- Email: [email protected]
Annexin 5 mediates a peroxide-induced Ca2+ influx in B cells
Brief Communication
1403
Annexin 5 mediates a peroxide-induced Ca2+ influx in B cells Helmut Kubista*, Tim E. Hawkins*, Darshana R. Patel†, Harry T. Haigler† and Stephen E. Moss* Annexin 5 is a Ca2+-binding protein, the function of which is poorly understood. Structural and electrophysiological studies have shown that annexin 5 can mediate Ca2+ fluxes across phospholipid membranes in vitro [1]. There is, however, no direct evidence for the existence of annexin 5 Ca2+ channels in living cells. Here, we show that annexin 5 inserts into phospholipid vesicle membranes at neutral pH in the presence of peroxide. We then used targeted gene disruption to explore the role of annexin 5 in peroxide-induced Ca2+ signaling in DT40 pre-B cells. DT40 clones lacking annexin 5 exhibited normal Ca2+ responses to both thapsigargin and B-cell receptor stimulation, but lacked the sustained phase of the response to peroxide. This late phase was due to Ca2+ influx from the extracellular space, demonstrating that annexin 5 mediates a peroxideinduced Ca2+ influx. Thus, peroxide induces annexin 5 membrane insertion in vitro, and peroxide-induced Ca2+ entry in vivo in DT40 cells requires annexin 5. Our results are consistent with a role for annexin 5 either as a Ca2+ channel, or as a signaling intermediate in the peroxideinduced Ca2+-influx pathway. Addresses: *Department of Physiology, University College London, Gower Street, London WC1E 6BT, UK. †Department of Physiology and Biophysics, University of California, Irvine, California 92697, USA. Correspondence: Stephen E. Moss E-mail: [email protected] Received: 17 June 1999 Revised: 16 September 1999 Accepted: 20 October 1999 Published: 22 November 1999 Current Biology 1999, 9:1403–1406 0960-9822/99/$ – see front matter © 1999 Elsevier Science Ltd. All rights reserved.
neutral pH was inhibited by Ca2+, this result demonstrates that annexin 5 can integrate into a lipid membrane at physiological pH in the presence of a natural agonist. Because physiological exposure to peroxide and/or intracellular acidification might induce membrane insertion and ion channel formation by annexin 5 in vivo, we developed a genetic model to investigate the requirement for annexin 5 in the generation of Ca2+ signals. Chicken DT40 clones were derived, in which expression of either annexin 5 or annexin 2 was ablated by targeted gene disruption. Changes in intracellular free Ca2+ concentration ([Ca2+]i) were recorded in annexin 5–/– and annexin 2–/– clones stimulated with thapsigargin, H2O2 or antibodies to the B-cell receptor (anti-IgM). As the annexin 2–/– DT40 cell responses resembled those of wild-type cells [5], these were used as the control throughout this study. Thapsigargin- and anti-IgM-induced Ca2+ responses were similar in annexin 5–/–, annexin 2–/– (see Supplementary material) and wild-type DT40 cells [5]. In contrast, Ca2+ elevations induced by H2O2 were transient in annexin 5–/– cells, but exhibited a slow decaying late phase in annexin 2–/– cells (Figure 2a,b). This observation was confirmed in four independently generated annexin 5–/– clones (Figure 2b). Thus, the absent late phase of the H2O2-induced Ca2+ response represents a true phenotype of annexin 5–/– cells. To investigate H2O2-induced Ca2+ release from internal stores, experiments were performed in Ca2+-free buffer. Release responses in annexin 5–/– and annexin 2–/– cells were transient and indistinguishable during the late phase Figure 1 pH
4.0
5.0
6.0
6.5
7.4
7.4
Control
Results and discussion Using the photoactivatable reagent 3-(trifluoromethyl)-3(m-[125I]iodophenyl)diazirine (125I-TID), which has been used to study the interaction of annexins with phospholipid bilayers (J. Isas, J. Cartailler, Y. Sokolov, D.R.P., R. Langen, H. Luecke, J. Hall, and H.T.H., unpublished observations), we found that annexin 5 and annexin 12 insert into membranes at low pH. Similar findings have been reported for other annexins [2–4]. A survey of physiological agonists led to the discovery that H2O2 induces insertion of annexin 5 into large unilammelar vesicles composed of phosphatidylserine and phosphatidylcholine, without the need for acidification (Figure 1). Although membrane insertion at
H2O2
Current Biology
EGTA
Ca2+
Labelling of annexin 5 with 125I-TID in the presence and absence of H2O2. Annexin 5 was incubated with phospholipid vesicles at the indicated pH in the absence (control) or presence of 3 mM H2O2 as described in the Supplementary material. The entire sample was then subjected to SDS–PAGE and radiography. The only radioactive bands were in the illustrated 35 kDa region of the gel, except for low molecular weight material at the dye front. The data shown are typical of three experiments, each of which gave the same pH-dependent response.
1404
Relative [Ca2+]i (arbitrary units)
(b) 1.0 1.0 0.8 0.8 * 0.6 0.6 0.4 0.4 0.2 0.2 0.0 0.0 H2O2 H2O2 –0.2 –0.2 0 100200300400500 0 100200300400500 Time (sec) Time (sec) Annexin 2–/– Annexin 5–/–
Relative [Ca2+]i (arbitrary units)
(a)
Current Biology Vol 9 No 23
(c)
(d)
0.6 0.4
0.6 0.4
–0.2 0
Fluorescence signal (%)
100 90 80 70 60 50 40
H2O2
0.2 0.0
100 200 300 Time (sec) (f) Mn2+ Annexin 2–/–
0 100 200 300 400 Time (sec) Control
–0.2 0
100 90 80 70 60 50 40
H2O2 100 200 300 Time (sec) Mn2+
Fluorescence signal (%)
0.2 0.0
(e)
0.8
Relative [Ca2+]i (arbitrary units)
Relative [Ca2+]i (arbitrary units)
0.8
Annexin 5–/–
0 100 200 300 400 Time (sec)
Thapsigargin
H2O2 Current Biology
(Figure 2c,d), indicating that annexin 5 influences neither the filling state of intracellular Ca2+ stores, nor the inositol trisphosphate (InsP3)-dependent release of Ca2+ from such stores, upon which the anti-IgM response has been shown to be entirely dependent [6]. We then used the Mn2+ quench technique to establish whether or not the sustained late phase of H2O2-induced Ca2+ responses in DT40 cells was due to Ca2+ influx across the plasma membrane, or the result of changes in efflux or uptake mechanisms (Figure 2e,f). In annexin 2–/– cells, H2O2 stimulation induced an increase in quench rate (%F/min = + 8.47 ± 1.49%) that was similar to the quench obtained with thapsigargin (+ 13.31 ± 0.47%, Figure 2e). In contrast, H2O2 failed to increase quench rates in annexin 5–/– cells (+ 2.74 ± 1.57%), whereas the quench induced by thapsigargin was normal (+ 10.16 ± 1.62%, Figure 2f). These results show that H2O2 increases the Ca2+ permeability of the plasma membrane in annexin 2–/– cells but not in annexin 5–/– cells. Additional experimental evidence that the annexin 5–/– cells have a defective Ca2+ influx is available in the Supplementary material. Thus, the late-phase Ca2+ response to all three agonists requires Ca2+ influx. Only the late-phase Ca2+ influx of the H2O2 response was lost in annexin 5–/– cells, however. As the response to thapsigargin was normal in annexin 5–/–
Figure 2 H2O2-induced Ca2+ responses are abnormal in cells lacking annexin 5. (a–d) Annexin 5–/– cells (red trace and symbols) and annexin 2–/– cells (green trace and symbols) were stimulated with H2O2 in (a,b) Ca2+containing and (c,d) Ca2+-free buffer. H2O2 was 2 mM in (a,b) and 3.5 mM in (c,d). (a,c) Typical traces. (b,d) Averaged responses (for three and four independently generated clones, respectively). Student t-test confirmed that, in the presence of extracellular Ca2+, the late phase of H2O2-induced Ca2+ responses was significantly different (p ≤ 0.05) between annexin 2–/– and annexin 5–/– cells at 125 sec after stimulation*. (e,f) Fura-2-loaded (e) annexin 2–/– and (f) annexin 5–/– cells were excited at the isosbestic wavelength of 360 nm. Mn2+ (0.5 mM) was added as indicated to unstimulated cells (blue traces), or to cells 225 sec (corresponding to –150 sec in the graphs, not shown) after stimulation with 600 nM thapsigargin (red traces) or 2 mM H2O2 (green traces). The instantaneous drop in fura-2 fluorescence, which was similar in annexin 2–/– cells and annexin 5–/– cells, is due to the quench of extracellular dye that has leaked out into the bath solution from loaded cells. The rate of quench was increased by thapsigargin in both annexin 2–/– cells and annexin 5–/– cells, demonstrating the opening of capacitative Ca2+ entry (CCE) channels. H2O2 also increased the quench rate in annexin 2–/– cells, but failed to do so in annexin 5–/– cells. Within the first 100 sec after addition of Mn2+, the quench rate in H2O2-stimulated annexin 2–/– cells was similar to the quench rate observed in the presence of thapsigargin, but thereafter declined to control levels. Fluorescence signals were normalised to the values before the addition of Mn2+ (100% fluorescence). Traces shown are the average of four identical experiments. Vertical bars represent SD.
