Classical Galactosemia - ACS Publications - American Chemical Society

Subscriber access provided by Fudan University

Article

Classical galactosemia: Insight into molecular pathomechanisms by differential membrane proteomics of fibroblasts under galactose stress Simon Staubach, Stefan Müller, Murat Pekmez, and Franz-Georg Hanisch J. Proteome Res., Just Accepted Manuscript • DOI: 10.1021/acs.jproteome.6b00658 • Publication Date (Web): 11 Jan 2017 Downloaded from http://pubs.acs.org on January 13, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of Proteome Research is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

Classical galactosemia: Insight into molecular pathomechanisms by differential membrane proteomics of fibroblasts under galactose stress

Simon Staubach 1, Stefan Müller 2, Murat Pekmez1,3, and Franz-Georg Hanisch 1,2,*

1

Institute of Biochemistry II, Medical Faculty, University of Cologne, Joseph-Stelzmann-Str. 52,

50931 Köln, Germany; 2 Center for Molecular Medicine Cologne, University of Cologne, RobertKoch-Str. 21, 50931 Köln, Germany; 3 Department of Molecular Biology and Genetics, Faculty of Science, Istanbul University, Istanbul 34134, Turkey

Running title:

Molecular pathomechanisms in classical galactosemia

Key words:

Classical galactosemia, GALT deficiency, lipid rafts, membrane glycoproteins, differential proteomics, iTRAQ

* Address correspondence to: Prof. Dr. Franz-Georg Hanisch, Institute of Biochemistry II, Medical Faculty, University of Cologne, Joseph-Stelzmann-Str. 52, 50931 Köln, Germany; Tel: +49 221 478 4493; Fax: +49 221 478 7788; e-mail: [email protected]

1

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Abbreviations CNS, central nervous system EGFR, epidermal growth factor receptor iTRAQ: isobaric tag for relative and absolute quantitation

2


Page 2 of 40

Page 3 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Abstract Classical galactosemia, a hereditary metabolic disease caused by the deficiency of galactose-1phosphate uridyltransferase (GALT; EC 2.7.712), results in an impaired galactose metabolism and serious long-term developmental affection of the CNS and ovaries, potentially related in part to endogenous galactose-induced protein dysglycosylation. In search for galactose-induced changes in membrane raft proteomes of GALT-deficient cells, we performed differential analyses of lipid rafts from patient-derived (Q) and sex- and age-matched control fibroblasts (H) in the presence or absence of the stressor. Label-based proteomics revealed of the total 454 (female) or 678 (male) proteins a proportion of about 12% in at least one of four relevant ratios as fold-changed. GALT(-) cell-specific effects in the absence of stressor revealed cell model-dependent affection of biological processes related to protein targeting to the plasma membrane (female) or to cellular migration (male). However, a series of common galactose-induced effects were observed, among them the strongly increased ER-stress marker GRP78 and calreticulin involved in N-glycoprotein quality control. The membrane-anchored N-glycoprotein receptor CD109 was concertedly decreased under galactose-stress together with cadherin-13, GLIPR1, glypican-1, and semaphorin-7A. A series of proteins showed opposite fold-changes in the two cell models, whereas others fluctuated in only one of the two models.

INTRODUCTION

3



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Galactose-1-phosphate uridyltransferase (GALT; EC 2.7.712), converts galactose-1-phosphate to UDP-galactose, the cosubstrate involved in enzymatic galactosylation of glycoproteins and glycolipids. Newborn children suffering from a GALT deficiency (classical galactosemia, refer to #5056 in Disease Database) run the risk of developing a potentially lethal acute hepatotoxic syndrome, which is induced by ingested galactose, and can resolve rapidly under a galactoserestricted diet. However, even under controlled ingestion of exogenous galactose in reduced amounts, the long-term prospects of classical galactosemia patients are not favourable, as they endogenously synthesize galactose at a certain level that exerts chronic cell toxic effects. Among long-term complications, which may arise from these chronic effects, a cognitive impairment, speech defects, motor function disturbances and a hypergonadotropic hypogonadism in most female patients point to the CNS and the ovary as the most affected organs1. There are cell-toxic effects directly induced by accumulated galactose-1-phosphate or galactitol, and others related to induced enzymatic inbalances in the glycosylation machinery that should lead to altered protein or lipid glycosylation with severe impact on cellular functions. Dysglycosylations with respect to an under-galactosylation of N-linked glycans were reported to occur on transferrin or glycopeptide hormones2. On the other hand, a later study revealed high individual fluctuations of aberrant N- and O-glycan patterns of plasma glycoproteins, but did not find any correlation of this aberrant glycosylation with the long-term outcome in patients3. We previously demonstrated that membrane-bound N-glycoproteins on urinary exosomes from galactosemic patients exhibit shifts from preponderant expression of high-mannose-type to complex-type glycans4, an observation which might be related in part to a subclinical kidney insufficiency5. Membrane receptors often belong to the large group of N-linked glycoproteins, like the human epidermal growth factor receptor (EGFR), which is known to show decreased membrane expression in GALT-deficient fibroblasts grown under galactose stress6. For a considerable number of membrane glycoproteins it has meanwhile been proven that glycans located in the stem region of the transmembrane proteins fulfill functions in the sorting for their apical targeting, and that proper N- and O-glycosylation of these glycoproteins is a precondition for their correct trafficking, in order to reach their definite destination7. Irrespective of the controversal

