Calcium-Dependent Translocation of S100A11 Requires Tubulin ¢Laments

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Biochimica et Biophysica Acta 1498 (2000) 220^232 www.elsevier.com/locate/bba

Calcium-dependent translocation of S100A11 requires tubulin ¢laments

Gabriela E. Davey a, Petra Murmann a, Mathias Hoechli b, Toshio Tanaka c, Claus W. Heizmann a;*

a Department of Pediatrics, Division of Clinical Chemistry and Biochemistry, University of Zurich, Steinwiesstrasse 75, CH-8032 Zurich, Switzerland b Electron Microscopy Central Laboratory, University of Zurich, Gloriastrasse 30, CH-8032 Zurich, Switzerland c Department of Molecular and Cellular Pharmacology, Mie University, School of Medicine, Tsu, Mie 514-8507, Japan

Received 11 September 2000; accepted 12 September 2000

Abstract

Protein translocation between different subcellular compartments might play a significant role in various signal transduction pathways. The S100 family is comprised of the multifunctional, small, acidic proteins, some of which translocate in the form of vesicle-like structures upon increase in intracellular Ca2 levels. Previously, cells were fixed before and after calcium activation in order to examine the possible relocation of S100 proteins. In this study, we were able to track the real-time translocation. We compared the localization of endogenous S100A11 to that of the S100A11-green fluorescent protein. The application of thapsigargin, an agent increasing intracellular Ca2 levels, resulted in the relocation of the S100A11. In contrast, addition of EGTA, which specifically binds Ca2, either inhibited the ongoing process of translocation or prevented its induction. Since translocation was not affected by treatment with brefeldin A, it appears that S100A11 relocates in an endoplasmic reticulum-Golgi-independent pathway. Furthermore, the depolymerization of actin filaments by amlexanox did not affect the capacity of S100A11 to translocate. However, the time course treatment with demecolcine, which depolymerizes tubulin filaments, resulted in cease of translocation, suggesting that the tubulin network is required for this process. ß 2000 Elsevier Science B.V. All rights reserved.

Keywords: Calcium-binding protein; S100C; Calgizzarin; Translocation; Microtubule; Green £uorescent protein

1. Introduction 1q21, the region where most S100 genes cluster, has led several research groups to examine and implicate A growing number of studies suggest that S100 S100 proteins as possible markers in tumor grading. proteins are the possible e¡ectors in a variety of Di¡erent S100 proteins are involved in speci¢c types physiological processes such as in£ammation, al- of cancers. For example, S100A4 induces metastasis lergy, or cancer [1^6]. The frequency of tumor-asso- in breast cancer cells [7,8] and in non-small cell lung ciated rearrangement on the human chromosome cancer [9], while S100B is a marker for metastatic melanoma [10,11]. In fact, the S100A11 might also possess a substantial diagnostic value; it was reported to be signi¢cantly overexpressed in colorectal * Corresponding author. Fax: +41 (1) 2667169; cancer, as compared to normal colorectal mucosa, or E-mail: [email protected] in malignant melanomas of the uvea, with respect to

