FK506-Binding Protein 2 Participates in Proinsulin Folding

Abstract

Apart from chaperoning, disulfide bond formation, and downstream processing, the molecular sequence of proinsulin folding is not completely understood. Proinsulin requires proline isomerization for correct folding. Since FK506-binding protein 2 (FKBP2) is an ER-resident proline isomerase, we hypothesized that FKBP2 contributes to proinsulin folding. We found that FKBP2 co-immunoprecipitated with proinsulin and its chaperone GRP94 and that inhibition of FKBP2 expression increased proinsulin turnover with reduced intracellular proinsulin and insulin levels. This phenotype was accompanied by an increased proinsulin secretion and the formation of proinsulin high-molecular-weight complexes, a sign of proinsulin misfolding. FKBP2 knockout in pancreatic β-cells increased apoptosis without detectable up-regulation of ER stress response genes. Interestingly, FKBP2 mRNA was overexpressed in β-cells from pancreatic islets of T2D patients. Based on molecular modeling and an in vitro enzymatic assay, we suggest that proline at position 28 of the proinsulin B-chain (P28) is the substrate of FKBP2's isomerization activity. We propose that this isomerization step catalyzed by FKBP2 is an essential sequence required for correct proinsulin folding.

Keywords: FKBP2; endoplasmic reticulum; proinsulin; proline isomerization.

Figure 1

The FKBP2 isomerase interacts with proinsulin and GRP94 and is expressed in β-cells. (A). INS1-E cells were transiently transfected to express GFP-tagged human proinsulin or GRP94, both of which were immunoprecipitated and analyzed by mass spectrometry for binding partners present in the endoplasmic reticulum. FKBP2’s interaction with proinsulin and GRP94 was confirmed by SDS-PAGE/Western blot analysis of the immunoprecipitates of myc-tagged FKBP2 expressed in INS1-E cells. PDIA 1, 3, 6: Protein disulfide-isomerase 1, 3, 6; BiP/GRP78: binding immunoglobulin protein/Glucose Regulated Protein 78; CgA, CgB: chromogranin A and B; CPE: carboxypeptidase E; ORP150: oxygen-regulated protein 150. (B). Graphic representation of the FKBP2 promoter with putative transcription factor binding sites and regulatory elements identified using PROMO from ALGGEN. Analyzed DNA sequence was retrieved from HGNC:3718. ERSE: ER stress response element; TFIIB/I: transcription factor IIB/I; ATF6 and 3: activating transcription factor 6 and 3; NF-kB/RelA: nuclear factor-kB/REL-associated protein; IRF-1/2: interferon regulatory factor 1/2; NF-AT1/2: nuclear factor of activated T cells; SP1: Specificity Protein 1; STAT1: signal transducer and activator of transcription 1; CREB: cAMP response element-binding protein; Ex: exon. (C). FKBP2 single-cell gene expression. t-SNE representations colored according to FKBP2 expression levels, n = 1554 from Segerstolpe et al., Cell Metabolism 2016. (D). Representative fluorescence microscopy images of the human adult pancreas stained with antibodies against FKBP2 (red), INS (insulin, green) labeling β-cells and SST (Somatostatin, white) labeling δ-cells. Scale bars = 100 μm. DAPI (blue) labels the nuclei. (E). FKBP2 mRNA expression analyzed by single-cell RNA sequencing of β-cells derived from healthy and T2D individuals. Lines represent median with 95% CI error bars. Statistical analysis was performed using unpaired two-tailed parametric t-tests. FKBP2 transcript expression is shown as log2 (counts per million + 1).