4


Page 4 of 40

Page 5 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


discussions on the raft concept there is mostly a consent regarding the usefullness of preparing membranous subdomains as detergent-resistant membranes8. These subdomains may not only form sorting platforms for the organization of cellular trafficking, but are discussed primarily as platforms, which organize receptor-mediated cell signalling. Based on these considerations we hypothesize that GALT-deficient cells grown under galactose stress should exhibit alterations in their membrane raft proteomes, which could be causative for perturbations in signalling cascades and could give hints to an understanding of the molecular pathomechanisms of the disease.

EXPERIMENTAL SECTION

Cells and Cell Culture - GALT-deficient and control fibroblasts were obtained from the NIGMS Human Genetic Cell Repository. Q fibroblasts (GM00528, female, 2 years at sampling; GM01704, male, 2 years at sampling) were homozygously affected by the mutation Gln188Arg (Q188R) in the galt gene. H fibroblasts (GM00969, female, 2 years at sampling; GM05659, male,

1 year at

sampling) are described as apparently healthy skin fibroblasts. Healthy and diseased fibroblasts were cultivated parallel in eight 300 cm2 flasks (4 flasks Q and 4 flasks H fibroblasts) to confluency using DMEM with stable glutamine/PAA, 10%FCS, PenStrep. Afterwards cells were grown for at least 3 days in reduced glucose medium (low glucose 1g/l) to minimize stress reactions associated with changing the medium supplementation to either 0.1% glucose or 0.1% galactose/0.1% glucose. The cells were finally grown in the presence or absence of stressor (galactose) for 48h (10% monosaccharide-depleted/dialysed fetal calf serum).

Lipid Raft Isolation - To prepare lipid rafts, confluent fibroblasts of eight culture flasks (300 cm2) were harvested using trypsin in PBS (two flasks of each sample: Q or H fibroblasts treated with either galactose or glucose). The cell samples from two flasks were combined in warm PBS (37°C) containing trypsin inhibitor (Applichem) followed by centrifugation (5 min/1000 rpm). The pellet was washed twofold

5



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

with cold PBS (4°C). The cell pellet was adjusted with cold PBS to a volume of 1.6 ml. A volume of 200 µl protease inhibitor (one Roche Protease Inhibitor tablet solved in 1 ml PBS) was added and the cells were dispersed by soft vortexing. Another volume of 200 µl Triton X-100 (10%) was added with vortexing to obtain a final volume of 2 ml (1% TritonX-100). The lipid raft extraction was performed by rotation for 1h at 4°C. Afterwards 1.8 g sucrose was added to the homogenate and filled up to a volume of 4 ml (final sucrose concentration 45%). The sample was rotated for another 30 min at 4°C, before it was poured into ultraclear centrifuge tubes (Beckman Instruments, Fullerton, CA). Finally, 3 mL of 35% sucrose and 1.5 mL of 5% sucrose were layered stepwise on top. Buoyant density centrifugation was performed at 198.000 xg (SW41 rotor/Beckman) for 2.5 h to float the lipid rafts. A volume of 3 mL corresponding to the light scattering band that contained the rafts was collected and centrifuged at 114.000 xg for 1.5 h to pellet the lipid raft fraction (Ti-60 rotor/Beckman). The pellet was transferred with cold PBS in a 1.5 ml Beckmann tube followed by 40.000 rpm (TLA55 rotor/Beckmann) centrifugation. The pellet was stored at -20°C.

Filter-aided Sample Preparation prior to iTRAQ Labeling – To extract proteins 100 µl 2% SDS in water were added to the lipid raft pellets. Subsequently SDS and lipids were removed by precipitating proteins with chloroform/methanol according to published protocols9. Protein pellets were redissolved in 20 µl 2 % SDS, 0.1 M dithiothreitol, 0.1 M triethylammonium bicarbonate (TEABC) pH 8.0. After boiling at 96 °C for 10 min samples were processed and digested with trypsin as described5. 0.1 M TrisHCl was replaced by 0.1 M TEABC pH 8.0 throughout. After centrifugation 60 µl of 10 % acetonitrile in water were added to the filter to take up residual sample and the second were combined with the first filtrates. Peptides in the filtrates were desalted by application to C18 Pepclean spin columns (Thermo) activated with 80 % acetonitrile and equilibrated to 0.1 M TEABC. The absorbed sample was washed two times with 100 µl 0.1 M TEABC and peptides were eluted with 20 µl 80 % acetonitrile in 0.1 M TEABC.

