Ca2+/calmodulin-dependent protein kinase II inhibition by heparin in mesangial cells
http://www.100md.com
《美国生理学杂志》
Department of Laboratory Medicine and Pathobiology, University of Toronto, Toronto, Ontario, Canada
ABSTRACT
Heparin exerts an antiproliferative effect in smooth muscle cells, and the Ca2+/calmodulin-dependent protein kinase (CaMK) signaling pathway is heparin sensitive. Here, we report that transfection with a truncated 326-amino acid fragment of CaMK-II increases basal activity of CaMK-II in mesangial cells. Ionomycin increased CaMK-II activity in both transfected and untransfected cells, with a concomitant increase in activated Ca2+/calmodulin. Heparin (1 μg/ml), but not chondroitin or dermatan sulfate, significantly attenuated both serum- or ionomycin-induced CaMK-II activity, and attendant c-fos mRNA expression, but did not affect upstream Ca2+/calmodulin. Autophosphorylation of Thr286 generates an autonomously active CaMK-II. Both serum and ionomycin increased phosphorylation at this site and increased CaMK-II activity in antiphosphothreonine immunoprecipitates. Heparin (1 μg/ml) did not inhibit phosphorylation of Thr286 (although much higher concentrations did). Replacement of Thr286 with Asp produces a constitutively active mutant that was insensitive to ionomycin but was inhibited by heparin maximally at 1 μg/ml. These results suggest that heparin at physiological concentrations acts at or downstream of CaMK-II to suppress its activity independent of an effect on autophosphorylation.
calmodulin kinase II; phosphorylation; c-fos; phosphothreonine; Ca2+ ionophore
PROLIFERATION OF VASCULAR smooth muscle cells is a prominent feature of atherosclerosis and of restenosis following angioplastic injury (32, 34). Similarly, proliferation of the smooth muscle-like mesangial cell is central to experimental and clinical glomerulonephritis (3, 15, 33). Thus both vascular smooth muscle and mesangial cells are potential targets for antiproliferative therapeutic intervention, and both have occasioned numerous studies on factors involved in regulating their proliferation. Heparin is well known to inhibit smooth muscle cell proliferation in vivo and under cell culture conditions (4, 6, 29, 41) and to suppress smooth muscle cell migration (16, 17, 28). These effects of exogenously administered heparin may reflect physiological properties of endogenous heparan sulfate molecules (17, 37), providing further impetus for studying the mechanisms involved. Numerous signal transduction mechanisms have been implicated as potentially contributing to the antiproliferative phenomenon (7, 21, 29, 30), but no single definitive mechanism has been established. Both protein kinase C-dependent MAPK signaling and MAPK-independent signals associated with cell cycle entry and immediate-early response gene induction are inhibited by heparin in mesangial cells (21). Among non-MAPK signals shown to be heparin sensitive, Ca2+/calmodulin-dependent protein kinase (CaMK) activity has been studied in cultured rat mesangial (20) and aortic smooth muscle (23) cells.
The CaMKs are a family of calmodulin-dependent kinases with a variety of substrates that serve to integrate Ca2+ signaling within the cell (8, 13, 14). Calmodulin binds Ca2+ to form the active Ca2+/calmodulin complex. Four CaMKs have been described (13, 14); CaMK-II is unique in being autoregulatory. On binding Ca2+/calmodulin, it undergoes autophosphorylation in a complex that then retains activity, until dephosphorylated, independently of calmodulin. This calmodulin-independent activity is referred to as autonomous activity. CaMK-II has been identified in mesangial cells by Western blotting (40), and it has been linked to c-fos induction in several cell lines (1, 24), including mesangial (40) and vascular smooth muscle (23) cells. It may thus be important for unwanted proliferation in these cells, and indeed it has been found that heparin does not inhibit CaMK-II in vascular smooth muscle cells selected for heparin resistance (i.e., that continue to proliferate in the presence of heparin) (23). Further linking CaMK-II to proliferation are observations that its inhibition suppresses cyclin D1 expression (25) and blocks cells in late G1 (25, 36) and that it is required for progression of oocytes through G2/M (19).
We demonstrated that heparin suppresses autonomous CaMK activity and CaMK signaling to induce c-fos in mesangial cells (20). However, heparin added to an in vitro CaMK assay was without effect (20), indicating that its effects on the CaMK pathway are mediated through other cellular components or at a level other than direct inhibition of autonomous CaMK-II. Furthermore, heparin was without effect on expression of total CaMK kinase activity elicited in vitro by addition of excess Ca2+ and calmodulin to cell extracts, suggesting inhibition is not at the level of calmodulin. Mishra-Gorur et al. (23) confirmed inhibition of CaMK-II activity in vascular smooth muscle cells and demonstrated a decreased phosphorylation of CaMK-II immunoprecipitates in the presence of heparin. Using phosphatase inhibitors, they proposed activation of protein phosphatases PP-1 and PP-A2 by heparin (23). Indeed, activation of phosphatases is an interesting hypothesis to account for heparin's suppressive effects on multiple, disparate kinase pathways. However, we recently showed that the decrease in serum-stimulated Erk MAPK activity that occurs in the presence of heparin is not dependent on activity of the Erk regulatory phosphatases (43). In particular, activity of the MAPK phosphatase-1 (MKP-1), a primary regulator of Erk in smooth muscle cells (9, 18), is actually decreased in the presence of heparin, apparently secondarily to a decrease in Erk-dependent MKP-1 induction (43). Therefore, if a unifying mechanism exists to explain heparin's effects on multiple kinases, it may lie elsewhere than in phosphatase activation. The present study was undertaken, in part, to determine the role decreased phosphorylation may play in heparin's suppression of CaMK-II activity and c-fos induction.
MATERIALS AND METHODS
Materials. FBS, heparin (bovine lung, lot no. 49H0648 used in all experiments), chondroitin and dermatan sulfates, cAMP, phosphodiesterase (PDE), trifluoperazine, and malachite green were all obtained from Sigma (St. Louis, MO). RPMI-1640 cell culture medium, trypsin, Trizol reagent, and LipoFectamine 2000 were from Invitrogen Canada (Burlington, Ontario). NuSerum IV (BD Biosciences, Mississauga, Ontario) was used as the stimulus wherever serum stimulation of quiescent cell cultures is referred to. Calmodulin, ionomycin, and the CaMK-II inhibitor KT-5926 were products of Calbiochem (San Diego, CA). Autocamtide-2 was from Bachem (Torrance, CA). ECL Plus Western blotting kits, nitrocellulose, and Hybond-N membranes were purchased from Amersham Biosciences (Baie D'urfé, Quebec). SR plasmids containing either the coding sequence for the first 326 amino acids of the native sequence of CaMK-II, designated CaMK-II326 and containing regulatory and catalytic domains, or coding a constitutively active mutant of 326 amino acids with a T286D replacement, were kindly provided by Dr. Howard Schulman (11). Anti-CaMK-II antibody (mouse monoclonal IgG1, cat. no. 05–532) was from Upstate (Charlottesville, VA). Anti-phospho-Thr antibody was from Santa Cruz Biotechnology (Santa Cruz, CA) and an antibody to phospho-Thr286-CaMK-II (rabbit polyclonal, cat. no. 44–674G) was a product of BioSource (Camarillo, CA).
Cell culture and transfection. Rat mesangial cell cultures were established and characterized as previously described (38). They were maintained in RPMI 1640 medium with 10% FBS, in a humidified atmosphere of 5% CO2 at 37°C, and subcultured by trypsinization. Experiments were performed on cells between passages 5 and 15, as described previously (20). For transfection, cells were passaged at a split ratio of 1:3 and grown to 80% confluence. Four micrograms of plasmid DNA per 10-cm plate (either empty vector SR, vector coding CaMK-II326, or the T286D mutant) were diluted with serum- and antibiotic-free RPMI-1640 and incubated for 5 min before addition of the lipofectamine reagent at a ratio of DNA:lipofectamine of 1:2–1:3 (wt/wt). The mixture was incubated for 15 min at room temperature and then overlaid onto prewashed mesangial cell layers. After incubation at 37°C for 5–6 h, the cells were returned to 10% FBS medium and grown overnight before serum deprivation (0.4% for 48 h) and subsequent stimulation with NuSerum or ionomycin.
Western blotting. For detection of total CaMK-II, cells were washed with chilled phosphate-buffered saline and lysed by two freeze-thaw cycles in 50 mM HEPES buffer containing 0.5% Nonidet P-40, containing protease and phosphatase inhibitors (1 mM Na3VO4, 25 mM NaF, 1 mM PMSF, and 1 μg/ml each of aprotinin, leupeptin, and pepstatin). They were then sonicated twice for 5 s at 200 W and centrifuged (13,400 g, 15 min). Cell extracts were separated by 12% SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes. Membranes were blocked with 5% nonfat dry milk in TBS and blotted overnight with 1 μg/ml anti-CaMK-II at 4°C. After being washed twice with water, membranes were blotted with horseradish peroxidase-conjugated rabbit anti-mouse IgG for 90 min at room temperature, washed twice with water, three times with TBS containing 0.01% Tween 20, and visualized with ECL Plus reagent according to the manufacturer's protocol.
