Loading...
Page 2
Loading...
Page 3
Loading...
Page 4
Loading...
Page 5
Loading...
Page 6
Loading...
Page 7
Loading...
Page 8
Loading...
Page 9
/lp/unpaywall/mechanisms-of-ph-regulation-in-the-regulated-secretory-pathway-HvptPRkO0a
References
References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.
- Publisher
- Unpaywall
- ISSN
- 0021-9258
- DOI
- 10.1074/jbc.m103917200
- Publisher site
- See Article on Publisher Site
Abstract
THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 276, No. 35, Issue of August 31, pp. 33027–33035, 2001 © 2001 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A. Received for publication, May 1, 2001, and in revised form, June 11, 2001 Published, JBC Papers in Press, June 11, 2001, DOI 10.1074/jbc.M103917200 Minnie M. Wu‡§, Michael Grabe¶, Stephen Adams**, Roger Y. Tsien**, Hsiao-Ping H. Moore‡, and Terry E. Machen‡ ‡‡ From the ‡Department of Molecular and Cell Biology, University of California, Berkeley, California 94720-3200, the ¶Department of Physics, University of California, Berkeley, California 94720-3112, and the **Department of Pharmacology and Howard Hughes Medical Institute, University of California, San Diego, La Jolla, California 92093-0647 A precise pH gradient between organelles of the reg- urements of organelle pH along the regulated secretory path- way. The “sorting for entry” model postulates that sorting ulated secretory pathway is required for sorting and processing of prohormones. We studied pH regulation in occurs when proteins encounter the ionic milieu of the trans- live endocrine cells by targeting biotin-based pH indica- Golgi network (3). In contrast, the “sorting by retention” model tors to cellular organelles expressing avidin-chimera asserts that aggregation serves to retain regulated proteins in proteins. In AtT-20 cells, we found that steady-state pH granules and does not occur until prohormones have entered decreased from the endoplasmic reticulum (ER) (pH ER acidic immature secretory granules (ISGs) and become proteo- 7.4 0.2, mean S.D.) to Golgi (pH 6.2 0.4) to lytically processed (5). mature secretory granules (MSGs) (pH 5.5 0.4). MSG Work by several groups (1, 6 –11) using a variety of tech- Golgi and MSGs required active H v-ATPases for acid- niques indicates that organelles of the secretory pathway, from ification. ER, Golgi, and MSG steady-state pH values ER to Golgi to secretory granules, become increasingly acidic. were also dependent upon the different H leak rates Indirect measurements of pH in isolated secretory granules of across each membrane. However, neither steady-state endocrine and neuroendocrine cells using either electron mi- pH nor rates of passive H leak were affected by MSG croscopy (measuring acidity based on accumulation of the weak Cl -free solutions or valinomycin, indicating that MSG base DAMP) or biochemical reactions (measuring acidity based membrane potential was small and not a determinant of on the extent of processing) (1, 6 –11) suggested that ISGs and pH . Therefore, our data do not support earlier sug- MSG MSGs were both acidic (pH 6.3–5.7; pH 5.5–5.0). In ISG MSG gestions that organelle acidification is primarily regu- live cell pH measurements using the green fluorescent protein lated by Cl conductances. Measurements of H leak derivative pHlorin (12), secretory granules of mast cells were rates, buffer capacities, and estimates of surface areas also acidic (pH 5.2). In “non-regulated” cells (Chinese hamster and volumes of these organelles were applied to a math- ovary, HeLa, HepG2, and Vero), pH experiments performed ematical model to determine the H permeability (P ) of each organelle membrane. We found that P using both DAMP on fixed cells (13, 14) and fluorescent probes de- creased progressively from ER to Golgi to MSGs, and in live, intact cells (15–21) showed that ER pH (pH 7.1– ER proper acidification of Golgi and MSGs required grad- 7.2) was similar to cytosolic pH (pH ), whereas Golgi pH was ual decreases in P and successive increases in the H acidic (pH 6.5– 6.2). In contrast, in “regulated” cells, exper- active H pump density. iments using DAMP detected no acidification of Golgi in pan- creatic islet cells (7, 8). It is unclear whether the conflicting Golgi pH data are due to differences in cell type or techniques Maintenance of lumenal pH in organelles of the secretory used. Therefore, although there is general agreement regard- pathway is required for proper sorting and proteolytic process- ing the acidity of MSGs, the exact pH of Golgi versus ISGs (and ing of prohormones. Even small pH differences between or- hence the site of sorting) in regulated secretory cells remains in ganelles can be critical in separating cellular events. For ex- question. ample, a difference of 0.5 pH can determine whether a Furthermore, although the dynamics of pH regulation in the prohormone is processed (1). In professional secretory cells (e.g. Golgi and ER of non-regulated cells has been extensively stud- endocrine and neuroendocrine), “regulated” secretory proteins ied, there has been no examination of the pH regulatory mech- have been hypothesized to be sorted from constitutively se- anisms of Golgi and MSGs in live, intact cells with regulated creted proteins by a process of pH- and calcium-dependent secretory pathways. pH studies of these organelles have been selective aggregation (2– 4). The exact site where aggregation limited to either fixed cells or isolated organelles in vitro. Data occurs is controversial, largely due to the lack of direct meas- from these studies suggested that Golgi and secretory granules controlled their acidic pH values by altering their conductances This is an open access article under the CC BY license. to Cl , which served as a counterion for the H v-ATPase (6, 9). * This work was supported in part by National Institutes of Health Experiments on isolated synaptic vesicles also supported a Grants DK51799 (to T. E. M.) and R24RR14891 (to H.-P. H. M.) and National Science Foundation Grant MCB-9983342 (to H.-P. H. M. and T. E. M.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby The abbreviations used are: ISGs, immature secretory granules; marked “advertisement” in accordance with 18 U.S.C. Section 1734 ER, endoplasmic reticulum; MSGs, mature secretory granules; solely to indicate this fact. v-ATPase, vacuolar ATPase; FCCP, carbonyl cyanide p-trifluoro- § Supported by National Institutes of Health training grants. methoxyphenylhydrazone; BCECF-AM, 2,7-bis-(2carboxyethyl-5- Supported by National Science Foundation Grant DMS9220719. (and-6)-carboxyfluorescein, acetoxymethylester; AV-POMC. avidin-pro- ‡‡ To whom correspondence should be addressed: 231 Life Sciences opiomelanocortin; DMEM, Dulbecco’s modified Eagle’s medium; ODE, Addition, University of California, Berkeley, CA 94720-3200. Tel.: 510- ordinary differential equations; DAMP, 3-(2,4-dinitroanilino)-3 amino- 642-2983; Fax: 510-643-6791; E-mail: machen@socrates.berkeley.edu. N-methyl dipropylamine. This paper is available on line at http://www.jbc.org 33027 33028 Organelle pH in the Regulated Secretory Pathway pellet was rinsed with cold DMEM, re-pelleted, and resuspended in 300 critical role for Cl in organelle acidification (22, 23). These l of cold DMEM. 100 g of AV-POMC or AV-KDEL DNA and 50 gof results led to the general hypothesis that acidic organelles a puromycin resistance gene-encoding plasmid (pSFPACEBv, Dr. Brian maintain distinct lumenal pH values by maintaining different Seed) were combined, phenol/chloroform-extracted, ethanol-precipi- permeabilities to Cl , the primary counterion (24). However, tated, and resuspended in 100 l of DMEM. To electroporate, the 300 l experiments in non-regulated cells demonstrated that although of cells were mixed with the 100 l of DNA, incubated for 5 min on ice, Cl and K did serve as counterions for H pumping, Cl and and transferred to a cold 0.4-cm gap cuvette. Cells were electroporated at 250 V, 0 ohm resistance, and 960 microfarads using a Bio-Rad Gene K conductances were large compared with the passive H Pulser. After electroporation, cells recovered for 10 min on ice before conductance, arguing against modulation of Cl and K con- being replated onto a 15-cm plate. 