A P387L Variant in Protein Tyrosine Phosphatase-1B (PTP-1B) Is Associated With Type 2 Diabetes and Impaired Serine Phosphorylation of PTP-1B In Vitro

The primary mutation analysis was performed on cDNA derived from muscle biopsies from 56 unrelated insulin-resistant type 2 diabetic patients as previously described (13) and on cDNA derived from subcutaneous fat biopsies from 56 unrelated obese patients (age 42.5 ± 11.7 years and BMI 43.2 ± 6.1 kg/m²). The association studies of the G381S, P387L, and 3′UTR+104insG variants were performed using 527 unrelated type 2 diabetic patients (58% men and 42% women), who were recruited from the outpatient clinic at Steno Diabetes Center. The mean age of the patients was 60.0 ± 11.0 years, and their mean BMI was 29.0 ± 5.0 kg/m². Twenty-eight percent of the patients were treated with diet, 57% with an oral hypoglycemic agent, and 15% with insulin. The association studies also comprised 542 unrelated age-matched glucose-tolerant control subjects (49% men and 51% women) who were traced in the central population register and living in the same area of Copenhagen as the type 2 diabetic patients examined at Steno Diabetes Center (n = 240) (14) and in the Copenhagen County Center of Preventive Medicine during 1994–1997 (n = 302) (15). The control subjects had a mean age of 56 ± 10 years and a mean BMI of 25 ± 8 kg/m². All participants underwent a 75-g oral glucose tolerance test. Diabetes was diagnosed in accordance to the 1985 World Health Organization criteria (16). Only normal glucose-tolerant subjects were used in the association study, and all study participants were Danish Caucasians by self-report. The studies were approved by the Ethics Committee of Copenhagen (KA-96008) and were carried out in accordance with Helsinki Declaration II. Before participating in the study, informed consent was obtained from all subjects.

Blood samples for measurement of serum levels of insulin, total cholesterol, HDL cholesterol, triglycerides, and plasma glucose were drawn after a 12-h overnight fast. Serum triglycerides, total serum cholesterol, serum HDL cholesterol, and serum LDL cholesterol were analyzed using enzymatic colorimetric methods (GPO-PAP and CHOD-PAP; Boehringer Mannheim, Mannheim, Germany). The plasma glucose concentration was analyzed by a glucose oxidase method (Granutest; Merck, Darmstadt, Germany); serum-specific insulin, excluding des(31,32) and intact proinsulin, was measured by enzyme-linked immunosorbent assay (Dako insulin kit K6219; Dako Diagnostics, Ely, U.K.). HbA_1c was measured by ion-exchange high-performance liquid chromatography (nondiabetic reference range: 4.1–6.4%).

We conducted a primary mutation analysis on cDNA prepared from total RNA extracted from 56 biopsies from the vastus lateralis muscle and 56 subcutaneous fat biopsies according to standard procedures as previously described (13,17). The PTP-1B cDNA was amplified by a primary polymerase chain reaction (PCR) segment covering parts of the 5′ and 3′ untranslated region and the coding region (primers are listed in Table 1 according to sequence published by Chernoff et al. [18] [GenBank accession no. M31724]). Secondary PCRs were performed with incorporation of [α-³²P]dCTP on 2 μl of the primary PCR in a total reaction volume of 25 μl. In this way, the primary PCR was reamplified in seven overlapping secondary PCR segments ranging in size from 267 to 313 bp (primers listed in Table 1). All PCRs were performed in 25-μl reaction volumes containing 1 × PCR buffer, 0.2 μmol/l of each primer, 0.2 mmol/l dNTP, 10 mCi/ml [α-³²P]dCTP, 0.3 units AmpliTaq DNA polymerase (Perkin Elmer, Forster City, CA), and 1.5 mmol/l MgCl₂. The cycling program was a denaturation step at 95°C for 2 min followed by 95°C for 30 s, annealing for 30 s (annealing temperature for each primer set is given in Table 1), 72°C elongation for 30 s in 35 cycles, and finally an extension at 72°C for 9 min (except for the primary PCR, which was elongated for 2 min using a GeneAMp 9600 thermal cycler [Perkin Elmer]). The secondary PCR samples were analyzed by single-strand conformational polymorphism (SSCP) and heteroduplex analysis, applying two different experimental conditions as previously reported (19). The variants identified were sequenced on both strands using an ABI377 automated sequencer (Perkin Elmer) or by dideoxysequencing using the Thermo Sequenase cycle sequencing kit (USB, Cleveland, OH). All variants identified were confirmed on genomic DNA.

