Genetic characterization of a new splice variant of the β2 subunit of the voltage-dependent calcium channel

This study reports a novel splice variant form of the voltage-dependent calcium channel ₂ subunit (_2g). This variant is composed of the conserved amino-terminal sequences of the _2a subunit, but lacks the -subunit interaction domain (BID), which is thought essential for interactions with the ₁ subunit. Gene structure analysis revealed that this gene was composed of 13 translated exons spread over 107 kb of the genome. The gene structure of the ₂ subunit was similar in exon-intron organization to the murine ₃ and human ₄ subunits. Electrophysiological evaluation revealed that _2a and _2g affected channel properties in different ways. The _2a subunit increased the peak amplitude, but failed to increase channel inactivation, while _2g had no significant effects on either the peak current amplitude or channel inactivation. Other subunits, such as ₃ and ₄, significantly increased the peak current and accelerated current inactivation.     

Structural studies on equine chorionic gonadotropin

The amino acid sequence of the subunit of equine chorionic gonadotropin (eCG, also pregnant mare serum gonadotropin, PMSG) has been determined. Overlapping peptides from tryptic and chymotrypic digests were isolated by a two-dimensional peptide mapping technique and sequenced by the Edman procedure. The proposed amino acid sequence of eCG is: (**Denotes carbohydrate attachment points.) This sequence differs significantly from that proposed by Rathnamet al. (1978) for equine follitropin subunit; in particular, their sequence lacked the first fourteen residues.For the subunit we have placed in sequence 104 amino acid residues by direct sequence determination and peptide overlap procedures; in addition, 37 residues have been placed provisionally by homology with the human chorionic gonadotropin (hCG) sequence and composition and/or sequence data for the peptides isolated in the present studies. Difficulties in the procurement of the hormone have stalled completion of the -subunit amino acid sequence determination. The data now available indicate that eCG -subunit is highly homologous to hCG subunit and the subunits of luteinizing hormone from the pituitary gland of the several species so far described. The proposed partial sequence of eCG is:     

Functional analysis of CK2β-derived syntheticfragments

Synthetic peptides reproducing the amino and carboxyl terminal region of CK2 subunit have been analyzed for their ability to mimic different properties of full length subunit. Peptide [1-77], containing both the autophosphorylation site and the down-regulatory domain 55-64, is readily phosphorylated by a subunit whose activity is concomitantly inhibited. Such inhibition is accompanied by a weak interaction detectable by BIAcore sensograms but not by far Western blots, and is not reversed by polylysine which conversely overcome inhibition of calmodulin phosphorylation by full length subunit. A strong interaction with is observed with [155-215] but not with its shorter derivative [170-215] as judged from far Western blotting and sucrose gradient ultracentrifugation analysis. Both peptides, however, affect the regular interaction between and subunits altering the autophosphorylation pattern and responsiveness to salt. [155-215], unlike [170-215] tends to aggregate more readily than full length subunit. This behaviour which is reminiscent of the homodimerization of full length subunit, would indicate that tight self-association of [155-215] crucially depends on residues in the 155-170 sequence. Failure of [1-77] fragment to mediate responsiveness to polybasic peptides and accentuated self-association propensity of [155-215] suggest that other structural elements between the sequences 1-77 and 155-215 are required in order to confer optimal functionality to the subunit.     

Inhibitory Effects of Bombusae concretio Salicea on Neuronal Secretion of Alzheimer's β-Amyloid Peptides, a Neurodegenerative Peptide

Alzheimer's disease (AD) is characterized by the age-related deposition of -amyloid (A) 40/42 peptide aggregates in vulnerable brain regions. Multiple levels of evidence implicate a central role for A in the pathophysiology of AD. A is generated by the regulated cleavage of a = 700 amino acid A precursor protein (APP). Full-length APP can undergo proteolytic cleavage either within the A domain to generate secreted sAPP or at the N-terminal and C-terminal domain(s) of A to generate amyloidogenic A peptides. Several epidemiological studies have reported that estrogen replacement therapy protects against the development of AD in postmenopausal women. The aim of this study was to elucidate the antioxidant neuroprotective mechanism of Bombusae concretio Salicea (BC). BC was effective protectants against oxidative glutamate toxicity in the murine neuroblastoma cells (N2a) and human neuroblastoma cells (SK-N-MC). BC exhibited similar protective properties against oxidative glutamate toxicity and H₂O₂ toxicity. BC exhibited an antioxidant activity at approximately 20 g/ml. BC of 5 g/ml was ineffective in preventing the oxidative modification of LDL. The half-maximal effective concentration for BC was 16 g/ml. These results suggested that BC supplementation in elderly men may be protective in the treatment of Alzheimer's disease (AD). We report here that treatment with BC increases the secretion of the nonamyloidogenic APP fragment, sAPP and decreases the secretion of A peptides from N2a cells and rat primary cerebrocortical neurons. These results raise the possibility that BC supplementation in elderly men may be protective in the treatment of AD.     