ITRAQ Labeling – Two independent quadruplex iTRAQ experiments (A and B) were performed: A, samples from Q and H fibroblasts of female donors grown in the presence of either galactose or

6


Page 6 of 40

Page 7 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


glucose; and B, samples from Q or H fibroblasts of male donors grown in the presence of either galactose or glucose. Two alternative iTRAQ labeling pattern were used in Experiment A: Q-Gal (114), H-Gal (115), Q-Glc (116), H-Glc (117), and in Experiment B: H-Gal (114), H-Glc (115), QGal (116), Q-Glc (117). Varying yields of lipid rafts were normalized on the basis of protein concentrations prior to the iTRAQ labeling. To 10 µl of sample 15 µl of 1 M TEABC were added and volumes were adjusted to 30 µl with water before iTRAQ labeling according to previously published protocols10. In brief, samples were reacted for 2 h in the dark, and after quenching with water (50 µl) combined. After concentration of the sample by centrifugal evaporation to about 20 µl and addition of 200 µl water (ultrapure), a peptide cleanup was performed on C18 cartridges, where TEABC was replaced by 0.1% formic acid throughout.

UHPLC-ESI-MS/MS analysis – Samples were run on a Q-Exactive Plus Orbitrap (Thermo Scientific) equipped with the EASY-nLC 1000 UHPLC. Separation of peptides was performed on an Acclaim PepMap RSLC 150 mm C18 column with 50 µm internal diameter (2 µm bead diameter, 100A) with flow rates of 200 nl/min in a gradient of 0.1% formic acid in water (buffer A) and 0.1% formic acid in acetonitril (B) increasing the content of buffer B from 2 % to 30% in 240 min, and from 40 to 100% in 5 min. The sample load was 5 µl. The mass spectrometer was operated in the data dependent mode with automatic switching between full scan MS and MS/MS acquisition. Survey full scan MS spectra (m/z 300−1800) were acquired in the Orbitrap with a resolution of 70 000 (m/z 200) after accumulation of ions to a target value of 3 × 106 based on predictive AGC from the previous full scan. Dynamic exclusion was set to 20 s. The 12 most intense multiply charged ions (z ≥ 2) were sequentially isolated and fragmented in the octopole collision cell by higher-energy collisional dissociation HCD with a maximum injection time of 120 ms. A mass range from 50 to M+50 Da was covered for MS2 (resolution of 17 500). A 2.5 Da isolation width was chosen.

Database Searches - Raw MS files were processed with MaxQuant in the Perseus framework (MaxQuant version 1.5.2.8), a freely available software suite. Peak list files were searched by the

7



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Andromeda search engine, incorporated into the Max-Quant framework. The database searching parameters included: enzyme name: Trypsin (full), max. missed cleavage sites: 2, min. peptide length: 7. Variable modifications were: methionine oxidation, and peptide N-terminus acetylation; fixed modifications were: cysteine carbamidomethylation. Values of the target FDR for proteins and peptides was 0.01.

Quantitative evaluation and calculation of thresholds – The raw files from ESI-MS/MS were loaded into MaxQuant and processed using parameter settings described in Supporting Information. Processing in MaxQuant included a correction of reporter ion intensities based on isotopic impurity, whereas a filtration of other interfering precursor ions (PIF) was ommitted. Corrected reporter ion intensities were used to calculate relevant ratios in excel before data were imported and quantitatively evaluated in Perseus (version: 1.5.2.4). Further processing in Perseus comprized the removal of decoy hits (“reverse”), “potential contaminants”, and “only identified by site”. Ratios were calculated on the basis of the summed intensities of reporter ions. Ratios were transformed to log2 values and normalized by subtraction of the medians. Values above or below 2 x standard deviation (SD) were considered as criterion to define thresholds of significant changes in protein abundancies. In detail, the 2 x SD was determined for each of the four relevant ratios given as log2 values in Table S-1. In Experiment A these values were 0.64 (Q-Gal/Q-Glc), 1.16 (QGal/H-Gal), 0.98 (Q-Glc/H-Glc), 1.11 (Q-Gal/H-Glc), in Experiment B 0.62 (Q-Gal/Q-Glc), 0.99 (QGal/H-Gal), 0.89 (Q-Glc/H-Glc), 1.12 (Q-Gal/H-Glc). General features of the differential study were revealed by correlation analyses for the two independent experiments and for relevant ratios within each experiment, and presented in scatter plots. Moreover, cluster analyses were performed in Perseus (Principal Component Analysis) to identify variations in similarities of expression patterns, when comparing the two cell models and growth conditions.

