Utilization of ATP-binding cassette exporter for hyperproduction of an exoprotein: construction of lipase-hyperproducing recombinant strains of Serratia marcescens.

The Serratia marcescens extracellular lipase (LipA) is an enzyme applicable to enantioselective hydrolysis of racemic substrates. The enzyme is secreted through an ATP-binding cassette (ABC) exporter, the Lip system, encoded by the lipBCD genes. The S. marcescens recombinant carrying pLIPE121, which encodes the lipA gene in pUC19, exhibited a higher LipA production level than the wild-type strain. However, the level was lower than expected, and secretion was suggested to be a bottleneck. lipBCD plasmids were introduced into S. marcescens recombinants harboring lipA plasmids and the effectiveness of the lipBCD plasmids in elevating LipA productivity was investigated. S. marcescens strains harboring both lipA and lipBCD plasmids showed sevenfold greater extracellular LipA activity than the strain harboring the lipA plasmid alone. A high level of extracellular LipA production (1,300 kU/ml) and high plasmid stability (enough to carry out large-scale cultivation) were observed under non-selective conditions. Addition of L-proline and Tween 80 was effective in increasing cell growth of the recombinant, which led to high LipA production. In batch cultivation using a 30-l jar fermentor, LipA production was achieved at a high level of 5,200 kU/ml. This is the first report describing utilization of ABC exporter for the overproduction of an industrially important extracellular protein.

Isolation and analysis of lipase-overproducing mutants of Serratia marcescens.

We have isolated a lipase-overproducing mutant, GE14, from Serratia marcescens 8000 after three rounds of N-methyl-N'-nitro-N-nitrosoguanidine mutagenesis. The mutant GE14 produced 95 kU/ml of extracellular lipase in the lipase medium, which was about threefold higher than that of produced by the original strain 8000. Enzymatic characteristics including specific activity of purified lipases from culture supernatants of GE14 and 8000 were almost same. The lipase gene (lipA) of GE14 contained two base substitutions; one in the promoter region and another in the N-terminal region of the lipA gene without an amino acid substitution. Promoter analysis using lipA-lacZ fusion plasmids revealed that these substitutions were responsible for the increase in the lipA expression level, independently. In contrast, no base substitution was found in the genes encoding the lipase secretion device, the Lip system. In addition, the genes coding for metalloprotease and the cell surface layer protein which are both secreted through the Lip system and associated with extracellular lipase production, also contained no base substitution. The strain GE14 carrying a high-copy-number lipA plasmid produced a larger amount of the extracellular lipase than the recombinant strains of 8000 and other mutants also did, indicating that GE14 was not only a lipase-overproducing strain, but also an advantageous host strain for overproducing the lipase by a recombinant DNA technique. These results suggest that the lipase-overproducing mutant GE14 and its recombinant strains are promising candidates for the industrial production of the S. marcescens lipase.

Domain separation precedes global unfolding of rhodanese.

The enzyme rhodanese was investigated for the conformational transition associated with its urea unfolding. When rhodanese was treated with 0 or 3 M urea, the activity was not significantly affected. 4.25 M urea treatment led to a time-dependent loss of activity in 60 min. Rhodanese was completely inactivated within 2 min in 6 M urea. The 1,1'-bi(4-anilino)naphthalene-5,5'-disulfonic acid fluorescence intensity was not significantly increased during 0, 3, and 6 M urea equilibrations, and the fluorescence was dramatically increased with 4.25 M urea, indicating that hydrophobic surfaces are exposed. After 0 and 3 M urea equilibration, rhodanese was not significantly proteolyzed with trypsin. Treatment with 4.25 M urea led to simultaneous formation of major 12-, 15.9-, 17-, and 21.2-kDa fragments, followed by progressive emergence of smaller peptides. The N termini of the 17- and 21.2-kDa bands were those of intact rhodanese. The N terminus of the 15.9-kDa band starts at the end of the interdomain tether. The 12-kDa band begins with either residue 183 or residue 187. The size and sequence information suggest that the 17- and 15.9-kDa bands correspond to the two domains. The 21.2- and 12-kDa bands appear to be generated through one-site tryptic cleavage. It is concluded that urea disrupts interaction between the two domains, increasing the accessibility of the interdomain tether that can be digested by trypsin. The released domains have increased proteolytic susceptibility and produce smaller peptides, which may represent subdomains of rhodanese.