24 h post-transfection, 0.75 g/ml of ductances as a mechanism for trans-Golgi network or Golgi pH puromycin (Sigma) was added to the cell medium to select for positive regulation (18, 21). In addition, the pH-dependent processing of transfectants. After 48 h of puromycin treatment, cells were replated secretogranin II measured in vitro was not stimulated by in- onto laminin-coated coverslips and allowed to recover 24 h before im- creasing the outside Cl concentration (11). Thus, the exact aging experiments were performed. Butyrate Induction of AV-POMC Expression—Because the expres- mechanisms regulating acidification along the regulated secre- sion level of AV-POMC in AtT-20 stable cell lines was too low for tory pathway are unknown. fluorescence imaging experiments, we used a butyrate incubation pro- To begin to address the controversies and the gaps in our tocol, described previously (28), to boost expression of AV-POMC. Bu- understanding of pH regulation in organelles of the regulated tyrate prevents histone deacetylation, inducing expression from viral secretory pathway, we targeted pH indicators to the ER, Golgi, promoters. AV-POMC stably expressing AtT-20 cells were plated on and MSGs to study the pH regulatory mechanisms in live, laminin-coated coverslips; 24 h after plating, cells were incubated in 6 mM sodium butyrate in normal growth medium for 15 h. After butyrate intact endocrine cells. We chose the AtT-20 mouse pituitary cell induction, cells were rinsed once in a large volume of DMEM and chased line because it has a well characterized regulated secretory in normal growth medium for at least 34 h before loading with Flubi pathway. The data presented in this paper represent the first dyes. The butyrate incubation had no effect on endogenous POMC or study of pH regulation in MSGs of live, intact cells, and the AV-POMC processing, indicating that MSG pH was unaffected by this first systematic comparison of the pH regulatory mechanisms protocol (data not shown). between the major organelles of the regulated secretory path- Fluorescent Labeling of Cytosol, ER, Golgi, and MSGs—Cell cytosol was labeled using BCECF-AM and fluorescein isothiocyanate-dextran and K conductances way. We tested for the importance of Cl as described previously (21, 29). in determining pH and pH by eliminating Cl from the G MSG We used AV-KDEL-expressing AtT-20 cells to monitor ER pH and solutions and by using the K ionophore valinomycin. In addi- AV-POMC-expressing AtT-20 cells to monitor the pH of both Golgi and tion, since H leaks appear to be crucial determinants of or- MSGs. To label the avidin-containing lumens of ER, Golgi, and MSGs, ganelle pH (18, 21, 25), we measured H leak rates and buffer the cell-permeable Flubi-2 diacetate (Flubida-2, 2mM) was mixed 1:1 capacities of the ER, Golgi, and MSGs and applied the results with Pluronic F-127 (20% w/v in dry Me SO) and then diluted to the desired final concentration with DMEM containing 2% fetal calf se- to a mathematical model (26) to calculate H permeabilities for rum. AV-KDEL- or AV-POMC-expressing cells were rinsed with each organelle. From our experimental data and mathematical DMEM, loaded with 2– 4 M Flubida-2 dye for 4 – 6 h, and then chased modeling, we found that the acidification step between the ER with normal growth medium for 0 –2 h (0 h of chase for Golgi measure- and Golgi in AtT-20 cells was similar to that in non-regulated ments; 2 h of chase for MSG measurements) at 37 °C. cells (21, 25) in that it required both an increase in active H Fluorescence Ratio Imaging of Cytosolic, ER, Golgi, and MSG pH— pump density and a reduction of H permeability from the ER Cytosol, ER, Golgi, and MSG pH values were measured in separate experiments using digitally processed fluorescence ratio imaging. Ratio to the Golgi. Meanwhile, the acidification step between Golgi imaging measurements were performed at room temperature as de- and MSGs required a decrease in H permeability in Golgi scribed previously (21, 30). pH data collected from the bright Flubi- versus MSG membranes. stained cell body of AtT-20 cells represented both AV-POMC-containing Golgi and some AV-POMC-containing ISGs. MSG pH was measured by EXPERIMENTAL PROCEDURES collecting data from only the brightly labeled tips of cell processes of Materials—Salts, amiloride, FCCP, fluorescein isothiocyanate-dex- AtT-20 cells. tran, monensin, sodium butyrate, nigericin, puromycin, and valinomy- Perfusion and Calibration Solutions and pH Calibration—Ringer’s, cin were from Sigma; all other organic chemicals were from Aldrich; NH Cl Ringer’s, and Cl -free, and sodium-free Ringer’s solutions solvents were from Fisher; and restriction enzymes were from New were all prepared as described previously (21). Calibration solu- England Biolabs (Beverly, MA). Bafilomycin was from Calbiochem; tions were also prepared as described previously (21), except the mouse laminin was from Life Technologies, Inc.; BCECF-AM and Plu- solutions were titrated to the following pH values: 8.2, 7.0, 6.5, 6.0, 5.4, ronic F-127 were from Molecular Probes, Inc. (Eugene, OR). and 4.9. In situ calibrations were performed and the data fit to calibra- Construction of ER, Golgi, and MSG-targeted Avidin-Chimera Pro- tion curves, and the calibration curves were used to convert ratio values teins—The avidin-KDEL (AV-KDEL)-encoding plasmid was con- to pH values exactly as described (21). structed as described (21). An avidin-pro-opiomelanocortin (AV- Determining H Leak Rates, Buffer Capacities, and H Permeabili- POMC)-encoding plasmid was constructed by polymerase chain ties—Rates of H leak out of ER, Golgi, MSG, and plasma membranes reaction amplification of avidin (Dr. Markku Kulomaa, University of were calculated by fitting the data to the single exponential equation: kt Jyvaskyla, Finland) using primers (5-cgcgggaagcttgccaccatggtgcacg- y A(1 e ) using GraphPad InPlot (Kelvin Gee, Irvine, CA). Rate caacctcc-3 and 5-cgcgggggatcctccttctgtgtgcgcag-3), which allowed iso- constants (k) and half-times (t1 ) were determined from the curve fits. lation of avidin by HindIII and BamHI digestion. A signal sequence- Since H leak rates across organelle membranes are affected by lacking POMC was polymerase chain reaction-amplified from mouse buffer capacity () and surface area-to-volume (S/V) ratio, differences in POMC (gift from Dr. Edward Herbert) using primers 5-cgcggggatccct- or S/V between organelles could account for differences in H leak ggtgcctggagagcagc-3 and 5-cgcgggggcggccgctcactggcccttcttgtg-3 and rates. We accounted for and S/V ratio values of each organelle by isolated by digestion with BamHI and NotI. Avidin and POMC were