Regulation of Cellular Levels of Sprouty2 Protein by Prolyl Hydroxylase Domain and von Hippel-Lindau Proteins

Background: Sprouty2 (Spry2) inhibits the actions of receptor tyrosine kinases (RTK) during development and disease.

Results: Stability of Spry2 is regulated by prolyl hydroxylation and binding to von Hippel-Lindau protein-associated E3 ligase.

Conclusion: PHD- and pVHL-mediated regulation of cellular levels of Spry2 modulates its ability to inhibit signaling by RTKs.

Significance: These findings provide new insights into modulation of levels of Spry2 to regulate RTK actions in disease.

Abstract

Sprouty (Spry) proteins modulate the actions of receptor tyrosine kinases during development and tumorigenesis. Decreases in cellular levels of Spry, especially Sprouty2 (Spry2), have been implicated in the growth and progression of tumors of the breast, prostate, lung, and liver. During development and tumor growth, cells experience hypoxia. Therefore, we investigated how hypoxia modulates the levels of Spry proteins. Hypoxia elevated the levels of all four expressed Spry isoforms in HeLa cells. Amounts of endogenous Spry2 in LS147T and HEP3B cells were also elevated by hypoxia. Using Spry2 as a prototype, we demonstrate that silencing and expression of prolyl hydroxylase domain proteins (PHD1–3) increase and decrease, respectively, the cellular content of Spry2. Spry2 also preferentially interacted with PHD1–3 and von Hippel-Lindau protein (pVHL) during normoxia but not in hypoxia. Additionally, Spry2 is hydroxylated on Pro residues 18, 144, and 160, and substitution of these residues with Ala enhanced stability of Spry2 and abrogated its interactions with pVHL. Silencing of pVHL increased levels of Spry2 by decreasing its ubiquitylation and degradation and thereby augmented the ability of Spry2 to inhibit FGF-elicited activation of ERK1/2. Thus, prolyl hydroxylase mediated hydroxylation and subsequent pVHL-elicited ubiquitylation of Spry2 target it for degradation and, consequently, provide a novel mechanism of regulating growth factor signaling.

Introduction

The four mammalian Sprouty proteins (Spry1–Spry4) are products of different genes and are orthologs of the originally discovered Drosophila Sprouty (1, 2). Spry² proteins regulate the actions of receptor tyrosine kinases and therefore play a major role in the development of different organs, including kidneys, lungs, limb buds, and process of angiogenesis (3–7). During development, the expression of Spry proteins is augmented at the centers of growth factor signaling such as periphery of limb buds and tips of developing trachea or blood vessels to oppose the actions of growth factors in a negative-feedback manner (3).

Previous findings have demonstrated that Spry proteins regulate cell migration and proliferation in response to a number of growth factors (8–13). In keeping with the anti-migratory and anti-proliferative actions of Spry proteins, the levels of Spry1 and Spry2 have been shown to be decreased in breast, hepatocellular, prostate, lung, and colon cancers (14–20), and overexpression of Spry2 decreases formation of lung tumors (21). Recent studies have also shown that c-Met or activated β-catenin expression and decrease in Spry2 function synergize to promote hepatocellular carcinomas (17, 22). Moreover, a direct correlation has been observed between decreased levels of Spry2 in hepatocellular carcinomas from patients and poor prognosis, including tumor metastasis (22). Likewise, a decrease in the level of Spry2 has also been suggested to contribute toward cardiac hypertrophy (23). It is now well established that Spry proteins inhibit growth factor-mediated downstream signaling by their interactions with Raf, Grb2, elevation of phosphatase and tensin homolog activity, and also inhibiting the activity of phospholipase C (10, 13, 20, 24–30).

Given the importance of Spry proteins in development as well as regulating pathological conditions such as tumorigenesis and cardiac hypertrophy, it is important to understand the mechanisms that regulate their levels. In this context, growth factors have been shown to increase the transcription of Spry proteins (3). Epigenetic regulation of the Spry2 promoter has also been shown to alter its transcription (14, 31). The levels of Spry proteins can also be regulated by post-translational mechanisms. Previous reports have shown that growth factor-mediated phosphorylation of Spry2 on Tyr-55 creates a binding site for the Src homology 2-like phosphotyrosine binding domain of the E3 ubiquitin ligase c-Cbl (c-Cbl is Casitas B-lineage lymphoma proto-oncogene) which can then ubiquitylate Spry2 and target it for proteosomal degradation (32–34). Similarly, the E3 ubiquitin ligase Siah2 can also polyubiquitylate Spry2 and target it for degradation (35, 36). Recently, we demonstrated that a HECT domain containing E3 ubiquitin ligase, Nedd4-1, associates with Ser-phosphorylated Spry2 and targets it for degradation (37). In tumor samples derived from patients with hepatocellular carcinomas, the decrease in content of Spry2 best correlated with an increase in Nedd4-1 expression but not c-Cbl or Siah2 (22). However, not all of the patient-derived hepatocellular carcinomas had elevated Nedd4-1 levels (22). Likewise, methylation of the Spry2 promoter or loss of heterozygosity did not account for decreased content of Spry2 in all hepatocellular carcinomas (22). These findings suggest that other mechanisms are also involved in modulating levels of Spry2.