Serratia marcescens S-layer protein is secreted extracellularly via an ATP-binding cassette exporter, the Lip system.

The Serratia marcescens Lip exporter belonging to the ATP-binding cassette (ABC) exporter is known to be involved in signal peptide-independent extracellular secretion of a lipase and a metalloprotease. Although the genes of secretory proteins and their ABC exporters are usually all reported to be linked in several gram-negative bacteria, neither the lipase nor the protease gene is located close to the Lip exporter genes, lipBCD. A gene (slaA) located upstream of the lipBCD genes was cloned, revealing that it encodes a polypeptide of 100 kDa and is partially similar to the Caulobacter crescentus paracrystalline cell surface layer (S-layer) protein. The Lip exporter-deficient mutants of S. marcescens failed to secrete the SlaA protein. Electron micrography demonstrated the cell surface layer of S. marcescens. The S-layer protein was secreted to the cultured media in Escherichia coli cells carrying the Lip exporter. Three ABC exporters, Prt, Has and Hly systems, could not allow the S-layer secretion, indicating that the S. marcescens S-layer protein is strictly recognized by the Lip system. This is the first report concerning secretion of an S-layer protein via its own secretion system.

D-lysine production from L-lysine by successive chemical racemization and microbial asymmetric degradation.

In order to develop a practical process for D-lysine production from L-lysine, successive chemical racemization and microbial asymmetric degradation were investigated. The racemization of L-lysine proceeded quantitatively at elevated temperatures. A sample of 1000 strains of bacteria, fungi, yeast and actinomyces were screened for the ability to degrade L-lysine asymmetrically. Microorganisms belonging to the Achromobacter, Agrobacterium, Candida, Comamonas, Flavobacterium, Proteus, Providencia, Pseudomonas and Yarrowia genera exhibited a high L-lysine-degrading activity. Comamonas testosteroni IAM 1048 was determined to be the best strain and used as a biocatalyst for eliminating the L isomer. The degradation rate of L-lysine with C. testosteroni IAM 1048 was influenced by pH, temperature and agitation speed. Under the optimal conditions, the L isomer in a 100-g/l mixture of racemic lysine was completely degraded within 72 h, with 47 g D-lysine/l left in the reaction mixture. Crystalline D-lysine, with a chemical purity greater than 99% and optical purity of 99.9% enantiomeric excess, was obtained at a yield of 38% from the reaction mixture by simple purification. An engineering analysis of L-lysine racemization and microbial degradation was carried out to establish the basis of process design for D-lysine production.

Paracoccus denitrificans aromatic amino acid aminotransferase: a model enzyme for the study of dual substrate recognition mechanism.

The gene for aromatic amino acid aminotransferase (ArAT) from Paracoccus denitrificans was cloned, sequenced, and overexpressed in Escherichia coli cells. The sequence differed from that reported previously [Takagi, T., Taniguchi, T., Yamamoto, Y., and Shibatani, T. (1991) Biotechnol. Appl. Biochem. 13, 112-119]. The enzyme (pdArAT) was purified to homogeneity, and characterized. It was similar to aspartate aminotransferase (AspAT) and ArAT of E. coli (ecArAT) in many respects, including gross protein structure and spectroscopic properties. pdArAT showed activities toward both dicarboxylic and aromatic substrates, and analysis of the binding of substrate analogs and quasisubstrates to the enzyme showed that both dicarboxylic and aromatic substrates take a similar orientation in the active site of pdArAT; these properties are essentially identical with those of ecArAT. As in the case of ecArAT, neutral amino acids with larger side chains are better substrates for pdArAT, suggesting that hydrophobic interaction between the substrate and the enzyme is important for the recognition of substrates with neutral side chains. pdArAT catalyzed transamination of phenylalanine and tyrosine far more efficiently (10(2)-fold in terms of kcat/Km) than those of straight-chain aliphatic amino acids with similar side-chain surface area, whereas ecArAT did not show significant preference for aromatic amino acids over aliphatic amino acids. This shows that the substrate-side-chain-binding pocket of pdArAT, as compared with the pocket of ecArAT, is well suited in shape for interaction with the phenyl and hydroxyphenyl rings of substrates. Thus, pdArAT is an ideal enzyme among ArATs for the study of the high-specificity recognition of two different kinds of substrates, the one having a carboxylic side chain and the other having an aromatic side chain.