calculating intrinsic H permeabilities, P , for ER, Golgi, and MSGs ligated into sIg-7-poly vector (Dr. Brian Seed, Massachusetts General using data from experiments such as those shown in Figs. 3C and 4 and Hospital) using HindIII, BamHI, and NotI sites (avidin replaced Ig). the model described below. In order to calculate H permeabilities, we Cell Culture and Transfection of AtT-20 Cells—AtT-20 cells were had to first measure of the cytosol, ER, Golgi, and MSGs of AtT-20 grown as described previously (1). Golgi and MSG pH experiments were cells. was calculated from the magnitudes of rapid increases in pH of performed on AtT-20 cells that were either transiently transfected (by each compartment during perfusion with solutions containing 20, 30, or electroporation, see below) or stably transfected (27) with AV-POMC 40 mM NH Cl or from rapid decreases in pH of compartments during DNA. ER pH experiments were performed on AtT-20 cells transiently perfusion with solutions containing 20 or 30 mM sodium acetate. Values transfected with AV-KDEL DNA by electroporation. of for ER, Golgi, and MSGs were determined using bafilomycin- 24 h before electroporation, AtT-20 cells were passaged to obtain a pretreated cells (500 nM, 2 h) so that the base-line pH values, prior to 50% confluent 15-cm dish. On the day of electroporation, cells were NH Cl or sodium acetate treatment, were similar for all the compart- trypsinized, rinsed with normal growth medium, and pelleted. The cell ments. For NH Cl experiments, using pK 9.0 for the NH 3 H 4 4 Organelle pH in the Regulated Secretory Pathway 33029 NH subsequent movement of counterions, a large membrane potential will reaction and assuming that NH equilibrates equally across all the 3 3 was calculated from the change in pH (extrapolated to membranes, build up across the organelle membrane and quickly limit further time 0) during the switch from Ringer’stoNH Cl according to (31) changes in pH. This physical insight suggests that organelles that Equation 1. undergo acidification must be permeable to some counterions. However, it is unclear which ions are permeable and how fast their movement is NH / pH (Eq. 1) compared with the movement of H . Schapiro and Grinstein (25) have convincingly shown that the Golgi is permeable to K ; they also con- For sodium acetate experiments, using pK 4.7 for the CH CO H 3 3 2 cluded that the fast dissipation of H gradients in the Golgi by prot- H CH CO reaction and assuming that CH CO H equilibrates 3 2 3 2 onophores implied that the counterion movement must be much faster equally across all the membranes, was calculated from the change in than the endogenous H leak. As discussed under “Results,” our pres- pH (extrapolated to time 0) during the switch from Ringer’s to sodium ent data support these ideas. Therefore, in predicting P we assumed acetate according to Equation 3, that both K and Cl were free to diffuse across the organelle mem- branes. We chose permeability coefficients for these ions of 10 cm/s CH CO / pH (Eq. 2) based upon plasma membrane measurements (35). Our numeric results After determining key physical characteristics for each organelle, such remain unaffected for counterion permeability values down to 10 as , a mathematical model for the movement of H across the organelle cm/s. Below this point, counterion movement becomes the rate-limiting was required in order to predict the H permeability of the membrane. step during alkalinization, which affects predicted P values. We employed a model that has been shown to be in quantitative agree- Equation 3 and the two additional equations for the passive flux of ment with a diverse range of organelle pH and membrane potential K and Cl combine to form a set of ordinary differential equations data (26). Our present modeling effort is a subset of this more complete (ODE). These ODEs are coupled by the algebraic constraint of Equation model in that we ignored H pumping by the H v-ATPase, since all the 5 and uniquely determine the time course of changes of ionic concen- permeability measurements were performed in the presence of bafilo- trations in the organelle(s) given an initial set of conditions (e.g. initial mycin. Given this, the measured rate of change of lumenal pH is related lumenal concentrations). The model was applied to data recorded using to the passive flux of H and the physical characteristics of the or- two distinct experimental protocols for inducing organelle alkaliniza- ganelle as shown in Equation 3, tion as follows: (i) as shown in Fig. 4, pH recovery after an NH acid-load of cytosol and organelles in the presence of bafilomycin and (ii) dpH 1 S as shown in Fig. 3C, bafilomycin-induced alkalinization. In both cases, J (Eq. 3) leak dt V cytosolic K and Cl were assumed to remain constant and to be equal to typical cytosolic values ([K ] 140 mM; [Cl ] 20 mM). Under where J is the total passive flux of H across a unit area of mem- leak Cl -free conditions, cytosolic and organelle [Cl ] were set to 0. pH brane, and is the organelle buffer capacity. We used surface areas, S, values were experimentally measured. Before alkalinization began, in and volumes, V, of ER and Golgi as determined from terminal tubule both protocols, all compartments reached steady-state pH, indicating and acinar cells of the rat submandibular gland (32); S and V of MSGs that net Cl and K fluxes were initially at equilibrium (i.e. 0), since were based on a 200-m diameter sphere (33). Although our H perme- Cl and K movements were fast. Additionally, the H concentration ability calculations were determined for AtT-20 (mouse anterior pitui- was initially at equilibrium in the acid-load experiments but not in the tary) cells, we used the S and V values from Taga et al. (32) because this bafilomycin-induced alkalinization experiments due to the presence of was the only published report where S and V for both ER and Golgi active H v-ATPases. For a given value of B, the initial lumenal con- membranes were measured in the same cell type. centrations of ions at equilibrium were uniquely determined from the We modeled the H leak, J , as simple passive diffusion. With this leak passive flux equations and Equation 5. With bafilomycin-induced alka- assumption, the leak depends on the membrane potential and the linization, the initial lumenal pH was a free parameter independent of concentration gradient as shown in Equation 4, the value of B. In summary, our simulations had two free parameters zU z U C C e (P , B). The initial organelle pH (pH ) in the bafilomycin-induced H o H O H C J P (Eq. 4) leak H zU alkalinization experiments was chosen to be the average steady-state 1 e pH of the organelle before the addition of bafilomycin. All other param- where P is the H permeability of the membrane; C is the concen- eters were recorded experimentally or determined from equilibrium H H tration of H in the organelle (O) or cytosol (C); z is the valence of the conditions. As mentioned above, the value B was constrained by the H ; and U F/(RT), where is the organelle membrane potential. data since it was the only free parameter in the model that accounted We computed in terms of the excess charge inside the organelle for equilibrium differences between pH and pH . c o membrane, which was treated as a parallel plate capacitor (Equation 5), The free parameters, P and B, were determined by using the model to fit the experimentally measured changes of pH (Figs. 5 and 