Figure 2

In silico modeling of proinsulin interaction with FKBP2, identification of proinsulin isomers, and demonstration of FKBP2 as a proinsulin isomerase. FKBP2 binding to its inhibitor FK506 ((A), crystal structure PDB: 4NNR) and proinsulin ((B), FKBP2 crystal structure PDB: 2PBC, proinsulin NMR structure PDB: 2KQP, blue) in the representative in silico models. In both structures, the residues involved in the interaction are highlighted in purple. The models were generated on the ZDOCK server. (C). The panel shows the frequency of proinsulin residues interacting with FKBP2 in the top ten generated interaction models between FKBP2 and proinsulin (2PBC and 2KQP structures, respectively). (D). Overlap between proinsulin interacting residues and conserved residues within FKBP2. The data displayed are the combined data from 100 FKBP2–proinsulin interaction models. The conserved residues are highlighted in dark purple. (E,F). Alignment of human proinsulin sequence (UniProt P01308), between positions 23 of the B-chain and position 1 of the C peptide with proinsulin (E) and IGF-1/2 (F) of Homo sapiens (HUMAN IGF1 P05019, IGF2 P01344), Rattus norvegicus (RAT INS1 P01322, INS2 P01323, IGF1 P08025, IGF2 P01346), Mus musculus (MOUSE INS1 P01325, INS2 P01326, IGF1 P05017, IGF2 P09535), Sus scrofa (PIG INS P01315, IGF1 P16545, IGF2 P23695), Bos Taurus (BOVIN INS P01317, IGF1 P07455, IGF2 P07456), Gallus gallus (CHICK INS P67970, IGF1 P18254, IGF2 P33717), and Xenopus laevis (XENLA INS1 P12706, INS2 P12707, IGF1A P16501). Sequences were retrieved from UniProt v. 2021_04. (G). Peptidyl-prolyl cis–trans isomerization assay of a peptide containing Pro28 (GERGFFYTP-F-pNA, B20–28) incubated with or without FKBP2. The substrate peptide was dissolved in LiCl/TFE buffer, and to initiate the reaction, chymotrypsin (CTRC) was added to digest and release the fluorophore (pNA) from the peptide in trans conformation. The absorbance was acquired at 405 nm. The signal from the time point with the highest absorbance in the reaction that contained FKBP2, n-Nitroanilide modified peptide and chymotrypsin was used to generate the graph. Data were analyzed by paired t-test between individually tested conditions.

Figure 3

Diminished intracellular proinsulin and insulin contents after FKBP2 KO. (A). SDS-PAGE and Western blot analysis of proinsulin and insulin expression levels in FKBP2 KO and FKBP2 WT cells (failed FKBP2 KO and INS1-E) after 2 h in 20 or 2 mM glucose-containing media. FKBP2 KO was achieved via CRISPR/Cas9 guide FKBP2 directed-RNA targeting (clonal cell lines shown 3–6 months after viral transduction; representative blots of n > 10 on the left and band quantification on the right). Proinsulin and insulin were visualized with an anti-insulin antibody. (B). SDS-PAGE and Western blot analysis of proinsulin and insulin expression levels upon exogenous expression of FKBP2 in FKBP2 KO clones and INS1-E control (Ctrl) cells. Cells were transfected with plasmids coding for myc-tagged FKBP2, cultured for 48 h, and lysed 2 h after the introduction of fresh culture media supplemented with 20 mM glucose. Representative blot of n = 4 are presented on the left and band quantification on the right. (C). SDS-PAGE and Western blot analysis of proinsulin from FKBP2 KO, FKBP2 WT (Ctrl 1), and INS1-E (Ctrl 2) cells after size exclusion chromatography (SEC) separation at the pH 7.4 condition of cell lysis and SEC separation. SEC experiments n = 3. Data for A-B represent means ± SD analyzed by non-paired (A) or paired (B) t-test of treatments versus control.