For phospho-Thr286-CaMK-II, cells were lysed in modified RIPA buffer (10 mM Tris, pH 7.4, with 100 mM NaCl, 1 mM EDTA, 1 mM EGTA, 20 mM Na4P2O7, 0.1% SDS, 0.5% sodium deoxycholate, 1% Triton X-100, 10% glycerol, 1 mM PMSF, 60 μg/ml aprotinin, 10 μg/ml leupeptin, 1 μg/ml pepstatin, 1 mM NaF, and 2 mM Na3VO4). Thirty micrograms of total protein were separated by 10% PAGE, transfered to polyvinyl difluoride membrane, and blocked in TBS with 0.1% Tween and 5% BSA. Overnight incubation with a 1:1,000 dilution of anti-phospho-Thr286 was followed by incubation with a 1:7,500 dilution of secondary antibody and visualization with ECL Plus reagent. Dephosphorylation was achieved by incubating cell lysates (30 μg total protein prepared without vanadate) with 400 units of phosphatase in phosphatase buffer (50 mM Tris, pH 7.5, with 1 mM EDTA, 5 mM DTT, and 2 mM MnCl2) for 30 min at 30°C before 10% SDS-PAGE.
Detection of c-fos mRNA expression by Northern blotting. Total RNA was isolated with Trizol reagent and 10 μg of total RNA were loaded on a 1.2% agarose/2.2 M formaldehyde gel, transferred to Hybond-N membrane overnight, and hybridized with a [-32P]dCTP-labeled c-fos fragment as before (42). The hybridization solution contained 50% formamide, 10% dextran sulfate, and 1% SDS. Membranes were washed with 2x SSC/0.1% SDS twice at room temperature for 10 min, twice at 50°C for 15 min, and then in 0.1x SSC/0.1% SDS twice at 50°C for 30 min before exposure to Kodak MS X-ray film.
Kinase activity assay. Autonomous CaMK activity and total activity were measured as described previously (20). Briefly, total cell proteins were harvested in lysis buffer as above for Western blotting and centrifuged (13,400 g, 15 min). For autonomous activity, 5–10 μg of lysate protein were incubated at 30°C for 3 min in 25 μl of 50 mM HEPES assay buffer containing 10 mM MgCl2, 0.1 mM ATP, 1 mM EGTA, 0.02 mM autocamtide peptide (KKALRRQETVDAL) as a substrate, and 5 μCi/ml [-32P]ATP. For total activity, EGTA was replaced by 4 mM CaCl2 and 2 μM calmodulin. Reactions were stopped by addition of trichloroacetic acid to 5% and spotted on P81 phosphocellulose filters, washed with 75 mM H3PO4, and counted by liquid scintillation. Immunoprecipitation of phosphorylated protein for kinase assay was carried out with an anti-phospho-Thr antibody. Cell extracts were prepared as for CaMK Western blots and 200 μg of total protein were incubated with 1 μg antibody overnight a 4°C. After being incubated with 100 μl of a 50% slurry of protein G (2 h, 4°C), the pellets were resuspended in equal volume of 2x kinase buffer, incubated at 30°C for 3 min, and loaded on P81 filter paper as above.
Calmodulin activity assay. We adapted a PDE-linked enzyme assay that has been developed for in vitro assessment of calmodulin activity during purification from mucus (10), as this assay has good sensitivity and specificity (5). To prepare samples, treated cells were placed in 100 μl 0.9% NaCl during freeze-thaw cycles, then scraped into 1.5-ml tubes, boiled at 100°C for 5 min, placed on ice for 10 min, and centrifuged at 13,400 g for 15 min at 4°C.
The calmodulin activity assay uses calmodulin-dependent PDE to convert cAMP to AMP with subsequent liberation of inorganic phosphate by alkaline phosphatase. Phosphate is then measured spectrophotometrically (A650) after reaction with malachite green as described (10, 27). Five micrograms of total cell protein were preincubated with 2 mU of PDE and 1.5 U alkaline phosphatase in 0.8 ml assay buffer without Ca2+ [40 mM Tris/40 mM imidazole (pH 7.4), 1 mM Mg(CH3COO)2] for 1 min at 30°C. After addition of 1 mM cAMP, the assay mixtures were incubated for a further 30 min at 30°C. Fifty-microliter aliquots of assay mixtures were then added to 900 μl molybdic acid reagent and mixed with 25 μl malachite green solution (1.26 mg/ml). After 2 min, the absorption was measured at 650 nm. Calmodulin activity was calculated based on a standard curve prepared with activated calmodulin.
The response of the malachite green detection component of the assay was linear over a range of potassium phosphate from 0.5 to 3 μg, allowing standard curves to be constructed with purified calmodulin that were linear (r2 > 0.95) over 0 to 125 ng calmodulin and an absorbance range (A650 nm) of 0 to 0.6. Conditions for the initial reaction with cytosolic extracts were chosen to give a final absorption in the range 0.2 to 0.4 (typically, 5 μg cell protein). Under these conditions, addition of excess, exogenous Ca2+ caused an approximate doubling of the signal, presumably representing activation of total cellular calmodulin, and the following controls all gave an absorption of less than 0.03: omission of PDE, omission of alkaline phosphatase, omission of cAMP with or without added Ca2+, and inclusion of EGTA. Furthermore, the Ca2+-calmodulin antagonist trifluoperazine (1 mM) suppressed the calmodulin standard curve to background absorbance, and standard curves were corrected for background by subtracting values in the presence of trifluoperazine.
Statistical analyses. Statistical differences between groups within an experiment were analyzed by one-way ANOVA with Tukey's post hoc test for pairwise comparisons.
RESULTS
As in our previous studies (20), treatment of quiescent rat mesangial cells with either serum or ionomycin resulted in a significant increase in autonomous CaMK activity (Fig. 1) that consistently reached about 15–20% of total activity (data not shown). This autonomous activity was partially inhibited by the CaMK-II inhibitor KT-5926 and by heparin (1 μg/ml). Endogenous CaMK-II was only weakly detected by Western blotting (Fig. 2), and to further confirm inhibition of CaMK-II by heparin, we transfected cells with wild-type CaMK-II326. The CaMK-II326 protein was strongly expressed (Fig. 2), resulted in a higher basal level of CaMK activity in quiescent cells, and was activated by ionomycin (Fig. 1). Again, both KT-5926 and heparin inhibited ionomycin-dependent activity in the CaMK-II326 transfectants. Mock transfection with empty SR plasmid did not affect basal CaMK activity or activation of endogenous CaMK-II by serum or ionomycin, compared with untransfected cells (compare black and open bars, Fig. 1).
c-fos mRNA is elevated 30 min after ionomycin treatment of quiescent mesangial cells (Fig. 3). Consistent with c-fos induction through CaMK-II (20), c-fos mRNA levels show the same pattern of response to KT-5926 and heparin as does autonomous CaMK activity; significant inhibition of c-fos expression occurs with each agent, and heparin decreases c-fos mRNA levels to the basal level observed in quiescent cells. Basal c-fos expression is increased in cells transfected with either empty plasmid or with plasmid containing CaMK-II326. The c-fos mRNA levels in quiescent cells transfected with the empty plasmid are not significantly affected by ionomycin or inhibitors (data not shown). However, ionomycin increases c-fos mRNA levels further in CaMK-II326-transfected cells, and heparin decreases these levels to below basal values (Fig. 3).
To determine whether heparin acts at a level proximal to CaMK in this signaling pathway, we measured calmodulin activity in control, transfected, and heparin-treated cells. Use of the phosphodiesterase-linked assay with cytosolic extracts is with caveats such as potential activation by Ca2+ release during cell extraction or interference from cellular inorganic phosphate (see DISCUSSION). Nevertheless, we were able to implement the assay in a reproducible manner that allows certain conclusions to be drawn. Stimulation of quiescent cells with serum or ionomycin caused increases in active Ca2+-calmodulin from 12.2 ng/μg cytosolic protein to 18.0 and 19.0 ng/μg, respectively (Table 1). Inclusion of heparin (1 μg/ml) with ionomycin was without effect (18.7 ng/μg). Similarly, heparin did not affect Ca2+-calmodulin activation by ionomycin in either mock transfected cells or cells transfected with CaMK-II326 (Table 1). Thus the phosphodiesterase-linked assay measures an increase in calmodulin activation in both control and transfected cells that is unresponsive to heparin; the influence of heparin is downstream of calmodulin activation.
View this table:
An initial event in CaMK-II activation following upstream activation of calmodulin is autophosphorylation on Thr286. We used a phospho-Thr286-specific antibody to examine whether 1 μg/ml heparin interfered with this event. Because higher concentrations of heparin have been reported to decrease net CaMK-II phosphorylation (23), we also compared the effects of 100 μg/ml heparin (Fig. 4). Treatment with phosphatase demonstrates that the antibody recognizes primarily a phosphorylated form of the protein. Heparin at 1 μg/ml is ineffective at decreasing ionomycin-dependent phosphorylation of Thr286. On the other hand, 100 μg/ml heparin significantly decreases phosphorylation, as previously observed (23).
To further examine the effects of heparin on phosphorylation, we used the alternative approach of measuring CaMK activity in immunoprecipitates prepared with an antibody directed against phosphothreonine. Despite inclusion of phosphatase inhibitors in the extraction solution, we were unable to achieve reproducible or fully autonomous activity in the immunoprecipitated fractions (data not shown), and we conclude that activated CaMK-II may be unstable under the conditions of immunoprecipitation. As an alternative indicator of CaMK-II phosphorylation, we measured total CaMK activity elicited by addition of Ca2+ and calmodulin in anti-phospho-Thr immunoprecipitates (Table 2). Ionomycin stimulation increased the CaMK-II activity of anti-phospho-Thr immunoprecipitates by about twofold. Heparin (1 μg/ml) was without significant effect, supporting the conclusion that this level of heparin inhibits CaMK-II activity independently of altering its level of phosphorylation. Consistent with decreased phosphorylation in the presence of 100 μg/ml heparin, this concentration of heparin decreased recovery of activity in the immunoprecipitate by about 45% (Table 2).