Because rapidly growing cells in developing organs and tumors experience hypoxia, we postulated that the cellular content of Spry proteins may be regulated by hypoxia. Our studies demonstrated that the levels of all expressed Spry isoforms were elevated in hypoxia. Using Spry2 as a prototype, we show that prolyl hydroxylase domain proteins (PHDs) and von Hippel-Lindau protein (pVHL), a recognition component of elongins B/C and cullin E3 ubiquitin ligase (38–43), post-translationally regulate the cellular content of Spry2. The PHD/pVHL-mediated regulation of Spry2 also alters its ability to inhibit FGF signaling. This is the first demonstration of a novel post-translational mechanism that regulates cellular content of Spry2 via PHDs and pVHL and thereby alters its ability to modulate growth factor receptor actions.

EXPERIMENTAL PROCEDURES

Chemicals and Reagents

Dimethyloxallyl glycine (DMOG) was purchased from Cayman Chemical (Ann Arbor, MI), and recombinant human fibroblast growth factor basic (FGF) was purchased from CHEMICON (Temecula, CA). Deferoxamine (DFO), cycloheximide, N-ethylmaleimide, and MG132 were obtained from Sigma. All primers were synthesized by Integrated DNA Technologies Inc. Antibodies used for Western analyses of the various proteins were from the following sources: PHD1 (Novus Biological or Oxycell), PHD3 (Novus Biological or Oxycell), PHD2 (AbCam), HIF1α (BD Transduction Laboratories), ERK1/2 (Millipore), phospho-p44/p42 ERK1/2 (T202/Y204, Cell Signaling), actin (MP Biomedicals, LLC, Ohio), β-PIX (Cell Signaling), HA tag (HA-HRP from Roche Applied Science), FLAG tag (FLAG-HRP from Sigma), c-Myc (Covance, monoclonal), and pVHL (Cell Signaling, polyclonal). All shRNA constructs were purchased from Open Biosystems (Huntsville, AL); the catalog numbers are indicated below.

DNA Constructs

The cloning of human Spry2 cDNA into the pHM6-HA vector has been previously described (8). The pCMV-3×-FLAG-PHD1 and pCMV-3×-FLAG-PHD3 plasmids were constructed from human cDNAs for EGLN2 (PHD1, MHS1010-7296053) and EGLN3 (PHD3, MHS1010–99621891) genes, respectively. The full-length gene sequence was PCR-amplified using forward (F) and reverse (R) primers that contained HindIII (boldface) and XbaI (underlined) restriction sites, respectively. The primer sequences were as follows: PHD1(F), 5′-GATACAAGCTTATGGACAGCCCGTGCCAGCCG-3′; PHD-3(F), 5′-GATACAAGCTTATGCCCCTGGGACACATCATGAGGCT-3′; PHD1(R), 5′-TGTTATCTAGACTAGGTGGGCGTAGGCGGCTGTGATAC-3′; and PHD3(R), TGTTTTCTAGATCAGTCTTCAGTGAGGGCAGATTCA-3′. Each PCR product was digested with HindIII/XbaI, purified, and ligated into the HindIII/XbaI site of pCMV-3×-FLAG expression vector (Sigma) to yield the pCMV-3×-FLAG-PHD1 or pCMV-3X-FLAG-PHD3 vectors. A pCS2+MT-6×-Myc-VHL construct was prepared from human cDNA for VHL genes (IHS1380-97652010). The full-length gene sequence was PCR-amplified using forward (F) and reverse (R) primers that contained EcoRI (boldface) and ApaI (underlined) restriction sites, respectively. Primer sequences were as follows: VHL(F), 5′-GATACGAATTCAATGCCCCGGAGGGCGGAGAAC-3′, and VHL(R), 5′-TGTTAGGGCCCTCAATCTCCCATCCGTTGATGTGC-3′. The PCR product was digested with EcoRI/ApaI, purified, and ligated into the EcoRI/ApaI site of pCS2+MT to yield the pCS2+MT-6×-Myc-VHL vector. pFLAG-PHD2 was a gift from Dr. Frank S. Lee, University of Pennsylvania, School of Medicine. pCMV-3×-FLAG-ubiquitin was kindly provided by Dr. Adriano Marchese, Loyola University Chicago. All other cDNAs were purchased from Open Biosystems (Huntsville, AL). All plasmid constructs were verified by sequencing.

