Skip to main content
NCBI home page
Cell Stress & Chaperones logoLink to Cell Stress & Chaperones
. 2020 Nov 27;26(2):377–386. doi: 10.1007/s12192-020-01183-0

FKBP22 from the psychrophilic bacterium Shewanella sp. SIB1 selectively binds to the reduced state of insulin to prevent its aggregation

Cahyo Budiman 1,2,3,✉, Carlmond Kah Wun Goh 3, Irma Isnafia Arief 2, Muhammad Yusuf 2
PMCID: PMC7925751  PMID: 33247372

Abstract

FKBP22 of a psychrophilic bacterium, Shewanella sp. SIB1 (SIB1 FKBP22), is a member of peptidyl-prolyl cis-trans isomerase (PPIase) and consists of N- and C-domains responsible for chaperone-like and PPIase catalytic activities, respectively. The chaperone-like activity of SIB1 FKBP22 was previously evidenced by its ability to prevent dithiothreitol (DTT)-induced insulin aggregation. Nevertheless, the mechanism by which this protein inhibits the aggregation remains unclear. To address this, the binding affinity of SIB1 FKBP22 to the native or reduced states of insulin was examined using surface plasmon resonance (SPR). The native and reduced states refer to insulin in the absence or DTT presence, respectively. The SPR sensorgram showed that SIB1 FKBP22 binds specifically to the reduced state of insulin, with a KD value of 37.31 ± 3.20 μM. This binding was facilitated by the N-domain, as indicated by the comparable KD values of the N-domain and SIB1 FKBP22. Meanwhile, the reduced state of insulin was found to have no affinity towards the C-domain. The KD value of SIB1 FKBP22 was slightly decreased by NaCl but was not severely affected by FK506, a specific FKBP inhibitor. Similarly, the prevention of DTT-induced aggregation by SIB1 FKBP22 was also modulated by the N-domain and was not affected by FK506. Further, the reduced and native states of insulin had no effect on the catalytic efficiency (kcat/KM) of SIB1 FKBP22 towards a peptide substrate. Nevertheless, the reduced state of insulin slightly reduced the catalytic efficiency towards refolding RNase T1, at up to 1.5-fold lower than in the absence of insulin. These results suggested that the binding event was mainly facilitated by hydrophobic interaction and was independent from its PPIase activity. Altogether, a possible mechanism by which SIB1 FKBP22 prevents DTT-induced insulin aggregation was proposed.

Keywords: FKBP, Psychrophilic, Chaperone, Peptidyl prolyl cis-trans isomerase (PPIase), Surface plasmon resonance

Introduction

A 22-kDa FK506-binding protein of a psychrophilic bacterium, Shewanella sp. SIB1 (SIB1 FKBP22), was assumed to be involved in the cold adaptation of this bacterium (Suzuki et al. 2004; Budiman et al. 2011). This protein belongs to peptidyl-prolyl cis-trans isomerase (PPIase) family of proteins due to its ability to catalyze the slow cis-trans isomerization of prolyl bonds (Suzuki et al. 2004). SIB1 FKBP22 shows amino acid sequence identities of about 60% and 70% to FKBP22 of Escherichia coli (Ec-FKBP22) and human FKBP12 (Hs-FKBP12), respectively (Suzuki et al. 2004). The three-dimensional model of this protein showed that this protein consists of N- and C-terminal domains responsible for catalytic PPIase activity and dimerization, respectively (Budiman et al. 2011; Suzuki et al. 2005). The dimerization of the N-domain allows this protein to form a V-shaped dimeric structure. The V-form platform is located at the N-domain’s dimeric interface, while the two C-domains are located at the tips of the V-shaped form (Fig. 1a).

Fig. 1.

Fig. 1

Open in a new tab

a The three-dimensional models of full-length SIB1 FKBP22, N-domain+, and C-domain+. The long α3 helices connecting N- and C-domains in SIB1 FKBP22 are shown in cyan. b Schematic representations of the primary structures of full-length SIB1 FKBP22, N-domain+, and C-domain+. A His-tag attached to the N-termini of the proteins is represented by a gray box. Numbers indicate the positions of the residues relative to the initiator methionine residue of the proteins without a His-tag. The figures are not to scale

Biochemical characterization has revealed that SIB1 FKBP22 exhibits the dual function of foldase and chaperone-like activities (Budiman et al. 2011; Budiman et al. 2012). SIB1 FKBP22 has demonstrated the foldase activity through catalysis of the slow refolding of a cis-prolyl bond–containing protein RNase T1. This activity was previously found to be dependent on its PPIase catalytic activity (Suzuki et al. 2004). Meanwhile, the chaperone-like activity of SIB1 FKBP22 was indicated by the binding of SIB1 FKBP22 to a folding intermediate of α-lactalbumin and prevention of dithiothreitol (DTT)-induced insulin aggregation (Budiman et al. 2012; Suzuki et al. 2005). The chaperone-like activity of SIB1 FKBP22 was further found to be independent of its PPIase catalytic activity (Budiman et al. 2012). The dual function of SIB1 FKBP22 leads to an assumption that this protein may be involved in cold adaptation of SIB1 cells through accelerating slow folding of cis-prolyl bond–containing proteins at low temperatures as well as preventing the protein aggregation due to cold shock stress (Budiman et al. 2011). We have previously proposed that the V-shaped form, which is facilitated by the N-domain, is essential for its chaperone-like activity, as this shape facilitates an efficient binding to the protein substrate (Budiman et al. 2009). Furthermore, the ability to prevent DTT-induced insulin aggregation has been speculated to be due to the ability of SIB1 FKBP22 to bind to insulin and shield it from DTT reduction (Budiman et al. 2012). Nevertheless, this speculation has never been experimentally evidenced.