BBAMCR 14682 27-11-00 G.E. Davey et al. / Biochimica et Biophysica Acta 1498 (2000) 220^232 221 normal uveal melanocytes [12^14]. Zhao and co- 2.2. Cell culture workers showed that S100A11 inhibits actin-activated myosin ATPase activity [15]. Furthermore, Human glioblastoma cell lines U-373MG and U- S100A11 is a component of the keratinocyte corni- 87MG (HTB-17 and HTB-14, respectively, available ¢ed envelope [16,17], and it interacts with annexin-I from ATCC) were maintained in 10% FCS-DMEM in a Ca2-dependent manner [18^20]. Recent studies with penicillin-streptomycin. The transfection of demonstrate the importance of Ca2 activation in GFP constructs was accomplished using the lipofect- protein translocation [21^24]. Consequently, the pro- amine procedure. Cells were seeded at 2^2.5U105 per cess of translocation may supply additional levels of 35 mm well and incubated overnight. Plasmid DNA signaling control, by providing a rate-limiting step (1 Wg) was added to DMEM (100 Wl). In a separate with respect to induction of the cascades [25]. Anal- tube, the lipofectamine (5 Wl) reagent (Life Technol- ysis of expression levels, subcellular localization, and ogies) was added to DMEM (100 Wl). These solutions the interactions with target proteins is essential in were combined and incubated for 15^30 min before establishing the speci¢c roles for di¡erent S100 pro- being diluted to 1 ml in DMEM and added to the teins. In order to study and characterize the potential cells. After 5 h of incubation, a solution of 10% fetal translocation of the S100A11, we have prepared calf serum (FCS) in DMEM was added to each well S100A11-green £uorescent fusion proteins and trans- and cultures were incubated overnight. On the sec- fected them into human glioblastoma cell lines, ond day, the transfection medium was changed to which respond to thapsigargin with a Ca2 increase the normal medium. Cells were either analyzed as and exhibit high levels of translocation. Previously, transient transfections or after they were submitted green £uoroscent protein (GFP) was successfully to selection with G418. fused to other calcium binding proteins [26,27]. The transfectants were maintained in the medium, Modi¢ed for brighter £uorescence and higher expres- which additionally contained G418. They were pas- sion in mammalian cells, it has become a great tool saged at 80^90% con£uence. To examine localization for the monitoring of gene expression and protein and translocation of the S100 endogenous and re- localization [28^30]. combinant fusion proteins by confocal microscopy, the cells (2^5U104) were plated on coverslips in 24-well dishes, 24 h prior to the experiment. The cells 2. Materials and methods to be stimulated were rinsed with DMEM, and then incubated for 5^90 min with thapsigargin (100 nM^1 2.1. Preparation of S100A11-GFP constructs WM) in the stimulation bu¡er (140 mM NaCl, 5 mM KCl, 1 mM MgCl2, 10 mM glucose, and 10 mM In order to fuse the S100A11 to GFP (enhanced HEPES, supplemented with 1 mM CaCl2) at 37³C. form, modi¢ed for brighter £uorescence, Clontech), In order to assess the viability of cells after pro- we designed primers which included a linker consist- longed treatment with thapsigargin (Sigma), cells ing of ¢ve glycines, 20 bases of the S100A11-coding (2.5U105) were seeded in 35 mm dishes (in tripli- sequence, and restriction sites compatible with the cate), incubated with thapsigargin for up to 90 min, GFP-containing vectors (XhoI-BamHI for pEGFP- trypsinized, stained with trypan blue and counted. N1 or BglII-SalI for pEGFP-C1), followed by an In order to examine translocation of S100A11 in- additional three bases. The primers were incorpo- side live cells, 5^10U105 cells expressing fusion pro- rated into S100A11 cDNA in the correct reading teins were seeded in 35 mm glass bottom dishes (Ma- frame by PCR. The modi¢ed sequences were di- tek) and treated with di¡erent drugs. In order to gested, puri¢ed, and ligated to the corresponding establish a time course for S100A11 translocation vectors. DH5K were transformed with the aliquots following the application of thapsigargin, cells were from the ligation mixtures. Plasmid DNA was puri- exposed to the drug while being monitored by micro- ¢ed from the resulting colonies, and the constructs scope time lapse photography. In order to determine were analyzed by restriction digestion and DNA se- whether sequestering of Ca2 could inhibit the trans- quencing. location, cells were incubated with 10 mM EGTA

BBAMCR 14682 27-11-00 222 G.E. Davey et al. / Biochimica et Biophysica Acta 1498 (2000) 220^232

