Glycoprotein-specific ubiquitin ligases recognize N-glycans in unfolded substrates

Misfolded or unassembled polypeptides in the endoplasmic reticulum (ER) are retro-translocated into the cytosol and degraded by the ubiquitin-proteasome system. We reported previously that the SCF(Fbs1,2) ubiquitin-ligase complexes that contribute to ubiquitination of glycoproteins are involved in the ER-associated degradation pathway. Here we investigated how the SCF(Fbs1,2) complexes interact with unfolded glycoproteins. The SCF(Fbs1) complex was associated with p97/VCP AAA ATPase and bound to integrin-beta1, one of the SCF(Fbs1) substrates, in the cytosol in a manner dependent on p97 ATPase activity. Both Fbs1 and Fbs2 proteins interacted with denatured glycoproteins, which were modified with not only high-mannose but also complex-type oligosaccharides, more efficiently than native proteins. Given that Fbs proteins interact with innermost chitobiose in N-glycans, we propose that Fbs proteins distinguish native from unfolded glycoproteins by sensing the exposed chitobiose structure.     

A neural-specific F-box protein Fbs1 functions as a chaperone suppressing glycoprotein aggregation

Fbs1 is an F-box protein present abundantly in the nervous system. Similar to the ubiquitously expressed Fbs2, Fbs1 recognizes N-glycans at the innermost position as a signal for unfolded glycoproteins, probably in the endoplasmic reticulum-associated degradation pathway. Here, we show that the in vivo majority of Fbs1 is present as Fbs1-Skp1 heterodimers or Fbs1 monomers but not SCF(Fbs1) complex. The inefficient SCF complex formation of Fbs1 and the restricted presence of SCF(Fbs1) bound on the endoplasmic reticulum membrane were due to the short linker sequence between the F-box domain and the sugar-binding domain. In vitro, Fbs1 prevented the aggregation of the glycoprotein through the N-terminal unique sequence of Fbs1. Our results suggest that Fbs1 assists clearance of aberrant glycoproteins in neuronal cells by suppressing aggregates formation, independent of ubiquitin ligase activity, and thus functions as a unique chaperone for those proteins.     

Fbs1 protects the malfolded glycoproteins from the attack of peptide:N-glycanase

Fbs1 is a cytosolic lectin putatively operating as a chaperone as well as a substrate-recognition subunit of the SCF(Fbs1) ubiquitin ligase complex. To provide structural and functional basis of preferential binding of Fbs1 to unfolded glycoproteins, we herein characterize the interaction of Fbs1 with a heptapeptide carrying Man3GlcNAc2 by nuclear magnetic resonance (NMR) spectroscopy and other biochemical methods. Inspection of the NMR data obtained by use of the isotopically labeled glycopeptide indicated that Fbs1 interacts with sugar-peptide junctions, which are shielded in native glycoprotein, in many cases, but become accessible to Fbs1 in unfolded glycoproteins. Furthermore, Fbs1 was shown to inhibit deglycosylation of denatured ribonuclease B by a cytosolic peptide:N-glycanase (PNGase). On the basis of these data, we suggest that Fbs1 captures malfolded glycoproteins, protecting them from the attack of PNGase, during the chaperoning or ubiquitinating operation in the cytosol.     

The Structural Differences between a Glycoprotein Specific F-Box Protein Fbs1 and Its Homologous Protein FBG3

