Cell Motility and the Cytoskeleton 38:215–228 (1997) Video-Rate Dynamics of Exocytotic Events Associated With Phagocytosis in Neutrophils Etsuko Suzaki,1 * Hideyuki Kobayashi,1 Yuka Kodama,1 Tsutomu Masujima,1 and Susumu Terakawa2 1Institute of Pharmaceutical Sciences, Hiroshima University School of Medicine, Hiroshima, Japan 2Photon Medical Research Center, Hamamatsu University School of Medicine, Hamamatsu, Japan Exocytotic responses associated with phagocytosis were investigated in a single neutrophil with a special reference to their dynamic properties and their spatiotemporal relationships with ionic and chemical responses during phagocytosis. The real-time sequence of phagocytosis-exocytosis was directly visualized by videoenhanced contrast differential interference contrast (VEC-DIC) microscopy. The actual release of contents from such a granule was proven by examining a cell loaded with quinacrine with a dual imaging system that allowed us to observe DIC and fluorescence images simultaneously at a high magnification. During the process of phagosome formation in a neutrophil engulfing an opsonized zymosan, the exocytotic response was observed first in a granule located near the cell surface initially attached to the zymosan, and then in other granules sequentially along pseudopodia surrounding the zymosan. When the phagocytosis was induced in a medium containing luminol, a chemiluminescence due to active oxidants was detected exclusively in the region of phagosome, suggesting that exocytosis took place on the phagosomal membrane and not on the plasma membrane. Changes in cytosolic free calcium concentration ([Ca21]1) were further measured using fura-2 under the dual imaging system. [Ca21]i transients were more closely related to the extension of pseudopodia for engulfing zymosan and not directly to the exocytosis. These findings lead to a conclusion that exocytosis associated with phagocytosis is initiated by attachment of the cell membrane to the invading organism and mediated by local activation of the phagosomal membrane. Cell Motil. Cytoskeleton 38:215–228, 1997. r 1997 Wiley-Liss, Inc. Key words: video microscopy; chemiluminescence; cytosolic free calcium; rabbit neutrophil INTRODUCTION Phagocytosis is one of the most important functions of the neutrophil to eliminate invading microorganisms. This function involves a series of subcellular activities taking place in a highly coordinated manner, i.e., an attachment of the cell membrane to a microorganism, ingestion of the organism with extended pseudopodia, formation of a phagosome, rearrangement of secretory granules, and release of granular contents into the forming phagosome by exocytosis. Various intracellular signals with molecular and ionic mediators underlie these responses in order to maintain their orchestration. As the r 1997 Wiley-Liss, Inc. exocytotic response is one of the smallest structural changes associated with the phagocytosis, only cinemicrophotographic [Hirsch and Cohn, 1960; Hirsch, 1962] and electron micrographic [Zucker-Franklin and Hirsch, 1964] studies of this response have been done to reveal fine Contract grant sponsor: Grant-in-Aid for Scientific Research from the Ministry of Education, Science and Culture, Japan *Correspondence to: Etsuko Suzaki, Department of Anatomy, Hiroshima University School of Medicine, 1-2-3 Kasumi, Minami-ku, Hiroshima 734, Japan. e-mail: firstname.lastname@example.org. Received 17 December 1996; accepted 23 June 1997 216 Suzaki et al. aspects and some dynamic nature of exocytosis. It was an early histochemical study that suggested sequential degranulation of specific and azurophilic granules in a time range of 0.5 to 10 min [Bainton, 1973], while a later immunocytochemical study suggested a release of bactericidal substances within 5 s after microorganism attachment [Pryzwansky et al., 1979]. Measurement of the membrane capacitance recently provided some insights into dynamics of exocytosis in a single neutrophil [Nüsse and Lindau, 1988]. However, this technique may not easily be applied to the exocytotic response during phagocytosis, because the membranous change for the phagocytosis itself is large and long lasting. Therefore, spatial and temporal relationships between exocytotic responses and other signaling activities involved in phagocytosis have remained obscure. To obtain exact knowledge about these processes, a more accurate and continuing measurement with a higher time resolution would be necessary. An intracellular Ca21 signal, for example, is reported to appear rapidly when a neutrophil attaches to microorganisms and then disappear oscillatory or monotonically [Jaconi et al., 1990]. If so, how is the phagosome formation regulated, and how is the exocytosis initiated? Does exocytosis also start as early as the microorganism attachment or later than the microorganism ingestion? How is the timing of exocytosis regulated to maximize the bactericidal effects in phagosome? In order to obtain clues to answer these questions, it is necessary to determine sequential relationships between differing activities involved in phagocytosis. The differential interference contrast (DIC) microscopy combined with contrast enhancement [Allen et al., 1981] by an advanced image-processing technique [Weiss and Galfe, 1992] has improved spatial and temporal resolutions significantly. Thus, it became a powerful tool for real-time observation of exocytosis in various secretory cells [Terakawa et al., 1991, 1993, 1995a,b; Terakawa and Suzuki, 1991; Kamijo et al., 1993; Sakurai and Terakawa, 1995]. In the present study, we revisited the problems related to dynamic nature of exocytosis in the neutrophil, employing the video-enhanced DIC microscopy. Particular attention was paid to spatial distribution and temporal sequence of exocytosis in reference to a zymosan-induced phagosome formation. The release of active substances from the granules has been visualized and quantified using a fluorescent probe [Kolber and Henkart, 1988; Kawasaki et al., 1991] and a chemiluminescent probe [Yanai, 1979; DeChatelet et al., 1982]. Accordingly, we employed a dual imaging system that allowed simultaneous visualization of a fluorescence image and a DIC image [Foskett and Melvin, 1989]. We improved resolution of the system to a submicron