Cell Culture

HEK293T and HeLa cells were cultured in high glucose DMEM supplemented with 10% FBS, penicillin (100 units/ml), and streptomycin (100 μg/ml). Cells were maintained under normoxic conditions at ambient O₂ levels (21% O₂) in a water-jacketed CO₂ incubator (Thermo Forma) at 37 °C, 5% CO₂. Cells grown under hypoxic conditions were maintained at 3% O₂ in a Coy Hypoxic Chamber (Grass Lake, Michigan) at 37 °C, 5% CO₂. All media used for hypoxia experiments was pre-equilibrated under hypoxic conditions for 24 h before use.

Stability of Spry2

HeLa cells stably expressing HA-tagged wild-type or P18A/P144A/P160A forms of Spry2 were seeded in 35-mm dishes at 300,000 cells/dish. After 24 h, media above cells were switched to normoxic or hypoxic media, and cells were grown for an additional 24 h. One dish from each set was lysed for t = 0 time point before the addition of 200 μm cycloheximide to stop protein synthesis. Thereafter, at the indicated times, cells were lysed in reducing Laemmli sample buffer, and amount of HA-Spry2 was analyzed by Western blotting.

Treatment with Hypoxia Mimetics

HeLa cells stably expressing HA-Spry2 were seeded at 300,000 cells/dish into 35-mm dishes. The next day, the media were changed, and PHD inhibitors (50 or 200 μm DMOG and 20 or 40 μm DFO) were added to cells cultured under normoxic conditions. After 24 h, cells were lysed, and levels of HA-Spry2 were monitored.

Expression of PHDs and pVHL

HeLa cells were seeded at 300,000 cells into 35-mm dishes and transfected the following day using the TransIT-HeLa MONSTER transfection kit (Mirus, Madison, WI) according to the manufacturer's instructions. Each dish received 1 μg of pCMV-3×-FLAG-PHD1, pCMV-3×-FLAG-PHD2, or pCMV-3×-FLAG-3-PHD3 constructs along with 250 ng of pHM6-HA-Spry2. After 24 h, the media were changed with that equilibrated under normoxia (21% O₂) or hypoxia (3% O₂), and cells were incubated for an additional day under normoxic or hypoxic conditions before collecting lysate. For studies with pVHL, HEK293T cells were seeded at 200,000 cells/dish and co-transfected 24 h later with 250 ng of pHM6-HA-Spry2 plus increasing amounts of pCS2+MT-6×-Myc-VHL vector (from 62.5 ng to 1.5 μg) using FuGENE 6 transfection reagent (Roche Applied Science) according to the manufacturer's instructions. Total plasmid concentration was kept constant by adding appropriate amounts of empty pCS2+MT vector. Lysates were prepared in reducing Laemmli buffer and analyzed by immunoblotting.

Silencing of Endogenous PHD1, PHD3, and pVHL

HEK293T cells (200,000/35-mm dish) were co-transfected with 250 ng of pHM6-HA-Spry2 and one of the shRNAs (1 μg each) against PHD1 (shRNA#26, catalog no. RHS3979-9589734, or shRNA#27, catalog no. RHS3979-9589735) or PHD3 (shRNA#47, catalog no. RHS3979-9569478, or shRNA#50, catalog no. RHS3979-9569481) using FuGENE 6 transfection reagent. Media were changed the following day, and cell lysates were collected after another 48 h. The nonsilencing pKLO shRNA served as control. To silence pVHL, HeLa cells stably expressing HA-Spry2 were transfected with 20 nm each of two different 25-mer siRNA duplex against pVHL of the following sequence: VHL#4, sense 5′-GCUCUACGAAGAUCUGGAAGACCAC-3′ and antisense 5′-GUGGUCUUCCAGAUCUUCGUAGAGCGA-3′. A mutant 27-mer Spry2 siRNA containing three ribonucleotide substitutions was used as a control as described previously (44). All siRNA were designed and purchased from Integrated DNA Technologies Inc.

Immunoprecipitation (IP) Assays

To IP FLAG-PHDs, 500,000 HEK293T cells were seeded in 60-mm dishes and co-transfected with 1.5 μg of pCMV-3×-FLAG-PHD1–3 and 1.5 μg of pHM6-HA-Spry2 constructs using FuGENE 6 transfection reagent. The following day, cells were exposed to normoxic (21% O₂) or hypoxic (3% O₂) media and incubated for an additional 24 h under normoxic or hypoxic conditions. Cells were washed one time with ice-cold PBS, lysed in Lysis Buffer as described previously (11), and rotated for 30 min at 4 °C. Lysates were spun at 13,000 × g for 10 min, and 500 μg were pre-cleared with protein G beads (Millipore) for 1 h rotating at 4 °C. FLAG-PHD1–3 was immunoprecipitated with 1 μg of anti-FLAG M2 antibody (Sigma) bound to protein G beads (Millipore) by rotating overnight at 4 °C. The next day, beads were washed three times with Lysis Buffer and heated at 95 °C for 5 min in 30 μl of reducing Laemmli buffer to collect bound proteins. To examine HA-Spry2 interaction with endogenous PHD1, HeLa cells stably expressing HA-Spry2 were seeded in 10-cm dishes at 1.5 × 10⁶ density and left under normoxic conditions for 24 h. Cells were then exposed to 16 h of normoxia (21% O₂) or hypoxia (3% O₂) and lysed. The IP procedure was as described above, except HA polyclonal antibody (Sigma) was used to IP HA-Spry2.

