Fluorescent dyes were conjugated to antibodies for use as microinjection markers during
various experiments. Conjugation reactions were performed as follows: lyophilised, reagent-
grade sheep IgG (Sigma) was reconstituted to a concentration of 2 mg/ml in 100 mM
Na2C0 3, pH 9.3. 1 ml o f IgG solution was then applied to a single pre-packed vial o f either
Cy3.5 or Cy5 fluorescent dye (Amersham Biosciences, cat: PA23500 & PA25000). The
reaction was then incubated at room temperature for 30 min with one brief additional mixing
step half way through. Conjugation reactions were always protected from light. Dye-
conjugated antibodies were then separated from free dye using the QuickSep dialysis system
(Tripple Red). Briefly, the 1 ml reaction volume was placed into a dialysis cassette
containing a 16,000 MW cut-off membrane and subsequently immersed in 400 ml of
phosphate buffered saline, pH 7.5 (PBS). Dialysis was preformed over 48 hours at 4 ®C with
periodic buffer changes. A magnetic stirrer ensured thorough mixing throughout. Dialysed
Cy3.5-IgG and CyS-IgG stocks were stored at -20 °C until required.
2,2,6 In vitro dye transfer assay fo r the study o f intercellular communication
Preparation o f experimental cultures: Freshly trypsinised NEBl kératinocytes and NIH 3T3 fibroblasts were seeded in 35 mm diameter glass-bottomed micro well dishes (MatTek)
containing 3 ml o f the appropriate tissue culture medium for the cell type in question. Cells
were allowed three days to grow to confluency before experimentation.
For all microinjection and imaging cells were transferred to fresh phenol red-free tissue
culture medium supplemented with 20 mM HEPES, pH 7.5 (experimental culture medium).
The presence o f phenol red in culture medium was not desirable during fluorescence imaging
due to the autofluorescent properties of this compound. HEPES was added to culture medium
to compensate for the low ambient levels o f CO2 during both microinjection and imaging,
therefore removing the need for more complicated CO2 delivery systems.
Nuclear microinjection: A confluent region o f cells within the central microwell o f the culture dish was selected and groups o f adjacent cells microinjected in the nucleus with
plasmid DNA encoding one o f the EGFP-Cx31 variants. Expression constructs were all
microinjected at a concentration o f 0.05 pg/pl. Following microinjection cells were
transferred to fresh tissue culture medium and returned to the incubator. Cells were allowed
4-5 hours to express the microinjected constructs.
Cytoplasmic microinjection and time-lapse microscopy: The cytoplasmic microinjection of fluorescent dye into cells was conducted on the same Zeiss Axiovert 135 TV inverted
microscope (Carl Zeiss) that was subsequently used for the imaging o f dye-transfer through
cells. For a detailed description o f the microscope specifications see section 2.2.3, page 19.
The culture dish was positioned on the stage of the microscope in a glass-topped, steel
humidity chamber (constructed in-house) that was designed to prevent evaporation during the
course o f time-lapse experiments. The lid of the chamber was removed during
microinjection. For all dye transfer experiments patches o f fluorescent cells were identified
under the microscope and single cells close to the patch perimeter were microinjected in the
cytoplasm with 2 mM Alexa 568 dye (Molecular Probes). Upon the successful cytoplasmic
microinjection of a single fluorescent cell the microinjection needle was quickly removed
from the microscope and the lid o f the steel humidity chamber replaced, completely
enclosing the culture dish. The occurrence of dye transfer between adjacent cells was then
determined using multi-channel digital time-lapse microscopy. The acquisition software was
used to acquire images every 20 seconds in three separate channels (phase contrast, EGFP
fluorescence, and Alexa 568 fluorescence) using the 20 x NA 0.5 objective, the CCD camera
and the multiple dichromatic mirror in conjunction with motorised excitation and emission
rounds). Although the exposure times used for imaging EGFP fluorescence tended to vary
between dye transfer experiments, the exposure time used for imaging Alexa 568
fluorescence was kept constant as fluorescence intensity o f the dye did not vary considerably.
