A new enzyme mixture to increase the yield and transplant rate of autologous and allogeneic human islet products.

BACKGROUND: The optimal enzyme blend that maximizes human islet yield for transplantation remains to be determined. In this study, we evaluated eight different enzyme combinations (ECs) in an attempt to improve islet yield. The ECs consisted of purified, intact or truncated class 1 (C1) and class 2 (C2) collagenases from Clostridium histolyticum (Ch), and neutral protease (NP) from Bacillus thermoproteolyticus rokko (thermolysin) or Ch (ChNP). METHODS: We report the results of 249 human islet isolations, including 99 deceased donors (research n=57, clinical n=42) and 150 chronic pancreatitis pancreases. We prepared a new enzyme mixture (NEM) composed of intact C1 and C2 collagenases and ChNP in place of thermolysin. The NEM was first tested in split pancreas (n=5) experiments and then used for islet autologous (n=21) and allogeneic transplantation (n=10). Islet isolation outcomes from eight different ECs were statistically compared using multivariate analysis. RESULTS: The NEM consistently achieved higher islet yields from pancreatitis (P<0.003) and deceased donor pancreases (P<0.001) than other standard ECs. Using the NEM, islet products met release criteria for transplantation from 8 of 10 consecutive pancreases, averaging 6510 ± 2150 islet equivalent number/gram (IEQ/g) pancreas and 694,681 ± 147,356 total IEQ/transplantation. In autologous isolation, the NEM yielded more than 200,000 IEQ from 19 of 21 pancreases (averaging 422,893 ± 181,329 total IEQ and 5979 ± 1469 IEQ/kg recipient body weight) regardless of the severity of fibrosis. CONCLUSIONS: A NEM composed of ChNP with CIzyme high intact C1 collagenase recovers higher islet yield from deceased and pancreatitis pancreases while retaining islet quality and function.

Effects of site-directed mutagenesis of the loop residue of the N-terminal domain Gly117 of thermolysin on its catalytic activity.

In the N-terminal domain of thermolysin, two polypeptide strands, Asn112-Ala113-Phe114-Trp115 and Ser118-Gln119-Met120-Val121-Tyr122, are connected by a short loop, Asn116-Gly117, to form an anti-parallel ß-sheet. The Asn112-Trp115 strand is located in the active site, while the Ser118-Tyr122 strand and the Asn116-Gly117 loop are located outside the active site. In this study, we explored the catalytic role of Gly117 by site-directed mutagenesis. Five variants, G117A (Gly117 is replaced by Ala), G117D, G117E, G117K, and G117R, were produced by co-expressing in Escherichia coli the mature and pro domains as independent polypeptides. The production levels were in the order G117E > wild type > G117K, G117R > G117D. G117A was hardly produced. This result is in contrast to our previous one that all 72 active-site thermolysin variants were produced at the similar levels whether they retained activity or not (M. Kusano et al. J. Biochem., 145, 103-113 (2009)). G117E exhibited lower activity in the hydrolysis of N-[3-(2-furyl)acryloyl]-glycyl-L-leucine amide and higher activity in the hydrolysis of N-carbobenzoxy-L-aspartyl-L-phenylalanine methyl ester than the wild-type thermolysin. G117K and G117R exhibited considerably reduced activities. This suggests that Gly117 plays an important role in the activity and stability of thermolysin, presumably by affecting the geometries of the Asn112-Trp115 and Ser118-Tyr122 strands.

Tissue dissociation enzyme neutral protease assessment.