IP experiments with 6×-Myc-VHL were as described above, except the cells were co-transfected with 1.5 μg of pHM6-HA-Spry2 and pCS2+MT-6×-Myc-VHL vector. The cells were then treated with 25 μm MG132 for 4 h before IP of 6×-Myc-VHL with the anti-c-Myc monoclonal (Covance, Emeryville, CA).

Ubiquitination Assays

HEK293T cells (500,000/60-mm dish) were cultured for 24 h prior to transfection with 2 μg of control GIPZ shRNA (catalog no. RHS4346) or VHLshRNA#4 (catalog no. RHS4430-98521242), 2 μg of pCMV-3×-FLAG-ubiquitin, and 750 ng of pHM6-HA-Spry2 using FuGENE 6. After 48 h of transfection, the media were changed, and the dishes were left for another 16 h. Cells were treated with MG132 (25 μm) for 4 h at 37 °C before washing two times with ice-cold PBS. A 500-μl aliquot of Lysis Buffer containing 25 μm MG132 and 5 mm N-ethylmaleimide was used to lyse cells. HA-Spry2 was immunoprecipitated from 500 μg of cell lysate using 1 μg of anti-HA monoclonal antibody (Covance) bound to protein G beads for 2 h at 4 °C. The subsequent wash and elution steps from beads were as described above.

For 6×-Myc-VHL add-back experiments, HEK293T cells were first transfected with 1.5 μg of control GIPZ shRNA or pVHL shRNA#4 for 24 h using FuGENE 6 reagent. A second transfection was then performed with 1.5 μg of pCMV-3×-FLAG-ubiquitin, 1.7 μg of pCS2+MT-6×-Myc-VHL, and 550 ng of pHM6-HA-Spry2. After 8 h, the media were changed, and cells were left to incubate for another 16 h. IP of HA-Spry2 was as described above.

Growth Factor Stimulation

HEK293T cells were seeded into 35-mm dishes at 200,000 cells per dish. The following day, cells were co-transfected with 1 μg of PHD1 shRNA#27 and 250 ng of pHM6-HA-Spry2 using FuGENE 6 transfection reagent and returned to the incubator. After 24 h, cells were washed once with PBS and serum-starved for 16 h before addition of FGF. Cells were stimulated with 150 ng/ml FGF for 5, 10, 30, or 60 min then lysed in reducing Laemmli buffer to stop stimulation. pVHL was knocked down with pVHL siRNA#4, and conditions for stimulation were as described for PHD1 shRNA, except the cells were stimulated with FGF for the indicated times.

RESULTS

Hypoxia Increases Cellular Levels of Spry2

To facilitate the interpretation of our data concerning the post-translational regulation of Spry levels without interference from changes in its transcription, our studies were performed with exogenously expressed Spry isoforms under the control of CMV promoter in HeLa cells. All of the results described below were obtained from more than one stably transfected HeLa clonal cell line or with transient transfection of Spry2 in HeLa cells. Hypoxia (3% O₂) increased the amounts of all moderately overexpressed HA-Spry isoforms (Fig. 1A). Using stably transfected HeLa cells that express moderate amounts of HA-Spry2, we found that this increase is due to a decrease in Spry2 degradation during hypoxia (Fig. 1B). Controls with constructs of other proteins (e.g. RSK1) in the same plasmid showed that the activity of the CMV promoter is not altered by hypoxia (supplemental Fig. S1A). Like the N-terminally tagged Spry2 in HeLa cells, the amount of endogenous Spry2 in LS174T cells (Fig. 1C) and HEP3B cells (supplemental Fig. S2) was also higher in hypoxia.

FIGURE 1.

Hypoxia elevates the levels of expressed Spry1–4 and endogenous Spry2. A, HeLa cells transfected to express HA-tagged Spry isoforms were cultured in normoxia (N) (21% O₂) or hypoxia (H) (3% O₂) for 48 h. Western blots with anti-HA (short (Li) and long (Da) exposures) and ERK1/2 (loading control) antibodies are shown. B, same as A, except HeLa cells expressing HA-Spry2 under hypoxic and normoxic conditions were treated with cycloheximide (CHX, 200 μm). At the indicated times, cells were lysed to analyze Spry2 and actin content by Western blots. Quantification of stability of Spry2 as a ratio of actin is shown in the lower panel. C, LS174T colon carcinoma cells grown in hypoxia or normoxia for the days indicated were analyzed for endogenous Spry2 and ERK1/2 (loading control) by Western blots. D, HA-Spry2-expressing HeLa cells grown in normoxia were treated with the indicated concentrations of PHD inhibitors DMOG and DFO for 24 h. The bar graph is quantification of data from three experiments. Bottom panel shows representative blots of Spry2 and actin (loading control) content.

