Search

Silencing oflevB, the second structural gene of the tricistronic levansucrase operon encoding the endolevanase LevB, decreases the level of levansucrase expression inBacillus subtilis. Conversely, independent expression oflevBgreatly stimulates operon expression. This autogenous effect is mediated by thelevBtranscript, which carries an internal sequence (5′-AAAGCAGGCAA-3′) involved in the enhancing effect.In vitro, thelevBtranscript displays an affinity for the N-terminal fragment of SacY (KD0·2 μM), the regulatory protein that prevents transcription termination of the levansucrase operon. This positive-feedback loop leads to an increase in the operon expression whenB. subtilisis growing in the presence of high sucrose concentrations. Under these conditions, extracellular levan synthesized by the fructosyl polymerase activity of levansucrase can be degraded mainly into levanbiose by the action of LevB. Levanbiose is neither taken up nor metabolized by the bacteria. This work modifies the present view of the status of levansucrase inB. subtilisphysiology.

YfkN isolated from the culture supernatant of Bacillus subtilis in the exponential phase of growth is a protein of 143.5 kDa that derives from a putative large precursor of 159.6 kDa processed at both the N- and C-terminal ends. Pulse-chase experiments indicated that the release occurs slowly with a half-time longer than 30 min, suggesting that the event is coupled with wall turnover. YfkN exhibits 2',3' cyclic nucleotide phosphodiesterase, 2' (or 3') nucleotidase and 5' nucleotidase activities. In vitro the protein is reduced by subtilisin digestion to a shorter polypeptide (68 kDa), displaying phosphodiesterase activity but devoid of any 5'nucleotidase activity. This proteolytic processing led us to localize the potential active sites of the various nucleotidase activities. When bacteria were grown in low phosphate medium, the exocellular production of the enzyme was enhanced, suggesting that it plays a role in phosphate metabolism. Comparison with nucleotidase databases suggests that yfkN resulted from gene fusion.

The refolding intermediate of Bacillus subtilis α-amylase is prone to aggregate at 37°C and pH 7 when the protein concentration is relatively high (≥ 1µM). Low concentrations of 2,2,2 trifluoroethanol greatly increased the rate of aggregation. Aggregation made the folding intermediate resistant to proteases and there was kinetic competition between aggregation and proteolytic degradation. Analysis by Fourier transform infrared spectroscopy indicated that the secondary structure of the refolding intermediate is sightly different under soluble or aggregated states. Aggregates were readily solubilized by guanidium chloride (D1/2 = 1.25 M), but disaggregation was slow when aggregates were resuspended in solutions of various foreign native proteins under physiological conditions of pH and temperature. This destabilizing effect resulting from protein-protein interactions may make the aggregation process reversible in vivo.

When Bacillus subtilis levanase (SacC), alpha-amylase (AmyE) and chitosanase (Csn) structural genes were expressed under the regulated control of sacR, the inducible levansucrase (SacB) leader region in a degU32(Hy) mutant, it was observed that the production yields of the various extracellular proteins were quite different. This is mainly due to differences in the stabilities of their corresponding mRNAs which lead to discrepancies between the steady-state level of mRNA of sacB and csn on the one hand and amyE and sacC on the other. In contrast to levansucrase mRNA, the decay curves of alpha-amylase and levanase mRNAs obtained by Northern blotting analysis did not match the decay curves of their functional mRNA. This suggested that only a part of the population of the amyE and sacC transcripts was fully translated, while the others were possibly poorly bound to ribosomes and thus were only partially translated or not at all and consequently submitted to rapid endonuclease degradation. This hypothesis was substantiated by the finding that the introduction of a Shine-Dalgarno sequence upstream from the ribosome-binding site in the sacC transcript resulted in a fourfold increase in both the half-life of this transcript and the production of levanase. An additional cause of low-level levanase production is the premature release of mRNA by the polymerase. It was attempted to correlate this event with internal secondary structures of sacC mRNA.

