Several organelles are present adjacent to the plasma membrane
Different organelles were visualized by expressing fluorescently tagged reporter proteins in living cells (
; ). The endoplasmic reticulum was labeled with a fusion protein of YFP and calreticulin (EYFP-ER; CLONTECH Laboratories, Inc.), and the Golgi apparatus was labeled with a fusion protein of CFP and human beta 1,4-galactosyltransferase (GalT-ECFP; Llopis et al., 1998
; Zaal et al., 1999
). For labeling post-Golgi vesicles, we expressed a fusion protein containing EGFP fused to the NH2
terminus of human growth hormone containing four FK506 modified (FM) binding domains (EGFP-FM4-hGH; Rivera et al., 2000
). The FM domains have been shown to cause aggregation of the newly synthesized fusion protein in the ER. When cells are treated with a ligand that specifically binds the FM domain, the fusion protein disaggregates, becoming free to exit the ER. The disaggregated protein then traverses the constitutive secretory pathway, passing via Golgi into post Golgi vesicles before being exocytosed (Rivera et al., 2000
). Early endosomes were labeled with ECFP fused to the COOH terminus of the v-SNARE vesicle-associated membrane protein (VAMP)8-ECFP (Nagamatsu et al., 2001
), whereas late endosomes were labeled with ECFP fused to the NH2
-terminal of Rab7 (Rab7-ECFP; Barbero et al., 2002
Figure 1. Distribution of cellular organelles in the evanescent field. CHO cells were transiently transfected with (A) Calreticulin-EYFP; (B) hGH-EGFP; (C) GalT-ECFP; (D) VAMP8-ECFP; (E) Rab7-ECFP; and (F) CD63-EGFP and imaged (more ...)
Fluorescent tags used to mark the various cellular compartments
For labeling the lumen and the membrane of lysosomes two complementary approaches were used. The membrane was labeled by transiently expressing fluorescent reporter proteins including Synaptotagmin VII-EGFP (Martinez et al., 2000
), lysosomal associated membrane protein (LAMP)1-EGFP (Kornfeld and Mellman, 1989
), CD63-EGFP (Blott et al., 2001
), and VAMP7-ECFP (Advani et al., 1999
). For labeling the lysosomal lumen, the entire endocytic pathway was first loaded with fluorescent dextran followed by a chase of 3–10 h. In these conditions, dextran was retained predominantly in the lysosomes (
A). The fluorescent dextran colocalized with VAMP8-ECFP (diagnostic of early endosomes) only shortly after incubation (unpublished data). 3 h after shifting the cells to dextran-free media, no fluorescent dextran colocalized with VAMP8-ECFP, very little colocalized with Rab7-CFP–labeled vesicles (diagnostic of late endosomes), whereas the bulk of dextran colocalized with vesicles labeled with the lysosomal markers Synaptotagmin VII-EGFP, VAMP7-ECFP, and CD63-EGFP ( A).
Figure 2. Lysosomes are located within 100 nm of the plasma membrane in several cell types. (A) Long-term loading with dextran specifically labels the lysosome. Endocytic compartments of CHO cells transiently transfected with VAMP-ECFP, Rab7-ECFP, SytVII-EGFP, (more ...)
To study only those fluorescently labeled compartments that are near the plasma membrane, cells were observed using TIR-FM. In TIR-FM, the excitatory evanescent field decays exponentially from the interface of the cell membrane and coverslip, thus illuminating a region 70–120 nm from the plasma membrane (see Materials and methods). Thus, fluorophores that are any further from the plasma membrane are very poorly excited. This allows us not only to monitor the vesicles and organelles near the basal cell membrane with a high signal to noise ratio but also minimize the photo-damage due to laser excitation.
Using epifluorescence and TIR-FM microscopy, we observed that compartments containing calreticulin-EYFP ( A, diagnostic of ER), human growth hormone–EGFP conjugated to four FM domains ( B, diagnostic of post Golgi derived vesicles), VAMP8-ECFP ( D, early endosomes), Rab7-ECFP ( E, late endosomes), and CD63-EGFP ( F, lysosomes), were present both deeper in the cell and within the evanescent field. In contrast, the galactosyl-transferase-ECFP ( C, diagnostic of Golgi apparatus; Sciaky et al., 1997
; Llopis et al., 1998
; Zaal et al., 1999
)-labeled compartments were observed only deeper inside the cell and were not visible in the evanescent field.