View this table:
Because higher concentrations of heparin reduced phosphorylation of Thr286 and decreased recovery of CaMK-II activity in anti-phospho-Thr immunoprecipitates, we compared the effects of 1 and 100 μg/ml on c-fos induction. The higher concentration did not give a significantly greater decrease in ionomycin-dependent c-fos induction than the lower (Fig. 5). We previously showed that 1 μg/ml heparin is sufficient to cause near-maximal suppression of [3H]thymidine incorporation in stimulated mesangial cells (38).
As a further approach to studying the role of phosphorylation in heparin's effect on CaMK signaling, we used a constitutively active phosphomimetic mutant of CaMK-II326 in which Thr286 is replaced with Asp, designated CaMK-II(T286D). Cells transfected with this mutant expressed 110 ± 40% (n = 7) basal CaMK activity compared with untransfected cells stimulated with serum (Fig. 6A). This was not significantly increased by ionomycin nor was it decreased by the CaMK-II inhibitor KT-5926, consistent with the autonomous activity reported for this modified protein (11). However, the activity was significantly decreased by 1 μg/ml heparin. Furthermore, quiescent cells expressing CaMK-II(T286D) had their basal c-fos mRNA levels decreased to 66.5 ± 17.4% by 15 min of exposure to 1 μg/ml heparin (Fig. 6B). Calmodulin activity was unaffected (Table 1). The 1-μg/ml concentration of heparin apparently produces maximal suppression in this model. In three further independent experiments comparing 1 and 100 μg/ml heparin in CaMK-II(T286D)-transfected cells, heparin again caused a decrease to 63.9 ± 7.7% of basal c-fos expression, and no further decrease was seen with 100 μg/ml (Fig. 7).
The effects observed here also raise the question of specificity. The closely related galactosamine-based glycosaminoglycan chondroitin sulfate had no effect on ionomycin-induced c-fos expression, nor did the even more cosely related iduronic acid-rich dermatan sulfate (Fig. 8).
DISCUSSION
CaMK-II is a broad-specificity calmodulin-dependent kinase with homology to, but distinct from, CaMK-I and CaMK-IV (14). The assembled CaMK-II complex (8, 14) consists of 10–12 subunits representing four different isoforms (, , , and ), each of which occurs in multiple splice variants. Each has an inhibitory regulatory domain that interacts with the catalytic domain; this interaction is disrupted by Ca2+/calmodulin to expose the active site. The -subunit is highly conserved among many species. In contrast to other calmodulin-dependent enzymes, CaMK-II has autophosphorylation sites in the regulatory domain whose phosphorylation breaks the inhibitory interaction of the catalytic domain and confers autonomous activity, i.e., activity independent of Ca2+ or calmodulin (13, 14, 35). In particular, Thr286 of CaMK-II is essential for autonomous activity; when it is mutated to leucine, the kinase remains ligand (calmodulin) dependent (11), like myosin light chain kinase (MLCK). Mutation of Thr286 to the phosphomimetic aspartate confers constitutive activity of about 10–20% that of native autonomous protein (2), as exploited in the present study. Thus phosphorylation of Thr286 is both necessary and sufficient for CaMK-II ativity. Although phosphorylation of other Thr residues occurs, notably Thr305 and Thr306, this is inhibitory and prevents activation of the kinase by binding of Ca2+/calmodulin (12). Among the numerous substrates of CaMK-II are transcription factors such as cAMP response element-binding protein (CREB), and it is likely through phosphorylation of this factor that its activation induces c-fos (31). Indeed, ionomycin-induced phosphorylation of CREB on Ser133 is accompanied by c-fos induction in mesangial cells (42).
We previously suggested that the rapid response of Erk and CaMK signaling to heparin supports a mechanism of action from the cell surface (20, 21, 38). For instance, internalization of negatively charged heparin by mesangial cells is slow (hours), compared with suppression of a CaMK-II signal 30 s after addition to cultured cells of heparin together with a stimulus such as ionomycin (20). This is not due to interference with ionomycin-dependent Ca2+ signaling, which is unaffected by heparin in these circumstances (20). And the effect shows some specificity for heparin; the closely related glycosaminoglycans, chondroitin and dermatan sulfates, are ineffective in suppressing signaling to c-fos. This points to a mechanism of competing signaling, or perhaps to disruption of proximal events at the cell surface, rather than direct interaction of heparin with downstream components of the kinase cascades. Furthermore, because autonomous activity assayed in vitro is decreased in extracts of cells that had previously been exposed to heparin, the inhibition of CaMK-II might involve events at, or upstream of, enzyme assembly and autophosphorylation. However, the present results appear to rule out this explanation.
In the present study, we could obtain no evidence for an effect of heparin on formation of active Ca2+/calmodulin. This result must be qualified by the limitations of the assay. Release of Ca2+ from stores during cell disruption could mask true levels of active calmodulin. Alternatively, Ca2+ could dissociate from calmodulin prior to assay. However, we do measure an increase in activity after treatment with ionomycin, which is sensitive to the calmodulin antagonist trifluoperazine. This suggests that activation can be measured with our assay, and the response is not affected by heparin. This is consistent with the earlier observation that heparin does not affect angiotensin-mediated mesangial cell contraction (20), a process also operating through calmodulin-dependent activation of MLCK. Together with ionomycin-dependent phosphorylation of Thr286 in the presence of 1 μg/ml heparin, these observations support a mechanism of action of heparin independent of calmodulin.
Activation of phosphatase(s) was proposed as an attractive mechanism whereby heparin might negatively regulate diverse kinase cascades (22), potentially at multiple sites. Subsequently, Mishra-Gorur et al. (23) presented evidence for decreased phosphorylation of the CaMK-II subunit in vascular smooth muscle labeled with [32P]phosphate and then treated with serum containing 100 or 500 μg/ml heparin (49 and 68% decrease, respectively, compared with heparin-free controls). Using several phosphatase inhibitors, which prevented decreased phosphorylation of CaMK-II otherwise occurring in the presence of heparin, they provided evidence for a role of protein phosphatases PP-1 and PP-2A. However, we recently showed that 1 μg/ml heparin did not suppress Erk activation by upregulating its phosphatase, MKP-1, but rather decreased MKP-1 secondarily to effects on Erk (43). In our mesangial cell system, this more physiological concentration of heparin shows near-maximal suppression of [3H]thymidine incorporation (38) and c-fos induction (Fig. 5). Therefore, we investigated in the present study whether dephosphorylation of CaMK-II was necessary for the effect of heparin at the lower concentration. Although high amounts of heparin did indeed decrease the amount of phospho-CaMK-II, 1 μg/ml did not, suggesting that at least some of heparin's antimitogenic activity is not due to effects on protein phosphatases. We also found that immunoprecipitates of phospho-Thr phosphoproteins contained increased amounts of CaMK-II activity following ionomycin stimulation, that was decreased by 100 μg/ml heparin (Table 2), again consistent with the findings of Mishra-Gorur et al. (23). Heparin at 1 μg/ml, on the other hand, was without effect on recovery of activity, indicating that the lower concentration of heparin did not affect total Thr phosphorylation. We can conclude that decreased phosphorylation is not necessary for suppression of CaMK-II activity, and indeed heparin treatment also decreases the constitutive activity of the phosphomimetic mutant T286D, which is not dependent on phosphorylation (at least of Thr286) for enzyme activity. As in control cells, 1 μg/ml heparin is as effective as 100 μg/ml in decreasing activity of the phosphomimetic T286D mutant. We conclude that heparin must elicit signals that disrupt or prevent an active form of the protein from occurring whether Thr286 is phosphorylated or not. Because the suppressed activity persists in the assay carried out in vitro in the absence of heparin, it is reasonable to suggest that heparin may rapidly mobilize an inhibitory factor, or alternatively disrupt association with an activating component. The existence of such factors is speculative but suggests models for the more general effects of heparin on signaling that should be evaluated. In this context, it is interesting to note that Src has also been implicated in induction of c-fos through the serum response element, downstream of CaMK-II signaling in mesangial cells (39), although the nature of the CaMK-II/Src interaction is unknown.
The inhibitor KT-5926 was initially introduced as an inhibitor of MLCK but actually has a lower Ki for CaMK-II (0.004 μM) than for MLCK (0.018 μM). It was shown to act directly at the catalytic site and competitively with ATP (26). Therefore, it is expected to inhibit CaMK-II326 as seen in Fig. 1. That it is not as effective at inhibiting c-fos induction in these transfected cells as it is in untransfected cells (Fig. 3) suggests heparin-sensitive pathways additional to CaMK may be activated in the transfected cells. Alternatively, KT-5926 could have differing inhibitory potencies for autocamtide-2 used as a substrate in assaying autonomous activity and as an endogenous substrate (e.g., CREB) that differ in native protein and truncated recombinant protein. This possibility has not been addressed.
In conclusion, heparin appears to elicit signals that suppress CaMK-II activity independently of phosphorylation of Thr286, although high concentrations of heparin may indeed decrease autophosphorylation of the protein. Activation of calmodulin by Ca2+ is not affected by heparin, and other calmodulin-dependent proteins such as MLCK may also be unaffected. Whether heparin exploits common mechanisms in regulating CaM kinases and MAP kinases remains to be determined.