Anti-(transforming growth factor β) antibodies with predefined specificity inhibit metastasis of highly tumorigenic human xenotransplants in nu/nu mice

Monoclonal antibodies (mAb) were prepared against conjugated transforming growth factor 1 (TGF1) peptides: amino acid positions 48–60 and positions 86–101. Two antibodies, mAb 16-3G1 [anti-(48–60)] and mAb 5-2G6 [anti-(86–101)] cross-reacted with native TGF1,-2 and-3 (16-3G1) or only with native TGF1 (5-2G6). Both mAb were used to characterize TGF-mediated effects on the metastatic potential in nude mice of human carcinoma cell line SLU-1 and its metastatic subline SLU-M1. Autocrine TGF1-mediated up-regulation of cell proliferation and its suppression by anti-TGF antibodies in vitro was recorded for SLU-M1 cells whereas SLU-1 cell proliferation in vitro appeared to be refractory to anti-TGF antibodies and exogenous TGF-1. However, the potential of s.c. tumours to develop distant metastases in nude mice was about the same for both cell lines. Development of primary tumours and distant metastases could be suppressed by treatment of mice with anti-TGF antibodies. Thus we assume that the metastatic potential of tumour cells is independent of TGF-mediated growth-regulation effects in vitro. The anti-TGF-induced suppression of tumour progression and metastasis in nude mice might rather result from stimulation of the immune surveillance. TGF-mediated autocrine down-regulation of MHC-unrestricted cytotoxicity of activated human monocytes and CD56⁺ LAK cells and its reversion by anti-TGF antibodies could be readily demonstrated. In all our experimental series, the neutralizing potential of both anti-TGF antibodies, though directed against opposite sites of the TGF1 molecule, was very similar.     

Identification of beta subunit of the rhesus monkey chorionic gonadotropin (rmCG)

Chorionic gonadotropin (CG) is a placental derived hormone that plays a crucial role in successful implantation and establishment of early pregnancy in the primates. The rhesus monkey was chosen as a model to understand the feasibility of developing human DNA immuno-contraceptive. The coding region of rhesus monkey CG -subunit (rmCG) was isolated by the TDRT-PCR method. The nucleotide sequence including the leader peptide was 499 nucleotide long and encoded 166 amino acids. In comparing with the previous known primates CG -subunits, the rmCG was the highest degree of homology with baboon CG -subunit at the deduced amino acid sequence (94%), 79.5% homology with human CG -subunit and 70.4% homology with marmoset monkey CG -subunit. The eukaryotic expression vector pCMV4-rmCG inserted full-coding cDNA sequence of rmCG was constructed, and the expression of rmCG -subunit in HeLa cells transient expressing system in vitro and BALB/c mice in vivo was determined. The results demonstrated that the recombinant PCMV4-rmCG eukaryotic expression vector could express rmCG -subunit in vitro and in vivo.     

Interaction of Amyloid Beta-Protein with Anionic Phospholipids: Possible Involvement of Lys28 and C-Terminus Aliphatic Amino Acids