A recent report has shown that hypoxia increases the transcription of Spry4 suggesting that the Spry4 promoter has some elements that are regulated by transcriptional factors activated in hypoxia (45). Because hypoxia increases HIF1α levels, we determined whether or not the stability of Spry2 in hypoxia was regulated by this transcription factor. Despite significant (∼80%) silencing of HIF1α, the amount of Spry2 in hypoxia was not altered (supplemental Fig. S3). Hence, elevations in HIF1α subunit do not regulate stability of Spry2 in hypoxia.

Prolyl Hydroxylase Domain Proteins (PHD1–3) Regulate Cellular Content of Spry2

In normoxia, HIFα subunits and β₂-adrenergic receptor are hydroxylated on Pro residues by PHD1–3 (38, 39, 46). The hydroxylated Pro residues and regions around them serve as binding sites for pVHL, which is the substrate recognition component of elongins B/C and cullin E3 ubiquitin ligase complex, thus facilitating the polyubiquitylation and proteosomal degradation of their substrates (38, 40–42, 47, 48). In hypoxia, the activities of the PHDs are attenuated, thereby decreasing Pro hydroxylation and the subsequent pVHL-mediated ubiquitylation and degradation of these proteins (49, 50). Therefore, we investigated whether PHDs regulate cellular levels of Spry2. The addition of the commonly used, nonselective PHD inhibitors, DMOG and DFO, mimicked the hypoxia-mediated increase in content of HA-Spry2 (Fig. 1D) as well as endogenous levels of Spry2 in HEP3B cells (supplemental Fig. S2). As observed with hypoxia, DMOG did not alter the activity of the CMV promoter because expression of RSK1 in the same plasmid (pHM6) was not altered (supplemental Fig. S1B). Consistent with the notion that PHDs regulate cellular content of Spry2, silencing of endogenous PHD1 and PHD3 augmented the amount of HA-Spry2 in normoxia (Fig. 2) and mimicked the actions of hypoxia; silencing of PHD1 or PHD3 did not alter the levels of the other PHD isoforms (Fig. 2). Moreover, expression of PHD1, PHD2, and PHD3 decreased the content of HA-Spry2 in normoxia and hypoxia (Fig. 3, A and B). Note that hypoxia attenuates, but does not abolish, the activities of endogenous and expressed PHDs (49, 50), explaining the diminution in levels of Spry2 by the expressed PHDs in hypoxia (Fig. 3). The ability of the expressed PHD1, PHD2, and PHD3 to decrease levels of Spry2 in cells was attenuated by the PHD inhibitor DMOG (supplemental Fig. S4), further demonstrating that PHDs regulate the cellular content of Spry2.

FIGURE 2.

Endogenous PHD3 and PHD1 regulate cellular content of Spry2. HeLa cells expressing HA-Spry2 were cultured under normoxic conditions and transfected with control (Cont) shRNA or two shRNAs against each of PHD3 and PHD1. The siRNA experiment was as described for shRNA except cells were transfected with 20 nm control or siRNA against PHD1. Three days later, cells were lysed and analyzed for HA-Spry2, PHDs, and ERK1/2. Representatives of three experiments each are shown.

FIGURE 3.

Expression of PHDs decreases cellular content of Spry2. HA-Spry2-expressing HeLa cells were transfected with empty vector or vector constructs that express FLAG-tagged PHD1, PHD2, or PHD3 and then grown in normoxia (N) (21% O₂) or hypoxia (H) (3% O₂) for 48 h. Cell lysates were analyzed for HA-Spry2, FLAG-PHDs, and ERK1/2 (loading control). Bar graphs show quantification of data from three experiments. Representative Western blots are shown in the bottom panels.

PHD1–3 Interact with Spry2

Next, we determined if PHDs interact with Spry2. Although hypoxia did not alter endogenous PHD1 abundance in HeLa cells, immunoprecipitates of Spry2 contained endogenous PHD1 in cells grown under normoxia but not in hypoxia (Fig. 4A). Similar experiments to examine the association of Spry2 with endogenous PHD2 and PHD3 were unsuccessful, perhaps because of the low endogenous levels of these proteins in HeLa cells. However, Spry2 was present in immunoprecipitates with overexpressed PHD1–3 isoforms, and this interaction was also stronger under normoxic conditions (Fig. 4, B and C). Notably, although in Fig. 4B (lighter exposure), Spry2 does not appear to interact with PHD2, when larger amounts of cell lysates are used to IP the PHDs and when blots are exposed longer, IPs of expressed PHD2 contain Spry2 (Fig. 4C). Additionally, although the Spry2/PHD interactions in hypoxia are decreased, they are not absent (Fig. 4C). By comparing the amounts of PHDs expressed and associated with Spry2, it would appear that Spry2 interactions with PHD3 are stronger and perhaps of greater stoichiometry than those with PHD1 and PHD2 (Fig. 4, B and C).