cells, annexin 5 expression is not required for CCE. The normal late-phase influx of the anti-IgM-induced Ca2+ response, which depends on InsP3-mediated release of Ca2+ from intracellular stores and CCE, also shows that these influx pathways are intact. The defect observed in the annexin 5–/– H2O2-induced response is thus not manifest in these other pathways. Indeed, H2O2 blocks both CCE and the anti-IgM-induced late phase [5]. Furthermore, loss of the InsP3- and store-mediated components of both the H2O2- and anti-IgM-induced responses in DT40 cells lacking the tyrosine kinase Syk only affects the Ca2+ influx component of the IgM response; the H2O2-induced Ca2+ influx remains intact, confirming that it uses an alternative mechanism [6,7]. The late-phase H2O2-induced Ca2+ entry is therefore annexin 5 dependent, mediated by a Ca2+ influx channel that is not involved in B-cell receptor or store-operated Ca2+ entry, and is activated rather than inactivated during oxidative stimulation. Peroxide-induced Ca2+ responses are due to the release of Ca2+ from intracellular stores and Ca2+ influx [8], but the identity of the H2O2-induced Ca2+ influx channel(s) is not known. To gain further insight into the role of annexin 5, we attempted to characterise this influx channel, candidates for which include CCE channels [9], non-selective cation channels [10,11] and L-type Ca2+ channels [12]. Our data show that none of these channels mediate H2O2-induced Ca2+ entry in DT40 cells. First, Ca2+ influx evoked by thapsigargin was blocked by H2O2 [5], ruling out CCE channels. Second, H2O2-induced non-selective cation currents in
Brief Communication
1405
Figure 3
KCl
1.00 H2O2 0.50 0.00 0 200 400 600 800 Time (sec)
7.5 7.4 7.3 7.2 7.1 7.0 6.9
H2O2 Anti-IgM Thapsigargin
(c) Relative [Ca2+]i (arbitrary units)
1.50
(b)
Annexin 2–/– Annexin 5–/–
Cytosolic pH
(a) Relative fluorescence (arbitrary units)
Effect of H2O2 on Vm and intracellular pH (pHi). (a) H2O2 (2 mM) was added as indicated. A decrease in fluorescence indicates hyperpolarisation, an increase indicates depolarisation. Vm was not affected by H2O2 stimulation in annexin 2–/– cells, whereas a hyperpolarisation occurred in annexin 5–/– cells. Cells were depolarised by five consecutive additions of 13 mM KCl (arrowheads) raising extracellular K+ concentration from 5.8 mM to 70.8 mM and Cl– concentration from 282 mM to 347 mM. Signals were normalised to the fluorescence increase obtained by the fifth addition of KCl and aligned to the maximum fluorescence value. (b) Effect of anti-IgM (400 ng/ml), thapsigargin (600 nM) and H2O2 (3.5 mM) on cytosolic pH, as determined using the pH indicator BCECF, in control DT40 cells. H2O2 induced a decrease in the cytosolic pH
100 300 500 Time (sec)
1.0 0.8 0.6 0.4 0.2 0.0 0
Control HCl 200 400 600 Time (sec) Current Biology
from a resting value of 7.25 to 7.05, whereas anti-IgM and thapsigargin were without effect. Identical results were obtained in annexin 5–/– cells. (c) Acidification by addition of HCl (4 mM) as indicated (blue trace) blocked the late phase of H2O2induced Ca2+ responses in annexin 2–/– cells
endothelial and CRI-G1 cells [10,11] have a reversal potential close to or at 0 mV. Accordingly, oxidant activation of these channels leads to depolarisation of the transmembrane voltage (Vm). To test whether similar channels exist in DT40 cells, we measured changes in Vm in annexin 5–/– and annexin 2–/– cells following stimulation with H2O2 (Figure 3a). Annexin 5–/– cells responded to H2O2 in all experiments by hyperpolarisation (n = 8), whereas the Vm of annexin 2–/– cells remained unchanged, slightly hyperpolarised, or depolarised. These data indicate that the conductance changes responsible for the hyperpolarisation in annexin 5–/– cells can also occur in annexin 2–/– cells. Nevertheless, the resting membrane potential (RMP) of annexin 2–/– cells is more negative and closer to the reversal potential for the H2O2induced membrane current. Because H2O2 activates K+ channels [8,13], we suggest that the RMP of annexin 2–/– cells is close to the K+ equilibrium potential (EK), whereas annexin 5–/– cells hyperpolarise towards EK only in response to H2O2 stimulation. These data exclude non-selective cation channels of the type responsible for peroxideinduced Ca2+ influx in CRI-G1 [11] and endothelial cells [10], and also rule out voltage-dependent L-type Ca2+ channels, this being the third proposed H2O2-activated Ca2+ channel. Oxidative activation of L-type channels is also unlikely, as H2O2 suppresses L-type Ca2+ currents [13]. Another potential Ca2+-entry channel in B cells is the antiIgM-activated voltage-insensitive ion channel, which shares pharmacological and serological similarities with L-type Ca2+ channels [14]. But this channel can also be ruled out as the mediator of H2O2-induced Ca2+ entry, because antiIgM-induced Ca2+ responses are blocked by H2O2 [5]. So is annexin 5 itself the Ca2+ influx channel? Unlike annexins 4 and 6, which inhibit ion-channel activity [15,16], annexin 5 has never been demonstrated to modulate the
(green trace). Addition of 4 mM HCl decreased the pHi from 7.3 to 6.9, as determined in separate experiments using BCECF (data not shown). This experiment was performed with buffer containing 250 µM sulfinpyrazone to block leakage of fura-2 from the cells.
Figure 4
Thapsigargin
SOC
Ca2+ Ca2+
Anti-IgM
Peroxide
Anx V
PLC-γ Syk
Ca2+ Ca2+
BCR
InsP3
Ca2+–ATPase Intracellular calcium store Current Biology
Ca2+ signaling pathways in DT40 pre-B cells. Summary of the Ca2+ responses to the three agonists tested in this study. Thapsigargin mobilises Ca2+ by inhibiting the plasma membrane Ca2+–ATPase, resulting in emptying of intracellular stores and activation of storeoperated Ca2+ channels (SOC) by the CCE pathway. The dashed line indicates that the mechanism of SOC activation is not well understood. Antibodies (anti-IgM) to the B-cell receptor (BCR) mobilise Ca2+ by activation of Syk [6] and InsP3-dependent release of Ca2+ from intracellular stores. Emptying of stores leads to SOC activation and Ca2+ influx. H2O2 has also been shown to mobilise intracellular Ca2+ through a Syk-dependent pathway [7], and Syk–/– cells fail to release intracellular Ca2+ in response to either peroxide or anti-IgM [6,7]. Peroxide, but not anti-IgM, stimulates a Ca2+ influx in Syk–/– cells, indicating the existence of a peroxide-sensitive Ca2+ channel that does not require emptying of intracellular stores for its activation [7]. We show here that peroxide inhibits SOC activation, and that peroxide-induced Ca2+ influx requires annexin 5. The observation that annexin 5 inserts into artificial membranes on peroxide treatment is consistent with the idea that a membraneinserted form of annexin 5 functions as a redox-sensitive Ca2+ channel. Green arrows, direct activation; red blocked arrows, inhibition; dashed lines, poorly understood or indirect pathways; PLC-γ, phospholipase C-γ.