GRANTS
This work was supported by operating grant NA-3520 from the Heart and Stroke Foundation of Ontario.
FOOTNOTES
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
REFERENCES
Antoine M, Gaiddon C, and Loeffler JP. Ca2+/calmodulin kinase types II and IV regulate c-fos transcription in the AtT20 corticotroph cell line. Mol Cell Endocrinol 120: 1–8, 1996.
Brickey DA, Bann JG, Fong YL, Brennan RG, and Soderling TR. Mutational analysis of the autoinhibitory domain of calmodulin kinase II. J Biol Chem 269: 29047–29054, 1994.
Carl M, Akagi Y, Weidner S, Isaka Y, Imai E, and Rupprecht HD. Specific inhibition of Egr-1 prevents mesangial cell hypercellularity in experimental nephritis. Kidney Int 63: 1302–1312, 2003.
Castellot JJ Jr and Karnovsky MJ. Heparin and the regulation of growth of the vascular wall. In: Vascular Smooth Muscle in Culture, edited by Campbell JH and Campbells GR. Boca Raton, FL: CRC, 1987, p. 93–115.
Chock SP and Huang CY. An optimized continuous assay for cAMP phosphodiesterase and calmodulin. Anal Biochem 138: 34–43, 1984.
Clowes AW and Clowes MM. Inhibition of smooth muscle cell proliferation by heparin molecules. Transplant Proc 21: 3700–3701, 1989.
Daum G, Hedin U, Wang YX, Wang T, and Clowes AW. Diverse effects of heparin on mitogen-activated protein kinase-dependent signal transduction in vascular smooth muscle cells. Circ Res 81: 17–23, 1997.
De Koninck P and Schulman H. Sensitivity of CaM kinase II to the frequency of Ca2+ oscillations. Science 279: 227–230, 1998.
Duff JL, Monia BP, and Berk BC. Mitogen-activated protein (MAP) kinase is regulated by the MAP kinase phosphatase (MKP-1) in vascular smooth muscle cells. Effect of actinomycin D and antisense oligonucleotides. J Biol Chem 270: 7161–7166, 1995.
Flik G, Van Rijs JH, and Wendelaar Bonga SE. Evidence for the presence of calmodulin in fish mucus. Eur J Biochem 138: 651–654, 1984.
Hanson PI, Kapiloff MS, Lou LL, Rosenfeld MG, and Schulman H. Expression of a multifunctional Ca2+/calmodulin-dependent protein kinase and mutational analysis of its autoregulation. Neuron 3: 59–70, 1989.
Hanson PI and Schulman H. Inhibitory autophosphorylation of multifunctional Ca2+/calmodulin-dependent protein kinase by site-directed mutagenesis. J Biol Chem 267: 17216–17224, 1992.
Hook SS and Means AR. Ca2+/CaM-dependent kinases: from activation to function. Annu Rev Pharmacol Toxicol 41: 471–505, 2001.
Hudmon A and Schulman H. Neuronal Ca2+/calmodulin-dependent protein kinase II: the role of structure and autoregulation in cellular function. Annu Rev Biochem 71: 473–510, 2002.
Johnson RJ. The glomerular response to injury: progression or resolution Kidney Int 45: 1769–1782, 1994.
Kohno M, Yokokawa K, Yasunari K, Minami M, Kano H, Mandal AK, and Yoshikawa J. Heparin inhibits human coronary artery smooth muscle cell migration. Metabolism 47: 1065–1069, 1998.
Koyama N, Kinsella MG, Wight TN, Hedin U, and Clowes AW. Heparan sulfate proteoglycans mediate a potent inhibitory signal for migration of vascular smooth muscle cells. Circ Res 83: 305–313, 1998.
Lai K, Wang H, Lee WS, Jain MK, Lee ME, and Haber E. Mitogen-activated protein kinase phosphatase-1 in rat arterial smooth muscle cell proliferation. J Clin Invest 98: 1560–1567, 1996.
Lorca T, Cruzaleugi FH, Fesquet D, Cavadore JC, Mery J, Means A, and Doree M. Calmodulin-dependent protein kinase II mediates inactivation of the MPF and CSF upon fertilization of Xenopus eggs. Nature 366: 270–273, 1993.
Miralem T and Templeton DM. Heparin inhibits Ca2+/calmodulin-dependent kinase II activation and c-fos induction in mesangial cells. Biochem J 330: 651–657, 1998.
Miralem T, Wang A, Whiteside CI, and Templeton DM. Heparin inhibits mitogen-activated protein kinase-dependent and -independent c-fos induction in mesangial cells. J Biol Chem 271: 17100–17106, 1996.
Mishra-Gorur K and Castellot JJ Jr. Heparin rapidly and selectively regulates protein tyrosine phosphorylation in vascular smooth muscle cells. J Cell Physiol 178: 205–215, 1999.
Mishra-Gorur K, Singer HA, and Castellot JJ Jr. Heparin inhibits phosphorylation and autonomous activity of Ca2+/calmodulin-dependent protein kinase II in vascular smooth muscle cells. Am J Pathol 161: 1893–1901, 2002.
Misra RP, Bonni A, Miranti CK, Rivera VM, Sheng M, and Greenberg ME. L-type voltage-sensitive calcium channel activation stimulates gene expression by a serum response factor-dependent pathway. J Biol Chem 269: 25483–25493, 1994.
Morris TA, DeLorenzo RJ, and Tombes RM. CaMK-II inhibition reduces cyclin D1 levels and enhances the association of p27kip1 with cdk2 to cause G1 arrest in NIH 3T3 cells. Exp Cell Res 240: 218–227, 1998.
Nakanishi S, Yamada K, Iwahashi K, Kuroda K, and Kase H. KT5926, a potent and selective inhibitor of myosin light chain kinase. Mol Pharmacol 37: 482–488, 1990.
Penney CL. A simple microassay for inorganic phosphate. Anal Biochem 75: 201–210, 1976.
Pukac L, Huangpu J, and Karnovsky MJ. Platelet-derived growth factor-BB, insulin-like growth factor-I, and phorbol ester activate different signaling pathways for stimulation of vascular smooth muscle cell migration. Exp Cell Res 242: 548–560, 1998.
Pukac LA, Carter JE, Ottlinger ME, and Karnovsky MJ. Mechanisms of inhibition by heparin of PDGF stimulated MAP kinase activation in vascular smooth muscle cells. J Cell Physiol 172: 69–78, 1997.
Pukac LA, Ottlinger ME, and Karnovsky MJ. Heparin suppresses specific second messenger pathways for protooncogene expression in rat vascular smooth muscle cells. J Biol Chem 267: 3707–3711, 1992.
Rosen LB, Ginty DD, and Greenberg ME. Calcium regulation of gene expression. In: Advances in Second Messenger and Phosphoprotein Research, edited by Meanss AR. New York: Raven, 1995, vol. 30, p. 225–253.
Ross R. Cell biology of atherosclerosis. Annu Rev Physiol 57: 791–804, 1995.
Saleh H, Schlatter E, Lang D, Pauels HG, and Heidenreich S. Regulation of mesangial cell apoptosis and proliferation by intracellular Ca2+ signals. Kidney Int 58: 1876–1884, 2000.
Schwartz SM, deBlois D, and O'Brien ERM. The intima. Soil for atherosclerosis and restenosis. Circ Res 77: 445–465, 1995.
Soderling TR. The Ca2+-calmodulin-dependent protein kinase cascade. Trends Biochem Sci 24: 232–236, 1999.
Tombes RM, Grant S, Westin EH, and Krystal G. G1 cell cycle arrest and apoptosis are induced in NIH 3T3 cells by KN-93, an inhibitor of CaMK-II. Cell Growth Differ 6: 1063–1070, 1995.
Wang A, Fan MY, and Templeton DM. Growth modulation and proteoglycan turnover in cultured mesangial cells. J Cell Physiol 159: 295–310, 1994.
Wang A and Templeton DM. Inhibition of mitogenesis and c-fos expression in mesangial cells by heparin and heparan sulfates. Kidney Int 69: 437–448, 1996.
Wang Y, Mishra R, and Simonson MS. Ca2+/calmodulin-dependent protein kinase II stimulates c-fos transcription and DNA synthesis by a Src-based mechanism in glomerular mesangial cells. J Am Soc Nephrol 14: 28–36, 2003.
Wang Y and Simonson MS. Voltage-insensitive Ca2+ channels and Ca2+/calmodulin-dependent protein kinases propogate signals from endothelin-1 receptors to the c-fos promoter. Mol Cell Biol 16: 5915–5923, 1996.
Wright TC Jr, Castellot JJ Jr, Diamond JR, and Karnovsky MJ. Regulation of cellular proliferation by heparin and heparan sulphate. In: Heparin: Chemical and Biological Properties, Clinical Applications, edited by Lane DA and Lindahls U. Boca Raton, FL: CRC, 1989, p. 295–316.
Zeng H, Liu Y, and Templeton DM. Ca2+/calmodulin-dependent and cAMP-dependent kinases in induction of c-fos in human mesangial cells. Am J Physiol Renal Physiol 283: F888–F894, 2002.