Figure 4

FKBP2 knockout does not induce endoplasmic reticulum stress but sensitizes cells to apoptosis. (A). mRNA levels of genes in the endoplasmic reticulum stress pathways were analyzed by quantitative reverse transcription-PCR (qRT-PCR) in Ctrl (FKBP2 WT cells) and FKBP2 KO cells. Data represent the mean ± SD analyzed by ordinary one-way ANOVA of treatments versus control, n > 6. (B). Apoptosis levels, representing internucleosomal degradation of genomic DNA, were analyzed in FKBP2 KO and Ctrl cells, n > 4. Where indicated, cells in A and B were exposed for 18 h to 0.7 µM of thapsigargin.

Figure 5

FKBP2 KO shortens proinsulin intracellular half-life and increases its secretion. (A,B). INS−1E control (Ctrl) and FKBP2 KO cells were cultured for 3 h in 2 mM glucose-containing media. subsequently, 100 µM of the protein synthesis inhibitor cycloheximide (CHX, upper panels) and 200 nM of the inhibitor of exocytosis Brefeldin A (BFA, bottom panels) were added to the culture media from the start of the experiment. Cells were lysed at indicated time points, analyzed via reducing SDS-PAGE, and proinsulin and insulin were visualized through Western blotting (WB) with an anti-proinsulin antibody, n = 5. B. WB data quantification. Top: graph p values were calculated by unpaired t-test of Ctrl + CHX vs. FKBP2 KO + CHX for each time point. Bottom: Table with averages of changes of Ctrl and FKBP2 KO cells over the course of experiment in proinsulin content (band intensity) in response to BFA treatment. (C). Accumulated secretion of proinsulin over the period of 4 h in 2 or 20 mM glucose by failed FKBP2 KO (Ctrl 1 and 2) and FKBP2 KO cells, analyzed by SDS-PAGE and WB (n = 10). In total, 500 µL of cell supernatants (adjusted to cell number) was concentrated using 10 kDa MWCO filters to remove salts and reduce the volume to 15 µL. Quantification of proinsulin bands was carried out with ImageJ (A,B) and normalized to GADPH (A) bands. Data are presented as means ± SD analyzed by unpaired t-test of treatments versus control. (D). The same cell types were tested for their ability to secrete insulin in response to the given glucose concentrations in KRBH buffer for a period of 30 min. Supernatants were analyzed by ELISAs specifically detecting mature insulin (left graph) or proinsulin (right graph) only. The bars represent the means ± SD.

Figure 6

FKBP2 KO increases intracellular levels of high-molecular-weight proinsulin complexes and non-soluble proinsulin fraction. (A). INS-1E cells were transfected with myc-tagged FKBP2, followed 48 h later with immunoprecipitation (via myc tag) and SDS-PAGE/Western blotting to detect proinsulin. Where indicated, cell lysates were pretreated prior to immunoprecipitation for 10 min with 100 mM of reducing agent 2-mercaptoethanol (2-ME), n = 3. (B). Non-reducing SDS-PAGE/Western blot analysis of failed FKBP2 KO (Ctrl 1 and 2) and FKBP2 KO cells cultured for 3 h in 20 and 2 mM glucose conditions. The presence of high-molecular-weight proinsulin complexes was evaluated with a proinsulin-specific antibody as in Arunagiri et al. [1], n = 10. (C). Void fractions of size exclusion chromatography of failed FKBP2 KO (Ctrl) and FKBP2 KO cells (cultured for 3 h in 20 mM glucose-containing media) were analyzed under non-reducing and reducing conditions with SDS-PAGE/Western blot to detect proinsulin, n = 3. (D). Insoluble fractions of failed FKBP2 KO (Ctrl 1 and 2) and FKBP2 KO cells cultured for 3 h in 20 and 2 mM glucose conditions were treated with 2% SDS containing loading buffer, boiled for 20 min and analyzed by SDS-PAGE/Western blot to detect proinsulin, n = 10. Proinsulin detection was performed with the CCI-17 monoclonal antibody. All Western blot quantifications were carried out with ImageJ. The data were analyzed by the unpaired Student’s t-test and are presented as means SD ±. (E). A schematic representation of the proinsulin isomers’ folding steps is proposed in this work.