Genotyping of the P387L variant was carried out by PCR amplification of a 327-bp segment covering exon 9 on genomic DNA using the primers (5′-3′sense primer) CATCTCTGCCCTCTGATTCC and (5′-3′ antisense primer) TGAGACTGGCTCAGATGCAC (T_anneal 60°C and 2.0 mmol/l MgCl₂). The mutation removes a Bsl I restriction enzyme site at position 101 in the segment. An additional control site is present at position 290 in the segment. Thus, enzyme digestion of PCR segments from wild-type carriers will produce three fragments: 37, 101, and 189 bp, whereas homozygous mutant carriers will produce only the 37-bp fragment and a 290-bp fragment. Heterozygous carriers will contain all fragments. Screening is performed by adding 5 units of Bsl I enzyme (New England Biolabs, Beverly, MA) into 20 μl of amplified PCR product, and after 4 h incubation at 55°C, products are loaded onto 3% agarose gels and visualized by staining with ethidium bromide.

Genotyping of the G381S variant was carried out by PCR amplification of a segment covering exon 9 on genomic DNA using the primers (5′-3′sense primer) TACCCATCTCTGCCCTCT and (5′-3′ antisense primer) GGTAGGATTCAGTTCTGTG (T_anneal 56°C and MgCl₂ 1.5 mmol/l) and PCR-SSCP and heteroduplex analysis. Genotyping of the 3′UTR+104insG variant was carried out by PCR amplification of a segment covering exon 10 and part of the 3′UTR region using the primers (5′-3′ sense primer) GTCTGGGCTCATCTGAACTGT and (5′-3′ antisense primer) GGACGGACGTTGGTTCTG (T_anneal 58°C and MgCl₂ 1.5 mmol/l) and by PCR-SSCP and heteroduplex analysis. SSCP genotyping revealed two intronic variants as described in results.

Fisher’s exact test was applied to test for differences in allele and genotype frequencies. P < 0.05 was considered significant. SPSS software for windows (version 10.0) was used to carry out descriptive analysis and analysis using a generalized linear model. The analysis included age and BMI as covariate and sex and genotype as a fixed factor. The normal distribution of the residuals was visually verified. The odds ratios from the case-control study were calculated using logistic regression while controlling for age. For comparison of the quantification of the radioactive incorporation into peptides, a two-tailed distribution paired Student’s t test was used on the log¹⁰-transformed data set from each of four independent assays. The values are expressed as means ± SD.

Incorporation of [γ-³²P]ATP into wild-type peptide (RRRGAQAASPAKGE: 387P) and mutant peptide (RRRGAQAASLAKGE: 387L) by the p34 cell division cycle (p34^cdc2) kinase was performed in a final reaction mixture containing 50 μl of 1 mmol/l of wild-type (387P) or mutant peptide (387L), 100 μmol/l ATP at a final specific activity of 100 μCi/μmol [γ-³²P]ATP (Amersham Pharmacia Biotech), and reaction buffer containing 50 mmol/l Tris-HCl, 10 mmol/l MgCl₂, 2 mmol/l dithiothreitol, 1 mmol/l EGTA, 0.01% Brij 35, pH 7.5, and 0.272 units recombinant p34^cdc2 protein kinase (New England Biolabs). The peptides were synthesized using Fmoc chemistry, and the resin-cleaved peptides were analyzed by mass spectral analysis (MALDI/TOF; Research Genetics, Invitrogen, Huntsville, AL). For negative and positive controls, a 0.33-μg/μl concentration of myelin basic protein (MBP; Sigma) was used in the final 50-μl reaction. The substrate and reaction mixtures were heated separately at 30°C for 1 min before mixing and then incubated at 30°C for 30 min. The reaction was stopped by the addition of 50 μl of 2 × Tricine SDS sample buffer (cat. no. LC1676; Novex, San Diego, CA) and 2.5% of 2-mercaptoethanol. The samples were subsequently heated for 2 min at 85°C, and then 25 μl of each sample was applied on a 16% tricine gel (Novex) and electrophoresed for 2 h at 110 V. The tricine gels were silver-stained to ensure peptide location using the protein silver stain kit PlusOne (Amersham Pharmacia Biotech). After gel exposure to phosphoimager screens, the radioactivity of the peptides was quantified by scanning in a Typhoon 8600 (Molecular Dynamics, Amersham Pharmacia Biotech). Quantification of the incorporation of [γ-³²P]ATP into wild-type and mutant peptide was examined in four independent assays (Fig. 1) using Image Quant 5.1 software (Molecular Dynamics, Amersham Pharmacia Biotech). The wild type from each assay was standardized to 100%, and the mutant was counted as a percentage of the wild type. SD is calculated from the log¹⁰-transformed data set (Fig. 2).