8


Page 8 of 40

Page 9 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Western Blot Analysis - Samples were run on SDS-polyacrylamid gels (4% stacking gel, 5-20% running gel) in a Mini Protean 3 cell (BioRad, Munich, Germany). The gel was equilibrated in transfer buffer (20 mM glycine, 24 mM Trizma-base, 10% methanol) before it was wet-blotted (BioRad, Munich, Germany) onto a nitrocellulose membrane (Whatman, Dassel, Germany) at 100 mA overnight. Thereafter, the membrane was blocked in Tris buffered saline (TBS) containing 5% non-fat dried milk and 0.1% Tween20 for 1 hour at room temperature before incubation with the primary antibody (1 hour at room temperature). The following antibodies were used for specific staining of membrane proteins: anti-AHNAK (H-153, Santa Cruz, rabbit polyclonal, sc-98373); anticadherin-13 (E-9, Santa Cruz, mouse monoclonal, sc-166875); anti-CD59 (B-3, Santa Cruz mouse monoclonal, sc-133171); anti-CD109 (B-9, Santa Cruz, mouse monoclonal, sc-271039); antiflotillin-2 (B-6, Santa Cruz, mouse monoclonal, sc-28320); anti GRP78 (H-129, Santa Cruz, rabbit polyclonal, sc-13968); anti-GAPDH (Fl-335, Santa Cruz, rabbit polyclonal, sc-25778). Immunoblots were incubated with rabbit anti-mouse IgG (P260, DAKO, Hamburg, Germany), or pig anti rabbit HRP-conjugated secondary antibody (P399, DAKO) and the respective proteins were detected using a Lumilight Kit (Roche) and super RX film (Fujifilm Europe, Düsseldorf, Germany). Between each incubation step, the membrane was washed 3-times with TBST (20 mM Tris-HCl, 137 mM NaCl, pH 7.6, 0.1% Tween20).

Figure 1

RESULTS Experimental Design – The present study was based on two sex- and age-matched in vitro cell models of human GALT-deficient fibroblasts and healthy controls. The cells were grown either in the presence of 0.1% glucose to reveal cell-specific fluctuations of the raft proteomes or in the presence of 0.1% galactose as the specific stressor for GALT-deficient fibroblasts. Two major sets of experiments were performed with lipid rafts prepared from each of these cell samples, 1) a differential proteomics analysis in two independent experiments (A, B) applying the iTRAQ

9



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

technology combined with LC-ESI mass spectrometry on a Q Exactive Orbitrap (Fig. 1) to identify fluctuating raft proteins and potentially affected molecular pathways; 2) a semi-quantitative westernblot analysis for validation of differentially expressed raft proteins. Experiments A and B (see Supporting Information in Table S-1) were included in data evaluation summarized in Tab. 1A and Tab. 1B.

General Assessment of Protein Identification in Experiments A and B and GO Term Annotation – In total, 694 proteins were identified across two biological and technical replicates (464 proteins in Experiment A, and 678 proteins in Experiment B). The venn diagram given in Fig. 1C shows that 448 of these proteins were common to the two data sets. This corresponds to 96.5% (Experiment A) or 66.1% (Experiment B) of total proteins in the respective data set. A proportion of 12.1% (Tab. 1A: 56 proteins in at least one of four ratios in Experiment A) or 11.9% (Tab. 1B: 81 proteins in at least one of four ratios in Experiment B) were fold-changed species (Experiment A: 26 decreased, and 30 increased proteins; Experiment B: 48 decreased, and 33 increased proteins). For GO Term Enrichment Analyses of Experiments A and B refer to Tables S2 and S-3, respectively. According to GO Term Enrichment Analysis of 20 proteins with fold-changes in both experiments (Table S-4) 12 proteins are assigned to “extracellular region part”, 7 are located at the “cell surface”, and 4 represent “anchored components of membranes”, whereas 6 proteins have a putative localisation in “adherence junctions”, and 3 are found in the “proton-transporting twosector ATPase complex” (see also Fig. 2B). Referring to the GO terms of affected biological processes, 12 proteins are assigned to pathways involved in the “regulation of cell communication”, 4 proteins to the “negative regulation of cellular responses to growth factor stimuli”, 4 proteins to the “endoplasmic reticulum unfolded protein response”, 3 proteins to the “ERnucleus signaling pathway”, and 3 proteins to “ATP hydrolysis coupled proton transport” (Tab. S-4, Fig. 2A). Figure 2A-C