Fibrillar amyloid beta-protein (A) is the major protein of amyloid plaques in the brains of patients with Alzheimer's disease (AD). The mechanism by which normally produced soluble A gets fibrillized in AD is not clear. We studied the effect of neutral, zwitterionic, and anionic lipids on the fibrillization of A 1-40. We report here that acidic phospholipids such as phosphatidic acid, phosphatidylserine, phosphatidylinositol (PI), PI 4-phosphate, PI 4,5-P₂ and cardiolipin can increase the fibrillization of A, while the neutral lipids (diacylglycerol, cholesterol, cerebrosides), zwitterionic lipids (phosphatidylcholine, phosphatidylethanolamine, sphingomyelin) and anionic lipids lacking phosphate groups (sulfatides, gangliosides) do not affect A fibrillization. A was found to increase the fluorescence of 1-acyl-2-[12-[ (7-nitro-2-1, 3-benzoxadiazol-4-yl) amino] dodecanoyl]-sn-glycero-3-phosphate (NBD-PA) in a concentration-dependent manner, while no change was observed with 1-acyl-2-[12-[(7-nitro-2-1, 3-benzoxadiazol-4-yl) amino] dodecanoyl]-sn-glycero-3-phosphoethanolamine (NBD-PE). Under similar conditions, other proteins such as apolipoprotein E, gelsolin and polyglutamic acid did not interact with NBD-PA. The order of interaction of amyloid -peptides with NBD-PA was A 1-43 = A 1-42 = A 17-42 > A 1-40 = A 17-40. Other A peptides such as A 1-11, A 1-16, A 1-28, A 1-38, A 12-28, A 22-35, A 25-35, and A 31-35 did not increase the NBD-PA fluorescence. These results suggest that phosphate groups, fatty acids, and aliphatic amino acids at the C-terminus end of A 1-40/A 1-42 are essential for the interaction of A with anionic phospholipids, while hydrophilic A segment from 1-16 amino acids does not participate in this interaction. Since positively charged amino acids in A are necessary for the interaction with negatively charged phosphate groups of phospholipids, it is suggested that Lys²⁸ of A may provide anchor for the phosphate groups of lipids, while aliphatic amino acids (Val-Val-Ile-Ala) at the C-terminus of A interact with fatty acids of phospholipids.     

LH-producing cells in the ovine pituitary

Summary An immunocytochemical technique using the peroxidase-antiperoxidase complex (PAP) was applied to identify and characterize the LH-secreting cells in the ovine pituitary at the ultrastructural level. These cells, round or oval in shape, possessing flattened cisternae of the rough endoplasmic reticulum, contain one class of secretory granules (mean diameter 250 nm) and large dense bodies (600 to 800 nm in diameter). LH molecules and the two subunits LH and LH were localized on the secretory granules and on the small granules near the Golgi complex. The large dense bodies, the cisternae of the endoplasmic reticulum and the saccules of the Golgi complex showed no reaction product.Abbreviations used in this Article O-LH ovine luteinizing hormone - b-LH bovine luteinizing hormone - p-LH porcine luteinizing hormone - p-LH porcine LH subunit - p-LH porcine LH subunit - O-FSH ovine follicle stimulating hormone - b-TSH bovine thyrotropic hormone - A-b LH antiserum to bovine LH - A-pLH antiserum to porcine LH subunit - A-pLH antiserum to porcine LH subunit     

Modification of domains of α and β subunits of F1-ATPase from the thermophylic bacterium PS3, in their isolated and associated forms,by 3′-O-(4-benzoyl)benzoyl adenosine 5′-triphosphate (BzATP)

Photoaffinity labeling by 3-O-(4-benzoyl)benzoyl adenosine 5-triphosphate (BzATP) of the adenine nucleotide binding site(s) on isolated and complexed and subunits of F₁-ATPase from the thermophilic bacterium PS3 (TF₁) is described. BzATP binds to both isolated and subunits, to complexed subunit but not to complexed subunit. Amino acid sequence determination of radiolabeled peptides obtained by proteolytic digestion of [-³²P]BzATP-labeled subunit indicates that residues on both the amino-terminal (residues A41-E67) and carboxy-terminal (residues Q422–Q476) were modified by BzATP. One of the residues in the carboxy-terminal modified by BzATP is most probably Q422. Although the binding stoichiometry of 1 mol of BzATP incorporated by either isolated or complexed subunit was maintained, the spatial conformation of the polypeptide determines which amino acid residue(s) is more accessible to the reactive radical. CNBr derived fragments G10-M64, E75-M233, and D390-M469 were labeled with the isolated subunit. With complexed subunit the label was found only in CNBr fragments: E75-M233 and G339-M389. The locations where the covalently bound BzATP was found, in the soluble and assembled subunits, indicate that different conformational states exist. In the isolated form of the and subunits the amino- and carboxy-termini can fold and reach the central domain of the polypeptide, the domain containing the adenine nucleotide binding site. When combines with to form the ₃₃ core complex the new conformation of the subunits is such that covalent labeling by BzATP of and of the amino terminal of subunit is excluded.     