1406
Current Biology Vol 9 No 23
Ca2+ channel activity of another protein. In contrast, there is evidence from both in vitro experiments and studies on mineralising chondrocytes that argues for the formation of a Ca2+-selective ion channel by annexin 5 [1,17]. Two models exist to explain Ca2+ conductance by annexin 5. The first model — ‘micro-electroporation’ — depends on Ca2+dependent membrane association by the annexin monomer, with increased permeability of the membrane to Ca2+ [18]. If Ca2+-dependent membrane association was sufficient to activate an annexin 5-dependent Ca2+ flux, one would expect to see an annexin 5-dependent Ca2+ influx in responses induced by agonists such as thapsigargin. The Ca2+ responses induced by anti-IgM and thapsigargin did not show any detectable annexin 5-dependent component in our study, however. Therefore, a rise in [Ca2+]i cannot be the sole factor that activates annexin 5-dependent Ca2+ entry. Indeed, at elevated Ca2+ levels, annexin 5 has stabilising effects on phospholipid membranes [3]. The second model invokes a Ca2+-independent form of membrane integration that has been reported for annexins 1, 5 and 12 at acidic pH [2–4]. To examine whether or not H2O2 elicits intracellular acidification in DT40 cells, we analysed the responses of annexin 2–/– cells to H2O2, thapsigargin and anti-IgM, using the pH-sensitive fluorescent dye BCECF. Of these agonists, only H2O2 induced acidification of the cytosol of DT40 cells (Figure 3b). Identical results were obtained with annexin 5–/– cells (data not shown). Thus, annexin 5 is not required for this component of the H2O2 response. To investigate the influence of pHi on the Ca2+ response, we artificially acidified the cytosol following H2O2 stimulation. Addition of 4 mM HCl decreased the pHi from 7.3 to 6.9, but rather than enhancing the late-phase Ca2+ responses to H2O2 in annexin 2–/– cells, acidification was inhibitory (Figure 3c). This result, taken together with the observation that more extreme acidification (pH 5.0) is required for membrane insertion of annexin 5, effectively rules out a causative role for pH changes in generating the annexin 5-dependent Ca2+ influx. We have shown here that peroxide-induced Ca2+ entry in the chicken pre-B cell line DT40 requires annexin 5, whereas anti-IgM- and thapsigargin-induced Ca2+ signals are annexin 5 independent (summarised in Figure 4). There are two ways in which annexin 5 could mediate peroxide-dependent Ca2+ entry. Annexin 5 either functions in a signaling pathway operated by H2O2 that leads to activation of a Ca2+ entry channel, or annexin 5 forms a Ca2+ channel itself. The in vitro structural, mutational and electrophysiological evidence [1,19] suggests a channelforming role for annexin 5. These findings, taken together with our observation that, in the presence of H2O2, annexin 5 inserts into artificial membranes, support the idea that, in DT40 cells in vivo, H2O2 induces Ca2+ entry through a channel formed by annexin 5.
Supplementary material Additional methodological details and figures showing that Ca 2+ responses mediated by thapsigargin and the B-cell receptor are normal in annexin 5–/– cells, and that H2O2-induced Ca2+ influx is lost in annexin 5–/– cells are available at http://currentbiology.com/supmat/supmatin.htm.
Acknowledgements We thank Jim Putney for helpful discussions and for critically reading the manuscript. This work was funded by the EC (Project ERB-BIO4CT960083) and the Medical Research Council.
References 1. Demange P, Voges D, Benz J, Liemann S, Gottig P, Berendes R, et al.: Annexin V: the key to understanding ion selectivity and voltage regulation? Trends Biochem Sci 1994, 19:272-276. 2. Langen R, Isas JM, Hubbell WL, Haigler HT: A transmembrane form of annexin XII detected by site-directed spin labeling. Proc Natl Acad Sci USA 1998, 95:14060-14065. 3. Kohler G, Hering U, Zschornig O, Arnold K: Annexin V interaction with phosphatidylserine-containing vesicles at low and neutral pH. Biochemistry 1997, 36:8189-8194. 4. Rosengarth A, Wintergalen A, Galla HJ, Hinz HJ, Gerke V: Ca2+-independent interaction of annexin I with phospholipid monolayers. FEBS Lett 1998, 438:279-284. 5. Kubista H, Hawkins T, Moss SE: Characterisation of calcium signalling in DT40 chicken B-cells. Biochim Biophys Acta 1998, 1448:299-310. 6. Takata M, Sabe H, Hata A, Inazu T, Homma Y, Nukada T, et al.: Tyrosine kinases Lyn and Syk regulate B cell receptor-coupled Ca2+ mobilization through distinct pathways. EMBO J 1994, 13:1341-1349. 7. Qin S, Inazu T, Takata M, Kurosaki T, Homma Y, Yamamura H: Cooperation of tyrosine kinases p72syk and p53/56lyn regulates calcium mobilization in chicken B cell oxidant stress signaling. Eur J Biochem 1996, 236:443-449. 8. Krippeit-Drews P, Kramer C, Welker S, Lang F, Ammon HPT, Drews G: Interference of H2O2 with stimulus-secretion coupling in mouse pancreatic β-cells. J Physiol Lond 1999, 514:471-481. 9. Volk T, Hensel M, Kox WJ: Transient Ca2+ changes in endothelial cells induced by low doses of reactive oxygen species: role of hydrogen peroxide. Mol Cell Biochem 1997, 171:11-21. 10. Koliwad SK, Kunze DL, Elliott SJ: Oxidant stress activates a nonselective cation channel responsible for membrane depolarization in calf vascular endothelial cells. J Physiol Lond 1996, 491:1-12. 11. Herson PS, Lee K, Pinnock RD, Hughes J, Ashford MLJ: Hydrogen peroxide induces intracellular calcium overload by activation of a non-selective cation channel in an insulin-secreting cell line. J Biol Chem 1999, 274:833-841. 12. Roveri A, Coassin M, Maiorino M, Zamburlini A, van Amsterdam FT, Ratti E, et al.: Effect of hydrogen peroxide on calcium homeostasis in smooth muscle cells. Arch Biochem Biophys 1992, 297:265-270. 13. Kourie JI: Interaction of reactive oxygen species with ion transport mechanisms. Am J Physiol 1998, 275:C1-C24. 14. Sadighi Akha AA, Willmott NJ, Brickley K, Dolphin AC, Galione A, Hunt SV: Anti-Ig-induced calcium influx in rat B lymphocytes mediated by cGMP through a dihydropyridine-sensitive channel. J Biol Chem 1996, 271:7297-7300. 15. Kaetzel MA, Chan HC, Dubinsky WP, Dedman JR, Nelson DJ: A role for annexin IV in epithelial cell function. Inhibition of calciumactivated chloride conductance. J Biol Chem 1994, 269:5297-5302. 16. Naciff JM, Behbehani MM, Kaetzel MA, Dedman JR: Annexin VI modulates Ca2+ and K+ conductances of spinal cord and dorsal root ganglion neurons. Am J Physiol 1996, 271:C2004-C2015. 17. Arispe N, Rojas E, Genge BR, Wu LN, Wuthier RE: Similarity in calcium channel activity of annexin V and matrix vesicles in planar lipid bilayers. Biophys J 1996, 71:1764-1775. 18. Voges D, Berendes R, Demange P, Benz J, Goettig P, Liemann S, Huber R, Burger A: Structure and function of the ion channel model system annexin V. In Advances in Enzymology and Related Areas of Molecular Biology. Edited by Meister A. John Wiley & Sons 1995, 71:209-239. 19. Burger A, Voges D, Demange P, Ruiz Perez C, Huber R, Berendes R: Structural and electrophysiological analysis of annexin V mutants. J Mol Biol 1994, 237:479-499.