The Skp1-Cul1-F-box protein (SCF) complex catalyzes protein ubiquitination in diverse cellular processes and is one of the best-characterized ubiquitin ligases. F-box proteins determine the substrate specificities of SCF ubiquitin ligases. Among these, Fbs1/FBG1/FBXO2, Fbs2/FBG2/FBXO6, and Fbs3/FBG5/FBXO27 recognize the N-glycans of glycoproteins, whereas FBG3/FBXO44 has no sugar-binding activity, despite the high sequence homology and conservation of the residues necessary for oligosaccharide binding between Fbs1–3 and FBG3. Here we determined the crystal structure of the Skp1–FBG3 complex at a resolution of 2.6 Å. The substrate-binding domain of FBG3 is composed of a 10-stranded antiparallel β-sandwich with three helices. Although the overall structure of FBG3 is similar to that of Fbs1, the residues that form the Fbs1 carbohydrate-binding pocket failed to be superposed with the corresponding residues of FBG3. Structure-based mutational analysis shows that distinct hydrogen bond networks of four FBG3 loops, i.e., β2-β3, β5-β6, β7-β8, and β9-β10, prevent the formation of the carbohydrate-binding pocket shown in Fbs1.     

Skp1 stabilizes the conformation of F-box proteins

The SCF ubiquitin ligase complex consists of four components, Skp1, Cul1, ROC1/Rbx1, and a variable subunit F-box protein, which serves as a receptor for target proteins. The F-box proteins consist of an N-terminal ∼40 amino acid F-box domain that binds to Skp1 and the C-terminal substrate-binding domain. We have reported previously that Fbs1 and Fbs2 are N-linked glycoprotein-specific F-box proteins. In addition, other three F-box proteins, Fbg3, Fbg4, and Fbg5, show high homology to Fbs1 and Fbs2, but their functions remain largely unknown. Here we report that Skp1 assists in correct folding of exogenously expressed F-box proteins. Fbs2 as well as Fbg3, Fbg4, and Fbg5 proteins formed SCF complexes but did not bind to N-glycoproteins when exogenously expressed alone. However, co-expression of Fbs2 and Fbg5 with Skp1 facilitated their binding to glycoproteins that reacted with ConA. Furthermore, Skp1 increased the cellular concentrations of F-box proteins by preventing aggregate formation. These observations suggest that Skp1 plays an important role in stabilizing the conformation of these F-box proteins, which increases their expression levels and substrate-binding.     

The Nogo receptor 2 is a novel substrate of Fbs1

Members of the Nogo66 receptor family (NgR) are closely associated with nerve growth inhibition and plasticity in the CNS. All three members, NgR1, NgR2 and NgR3, are GPI anchored and highly glycosylated proteins. The binding and signaling properties of NgR1 are well described, but largely unknown for NgR2. At present the only known ligands are myelin associated glycoprotein (MAG) and amyloid beta precursor protein (APP). Despite the requirement of co-receptors for signaling no other binding partner has been uncovered. To learn more about the interactome of NgR2 we performed pull down experiments and were able to identify F-box protein that recognizes sugar chain 1 (Fbs1) as binding partner. We confirmed this finding with co-immunoprecipitations and in vitro binding assays and showed that the binding is mediated by the substrate recognition domain of Fbs1. As a substrate recognition protein of the SCF complex, Fbs1 binding leads to polyubiquitination and finally degradation of its substrates. This is the first time a member of the Nogo receptor family has been connected with an intracellular degradation pathway, which has not only implications for its production, but also for amyloid deposition in Alzheimer's disease.     

F-box proteins that contain sugar-binding domains

F-box proteins are the substrate-recognition subunits of the SCF (Skp1-Cul1-F-box protein) complex, which is the largest known class of E3 ubiquitin ligases. They play important roles in ubiquitin-dependent proteolysis in eukaryotes. The human genome contains about 70 genes for F-box proteins, and at least five homologous F-box proteins containing a conserved motif in their C-termini are thought to recognize sugar chain of N-linked glycoproteins. Among theses, Fbs1 and Fbs2 are perhaps involved in the endoplasmic reticulum-associated degradation pathway. In this review, I focus on the in vivo function of Fbs1 and homologous proteins, novel intracellular oligosaccharide recognition molecules involved in the quality control system.     

