Plankton classification with high-throughput submersible holographic microscopy and transfer learning.

BACKGROUND: Plankton are foundational to marine food webs and an important feature for characterizing ocean health. Recent developments in quantitative imaging devices provide in-flow high-throughput sampling from bulk volumes-opening new ecological challenges exploring microbial eukaryotic variation and diversity, alongside technical hurdles to automate classification from large datasets. However, a limited number of deployable imaging instruments have been coupled with the most prominent classification algorithms-effectively limiting the extraction of curated observations from field deployments. Holography offers relatively simple coherent microscopy designs with non-intrusive 3-D image information, and rapid frame rates that support data-driven plankton imaging tasks. Classification benchmarks across different domains have been set with transfer learning approaches, focused on repurposing pre-trained, state-of-the-art deep learning models as classifiers to learn new image features without protracted model training times. Combining the data production of holography, digital image processing, and computer vision could improve in-situ monitoring of plankton communities and contribute to sampling the diversity of microbial eukaryotes. RESULTS: Here we use a light and portable digital in-line holographic microscope (The HoloSea) with maximum optical resolution of 1.5 µm, intensity-based object detection through a volume, and four different pre-trained convolutional neural networks to classify > 3800 micro-mesoplankton (> 20 µm) images across 19 classes. The maximum classifier performance was quickly achieved for each convolutional neural network during training and reached F1-scores > 89%. Taking classification further, we show that off-the-shelf classifiers perform strongly across every decision threshold for ranking a majority of the plankton classes. CONCLUSION: These results show compelling baselines for classifying holographic plankton images, both rare and plentiful, including several dinoflagellate and diatom groups. These results also support a broader potential for deployable holographic microscopes to sample diverse microbial eukaryotic communities, and its use for high-throughput plankton monitoring.

Osmotic modulation of slowly activating IKs in guinea-pig ventricular myocytes.

AIMS: The aims of the study were to determine the effects of anisosmotic bathing solution on selected properties of I(Ks), the slowly activating delayed-rectifier K(+) current important for repolarization of the action potential in cardiac cells. METHODS AND RESULTS: Guinea-pig ventricular myocytes were voltage-clamped using either the ruptured-patch or perforated-patch technique, and the amplitude, time course, and voltage dependence of I(Ks) were determined before [isosmotic (1T)] and during superfusion of hyposmotic (<1T) or hyperosmotic (>1T) bathing solution. Hyposmotic solution increased the amplitude of I(Ks), and hyperosmotic solution decreased it. Anisosmotic-induced changes in I(Ks) amplitude were complete in 2-5 min, well-maintained, reversible, and not accompanied by significant changes in I(Ks) time course and voltage dependence. There was little difference in the results obtained with the ruptured-patch technique and those obtained with the perforated-patch technique. The amplitude of I(Ks) was sensitive to small (±10%) changes in osmolarity, maximally increased by hyposmotic solution with T < 0.7, and strongly decreased by hyperosmotic solution with T > 1.5. Experimental data on a plot of relative (1T = 1.0) I(Ks) amplitude vs. the reciprocal of relative osmolarity are well-described by a Hill equation that has a lower asymptote of 0.0, an upper asymptote of 2.0, and a slope factor of 1.87 ± 0.07. CONCLUSION: Modulation of I(Ks) amplitude by anisosmotic solution is independent of patch configuration, unaccompanied by changes in current gating, and well-described by a Hill dose-response relation that predicts relatively strong responses of I(Ks) to small perturbations in external osmolarity.

Regulation of wild-type and mutant KCNQ1/KCNE1 channels by tyrosine kinase.