level, and correlated spatiotemporal patterns of granule dis- charge as well as an intracellular Ca21 signal with the DIC images of exocytosis. Here, we describe properties of phagocytosis-exocytosis sequence observed by these video-rated imaging techniques and present evidence that active substances are released from granules to the phagosome with a maximum efficiency for the host defense mechanism. Furthermore, we present a clue to the intracellular signal other than [Ca21]i that may specifically regulate exocytotic responses associated with phagocytosis. MATERIALS AND METHODS Isolation of Neutrophils Neutrophils were isolated from the peripheral blood of the rabbit using a combined dextran/Ficoll-Conray sedimentation procedure [Yanai, 1979]. Contaminating erythrocytes were removed by hypotonic lysis, and neutrophils were finally suspended in a density of 2.5 3 106/ml in Ca21-free Hanks’ balanced salt solution (HBSS(-)). Zymosans Zymosan A from Saccharomyces cerevisiae (Sigma Chemical Co., St. Louis, MO) was opsonized as previously described [Wardlaw and Pillemer 1956]. Briefly, zymosan particles (40 mg) suspended in veronal buffer (20 mM, pH 7.4) were boiled, incubated with fresh rabbit serum at 37°C for 30 min, washed twice with Hanks’ balanced salt solution (HBSS), and resuspended in the same buffer. The opsonized zymosans (44 µg/ml) in HBSS were placed in a chamber and settled on a silicone-coated coverslip for 10 min. The excess amount of solution was removed immediately before use. Observations of Phagocytosis and Exocytosis The neutrophil was prewarmed at 37°C for 3 min and placed in HBSS in a Flexiperm-Disc (Heraeus Biotechnology, Hanau, Germany) of which the bottom was made of a silicone-coated coverslip. Opsonized zymosans were then added to the neutrophil preparation at a concentration of 150 µg/ml. They were observed under inverted microscopes (Axiovert 35 and 135-TV, Zeiss, Oberkochen, Germany) equipped with 3100 DIC objective lenses and 32.5 insertion lenses. Images were obtained with a CCD camera (SSC-M350, Sony, Tokyo). The contrast of video image was enhanced by about 5-fold with a digital image processor (PIP-4000, ADS, Osaka). The process of phagocytosis was monitored and recorded with a videotape recorder (AG-7735, Panasonic, Tokyo). Phagocytosis and exocytosis of a neutrophil were investigated in detail from the recorded images. The exocytotic process of a single granule was analyzed in detail from the change in light intensity of a small spot in Phagocytosis-Exocytosis Sequence in Neutrophils the recorded DIC images where exocytosis of a granule occurred. For this calculation, a series of images were digitized at an 8-bit resolution and stored in the frame memories. For further analyzing exocytosis, the recorded images were processed through the sequential subtraction (time differentiation) mode, (ARGUS-100 and ARGUS20, Hamamatsu Photonics, Hamamatsu [Terakawa et al., 1991]) by which the rapid exocytosis was selectively distinguished from the slower granule movements or other static configurations. The time differentiation processing was done by subtracting an image from the previous image consecutively; i.e., image 2 was subtracted from image 1, and image 3 was subtracted from image 2, and so on. Resultant images were continuously displayed in real time with the contrast enhanced by about 2-fold. The exocytotic responses were extracted as bright and dark spots, and were counted for constructing a frequency histogram of a single cell. The sites of exocytosis were traced also from this differential display of the video record. Simultaneous Observations of DIC and Quinacrine Fluorescence Images Neutrophils were incubated with quinacrine (20 µM) in HBSS at 37°C for 10 min. After the incubation, the cells were placed on the coverslip and were observed under an inverted microscope (Axiovert 35, Zeiss) equipped with a 3100 DIC objective lens and a 32.5 insertion lens. The cells were maintained on the microscope stage at 32–36°C by continuous perfusion with an oxygenated and warmed HBSS. Both DIC and fluorescence images were obtained with a dual imaging system [Foskett and Melvin, 1989]. For the DIC image, a long wavelength of 650 nm was used. This light was separated from the fluorescence by a dichroic mirror (half-reflection wavelength 5 580 nm), detected by an intensified CCD camera (WV-150, Panasonic), and the image was enhanced with an image processor (PIP-4000, ADS). For the fluorescence image, a combination of a blue excitation filter (bandpass at 470 nm), a dichroic mirror (half-reflection wavelength 5 490 nm), and a short cut filter (half-pass at 520 nm) were used. The fluorescence image was detected with another intensified CCD camera (DAS-512, Imagista, Tokyo) equipped with a frame memory by which fluorescence intensity was accumulated for 66 ms. The fluorescence image was then enhanced 2- to 3-fold with an analog amplifier of video signal (410, Suncomm Technical, Tokyo). Both DIC and fluorescence images during the phagocytosis-exocytosis sequence were recorded simultaneously with two video tape recorders (AG-7735 and AG-7500, Panasonic). For analyzing exocytosis further, the recorded images were processed through the sequential subtraction (time differentiation) mode [Terakawa et al., 1991]. 217 Chemiluminescence Imaging Neutrophils were settled on a coverslip in a Flexiperm-Disc placed on an inverted microscope (DiaphotTMD, Nikon, Tokyo) equipped with a 3100 objective lens and an incubation box (37°C). After 5 min of prewarming, a reaction medium containing zymosans (300 µg/ml) and luminol (10 µg/ml, Nakarai Chemical Co., Tokyo) in HBSS was added to the chamber. The neutrophils were observed initially in the brightfield mode. As soon as the initiation of the phagocytosis was confirmed, all illuminations were turned off, and a chemiluminescence was recorded. The chemiluminescence derived from active oxidants was detected with a photon-counting camera (C2400-20, Hamamatsu Photonics) coupled with an image processor (ARGUS-100, Hamamatsu Photonics). Two-dimensional images of chemiluminescence were accumulated for 2 min in a frame memory. Brightfield images obtained before and after the chemiluminescence measurement were also recorded. The chemiluminescence image superimposed on the