Zhao Y, Xiao W, and Templeton DM. Suppression of mitogen-activated protein kinase phosphatase-1 (MKP-1) by heparin in vascular smooth muscle cells. Biochem Pharmacol 66: 769–776, 2003.(Weiqun Xiao, Ying Liu, an)
ABSTRACT
Heparin exerts an antiproliferative effect in smooth muscle cells, and the Ca2+/calmodulin-dependent protein kinase (CaMK) signaling pathway is heparin sensitive. Here, we report that transfection with a truncated 326-amino acid fragment of CaMK-II increases basal activity of CaMK-II in mesangial cells. Ionomycin increased CaMK-II activity in both transfected and untransfected cells, with a concomitant increase in activated Ca2+/calmodulin. Heparin (1 μg/ml), but not chondroitin or dermatan sulfate, significantly attenuated both serum- or ionomycin-induced CaMK-II activity, and attendant c-fos mRNA expression, but did not affect upstream Ca2+/calmodulin. Autophosphorylation of Thr286 generates an autonomously active CaMK-II. Both serum and ionomycin increased phosphorylation at this site and increased CaMK-II activity in antiphosphothreonine immunoprecipitates. Heparin (1 μg/ml) did not inhibit phosphorylation of Thr286 (although much higher concentrations did). Replacement of Thr286 with Asp produces a constitutively active mutant that was insensitive to ionomycin but was inhibited by heparin maximally at 1 μg/ml. These results suggest that heparin at physiological concentrations acts at or downstream of CaMK-II to suppress its activity independent of an effect on autophosphorylation.
calmodulin kinase II; phosphorylation; c-fos; phosphothreonine; Ca2+ ionophore
PROLIFERATION OF VASCULAR smooth muscle cells is a prominent feature of atherosclerosis and of restenosis following angioplastic injury (32, 34). Similarly, proliferation of the smooth muscle-like mesangial cell is central to experimental and clinical glomerulonephritis (3, 15, 33). Thus both vascular smooth muscle and mesangial cells are potential targets for antiproliferative therapeutic intervention, and both have occasioned numerous studies on factors involved in regulating their proliferation. Heparin is well known to inhibit smooth muscle cell proliferation in vivo and under cell culture conditions (4, 6, 29, 41) and to suppress smooth muscle cell migration (16, 17, 28). These effects of exogenously administered heparin may reflect physiological properties of endogenous heparan sulfate molecules (17, 37), providing further impetus for studying the mechanisms involved. Numerous signal transduction mechanisms have been implicated as potentially contributing to the antiproliferative phenomenon (7, 21, 29, 30), but no single definitive mechanism has been established. Both protein kinase C-dependent MAPK signaling and MAPK-independent signals associated with cell cycle entry and immediate-early response gene induction are inhibited by heparin in mesangial cells (21). Among non-MAPK signals shown to be heparin sensitive, Ca2+/calmodulin-dependent protein kinase (CaMK) activity has been studied in cultured rat mesangial (20) and aortic smooth muscle (23) cells.
The CaMKs are a family of calmodulin-dependent kinases with a variety of substrates that serve to integrate Ca2+ signaling within the cell (8, 13, 14). Calmodulin binds Ca2+ to form the active Ca2+/calmodulin complex. Four CaMKs have been described (13, 14); CaMK-II is unique in being autoregulatory. On binding Ca2+/calmodulin, it undergoes autophosphorylation in a complex that then retains activity, until dephosphorylated, independently of calmodulin. This calmodulin-independent activity is referred to as autonomous activity. CaMK-II has been identified in mesangial cells by Western blotting (40), and it has been linked to c-fos induction in several cell lines (1, 24), including mesangial (40) and vascular smooth muscle (23) cells. It may thus be important for unwanted proliferation in these cells, and indeed it has been found that heparin does not inhibit CaMK-II in vascular smooth muscle cells selected for heparin resistance (i.e., that continue to proliferate in the presence of heparin) (23). Further linking CaMK-II to proliferation are observations that its inhibition suppresses cyclin D1 expression (25) and blocks cells in late G1 (25, 36) and that it is required for progression of oocytes through G2/M (19).
We demonstrated that heparin suppresses autonomous CaMK activity and CaMK signaling to induce c-fos in mesangial cells (20). However, heparin added to an in vitro CaMK assay was without effect (20), indicating that its effects on the CaMK pathway are mediated through other cellular components or at a level other than direct inhibition of autonomous CaMK-II. Furthermore, heparin was without effect on expression of total CaMK kinase activity elicited in vitro by addition of excess Ca2+ and calmodulin to cell extracts, suggesting inhibition is not at the level of calmodulin. Mishra-Gorur et al. (23) confirmed inhibition of CaMK-II activity in vascular smooth muscle cells and demonstrated a decreased phosphorylation of CaMK-II immunoprecipitates in the presence of heparin. Using phosphatase inhibitors, they proposed activation of protein phosphatases PP-1 and PP-A2 by heparin (23). Indeed, activation of phosphatases is an interesting hypothesis to account for heparin's suppressive effects on multiple, disparate kinase pathways. However, we recently showed that the decrease in serum-stimulated Erk MAPK activity that occurs in the presence of heparin is not dependent on activity of the Erk regulatory phosphatases (43). In particular, activity of the MAPK phosphatase-1 (MKP-1), a primary regulator of Erk in smooth muscle cells (9, 18), is actually decreased in the presence of heparin, apparently secondarily to a decrease in Erk-dependent MKP-1 induction (43). Therefore, if a unifying mechanism exists to explain heparin's effects on multiple kinases, it may lie elsewhere than in phosphatase activation. The present study was undertaken, in part, to determine the role decreased phosphorylation may play in heparin's suppression of CaMK-II activity and c-fos induction.
MATERIALS AND METHODS
Materials. FBS, heparin (bovine lung, lot no. 49H0648 used in all experiments), chondroitin and dermatan sulfates, cAMP, phosphodiesterase (PDE), trifluoperazine, and malachite green were all obtained from Sigma (St. Louis, MO). RPMI-1640 cell culture medium, trypsin, Trizol reagent, and LipoFectamine 2000 were from Invitrogen Canada (Burlington, Ontario). NuSerum IV (BD Biosciences, Mississauga, Ontario) was used as the stimulus wherever serum stimulation of quiescent cell cultures is referred to. Calmodulin, ionomycin, and the CaMK-II inhibitor KT-5926 were products of Calbiochem (San Diego, CA). Autocamtide-2 was from Bachem (Torrance, CA). ECL Plus Western blotting kits, nitrocellulose, and Hybond-N membranes were purchased from Amersham Biosciences (Baie D'urfé, Quebec). SR plasmids containing either the coding sequence for the first 326 amino acids of the native sequence of CaMK-II, designated CaMK-II326 and containing regulatory and catalytic domains, or coding a constitutively active mutant of 326 amino acids with a T286D replacement, were kindly provided by Dr. Howard Schulman (11). Anti-CaMK-II antibody (mouse monoclonal IgG1, cat. no. 05–532) was from Upstate (Charlottesville, VA). Anti-phospho-Thr antibody was from Santa Cruz Biotechnology (Santa Cruz, CA) and an antibody to phospho-Thr286-CaMK-II (rabbit polyclonal, cat. no. 44–674G) was a product of BioSource (Camarillo, CA).
Cell culture and transfection. Rat mesangial cell cultures were established and characterized as previously described (38). They were maintained in RPMI 1640 medium with 10% FBS, in a humidified atmosphere of 5% CO2 at 37°C, and subcultured by trypsinization. Experiments were performed on cells between passages 5 and 15, as described previously (20). For transfection, cells were passaged at a split ratio of 1:3 and grown to 80% confluence. Four micrograms of plasmid DNA per 10-cm plate (either empty vector SR, vector coding CaMK-II326, or the T286D mutant) were diluted with serum- and antibiotic-free RPMI-1640 and incubated for 5 min before addition of the lipofectamine reagent at a ratio of DNA:lipofectamine of 1:2–1:3 (wt/wt). The mixture was incubated for 15 min at room temperature and then overlaid onto prewashed mesangial cell layers. After incubation at 37°C for 5–6 h, the cells were returned to 10% FBS medium and grown overnight before serum deprivation (0.4% for 48 h) and subsequent stimulation with NuSerum or ionomycin.
Western blotting. For detection of total CaMK-II, cells were washed with chilled phosphate-buffered saline and lysed by two freeze-thaw cycles in 50 mM HEPES buffer containing 0.5% Nonidet P-40, containing protease and phosphatase inhibitors (1 mM Na3VO4, 25 mM NaF, 1 mM PMSF, and 1 μg/ml each of aprotinin, leupeptin, and pepstatin). They were then sonicated twice for 5 s at 200 W and centrifuged (13,400 g, 15 min). Cell extracts were separated by 12% SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes. Membranes were blocked with 5% nonfat dry milk in TBS and blotted overnight with 1 μg/ml anti-CaMK-II at 4°C. After being washed twice with water, membranes were blotted with horseradish peroxidase-conjugated rabbit anti-mouse IgG for 90 min at room temperature, washed twice with water, three times with TBS containing 0.01% Tween 20, and visualized with ECL Plus reagent according to the manufacturer's protocol.
