- Record: found
- Abstract: found
- Article: found
Going with the membrane flow: the impact of polarized secretion on bulk plasma membrane flows
article-commentary
Publication date (Electronic):
06 December 2021
Journal:
The Febs Journal
Publisher:
John Wiley and Sons Inc.
Keywords:
cell polarity,
endocytosis,
exocytosis,
membrane,
patterning,
secretion,
self‐organization
Read this article at
There is no author summary for this article yet. Authors can add summaries to their articles on ScienceOpen to make them more accessible to a non-specialist audience.
Abstract
Even the simplest cells show a remarkable degree of intracellular patterning. Like developing multicellular organisms, single cells break symmetry to establish polarity axes, pattern their cortex and interior, and undergo morphogenesis to acquire sometimes complex shapes. Symmetry‐breaking and molecular patterns can be established through coupling of negative and positive feedback reactions in biochemical reaction‐diffusion systems. Physical forces, perhaps best studied in the contraction of the metazoan acto‐myosin cortex, which induces cortical and cytoplasmic flows, also serve to pattern‐associated components. A less investigated physical perturbation is the in‐plane flow of plasma membrane material caused by membrane trafficking. In this review, we discuss how bulk membrane flows can be generated at sites of active polarized secretion and growth, how they affect the distribution of membrane‐associated proteins, and how they may be harnessed for patterning and directional movement in cells across the tree of life.
Abstract
Sites of polarized secretion induce lateral bulk plasma membrane flows. Peripheral proteins with fast unbinding and lateral diffusion dynamics can repopulate secretion sites. By contrast, peripheral proteins with long‐enough membrane residence time couple to these flows and are displaced from secretion zones. This phenomenon can be harnessed by the cell for self‐organization.
Related collections
Most cited references45
- Record: found
- Abstract: found
- Article: found
Membrane Lipid Composition: Effect on Membrane and Organelle Structure, Function and Compartmentalization and Therapeutic Avenues
Biological membranes are key elements for the maintenance of cell architecture and physiology. Beyond a pure barrier separating the inner space of the cell from the outer, the plasma membrane is a scaffold and player in cell-to-cell communication and the initiation of intracellular signals among other functions. Critical to this function is the plasma membrane compartmentalization in lipid microdomains that control the localization and productive interactions of proteins involved in cell signal propagation. In addition, cells are divided into compartments limited by other membranes whose integrity and homeostasis are finely controlled, and which determine the identity and function of the different organelles. Here, we review current knowledge on membrane lipid composition in the plasma membrane and endomembrane compartments, emphasizing its role in sustaining organelle structure and function. The correct composition and structure of cell membranes define key pathophysiological aspects of cells. Therefore, we explore the therapeutic potential of manipulating membrane lipid composition with approaches like membrane lipid therapy, aiming to normalize cell functions through the modification of membrane lipid bilayers.
- Record: found
- Abstract: found
- Article: not found
Cargo and Dynamin Regulate Clathrin-Coated Pit Maturation
Dinah Loerke, Marcel Mettlen, Defne Yarar … (2009)
Introduction Clathrin-mediated endocytosis (CME) is the major endocytic pathway in eukaryotic cells. It occurs via clathrin-coated pits (CCPs) that are assembled from cytosolic coat proteins. CCPs capture transmembrane cargo molecules, invaginate, and then pinch off to form clathrin-coated vesicles (CCVs). CME is a constitutive, yet highly regulated process. Biochemical assays of endocytosis score ligand uptake and measure only the ensemble average of successful internalization events, thereby obscuring critical, rate-limiting early stages of maturation and alternative outcomes that might cause variability in individual CCP dynamics. Indeed, live cell imaging has revealed striking heterogeneity in the dynamic behavior of plasma membrane–associated CCPs [1–5]. An important parameter for analyzing CCP heterogeneity is their lifetimes. The lifetime of an individual CCP at the plasma membrane, i.e., the time required for (1) coat initiation, (2) coat propagation, (3) neck constriction, and (4) vesicle budding, is critical for understanding CME. Changes in lifetimes caused by specific molecular perturbations can reveal mechanisms that regulate each of these steps. However, selective probing of all stages of CCP maturation is only possible by mild perturbation of the underlying molecular processes. Detection and interpretation of these necessarily milder phenotypes requires sensitive and comprehensive analysis of individual CCP lifetimes and behavior. To this end, we have employed total internal reflection fluorescence microscopy (TIR-FM), the premier assay to detect early intermediates in CCV formation and visualize the dynamics of CCPs in living cells [1,3–9]. By selectively exciting fluorophores associated with molecular components of CCPs at the ventral plasma membrane, TIR-FM provides excellent signal-to-background ratio and high time resolution. In spite of these strengths, it has remained a challenge to extract reliable measurements of CCP lifetimes from TIR-FM videos. Lifetime measurements are notoriously susceptible to tracking errors, which