brightfield image was taken to indicate the loci of oxidant production. Simultaneous Observations of DIC and Fura-2 Fluorescence Images Neutrophils were prewarmed at 37°C for 10 min and incubated with fura-2/AM (2 µM) in HBSS (-) which contained 0.5% bovine serum albumin (BSA) for 30 min. After the incubation, the cells were placed on a zymosanattached coverslip in a chamber. The cells were maintained on the microscope stage at 32–36°C by continuous perfusion with an oxygenated and warmed HBSS, and were observed using the dual imaging system. For estimating intracellular Ca21 concentration ([Ca21]i), fluorescence images were taken with an intensified CCD camera (DAS-512, Imagista) at an emission wavelength of 510 nm and alternative excitations at 380 nm (yielding a [Ca21]i-dependent fluorescence) and 360 nm (yielding a [Ca21]i-independent fluorescence). Interference filters of a 10-nm bandwidth were used to select the wavelengths. These images were accumulated for 0.5 s alternately and continuously during the phagocytosis-exocytosis sequence, and were further processed by an image processor (ARGUS-100, Hamamatsu Photonics) to calculate the ratio value of the fluorescence intensities (I360 nm/I380 nm). The DIC image was obtained simultaneously with these fluorescence images by illuminating at 650 nm (bandwidth of interference filter 5 40 nm) and separating the light with a dichroic filter (half-reflection wavelength 5 570 nm). Another intensified CCD camera (WV150, Panasonic) was used for detecting the DIC image. 218 Suzaki et al. RESULTS Phagosome Formation and Exocytotic Responses Under the video microscope with the proper contrast enhancement, a whole sequence of the phagocytosisexocytosis response was clearly visible (Fig. 1). Neutrophils attached easily to the bottom of the microscopic chamber made of a coverslip, and moved slowly in various directions. They spread and retracted pseudopodia or lamellipodia repeatedly, and transported granules back and forth intracellularly. Zymosans, when added to the chamber, sank to the bottom, but did not attach to it. Instead, they fluctuated in a Brownian movement near the bottom. A glass-attached neutrophil approached them in 20–30 s, and formed a contact with them using a short hand-like structure. An exact moment of the firm attachment to a zymosan could be accurately determined by its abrupt cessation of the Brownian movement. After a certain lag period following the capture of zymosan, the neutrophil extended new pseudopodia around the zymosan and engulfed it into an intracellular space, i.e., a phagosome (Fig. 1). By changing the focal plane up and down, it was confirmed that the zymosan was totally surrounded by the cytoplasm of the neutrophil. The lag period from the neutrophil-zymosan attachment to the emergence of pseudopodia was 34.1 6 3.4 s (mean 6 S.E.M., n 5 30) and the time from the latter to a complete closure of a phagosome was 36.4 6 3.2 s (n 5 30, Fig. 2). During the phagosome formation, a rather rapid transport of granules was observed from the tail to the head of a neutrophil, i.e., in the direction towards the captured zymosan. The nucleus of the neutrophil also moved together with the granules to some extent toward the phagosome. Soon after these morphological changes, many of the granules located around the phagosome stopped moving and then abruptly disappeared one by one. This was a very characteristic response which gave an impression of ‘‘popping’’ or ‘‘flickering’’ of granules, in a resemblance to a burst of bubbles. We referred to this response as ‘‘popping.’’ The images obtained through the sequential subtraction mode selectively distinguished the popping of a granule from other configurations. We refer to this response as exocytotic response. The clarity of these images was high enough to determine the sequential relationships of the exocytotic responses with other responses (Fig. 2). In many cases, the initial exocytotic responses were observed while the pseudopodia still extended to engulf the zymosan. Later, similar responses continued also after the complete closure of the phagosome. These responses were release of granular contents into the phagosome which was isolated partially at the beginning and then completely from the extracellular space. The percentage of enclosure of a zymosan into the phagosome was calculated as a ratio of the pseudopodial length to the circumference of zymosan. This percentage, measured at the moment of the first exocytotic response, varied from 27% to 100% and was 69% 6 21% (mean 6 S.D., n 5 24) on average. Timewise, the first exocytotic response occurred at 59.1 6 4.3 s (n 5 30) after the capture of a zymosan. Subsequently, many exocytotic responses followed and continued for a period of 38.1 6 3.2 s (mean 6 S.E.M., n 5 30; Fig. 2). Thus, it took 97.2 6 5.5 s (n 5 30) for a whole sequence of phagocytosisexocytosis responses to complete. Visualization of Exocytotic Discharge With a Fluorescent Probe Quinacrine-loaded neutrophils were observed using the dual imaging system. In order to avoid a heat absorption or a photodynamic action, we started illuminating the cell with an excitation light after observing the initial sign of the phagocytosis in the DIC image. In the fluorescence image, the cells showed many bright granules of 0.4–1.0 µm in diameter. Many of these granules surrounded the zymosan which appeared dark as a result of negative staining. Exactly when a granule showed ‘‘popping’’ in the DIC image and in the time differential image, a bright granule abruptly disappeared at the same site in the fluorescence image (Fig. 3). The disappearance was so quick and discrete that it was easy to distinguish from the fading of fluorescence due to photobleaching. These observations indicated that the popping of a granule in the DIC image was in fact the image of exocytosis accompanied with the discharge of granular contents. The rate of fluorescence change indicated that about 66 ms were necessary for a single granule to discharge its fluorescent content. Dynamics of Exocytosis Dynamic aspects of exocytosis were further analyzed at a time resolution of the video rate. One of the major factors that alters the light intensity of a small spot in the DIC image is a change in refractive index resulting from a change