SecA, the translocation ATPase of the preprotein translocase, accounts for 0.25% of the total protein in a degU32 (Hy) Bacillus subtilis strain in logarithmic phase. The SecA level remained constant irrespective of the demand for exoprotein production but dropped about 12-fold during the late stationary phase. Modulation of the level of functional SecA during the exponential phase of growth affected differently the secretion of levansucrase and α-amylase overexpressed under the control of the sacB leader region. The level of SecA was reduced in the presence of sodium azide and in the div341 thermosensitive mutant at nonpermissive temperatures. Overproduction of SecA was obtained with a multicopy plasmid bearing secA . The gradual decrease of the SecA level reduced the yield of secreted levansucrase with a concomitant accumulation of unprocessed precursor in the cells, while an increase in the SecA level resulted in an elevation of the production of exocellular levansucrase. In contrast, α-amylase secretion was almost unaffected by high concentrations of sodium azide or by very low levels of SecA. Secretion defects were apparent only under conditions of strong SecA deprivation of the cell. These data demonstrate that the α-amylase and levansucrase precursors markedly differ in their dependency on SecA for secretion. It is suggested that these precursors differ in their binding affinities for SecA.

The Bacillus subtilis levanase structural gene sacC was expressed under the regulated control of sacR, the inducible levansucrase leader region, in a degU32(Hy) strain. In this genetic context, exocellular levanase is overproduced (0.5% of total protein) during the exponential phase of growth upon induction by sucrose at 37 degrees C and pH 7. No precursor form that comprised a signal peptide was detected in pulse-chase experiments. The subsequent release of the cell-associated processed protein is a slow event (t(1,2) = 80+/-10 s). The unfolding-folding transition of pure levanase monitored in vitro by the resistance to proteolysis was achieved within the same time range (t(1/2) = 50 s) under the same conditions of pH and temperature. Calcium ions, which modulate the rate and the yield of refolding, have a low affinity for the protein. Comparison of these results with those obtained previously with levansucrase and alpha-amylase overproduced in the same genetic and physiological context suggests that the precursor processing is more efficient in levanase and alpha-amylase than in levansucrase. This discrepancy could lie in information borne by the signal peptide sequence of these exoproteins. However, the rate of the ultimate stage of release of these three proteins, which includes the passage through the cell wall, is correlated with the rate of folding and appears to be independent of their molecular size.

The Bacillus subtilis α-amylase gene, amyE, was expressed under the regulated control of sacR, the levansucrase leader region. The gene fusion including the complete amyE coding sequence with the signal peptide sequence was integrated into the chromosome of a degU32(Hy) strain deleted of the sacB DNA fragment. In this genetic context, α-amylase is produced in the culture supernatant at a high level (2% of total protein) during the exponential phase of growth upon induction by sucrose. Pulse-chase experiments showed that the rate-limiting step (t 1/2 = 120 s) of the secretion process is the release of a cell-associated precursor form whose signal peptide has been cleaved. The efficiency of this ultimate step of secretion decreased dramatically in the presence of a metal chelator (EDTA) or when the cells were converted to protoplasts. The hypothesis that this step is tightly coupled with the folding process of α-amylase occurring within the cell wall environment was substantiated by in vitro folding studies. The unfolding-folding transition, monitored by the resistance to proteolysis, was achieved within the same time range (t 1/2 = 60 s) and required the presence of calcium. This metal requirement could possibly be satisfied in vivo by the integrity of the cell wail. The t 1/2 of the α-amylase release step is double that of levansucrase, although their folding rates are similar. This perhaps indicates that the passage through the cell wall may depend on parietal properties (e.g. metal ion binding and porosity) and on certain intrinsic properties of the protein (molecular mass and folding properties).

We compared the ability of signal sequences from various Bacillus or yeast secreted proteins to direct Bacillus subtilis levansucrase into the secretion pathway of the yeast Saccharomyces cerevisiae. The efficiency of these sequences correlated with the overall hydrophobicity of their h-domain and was independent of their origin. Furthermore, the net charge of the proximal protein sequence downstream from the signal sequence contributed to the competence of the heterologous proteins to be secreted by yeast. Modification of this net charge allowed the protein to be translocated under the control of the yeast invertase signal sequence. Moreover, glycosylation of levansucrase did not modify significantly the fructosyl polymerase activity.