FIGURE 4.

Endogenous PHD1 and expressed PHDs interact with Spry2. A, HeLa cells stably expressing HA-Spry2 were grown in normoxia (N) (21% O₂) or hypoxia (H) (3% O₂) for 48 h. HA-Spry2 was immunoprecipitated (IP), and co-immunoprecipitated of PHD1 was monitored; nonspecific IgG was used as control. Cell lysates (WCL) were analyzed for PHD1 and βPix (loading control). IB, immunoblot. B, HeLa cells stably expressing HA-Spry2 were transfected to express FLAG-tagged PHD1, PHD2, or PHD3 and then cultured in normoxia (N) or hypoxia (H) for 24 h. The amount of HA-Spry2 that co-immunoprecipitated with FLAG-PHDs was monitored by Western blots. Western of input shows the expression of Spry2 (HA) and PHDs (FLAG) in the 13,000 × g supernatant used for IP. C, same as B, except that the larger amounts of cell lysates were used to detect the co-IP of Spry2 with PHD2. Representatives of three similar experiments are shown.

The levels of Spry2 shown in Fig. 4B reflect its amount in 13,000 × g supernatants of lysates that were made in the nondenaturing buffer for immunoprecipitation. These levels do not represent the total cellular Spry2 content because the pellets from these procedures contain a significant amount of Spry2 (data not shown). It is for this reason that changes in Spry2 levels in hypoxia and in the presence of expressed PHDs are not the same as those observed in Fig. 3.

Because the activity of PHDs is decreased in hypoxia, the greater interaction of Spry2 with PHDs in normoxia (Fig. 4, B and C) suggests that active PHDs interact with Spry2. This observation together with the ability of PHD inhibitors (Fig. 1D and supplemental Fig. S2) or silencing of endogenous PHD1 (Fig. 2) to increase the levels of Spry2 and the ability of expressed PHDs to decrease the content of Spry2 (Fig. 3) strongly suggest that PHDs regulate the stability of Spry2 by hydroxylating it on Pro residue(s).

pVHL Interacts with Spry2 and Regulates Cellular Content of Spry2

Because prolyl hydroxylation sites on proteins such as the HIFα subunits and β₂-adrenergic receptors serve as binding sites for pVHL (38, 40, 42), we next investigated whether Spry2 could interact with pVHL. As shown in Fig. 5, pVHL preferentially interacts with Spry2 in normoxia (Fig. 5A). Additionally, like PHDs, silencing of endogenous pVHL increased levels of Spry2 (Fig. 5B), and expression of pVHL decreased levels of Spry2 (Fig. 5C). To demonstrate that pVHL is involved in poly-ubiquitylation of Spry2, we silenced pVHL in HEK293T cells and examined the ubiquitylation of Spry2. As shown in Fig. 5D (top panel), silencing of pVHL by shRNAs significantly diminished ubiquitylation of Spry2. Conversely, overexpression of pVHL augmented ubiquitylation of Spry2 (supplemental Fig. S5).

FIGURE 5.

Interactions of Spry2 with pVHL in normoxia and ubiquitylation of Spry2 by pVHL to regulate cellular content of Spry2. A, HEK293T cells expressing HA-Spry2 and 6×Myc-pVHL or empty plasmid (−) were grown in normoxia or hypoxia and treated with MG132 (25 μm) for 4 h. pVHL was immunoprecipitated (IP) with anti-Myc antibody, and co-IP of HA-Spry2 was monitored. Spry2 and pVHL levels in whole cell lysates (WCL) are also shown. IB, immunoblot. B, endogenous pVHL was silenced by two siRNAs (#1 and #4) in HEK293T cells expressing HA-Spry2. High and low molecular weight pVHL isoforms and actin (loading control) are shown. C, HEK293T cells expressing HA-Spry2 were transfected with the indicated amounts of Myc-pVHL cDNA. ERK1/2, loading control. Graph on right shows quantification of increase in Myc-pVHL expression and decrease in HA-Spry2 as a function of ERK1/2 at different concentrations of Myc-pVHL expressing plasmid. D, HEK293T cells were transfected with control and pVHL-specific shRNAs 24 h prior to transfection with HA-Spry2 and FLAG-ubiquitin as described under “Experimental Procedures.” Following immunoprecipitation of HA-Spry2 using the method that dissociates associated proteins (37), the immunoprecipitates were analyzed for HA-Spry2 and FLAG-ubiquitin. WCL, whole cell lysate. The poly-ubiquitylated Spry2 migrates as multiple bands. A representative of three experiments is shown.