The objective of the study was to investigate the role of tyrosine phosphorylation in the regulation of KCNQ1/KCNE1 channels. Large whole-cell time- and voltage-dependent K(+) currents were present in human embryonic kidney 293 cells cotransfected with human KCNQ1 and KCNE1 but not in control nontransfected cells. The time- and voltage-dependent current had biophysical properties typical of cardiac KCNQ1/KCNE1 current and was almost completely abolished by KCNQ1 blocker chromanol 293B (50 microM). Both KCNQ1/KCNE1 and KCNQ1 current were inhibited in a voltage-independent manner by tyrosine kinase (PTK) inhibitor tyrphostin A25 (100 microM), but not by PTK-inactive tyrphostin A1 (100 microM), suggesting involvement of tyrosine phosphorylation in maintaining channel activity. This view was strengthened by the finding that phosphotyrosyl phosphatase inhibitor monoperoxo(picolinato)-oxo-vanadate(V) (200 microM) reversed the inhibition of current by tyrphostin A25. However, the channel-pertinent tyrosine phosphorylation modulated by these compounds does not appear to be on the channel itself because inhibition of current by tyrphostin A25 was unaffected by single and multiple mutations of KCNQ1 cytoplasmically accessible tyrosine residues. Inhibition by tyrphostin A25 was unaffected by intracellularly applied diC8 phosphatidylinositol-4,5-bisphosphate (diC8 PIP(2); 25 microM), and based on the results obtained from cell surface biotinylation experiments, it was not due to loss of channels from the membrane. We conclude that tyrphostin A25 inhibits KCNQ1/KCNE1 current by lowering tyrosine phosphorylation on unidentified nonchannel protein(s) that directly or indirectly regulate the open probability of the KCNQ1 pore in a PIP(2)-independent manner.

Involvement of tyrosine kinase in the hyposmotic stimulation of I Ks in guinea-pig ventricular myocytes.

The objective of this study was to investigate the involvement of tyrosine phosphorylation in the hyposmotic stimulation of cardiac I Ks, a slowly activating delayed-rectifier K+ current that promotes repolarization of the action potential. The current was recorded from whole-cell-configured guinea-pig ventricular myocytes before, during, and after their exposure to solution whose osmolarity was 0.75 times normal. Exposure to hyposmotic solution caused a near-doubling of the amplitude of I Ks, with little change in the voltage dependence of current activation. Stable, hyposmotically stimulated I Ks (I Ks,Hypo) was decreased by broadspectrum tyrosine kinase (TK) inhibitors tyrphostin A23 (IC50 approximately 5 microM) and tyrphostin A25 (IC50 15.8 +/- 1.6 microM) but not by TK-inactive tyrphostin analogs, suggesting that tyrosine phosphorylation is important for maintenance of the current. In agreement with that view, we found that the TK-inhibitor action on I Ks,Hypo was strongly antagonized by vanadate compounds known to inhibit phosphotyrosyl phosphatase. When myocytes were pretreated with TK inhibitors, the stimulation of I Ks was attenuated in a concentration-dependent manner. The attenuation was not due to concomitant attenuation of a stimulation of tyrosine phosphorylation because neither the stimulation of I Ks nor its rate of decay following removal of hyposmotic solution was affected by pretreatment with vanadates. We suggest that the stimulation of I Ks by hyposmotic solution is dependent on a basal tyrosine phosphorylation that modulates a swelling-induced I Ks-stimulatory signal and/or the receptivity of Ks channels to that signal.

Contribution of KCNQ1 to the regulatory volume decrease in the human mammary epithelial cell line MCF-7.

Using the human mammary epithelial cell line MCF-7, we have investigated volume-activated changes in response to hyposmotic stress. Switching MCF-7 cells from an isosmotic to a hyposmotic solution resulted in an initial cell swelling response, followed by a regulatory volume decrease (RVD). This RVD response was inhibited by the nonselective K(+) channel inhibitors Ba(2+), quinine, and tetraethylammonium chloride, implicating K(+) channel activity in this volume-regulatory mechanism. Additional studies using chromonol 293B and XE991 as inhibitors of the KCNQ1 K(+) channel, and also a dominant-negative NH(2)-terminal truncated KCNQ1 isoform, showed complete abolition of the RVD response, suggesting that KCNQ1 plays an important role in regulation of cell volume in MCF-7 cells. We additionally confirmed that KCNQ1 mRNA and protein is expressed in MCF-7 cells, and that, when these cells are cultured as a polarized monolayer, KCNQ1 is located exclusively at the apical membrane. Whole cell patch-clamp recordings from MCF-7 cells revealed a small 293B-sensitive current under hyposmotic, but not isosmotic conditions, while recordings from mammalian cells heterologously expressing KCNQ1 alone or KCNQ1 with the accessory subunit KCNE3 reveal a volume-sensitive K(+) current, inhibited by 293B. These data suggest that KCNQ1 may play important physiological roles in the mammary epithelium, regulating cell volume and potentially mediating transepithelial K(+) secretion.