(which speci¢cally binds Ca2). A time course was In order to mount slides, a small drop of mounting derived in order to establish at what point the po- medium (Molviol, Hoechst) containing 0.75% tential e¡ect occurs. In order to determine whether n-propyl-gallate as an anti-bleaching agent was the translocation of S100 proteins is linked to the placed on the slide. Mounted slides were left to dry endoplasmic reticulum (ER)-Golgi-dependent path- for 24 h at room temperature and stored in the dark way, cells were treated with brefeldin A, which re- at 4³C until viewed. versibly blocks this route (0.5^5 Wg/ml, Calbiochem), and analyzed by microscope time lapse photography. 2.4. Confocal laser scanning and wide-¢eld In order to resolve whether the relocation of S100 microscopy proteins is dependent on actin cytoskeleton or microtubules, amlexanox (Takeda Chemical Industries), The samples with ¢xed cells were analyzed using a which reversibly collapses actin stress ¢bers (1 WM), Zeiss Axioplan £uorescence microscope (under 63U or demecolcine (Sigma), which reversibly disrupts the oil objective lens) equipped with a confocal scanning tubulin ¢laments (1 WM), was added to the cells. unit MRC-600 (Bio-Rad) and an argon-krypton laser Similarly, a time course was derived. with an excitation wavelength of 568 nm for Cy3, and 488 nm for enhanced GFP. Subsequently, the 2.3. Immuno£uorescent staining images were processed using Graphic Converter (version 3.8) and Photoshop (Adobe System, version 5.5) In order to stain cells with anti-S100A11, they software on a Macintosh computer. were rinsed with DMEM and ¢xed with 3.7% form- The samples with live cells were analyzed by time aldehyde in DMEM for 15 min at 37³C. Cells were lapse photography, using a wide-¢eld Leica DM permeabilized with methanol for 10 min at room IRBE microscope equipped with a Hamamatsu cam- temperature, washed in 5% horse serum (HS)- era controller (C4742-95) and connected to the Mac- DMEM, and incubated with a peptide anti-human intosh computer. The imaging and recording were S100A11 (dilution 1/500), as described [31], for 1 h operated by Open Lab software, version 2.1.1. The at 37³C. As a control, cells were incubated with the 25^50 images (depending on the quickness of the antibodies pre-absorbed with S100A11 antigen. After cellular response) were recorded one every 4 s; cells were washed twice for 5 min in 5% HS-DMEM, each exposure was set between 50 and 300 ms (de- they were incubated with Cy3-conjugated anti-rabbit pending on the intensity of the signal, but constant in IgG (H+L) (goat, Jackson Immuno-Research Labo- each series) using a set ¢lter for GFP (Leica), and ratories) for 45 min at 37³C. Cells were washed twice neutral density ¢lters to minimize bleaching. The ob- for 5 min in 5% HS-DMEM, and once in PBS (pH tained images were processed using Photoshop 9). (Adobe System, version 5.5) or converted into quick The cells to be stained with the monoclonal Cy3- time ¢les using a public domain NIH Image program conjugated anti-actin (mouse IgG2a isotype, recog- (developed at the US National Institutes of Health nizes an epitope located at the C-terminus, which and available on the Internet at http://rsb.info.nih. is conserved in all actin isoforms, dilution 1/50, Sig- gov/nih.-image/). ma) or anti-L-tubulin (mouse IgG1 isotype, localizes L-tubulin in most species, including human, dilution 1/100, Sigma) were ¢xed with formaldehyde (3.7% 3. Results in PBS with Ca2/Mg2, at room temperature for 20^30 min) and washed twice with PBS. The ¢xed 3.1. S100A11-GFP constructs and their subcellular cells were treated with NH4Cl (50 mM in PBS), localization washed with PBS, and permeabilized with Triton X-100 (0.1% in PBS) as described [32]. Cells were The modi¢ed S100A11 insert contained a ¢ve gly- incubated with antibodies against actin or tubulin cine linker, which allowed for more rotational free- for 30 min and washed 4 times with PBS and once dom between the S100A11 protein and GFP. The with water. S100A11 fused to the N-terminus of GFP (through

BBAMCR 14682 27-11-00 G.E. Davey et al. / Biochimica et Biophysica Acta 1498 (2000) 220^232 223

Fig. 1. Preparation of S100A11-GFP constructs. S100A11 cDNA was modi¢ed by PCR and then ligated into the corresponding vec- tor. (a) S100A11 was fused to the N-terminus of GFP, thus its stop codon was eliminated and the ¢ve glycine linker was added at the 3P end. (b) S100A11 was fused to the C-terminus of the GFP, thus its start codon was removed and the ¢ve glycine linker was added at the 5P end. Compatible restriction sites were added at the 5P and 3P ends of both primers, followed by an additional three base pairs, which allowed for e¤cient restriction digestion of the modi¢ed insert.

Fig. 2. Comparison of the localization of the endogenous S100A11 and the A11N or A11C fusion proteins in U-373MG cells. (a) Cells were ¢xed and stained with anti-S100A11 and Cy3-conjugated goat anti-rabbit IgG (b: negative control, antibody pre-absorbed with S100A11 antigen). (c,d) Cells were transfected with A11N construct (c) or with A11C construct (d). All samples were analyzed by laser-scanning confocal microscopy.

BBAMCR 14682 27-11-00 224 G.E. Davey et al. / Biochimica et Biophysica Acta 1498 (2000) 220^232 ligation to pEGFP-N1, Fig. 1a) was named A11N with either construct were ¢xed and directly analyzed and that to the C-terminus of GFP (through ligation (Fig. 2c,d). The staining of the fusion proteins was to pEGFP-C1, Fig. 1b) was named A11C. We trans- similar to that of the endogenous S100A11 and fected the parental cells 6^7 times with the prepared mostly appeared in the nucleus and perinuclear area. constructs. The transfection e¤ciency was about 50^70%. The viability of transient transfectants was 3.2. S100A11 relocates in a Ca2+-dependent manner s 90%. The experiments were performed within a couple of days after transfection. Also, the transfec- In order to determine the e¡ects of intracellular tants were selected and maintained in G418. The Ca2 increase, we stimulated cells for 10 min with Western blot of the total cell extracts showed that thapsigargin (0.5 WM, a non-phorbol ester, depleting 2 the levels of endogenous S100A11 were una¡ected in IP3-sensitive Ca stores of the ER, and thus increas- the transfectants. The expression of S100A11-GFP ing cytosolic Ca2) [33^35]. Then we ¢xed and fusion protein did not have a notable e¡ect on cell stained cells with anti-S100A11 and the Cy3-conju- growth rates or morphology. Transient and stable gated secondary antibody. We observed vesicle-like transfectants gave similar results. structures localized apart from where most of the The localization of A11N and A11C was com- signal was concentrated. Scanning of the cell at the pared to the endogenous expression. The untrans- di¡erent Z-levels using a confocal microscope re- fected cells were ¢xed, stained with anti-S100A11 vealed that S100A11 was relocated toward the pe- or the pre-absorbed antibody (Fig. 2a,b), and then riphery of the cell (Fig. 3a^d). In order to verify viewed by confocal microscopy. The cells transfected these events in real time, we seeded the A11N or