Nicotinic acetylcholine receptors of Drosophila: three subunits encoded by genomically linked genes can co-assemble into the same receptor complex

The second beta-like subunit (SBD) is a putative structural subunit of Drosophila melanogaster nicotinic acetylcholine receptors (nAChRs). Here we have produced specific antibodies against SBD to study, which other nAChR subunits can co-assemble with SBD in receptor complexes of the Drosophila nervous system. Immunohistochemical studies in the adult optic lobe revealed that SBD has a distribution similar to that of the alpha-subunit ALS in the synaptic neuropil. The subunits ALS, D(alpha)2 and SBD can be co-purified by alpha-bungarotoxin affinity chromatography. Moreover, anti-SBD antibodies co-precipitate ALS and D(alpha)2 and, vice versa, ALS and D(alpha)2 antibodies co-immunoprecipitate SBD protein. Two-step immunoaffinity chromatography with immobilized antibodies against ALS and D(alpha)2 revealed the existence of nAChR complexes that include ALS, D(alpha)2 and SBD as integral components. Interestingly, the genes encoding these three subunits appear to be directly linked in the Drosophila genome at region 96 A of the third chromosome. In addition, SBD appears to be a component of a different receptor complex, which includes the ARD protein as an additional beta-subunit, but neither ALS nor D(alpha)2 nor the third alpha-subunit D(alpha)3. These findings suggest a considerable complexity of the Drosophila nicotinic receptor system.     

Protein and carbohydrate moieties of a preparation of β-lactamase II

1. A crystalline preparation of beta-lactamase II has been separated into two moieties by gel filtration on a column of Sephadex G-100. 2. The first moiety consisted mainly of carbohydrate and showed virtually no beta-lactamase activity. 3. The second moiety was a protein of molecular weight 22500, which was enzymically active. 4. The protein moiety, like the original protein-carbohydrate complex, required Zn(2+) for beta-lactamase activity. It did not differ significantly from the complex in its behaviour to a number of cephalosporin substrates, but was less stable to heat than the complex. 5. About 30% of the total beta-lactamase activity was lost when the protein-carbohydrate complex was separated into the two moieties. This activity was regained when the protein and carbohydrate moieties were mixed, but the mixture did not show the heat stability of the original complex.     

A fluorescent glycolipid-binding peptide probe traces cholesterol dependent microdomain-derived trafficking pathways

Background

The uptake and intracellular trafficking of sphingolipids, which self-associate into plasma membrane microdomains, is associated with many pathological conditions, including viral and toxin infection, lipid storage disease, and neurodegenerative disease. However, the means available to label the trafficking pathways of sphingolipids in live cells are extremely limited. In order to address this problem, we have developed an exogenous, non-toxic probe consisting of a 25-amino acid sphingolipid binding domain, the SBD, derived from the amyloid peptide Aβ, and conjugated by a neutral linker with an organic fluorophore. The current work presents the characterization of the sphingolipid binding and live cell trafficking of this novel probe, the SBD peptide. SBD was the name given to a motif originally recognized by Fantini et al [1] in a number of glycolipid-associated proteins, and was proposed to interact with sphingolipids in membrane microdomains.

Methodology/Principal Findings

In accordance with Fantini''s model, optimal SBD binding to membranes depends on the presence of sphingolipids and cholesterol. In synthetic membrane binding assays, SBD interacts preferentially with raft-like lipid mixtures containing sphingomyelin, cholesterol, and complex gangliosides in a pH-dependent manner, but is less glycolipid-specific than Cholera toxin B (CtxB). Using quantitative time-course colocalization in live cells, we show that the uptake and intracellular trafficking route of SBD is unlike that of either the non-raft marker Transferrin or the raft markers CtxB and Flotillin2-GFP. However, SBD traverses an endolysosomal route that partially intersects with raft-associated pathways, with a major portion being diverted at a late time point to rab11-positive recycling endosomes. Trafficking of SBD to acidified compartments is strongly disrupted by cholesterol perturbations, consistent with the regulation of sphingolipid trafficking by cholesterol.