S1
Supplementary material Annexin 5 mediates a peroxide-induced Ca2+ influx in B cells Helmut Kubista, Tim E. Hawkins, Darshana R. Patel, Harry T. Haigler and Stephen E. Moss Current Biology 22 November 1999, 9:1403–1406 Figure S1
(a)
Time (sec)
Time (sec) Annexin 2–/–
Ca2+ responses mediated by thapsigargin and the B-cell receptor are normal in cells lacking annexin 5. DT40 cells were loaded with fura-2 and stirred at 40oC in a cuvette in a spectrofluorimeter. (a) Thapsigargin (600 nM) and (b) anti-IgM (400 ng/ml) were added at 100 sec as indicated by the arrows, and fluorescence changes (∆R) were monitored continuously for a further 500 sec. Red trace, annexin 5–/– cells; green trace, annexin 2–/– cells.
(b)
Supplementary materials and methods
Labeling with 125I-TID Large unilammelar vesicles composed of phosphatidylserine and phosphatidylcholine (2:1 molar ratio) were prepared according to the Reeves/Dowben protocol [S2]. The topography of membrane-bound annexin 5 was probed with 125I-TID as in [S3,S4] and (J. Isas, J. Cartailler, Y. Sokolov, D.R.P., R. Langen, H. Luecke, J. Hall and H.T.H., unpublished data). Recombinant human annexin 5 (12 µg) was incubated with vesicles (1:1000 molar ratio protein:phospholipid) in a solution (30 µl final volume) containing either EGTA (500 µM) or CaC03 (500 µM) at the indicated pH. The following buffers (100 mM) were used: sodium acetate for pH 4.0–5.6; MES for pH 6.0–6.5; and Tris:HCl for pH 7.4. The 125I-TID solutions were diluted with ethanol to a concentration of 1 µCi/l. After incubating protein and phospholipid for 10 min in either the presence or absence of H2O2 (3 mM), 1 µCi 125I-TID was added and the solution was incubated for 20 min in the dark. The sample was then subjected to a 30 min irradiation from an ultraviolet light (Minerallight Lamp Model UVSL-25, 366 nm) at a distance of 5 cm to generate the reactive carbene. All incubations were at room temperature. After the incubation, Laemmli SDS sample buffer was added and the samples were analysed by SDS–PAGE and the dried gels were exposed to Phosphor Screens (Molecular Dynamics) and the data were analysed using Image Quant software.
Ca2+
1.6 1.4
EGTA
1.2 H2O2
1.0
Control (no H2O2)
0.8 0 100 200 300 400 500 600 Time (sec) 2.0 Annexin 5–/– 1.8 Ca2+ 1.6 1.4
EGTA
1.2 H2O2
1.0
Control (no H2O2)
0.8 0.6
Materials and cell culture (c) Relative [Ca2+]i (arbitrary units)
Source of materials as well as generation and culture of annexin gene knock-out DT40 clones was as or will be described elsewhere ([S1]; T.E.H. and S.E.M., unpublished data). Phosphatidylserine (bovine brain, catalogue #840032) and phosphatidylcholine (egg yolk, catalogue #840051) were obtained from Avanti Polar Lipids (Alabaster, Alabama). The photoactivatable probe 3-(trifluoromethyl)-3-(m[125I]iodophenyl)diazirine (125I-TID) was purchased from Amersham.
Annexin 2–/–
1.8
0.6
Annexin 5–/–
[Ca2+]i (arbitrary units)
Current Biology
2.0 [Ca2+]i (arbitrary units)
Anti-IgM
0 10 0 20 0 30 0 40 0 50 0
Thapsigargin
1.25 1.00 0.75 0.50 0.25 0.00 –0.25
Relative [Ca2+]i (arbitrary units)
(b)
0 10 0 20 0 30 0 40 0 50 0 60 0
Relative [Ca2+]i (arbitrary units)
(a) 1.25 1.00 0.75 0.50 0.25 0.00 –0.25
Figure S2
0 100 200 300 400 500 600 Time (sec) 1.0 Annexin 2–/– 0.8 0.6 0.4 0.2
Annexin 5–/–
0 250 350 450 550 650 Time (sec) Current Biology
H2O2-induced Ca2+ influx is lost in annexin 5–/– cells. EGTA (1 mM), H2O2 (2 mM) and Ca2+ (2.5 mM) were added as indicated to (a) annexin 2–/– cells (green trace) and (b) annexin 5–/– cells (red trace) in nominally Ca2+-free buffer. The black traces (control) in (a,b) were obtained in experiments on the respective cells by omission of H2O2. Ca2+ was added 250 sec after H2O2 stimulation and evoked pronounced intracellular Ca2+ elevations in annexin 2–/– cells compared with the control, but failed to do so in annexin 5–/– cells. (c) Responses obtained by Ca2+ re-addition at various time points (160, 250 and 360 sec) after H2O2 stimulation from annexin 2–/– cells (green traces) and annexin 5–/– cells (red traces) are shown superimposed. Ca2+ addition to the EGTA-containing buffer in these experiments was calculated to give a final free Ca2+ concentration of 1.5 mM. Experiments shown here were performed with buffer containing 250 µM sulfinpyrazone. This result supports the data in Figure 2e,f in the paper, showing that the defect in the response of annexin 5–/– cells to peroxide is due to a failure in Ca2+ influx.
Spectrofluorimetry All experiments were performed with an LS50 B fluorimeter (Perkin Elmer) at 40°C on stirred cell suspensions in HBSS (Hanks’ Balanced
Salts Solution), which was supplemented with HEPES, sodium bicarbonate and contained: 1.26 mM CaCl2, 0.81 mM MgSO4, 5.4 mM KCl,
S2
Supplementary material
0.44 mM KH2PO4, 136.9 mM NaCl, 0.34 mM Na2HPO4, 4.2 mM NaHCO3, 5.5 mM D-glucose and 10 mM HEPES. For Ca2+-free experiments, 3 mM EGTA was added. Agonists were added from concentrated stocks to obtain the desired concentrations. Fura-2 Ca2+ measurements of DT40 cells were performed as described in detail in a previous paper [S1]. In some experiments (as indicated in the figure legends), HBSS was supplemented with sulfinpyrazone (250 µM) to minimise loss of indicator dye. Transmembrane voltage was measured using bis-(1,3-dibutylbarbituric acid) trimethine oxonol (DiBAC4(3), Molecular Probes); 106 cells per ml were incubated for 10 min with 200 nM DiBAC4(3), which was added to the fluorimeter cuvette prior to the actual experiment. DiBAC4(3) fluorescence was measured with excitation at 490 nm and emission at 516 nm. Measurements of pH were performed by loading 106 cells per ml with 200 nM of the acetoxymethyl ester of 2′,7′-bis-(2-carboxyethyl)-5-(and-6)-carboxyfluorescin (BCECF, Molecular Probes) for 30 min at 40°C. After loading, cells were washed twice with HBSS, whereafter the fluorescence ratio (Em1/Em2) was determined by excitation 1 at 490 nm, excitation 2 at 440 nm and emission (Em) at 535 nm; pH was calibrated from the BCECF fluorescence ratio as described elsewhere [S5].
Manganese quench DT40 cells were loaded with Fura-2/AM as described for measurements of intracellular free Ca2+ [S1]. After loading, cells were washed and finally resuspended in HBSS containing 250 µM sulfinpyrazone. Suspensions of cells were excited at the dye’s isosbestic wavelength of 360 nm and emission was monitored at 510 nm; 0.5 mM Mn2+ was added as indicated to unstimulated cells or to cells 225 sec after stimulation with thapsigargin or H2O2.
Supplementary references S1. Kubista H, Hawkins T, Moss SE: Characterisation of calcium signalling in DT40 chicken B-cells. Biochim Biophys Acta 1998, 1448:299-310. S2. Reeves JP, Dowben RM: Formation and properties of thin-walled phospholipid vesicles. J Cell Physiol 1969, 73:49-60. S3. Brunner J, Semenza G: Selective labeling of the hydrophobic core of membranes with 3-(trifluoromethyl)-3-(m-[125I]iodophenyl) diazirine, a carbene-generating reagent. Biochemistry 1981, 20:7174-7182. S4. Hoppe J, Brunner J, Jorgensen BB: Structure of the membraneembedded F0 part of F1F0 ATP synthase from Escherichia coli as inferred from labeling with 3-(trifluoromethyl)-3-(M[125I]iodophenyl) diazirine. Biochemistry 1984, 23:5610-5616. S5. James-Kracke MR: Quick and accurate method to convert BCECF fluorescence to pHi: calibration in three different types of cell preparations. J Cell Physiol 1992, 151:596-603.
