pH Induced Conformational Transitions in the Transforming Growth Factor β-Induced Protein (TGFβIp) Associated Corneal Dystrophy Mutants

1 Scientific RepoRts | 6:23836 | DOI: 10.1038/srep23836 www.nature.com/scientificreports pH Induced Conformational Transitions in the Transforming Growth Factor β-Induced Protein (TGFβIp) Associated Corneal Dystrophy Mutants Elavazhagan Murugan1,2, Anandalakshmi Venkatraman1, Zhou Lei3, Victoria Mouvet1, Rayne Rui Yi Lim1, Nandhakumar Muruganantham4, Eunice Goh4, Gary Swee Lim Peh1,2, Roger W. Beuerman2,4,5,6, Shyam S. Chaurasia1,2,5, Lakshminarayanan Rajamani2,4,5 & Jodhbir S. Mehta1,2,5,6 Most stromal corneal dystrophies are associated with aggregation and deposition of the mutated transforming growth factor-β induced protein (TGFβIp). The 4th_FAS1 domain of TGFβIp harbors ~80% of the mutations that forms amyloidogenic and non-amyloidogenic aggregates. To understand the mechanism of aggregation and the differences between the amyloidogenic and non-amyloidogenic phenotypes, we expressed the 4th_FAS1 domains of TGFβIp carrying the mutations R555W (nonamyloidogenic) and H572R (amyloidogenic) along with the wild-type (WT). R555W was more susceptible to acidic pH compared to H572R and displayed varying chemical stabilities with decreasing pH. Thermal denaturation studies at acidic pH showed that while WT did not undergo any conformational transition, the mutants exhibited a clear pH-dependent irreversible conversion from αβ conformation to β-sheet oligomers. The β-oligomers of both mutants were stable at physiological temperature and pH. Electron microscopy and dynamic light scattering studies showed that β-oligomers of H572R were larger compared to R555W. The β-oligomers of both mutants were cytotoxic to primary human corneal stromal fibroblast (pHCSF) cells. The β-oligomers of both mutants exhibit variations in their morphologies, sizes, thermal and chemical stabilities, aggregation patterns and cytotoxicities. Corneal Dystrophies are inherited protein aggregation disorders characterized by the deposition of misfolded proteins aggregates in various layers of the cornea1–3. Most dystrophies in the corneal stromal region are associated with the mutations in the transforming growth factor β -induced protein (TGFβ Ip). TGFβ Ip aggregation and deposition occurs only in the cornea, though it is present in abundance in various connective tissues4–7. The mature TGFβ Ip, a 660aa protein has an N-terminal cysteine-rich EMILIN-like (EMI) domain, four fascicilin-like (FAS1) domains and an integrin-binding RGD motif at the C-terminus8. TGFβ Ip-associated corneal dystrophies are phenotypically heterogeneous, inherited in an autosomal dominant manner1,9,10 and are classified as lattice, granular, combined lattice and granular, Reis-Buckler and Thiel-Behnke corneal dystrophies1,2,11,12. So far, 64 single amino acid mutations associated with distinct phenotypes have been reported4,13,14. Among the four FAS1 domains of TGFβ Ip, the 1st and 4th FAS1 domains carry the disease related mutations, with ~80% of the mutations residing in the 4th_FAS1 domain11. In lattice corneal dystrophies (LCD), the protein aggregates appear as 1Tissue Engineering and Stem Cell Group, Singapore Eye Research Institute, Singapore. 2Duke-NUS Graduate Medical School, Singapore. 3Proteomics and Microanalysis laboratory, Singapore Eye Research Institute, Singapore. 4Ocular Chemistry and Anti-Infectives, Singapore Eye Research Institute, Singapore. 5Department of Ophthalmology, Yong Loo Lin School of Medicine, NUS, Singapore. 6Singapore National Eye Centre, Singapore. Correspondence and requests for materials should be addressed to J.S.M. (email: [email protected]) or L.R. (email: [email protected]) Received: 17 June 2015 Accepted: 12 February 2016 Published: 31 March 2016 OPEN www.nature.com/scientificreports/ 2 Scientific RepoRts | 6:23836 | DOI: 10.1038/srep23836 lattice lines or amyloid fibrils (amyloidogenic). In granular corneal dystrophies (GCD), they appear as granular, non-amyloidogenic deposits11–16. Both phenotypes display significant differences in morphology, aggregation and tinctorial properties. However, the mechanisms adopted by these phenotypes in forming highly distinct ultrastructures, remains to be elucidated. It has been reported that the ability to form highly ordered aggregates such as amyloids, resides within the polypeptide chains rather than the whole protein17. Hence, we chose to explore the crucial 4th_FAS1 domain (135aa) as a representative of the full-length TGFβ Ip. Bioinformatics analyses of the aggregation propensities of various regions of TGFβ Ip have also shown that the 4th_FAS1 domain harbors regions of high aggregation propensities18,19. Also, homology-based modelling studies have shown that the 4th_FAS1 domain displays the properties of the full-length TGFβ Ip 20. In our study, we aim to delineate the differences between the