Neutral proteases, essential components of purified tissue dissociation enzymes required for successful human islet isolation, show variable activities and effects of substrate on their activities. Initially we used a spectrophotometric endpoint assay with azocasein substrate to measure neutral protease activity. After critical review of the results, we observed these data to be inconsistent and not correlating expected differences in specific activities between thermolysin and Bacillus polymyxa proteases. This observation led to the development of a fluorescent microplate assay using fluorescein isothyocyanate-conjugated bovine serum albumin (FITC-BSA) as the substrate. This simpler, more flexible method offered a homogeneous, kinetic enzyme assay allowing determination of steady state reaction rates of sample replicates at various dilutions. The assay had a linear range of 4- to 8-fold and interassay coefficients of variation for B polymyxa protease and thermolysin of <9% and <15%, respectively, which were lower than those using the spectrophotometric endpoint assay, namely, 54% and 36%, respectively. This format allowed for incorporation of enzyme inhibitors, as illustrated by addition of sulfhydryl protease inhibitors, which, consistent with earlier reports, strongly indicated that the main contaminant in purified collagenase preparations was clostripain. Determination of the specific activities for several purified neutral proteases showed that the B polymyxa and Clostridium histolyticum proteases had approximately 40% and 15% specific activities, respectively, of those obtained with purified thermolysin, indicating the different characteristics of neutral protease enzymes for cell isolation procedures.

Comparison of human islet isolation outcomes using a new mammalian tissue-free enzyme versus collagenase NB-1.

BACKGROUND: After the discontinuation of the manufacturing Liberase HI because of a small potential for prion disease transmission, Roche Diagnostics (Indianapolis, IN) developed a new enzyme product (Liberase MTF [mammalian tissue free]), which is similar to Liberase HI with the exception that no mammalian tissue is used in the manufacture of the collagenase component. We report our experience using the MTF enzyme in clinical islet isolations compared with Serva NB-1 with modified enzyme delivery method. METHODS: Islets were isolated from 41 pancreata using MTF enzyme (n=17) or NB-1 enzyme (n=24). NB-1 enzymes were delivered using a modified (nonsimultaneous) enzyme delivery method whereas isolations using MTF used the standard method of simultaneous collagenase and thermolysin perfusion. Islets were purified on a COBE 2991 Cell Blood Processor and subsequently cultured. RESULTS: The average islet mass after purification was 392+/-36 x 10 islet equivalent (IE) for MTF versus 371+/-40 x 10 IE for Serva NB-1 (P=0.63). Post-IE/cm of tissue was 110+/-9 x 10 IE/cm and 91+/-11 x 10 IE/cm for MTF and NB-1, respectively (P=0.07). The isolation success rate (>400,000 IE) for MTF was 53% compared with 33% for Serva (P=0.33). CONCLUSION: We conclude that MTF may be successfully used for high-yield human islet isolation and clinical transplantation and provides similar quality islets to those derived using NB-1.

Effects of the mutational combinations on the activity and stability of thermolysin.

We have previously indicated that three single mutations (Leu144-->Ser, Asp150-->Glu, and Ile168-->Ala) in the site-directed mutagenesis of thermolysin increase the activity and two single (Ser53-->Asp and Leu155-->Ala) and one triple (Gly8-->Cys/Asn60-->Cys/Ser65-->Pro) mutations increase the stability. In the present study, aiming to generate highly active and stable thermolysin variants, we combined these mutations and analyzed the effect of combinations on the activity and stability of thermolysin. The combination of the mutations of Leu144-->Ser and Asp150-->Glu yielded the most significant increase in the hydrolytic activities for N-[3-(2-furyl)acryloyl]-Gly-L-Leu amide (FAGLA) and N-carbobenzoxy-L-Asp-L-Phe methyl ester (ZDFM), while that of Leu144-->Ser and Ile168-->Ala abolished the activity. The combination of Ser53-->Asp and Leu155-->Ala yielded the greatest increase in the thermal stability, while that of Ser53-->Asp and Gly8-->Cys/Asn60-->Cys/Ser65-->Pro increased the stability as high as the individual mutations do. The combination of three mutations of Leu144-->Ser, Asp150-->Glu, and Ser53-->Asp yielded a variant L144S/D150E/S53D with improved activity and stability. Its k(cat)/K(m) values in the hydrolysis of FAGLA and ZDFM were 8.6 and 10.2 times higher than those of wild-type thermolysin (WT), respectively, and its rate constant for thermal inactivation at 80 degrees C was 60% of that of WT.

Processing of protealysin precursor.

