Our major goals are to study de-novo formation, gating and regulation of gap junction (GJ) channels and unapposed/undocked hemichannels, which mediate direct cell-cell communication and paracrine signaling, under normal and pathological conditions.
- Studies of de-novo formation of gap junction channels: Bukauskas & Weingart, 1993 & 1994; Weingart & Bukauskas 1993; Bukauskas et al., 1995; Valiunas et al., 1997.
- Voltage and chemical gating of gap junction channels in cells expressing different Cx isoforms; fast and slow/loop gating mechanisms: Bukauskas & Weingart, 1994; Bukauskas & Peracchia 1997; Bukauskas et al., 1995, 2000, 2001, 2002; Teubner et al., 2000; Kreuzberg et al., 2005; Abrams et al., 2006; Palacios-Prado, 2009, 2010; reviewed in: Bukauskas & Peracchia, 2000, Verselis & Bukauskas, 2002, Bukauskas & Verselis, 2004 and Kreuzberg et al., 2006.
- Development of stochastic and Markovian multi-state models describing voltage-gating of GJ channels and hemichannels combined with methods of global optimization for automated define gating parameters from experimental measurements of voltage gating: Paulauskas et al., 2009; Sakalauskaite et al., 2011; Paulauskas et al., 2012.
- Permeability of gap junction channels and its regulation: Bukauskas et al., 2002; Rackauskas et al, 2007; Palacios-Prado et al., 2009; Palacios-Prado & Bukauskas, 2012; Bukauskas, 2012.
- Studies of unapposed/nonjunctional hemichannels: Trexler et al., 1999; Kronengold et al., 2003; Contreras et al., 2002 & 2003; Srinivas et al., 2005; Bukauskas et al., 2006; reviewed in: Bennett et al., 2003; Contreras et al., 2004; Saez et al., 2005.
- Biophysical studies of connexin mutants related to hereditary diseases: Abrams el al., 2003; Beltramello et al., 2005; Bicego et al., 2006; Dobrowolski et al., 2008.
MethodsA dual whole-cell voltage clamp to study electrical cell-cell coupling, voltage gating and function of unapposed hemichannels.
Fluorescence imaging for:
- Intracellular pH and Ca2+ measurements using ratiometric pH-, Ca2+ and Mg2+ - sensitive probes as well as derivatives of green fluorescence protein sensitive to pH.
- The single GJ channel permeability studies to fluorescent dyes depending on their molecular masses and the net charges.
- Time-lapse imaging to study de-novo formation and turnover of gap junction channels as well as function of unapposed hemichannels.
Movies shown below illustrate our latest research directions as well as some unexpected behavior of GJ channels that still do not have explanation, while their more close examination may lead to better understanding of their the function.
Stochastic multistate modeling of voltage-gating of GJ channels and hemichannels and global optimization of their gating parameters.
Two movies demonstrate automated fitting of experimental steady-state junctional conductance (gj) dependence on transjunctional voltage (Vj) (gj,ss-Vj plot in black) to a stochastic 16-state model (S16SM) describing Vj-gating of Cx45/Cx45 homotypic (Movie 1) and Cx43/Cx45 heterotypic (Movie-2) GJ channels containing both fast and slow gates per hemichannel (details in Paulauskas et al., 2012; see also a S16SM model in action on the "Model" page of this website). The table on the right/top show changes of parameters, ascribed to fast and slow gates of the ‘left’ and ‘right’ hemichannels, during global optimization. The plot on the bottom/right shows the dynamics of modulus difference between simulated (red) and experimental (black) gj,ss-Vj plots during global optimization. A blue gj,ss-Vj plot shows intermediate best fits that gradually approach the experimental gj,ss-Vj plot. The fitting process was automatically controlled using Exkor Global Optimization algorithm and ended by estimating gating parameters (top/right table) that characterize properties for each of four gates: A – maximal derivative, dgj,ss/dVj; V0 - voltage at which probabilities for the gate to be in the open and closed states are equal, Go – unitary conductance of the open state, Gc - unitary conductance of the residual state, Ro and Rc - rectification coefficients of Go and Gc, respectively, depending on voltage across the gate. Red gj,ss-Vj plots were simulated assuming that the junction contains 1000 channels.
