Condensation of the β-cell secretory granule luminal cargoes pro/insulin and ICA512 RESP18 homology domain

Abstract

ICA512/PTPRN is a receptor tyrosine-like phosphatase implicated in the biogenesis and turnover of the insulin secretory granules (SGs) in pancreatic islet beta cells. Previously we found biophysical evidence that its luminal RESP18 homology domain (RESP18HD) forms a biomolecular condensate and interacts with insulin in vitro at close-to-neutral pH, that is, in conditions resembling those present in the early secretory pathway. Here we provide further evidence for the relevance of these findings by showing that at pH 6.8 RESP18HD interacts also with proinsulin-the physiological insulin precursor found in the early secretory pathway and the major luminal cargo of β-cell nascent SGs. Our light scattering analyses indicate that RESP18HD and proinsulin, but also insulin, populate nanocondensates ranging in size from 15 to 300 nm and 10e2 to 10e6 molecules. Co-condensation of RESP18HD with proinsulin/insulin transforms the initial nanocondensates into microcondensates (size >1 μm). The intrinsic tendency of proinsulin to self-condensate implies that, in the ER, a chaperoning mechanism must arrest its spontaneous intermolecular condensation to allow for proper intramolecular folding. These data further suggest that proinsulin is an early driver of insulin SG biogenesis, in a process in which its co-condensation with RESP18HD participates in their phase separation from other secretory proteins in transit through the same compartments but destined to other routes. Through the cytosolic tail of ICA512, proinsulin co-condensation with RESP18HD may further orchestrate the recruitment of cytosolic factors involved in membrane budding and fission of transport vesicles and nascent SGs.

Keywords: insulin; mesoscopic clusters; nanocondensates; proinsulin; protein secretion; protein sorting; protein tyrosine phosphatase; protein-protein interactions; secretory granule biogenesis; β cell.

FIGURE 1

ICA512 domain layout and RESP18HD sequence. (a) The SG luminal segment of ICA512 (UniProtKB Q16849) comprises residues 1–575 and includes a signal peptide (SP), the NTF, and the MPE. The transmembrane domain (TM) comprises residues 576–600. The cytoplasmic region, residues 601–979, is made of the juxtamembrane intracellular domain (IC) and the cytoplasmic cleaved fragment (CCF). Most of the cytoplasmic cleaved fragment corresponds to the pseudo‐phosphatase (catalytically inactive) domain (PTP), from which the entire protein is named. Scissors mark characterized processing sites. Convertase‐mediated cleavage at residue 448 generates the transmembrane fragment (ICA512 TMF, residues 449–979, corresponding to MPE–TM–IC), which is transiently inserted into the plasma membrane upon exocytosis. ICA512 RESP18HD constitutes the N‐terminal portion of ICA512 NTF. (b) Sequence alignment of ICA512 RESP18HD, Resp18 (UniProtKB Q5W5W9), and phogrin RESP18HD (UniProtKB Q92932). Conserved residues and cysteine residues are indicated in orange and white, respectively. An ICA512 RESP18HD C‐terminal cleavage site between Arg‐124 and Asp‐125 in ICA512 RESP18HD is indicated with a down arrow. Most of the IDD accounting for ICA512 RESP18HD condensing activity is the product of exon‐4, which encodes for ICA512 residues 94–126 (underlined). IDD, intrinsically disordered domain; MPE, membrane‐proximal ectodomain; NTF, N‐terminal fragment; SG, secretory granule.

FIGURE 2

Co‐condensation of RESP18HD124 and insulin/proinsulin monitored by SDS–PAGE and Coomassie‐blue staining. Proteins were incubated or co‐incubated at 20°C and pH 6.8. Pellets (P) and supernatants (S) separated by centrifugation from the incubates were TCA‐precipitated and quantitatively loaded into the SDS–PAGE lanes. Arrowheads and asterisks indicate the electrophoretic mobility of RESP18HD124 and insulin/proinsulin, respectively.

FIGURE 3

Co‐condensation of RESP18HD124 with insulin or proinsulin. The reaction was monitored by light scattering (apparent absorbance at 400 nm). RESP18HD124 (2 μM), insulin (8 μM), and proinsulin (8 μM) were incubated or co‐incubated at 20°C and pH 6.8. Averages of three independent experiments are shown.

FIGURE 4

DTT driven co‐condensation of insulin and proinsulin. The reaction protocol of Figure 3 was modified as to include a large excess of insulin or proinsulin (100 μM) over RESP18HD (8 μM) and 1 mM DTT in the reaction medium. Averages of three independent experiments are shown. DTT, dithiothreitol.

FIGURE 5

DTT‐driven condensation of insulin and RESP18HD‐NEM. RESP18HD124‐NEM is a variant of RESP18HD124 fully S‐alkylated with N‐ethylmaleimide. The reaction conditions were the same as in Figure 4. The well‐known reaction between insulin, TRX, and DTT was included as a control (see the main text). Averages of three independent experiments are shown. DTT, dithiothreitol; TRX, thioredoxin.

FIGURE 6

DLS size distribution of RESP18HD124. Measurements at pH 4.5 correspond to the condition of RESP18HD124 storage. Measurements at pH 6.8 correspond to the condensation assay condition and incubation times of Figure 3. Twelve or more replicates for each sample were automatically processed by the Zetasizer's software. Averages of two or more, independent experiments are shown.

FIGURE 7

DLS analysis of insulin and co‐incubates of RESP18HD124 and insulin. Samples were incubated at pH 6.8, the condensation assay condition. Twelve or more replicates for each sample were automatically processed by the Zetasizer's software. Averages of two or more, independent experiments are shown. DLS, dynamic light scattering.

FIGURE 8

DLS analysis of proinsulin and RESP18HD124 + proinsulin. Samples were incubated at pH 6.8, in the condensation assay conditions. Twelve or more replicates for each sample were automatically processed by the Zetasizer's software. Averages of two or more, independent experiments are shown. DLS, dynamic light scattering.

FIGURE 9

DLS analysis of proinsulin and proinsulin–Zn²⁺. Samples were incubated at the indicated pHs. Proinsulin and Zn²⁺ concentrations were 20 μM and 100 μM, respectively. Twelve or more replicates for each sample were automatically processed by the Zetasizer's software. Averages of two or more, independent experiments are shown. DLS, dynamic light scattering.

FIGURE 10

DLS analysis of proinsulin and proinsulin–Ca²⁺. Samples were incubated at the indicated pHs. Proinsulin and Ca²⁺ concentrations were 20 μM and 100 μM, respectively. Twelve or more replicates for each sample were automatically processed by the Zetasizer's software. Averages of two or more, independent experiments are shown. DLS, dynamic light scattering.