The β-1,3-glucan-binding protein from the crayfish Pacifastacus leniusculus,when reacted with a β-1,3-glucan,induces spreading and degranulation of crayfish granular cells

Summary A -1,3-glucan-binding protein (GBP) was purified from crayfish plasma, and incubated with laminarin (L), a -1,3-glucan. The GBP reacted with laminarin (GBP-L) induced strong spreading and partial degranulation of isolated and separated crayfish granular haemocytes. However, neither the GBP nor laminarin alone induced any changes in the crayfish granular cells. When monolayers of granular haemocytes were incubated with 20 g of GBP-L, more than 82% of the haemocytes were affected. The activity of GBP-L on granular cells was dose-dependent and a plateau was reached at 10 g of GBP-L. The degranulation of crayfish haemocytes induced by GBP-L seemed to occur by a regulated exocytosis, since it was strongly inhibited by specific blockers of this process such as SITS or calmidazolium. Monospecific anti-GBP antibodies also totally blocked the effect of GBP-L on crayfish granular cells. Indirect immunofluoresence staining demonstrated that the GBP-L could bind to the surface of granular cells, whereas GBP did not bind or bound very weakly to the haemocyte surface.     

Relationship of amino acid sequence to immunological activity — syntheses and structure-activity relationships of thymosinβ 4 family

Summary Four peptides related to thymosin ₄ family and its six fragments were synthesized by the solution method. Among them, the four peptides related to thymosin ₄ family and its two fragments were found to have restoring activity on the impaired blastogenic response of T-lymphocytes isolated from uremic patients, but the other four fragments had no effect. Studies on the structureactivity relationships suggest that the tricosapeptide moiety corresponding to amino acids 16–38 of thymosin ₄ is found to be an important moiety on impaired immunological deficiency.Amino acids and their derivatives used in this investigations were of the L-configuration. The following abbreviations are used: DMF, dimethylformamide; DMSO, dimethylsulfoxide; Boc, tert-butoxycarbonyl; Z, benzyloxycarbonyl; NP, p-nitrophenyl; ONp, p-nitrophenyl ester; OBzl, benzyl ester; Bzl, benzyl; Troc,,,-tricloroethoxycarbonly; Su, N-hydroxysuccinimide; NMM, N-methylmorpholine; OSu, N-hydroxysuccinimide ester; WSCI, l-ethyl-3; (3-dimethylaminopropyl) carbodiimide; HOBT, N-hydroxybenzotriazole; TFA, trifluoroacetic acid; TLC, thin-layer chromatography; E-rosette, a rosette with sheep erythrocytes; EDTA, ethylenediaminetetraacetic acid; PHA, phytohemagglutinin; HPLC, high-performance liquid chromatography; ONb, p-nitrobenzyl ester; DCC, dicyclohexyl-carbodiimide.     

β-Amyloid Peptide Binds Equivalently to Binary and Ternary α<Subscript>2</Subscript>-Macroglobulin–protease Complexes

₂-Macroglobulin (₂M) is a protease inhibitor that has separate binding sites for transforming growth factor- (TGF-) and -amyloid peptide (A), both of which have been identified in the ₂M sequence. In the 3D-structure of ₂M, TGF- occupies the ₂M central cavity, overlapping with the space that can accommodate up to two molecules of protease. As a result, ternary ₂M–protease complexes (2 mol protease/mol ₂M) have been reported to not bind TGF-. The goal of the present study was to test whether binding of A to ₂M is controlled by steric constraints imposed by associated proteases, similarly to TGF-. We confirmed that binary ₂M–trypsin complex (1 mol trypsin/mol ₂M) binds increased amounts of TGF-1, compared with native ₂M, while ternary ₂M–trypsin complex binds substantially decreased amounts of TGF-1. By contrast, A-binding to binary and ternary ₂M–trypsin complex was equivalent. In both cases, binding was substantially increased compared with the negligible level observed with native ₂M. Plasmin is a large protease (Mr ~82,000) that substantially occupies the ₂M central cavity; however, ₂M–plasmin complex also bound increased amounts of A, compared with native ₂M. We conclude that A accesses its binding site, in ₂M, from outside the ₂M central cavity. The TGF--and A-binding sites are spatially separated not only in the primary sequence of ₂M, but also in the 3D-structure.     

Primary structure of the marmoset (Saguinus fuscicollis) hemoglobin. I. Use of tryptic maleylated peptides in the solubilization and sequence elucidation of the α- and β-chains

The primary structure of adult marmoset hemoglobin has been determined. The - and -chains of HbA were separated on a CM23 column in 8 M urea using a sodium phosphate gradient. Tryptic digests of the - and -chains were fractionated on a Dowex 50W-X2 column using a pH and pyridine acetate gradient. Large peptide fragments were obtained by the cyanogen bromide cleavage of the - and -chains, as well as by tryptic digestion of the maleylated - and -chains. The sequence was derived from the amino acid compositions and sequences of the individual tryptic peptide, automated sequence determination of intact - and -chains, as well as automated sequence determination of cyanogen bromide fragments and tryptic maleylated peptides derived from the - and -chains. The complete structure of marmoset adult hemoglobin is closely homologous to that of other primate hemoglobins. The sequence of the marmoset -chain differs from the -chain of human HbA at positions 8, 19, 23, 68, and 116. The -chain from marmoset HbA differs from the -chain of human HbA at positions 5, 13, 21, 50, 87, and 125.This work was supported in part by funds from a Physicians' Medical Education and Research Foundation Grant of the University of Tennessee Memorial Research Hospital and by NIH General Research Support Grant FR-5541 to the institution.     

A glycine-rich sequence in the catalytic site of F-type ATPase

Affinity labeling and genetic studies on the glycine-rich sequence of the subunit ofE. coli F-type ATPase are discussed. A model of the structure of the enzyme near the phosphate moiety is proposed.     

Effect of 17β-estradiol on cells stained for FSHβ and/or LHβ in the dog pituitary gland

Summary The effect of long-term treatment (52 weeks) with high doses of 17-estradiol (1.28 mg/kg/week intramuscularly) on gonadotrophs was studied in the pituitary gland of the beagle bitch. For immunochemical staining the immunoperoxidase technique and antisera to the specific beta () subunits of FSH and LH were employed. For control purposes antisera to the following hormones were also used: bovine TSH, canine GH, canine PRL and porcine ACTH¹. In the pars distalis and pars tuberalis of control bitches, in addition to the cells which react solely with antisera to either LH or FSH, most cells were reactive to both antisera. The cells stained for FSH were less numerous than those shown to contain LH. TSH, PRL, GH and ACTH/MSH were localized in distinctly different cell types in the pars distalis of all control animals. In the treated bitches, almost complete regression of cells classically identified as gonadotrophs and stained for LH was observed. On the other hand, using the antiserum to FSH, selective immunochemical staining was localized in cells fitting the morphological characteristics of TSH cells. All these cells were also stained for TSH. However, a few cells were also shown to react solely with the antiserum to TSH. These cells, which seem to contain both TSH and FSH, were further clearly differentiated from PRL, GH and ACTH/MSH cells on the basis of their cytological features, intraglandular distribution and by immunochemical double staining. These observations support the concept that the one cell-one hormone theory may not necessarily apply to the glycoprotein hormones of the dog pituitary gland.Abbreviations of Hormones cited in this Paper ACTH Adrenocorticotropin - FSH Follicle Stimulating Hormone - GH Growth Hormone - LH Luteinizing Hormone - MSH Melanocyte Stimulating Hormone - PRL Prolactin - TSH Thyrotropin The authors are grateful to Mrs. K. Oertel for carrying out the experimental work on animals, to Mrs. B. Schilk and Miss U. Tüshaus for their excellent technical assistance, and to Dr. P. Günzel for his advice and encouragement     

Nucleotide and derived amino acid sequence of the cyanogenic β-glucosidase (linamarase) from white clover (Trifolium repens L.)

The nucleotide sequence and derived amino acid sequence of two different -glucosidase cDNA clones were determined. One clone (TRE104) was identified as the cyanogenic -glucosidase by homology with the N-terminal and internal peptide amino acid sequence of the purified enzyme. The biological function of the other -glycosidase (TRE361) is not known. Co-segregation of genomic restriction fragments uniquely identified by each cDNA clone shows that these two genes are linked in the white clover genome. Both TRE104 and TRE361 fragments co-segregate with cyanogenic -glucosidase activity. Extensive homology was found between the white clover -glucosidase sequences and a group of prokaryote and mammalian -glycosidases. This group of sequences has no homology with a separate set of -glucosidase genes isolated from fungi and the thermophilic bacterium Clostridium thermocellum.     