Role of kinases and G-proteins in the hyposmotic stimulation of cardiac IKs.

Exposure of cardiac myocytes to hyposmotic solution stimulates slowly-activating delayed-rectifying K(+) current (I(Ks)) via unknown mechanisms. In the present study, I(Ks) was measured in guinea-pig ventricular myocytes that were pretreated with modulators of cell signaling processes, and then exposed to hyposmotic solution. Pretreatment with compounds that (i) inhibit serine/threonine kinase activity (10-100 microM H89; 200 microM H8; 50 microM H7; 1 microM bisindolylmaleimide I; 10 microM LY294002; 50 microM PD98059), (ii) stimulate serine/threonine kinase activity (1-5 microM forskolin; 0.1 microM phorbol-12-myristate-13-acetate; 10 microM acetylcholine; 0.1 microM angiotensin II; 20 microM ATP), (iii) suppress G-protein activation (10 mM GDPbetaS), or (iv) disrupt the cytoskeleton (10 microM cytochalasin D), had little effect on the stimulation of I(Ks) by hyposmotic solution. In marked contrast, pretreatment with tyrosine kinase inhibitor tyrphostin A25 (20 microM) strongly attenuated both the hyposmotic stimulation of I(Ks) in myocytes and the hyposmotic stimulation of current in BHK cells co-expressing Ks channel subunits KCNQ1 and KCNE1. Since attenuation of hyposmotic stimulation was not observed in myocytes and cells pretreated with inactive tyrphostin A1, we conclude that TK has an important role in the response of cardiac Ks channels to hyposmotic solution.

Insensitivity of cardiac delayed-rectifier I(Kr) to tyrosine phosphorylation inhibitors and stimulators.

1. The rapidly activating delayed-rectifying K+ current (I(Kr)) in heart cells is an important determinant of repolarisation, and decreases in its density are implicated in acquired and inherited long QT syndromes. The objective of the present study on I(Kr) in guinea-pig ventricular myocytes was to evaluate whether the current is acutely regulated by tyrosine phosphorylation. 2. Myocytes configured for ruptured-patch or perforated-patch voltage-clamp were depolarised with 200-ms steps to 0 mV for measurement of I(Kr) tail amplitude on repolarisations to -40 mV. 3. I(Kr) in both ruptured-patch and perforated-patch myocytes was only moderately (14-20%) decreased by 100 microM concentrations of protein tyrosine kinase (PTK) inhibitors tyrphostin A23, tyrphostin A25, and genistein. However, similar-sized decreases were induced by PTK-inactive analogues tyrphostin A1 and daidzein, suggesting that they were unrelated to inhibition of PTK. 4. Ruptured-patch and perforated-patch myocytes were also treated with promoters of tyrosine phosphorylation, including phosphotyrosyl phosphatase (PTP) inhibitor orthovanadate, exogenous c-Src PTK, and four receptor PTK activators (insulin, insulin-like growth factor-1, epidermal growth factor, and basic fibroblast growth factor). None of these treatments had a significant effect on the amplitude of I(Kr). 5. We conclude that Kr channels in guinea-pig ventricular myocytes are unlikely to be regulated by PTK and PTP.

Tyrosine kinase and phosphatase regulation of slow delayed-rectifier K+ current in guinea-pig ventricular myocytes.