Insulin is composed of two peptide chains referred to as the A and B chains. These chains are linked together by two disulfide bonds, and an additional disulfide is formed within the A chain (Chang et al. 2003; Kurouski et al. 2012; Jia et al. 2003). This protein is soluble in water and is presented mainly as a hexamer at neutral pH. Disruption of these disulfide bonds leads to separation of the A and B chains, whereby B chain, which is relatively hydrophobic, aggregates and precipitates. By contrast, the relatively hydrophilic A chain is stable and stays in solution as an unstructured entity. Insulin is especially interesting in this aspect, because the aggregation of insulin, induced by disorder in its native structure, can cause numerous pathological states (Bumagina et al. 2010; Vekshin 2008). Zako et al. (2009) reported that reducing all disulfide bonds of the native state of insulin leads to structurally and morphologically different insulin fibrils that mimic diseases associated with amyloid fibril structures. This suggests that the DTT reduction leads to the conversion of the native state of insulin to the denatured state, likely through a folding intermediate state (Yan et al. 2003; Tang et al. 2007). Accordingly, insulin has been widely used as a model substrate in folding studies and chaperone mechanism (Bumagina et al. 2010; Reddy et al. 2002; Ecroyd and Carver 2008; Tomoyasu et al. 2010). Hua (2010) also stated that as a small protein, insulin is a good model for basic structural research, especially for studying protein folding intermediates with a variety of disulfide pairings.

The mechanism by which a chaperone prevents insulin aggregation has been previously proposed. Lindner et al. (1998) explored the interaction of reduced insulin with the molecular chaperone α-crystallin extensively and concluded that α-crystallin interacted with and stabilized an intermediately folded (molten globule) form of the insulin B chain. In addition, Vekshin (2008) proposed that α-crystallin prevented insulin aggregation through binding to the non-native state of insulin. The non-native state of insulin might refer to the reduced state of insulin, in which DTT reduced the disulfide bonds of this protein. The mechanism by which SIB1 FKBP22 prevents DTT-induced insulin aggregation that follows α-crystallin is therefore an interesting avenue to be explored.

In this report, the binding of SIB1 FKBP22 to the reduced and native states of insulin was analyzed using surface plasmon resonance. Further, the involvement of PPIase activity in the binding event was also investigated. The results have allowed us to propose a possible mechanism by which SIB1 FKBP22 prevents DTT-induced insulin aggregation. This report may also solidify the understanding of the mechanism by which chaperones prevent aggregation of proteins.

Materials and methods

Overproduction and purification

E. coli BL21(DE3) (F− ompT hsdSB(rB−mB−) gal (λ cI857 ind1 Sam7 nin5 lacUV5-T7gene1) dcm (DE3)) (Novagen Inc., Madison, WI, USA) was used as a host strain for the overproduction of N-terminal His-tagged forms of full-length SIB1 FKBP22, N-domain+ (residues 9–93), and C-domain+ (residues 95–205). While SIB1 FKBP22 refers to the full-length form of this protein, N-domain+ and C-domain+ correspond to the variants of SIB1 FKBP22 N- or C-domain containing the α3-helix segment, respectively. The plasmids for expression of SIB1 FKBP22, N-domain+, or C-domain+ were prepared as previously described (Suzuki et al., 2004; Suzuki et al., 2005) and designated as pSIB1-FKBP22, pSIB1-Ndo, and pSIB1-Cdo, respectively. Transformation of the E. coli cells with pSIB1-FKBP22 or pSIB1-Ndo or pSIB1-Cdo and overproduction and purification of the recombinant proteins were carried out as previously described (Suzuki et al. 2004; Budiman et al. 2009). Production levels of the recombinant proteins in E. coli cells and their purities were analyzed by SDS-PAGE using 15% polyacrylamide gel, followed by staining with Coomassie brilliant blue (Laemmli 1970). Protein concentrations of SIB1 FKBP22 and its derivatives that were determined from the UV absorption on the basis of the absorbance at 280 nm of a 0.1% (1 mg/mL) solution are 0.68 for SIB1-FKBP22, 0.75 for C-domain+, and 0.17 for N-domain+. These values were calculated based on Goodwin and Morton (1946).

Surface plasmon resonance

The interaction between the protein sample and insulin was monitored by surface plasmon resonance using the Biacore X instrument (Biacore, Uppsala, Sweden). Immobilization of the His-tagged form of SIB1 FKBP22 to Ni2+-chelated NTA (nitrilotriacetic acid) sensor chip (Biacore) was carried out as described previously (Suzuki et al. 2005; Budiman et al. 2009). The native or reduced state of insulin was prepared in 20 mM sodium phosphate (pH 8.0) in the absence or the presence of 20 mM DTT (final concentration), respectively, with 1 mM EDTA. Various concentrations of the native or reduced state of insulin were then injected onto the surface of the sensor chip on which the His-tagged SIB1 FKBP22 was immobilized at 10 °C with a flow rate of 10 μL/min. Binding surfaces were regenerated by washing with 0.5 M EDTA. To determine the dissociation constant, KD, the concentration of reduced state of insulin injected onto the sensor chip was varied from 1 to 100 μM. The plot of the equilibrium binding responses acted as a function of the concentrations of the reduced state of insulin.

To determine the effect of NaCl or FK506 (a specific inhibitor of FKBP), surface plasmon resonance experiments were conducted as described above in the presence of NaCl (50, 100, 150, and 200 mM) or FK506 (80 nM and 160 nM). The plot of the equilibrium binding responses in the presence of NaCl or FK506 was then derived for KD calculation. The KD values obtained from the analysis with NaCl or FK506 were then compared to the KD values obtained from the equilibrium binding responses in the absence of NaCl or FK506 (a control).