10


Page 10 of 40

Page 11 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


General Assessment of Protein Quantification in the Two Cell Models and Under Different Growth Conditions- Correlation analyses were performed for the two independent experiments (A/B) and for selected ratios within each experiment. The two experiments based on distinct cell models do not correlate well (refer to the scatter plots shown in Fig. S-1 and S-2), as is reflected in the results from GO Term Enrichment Analyses performed on fold-changed proteins in the absence of stressor (see below). When correlating data within each experiment, the scatter plots in both experiments show similar features with strongly increased variabilities induced by the stressor in Q cells. This holds in particular true when data corresponding to the Q-Gal/H-Glc ratio with expected high variability are correlated with those corresponding to the H-Gal/H-Glc ratio with expected minor variability (Fig. S-3A,B). To identify variations between the two cell models (Q/H) and the two growth conditions (Gal/Glc) via detection of similarities or dissimilarities among the expression patterns we performed a Principal Component Analysis (PCA) of the two experiments. PCA revealed that the results from Q/H experiments in A and B formed separate groups, whereas the Gal/Glc experiments in A and B clustered in a single group (refer to Figure S-4 in Supporting Information). The same clustering features are revealed in a heatmap for the two experiments and growth conditions (Fig. S-4).

Alterations in Human Fibroblast Raft Proteomes in GALT-deficient Q vs. Healthy H Cells in the Absence of the Stressor–

To address a differential protein expression associated with

enzymatic inbalances induced by the GALT-deficiency, we investigated GALT-deficient (Q) and healthy fibroblasts (H) grown in the absence of exogenously administered galactose (samples QGlc and H-Glc). In Experiment A 23 proteins and in Experiment B 33 proteins showed significantly changed ratios in GALT-deficient cells grown in the presence of glucose and should reflect accordingly cell model-specific fluctuations independent of the stressor.

Tab. 1

11



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

In the Q-Glc vs. H-Glc ratios listed for both experiments in Tab. 1 the patterns of significantly fold-changed proteins were inconsistent throughout. Mostly, proteins with altered expression in Experiment A did not fluctuate in Experiment B and vice versa. A small proportion of proteins showed fold-changes in both experiments, however in different directions (refer to ATP synthase F(0) complex subunit C1, cadherin 13, gremlin-1, and transmembrane glycoprotein NMB). In Experiment A highest enrichment of GO terms was found for “protein targeting to the plasma membrane” (refer to GO terms listed according to decreasing fold-enrichment in Table S2B), whereas in Experiment B GO term enrichment analyses point to an affection of biological processes related to “cell migration” and to “protein localization” (Table S-3B). These findings indicate GALT(-) cell model-specific differences in cellular perturbations associated with endogenous galactose levels in fibroblasts from a female (Experiment A) or a male donor (Experiment B). These “cell model-specific background” patterns need to be discriminated from protein fluctuations associated with exogenously administered galactose (see below). A sexdependent difference in cellular protein expression associated with GALT-deficiency cannot be postulated on the basis of only one female or male fibroblast model.

Raft Proteins Differentially Expressed in Q cells under Galactose Stress – The central question of this study refers to those alterations in membrane raft proteomes that are induced directly or indirectly in GALT-deficient fibroblasts by the stressor galactose. These effects were mimicked by cultivating Q and H cells in the presence of 0.1% galactose (in replacement of 0.1% glucose). ITRAQ analysis was performed in two independent quadruplex analyses combining the samples from Q-Glc, H-Glc, Q-Gal, and H-Gal (Table S-1). This paragraph refers to the comparative analysis of Q-Gal vs. H-Gal and Q-Gal vs. Q-Glc. In Experiment A 33 proteins were revealed as regulated under galactose stress, and 48 proteins in Experiment B (Tab. 1A,B). Also under galactose stress the patterns of fold-changed proteins were largely divergent in both experiments. As more proteins were identified in Experiment B a considerable number of proteins was found to fluctuate only in this experiment (see for example sialin, CD63, CD44, tenascin,

12


Page 12 of 40

Page 13 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


annexin A1 and A5). A concerted, specifically galactose-induced change in expression was revealed for the proteins CD109, CDH13, glypican-1, GRP78, calreticulin, glioma pathogenesis related protein 1, prelamin A/C, protein phosphatase 1 regulatory subunit, semaphorin-7A, and transcription factor IIIB 90 kDa subunit. Fold-changed expression in opposite directions was found for 9 proteins: ATP synthase F(0) complex subunit C1, dipeptidyl peptidase 4, gremlin-1, neprilysin, ragulator complex protein LAMTOR4, tetraspanin-6, transmembrane glycoprotein NMB, V-type proton ATPase subunit d1, V-type proton ATPase subunit S1. Cellular topologies of the 20 proteins with fold-changes in both experiments are listed together with information on their lipid raft association and N-/O-glycosylation in Tab. 2. The ER and plasma membrane localized GRP78 represents a known marker for cellular stress responses11, which was induced by the stressor galactose in both cell models of GALT-deficient cells (Tab. 1). The same holds true for CALR (calreticulin, Tab. 1) that fulfills functions in the context of quality control mechanisms in the ER by binding to mis-folded or dys-glycosylated Nglycoproteins12. Neither caveolins nor flotillins were found among fluctuating proteins in quantitative proteomics (Experiment A and B), however flotillin 2 was shown to decrease in expression by westernblot, see Fig. 3A). Like caveolin, the flotillins act in lipid raft formation as scaffolding proteins within membrane rafts. In this way, they are involved in a variety of cellular sorting and trafficking processes. Moreover they fulfill functions in the organisation of signaling platforms. CD109, a membrane raft-integrated glycoprotein receptor with functions in cell signaling was revealed as concertedly decreased in both experiments (Tab.1A,B). As a TGF-β co-receptor, CD109 binds to the receptor and promotes its caveolar internalization and proteasomal degradation13 with severe consequences on cell growth, differentiation, migration and extracellular matrix deposition. CD44 (see Experiment B) exerts negative regularory functions in programmed cell death and is involved in positive regulation of gene expression mediated by wnt-receptor signaling via the MAPK/ERK pathway. CD44 has been implicated as a stem cell marker in several malignancies of hematopoietic and epithelial origin14.

