Tyrosines 868, 966, and 972 in the kinase domain of JAK2 are autophosphorylated and required for maximal JAK2 kinase activity

Abstract

Janus kinase 2 (JAK2) is activated by a majority of cytokine family receptors including receptors for GH, leptin, and erythropoietin. To identify novel JAK2-regulatory and/or -binding sites, we set out to identify autophosphorylation sites in the kinase domain of JAK2. Two-dimensional phosphopeptide mapping of in vitro autophosphorylated JAK2 identified tyrosines 868, 966, and 972 as sites of autophosphorylation. Phosphorylated tyrosines 868 and 972 were also identified by mass spectrometry analysis of JAK2 activated by an erythropoietin-bound chimeric erythropoietin receptor/leptin receptor. Phosphospecific antibodies suggest that the phosphorylation of all three tyrosines increases in response to GH. Compared with wild-type JAK2, which is constitutively active when overexpressed, JAK2 lacking tyrosine 868, 966, or 972 has substantially reduced activity. Coexpression with GH receptor and protein tyrosine phosphatase1B allowed us to investigate GH-dependent activation of these mutated JAK2s in human embryonic kidney 293T cells. All three mutated JAK2s are activated by GH, although to a lesser extent than wild-type JAK2. The three mutated JAK2s also mediate GH activation of signal transducer and activator of transcription 3 (Stat3), signal transducer and activator of transcription 5b (Stat5b) and ERK1, but at reduced levels. Coexpression with Src-homology 2B1beta (SH2B1beta), like coexpression with GH-bound GH receptor, partially restores the activity of all three JAK2 mutants. Based on these results and the crystal structure of the JAK2 kinase domain, we hypothesize that small changes in the conformation of the regions of JAK2 surrounding tyrosines 868, 966, and 972 due to e.g. phosphorylation, binding to a ligand-bound cytokine receptor, and/or binding to Src-homology 2B1, may be essential for JAK2 to assume a maximally active conformation.

Figure 1

JAK2 is autophosphorylated on tyrosines 868, 966, 972, and 1008. 293T cells transiently expressing SH2B1β (504-670) (5 μg cDNA) and JAK2 or the indicated JAK2 Y→F mutant (2 μg cDNA) were lysed, and JAK2 was immunoprecipitated using αJAK2. Immobilized JAK2 was incubated with [γ-³²P]ATP for 30 min. The ³²P-labeled JAK2 was subjected to 2D phosphopeptide mapping. Insets are enlargements of the boxed regions containing spots of interest. Numbered solid arrows in maps A, L, and P indicate the position of spots (1, 2, and 3) that are missing (open arrows) in maps B, K, and Q of JAK2 Y868F, JAK2 Y966F, or JAK2 Y972F, respectively. The incorporation of ³²P into JAK2 Y868F (map B) was minimal when compared with the other JAK2 constructs (data not shown). Therefore, in map B, background spots (determined by their presence in control cells not expressing ectopic JAK2) are more evident than in the corresponding map A of wild-type (WT) JAK2. The most prominent spot in the inset of map B is an example of such a background spot. The solid arrow in map R indicates the position of spot 5 that corresponds to the peptide containing phosphorylated Tyr 1008 (theoretically VLPQDKEpY₁₀₀₇pY₁₀₀₈K). In map S of JAK2 Y1008F, this peptide shifts to spot 7 (theoretically VLPQDKEpY₁₀₀₇F₁₀₀₈K). An additional cleavage product (theoretically EpY₁₀₀₇F₁₀₀₈K) is also present (spot 6). Boxes surround the maps that were performed in a single individual experiment. To determine with confidence that mutation of a tyrosine to phenylalanine leads to the disappearance of a specific spot, only JAK2 and JAK2 mutants from tryptic digests that were run simultaneously were compared. Between two and five maps were obtained for each of the JAK2 Y→F mutants.

Figure 2

Phosphorylation of tyrosines 868 and 966 in JAK2 is detected by MS. The 293 cells overexpressing ELR and JAK2 were stimulated with Epo for 30 min. JAK2 was immunoprecipitated with αJAK2(758), resolved by SDS-PAGE, reduced, alkylated, and digested with trypsin. MS/MS spectrum corresponding to the doubly charged JAK2 tryptic peptides (A) (p)Y₈₆₈DPLQDNTGEVVAVK (m/z 892.92) and (B) GME(p)Y₉₆₆LGTK (m/z 503.76) were obtained. Sequence assignments with the expected m/z value for the b-type and y-type ions are noted. Underlined m/z values indicate fragment ions present in the MS/MS spectrum. Y-type ions (y), along with associated neutral losses of ammonia (y*), and water (y⁰) ions are indicated. Similarly, b-type ions (b), along with neutral losses of carbon monoxide (a) ions are identified. In Fig. 2A, [M+2H]²⁺ denotes the doubly charged peptide ion corresponding to (p)Y₈₆₈DPLQDNTGEVVAVK. [M+H-2H₂O-80]⁺ denotes the singly charged ion corresponding to (p)Y₈₆₈DPLQDNTGEVVAVK, which has sustained neutral losses of water and phosphate. In Fig. 2B, the fragment corresponding to the PY-specific immonium ion (m/z = 216.043) is designated with a square.

Figure 3

αpY868 JAK2, αpY966 JAK2, and αpY972 JAK2 identify tyrosines 868, 966, and 972, respectively, as sites of phosphorylation in JAK2. 293T cells transiently transfected with SH2B1β cDNA (1 μg) and the cDNAs (1 μg) for JAK2, JAK2 Y868F, JAK2 Y966F, JAK2 Y972F, JAK2 Y1007F, or vector were lysed. Proteins were resolved on SDS-PAGE gels and blotted [immunoblotted (IB)] with the indicated antibodies (n = 2). The migration of JAK2 is indicated.