To determine the domain responsible for binding to reduced insulin, surface plasmon resonance experiments were also conducted for the N-domain+ and C-domain+. For this purpose, immobilization of the His-tagged form of N-domain+ or C-domain+ to Ni2+-chelated NTA (nitrilotriacetic acid) sensor chip (Biacore) was carried out as SIB1 FKBP22, followed by injection of various concentrations of the reduced state of insulin as describe above. No NaCl or FK506 was added in this experiment. The plot of the equilibrium binding responses of the domains were then derived for KD calculation followed by comparison with the KD obtained from the equilibrium binding responses of SIB1 FKBP22 in the absence of NaCl or FK506 (as a control).

In all experiments, the KD value was determined using a steady-state affinity program of BIA evaluation software (Biacore). For this purpose, the curve was fitted into a single binding-site affinity model under the equation of =Bmax.XKD+X, where Y, Bmax, X, and KD are the specific binding, the concentration of ligand, and the dissociation constant, respectively.

Insulin aggregation assay

The insulin aggregation assay was performed as described previously (Budiman et al. 2012; Guha et al. 1998) with slight modifications. Insulin was dissolved in 20 mM sodium phosphate (pH 8.0) at 0.2 mg/mL. The reduction of disulfide bonds of insulin was initiated by adding dithiothreitol (DTT) to the final concentration of 20 mM. Aggregation of insulin in the presence of 0.2 mg/mL of SIB1 FKBP22 in the absence or the presence of 80 or 160 nM of FK506 was measured at 10 °C by monitoring the change in light scattering at 465 nm as a function of time using a Hitachi F-2500 spectrofluorometer (Hitachi High-Technologies Co.).

In addition, to determine the domain responsible for insulin aggregation prevention, the ability of N-domain+ or C-domain+ to prevent DTT-induced insulin aggregation was also measured. For this purpose, the light scattering changes of DTT-induced insulin aggregation were monitored in the presence full-length SIB1 FKBP22, N-domain+, or C-domain+ at the same concentration of insulin (0.2 mg/mL). Under these conditions, the molar ratios of full-length SIB1 FKBP22, N-domain+, or C-domain+ to insulin were about 1:3.9, 1:4.3, and 1:3.0, respectively. The detailed assay followed the condition for SIB1 FKBP22, with no FK506 present in the cocktail.

PPIase activities of SIB1 FKBP22 towards peptide and protein substrates

The PPIase activity towards a peptide substrate of N-succinyl-Ala-Leu-Pro-Phe-p-nitroanilide (Suc-ALPF-pNA) (Wako Pure Chemical) was determined in 35 mM Hepes buffer, pH 7.8, at 10 °C in a protease-coupled assay (Budiman et al. 2018). FK506 stock (Sigma-Aldrich, St. Louis, MO) was prepared in 50% ethanol. The native or reduced state of insulin was prepared in 20 mM sodium phosphate (pH 8.0) in the absence or presence of 20 mM DTT (final concentration), respectively, with 1 mM EDTA. The cocktail was incubated with various concentrations of the native or reduced state of insulin for 5 min before the addition of chymotrypsin to start the reaction. The final concentration of native or reduced states of insulin was 100 μM. The reaction in the absence of insulin was performed as a control. PPIase activity was measured through pNA released under a Hitachi U-2010 UV/Vis spectrophotometer (Hitachi High-Technologies Co., Tokyo, Japan) at 390 nm. The catalytic efficiency (kcat/KM) was calculated from the relationship kcat/KM = (kp − kn)/ E, where E represents the concentration of the enzyme and kp and kn represent the first-order rate constants for the release of pNA from the substrate in the presence and absence of the enzyme, respectively (Harrison and Stein 1990).

The PPIase activity towards a protein substrate was determined using RNase T1 as a model substrate under a refolding assay, according to Budiman et al. (2012). RNase T1 (16 μM) (Funakoshi, Tokyo, Japan) was first unfolded by incubating it in 20 mM sodium phosphate (pH 8.0) containing 0.1 mM EDTA and 6.2 M guanidine hydrochloride at 10 °C overnight. Refolding was then initiated by diluting this solution 80-fold with 20 mM sodium phosphate (pH 8.0) containing 100 mM NaCl in the presence or absence of the SIB1 FKBP22 and reduced or native state of insulin. The final concentrations of RNase T1 and SIB1 FKBP22 were 0.2 μM and 5 nM, respectively. Meanwhile, the final concentration of native or reduced state of insulin was 100 μM. The refolding reaction was monitored by measuring the increase in tryptophan fluorescence with an F-2000 spectrofluorometer (Hitachi High-Technologies). The excitation and emission wavelengths were 295 and 323 nm, respectively, and the bandwidth was 10 nm. The refolding curves were analyzed with a double exponential fit (Ramm and Plucktun 2000). The kcat/KM values were calculated as previously described (Suzuki et al. 2004).

Data analysis

The data were presented as mean ± standard deviation (SD) from at least three independent replications. The differences between the means were descriptively analyzed.

Results

Protein design and preparation

The three-dimensional models and the primary structures of the proteins used in this study are shown in Fig. 1. The two truncation proteins used here (Fig. 1), referred to as the N-domain+ and the C-domain+, each included the long α3-helix because this helix is essential for the folding of the two domains (Suzuki et al. 2004). All proteins accumulated in E. coli cells in a soluble form upon overproduction and were purified to give a single band on SDS-PAGE (data not shown). The amount of protein purified from 1 L culture was 7–10 mg for all proteins.