amyloidogenic and the non-amyloidogenic mutants and between the mutants and wild-type (WT) TGFβ Ip to explain the physiological variations exhibited by these phenotypes. We had previously reported the cloning, expression and purification of the 4th_FAS1 domain of four TGFβ Ip mut ant s21. Our studies indicated that under physiological pH, the mutants are more stable than the WT. Here, we chose to examine the effects of factors like temperature and pH on the properties of the 4th_ FAS1 domains of TGFβ Ip harboring mutations of the amyloidogenic and non-amyloidogenic phenotypes. The amyloidogenic H572R (LCDI/IIIA), discovered in Thai22 and Chilean populations23 with ages of onset ranging around mid-twenties, is characterized by central sub-epithelial needle-like lattice lines and polymorphic anterior stromal opacities. The highly ubiquitous non-amyloidogenic R555W (GCDI/II)24–26 appears as rod shaped or granular bodies with sharp borders found in the central corneal stroma27,28. We aimed to examine the effects of these charge modifying mutations on the domains, their aggregation, their sensitivity to pH and temperature. Aggregating proteins are found to be more susceptible to acidic pH and increase in temperature29,30. Even in TGFβ Ip, it has been shown that the R124H mutation induces localization of TGFβ Ip to lysosomes with an acidic environment31. In the present study, we have investigated the effects of acidic pH, denaturants and temperature on the secondary structure and conformational stability of the domains. The cytotoxicities of the aggregates were also studied in primary human corneal stromal fibroblasts (pHCSF). Results Effects of biochemical and biophysical factors on WT and mutants. The native 4th_FAS1 domain (Fig.1a) of TGFβ Ip (Genbank_ID-NM_000358; Protein_ID:-NP_000349) and the mutants R555W and H572R were cloned and purified (Fig.1b) as described previously21. The estimated pI values of the WT, R555W and H572R domains were 6.53, 6.32 and 6.65 respectively. The amino acid substitutions (R→ W and H→ R) were associated with changes in charge and hydrophobicity as listed in the table (Table1). Effect of pH on the secondary structures of the WT and mutants. The CD spectra for the WT and the mutants at pH 7 (Fig.1c–e) showed negative minima in the n – π * region (222 nm) and a weak shoulder at the π –π * region (207 nm) corresponding to their mixed α -helical and β -sheet conformations. Under acidic conditions, there were discernible differences between the WT and mutants in their secondary structure. The non-amyloidogenic R555W was more sensitive to pH compared to the WT and amyloidogenic H572R. The WT (Fig.1c) and H572R (Fig.1e) remained unchanged under neutral and acidic pH. The R555W mutant displayed an increase in the CD intensity at 222 nm and 207 nm with decrease in pH (Fig.1d). CD intensities at 222 nm at varying pH values (Fig.1f) showed that R555W, was more sensitive to pH and the pH-response of H572R was similar to the WT protein. We also incubated the mutants in acidic pH for 1 week and followed their aggregation/ oligomerization by ThT fluorescence (Supplementary Fig. S1). However, no significant conversion was observed as seen from the corresponding CD spectra for WT and H572R. At pH 2.75, R555W showed a partial conversion to β -sheet. Compared to the amyloid fibril peptide pN622K, almost no increase in fluorescence was observed for WT and R555W. The slight increase in fluorescence corresponding to pH 2.75 and pH 3.25 for H572R did not show a corresponding conversion in the CD spectra. Effect pH on thermal denaturation of WT and mutants. Conformational transition of the domains undergoing thermal denaturation was examined by heating them from 20 °C to 70 °C, at neutral and acidic pH conditions. At pH 7 and pH 8, for both the mutants and the WT, no well-defined transition was observed with increasing temperatures (Fig.2a–f). For both the mutants, the amplitude of the negative minima at 222 nm decreased upon heating (Fig.2c–f) and a weak hysteresis was observed when cooled. With acidic pH, while the WT showed no apparent changes in the secondary structure with increasing temperature (Fig.3a–c), both the mutants displayed a clear transition from monomeric α /β -structure to β -sheet (Fig.3d–i). Single wavelength scan at 222 nm indicated a clear sigmoidal transition for both mutants under acidic conditions when compared to the WT, which was unperturbed by the changes in pH and temperature (Fig.4a–e). Thermal denaturation experiments were also done by heating the domains from 20 °C to 90 °C (Supplementary Fig. S2). Though previous studies have observed denaturation of WT above 60 °C 32, we did not observe any conformational transitions to β -sheet even after heating to 90 °C. The mid-point of the normalized sigmoidal curve defines the transition temperature (Tt) wherein the conversion of a monomeric α /β -structure to the β -structured