13



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The membrane-anchored endopeptidase NEP (neprilysin, enkephalinase, skin fibroblast elastase) shows inconsistent up- or down-regulation. The protein is listed under the GO term “Cellular Component” as localized in dendrites, axons and synaptic vesicles. The peptidase DPP4 (dipeptidyl peptidase 4) is involved in the pericellular proteolysis of ECM proteins, in the migration and invasion of endothelial cells during development and in wound healing processes15. Relevant GO terms refer to the lamellipodia (cellular components) or to locomotion and cell motility (biological process). GLIPR1 (glioma-pathogenesis related protein 1) is preferentially expressed in normal tissue of lung and kidney, but strongly increased in its expression in brain tumors of glial origin (uniprot). GPC1 (glypican-1), a cell surface proteoglycan, is particularly involved in Schwann cell myelinization and in the misfolding of normal prion proteins to the infectious prion form16. SEMA7A (semaphorin A7) plays roles in a variety of biological processes, in particular in integrinmediated signaling, in cell migration, and in the promotion of axon growth in the embryonic olfactory bulb17. CDH13 (cadherin-13), shows antagonistic expression changes, when comparing galactoseinduced effects in Q vs. H cells (supported by westernblot analysis in Fig. 3B). However, it shows either unchanged (female) or decreased expression (male), when Gal vs. Glc-induced effects are compared in the Q cells. Cadherin-13 belongs to a family of calcium-dependent cell adhesion proteins that mediate the formation of cell contacts by homophilic interaction. In this way the cadherins eventually contribute to the sorting of heterogeneous cell types and are claimed to negatively regulate neural cell growth18. GREM1 (gremlin-1) is a cytokine with morphogen activity that is particularly involved in metanephric kidney morphogenesis. It binds to bone morphogenic protein 4 and down-regulates BMP4-mediated signaling19. GPNMB (glycoprotein NMB) binds to heparin and may play a role in bone mineralization and osteoblast differentiation (uniprot).

14


Page 14 of 40

Page 15 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


LAMTOR4 (LTOR4) is a component of the Ragulator complex and activates the signaling complex mTORC1, which promotes cell growth in response to growth factors. As a guanine nucleotide exchange factor it activates the small GTPases Rag20. PPP1R12A, protein phosphatase 1 regulatory subunit 12A (MYPT1) represents an essential regulator of protein phosphatase 1C (PPP1C), is able to bind to myosin, and as part of the PPP1C complex it dephosphorylates the serine/threonine-protein kinase PLK121.

Fig. 3

Validation of iTRAQ-identified Candidate Proteins by Western Blotting - To validate candidate proteins revealed in the above reported differential proteomics experiments as regulated and to identify further putative candidates, a selected panel of antibodies was used in semiquantitative western blot analyses using GAPDH as a reference protein. Two different experiments are shown in Fig. 3A, 3B, one referring mostly to proteins fluctuating in Experiment B (extended by genuine raft proteins and other candidate proteins, like AHNAK) (Fig. 3A), another referring to both experiments, but with restriction to cadherin-13 (Fig. 3B).

The high-molecular mass protein

AHNAK (>600 kD) did not enter 5-15% polyacrylamide gels and was accordingly analysed in a dot blot setting. As shown in Fig. 3A of the four cellular lysates (Q-Glc, H-Glc, Q-Gal, H-Gal) only the GALT-deficient fibroblasts grown in the presence of galactose (Q-Gal) revealed strongly increased staining intensity. A similar pattern of staining was observed for the cellular stress marker GRP78 (Fig. 3A), which was not detectable in fibroblasts grown in the presence of glucose (Q-Glc, H-Glc) or in healthy control fibroblasts grown in the presence of galactose (H-Gal), but revealed strong staining in galactose-stressed GALT-deficient cells (Q-Gal). Reciprocal results were obtained for the membrane glycoprotein receptors CD109 (down-regulated according to Experiments A and B) and CD59 (no fold-change expression in both experiments), both showing significant downregulation in Q-Gal cells. As CAV1 antibodies tested negative for the membrane raft preparations under study, we included another scaffolding protein, flotillin-2, as a potential marker for galactose

15



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

effects exerted on raft formation. Also this protein showed decreased expression in GALT-deficient fibroblasts, however independent of supplementing the culture medium with glucose or galactose. Cadherin-13 fluctuated under galactose stress gender-dependently as its expression increased in the male and decreased in the female GALT(-) cell model (Fig. 3B).