Binding SIB1 FKBP22 to insulin

To examine which state of insulin binds to SIB1 FKBP22, the binding of SIB1 FKBP22 to the two states of insulin was analyzed using surface plasmon resonance (Biacore). The native or reduced state of insulin was injected onto the sensor chip, on which SIB1 FKBP22 was immobilized. The amount of the protein immobilized onto the sensor chip was equivalent to 1200 resonance units (RU), or 69 ng/mm2. When the reduced state of insulin was injected onto the sensor chip, an increase in the RU value was observed, indicating the reduced state of insulin bound to the immobilized-SIB1 FKBP22 on the surface. In contrast, when the native state of insulin was injected onto the sensor chip, there was no increasing RU value obtained, suggesting no interaction between the native state of insulin and SIB1 FKBP22. The sensorgrams obtained by injecting 100 μM of the reduced or native states of insulin, as representatives, onto these sensor chips are shown in Fig. 2a.

Fig. 2.

Fig. 2

Open in a new tab

a Representative of sensorgrams from Biacore X showing the binding of 100 μM of reduced (solid line) and non-reduced (dotted line) insulin to immobilized SIB1 FKBP22. Injections were done at time zero for 60 s. b The equilibrium binding responses of SIB1 FKBP22 to reduced insulin are shown as a function of the concentration of reduced insulin. The dotted line represents the fitting curve of a single binding-site affinity model

As Fig. 2a clearly shows that only the reduced state of insulin binds to SIB1 FKBP22, we then performed quantitative measurement on the binding affinity between these two proteins. Nevertheless, as the association and dissociation of the reduced state of insulin were too fast to accurately determine the kinetic constants such as kon and koff, the dissociation constant, KD, was determined by measuring equilibrium binding responses at various concentrations of the reduced state of insulin. The plots of the equilibrium binding responses as a function of the concentration of the reduced state of insulin gave saturation curves, as shown in Fig. 2b, which fits a single binding affinity model well. Therefore, the KD value for binding the reduced state of insulin to SIB1 FKBP22 was determined to be 37.31 ± 3.2 μM.

Interestingly, N-domain+ and SIB1 FKBP22 showed comparable binding responses to 100 μM of the reduced state of insulin (Fig. 3a). Meanwhile, no increasing RU was observed when the reduced state of insulin, up to 500 μM, was injected onto immobilized C-domain+. This suggested that no binding event of C-domain+ and the reduced state of insulin was detected. Further, the equilibrium binding responses of N-domain+ and SIB1 FKBP22 (Fig. 3b) were also found to be similar. The KD value of the N-domain to the reduced state of insulin was 32.08 ± 6.21 μM, comparable to the KD value of SIB1 FKBP22 (Table 1). These results indicated that the N-domain modulates the binding affinity of SIB1 FKBP22 to the reduced state of insulin.

Fig. 3.

Fig. 3

Open in a new tab

a Representative of sensorgrams from Biacore X showing the binding of 100 μM of reduced insulin to SIB1 FKBP22 (solid line) or N-domain+ (dashed line). Meanwhile, the sensorgram of 500 μM of reduced insulin to C-domain+ is shown as dotted line. Injections were done at time zero for 60 s. b The equilibrium binding responses of SIB1 FKBP22 (black circle) or N-domain+ (white circle) to reduced insulin is shown as a function of the concentration of reduced insulin. The dotted and solid lines represent the fitting curve of SIB1 FKBP22 and N-domain+, respectively

Table 1.

Binding affinity of SIB1 FKBP22 and its domains to reduced insulin

Protein KD (μM)
SIB1 FKBP22 37.31 ± 3.20
N-domain+ 32.08 ± 6.21
C-domain+ Not detected
Open in a new tab

Effect of NaCl on binding to the reduced state of insulin

To confirm whether the binding between SIB1 FKBP22 and the reduced state of insulin is mainly facilitated by hydrophobic interaction, surface plasmon resonance was performed in various concentrations of NaCl. The presence of NaCl may disrupt electrostatic networks between SIB1 FKBP22 and the reduced state of insulin. The equilibrium binding responses as a function of the concentration of the reduced state of insulin in the presence of various concentrations of NaCl are shown in Fig. 4. The calculated KD values, as summarized in Table 2, showed that the values slightly decreased in the presence of 50 mM NaCl and kept decreasing as the NaCl concentration increased up to 200 mM. These results suggested that hydrophobic interactions are dominant forces for the association of the reduced state of insulin to SIB1 FKBP22.

Fig. 4.

Fig. 4

Open in a new tab

The equilibrium binding responses of SIB1 FKBP22 to reduced insulin in the absence (black circle) or in the presence of 50 mM (white circle), 100 (black triangle down), 150 mM (white triangle up), and 200 mM (black square) of NaCl with the fitting curve shown in solid black line, solid gray line, dashed black line, dotted gray line, and dotted black line, respectively

Table 2.

Binding affinity of SIB1 FKBP22 to reduced insulin in the presence of NaCl

NaCl (mM) KD (μM)
0 37.31 ± 3.20
50 24.93 ± 2.61
100 19.27 ± 2.83
150 11.03 ± 1.04
200 8.20 ± 0.59
Open in a new tab

Effect of FK506 on binding to the reduced state of insulin

FK506 is a 23-membered macrolide lactone which is known to reduce PPIase activity in T cells by binding to the immunophilin FKBP12. This drug is a standard component of immunosuppressive regimens currently in use for organ and reconstructive tissue transplants (Tung 2010). FK506 was reported to bind to FKBP family groups, including SIB1 FKBP22 specifically, and inhibits their catalytic PPIase activity (Fanghänel and Fischer 2004; Budiman et al. 2018). Thus, it is interesting to confirm whether the binding to the reduced state of insulin and the ability to prevent DTT-induced aggregation of insulin are affected by FK506. In this study, two concentrations of FK506 were examined (80 nM and 160 nM). The 80 nM concentration represents the IC50 value of FK506 towards SIB1 FKBP22, while at the concentration of 160 nM of FK506, the PPIase activity of SIB1 FKBP22 was completely inhibited (Budiman et al. 2018). Figure 5 showed that the equilibrium binding responses of the reduced state of insulin and SIB1 FKBP22 in the absence and the presence of different concentrations of FK506 were found to be similar. Furthermore, Table 3 also showed that the KD values of SIB1 FKBP22 to the reduced state of insulin were comparable for different concentrations of FK506. This suggested that the binding event to the reduced state of insulin was not seriously affected by FK506.