1403
Annexin 5 mediates a peroxide-induced Ca2+ influx in B cells Helmut Kubista*, Tim E. Hawkins*, Darshana R. Patel†, Harry T. Haigler† and Stephen E. Moss* Annexin 5 is a Ca2+-binding protein, the function of which is poorly understood. Structural and electrophysiological studies have shown that annexin 5 can mediate Ca2+ fluxes across phospholipid membranes in vitro [1]. There is, however, no direct evidence for the existence of annexin 5 Ca2+ channels in living cells. Here, we show that annexin 5 inserts into phospholipid vesicle membranes at neutral pH in the presence of peroxide. We then used targeted gene disruption to explore the role of annexin 5 in peroxide-induced Ca2+ signaling in DT40 pre-B cells. DT40 clones lacking annexin 5 exhibited normal Ca2+ responses to both thapsigargin and B-cell receptor stimulation, but lacked the sustained phase of the response to peroxide. This late phase was due to Ca2+ influx from the extracellular space, demonstrating that annexin 5 mediates a peroxideinduced Ca2+ influx. Thus, peroxide induces annexin 5 membrane insertion in vitro, and peroxide-induced Ca2+ entry in vivo in DT40 cells requires annexin 5. Our results are consistent with a role for annexin 5 either as a Ca2+ channel, or as a signaling intermediate in the peroxideinduced Ca2+-influx pathway. Addresses: *Department of Physiology, University College London, Gower Street, London WC1E 6BT, UK. †Department of Physiology and Biophysics, University of California, Irvine, California 92697, USA. Correspondence: Stephen E. Moss E-mail: [email protected] Received: 17 June 1999 Revised: 16 September 1999 Accepted: 20 October 1999 Published: 22 November 1999 Current Biology 1999, 9:1403–1406 0960-9822/99/$ – see front matter © 1999 Elsevier Science Ltd. All rights reserved.
neutral pH was inhibited by Ca2+, this result demonstrates that annexin 5 can integrate into a lipid membrane at physiological pH in the presence of a natural agonist. Because physiological exposure to peroxide and/or intracellular acidification might induce membrane insertion and ion channel formation by annexin 5 in vivo, we developed a genetic model to investigate the requirement for annexin 5 in the generation of Ca2+ signals. Chicken DT40 clones were derived, in which expression of either annexin 5 or annexin 2 was ablated by targeted gene disruption. Changes in intracellular free Ca2+ concentration ([Ca2+]i) were recorded in annexin 5–/– and annexin 2–/– clones stimulated with thapsigargin, H2O2 or antibodies to the B-cell receptor (anti-IgM). As the annexin 2–/– DT40 cell responses resembled those of wild-type cells [5], these were used as the control throughout this study. Thapsigargin- and anti-IgM-induced Ca2+ responses were similar in annexin 5–/–, annexin 2–/– (see Supplementary material) and wild-type DT40 cells [5]. In contrast, Ca2+ elevations induced by H2O2 were transient in annexin 5–/– cells, but exhibited a slow decaying late phase in annexin 2–/– cells (Figure 2a,b). This observation was confirmed in four independently generated annexin 5–/– clones (Figure 2b). Thus, the absent late phase of the H2O2-induced Ca2+ response represents a true phenotype of annexin 5–/– cells. To investigate H2O2-induced Ca2+ release from internal stores, experiments were performed in Ca2+-free buffer. Release responses in annexin 5–/– and annexin 2–/– cells were transient and indistinguishable during the late phase Figure 1 pH
4.0
5.0
6.0
6.5
7.4
7.4
Control
Results and discussion Using the photoactivatable reagent 3-(trifluoromethyl)-3(m-[125I]iodophenyl)diazirine (125I-TID), which has been used to study the interaction of annexins with phospholipid bilayers (J. Isas, J. Cartailler, Y. Sokolov, D.R.P., R. Langen, H. Luecke, J. Hall, and H.T.H., unpublished observations), we found that annexin 5 and annexin 12 insert into membranes at low pH. Similar findings have been reported for other annexins [2–4]. A survey of physiological agonists led to the discovery that H2O2 induces insertion of annexin 5 into large unilammelar vesicles composed of phosphatidylserine and phosphatidylcholine, without the need for acidification (Figure 1). Although membrane insertion at
H2O2
Current Biology
EGTA
Ca2+
Labelling of annexin 5 with 125I-TID in the presence and absence of H2O2. Annexin 5 was incubated with phospholipid vesicles at the indicated pH in the absence (control) or presence of 3 mM H2O2 as described in the Supplementary material. The entire sample was then subjected to SDS–PAGE and radiography. The only radioactive bands were in the illustrated 35 kDa region of the gel, except for low molecular weight material at the dye front. The data shown are typical of three experiments, each of which gave the same pH-dependent response.
1404
Relative [Ca2+]i (arbitrary units)
(b) 1.0 1.0 0.8 0.8 * 0.6 0.6 0.4 0.4 0.2 0.2 0.0 0.0 H2O2 H2O2 –0.2 –0.2 0 100200300400500 0 100200300400500 Time (sec) Time (sec) Annexin 2–/– Annexin 5–/–
Relative [Ca2+]i (arbitrary units)
(a)
Current Biology Vol 9 No 23
(c)
(d)
0.6 0.4
0.6 0.4
–0.2 0
Fluorescence signal (%)
100 90 80 70 60 50 40
H2O2
0.2 0.0
100 200 300 Time (sec) (f) Mn2+ Annexin 2–/–
0 100 200 300 400 Time (sec) Control
–0.2 0
100 90 80 70 60 50 40
H2O2 100 200 300 Time (sec) Mn2+
Fluorescence signal (%)
0.2 0.0
(e)
0.8
Relative [Ca2+]i (arbitrary units)
Relative [Ca2+]i (arbitrary units)
0.8
Annexin 5–/–
0 100 200 300 400 Time (sec)
Thapsigargin
H2O2 Current Biology
(Figure 2c,d), indicating that annexin 5 influences neither the filling state of intracellular Ca2+ stores, nor the inositol trisphosphate (InsP3)-dependent release of Ca2+ from such stores, upon which the anti-IgM response has been shown to be entirely dependent [6]. We then used the Mn2+ quench technique to establish whether or not the sustained late phase of H2O2-induced Ca2+ responses in DT40 cells was due to Ca2+ influx across the plasma membrane, or the result of changes in efflux or uptake mechanisms (Figure 2e,f). In annexin 2–/– cells, H2O2 stimulation induced an increase in quench rate (%F/min = + 8.47 ± 1.49%) that was similar to the quench obtained with thapsigargin (+ 13.31 ± 0.47%, Figure 2e). In contrast, H2O2 failed to increase quench rates in annexin 5–/– cells (+ 2.74 ± 1.57%), whereas the quench induced by thapsigargin was normal (+ 10.16 ± 1.62%, Figure 2f). These results show that H2O2 increases the Ca2+ permeability of the plasma membrane in annexin 2–/– cells but not in annexin 5–/– cells. Additional experimental evidence that the annexin 5–/– cells have a defective Ca2+ influx is available in the Supplementary material. Thus, the late-phase Ca2+ response to all three agonists requires Ca2+ influx. Only the late-phase Ca2+ influx of the H2O2 response was lost in annexin 5–/– cells, however. As the response to thapsigargin was normal in annexin 5–/–
Figure 2 H2O2-induced Ca2+ responses are abnormal in cells lacking annexin 5. (a–d) Annexin 5–/– cells (red trace and symbols) and annexin 2–/– cells (green trace and symbols) were stimulated with H2O2 in (a,b) Ca2+containing and (c,d) Ca2+-free buffer. H2O2 was 2 mM in (a,b) and 3.5 mM in (c,d). (a,c) Typical traces. (b,d) Averaged responses (for three and four independently generated clones, respectively). Student t-test confirmed that, in the presence of extracellular Ca2+, the late phase of H2O2-induced Ca2+ responses was significantly different (p ≤ 0.05) between annexin 2–/– and annexin 5–/– cells at 125 sec after stimulation*. (e,f) Fura-2-loaded (e) annexin 2–/– and (f) annexin 5–/– cells were excited at the isosbestic wavelength of 360 nm. Mn2+ (0.5 mM) was added as indicated to unstimulated cells (blue traces), or to cells 225 sec (corresponding to –150 sec in the graphs, not shown) after stimulation with 600 nM thapsigargin (red traces) or 2 mM H2O2 (green traces). The instantaneous drop in fura-2 fluorescence, which was similar in annexin 2–/– cells and annexin 5–/– cells, is due to the quench of extracellular dye that has leaked out into the bath solution from loaded cells. The rate of quench was increased by thapsigargin in both annexin 2–/– cells and annexin 5–/– cells, demonstrating the opening of capacitative Ca2+ entry (CCE) channels. H2O2 also increased the quench rate in annexin 2–/– cells, but failed to do so in annexin 5–/– cells. Within the first 100 sec after addition of Mn2+, the quench rate in H2O2-stimulated annexin 2–/– cells was similar to the quench rate observed in the presence of thapsigargin, but thereafter declined to control levels. Fluorescence signals were normalised to the values before the addition of Mn2+ (100% fluorescence). Traces shown are the average of four identical experiments. Vertical bars represent SD.