We investigated whether the concentration of H+ or metal ions such as Ca2+, or both, on the external side of the cytoplasmic membrane is involved in coupling of folding and secretion of Bacillus subtilis levansucrase by studying the modulation of each isolated event. In vitro at 30°C, in the absence of Ca2+, the equilibrium between the unfolded and the folded states of levansucrase was rapidly and totally displaced toward the folded state by a small pH shift from 7·4 to 6·0. Ca2+ (> 5 mM) acted as a catalyst of folding at pH ≥ 7. In vivo pulse-chase experiments at 30°C showed that, in the absence of Ca2+, the rate and the yield of the second step of levansucrase secretion were strongly dependent on the external pH. In acidic growth medium (pH 5·8), secretion was efficient. In contrast, at pH ≥ 7, the presence of Ca2+ was essential for secretion. In bacteria grown at high temperature (48°C), both external acidic pH and Ca2+ were required for efficient secretion. Moreover, a levansucrase variant altered in its calcium affinity was efficiently secreted only under acidic growth conditions. Depending on the culture conditions, the differences in H+ or Ca2+ concentrations which are maintained between the opposite sides of the energized cytoplasmic membrane could be adequate to catalyse conformational transition which could play a critical role in the second step of the protein release. These environmental parameters could also affect the yield of secretion of some other secretory proteins, leading to the hypothesis that several different secretion mechanisms could coexist in B. subtilis.

Levansucrase, a Bacillus subtilis extracellular enzyme, was not secreted in the culture medium when produced in yeast. The protein accumulated inside the cell in its precursor form which represented 0·3% of total proteins. The absence of any post-translational modifications, such as signal sequence cleavage or addition of N-linked sugars, indicated that this protein did not enter the reticulum secretion pathway. Direct observation of the cells by confocal laser scanning microscopy showed that levansucrase was associated with the cytoplasmic membrane. Subcellular fractionation experiments revealed that levansucrase precursor form is associated with membranes through weak ionic interactions. The purified precursor displayed the same catalytic properties as levansucrase secreted by B. subtilis. Thus yeast could be used as a source of levansucrase precursor allowing its isolation as a pure form on a milligram scale.

A synthetic oligodeoxynucleotide encoding the vasopressin peptide was ligated to the 3' terminal codon of sacB, the structural gene of levansucrase. This gene fusion was integrated into the chromosome of a Bacillus subtilis strain able to overproduce levansucrase. The extracellular production of the hybrid protein, consisting of the whole levansucrase primary sequence plus the nine amino acids of the vasopressin peptide added at the C-terminal end, represented 50-55% of that found for the wild-type levansucrase (20 mg l-1). The purified hybrid protein displayed the same conformational stability, protease insensitivity and enzymic properties as the wild-type levansucrase. However, the rate and the yield of the unfolding-folding transition at the pH and temperature used for bacterial growth were lower in the case of the hybrid protein; the latter also required a higher iron concentration to be completely folded.

The levansucrase (sucrose:2,6-beta-D-fructan 6-beta-D-fructosyltransferase, EC 2.4.1.10) structural gene from a Bacillus subtilis mutant strain displaying a low polymerase activity was sequenced. Only one missense mutation changing Arg331 to His was responsible for this modified catalytic property. From this allele we created new mutations by directed mutagenesis, which modified the charge and polarity of site 331. Examination of the kinetics of the purified levansucrase variants revealed that transfructosylation activities are affected differently by the substitution chosen. His331→Arg completely restored the properties of the wild-type enzyme. The most striking feature of the other variants, namely Lys331, Ser331 and Leu331, was that they lost the ability of the wild-type enzyme to synthesize levan from sucrose alone. They were only capable of catalysing the first step of levan chain elongation, which is the formation of the trisaccharide ketose. The variant His331→Lys presented a higher kcat. for sucrose hydrolysis than the wild-type, and only this hydrolase activity was preserved in a solvent/water mixture in which the wild-type acted as a true polymerase. The two other substitutions reduced the efficiency of transfructosylation activities of the enzyme via the decrease of the rate of fructosyl-enzyme intermediate formation. For all variants, the sucrose affinity was slightly affected. This strong modulation of the enzyme specificities from a single amino acid substitution led us to postulate the hypothesis that bacterial levansucrases and plant fructosyltransferases involved in fructan synthesis may possess a common ancestral form.