oligomers was observed. The transition for the domains at different pH conditions was similar to the samples heated to 70 °C and the Tt lied between 35–58 °C. Hence all the subsequent thermal denaturation experiments were performed by heating the domains upto 70 °C. The non-amyloidogenic R555W showed a marked sensitivity to pH and displayed a higher thermal instability compared to the H572R. Though H572R mutant displayed a clear pH-dependent conversion to β -structure when heated, the Tt was higher than R555W. A difference in Tt of 5–12 °C is observed at various pH conditions. Mild precipitation was observed in the samples after heating, which correlates with the change in www.nature.com/scientificreports/ 3 Scientific RepoRts | 6:23836 | DOI: 10.1038/srep23836 intensities observed in the CD spectra. When the concentration of the sample was increased (from 0.6 mg/ml to 1.2 mg/ml~75 μM), the turbidity and precipitation increases. Urea denaturation of non-amyloidogenic R555W. The tryptophan residue in R555W allowed us to measure the emission fluorescence. Examination of the emission fluorescence at ~332 nm of R555W in acidic pH showed a significant decrease in emission intensity with decreasing pH, however the emission maxima remained unchanged (Fig.5a). To obtain a better insight into the effect of pH on R555W, we monitored the urea-induced unfolding of R555W. Increasing the urea concentration progressively shifts the emission maxima (~332 nm) to longer wavelengths (~352 nm), suggesting a clear transition from folded to unfolded conformations (5b–e). To confirm refolding of unfolded R555W, the unfolded protein at different pH conditions (pH 3.0, pH 4.5, pH 5.5 and pH 7.0) was diluted appropriately and emission spectra were recorded. The emission maxima (λ max) plotted with the unfolded and refolded domains were superimposable (Supplementary Fig. S3). (Fig.5b–e). A clear reversal in fluorescence maxima from ~352 nm to ~332 nm was observed thereby allowing us to estimate the thermodynamic stability of the mutant protein in various pH. Figure5f–i shows urea denaturation curves plotted as ‘fraction unfolded (yU) vs increasing urea concentrations’ as monitored by the changes in emission maxima Figure 1. Biochemical and biophysical properties of the native 4th_FAS1 domains of the wild-type and mutant TGFβIp. (a) Schematic representation of the domain arrangement and boundaries of the full-length TGFβ Ip and 4th_FAS1 domains of the wild-type, non-amyloidogenic (R555W) and amyloidogenic (H572R) mutants used in the study. (b) SDS-PAGE gel showing the purified fractions of the 4th_FAS1 domains of the WT, R555W and H572R mutants. (c–e) Far UV CD spectra of the 4th_FAS1 domains of WT (d), R555W (e) and H572R (f) incubated for 16 hours at acidic pH conditions (pH 3, 4.5, 5.5 and 7). The R555W mutant displayed clear changes in the CD spectra at 222 nm and 207 nm with decrease in pH (f). The CD intensity at 222 nm decreased with decrease in pH confirming the unfolding of the secondary structures. The CD spectra for the WT (c) and H572R (e) mutant remained almost unchanged. Plotting the intensities at 222 nm at varying pH (g) showed that the non-amyloidogenic phenotype, R555W, was more sensitive to pH and the amyloidogenic phenotype, H572R, remained more stable to pH changes at room temperature. www.nature.com/scientificreports/ 4 Scientific RepoRts | 6:23836 | DOI: 10.1038/srep23836 (Δλ max) at various pH values (Supplementary Fig. S3) for the R555W mutant. The thermodynamic parameters derived from urea denaturation are shown in the table (Table2). A significant decrease in free energy of unfolding from 13.8 ± 0.8 kJ/mole to 7.108 ± 1.5 kJ/mole was observed from pH 7.0 to pH 4.5 for the R555W mutant. However, for the H572R mutant, fluorescence studies could not be performed because of the absence of a tryptophan residue. Characterization of the β-oligomers of amyloidogenic and non-amyloidogenic mutants. The amyloidogenic and non-amyloidogenic mutants displayed a clear transition to an all β -sheet conformation when heated under acidic conditions. Conversion to an all β -sheet conformation could indicate β -oligomer formation17. A detailed investigation of the β -oligomers from two mutants was therefore performed to validate and characterize the proposed β -oligomers. Confirmation of β-oligomers formed by the mutants and characterization of the β-oligomers by TEM and DLS. TEM examination of the β -oligomers revealed that while no particles were visible for the WT (data not shown), both the mutants displayed particles validating our proposal. The β -oligomers of the H572R and R555W displayed varying sizes and morphologies (Fig.6a,b). The β -oligomers of the non-amyloidogenic R555W were homogeneous, displayed smoother edges and measured ~4–8 nm (mean diameter ~5.1 ± 1.79 nm) (Fig.6a). The β -oligomers of the amyloidogenic H572R were more heterogeneous, displayed rugged edges