To identify the sites on Spry2 that are hydroxylated, overexpressed HA-Spry2 from HEK293T cells grown in normoxia was immunoprecipitated. Mass spectrometric analyses of this immunoprecipitated HA-Spry2 demonstrated that it was hydroxylated on Pro-18, Pro-144, and Pro-160 (supplemental Fig. S6). Consistent with this finding, the substitution of Pro residues 18, 144, and 160 with Ala increased the stability of Spry2 (Fig. 6A). Additionally, the P18A/P144A/P160A form of Spry2 did not associate with pVHL (Fig. 6B). These findings strongly suggest that Pro residues 18, 144, and 160 on Spry2 are hydroxylated by PHDs and form the pVHL-binding site(s).

FIGURE 6.

Substitution of Pro-18, Pro-144, and Pro-160 with Ala on Spry2 increases its stability and abrogates its interactions with pVHL. A, HA-Spry2 and its P18A/P144A/P160A mutant (350 ng each plasmid) were expressed in HEK293T cells. Following cycloheximide (200 μm) addition, cells were lysed at different indicated times, and the amounts of HA-Spry2 and actin were monitored by Western analyses. A representative of three similar experiments is shown. B, HEK293T cells (500,000/60-mm dish) were transfected with 2 μg each of plasmids pHM6-HA-Spry2 or its P18A/P144A/P160A mutant together with 0.5 μg of pCS2+MT-6 × c-Myc-pVHL or empty vector. Twenty hours after transfection, cells were treated with 100 μm MG132 for 4 h and then harvested for immunoprecipitation with anti-c-Myc antibody. HA-Spry2 and c-Myc-pVHL in the IP were probed with anti-HA or anti-c-Myc antibody. A representative of three similar experiments is shown. WCL, whole cell lysate.

Silencing of PHD1 or pVHL Augment the Levels of Spry2 and Its Ability to Inhibit FGF Signaling

Finally, we examined whether the hydroxylation-mediated alterations in the content of Spry2 regulate its ability to inhibit growth factor actions. For this purpose, we determined whether silencing of endogenous PHD1 altered the ability of Spry2 to inhibit FGF-mediated ERK1/2 activation. As shown in Fig. 7A, expression of Spry2 in HeLa cells inhibited the activation of ERK1/2. Additionally, silencing of endogenous PHD1 did not alter ERK1/2 activation in HeLa cells not expressing Spry2 (Fig. 7B). However, in HeLa cells expressing Spry2, silencing of PHD1 increased the levels of Spry2 and further diminished FGF-elicited ERK1/2 activation (Fig. 7, C and D). Together, these data show that silencing of PHD elevates the cellular content of Spry2 and its ability to inhibit FGF-elicited ERK1/2 activation.

FIGURE 7.

Silencing of PHD1 increases Spry2 and its ability to inhibit FGF-elicited ERK1/2 phosphorylation. A, control and HA-Spry2-expressing HeLa cells grown in normoxia were serum-starved and treated with FGF (150 ng/ml) for various times. Cells were lysed in Laemmli sample buffer, and the phosphorylation of ERK1/2 was monitored by Western analyses. B and C, same as A, except the ability of FGF to activate ERK1/2 was monitored when PHD1 was silenced in HeLa cells not expressing Spry2 (B) or in HeLa cells expressing HA-Spry2 (C). The lower panel in C shows knockdown of PHD1 expression. Total ERK1/2 was used for loading in all panels and for normalizing phospho-ERK1/2 quantification shown in D, Representative of three experiments is shown.

As observed with silencing of PHD1, the silencing of endogenous pVHL with siRNA in HEK293T cells expressing HA-Spry2 also elevated the amount of Spry2 and diminished FGF-elicited activation of ERK1/2 (Fig. 8). Hence, the regulation of cellular content of Spry2 by the PHD/pVHL-mediated hydroxylation and ubiquitylation regulates the ability of growth factors such as FGF to activate downstream signaling.

FIGURE 8.

Silencing of pVHL elevates cellular content of Spry2 and augments inhibition of FGF-mediated ERK1/2 activation. Same as in Fig. 7 except that pVHL was silenced using siRNA#4 in HEK293T cells. A representative of three experiments is shown.