Conclusions/Significance

The current work presents the characterization and trafficking behavior of a novel sphingolipid-binding fluorescent probe, the SBD peptide. We show that SBD binding to membranes is dependent on the presence of cholesterol, sphingomyelin, and complex glycolipids. In addition, SBD targeting through the endolysosomal pathway in neurons is highly sensitive to cholesterol perturbations, making it a potentially useful tool for the analysis of sphingolipid trafficking in disease models that involve changes in cholesterol metabolism and storage.     

Skin telocytes versus fibroblasts: two distinct dermal cell populations

It is already accepted that telocytes (TCs) represent a new type of interstitial cells in human dermis. In normal skin, TCs have particular spatial relations with different dermal structures such as blood vessels, hair follicles, arrector pili muscles or segments of sebaceous and/or eccrine sweat glands. The distribution and the density of TCs is affected in various skin pathological conditions. Previous studies mentioned the particular (ultra)structure of TCs and also their immunophenotype, miR imprint or proteome, genome or secretome features. As fibroblast is the most common intersitital cell (also in human dermis), a dedicated comparison between human skin TCs and fibroblasts (Fbs) was required to be performed. In this study, using different techniques, we document several points of difference between human dermis TCs and Fbs. By transmission electron microscopy (TEM) and scanning electron microscopy (SEM), we demonstrated TCs with their hallmark cellular prolongations – telopodes. Thus, we showed their ultrastructural distinctiveness from Fbs. By RayBio Human Cytokine Antibody Array V analyses performed on the supernatant from separately cultured TCs and Fbs, we detected the cytokine profile of both cell types, individually. Two of 79 detected cytokines – epithelial‐derived neutrophil‐activating peptide 78 and granulocyte chemotactic protein‐2 – were 1.5 times higher in the supernatant of TCs (comparing with Fbs). On the other hand, 37 cytokines were at least 1.5 higher in Fbs supernatant (comparing with TCs), and among them six cytokines – interleukin 5, monocyte chemotactic protein‐3 (MCP‐3), MCP‐4, macrophage inflammatory protein‐3, angiogenin, thrombopoietin – being 9.5 times higher (results also confirmed by ELISA testing). In summary, using different techniques, we showed that human dermal TCs and Fbs are different in terms of ultrastructure and cytokine profile.     

Interaction of E. coli outer-membrane protein A with sugars on the receptors of the brain microvascular endothelial cells

Esherichia coli, the most common gram-negative bacteria, can penetrate the brain microvascular endothelial cells (BMECs) during the neonatal period to cause meningitis with significant morbidity and mortality. Experimental studies have shown that outer-membrane protein A (OmpA) of E. coli plays a key role in the initial steps of the invasion process by binding to specific sugar moieties present on the glycoproteins of BMEC. These experiments also show that polymers of chitobiose (GlcNAcbeta1-4GlcNAc) block the invasion, while epitopes substituted with the L-fucosyl group do not. We used HierDock computational technique that consists of a hierarchy of coarse grain docking method with molecular dynamics (MD) to predict the binding sites and energies of interactions of GlcNAcbeta1-4GlcNAc and other sugars with OmpA. The results suggest two important binding sites for the interaction of carbohydrate epitopes of BMEC glycoproteins to OmpA. We identify one site as the binding pocket for chitobiose (GlcNAcbeta1-4GlcNAc) in OmpA, while the second region (including loops 1 and 2) may be important for recognition of specific sugars. We find that the site involving loops 1 and 2 has relative binding energies that correlate well with experimental observations. This theoretical study elucidates the interaction sites of chitobiose with OmpA and the binding site predictions made in this article are testable either by mutation studies or invasion assays. These results can be further extended in suggesting possible peptide antagonists and drug design for therapeutic strategies.     