typically break CCP trajectories into two or more subtrajectories, leading to systematic bias of lifetimes towards shorter values. As a result, tracking has previously been accomplished either manually for a low number of well-discernable, high-intensity CCPs [1,6], or using semiautomated tracking restricted to isolated CCPs, for which no close neighbors are likely to confuse the tracking algorithm [2,4]. Both approaches sample the behavior of arbitrary and typically small subpopulations with relatively uniform properties. To solve these problems and to better exploit the heterogeneity of CCP dynamics as a source of mechanistic information, we have employed particle-tracking software [10] capable of detecting and tracking all CCPs visualized by TIR-FM in an unbiased fashion. Automated detection and tracking enabled analysis of several tens of thousands of trajectories per condition, 100 times more than previous studies, thus providing a comprehensive and accurate measurement of CCP lifetime distributions. Results Three Kinetically Distinct Subpopulations of CCPs We used TIR-FM and our automated tracking assay [10] (see Materials and Methods, Figure S1, and Videos S1, S2, and S4) to obtain large and unbiased datasets of CCP dynamics in well-characterized BSC1 cells expressing a fully functional enhanced green fluorescent protein (EGFP)-tagged clathrin light chain a (LCa-EGFP; Figure 1A and 1B) [2]. To capture both fast events at the timescale of seconds and slower events at the timescale of minutes, we combined data from time-lapse sequences taken at a frame rate of 0.4 s for ≥3 min with data from sequences taken at a frame rate of 2 s for 10–15 min (see Materials and Methods). Objects not detected for at least five consecutive frames were not counted, to exclude transient, highly motile structures. Nonetheless, CCPs displayed a nearly exponential decay of lifetimes (Figure 1C), indicating that a large number appear and disappear on the timescale of a few seconds. To further analyze CCP lifetimes, the raw data were fitted to a series of models that differed in the number and types of subdistributions. Our goal was to identify the minimal number of kinetically distinct subpopulations that could account for the overall lifetime distribution observed (see Figure S2A). Model selection was achieved by three strategies. The first two involved minimization of the Bayesian Information Criterion (BIC) [11,12], which defines the optimal tradeoff between the goodness of fit of the model and the number of free model parameters. The application of the BIC requires a priori knowledge of the distribution of experimental errors, which is unknown for lifetime data. Thus, we performed BIC minimization, first, in a fit of the cumulative lifetime histogram, i.e., with the lifetime and the cumulative frequency as the independent and dependent, error-perturbed variables, respectively; and second, in a fit of the inverse of the cumulative histogram, i.e., with the percentage rank as the independent and the lifetime as the dependent error-perturbed variables. In both cases, we approximated the distribution of the fitting errors as normal. The third strategy for model selection involved a nonparametric test of the distribution of the fitting residuals, which did not require a priori assumptions (see Figure S2). Figure 1 Automated Tracking and Statistical Analyses Detect Three Subpopulations of CCPs at the Plasma Membrane (A) A single frame from a 10-min video of BSC1 cells stably expressing EGFP-labeled clathrin LCa-EGFP (the scale bar indicates 2 μm). (B) Overlay of all trajectories observed during the video. (C) Lifetime distribution for CCPs determined in fast- (grey) and slow-acquisition (black) TIR-FM time-lapse videos. The distribution was determined from the elapsed time between the appearance and disappearance of fluorescent structures present in the time series. n = 43,568 CCP trajectories from 23 cells. (D) Identification of three subpopulations of CCPs. Black dots and smoothed black line show distribution of all CCP lifetimes detected by TIR-FM. These data were best fit by three kinetically distinct subpopulations termed early abortive (blue line), late abortive (red line), and productive CCPs (green line). (E and F) Relative contributions and lifetimes of subpopulations. Error bars represent cell-to-cell variation; the height of the lifetime bar in (F) denotes the t50-spread of the distribution, i.e., the range around the characteristic lifetime that contains 50% of the data. Insets in (C, D, and F) show the data at different scales to emphasize short-lived CCP populations. All three strategies identified three statistically significant subpopulations (Figure 1D) with distinct time constants, but broad and overlapping lifetime distributions (Figure 1F). Importantly, at the noise level of our TIR-FM images, accurate assignment of these subpopulations required analysis of >5,000 trajectories (see Figure S2C), and hence, our results strongly relied on the accurate and automatic tracking of all CCPs in multiple videos per experimental condition (see Materials and Methods). Our data contained two short-lived subpopulations with time constants of 5.2 ± 0.1 s and 15.9 ± 1 s, respectively (±jackknifed cell-to-cell error, see Table S1) that were best fit with Rayleigh distributions, i.e., the shortest and longest lifetimes within the population occur less frequently than the intermediate ones. This suggests that our time sampling of 0.4 s per frame was sufficient to capture all events of significant clathrin coat accumulation. A single longer-lived subpopulation (time constant 86.9 ± 5.8 s) was best