6, pHG F V data points). For every data set, the model ODEs were solved to find K Cl dpH B (Eq. 5) O O 0 which values of P and B gave the best fit to the experimental data. pH Example model fits, from which P values were determined, are shown in Figs. 5 and 6 (solid curves). The search for the best model fit where C is the total capacitance of the membrane (calculated assum- 2 was performed with a Nelder-Mead algorithm, and the ODEs were ing capacitance 1 microfarad/cm ); [K ] and [Cl ] are the concen- O O solved with a stiff method in both Matlab (Mathworks, Natick, MA) and trations of K and Cl in the organelle (based on previous experiments, Berkeley Madonna (George Oster and Robert Macey, University of we assumed that K and Cl diffuse across organelle membranes California, Berkeley). according to equations of the form (4)); the integral term represents the In experiments in which P was calculated from rates of bafilomy- total H in the organelle lumen (both buffered and free), and B (a cin-induced alkalinization (e.g. Figs. 3C and 6), we assumed pH was constant) is the concentration of charged species trapped in the or- c constant, since bafilomycin had no effect on pH (data not shown). In ganelle. In the absence of any membrane-energizing enzymes, one c experiments in which P was calculated from rates of recovery of pH expects H to diffuse across the organelle membrane until the lumenal H o and pH following an acid-load, the pH recovery was directly depend- pH and the cytosolic pH are equal. However, this is often not the case c o ent upon the recovery of pH (see Figs. 4 and 5). To determine P using for organelles that have been treated with bafilomycin (16, 21); further- c H these experimental data, knowledge of the instantaneous H gradient more, at rest, the distribution of H across a membrane can be manip- across the organelle membrane was required. Because our experimen- ulated by the distribution of other ionic species (34). This discrepancy tal system did not permit us to measure pH and pH in the same cells can be attributed to trapped, negatively charged species inside the c o simultaneously, we undertook the computationally intensive task of organelle that, in the absence of an active H v-ATPase, lead to an fitting all pH recovery data against all separately recorded pH recov- accumulation of H . Our model accounts for the difference in steady- o c ery data. For calculations of P in Cl -free solutions, Golgi and MSG state pH between organelle lumen and cytosol only through changes in data sets were fit against corresponding measurements of pH in Cl - the parameter B. Average values of B varied between 50 and 200 mM. free solutions. In each of the data runs (e.g. 782 runs in the case of the Average B values determined from the model for ER, Golgi, and MSGs did not significantly contribute to the different steady-state pH values ER data; 17 pH data sets 46 pH data sets), a fitness parameter was ER c of these compartments. used in order to determine the likelihood that the pH data were a From our buffer capacity measurements, we know that a significant result of a particular pH data set. A root mean square fit of the model number of H are transported into/out of the organelle during acidifi- to the pH data was used as a measure of fitness, and its inverse, w, was cation/alkalinization. If this change in charge is not off-set by the used as a weight for computing averages and S.E. (Equation 6), 33030 Organelle pH in the Regulated Secretory Pathway FIG.2. Average ratio versus pH calibration curve for Flubi-2 in AtT-20 secretory granules. Calibration solutions containing nigeri- cin and monensin (10 M each) and titrated to different pH values (8.2, 7.0, 6.0, 5.5, 5.0, and 4.9) were used to generate calibration curves that were used to convert 490 nm/440 nm excitation ratio values (emission 520 –560 nm) to pH values. The solid line shows the average pH MSG S.D. obtained for 24 cells. Average calibration curve error bars are from S.E. values. Flubi-2 for Measuring pH in Secretory Granules, Golgi, and ER of Intact AtT-20 Cells—The pH dependence of the 490/440 nm excitation ratio of Flubi-2 in situ within secretory granules of living AtT-20 cells is shown in Fig. 2. Secretory granules FIG.1. Avidin-POMC follows the regulated secretory path- were labeled with Flubi-2, and calibration solutions containing way in AtT-20 cells. A, an AV-POMC chimera protein was con- nigericin and monensin were perfused onto cells at the end of structed by fusing amino acids 1–132 of chicken avidin (signal peptide each experiment. An average calibration curve was generated intact, stop codon removed) with amino acids 27–235 of mouse POMC by plotting 490/440 nm fluorescence ratio versus pH of the (signal peptide removed). sp, signal peptide. AtT-20 cells expressing AV-POMC were loaded with 3 M Flubi-2 (4.5 h) and chased for either external solution. pH was determined by averaging the MSG 1.5 (B)or3h(C) before viewing. After the longer chase (C), Flubi-2 steady-state pH of MSGs of cells bathed in pH 7.4 Ringer’s was chased out of the Golgi (arrow in B) and into ISGs and MSGs solution. pH (mean S.D.) was 5.5 0.4 (Fig. 2, solid line). MSG (punctate cell body and tip staining). Arrow, cell body staining (Golgi We used similar methods to determine ER and Golgi pH and some ISGs); arrowheads, tips of cell processes where MSGs accumulate. Scale bars,10 m. values in AtT-20 cells by averaging the steady-state pH values of each compartment of cells bathed in pH 7.4 Ringer’s solution. pH of AtT-20 cells (7.4 0.2) was similar to pH in HeLa 1 1 ER ER i i 2 R pH pH or w (Eq. 6) observed mode cells (7.2 0.2) (21). In AtT-20 cells, as in HeLa cells, pH was N R ER i1 slightly lower than cytosolic pH (AtT-20 pH 7.6 0.2; HeLa pH 7.4 0.2). In AtT-20 cells, pH was 6.2 0.4, similar to where the sum was over all N experimentally measured time points. c G Chauvenet’s criteria were used to exclude three highly improbable values measured for the Golgi in non-regulated cells (15, 16, predicted permeability values (40). Two sets were rejected because the 21). Thus, the pH of the Golgi compartments in live, intact organelle pH was more alkaline than any cytosolic pH values, and one endocrine cells appears acidic and similar to the Golgi pH in additional set was rejected due to technical difficulties in finding a non-regulated cells. Although the Flubi-2 staining in the cell reliable best fit. body (Fig. 1B) may consist of signals from both Golgi and ISGs, RESULTS AND DISCUSSION the majority of these signals come from the Golgi compart- ments and not the ISGs. This latter conclusion is based on the Targeting Avidin-Chimera Proteins and Flubi Dyes to the ER, Golgi, and MSGs of AtT-20 Cells—ER pH measurements observation that brefeldin A, which causes Golgi but not trans- Golgi network and post-Golgi organelles to redistribute to the in AtT-20 cells were performed exactly as in HeLa cells (21), where Flubi-2 was loaded into cells expressing avidin-KDEL ER (36), causes complete dispersal of the perinuclear POMC (AV-KDEL). For AtT-20 Golgi and MSG pH measurements, an staining to the ER (data not shown). Future experiments using avidin-POMC (AV-POMC) chimera protein was expressed, and avidin constructs targeted specifically to individual cisternae of cells were loaded with Flubi dye to monitor pH. In AtT-20 cells, the Golgi complex and the ISGs will be necessary to resolve the POMC is a regulated secretory protein that is transported from likely small pH differences between these compartments. the ER to the Golgi and packaged into ISGs that mature into Are Cl or K Conductances Determinants of Steady-state MSGs. We constructed an AV-POMC chimera (Fig. 1A) that pH or the H Leak Out of MSGs?