Fig. 3. Translocation of the S100A11 cells after thapsigargin treatment analyzed by confocal microscopy. The U-87MG cells were treated for 10 min with 0.5 WM thapsigargin, followed by ¢xing and staining with anti-S100A11 and Cy3-conjugated anti-rabbit IgG. The cell was laser scanned at di¡erent depths from cell top to cell bottom (a^d). The arrows point to the vesicles, which translocated from the nucleus to the periphery of the cell.

BBAMCR 14682 27-11-00 G.E. Davey et al. / Biochimica et Biophysica Acta 1498 (2000) 220^232 225

Fig. 4. Ca2-dependent S100A11 translocation cells analyzed by microscope time lapse photography. The U-87MG cells transfected with A11N were seeded on glass bottom plates. After addition of thapsigargin (0.5 WM), a relocation of the S100A11 was observed already 1 min after treatment. In panel b (1.5 min) and c (2.5 min), formation and relocation of new vesicles are marked by arrows. (d) A cell undergoing translocation is treated with 10 mM EGTA. (e) 1.5 min after addition of EGTA the size and the number of vesicles sharply decreased. (f) 2.5 min after treatment any noticeable translocation ceased.

BBAMCR 14682 27-11-00 226 G.E. Davey et al. / Biochimica et Biophysica Acta 1498 (2000) 220^232

A11C transfectants on the glass bottom plates and viewed them using wide-¢eld microscopy. Most of the experiments were performed using both constructs in the two cell lines (transient and stable transfections), yielding very similar results. The ¢g- ures shown are examples of the most representative data. Initial observation revealed that a small number of the S100A11-GFP positive cells was undergoing spontaneous S100A11 translocation. In fact, this process would stop in one cell and begin in another in 10^20% of the population at any time point. Ad- dition of 0.5 WM thapsigargin, monitored by time lapse (50 frames, one every 4 s), caused relocation of S100A11 in 80^90% of the cells (within 1^2 min), suggesting that this process occurs normally and can be intensi¢ed by an increase in the intracellular Ca2 concentration (Fig. 4a^c). The levels of total S100A11 (endogenous and fusion proteins) were higher in the transfected cells, but this did not a¡ect the translocation process. In un- transfected cells stimulated with thapsigargin, we observed formation and relocation of similar vesicles, as in the cells transfected with the fusion protein. Transient and stable transfectants gave similar results with respect to translocation. Viability of cells after 0, 15, 45, and 90 min of treatment with 1 WM thapsigargin assessed by staining of the triplicate samples with trypan blue, showed on average no increase in cell death. (Note that the concentration and the length of treatment used for the viability test were increased as compared to cell stimulation.) When cells undergoing S100A11 translocation were treated with 10 mM EGTA, which speci¢cally chelates Ca2, the translocation process was com- pletely inhibited within 2 min (Fig. 4d^f). There was no decrease in cell viability after 60 min treatment with EGTA. Although the initial Ca2 peak is due to calcium release from the intracellular stores, 2 2 the continuous elevation in Ca levels requires Ca Fig. 5. S100A11 translocation independent of the ER-Golgi in£ux, which can be e¡ectively blocked by chelation classical pathway. The U-373MG cells transfected with A11N with EGTA [36^38]. In fact, Losavio and Muchnik were seeded on glass bottom plates. After addition of brefeldin [39] found that EGTA or the permeant BAPTA A(5Wg/ml) cells were monitored by microscope time lapse pho- yielded similar overall results. Pretreatment of un- tography. After 45 min, cells were treated with thapsigargin (a^ stimulated cells with EGTA signi¢cantly reduced c). The arrows show the formation or relocation of the vesicles with respect to the previous frame. the response of cells to thapsigargin; however, if the EGTA-containing medium was replaced with normal medium, cells resumed the process of trans-