Structural basis for the alteration of coenzyme specificity in a malate dehydrogenase mutant

To elucidate the structural basis for the alteration of coenzyme specificity from NADH toward NADPH in a malate dehydrogenase mutant EX7 from Thermus flavus, we determined the crystal structures at 2.0 A resolution of EX7 complexed with NADPH and NADH, respectively. In the EX7-NADPH complex, Ser42 and Ser45 form hydrogen bonds with the 2'-phosphate group of the adenine ribose of NADPH, although the adenine moiety is not seen in the electron density map. In contrast, although Ser42 and Ser45 occupy a similar position in the EX7-NADH complex structure, both the adenine and adenine ribose moieties of NADH are missing in the map. These results and kinetic analysis of site-directed mutant enzymes indicate (1) that the preference of EX7 for NADPH over NADH is ascribed to the recognition of the 2'-phosphate group by two Ser and Arg44, and (2) that the adenine moiety of NADPH is not recognized in this mutant.     

Preparation and kinetic properties of 5-ethylphenazine-lactate-dehydrogenase-NAD+ conjugate, a semisynthetic lactate oxidase showing a hide-and-seek effect.

5-Ethylphenazine-lactate-dehydrogenase-NAD+ conjugate (EP(+)-LDH-NAD+) was prepared by linking poly(ethylene glycol)-bound 5-ethylphenazine and poly(ethylene glycol)-bound NAD+ to lactate dehydrogenase. The average number of the ethylphenazine moieties bound per molecule of enzyme subunit was 0.46, and that of the NAD+ moieties was 0.32. This conjugate is a semisynthetic enzyme having lactate oxidase activity using oxygen or 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) as an electron acceptor; to make such conjugates seems to be a general method for artificially converting a dehydrogenase into an oxidase. When the concentration of oxygen or MTT is varied, the oxidase activity fits the Michaelis-Menten equation with the following kinetic constants: for the reaction system with oxygen, the turnover number per subunit is 2.3 min-1 and Km for oxygen is 1.91 mM; and for the system with MTT, the turnover number is 0.25 min-1 and Km for MTT is 0.076 mM. At the initial steady state of the oxidase reaction, only 2.1% of the NAD+ moieties of the conjugate are in the free state (i.e. not bound in the coenzyme-binding site of the lactate dehydrogenase moiety) and the rest are hidden in the coenzyme site; almost all the NAD+ moieties are in the reduced state. The apparent intramolecular rate constant for the reaction between a free NADH moiety and an oxidized ethylphenazine moiety is 2.3 s-1 and 2.1 s-1 for the systems with oxygen and with MTT, respectively. The apparent effective concentration of the free NADH moiety for the ethylphenazine moiety is 5.5 microM and is much smaller than that (0.34 mM) of the ethylphenazine moiety for the free NADH moiety; this difference is due to the effect of hiding the NADH moiety in the binding site, as the hidden NADH moiety cannot react with the ethylphenazine moiety.     

The binding characteristics and utilization of Aleuria aurantia,Lens culinaris and few other lectins in the elucidation of fucosyltransferase activities resembling cloned FT VI and apparently unique to colon cancer cells