—Early models of pH MSG exhibited similar secretion patterns as endogenous POMC in regulation in secretory granules suggested that the assumed AtT-20 cells (data not shown). Fig. 1B shows that live AtT-20 inside positive organelle membrane potential limited H accu- cells loaded with Flubi-2 (4.5 h) and chased briefly (2h) mulation by the H v-ATPase and that a Cl conductance was showed both cell body staining (Golgi and some ISGs) and required to shunt this potential to allow adequate acidification staining at the tips of cell processes (MSGs). Longer chase of organelles (9). We examined the potential role of Cl con- times (3 h) resulted in the disappearance of Flubi-2 from the ductance in regulating pH in AtT-20 cells by replacing all MSG Golgi, with increased punctate staining throughout the cell Cl with gluconate in the extracellular solution of intact Flubi- body, particularly concentrated at cell tips, indicating that 2-loaded, AV-POMC-expressing AtT-20 cells. We assumed that Flubi-2 had been chased out of the Golgi into MSGs (Fig. 1C). cytosolic Cl would be depleted after a long (20 min) incuba- Since MSGs are spatially separated from the Golgi, we were tion in Cl -free solution, since for plasma membrane Cl per- 5 6 able to use AV-POMC-expressing AtT-20 cells to monitor the meabilities between 10 and 10 cm/s (35), the half-time for pH of both proximal (Golgi) and distal (MSGs) compartments of cytosolic Cl depletion from a 1000-m cell is 100 –200 s. This the regulated pathway by restricting data collection to either calculation follows from Equation 4 (see “Experimental Proce- the cell body or cell tips. dures”) but neglects the inside negative membrane potential Organelle pH in the Regulated Secretory Pathway 33031 determining the acidity in MSGs and Golgi using a different approach. As shown in Fig. 3C, cells were treated first with 500 nM bafilomycin (H v-ATPase inhibitor) which caused pH MSG to alkalinize. This result showed that pH was maintained MSG by constantly active, bafilomycin-sensitive, H v-ATPases op- posing H leaks. The rate of H leak out of the acidic MSGs into the pH 7.4 cytosol is a function of the H permeability, transmembrane pH gradient, and membrane potential (which, in the presence of bafilomycin, is a function of the conductances to the major ionic constituents of the cytosol, K , and Cl ). If the K conductance were limiting the rate of H leak through effects on membrane potential, then addition of valinomycin (K ionophore) should increase both the K conductance and also the rate of H leak out of MSGs. As shown in Fig. 3C, addition of 10 M valinomycin did not affect the rate of bafilo- mycin-induced MSG alkalinization; the rate of alkalinization increased only upon treatment with the protonophore FCCP (20 M). This result indicated that in the presence of valinomy- cin, the H conductance was limiting the rate of H leak out of the MSGs. Similar experiments were performed on cells bathed in Cl -free solutions for 30 min to reduce cell [Cl ] (Fig. 3D). By incubating the cells in Cl -free solution, we expected that the major counterion conductance (for H ) across MSG mem- branes would now be due solely to K . If the K conductance were smaller than the H conductance, then during bafilomy- cin treatment H would leak out of the MSGs only as fast as K leaked into the MSGs, and the rate of alkalinization would be increased by valinomycin. However, valinomycin had no effect on the rate of H leak out of MSGs in the absence of Cl (Fig. 3D). Again, the H leak out of the MSGs increased only when the cells were treated with FCCP. The simplest conclusion from these experiments was that the conductances to both Cl and K were larger than the H conductance for both MSGs and Golgi. Because the conduct- ance to K was large, removing Cl (eliminating Cl as a counterion) had no effect on either the ability of MSGs to acidify their lumens in control cells or on the rates of passive FIG.3. AtT-20 MSG steady-state pH and the H leak out are not H flux out of the lumen in bafilomycin-treated cells. The K regulated by Cl or K counterion conductances. Removal of Cl (0 Cl Ringer’s, replaced with gluconate) from the extracellular solution conductance appeared to be larger than the H conductance had no effect on pH or pH in intact cells, as shown by a typical G MSG because valinomycin did not affect H flux in Cl -free solutions experiment (A) and by quantitating the results from numerous exper- (when the rate of H exit from the MSGs is determined by the iments (B). C, treatment with bafilomycin (Baf)(H v-ATPase inhibitor, 500 nM) caused pH to alkalinize. Addition of the K ionophore, H permeability and the rate of K entry). Thus, it appeared MSG valinomycin (Val) (10 M), did not affect the rate of alkalinization; that the Golgi compartments in both regulated AtT-20 cells and however, FCCP (F) (protonophore, 20 M) caused rapid alkalinization of non-regulated HeLa cells (21, 25) had similar pH values and pH . F, FCCP; R, Ringer’s. Similar experiments in Cl -free solutions MSG were also highly conductive to both K and Cl . gave similar results (D). The experiments shown in C and D were Our conclusion that MSGs did not require Cl to generate repeated three times with the same results. normal lumenal acidity contradicts previous conclusions (9, 22). However, it should be noted that these previous experi- that is likely to exist in AtT-20 cells. In all likelihood, the ments were non-quantitative pH measurements using acridine membrane potential and also the presence of neutral transport- orange in isolated synaptic vesicles in vitro; furthermore, the ers (e.g. NaKCl cotransport) will cause the half-time for Cl dose-dependent requirement for Cl in acidification of the se- depletion to be shorter than 100 –200 s, ensuring that cell [Cl] cretory vesicles was accomplished by changing the concentra- will probably be lower than 1 mM after about 10 min. We tions of both K and Cl , and the role of Cl alone or K alone expected that if Cl were required to provide a counterion for was not determined. The contradictory Cl results may also the accumulation of lumenal H , then incubation in Cl -free have been due to different pH regulatory mechanisms present solution for 30 min would reduce the acidity of MSGs. Similar in organelles of the endocytic pathway versus organelles of the to previous experiments on the Golgi of Chinese hamster ovary secretory pathway. Early studies of pH regulation in the endo- and HeLa cells (18, 21, 25), treatment with Cl -free solution cytic pathway indicated that endosome acidification could be did not affect steady-state pH over the course of 30 min MSG regulated by altering the membrane potential by changing the (Fig. 3A), indicating that Cl was not required to maintain Cl conductance or the Na/K-ATPase activity (37, 38). Perhaps acidic MSGs. A summary of results from experiments such as the one in Fig. 3A is shown in Fig. 3B.InCl the conflicting Cl data reflect two sets of organelles with -containing distinct pH regulatory mechanisms as follows: (i) organelles of solutions pH (mean S.D.) was 5.5 0.4, nearly the same MSG the endocytic recycling pathway (including synaptic vesicles), as pH in Cl -free solutions, 5.6 0.4; pH was 6.3 0.3 in MSG G Cl -free solutions, insignificantly different from pH obtained where membrane potential and thus Cl are important regu- lators of pH, and (ii) organelles of the biosynthetic secretory in Cl -containing solutions, 6.2 0.4. We also tested for the role of K and Cl conductances in pathway (ER, Golgi, secretory granules), where organelle pH is 33032 Organelle pH in the Regulated Secretory Pathway TABLE I Half-times for pH recoveries of the cytosol, ER, Golgi, and MSGs of AtT-20 cells The data for the pH recoveries of the cytosol, ER, Golgi, and MSG from experiments such as those shown in Fig. 4B were fit to the single kt exponential equation: y A(1 e ). From each curve fit, a rate constant (k) and t value, representing the half-time for the pH re- 1/2 covery of the cytosol or organelle, were obtained. The t values are 1/2 presented as mean S.E. AtT-20 compartment t n 1/2 (s) Cytosol 40 230 ER 59 517 Golgi 283 87 11 MSG 734 184 10 indicating a large ER H leak that quickly equilibrated pH ER and pH (cytosol t1 40 s). This result is very similar to results c ⁄2 obtained on HeLa cells (21). In AtT-20 cells, just as in HeLa FIG.4. pH recoveries of AtT-20 cell compartments after an acid cells, more distal organelles had slower pH recoveries com- load. The pH recoveries of cytosol, ER, Golgi, and MSGs were measured pared with ER and cytosol. The pH recovery out of the Golgi (t1 ⁄2 in either BCECF-AM-loaded or Flubi-2-loaded, bafilomycin-pretreated, 283 s) was almost 5 times slower than the ER; the pH intact cells by acid-loading the cytosol (Cyto) and organelles with an recovery of MSGs (t1 734 s) was 2.5 times slower than that ⁄2 NH /NH pulse, followed by a washout with sodium-free Ringer’s(0Na 3 4 R) before allowing the pH values to recover in sodium-containing Ring- of the Golgi and 12 times slower than that of the ER. er’s(Ringer’s). This protocol is illustrated in A with typical results for These data suggested that organelles in the regulated secre- cytosol (black) and MSG (green) pH. B, only the pH recovery traces for tory pathway exhibited decreasing H leak rates, which may each organelle are shown. be a major contributor to the decreasing steady-state pH values of these organelles. However, the decreasing H leak rates we not regulated by Cl or K conductances. measured could also have been the result of increasing buffer Measuring H Leak Rates and Intrinsic H Permeabilities capacities of organelles along the secretory pathway, or these across ER, Golgi, and MSG Membranes in AtT-20 Cells—The results could have been due to differences in the surface area- data presented above were consistent with the hypothesis that to-volume (S/V) ratios of secretory pathway organelles. There- the MSG membrane was likely to have large conductances to fore, to account for buffer capacities and S/V ratios, we deter- both K and Cl but lower conductance/permeability to H . mined the intrinsic H permeability, P , for each organelle Under these conditions, pH was limited by the magnitude membrane (ER, Golgi, and MSG). MSG of the H “leak” or “permeability” pathway. Based on our Buffer capacities (mean S.E. mM/pH) were determined (as previous work on ER and Golgi pH regulation in HeLa cells (21) described under “Experimental Procedures”) for the cytosol and the finite H permeability of phospholipid bilayers (10 (23 3, n 38), ER (17 3, n 21), Golgi (26 6, n 10), cm/s, (39)), we hypothesized that the pH differences between and MSG (20 6, n 10) compartments. By using these mean the ER, Golgi, and secretory granules in AtT-20 cells might be buffer capacity values and previously estimated surface areas generated by gradually decreasing the H leak or permeability and volumes (32, 33), the data from acid-load recovery experi- along the secretory pathway. To gain insights into H leak ments (Fig. 4B) were fit (Fig. 5) using the mathematical or- rates across AtT-20 ER, Golgi, and MSG membranes, we fol- ganelle pH model (described under “Experimental Procedures”) lowed the same acid-loading protocol we used to measure rates to determine P for the ER, Golgi, and MSG membranes. Fig. of pH recoveries for organelles of HeLa cells (21). AtT-20 pH 5 shows typical organelle pH recovery data (colored data measurements were performed on cells loaded with the well points), the corresponding best pH recovery (black), and the known cytosolic pH dye, BCECF-AM (2 M loaded 1 h, 1-h model curve fits (colored solid curves, determined using Equa- chase). AtT-20 cells expressing either AV-KDEL or AV-POMC tions 3–5). P values are listed in Table II, which shows that were loaded with Flubi-2 (at least 4 h), chased (0 –2 h), and P decreased progressively from the ER (P 51 10 H H treated with 500 nM bafilomycin (at least 2 h) before measuring cm/s) to Golgi (P 21 10 cm/s) to MSGs (P 3 H H 4 4 pH ,pH ,pH , and pH . AtT-20 cells were acid-loaded by 10 cm/s). The P that we estimated for ER (51 10 cm/s) c ER G MSG H incubating in 40 mM NH Cl for 20 min. When the NH Cl should be considered a lower estimate since these measure- 4 4 solution was replaced with sodium-free Ringer’s, pH ,pH , ments may have been limited by the rates of pH recovery, c ER c pH , and pH remained acidic until sodium-containing which were controlled by plasma membrane pH regulatory G MSG Ringer’s was returned to the chamber, at which point pH mechanisms. The P value for each organelle was statistically c H recovered rapidly to pH 7.4, and the ER, Golgi, and MSGs different (p 0.07) from the other organelles. leaked out H to recover their pH values. This protocol is In addition to using pH recovery data from acid-loaded cells illustrated in Fig. 4A, which shows typical pH traces for cytosol and organelles, we also estimated P for Golgi and MSG (black) and MSG (green) compartments. Fig. 4B shows the membranes using H leak data obtained from bafilomycin- representative pH recoveries for each compartment, cytosol treated cells. By using H leak data from experiments such as (black), ER (red), Golgi (blue), and MSGs (green), as the outside the one shown in Fig. 3C, we fit the bafilomycin-induced alka- solution was switched from sodium-free to sodium-containing linization rate using the mathematical organelle pH model as Ringer’s. shown in Fig. 6. The main difference between the bafilomycin- The data for the pH recoveries of AtT-20 cytosol, ER, Golgi, induced alkalinization experiments and the acid-load pH re- and MSGs were fit to single exponential equations to obtain covery experiments was that using the former protocol pH half-times (t1 ) for the pH recovery of each compartment. The remained constant (data not shown) since bafilomycin had no ⁄2 results from these calculations have been summarized in Table effect on pH , whereas using the latter protocol pH acidified c c I. ER membranes had the fastest pH recovery (t1 59 s), and changed as pH changed (e.g. Fig. 4). For both protocols, ⁄2 o Organelle pH in the Regulated Secretory Pathway 33033 FIG.6. Fitting bafilomycin-induced alkalinization data to a mathematical pH model to determine H permeability. Repre- sentative model fits (curves) of Golgi and MSG bafilomycin-induced H leak data (data points) are shown. Bafilomycin had no effect on pH ; thus we assumed a constant pH (7.58 0.19, n 187 cells) for c c the model fits. For the model fits shown, P (G) 1.6 10 cm/s, FIG.5. Fitting pH recovery data to a