BBAMCR 14682 27-11-00 G.E. Davey et al. / Biochimica et Biophysica Acta 1498 (2000) 220^232 227 location (data not shown). Therefore, it appears that volved in the process of translocation, the cells the induction and the inhibition of translocation are were incubated with amlexanox [1], an anti-allergy both Ca2-dependent and reversible. drug which reversibly depolymerizes actin cytoskeleton, or demecolcine, which reversibly depolymerizes 3.3. Translocation of S100A11 is independent of the microtubules. We derived the time course guidelines ER-Golgi classical pathway based on the depolymerization kinetics of actin or tubulin ¢laments. We have treated cells with corre- In order to determine whether the observed trans- sponding drugs for di¡erent time periods (0, 30, 60, location depends on the ER-Golgi pathway, the cells and 120 min), ¢xed them, and analyzed them by expressing fusion proteins were seeded on the glass confocal microscopy. Untreated cells are shown in bottom plates and incubated with brefeldin A, which Fig. 6a,c. Amlexanox had already a noticeable e¡ect blocks the classical translocation/secretion pathway. after 30 min, which was even more conspicuous after After addition of up to 5 Wg/ml of brefeldin A 60 min (Fig. 6b) and lasted trough 120 min. Deme- [40,41], cells were monitored by time lapse (50 colcine appeared to have a very pronounced e¡ect frames, one every 4 s) at 15 min, 30 min, and after 30 min (Fig. 6d), which lasted for 60 min. How- 45 min, at which point the cells were treated with ever, the microtubules began to slowly re-polymerize 0.5 WM thapsigargin (for selected frames see Fig. after 2 h. 5a^c). This led to an even more intensi¢ed activity. We used cells undergoing spontaneous transloca- A time course lasting for up to 2 h showed no in- tion to make this assessment more physiological. The hibitory e¡ect on translocation. cells were treated with 1 WM amlexanox and recorded a time lapse (25 frames, one every 4 s) at 3.4. S100A11 translocation depends on tubulin 15 min, 45 min, and 120 min, at which point the cells network, but not on actin ¢lament were also stimulated with thapsigargin (for selected frames see Fig. 7a^c). Throughout the time course of In an e¡ort to identify cellular components in- the experiment, the treatment of cells with amlexa-

Fig. 6. Amlexanox and demecolcine depolymerize actin and tubulin ¢laments, respectively. Cells were seeded on glass coverslips in a 24-well dish and left untreated (a,c) or incubated with 1 WM amlexanox for 60 min (b) or 1 WM demecolcine for 30 min (d). Then the cells were ¢xed and stained with Cy3-conjugated anti-actin (a,b) or anti-tubulin (c,d). The samples were analyzed by confocal microscopy.

BBAMCR 14682 27-11-00 228 G.E. Davey et al. / Biochimica et Biophysica Acta 1498 (2000) 220^232

nox had no noticeable e¡ect on S100A11 translocation. In order to accentuate the translocation we added thapsigargin, which intensi¢ed the activity. However, when cells undergoing translocation of S100A11, monitored by time lapse (50 frames, one every 4 s, for selected frames see Fig. 8a,b), were incubated with 1 WM demecolcine, the process of translocation was notably inhibited in less than 30 min (Fig. 8c). Furthermore, stimulation of cells with thapsigargin at 45 min did not rescue cells (Fig. 8d^f). However, the removal of demecolcine would result in restoration of the microtubules and translocation. Hence, the translocation of S100A11 specifically requires tubulin ¢laments and is independent of the actin cytoskeleton or cell shape.

4. Discussion

The indisputable relationship between the deregu- lated expression of S100 proteins and cancer, Alz- heimer's disease, or in£ammation strongly suggests their profound involvement in these processes [1^6]. However, despite the overwhelming correlation between the S100 proteins and di¡erent pathological processes, many questions remain unanswered. Hence, the analysis of the cellular processes in which they are involved should provide a step toward de- picting their signaling pathways. Previously, in order to examine possible relocation of the S100 proteins, cells were ¢xed before and after calcium activation, o¡ering limited £exibility with respect to drug ma- nipulation and con¢ning analysis to the general ef- fects on the cell population. Our system allows for tracking of the real-time translocation and determin- ing the outcome produced by di¡erent drugs in the same cell. We have found that reversible translocation oc- curred in unstimulated glioblastoma cells which al- 2 Fig. 7. S100A11 translocation is independent of actin ¢laments. ready exhibit high Ca levels, suggesting that The U-87MG cells transfected with A11N were seeded on glass unknown Ca2 triggers regulate subcellular localiza- bottom plates, treated with 1 WM amlexanox and monitored by tion of S100A11 at di¡erent times. However, thapsi- microscope time lapse photography. At 120 min, cells were gargin, which increases cytosolic Ca2, e¡ectively in- stimulated with 0.5 WM thapsigargin (a). (b) A cell treated for duced or intensi¢ed this process, con¢rming earlier 1.5 min. (c) A cell was incubated for 3 min. Arrows mark the change in vesicle formation. ¢ndings [22,23]. These results are supported by the reversal of translocation by EGTA, a speci¢c Ca2 chelator. The fact that in resting cells S100 proteins are localized to the speci¢c cellular compartments