and were larger, measuring 10–40 nm (mean diameter ~19.1 ± 4.9 nm) (Fig.6b). The β -oligomers formed from the amyloidogenic and the non-amyloidogenic phenotypes are distinctly different from each other. Dynamic light scattering (DLS) allows the examination of apparent hydrodynamic radius (RH) of a protein in solution33,34. DLS analysis on the β -oligomers prepared under acidic conditions show that H572R β -oligomers exhibit large variations in their RH (~89.95 nm|pH 3.0, ~69 nm|pH 4.5 and ~155 nm|pH 5.5), while R555W β -oligomers are more uniform and almost similar across various acidic pH conditions (~39.58 nm|pH 3.0, ~51.9 nm|pH 4.5 and ~68.2 nm|pH 5.5). The distribution curves from the % intensity plots also show the homogeneity of the non-amyloidogenic R555W, and relative heterogeneity of the amyloidogenic H572R. This is in conjunction with the results obtained from TEM, where we see more homogenous and smaller β -oligomers from the non-amyloidogenic R555W mutant and heterogeneous and relatively larger β -oligomers from the amyloidogenic H572R mutant. The β-oligomers of the amyloidogenic H572R shows stronger binding to Thioflavin T (ThT). Amyloid fibrils bind to the dye ThT and display an emission fluorescence at 485 nm35,36. In Alzheimer’s disease, ThT binds to Aβ -oligomers themselves, and has been proposed for early diagnosis37. We wanted to test if the TGFβ Ip β -oligomers bind to ThT. The amyloid forming TGFβ Ip peptide pN622K (pN622K611–633) displayed a high fluorescence intensity when bound to ThT (Fig.6c). The fluorescence displayed by WT was comparable to the background fluorescence. Relatively, the β -oligomers of H572R showed significant fluorescence on binding to ThT (**P < 0.01) and atleast ~3 times more fluorescence compared to R555W. While the fluorescence intensities were much lower compared to the fibril forming peptide, it was significant that the β -oligomers of the amyloidogenic H572R were able to bind to ThT and this could aid further characterization of the β -oligomers. The mutants display differences in their ‘aggregation hotspots’. To determine the regions with high aggregation propensities or ‘aggregation hotspots’ within the mutants and the β -oligomers, the domains were digested with trypsin and the resulting peptides were examined using LC-MS/MS. The peptide map generated (Supplementary data) following the insilico trypsin digestion displayed a series of peptides (Fig.6d) formed following trypsin digestion. We aimed to identify the regions that were probably buried within the β -oligomers and hence resisted trypsin digestion. For WT, three short peptides (549ALPPR553, 591SLQGDK596 603NNVVSVNK610) were not detectable after tryptic digestions. For R555W, two short peptides (558LLGDAK563, 591SLQGDK596 were absent. However, for the H572R, peptides in the region E611-L632 were not observed in addition to a short peptide (591SLQGDK596) that was absent in the WT and R555W. Mapping the generated peptides to the 4th_FAS1 domain displayed interesting results (Fig.6e). The entire C-terminal region encompassing the residues 603NNVVSVNK610 and 611EPVAEPDIMATNGVVHVITNVL632 peptides was absent in the H572R native protein and β -oligomers. This region was intact in WT and R555W. A long stretch of residues between E534 and K563 containing the peptides 534EGVYTVFAPTNEAFR548, 549ALPPR553, 554EWSR557, 558LLGDAK563 was not observed in R555W β -oligomers. The peptide 591SLQGDK596 was absent in all the proteins. The segment WT R555W H572R Charge of the mutation –R ( + 1) → W(0) H (0) → R(+1) (at pH 7.0) H (+1) → R(+1) (at pH < 6.0) Net Charge (at pH 7.0) −0.6 − 1.6 0.2 Net Charge (at pH 5.5) 1.9 0.9 2.0 Change in hydrophobicity due to the mutation (17) – 3.95 → − 2.13 (at pH 7.0 and pH < 6.0) 0.64 → 3.95 (at pH 7.0) 2.87 → 3.95 (pH < 6.0) Table 1. Effects of mutation on the charge and hydrophobicity. The change in the individual charges and overall net charge in the mutants. The change in hydrophobicities were calculated using the equation described previously (17). www.nature.com/scientificreports/ 5 Scientific RepoRts | 6:23836 | DOI: 10.1038/srep23836 549ALPPR553 was absent in H572R β -oligomers. The β -oligomers of R555W, displayed a region extending between E534 and A562 that could possibly be buried. It is interesting to note that the mutation R555W resides within this region. The β -oligomers of H572R did not reveal such a region within the domain. It is likely that different regions of the 4th_FAS1 domain may be involved in the formation and stabilization of these β -oligomers. The β-oligomers remained stable at physiological conditions. For studying the thermal stability of the β -oligomers of the R555W (Fig.7a,b) and H572R (Fig.7c,d) the mutant domains at pH 5.5 were heated to 70 °C and cooled back to 20 °C. The β -oligomers remained in their β -sheet conformation at 20 °C showing no reversibility to the native α /β -conformation. To