DISCUSSION This study is the first differential proteomics-based investigation of classical galactosemia, which attempts to identify potential molecular pathomechanisms and marker protein candidates that are associated with a GALT-deficiency and with effects exerted by galactose-induced cell stress. The severe long-term toxic effects by endogenous galactose were studied in a cellular model (GALTdeficient fibroblasts) applying the iTRAQ technology for relative protein quantification and a validation of selected candidate proteins by western blotting. Primary fibroblasts from patients were chosen as alternative to immortalized lymphoblasts, both representing the only available cellular models of the disease.

Fig. 4

Molecular Pathways Affected in GALT-deficient Q cells under Galactose Stress – The proteins listed in Tab. 1 with fold-changed expression in GALT-deficient cells from both, female and male donors can be classified according to their GO terms into groups with common molecular pathway assignments (Fig. 4). Three proteins with cell model-specific, stressor-independent fold-changes belong to a group of ATPases and the GO term “ATP hydrolysis coupled proton transport”. Importantly in this respect is a cell model-specific effect of glucose depriviation under galactose stress, which results in the female fibroblast Q model in a concerted decrease of enzymes involved in electron transport and increased mitochondrial ATPases, whereas the male fibroblast Q cells react in the opposite way. Associated with these effects is an increase in LAMTOR4 expression in the female model and a

16


Page 16 of 40

Page 17 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


corresponding decrease in male fibroblasts, where besides LAMTOR4 also LAMTOR1 and LAMTOR2 are reduced in expression. These proteins belong to the pentameric ragulator complex, which recruits Rag GTPases and indirectly mTORC1 (mammalian target of rapamycin complex 1 ) to lysosomal membranes22. mTORC1 signaling couples nutrient abundance to cell growth and proliferation and functions as a nutrient/energy/redox sensor that controls protein synthesis. The inconsistent changes in expression of LAMTOR4 point to differences in nutrient consumption and energy economy in the two cell models of GALT deficiency, as was reflected in the GO Term Enrichment Analyses Another group with consistently increased proteins (ER unfolded protein response) covers the ER resident membrane proteins GRP78 and CALR, which are involved in the functionally linked processes “protein folding and protein quality control”, and can be regarded as stress markers that indicate activation of the Unfolded Protein Response (UPR) (Fig. 4). GRP78 is an established ER stress marker and concertedly increased in GALT-deficient cells under galactose stress. The other increased protein within this group, CALR, is involved in the quality control of N-glycoproteins and known to bind to dysglycosylated molecules. Increased expression of this protein could be a direct response to disturbances in the galactose metabolism and the affection of N-glycan biosynthesis. GALT-deficiency causes a shortage of UDP-Gal, the co-substrate of galactosyltransferases, and could hence influence N-glycosylation of proteins and their sorting and trafficking to the plasma membrane. In agreement with this, some raft-associated N-glycoproteins listed in table 1 show concerted decrease in expression under galactose stress. In compensation for a decrease in UDP-Gal cosubstrate concentrations an epimerization of UDP-Glc to UDP-Gal could affect the glucose catabolism. D-Glucose starvation is known to activate the Unfolded Protein Response (UPR) (Fig. 4). Sustained activation of the UPR has been implicated in the progression of several neurodegenerative diseases23. Cell motility- or migration-associated proteins cover DPP1, and NEP. These N-glycoproteins represent proteases involved in the proteolysis of extracellular matrix components and promote in this way cell migration. Interestingly, the expression of both proteins appears to be differently

17



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

regulated in the two cellular models under study, as in fibroblasts from female donors a diseaseassociated increase of the peptidases contrasts with a decrease in the corresponding male model.