in density of substances in that spot. Fig. 1. Video micrographs of a neutrophil showing a whole sequence of phagocytosis and exocytosis. a: Onset of phagocytosis. A neutrophil (N) attached to a zymosan (Z) using a hand-like structure (H) for more than 30 s. b: Initiation of phagosome formation. Pseudopodia started to extend after a lag period. The front line of pseudopodia lay between the arrows. The image was taken 10 s after a. c: Phagosome formation. Pseudopodia extended (arrows) further around the zymosan. The image was taken 23 s after a. d: Completion of phagosome formation. Tips of pseudopodia fused together to form a phagosome. The image was taken 46 s after a. Nc, nucleus. e–h: High magnification images of the same neutrophil at the stage shown in d. A granule undergoing exocytosis is indicated by arrows. The interval between images was 33 ms. Inset: One of the time differential images made by sequential subtraction of g from f. A small area including the granule indicated by the arrow is shown at the same magnification. The exocytosed granule was selectively extracted. Phagocytosis-Exocytosis Sequence in Neutrophils Figure 1. 219 220 Suzaki et al. Fig. 2. Time course of the phagocytosis-exocytosis sequence. Relationships between a neutrophil and a zymosan in three stages of the phagocytosis are schematically drawn on the bottom: left, attachment to a zymosan; middle, extension of pseudopodia; and right, the completion of phagosome formation. Average periods (n 5 30) of the lag time, phagosome formation, and exocytosis are shown on the top with their onset time relative to each other. Time zero indicates the time of attachment of the neutrophil to a zymosan. Therefore, the observed change in light intensity would reflect the process of discharge of the granule contents and flattening of the omega-shaped membrane. Figure 4 shows representative time courses of light intensity change obtained from four granules that underwent exocytosis. The light intensity changed from a stable level to a sharp peak within 33–66 ms and then to the original level in about 200 ms. These kinetics of exocytotic response measured in a single granule were faster than those observed in some exocrine cells [Terakawa et al., 1993, 1995a; Kamijo et al., 1993]. Site of Exocytotic Responses For counting the frequency and determining the exact site of individual exocytotic responses, the images processed through the sequential subtraction (time differentiation) mode were analyzed in detail. Through this processing, a secretory granule showed abrupt and rapid change in light intensity at the exocytotic moment and eventually disappeared (Figs. 1, inset; and 3c). The frequency of such discrete changes of individual granules was counted in a neutrophil as a measure of their exocytotic activity (Fig. 5). This activity reached a peak 5–20 s after the appearance of the first exocytotic response. Granules ranging from 14 to 95 in number (36 6 17 granules, mean 6 S.D.; n 5 30) underwent exocytosis during a single sequence of phagocytosis. Occasionally, a neutrophil captured 2, 3 or 4 zymosans sequentially at different portions, forming several phagosomes at some intervals. In such cases, the exocytotic activity was always higher during the first phagosome formation than the later. In the neutrophil that engulfed a zymosan, exocytosis occurred only at the periphery of the phagosome and never on the cell surface directly facing the medium. This was in sharp contrast to the exocytotic responses induced by application of chemical stimulants like concanavalin A and cytochalasin E. These stimulants induced similar popping responses of granules on the surface of a neutrophil without any preferred loci (data not shown). Phagocytosis-Exocytosis Sequence in Neutrophils Fig. 3. Video micrographs of DIC and fluorescence images simultaneously obtained at the moment of an exocytotic response in a neutrophil. The cell was previously loaded with quinacrine. a: Fluorescence image. a1: A whole view of a neutrophil engulfing two zymosans (asterisks) at both ends. a2–a5: Sequential video images enlarged from the area indicated by the white frame in a1. b: DIC image. b1: A whole view of the same neutrophil. Asterisks indicate zymosans. b2–b5: Sequential video images enlarged from the area 221 indicated by the black frame in b1. c: Differential image made from the DIC image. Image c1–2 was made by subtracting image b2 from the previous image, image c2–3 by subtracting b3 from image b2, and so on. A bright spot in a quinacrine-loaded neutrophil abruptly disappeared (white arrows in a) when a granule showed an abrupt popping response in DIC image (black arrows in b and c). The interval between each micrograph was 66 ms. delayed, the first exocytosis was observed at about 3 µm apart from this area. As the pseudopodia extended to engulf the zymosan, the region active for exocytosis shifted from this proximal area to the distal area along the periphery of the phagosome (Fig. 7). Although many granules were accumulated around a phagosome, only a fraction of them underwent exocytosis at its periphery. Luminol-Dependent Chemiluminescence Fig. 4. Time course of a single exocytotic response observed during the phagocytosis. The relative changes in light intensity of a single granule were measured from the video records and plotted at a time resolution of video rate. Four representatives are shown. The decrease in light intensity reflects extrusion of granule contents, and the recovery phase reflects retrieval of the omega-figured membrane. When a neutrophil initially attached to a zymosan, a small portion of the cell membrane, about 2 µm in length, served as an attachment site. The first exocytosis always appeared near this place where the neutrophil first recognized the foreign object. Figure 6 shows the localization relationship between the site of initial attachment and the site of the first exocytosis. Mostly, the first exocytosis occurred within 1 µm of the initial attachment site. In some cases in which the initiation of exocytosis was When phagocytosis of a neutrophil was induced by a zymosan in a medium containing luminol, a significant chemiluminescence was observed. About 30 photons were counted from a single cell during a 2-min period of its highest activity. This