test their stabilities at physiological pH, the β -oligomers formed at pH 5.5 were reconstituted to pH 7 and examined (Fig.7e,f). The samples were also incubated for 4 weeks at pH 7 and the CD spectra were recorded. In both cases, the β -oligomers remained in their stable β -sheet Figure 2. Thermal denaturation of the 4th_FAS1 domains of the WT and mutants at neutral and basic pH. (a,b) Far UV CD spectra of the 4th_FAS1 domain of WT at pH 7.0 and pH 8.0 before heating (black), after heating to 70 °C (red) and cooling back to 20 °C (blue). (c,d) Far UV CD spectra of the 4th_FAS1 domains of R555W at pH 7 (c) and pH 8 (d) before heating (black) and after heating to 70 °C (red) and cooling back to 20 °C (blue). (e,f) Far UV CD spectra of the 4th_FAS1 domains of H572R a pH 7.0 (e) and pH 8.0 (f) before heating (black) and after heating to 70 °C (red) and cooling back to 20 °C (blue). The WT and the mutants did not display any significant changes in structure at pH 7.0 and pH 8.0. www.nature.com/scientificreports/ 6 Scientific RepoRts | 6:23836 | DOI: 10.1038/srep23836 conformation. This demonstrates that the β -oligomers were stable to thermal changes and at physiological conditions allowing for further examination. Cytotoxicity of β-oligomers. The toxicity of the β -oligomers on primary human corneal stromal fibroblast (pHCSF) cells was monitored using xCELLigence system. The WT domain did not interfere with the cell adhesion and proliferation and exhibited little or no toxic effect on the seeded cells (Fig.8a). R555W also displayed no cytotoxicity. H572R, however, displayed higher cytotoxicity (**P < 0.01) on the fibroblasts compared to R555W (Fig.8b). Interestingly, β -oligomers from both R555W and H572R were cytotoxic (**P < 0.01). While, the β -oligomers of R555W decreased the cell proliferation and the cytotoxic effect was visible after 12 hours, the β -oligomers of H572R displayed the maximum cytotoxic effect as no proliferation was observed and the xCELLigence showed minimum cell index. These results suggest that amyloidogenic mutant displayed significant cytotoxic effect both in the native as well as in the β -oligomeric forms whereas the non-amyloidogenic mutant was cytotoxic in the β -oligomeric form only. To obtain detailed information on the cytotoxicities, we prepared the β -oligomers of the two mutants at various acidic pH (pH 3.0, pH 4.5, pH 5.5). We also included the insoluble Figure 3. Thermal denaturation of the 4th_FAS1 domains of the WT and mutants at acidic pH. (a–c) Far UV CD spectra of the 4th_FAS1 domain of WT at pH 3.0 (a), pH 4.5 (b) and pH 5.5 (c) before heating (black) and after heating to 70 °C (red) and cooling back to 20 °C (blue). (d–f) Far UV CD spectra of the 4th_FAS1 domain of R555W at pH 3 (d), pH 4.5 (e) and pH 5.5 (f) before heating (black) and after heating to 70 °C (red) and cooling back to 20 °C (blue). (g–i) Far UV CD spectra of the 4th_FAS1 domain of H572R at pH 3.0 (g), pH 4.5 (h) and pH 5.5 (i) before heating (black) and after heating to 70 °C (red) and cooling back to 20 °C (blue). While the WT did not show any changes in structure, both the mutants displayed a very clear transition to β -sheet under acidic conditions. www.nature.com/scientificreports/ 7 Scientific RepoRts | 6:23836 | DOI: 10.1038/srep23836 aggregates obtained after centrifuging the mildly precipitated samples. When we examined this precipitates using CD, we found similar β -sheet curves as we found for the soluble oligomers (data not shown). Hence, we investigated these “insoluble aggregates” along with the soluble β -oligomers. The populations were reconstituted to pH 7 using buffer exchange. pHCSFs from 3 different donors (n = 3) were treated with the β -oligomers and examined (Fig.8c). The WT shows almost no cytotoxicity, similar to the control. The soluble β -oligomers derived from both the mutants at various acidic pH conditions displayed potent cytotoxic effect (**P < 0.01) compared to the insoluble aggregates and controls. The insoluble aggregates were also cytotoxic (*P < 0.05) but were relatively lesser compared to the soluble β -oligomers. The results were also validated using an MTT assay (Fig.8d). Similar to xCELLigence, we could see that soluble β -oligomers derived from both the mutants displayed potent cytotoxic effect (**P < 0.01) compared to the controls and insoluble aggregates. Discussion The most significant aspect of the TGFβ Ip associated corneal dystrophies is that the single amino acid substitutions in the mutants are responsible for clinically distinct phenotypes. Here, we examined the properties of Figure 4. Difference in thermal denaturation induced transition between non-amyloidogenic and amyloidogenic mutants at acidic pH. (a–e) Variable temperature CD curves at 222 nm of WT (black), R555W (red) and H572R (blue) proteins heated from 20 °C to 70 °C at various pH (3.0 [a], 4.5 [b], 5.5 [c], 7.0 [d] and 8.0 [e]) and