Protealysin, a protease previously described by us in Serratia proteamaculans, belongs to the group of thermolysin-like proteases (TLPs) that differ from classical TLPs by the precursor structural organization. The propeptide of protealysin precursor has no significant structural similarity to the propeptides of most TLPs. The functions of protealysin-like precursors and mechanisms of their action remain unclear. We studied the pathway of protealysin precursor processing in vitro using standard approaches: modification of the catalytic site and monitoring immobilized precursor maturation. The Glu(113) --> Ala substitution inhibited the precursor maturation, which pointed to the autocatalytic processing. The mutant precursor exposure to active protealysin converted it to the mature enzyme, thus, indicating the intermolecular processing. Intermolecular processing of the mutant protein by other proteases such as thermolysin or subtilisin is also possible. The intact protealysin precursor was efficiently autoprocessed in solution but not after immobilization. These data indicate that the processing of protealysin precursor differs from that of classical TLPs. The protealysin propeptide is cleaved by an autocatalytic or heterocatalytic intermolecular mechanism and is most likely not removed intramolecularly.

Deterioration and variability of highly purified collagenase blends used in clinical islet isolation.

BACKGROUND: The quality and stability of enzyme blends used in islet cell processing are critical for successful human islet isolation. A wide variability in enzymatic activity among lots of Liberase HI has been reported. This study examines the interlot and intralot variability of Liberase HI and the over-time deterioration of enzyme quality based on the analysis of islet isolation outcomes. METHODS: The data of 169 human isolations processed for clinical islet transplantation, using five different lots of Liberase HI, were retrospectively analyzed. Inter- and intralot variables in the islet isolation were assessed over a 15-month period. RESULTS: The analysis revealed significant interlot differences in the digestion time, prepurification islet counts, percent recovery, viability, and glucose stimulation insulin index. Moreover, a significant decrease in the pre- and postpurification islet yield per pancreas weight (IEQ/g) in isolations processed by two different enzyme lots used over a 15-month period was observed, suggesting a progressive deterioration of enzyme quality. CONCLUSIONS: Our data demonstrate a significant lot-to-lot related variability in islet isolation outcomes. In addition, the over-time decline in isolation outcomes processed using a single enzyme lot was observed even when the enzyme blends were used within the expiration dating specified by the manufacturer.

Engineering, expression, purification, and production of recombinant thermolysin.

Thermolysin [EC 3.4.24.27] is a thermostable neutral zinc metalloproteinase originally identified in the culture broth of Bacillus thermoproteolyticus Rokko. Since the discovery in 1962, the enzyme has been extensively studied regarding its structure and catalytic mechanism. Today, thermolysin is a representative of zinc metalloproteinase and an attractive target in protein engineering to understand the catalytic mechanism, thermostability, and halophilicity. Thermolysin is used in industry, especially for the enzymatic synthesis of N-carbobenzoxy L-Asp-L-Phe methyl ester (ZDFM), a precursor of an artificial sweetener, aspartame. Generation of genetically engineered thermolysin with higher activity in the synthesis of ZDFM has been highly desired. In accordance with the expansion of studies on thermolysin, various strategies for its expression and purification have been devised and successfully used. In this review, we aim to outline recombinant thermolysins associated with their engineering, expression, purification, and production.

Improvement in performance of affinity gels containing Gly-D-Phe as a ligand to thermolysin due to increasing the spacer chain length.

The aim of this study was to improve the performance of affinity gels containing glycyl-D-phenylalanine (Gly-D-Phe) as a ligand to thermolysin. Gly-D-Phe was immobilized to the resin through spacers of varying chain lengths. The resulting affinity gels had spacer chain lengths of 2 carbon atoms and 11 and 13 carbon-and-oxygen atoms (designated T2, T11, and T13), and were characterized for their binding abilities to thermolysin. Measurement of adsorption isotherms showed that the association constants to thermolysin were in the order T13 > T11 > T2. In affinity column chromatography, in which 5 mg thermolysin was applied onto 1-ml volumes of the gels, the adsorption ratios of thermolysin were also in the order T13 > T11 > T2. These results indicate that the performance of affinity gels is improved by increasing the spacer chain length to 13 carbon-and-oxygen atoms.