The time elapsed from the injection o f dye to commencement o f image acquisition was
approximately a minute. The procedure for studying intercellular dye transfer is summarised
in Figure 3, page 19.
2,2,7 Measurement o f dye transfer rate
The time-lapse acquisition of Alexa 568 fluorescence images over the course o f dye transfer
experiments provided accurate information concerning the temporal dynamics of dye spread
through cell populations expressing different Cx31 variants. Image processing software was
specially developed in order to numerically quantify this process, enabling the statistical
comparison o f the effects of different Cx31 mutations on the rate o f dye transfer. By
comparing the various rates at which dye transferred through cell populations expressing the
wild type Cx31 protein with those found for cell populations expressing particular Cx31
variants, significances could be derived that reflected the relative degree to which different
Cx31 mutations impeded or enhanced intercellular communication.
Preparation o f film sequences fo r image analysis: Prior to image analysis film sequences were cropped (IMD analysis. Kinetic Imaging) to incorporate a region that contained the
EGFP expressing cells and dispersing dye. Interactive tracking (Motion Analysis, Kinetic
Imaging) was then used to mark, using the mouse pointer, the position o f the cell originally
Figure 3. Summary o f the in vitro dye transfer assay for the assessment connexin function
(A) Groups o f adjacent cells within a monolayer culture were microinjected in the nucleus with 0.05 jug/pi plasmid DNA encoding one o f the CxSl-EGFP fusion constructs under investigation. (B) Cultures were left fo r 4 — 5 hours to allow the expression o f microinjected constructs. This resulted in the formation o f tight patches o f fluorescent cells. (C) A single expressing cell close to the patch perimeter was then selected and microinjected in the cytoplasm with 2 mM Alexa 568 dye. Multi-channel time-lapse microscopy was then used to monitor the spread o f Alexa dye through the surrounding cell population: Images were acquired at 20 s time intervals in three separate channels (phase contrast, EGFP, and Cy3.5 fluorescence). Each fd m sequences consisted o f 100 acquisition rounds (33 min duration).
Figure 4. Summary o f the in vitro assay fo r the assessment o f NIH 3T3 cell death
(A) Groups o f adjacent cells within a monolayer culture were microinjected in the nucleus with a mixture o f 1 pg/pi Cy3.5-IgG and 0.05 pg/pl plasmid DNA encoding one o f the EGFP Cx31 fusion constructs under investigation. Cultures were allowed 20 min to recover from microinjection. (B) Propidium iodide was added to the culture medium to a final concentration o f 0.5 pg/ml and the microinjected cells then relocated under the microscope via detection o f their Cy3.5-IgG labelled nuclei. (C) Multi-channel time-lapse recording was then used to monitor the onset o f protein expression and any subsequent change in cell behaviour: Sequential phase contrast, EGFP, Cy3.5, and propidium iodide fluorescence images were acquired every 60 s in order to monitor the effects o f protein expression on cell viability. The concentration o f propidium iodide within the medium was sufficiently low that it could only be detected upon accumulation within the nuclei o f dead cells. Cells were imaged fo r a 100 min duration.
and was defined as the centroid of the microinjected cell within the first frame o f the film
sequence. Cell movement within the monolayer was minimal over the course o f time-lapse
experiments and so it was considered unnecessary to score the cell position for each fi'ame
within the film sequence.
Image analysis: A single numerical value representing the normalised mean distance o f dye transfer {NMD) was generated for each Alexa 568 fluorescence image within a film sequence and collectively these values then used to calculate the mean rate o f dye transfer for a
particular experiment. Alexa 568 fluorescence images were extracted from film sequences,
binned to a pixel size of 2.6 x 2.6 pm ( 4 x 4 pixel binning) in order to remove high fi*equency
noise, and thresholded in order to remove background. Single NMD values were then calculated for each image as: