HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 38, 35240-35247, September 20, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/38/35240    most recent M201297200v1 Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Linden, H. M. Articles by Kaushansky, K. Search for Related Content PubMed PubMed Citation Articles by Linden, H. M. Articles by Kaushansky, K. Social Bookmarking                      What's this?

The Glycan Domain of Thrombopoietin (TPO) Acts in trans to Enhance Secretion of the Hormone and Other Cytokines^*

Hannah M. Linden and Kenneth Kaushansky

From the Division of Hematology, University of Washington School of Medicine, Seattle, Washington 98195

Received for publication, February 7, 2002, and in revised form, June 21, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Thrombopoietin (TPO), the primary regulator of platelet production, is composed of an amino-terminal 152 amino acids, sufficient for activity, and a carboxyl-terminal region rich in carbohydrates (183 residues) that enhances secretion of the molecule. Full-length TPO is secreted at levels 10-20-fold greater than truncated TPO. By introducing into mammalian cells a novel cDNA encoding the TPO secretory leader linked to its carboxyl-terminal domain (TPO glycan domain (TGD)), we tested whether TGD could function in trans to enhance secretion of TPO. The artificial TGD was secreted, inactive in proliferation assays, and did not inhibit TPO activity. However, when co-transfected with a cDNA encoding truncated TPO, TGD enhanced secretion 4-fold, measured by specific bioassay and immunoassay. TGD also enhanced secretion of granulocyte monocyte colony-stimulating factor and stem cell factor but did not affect the production of erythropoietin, interleukin-3, growth hormone, or of full-length TPO. To localize TGD function, we added an endoplasmic reticulum (ER) retention signal to TGD and, separately, deleted the secretory leader. Deletion of the secretory leader attenuated the secretory function of TGD, whereas addition of the ER retention signal did not alter its function. To investigate the physiologic role of TGD in folding and proteasomal protection, we tested full-length and truncated TPO in assays of protein refolding, and we examined protein stability in the presence of proteasome inhibitors. We found that truncated TGD re-folds readily and that proteasome-mediated degradation contributes to the poor secretion of truncated TPO. We conclude that TGD enhances secretion of TPO and can additionally function as an inter-molecular chaperone, in part because of its ability to prevent degradation of the hormone. The cellular location of TGD action is likely to be within the ER or earlier in the secretory pathway.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Thrombopoietin (TPO)¹ is the principal cytokine that regulates megakaryocyte development and platelet production (1). TPO acts at both early (2) and late stages of megakaryopoiesis (3, 4), alone and in synergy with other cytokines (5). The hormone also acts in synergy with erythropoietin (EPO) to stimulate erythropoiesis (6). Subsequent studies have revealed TPO to be both necessary and sufficient for full MK1 maturation (6). As a therapeutic agent TPO has been shown to speed platelet recovery following myelosuppressive therapy in cancer patients receiving chemotherapy (7, 8). In addition, the biological effects of TPO are not limited to the MK lineage; the growth of erythroid (burst forming unit-erythrocyte) and myeloid (colony forming unit-granulocyte macrophage) colony-forming cells is also expanded by TPO in vitro, and its use in normal and myelosuppressed mice and non-human primates leads to enhanced recovery of multiple hematopoietic lineages (9, 10). Finally, TPO affects the survival and proliferation of primitive hematopoietic stem cells in vitro (11, 12) and in vivo (13). In this manner, clinical benefit may be derived from the administration of TPO to enhance peripheral blood stem cell collection (14) and the ex vivo expansion of primitive hematopoietic cells (15). TPO may thus offer patients with primary and acquired hematological disorders a significant therapeutic benefit.

Thrombopoietin is a 335-amino acid polypeptide produced in multiple organs, first cloned as a ligand for the orphan hematopoietic cytokine receptor proto-oncogene c-mpl. Like EPO, GH, and other members of the hematopoietic cytokine family, the amino-terminal region of TPO is predicted to fold into a left-handed four-helix bundle protein. The amino acids of this domain of TPO (residues 1-152) share greater homology with EPO than does any other pair of hematopoietic cytokines (22% identity, and an additional 24% sequence similarity (16)). Our laboratory and others (17-19) have demonstrated that the amino-terminal (or cytokine-like) domain of TPO is adequate for receptor binding, signaling, and supporting cellular proliferation.

In addition to the four-helical bundle domain, and unlike all other known members of the hematopoietic cytokine family, the TPO gene encodes a carboxyl-terminal polypeptide extension (residues 153-335), a serine- and a proline-rich domain remarkable for abundant carbohydrate modification, and a lack of homology to other known proteins. Glycosylation of this carboxyl-terminal (or glycan) domain has been experimentally determined (residues 153-246) and predicted (residues 246-332) in human TPO (20). The glycosylation of TPO accounts for approximately one-half of its observed 70-kDa molecular mass. Not surprisingly, the inter-species (mouse-human) homology of TPO is greatest in the amino-terminal receptor-binding domain (93%); however, the carboxyl-terminal domain retains 74% homology, suggesting that it also serves an important physiologic function.

In previous work our laboratory and others (19, 21, 22) have shown that the TPO glycan domain (TGD) functions to enhance secretion; in our hands deletion of the TGD reduced TPO secretion ~20-fold. Interestingly, several bacterial proteases display a two-domain structure, a protease domain, and a proregion essential for secretion, shown to facilitate the folding of the parent compound prior to extracellular proteolytic cleavage (reviewed in Refs. 23 and 24). Furthermore, several of these proregions have been studied and shown to function both in cis (covalently linked to their corresponding enzyme) and in trans (when co-expressed from a different gene). Co-transfection of cDNAs encoding a normal proregion with a cDNA encoding a protein either lacking a proregion or with a mutated proregion can rescue the protein from aggregation or destruction and hence facilitate protein secretion (25-30). However, in some instances, the proregion is unable to rescue the parent protein to enhance secretion in trans (31, 32) or has only a modest effect (33). The function of mammalian proregions and their ability to function with an extended range of target proteins among protein families have been less well defined. Here we report that the TGD functions in trans to substantially enhance secretion of the truncated protein and of some, but not all, other cytokines.

The specific role(s) of proregions as well as their subcellular site of action is diverse. Notably, in another member of the hematopoietic cytokine family, the prolactin proregion is essential for secretion; mutation of the proregion results in accumulation of aggregates in the endoplasmic reticulum (34). The proregion of human neutrophil defensin is essential for secretion; the proregion of this protein is anionic (35, 59, 60), can neutralize the activity of the protein prior to cleavage, and contains a region that is essential for proper subcellular sorting to granule-like vacuoles in cells. Similarly, the proregion of von Willebrand factor (vWF) promotes inter-dimer disulfide bond formation, multimer formation, and targets vWF to storage granules in several in vitro studies (36, 37). These conclusions were confirmed by the finding that a naturally occurring mutation in the proregion of vWF resulted in impaired multimerization and secretion. In the studies reported here, we show that the activity of TGD appears to take place within the endoplasmic reticulum or earlier in the secretory pathway.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

PCR Mutagenesis-- The cDNA-encoding murine TPO (described previously in Ref. 16) was used as a template for PCR-based mutagenesis. To facilitate iodination and purification, PCR-based site-directed mutagenesis was used to add a poly(Tyr) and poly(His) terminus and to mutate an Arg-Arg (Arg-153Arg-154) potential cleavage site to Gln-Gln. Oligonucleotides were designed to generate two forms of murine TPO, full-length (TPO 1-335) and a truncated form (TPO 1-152), with a poly(Tyr) and poly(His) carboxyl terminus using a strategy we have described previously (38). The secretory efficiency (as measured by ELISA) and function (as measured by MTT assay, see below) of TPO 1-335 did not differ from that of wild type mTPO, confirming that mutation of the Arg-Arg site and addition of the poly(Tyr) and poly(His) tail were functionally silent. Site-directed mutagenesis was also used to generate a deletion mutein of TPO in which the entire receptor binding domain (residues 4-174) was removed, termed the TPO glycan domain (TGD); we then modified this construct further by using site-directed mutagenesis to construct TGD variants with a Lys-Asp-Glu-Leu endoplasmic retention signal at the carboxyl terminus (TGD-KDEL) and one lacking the secretory leader (TGD-SL). Fig. 1 illustrates the cDNA constructs utilized in this study. Each mTPO construct was cloned into the mammalian expression vector pDX (39).

View larger version (17K): [in this window] [in a new window]   Fig. 1.   Diagram of TPO muteins. The TPO and TGD constructs used in our studies are illustrated. In the center of the figure is TPO 1-335, the native murine (m) TPO, with glycan domain derivatives (lacking the entire receptor binding domain) immediately below, and two truncated yet biologically active TPO forms above. Locations of O- and N-linked residues as identified and proposed by Hoffman et al. (20) are denoted by the letters O and N. The filled diamond region denotes the secretory leader, the unfilled region denotes the receptor binding or erythropoietin (EPO)-like domain, and the gray shaded region denotes the glycan domain. The TGD mutein includes the secretory leader of TPO, the first three coding residues of the RBD, followed by residues 173-335 of mTPO, the majority (all but the first 20 residues) of the glycan domain. TGD-SL encodes the same sequence as TGD, but the secretory leader is deleted and replaced by a single Met initiation codon, and TGD-KDEL encodes the same sequence as TGD but adds an endoplasmic retention signal (Lys-Asp-Glu-Leu) to the carboxyl terminus of the construct. TPO 1-238 encompasses the RBD plus the first 86 residues of the TGD. TPO 1-152 encompasses the entire RBD, a spacer (show in diagonal markings, containing Ser, Ala, and Gly residues without identity or homology to TPO) and a poly(Tyr) and poly(His) (pYpH) tail. Note that TPO 1-152 shares no glycan domain residues with TGD. Restriction mapping and DNA sequencing verified the cDNA sequence of each TPO glycomutein and truncation, and the cDNA was subcloned into the mammalian expression vector pDX (39).

Transfection of Cell Lines-- Cytokine or TGD cDNA expression vectors were co-transfected using LipofectAMINE® (Invitrogen), Lipofectin® (Invitrogen), or calcium phosphate into the rodent fibroblast cell lines BHK with a second plasmid encoding RSV-CAT (chloramphenicol acetyltransferase) to control for transfection efficiency, used at one-tenth the concentration of the cytokine expression vector (40). Cells were then cultured in Dulbecco's modified Eagle's medium with 2% fetal calf serum and penicillin, streptomycin, and fungizone. In co-transfection experiments with TGD constructs, equal amounts of cytokine cDNA were transfected with TGD cDNA (e.g. 1 µg each for a 2-ml plate). At 24-72 h post-transfection, supernatants were collected and cells counted and lysed. Supernatants were assayed for TPO (see below) and lysates for CAT activity (see below). For all transient transfection experiments, at least two different plasmid preps were utilized. To develop stable transfected cell lines, the TGD encoding cDNA was co-transfected with a plasmid encoding dihydrofolate resistance, at one-tenth the concentration of the cytokine-containing plasmids. Cells were selected and grown in 1 µM methotrexate-containing media, and the supernatant was assayed for TGD secretion and RNA by RT-PCR to confirm the stable expression of TGD.

Metabolic Labeling-- In some experiments cDNA expression vectors were used to produce metabolically labeled proteins. Six hours following transient transfection, cultures were trypsinized and split into two plates to allow metabolic labeling of half the transfected cells, and the remaining cells were left unlabeled in standard culture medium. Eighteen hours following transfection adherent cells were washed, and the supernatant was replaced with Met- and Cys-deficient media (Invitrogen) for 1 h. The media were again replaced with Met- and Cys-deficient media containing 1% dialyzed fetal calf serum, antibiotics, and 50 µCi/ml ³⁵S-labeled Met and Cys (PerkinElmer Life Sciences) and incubated overnight. Supernatants were then collected and clarified, and cells were trypsinized, counted, and lysed for CAT assays. The parallel non-labeled culture supernatant was evaluated for mTPO by ELISA (see below).

For experiments using reticulocyte lysates (TNT® SP6 Coupled Reticulocyte Lysate System, Promega), standard protocols supplied by the manufacturer were followed (using Redivue^{TM 35}S from Amersham Biosciences for metabolic labeling). The constructs shown in Fig. 1 were cloned into pcDNA3 (Invitrogen) downstream of the Sp6 promoter. Reticulocyte lysate transcripts were electrophoretically size-fractionated on a gradient gel (NOVEX^TM, 4-20% acrylamide), which was then dried, exposed to film overnight, and then to a PhosphorImaging screen for 2 h.

Immunoprecipitation-- Two anti-peptide antibodies were generously provided by T. Kato at Kirin Pharmaceuticals; each anti-peptide antibody was directed against a region of the receptor-binding domain of rat TPO. RT1 reacts with a 20-residue region of the putative first helix (⁹PRLLNKLLRDSYLLHSRLSQ²⁸) and RT2 is directed against a 21-residue segment of the putative AB loop (⁴⁶FSLGEWKTQTEOSKAQDILGA⁶⁶). Murine and rat sequences are highly conserved in this region differing at only 2 and 1 residues, respectively (anti-peptide antibodies directed against similar regions of human TPO have been described previously (41)). Immunoprecipitation of metabolically labeled tissue culture supernatants was performed using standard protocols. Briefly, 1 ml of metabolically labeled tissue culture supernatant was pre-cleared by incubation with 20 µl of protein A-agarose beads (Santa Cruz Biotechnology) for 1/2 h; phenylmethylsulfonyl fluoride, leupeptin, and aprotinin were added to diminish proteolysis (as described previously (42)), and 1% Tween 20 was used to diminish nonspecific binding. Each supernatant was then incubated overnight with 10 µg/ml of the anti-TPO peptide antibody at 4 °C. Protein A-agarose beads were added to precipitate the antibody-TPO complex, and the beads were washed with RIPA buffers as described previously (19) and then denatured and eluted from staphylococcus A beads by boiling in a loading buffer with 2% -mercaptoethanol. The immunoprecipitated TPO forms were size-fractionated by electrophoresis through 10% polyacrylamide gels, soaked in Amplify^TM enhancer (Amersham Biosciences), dried down, and exposed to PhosphorImaging screens overnight.

Biological Activity Assay-- Baf/3 cells transfected with the mouse Mpl receptor were used to assay for TPO activity (16). Cells were grown and maintained in IL-3, washed, plated at a concentration of 10,000 cells per well in 100 µl of media in 96-well plates, and incubated for 36 h with serial dilutions of tissue culture supernatants. Wild type mTPO supernatant of known concentration was used to generate a standard curve and determine maximal proliferation of Baf/Mpl cells for each assay. 3[4,5-Dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide, (MTT Sigma) was then added and was followed 5 h later by a lysis buffer. Optical density was measured by using an ELISA plate reader (EL-340 Biotek Instruments) to measure the absorbance at 570-630 nm in order to assess the intracellular conversion of tetrazolium to formazan. Sample activity was determined by comparison to the standard at half-maximal proliferation. The cell line used for these experiments was more sensitive to IL-3 than to TPO; however, as measurement of response to TPO was our desired effect, we have based our results on percentage of mTPO-induced maximal activity. Similarly, MO7e cells, which constitutively express c-Kit and GM-CSF (and a low level of Mpl) receptors and proliferate in response to SCF and GM-CSF, were used to assay for these cytokines. To test SCF activity selectively, Baf/murine Kit receptor cells were also utilized in MTT assays. Baf/EPO receptor cells were used to measure EPO activity, and Baf/3 cells were used to measure IL-3 activity in similar assays. In a similar fashion, megakaryocyte assays were used to quantify SDF-1, as described previously (43).

Folding Assay-- The activation of mitogen-activated protein kinase (MAPK) in Baf/Mpl cells was utilized to read out evidence of complete refolding of denatured TPO, in a similar fashion to the assay employed by Baker et al. (44) to demonstrate refolding of -lytic protease. Full-length TPO was generously provided by Dr. Don Foster (Zymogenetics Inc., Seattle, WA), and truncated hormone (containing the amino-terminal 162 residues) was generously provided by Dr. Takashi Kato (Kirin Pharmaceuticals, Takasaki, Japan). Equipotent solutions of truncated and full-length hormone were verified by equal activity in MTT assays using Baf/Mpl cells, as described above. Each protein was then denatured in 6 M guanidine hydrochloride (GdnHCl) at room temperature overnight and heated to 65 °C for 10 min to fully denature the protein (as described previously (45)) A roughly 0.8 M solution of the protein was then diluted into cold phosphate-buffered saline with 0.01% bovine serum albumin at room temperature to effectively dilute the protein concentration and GdnHCl 200-fold. The protein was then allowed to refold over a time course of 0-30 min, and aliquots of the refolded protein were then immediately added to starved Baf/Mpl cells, resulting in a final concentration of 0.015 mM GdnHCl. 1 × 10⁶ starved Baf/Mpl cells in serum-free media were incubated with diluted protein samples for 2 min in a total volume of 10 ml at 37 °C and then immediately poured into 40 ml of ice-cold phosphate-buffered saline; the cells were pelleted and lysed in freshly prepared lysis buffer (46) and frozen at 0 °C. Whole cell lysates were size-fractionated by electrophoresis through 8% polyacrylamide gels and transferred to nitrocellulose, blocked with 3% bovine serum albumin, and probed for phosphorylated (activated) MAPK (Cell Signaling antibody number 9101L), and MAPK protein (Cell Signaling antibody number 9109), as we have described previously (47). Repeated optimization assays demonstrated that a minimum of 2 min at 37 °C was necessary for visualization of MAPK phosphorylation and that the addition of trace quantities of GdnHCl to native protein and starved cells did not impair MAPK activation.

Proteasome Inhibitor Assay-- The proteasome inhibitor ALLN was used to determine whether intracellular protein degradation might account for the reduced secretion of truncated forms of TPO. COS cells were tested for tolerance to ALLN (Calbiochem 108909 Calpain Inhibitor dissolved in Me₂SO), and a time course of TPO expression following transient transfection was examined to determine the optimal timing of metabolic labeling and proteasome inhibition. COS cells were transiently transfected as described above, with truncated and full-length TPO in parallel; serum-free, liposome-, and TPO plasmid-containing media were washed off the cells, and they were allowed to recover at 37 °C overnight in media with 2% fetal calf serum. The following day the cells were washed with phosphate-buffered saline and Cys- and Met-deficient media were added with radiolabeled ³⁵S as described above. Following a 4-h incubation with labeled culture medium, ALLN or diluent (Me₂SO) was then added to the culture to achieve a final concentration of 50 mM, and cells were incubated additionally at 37 °C for 12 h. Supernatants were collected, and cells were harvested and lysed as described above. Similarly, supernatants and cell lysates were analyzed by immunoprecipitation (IP), size fractionation, and PhosphorImaging quantitation as described above.

Reporter Gene Assay-- Standard protocols were used to assay CAT activity (40, 48) and thereby correct for transfection efficiency. Cellular lysates were diluted in 0.25 M Tris and incubated with acetyl-CoA and [¹⁴C]chloramphenicol. Following ethyl acetate extraction, thin layer chromatography was performed to allow quantitation of chloramphenicol acetylation. If transfection efficiency varied by more than 4-fold between samples, we did not include the results from corresponding supernatants in our analysis.

Immune Assay-- Supernatants were tested in paired samples using a commercial ELISA for mTPO (MTP00, Quantikine^TM M from R & D Systems, Minneapolis, MN); in this immunoassay the lower limit is 62.5 pg/ml. The antibodies for this ELISA were also purchased separately and their TPO-binding epitopes were mapped, using purified truncated forms of mTPO, to identify which region(s) of TPO they detect. Both the monoclonal catch antibody and the polyclonal detection antibody detected mTPO by IP and standard Western blot techniques. The monoclonal antibody detected full-length TPO and a truncated TPO form 50 kDa in length but failed to detect shorter forms of TPO, where the polyclonal antibody detected 18-, 20-, 30-, 50-, and 70-kDa forms of TPO, folded and denatured, suggesting that at least part of the epitope of the monoclonal antibody maps to a segment of TPO in the region of amino acids 200-250. Based on the mapping experiments of the antibodies utilized by the ELISA, we did not use the assay to measure truncated TPO proteins (e.g. TPO 1-152), but we assumed that the commercial assay will measure 1 ng of TPO 1-335 or 1 ng of TGD roughly the same given the steric constraints of the antibodies and the relative sizes of the mTPO derivative proteins. Commercial assays for EPO, granulocyte-macrophage colony-stimulating factor (GM-CSF; R & D Systems), as well as human GH (chemiluminescent assay for hGH, Nichols Institute Diagnostics, San Juan Capistrano, CA) were utilized to measure the corresponding cytokines.

RT-PCR-- RNA was prepared from established BHK cell lines using standard protocols. RT-PCR was performed as described previously (49) using primers specific for the glycan domain of murine TPO, and for a housekeeping gene (glyceraldehyde-3-phospate dehydrogenase) as a control. Cell lines were established following transfection of the cDNA constructs shown in Fig. 1 and shown to have levels of mRNA well above the background of the parent cell line.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Our previous studies demonstrated that deletion of the carboxyl-terminal domain of TPO reduced secretion of the molecule by 10-20-fold. To explore the function of this discrete region of the protein on the synthesis, processing, and secretion of TPO, we expressed a protein encompassing only the TPO secretory leader and the carboxyl-terminal domain of the protein (TGD) in mammalian cells and assessed its effect on TPO production.

TGD Is Secreted and Is Inactive as a Thrombopoietin-- Although transiently transfected cells did not secrete enough TGD to be measured by our immunologic methods (ELISA or by ³⁵S labeling and IP), stably transfected cell lines engineered to express TGD (with and without the addition of 4 Met residues) expressed and secreted enough protein to be readily detected by the commercial ELISA, 200-500 pg/ml. We tested the supernatant from our highest TGD expressing cell line in MTT assays alone and with TPO for inhibitory or stimulatory effects. TGD alone did not stimulate proliferation of the TPO receptor bearing cells, Baf/Mpl, or of MO7e cells. To test for inhibitory effects of TGD, we tested the TGD supernatant in proliferation assays in which half-maximal amounts of TPO had been added; addition of TGD (in over 25-fold molar excess) did not inhibit TPO 1-152 or TPO 1-335-induced activity. TGD also failed to inhibit IL-3 stimulation of Baf3 cells or GM-CSF-induced proliferation of MO7e cells in similar assays.

TGD Enhances Secretion of TPO, GM-CSF, and SCF in Trans-- To test TGD for its effects on TPO production, we transiently co-transfected TGD with cDNA expression vectors for full-length and truncated TPO, GM-CSF, SCF, IL-3, EPO, hGH, and SDF-1, and we measured expression of each cytokine using bioassays, immunoassays, and metabolic labeling experiments. Each experiment was performed using either TGD co-transfection or co-transfection of a sham plasmid of the same parent vector encoding an irrelevant cytokine. TGD enhanced secretion of truncated TPO (TPO 1-152) 4-fold, measured by specific bioassay (Table I). By using ³⁵S metabolic labeling, we also found a similar level of enhanced secretion of TPO 1-152 in the supernatants of cells transiently transfected with TGD and TPO 1-152, compared with supernatants from cells transiently transfected with TPO 1-152 and a sham plasmid (Fig. 2). TGD also enhanced secretion of immunoreactive and bioactive hGM-CSF 3-4-fold, and bioactive mSCF 4-5-fold but did not enhance secretion of bioactive or immunoreactive mEPO, bioactive IL-3, or immunoreactive hGH (see Table I).

View larger version (40K): [in this window] [in a new window]   Fig. 2.   Secretion of TPO and truncated forms with and without TGD. Shown are representative PhosphorImaging exposures of immunoprecipitation experiments (using an anti-peptide antibody directed against the receptor binding domain of TPO) of 35S-labeled TPO from supernatants of BHK cells transiently transfected with TPO and either a sham cytokine or TGD. The precipitates were size-fractionated by polyacrylamide electrophoresis, dried down, and exposed to a PhosphorImaging screen for 2-24 h. Each truncated TPO form is shown in comparison with the full-length hormone. A shows TPO 1-335 co-transfected with a sham cytokine (1st lane) and with TGD (2nd lane). B shows TPO 1-335 (1st lane), TPO 1-238 with sham cytokine (2nd lane), TPO 1-238 co-transfected with TGD (3rd lane), TPO 1-152 with sham co-transfection (4th lane), and TPO 1-152 with co-transfection of TGT (5th lane). TPO 1-238 alone and with TGD is more apparent at differing exposures and was well above background by PhosphorImaging techniques. TPO 1-152 co-transfected without TGD was barely detectable above background by PhosphorImaging; however, crude unlabeled supernatant collected in parallel from all TPO forms, with and without TGD was active in proliferation assays. Secretion rates (normalized for cell count, transfection efficiency, and number of radiolabeled residues per molecule of protein) relative to TPO 1-335 measured by PhosphorImaging (Amersham Biosciences PhosphorImager, using ImageQuant) were 16.5% for TPO 1-238 plus sham cytokine, 36% for TPO 1-238 plus TGD, 0.75% for TPO 1-152 and sham cytokine, and 4.75% for TPO 1-152 plus TGD.

To test whether the function of TGD persisted when the domain is embedded in its native state (i.e. as part of the full-length hormone), we examined the effects of co-expressing TPO 1-335 with GM-CSF or SCF. The secretion of GM-CSF (as measured by immunoassay) and SCF (measured by Baf3/kit cells) was no different with and without co-transfection of TPO 1-335 (data not shown).

To determine whether TGD could enhance secretion of molecules already containing some or all of the TGD region, we tested whether co-transfection of a TGD cDNA expression vector augmented secretion of a series of TPO truncation muteins, TPO 1-335, TPO 1-238, and TPO 1-152. As shown in Fig. 2 and Table I, TGD enhanced secretion of TPO fully devoid of TGD (TPO 1-152), but inclusion of 64 residues (TPO 1-238) reduced and inclusion of the full-length proregion TPO 1-335 eliminated the favorable effect of TGD on TPO secretion.

To begin to identify the cellular site at which TGD enhances TPO secretion, similar co-transfection experiments were conducted with a TGD that remains tethered in the endoplasmic reticulum by virtue of a tetrapeptide localization signal (KDEL (50)) or one that fails to enter the secretory pathway by elimination of the SL sequence (Fig. 1). As anticipated, the mTPO ELISA did not detect TGD in the culture supernatants from these cell lines or from transient transfections. However, these cell lines were shown to have abundant TGD message by RT-PCR (data not shown). The TPO 1-152 expression vector was co-transfected with TGD-KDEL or a sham vector and into BHK cells, and TPO secretion was monitored. We found that targeting TGD to the endoplasmic reticulum did not affect TPO 1-152 secretion (see Table II). In contrast, elimination of the SL from TGD substantially reduced the capacity of the domain to enhance secretion.

TGD Constructs in Cell Lines Confirm the Results of Transient Co-transfection Assays-- The above results were obtained using transient co-transfection assays. To examine further the capacity of TGD to affect TPO secretion, four BHK cell lines stably expressing TGD at a low level (TGD-L), high level (TGD-H), without a secretory leader (TGD-SL), and one with an endoplasmic reticulum retention signal (TGD-KDEL) were established and then transiently transfected with several cytokine expression vectors, and the levels of the encoded cytokines were compared with control BHK cells. As expected, TGD could be detected at low and high levels from the TGD lines by ELISA; however, no immunoreactive material was detected in the supernatants of BHK cells stably expressing TGD-SL, TGD-KDEL, or control BHK cells prior to transfection with TPO expression vectors. We found that transiently transfected TPO 1-335 was secreted equally well from all 5 types of cell lines, as measured by bioassay (see Table III). However, transient transfection of TPO 1-152 resulted in a 1.8-fold greater secretion of TPO 1-152 in TGD-L BHK cells, 4.1-fold greater in TGD-H BHK cells, and 4-fold greater in TGD-KDEL than in control BHK cells. In contrast, TGD-SL cells expressed only 1.4-fold more TPO 1-152 than control BHK cells in these experiments. Similarly, transient transfection of SCF and GM-CSF expression vectors into these same cells led to 3.8- and 4.1-fold enhanced secretion of GM-CSF and 2.7- and 3.0-fold enhanced secretion of SCF in TGD cells and TGD-KDEL cells, respectively (see Table IV). In contrast, in TGD-SL cells, only 2.3- and 1.5-fold enhanced secretion was appreciated of GM-CSF and SCF, respectively.

In Vitro Translation of TPO-- All of the above experiments were conducted in cell lines, either transiently or in a stably expressing subline of BHK cells. To determine whether the reduced secretion characteristic of TPO 1-152 (compared with TPO 1-335) is an intrinsic property of the encoded polypeptide, or related to a post-translational processing event related to the secretory pathway, we utilized a rabbit reticulocyte lysate to translate TPO 1-152, TPO 1-238, and TPO 1-335 proteins. As shown in Table V and Fig. 3, we found that in contrast to both the 10-20-fold enhanced secretion observed between full-length and truncated TPO (19), and the 4-fold enhanced secretion of TPO 1-152 co-transfected with TGD or TGD-KDEL in intact mammalian cell lines), TPO 1-335 was translated only 2-fold better than TPO 1-152 in rabbit reticulocyte lysates. Thus we conclude that enhanced mRNA translation alone does not account for the enhanced secretion of TPO 1-152 with TGD in cis or in trans.

View larger version (39K): [in this window] [in a new window]   Fig. 3.   Production of TPO and truncated TPO by reticulocyte lysates. Rabbit reticulocyte lysates were utilized to express TPO 1-335 (1st lane, predicted molecular mass of 40 kDa), TPO 1-238 (2nd lane, predicted molecular mass of 28 kDa), and TPO 1-152 (4th lane, predicted molecular mass of 23 kDa) in the presence of 35S, and then electrophoretically sized on a gradient gel (4-20%), and quantitated by PhosphorImaging techniques. Correcting for the anticipated number of incorporated radiolabeled residues, the relative expressions of TPO 1-335, TPO 1-238, and TGD are 100, 44, and 20%.

Folding of TPO with and without TGD-- To test potential physiologic roles of TGD, we studied the rate of refolding of truncated and full-length TPO in a surrogate folding assay based on the capacity of native TPO to phosphorylate MAPK in Baf/Mpl cells. Solutions of truncated and full-length TPO were tested in activity assays of Baf/Mpl cells to verify equal potency. Separately, the two proteins were denatured and allowed to refold in parallel for each assay, and a single starved batch of Baf/mMpl cells was utilized for read-out of protein activity. The 2-min incubation of proteins with cells resulted in a net time course from 2 to 32 min. These assays were performed 5 times, with fresh protein preparations and freshly starved Baf/Mpl cells. As shown in Fig. 4 the full-length protein re-folded at the same rate or slightly more slowly than the truncated protein.

View larger version (33K): [in this window] [in a new window]   Fig. 4.   Folding assay of full-length and truncated TPO. Purified concentrated TPO 1-335 and TPO 1-162 were separately denatured and allowed to refold over the time course as shown from 2 to 32 min. Control TPO (not previously denatured, but less concentrated) is shown in the 1st lane. Activity, manifest by initiation of MAPK phosphorylation, was used as a surrogate measure of TPO refolding and is seen as two bright bands at 44 and 42 kDa.

Proteasome Inhibition Effects on Secretion of TPO Forms-- To determine whether proteasome degradation is a significant mechanism of intracellular TPO destruction prior to secretion in our culture system, transiently transfected COS cells were incubated for 12 h in the presence of ALLN or vehicle alone (Me₂SO), both using standard cultures and radiolabeling conditions. As ALLN and cell lysates are inhibitory in biological activity assays, only immune assays were used to measure the effect of proteasome inhibition on TPO production. As shown in Fig. 5, ALLN did not significantly increase TPO 1-335 production nor increase the amount of TPO 1-152 secreted into the culture medium. These findings were confirmed by ELISA for TPO 1-335 production, and by activity assays of dialyzed supernatants from transient transfections of both TPO forms incubated with ALLN (data not shown). Interestingly, the addition of a proteasome inhibitor allowed detection of TPO 1-152 from lysates of transiently transfected cells. Although we have detected TPO forms in transfected and selected cell lines, we have been unable to detect TPO forms in transiently transfected cell lines without the addition of a proteasome inhibitor. As shown in Fig. 5, ALLN did not allow detection of TPO 1-335 in lysates.

View larger version (25K): [in this window] [in a new window]   Fig. 5.   Proteasome inhibition assay. Immunoprecipitates from 35S-labeled supernatants and lysates from COS cells transiently transfected with TPO 1-335 and TPO 1-152 and treated with diluent alone or the proteasome inhibitor ALLN are shown fractionated on separate gels. The immunoprecipitates were size-fractionated by polyacrylamide electrophoresis, dried down, and exposed a to PhosphorImaging screen for 2-24 h.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Our laboratory and others (19, 21) have previously demonstrated that the carboxyl terminus of TPO enhances secretion of the hormone and in cis modestly inhibits the activity of the protein, as would be the anticipated role of a proregion. In this work we have further characterized the function and cellular localization at which the glycan domain acts. We demonstrated that the glycan domain can function independently to enhance secretion of the receptor-binding domain of TPO and that the secretion of TGD into the culture medium is not necessary for the protein to exert its effects, as tethering of TGD to the endoplasmic reticulum does not diminish its activity. As we could not detect TGD immunologically from either transient expression or established cell lines expressing the tethered construct, TGD-KDEL, we believe this protein was effectively restricted to the ER and not secreted. Therefore, we suspect that the cellular location of the activity of TGD is within the endoplasmic reticulum, as in vitro translation does not differ substantially between TPO 1-335 and TPO 1-152, and the addition of an endoplasmic retention signal does not abrogate TGD function. Given the observed cellular location of activity, putative roles of the glycan domain in enhancing secretion of TPO include enhancing the folding of the receptor binding domain, as has been shown with several bacterial proregions, and/or reducing ER-associated degradation.

The major finding in this paper, that TGD can augment the secretion of the ligand-binding domain of TPO in trans, could be because of artifacts of our experimental detection systems. However, we do not believe this is the case. TGD alone does not stimulate proliferation of the cell lines used in our bioassay studies (Baf3/Mpl or MO7e) nor does it inhibit stimulation of growth by a number of ligands, IL-3, TPO, SCF, and GM-CSF. Hence, we do not believe that the TGD has interfered with the measurement of the secretion of these ligands in activity assays. Moreover, we have employed three types of assays (bioassay, immunoprecipitation of radiolabeled supernatants, and ELISA using different antibodies) to confirm our findings, and the effects of TGD on protein secretion are relatively specific, both within the cytokine family and among different mutants of TPO. For example, secretion of TPO 1-152 was enhanced but TPO 1-335 was not, and TGD failed to enhance secretion of hGH and EPO (both members of the hematopoietic growth factor family) or the unrelated cytokine SDF-1 (data not shown).

The mechanism by which TGD enhances secretion was also studied; we examined two potential actions of TGD, enhancing refolding of denatured TPO and protecting the protein from proteasome-mediated degradation.

Enhanced folding or stabilization of an intermediate state has been seen with several proregions (23, 51) and elegantly demonstrated in vitro, for subtilisin B and -lytic protease (44, 52). We performed similar studies, and we found that the unfolded receptor-binding domain was able to re-fold as readily as the full-length hormone, suggesting that the influence of TGD on folding is not significant. However, the kinetic limits of our refolding assay (the soonest we can assay is at 2 min) are significant. It is possible that with an assay that could follow folding more rapidly and/or at lower temperatures, significant differences would emerge between truncated TPO and the full-length hormone, TGD, in cis. Nonetheless, the lack of difference at physiologic temperatures between the two TPO forms suggests that the impact of TGD on protein folding is, at most, minor.

An additional candidate mechanism to help explain the beneficial effects of TGD on TPO secretion is that the former protects the parent hormone from proteasome-mediated proteolysis and thereby leads to greater protein secretion. ERAD occurs when misfolded proteins are translocated to the cytosol for proteasomal degradation (53). We tested the possibility that ERAD is decreased by TGD by examining the effect of proteasome inhibition on COS cells transiently transfected with full-length and truncated TPO. We found that the TPO 1-152, lacking TGD, was rescued by proteasome blockade, as it was detected in lysates. However, the truncated TPO was not better secreted in the presence of ALLN, suggesting that once the poorly folded protein is translocated into the cytosol, even if it is protected from degradation, it cannot be further rescued for secretion. It is also possible that our failure to detect increased secretion in the presence of a proteasome inhibitor could be due to a general reduction in metabolic function in cells treated with ALLN. Interestingly, as glycosylation has been shown to retard ERAD (54), we can speculate that TPO 1-335, with its heavily glycosylated TGD intact, deters ERAD which occurs more briskly when the TGD is not present in TPO 1-152. The limitations of our experimental methods preclude elucidation of the mechanism by which TGD prevents proteasomal degradation. Nonetheless, it is apparent from our experiments that one role of TGD is to prevent the cytokine from ERAD, and protection from this intracellular fate is a prerequisite to secretion.

The ability of a proregion to exert effects (enhanced secretion, intracellular targeting, folding, and inhibition) in trans has been demonstrated by several groups for different proteins (30, 51, 52, 55, 56). For bacterial proteases, deletion of the proregion (or mutation of the active region) results in absolute loss of secretion, unlike for TPO, in which progressive deletion of the proregion resulted in progressive loss of secretory efficiency but not an absolute loss of secretion. The range of restoration of activity by trans-complementation rescue is from 5% for yeast proteinase A (28) to 80% for aqualysin I (57), with most examples, such as Rhizopus niveus aspartic proteinase-I, Rhizopus oryzae lipase, and alkaline extracellular protease, restoring function by 30-50% (26, 27, 29). Hence, our data for enhanced secretion of truncated TPO with its proregion in trans are consistent with the enhanced levels of secretion seen with bacterial protease proregions. With the exception of TGF- and activin A, no other examples of mammalian proregions functioning in trans to enhance secretion have been reported previously. Moreover, for TGF- and activin A the restoration of secretion was far more modest than that seen in our studies with TPO. In studies similar to ours testing two members of the TGF- superfamily, Gray and Mason (33) found no secretion of activin A and TGF-1 when each protein was expressed without its respective proregion, and with the proregion expressed in trans, we observed only modest secretion (2.4 and 9%, respectively, of that when the protein is expressed in its native state with its proregion in cis).

Another characteristic feature of bacterial protease proregions is their capacity to inhibit the activity of the parent protein, exerting a regulatory function to limit activity, both prior to and upon cleavage from the parent protein. The TPO proregion also appears to inhibit activity of the full-length protein, but only weakly (~4-fold) in cis and does not appear to inhibit the protein in trans. It is possible that we have not produced enough TGD to reach a maximal inhibitory effect, but the molar ratio of TGD to TPO 1-152 in our inhibition experiments likely exceeds 100:1.

Our studies confirm the activity of the glycan domain of TPO by demonstrating an effect of TGD in trans. However, when expressed as part of a full-length TPO molecule, TGD does not enhance secretion of other cytokines nor can it additionally enhance secretion of TPO 1-335. This suggests that the functional ratio of TGD to cytokine is 1:1. Although cleavage of TPO can be demonstrated in vitro (58), it is unclear if TPO is cleaved in vivo allowing truncated TPO and TGD to co-exist under natural biologic conditions. We suspect that if TPO 1-152 and TGD were generated in vivo by proteolysis, the resulting TPO would be slightly more potent, and TGD as a disparate protein would not inhibit truncated TPO. Our experimental conditions preclude more detailed studies of the association between TPO 1-152 and TGD. From our experiments in which TGD constructs are tethered to the ER, we infer that the two proteins are associated within the ER. The association between the secreted proteins is weak, as we do not observe both proteins by co-immunoprecipitation (data not shown) and TGD does not reduce the activity of TPO, truncated TPO, or other cytokines.

In summary, we have previously identified a role for the evolutionarily unique glycan domain of thrombopoietin. Here we have demonstrated that the function of the TGD is robust enough to persist when the region is separated from the parent protein, and we have begun to characterize the promiscuity of this function and the cellular location of the activity. TGD functions to enhance secretion of some but not all members of the hematopoietic growth factor family. It is puzzling that this effect was not manifest in our experiments with EPO, the closest structural homologue of TPO. The function of TGD is minimal at the level of protein translation, as our experiments with reticulocyte lysates and a TGD cDNA expression vector lacking a secretory leader failed to show significant changes in protein production or secretion, respectively. However, TGD secretion, or progress beyond the endoplasmic reticulum, is not necessary for its function, as the addition of an endoplasmic retention signal did not diminish the effects of TGD expression in trans. Further studies in vitro may determine the precise mechanisms of TGD activity.

	FOOTNOTES

^* This work was supported by National Institutes of Health Grants K08DK02665-02 (to H. L.), R01DK49855, and R01CA31615 (to K. K.).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.

Present address and to whom correspondence should be addressed: UCSD Medical Center, 402 Dickinson St., Ste. 380, San Diego, CA 92103-8811. Tel.: 619-543-2259; E-mail: kkaushansky@ucsd.edu.

Published, JBC Papers in Press, July 5, 2002, DOI 10.1074/jbc.M201297200

	ABBREVIATIONS

The abbreviations used are: TPO, thrombopoietin; RBD, receptor binding domain; TGD, TPO glycan domain; EPO, erythropoietin; GM-CSF, granulocyte monocyte colony-stimulating factor; SCF, stem cell factor or Kit ligand; IL-3, interleukin-3; MK, megakaryocyte; TGF, transforming growth factor; SDF-1, stromal cell-derived factor 1; ERAD, endoplasmic reticulum-associated degradation; BHK, baby hamster kidney; MTT, 3[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide; ELISA, enzyme-linked immunosorbent assay; ER, endoplasmic reticulum; GdnHCl, guanidine hydrochloride; MAPK, mitogen-activated protein kinase; CAT, chloramphenicol acetyltransferase; SL, secretory leader; GH, growth hormone; hGH, human GH; IP, immunoprecipitation; RT, reverse transcriptase; vWF, von Willebrand factor; mTPO, murine TPO.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This Article Abstract Full Text (PDF) All Versions of this Article: 277/38/35240    most recent M201297200v1 Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Linden, H. M. Articles by Kaushansky, K. Search for Related Content PubMed PubMed Citation Articles by Linden, H. M. Articles by Kaushansky, K. Social Bookmarking                      What's this?