The refolding of levansucrase denatured by urea was studied as a possible model for the second step of the secretion pathway of this protein. The folding-unfolding transition was monitored by measuring intrinsic fluorescence and resistance to proteolysis. Both methods provided the same estimation for the unfolding free energy of levansucrase, delta GD, which was 30.1 +/- 1.7 kJ.mol-1 (7.2 +/- 0.4 kcal.mol-1) at pH 7 in 0.1 M-potassium phosphate buffer. The rate of refolding was greatly enhanced by Fe3+, whereas the Fe3+ chelator EDTA prevented correct refolding. Fe3+ allowed the protein to reach its folded form in medium in which the dielectric constant had been lowered by ethanol. The efficiency in vivo of the export of levansucrase bearing an amino acid modification which blocks the second step of the translocation pathway was greatly increased by high concentrations of Fe3+ in the culture medium. Assuming that the protein folding governs the second step of the secretion process of levansucrase, we discuss from an irreversible thermodynamic point of view the possible role of Fe3+ in the efficient coupling of the two events.

Aims: To develop a strategy to increase the stability of transcripts of structural genes expressed under the control of sacR, the leader region of Bacillus subtilis levansucrase gene. Methods and Results: Insertion of Shine Dalgarno like sequences in the 5′-untranslated sacR region controlling the expression of sacB. Depending on the number of stabilizing sequences inserted and the position of these sequences with respect to the translation start codon, it was observed that the mRNA stability and the final protein production could be increased or decreased. Conclusions: This mRNA stabilization can be used to increase exocellular protein production in the degU32 (Hy) mutant. Significance and Impact of the Study: This approach can be applied to the expression of heterologous genes of biotechnological interest.

We characterized the reversible folding—unfolding transition of Bacillus subtilis exocellular chitosanase from either thermal or urea denaturation of the protein. The transitions were monitored in each case by intrinsic fluorescence changes and resistance to proteolysis. Unfolding and refolding kinetics and differential scanning calorimetry analysis suggested a two-state equilibrium. The equilibrium between the folded and unfolded states was rapidly displaced towards the folded state in the presence of a low concentration of calcium (2—20mM). The binding titration curve indicated that chitosanase possesses one weak Ca2+-binding site (with an equilibrium affinity constant, KA, of 0.3×103M-1). These results support the hypothesis that this metal ion, which is accumulated in the cell wall environment of B. subtilis, is an effector that influences folding and stability of newly translocated proteins.

Inoculation of wheat roots with Paenibacillus (formerly Bacillus) polymyxa CF43 increases the mass of root-adhering soil. We tested the role of levan, a fructosyl polymer produced by strain CF43, in the aggregation of soil adhering to wheat roots. The P. polymyxa gene homologous to the Bacillus subtilis sacB gene encoding levansucrase was cloned and sequenced. The corresponding gene product synthesises high molecular weight levan. A P. polymyxa mutant strain, SB03, whose sacB gene is disrupted, was constructed using heterogramic conjugation. Effects of wheat inoculation with the wild type and the mutant strain were compared using two different cultivated silt loam soils in four independent pot experiments. Roots of wheat plantlets inoculated with CF43 or SB03 were colonized after 7–14 days at the same level, and root and shoot masses were not significantly different from those of the non-inoculated control plants. The ratio of root-adhering soil dry mass to root tissue dry mass was significantly higher for plants inoculated with strain CF43 than for those inoculated with mutant strain SB03: + 30% in Orgeval soil and + 100% in Dieulouard soil. Thus the levan produced by P. polymyxa is implicated in the aggregation of root-adhering soil on wheat.