Optical resolution of a 1,5-benzothiazepine derivative, a synthetic intermediate of diltiazem, by preferential crystallization and diastereomeric salt formation.

Practical preparation methods of an optically active intermediate of diltiazem, (+)-(2S,3S)-5-[2-(dimethylamino)ethyl]-2,3-dihydro-3-hydroxy- 2-(4-methoxyphenyl)-1,5-benzothiazepin-4(5H)-one [(+)-7], have been developed by the use of physicochemical and chemical resolutions. 1) The salt of (+/-)-7 with 3-amino-4-hydroxy-benzenesulfonic acid (AHS), was found to exist as a conglomerate and could be reproducibly resolved into (+)-7.AHS and (-)-7.AHS of 94-98% ee by a preferential crystallization procedure. 2) (+)-(1R)-3-Bromocamphor-9-sulfonic acid [(+)-BCS] was found to be an efficient resolving agent for (+/-)-7 and the diastereomeric resolution provided (+)-7.(+)-BCS.2H2O salt in > 43% yield and > 97% ee by fractional crystallization. It is presumed that the crystal water of (+)-7.(+)-BCS.2H2O plays an important role in the selective crystallization during this efficient resolution.

Alteration of rat liver 20S proteasome activities by age and food restriction.

The effects of age and food restriction on proteasome function in rat liver supernatant (100,000 x g) were investigated. The cellular level of the proteasome has been quantitated by using Western blot analysis. The level of the proteasome was not affected by either age or food restriction. The three best-characterized proteasomal peptidase activities, chymotrypsin-like (ChT-L), trypsin-like (T-L), and peptidylglutamyl peptide hydrolyzing (PGPH) activities, were measured in the presence and absence of the proteasomal activator, sodium dodecyl sulfate (SDS). Basal ChT-L, T-L, and PGPH activities were not markedly affected by either age or food restriction. SDS-stimulated ChT-L and T-L activities increased approximately 15% and approximately 30%, respectively, between 7 and 26 months of age, and the increase of both activities was prevented by food restriction. In marked contrast, the SDS-stimulated PGPH activity decreased approximately 40% with age. Food restriction, while not preventing the age-related decline, maintained higher levels of SDS-stimulated PGPH activity at all ages. The proteolytic activity of the proteasome toward casein was not altered by either age or food restriction. In conclusion, the cellular level of the proteasome as well as the caseinolytic activity of the proteasome appear to be unaffected by either age or food restriction. It appears unlikely that the proteasome activity changes are related to the reported age-associated decline of protein degradation. Simultaneously, proteasomal peptidase activities appear to be differentially regulated by both age and food restriction. It suggests more subtle age-related changes in proteasome function, which could include an effect on proteasomal subunit composition.

Effect of age and food restriction on alkaline protease activity in rat liver.

The effect of age and food restriction on the hepatic alkaline protease activity of 100,000 x g supernatant has been investigated using 7-, 16-, and 26-month-old Fischer 344 rats. The proteasome, a major component of alkaline protease activity, is activated by sodium dodecyl sulfate (SDS) and this property was exploited to gain insight into the effects of age and food restriction on proteasome activity. Three alkaline protease activities, chymotrypsin-like (ChT-L), trypsin-like (T-L), and peptidylglutamyl peptide hydrolyzing (PGPH) activities were measured. These activities are also commonly used as measurement of proteasomal activities. Basal ChT-L and PGPH activities were not markedly altered by either age or food restriction. The level of T-L activity did not change with age, but was decreased by food restriction. SDS-activated ChT-L activity increased 15% between 7 and 26 months of age and this increase was blocked by food restriction. SDS-activated PGPH activity decreased 40% and the decrease was ameliorated by food restriction. In conclusion, we have shown that the alteration of alkaline protease activities by age and food restriction is not uniform and that the changes observed are likely due to alterations of proteasomal activity. The lack of uniformity in these alterations indicates that any assessment of alkaline protease activity requires the measurement of more than one of the enzymatic activities. Lastly, the first evidence suggesting that age and food restriction can modulate proteasomal activity is presented.