period roughly paralleled with that of exocytosis. Luminol-dependent chemiluminescence activities measured by photon-counting imaging were certainly localized to a narrow region in the cell (Fig. 8). Superposition of this chemiluminescence image on the brightfield image revealed that chemiluminescence activity was high in the region of the phagosome. There was a tendency that the highest chemiluminescence appeared along the periphery of the zymosan. [Ca21]i Measurement With Fura-2 Illumination of fura-2-loaded cells with the UVlight tended to suppress their exocytotic responses. Because of this, continuous observation of fura-2 fluorescence in the neutrophil through the whole phagocytosis-exocytosis 222 Suzaki et al. Fig. 5. Frequency of exocytosis during the phagocytosis of a single zymosan in a neutrophil. The number of exocytosis in every 5 s was counted in 30 neutrophils after processing the video with the time differentiation mode (see Figs. 1, inset; and 3c). The time 0 is set at the first appearance of exocytosis. sequence was very difficult. In practice, we started the fluorescence illumination when extension of pseudopodia for engulfing the zymosan was recognized in the DIC image. Relationship between the exocytosis and [Ca21]i increase was thus analyzed from simultaneous video recordings. Figure 9 shows DIC images obtained simultaneously with [Ca21]i images during formation of a zymosan-engulfing phagosome. A rapid increase in [Ca21]i was observed in the whole cytoplasm when pseudopodia started to extend. [Ca21]i was not particularly high in the region where the exocytotic response was frequent (in more than 10 cells). The time course of [Ca21]i changes was calculated from the fluorescence images (Fig. 10). The high [Ca21]i level was maintained for about 30 s and then gradually decreased to the basal level (Fig. 10). The rising phase of the [Ca21]i response occurred simultaneously with the initiation of phagosome formation but not with that of exocytosis. In a neutrophil shown in Figures 9 and 10, 37 exocytosis were observed in the DIC image at the late stage of the [Ca21]i response which was already gradually decreasing and subsiding (Fig. 10). On the contrary, a control cell which did not show phagocytosis maintained almost the same [Ca21]i level (Fig. 10). DISCUSSION Phagocytosis Many studies have already shown phase-contrast and fluorescence pictures of neutrophils undergoing Fig. 6. Locational relationship between the initial attachment site to a zymosan and the site of the first exocytosis occurrence during the phagocytosis in a neutrophil. The first exocytosis in 15 neutrophils was traced (open circles) after aligning both the attachment site on the pseudopodium (the hand structure) and the position of zymosan. Phagocytosis-Exocytosis Sequence in Neutrophils Fig. 7. Localization of exocytosis during phagocytosis of a neutrophil. Top: Site of exocytosis in reference to the location of zymosan traced sequentially from the video images. Small circles indicate the site where each exocytosis was observed. Thirty-four exocytotic responses were observed in this example. Every 10 responses are sequentially grouped depending on their relative time of response (in phagocytosis [Hirsch, 1962; Pryzwansky et al., 1979; Sawyer et al., 1985; Jaconi et al., 1990]. These pictures have been widely accepted as the images of particle ingestion. It was demonstrated by these pictures that a neutrophil extended pseudopodia around a particle [Pry- 223 seconds), and are represented by different symbols. Bottom: Distance of the responding site measured from the site of the first exocytosis. The ordinate represents the order of responses, and the abscissa represents the distance with directions (right and left in reference with the longitudinal axis of neutrophil). zwansky et al., 1979], and that the particle was isolated from outer medium and was engulfed within a neutrophil [Jaconi et al., 1990]. Our DIC and fluorescence images were fully consistent with these observations. By changing the focal plane up and down, it was confirmed that the 224 Suzaki et al. Fig. 8. Luminol-dependent chemiluminescence observed in neutrophils during phagocytosis of zymosans. a: Brightfield image of neutrophils. Zymosans ingested by neutrophils are indicated by arrows, and a lymphocyte by an asterisk. b: Photon counting image of the same neutrophils immersed in luminolcontaining solution. c: Brightfield image and photon counting image superimposed. White dots represent detection of single photons derived from active oxidants. The brightness of the dots is proportional to the number of photons detected. cytoplasm of a neutrophil was present all around a zymosan. In the fluorescence image, the quinacrineloaded granules surrounded the dark zymosan area. Therefore, we conclude that the observed images really reflect the ingestion of particles by phagocytosis as shown by the early studies. The video (VEC-DIC) microscopy allowed us to observe exocytotic events associated with phagocytosis at the highest possible magnification at a time resolution of the video rate (33 ms in NTSC format). By using this technique, we have succeeded in visualizing ‘‘popping’’ responses of the granules, and concluded that these popping responses are exocytosis in the neutrophil (in the sense that granule contents are released transmembranously to a large fluid space) because of the following: (1) the responses involved decrease in refractive index, and thus decrease in density inside the granule (Figs. 1 and 4); (2) they were associated with a loss of dye from the granules as evidenced by the fluorescence microscopy (Fig. 3); (3) they were clearly distinguishable from the translocational movement of granules in neutrophils (Kawai et al., submitted for publication); (4) they were extremely rapid in comparison to the phagocytotic movement (Fig. 4); (5) after such responses, granules were not recovered by readjustment of the microscope focus (indicating degranulation); (6) they were induced only when phagocytosis took place and were never observed without stimulation; (7) they were similar to responses observed in neutrophils stimulated with concanavalin A and cytochalasin E (our unpublished observation); and (8) they were similar to exocytotic responses found in other