For phospho-Thr286-CaMK-II, cells were lysed in modified RIPA buffer (10 mM Tris, pH 7.4, with 100 mM NaCl, 1 mM EDTA, 1 mM EGTA, 20 mM Na4P2O7, 0.1% SDS, 0.5% sodium deoxycholate, 1% Triton X-100, 10% glycerol, 1 mM PMSF, 60 μg/ml aprotinin, 10 μg/ml leupeptin, 1 μg/ml pepstatin, 1 mM NaF, and 2 mM Na3VO4). Thirty micrograms of total protein were separated by 10% PAGE, transfered to polyvinyl difluoride membrane, and blocked in TBS with 0.1% Tween and 5% BSA. Overnight incubation with a 1:1,000 dilution of anti-phospho-Thr286 was followed by incubation with a 1:7,500 dilution of secondary antibody and visualization with ECL Plus reagent. Dephosphorylation was achieved by incubating cell lysates (30 μg total protein prepared without vanadate) with 400 units of phosphatase in phosphatase buffer (50 mM Tris, pH 7.5, with 1 mM EDTA, 5 mM DTT, and 2 mM MnCl2) for 30 min at 30°C before 10% SDS-PAGE.
Detection of c-fos mRNA expression by Northern blotting. Total RNA was isolated with Trizol reagent and 10 μg of total RNA were loaded on a 1.2% agarose/2.2 M formaldehyde gel, transferred to Hybond-N membrane overnight, and hybridized with a [-32P]dCTP-labeled c-fos fragment as before (42). The hybridization solution contained 50% formamide, 10% dextran sulfate, and 1% SDS. Membranes were washed with 2x SSC/0.1% SDS twice at room temperature for 10 min, twice at 50°C for 15 min, and then in 0.1x SSC/0.1% SDS twice at 50°C for 30 min before exposure to Kodak MS X-ray film.
Kinase activity assay. Autonomous CaMK activity and total activity were measured as described previously (20). Briefly, total cell proteins were harvested in lysis buffer as above for Western blotting and centrifuged (13,400 g, 15 min). For autonomous activity, 5–10 μg of lysate protein were incubated at 30°C for 3 min in 25 μl of 50 mM HEPES assay buffer containing 10 mM MgCl2, 0.1 mM ATP, 1 mM EGTA, 0.02 mM autocamtide peptide (KKALRRQETVDAL) as a substrate, and 5 μCi/ml [-32P]ATP. For total activity, EGTA was replaced by 4 mM CaCl2 and 2 μM calmodulin. Reactions were stopped by addition of trichloroacetic acid to 5% and spotted on P81 phosphocellulose filters, washed with 75 mM H3PO4, and counted by liquid scintillation. Immunoprecipitation of phosphorylated protein for kinase assay was carried out with an anti-phospho-Thr antibody. Cell extracts were prepared as for CaMK Western blots and 200 μg of total protein were incubated with 1 μg antibody overnight a 4°C. After being incubated with 100 μl of a 50% slurry of protein G (2 h, 4°C), the pellets were resuspended in equal volume of 2x kinase buffer, incubated at 30°C for 3 min, and loaded on P81 filter paper as above.
Calmodulin activity assay. We adapted a PDE-linked enzyme assay that has been developed for in vitro assessment of calmodulin activity during purification from mucus (10), as this assay has good sensitivity and specificity (5). To prepare samples, treated cells were placed in 100 μl 0.9% NaCl during freeze-thaw cycles, then scraped into 1.5-ml tubes, boiled at 100°C for 5 min, placed on ice for 10 min, and centrifuged at 13,400 g for 15 min at 4°C.
The calmodulin activity assay uses calmodulin-dependent PDE to convert cAMP to AMP with subsequent liberation of inorganic phosphate by alkaline phosphatase. Phosphate is then measured spectrophotometrically (A650) after reaction with malachite green as described (10, 27). Five micrograms of total cell protein were preincubated with 2 mU of PDE and 1.5 U alkaline phosphatase in 0.8 ml assay buffer without Ca2+ [40 mM Tris/40 mM imidazole (pH 7.4), 1 mM Mg(CH3COO)2] for 1 min at 30°C. After addition of 1 mM cAMP, the assay mixtures were incubated for a further 30 min at 30°C. Fifty-microliter aliquots of assay mixtures were then added to 900 μl molybdic acid reagent and mixed with 25 μl malachite green solution (1.26 mg/ml). After 2 min, the absorption was measured at 650 nm. Calmodulin activity was calculated based on a standard curve prepared with activated calmodulin.
The response of the malachite green detection component of the assay was linear over a range of potassium phosphate from 0.5 to 3 μg, allowing standard curves to be constructed with purified calmodulin that were linear (r2 > 0.95) over 0 to 125 ng calmodulin and an absorbance range (A650 nm) of 0 to 0.6. Conditions for the initial reaction with cytosolic extracts were chosen to give a final absorption in the range 0.2 to 0.4 (typically, 5 μg cell protein). Under these conditions, addition of excess, exogenous Ca2+ caused an approximate doubling of the signal, presumably representing activation of total cellular calmodulin, and the following controls all gave an absorption of less than 0.03: omission of PDE, omission of alkaline phosphatase, omission of cAMP with or without added Ca2+, and inclusion of EGTA. Furthermore, the Ca2+-calmodulin antagonist trifluoperazine (1 mM) suppressed the calmodulin standard curve to background absorbance, and standard curves were corrected for background by subtracting values in the presence of trifluoperazine.
Statistical analyses. Statistical differences between groups within an experiment were analyzed by one-way ANOVA with Tukey's post hoc test for pairwise comparisons.
RESULTS
As in our previous studies (20), treatment of quiescent rat mesangial cells with either serum or ionomycin resulted in a significant increase in autonomous CaMK activity (Fig. 1) that consistently reached about 15–20% of total activity (data not shown). This autonomous activity was partially inhibited by the CaMK-II inhibitor KT-5926 and by heparin (1 μg/ml). Endogenous CaMK-II was only weakly detected by Western blotting (Fig. 2), and to further confirm inhibition of CaMK-II by heparin, we transfected cells with wild-type CaMK-II326. The CaMK-II326 protein was strongly expressed (Fig. 2), resulted in a higher basal level of CaMK activity in quiescent cells, and was activated by ionomycin (Fig. 1). Again, both KT-5926 and heparin inhibited ionomycin-dependent activity in the CaMK-II326 transfectants. Mock transfection with empty SR plasmid did not affect basal CaMK activity or activation of endogenous CaMK-II by serum or ionomycin, compared with untransfected cells (compare black and open bars, Fig. 1).
c-fos mRNA is elevated 30 min after ionomycin treatment of quiescent mesangial cells (Fig. 3). Consistent with c-fos induction through CaMK-II (20), c-fos mRNA levels show the same pattern of response to KT-5926 and heparin as does autonomous CaMK activity; significant inhibition of c-fos expression occurs with each agent, and heparin decreases c-fos mRNA levels to the basal level observed in quiescent cells. Basal c-fos expression is increased in cells transfected with either empty plasmid or with plasmid containing CaMK-II326. The c-fos mRNA levels in quiescent cells transfected with the empty plasmid are not significantly affected by ionomycin or inhibitors (data not shown). However, ionomycin increases c-fos mRNA levels further in CaMK-II326-transfected cells, and heparin decreases these levels to below basal values (Fig. 3).
To determine whether heparin acts at a level proximal to CaMK in this signaling pathway, we measured calmodulin activity in control, transfected, and heparin-treated cells. Use of the phosphodiesterase-linked assay with cytosolic extracts is with caveats such as potential activation by Ca2+ release during cell extraction or interference from cellular inorganic phosphate (see DISCUSSION). Nevertheless, we were able to implement the assay in a reproducible manner that allows certain conclusions to be drawn. Stimulation of quiescent cells with serum or ionomycin caused increases in active Ca2+-calmodulin from 12.2 ng/μg cytosolic protein to 18.0 and 19.0 ng/μg, respectively (Table 1). Inclusion of heparin (1 μg/ml) with ionomycin was without effect (18.7 ng/μg). Similarly, heparin did not affect Ca2+-calmodulin activation by ionomycin in either mock transfected cells or cells transfected with CaMK-II326 (Table 1). Thus the phosphodiesterase-linked assay measures an increase in calmodulin activation in both control and transfected cells that is unresponsive to heparin; the influence of heparin is downstream of calmodulin activation.
View this table:
An initial event in CaMK-II activation following upstream activation of calmodulin is autophosphorylation on Thr286. We used a phospho-Thr286-specific antibody to examine whether 1 μg/ml heparin interfered with this event. Because higher concentrations of heparin have been reported to decrease net CaMK-II phosphorylation (23), we also compared the effects of 100 μg/ml heparin (Fig. 4). Treatment with phosphatase demonstrates that the antibody recognizes primarily a phosphorylated form of the protein. Heparin at 1 μg/ml is ineffective at decreasing ionomycin-dependent phosphorylation of Thr286. On the other hand, 100 μg/ml heparin significantly decreases phosphorylation, as previously observed (23).
To further examine the effects of heparin on phosphorylation, we used the alternative approach of measuring CaMK activity in immunoprecipitates prepared with an antibody directed against phosphothreonine. Despite inclusion of phosphatase inhibitors in the extraction solution, we were unable to achieve reproducible or fully autonomous activity in the immunoprecipitated fractions (data not shown), and we conclude that activated CaMK-II may be unstable under the conditions of immunoprecipitation. As an alternative indicator of CaMK-II phosphorylation, we measured total CaMK activity elicited by addition of Ca2+ and calmodulin in anti-phospho-Thr immunoprecipitates (Table 2). Ionomycin stimulation increased the CaMK-II activity of anti-phospho-Thr immunoprecipitates by about twofold. Heparin (1 μg/ml) was without significant effect, supporting the conclusion that this level of heparin inhibits CaMK-II activity independently of altering its level of phosphorylation. Consistent with decreased phosphorylation in the presence of 100 μg/ml heparin, this concentration of heparin decreased recovery of activity in the immunoprecipitate by about 45% (Table 2).