The three genes lipB, lipC, and lipD involved in the extracellular secretion of the Serratia marcescens lipase which lacks an N-terminal signal peptide.

The extracellular lipase of Serratia marcescens Sr41, lacking a typical N-terminal signal sequence, is secreted via a signal peptide-independent pathway. The 20-kb SacI DNA fragment which allowed the extracellular lipase secretion was cloned from S. marcescens by selection of a phenotype conferring the extracellular lipase activity on the Escherichia coli cells. The subcloned 6.5-kb EcoRV fragment was revealed to contain three open reading frames which are composed of 588, 443, and 437 amino acid residues constituting an operon (lipBCD). Comparisons of the deduced amino acid sequences of the lipB, lipC, and lipD genes with those of the Erwinia chrysanthemi prtDEC, prtEEC, and prtFEC genes encoding the secretion apparatus of the E. chrysanthemi protease showed 55, 46, and 42% identity, respectively. The products of the lipB and lipC genes were 54 and 45% identical to the S. marcescens hasD and hasE gene products, respectively, which were secretory components for the S. marcescens heme-binding protein and metalloprotease. In the E. coli DH5 cells, all three lipBCD genes were essential for the extracellular secretion of both S. marcescens lipase and metalloprotease proteins, both of which lack an N-terminal signal sequence and are secreted via a signal-independent pathway. Although the function of the lipD gene seemed to be analogous to those of the prtFEC and tolC genes encoding third secretory components of ABC transporters, the E. coli TolC protein, which was functional for the S. marcescens Has system, could not replace LipD in the LipB-LipC-LipD transporter reconstituted in E. coli. These results indicated that these three proteins are components of the device which allows extracellular secretion of the extracellular proteins of S. marcescens and that their style is similar to that of the PrtDEF(EC) system.

Divergence of a flagellin protein in Serratia marcescens.

A gene (hag) encoding the flagellin (Fla) protein was cloned from Serratia marcescens (Sm) 8000, the wild-type strain of Sr41. The hag gene codes for a 348-amino-acid (aa) protein of 36.7 kDa. The predicted aa sequence showed 79% homology compared with the Fla of Sm 274 which has been reported previously [Harshey et al., Gene 79 (1989) 1-8]. Dot-matrix analysis of the Sm 8000 Fla showed that the N- and C-terminal regions of this protein were highly similar to those of other bacterial Fla. However, the aa sequence of the middle portion was quite different from that of a variant strain of the same species, Sm 274.

Sodium dodecyl sulfate (SDS) activation of the 20S proteasome in rat liver.

The peptidase activity of the 20S proteasome (multicatalytic protease complex) was examined in the 100,000g supernatant fraction prepared from rat liver tissue. Fluorogenic substrates for three proteasome peptidase activities were selected on the basis of (i) observation of an accelerated degradation in the presence of sodium dodecyl sulfate (SDS) and (ii) preferential degradation by the proteasome. Peptidase activities were assayed using an immunoprecipitation technique utilizing polyclonal antibodies raised against the purified rat proteasome. The ability to demonstrate SDS activation of the proteasome is shown to be dependent upon the choice of substrate. In addition, among the cytosolic peptidases, the property of SDS activation appears to be unique to the proteasome. SDS activation profiles were determined for each peptidase activity. Chymotrypsin-like and peptidylglutamyl peptide-hydrolyzing activities exhibit a broad plateau of activation between 0.04 and 0.05% SDS. Trypsin-like activity exhibits a sharp peak of activation at an SDS concentration of 0.04%. The SDS activation profile can be altered by changing the protein (proteasome) concentration, i.e., increasing protein (proteasome) concentration of the reaction mixture produces a marked rightward shift of the activation profile. On the other hand, changing the substrate concentration does not alter the profile. In conclusion, a technique for measuring proteasome peptidase activity in the 100,000g supernatant has been described. This approach increases the ease of measurement of peptidase activity and provides data which may more closely reflect the in vivo activity of the proteasome.

[Application of quantitative assay for endotoxin using immobilized histidine and filtration plate to parenteral drugs].