cells, annexin 5 expression is not required for CCE. The normal late-phase influx of the anti-IgM-induced Ca2+ response, which depends on InsP3-mediated release of Ca2+ from intracellular stores and CCE, also shows that these influx pathways are intact. The defect observed in the annexin 5–/– H2O2-induced response is thus not manifest in these other pathways. Indeed, H2O2 blocks both CCE and the anti-IgM-induced late phase [5]. Furthermore, loss of the InsP3- and store-mediated components of both the H2O2- and anti-IgM-induced responses in DT40 cells lacking the tyrosine kinase Syk only affects the Ca2+ influx component of the IgM response; the H2O2-induced Ca2+ influx remains intact, confirming that it uses an alternative mechanism [6,7]. The late-phase H2O2-induced Ca2+ entry is therefore annexin 5 dependent, mediated by a Ca2+ influx channel that is not involved in B-cell receptor or store-operated Ca2+ entry, and is activated rather than inactivated during oxidative stimulation. Peroxide-induced Ca2+ responses are due to the release of Ca2+ from intracellular stores and Ca2+ influx [8], but the identity of the H2O2-induced Ca2+ influx channel(s) is not known. To gain further insight into the role of annexin 5, we attempted to characterise this influx channel, candidates for which include CCE channels [9], non-selective cation channels [10,11] and L-type Ca2+ channels [12]. Our data show that none of these channels mediate H2O2-induced Ca2+ entry in DT40 cells. First, Ca2+ influx evoked by thapsigargin was blocked by H2O2 [5], ruling out CCE channels. Second, H2O2-induced non-selective cation currents in
Brief Communication
1405
Figure 3
KCl
1.00 H2O2 0.50 0.00 0 200 400 600 800 Time (sec)
7.5 7.4 7.3 7.2 7.1 7.0 6.9
H2O2 Anti-IgM Thapsigargin
(c) Relative [Ca2+]i (arbitrary units)
1.50
(b)
Annexin 2–/– Annexin 5–/–
Cytosolic pH
(a) Relative fluorescence (arbitrary units)
Effect of H2O2 on Vm and intracellular pH (pHi). (a) H2O2 (2 mM) was added as indicated. A decrease in fluorescence indicates hyperpolarisation, an increase indicates depolarisation. Vm was not affected by H2O2 stimulation in annexin 2–/– cells, whereas a hyperpolarisation occurred in annexin 5–/– cells. Cells were depolarised by five consecutive additions of 13 mM KCl (arrowheads) raising extracellular K+ concentration from 5.8 mM to 70.8 mM and Cl– concentration from 282 mM to 347 mM. Signals were normalised to the fluorescence increase obtained by the fifth addition of KCl and aligned to the maximum fluorescence value. (b) Effect of anti-IgM (400 ng/ml), thapsigargin (600 nM) and H2O2 (3.5 mM) on cytosolic pH, as determined using the pH indicator BCECF, in control DT40 cells. H2O2 induced a decrease in the cytosolic pH
100 300 500 Time (sec)
1.0 0.8 0.6 0.4 0.2 0.0 0
Control HCl 200 400 600 Time (sec) Current Biology
from a resting value of 7.25 to 7.05, whereas anti-IgM and thapsigargin were without effect. Identical results were obtained in annexin 5–/– cells. (c) Acidification by addition of HCl (4 mM) as indicated (blue trace) blocked the late phase of H2O2induced Ca2+ responses in annexin 2–/– cells
endothelial and CRI-G1 cells [10,11] have a reversal potential close to or at 0 mV. Accordingly, oxidant activation of these channels leads to depolarisation of the transmembrane voltage (Vm). To test whether similar channels exist in DT40 cells, we measured changes in Vm in annexin 5–/– and annexin 2–/– cells following stimulation with H2O2 (Figure 3a). Annexin 5–/– cells responded to H2O2 in all experiments by hyperpolarisation (n = 8), whereas the Vm of annexin 2–/– cells remained unchanged, slightly hyperpolarised, or depolarised. These data indicate that the conductance changes responsible for the hyperpolarisation in annexin 5–/– cells can also occur in annexin 2–/– cells. Nevertheless, the resting membrane potential (RMP) of annexin 2–/– cells is more negative and closer to the reversal potential for the H2O2induced membrane current. Because H2O2 activates K+ channels [8,13], we suggest that the RMP of annexin 2–/– cells is close to the K+ equilibrium potential (EK), whereas annexin 5–/– cells hyperpolarise towards EK only in response to H2O2 stimulation. These data exclude non-selective cation channels of the type responsible for peroxideinduced Ca2+ influx in CRI-G1 [11] and endothelial cells [10], and also rule out voltage-dependent L-type Ca2+ channels, this being the third proposed H2O2-activated Ca2+ channel. Oxidative activation of L-type channels is also unlikely, as H2O2 suppresses L-type Ca2+ currents [13]. Another potential Ca2+-entry channel in B cells is the antiIgM-activated voltage-insensitive ion channel, which shares pharmacological and serological similarities with L-type Ca2+ channels [14]. But this channel can also be ruled out as the mediator of H2O2-induced Ca2+ entry, because antiIgM-induced Ca2+ responses are blocked by H2O2 [5]. So is annexin 5 itself the Ca2+ influx channel? Unlike annexins 4 and 6, which inhibit ion-channel activity [15,16], annexin 5 has never been demonstrated to modulate the
(green trace). Addition of 4 mM HCl decreased the pHi from 7.3 to 6.9, as determined in separate experiments using BCECF (data not shown). This experiment was performed with buffer containing 250 µM sulfinpyrazone to block leakage of fura-2 from the cells.
Figure 4
Thapsigargin
SOC
Ca2+ Ca2+
Anti-IgM
Peroxide
Anx V
PLC-γ Syk
Ca2+ Ca2+
BCR
InsP3
Ca2+–ATPase Intracellular calcium store Current Biology
Ca2+ signaling pathways in DT40 pre-B cells. Summary of the Ca2+ responses to the three agonists tested in this study. Thapsigargin mobilises Ca2+ by inhibiting the plasma membrane Ca2+–ATPase, resulting in emptying of intracellular stores and activation of storeoperated Ca2+ channels (SOC) by the CCE pathway. The dashed line indicates that the mechanism of SOC activation is not well understood. Antibodies (anti-IgM) to the B-cell receptor (BCR) mobilise Ca2+ by activation of Syk [6] and InsP3-dependent release of Ca2+ from intracellular stores. Emptying of stores leads to SOC activation and Ca2+ influx. H2O2 has also been shown to mobilise intracellular Ca2+ through a Syk-dependent pathway [7], and Syk–/– cells fail to release intracellular Ca2+ in response to either peroxide or anti-IgM [6,7]. Peroxide, but not anti-IgM, stimulates a Ca2+ influx in Syk–/– cells, indicating the existence of a peroxide-sensitive Ca2+ channel that does not require emptying of intracellular stores for its activation [7]. We show here that peroxide inhibits SOC activation, and that peroxide-induced Ca2+ influx requires annexin 5. The observation that annexin 5 inserts into artificial membranes on peroxide treatment is consistent with the idea that a membraneinserted form of annexin 5 functions as a redox-sensitive Ca2+ channel. Green arrows, direct activation; red blocked arrows, inhibition; dashed lines, poorly understood or indirect pathways; PLC-γ, phospholipase C-γ.