View this table:
Because higher concentrations of heparin reduced phosphorylation of Thr286 and decreased recovery of CaMK-II activity in anti-phospho-Thr immunoprecipitates, we compared the effects of 1 and 100 μg/ml on c-fos induction. The higher concentration did not give a significantly greater decrease in ionomycin-dependent c-fos induction than the lower (Fig. 5). We previously showed that 1 μg/ml heparin is sufficient to cause near-maximal suppression of [3H]thymidine incorporation in stimulated mesangial cells (38).
As a further approach to studying the role of phosphorylation in heparin's effect on CaMK signaling, we used a constitutively active phosphomimetic mutant of CaMK-II326 in which Thr286 is replaced with Asp, designated CaMK-II(T286D). Cells transfected with this mutant expressed 110 ± 40% (n = 7) basal CaMK activity compared with untransfected cells stimulated with serum (Fig. 6A). This was not significantly increased by ionomycin nor was it decreased by the CaMK-II inhibitor KT-5926, consistent with the autonomous activity reported for this modified protein (11). However, the activity was significantly decreased by 1 μg/ml heparin. Furthermore, quiescent cells expressing CaMK-II(T286D) had their basal c-fos mRNA levels decreased to 66.5 ± 17.4% by 15 min of exposure to 1 μg/ml heparin (Fig. 6B). Calmodulin activity was unaffected (Table 1). The 1-μg/ml concentration of heparin apparently produces maximal suppression in this model. In three further independent experiments comparing 1 and 100 μg/ml heparin in CaMK-II(T286D)-transfected cells, heparin again caused a decrease to 63.9 ± 7.7% of basal c-fos expression, and no further decrease was seen with 100 μg/ml (Fig. 7).
The effects observed here also raise the question of specificity. The closely related galactosamine-based glycosaminoglycan chondroitin sulfate had no effect on ionomycin-induced c-fos expression, nor did the even more cosely related iduronic acid-rich dermatan sulfate (Fig. 8).
DISCUSSION
CaMK-II is a broad-specificity calmodulin-dependent kinase with homology to, but distinct from, CaMK-I and CaMK-IV (14). The assembled CaMK-II complex (8, 14) consists of 10–12 subunits representing four different isoforms (, , , and ), each of which occurs in multiple splice variants. Each has an inhibitory regulatory domain that interacts with the catalytic domain; this interaction is disrupted by Ca2+/calmodulin to expose the active site. The -subunit is highly conserved among many species. In contrast to other calmodulin-dependent enzymes, CaMK-II has autophosphorylation sites in the regulatory domain whose phosphorylation breaks the inhibitory interaction of the catalytic domain and confers autonomous activity, i.e., activity independent of Ca2+ or calmodulin (13, 14, 35). In particular, Thr286 of CaMK-II is essential for autonomous activity; when it is mutated to leucine, the kinase remains ligand (calmodulin) dependent (11), like myosin light chain kinase (MLCK). Mutation of Thr286 to the phosphomimetic aspartate confers constitutive activity of about 10–20% that of native autonomous protein (2), as exploited in the present study. Thus phosphorylation of Thr286 is both necessary and sufficient for CaMK-II ativity. Although phosphorylation of other Thr residues occurs, notably Thr305 and Thr306, this is inhibitory and prevents activation of the kinase by binding of Ca2+/calmodulin (12). Among the numerous substrates of CaMK-II are transcription factors such as cAMP response element-binding protein (CREB), and it is likely through phosphorylation of this factor that its activation induces c-fos (31). Indeed, ionomycin-induced phosphorylation of CREB on Ser133 is accompanied by c-fos induction in mesangial cells (42).
We previously suggested that the rapid response of Erk and CaMK signaling to heparin supports a mechanism of action from the cell surface (20, 21, 38). For instance, internalization of negatively charged heparin by mesangial cells is slow (hours), compared with suppression of a CaMK-II signal 30 s after addition to cultured cells of heparin together with a stimulus such as ionomycin (20). This is not due to interference with ionomycin-dependent Ca2+ signaling, which is unaffected by heparin in these circumstances (20). And the effect shows some specificity for heparin; the closely related glycosaminoglycans, chondroitin and dermatan sulfates, are ineffective in suppressing signaling to c-fos. This points to a mechanism of competing signaling, or perhaps to disruption of proximal events at the cell surface, rather than direct interaction of heparin with downstream components of the kinase cascades. Furthermore, because autonomous activity assayed in vitro is decreased in extracts of cells that had previously been exposed to heparin, the inhibition of CaMK-II might involve events at, or upstream of, enzyme assembly and autophosphorylation. However, the present results appear to rule out this explanation.
In the present study, we could obtain no evidence for an effect of heparin on formation of active Ca2+/calmodulin. This result must be qualified by the limitations of the assay. Release of Ca2+ from stores during cell disruption could mask true levels of active calmodulin. Alternatively, Ca2+ could dissociate from calmodulin prior to assay. However, we do measure an increase in activity after treatment with ionomycin, which is sensitive to the calmodulin antagonist trifluoperazine. This suggests that activation can be measured with our assay, and the response is not affected by heparin. This is consistent with the earlier observation that heparin does not affect angiotensin-mediated mesangial cell contraction (20), a process also operating through calmodulin-dependent activation of MLCK. Together with ionomycin-dependent phosphorylation of Thr286 in the presence of 1 μg/ml heparin, these observations support a mechanism of action of heparin independent of calmodulin.
Activation of phosphatase(s) was proposed as an attractive mechanism whereby heparin might negatively regulate diverse kinase cascades (22), potentially at multiple sites. Subsequently, Mishra-Gorur et al. (23) presented evidence for decreased phosphorylation of the CaMK-II subunit in vascular smooth muscle labeled with [32P]phosphate and then treated with serum containing 100 or 500 μg/ml heparin (49 and 68% decrease, respectively, compared with heparin-free controls). Using several phosphatase inhibitors, which prevented decreased phosphorylation of CaMK-II otherwise occurring in the presence of heparin, they provided evidence for a role of protein phosphatases PP-1 and PP-2A. However, we recently showed that 1 μg/ml heparin did not suppress Erk activation by upregulating its phosphatase, MKP-1, but rather decreased MKP-1 secondarily to effects on Erk (43). In our mesangial cell system, this more physiological concentration of heparin shows near-maximal suppression of [3H]thymidine incorporation (38) and c-fos induction (Fig. 5). Therefore, we investigated in the present study whether dephosphorylation of CaMK-II was necessary for the effect of heparin at the lower concentration. Although high amounts of heparin did indeed decrease the amount of phospho-CaMK-II, 1 μg/ml did not, suggesting that at least some of heparin's antimitogenic activity is not due to effects on protein phosphatases. We also found that immunoprecipitates of phospho-Thr phosphoproteins contained increased amounts of CaMK-II activity following ionomycin stimulation, that was decreased by 100 μg/ml heparin (Table 2), again consistent with the findings of Mishra-Gorur et al. (23). Heparin at 1 μg/ml, on the other hand, was without effect on recovery of activity, indicating that the lower concentration of heparin did not affect total Thr phosphorylation. We can conclude that decreased phosphorylation is not necessary for suppression of CaMK-II activity, and indeed heparin treatment also decreases the constitutive activity of the phosphomimetic mutant T286D, which is not dependent on phosphorylation (at least of Thr286) for enzyme activity. As in control cells, 1 μg/ml heparin is as effective as 100 μg/ml in decreasing activity of the phosphomimetic T286D mutant. We conclude that heparin must elicit signals that disrupt or prevent an active form of the protein from occurring whether Thr286 is phosphorylated or not. Because the suppressed activity persists in the assay carried out in vitro in the absence of heparin, it is reasonable to suggest that heparin may rapidly mobilize an inhibitory factor, or alternatively disrupt association with an activating component. The existence of such factors is speculative but suggests models for the more general effects of heparin on signaling that should be evaluated. In this context, it is interesting to note that Src has also been implicated in induction of c-fos through the serum response element, downstream of CaMK-II signaling in mesangial cells (39), although the nature of the CaMK-II/Src interaction is unknown.
The inhibitor KT-5926 was initially introduced as an inhibitor of MLCK but actually has a lower Ki for CaMK-II (0.004 μM) than for MLCK (0.018 μM). It was shown to act directly at the catalytic site and competitively with ATP (26). Therefore, it is expected to inhibit CaMK-II326 as seen in Fig. 1. That it is not as effective at inhibiting c-fos induction in these transfected cells as it is in untransfected cells (Fig. 3) suggests heparin-sensitive pathways additional to CaMK may be activated in the transfected cells. Alternatively, KT-5926 could have differing inhibitory potencies for autocamtide-2 used as a substrate in assaying autonomous activity and as an endogenous substrate (e.g., CREB) that differ in native protein and truncated recombinant protein. This possibility has not been addressed.
In conclusion, heparin appears to elicit signals that suppress CaMK-II activity independently of phosphorylation of Thr286, although high concentrations of heparin may indeed decrease autophosphorylation of the protein. Activation of calmodulin by Ca2+ is not affected by heparin, and other calmodulin-dependent proteins such as MLCK may also be unaffected. Whether heparin exploits common mechanisms in regulating CaM kinases and MAP kinases remains to be determined.