We improved the Limulus amebocyte lysate (LAL) test for endotoxins in parenteral drugs using immobilized histidine and a filtration plate. In order to deal with many samples at the same time and to apply to a routine assay, we used a filtration plate having 96 wells instead of a filter unit with a working volume of 2 ml. LAL test-affecting substances which are contained in a parenteral drug were separated from endotoxins by adsorbing endotoxins on immobilized histidine in the well of a filtration plate. Then the absorbed endotoxins were allowed to react with LAL reagent in the same well. We defined that this method had the higher precision than conventional methods and was not influenced by the concentrations of endotoxin and parenteral drugs. Hence we examined the recovery of endotoxin spiked to 23 kinds of parenteral drugs by this method, as a result, 100 +/- 25% of recovery was obtained from 17 kinds of them.

The lipA gene of Serratia marcescens which encodes an extracellular lipase having no N-terminal signal peptide.

The lipA gene encoding an extracellular lipase was cloned from the wild-type strain of Serratia marcescens Sr41. Nucleotide sequencing showed a major open reading frame encoding a 64.9-kDa protein of 613 amino acid residues; the deduced amino acid sequence contains a lipase consensus sequence, GXSXG. The lipase had 66 and 56% homologies with the lipases of Pseudomonas fluorescens B52 and P. fluorescens SIK W1, respectively, but did not show any overall homology with lipases from other origins. The Escherichia coli cells carrying the S. marcescens lipA gene did not secrete the lipase into the medium. The S. marcescens lipase had no conventional N-terminal signal sequence but was also not subjected to any processing at both the N-terminal and C-terminal regions. A specific short region similar to the regions of secretory proteins having no N-terminal signal peptide was observed in the amino acid sequence. Expression of the lipA gene in S. marcescens was affected by the carbon source and the addition of Tween 80.

Molecular cloning of the L-phenylalanine transaminase gene from Paracoccus denitrificans in Escherichia coli K-12.

The L-phenylalanine transaminase gene of Paracoccus denitrificans was cloned by a shotgun method using the Escherichia coli K-12 mutant DG30, which lacks three distinct transaminase genes. Plasmid pPAP142 was constructed by inserting a 2.2-kb fragment carrying the transaminase gene into pUC18. Strain E. coli K-12 HB101 cells harboring the plasmid produced 20-fold to 30-fold more transaminase than wild type P. denitrificans cells. The nucleotide sequence of the 2.2-kb fragment was determined, revealing that the deduced amino acid sequence of the transaminase of P. denitrificans is similar to that of other transaminases.

Basic study of fluoride-releasing adhesive resin for prevention of root caries: 1. Transfer of fluorine from the resin to dentin.

Release of fluoride ion from a newly-developed fluoride-releasing resin (F-resin) into phosphate buffer and incorporation of fluorine by the dentin were studied in vitro. The rate of fluoride release (amount of daily release) from a pellet of cured F-resin showed rapid decrease to 1/2 of the initial value by day 6, and then moderate decrease to 1/4 of the initial by day 30. However, the pellet continued to release fluoride at a low but constant rate for more than 520 days. In the first 30 days after F-resin was applied to the dentin surface, large amount of fluorine was concentrated within 100 mu of the subsurface dentin. With longer incubation period, the peak concentration of fluorine at the dentin surface decreased but the depth of penetration increased to about 180 mu by day 180.

Basic study of fluoride-releasing adhesive resin for prevention of root caries: 2. Enhancement of acid resistance of the dentin.

The human root dentin was covered with a fluoride-releasing resin (F-resin) and incubated for various periods up to 180 days in phosphate buffer. After removal of the F-resin, resistance of the dentin to decalcification by acetic acid--sodium acetate buffer was assessed by measuring both the amount of calcium dissolved during decalcification and the knoop hardness of dentin after decalcification. The acid resistance progressively increased for 90 days when assessed by measurement of calcium dissolution and for 30 days by measurement of the knoop hardness. The percent inhibition of calcium dissolution achieved by 30 days of incubation was about 90% of the maximal inhibition obtained by incubation in concentrated sodium fluoride solution (2.4 x 10(-2) M). The enhancement of acid resistance was not induced by the resin infiltrated layer of the surface dentin but by the fluoride released from the F-resin into the dentin.