Human colon carcinoma cell fucosyltransferase (FT) in contrast to the FTs of several human cancer cell lines, utilized GlcNAcbeta1,4GlcNAcbeta-O-Bn as an acceptor, the product being resistant to alpha1,6-L-Fucosidase and its formation being completely inhibited by LacNAc Type 2 acceptors. Further, this enzyme was twofold active towards the asialo agalacto glycopeptide as compared to the parent asialoglycopeptide. Only 60% of the GlcNAc moieties were released from [14C]fucosylated asialo agalacto triantennary glycopeptide by jack bean beta-N-acetylhexosaminidase. These alpha1,3-L-fucosylating activities on multiterminal GlcNAc residues and chitobiose were further examined by characterizing the products arising from fetuin triantennary and bovine IgG diantennary glycopeptides and their exoglycosidase-modified derivatives using lectin affinity chromatography. Utilization of [14C]fucosylated glycopeptides with cloned FTs indicated that Lens culinaris lectin and Aleuria aurantia lectin (AAL) required, respectively, the diantennary backbone and the chitobiose core alpha1,6-fucosyl residue for binding. The outer core alpha1,3- but not the alpha-1,2-fucosyl residues decreased the binding affinity of AAL. The AAL-binding fraction from [14C]fucosylated asialo fetuin, using colon carcinoma cell extract, contained 60% Endo F/PNGaseF resistant chains. Similarly AAL-binding species from [14C]fucosylated TFA-treated bovine IgG using colon carcinoma cell extract showed significant resistance to endo F/PNGaseF. However, no such resistance was found with the corresponding AAL non- and weak-binding species. Thus colon carcinoma cells have the capacity to fucosylate the chitobiose core in glycoproteins, and this alpha1,3-L-fucosylation is apparently responsible for the AAL binding of glycoproteins. A cloned FT VI was found to be very similar to this enzyme in acceptor substrate specificities. The colon cancer cell FT thus exhibits four catalytic roles, i.e., alpha1,3-L-fucosylation of: (a) Galbeta1,4GlcNAcbeta-; (b) multiterminal GlcNAc units in complex type chain; (c) the inner core chitobiose of glycopeptides and glycoproteins; and (d) the nonreducing terminal chiotobiose unit.     

Crystal Structure of the Stress-Inducible Human Heat Shock Protein 70 Substrate-Binding Domain in Complex with Peptide Substrate

The HSP70 family of molecular chaperones function to maintain protein quality control and homeostasis. The major stress-induced form, HSP70 (also called HSP72 or HSPA1A) is considered an important anti-cancer drug target because it is constitutively overexpressed in a number of human cancers and promotes cancer cell survival. All HSP70 family members contain two functional domains: an N-terminal nucleotide binding domain (NBD) and a C-terminal protein substrate-binding domain (SBD); the latter is subdivided into SBDα and SBDβ subdomains. The NBD and SBD structures of the bacterial ortholog, DnaK, have been characterized, but only the isolated NBD and SBDα segments of eukaryotic HSP70 proteins have been determined. Here we report the crystal structure of the substrate-bound human HSP70-SBD to 2 angstrom resolution. The overall fold of this SBD is similar to the corresponding domain in the substrate-bound DnaK structures, confirming a similar overall architecture of the orthologous bacterial and human HSP70 proteins. However, conformational differences are observed in the peptide-HSP70-SBD complex, particularly in the loop L_{α, β} that bridges SBDα to SBDβ, and the loop L_L,1 that connects the SBD and NBD. The interaction between the SBDα and SBDβ subdomains and the mode of substrate recognition is also different between DnaK and HSP70. This suggests that differences may exist in how different HSP70 proteins recognize their respective substrates. The high-resolution structure of the substrate-bound-HSP70-SBD complex provides a molecular platform for the rational design of small molecule compounds that preferentially target this C-terminal domain, in order to modulate human HSP70 function.     

Molecular design of an acid–base cooperative catalyst for RNA cleavage based on a dizinc complex

The effects of donor groups of dizinc complexes, formed from a 2:1 mixture of Zn(II) and a dinucleating ligand, on adenylyl(3'-5')adenosine (ApA) cleavage have been studied. Two dinucleating ligands were used: one had two 2-pyridylmethyl and two 2-hydroxyethyl moieties on the 1,3-diamino-2-propanol linker moiety (2), and the other had two 2-pyridylmethyl and two carboxymethyl moieties on the 1,3-diamino-2-propanol linker moiety (3(2-)). The dizinc complex with2 [(Zn(2+))(2)-2] showed higher activities toward ApA cleavage than the dizinc complex using an analogous dinucleating ligand having four 2-pyridylmethyl donor moieties [(Zn(2+))(2)-1] at pH 5-8. The former showed a bell-shaped pH-rate constant profile, whereas the latter showed a sigmoidal pattern. The differences in the pH-rate constant profile are attributable to the various distributions of the monohydroxo-dizinc species, i.e. dideprotonated species, which are responsible for ApA cleavage. The monohydroxo species of (Zn(2+))(2)-2 has two acidic protons, which are not present in the corresponding monohydroxo species of (Zn(2+))(2)-1. The existence of both intracomplex acid (ROH or H(2)O) and base catalysts (RO(-) or OH(-)) in (Zn(2+))(2)-2 can explain its higher activity toward ApA cleavage than that of (Zn(2+))(2)-1. In contrast, (Zn(2+))(2)-3(2-) showed lower activity toward ApA cleavage at pH 7.0, which can be ascribed to the absence of the monohydroxo-dizinc species under these conditions.     