1406
Current Biology Vol 9 No 23
Ca2+ channel activity of another protein. In contrast, there is evidence from both in vitro experiments and studies on mineralising chondrocytes that argues for the formation of a Ca2+-selective ion channel by annexin 5 [1,17]. Two models exist to explain Ca2+ conductance by annexin 5. The first model — ‘micro-electroporation’ — depends on Ca2+dependent membrane association by the annexin monomer, with increased permeability of the membrane to Ca2+ [18]. If Ca2+-dependent membrane association was sufficient to activate an annexin 5-dependent Ca2+ flux, one would expect to see an annexin 5-dependent Ca2+ influx in responses induced by agonists such as thapsigargin. The Ca2+ responses induced by anti-IgM and thapsigargin did not show any detectable annexin 5-dependent component in our study, however. Therefore, a rise in [Ca2+]i cannot be the sole factor that activates annexin 5-dependent Ca2+ entry. Indeed, at elevated Ca2+ levels, annexin 5 has stabilising effects on phospholipid membranes [3]. The second model invokes a Ca2+-independent form of membrane integration that has been reported for annexins 1, 5 and 12 at acidic pH [2–4]. To examine whether or not H2O2 elicits intracellular acidification in DT40 cells, we analysed the responses of annexin 2–/– cells to H2O2, thapsigargin and anti-IgM, using the pH-sensitive fluorescent dye BCECF. Of these agonists, only H2O2 induced acidification of the cytosol of DT40 cells (Figure 3b). Identical results were obtained with annexin 5–/– cells (data not shown). Thus, annexin 5 is not required for this component of the H2O2 response. To investigate the influence of pHi on the Ca2+ response, we artificially acidified the cytosol following H2O2 stimulation. Addition of 4 mM HCl decreased the pHi from 7.3 to 6.9, but rather than enhancing the late-phase Ca2+ responses to H2O2 in annexin 2–/– cells, acidification was inhibitory (Figure 3c). This result, taken together with the observation that more extreme acidification (pH 5.0) is required for membrane insertion of annexin 5, effectively rules out a causative role for pH changes in generating the annexin 5-dependent Ca2+ influx. We have shown here that peroxide-induced Ca2+ entry in the chicken pre-B cell line DT40 requires annexin 5, whereas anti-IgM- and thapsigargin-induced Ca2+ signals are annexin 5 independent (summarised in Figure 4). There are two ways in which annexin 5 could mediate peroxide-dependent Ca2+ entry. Annexin 5 either functions in a signaling pathway operated by H2O2 that leads to activation of a Ca2+ entry channel, or annexin 5 forms a Ca2+ channel itself. The in vitro structural, mutational and electrophysiological evidence [1,19] suggests a channelforming role for annexin 5. These findings, taken together with our observation that, in the presence of H2O2, annexin 5 inserts into artificial membranes, support the idea that, in DT40 cells in vivo, H2O2 induces Ca2+ entry through a channel formed by annexin 5.
Supplementary material Additional methodological details and figures showing that Ca 2+ responses mediated by thapsigargin and the B-cell receptor are normal in annexin 5–/– cells, and that H2O2-induced Ca2+ influx is lost in annexin 5–/– cells are available at http://currentbiology.com/supmat/supmatin.htm.
Acknowledgements We thank Jim Putney for helpful discussions and for critically reading the manuscript. This work was funded by the EC (Project ERB-BIO4CT960083) and the Medical Research Council.
References 1. Demange P, Voges D, Benz J, Liemann S, Gottig P, Berendes R, et al.: Annexin V: the key to understanding ion selectivity and voltage regulation? Trends Biochem Sci 1994, 19:272-276. 2. Langen R, Isas JM, Hubbell WL, Haigler HT: A transmembrane form of annexin XII detected by site-directed spin labeling. Proc Natl Acad Sci USA 1998, 95:14060-14065. 3. Kohler G, Hering U, Zschornig O, Arnold K: Annexin V interaction with phosphatidylserine-containing vesicles at low and neutral pH. Biochemistry 1997, 36:8189-8194. 4. Rosengarth A, Wintergalen A, Galla HJ, Hinz HJ, Gerke V: Ca2+-independent interaction of annexin I with phospholipid monolayers. FEBS Lett 1998, 438:279-284. 5. Kubista H, Hawkins T, Moss SE: Characterisation of calcium signalling in DT40 chicken B-cells. Biochim Biophys Acta 1998, 1448:299-310. 6. Takata M, Sabe H, Hata A, Inazu T, Homma Y, Nukada T, et al.: Tyrosine kinases Lyn and Syk regulate B cell receptor-coupled Ca2+ mobilization through distinct pathways. EMBO J 1994, 13:1341-1349. 7. Qin S, Inazu T, Takata M, Kurosaki T, Homma Y, Yamamura H: Cooperation of tyrosine kinases p72syk and p53/56lyn regulates calcium mobilization in chicken B cell oxidant stress signaling. Eur J Biochem 1996, 236:443-449. 8. Krippeit-Drews P, Kramer C, Welker S, Lang F, Ammon HPT, Drews G: Interference of H2O2 with stimulus-secretion coupling in mouse pancreatic β-cells. J Physiol Lond 1999, 514:471-481. 9. Volk T, Hensel M, Kox WJ: Transient Ca2+ changes in endothelial cells induced by low doses of reactive oxygen species: role of hydrogen peroxide. Mol Cell Biochem 1997, 171:11-21. 10. Koliwad SK, Kunze DL, Elliott SJ: Oxidant stress activates a nonselective cation channel responsible for membrane depolarization in calf vascular endothelial cells. J Physiol Lond 1996, 491:1-12. 11. Herson PS, Lee K, Pinnock RD, Hughes J, Ashford MLJ: Hydrogen peroxide induces intracellular calcium overload by activation of a non-selective cation channel in an insulin-secreting cell line. J Biol Chem 1999, 274:833-841. 12. Roveri A, Coassin M, Maiorino M, Zamburlini A, van Amsterdam FT, Ratti E, et al.: Effect of hydrogen peroxide on calcium homeostasis in smooth muscle cells. Arch Biochem Biophys 1992, 297:265-270. 13. Kourie JI: Interaction of reactive oxygen species with ion transport mechanisms. Am J Physiol 1998, 275:C1-C24. 14. Sadighi Akha AA, Willmott NJ, Brickley K, Dolphin AC, Galione A, Hunt SV: Anti-Ig-induced calcium influx in rat B lymphocytes mediated by cGMP through a dihydropyridine-sensitive channel. J Biol Chem 1996, 271:7297-7300. 15. Kaetzel MA, Chan HC, Dubinsky WP, Dedman JR, Nelson DJ: A role for annexin IV in epithelial cell function. Inhibition of calciumactivated chloride conductance. J Biol Chem 1994, 269:5297-5302. 16. Naciff JM, Behbehani MM, Kaetzel MA, Dedman JR: Annexin VI modulates Ca2+ and K+ conductances of spinal cord and dorsal root ganglion neurons. Am J Physiol 1996, 271:C2004-C2015. 17. Arispe N, Rojas E, Genge BR, Wu LN, Wuthier RE: Similarity in calcium channel activity of annexin V and matrix vesicles in planar lipid bilayers. Biophys J 1996, 71:1764-1775. 18. Voges D, Berendes R, Demange P, Benz J, Goettig P, Liemann S, Huber R, Burger A: Structure and function of the ion channel model system annexin V. In Advances in Enzymology and Related Areas of Molecular Biology. Edited by Meister A. John Wiley & Sons 1995, 71:209-239. 19. Burger A, Voges D, Demange P, Ruiz Perez C, Huber R, Berendes R: Structural and electrophysiological analysis of annexin V mutants. J Mol Biol 1994, 237:479-499.