Fig. 5.

Fig. 5

Open in a new tab

The equilibrium binding responses of SIB1 FKBP22 to reduced insulin in the absence (black circle) or in the presence of 80 nM (white circle) and 160 nM (black triangle down) of FK506 with the fitting curve shown in solid black line, solid gray line, dotted black line, and solid gray line, respectively

Table 3.

Binding affinity of SIB1 FKBP22 to reduced insulin in the presence of NaCl

FK506 (nM) KD (μM)
0 37.31 ± 3.20
80 36.50 ± 5.30
160 39.53 ± 3.68
Open in a new tab

Prevention of DTT-induced insulin aggregation

Figure 6 showed that the light scattering of DTT-induced insulin aggregation in the presence of SIB1 FKBP22 was not severely affected by FK506. This suggested that the ability of SIB1 FKBP22 to prevent aggregation is independent of binding to FK506. To note, our attempts to evidence the binding event of FK506 to SIB1 FKBP22 through surface plasmon resonance were unsuccessfully obtained, which might be due to the small size of this molecule. Nevertheless, our previous study on the inhibition property of FK506 to SIB1 FKBP22 has convincingly implied that this molecule bound to SIB1 FKBP22, particularly at the catalytic site (Budiman et al. 2018).

Fig. 6.

Fig. 6

Open in a new tab

Representative of the relative changes on the light scattering of DTT-induced insulin aggregation measured in the absence (black) or in the presence of SIB1 FKBP22 with 0 nM (blue), 80 nM (gray), and 160 nM (red) of FK506. The insulin concentration was 0.2 mg/mL as detailed in “Materials and methods” section

It is interesting to note that Fig. 7 shows that N-domain+ significantly reduced the light scattering of DTT-induced insulin aggregation to the comparable level that of in the presence of SIB1 FKBP22. However, there was no reduction in the light scattering of DTT-induced insulin aggregation in the presence of C-domain+ (up to 2 mg/mL). These results suggested that N-domain+ is essential to generate the ability of SIB1 FKBP22 to prevent DTT-induced insulin aggregation.

Fig. 7.

Fig. 7

Open in a new tab

Representative of the relative changes on the light scattering of DTT-induced insulin aggregation measured in the absence (black) or in the presence of SIB1 FKBP22 (brown), N-domain+ (green), and C-domain+ (gray). The concentrations of SIB1 FKBP22 (brown), N-domain+ (green), and C-domain+ (gray) were adjusted to be the same as insulin (0.2 mg/mL)

Effect of insulin on the PPIase activity towards a peptide substrate

To confirm whether the presence of insulin interferes with the PPIase activity of SIB1 FKBP22 towards a peptide substrate, the activity towards Suc-Ala-Leu-Pro-Phe-pNA was performed in the absence or the presence of the native or reduced state of insulin. The PPIase activity towards the peptide substrate refers to catalysis of the cis-trans isomerization of a prolyl bond of the peptide by SIB1 FKBP22. Figure 8 shows that the first-order rate constant of the peptide substrate’s slow isomerization in the presence of 100 μM of the reduced or native state of insulin was similar to that in the absence of insulin. The calculated kcat/KM value of SIB1 FKBP22 towards the peptide substrate in the absence of the native or reduced state of insulin was found to be 0.52 ± 0.04 μM−1 s−1. This value was comparable to that in the presence of 100 μM of the native or reduced state of insulin (0.49 ± 0.08 μM−1 s−1) or 100 μM of the native state of insulin (0.51 ± 0.03 μM−1 s−1). These results indicated that insulin has no effect on the PPIase activity of SIB1 FKBP22 towards a peptide substrate.

Fig. 8.

Fig. 8

Open in a new tab

The first-order reaction rates (k) of cis-trans isomerization of Suc-Ala-Leu-Pro-Phe-pNA substrate by SIB1 FKBP22 in the absence (black circle) or in the presence of 100 μM of reduced (white circle) or native (gray circle) insulin

Effect of insulin on the PPIase activity towards a protein substrate

Figure 9 shows the refolding reaction of a protein substrate of RNase T1 in the absence (spontaneous) or the presence of SIB1 FKBP22 with or without insulin. This experiment was performed to determine whether native or reduced insulin has any effect on the catalytic activity of SIB1 FKBP22 to accelerate the spontaneous refolding rate of RNase T1. The refolding rate of RNase T1 in vitro is known to be the slow cis-trans isomerization of its peptidyl-prolyl bonds (Schmid 1993). The curves in Fig. 9 reflected the increase of Trp fluorescence during the refolding of denatured RNase T1 in the presence or absence of SIB1 FKBP22, with or without insulin. Figure 9 shows that the increase of Trp fluorescence during the refolding of RNase T1 was observed when SIB1 FKBP22 was added into the reaction. This indicated that SIB1 FKBP22 was able to accelerate the spontaneous refolding reaction of RNase T1. The calculated catalytic efficiency (kcat/KM) of SIB1 FKBP22 to accelerate the refolding reaction of RNase T1 was calculated to be 0.44 ± 0.01 μM−1 s−1. Interestingly, when the refolding reaction of RNase T1 was measured in the presence of RNase T1 and native insulin, the refolding reaction curve was comparable to that without addition of native insulin (Fig. 9), with the calculated kcat/KM value of 0.41 ± 0.02 μM−1 s−1. This suggested that the presence of native insulin has no effect on the catalytic activity of SIB1 FKBP22 towards RNase T1. By contrast, when the refolding reaction of RNase T1 was measured in the presence of SIB1 FKBP22 and the reduced state of insulin, the increase of Trp fluorescence during the refolding was found to be slightly slower than that in the presence of SIB1 FKBP22 (with or without native insulin), yet faster than the spontaneous refolding reaction (Fig. 9). The kcat/KM of the refolding reaction in the presence of reduced insulin was calculated to be 0.28 ± 0.05 μM−1 s−1. This suggested that the presence of reduced state of insulin affected the catalytic activity of SIB1 FKBP22 against RNase T1.