mathematical pH model P (MSG) 0.3 10 cm/s. Values of P were determined from H H to determine H permeability. Representative model fits (red, blue, the best model fit to the experimental data and are summarized in and green curves) of ER, Golgi, and MSG acid-load pH recovery data Table III. (red, blue, and green data points) are shown with their corresponding cytosolic runs (black) that gave rise to these fits. For the model fits 4 4 shown, P (ER) 44.3 10 cm/s, P (G) 12 10 cm/s, and H H P (MSG) 2.8 10 cm/s. Values of P were determined from the H H best model fit to the experimental data and are summarized in Table II. TABLE II H -permeabilities determined from rates of pH recoveries of the ER, Golgi, and MSGs of AtT-20 cells Data from pH recoveries (Figs. 4B and 5), measured buffer capacities (), and previously determined surface area-to-volume ratios (32, 33) were incorporated to determine the intrinsic H permeabilities (P )of AtT-20 ER, Golgi, and MSG membranes using the Berkeley Madonna (G. Oster and R. Macey, University of California, Berkeley) modeling program. and P values are presented as mean S.E. P values for H H each organelle were compared using the Welch test: ER versus Golgi (p 0.07), Golgi versus MSG (p 0.07), and ER versus MSG (p 0.02). AtT-20 Surface (m )/ P 3 4 membrane volume (m ) ( 10 cm/s) FIG.7. A model for pH regulation in the regulated secretory pathway. pH (mean S.D.) in AtT-20 cells was 7.4 0.2 (n 13); mM/pH ER pH was 6.2 0.4 (n 27), and pH was 5.5 0.4 (n 24). pH and ER (n 16) 17 3 1110/35.4 51 11 G MSG G pH were maintained by constantly active, bafilomycin-sensitive, H MSG Golgi (n 10) 26 6 514/26.4 21 9 v-ATPases that opposed H leaks. MSG membranes had large conduc- MSG (n 10) 20 6 0.126/0.00419 3 1 tances to both Cl and K compared with H . Our data indicate that, to generate the stepwise acidification from ER to Golgi to MSGs, the density of active H pumps must progressively increase, while the H TABLE III permeability gradually decreases. H permeabilities determined from rates of bafilomycin-induced alkalinization of the ER, Golgi, and MSGs of AtT-20 cells the pH gradient between the cytosol and organelle lumen was Data from bafilomycin-induced alkalinization rates measured in Cl - taken into account in the modeling (see Equation 4 under containing (e.g. Fig. 3C) and Cl -free (e.g. Fig. 3D) solutions, together “Experimental Procedures”). By using the bafilomycin-induced with measured buffer capacities and previously determined surface alkalinization protocol, we calculated P values of Golgi and area-to-volume ratios (see Table II), were incorporated to determine the H intrinsic H permeabilities (P ) of AtT-20 Golgi and MSG membranes MSG membranes to be lower than those calculated using acid- using the Berkeley Madonna modeling program. P values are pre- H load pH recovery data, but the P for MSGs was consistently sented as mean S.E. Golgi P values determined in the presence and lower than the P of Golgi membranes. These data are sum- absence of Cl were statistically the same according to the Student’s t marized in the first two columns of Table III. test. MSG P values determined in the presence and absence of Cl The data from Fig. 3 indicated that Cl was not required to were also statistically the same according to the Student’s t test. maintain an acidic steady-state pH . We further tested for MSG AtT-20 membrane P Cl -free P H H the potential role of Cl conductance in controlling Golgi and ( 10 cm/s) MSG membrane potential by determining the H permeability Golgi 1.3 0.4 (n 5) 1.6 0.6 (n 6) in Cl -free solutions. In the presence of bafilomycin, the H MSG 0.35 0.1 (n 9) 0.6 0.35 (n 7) leak across MSG membranes will be a function of membrane 33034 Organelle pH in the Regulated Secretory Pathway potential and pH gradient, and if Cl conductance were impor- increase in the active H pump density (assuming the H pump activity of each H tant for shunting membrane potential, then we would expect v-ATPase is constant) or a 10-fold that the calculated P would be smaller for experiments in decrease in P between the ER and Golgi. Generating the 0.7 H H Cl -free versus Cl -containing solutions. To determine unit pH drop between Golgi (pH 6.2) and MSGs (pH 5.5) re- quired either a 5-fold increase in active H pump density or a whether P of MSGs was Cl -dependent, we used data from Cl -free bafilomycin-induced alkalinization experiments (e.g. 5-fold decrease in P between Golgi and MSGs. Since our calculated P values (Table II) for the ER and Fig. 3D and Fig. 6) to determine the P of MSG membranes in Golgi differed by 2-fold rather than 10-fold, an increase in H the absence of Cl (Table III). The P in Cl -free solutions 4 4 pump activity must have accompanied the 2-fold decrease in was 1.6 10 cm/s for Golgi membranes and 0.6 10 cm/s to generate the lower pH of the Golgi compared with the for MSG membranes, in both cases greater than the H per- H ER. These calculations, together with previous data showing meabilities measured in Cl -containing solutions (Golgi 4 4 that bafilomycin treatment had no effect on steady-state pH 1.3 10 cm/s MSG 0.35 10 cm/s), indicating that Cl ER of Vero and HeLa cells (17, 21), indicate that for AtT-20 cells, was not required for the H leak out of either the Golgi or the Golgi had a higher density of active H v-ATPases com- MSGs in AtT-20 cells. pared with the ER. Based on the acid-load pH recovery data Summary, a Model for pH Regulation in the Regulated Se- (Table II), our estimated P values for Golgi and MSGs dif- cretory Pathway—This study constitutes the first investigation fered by 7-fold. When we used the rate of alkalinization due to into the dynamics of pH regulation in organelles of the secre- bafilomycin treatment (Table III) to calculate P , P for H H tory pathway of live, intact endocrine cells, where the organelle Golgi and MSGs differed by only 4-fold. It is unclear why the acidification process is crucial for the sorting and processing of different protocols (pH recovery after an acid-load versus bafilo- regulated secretory hormones. We determined that, just as for mycin-induced alkalinization) for measuring the H leak out of the ER and Golgi of HeLa cells, the steady-state pH values of organelles produced different H permeability values for Golgi AtT-20 ER, Golgi, and MSGs appeared to be controlled primar- and MSGs. The main difference between the two protocols was ily by rates of H v-ATPase pumping and by the magnitude of the effect on pH . In the bafilomycin-induced H leak protocol, H leaks. Our data and mathematical modeling showed that pH was 7.4 and remained constant throughout the protocol. the membrane potential in Golgi and MSGs of AtT-20 cells was In the NH /NH acid-load protocol, pH acidified to 6.5 and 3 4 c small and not perturbed by large changes in Cl and K then alkalinized throughout most of the protocol. In the acid- conductances; these results indicated that membrane potential load protocol, pH recovery was directly dependent upon the was not a determinant of steady-state Golgi and MSG pH recovery of pH . The different P values measured using the c H values. different protocols could be explained by a pH-dependent H We previously found in HeLa cells that the rate of H leak leak across the organelle membrane which would increase as out of the Golgi was three times slower than the H leak out of pH decreased. In the acid-load protocol, the acidic initial pH c c the ER (21). In AtT-20 