the CD intensities at 222 nm were plotted as a function of temperature. The baseline subtracted curves of the WT (black), R555W (red) and H572R (blue) proteins show that while there was no transition observed in the WT in all the conditions as observed from the unchanged straight line in black, little or no changes were seen in pH 7 and pH 8 for the mutants. However, clear transitions to β -sheet were observed at acidic pH (pH 3, pH 4.5 and pH 5.5) for both the mutants. In all cases, we observe transition (Tt) is higher for R555W compared to H572R. A clear shift in their thermal denaturation curves between the mutants at acidic pH (pH 3.0, 4.5 and 5.5) is observed. A difference in Tt of 5–12 °C is observed at various pH conditions. www.nature.com/scientificreports/ 8 Scientific RepoRts | 6:23836 | DOI: 10.1038/srep23836 two mutants, the amyloidogenic H572R22,23, and non-amyloidogenic R555W24–26. Previous attempts to study the differences between the WT and the mutants at physiological conditions have revealed minimal information on the mechanism of their aggregation38. To elucidate the inherent variations between these mutants, we examined the effects of various biophysical and biochemical factors like temperature and pH on them. In amyloid forming proteins such as β -microglobulin39, thermal unfolding leads to formation of β -oligomers. Results from our thermal denaturation studies show clear differences between both the phenotypes and the WT protein at acidic pH. Our results suggest that changes in the physiological pH can influence the biochemical and biophysical characteristics of TGFβ Ip mutants. A clear conversion from α /β -structure to β -sheet was observed when the mutants were heated under acidic conditions (< pH 5.5). TEM and DLS studies confirmed the formation of β -oligomers by both phenotypes that remain irreversible and stable at physiological pH and temperature. For the first time we have shown that both the amyloidogenic and non-amyloidogenic phenotypes displayed marked differences Figure 5. pH sensitivity and stability of the non-amyloidogenic (R555W) phenotype. (a) Fluorescence emission spectra of R555W with decrease in pH. There was a clear decrease in emission maximum at 332 nm with decrease in pH (indicated by the black arrow). (b–e) Fluorescence emission spectra of R555W showing the reversibility to folded state after removal of urea. The R555W mutant was incubated with increasing concentrations of urea from 0.25 M to 8 M at various acidic conditions (pH 3, pH 4.5, pH 5.5) and pH 7 and the emission fluorescence before and after urea incubation was measured. The emission spectra before urea incubation 332 nm (black), after incubating with 8 M urea (red) and after removing urea by buffer exchange (blue). Unfolding of the protein is seen by the shifting of peaks (black arrow) from 332 nm to ~ 352 nm. The refolding of the protein after removal of urea is seen by the return of the emission maximum to ~332 nm (green arrow). (f–i) Investigation of the stability the non-amyloidogenic phenotype using urea denaturation studies. The R555W mutant was incubated with increasing concentrations of Urea from 0.25 M to 8 M at various acidic pH (pH 3, pH 4.5, pH 5.5) and pH 7, and the emission fluorescence was measured. The denaturation plots of ‘fraction unfolded vs urea concentration’ were plotted and fit into a two state model, with the parameters calculated as described in the methods section. pH Δ G H 2O Z, kJ/mole Cm, M m, kJ/mole/M ΔΔGH2O , kJ/mole 3.0 8.417 ± 1.03 4.2 ± 0.15 3.14 ± 0.32 − 5.383 4.5 7.108 ± 1.5 4.74 ± 0.91 1.85 ± 0.25 − 6.692 5.5 11.28 ± 2.02 3.93 ± 0.21 2.78 ± 0.48 − 2.52 7.0 13.8 ± 0.8 3.09 ± 0.1 3.8 ± 0.3 – Table 2. Urea denaturation studies of the non-amyloidogenic phenotype. The R555W 4th FAS1 domain was incubated with increasing concentrations of urea from 0.25 M-8 M at various acidic pH conditions and the emission fluorescence was measured. The free energies at various pH conditions were calculated and tabulated. www.nature.com/scientificreports/ 9 Scientific RepoRts | 6:23836 | DOI: 10.1038/srep23836 Figure 6. Characterization amyloidogenic and non-amyloidogenic β-oligomers. (a–b) Transmission Electron Microscopy. TEM images of the β -oligomers of the 4th_FAS1 domains of R555W (a) and H572R (b) mutants were acquired with a JEOL JEM-1010 transmission electron microscope using Digital Micrograph™ 1.81.78 for GMS 1.8.0. The β -oligomers of the amyloidogenic phenotype were larger measuring between 10–40 nm, mean size ~19.1 nm ± 4.9 nm (a) compared to the non-amyloidogenic β -oligomers that measured 4–8 nm, with a mean size ~5.1 nm ± 1.79 nm (b). Inset figures – particle size distribution of the β -oligomers. www.nature.com/scientificreports/ 16 Scientific RepoRts | 6:23836 | DOI: 10.1038/srep23836 13. Lakshminarayanan, R. et al. Structure-Function Relationship of TGFβ Ip (BIGH3) Associated Mutants in Corneal Dystrophy. Brit. J. Ophthalmol. 