Fig. 9.

Fig. 9

Open in a new tab

The representative of the refolding reactions of RNase T1 in the presence or the absence of SIB1 FKBP22 with or without insulin. The refolding reactions in the presence of SIB1 FKBP22 with no reduced or native insulin are shown as a solid black line. Meanwhile, the reactions in the presence of SIB1 FKBP22 with the addition of 100 μM of reduced or native insulin are shown as dotted gray and dotted black lines, respectively. The spontaneous refolding reaction is also shown (solid gray line) as a control which refers to the reaction in the absence of SIB1 FKBP22 and insulin

Discussion

This study demonstrated that SIB1 FKBP22 specifically binds to the reduced state of insulin, but not to the native state. This finding is understandable as the hydrophobic region of the denatured state of insulin is believed to be more exposed than its native state (Vekshin 2008). The binding of chaperones to their client proteins was also reported to be facilitated by hydrophobic interaction (England and Pande 2008). Suzuki et al. (2005) reported that the binding of SIB1 FKBP22 to a folding intermediate of protein was dominated by hydrophobic interaction. Interestingly, NaCl was found to slightly decrease the KD value, which confirmed our previous hypothesis that hydrophobic interaction dominated the binding between SIB1 FKBP22 and the reduced state of insulin. Electrostatic interaction might exist in the complex yet does not play essential roles in the interaction. This result is in good agreement with the previous report showing that NaCl (up to 200 mM) also decreased the binding affinity of SIB1 FKBP22 to reduced α-lactalbumin (Suzuki et al. 2005). Similarly, NaCl did not lower the binding affinity of β-casein towards α-lactalbumin (Mossallatpour and Ghahghaei 2015).

To note, the KD value of SIB1 FKBP22 was found to be remarkably higher than that of interaction between SIB1 FKBP22 and the reduced α-lactalbumin (6.5 ± 0.38 μM) as reported previously (Suzuki et al. 2005). This indicated that the binding affinity of SIB1 FKBP22 to the reduced state of insulin was lower than that of reduced α-lactalbumin. This might be due to the smaller size of the insulin B chain, as compared to α-lactalbumin. Therefore, the hydrophobic region facilitating the binding was not as strong as that in α-lactalbumin.

Further, this study also showed that FK506, an inhibitor of PPIase activity of SIB1 FKBP22, did not affect the prevention of DTT-induced insulin aggregation by SIB1 FKBP22. This suggested that the binding site for FK506 and insulin at SIB1 FKBP22 are spatially different. Earlier, Budiman et al. (2018) reported that FK506 binds explicitly to the catalytic C-domain. Consequently, the binding site for insulin is therefore not located at the C-domain. This was evidenced by the lack of binding responses between C-domain+ and the reduced state of insulin under surface plasmon resonance (Fig. 3). Instead of the C-domain, the N-domain, which was previously reported to be the binding site for protein substrate (Budiman et al. 2009), is predicted to be the binding site of insulin. In the current study, it is evident that the reduced state of insulin binds selectively to N-domain, not C-domain (Fig. 3). The preferences of the reduced state of insulin to bind to the N-domain might be due to higher hydrophobicity of the N-domain than of the C-domain (Budiman et al. 2009). Further, the comparable activities of N-domain+ and SIB1 FKBP22 to prevent DTT-induced insulin aggregation (Fig. 7) confirmed the essential role of the N-domain for binding to the reduced state of insulin and prevention of DTT-induced insulin aggregation.

To note, Figs. 3 and 7 collectively confirm that the C-domain has no role in the binding to insulin or prevention of DTT-induced insulin aggregation. Consequently, the prevention of DTT-induced insulin aggregation of SIB1 FKBP22 is completely independent of its PPIase activity. The independence of the catalytic site on the binding event to the reduced state of insulin was also supported by finding of no effect of insulin on the PPIase activity towards the peptide substrate (Fig. 8). As previously demonstrated (Budiman et al. 2018, 2009), the peptide substrate and FK506 binding sites are localized at the C-domain and, therefore, should not be interfered with by binding of the reduced state of insulin. This result supports the previous finding that the aggregation prevention and PPIase activity of SIB1 FKBP22 are not related (Budiman et al. 2011, 2012).

Interestingly, the reduced state of insulin was found to slightly affect the PPIase activity towards RNase T1 (Fig. 9). This might be due to the binding competition between the reduced state of insulin and RNase T1 to the N-domain of SIB1 FKBP22. Earlier, the N-domain was proposed to facilitate the proper binding of SIB1 FKBP22 to RNase T1 (Budiman et al. 2012). Nevertheless, the effect of the reduced state of insulin was not so extreme, which might be due to the weaker and more transient binding of the reduced state of insulin as compared to RNase T1. Size-wise, RNase T1 is larger than insulin and, therefore, able to provide more hydrophobic patches to support the binding. Meanwhile, no effect on the native state of insulin to the PPIase activity is understandable because the native state of insulin has no interaction with SIB1 FKBP22.