approximated by an exponentially decaying distribution. Given its long time constant, accurate measurements of the mean lifetime of this population required imaging for ≥10 min. The longer-lived subpopulation was designated the “productive population,” because (1) its kinetics match those of surface-bound transferrin internalization measured biochemically (t 1/2 ≈ 104 s; Figure S3A), and (2) manual tracking of 450 CCPs (for which internalization was confirmed by sequential disappearance from the TIR-FM and the epifluorescence microscopy (EPI-FM) field [5,6]) also yielded a mean lifetime of approximately 100 s (Figure S3B). Accordingly, we hypothesized that the two shorter-lived species correspond to transient, nonproductive events and therefore termed them “early abortive” and “late abortive” CCPs, respectively. In BSC1 cells, the productive population constituted only 38.6 ± 3.4% of total CCPs at the plasma membrane, with early and late abortive CCPs representing 38.1 ± 3.1% and 21.9 ± 1.4%, respectively (±jackknifed cell-to-cell error; Figure 1E, Table S2). Taking into account the different mean lifetimes and relative contributions of the three individual subpopulations, the mean lifetime of all CCPs is 39 s, much shorter than the half-time of transferrin (Tfn) uptake. This further supports the hypothesis that the two short-lived subpopulations are abortive and do not contribute to Tfn uptake. To determine whether the existence of these three subpopulations was affected by the nature of the fluorescent tag or the cell type used, we performed lifetime analyses on BSC1 and HeLa cells transiently expressing LCa-tomato and NIH3T3 cells stably expressing LCa-DsRed. We consistently observed one long-lived and two short-lived populations (Table S1B), suggesting that this categorization into kinetically distinct CCP populations is a universal phenomenon of CME. As described by others [4,13], however, we also observed in both HeLa cells and NIH3T3 cells, a higher proportion of larger clathrin-coated structures (CCSs) from which multiple CCSs emerge and disappear. These so-called “nonterminal” events are rarely detected in BSC1 cells, consistent with previous findings [2]. We also analyzed CCP lifetimes by tracking the adaptor protein, AP2, in BSC1 cells expressing the EGFP-tagged σ2 subunit (σ2-EGFP), shown not to interfere with AP2 function [2] (Videos S3 and S5). Our model selection again identified three kinetically distinct populations of σ2-containing CCPs (Figure 2C and 2D, Tables S1 and S2); the preference for three versus two subpopulations was weaker but still highly significant (p 30-fold in nearly all cells (Figure 3A–3C). There have been conflicting observations regarding the effect of TfnR overexpression on surface density of CCPs [18–20]. We confirmed previous biochemical measurements [20] showing that at high levels of TfnR overexpression, the endocytic machinery becomes saturated, i.e., the bulk endocytic efficiency of TfnR declines (Figure S3A), thus obscuring any effect that cargo might have on CCP dynamics. In addition, we did not observe an increase in membrane recruitment of either clathrin or the TfnR adaptor protein AP2, as measured by subcellular fractionation and western blot analysis (Figure S3C and S3D) or by TIR-FM (Figure 3D and 3E). Figure 3 Cargo Overexpression in a Tetracycline (tet)-Repressed System BSC1 cells, coinfected with adenoviruses coding for a tet-repressible transcription activator and the human transferrin receptor (TfnR), were cultured in the presence (+) or absence (−) of tet. (A) Validation of infection efficiency by adenoviruses: cells were loaded with Tfn-Alexa 568, and analyzed by epifluorescence. The scale bars represent 10 μm. (B) Whole-cell lysates were analyzed by SDS-PAGE and western blot. (C) Quantitative analysis of surface-bound TfnRs in BSC1 cells stably expressing LCa-EGFP and σ2-EGFP under control (+ tet) or cargo overexpressing (− tet) conditions. Error bars represent standard deviations of three independent experiments. (D) CCPs, visualized by LCa-EGFP and TIR-FM, in control (ctrl) and TfnR overexpressing (cargo o/x) cells. The scale bar represents 5 μm. (E) Average pit density in LCa-EGFP– and σ2-EGFP–expressing cells under control and cargo overexpressing conditions. In cells overexpressing TfnRs, we could no longer detect a significant (p > 0.05) late abortive subpopulation labeled with LCa-EGFP, and this population was completely undetectable in σ2-EGFP labeled cells. The contribution of the productive population increased to 67.9 ± 7.9% of LCa-EGFP–labeled CCPs (Figure 2A) and 76.8 ± 5.6% of σ2-EGFP–containing CCPs (Figure 2C). From this, we conclude that CCPs mature with their highest efficiency when a threshold amount of AP2 and cargo are incorporated into the nascent clathrin lattice. In contrast, the relative contribution of early abortive CCPs is unchanged by cargo overexpression further supporting the notion that these are transient structures assembled in a cargo-independent manner. The mean lifetime of the productive population was only slightly decreased upon cargo overexpression when compared to control conditions (Kolmogorov-Smirnov test [KS-test] p = 0.03, Table S1) in LCa-EGFP–expressing cells (Figure 2B) and showed no decrease in σ2-EGFP–expressing cells (Figure 2D). Thus, we conclude that elevated TfnR concentrations result in more efficient CCP maturation, but that TfnR concentration is not rate-limiting for CCV formation. The approximately 2-fold increase in efficiency of CCP maturation in the presence of an approximately 40-fold increase in cargo concentration points to the existence of other limiting factors