secretory cells, namely chromaffin cells [Terakawa et al., 1991, 1993, 1995b], pancreatic b-cells [Sakurai and Terakawa, 1995], colonic goblet cells [Terakawa and Suzuki, 1991], salivary acinar cells [Segawa et al., 1991], nasal epithelial goblet cells [Kamijo et al., 1993], and gastric mucosal cells [Terakawa et al., 1995a]. Our technique therefore attained a spatiotemporal resolution high enough to study exocytosis associated with the phagocytosis, and thus provided an insight into the signaling pathway for such exocytosis in the neutrophil. Spatiotemporal observations described in the Results suggest that attachment of the cell membrane to invading organisms may send some signals for initiation of exocytosis. Efficient Sequence of the Granule Fusion A previous study showed that the two types of granules in neutrophils discharge sequentially, specific granules fusing with the phagosome before azurophils [Bainton, 1973]. Others reported that they fuse at random [Zucker-Franklin and Hirsch, 1964; Pryzwansky et al., 1979]. As the sequence of granule discharge roughly parallels the change in pH, thereby providing optimal conditions for coordinated activity of granule contents, the former view is attractive to have been widely accepted. However, both views have not yet been evidenced substantially. In our observation, both relatively small and relatively large granules seemed to undergo exocytosis randomly. Although we could not distinguish specific granules from azurophilic granules precisely in the DIC image, a wide range of the granule size suggests random exocytotic sequence of the two classes of gran- Phagocytosis-Exocytosis Sequence in Neutrophils 225 Fig. 10. Time courses of [Ca21]i transient in neutrophils with (filled circle) and [Ca21]i without (open circle) phagocytosis. The ordinate represents fluorescence ratio (I360 nm/I380 nm), and the abscissa experimental time with the origin at the moment of pseuodpodial extension. Data shown with filled circles were measured from the whole area of the fura-2-loaded neutrophil shown in Figure 9. The phagosome was formed during the period indicated by the open bar, and exocytosis was observed during the period indicated by the filled bar. From the data obtained by ordinary fluorescence ratiometry without DIC imaging, the resting level of [Ca21]i was estimated to be 270 nM, and the peak value of [Ca21]i to be 600 nM. Fig. 9. Sequential video micrographs of a neutrophil showing its shape and [Ca21]i transient during phagocytosis of a zymosan observed by the simultaneous imaging system. a–f; DIC images (upper) and coincidental fluorescence ratio (I360 nm/I380 nm) images (lower) observed after loading the neutrophil with fura-2. The brightness in the ratio images indicates an increase in [Ca21]i. Pseudopodia started extending at time zero. Images were taken at 5 s (a), 25 s (b), 45 s (c), 55 s (d), 75 s (e), and 85 s (f) after the pseudopodial extension. The arrow in d indicates the site of fusion of extended pseudopodia. Exocytotic responses were observed in the period including d and e. ules. Even random exocytosis may be able to achieve adequate pH conditions for each granule content to function, as mature neutrophils contain much smaller population of azurophils than specific granules. It should be emphasized that a considerable lag period exists after attachment of a zymosan and before actual initiation of exocytosis. This lag period, commonly found in phagosome-lysosome fusion in macrophages [Zimmerli et al., 1996, see below], is probably very important in phagocytosis. Since a neutrophil attaches to a zymosan at any part of the cell membrane during random movements, it probably reorients cytosolic organelles including granules and the nucleus to positions most proper for the defense reaction. Exocytosis, occurring with this latency, continues only for the late 38-s period during the whole phagocytosis process that lasts for 97 s (Fig. 2). Oxygen radicals are released to phagosome during this period (Fig. 8); see also Suzuki et al. . The first exocytosis takes place when the zymosan is engulfed by about 70% into the phagosome. These are very effective tactics for the bactericidal action, because the secretions from the granules can challenge the foreign object at the earliest timing without decreasing their concentration in a narrow space between the phagosomal membrane and the object. The long lag time makes this possible. Although the first exocytosis occurs before a complete closure of the phagosome, this exocytosis seems to be properly regulated to occur at an optimal time in a localized region close to the bottom of phagosome where the possibility for the secretions to leak outside (i.e., regurgitation) is minimum. Spatial Regulation In addition to the temporal regulation, exocytosis was also regulated spatially. The first exocytosis appears to be restricted to the place where a neutrophil first recognized the foreign object (Fig. 6). Furthermore, only the granules located at the periphery of a phagosome proceed to exocytosis. The direction of secretion is also restricted. The photon-counting imaging provides evidence that the secretions are simultaneous with release of active oxidants, and suggests that they are released into the phagosomal space (Fig. 8). These exocytotic responses are different from those induced by agonists such 226 Suzaki et al. as cytochalasin B [Zurier et al., 1973], formyl-methionylleucyl-phenylalanine (fMLP) [Lew et al., 1987], and leukotriene B4 (LTB4) [Showell et al., 1982; Lew et al., 1987] in the sense that secretions never took place on the plasma membrane but on the phagosomal membrane in spite of the identical property of the both membranes. When two or more zymosans were ingested sequentially at different parts of a neutrophil, exocytotic responses appeared in groups separated spatially and temporally. Thus, a highly organized regulatory mechanism is underlying the initiation of this class of exocytosis. Probably, immune receptors or some adhesion molecules on ingesting pseudopodia [Griffin et al., 1976] recognize the foreign object in an ordered manner, and play a key role in sending internal signals for exocytosis with a site-dependent proper delay time. Although there are many granules around a phagosome, the location of