The involvement in signaling cascades and hence the direct or indirect exertion of a multitude of regulatory functions is indicated by concertedly decreased raft N-glycoproteins CD109, GPC1 and CD44 (the latter only in the male model) (Fig. 4). Lipid raft microdomains of membranes organize signaling molecules into functional complexes, and the central organizing proteins are those, which provide a scaffolding domain (CSD). Caveolin-1 in caveolar rafts serves this function and interacts via its CSD with G-proteins, eNOS, adenylate cyclase isoforms, and a series of tyrosine kinases (Src family members, MAPK, PKA, PKC). Hence it plays a central role in a variety of signaling events. The GPI-anchored N-glycoprotein CD109 acts as a serine protease inhibitor, which is known to interact with TGFBR1 and to modulate negatively TGFB1 signaling with severe impact on the control of a variety of growth factors. GPC1 sequesters FGF2 in lipid rafts, prevents in this way its binding to FGFR and inhibits the FGF-mediated signaling. As both raft proteins CD109 and GPC1 negatively regulate growth factor-mediated signaling, their concerted down-regulation may indicate enhanced responsiveness of GALT-deficient cells under Gal stress to growth factors TGF and FGF. Expression of semaphorin-7A decreases under galactose stress in both cellular models, and may be related to a negative affection of integrin-β1-mediated signaling. This finding could have impact in the context of CNS development, as SEMA7A promotes axon outgrowth through integrins and MAPKs17. SEMA7A, together with CD109, GPC1, GLIPR1, CD44, and TSPAN6, belongs to a group of N-glycoproteins with plasma membrane localization and functions in the regulation of cell signaling (Fig. 5). The same holds true for cadherin-13. This N-glycoprotein, also known as H-cadherin or Tcadherin, was found differentially expressed in the two cell models of galactosemia, as in the male model galactose induced increased expression relative to the control cells (however a decrease relative to Q cells grown in the presence of glucose), wheras in the female Q cells the reverse effect was observed. As a matter of fact, CDH13 had not been described in human skin fibroblasts

18


Page 18 of 40

Page 19 of 40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


before. As the GALT(-) fibroblasts used in this study were derived from skin biopsies of galactosemia patients, it is noteworthy to mention that only keratinocytes were proven to express CDH1324. The protein generally is highly expressed in heart and in the cerebral cortex, medulla, hippocampus, amygdala, thalamus and substantia nigra. No expression was detected in cerebellum or spinal cord. An important aspect of this cadherin isoform refers to its growth inhibitory function25, and to its lower expression in the developing brain than in adult brain18. In neuroblastoma cells CDH13 exerts its growth inhibitory function by suppression of the mitogenic proliferative response to the epidermal growth factor18. AHNAK was not significantly fold-changed in Experiments A or B, however it fluctuated strongly in the immunoblot analysis. The functional significance of the protein in the context of this study refers to the GO term “nervous system development”, the potential requirement of AHNAK for neuronal cell differentiation, and to its interaction with dysferlin, which is a key calcium ion sensor involved in the Ca2+-triggered synaptic vesicle-plasma membrane fusion26. It may also be involved in endocytotic processes, since together with annexin-2 it mediates membrane fusion.

CONCLUSIONS A cellular model of GALT-deficiency underlying classical galactosemia revealed fluctuating protein expression under galactose stress. Importantly, CD109 and GPC1 were found as proteins with decreased membrane expression and point to accelerated TGFB1 and FGF signaling of galactosemic cells under galactose stress. Furtheron, a concerted activation of the Unfolded Protein Response became evident as a characteristic feature of the GALT(-) cell models that points to accumulating dys-glycosylation of N-glycoproteins with impact on targeted transport of membrane receptors (Fig. 5)7. Particularly, in the male cell model a concerted decrease of plasma membrane expression of N-glycoproteins became evident that play roles in the modulation of cell migration or in the regulation of cell signaling (Fig. 5B). The mechanism of decreased membrane expression, as depicted in Fig. 5A, is still hypothetical. It claims that at least for some of the

19



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

glycosylated membrane proteins a dys-glycosylation might be causative for decreased membrane expression, rather than a down-regulation of expression on the transcriptional level. A general criticism of the present study might refer to the used cellular model as it cannot be expected to reflect galactose-induced protein fluctuations in nervous tissue. On the other hand, a considerable number of proteins identified as regulated under galactose stress are GO-annotated as involved in functional networks of neuronal tissue. We interpret this finding as an indication that galactose-stress in GALT-deficient cells affects pathways common to different cell types, but do not draw any specific conclusions with respect to molecular pathomechanisms potentially involved in the disease-associated cognitive impairment, speech defects, or motor function disturbances observed in galactosemic patients.

AUTHOR INFORMATION Corresponding author: Prof. Dr. Franz-Georg Hanisch, Institute of Biochemistry II, Medical Faculty, University of Cologne, Joseph-Stelzmann-Str. 52, 50931 Köln, Germany; Tel: +49 221 478 4493; Fax: +49 221 478 7788; e-mail: [email protected]

ACKNOWLEDGEMENTS This study was supported by the Galactosemia Initiative Germany and by a grant from KölnFortune (both to FGH).

CONFLICT OF INTEREST STATEMENT Authors declared to have no conflict of interests.

SUPPORTING INFORMATION AVAILABLE Supplementary Tables Table S-1 Experiment A and B confident level (

Classical Galactosemia - ACS Publications - American Chemical Society

Recommend Documents