GRANTS
This work was supported by operating grant NA-3520 from the Heart and Stroke Foundation of Ontario.
FOOTNOTES
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
REFERENCES
Antoine M, Gaiddon C, and Loeffler JP. Ca2+/calmodulin kinase types II and IV regulate c-fos transcription in the AtT20 corticotroph cell line. Mol Cell Endocrinol 120: 1–8, 1996.
Brickey DA, Bann JG, Fong YL, Brennan RG, and Soderling TR. Mutational analysis of the autoinhibitory domain of calmodulin kinase II. J Biol Chem 269: 29047–29054, 1994.
Carl M, Akagi Y, Weidner S, Isaka Y, Imai E, and Rupprecht HD. Specific inhibition of Egr-1 prevents mesangial cell hypercellularity in experimental nephritis. Kidney Int 63: 1302–1312, 2003.
Castellot JJ Jr and Karnovsky MJ. Heparin and the regulation of growth of the vascular wall. In: Vascular Smooth Muscle in Culture, edited by Campbell JH and Campbells GR. Boca Raton, FL: CRC, 1987, p. 93–115.
Chock SP and Huang CY. An optimized continuous assay for cAMP phosphodiesterase and calmodulin. Anal Biochem 138: 34–43, 1984.
Clowes AW and Clowes MM. Inhibition of smooth muscle cell proliferation by heparin molecules. Transplant Proc 21: 3700–3701, 1989.
Daum G, Hedin U, Wang YX, Wang T, and Clowes AW. Diverse effects of heparin on mitogen-activated protein kinase-dependent signal transduction in vascular smooth muscle cells. Circ Res 81: 17–23, 1997.
De Koninck P and Schulman H. Sensitivity of CaM kinase II to the frequency of Ca2+ oscillations. Science 279: 227–230, 1998.
Duff JL, Monia BP, and Berk BC. Mitogen-activated protein (MAP) kinase is regulated by the MAP kinase phosphatase (MKP-1) in vascular smooth muscle cells. Effect of actinomycin D and antisense oligonucleotides. J Biol Chem 270: 7161–7166, 1995.
Flik G, Van Rijs JH, and Wendelaar Bonga SE. Evidence for the presence of calmodulin in fish mucus. Eur J Biochem 138: 651–654, 1984.
Hanson PI, Kapiloff MS, Lou LL, Rosenfeld MG, and Schulman H. Expression of a multifunctional Ca2+/calmodulin-dependent protein kinase and mutational analysis of its autoregulation. Neuron 3: 59–70, 1989.
Hanson PI and Schulman H. Inhibitory autophosphorylation of multifunctional Ca2+/calmodulin-dependent protein kinase by site-directed mutagenesis. J Biol Chem 267: 17216–17224, 1992.
Hook SS and Means AR. Ca2+/CaM-dependent kinases: from activation to function. Annu Rev Pharmacol Toxicol 41: 471–505, 2001.
Hudmon A and Schulman H. Neuronal Ca2+/calmodulin-dependent protein kinase II: the role of structure and autoregulation in cellular function. Annu Rev Biochem 71: 473–510, 2002.
Johnson RJ. The glomerular response to injury: progression or resolution Kidney Int 45: 1769–1782, 1994.
Kohno M, Yokokawa K, Yasunari K, Minami M, Kano H, Mandal AK, and Yoshikawa J. Heparin inhibits human coronary artery smooth muscle cell migration. Metabolism 47: 1065–1069, 1998.
Koyama N, Kinsella MG, Wight TN, Hedin U, and Clowes AW. Heparan sulfate proteoglycans mediate a potent inhibitory signal for migration of vascular smooth muscle cells. Circ Res 83: 305–313, 1998.
Lai K, Wang H, Lee WS, Jain MK, Lee ME, and Haber E. Mitogen-activated protein kinase phosphatase-1 in rat arterial smooth muscle cell proliferation. J Clin Invest 98: 1560–1567, 1996.
Lorca T, Cruzaleugi FH, Fesquet D, Cavadore JC, Mery J, Means A, and Doree M. Calmodulin-dependent protein kinase II mediates inactivation of the MPF and CSF upon fertilization of Xenopus eggs. Nature 366: 270–273, 1993.
Miralem T and Templeton DM. Heparin inhibits Ca2+/calmodulin-dependent kinase II activation and c-fos induction in mesangial cells. Biochem J 330: 651–657, 1998.
Miralem T, Wang A, Whiteside CI, and Templeton DM. Heparin inhibits mitogen-activated protein kinase-dependent and -independent c-fos induction in mesangial cells. J Biol Chem 271: 17100–17106, 1996.
Mishra-Gorur K and Castellot JJ Jr. Heparin rapidly and selectively regulates protein tyrosine phosphorylation in vascular smooth muscle cells. J Cell Physiol 178: 205–215, 1999.
Mishra-Gorur K, Singer HA, and Castellot JJ Jr. Heparin inhibits phosphorylation and autonomous activity of Ca2+/calmodulin-dependent protein kinase II in vascular smooth muscle cells. Am J Pathol 161: 1893–1901, 2002.
Misra RP, Bonni A, Miranti CK, Rivera VM, Sheng M, and Greenberg ME. L-type voltage-sensitive calcium channel activation stimulates gene expression by a serum response factor-dependent pathway. J Biol Chem 269: 25483–25493, 1994.
Morris TA, DeLorenzo RJ, and Tombes RM. CaMK-II inhibition reduces cyclin D1 levels and enhances the association of p27kip1 with cdk2 to cause G1 arrest in NIH 3T3 cells. Exp Cell Res 240: 218–227, 1998.
Nakanishi S, Yamada K, Iwahashi K, Kuroda K, and Kase H. KT5926, a potent and selective inhibitor of myosin light chain kinase. Mol Pharmacol 37: 482–488, 1990.
Penney CL. A simple microassay for inorganic phosphate. Anal Biochem 75: 201–210, 1976.
Pukac L, Huangpu J, and Karnovsky MJ. Platelet-derived growth factor-BB, insulin-like growth factor-I, and phorbol ester activate different signaling pathways for stimulation of vascular smooth muscle cell migration. Exp Cell Res 242: 548–560, 1998.
Pukac LA, Carter JE, Ottlinger ME, and Karnovsky MJ. Mechanisms of inhibition by heparin of PDGF stimulated MAP kinase activation in vascular smooth muscle cells. J Cell Physiol 172: 69–78, 1997.
Pukac LA, Ottlinger ME, and Karnovsky MJ. Heparin suppresses specific second messenger pathways for protooncogene expression in rat vascular smooth muscle cells. J Biol Chem 267: 3707–3711, 1992.
Rosen LB, Ginty DD, and Greenberg ME. Calcium regulation of gene expression. In: Advances in Second Messenger and Phosphoprotein Research, edited by Meanss AR. New York: Raven, 1995, vol. 30, p. 225–253.
Ross R. Cell biology of atherosclerosis. Annu Rev Physiol 57: 791–804, 1995.
Saleh H, Schlatter E, Lang D, Pauels HG, and Heidenreich S. Regulation of mesangial cell apoptosis and proliferation by intracellular Ca2+ signals. Kidney Int 58: 1876–1884, 2000.
Schwartz SM, deBlois D, and O'Brien ERM. The intima. Soil for atherosclerosis and restenosis. Circ Res 77: 445–465, 1995.
Soderling TR. The Ca2+-calmodulin-dependent protein kinase cascade. Trends Biochem Sci 24: 232–236, 1999.
Tombes RM, Grant S, Westin EH, and Krystal G. G1 cell cycle arrest and apoptosis are induced in NIH 3T3 cells by KN-93, an inhibitor of CaMK-II. Cell Growth Differ 6: 1063–1070, 1995.
Wang A, Fan MY, and Templeton DM. Growth modulation and proteoglycan turnover in cultured mesangial cells. J Cell Physiol 159: 295–310, 1994.
Wang A and Templeton DM. Inhibition of mitogenesis and c-fos expression in mesangial cells by heparin and heparan sulfates. Kidney Int 69: 437–448, 1996.
Wang Y, Mishra R, and Simonson MS. Ca2+/calmodulin-dependent protein kinase II stimulates c-fos transcription and DNA synthesis by a Src-based mechanism in glomerular mesangial cells. J Am Soc Nephrol 14: 28–36, 2003.
Wang Y and Simonson MS. Voltage-insensitive Ca2+ channels and Ca2+/calmodulin-dependent protein kinases propogate signals from endothelin-1 receptors to the c-fos promoter. Mol Cell Biol 16: 5915–5923, 1996.
Wright TC Jr, Castellot JJ Jr, Diamond JR, and Karnovsky MJ. Regulation of cellular proliferation by heparin and heparan sulphate. In: Heparin: Chemical and Biological Properties, Clinical Applications, edited by Lane DA and Lindahls U. Boca Raton, FL: CRC, 1989, p. 295–316.
Zeng H, Liu Y, and Templeton DM. Ca2+/calmodulin-dependent and cAMP-dependent kinases in induction of c-fos in human mesangial cells. Am J Physiol Renal Physiol 283: F888–F894, 2002.
Zhao Y, Xiao W, and Templeton DM. Suppression of mitogen-activated protein kinase phosphatase-1 (MKP-1) by heparin in vascular smooth muscle cells. Biochem Pharmacol 66: 769–776, 2003.(Weiqun Xiao, Ying Liu, an)