Altogether, a possible mechanism by which SIB1 FKBP22 prevents DTT-induced insulin aggregation is explained as follows. In the presence of DTT, three disulfide bonds of insulin are reduced by producing separated A and B chains, in which the B chain tends to aggregate due to non-specific hydrophobic interactions. Previous reports indicated that the aggregation was mainly due to instability of the B chain, or the segments of it (Swiontek et al. 2019). Ivanova et al. (2009) added that the B chain, or a segment of it, is also predicted to be the primary determinant of insulin fibrillation. Lindner et al. (1998) also confirmed that upon reduction of its disulfide bonds, the insulin A chain remains in solution, but the B chain aggregates and precipitates. Prior to its aggregation, the B chain adopts an intermediately folded, molten globule-like form that is unstable. Once DTT reduces insulin, SIB1-FKBP22 binds to and forms a complex with, presumably, the molten globule-like form of the B chain. A similar idea was also proposed by Lindner et al. (1998) in which α-crystallin, as determined by real-time NMR spectroscopy, stabilized the unstable molten globule-like form (probably the monomer) of its client protein to prevent the aggregation. Unfortunately, our attempts to collect complex of SIB1 FKBP22 and insulin were so far unsuccessful, which probably due to their transient and fast interaction. Besides, further investigation of the specific binding of SIB1 FKBP22 to the B chain alone and its aggregation prevention remains challenging. This is due to difficulties in isolation of the B chain alone as this chain easily aggregated once it was separated from the A chain.

Conclusion

The current study confirmed that the prevention of DTT-induced insulin aggregation by SIB1 FKBP22 is modulated by the interaction between SIB1 FKBP22 to the reduced state of insulin instead of its native state. The binding event was found to be independent from the PPIase activity of SIB1 FKBP22 and facilitated by its N-domain as a binding site. This finding supports previous evidence that chaperone-like activities of SIB1 FKBP22 are independent of its PPIase activity. Besides, a possible mechanism by which SIB1 FKBP22 prevents the insulin aggregation proposed in this study may also apply to other chaperones exhibiting DTT-induced insulin aggregation prevention activity.

Acknowledgments

The authors thank NUS-Enzyme Research Unit for the access of the instrument.

Funding

This study was supported by the research grants of SLB0089-ST-2014 (C.B) and KRB0427-FR (I.I.A, C.B).