95(10), 1457–62 (2011). 14. Ołdak, M. et al. Late-onset lattice corneal dystrophy without typical lattice lines caused by a novel mutation in the TGFBI gene. Cornea 33(3), 294–9 (2014). 15. Klintworth, G. K., Valnickova, Z. & Enghild, J. J. Accumulation of beta ig-h3 gene product in corneas with granular dystrophy. Am J Pathol 152, 743–748 (1998). 16. Klintworth, G. K., Bao, W. & Afshari, N. A. Two mutations in the TGFBI (BIGH3) gene associated with lattice corneal dystrophy in an extensively studied family. Invest Ophthalmol Vis Sci. 45, 1382–8 (2004). 17. Chiti, F., Stefani, M., Taddei, N., Ramponi, G. & Dobson, C. M. Rationalization of the effects of mutations on peptide and protein aggregation rates. Nature 424(6950), 805–8 (2003). 18. Yuan, C., Berscheit, H. L. & Huang, A. J. Identification of an amyloidogenic region on keratoepithelin via synthetic peptides. FEBS Lett 581, 241–247 (2007). 19. Schmitt-Bernard, C. F. et al. BIGH3 (TGFBI) Arg124 mutations influence the amyloid conversion of related peptides in vitro. Eur J Biochem 269, 5149–5156 (2002). 20. Clout, N. J. & Hohenester, E. A model of FAS1 domain 4 of the corneal protein beta(ig)-h3 gives a clearer view on corneal dystrophies. Mol Vis 9, 440–448 (2003). 21. Elavazhagan, M. et al. Expression, purification and characterization of fourth FAS1 domain of TGFβ Ip-associated corneal dystrophic mutants. Protein Expr Purif. 84(1), 108–15 (2012). 22. Atchaneeyasakul, L. O. et al. Study Group. A novel H572R mutation in the transforming growth factor-beta-induced gene in a Thai family with lattice corneal dystrophy type I. Jpn J Ophthalmol. 50(5), 403–8 (2006). 23. Romero, P., Moraga, M. & Herrera, L. Different phenotypes of lattice corneal dystrophy type I in patients with 417C > T (R124C) and 1762A > G (H572R) mutations in TGFBI (BIGH3). Mol Vis. 16, 1601–9 (2010). 24. Munier, F. L. et al. BIGH3 mutation spectrum in corneal dystrophies. Invest Ophthalmol Vis Sci. 43, 949–954 (2002). 25. Fujiki, K., Nakayasu, K. & Kanai, A. Corneal dystrophies in Japan. J Hum Genet 46, 431–435 (2001). 26. Chakravarthi, S. V. V. K., Kannabiran, C., Sridhar, M. S. & Vemuganti, G. K. TGFBI gene mutations causing lattice and granular corneal dystrophies in Indian patients. Invest Ophthalmol Vis Sci 46, 121–125 (2005). 27. Dighiero, P. et al. Clinical, histologic, and ultrastructural features of the corneal dystrophy caused by the R124L mutation of the BIGH3 gene. Ophthalmology 107(7), 1353–7 (2000). 28. Kocak-Altintas, A. G., Kocak-Midillioglu, I., Akarsu, A. N. & Duman, S. BIGH3 gene analysis in the differential diagnosis of corneal dystrophies. Cornea 20(1), 64–8 (2001). 29. Wu, J. W. et al. Comparative analysis of human γ D-crystallin aggregation under physiological and low pH conditions. PLoS One. 12, 9(11) (2014). 30. Noormägi A., Valmsen K., Tõugu V. & Palumaa P. Insulin Fibrillization at Acidic and Physiological pH Values is Controlled by Different Molecular Mechanisms. Protein J. 2015 Oct 22 [Epub ahead of print]. 31. Kim, B. Y. et al. Corneal dystrophy-associated R124H mutation disrupts TGFBI interaction with Periostin and causes mislocalization to the lysosome. J Biol Chem. 284(29), 19580–91 (2009). 32. Andreasen, M. et al. Polymorphic fibrillation of the destabilized fourth fasciclin-1 domain mutant A546T of the Transforming growth factor-β -induced protein (TGFBIp) occurs through multiple pathways with different oligomeric intermediates. J Biol Chem. 287(41), 34730–42 (2012). 33. Ahmad B., Winkelmann J., Tiribilli B. & Chiti F. Searching for conditions to form stable protein oligomers with amyloid-like characteristics: The unexplored basic pH. Biochim Biophys Acta. 1804(1), 223–34 (2010). 34. Campioni, S. et al. Conformational properties of the aggregation precursor state of HypF-N. J Mol Biol. 379(3), 554–67 (2008). 35. Biancalana, M. & Koide, S. Molecular mechanism of thioflavin-T binding to amyloid fibrils. Biochim. Biophys. Acta. 1804, 1405–1412 (2010). 36. Groenning, M. Binding mode of thioflavin T and other molecular probes in the context of amyloid fibrils. Current status. J. Chem. Biol. 3, 1–18 (2010). 37. Maezawa, I. et al. Congo red and thioflavin-T analogs detect Abeta oligomers. J Neurochem. 104(2), 457–68 (2008). 38. Grothe, H. L. et al. Altered protein conformation and lower stability of the dystrophic transforming growth factor beta-induced protein mutants. Mol Vis. 19, 593–603 (2013). 39. Sasahara, K., Yagi, H., Naiki, H. & Goto, Y. Heat-induced Conversion of β 2-Microglobulin and Hen Egg-white Lysozyme into Amyloid Fibrils J. Mol. Biol. 372, 981–991 (2007). 40. Cardoso, I. et al. Transthyretin fibrillogenesis entails the assembly of monomers: a molecular model for in vitro assembled transthyretin amyloid-like fibrils. J Mol Biol 317, 683–95 (2002). 41. Smith, D. P., Jones, S., Serpell, L. C., Sunde, M. & Radford, S. E. A systematic investigation into the effect of protein destabilisation on beta 2-microglobulin amyloid formation. J Mol Biol 330, 943–54 (2003). 