A derivative of the α-amylase from Bacillus licheniformis (AmyL) engineered to give an active enzyme with increased net positive charge is secreted by Bacillus subtilis with a yield that is significantly lower than that of the native enzyme. This reduction in yield is the result of increased proteolysis during or shortly after translocation through the cytoplasmic membrane. When we compared the overall rate of folding of the engineered derivative (AmyLQS50.5) with that of AmyL it exhibited a greater dependency on Ca2+ ions for in vitro folding. When the concentration of Ca2+ in the growth medium was increased, so too did the relative yield of AmyLQS50.5. We discuss the importance of secretory protein folding at the membrane/cell wall interface with respect to the yield of native and heterologous proteins from B. subtilis.

Bacillus subtilis exocellular alpha-amylase is reversibly refolded after denaturation by guanidine hydrochloride at pH 7 and 37 degrees C. The unfolding-folding transition monitored by intrinsic fluorescence changes and resistance to proteolysis was resolved into a two-state transition. The first step (t1/21 s) led from D, the totally unfolded state, to C, a stable partially structured state of the protein. This folding intermediate was devoid of any enzyme activity and partially resistant to protease degradation. Calcium was required for the transition from C to N, the native state. This metal did not remain associated with the native form and could be replaced by barium or strontium, but not by magnesium. We discuss the hypothesis that C, the folding intermediate whose further transformation is under kinetic control, is the competent state involved in the secretion process of alpha-amylase.

The Acetobacter diazotrophicus SRT4 gene encoding levansucrase (EC 2.4.1.10) (IsdA) was isolated from a genomic library. The nucleotide sequence of a 2.3 kb DNA fragment sufficient for complementation of a levansucrase-deficient mutant (obtained by EMS treatment) was determined. The IsdA gene (1751 bp) coded for a polypeptide of molecular mass 64.9 kDa with an isoelectric point of 5.2. The N-terminal amino acid sequence of the extracellular levansucrase indicated the presence of a precursor protein with a putative signal sequence of 51 residues which is possibly cleaved in two successive steps. Expression of the IsdA gene from the lac promoter in Escherichia coli resulted in the production of a protein with levansucrase activity. The deduced amino acid sequence of the IsdA gene was 48% and 46% identical with the levansucrases from the Gram-negative bacteria Zymomonas mobilis and Erwinia amylovora, respectively, but only 28-31% identical with levansucrases from Gram-positive bacteria. Multiple alignments of published levansucrase sequences from Gram-negative and Gram-positive bacteria revealed eight conserved motifs. A comparison of the catalytic properties and the sequence of the A. diazotrophicus levansucrase with those of the Bacillus subtilis levansucrase suggested that one of these motifs may be involved in the specificity of the synthesized product. Disruption of the IsdA gene in the genome of A. diazotrophicus resulted in a mutant lacking both levansucrase activity and the ability to utilize sucrose as a carbon source, suggesting that levansucrase is the key enzyme in sucrose metabolism of A. diazotrophicus.

Immunoblot analysis of Bacillus subtilis cell extracts with polyclonal antibodies, raised against purified exocellular alpha-amylase, revealed one protein species of 82,000 Da. This protein was found even in cells in which the amyE gene, encoding exocellular alpha-amylase, was disrupted. Isolated from the membrane fraction, the 82,000-M(r) protein displayed an alpha-amylase activity in vitro.

