and E in the
mitochondria: DNA and the membranes. Ability of
E to interact with the mitochondrial membranes was
demonstrated by others. Potential ability of DR to
interact with biomembranes was shown on artificial
membranes.
Upon combined addition of DR+DAPI
simultaneously, in both the nuclei and in the
mitochondria, the Red/Green fluorescence ratio was
higher in comparison with the application of DAPI
alone, but it was lower as compared to exposing the
cells to DR only. If the cells were first incubated
with DR, and DAPI was added later, the Red/Green
ratio in the nuclei was higher than in case of
simultaneous addition of the NAI, but still lower
than after the addition of DR alone. Similar results
were obtained for the pair E+DAPI with the
difference that the order of the NAI addition did not
significantly influence the final result. At the same
experimental conditions, there were no appreciable
changes in the Blue/Green ratio. So, quantitative
image analysis revealed appearance of the red
fluorescence component in the regions of DAPI
stained DNA upon combined application of DAPI
with DR and E. This is an indication of a
colocalization of DAPI with DR and E in the nuclei
and in the mitochondria.
In the sites of the colocalization of DR and E
with DAPI, the red component of their fluorescence
was less as compared to the application of DR and E
alone. These data can be interpreted as a competition
of DAPI with DR and E for the same DNA binding
sites. It should be mentioned here that, although NAI
can intercalate between base pairs, they may also be
bound to some other parts of DNA, e.g. in the minor
groove. The detailed mechanism of the competition
of DAPI, DR and E for binding sites in DNA is not
known as of yet. For more details see (Puchkov,
2011, in the press).
5 CONCLUSIONS
1. Three computer image analysis algorithms of
the ImageJ software for the fluorescence microscopy
data treatment have been developed.
2. The first, based on the use of “RGB-split” and
“Analyze Particles” options, was shown to be
successfully applied in a new method for rapid yeast
viability assessment. This method could be of
potential use for rapid viability evaluation of other
microbial cultures, too.
3. The second enables the determination of the size
and of the center of mass locations of the fluorescing
particles, <1 µm in size, by their space fluorescence
intensity distribution assessment. Using this
algorithm, the apparent viscosity in the vacuoles of
the individual yeast cells was evaluated by Brownian
motion measurements. The methodology developed
in this work may be of use for some other studies of
moving microscopic fluorescing particles.
4. The third algorithm, the “pseudospectral
analysis”, was designed for locating intracellular
sites/targets of fluorochromes in the yeast cells. The
intact (not fixed) yeast cells of S. cerevisiae can be
used as a model for intracellular locating of the
nucleic acid intercalators by fluorescence
microscopy combined with this computer image
analysis algorithm. The model and the approach
presented herein may be of help in the new DNA-
targeted drug discovery and in preliminary studies of
their interaction with eukaryotic cells.
ACKNOWLEDGEMENTS
Participation of Miss Millicent McCarren in the
experiments and discussions of the studies described
in the 4
th
Section of the paper is highly appreciated.
REFERENCES
Puchkov, E. O., 2006. ‘Viability assessment of
ethanol-producing yeast by computer-aided
fluorescence microscopy’, Microbiology, vol. 75, no.
2, pp. 154-160.
Puchkov, E. O., 2010. ‘Brownian motion of polyphosphate
complexes in yeast vacuoles: characterization by
fluorescence microscopy with image analysis’, Yeast,
vol. 27, no. 6, pp. 309-315.
IMAGE ANALYSIS COMBINED FLUORESCENCE MICROSCOPY - Examples of ImageJ Software Application in
Yeast Studies
261