With the exception of the lysosomes and Golgi apparatus, all of the other organelles mentioned above, including the endoplasmic reticulum (Lysakowski et al., 1999
; Nagata, 2001
), post-Golgi vesicles (Schmoranzer et al., 2000
) and endosomes (Lampson et al., 2001
) have been observed within 100 nm from the plasma membrane. Because the existence of a population of lysosomes adjacent to the plasma membrane has not been characterized, we tested if additional lysosomal markers were observed in the evanescent field ( B). With each of the fluorescently tagged lysosomal membrane proteins used, synaptotagmin VII, Lamp1, and VAMP7, we observed a population of vesicles adjacent to the plasma membrane, within the evanescent field ( B). Further, to examine if a peripheral population of lysosomes was typical of only CHO cells, we repeated the experiment in various cell lines ( C). All the cell lines examined (NIH3T3 fibroblasts, WM239 melanoma, MDCK epithelial cells, normal rat kidney (NRK) fibroblasts, HeLa cells, and murine embryonic primary fibroblasts) possessed a fraction of dextran labeled lysosomes adjacent to the plasma membrane.
Increase in calcium triggers exocytosis of lysosomes, but not of other organelles
To assess the effect of calcium on the movement and distribution of the fluorescently tagged compartments in the evanescent field, and on their ability to exocytose, cellular calcium was increased using the calcium ionophore A23187 (10 μM; Bennett et al., 1979
). After the ionophore treatment, cells were continuously imaged (at 5–10 frames/s) using TIR-FM. Observations were limited to an interval of <10 min, because treatment with the calcium ionophore A23187 often caused cells to round up after longer periods, leading to their disappearance from the evanescent field. Within this interval, the vesicles were analyzed for movement and fusion based on total fluorescence intensity, peak intensity, and the width squared ([width]2
) of the spread of fluorescence. As demonstrated previously (Schmoranzer et al., 2000
), movement of a fluorescent vesicle perpendicular to the coverslip alters the excitation by evanescent field, thereby resulting in changes in the fluorescence emission intensity. A fusion event, leading to the delivery of all membrane proteins of a vesicle to the plasma membrane, is determined by two criteria. First, there is an increase in the peak fluorescence intensity and the total fluorescence intensity (as all fluorophores are delivered to the plasma membrane, and hence better excited by the evanescent wave). Second, there is an increase in the width of the spread of fluorescence (as the vesicular membrane proteins diffuses into the plasma membrane; Schmoranzer et al., 2000
). Upon fusion the rate at which the (width)2
of the fluorescence increases is linear with time, the slope of which is equal to the diffusion constant for the membrane protein. If a vesicle lysed, the (width)2
would increase significantly faster and there would be no net delivery of fluorophores to the plasma membrane.
The organelles and vesicles we examined after stimulation of the cells with calcium ionophore fell into three groups. In one group, the vesicles stopped moving and became stationary adjacent to the membrane, or were stationary during the entire period of observation. In the second group, the vesicles showed synchronous increases and decreases in the total and peak fluorescent intensities, with no significant change in the width. The third group of vesicles showed synchronous increase in total and peak intensities, and a concomitant increase in the width of fluorescence. These results are consistent with the first group of vesicles docking with the plasma membrane but not fusing during the observation period; the second group of vesicles moving in and out of the plane of the evanescent field, without fusing with the plasma membrane; and the third group fusing to the plasma membrane.
In untreated cells, we did not observe the Golgi apparatus in the evanescent field (n = 4; C). The endoplasmic reticulum (n = 5), early endosomes (n = 5), late endosomes (n = 6), and lysosomes (n = 23) were present in the evanescent field (). Based on the above mentioned criteria, these compartments fell into the first or the second group of vesicles: no exocytosis was observed. Post-Golgi transport vesicles were also seen in the evanescent field (n = 6) in CHO cells ( B), which exocytosed at the rate of 6 ± 2/min (n = 4). Treatment with calcium ionophore did not affect the rate of fusion of the post-Golgi vesicles. The Golgi apparatus, ER, early endosomes, and late endosomes were never observed to exocytose in the presence or absence of calcium ionophore (). Further, calcium affected neither the movement nor the number of any of these organelles in the evanescent field.