Interactions of protein kinase CK2β subunit within the holoenzyme and with other proteins

Protein kinase CK2 is a ubiquitous, highly conserved protein kinase with a tetrameric 22 structure. For the formation of this tetrameric complex a - dimer seems to be a prerequisite. Using the two-hybrid system and a series of CK2 deletion mutants, we mapped domains involved in - and - interactions. We also detected an intramolecular b interaction within the amino acid stretch 132-165.Using CK2 as a bait in a two-hybrid library screening several new putative cellular partners have been identified, among them the S6 kinase p90rsk, the putative tumor suppressor protein Doc-1, the Fas-associated protein FAF1, the mitochondrial translational initiation factor 2 and propionyl CoA carboxylase subunit.     

Characterization of cDNA clones for rye endosperm β-amylase and analysis of β-amylase deficiency in rye mutant lines

Summary In order to characterize a -amylase deficiency in the endosperm of mutant rye lines, homologous cDNA probes were prepared. A rye cDNA library was constructed from a normal line and screened with a barley -amylase probe. Three partial cDNA clones specific for endosperm -amylase in rye were isolated and characterized. The largest of these clones was used to investigate the expression of endosperm -amylase in mutant and normal lines by Northern hybridization. These experiments, as well as in vitro translation experiments, demonstrate the absence of endosperm -amylase mRNA in mutant lines. Sequencing of three different cDNA clones revealed a single nucleotide difference, which suggests that two genes encoding endosperm -amylase genes might exist in rye. From Southern blots we anticipate that these two genes are tightly linked. Results of these experiments and previous data indicating that the mutation was located within the -amylase locus on chromosome 5 are consistent with the hypothesis that the mutation results from a deletion simultaneously affecting the two genes. However, due to extensive polymorphism within normal lines used as control, additional experiments will be required to further substantiate this conclusion. The deduced amino acid sequence reveals the occurrence of three short glycine-rich repeats containing 11 or 12 residues close to the carboxyl terminus of the protein. A comparison of the nucleotide sequence between rye and previously described barley cDNA clones revealed ca. 90% homology at the amino acid level, except in this C-terminal repeated part, where it drops to 45%.     

Reevaluation of the amino acid sequence of porcine follitropin

We have reevaluated the sequence of porcine follicle-stimulating hormone (pFSH) with more recent protein-sequencing methodology. This has led to revision of the earlier proposed sequence. As with almost all reported gonadotropin -subunits, NH₂-terminal heterogeneity was found in the porcine FSH -subunit (FSH), starting with residue Phe (1), Asp (3), Gly (4), or Thr (7). In the -subunit, there were found to be at least two molecular species, starting with residue Asn (1) (minor 20%) or Cys (3) (major 80%) as NH₂-terminal and ending at residue Glu (108) as COOH-terminal. The net effect of the present revisions is to increase the homology of pFSH with other reported follitropin sequences. Apparent differences in the half-cystine placements in a previous proposal for pFSH compared with other species of FSH are no longer tenable. The half-cystine placements thus remain a constant structural feature throughout the gonadotropin hormones (choriogonadotropin, follitropin, and lutropin).     

Nucleotide sequence of the Clostridium thermocellum bgIB gene encoding thermostable beta-glucosidase B: homology to fungal beta-glucosidases

Summary The nucleotide sequence of the bglB gene, coding for the thermostable -glucosidase B of Clostridium thermocellum was determined. The coding region of 2265 bp was identified by comparison with the N-terminal amino acid sequence of -glucosidase B purified from Escherichia coli. The derived amino acid sequence corresponding to a polypeptide of M_r 84100 was confirmed by sequencing of the C-terminal peptide generated by cleavage with cyanogen bromide. The protein bears no resemblance to other bacterial -glucosidase sequences. However, extensive regions of homology were identified between the C. thermocellum enzyme and fungal -glucosidases. The N-terminal homologous region contains an amino acid sequence very similar to the active site of -glucosidase A₃ from Aspergillus wentii. The striking sequence similarities between C. thermocellum -glucosidase B and Kluyveromyces fragilis -glucosidase suggest the possibility of a genetic exchange between thermophilic anaerobic bacteria and yeasts.