Primary analysis of the coding region of the PTP-1B gene, including the 5′-UTR from position −42 before translation start site and the 3′UTR to position +138 after the stop codon, revealed five variants. There were four silent variants, (NT CGA→CGG) R199R, (NT CCC→CCT) P303P, 3′UTR+104insG (insertion variant), and 3′UTR+86T→G (transversion variant), and one missense variant (NT CCA→CTA) P387L. Furthermore, during SSCP genotyping on genomic DNA using intronic primers, two missense variants, (NT GGT→AGT) G381S and (NT ACG→ATG) T420M, and two single-intron variants, IVS9+57C→T and IVS9+58G→A, were found. The T420M, 3′UTR+86T→G, IVS9+57C→T, and IVS9+58G→A variants were only found in one case each.

The prevalence of the G381S, P387L, and 3′UTR+104insG variants were further evaluated in a case-control study of type 2 diabetic patients and matched glucose-tolerant control subjects. As presented in Table 2, the allelic frequency of the G381S variant was 0.4% in type 2 diabetic patients and 1.1% in control subjects (P = 0.143). The allelic frequency of the 3′UTR+104insG insertion variant was 6.9% among type 2 diabetic patients and 10.2% among control subjects (P = 0.20) (Table 2). The observed genotype frequencies of these variants were in Hardy-Weinberg equilibrium. Furthermore, the 3′UTR+104insG variant was not found to be associated with BMI, fasting serum insulin, or fasting plasma glucose among control subjects (data not shown).

The allelic frequency of the P387L variant was 1.4% (95% CI 0.4–2.4) among 527 type 2 diabetic patients and 0.5 (−0.1 to 1.1) among 542 glucose-tolerant control subjects. Thus, the P387L variant was significantly associated with type 2 diabetes (P = 0.037) (Table 2), with carriers having an age-corrected relative risk of type 2 diabetes of 3.7 (1.26–10.93, P = 0.02). When comparing heterozygous carriers of the P387L variant with the remaining group of wild-type carriers, the phenotypical and biochemical characteristics were not significantly different between the two groups (Table 3).

The P387L variant is situated next to a serine residue that is a target of the proline-directed p34^cdc2 protein kinase. We investigated the incorporation of [γ-³²P]-labeled radioactivity into wild-type peptide (387P) versus mutant peptide (387L) in an in vitro kinase assay (Fig. 1). The assay results showed significantly lower p34^cdc2 kinase proline–directed phosphorylation of the mutant compared with the wild-type peptides (387P: 100% vs. 387L: 28.4 ± 5.8%; P = 0.0012) (Fig. 2).