Footnotes

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

  1. Budiman C, Bando K, Angkawidjaja C, Koga Y, Takano K, Kanaya S. Engineering of monomeric FK506-binding protein 22 with peptidyl prolyl cis-trans isomerase. Importance of a V-shaped dimeric structure for binding to protein substrate. FEBS J. 2009;276:4091–4101. doi: 10.1111/j.1742-4658.2009.07116.x. [DOI] [PubMed] [Google Scholar]
  2. Budiman C, Koga Y, Takano K, Kanaya S. FK506-binding protein 22 from a psychrophilic bacterium, a cold shock-inducible peptidyl prolyl isomerase with the ability to assist in protein folding. Int J Mol Sci. 2011;12:5261–5284. doi: 10.3390/ijms12085261. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Budiman C, Lindang HU, Cheong BE, Rodrigues KF. Inhibition and substrate specificity properties of FKBP22 from a psychrotrophic bacterium, Shewanella sp. SIB1. Protein J. 2018;37:270–279. doi: 10.1007/s10930-018-9772-z. [DOI] [PubMed] [Google Scholar]
  4. Budiman C, Tadokoro T, Angkawidjaja C, Koga Y, Kanaya S. Role of polar and nonpolar residues at the active site for PPIase activity of FKBP22 from Shewanella sp. SIB1. FEBS J. 2012;279:976–986. doi: 10.1111/j.1742-4658.2012.08483.x. [DOI] [PubMed] [Google Scholar]
  5. Bumagina Z, Gurvits B, Artemova N, Muranov K, Kurganov B. Paradoxical acceleration of dithiothreitol-induced aggregation of insulin in the presence of a chaperone. Int J Mol Sci. 2010;11(11):4556–4579. doi: 10.3390/ijms11114556. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Chang SG, Choi KD, Jang SH, Shin HC. Role of disulfide bonds in the structure and activity of human insulin. Mol Cells. 2003;16(3):323–330. [PubMed] [Google Scholar]
  7. Ecroyd H, Carver JA. The effect of small molecules in modulating the chaperone activity of αB-crystallin against ordered and disordered protein aggregation. FEBS J. 2008;275:935–947. doi: 10.1111/j.1742-4658.2008.06257.x. [DOI] [PubMed] [Google Scholar]
  8. England JL, Pande VS. Potential for modulation of the hydrophobic effect inside chaperonins. Biophys J. 2008;95:3391–3399. doi: 10.1529/biophysj.108.131037. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Fanghänel J, Fischer G. Insights into the catalytic mechanism of peptidyl prolyl cis/trans isomerases. Front Biosci. 2004;9:3453–3478. doi: 10.2741/1494. [DOI] [PubMed] [Google Scholar]
  10. Goodwin TW, Morton RA. The spectrophotometric determination of tyrosine and tryptophan in proteins. Biochem J. 1946;40:628–632. doi: 10.1042/bj0400628. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Guha S, Manna TK, Das KP, Bhattacharyya B. Chaperone-like activity of tubulin. J Biol Chem. 1998;273:30077–30080. doi: 10.1074/jbc.273.46.30077. [DOI] [PubMed] [Google Scholar]
  12. Harrison RK, Stein RL. Mechanistic studies of peptidyl prolyl cis-trans isomerase: evidence for catalysis by distortion. Biochemistry. 1990;29:1684–1689. doi: 10.1021/bi00459a003. [DOI] [PubMed] [Google Scholar]
  13. Hua Q. Insulin: a small protein with a long journey. Protein Cell. 2010;1(6):537–551. doi: 10.1007/s13238-010-0069-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Ivanova MI, Sievers SA, Sawaya MR, Wall JS, Eisenberg D. Molecular basis for insulin fibril assembly. Proc Natl Acad Sci. 2009;106(45):18990–18995. doi: 10.1073/pnas.0910080106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Jia XY, Guo ZY, Wang Y, Xu Y, Duan SS, Feng YM. Peptide models of four possible insulin folding intermediates with two disulfides. Protein Sci. 2003;12(11):2412–2419. doi: 10.1110/ps.0389303. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Kurouski D, Washington J, Ozbil M, Prabhakar R, Shekhtman A, Lednev IK. Disulfide bridges remain intact while native insulin converts into amyloid fibrils. PLoS One. 2012;7(6):e36989. doi: 10.1371/journal.pone.0036989. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Laemmli UK. Clevage of structural proteins during the assembly of the head of bacteriophage T4. Nature. 1970;227:680–685. doi: 10.1038/227680a0. [DOI] [PubMed] [Google Scholar]
  18. Lindner RA, Kapur A, Mariani M, Titmuss SJ, Carver JA. Structural alterations of α-crystallin during its chaperone action. Eur J Biochem. 1998;258:170–183. doi: 10.1046/j.1432-1327.1998.2580170.x. [DOI] [PubMed] [Google Scholar]
  19. Mossallatpour S, Ghahghaei A. The effect of hoffmeister salts on the chaperoning action of β-casein in preventing aggregation of reduced α-lactalbumin. Biom J. 2015;1:58–68. [Google Scholar]
  20. Ramm K, Pluckthun A. The periplasmic Escherichia coli peptidylprolyl cis,trans-isomerase FkpA. II. Isomerase-independent chaperone activity in vitro. J Biol Chem. 2000;275:17106–17113. doi: 10.1074/jbc.M910234199. [DOI] [PubMed] [Google Scholar]
  21. Reddy GB, Narayanan S, Reddy PY, Surolia I. Suppression of DTT-induced aggregation of abrin by KA- and KB-crystallins: a model aggregation assay for K-crystallin chaperone activity in vitro1. FEBS Lett. 2002;522:59–64. doi: 10.1016/S0014-5793(02)02884-3. [DOI] [PubMed] [Google Scholar]
  22. Schmid FX. Prolyl isomerase: enzymatic catalysis of slow protein-folding reactions. Annu Rev Biophys Biomol Struc. 1993;22:123–143. doi: 10.1146/annurev.bb.22.060193.001011. [DOI] [PubMed] [Google Scholar]
  23. Suzuki Y, Haruki M, Takano K, Morikawa M, Kanaya S. Possible involvement of an FKBP family member protein from a psychrotrophic bacterium Shewanella sp. SIB1 in cold-adaptation. Eur J Biochem. 2004;271:1372–1381. doi: 10.1111/j.1432-1033.2004.04049.x. [DOI] [PubMed] [Google Scholar]
  24. Suzuki Y, Takano K, Kanaya S. Stabilities and activities of the N- and C-domains of FKBP22 from a psychrotrophic bacterium overproduced in Escherichia coli. FEBS J. 2005;272:632–642. doi: 10.1111/j.1742-4658.2004.04468.x. [DOI] [PubMed] [Google Scholar]
  25. Suzuki Y, Win OY, Koga Y, Takano K, Kanaya S. Binding analysis of a psychrotrophic FKBP22 to a folding intermediate of protein using surface plasmon resonance. FEBS Lett. 2005;579:5781–5784. doi: 10.1016/j.febslet.2005.09.067. [DOI] [PubMed] [Google Scholar]
  26. Swiontek M, Fraczyk J, Wasko J, Chaberska A, Pietrzak L, Zj K, Szymanski L, Wiak S, Kolesinska B. Search for new aggregable fragments of human insulin. Molecules. 2019;24:1–20. doi: 10.3390/molecules24081600. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Tang YH, Wang S, Chen Y, Xu GJ, Feng YM. In vitro insulin refolding: characterization of the intermediates and the putative folding pathway. Life Sci. 2007;50:717–725. doi: 10.1007/s11427-007-0092-3. [DOI] [PubMed] [Google Scholar]
  28. Tomoyasu T, Tabata A, Nagamune H. Investigation of the chaperone function of the small heat shock protein-AgsA. BMC Biochem. 2010;11(27):1–14. doi: 10.1186/1471-2091-11-27. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Tung TH. Tacrolimus (FK506): safety and applications in reconstructive surgery. HAND. 2010;5:1–8. doi: 10.1007/s11552-009-9193-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Vekshin NL. How α-crystallin prevents the aggregation of insulin. Biochemistry. 2008;73:458–462. doi: 10.1134/s0006297908040111. [DOI] [PubMed] [Google Scholar]
  31. Yan H, Guo ZY, Gong XW, Xi D, Feng YM. A peptide model of insulin folding intermediate with one disulfide. Protein Sci. 2003;12:768–775. doi: 10.1110/ps.0237203. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Zako T, Sakono M, Hashimoto N, Ihara M, Maeda M. Bovine insulin filaments induced by reducing disulfide bonds show a different morphology, secondary structure, and cell toxicity from intact insulin amyloid fibrils. Biophys J. 2009;96(8):3331–3340. doi: 10.1016/j.bpj.2008.12.3957. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Cell Stress & Chaperones are provided here courtesy of Elsevier

RESOURCES