Figure 4

In response to GH, endogenous JAK2 is phosphorylated at tyrosines 868, 966, and 972. The 3T3-F442A cells were incubated with vehicle or 500 ng GH/ml (23 nm) for the indicated times. The cells were lysed. JAK2 was immunoprecipitated using αJAK2 and blotted [immunoblotted (IB)] with the indicated antibodies (n = 2 for 0, 10 min; n = 1 for 3, 30, 60 min; n = 3 for equivalent lysate blots, data not shown). The migration of JAK2 is indicated. IP, Immunoprecipitation.

Figure 5

Mutating tyrosines 868, 966, or 972 in JAK2 to phenylalanine decreases the ability of JAK2 to autophosphorylate when JAK2 is expressed alone. Coexpression with SH2B1β partially restores kinase activity. A, 293T cells transiently transfected with the cDNA (1 μg) for JAK2 or the indicated JAK2 Y→F mutant were blotted [immunoblot (IB)] with the indicated antibodies (n = 2). The migration of JAK2 is indicated. B, 293T cells transiently transfected with the cDNA (1 μg) for JAK2 or the indicated JAK2 Y→F mutant and the indicated amount of the cDNA for SH2B1β were blotted (IB) with the indicated antibodies (n = 2). The migration of JAK2 and SH2B1β is indicated.

Figure 6

PTP1B, but not PLCγ1, is necessary for GH-dependent phosphorylation of overexpressed JAK2 in 293T cells. 293T cells were transiently transfected with various combinations of the cDNAs encoding GH receptor (400 ng), JAK2 (320 ng), PLCγ1 (100 ng), and PTP1B (150 ng) as indicated. The cDNA for (A) Stat5b (100 ng) or (B) Stat3 (100 ng) was also transfected. The cells were incubated with vehicle or 500 ng GH/ml (23 nm) for 15 min. Cell lysates were resolved by SDS-PAGE and blotted [immunoblot (IB)] with the indicated antibodies (n = 2). The migration of JAK2 is indicated.

Figure 7

Basal and GH-stimulated activation of JAK2 Y868F, JAK2 Y966F, or JAK2 Y972F is decreased compared with wild-type JAK2. The 293T cells were transiently transfected with cDNA encoding GH receptor (400 ng), PTP1B (150 ng), PLCγ1 (100 ng), Stat3 (100 ng), myc-ERK1 (100 ng), and either JAK2, JAK2 Y868F, JAK2 Y966F, JAK2 Y972F, or JAK2 Y1007F (320 ng) as indicated. The cells were incubated with vehicle or 500 ng GH/ml (23 nm) for 15 min. A, Cell lysates were resolved by SDS-PAGE and blotted [immunoblot (IB)] with the indicated antibodies (n = 5). The migration of JAK2 and Stat3 is indicated. B, JAK2 was immunoprecipitated using αJAK2 and subjected to an in vitro kinase assay in the presence of 1 μm ATP and 500 μm of a peptide containing Tyr699 of Stat5. Background equal to the average ³²P (3965 cpm) incorporated into the control cells [(−) JAK2 cDNA, 0 min GH] was subtracted from all conditions (n = 3). Error bars represent the sem. GH caused a statistically significant (P < 0.05) increase in activity of JAK2 WT, JAK2Y868F, JAK2Y966F, and JAK2Y972F. When compared with JAK2 WT, both the basal and GH-stimulated activity of all of the JAK2 Y→F mutants was decreased to a statistically significant extent. WT, Wild type.

Figure 8

Effect of mutating tyrosines 868, 966, and 972 on the ability of JAK2 to phosphorylate Stat5b and ERK1. The 293T cells were transfected with cDNA encoding GH receptor (400 ng), Stat5b (100 ng), PLCγ1 (100 ng), PTP1B (150 ng), and either JAK2 WT, JAK2 Y868F, JAK2 Y966F, JAK2 Y972F, or JAK2 Y1007F (320 ng) as indicated. Cells were incubated with vehicle or with 500 ng GH/ml (23 nm) for 15 or 60 min and then lysed. Proteins were resolved on SDS-PAGE gels and blotted [immunoblot (IB)] with the indicated antibodies (n = 4). The migration of JAK2, Stat5b, and ERK is indicated. WT, Wild type.

Figure 9

Position of tyrosines 868, 966, and 972 in the structure of JAK2 [2B7A (12)]. Tyr 966 and Tyr 972 are in the C-lobe of the JAK2 kinase domain. Tyr 966 is in α-helix E (green) and Tyr 972 in β-strand 6 (dark green). Tyr 972 is in close proximity to α-helix C (mauve) in the N-lobe. The α-helix C plays a critical role in positioning the ATP in the active site. When the kinase domain is in the active conformation, Tyr 972 is in close proximity to the activation loop (dark blue). Tyr 868 is in the N-lobe of the kinase domain in β-strand 2. The top of the catalytic cleft is formed by β-strand 1, β-strand 2, and β-strand 3 (brown). The Gly-X-Gly-X-X-Gly 861 ATP binding motif is in the turn between β-strand 1 and β-strand 2. Lys 882, which binds to the α- and β-phosphates of the bound ATP, is in β-strand 3. Mutations in the region surrounding Tyr 868 have been linked to acute megakaryoblastic leukemia (Thr 875Arg) (43) and pediatric acute lymphoblastic leukemia (Arg 867Gln or Asp 873Arg) (44). The substrate tyrosine (light gold) is modeled into the substrate-binding site based on the closely related structure of the insulin receptor [1IR3 (56)]. The N-lobe is in brown-mauve tints. The C-lobe is in blue-green tints.