S1
Supplementary material Annexin 5 mediates a peroxide-induced Ca2+ influx in B cells Helmut Kubista, Tim E. Hawkins, Darshana R. Patel, Harry T. Haigler and Stephen E. Moss Current Biology 22 November 1999, 9:1403–1406 Figure S1
(a)
Time (sec)
Time (sec) Annexin 2–/–
Ca2+ responses mediated by thapsigargin and the B-cell receptor are normal in cells lacking annexin 5. DT40 cells were loaded with fura-2 and stirred at 40oC in a cuvette in a spectrofluorimeter. (a) Thapsigargin (600 nM) and (b) anti-IgM (400 ng/ml) were added at 100 sec as indicated by the arrows, and fluorescence changes (∆R) were monitored continuously for a further 500 sec. Red trace, annexin 5–/– cells; green trace, annexin 2–/– cells.
(b)
Supplementary materials and methods
Labeling with 125I-TID Large unilammelar vesicles composed of phosphatidylserine and phosphatidylcholine (2:1 molar ratio) were prepared according to the Reeves/Dowben protocol [S2]. The topography of membrane-bound annexin 5 was probed with 125I-TID as in [S3,S4] and (J. Isas, J. Cartailler, Y. Sokolov, D.R.P., R. Langen, H. Luecke, J. Hall and H.T.H., unpublished data). Recombinant human annexin 5 (12 µg) was incubated with vesicles (1:1000 molar ratio protein:phospholipid) in a solution (30 µl final volume) containing either EGTA (500 µM) or CaC03 (500 µM) at the indicated pH. The following buffers (100 mM) were used: sodium acetate for pH 4.0–5.6; MES for pH 6.0–6.5; and Tris:HCl for pH 7.4. The 125I-TID solutions were diluted with ethanol to a concentration of 1 µCi/l. After incubating protein and phospholipid for 10 min in either the presence or absence of H2O2 (3 mM), 1 µCi 125I-TID was added and the solution was incubated for 20 min in the dark. The sample was then subjected to a 30 min irradiation from an ultraviolet light (Minerallight Lamp Model UVSL-25, 366 nm) at a distance of 5 cm to generate the reactive carbene. All incubations were at room temperature. After the incubation, Laemmli SDS sample buffer was added and the samples were analysed by SDS–PAGE and the dried gels were exposed to Phosphor Screens (Molecular Dynamics) and the data were analysed using Image Quant software.
Ca2+
1.6 1.4
EGTA
1.2 H2O2
1.0
Control (no H2O2)
0.8 0 100 200 300 400 500 600 Time (sec) 2.0 Annexin 5–/– 1.8 Ca2+ 1.6 1.4
EGTA
1.2 H2O2
1.0
Control (no H2O2)
0.8 0.6
Materials and cell culture (c) Relative [Ca2+]i (arbitrary units)
Source of materials as well as generation and culture of annexin gene knock-out DT40 clones was as or will be described elsewhere ([S1]; T.E.H. and S.E.M., unpublished data). Phosphatidylserine (bovine brain, catalogue #840032) and phosphatidylcholine (egg yolk, catalogue #840051) were obtained from Avanti Polar Lipids (Alabaster, Alabama). The photoactivatable probe 3-(trifluoromethyl)-3-(m[125I]iodophenyl)diazirine (125I-TID) was purchased from Amersham.
Annexin 2–/–
1.8
0.6
Annexin 5–/–
[Ca2+]i (arbitrary units)
Current Biology
2.0 [Ca2+]i (arbitrary units)
Anti-IgM
0 10 0 20 0 30 0 40 0 50 0
Thapsigargin
1.25 1.00 0.75 0.50 0.25 0.00 –0.25
Relative [Ca2+]i (arbitrary units)
(b)
0 10 0 20 0 30 0 40 0 50 0 60 0
Relative [Ca2+]i (arbitrary units)
(a) 1.25 1.00 0.75 0.50 0.25 0.00 –0.25
Figure S2
0 100 200 300 400 500 600 Time (sec) 1.0 Annexin 2–/– 0.8 0.6 0.4 0.2
Annexin 5–/–
0 250 350 450 550 650 Time (sec) Current Biology
H2O2-induced Ca2+ influx is lost in annexin 5–/– cells. EGTA (1 mM), H2O2 (2 mM) and Ca2+ (2.5 mM) were added as indicated to (a) annexin 2–/– cells (green trace) and (b) annexin 5–/– cells (red trace) in nominally Ca2+-free buffer. The black traces (control) in (a,b) were obtained in experiments on the respective cells by omission of H2O2. Ca2+ was added 250 sec after H2O2 stimulation and evoked pronounced intracellular Ca2+ elevations in annexin 2–/– cells compared with the control, but failed to do so in annexin 5–/– cells. (c) Responses obtained by Ca2+ re-addition at various time points (160, 250 and 360 sec) after H2O2 stimulation from annexin 2–/– cells (green traces) and annexin 5–/– cells (red traces) are shown superimposed. Ca2+ addition to the EGTA-containing buffer in these experiments was calculated to give a final free Ca2+ concentration of 1.5 mM. Experiments shown here were performed with buffer containing 250 µM sulfinpyrazone. This result supports the data in Figure 2e,f in the paper, showing that the defect in the response of annexin 5–/– cells to peroxide is due to a failure in Ca2+ influx.
Spectrofluorimetry All experiments were performed with an LS50 B fluorimeter (Perkin Elmer) at 40°C on stirred cell suspensions in HBSS (Hanks’ Balanced
Salts Solution), which was supplemented with HEPES, sodium bicarbonate and contained: 1.26 mM CaCl2, 0.81 mM MgSO4, 5.4 mM KCl,
S2
Supplementary material
0.44 mM KH2PO4, 136.9 mM NaCl, 0.34 mM Na2HPO4, 4.2 mM NaHCO3, 5.5 mM D-glucose and 10 mM HEPES. For Ca2+-free experiments, 3 mM EGTA was added. Agonists were added from concentrated stocks to obtain the desired concentrations. Fura-2 Ca2+ measurements of DT40 cells were performed as described in detail in a previous paper [S1]. In some experiments (as indicated in the figure legends), HBSS was supplemented with sulfinpyrazone (250 µM) to minimise loss of indicator dye. Transmembrane voltage was measured using bis-(1,3-dibutylbarbituric acid) trimethine oxonol (DiBAC4(3), Molecular Probes); 106 cells per ml were incubated for 10 min with 200 nM DiBAC4(3), which was added to the fluorimeter cuvette prior to the actual experiment. DiBAC4(3) fluorescence was measured with excitation at 490 nm and emission at 516 nm. Measurements of pH were performed by loading 106 cells per ml with 200 nM of the acetoxymethyl ester of 2′,7′-bis-(2-carboxyethyl)-5-(and-6)-carboxyfluorescin (BCECF, Molecular Probes) for 30 min at 40°C. After loading, cells were washed twice with HBSS, whereafter the fluorescence ratio (Em1/Em2) was determined by excitation 1 at 490 nm, excitation 2 at 440 nm and emission (Em) at 535 nm; pH was calibrated from the BCECF fluorescence ratio as described elsewhere [S5].
Manganese quench DT40 cells were loaded with Fura-2/AM as described for measurements of intracellular free Ca2+ [S1]. After loading, cells were washed and finally resuspended in HBSS containing 250 µM sulfinpyrazone. Suspensions of cells were excited at the dye’s isosbestic wavelength of 360 nm and emission was monitored at 510 nm; 0.5 mM Mn2+ was added as indicated to unstimulated cells or to cells 225 sec after stimulation with thapsigargin or H2O2.
Supplementary references S1. Kubista H, Hawkins T, Moss SE: Characterisation of calcium signalling in DT40 chicken B-cells. Biochim Biophys Acta 1998, 1448:299-310. S2. Reeves JP, Dowben RM: Formation and properties of thin-walled phospholipid vesicles. J Cell Physiol 1969, 73:49-60. S3. Brunner J, Semenza G: Selective labeling of the hydrophobic core of membranes with 3-(trifluoromethyl)-3-(m-[125I]iodophenyl) diazirine, a carbene-generating reagent. Biochemistry 1981, 20:7174-7182. S4. Hoppe J, Brunner J, Jorgensen BB: Structure of the membraneembedded F0 part of F1F0 ATP synthase from Escherichia coli as inferred from labeling with 3-(trifluoromethyl)-3-(M[125I]iodophenyl) diazirine. Biochemistry 1984, 23:5610-5616. S5. James-Kracke MR: Quick and accurate method to convert BCECF fluorescence to pHi: calibration in three different types of cell preparations. J Cell Physiol 1992, 151:596-603.