BBAMCR 14682 27-11-00 G.E. Davey et al. / Biochimica et Biophysica Acta 1498 (2000) 220^232 229

Fig. 8. Tubulin-dependent S100A11 translocation. U-87MG cells transfected with A11N were seeded on glass bottom plates, treated with demecolcine (1 WM), and monitored by microscope time lapse photography. The di¡erences in vesicle appearance between panel a (directly after addition of the drug) and panel b (1.5 min later) are marked by arrows. Panel c shows the disappearance of translocation after 30 min. (d) Cells were stimulated with thapsigargin (0.5 WM), which did not induce translocation 1.5 min later (e) or 3 min later (f).

BBAMCR 14682 27-11-00 230 G.E. Davey et al. / Biochimica et Biophysica Acta 1498 (2000) 220^232 from which they relocate upon Ca2 activation sug- network might be the primary component interacting gests that the translocation might be the key mecha- with the S100 proteins, which regulates their trans- nism in the assembly of signaling complexes, regulat- location, implying a novel route, possibly related to ing it in a temporal and spatial manner. Ca2- the alternative secretion pathway utilized by cyto- dependent association and targeting of S100A11 to kines [45]. early endosomes by annexin I [18^20] could feasibly Moreover, phosphorylation of S100 proteins ap- depend on translocation. Similarly, the Ca2-depen- pears to be another mechanism regulating these pro- dent inhibition of actin-activated myosin ATPase [15] teins. Vogl et al. [46] found that S100A9 through might be linked to S100A11 relocation. phosphorylation by MAP kinase p38 might control Considering that Ca2 homeostasis is often monocyte transmigration. A recent report by Saka- a¡ected in cancer cells, it is not surprising that guchi et al. suggests that in normal con£uent cells the Ca2-binding S100 proteins may serve as vehicles S100A11 becomes phosphorylated on Thr 10, disso- for abnormal Ca2 signaling. Perhaps the anomalous ciates from actin ¢laments, and becomes translocated (or poorly timed) elevation of Ca2 concentration to the nucleus, where it upregulates p21 and p16 could lead to the increased translocation of the leading to growth arrest. They found that phosphor- S100 proteins and the undesired interaction with ylation was required for nuclear transport, but it was target proteins, which in turn may potentiate the not necessary for growth inhibition. Although they pathological condition. The scheme presented by reported that in immortalized cells, which they exam- Meyer and Shen [25], in which translocation in nor- ined, S100A11 was not phosphorylated or trans- mal cells provides ¢ne tuning to the signaling cas- ported to the nucleus, we found high levels of cades, supports this model. The regulation of the S100A11 in the nuclear and perinuclear areas of number of signaling proteins at one site by a con- the glioblastoma cells. Furthermore, earlier studies trolled release from another site prevents formation have shown overexpression of S100A11 in colorectal of signaling complexes in the absence of su¤cient cancer and ocular melanoma [12^14]. Perhaps, one stimulus. way to resolve these seemingly con£icting results is to On the other hand, the observed translocation view the S100A11-mediated growth arrest as a multi- seems to occur in a distinct manner. Brefeldin A, step process in which S100A11 phosphorylation/ which blocks the classical ER-Golgi secretion path- translocation is one of the many points that can be way, does not seem to have any e¡ect on the trans- targeted in cellular transformation. Conceivably, in location of S100A11, which is in agreement with some transformed cells, S100A11 can still be trans- Rammes et al. [40] reporting that S100A8/A9 are located to the nucleus, but another signal contribut- secreted independently of this route. This would be ing to induction of p16 and p21 expression is absent. expected since S100 proteins lack a signal sequence In fact, the cell may attempt without success to com- for secretion via the classical pathway. In an attempt pensate for lack of contact inhibition by upregulating to identify the cellular components involved in the the S100A11 expression, as shown in di¡erent can- translocation of S100A11, we have collapsed the ac- cers. On the other hand, S100A11 could have several tin cytoskeleton or tubulin network. The depolymer- di¡erent functions. For example, another member of ization of actin ¢laments did not inhibit S100A11 the S100 family, S100B, was found to be a glial mi- translocation. In fact, Sakaguchi et al. showed that togen, but under special conditions it can induce ap- dissociation of S100A11 from actin is a prerequisite optosis. In fact, S100B was recently reported to in- for its transport to the nucleus [42]. However, the teract with S100A11 in a manner similar to those of microtubules were essential for this process. Roth target proteins [47]. Furthermore, the observed local- and co-workers [40] have demonstrated the involve- ization of vesicles containing S100A11 at the cell ment of microtubules in the secretion of the S100A8/ membrane implies the possibility of secretion. A9, while Donato's group has shown a co-localiza- Although this is not certain, the recent discovery of tion of S100B and microtubules [43] and the possible RAGE, a receptor for S100A12 and S100B [2], sup- S100B-mediated disassembly of tubulin in situ [44]. ports the extracellular role for the S100 protein fam- Combining these results suggests that the tubulin ily.