Effect of ionophore induced calcium increase on exocytosis of organelles in CHO cells
The only compartment that exhibited increased exocytosis in response to an increase in calcium was the lysosome (). In contrast to unstimulated cells (n = 8) where almost none of the CD63-EGFP labeled lysosomes were observed to undergo exocytosis, after the addition of ionophore, many CD63-labeled lysosomes were visualized fusing with the plasma membrane ()
. The total intensity, peak intensity, and the (width)2 of the fluorescence for one of these fusion events is plotted as a function of time in C. These data allow us to clearly distinguish three temporal phases, designated as stationary, rise, and spread.
Figure 3. Lysosomal membrane and lumenal markers confirm that calcium induces exocytosis of lysosomes. NRK cells labeled with either CD63-EGFP (A–C) or CD63-EGFP (D) and 10 kD Texas red–dextran were treated with 10 μm A23187 ionophore and (more ...)
To demonstrate that exocytosing vesicles were bona-fide lysosomes and not biosynthetic transport vesicles carrying newly synthesized CD63-EGFP, or mislocalized protein as a consequence of overexpression, we loaded the lysosomes of CD63-EGFP–transfected cells, as described above, with Texas red–dextran. Most of the CD63-EGFP–labeled vesicles also contained Texas red–dextran ( A and 3 D), and addition of ionophore led to the exocytosis of vesicles containing both fluorophores ( D; Video 2, available at http://www.jcb.org/cgi/content/full/jcb.200208154/DC1
). Analysis of cells containing lysosomes loaded with Texas red–dextran showed that none of the lysosomes exocytose in untreated cells. Similar to the CD63-EGFP–transfected cells, after the addition of ionophore, large-scale exocytosis of dextran-loaded lysosomes was observed ( C, e, and 4, A–D; Videos 1 and 3, available at http://www.jcb.org/cgi/content/full/jcb.200208154/DC1
). Unlike CD63-EGFP, which is a membrane protein, dextran is present in the lumen of the lysosome. Thus, after fusion, whereas CD-63 diffuses in a two-dimensional plane, dextran diffuses in three dimensions, and hence at a much faster rate. This is reflected in the (width)2
and total intensity plots of the fluorescence of the exocytosing dextran (
D), which shows a much faster diffusion of fluorescence compared to diffusion of CD63-EGFP fluorescence ( C). Using the above quantitative criterion established for vesicle fusion (Schmoranzer et al., 2000
), we ascertained that these events were not lysis, but genuine membrane fusion events.
Figure 4. Characterization of lysosomal exocytosis by monitoring the release of a lumenal marker. Lysosomes in NRK cells were loaded with FITC dextran and cells were treated with 10 μm A23187 calcium-ionophore. Although the epifluorescence image (A) predominantly (more ...)
Extent of lysosomal exocytosis is sensitive to cell type and calcium level
To examine if calcium induced exocytosis of lysosomes was unique to CHO and NRK cells, we loaded HeLa, WM239 (melanoma), and murine embryonic primary fibroblast (primary isolates) cells with fluorescent dextran ( C). The cells were then treated with 10 μM calcium ionophore. The extent of lysosomal exocytosis varied depending on the cell type used and also on the amount of calcium in the media (). Among all the cell lines tested, calcium increase induced maximum lysosomal exocytosis in the primary fibroblasts (see B; Table III; Video 1). The weakest response to ionophore induced calcium increase was in HeLa cells. In these cells, the calcium ionophore-induced lysosomal exocytosis was barely detectable until the extracellular calcium was raised to 5 mM or 10 mM, at which point the rate of lysosomal exocytosis was similar to that in CHO cells in media containing 1.2 mM extracellular calcium ().
Cell-type dependence of the extent of calcium induced lysosomal exocytosis
Figure 6. Predocked lysosomes are primarily responsible for calcium induced exocytosis. Cells loaded with fluorescent dextran were treated with 10 μM calcium-ionophore, 0.2 U/ml thrombin, or 20 nM bombesin. The cells were continuously imaged using TIR-FM. (more ...)