Movie 1. Illustration of coarse global optimization of gating parameters of Cx45 homotypic gap junction channels from experimental gj,ss-Vj dependence (black) assuming that parameters of fast and slow gates in both hemichannels are identical. Simulated gj,ss-Vj plots (red) are comprised of 630 iterations.
Movie 2. Illustration of global optimization of gating parameters of Cx43/Cx45 heterotypic gap junction channels from experimental gj,ss-Vj dependence (black) assuming that parameters of fast and slow gates in both hemichannels differ. Simulated gj,ss-Vj plots (red) are comprised of 1680 iterations.
Reverberations of excitation in a 15x15 cluster caused by a single stimulus.
Two movies demonstrate reverberation of excitation in a cluster of 225 neurons (15 rows x 15 columns). Changes of the membrane potentials in all cells of the cluster at a particular moment are represented in colors according to the bar shown on the right. To simulate excitability of neurons (represented as nodes in the grid), we used a Hodgkin-Huxley model (Hodgkin and Huxley, 1952). Cells were connected through gap junctions (represented as edges in the grid), whose junctional currents, voltages and conductances were simulated using a 16-state model of voltage gating (Paulauskas et al., 2012). A moderate I-V retification in just a few gap junctions enabled unidirectional transfer of action potentials, and consequentially, the reverberation of exitation in 2-D clusters. In both movies the spread of excitation was initiated by a stimulus of external current applied to a single cell.
Movie 3. The reverberation was initiated by a single external stimulus (Iext ) applied to cell-1/1 (red circle). Initially, excitation spreads upwards to cells 2/1 and 3/1 but not to cells 1/2 and 2/2 due to a local I-V rectification of electrical synapses. Then, excitation returns back to cell-1/1 forming the core of the reverberator. From here, excitation spreads repeatedly over the cluster in the form of spiral waves. Shown below is a record of the transmembrane potential of cell-1/1. The red dot on the Vm record moves in synchrony with changes of isochrones.
Movie 4. The same as in Movie 3, but the reverberation was initiated by a single stimulus applied to cell-8/8 (red circle). Initially, excitation spreads in all direction but not upwards/downwards and to the left but not to cells 9/9, 8/9 and 7/9 due to a local I-V rectification of electrical synapses shown by the red dotted line. Then, excitation returns back to cell-8/8 forming two reverberators. From here, excitation spreads repeatedly over the cluster in the form of two spiral waves colliding with each other. Shown below is a record of the transmembrane potential o cell-8/8.
Movie 5. Simulation of arrhythmias mediated by electrical anisotropy in a cluster composed of 13,231 cardiomyocytes (131 rows x 101 columns). Excitability of cardiomyocytes was described according with Noble (1962). Cardiomyocytes were interconnected through dynamically modulatable GJ channels simulated using a 36-state model of voltage gating (Snipas et al., 2016). All parameters describing excitability and voltage-gating were identical among cardiomyocytes, however values of junctional conductance was 3 fold higher along Y axis than along X axis resulting to electrical anisotropy at 3/1 ratio. Reverberation was initiated by two external stimuli applied to cell-51/10 (red circle). The first stimulus, applied at 10th ms, initiated excitation that spreads through the cluster elliptically due to electrical anisotropy. Then, the second stimulus, applied at 200 ms, caused a spread of excitation only along Y axis and was blocked across it. After making left and right turns (at ~240 ms) excitation returned to stimulation point at ~250 ms. Then excitation once again spreads along Y axis until formation of two spiral waves at ~700 ms. Furthermore, reverberators multiplied resulting to virtually chaotic spread of excitation, resembling fibrillation. Shown below is a record of the membrane potential (Vm) of cell-51/10. The red circle on Vm record moves in synchrony with changes of isopotential areas where colors reflect the Vm of cardiomyocytes at any given time according to the colored bar on the right.
Dynamics of junctional plaques (JPs) and internalized annular (vesicular) junctions in living cells.
Time-lapse imaging of HeLa cells expressing Cx43 fused with EGFP during a 4 h period. Images are superimposed from phase-contrast (red) and Cx43-EGFP fluorescence (green) images. Green curved lines are junctional plaques representing clusters of GJ channels between neighboring cells. Vesicles lined with green are internalized junctional plaques in the process of degradation.
JPs between quiescent and dividing cells bow into the quiescent cell.