BBAMCR 14682 27-11-00 G.E. Davey et al. / Biochimica et Biophysica Acta 1498 (2000) 220^232 231

Acknowledgements [16] N.A. Robinson, S. Lapic, J.F. Welter, R.L. Eckert, J. Biol. Chem. 272 (1997) 12035^12046. This work was supported by Swiss National Sci- [17] N.A. Robinson, R.L. Eckert, J. Biol. Chem. 273 (1998) 2721^2728. ence Foundation. We would like to thank Takeda [18] W.S. Mailliard, H.T. Haigler, D.D. Schlaepfer, J. Biol. Chemical Industries Ltd. (Osaka, Japan) for the gen- Chem. 271 (1996) 719^725. erous supply of amlexanox. We are very grateful to [19] J. Seemann, K. Weber, V. Gerke, FEBS Lett. 413 (1997) Thomas Ba«chi for making possible the use of con- 185^190. focal and wide-¢eld microscopy and Ernst Niggli for [20] S. Rety, D. Osterloh, J.P. Arie, S. Tabaries, J. Seeman, F. Russo-Marie, V. Gerke, A. Lewit-Bentley, Struct. Fold Des. the calcium measurements. We thank Johann Peter 8 (2000) 175^184. Hossle, Beat Scha«fer, and Karl Frei for their contri- [21] C. Kerkho¡, M. Klempt, C. Sorg, Biochim. Biophys. Acta butions. 1448 (1998) 200^211. [22] A. Mandinova, D. Atar, B.W. Scha«fer, M. Spiess, U. Aebi, C.W. Heizmann, J. Cell Sci. 111 (1998) 2043^2054. [23] A. Mueller, T. Ba«chi, M. Hochli, B.W. Scha«fer, C.W. Heiz- mann, Histochem. Cell Biol. 111 (1999) 453^459. References [24] R.J. Passey, K. Xu, D.A. Hume, C.L. Geczy, J. Leukocyte Biol. 66 (1999) 549^556. [1] C.M. Carreira, T.M. LaVallee, F. Tarantini, A. Jackson, [25] T. Meyer, K. Shen, Trends Cell Biol. 10 (2000) 238^244. J.T. Lathrop, B. Hampton, W.H. Burgess, T. Maciag, [26] A. Miyawaki, J. Llopis, R. Heim, J.M. McCa¡ery, J.A. J. Biol. Chem. 273 (1998) 22224^22231. Adams, M. Ikura, R.Y. Tsien, Nature 388 (1997) 882^887. [2] M.A. Hofmann, S. Drury, C. Fu, W. Qu, A. Taguchi, Y. Lu, [27] N. Garamszegi, Z.P. Garamszegi, M.S. Rogers, S.J. DeMar- C. Avila, N. Kambham, A. Bierhaus, P. Nawroth, M.F. co, E.E. Strehler, Biotechniques 23 (1997) 864^866, 868^870, Neurath, T. Slattery, D. Beach, J. McClary, M. Nagashima, 872. J. Morser, D. Stern, A.M. Schmidt, Cell 97 (1999) 889^ [28] B.P. Cormack, R.H. Valdivia, S. Falkow, Gene 173 (1996) 901. 33^38. [3] T. Shishibori, Y. Oyama, O. Matsushita, K. Yamashita, H. [29] E. Oancea, T. Meyer, Cell 95 (1998) 307^318. Furuichi, A. Okabe, H. Maeta, Y. Hata, R. Kobayashi, [30] N.J. Hack, B. Billups, P.B. Guthrie, J.H. Rogers, E.M. Biochem. J. 338 (1999) 583^589. Muir, T.N. Parks, S.B. Kater, J. Neurosci. Methods 95 [4] C.W. Heizmann, J.A. Cox, Biometals 11 (1998) 383^397. (2000) 177^184. [5] R. Donato, Biochim. Biophys. Acta 1450 (1999) 191^231. [31] H. Inada, M. Naka, T. Tanaka, G. Davey, C.W. Heizmann, [6] C.W. Heizmann, B.W. Scha«fer, Encycl. Mol. Med. (2000) in Biochem. Biophys. Res. Commun. 263 (1999) 135^138. press. [32] X. Zhu, K. Roovers, G. Davey, R.K. Assoian, in: J.