Within the first 35 s of the addition of ionophore, no lysosomal exocytosis was seen in any of the cell lines tested (n
= 42). To test if this delayed response was related to the kinetics of calcium elevation or a later step, we monitored the time course of ionophore-induced increase in calcium near the plasma membrane using TIR-FM ()
. Fluo-3-AM and Fura red–AM fluorescence emission ratio imaging has previously been used to monitor cytosolic calcium levels (Floto et al., 1995
). Thus, we loaded CHO cells with both the dyes and simultaneously imaged their emission using a split optics setup (described in Materials and methods). Within 10 s of ionophore addition, the free calcium concentration near the plasma membrane reached the maximum value, and then declined to below the prestimulation levels over the next 150 s (). In contrast, the time elapsed to the first observed lysosomal exocytosis ranged from 35 to 100 s after the addition of ionophore (n
= 42), indicating that under these conditions, there is a time lag between increase in calcium level near the plasma membrane and initiation of lysosomal exocytosis. Once exocytosis was initiated, many more lysosomes fused with the plasma membrane within the next 3–7 min. The lysosomes remaining within the evanescent field after this period (94 ± 1% of the total in CHO cells) did not undergo exocytosis. Further addition of the calcium ionophore A23187 had no effect on this residual population of lysosomes.
Figure 5. Lysosomal exocytosis initiates with a delay after the increase in intracellular calcium. CHO cells were loaded for 20 min with Fluo-3-AM and Fura red–AM, after which 10μm A23187 calcium-ionophore was added (T0s) and the cells were imaged (more ...)
To test whether a physiological stimulus that mobilized calcium from intracellular stores would also lead to lysosomal exocytosis, cells were treated with thrombin (0.2 U/ml) or bombesin (20 nM). Bombesin and thrombin are agonists of surface receptors linked to PLC and phosphoinositide hydrolysis, resulting in IP3 formation and mobilization of Ca2+
from intracellular stores (Neylon et al., 1992
; Van Lint et al., 1993
). Both these agents led to exocytosis of dextran-loaded lysosomes in CHO and in murine embryonic primary fibroblasts. Treatment with calcium channel agonists gave a similar result to the ionophore: there was a significantly higher lysosomal exocytosis in primary fibroblasts (16 to 17%) than in CHO cells (2 to 3%;
, compare A and B).
Calcium induces exocytosis of lysosomes present at the plasma membrane
Based on the spatial localization, there appears to be two populations of lysosomes in mammalian cells which can be distinguished in epifluorescence and TIR fluorescence microscopy ( and ). Although many lysosomes are present in the perinuclear region near the microtubule-organizing center, a smaller number of lysosomes are present as a docked or motile population near the plasma membrane. Because in CHO cells there was a 35–100-s delay between the increase in the intracellular calcium concentration near the plasma membrane and lysosomal exocytosis, we examined the possibility that calcium increase leads to the recruitment of lysosomes from the deeper perinuclear regions of the cells to the sites of exocytosis at the plasma membrane.
To identify which of the two populations of lysosomes, perinuclear or plasma membrane apposed, was responsible for exocytosis we tracked individual lysosomes in CHO cells and murine embryonic primary fibroblasts. Both in the CHO cells ( A) and primary fibroblasts ( B), >80% of the lysosomes that exocytosed were present in the evanescent field before the increase in cytosolic calcium. Moreover, whereas in untreated CHO cells 35% of the lysosomes present in the evanescent field were motile, addition of calcium ionophore led to a rapid loss of movement of these lysosomes (n = 21) ( C). Increase in calcium did not lead to a significant increase in the recruitment of lysosomes to the vicinity of the plasma membrane, but instead (due to exocytosis), over a period of 5–10 min there was a 4 ± 1% decrease in the total number of lysosomes in the evanescent field. Further, the majority of lysosomes in CHO cells (210 out of 270, [~81%]) that underwent exocytosis did not move before undergoing fusion. Similar observations were made with the primary fibroblasts. Thus, addition of the calcium ionophore seems to cause the exocytosis of lysosomes that are apparently docked at the plasma membrane.