granules which undergo actual exocytosis is limited to the peri-phagosomal region closer to the cell body and the nucleus. These are also advantageous for the effective bactericidal action. Nature of Granule Fusion During phagocytosis, a series of exocytotic responses is limited only to the phagosomal membrane and no additional exocytotic response occurs at the other part of the cell membrane. Some of these responses occurring in the early stage of phagosome formation are indeed exocytosis as the phagosomal space is still open to the extracellular space and granule contents may be regurgitated to extracellular medium (Fig. 11). However, the same responses occurring in the later phase are not exocytosis but rather lysosomal fusion with the phagosome, because its membrane is completely closed within the cell. Since it is not clear whether or not the content of granules involved in these responses is really lysozyme, we propose to name these responses, simply based on morphological observation, as ‘‘encytosis’’ (from Greek en meaning inward). This change in the nature of the granule fusion, well-defined by the present technique, is very characteristic of the membrane fusion during phagocytosis in neutrophils. Signal for Exocytosis The time difference between the initiation of phagosome formation and of exocytosis was 25.0 s on the average. This value corresponds well with the time difference between the completion of phagosome formation and that of exocytosis which averaged out to 26.7 s. In addition, the localization where exocytosis occurs diffuses more distal from the site of the first exocytotic occurrence as the pseudopodial extension proceeds. These facts suggest that the formation of phagosome is closely related to the exocytosis occurrence, and that about 25 s are needed to cause the activations of key molecules for the membrane fusion after receiving the stimulus. These findings also suggest that the attachment to a zymosan itself is not an immediate cause of exocytosis, but chemical reactions are involved before directly triggering exocytosis in a localized area. As far as this final step of stimulation-secretion sequence in neutrophils is concerned, the temporal and spatial coordination of exocytotic events should be taken into consideration in order to understand their molecular mechanism. Degranulation in response to activation of receptors for chemoattractants such as fMLP and LTB4 has been shown to be a very sensitive [Ca21]i-dependent process in human neutrophils [Lew et al., 1987]. On the contrary, recent studies on a wide range of secretory cell types indicate that [Ca21]i-independent exocytosis also occurs. Especially in those related to phagosome-lysosome fusion, the fusion between granules and the plasma membrane is mediated by a multifactorial signaling system involving [Ca21]i increases, protein kinase C activation [Di Virgilio et al., 1984], and GTP binding to its regulatory proteins [Barrowman et al., 1986]. What is the signaling system for exocytosis on the phagosomal membrane? [Ca21]i elevation was shown to be a necessary and sensitive signal that triggers the phagosomegranule fusion event [Jaconi et al., 1990], and shown to be a localized response in the cytosol surrounding the phagosome during phagocytosis [Sawyer et al., 1985]. However, we found that [Ca21]i increases in the whole cytoplasm and the initiation of [Ca21]i increase corresponds well with the timing of the pseudopodial extension, which agrees well with the finding by Theler et al. . Our findings indicate that [Ca21]i elevation is closely related to the cytoskeletal activation for phagosome formation [Downey et al., 1990], but not to the machinery determining the timing and localization of the exocytosis. It is noteworthy here that phagosomelysosome fusion in macrophages is [Ca21]i-independent [Zimmerli et al., 1996]. Since [Ca21]i is not completely decreased to the basal level during the exocytosis, it may be necessary as an environmental factor to cause phagosome-granule fusion, but not as an immediate regulatory factor leading to exocytosis. There must be more closely related molecules for the phagosome-granule fusion event other than [Ca21]i. We infer that attachment of the membrane to foreign objects activates some membrane receptors, giving rise to a molecular signal immediately beneath the membrane. Opsonized zymosans bear IgG and C3b complements on their cell wall. Neutrophils possess membrane receptors for Fc portion of immunoglobulin and for C3b complements. During phagocytosis, therefore, a zymosan is gradually wrapped in a neutrophil by a zip-up mechanism based on binding of these ligands to receptors. It is highly possible that the membrane recep- Phagocytosis-Exocytosis Sequence in Neutrophils 227 Fig. 11. Phagocytosis-exocytosis sequence in the neutrophils. A neutrophil (red) is activated by attachment to a zymosan (green). It engulfs the zymosan by extending pseudopodia. Granules translocate near the zymosan (white arrows). The first exocytotic response (red granule) occurs at the bottom of phagosome when pseudopodia are still extending (black arrows) and the phagosomal space is open to the outside. This response as well as other few responses (yellow granules) may cause secretion to the environment by regurgitation (dashed red arrow). The following exocytotic responses occur in order of numbers indicated. Some granules (blue ones) fuse with the phagosomal membrane after complete closure of the phagosome. Thus, this response should be referred to as lysosomal fusion or encytosis (see text), although the molecular process must be the same as the one underlying the first exocytosis. 228 Suzaki et al. tors of neutrophils are thus activated sequentially, causing a sequential exocytosis along the phagosomal membrane. ACKNOWLEDGMENTS E.S. is a recipient of a JSPS Fellowship for Japanese Junior Scientists. This work was partly supported by Grant-in-Aid for Scientific Research from the Ministry of Education, Science and Culture, Japan. REFERENCES Allen, R.D., Allen, N.S., and Travis, J.L. (1981): Video-enhanced contrast, differential interference contrast (AVEC-DIC) microscopy: A new method capable of analyzing microtubule-related motility in the reticulopodial network of Allogromia laticollaris. Cell Motil. 