and is consistent with the saturation of endocytic efficiency observed biochemically (Figure S3A). Cargo-Enhanced CCP Maturation Requires AP2 Complexes To further probe the role of AP2 adaptors in pit maturation, we decreased cellular AP2 levels by approximately 50% through short-term small interfering RNA (siRNA)-mediated knockdown of the μ2-subunit (unpublished data) [21]. AP2 depletion reduced the densities of all classes of pits (from 0.477 ± 0.012 μm2 in control to 0.276 ± 0.069 μm2 in the knockdown), although the relative contributions of the three populations and their lifetimes remained unchanged (Figure 2A and 2B). This suggests that AP2-dependent nucleation events lead proportionally to both short-lived abortive and productive pits, implying a precursor/product relationship. In contrast to control cells, TfnR overexpression in AP2-depleted cells did not lead to an increase in the relative contribution of the productive population (Figure 2A). We conclude that the increase in CCP maturation efficiency with increased cargo concentration requires AP2 and/or is limited by AP2 concentration. Dynamin Controls Rate-Limiting Early Stages of CME Given that TfnR/cargo concentration only marginally affects lifetimes of CCPs, we next investigated which factor(s) might regulate this aspect of CCP dynamics. The self-assembling GTPase dynamin has been suggested to play a dual role in CME, both as a regulator and/or fidelity monitor during early, rate-limiting steps in endocytosis, and as a well-documented component of the fission apparatus late in CCV formation [22–24]. Dynamin is recruited along with clathrin and AP2 [2], and the early activities of dynamin presumably occur while it is associated with coated pits in its unassembled state, utilizing its basal GTP binding and hydrolysis activities [22,23]. In contrast, dynamin function in membrane fission requires its self-assembly and assembly-stimulated GTPase activities [22,23,25] and occurs subsequent to a burst of recruitment at late stages of CCP formation [5–7]. We sought direct evidence for dynamin's dual role in CME by examining the effects of well-characterized dynamin mutants on CCP dynamics by siRNA-mediated knockdown of dynamin-2 and reintroduction of siRNA-resistant wild-type (WT) or mutant dynamin-1. Because dominant-negative dynamin mutants block endocytosis and lead to clustering of nonproductive CCPs (unpublished data), we focused our analysis on three well-characterized hypoactive dynamin mutants: (1) Dyn1K694A is impaired in self-assembly and hence specifically in assembly-stimulated GTPase activity [22]. Overexpression of dyn1K694A was shown to increase rates of CME [22]. (2) Dyn1S61D exhibits WT GTP binding, but reduced basal and assembly-stimulated GTP hydrolysis rates [26]. Overexpression of dyn1S61D was shown to reduce rates of CME [26]. (3) Dyn1T141A exhibits reduced GTP binding in the unassembled, basal state, but increased basal GTP hydrolysis rates. It also exhibits a slight increase in assembly-stimulated GTPase activity and overexpression of dyn1T141A slightly increases the rate of CME [26]. The effects of dynamin-2 knockdown and expression of these mutants on the relative contributions of the different subpopulations were minor compared to the effect of cargo overexpression (Figure 4A); however, after dynamin-2 knockdown we now detect a substantial increase in the fraction of long-lived (>10 min), “persistent” CCPs that were negligible in control BSC1 cells (Figures 4B, S1E, and S1F). Reintroduction of WT dynamin-1 (dyn1WT) or dyn1K694A reduced the number of persistent CCPs, whereas reintroduction of dyn1S61D or dyn1T141A mutants increased their numbers. Figure 4 Dynamin Regulates CCP Maturation and CCV Budding Cells were either left untreated (ctrl) or transfected with dynamin-2–specific siRNA, after which we reintroduced siRNA-resistant WT or mutant dynamin-1 as indicated. (A) The relative contributions of CCP subpopulations are largely unaffected by the treatments shown. (B) The offset fraction of very long-lived CCPs, termed “persistent,” is increased in siRNA-treated cells. (C) The lifetimes of CCP subpopulations are significantly affected by siRNA treatment and overexpression of dynamin WT and dynamin mutants. A single asterisk (*) and triple asterisks (***) indicate confidence levels of p 99%) to represent productive events. The typical intensity time course is skewed (see example in Figure 5A) and can be divided into three distinct phases: (1) an “assembly phase” corresponding to the initial fast increase of signal intensity, which occurred during the first approximately 20 s of detectable clathrin lattice assembly; followed by (2) a “maturation phase,” during which LCa-EGFP signal intensity plateaus or increases only moderately; followed by (3) a “departure phase,” characterized by a sudden final drop of signal intensity. To measure the duration of these phases, the intensity time courses of individual CCP trajectories were fitted with a smoothing spline (Figure 5A) to identify the points where the approximated slope drops below a set threshold. siRNA-mediated knockdown of dynamin-2 markedly increased the duration of the maturation phase (t-test, p 5,000, a criterion met in each of our analyses, but which greatly exceeds the 100–600 selected events previously assessed in other studies [1–5]. Importantly, subsequent molecular perturbations identified certain conditions in which the contribution of subpopulations to the overall lifetime distribution changed while their time constants were unaffected, whereas other conditions left contributions unchanged while significant shifts occurred in the time