The objective of this study was to investigate the involvement of tyrosine phosphorylation in the regulation of the cardiac slowly activating delayed-rectifier K(+) current (I(Ks)) that is important for action potential repolarization. Constitutive I(Ks) recorded from guinea-pig ventricular myocytes was suppressed by broad-spectrum tyrosine kinase (TK) inhibitors tyrphostin A23 (IC(50), 4.1+/-0.6 microm), tyrphostin A25 (IC(50), 12.1+/-2.1 microm) and genistein (IC(50), 64+/-4 microm), but was relatively insensitive to the inactive analogues tyrphostin A1, tyrphostin A63, daidzein and genistin. I(Ks) was unaffected by AG1478 (10 microm), an inhibitor of epidermal growth factor receptor TK, and was strongly suppressed by the Src TK inhibitor PP2 (10 microm) but not by the inactive analogue PP3 (10 microm). The results of experiments with forskolin, H89 and bisindolylmaleimide I indicate that the suppression of I(Ks) by TK inhibitors was not mediated via inhibition of (I(Ks)-stimulatory) protein kinases A and C. To evaluate whether the suppression was related to lowered tyrosine phosphorylation, myocytes were pretreated with TK inhibitors and then exposed to the phosphotyrosyl phosphatase inhibitor orthovanadate (1 mm). Orthovanadate almost completely reversed the suppression of I(Ks) induced by broad-spectrum TK inhibitors at concentrations around their IC(50) values. We conclude that basal I(Ks) is strongly dependent on tyrosine phosphorylation of Ks channel (or channel-regulatory) protein.

CESE: Cell Electrophysiology Simulation Environment.

UNLABELLED: Cell electrophysiology simulation environment (CESE) is an integrated environment for performing simulations with a variety of electrophysiological models that have Hodgkin-Huxley and Markovian formulations of ionic currents. CESE is written in Java 2 and is readily portable to a number of operating systems. CESE allows execution of single-cell models and modification and clamping of model parameters, as well as data visualisation and analysis using a consistent interface. Model creation for CESE is facilitated by an object-oriented approach and use of an extensive modelling framework. The Web-based model repository is available. AVAILABILITY: CESE and the Web-based model repository are available at http://cese.sourceforge.net/.

Cardiac Na+-Ca2+ exchanger current induced by tyrphostin tyrosine kinase inhibitors.

Tyrosine kinase (TK) inhibitors genistein and tyrphostin A23 (A23) inhibited Ca(2+) currents in guinea-pig ventricular myocytes investigated under standard whole-cell conditions (K(+)-free Tyrode's superfusate; EGTA-buffered (pCa-10.5) Cs(+) dialysate). However, the inhibitors (100 microM) also induced membrane currents that reversed between -40 and 0 mV, and the objective of the present study was to characterize these currents. Genistein-induced current behaved like Cl(-) current, and was unaffected by either the addition of divalent cations (0.5 mM Cd(2+); 3 mM Ni(2+)) that block the Na(+)-Ca(2+) exchanger (NCX), or the removal of external Na(+) and Ca(2+). A23-induced current was independent of Cl(-) driving force, and strongly suppressed by addition of Cd(2+) and Ni(2+), and by removal of either external Na(+) or Ca(2+). These and other results suggested that A23 activated an NCX current driven by submembrane Na(+) and Ca(2+) concentrations higher than those in the bulk cytoplasm. Improved control of intracellular Na(+) and Ca(2+) concentrations was obtained by suppressing cation influx (10 microM verapamil) and raising dialysate Na(+) to 7 mM and dialysate pCa to 7. Under these conditions, stimulation by A23 was described by the Hill equation with EC(50) 68 +/- 4 microM and coefficient 1.1, tyrphostin A25 was as effective as A23, and TK-inactive tyrphostin A1 was ineffective. Phosphotyrosyl phosphatase inhibitor orthovanadate (1 mM) antagonized the action of 100 microM A23. The results suggest that activation of cardiac NCX by A23 is due to inhibition of genistein-insensitive TK.