Considering the crucial role of PTP-1B in regulating insulin signaling and the clear physiological effect on insulin-stimulated glucose uptake associated with the lack of PTP-1B in mice, PTP-1B is a credible candidate to be involved in the assumed polygenic basis of type 2 diabetes in humans. In addition, PTP-1B is located on the chromosomal region 20q13.1–13.2 on the long arm of chromosome 20, which has been linked to quantitative trait loci for obesity and fasting serum insulin levels (20) as well as type 2 diabetes (21). In the present study, we have identified a rare P387L variant of the PTP-1B gene that is associated with type 2 diabetes in the Danish population and with impaired in vitro serine phosphorylation of a PTP-1B peptide. We found that the variant has an almost three times higher prevalence among type 2 diabetic patients, conferring an effective relative risk of 3.7 for carriers of this variant.

The proline at position 387 in the COOH-terminal regulatory region of the protein is conserved between mouse, rat, and humans and, in the human sequence, is located next to a serine residue (position 386). It is part of a consensus sequence known to be phosphorylated in vitro by the proline-directed kinase p34^cdc2 (22,23), in vivo in states of mitosis (24,25) and, perhaps more importantly, in response to various stress stimuli (26). Replacing the proline by a leucine reduced the phosphorylation of the serine by 70% in our in vitro studies of the relevant peptide. Testing the replacement on the NetPhos 2.0 Phosphorylation Prediction database further confirmed that the predicted likelihood of serine 386 being a phosphorylation target was reduced from 0.884 to 0.135 (range from 0 to 1.0) when replacing proline 387 with leucine (27).

Because a lack of PTP-1B is associated with increased insulin sensitivity, a PTP-1B variant associated with type 2 diabetes would be hypothesized to cause increased PTP-1B levels and phosphatase activity. However, the direct consequences on enzyme function of the P387L variant is still unclear, and Ser386 phosphorylation can be associated with both increased (24) as well as decreased (25) PTP-1B enzymatic activity. Moreover, human studies have also shown conflicting results in relating PTP-1B mRNA levels and phosphatase activity to insulin sensitivity and obesity (10,28,29,30,31), and in vitro studies suggest that the expression levels of PTPases may not be the primary determinant of their insulin receptor phosphorylation activity (32), indicating that other factors may regulate PTPase activity. This is also suggested by the surprising localization of the PTP-1B to the endoplasmic reticulum by a COOH-terminal motif (7). Therefore, it is likely that the in vivo action of the PTP-1B protein is regulated by both the phosphorylation status as well as by restricted accessibility to intracellular substrates, and it might be hypothesized that the P387L variant and potentially the phosphorylation status at position Ser386 may affect the activity or localization of the protein. Thus, a recent study has shown that increased phosphoserine content of PTP-1B in vivo was followed by increased PTP-1B activity and indicated that insulin and protein kinase A have a critical role in the regulation of PTP-1B (33). However, the mechanisms of COOH-terminal PTP-1B regulation have not been fully elucidated, and further studies are needed to evaluate the functional consequences of the P387L variant.

The search for variability in the human PTP-1B gene also identified several additional variants, of which only the silent variant P303P has been previously identified and reported to have no association with type 2 diabetes (34). The association study of the missense variant G381S and the insertion variant 3′UTR+104insG showed no significant association with type 2 diabetes, although the 3′UTR+104insG variant shows a trend toward an association. These findings need to be tested in larger association studies.

In conclusion, we identified a rare P387L variant in the PTP-1B gene that was associated with type 2 diabetes in the examined Danish population. The variant peptide exhibited reduced in vitro serine phosphorylation. Testing for association of the variant to diabetes in other populations as well as further studies of the effect of the variant on PTP-1B function are needed.

The present study was supported in by the Poul and Erna Sehested-Hansen Foundation, the Danish Council for Health Science (“Growth and Regeneration”), the Danish Diabetes Association, the Danish Medical Research Council (grant nos. 9900810 and 9902592), the Velux Foundation, the Danish Heart Foundation, and the EEC (grants BMH4-CT98-3084 and QLRT-CT-1999-00546).

The authors thank Helle Fjordvang, Annemette Forman, Lene Aabo, Inge Lise Wantzin, and Christina B.P. Søholm for dedicated and careful technical assistance, Søren N. Jacobsen (Novo Nordisk, Bagsvaerd, Denmark) for helpful advice on the kinase assay, Bendix Carstensen for statistical assistance, and Grete Lademann for secretarial support.