Reaction of the octahedral antitumor complex trans-RuCl2 (DMSO)4 with 2'deoxyguanosine.

The reaction of the antitumor octahedral complex trans-RuCl2(DMSO)4 with 2'deoxyguanosine leads to the reversible formation of two diastereoisomeric monoadducts and one biadduct. This shows that it is possible to accommodate two purine bases in a cis configuration in an octahedral transition metal complex which exhibits antiblastic activity. All the product compounds are characterized by a guanine moiety coordination via the N7 atom. A marked decrease (about two pK units) is observed for the N1H pKa of the coordinated guanine moieties. The reversibility of the monodentate binding could explain the low toxicity of the ruthenium(II) complexes.     

Thermodynamics of reversible and irreversible unfolding and domain interactions of glucoamylase from Aspergillus niger studied by differential scanning and isothermal titration calorimetry.

The stability of three forms of glucoamylase from Aspergillus niger has been investigated by differential scanning and isothermal titration calorimetry: Glucoamylase 1 (GA1), which consists of a catalytic domain and a starch-binding domain (SBD) connected by a heavily O-glycosylated linker region; glucoamylase 2 (GA2), which lacks SBD; and a proteolytically cleaved glucoamylase (GACD), which contains the catalytic domain and part of the linker region. The structures of the catalytic domain with part of the linker region and of SBD are known from crystallography and NMR, respectively, but the precise spatial arrangement of the two domains in GA1 is unknown. To investigate the stability of the three glucoamylase forms, we unfolded the enzymes thermally by differential scanning calorimetry (DSC). Aggregation occurs upon heating GA1 and GA2 at pH values between 2.5 and 5.0, whereas no aggregation is observed at higher pH (5.5-7.5). At all pH values, the catalytic domain of GA1 and GA2 unfolds irreversibly, while SBD unfolds reversibly in the pH range 5. 5-7.5 where aggregation does not occur. The unfolding of the catalytic domain of all glucoamylase forms seems to follow an irreversible one-step mechanism with no observable reversible intermediates on the experimental time scale. SBD of GA1 unfolds reversibly, and the ratio between the van't Hoff and calorimetric enthalpies is 1.4 +/- 0.1. Assignment of peaks of the DSC profile to the domains at pH 7.5 is achieved by using two different ligands: Acarbose, a very strong inhibitor that binds exclusively to the catalytic domain, and beta-cyclodextrin, a small starch analogue of which 2 molecules bind solely to the two binding sites present in SBD. Differences are seen in the unfolding processes of GA1 and GA2 since the former unfolds with one peak at all pH values, while the calorimetric trace of the latter can be resolved into more peaks depending on pH and the chemical composition of the buffers. In general, peaks corresponding to unfolding of GA2 are more complex than the peaks of GA1 and GACD. Some part of GA2 unfolds before the rest of the molecule which may correspond to the linker region or a particular early unfolding part of the catalytic domain. This leads to the conclusion that the structure of the GA2 molecule has a larger cooperative unfolding unit and is less stable than the structures of GA1 and GACD and that the C-terminal part of the linker region has a destabilizing effect on the catalytic domain.