Time-lapse imaging of HeLa43-EGFP cells during a 20 hour period. Images are superimposed from phase-contrast (red) and Cx43-EGFP fluorescence (green) images. The dynamics of junctional plaques (JPs) during proliferation of cells can be seen.
A 3-D reconstruction of Cx43-CFP/Cx43-YFP GJs.
A 3-D reconstructed movie of four HeLa cells expressing Cx43-CFP (green) or Cx43-YFP (red) that resulted from 100 Z-series images. Cx43-CFP/Cx43-YFP gap junction channels form a junctional plaque (yellow) in apposition between two cells.
Studies of unapposed/nonjunctional hemichannels by examining dye uptake
Left: Time-lapse imaging during a ~2 h period starting shortly after application of DAPI and propidium iodide. Images in the movie are colour composites of images corresponding to DAPI (blue) and propidium iodide (red). HeLaCx30.2-EGFP cells, but not parental cells, show intense uptake of DAPI. The absence of labeling in red indicates that uptake of propidium iodide was below detection levels and the integrity of the plasma membrane was not compromised. Right: Continuation of Left, recorded from 120 to 350 min, when only EtBr was present in the medium. Cells remain labeled with DAPI (blue) and exhibit continuous uptake of EtBr (in red).
Gap junctional worms
Time-lapse imaging for ~3h of HeLaCx43-EGFP cells during application of calmodulin inhibitor, W-7 (~50 µM). Images in the movie are overlays of phase contrast (red) and EGFP (green) images. Due to W-7 action, each “worm?? or cytoplasmic extension (CE) originates in the junctional plaque (JP) area and it remains inside the CE. CEs are highly transparent, probably due to the reduced presence of organelles, and cells retain coupling that is reversibly blocked by heptanol. The CEs retain a nearly constant diameter throughout their length, which is close to the diameter of the JP, presumably due to elements of the cytoskeleton connected to both sides of the JP in almost equal amounts. We expect that formation of CEs is related to a property of W-7 and other calmodulin inhibitors to bind to the inner leaflet of the plasma membrane reducing its negative electrostatic surface potential, which allows electrostatically bound proteins to translocate from the membrane to the cytoplasm (Sengupta et al., 2007) and make the plasma membrane more fluid.
Triplets! Is it truth or fiction?
It is incontestably accepted that cells divide into two sister cells. This movie shows one of eight observations (approximately ~1 of 1000 divisions) obtained as a ‘by-product’ during time-lapse imaging where the HeLa cell expressing Cx43-EGFP (in green) divides into three sister cells (the lower-left corner). It took ~5.2 h from detachment of the cell from the coverslip until a full division, which was ~1.4 longer than the division into doublets, seen on the top of images. Interestingly, it took ~1 h to form a junctional plaque between two sister cells of the triplet, which indicates how long it takes GJs to form de novo.
Cheese-like gap junctions
The movie ‘A’ shows time-lapse imaging of HeLaCx43-EGFP cells during application of hypertonic (x1.5 by adding polyethylene glycol, PEG-200) solution (at 12 s) and during washout (at 41 s). Shown are large junctional plaques (JP) viewed en face in regions where cells overlap. The hypertonic solution induces the formation of 'holes,' or connexin depleted regions, CDRs; similar CDRs appeared transiently under a temporal exposure to hypertonic solution made by the addition of NaCl sucrose, or by exchanging hypotonic solution to normal. As CDRs form, the mean fluorescence intensity for the entire JP remains essentially constant, whereas fluorescence intensity between the CDRs rises indicating that the packing density of GJ channels in these regions increases. During washout, CDRs gradually reduce in size. Gap junctional cell-cell coupling remained during application of hypertonic solutions and can be blocked transiently with alkanols.
Formation of CDRs can be observed during cell exposure to toxins of the AB5 family, such as Vibrio cholera (ChTX), Shigella dysenteriae, and enterohemorrhagic Escherichia coli, as reported by Majoul et al., 2013 (PNAS). Movie ‘B’ shows formation of CDRs during an exposure of HeLaCx43-EGFP cells to Cholera toxin at 25 s and retraction of CDRs during washout, starting at 65 s of imaging.
Hypertonic solutions and toxins cause an osmotic shock leading to cell shrinkage, as seen in Movie ‘C’, which illustrates a reversible shrinkage of RIN cells exposed to Cholera toxin during the 20 - 60 s time period.