constants. This indicates the orthogonality of the two parameters we extracted and establishes that they can be independently determined to probe distinct mechanistic aspects of CCP maturation. In control BSC1 cells expressing LCa-EGFP, we detected three CCP subpopulations (early abortive, late abortive, and productive), with distinct time constants (∼5 s, ∼15 s, and ∼90 s, respectively). A previous analysis of CCP dynamics in the same cells suggested an average lifetime of approximately 46 s, assuming a single population of CCPs [2]. Taking into consideration the different contributions of these subpopulations to the entire ensemble of CCPs, we obtain a value of 39 s, consistent with this previous data. We suspect that the slightly lower ensemble lifetime in our measurements may be associated with a more systematic exclusion of the very faint, short-lived early abortive CCPs in the previous study [2]. The longer CCP lifetimes (60–90 s) consistently reported by others [1,6] reflect selection and analysis of a single subset of typically productive CCPs. The functional assignment of productive CCPs rested on the agreement of the lifetimes of the long-lived subpopulation with biochemical rates of TfnR uptake, as well as with the lifetimes of LCa-EGFP structures tracked manually in quasi-simultaneous TIRF- and epifluorescent images showing CCV internalization. Therefore, we interpret the short-lived CCPs to represent abortive events, but it is also conceivable that they could represent clathrin-coated intracellular structures, e.g., endosomes, that transiently move as visitors through the evanescent field of the TIRF microscope. However, for the following reasons, we think this possibility is unlikely: (1) Although we occasionally observe fast-moving visitors in LCa-EGFP–labeled cells, their displacement between consecutive frames is so much above average that particle correspondences are generally outside the tracking algorithm's self-adaptive search range [10]. Thus, the trajectories of visitors are typically fragmented into short sub-trajectories (less than three frames), similar to trajectories associated with false-positive detections. To exclude both types of false structures from the lifetime analysis, trajectories shorter than five frames were removed from the dataset (see Materials and Methods). (2) Early and late abortive events detected with LCa-EGFP are also detected by our statistical model selection following σ2-EGFP, a selective marker of plasma-membrane–associated CCPs. In addition, the early abortive σ2-labeled CCPs have virtually the same lifetime as early abortive LCa-labeled CCPs, giving us confidence that these are bona fide plasma membrane–associated structures. (3) The relative contributions of both abortive and productive CCPs are affected by transferrin receptor overexpression, and their lifetimes are affected by dynamin. There is no reason why these parameters should be affected for visitors unrelated to CCPs. (4) The strongest indication that the vast majority of the structures we have studied are plasma membrane–associated CCPs comes from the AP2 depletion experiments in which the numbers of all three subpopulations are proportionally decreased. This would not be expected if the shorter-lived species were derived from internal membranes. Furthermore, AP2 depletion prevents the shift to productive CCPs induced by TfnR overexpression. Thus, although it is possible that there is a minor contamination of CCPs by clathrin-coated internal membrane vesicles or clathrin-coated nonendocytic structures, their contribution appears not to be significant enough to affect our findings. Having identified three kinetically distinct subpopulations of CCPs, we next showed that their relative distributions and lifetimes could be affected by systematically manipulating cargo concentration, adaptor protein levels and the level and activity of the GTPase dynamin. Based on the shifts in contribution and lifetime of the three subpopulations we propose that CME might be governed by an endocytosis checkpoint or restriction point, which is regulated, in part, by dynamin. The following observations support the existence of this checkpoint: (1) the identification of abortive and productive CCPs (see also [2]), (2) the finding that AP2-containing (and presumably cargo-enriched) CCPs are more likely to be productive, (3) the finding that cargo load enhances the efficiency of CCV formation leading to more productive CCPs at the expense of abortive ones, and (4) the finding that this effect of cargo concentration is dependent on or limited by AP2 adaptor concentrations. Progression through a restriction or checkpoint requires the tight interaction of a monitor and an activator system. As the major GTPase involved in CME, dynamin was a prime candidate to regulate the endocytosis checkpoint. Two models have been proposed for dynamin function in endocytosis: as a regulatory molecule [27] and as a component of the fission machinery [28–30]. However, these are not mutually exclusive, and recent data support aspects of both [23–26]. Our data on the effects of dynamin depletion and dynamin mutants on CCP dynamics provide several lines of evidence that unassembled Dyn•GTP acts early and controls progression through the endocytosis checkpoint: (1) siRNA reduction of dynamin decreases the rate of both productive CCV formation and turnover of abortive CCPs, whereas overexpression of WT dynamin accelerates each of these rates; (2) a mutant defective in self-assembly (K694A) that is predicted to increase cellular levels of unassembled Dyn•GTP further increases the rates of abortive CCP turnover; and (3) dynamin GTPase domain mutants predicted to be defective in basal GTP binding (T141A) or hydrolysis (S61D) selectively reduce the rates of abortive CCP turnover. Importantly, these conclusions rely on the use of well-characterized, hypomorphic dynamin mutants that mildly alter the kinetics of CME, yet have robust and readily detectable effects when assessed by large-scale image analysis. Strong dominant-negative dynamin mutants that stop CME lead to the accumulation of aggregated CCPs, thus limiting their usefulness for mechanistic interpretation. Dynamin is also positioned to act as a monitor of factors that satisfy restriction point requirements through its many SH3 domain–containing binding partners. These have additional domains that recognize coat proteins (e.g., amphiphysin, SNX9,) membrane curvature (e.g., amphiphysin, endophilin, SNX9), and/or cargo molecules (e.g., SNX9, grb2, TTP). It is known that these proteins can differentially affect dynamin's basal GTPase activity and assembly properties [31–33]. Hence, they provide a potential mechanism for regulating dynamin function in response to these parameters of CCP maturation. Dynamin has also been shown to interact with auxilin and hsc70 [34], thus providing a potential mechanism for the dynamin-dependent turnover of abortive CCPs that we have observed. A model describing these results is illustrated in Figure 6. In this model, productive CCP formation is a stochastic event initiated by the cargo-independent association of AP2 at the plasma membrane, which nucleates clathrin assembly. If a critical mass of the additional factors required for CCP stabilization is not reached during this assembly phase, these structures, which correspond to early and late abortive CCPs, fail to pass through the restriction point and disassemble. We propose that dynamin regulates the checkpoint and controls entry into and progression through the CCP maturation phase. The basal GTP binding and hydrolysis activities may enable unassembled dynamin to function either as a sensor or timer of CCP assembly, through its SH3 domain–containing partners, and thus directly or indirectly control the termination or progression of early endocytic intermediates. Further work will be needed to test this hypothesis and to determine the mechanistic details of how dynamin may monitor CCP assembly and maturation. Figure 6 Model of Endocytosis Checkpoint That Controls Clathrin-Coated Pit Maturation CCVs are formed from productive CCPs, which have undergone a maturation process. An endocytosis checkpoint gates the maturation process, and CCPs that do not progress beyond this restriction point abort by sequential disassembly of AP2 and clathrin. Progression through the restriction point is dependent on the concentration of cargo receptors, AP2 adaptors, and probably other factors. The GTP binding and hydrolysis activities of unassembled dynamin, as revealed by analysis of dynamin GTPase mutants, control the rate and extent of progression through this checkpoint. Dynamin assembly and assembly-stimulated GTPase activities required at late stages of CCV formation are not rate-limiting. See text for details. In sum, we propose that the presence of sufficient cargo, a threshold concentration of AP2 adaptors and perhaps other parameters such as the recruitment of endocytic accessory factors, acquisition of membrane curvature, etc., are detected by dynamin to signal progression beyond the endocytosis checkpoint. Whereas three kinetically distinct subpopulations are detected with statistical significance in our analyses, the lifetime distributions—particularly of the productive population—remain very broad. Thus, we expect that there are other aspects of functional heterogeneity and other factors regulating the endocytosis checkpoint masked within this subpopulation. Future studies involving mild perturbation of other endocytic accessory factors together with comprehensive quantitative analysis of CCP dynamics should provide further mechanistic insight into this functional heterogeneity. Materials and Methods Total internal reflection fluorescence microscopy. TIR-FM was performed on BSC1 monkey kidney epithelial cells stably expressing rat brain clathrin LCa-EGFP or the AP2 rat brain σ2-adaptin fused to EGFP (provided by Dr. T. Kirchhausen, Harvard Medical School, and described in [2]) using a 100 × 1.45 NA objective (Nikon) mounted on a Nikon TE2000U inverted microscope (Nikon). CCP lifetimes range from a few seconds to several minutes. To fully capture this range of dynamics would require image sampling over minutes at a high frame rate ( 200 s, and five to 21 videos with a frame rate 2 s/frame, each from different cells, were acquired over approximately 10 min, using a 14-bit mode operated Hamamatsu Orca II-ERG camera. Image and lifetime analysis. CCPs were detected using à-trous wavelet transform decomposition of the image [35]. Tracking of CCP was accomplished using spatially and temporally global particle assignment described in detail elsewhere [10]. The histograms of CCP lifetimes extracted from the two TIR-FM video categories were merged for the final lifetime analysis by normalizing the relative contribution of the CCP population with a lifetime in the range 30 to 50 s. Thus, the integrals of the measured lifetime probability density functions gi, [0.4] from all N [0.4] videos sampled at 0.4 s/frame and the integrals of the measured lifetime functions fj, [2] from all M [2] videos sampled at 2 s/frame were set to equal values: From the merged lifetime histograms, the underlying distributions of multiple CCP populations with different lifetime dynamics were extracted