Báo cáo khoa học: HIP/PAP, a C-type lectin overexpressed in hepatocellular carcinoma, binds the RIIa regulatory subunit of cAMP-dependent protein kinase and alters the cAMP-dependent protein kinase signalling

doi:10.1111/j.1432-1033.2004.04302.x

Eur. J. Biochem. 271, 3812–3820 (2004) (cid:1) FEBS 2004

HIP/PAP, a C-type lectin overexpressed in hepatocellular carcinoma, binds the RIIa regulatory subunit of cAMP-dependent protein kinase and alters the cAMP-dependent protein kinase signalling

France Demaugre1, Yannick Philippe1, Sokavuth Sar1, Bernard Pileire2, Laurence Christa1, Chantal Lasserre1 and Christian Brechot1 1INSERM U370 CHU Necker Enfants Malades, Paris, France; 2Laboratory of Biochemistry, CHU Antilles-Guyane Point a` Pitre, Guadeloupe, France

via the Golgi apparatus we showed that a fraction of HIP/PAP escaped the secretory apparatus and was recovered in the cytosol. Basal PKA activity was in- creased in HIP/PAP expressing cells, suggesting that HIP/PAP may alter PKA signalling. Indeed, we showed, using a thymidine kinase-luciferase reporter plasmid in which a cAMP responsive element was inserted upstream of the thymidine kinase promoter, that luciferase activity was enhanced in HIP/PAP expressing cells. Thus our findings suggest a novel mechanism for the biological activity of the HIP/PAP lectin.

Keywords: C-type lectin; HIP/PAP; PKA; phosphorylation; liver. HIP/PAP is a C-type lectin overexpressed in hepatocel- lular carcinoma (HCC). Pleiotropic biological activities have been ascribed to this protein, but little is known about the function of HIP/PAP in the liver. In this study, therefore, we searched for proteins interacting with HIP/PAP by screening a HCC cDNA expression library. We have identified the RIIa regulatory subunit of cAMP-dependent protein kinase (PKA) as a partner of HIP/PAP. HIP/PAP and RIIa were coimmunoprecipi- tated in HIP/PAP expressing cells. The biological rele- vance of the interaction between these proteins was established by demonstrating, using fractionation meth- ods, that they are located in a same subcellular com- partment. Indeed, though HIP/PAP is a protein secreted

cells) in the intestine [8]. Moreover in rats, the HIP/PAP homologue (PAP 1/peptide 23/Reg 2), is expressed in pituitary and uterine cells under the influence of growth hormone releasing hormone and oestradiol, respectively [9,10], and by motor neurones in vivo during their regener- ation and in vitro when incubated with ciliary neurotrophic factor-related cytokines [11,12].

The HIP/PAP-encoding gene has been shown to be overexpressed in human hepatocellular carcinoma (HCC) [1] and in the pancreas during acute pancreatitis [2]. HIP/ PAP has been characterized as a protein belonging to the group 7 of C-type lectins [3,4]. HIP/PAP cDNA encodes a 175 amino acid protein containing only one carbohydrate- binding domain (CRD) linked to an N-terminal sequence, part of which is cleaved during its maturation and secretion [5]. In humans, HIP/PAP protein is not expressed in normal liver but is overexpressed in 75% of HCC, in cholangio- carcinoma and in reactive ductular cells in nonmalignant liver [6]. HIP/PAP expression in HCC does not result from the re-expression of a fetal marker. Indeed, analysis of mouse embryos has revealed that HIP/PAP is not expressed in the liver during development [7]. HIP/PAP has also been detected in the pancreas and in a subset of cells (Paneth

Correspondence to F. Demaugre, INSERM U370 CHU Necker Enfants Malades, 156 rue de Vaugirard, 75015 Paris, France. Fax: +33 1 40615581, Tel.: + 33 1 40615343, E-mail: demaugre@necker.fr Abbreviations: CRD, carbohydrate-binding domain; CRE, cAMP response element; HCC, hepatocellular carcinoma; HMK peptide, peptide phosphorylatable by heart muscle kinase; PKA, cAMP- dependent protein kinase; SERCA 2, sarco/endoplasmic reticulum Ca2+ATPase 2. (Received 19 March 2004, revised 9 July 2004, accepted 23 July 2004)

Little is known about the physiopathological significance of HIP/PAP expression. In the pancreas, there is evidence that HIP/PAP may participate in the antiapoptotic pro- gramme developed by acinar cells during acute pancreatitis [13]; indeed, HIP/PAP was reported to protect pancreatic AR4–2 J cells against apoptosis induced by oxidative stress [14]. In pituitary cells, PAP1/peptide 23 was reported to act as a growth factor [10,15] and it has been shown that PAP1 (referred to as Reg 2) is an important neurotrophic factor for motor neurones in vitro and in vivo in the rat [11,12]. In liver recombinant HIP/PAP has been shown to promote the adhesion of rat hepatocytes and to bind elements of the extracellular matrix [8]. Moreover HIP/PAP has been recently reported to combine mitogenic and antiapoptotic functions regarding hepatocytes and to enhance liver regeneration [16]. Nothing is known concerning the possible role of HIP/PAP during liver carcinogenesis. Thus, identi- fication of the proteins interacting with HIP/PAP liver should help to understand the function(s) of HIP/PAP during hepatic carcinogenesis.

In this study we have identified the RIIa regulatory subunit of cAMP-dependent protein kinase (PKA) as being

HIP/PAP alters PKA signalling (Eur. J. Biochem. 271) 3813

(cid:1) FEBS 2004

a partner of HIP/PAP, and we have demonstrated that PKA activity is enhanced in HIP/PAP expressing cells. The inserts amplified by PCR using Advantage cDNA polymerase and the kgt11 insert screening amplimer set (Clontech) were directly sequenced.

Materials and methods

Cell culture and transfection Plasmid constructs

serum, 100 lgÆmL)1

Chang cells (CCL13, ATCC) seeded in 100 mm Petri dish were maintained in DMEM supplemented with 7% (v/v) fetal bovine streptomycin and 100 lgÆmL)1 penicillin. Cells plated at a density of 1.5 · 106 cells per 100 mm diameter dish were transfected with appropriate vectors (20 lg ADN) using the calcium precipitation method, and further cultured for 48 h unless indicated. For the isolation of stable transformants Chang, cells transfected with pcDNA-HIP/PAP were cultured for 4 weeks with 800 lgÆmL)1 neomycin and screened for HIP/ PAP by immunoblot. Proteins were quantified using the BioRad protein Assay.

The HIP/PAP(29–175) coding sequence amplified by PCR using human HIP/PAP cDNA as a template [1] was subcloned at the EcoRI site in the bacterial expression plasmid pAR(deltaRI)[59/60] [17]. This plasmid allowed the production of HIP/PAP in fusion at the N-terminal extremity, with Flag and heart muscle kinase (HMK) peptides which allowed, respectively, the purification of chimeric HIP/PAP and its phosphorylation by bovine heart PKA. The sense primer (5¢-GTCGAATTCCAAGGTG AAGAACCCCAG-3¢) was located at nucleotides 63–90 of the coding sequence, and the antisense primer (5¢-TG CTGAATTCCCTCCCTCCTGCACTAGTCAG-3¢) over- lapped the stop codon. DNA sequencing confirmed the restored open reading frame of the fusion construct. Analysis of HIP/PAP in transiently HIP/PAP expressing Chang cells

The complete HIP/PAP(1–175) sequence, amplified using the same template, was subcloned at EcoRI and XhoI sites in pcDNA3.1, and in pcDNA3.1/myc-His (Invitrogen). The QuickChange Site-directed Mutagenesis Kit (Stratagene) was used to switch serines 73 and 138 and threonine 153 of the HIP/PAP protein for alanines. Oligonucleotides cas- settes containing the desired mutations were inserted into pcDNA3-HIP/PAPmyc-His as indicated by the manufac- turer. Direct sequencing confirmed the sequence of the inserts.

Production, purification and labelling of Flag-HMK- HIP/PAP(29–175)

Effect of brefeldin A. Twenty-four hours post transfection with pcDNA-HIP/PAP, cells were seeded in 60-mm Petri dishes and further grown for 24 h before 10 lM brefeldin A was added to the culture medium. At the end of incubation, cells lysed in buffer A (10 mM KH2PO4 pH 7.4, 150 mM NaCl, 10 mM EDTA, 1% (v/v) Triton X-100 and 2 lgÆmL)1 aprotinin, 1 lgÆmL)1 pepsatin, 2 lgÆmL)1 leu- peptin, 0.1 lgÆmL)1 phenylmethylsulfonyl fluoride, 10 mM sodium fluoride, 2 mM sodium orthovanadate, 1 lM oka- daic acid) and the culture medium were resolved in 13% SDS/PAGE and analyzed for HIP/PAP by Western blotting using anti-HIP/PAP Ig [4]. The blots were revealed using an enhanced chemiluminecence system, according to the manufacturer’s instructions (Amersham Life Science).

labelled

Chimeric HIP/PAP was produced in BL21 (DE3) Escheri- chia coli transformed with pAR(deltaRI)[59/60]-HIP/ PAP(29–175) using conventional methods. At the end of the culture the bacteria were lysed at 4 (cid:2)C with 10 lgÆmL)1 lysozyme in 50 mM Tris pH 8.0, 2 mM EDTA, 300 mM KCl, 0.2% (v/v) Triton X-100 and 0.1 lgÆmL)1 phenyl- methylsulfonyl fluoride, and centrifuged. Chimeric HIP/ PAP was purified from the supernatant using affinity chromatography with monoclonal M2 anti-Flag agarose (Sigma). Chimeric HIP/PAP was using [32P]ATP[cP] and the catalytic subunit of PKA as described [17] and cleared from unincorporated [32P]ATP[cP] using Sephadex G25 chromatography.

Effect of PKA overexpression. Forty hours post transfec- tion with 18 lg of either the wild or mutated forms of pcDNA-HIP/PAPmyc and 2 lg pCaEV encoding for the catalytic subunit of PKA [20] when indicated, cells were lysed with buffer A. Cellular lysates (100 lg protein) were incubated overnight at 4 (cid:2)C with 2 lg monoclonal anti-myc and then for 2 h with 10 lL protein G Sepharose beads (Amersham Life Science). Immune complexes washed with buffer A were released from beads using Laemmli buffer and analyzed by Western blotting for HIP/PAP using polyclonal antibody anti-HIP/PAP and for phosphorylated serine using polyclonal anti-phosphoserine (Zymed Labor- atories). Screening of a human HCC cDNA kgt11 library with [32P]Flag-HMK- HIP/PAP(29–175)

Cell fractionation

HIP/PAP expressing and control Chang cells were fraction- ated between soluble and particulate fractions as described [21]. Sarco/endoplasmic reticulum Ca2+ATPase 2 (SERCA 2), an integral protein of the endoplasmic reticulum [22], calreticulin, a protein of the endoplasmic reticulum lumen [23], HIP/PAP, the RIIa and the Ca subunits of PKA were checked by immunoblotting in both the 100 000 g pellet solubilized with buffer A and the supernatant using An amplified human HCC cDNA library, inserted in kgt11 (provided by C. Lasserre), was plated with Y1090 E. coli and induced with isopropyl thio-b-D-galactoside, as des- cribed previously [18]. At the end of culture, nitrocellulose filters subjected to a denaturation-renaturation cycle [19] were hybridized overnight at 4 (cid:2)C with 32P-labelled chimeric HIP/PAP at a final concentration of 100 000–300 000 cpmÆmL)1 as described [17]. Plaques hybridized with the probe were grown until they were purified. Phage DNA was purified using the kgt11 DNA purification kit (Stratagene).

3814 F. Demaugre et al. (Eur. J. Biochem. 271)

(cid:1) FEBS 2004

anti-HIP/PAP, anti-RIIa and anti-Ca (Transduction Laboratories, Lexington, KY, USA), anti-(SERCA 2) (clone IID8; Tebu, Paris, France) and anti-calreticulin (ABR Golden Co.) Igs.

vector were seeded at a density of 2 · 106 cells per 100 mm Petri dish 30 h before the assays. They were lysed in 20 mM Tris, pH 7.5, containing 1 mM EDTA, 1 mM dithiothreitol, and protease and phosphatase inhibitors (see above), and centrifuged at 3000 g. Supernatants were assayed immedi- ately for kinase activity as described previously [24]. Co-immunoprecipitation experiments

Reporter gene assays HIP 9 and PC8 clones seeded at a density of 2 · 105 cells per 35 mm diameter dish were transfected with 5 lg of total DNA including either 2 lg of TK-LUC reporter plasmid or 2 lg of CRE-TK-LUC reporter plasmid [25] and when indicated 0.5 lg of pCa EV [20]. Cells were lysed 48 h post- transfection. Luciferase activity was measured by a standard assay with a Lumat LB9501 luminometer (Fisher Bioblock Scientific, Illkirch, Cedex, France).

Statistical analysis

Forty-eight hours post transfection with either pcDNA- HIP/PAP or the empty vector Chang cells were lysed in 10 mM Tris pH 7.5, 2.5 mM MgCl2, 10 mM KCl, 0.5 mM dithiothreitol, 0.05% (v/v) NP40, and protease and phos- phatase inhibitors (see above). Extracts (400 lg protein) clarified by centrifugation at 6000 g, were incubated over- night with 2 lg of either polyclonal anti-RIIa (Santa Cruz Biotechnology, Santa Cruz, CA, USA) or control serum, in lysis buffer. The immune complexes were recovered with 10 lL of protein G Sepharose, washed with lysis buffer adjusted to 100 mM KCl and 0.1% (v/v) NP40. Proteins were released from beads using 50 lL of Laemmli buffer. One sample (45 lL) was analyzed by Western blotting for HIP/PAP by 13% (w/v) SDS/PAGE and the other (5 lL) for RIIa by 9% (w/v) SDS/PAGE, using anti-RIIa mAb (Transduction Laboratories).

Immunofluorescence and confocal analysis

Using the nonparametric Kolmogorov–Smirnov test and the Levene test, it was established that the distribution of data obtained with different clones was normal. Student’s t-test was used to compare mean values of enzymatic activities measured under different conditions. Similar levels of statistical significance were obtained when HIP/PAP effects were analyzed in individual control and HIP/PAP clones or in pooled clones.

Results

Identification of the RIIa regulatory subunit of PKA as a partner of HIP/PAP

After transfection with pcDNA-HIP/PAP, cells grown on glass coverslips were fixed with 4% (v/v) paraformalde- hyde and permeabilized with methanol at 4 (cid:2)C. They were then incubated with anti-RIIa mAb and polyclonal anti-(WAP-HIP/PAP) [5] for 1 h at room temperature. Immunodetection was carried out using fluorescein iso- thiocyanate-conjugated anti-rabbit Ig for HIP/PAP and/or cyanin-5 conjugated anti-mouse Ig for RIIa detection. Monoclonal antibody CTR433 (a gift from M. Bornens, Curie Institute, Paris, France) associated with cyanin-5- conjugated anti-mouse Ig was used for labelling of median Golgi. The coverslips were analyzed using laser confocal scanning microscopy. Fluorochrome-conjugated secon- dary antibodies were from Jackson (West Grove, PA, USA).

Phosphorylation of recombinant HIP/PAP by PKA

Recombinant HIP/PAP [4] was incubated at 30 (cid:2)C in 80 lL, with 100 lM [32P]ATP[cP] (specific activity, 15 000 cpmÆ pmol)1) and 25 units of bovine heart PKA in 20 mM Tris pH 7.5, 100 mM NaCl, 12 mM MgCl2. Control incubations performed without recombinant HIP/PAP were conducted in parallel. At indicated times, 5 lL of incubation mixtures were spotted on phosphocellulose filters (Whatman P81) which were then washed in phosphoric acid and dried as described [24]. Radioactivity was measured by liquid scintillation with Econofluor. Incubation mixtures (2 lL) were also analyzed using SDS/PAGE, and [32P]HIP/PAP was detected by autoradiography of the wet gel.

Protein kinase assays

In order to assess the biological consequences of HIP/PAP expression in hepatocellular carcinoma, we looked for proteins capable of interacting with this protein by screening a human HCC cDNA expression library in kgt11 using [32P]chimeric HIP/PAP as a probe. For this purpose, we cloned HIP/PAP(29–175) in the pAR[DRI] vector. Of the 750 000 plaques analyzed, two of them hybridized with the probe. The sequences of the inserted cDNA were identical. In frame with the kgt11 Lac Z coding sequence they contained 1500 bp DNA, 1120 bp of which encoded for the C-terminal portion of the RIIa regulatory subunit of PKA. No hepatic cell line expressing HIP/PAP was available. Thus we have established hepatic cell models expressing HIP/PAP through their transfection with pcDNA-HIP/ PAP in order to validate HIP/PAP–RIIa interaction. HIP/ PAP was expressed more efficiently in Chang cells. Experi- ments were therefore performed using this cell line. HIP/ PAP was recovered in the serum of patients with hepato- in an in vivo cellular carcinoma which suggested that, setting, HIP/PAP was secreted by liver cells [6]. A similar pattern was observed in HIP/PAP-expressing Chang cells (Fig. 1A). HIP/PAP was recovered in the cells and the culture medium, and brefeldin A, an inhibitor of protein secretion [26], reduced HIP/PAP expression in the culture medium which indicated that HIP/PAP was secreted via a pathway involving the Golgi apparatus.

Expression of HIP/PAP and RIIa in Chang cells was analyzed using immunofluorescence methods (Fig. 2). As Two independent clones of stably expressing HIP/PAP Chang cells (HIP 9 and HIP 4) and two independent control clones (PC4 and PC8) stably transfected with the empty

HIP/PAP alters PKA signalling (Eur. J. Biochem. 271) 3815

(cid:1) FEBS 2004

Fig. 1. HIP/PAP expression in Chang cells. Experiments were performed with Chang cells transiently expressing HIP/PAP. (A) Effect of brefeldin A on HIP/PAP distribution in cell culture. After incubation for 2 h with or without 10 lM brefeldin A, lysed cells and culture media were analyzed for HIP/PAP by Western blotting. (B) Fractionation experiments. Pellets and supernatants recovered after centrifugation at 100 000 g of homo- genates from control (Neo) and HIP/PAP-expressing cells were analyzed by Western blotting for HIP/PAP [13% (w/v) SDS/PAGE] and, for SERCA 2, RIIa and Ca subunits of PKA, and calreticulin [9% (w/v) SDS/PAGE]. (C) Co-immunoprecipitation of HIP/PAP with RIIa. Cell lysates (400 lg protein) were incubated overnight with control serum (1), polyclonal anti-RIIa (2) or without serum (3). The resulting immune complexes recovered with protein G Sepharose, were analyzed for HIP/PAP and RIIa by Western blot using polyclonal anti-HIP/PAP and mAb anti-RIIa.

to that of HIP/PAP when experiments were conducted with control cells (results not shown).

HIP/PAP is phosphorylated by PKA

previously observed in other HIP/PAP expressing cell lines [12,27] the immunostaining generated by anti-HIP/PAP Ig was cytoplasmic and mostly present in the juxta nuclear area (Fig. 2Aa). It partially colocalized with CTR433 (Fig. 2B) a marker of median Golgi [28]. Immunostaining generated by anti RIIa antibody was not altered in HIP/ PAP expressing cells. As observed in other cell lines [29], it was mostly juxta nuclear in control and in HIP/PAP expressing cells. Detailed confocal analysis (Fig. 2C) showed that these proteins partly colocalized, suggesting their presence in a same subcellular compartment.

The locations of HIP/PAP and RIIa were further analyzed using a fractionation method (Fig. 1B). The regulatory RIIa and the catalytic Ca subunits of PKA were detected in the 100 000 g ultracentrifugation pellet and in the supernatant indicating their presence in both soluble and particulate forms in Chang cells as reported for other cell lines [30]. HIP/PAP was recovered associated to membranes in the pellet confirming its presence in the secretory apparatus, but also in the supernatant (23 and 28% of total HIP/PAP in two independent experiments). Presence of HIP/PAP in the soluble fraction did not result from a significant contamination of this fraction with elements of the endoplasmic reticulum, as SERCA 2, an integral protein of endoplasmic reticulum, and calreticulin, protein of the reticulum lumen, were only detected in the centrifugation pellet.

The antibodies we raised against HIP/PAP [4,5] are not suitable for immunoprecipitation experiments. Thus, using polyclonal anti-RIIa, we tested whether HIP/PAP could be coimmunoprecipitated with RIIa (Fig. 1C). HIP/PAP was recovered in the precipitate if the experiment was performed with anti-RIIa Ig, but not with a control serum. We did not detect any protein with an electrophoretic mobility similar Analysis of the HIP/PAP protein sequence revealed the presence of three potential PKA phosphorylation sites (serines 73 and 138, and threonine 153). In vitro, recombinant HIP/PAP was phosphorylated by PKA (Fig. 3A). It has been determined that phosphorylation was more efficient at 30 (cid:2)C than at lower or higher temperature (results not shown). Thus time course of recombinant HIP/PAP phos- phorylation by PKA was studied at this temperature. HIP/ PAP phosphorylation increased with the incubation time and reached a plateau. After a 2 h incubation, 0.75 mol of 32PO4 was bound to 1 mol of recombinant HIP/PAP (Fig. 3B). Whether HIP/PAP expressed in Chang cells might be phosphorylated by PKA was studied in cells transfected with pcDNA-HIP/PAPmyc. Cellular lysates were immuno- precipitated with monoclonal anti-myc Ig and the precipi- tates were further analyzed by Western blot using first polyclonal anti-HIP/PAP and then anti-phosphoserine Ig, after stripping of the membrane (Fig. 3C,D). HIP/PAP was detected by anti-HIP/PAP as a single band. When PKA was overexpressed, this antibody labelled two faint additional bands with reduced electrophoretic mobility. Anti-phospho- serine Ig labelled one protein with electrophoretic migration similar to that of the upper one detected by anti-HIP/PAP. In contrast, no extra band was detected in cells expressing the mutated form of HIP/PAPmyc where the three potential PKA phosphorylation sites were mutated to alanine. Anti- phosphothreonine did not detect any band labelled by anti- HIP/PAP in cells expressing either the wild or the mutated forms of HIP/PAPmyc (results not shown).

3816 F. Demaugre et al. (Eur. J. Biochem. 271)

(cid:1) FEBS 2004

Fig. 2. Immunofluorescence analysis of RIIa and HIP/PAP subcellular location in HIP/PAP expressing Chang cells. (A) Transiently HIP/PAP expressing cells were processed for immunofluorescence using the antibody against HIP/PAP labelled with FITC (a) or antibodies against RIIa labelled with cyanin-5 (b). Part (c) depicts a phase con- trast image of the analyzed cells. (B) Colocalization of HIP/PAP with a marker of median Golgi (CTR433). Cells were processed for double immunofluorescence using antibodies against HIP/PAP labelled with FITC (green; a) and CTR433, labelled with cyanin-5 (b). Colocaliza- tion of HIP/PAP and CTR433 is visible as yellow staining (c) when the colour images merge. (C) Colocalization of HIP/PAP with RIIa. Cells were processed for double immunofluorescence using anti HIP/PAP Ig labelled with FITC (a) and anti-RIIa Ig labelled with cyanin-5 (b). The yellow staining (c) observed when the colour images merge and the cytofluorogramme (d) demonstrate the colocalization of HIP/PAP with RIIa. Staining was analyzed by confocal laser scanning micros- copy. Image is an optical section of 0.3 lm along the z-axis.

PKA activity in Chang cells expressing HIP/PAP

Fig. 3. HIP/PAP is a substrate for PKA. (A) Recombinant HIP/PAP was incubated for 30 min at 30 (cid:2)C with the catalytic subunit of PKA and 100 lM [32P]ATP[cP] in 80 lL as described in the Materials and methods. Aliquots of incubation mixtures (2 lL) were analyzed by SDS/PAGE. [32P]HIP/PAP was detected by autoradiography (1 h at room temperature) of the gel. T, control reaction performed without HIP/PAP. (B) Time course of HIP/PAP phosphorylation. Recom- binant HIP/PAP (60 pmol) was incubated at 30 (cid:2)C with PKA and 100 lM [32P]ATP[cP] in 80 lL as described in the Materials and methods. Control incubations were performed in parallel without recombinant HIP/PAP. At indicated times, 5 lL of incubation mix- tures were spotted on phosphocellulose filters, which were treated as indicated in Materials and methods. The incorporated radioactivity was determined by scintillation counting. (C) and (D) Chang cells were cotransfected with 18 lg of either the mutant or the wild type HIP/ PAPmyc expressing vector (empty vector called Neo was used in controls), and 2 lg of PKA expressing vector when indicated. Forty- eight hours post-transfection, cells were lysed and immunoprecipitated with anti-myc mAb. Immune complexes recovered with protein G Sepharose were analyzed for by Western blotting for HIP/PAP using polyclonal anti-HIP/PAP (C) and for phosphorylated protein using polyclonal anti-phosphoserine (D). Molecular masses indicated on the right of the figures are deduced from the electrophoretic migration of molecular mass markers run in parallel with the samples.

We investigated PKA activity in two clones isolated from a Chang cell line stably expressing HIP/PAP (HIP9 and HIP4 clones), and in two clones of Chang cells stably transfected with the empty vector as controls (PC4 and PC8 clones). Protein kinase activity assayed with kemptide, a specific substrate of PKA was measured with or without 8-bromo-cAMP and PKI, respectively, activa- tor and inhibitor of PKA in order to estimate basal and overall PKA activities. Endogenous phosphotransferase activity measured without kemptide did not differ between the two groups of cells (data not shown). For the sake of convenience (see Material and methods), pooled data from the two groups of cells are presented in Fig. 4. No difference was observed between the two groups of cells when the assays were conducted with 2 lM 8-bromo-

HIP/PAP alters PKA signalling (Eur. J. Biochem. 271) 3817

(cid:1) FEBS 2004

Fig. 5. Reporter gene assays. HIP9 and PC4 clones were transfected with 5 lg DNA including TK-LUC (2 lg) or CRE-TK-LUC (2 lg) and 0.5 lg CaEV (0.5 lg) when indicated. Luciferase activity was assayed 48 h post-transfection. In each experiment, transfections were performed in triplicate for the different studied conditions. Results are expressed as mean ± SEM of four independent experiments. Student’s t-test was used to compare mean values activities determined in PC4 and HIP9.

luciferase activity by about 65% when cells were transfected with CRE-TK-LUC. That effect was no more observed when cells were cotransfected with CRE-TK-LUC and the pCaEV vector encoding for the catalytic subunit of PKA. Thus, taken together, these data indicated that HIP/PAP expression enhanced native PKA activity in Chang cells.

Discussion

Fig. 4. Protein kinase activity in HIP/PAP expressing Chang cells. Protein kinase activity was assayed with 50 lM kemptide as the sub- strate in the presence or absence of 2 lM 8-bromo cAMP and 100 lM PKI, in two clones of Chang cells stably expressing HIP/PAP (called HIP9 and HIP4) and two clones of Chang cells stably transfected with the empty vector (control clones called PC4 and PC8). Each assay was performed in triplicate. Data were obtained from eight independent experiments. (A) Protein kinase activities measured in the different conditions. (B) PKA activities: data obtained in presence of PKI were subtracted from the kinase activities measured without effector (basal PKA activity) or with 8-bromo-cAMP (overall PKA activity). Results are expressed as mean ± SEM. Student’s t-test was used to compare mean values of enzymatic activities measured under different condi- tions. NS, not statistically significant.

HIP-encoding gene has been identified by our group as a gene over-expressed in tumourous but not in normal hepatocytes. The subsequent finding that this gene was identical to the PAP I/peptide 23/Reg2-encoding gene, which controls pancreatic, pituitary and motor neurone viability and proliferation, has led to the hypothesis that this C-type lectin may play an important physiological and/or physiopathological role. The biological function of this protein in the liver is unknown. To address this issue, we therefore looked for proteins capable of interacting with HIP/PAP in hepatocellular carcinoma cells. By screening a HCC cDNA library expressed in E. coli with [32P]Flag- HMK-HIP/PAP(29–175) as a probe, we identified the regulatory RIIa subunit of PKA as being a partner of HIP/PAP. cAMP (optimal concentration to activate PKA in both groups of cells, data not shown) or with 100 lM PKI, inhibitor of PKA [31]. On the other hand phosphotrans- ferase activity assayed without any effector of PKA was increased by about 20% in HIP/PAP-expressing cells suggesting that HIP/PAP expression did not alter overall PKA activity but enhanced basal PKA activity. This effect was better disclosed when the phosphotransferase activities measured in presence of PKI, which may not be attributed to PKA, were subtracted from the data obtained in absence and presence of 8-bromo-cAMP.

The demonstration of the biological relevance of the HIP/ PAP–RIIa interaction in HIP/PAP expressing cells required to establish that the two proteins may be located in a same subcellular compartment where they might interact. Indeed there was no evidence that the RIIa regulatory subunit of PKA is expressed anywhere other than the cytosol and the cytoplasmic surfaces of membranes [29]. On the other hand accurate subcellular distribution of HIP/PAP had not been studied and thus it was considered that HIP/PAP, protein secreted via the Golgi apparatus, was probably exclusively expressed in the luminal compartment of the secretory apparatus. We showed, using immunofluorescence studies To further document the enhanced basal PKA activity observed in HIP/PAP expressing cells we examined the effects of HIP/PAP upon the expression of a gene whose promoter is under the control of PKA. The cAMP response element (CRE) present in the promoter of cyclin A2 has been shown to respond to PKA [25]. Thus using a thymidine kinase-luciferase reporter plasmid (TK-LUC) in which one copy of the cyclin A2 CRE was inserted upstream of the TK promoter (CRE-TK-LUC) we examined if the TK promo- ter was activated in HIP/PAP expressing cells. As shown in Fig. 5, expression of HIP/PAP did not alter luciferase activity in cells transfected with TK-LUC but increased

3818 F. Demaugre et al. (Eur. J. Biochem. 271)

(cid:1) FEBS 2004

and fractionation experiments, that a fraction of the cellular pool of HIP/PAP escaped the secretory pathway. Similar observations concerning the hepatitis C virus protein E2 have been recently reported [32]. E2 has previously been considered as a protein with an exclusive location in the endoplasmic reticulum [33], but in that study it was demonstrated that it also exists in the cytosol where it impairs cellular functions [32]. Thus, HIP/PAP and RIIa are both present as soluble forms in the cytosol of cells where they may interact. We have shown that they were coimmunoprecipitated in HIP/PAP-expressing cells. Thus our finding indicates that the location of HIP/PAP and RIIa is consistent with the relevance of their interaction.

in liver [37], is not altered in HIP/PAP expressing cells (results not shown). Thus, the enhanced native PKA activity may result from the impaired association of catalytic and regulatory PKA subunits. PAP 1 (referred to as Reg 2) prevents neuronal cell death using both autocrine and paracrine ways in rat [12]. Thus two nonexclusive hypothesis may be put forward to explain the effects of HIP/PAP upon PKA. HIP/PAP has been reported to promote hepatocyte adhesion [8]. Thus through its interaction with a yet unidentified receptor, it could activate adenylcyclase and thus increase cellular cAMP levels and native PKA activity. On the other hand, HIP/PAP via its interaction with RIIa might impair the association of PKA catalytic subunits with the RIIa dimer, thus increasing PKA native activity without altering overall PKA activity.

HIP/PAP has been classified in the group 7 of C-type lectins because it binds lactose and contains only one CRD [3,4]. The HIP/PAP sequence (the 146 C-terminal amino acids) present in the probe used to screen the cDNA library encompasses the CRD. E. coli does not express enzymes involved in glycosylation. Thus the interaction between HIP/PAP and RIIa is not dependent on sugar residues, suggesting that the CRD might bind both nonglycosylated and glycosylated proteins.

Whether the biological functions of HIP/PAP results from its effects upon PKA remains to be established. It is noteworthy that links between HIP/PAP and the PKA-dependent pathways have already been suggested previously. In the rat the stimulatory effect of PAP 1 on Schwann cell proliferation was reported to involve cAMP and therefore probably, PKA-dependent pathways [11]. In liver, PKA is an important regulator of numerous metabolic functions. It has been involved in the protec- tion of hepatocytes against apoptosis [38–40] and in the control of their proliferation [41–44]. Recently, it was shown that, in transgenic mice expressing human HIP/ PAP in the liver, HIP/PAP enhances liver regeneration and acts as a hepatic cytokine that combines mitogenic and anti-apoptotic functions using pathways involving PKA [16].

In conclusion, our findings lead us to propose PKA as a target for HIP/PAP, a C-type lectin and thus offer a novel mechanism for its biological activity.

Acknowledgements

We are grateful to Dr Michael Blanar for generously providing the pAR(DRI)[59/60] plasmid. We thank D. Kremsdorf and P. Soussan for helpful discussions. This work was supported by a grant from ARC number 5156 (France).

References

1. Lasserre, C., Christa, L., Simon, M.T., Vernier, P. & Brechot, C. (1992) A novel gene (HIP) activated in human primary liver cancer. Cancer Res. 52, 5089–5095.

2. Orelle, B., Keim, V., Masciotra, L., Dagorn, J.C. & Iovanna, J.L. (1992) Human pancreatitis-associated protein: messenger RNA cloning and expression in pancreatic diseases. J. Clin. Invest. 90, 2284–2291.

3. Drickamer, K. (1993) Recognition of complex carbohydrates by Ca2+-dependent animal lectins. Biochem. Soc. Trans. 21, 456– 459.

4. Christa, L., Felin, M., Morali, O., Simon, M.T., Lasserre, C., Brechot, C. & Seve, A.P. (1994) The human HIP gene, over- expressed in primary liver cancer encodes for a C-type carbohy- drate binding protein with lactose binding activity. FEBS Lett. 337, 114–118.

5. Christa, L., Pauloin, A., Simon, M.T., Stinnakre, M.G., Fontaine, M.L., Delpal, S., Ollivier-Bousquet, M., Brechot, C. & Devinoy, E. (2000) High expression of the human hepatocarcinoma-intes- tine-pancreas/pancreatic-associated protein (HIP/PAP) gene in

HIP/PAP may be a target for PKA-dependent phos- phorylation. Three potential PKA phosphorylation sites (serines 73 and 138 and threonine 153) are detected in the sequence of HIP/PAP. In vitro PKA was shown to phosphorylate recombinant HIP/PAP. Analysis of HIP/ PAP-expressing Chang cells has allowed us to determine that PKA phosphorylated a serine in the HIP/PAP protein. Indeed antiphosphoserine antibody recognized in PKA- overexpressing cells an HIP/PAP form when cells expressed wild HIP/PAP but not when PKA-phosphorylation sites of this protein were mutated to alanine. PKA-dependent phosphorylation of recombinant HIP/PAP did not alter its electrophoretic mobility (results not shown). On the other hand the antiphosphoserine antibody recognized in HIP/ PAP expressing cells, an HIP/PAP form whose electropho- retic migration was reduced, which suggests that HIP/PAP may be the target of additional post-translational modifi- cations altering its electrophoretic mobility. There is no evidence that PKA may phosphorylate proteins present in the luminal compartment of the secretory pathway. Thus it is likely that PKA phosphorylates the fraction of HIP/PAP escaping the secretory pathway. Whether phosphorylation alters HIP/PAP properties remains to be investigated. It has to be noted that the PKA-dependent phosphorylation pattern remains unexplored and has to be determined to understand properties of HIP/PAP. However, as demon- strated for other lectins such as galectin 3 [34–36] HIP/PAP phosphorylation might alter its biological properties.

PKA regulatory subunits control the release of catalytic subunits from the inactive tetramer complex upon binding of cAMP to the regulatory subunit-dimer. Thus, we examined whether PKA activity was altered in cells expressing HIP/PAP. Two independent methods were used to address this question: assay of PKA activity and study of the expression of a gene whose promoter contains a sequence responding to PKA. These approaches gave consistent results and allowed us to conclude that HIP/ PAP did not alter overall PKA activity but increased native PKA activity. The expression of the Ca catalytic, and the RIa and RIIa regulatory subunits of PKA, well represented

HIP/PAP alters PKA signalling (Eur. J. Biochem. 271) 3819

(cid:1) FEBS 2004

20. Foulkes, N.S., Borrelli, E. & Sassone-Corsi, P. (1991) CREM gene: use of alternative DNA-binding domains generates multiple antagonists of cAMP-induced transcription. Cell 64, 739–749.

the mammary gland of lactating transgenic mice. Secretion into the milk and purification of the HIP/PAP lectin. Eur J. Biochem. 267, 1665–1671.

21. Heilmann, C., Spamer, C. & Gerok, W.

6. Christa, L., Simon, M.T., Brezault-Bonnet, C., Bonte, E., Carnot, F., Zylberberg, H., Franco, D., Capron, F., Roskams, T. & Bre- chot, C. (1999) Hepatocarcinoma-intestine-pancreas/pancreatic associated protein (HIP/PAP) is expressed and secreted by proliferating ductules as well as by hepatocarcinoma and cho- langiocarcinoma cells. Am. J. Pathol. 155, 1525–1533.

(1985) Reaction mechanism of the calcium-transport ATPase in endoplasmic reticulum of rat liver: demonstration of different reactive forms of the phosphorylated intermediate. J. Biol. Chem. 260, 788–794. 22. MacLennan, D.H., Rice, W.J. & Green, N.M. (1997) The mechanism of Ca2+ transport by sarco (endo) plasmic reticulum Ca2+-ATPases. J. Biol. Chem. 272, 28815–28818.

23. Michalak, M., Corbett, E.F., Mesaeli, N., Nakamura, K. & Opas, M. (1999) Calreticulin: one protein, one gene, many functions. Biochem. J. 344 Part 2, 281–292.

7. Lasserre, C., Colnot, C., Brechot, C. & Poirier, F. (1999) HIP/PAP gene, encoding a C-type lectin overexpressed in primary liver cancer, is expressed in nervous system as well as in intestine and pancreas of the postimplantation mouse embryo. Am. J. Pathol. 154, 1601–1610.

24. Tortora, G., Yokozaki, H., Pepe, S., Clair, T. & Cho-Chung, Y.S. (1991) Differentiation of HL-60 leukemia by type I regulatory subunit antisense oligodeoxynucleotide of cAMP- dependent protein kinase. Proc. Natl Acad. Sci. USA 88, 2011– 2015.

8. Christa, L., Carnot, F., Simon, M.T., Levavasseur, F., Stinnakre, M.G., Lasserre, C., Thepot, D., Clement, B., Devinoy, E. & Brechot, C. (1996) HIP/PAP is an adhesive protein expressed in hepatocarcinoma, normal Paneth, and pancreatic cells. Am. J. Physiol. 271, G993–G1002.

25. Desdouets, C., Matesic, G., Molina, C.A., Foulkes, N.S., Sassone- Corsi, P., Brechot, C. & Sobczak-Thepot, J. (1995) Cell cycle regulation of cyclin A gene expression by the cyclic AMP- responsive transcription factors CREB and CREM. Mol. Cell Biol. 15, 3301–3309.

9. Chakraborty, C., Katsumata, N., Myal, Y., Schroedter, I.C., Brazeau, P., Murphy, L.J., Shiu, R.P. & Friesen, H.G. (1995) Age- related changes in peptide-23/pancreatitis-associated protein and pancreatic stone protein/reg gene expression in the rat and regu- lation by growth hormone-releasing hormone. Endocrinology. 136, 1843–1849.

26. Oda, K., Hirose, S., Takami, N., Misumi, Y., Takatsuki, A. & Ikehara, Y. (1987) Brefeldin A arrests the intracellular transport of a precursor of complement C3 before its conversion site in rat hepatocytes. FEBS Lett. 214, 135–138.

10. Tachibana, K., Marquardt, H., Yokoya, S. & Friesen, H.G. (1988) Growth hormone-releasing hormone stimulates and somatostatin inhibits the release of a novel protein by cultured rat pituitary cells. Mol. Endocrinol. 2, 973–978.

11. Livesey, F.J., O’Brien, J.A., Li, M., Smith, A.G., Murphy, L.J. & (1997) A Schwann cell mitogen accompanying

27. Blouin, R., Grondin, G., Beaudoin, J., Arita, Y., Daigle, N., Talbot, B.G., Lebel, D. & Morisset, J. (1997) Establishment and immunocharacterization of an immortalized pancreatic cell line derived from the H-2Kb-tsA58 transgenic mouse. In Vitro Cell Dev. Biol. Anim. 33, 717–726.

Hunt, S.P. regeneration of motor neurons. Nature 390, 614–618.

28. Jasmin, B.J., Cartaud, J., Bornens, M. & Changeux, J.P. (1989) Golgi apparatus in chick skeletal muscle: changes in its distribu- tion during end plate development and after denervation. Proc. Natl Acad. Sci. USA 86, 7218–7222.

12. Nishimune, H., Vasseur, S., Wiese, S., Birling, M.C., Holtmann, B., Sendtner, M., Iovanna, J.L. & Henderson, C.E. (2000) Reg-2 is a motoneuron neurotrophic factor and a signalling intermediate in the CNTF survival pathway. Nat. Cell Biol. 2, 906–914.

29. Keryer, G., Skalhegg, B.S., Landmark, B.F., Hansson, V., Jahn- sen, T. & Tasken, K. (1999) Differential localization of protein kinase A type II isozymes in the Golgi-centrosomal area. Exp. Cell Res. 249, 131–146.

13. Dusetti, N.J., Ortiz, E.M., Mallo, G.V., Dagorn, J.C. & Iovanna, J.L. (1995) Pancreatitis-associated protein I (PAP I), an acute phase protein induced by cytokines: identification of two func- tional interleukin-6 response elements in the rat PAP I promoter region. J. Biol. Chem. 270, 22417–22421.

30. Martin, M.E., Hidalgo, J., Vega, F.M. & Velasco, A. (1999) Tri- meric G proteins modulate the dynamic interaction of PKAII with the Golgi complex. J. Cell Sci. 112, 3869–3878.

14. Ortiz, E.M., Dusetti, N.J., Vasseur, S., Malka, D., Bodeker, H., Dagorn, J.C. & Iovanna, J.L. (1998) The pancreatitis-associated protein is induced by free radicals in AR4–2J cells and confers cell resistance to apoptosis. Gastroenterology 114, 808–816.

31. Cheng, H.C., van Patten, S.M., Smith, A.J. & Walsh, D.A. (1985) An active twenty-amino-acid-residue peptide derived from the inhibitor protein of the cyclic AMP-dependent protein kinase. Biochem. J. 231, 655–661.

15. Katsumata, N., Chakraborty, C., Myal, Y., Schroedter, I.C., Murphy, L.J., Shiu, R.P. & Friesen, H.G. (1995) Molecular cloning and expression of peptide 23, a growth hormone-releasing hormone-inducible pituitary protein [see comments]. Endocrinol- ogy 136, 1332–1339.

32. Pavio, N., Taylor, D.R. & Lai, M.M. (2002) Detection of a novel unglycosylated form of hepatitis C virus E2 envelope protein that is located in the cytosol and interacts with PKR. J. Virol. 76, 1265– 1272.

33. Cocquerel, L., Meunier, J.C., Pillez, A., Wychowski, C. & Dubuisson, J. (1998) A retention signal necessary and sufficient for endoplasmic reticulum localization maps to the transmembrane domain of hepatitis C virus glycoprotein E2. J. Virol. 72, 2183– 2191.

16. Simon, M.T., Pauloin, A., Normand, G., Lieu, H.T., Mouly, H., Pivert, G., Carnot, F., Tralhao, J.G., Brechot, C. & Christa, L. (2003) HIP/PAP stimulates liver regeneration after partial hepatectomy and combines mitogenic and anti-apoptotic functions through the PKA signaling pathway. FASEB J. 17, 1441–1450.

17. Blanar, M.A. & Rutter, W.J. (1992) Interaction cloning: identifi- cation of a helix-loop-helix zipper protein that interacts with c-Fos. Science 256, 1014–1018.

34. Mazurek, N., Conklin, J., Byrd, J.C., Raz, A. & Bresalier, R.S. (2000) Phosphorylation of the beta-galactoside-binding protein galectin-3 modulates binding to its ligands. J. Biol. Chem. 275, 36311–36315.

18. Singh, H., Clerc, R.G. & LeBowitz, J.H. (1989) Molecular cloning of sequence-specific DNA binding proteins using recognition site probes. Biotechniques 7, 252–261.

35. Huflejt, M.E., Turck, C.W., Lindstedt, R., Barondes, S.H. & Leffler, H. (1993) L-29, a soluble lactose-binding lectin, is phos- phorylated on serine 6 and serine 12 in vivo and by casein kinase I. J. Biol. Chem. 268, 26712–26718.

36. Cowles, E.A., Agrwal, N., Anderson, R.L. & Wang, J.L. (1990) Carbohydrate-binding protein 35. Isoelectric points of the poly-

19. Vinson, C.R., LaMarco, K.L., Johnson, P.F., Landschulz, W.H. & McKnight, S.L. (1988) In situ detection of sequence-specific DNA binding activity specified by a recombinant bacteriophage. Genes Dev. 2, 801–806.

3820 F. Demaugre et al. (Eur. J. Biochem. 271)

(cid:1) FEBS 2004

deoxyribonucleic acid synthesis in liver. J. Biol. Chem. 250, 3602– 3606.

peptide and a phosphorylated derivative. J. Biol. Chem. 265, 17706–17712.

37. Skarpen, E., Thoresen, G.H., Tasken, K., Samuelsen, J.T., Jahn- sen, T., Schwarze, P.E. & Huitfeldt, H.S. (1998) Localization of cAMP-dependent signal transducers in early rat liver carcino- genesis. Histochem. Cell Biol. 109, 203–209.

42. Desdouets, C., Thoresen, G.H., Senamaud-Beaufort, C., Christ- offersen, T., Brechot, C. & Sobczak-Thepot, J. (1999) cAMP- dependent positive control of cyclin A2 expression during G1/S transition in primary hepatocytes. Biochem. Biophys. Res. Com- mun. 261, 118–122.

38. Fladmark, K.E., Gjertsen, B.T., Doskeland, S.O. & Vintermyr, O.K. (1997) Fas/APO-1 (CD95)-induced apoptosis of primary hepatocytes is inhibited by cAMP. Biochem. Biophys. Res. Com- mun. 232, 20–25.

39. Webster, C.R. & Anwer, M.S. (1998) Cyclic adenosine mono- phosphate-mediated protection against bile acid-induced apopto- sis in cultured rat hepatocytes. Hepatology 27, 1324–1331.

43. Vintermyr, O.K., Mellgren, G., Boe, R. & Doskeland, S.O. (1989) Cyclic adenosine monophosphate acts synergistically with dexa- methasone to inhibit the entrance of cultured adult rat hepato- cytes into S-phase: with a note on the use of nucleolar and extranucleolar [3H]-thymidine labelling patterns to determine rapid changes in the rate of onset of DNA replication. J. Cell Physiol. 141, 371–382.

40. Li, J., Yang, S. & Billiar, T.R. (2000) Cyclic nucleotides suppress tumor necrosis factor alpha-mediated apoptosis by inhibiting caspase activation and cytochrome c release in primary hepato- cytes via a mechanism independent of Akt activation. J. Biol. Chem. 275, 13026–13034.

44. Mellgren, G., Bruland, T., Doskeland, A.P., Flatmark, T., Vintermyr, O.K. & Doskeland, S.O. (1997) Synergistic anti- proliferative actions of cyclic adenosine 3¢,5¢-monophosphate, interleukin-1beta, and activators of Ca2+/calmodulin-dependent protein kinase in primary hepatocytes. Endocrinology 138, 4373– 4383.

41. Short, J., Tsukada, K., Rudert, W.A. & Lieberman, I. (1975) Cyclic adenosine 3¢,5¢-monophosphate and the induction of