42. Aso, Y., Shiraki, K. & Takagi, M. Systematic analysis of aggregates from 38 kinds of non disease-related proteins: identifying the intrinsic propensity of polypeptides to form amyloid fibrils. Biosci Biotechnol Biochem 71, 1313–21 (2007). 43. Pfefferkorn, C. M., McGlinchey, R. P. & Lee, J. C. Effects of pH on aggregation kinetics of the repeat domain of a functional amyloid, Pmel17. Proc Natl Acad Sci USA 107(50), 21447–52 (2010). 44. Karring, H. et al. Differential expression and processing of transforming growth factor beta induced protein (TGFBIp) in the normal human cornea during postnatal development and aging. Exp Eye Res. 90(1), 57–62 (2010). 45. Brunk, U. T. & Ericsson, J. L. Cytochemical evidence for the leakage of acid phosphatase through utrastructurally intact lysosomal membranes. Histochem J 4, 479–491 (1972). 46. Schneider, L., Korber, A., Grabbe, S. & Dissemond, J. Influence of pH on wound-healing: a new perspective for wound therapy ? Archives of Dermatological Research 298, 413–420 (2007). 47. Huynh, K. K. & Grinstein, S. Regulation of vacuolar pH and its modulation by some microbial species. Microbiol Mol Biol Rev. 71, 452–462 (2007). 48. Runager, K. et al. Human phenotypically distinct TGFBI corneal dystrophies are linked to the stability of the fourth FAS1 domain of TGFBIp. J. Biol Chem. 286(7), 4951–8 (2011). 49. Underhaug, J. et al. Mutation in transforming growth factor beta induced protein associated with granular corneal dystrophy type 1 reduces the proteolytic susceptibility through local structural stabilization. Biochimica et Biophysica Acta 1834, 2812–2822 (2013). 50. Klein, W. L., Krafft, G. A. & Finch, C. Targeting small Aβ oligomers: the solution to an Alzheimer’s disease conundrum? Trends Neurosci. 24, 219–224 (2001). 51. Ferreira, S. T., Vieira, M. N. & De Felice, F. G. Soluble protein oligomers as emerging toxins in Alzheimer’s and other amyloid diseases. IUBMB Life 59, 332–345 (2007). 52. Volles, M. J. & Lansbury, P. T. Jr. Vesicle permeabilization by protofibrillar alpha-synuclein is sensitive to Parkinson’s disease-linked mutations and occurs by a pore-like mechanism. Biochemistry 41(14), 4595–602 (2002). 53. Sokolov, Y. et al. Soluble amyloid oligomers increase bilayer conductance by altering dielectric structure. J Gen Physiol. 128(6), 637–47 (2006). 54. Valincius, G. et al. Soluble amyloid beta-oligomers affect dielectric membrane properties by bilayer insertion and domain formation: implications for cell toxicity. Biophys J. 95(10), 4845–61 (2008). www.nature.com/scientificreports/ 17 Scientific RepoRts | 6:23836 | DOI: 10.1038/srep23836 55. Srinivasan, S. et al. Pathogenic serum amyloid A 1.1 shows a long oligomer-rich fibrillation lag phase contrary to the highly amyloidogenic non-pathogenic SAA2.2. J Biol Chem. 288(4), 2744–55 (2013). 56. Pace, C. N. & Sholtz, J. M. Protein Structure: A Practical Approach (Creighton T. E. ed) pp. 299–321, IRL Press, Oxford, UK. 57. Benjwal, S., Verma, S., Rohm, K. H. & Gursky, O. Monitoring protein aggregation during thermal unfolding in circular dichroism experiments. Protein science: a publication of the Protein Society 15, 635–9 (2006). 58. Bullock, A. N. et al. Thermodynamic stability of wild-type and mutant p53 core domain. Proc Natl Acad Sci USA 94, 14338–42 (1997). 59. Pace, C. N. & Shaw, K. L. Linear extrapolation method of analyzing solvent denaturation curves. Proteins Suppl 4, 1–7 (2000). 60. Urcan, E. et al. Real-time xCELLigence impedance analysis of the cytotoxicity of dental composite components on human gingival fibroblasts. Dental Materials 26, 51–58 (2010). Acknowledgements The authors would like to thank Dr. Gary Yam for his helpful insights. The project was funded by NMRC Singapore (CS-IRG grant R978/87/2012). Author Contributions E.M., L.R. and J.S.M. conceived the experiments and wrote the paper. E.M. cloned the genes, expressed the proteins, performed the CD fluorescence, TEM experiments and prepared the oligomers for MS experiments and cytotoxicity studies. A.V., V.M. and N.M. assisted in the expression and purification of the proteins. Z.L. performed the MS experiments. N.M. and E.G. assisted with the CD experiments on thermal denaturation. R.R.Y.L., G.S.L.P. and S.S.C. performed the cytotoxicity assays. R.W.B. and S.S.C. provided inputs to the conception of the work and editing the article. Additional Information Supplementary information accompanies this paper at http://www.nature.com/srep Competing financial interests: The authors declare no competing financial interests. How to cite this article: Murugan, E. et al. pH Induced Conformational Transitions in the Transforming Growth Factor β-Induced Protein (TGFβIp) Associated Corneal Dystrophy Mutants. Sci. Rep. 6, 23836; doi: 10.1038/srep23836 (2016). This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/