using both parametric and nonparametric model selection as described in detail in Text S1. For the intensity analysis, we extracted the trajectories of CCPs that were long-lived (>60 s) and isolated (no nearest neighbors within six pixels); this criterion ensured that the chosen CCPs were in fact “productive,” i.e., that they underwent full maturation and internalization, and that there was no crossover to neighbors, both in terms of the physical material of the CCP and in terms of the tracked CCP trajectories, both of which can lead to artifacts in the intensity time course. Online supplementary material. Text S1 contains a more detailed description of methods including particle tracking, lifetime analysis, cell culture, adenoviral infection, cell fractionation, western blotting, immunofluorescence, siRNA transfection, and biochemical measurement of endocytic uptake. Tables S1 and S2 summarize the mean lifetimes and relative contributions of CCP subpopulations in each experimental condition. Table S3 summarizes the intensity phase data. Figure S1 shows data relating to the validation of tracking and lifetime analysis. Figure S2 shows three statistical methods used to identify CCP subpopulations. Figure S3 shows the effect of TfnR overexpression on (1) endocytosis efficiency and (2) subcellular distribution of AP2 and clathrin, and a western blot of dynamin knock-down and corresponding quantification. Videos S1–S6 show examples of particle detection and tracking in simulated videos (Video S6) and those obtained imaging live cells (Videos S1–S5). Supporting Information Figure S1 Validation of Tracking and Lifetime Analysis (A) Videos with CCP density and signal-to-noise similar to real data were simulated with a Dirac lifetime distribution (ground truth; lifetime = 90 frames). Detection and tracking of the CCPs reproduces a Dirac peak at 90 frames, but also generates artificially short lifetimes due to detection or gap-closing failures. (B) Videos were simulated with a Gaussian lifetime distribution (τ0 = 90, σ = 5). Notice that the ground truth distribution jitters because of the limitations of the random number generator. A fit to the lifetime distribution of the detected and tracked trajectories yields τ0 = 89.5, σ = 5.3, statistically indistinguishable from the ground truth (KS-test p = 0.36) (C) Videos were simulated with a lifetime distribution reflecting an exponential decay process (τ = 40). A fit to the lifetime distribution of the detected and tracked trajectories yields τ = 39.8, after correction for the cutoff probability of long trajectories due to the finite video length. Ground truth and reconstructed distribution are statistically indistinguishable (KS-test p = 0.99). (D) Histogram showing the frequency in measured lifetime data of closed gaps of varying lengths. (E) Original lifetime histogram (top) and cumulative lifetime histogram (bottom) of an example dataset. (F) Zoom-in of long lifetimes showing the cutoff of the cumulative histogram (dark blue, solid line) at 300 s and the continuation of the fitted cumulative probability density function (light blue, dashed line). Offset b300 is determined by the relative occurrence of objects of lifetime >300 s. The fit of lifetime distribution models explicitly accounts for a plateau offset bin, which corresponds to the offset of the fitted function at infinity. The plateau offset defines the “persistent” population of long-lived pits that are functionally distinct from the productive population. (234 KB TIF) Click here for additional data file. Figure S2 Separation of Subpopulations with Different Lifetime Distributions (A) BIC test on fit of cumulative lifetime histogram with a mixture-model of Weibull distributions. Blue: fit of lifetime (independent variable) versus cumulative frequency (dependent variable); green: fit of rank (independent variable) versus lifetime (dependent variable). (B) KS-test on residuals of mixture-model fit with different number of subpopulations. Red horizontal line defines p-value of a 95% confidence interval for significantly different residual distributions as determined by bootstrap analysis. All three statistical methods identify three CCP subpopulations in control cells. (C) BIC test on different mixture models with specified shape constants for different numbers of trajectories contributing to the lifetime histogram. (157 KB TIF) Click here for additional data file. Figure S3 Characteristics of Clathrin-Mediated Endocytosis in BSC1 Cells and Effects of Overexpression of Transferrin Receptors (A) Receptor-mediated endocytosis of transferrin is saturable. BSC1 were infected with adenovirus coding for human TfnR in a tet-repressible system. After overnight induction in the absence of tet, the internalization of transferrin was followed in cells overexpressing the transferrin receptor (cargo o/x) and compared to control cells, cultured in the presence of tet (ctrl). After incubation for the indicated times at 37 °C, the amount of internalized transferrin was determined and expressed as a percentage of total surface-bound TfnR. (B) Manual lifetime quantification of productive CCPs. Lifetimes of 450 readily detectable CCPs were determined from the first appearance to their departure. Endocytosis was confirmed by detecting sequential departure from the TIRF and then epifluorescent field. (C and D) Membrane recruitment of coat components. Cells have been cultured as in (A) and analyzed by western blot after subcellular fractionation, S indicates the cytosolic fraction; P the membrane fraction. β-tubulin served as loading control. The cell fractionation experiment has been repeated at least three