DISCUSSION

Here, by examining the stability of Spry proteins in hypoxia for the first time, we have unraveled a new post-translational mechanism by which cellular content of Spry2 is regulated. Hypoxia increased the amounts of all ectopically expressed Spry isoforms under the control of the CMV promoter. The effects of hypoxia were not related to the activity of the CMV promoter because the levels of p90 RSK1 expressed in the same vector did not change with hypoxia. Additionally, endogenous levels of Spry2 in LS147T and HEP3B cells were also elevated by hypoxia. Using Spry2 as a prototype, and by inhibiting PHDs with pharmacological inhibitors or silencing of PHDs, we show that levels of both endogenous and ectopically expressed Spry2 are altered in hypoxia by changes in PHD activity. This notion is further supported by the following findings. First, when PHDs are more active to hydroxylate their substrates in normoxia, the interactions between PHDs and Spry2 are greater than in hypoxia (Fig. 4, A–C). This suggests that the enzyme/substrate interactions are more easily discerned when PHDs are active. Second, overexpression of PHDs in normoxia or hypoxia decreases content of Spry2 (Fig. 3, A and B). In this context, it should be noted that PHD activity is diminished, but not abrogated, in hypoxia, and overexpressed PHDs still retain some activity (49, 50). This explains the ability of PHDs to decrease content of Spry2 in hypoxia and also interact, albeit to a lesser extent, with Spry2. Further evidence for a PHD-mediated hydroxylation-dependent regulation of cellular content of Spry2 comes from the following additional three pieces of experimental evidence. First, pVHL that associates with hydroxylated Pro residues on proteins (38, 40, 42) preferentially interacts with Spry2 in normoxia when the PHDs would be more active and more of Spry2 would be expected to be hydroxylated (Fig. 5A). Second, the interactions between the hydroxylation site mutant (P18A/P144A/P160A) form of Spry2 and pVHL is diminished even in normoxia (Fig. 6B). Third, the Pro hydroxylation site mutant (P18A/P144A/P160A) of Spry2 is more stable and is degraded to a lesser extent than its wild-type counterpart (Fig. 6A). That Spry2 is ubiquitylated by the pVHL associated E3 ligase is demonstrated by our findings that silencing of endogenous pVHL decreases the ubiquitylation of Spry2 (Fig. 5D) and overexpression of pVHL increases its ubiquitylation (supplemental Fig. S5).

Thus far, three ubiquitin ligases, c-Cbl, Siah2, and Nedd4-1, have been shown to ubiquitylate and regulate the content of Spry2 in cells (33–37). The amounts of c-Cbl and Siah2 were not altered by hypoxia, whereas Nedd4-1 was modestly decreased by hypoxia (data not shown). Currently, neither the mechanisms that regulate Nedd4-1 content in hypoxia nor the contribution of Nedd4-1 toward elevating Spry2 content in hypoxia are known and form the subject of future investigations.

In this study, we present an additional mechanism involving PHDs and pVHL-associated E3 ligase in the ubiquitylation and regulation of levels of Spry2 protein. Notably, all four aforementioned ligases may not be involved in the regulation of content of Spry2 in the same cell or tissue type. For instance, in tumor samples derived from patients with hepatocellular carcinomas, the decreased levels of Spry2 do not correlate with amounts of either c-Cbl or Siah2 (22). Rather, there is an increase in Nedd4-1 in a subset of these hepatocellular carcinomas with low levels of Spry2 (22). Moreover, we have previously shown that overexpression of Nedd4-1 in cells does not completely diminish the amount of Spry2 (37). Likewise, here we show that when pVHL is overexpressed, levels of Spry2 decline and plateau off with ∼33% of Spry2 still remaining (Fig. 5C). Hence, it is possible that different pools of Spry2 within the same cell are accessible to different E3 ubiquitin ligases, and depending upon the cell type and repertoire of E3 ubiquitin ligases that are expressed, Spry2 may be ubiquitylated and regulated by one or more of the different enzymes.

The novel regulatory paradigm that we have delineated using Spry2 may be applicable to all Spry isoforms because their amounts were elevated by hypoxia. In this respect, Pro-144 is conserved in all human Spry isoforms, whereas Pro-18 and Pro-160 are conserved only in Spry1 and Spry3, respectively. Notably, however, there are no clear consensus sequences for hydroxylation by the PHDs (51). Therefore, the lack of conservation of Pro residues 18 and 160 on Spry2 in some of the Spry isoforms may involve the hydroxylation of other Pro sites. This remains to be determined.

Interestingly, the levels of pVHL are increased in a subset of hepatocellular carcinomas (52). Thus, just as Nedd4-1 elevation in a subgroup of hepatocellular carcinomas results in decreased content of Spry2 and a poor outcome (22), it is possible that pVHL elevation in certain hepatocellular carcinomas may result in decreased levels of Spry2 contributing to the growth and progression of these tumors. In von Hippel-Lindau patients, pVHL deficiency leads to development of hemangiomas and clear cell renal carcinomas (53, 54). However, nothing is presently known concerning whether or not pVHL deficiency in these patients alters the levels of Spry isoforms and whether Spry proteins modulate renal clear cell carcinomas. One could speculate that as shown in certain cell types (44, 55–59), increases in Spry levels in renal cell carcinomas due to pVHL deficiency may protect growth factor receptors from c-Cbl-mediated degradation, augment their levels, and thereby contribute toward the disease. This and other possibilities remain to be experimentally determined.

In summary, we present here the first evidence for a novel mode of regulation of cellular content of Spry2 by mechanisms that involve its prolyl hydroxylation by PHDs and subsequent recognition and ubiquitylation by pVHL and its associated E3 ubiquitin ligase. Moreover, we demonstrate that silencing of PHDs or pVHL by augmenting cellular level of Spry2 inhibits FGF signaling to a greater extent.