A new method for the extracellular production of recombinant thermolysin by co-expressing the mature sequence and pro-sequence in Escherichia coli.

Thermolysin, a representative zinc metalloproteinase from Bacillus thermoproteolyticus, is synthesized as inactive pre-proenzyme and receives autocatalytic cleavage of the peptide bond linking the pro- and mature sequences. The conventional expression method for recombinant thermolysin requires the autocatalytic cleavage, so that production of a mutant thermolysin is affected by its autocatalytic digestion activity. In this study, we have established a new expression method that does not require the autocatalytic cleavage. The mature sequence of thermolysin containing an NH(2)-terminal pelB leader sequence and the pre-prosequence of thermolysin were co-expressed constitutively in Escherichia coli as independent polypeptides under the original promoter sequences in the npr gene which encodes thermolysin. Unlike the conventional expression method, not only the wild-type thermolysin but also mutant thermolysins [E143A (Glu143 is replaced with Ala), N112A, N112D, N112E, N112H, N112K and N112R] were produced into the culture medium. The wild-type enzyme expressed in the present method was indistinguishable from that expressed in the conventional method based on autocatalytic cleavage, as assessed by hydrolysis of N-[3-(2-furyl)acryloyl]-glycyl-L-leucine amide and N-carbobenzoxy-L-aspartyl-L-phenylalanine methyl ester. The present method should be useful especially for preparation of active-site mutants of thermolysin, which might have suppressed autocatalytic digestion activity. The results also demonstrate clearly that the covalent linking between the pro- and mature sequences is not necessary for the proper folding of the mature sequence by the propeptide in thermolysin.

Cloning, sequencing, expression, and characterization of protealysin, a novel neutral proteinase from Serratia proteamaculans representing a new group of thermolysin-like proteases with short N-terminal region of precursor.

The gene of Serratia proteamaculans proteinase, protealysin, was cloned, sequenced, and expressed in Escherichia coli. The gene encoded a precursor of 341 amino acids (AAs) with a significant homology to thermolysin-like proteinases (TLPs). The molecular weight of the purified mature active recombinant protein was 32 kDa, the N-terminal amino acid sequence was AKTSTGGEVI. The optimum pH for azocasein hydrolysis by protealysin was seven and it was completely inhibited by o-phenanthroline. The enzyme hydrolyzed 3-(2-furyl)acryloyl-glycyl-L-leucine amide, the standard substrate for TLPs, with k(cat)/K(m) ratio of (2.52 +/- 0.02) x 10(2) M(-1) s(-1). Protealysin maturation removes 50 AA from the N-terminus of the precursor. The removed region had no similarity with the preprosequence of thermolysin (232 AA) but was homologous to some other TLPs. These proteins shared a conserved 7-AA motif near the initial methionine. Such motif was also found in some nonproteolytic putative proteins; two of them were eukaryotic.

Characterization of Gly-D-Phe, Gly-L-Leu, and D-Phe as affinity ligands to thermolysin.

In this study, glycyl-D-phenylalanine (Gly-D-Phe), glycyl-L-leucine (Gly-L-Leu), and D-phenylalanine (D-Phe) were characterized for their abilities as affinity ligands to thermolysin. Each of the ligands was immobilized to the resin. The optimum pH for adsorption of thermolysin is 5.0-6.0 for each of the ligands. By the affinity column chromatography in which 2mg thermolysin was applied onto 4 ml volume of the resins at pH 5.5, the adsorption ratios based on casein hydrolysis activity were 100% for each of the ligands. However, the adsorption ratios of the resins containing Gly-L-Leu and D-Phe, unlike that of Gly-D-Phe, were progressively decreased with increasing the amounts of thermolysin applied to the column. Measurement of adsorption isotherms showed that the association constant to thermolysin at pH 5.5 of the resins containing Gly-D-Phe was (3.3+/-0.8)x10(5)M(-1), while those of Gly-L-Leu and D-Phe were approximately ten times less. This result is coincident with the observations of performances in affinity column chromatography. On the other hand, maximum thermolysin binding capacities were almost the same among the resins examined. These results indicate that Gly-D-Phe is more suitable than Gly-L-Leu and D-Phe as an affinity ligand for purification of thermolysin.