cells, the rates of H leak steadily would result in a much faster P than in the bafilomycin decreased from ER to Golgi to MSGs, with MSGs having a 12 protocol, where pH is neutral. Since both protocols had their times slower leak rate than ER membranes. After accounting advantages and disadvantages, we chose to present the data for buffer capacities and S/V ratios, and with the realization obtained by both protocols rather than select one data set over that the calculated P of the ER was likely to be somewhat the other. Most likely, the difference in P between Golgi and underestimated (see above), we conclude that P of the ER MSGs in AtT-20 cells lies somewhere between 7- and 4-fold. was twice as large as that of the Golgi which was 4 –7 times Both data sets are consistent with a gradual decrease in or- greater than MSGs. The variability of our calculated P val- ganelle H permeability from ER to Golgi to secretory gran- ues stemmed in part from the variability in the experimentally ules. Our working model for organelle pH regulation along the determined buffer capacity () values. The experimental deter- regulated secretory pathway is illustrated in Fig. 7. Based on mination of is prone to large variabilities due to different our experimental and modeling results, we conclude that the rates of mixing of NH /NH solutions into the chamber, possi- 3 4 decreasing pH values of organelles of the regulated secretory ble contributions of pH regulatory mechanisms in the organelle pathway is established by gradually increasing the density of and plasma membranes, and the finite permeability of mem- active H pumps from ER to Golgi while concomitantly de- branes to NH . We tried to limit these factors as much as creasing the H permeability from ER to Golgi to MSGs. possible by performing experiments in sodium-free solutions following bafilomycin treatment, but it is nearly impossible to Acknowledgments—We thank J. Llopis (University of California, San Diego, currently at the Universidad de Castilla, La Mancha) for the eliminate these complicating factors completely. Furthermore, AV-KDEL plasmid, M. Kulomaa (University of Jyvaskyla, Finland) for the buffer capacity experiments were performed separately the avidin clone, and K. Teter for the AV-POMC construct and the from the experiments used to determine P . It is likely that H AtT-20 AV-POMC stable cell line. We thank Juan Llopis and members of the Machen and Moore labs for helpful discussions. our error bars for P would have been smaller had we been able to perform the buffer capacity experiments and H leak REFERENCES experiments in the same cells and then use the value for a 1. Schmidt, W. K., and Moore, H. P. (1995) Mol. Biol. Cell 6, 1271–1285 particular cell to calculate the P for that cell. 2. Gerdes, H. H., Rosa, P., Phillips, E., Baeuerle, P. A., Frank, R., Argos, P., and Are the differences in H permeabilities sufficient to account Huttner, W. B. (1989) J. Biol. Chem. 264, 12009 –12015 3. Chanat, E., and Huttner, W. B. (1991) J. Cell Biol. 115, 1505–1519 for the different steady-state pH values of the ER, Golgi, and 4. Colomer, V., Kicska, G. A., and Rindler, M. J. (1996) J. Biol. Chem. 271, 48 –55 MSGs? Alternatively, are concomitant increases in H pump- 5. Arvan, P., and Castle, D. (1998) Biochem. J. 332, 593– 610 6. Hutton, J. C. (1982) Biochem. J. 204, 171–178 ing activity required to establish the stepwise decreases in pH 7. Orci, L., Ravazzola, M., Amherdt, M., Madsen, O., Perrelet, A., Vassalli, J. D., between these organelles? To answer these questions, we used and Anderson, R. G. (1986) J. Cell Biol. 103, 2273–2281 the previously published organelle pH model (26) that de- 8. Orci, L., Ravazzola, M., Storch, M. J., Anderson, R. G., Vassalli, J. D., and Perrelet, A. (1987) Cell 49, 865– 868 scribes both H pumping and H leaking. By using this model, 9. Barasch, J., Gershon, M. D., Nunez, E. A., Tamir, H., and al-Awqati, Q. (1988) along with our experimentally determined Golgi and MSG J. Cell Biol. 107, 2137–2147 10. Orci, L., Halban, P., Perrelet, A., Amherdt, M., Ravazzola, M., and Anderson, buffer capacities and estimated Golgi and MSG P and S/V R. G. (1994) J. Cell Biol. 126, 1149 –1156 values (32, 33), we found that generating the 1.0 unit pH drop 11. Urbe, S., Dittie, A. S., and Tooze, S. A. (1997) Biochem. J. 321, 65–74 from the ER (pH 7.4) to Golgi (pH 6.2) required either a 10-fold 12. Miesenbo ¨ ck, G., De Angelis, D. A., and Rothman, J. E. (1998) Nature 394, Organelle pH in the Regulated Secretory Pathway 33035 192–195 26. Grabe, M., and Oster, G. (2001) J. Gen. Physiol. 117, 329 –343 13. Anderson, R. G. W., and Pathak, R. K. (1985) Cell 40, 635– 643 27. Chavez, R. A., Chen, Y. T., Schmidt, W. K., Carnell, L., and Moore, H. P. (1994) 14. Schwartz, A. L., Strous, G. J., Slot, J. W., and Geuze, H. J. (1985) EMBO J. 4, Methods Cell Biol. 43, 263–288 899 –904 28. Grote, E., Hao, J. C., Bennett, M. K., and Kelly, R. B. (1995) Cell 81, 581–589 15. Seksek, O., Biwersi, J., and Verkman, A. S. (1995) J. Biol. Chem. 270, 29. McNeil, P. L., Murphy, R. F., Lanni, F., and Taylor, D. L. (1984) J. Cell Biol. 4967– 4970 98, 1556 –1564 16. Kim, J. H., Lingwood, C. A., Williams, D. B., Furuya, W., Manolson, M. F., and 30. Teter, K., Chandy, G., Quin ˜ ones, B., Pereyra, K., Machen, T., and Moore, H. P. Grinstein, S. (1996) J. Cell Biol. 134, 1387–1399 (1998) J. Biol. Chem. 273, 19625–19633 17. Kim, J. H., Johannes, L., Goud, B., Antony, C., Lingwood, C. A., Daneman, R., 31. Roos, A., and Boron, W. F. (1981) Physiol. Rev. 61, 296 – 434 and Grinstein, S. (1998) Proc. Natl. Acad. Sci. U. S. A. 95, 2997–3002 32. Taga, R., Alvares, E. P., and Sesso, A. (1993) Arch. Histol. Cytol. 56, 517–523 18. Demaurex, N., Furuya, W., D’Souza, S., Bonifacino, J. S., and Grinstein, S. 33. Gumbiner, B., and Kelly, R. B. (1981) Proc. Natl. Acad. Sci. U. S. A. 78, (1998) J. Biol. Chem. 273, 2044 –2051 318 –322 19. Llopis, J., McCaffery, J. M., Miyawaki, A., Farquhar, M. G., and Tsien, R. Y. 34. Langridge-Smith, J. E., and Dubinsky, W. P. (1986) J. Membr. Biol. 94, (1998) Proc. Natl. Acad. Sci. U. S. A. 95, 6803– 6808 197–204 20. Farinas, J., and Verkman, A. S. (1999) J. Biol. Chem. 274, 7603–7606 35. Hartmann, T., and Verkman, A. S. (1990) Biophys. J. 58, 391– 401 21. Wu, M. M., Llopis, J., Adams, S., McCaffery, J. M., Kulomaa, M. S., Machen, 36. Chege, N. W., and Pfeffer, S. R. (1990) J. Cell Biol. 111, 893– 899 T. E., Moore, H. P., and Tsien, R. Y. (2000) Chemistry and Biology 7, 37. Cain, C. C., Sipe, D. M., and Murphy, R. F. (1989) Proc. Natl. Acad. Sci. 197–209 U. S. A. 86, 544 –548 22. Maycox, P. R., Deckwerth, T., Hell, J. W., and Jahn, R. (1988) J. Biol. Chem. 38. Fuchs, R., Schmid, S., and Mellman, I. (1989) Proc. Natl. Acad. Sci. U. S. A. 86, 263, 15423–15428 539 –543 23. Stobrawa, S. M., Breiderhoff, T., Takamori, S., Engel, D., Schweizer, M., 39. Paula, S., Volkov, A. G., Van Hoek, A. N., Haines, T. H., and Deamer, D. W. Zdebik, A. A., Bosl, M. R., Ruether, K., Jahn, H., Draguhn, A., Jahn, R., and (1996) Biophys. J. 70, 339 –348 Jentsch, T. J. (2001) Neuron 29, 185–196 24. Al-Awqati, Q. (1986) Annu. Rev. Cell Biol. 2, 179 –199 40. Taylor, J. R. (1982) An Introduction to Error Analysis (Commins, E. D., ed) pp. 25. Schapiro, F. B., and Grinstein, S. (2000) J. Biol. Chem. 275, 21025–21032 141–152, University Science Books, Mill Valley, CA
Journal
Journal of Biological Chemistry – Unpaywall
Published: Aug 1, 2001