-L. [7] M. Pedrocchi, B.W. Scha«fer, H. Mueller, U. Eppenberger, Guan (Ed.), Signaling through the Cell Adhesion Molecules, C.W. Heizmann, Int. J. Cancer 57 (1994) 684^690. CRC Press, Boca Raton, FL, 1999, pp. 129^140. [8] A.M. Platt-Higgins, C.A. Renshaw, C.R. West, J.H. Win- [33] O. Thastrup, P.J. Cullen, B.K. Drobak, M.R. Hanley, A.P. stanley, S. De Silva Rudland, R. Barraclough, P.S. Rudland, Dawson, Proc. Natl. Acad. Sci. USA 87 (1990) 2466^ Int. J. Cancer 89 (2000) 198^208. 2470. [9] K. Kimura, Y. Endo, Y. Yonemura, C.W. Heizmann, B.W. [34] A. Diarra, R. Sauve, P£ug. Arch. 422 (1992) 40^47. Scha«fer, Y. Watanabe, T. Sasaki, Int. J. Oncol. 16 (2000) [35] J. Gromada, T.D. Jorgensen, S. Dissing, FEBS Lett. 360 1125^1131. (1995) 303^306. [10] A.J. Cochran, H.F. Lu, P.X. Li, R. Saxton, D.R. Wen, Mel- [36] J. Hartmann, A. Verkhratsky, J. Physiol. 513 (1998) 411^ anoma Res. 3 (1993) 325^330. 424. [11] R. Boni, G. Burg, A. Doguoglu, E.C. Ilg, B.W. Scha«fer, B. [37] A. Losavio, S. Muchnik, Life Sci. 66 (2000) 2543^2556. Muller, C.W. Heizmann, Br. J. Dermatol. 137 (1997) 39^43. [38] P. Meine, A. Teti, F. Pugliese, G.A. Cincotti, Kidney Int. 46 [12] M. Tanaka, K. Adzuma, M. Iwami, K. Yoshimoto, Y. Mon- (1994) 122^128. den, M. Itakura, Cancer Lett. 89 (1995) 195^200. [39] K. Tornquist, J. Cell. Physiol. 150 (1992) 90^98. [13] P.R. Van Ginkel, R.L. Gee, T.M. Walker, D.N. Hu, C.W. [40] A. Rammes, J. Roth, M. Goebeler, M. Klempt, M. Hart- Heizmann, A.S. Polans, Biochim. Biophys. Acta 1448 (1998) mann, C. Sorg, J. Biol. Chem. 272 (1997) 9496^9502. 290^297. [41] M. Jareb, G. Banker, J. Neurosci. 17 (1997) 8955^8963. [14] J. Stulik, K. Koupilova, J. Osterreicher, J. Knizek, A. Ma- [42] M. Sakaguchi, M. Miyazaki, Y. Inoue, T. Tsuji, H. Kouchi, cela, J. Bures, P. Jandik, F. Langr, K. Dedic, P.R. Jungblut, T. Tanaka, K. Yamada, M. Namba, J. Cell Biol. 149 (2000) Electrophoresis 20 (1999) 3638^3646. 1193^1206. [15] X.Q. Zhao, M. Naka, M. Muneyuki, T. Tanaka, Biochem. [43] G. Sorci, A.L. Agneletti, R. Bianchi, R. Donato, Biochim. Biophys. Res. Commun. 267 (2000) 77^79. Biophys. Acta 1448 (1998) 277^289.

BBAMCR 14682 27-11-00 232 G.E. Davey et al. / Biochimica et Biophysica Acta 1498 (2000) 220^232

[44] G. Sorci, A.L. Agneletti, R. Donato, VI European Sympo- Sorg, J. Roth, VI European Symposium on Calcium Binding sium on Calcium Binding Proteins in Normal and Trans- Proteins in Normal and Transformed Cells, Paris, 2000, formed Cells, Paris, 2000, Abstr. P93. Abstr. P46. [45] A. Muesch, E. Hartmann, K. Rohde, A. Rubartelli, R. Sitia, [47] J.C. Deloulme, N. Assard, G.C. Ouengue Mbele, J. Baudier, T.A. Rapoport, Trends Biochem. Sci. 15 (1990) 86^88. VI European Symposium on Calcium Binding Proteins in [46] T. Vogl, M.P. Manitz, S. Ludwig, A. Strey, M. Goebler, C. Normal and Transformed Cells, Paris, 2000, Abstr. P28.

BBAMCR 14682 27-11-00