Extracellular production of recombinant thermolysin expressed in Escherichia coli, and its purification and enzymatic characterization.

Thermolysin is a representative zinc metalloproteinase derived from Bacillus thermoproteolyticus and a target in protein engineering to understand the catalytic mechanism and thermostability. Extracellular production of thermolysin has been achieved in Bacillus, but not in Escherichia coli, although it is the most widely used as a host for the production of recombinant proteins. In this study, we expressed thermolysin as a single polypeptide pre-proenzyme in E. coli under the original promoter sequences in the npr gene, the gene from B. thermoproteolyticus, which encodes thermolysin. Active mature thermolysin (34.6 kDa) was secreted into the culture medium. The recombinant thermolysin was purified to homogeneity by sequential column chromatography procedures of the supernatant with hydrophobic-interaction chromatography followed by affinity chromatography. The purified recombinant product is indistinguishable from natural thermolysin from B. thermoproteolyticus as assessed by hydrolysis of N-[3-(2-furyl)acryloyl]-glycyl-L-leucine amide and N-carbobenzoxy-L-asparatyl-L-phenylalanine methyl ester. The results demonstrate that our expression system should be useful for structural and functional analysis of thermolysin.

An exocellular thermolysin-like metalloprotease produced by Vibrio fluvialis: purification, characterization, and gene cloning.

An exocellular metalloprotease produced by Vibrio fluvialis, an enteropathogenic vibrio, was purified and characterized. The metalloprotease (V. fluvialis protease [VFP]) was found to have very similar characteristics to V. vulnificus protease, including a molecular mass of 45kDa, sensitivity to chelating agents or competitive inhibitors for thermolysin-like metalloproteases, and the substrate specificity. The structural gene for VFP was also cloned, and its nucleotide sequence was determined. The deduced amino acid sequence confirmed that VFP was a member of the thermolysin family. VFP, like V. vulnificus protease, showed the haemagglutinating, permeability-enhancing and haemorrhagic activities in addition to the proteolytic activity toward oligopeptide, casein or elastin.

Evaluation of a purified enzyme blend for the recovery and function of canine pancreatic islets.

Recently developed technologies enabling the production of a reproducible, purified enzyme blend for optimal human pancreatic islet isolation has renewed interest in clinical islet transplantation. The canine model has been an ideal preclinical model for the development of islet transplantation protocols. As seen in other species, the application of crude collagenase for isolating canine islets resulted in highly variable islet yields, extensive islet fragmentation, and variable islet functionality. We compared the function of commercially available crude collagenases with that of Liberase-CI purified enzyme blend for canine islet isolation. We also compared two manufacturing runs of Liberase-CI enzyme (lots 1 and 2) to demonstrate reproducibility of islet recovery and function. We report on the improved recovery and function of islets isolated using Liberase-CI enzyme. No difference in dog age, mean body weight, or pancreas weight were observed between the experimental groups. We observed a significantly higher postpurification recovery of islet equivalent number (IE) from pancreases processed using two lots of Liberase-CI enzyme (189 +/- 20 x 10(3) IE, n = 4) and lot 2 (234 +/- 39 x 10(3) IE, n = 7) than that obtained from pancreases processed with Sigma Type V (116.8 +/- 27 x 10(3) IE, n = 5), Serva collagenase (49 +/- 11.6 x 10(3) IE, n = 5, p < 0.05) or Boehringer-Mannheim (BM) Type P collagenase (85.4 +/- 25 x 10(3) IE, n = 5, p < 0.05, ANOVA). No significant differences were observed in islet yield recovery from pancreases processed using the two production lots of Liberase-CI enzyme. Islet survival after 48 h in culture at 37 degrees C was significantly higher from islets isolated using Liberase-CI enzyme (88 +/- 3.7% survival) when compared to Sigma Type V (81.8 +/- 3.3%), Serva (71.7 +/- 2.8%), and BM Type P (77 +/- 7.2%) (p < 0.05). Islet functional testing in vitro demonstrated islets isolated using crude collagenase had an increased insulin basal release and a reduced insulin stimulated response when compared with islets isolated using the two lots of Liberase-CI enzyme. The calculated stimulation index was 7.8 +/- 1.7, 3.1 +/- 0.6, and 4.8 +/- 1.1 for Sigma Type V, Serva, and BM Type P isolated islets, respectively, compared to 15.7 +/- 1.6 and 16.2 +/- 1.9 for islets isolated with Liberase-CI enzyme production lots 1 and 2, respectively (p < 0.05). This evaluation demonstrates that a purified enzyme blend can significantly improve islet recovery and function. It also demonstrates the manufacturing reproducibility of Liberase-CI enzyme lots resulting in the isolation of canine islets with the same degree of efficacy. A blend of purified enzymes, specifically formulated for canine islet isolation, can consistently yield large numbers of islets that survive longer in culture and demonstrate an improved insulin response in vitro.