1:291–302. Bainton, D.F. (1973): Sequential degranulation of the two types of polymorphonuclear leukocyte granules during phagocytosis of microorganisms. J. Cell Biol. 58:249–264. Barrowman, M.M., Cockcroft, S., and Gomperts, B.D. (1986): Two roles for guanine nucleotides in the stimulus-secretion sequence of neutrophils. Nature 319:504–507. DeChatelet, L.R., Long, G.D., Shirley, P.S., Bass, D.A., Thomas, M.J., Henderson, F.W., and Cohen, M.S. (1982): Mechanism of the luminol-dependent chemiluminescence of human neutrophils. J. Immunol. 129:1589– 1593. Di Virgilio, F., Lew, D.P., and Pozzan, T. (1984): Protein kinase C activation of physiological processes in human neutrophils at vanishingly small cytosolic Ca21 levels. Nature 310:691–693. Downey, G.P., Chan, C.K., Trudel, S., and Grinstein, S. (1990): Actin assembly in electropermeabilized neutrophils: Role of intracellular calcium. J. Cell Biol. 110:1975–1982. Foskett, J.K., and Melvin, J.E. (1989): Activation of salivary secretion: Coupling of cell volume and [Ca21]i in single cells. Science 244:1582–1585. Griffin, F.M., Griffin, J.A., Jr., and Silverstein, S.C. (1976): Studies on the mechanism of phagocytosis: II. the interaction of macrophages with anti-immunoglobulin IgG-coated bone marrowderived lymphocytes. J. Exp. Med. 144:788–809. Hirsch, J.G. (1962): Cinemicrophotographic observations on granule lysis in polymorphonuclear leucocytes during phagocytosis. J. Exp. Med. 116:827–834. Hirsch, J.G., and Cohn, Z.A. (1960): Degranulation of polymorphonuclear leucocytes following phagocytosis of microorganisms. J. Exp. Med. 112:1005–1114. Jaconi, M.E.E., Lew, D.P., Carpentier, J.-L., Magnusson, K.E., Sjögren, M., and Stendahl, O. (1990): Cytosolic free calcium elevation mediates the phagosome-lysosome fusion during phagocytosis in human neutrophils. J. Cell Biol. 110:1555–1564. Kamijo, A., Terakawa, S., and Hisamatsu, K. (1993): Neurotransmitterinduced exocytosis in goblet and acinar cells of rat nasal mucosa studied by video microscopy. Am. J. Physiol. 265:200–209. Kawasaki, Y., Saitoh, T., Okabe, T., Kumakura, K., and OharaImaizumi, M. (1991): Visualization of exocytotic secretory processes of mast cells by fluorescence techniques. Biochim. Biophys. Acta. 1067:71–80. Kolber, M.A., and Henkart, P.A. (1988): Quantitation of secretion by rat basophilic leukemia cells by measurements of quinacrine uptake. Biochim. Biophys. Acta. 939:459–466. Lew, D.P., Monod, A., Waldvogel, F.A., and Pozzan, T. (1987): Role of cytosolic free calcium and phospholipase C in leukotriene-B4stimulated secretion in human neutrophils: comparison with the chemotactic peptide formyl-methionyl-leucyl-phenylalanine. Eur. J. Biochem. 162:161–168. Nüsse, O., and Lindau, M. (1988): The dynamics of exocytosis in human neutrophils. J. Cell Biol. 107:2117–2123. Pryzwansky, K.B., MacRae, E.K., Spitznagel, J.K., and Cooney, M.H. (1979): Early degranulation of human neutrophils: Immunocytochemical studies of surface and intracellular phagocytic events. Cell. 18:1025–1033. Sakurai, T., and Terakawa, S. (1995): Insulin secretion from pancreatic b-cells directly visualized as exocytosis by video microscopy. Bioimages. 3:85–92. Sawyer, D.W., Sullivan, J.A., and Mandell, G.L. (1985): Intracellular free calcium localization in neutrophils during phagocytosis. Science 230:663–666. Segawa, A., Terakawa, S., Yamashina, S., and Hopkins, C.R. (1991): Exocytosis in living salivary glands: Direct visualization by video-enhanced microscopy and confocal laser microscopy. Eur. J. Cell Biol. 54:322–330. Showell, H.J., Naccache, P.H., Borgeat, P., Picard, S., Vallerand, P., Becker, E.L., and Sha’afi, R.I. (1982): Characterization of the secretory activity of leukotriene B4 toward rabbit neutrophils. J. Immunol. 128:811–816. Suzuki, K., Hosaka, S., Morikawa, K., Suzuki, M., Suzuki, S., Miyoshi, M., Fujita, M., and Mizuno, S. (1993): Bioimaging of myeloperoxidase release into phagolysosome of viable polymorphonuclear leukocytes and reaction of the enzyme with microspheres using a highly sensitive analyzer system, IMRAS. Bioimages 1:13–20. Terakawa, S., and Suzuki, Y. (1991): Exocytosis in colonic goblet cells visualized by video-enhanced light microscopy. Biochem. Biophys. Res. Commun. 176:466–472. Terakawa, S., Fan, J.-H., Kumakura, K., and Ohara-Imaizumi, M. (1991): Quantitative analysis of exocytosis directly visualized in living chromaffin cells. Neurosci. Lett. 123:82–86. Terakawa, S., Manivannan, S., and Kumakura, K. (1993): Evidence against the swelling hypothesis for initiation of exocytosis in terminals of chromaffin cell processes. J. Physiol. (Paris) 87:209–213. Terakawa, S., Hanajima, K., Kamijo, A., Kitayama, Y., Kawakami, K., Baba, S., Sakurai, T., and Suzuki, Y. (1995a): Exocytosis in goblet cells: A comparative study by video microscopy. Jpn. J. Physiol. 45, Suppl. 2, S41. Terakawa, S., Kumakura, K., and Duchen, M.R. (1995b): Spatiotemporal analysis of quantal secretory events from single bovine adrenal chromaffin cells in culture. J. Physiol. 487:59P–60P. Theler, J.-M., Lew, D.P., Jaconi, M.E., Krause, K.-H., Wollheim, C.B., and Schlegel, W. (1995): Intracellular pattern of cytosolic Ca21 changes during adhesion and multiple phagocytosis in human neutrophils. Dynamics of intracellular Ca21 stores. Blood 85:2194–2201. Wardlaw, A.C., and Pillemer, L. (1956): The properdin system and immunity: V. The bactericidal activity of the properdin system. J. Exp. Med. 103:553–575. Weiss, D.G., and Galfe, G. (1992): Video-microscopic techniques to study the living cytoplasm. In D.-P. Häder, (ed.): ‘‘Image Analysis in Biology.’’ Boca Raton: CRC Press, pp. 135–158. Yanai, M. (1979): Measurement of chemiluminescence of human polymorphonuclear leukocytes. Rinsho Ketsueki (Clinical Hematology). 20:372–380. (in Japanese) Zimmerli, S., Majeed, M., Gustavsson, M., Stendahl, O., and Sanan, D.A. (1996): Phagosome-lysosome fusion is a calciumindependent event in macrophages. J. Cell Biol. 132:49–61. Zucker-Franklin, D., and Hirsch, J.G. (1964): Electron microscope studies on the degranulation of rabbit peritoneal leukocytes during phagocytosis. J. Exp. Med. 120:569–576. Zurier, R.B., Hoffstein, S., and Weissmann, G. (1973): Cytochalasin B: Effect on lysosomal enzyme release from human leukocytes. Proc. Nat. Acad. Sci. USA. 70:844–848.