Transient outward current carried by inwardly rectifying K+ channels in guinea pig ventricular myocytes dialyzed with low-K+ solution.

There have been periodic reports of nonclassic (4-aminopyridine insensitive) transient outward K+ current in guinea pig ventricular myocytes, with the most recent one describing a novel voltage-gated inwardly rectifying type. In the present study, we have investigated a transient outward current that overlaps inward Ca2+ current (I(Ca,L)) in myocytes dialyzed with 10 mM K+ solution and superfused with Tyrode's solution. Although depolarizations from holding potential (Vhp) -40 to 0 mV elicited relatively small inward I(Ca,L) in these myocytes, removal of external K+ or addition of 0.2 mM Ba2+ more than doubled the amplitude of the current. The basis of the enhancement of I(Ca,L) was the suppression of a large transient outward K+ current. Similar enhancement was observed when Vhp was moved to -80 mV and test depolarizations were preceded by short prepulses to -40 mV. Investigation of the time and voltage properties of the outward K+ transient indicated that it was inwardly rectifying and unlikely to be carried by voltage-gated channels. The outward transient was attenuated in myocytes dialyzed with high-Mg2+ solution, accelerated in myocytes dialyzed with 100 microM spermine solution, and abolished with time in myocytes dialyzed with ATP-free solution. These and other findings suggest that the outward transient is a component of classic "time-independent" inwardly rectifying K+ current.

Selective block of swelling-activated Cl- channels over cAMP-dependent Cl- channels in ventricular myocytes.

The objective of this study on guinea-pig and rabbit ventricular myocytes was to evaluate the sensitivities of swelling-activated Cl- current (ICl(swell)) and cAMP-dependent cystic fibrosis transmembrane regulator (CFTR) Cl- current (ICl(CFTR)) to block by dideoxyforskolin and verapamil. The currents were recorded from whole-cell configured myocytes that were dialysed with a Cs+-rich pipette solution and superfused with either isosmotic Na+-, K+-, Ca2+-free solution that contained 140 mM sucrose or hyposmotic sucrose-free solution. Forskolin-activated ICl(CFTR) was inhibited by reference blocker anthracene-9-carboxylic acid but unaffected by < or = 200 microM dideoxyforskolin and verapamil. However, dideoxyforskolin and verapamil had strong inhibitory effects on outwardly-rectifying, inactivating, distilbene-sensitive ICl(swell); IC50 values were approximately 30 microM, and blocks were voltage-independent and reversible. The results establish that dideoxyforskolin and verapamil can be used to distinguish between ICl(CFTR) and ICl(swell) in heart cells, and expand the pharmacological characterization of cardiac ICl(swell).

Block of cardiac delayed-rectifier and inward-rectifier K+ currents by nisoldipine.

The objective of this study was to determine the concentration-dependent effects of nisoldipine, a dihydropyridine Ca2+ channel blocker, on K+ currents in guinea-pig ventricular myocytes. Myocytes in the conventional whole-cell configuration were bathed in normal Tyrode's solution or K+-free Tyrode's solution for the measurement of the effects of 0.01-100 microM nisoldipine on rapidly activating delayed-rectifier K+ current (I(Kr)), slowly activating delayed-rectifier K+ current (I(Ks)), inwardly rectifying K+ current (I(K1)), and reference L-type Ca2+ current (I(Ca,L)). Nisoldipine inhibited I(Kr) with an IC(50) of 23 microM, and I(Ks) with an IC(50) of 40 microM. The drug also had weak inhibitory effects on inward- and outward-directed I(K1); the IC(50) determined for outward-directed current was 80 microM. Investigation of nisoldipine action on I(Ks) showed that inhibition occurred in the absence of previous pulsing, and with little change in the time courses of activation and deactivation. However, the drug-induced inhibition was significantly weaker at >or =+30 mV than at +10 mV.5 We estimate that nisoldipine is about 30 times less selective for delayed-rectifier K+ channels than for L-type Ca2+ channels in fully polarised guinea-pig ventricular myocytes, and several orders less selective in partially depolarised myocytes.