Autolysis of thermolysin. Isolation and characterization of a folded three-fragment complex.

Incubation of the neutral metalloendopeptidase thermolysin at pH 7.2 in the presence of EDTA and/or low concentrations of calcium ions produces fast enzyme inactivation as a result of autolysis. The 'nicked' protein is a folded species composed of three tightly associated protein fragments. Dissociation of this complex can be achieved under denaturing conditions, such as gel filtration on a column equilibrated with 5 M guanidine hydrochloride or reverse-phase high-performance liquid chromatography (HPLC) at acidic pH. The positions of the peptide bond cleavages were defined by isolation of the individual fragments by HPLC and their characterization by amino acid analysis after acid hydrolysis, end-group determination and partial amino acid sequencing. The results of these analyses indicated that the nicked protein is composed of fragments 1-196, 197-204 and 205-316 and thus that the corresponding sites of limited proteolysis occur at the polypeptide chain loop involved in the binding of Ca(4) in native thermolysin [Matthews, B. W., Weaver, L. H. and Kester, W. R. (1974) J. Biol. Chem. 249, 8030-8044]. The overall conformational properties of nicked thermolysin are quite similar to those of the intact protein, as judged by spectroscopic measurements and by the fact that rabbit antibodies against native thermolysin recognize and precipitate the nicked protein in immunodiffusion assays. The nicked protein was much less stable to heat and unfolding agents than intact thermolysin. These results contribute to a better knowledge of the molecular mechanism of stabilization of native thermolysin by the four bound calcium ions and demonstrate that the function of Ca(4) is to stabilize the loop 190-205 on the surface of the molecule against autolysis.

Limited proteolysis of thermolysin by subtilisin: isolation and characterization of a partially active enzyme derivative.

Incubation of the neutral metalloendopeptidase thermolysin at pH 9-10 in the presence of 10 mM CaCl2 for 2 days at room temperature with subtilisin at a 50:1 molar ratio leads to a derivative possessing lower (approximately 3%) but intrinsic catalytic activity. This derivative, called thermolysin S, was isolated by gel filtration in approximately 80% yield and then separated from some residual intact thermolysin by an affinity chromatographic step on Sepharose-Gly-D-Phe. It was found that thermolysin S results from a tight association of two polypeptide fragments of apparent Mr of 24000 and 10000. Dissociation of the complex was achieved under strong denaturing conditions, such as gel filtration on a column equilibrated and eluted with 5 M guanidine hydrochloride. The positions of the clip sites were defined by amino acid analysis, end-group determination, and amino acid sequencing of the isolated fragments and shown to lie between Thr-4 and Ser-5, between Thr-224 and Gln-225, and also between Gln-225 and Asp-226. Thermolysin S, which is therefore a stable complex of fragments 5-224(225) and 225(226)-316, shows a shift in optimum pH of about 1 unit toward the acid range with respect to intact thermolysin and a Km essentially unchanged, with furylacryloyl-Gly-Leu-NH2 as substrate. Inhibitors of thermolysin such as ethoxyformic anhydride and Zn2+ ions inactivate also the nicked enzyme.(ABSTRACT TRUNCATED AT 250 WORDS)