times, and a typical experiment is shown. (D) Bar graph of quantitation showing the percentage of membrane-associated clathrin and AP2 in control and TfnR-overexpressing cells. Error bars represent standard deviations of three independent experiments. (E and F) si-RNA–mediated Dynamin knock-down. Immunoblot showing dynamin-2 content in BSC1 cells transfected with control siRNA (ctrl) or dynamin-2–specific siRNA (+ dyn siRNA) and quantification of a representative western blot. Actin served as loading control. (647 KB TIF) Click here for additional data file. Table S1 Mean Lifetimes of Clathrin-Coated Pits Detected by TIR-FM Note: k and τ are the shape factor and the time constant (characteristic lifetime) of the population's corresponding Weibull function f(t) = a ·(k/τ) k− 1 · exp(−t/τ) k ; ± t50 denotes the interval on both sides of τ containing 50% of the data in the distribution, i.e. is the jackknifed error (standard deviation [sd]) of the time constant τ, representing cell-to-cell variation. τ m is the mean lifetime (expectancy value) of the distribution, where τ m = τ · Γ(1 + 1/k), so that τ m = τ for k = 1 and fork = 2. ntraj is the number of trajectories measured for the condition, ncells the number of cells. (115 KB DOC) Click here for additional data file. Table S2 Relative Contributions of Clathrin-Coated Pit Subpopulations Detected by TIR-FM (78 KB DOC) Click here for additional data file. Table S3 Intensity Phases of Productive CCPs Detected by TIR-FM (55 KB DOC) Click here for additional data file. Text S1 Details of Methodology Methodology includes cell culture, microscopy, biochemistry, particle detection and tracking, validation, lifetime analysis, and statistical methods. (156 KB DOC) Click here for additional data file. Video S1 Clathrin (LCa-EGFP), Frame Rate 0.4 s Left panel: original image with overlaid detected particle positions and one-frame dragtail of trajectory (red). Right panel: binary image (mask) of à-trous particle detection results with overlaid particle positions (red circles) and closed gaps in trajectory (blue crosses). Note that contrast is adjusted automatically, which may reduce visibility of dimmer objects. Image size 10 × 10 μm, pixel size approximately 67 nm. (6.37 MB MOV) Click here for additional data file. Video S2 Clathrin (LCa-EGFP), Frame Rate 2 s Left and right panels are as in Video S1. (6.35 MB MOV) Click here for additional data file. Video S3 AP2 (σ2-EGFP), Frame Rate 2 s Left and right panels are as in Video S1. (4.98 MB MOV) Click here for additional data file. Video S4 Clathrin (LCa-EGFP) Zoom-In, Frame Rate 2 s Left and right panels are as in Video S1. (2.33 MB MOV) Click here for additional data file. Video S5 AP2 (σ2-EGFP) Zoom-In, Frame Rate 2 s Left and right panels are as in Video S1. (2.09 MB MOV) Click here for additional data file. Video S6 Simulation of CCS Video with Intensity Parameters Chosen to Resemble Measured Data Lifetimes correspond to superposition of three subpopulations as for clathrin control conditions. Left and right panels are as in Video S1. (2.19 MB MOV) Click here for additional data file.
- Record: found
- Abstract: not found
- Article: not found
Cell Membranes Resist Flow
Zheng Shi, Zachary Graber, Tobias Baumgart … (2018)
Author and article information
Contributors
Veneta Gerganova: Veneta.Gerganova@med.lu.se
Sophie G. Martin:
ORCID: https://orcid.org/0000-0002-5317-2557
Sophie.Martin@unil.ch
Journal
Journal ID (nlm-ta): FEBS J
Journal ID (iso-abbrev): FEBS J
Journal ID (doi): 10.1111/(ISSN)1742-4658
Journal ID (publisher-id): FEBS
Title:
The Febs Journal
Publisher:
John Wiley and Sons Inc.
(Hoboken
)
ISSN
(Print):
1742-464X
ISSN
(Electronic):
1742-4658
Publication date
(Electronic):
06
December
2021
Publication date
(Print):
February
2023
Volume: 290
Issue: 3
(
doiID:
10.1111/febs.v290.3
)
Pages: 669-676
Affiliations
Author notes
[*]
[*
]
Correspondence
V. Gerganova and S. G. Martin, Department of Fundamental Microbiology, University of Lausanne, Lausanne, Switzerland
Tel: +41 21 692 3931
E‐mails: Veneta.Gerganova@ 123456med.lu.se ; Sophie.Martin@ 123456unil.ch
V. Gerganova and S. G. Martin, Department of Fundamental Microbiology, University of Lausanne, Lausanne, Switzerland
Tel: +41 21 692 3931
E‐mails: Veneta.Gerganova@ 123456med.lu.se ; Sophie.Martin@ 123456unil.ch
Author information
Sophie G. Martin
https://orcid.org/0000-0002-5317-2557
Article
Publisher ID: FEBS16287DOI: 10.1111/febs.16287
PMC ID: 10078680
PubMed ID: 34797957
SO-VID: d7c4326e-913b-4480-8422-974ff3b5401c
Copyright © © 2021 The Authors. The
FEBS Journal published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical
Societies
License:
This is an open access article under the terms of the http://creativecommons.org/licenses/by/4.0/ License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.
History
Date
revision received
: 07
November
2021
Date
received
: 05
October
2021
Date
accepted
: 18
November
2021
Page count
Figures: 3,
Tables: 0,
Pages: 676,
Words: 5615
Funding
Funded by: SGM
, doi 10.13039/501100000342;
Funded by: ERC
, doi 10.13039/501100000781;
Funded by: Swiss National Science Foundation
, doi 10.13039/501100001711;
Award ID: 310030B_176396
Funded by: Universite de Lausanne
Categories
Subject:
Viewpoint
Subject:
Viewpoints
Custom metadata
source-schema-version-number 2.0
cover-date February 2023
details-of-publishers-convertor Converter:WILEY_ML3GV2_TO_JATSPMC version:6.2.7 mode:remove_FC converted:06.04.2023
ScienceOpen disciplines: Molecular biology
Data availability:
ScienceOpen disciplines: Molecular biology
Keywords: cell polarity, endocytosis, exocytosis, membrane, patterning, secretion, self‐organization