Acetobacter diazotrophicus, a nitrogen-fixing bacterium associated with sugar cane, secretes a levansucrase (sucrose-2,6-beta-D-fructan 6-beta-D-fructosyltransferase; EC 2.4.1.10). This enzyme is constitutively expressed and represents more than 70% of the total proteins secreted by strain SRT4. The purified protein consists of a single 58 kDa polypeptide with an isoelectric point of 5.5. Its activity is optimal at pH 5.0. It catalyses transfructosylation from sucrose to a variety of acceptors including water (sucrose hydrolysis), glucose (exchange reaction), fructan (polymerase reaction) and sucrose (oligofructoside synthesis). In vivo the polymerase activity leads to synthesis of a high-molecular-mass fructan of the levan type. A. diazotrophicus levansucrase catalyses transfructosylation via a Ping Pong mechanism involving the formation of a transient fructosyl-enzyme intermediate. The catalytic mechanism is very similar to that of Bacillus subtilis levansucrase. The kinetic parameters of the two enzymes are of the same order of magnitude. The main difference between the two enzyme specificities is the high yield of oligofructoside, particularly 1-kestotriose and kestotetraose, accumulated by A. diazotrophicus levansucrase during sucrose transformation. We discuss the hypothesis that these catalytic features may serve the different biological functions of each enzyme.

Summary A weak Ca2+-binding site (Ka= 0.8× 103 M−1, at pH7) was identified in the mature part of levansucrase. An amino acid substitution (Thr-236 lle) in this site alters simultaneously the affinity for calcium, the folding transition and the efficiency of the secretion process of levansucrase. Moreover, the ability of the Bacillus subtilis cell wall to concentrate calcium ions present in the culture medium was studied. We confirm the results of Beveridge and Murray who showed that the concentration factor is about 100 to 120 times. This property preserves a high concentration of Ca2+ (>2 mM) on the external side of the cytoplasmic membrane, even in the absence of further Ca2+ supplementation in the growth medium. Such local conditions allow the spontaneous unfolding folding transition of levansucrase en route for secretion. Since several exocellular proteins of B. subtilis are calcium-binding proteins, we propose that the high concentration of calcium ion in the microenvironment of the cell wall may play a key role in the ultimate step of their secretion process.

The catalytic properties of levansucrase bound to hydroxyapatite were studied as a possible model for enzyme behaviour when associated in vivo to matrices such as the cell wall of bacteria or tooth surfaces. The activity of the immobilised enzyme was mainly directed towards its polymerase activity. The yield of levan reached 85%. The kcat of the enzyme for sucrose transformation was increased and the Km for this substrate was unmodified. These properties allow the design of a system for the large-scale production of high-molecular-weight branched-chain levan in vitro in high yield.

A significant modification of the transfructosylation specificities of Bacillus subtilis levansucrase was obtained by two methods. Firstly, in concentrated solution of organic solvents this enzyme displays only its polymerase activity. Secondly, amino acid substitutions of the catalytically essential residue Arg331 generate levansucrase variants which have lost the ability of the wild-type enzyme to synthesize levan from sucrose, but have preserved the sucrose hydrolytic activity and the ability to synthesize the trisaccharide kestose. The latter reaction was investigated using a Lys331-variant. The aim of the present study is to contribute to a better understanding of fructan synthesis in bacteria and to provide a way to investigate the hypothesis that the enzymes involved in fructan synthesis by bacteria and plants have a common ancestor.

It has been generally accepted that the structural sacB gene of Bacillus subtilis levansucrase encodes a 50 000 Da extracellular protein. However, examination of the DNA sequence of the sacB flanking regions shows a putative open reading frame coding for a 20 amino acid peptide downstream immediately following the terminal TAA stop codon. By site-directed mutagenesis we have changed this stop codon to a glutamine codon. This stop codon readthrough leads to the synthesis and secretion by B. subtilis of a levansucrase possessing an extended polypeptide chain. The extended levansucrase has a molecular weight of 53 000 with a new carboxyl-terminus, rich in basic and hydrophobic amino acids and possessing one cysteine residue. This enzyme synthesizes fructosyl polymer levan of higher molecular weight than the shorter levansucrase. The increase in molecular weight was achieved by increasing the number of branches. These results suggest that the C-terminal part of the enzyme plays a specific role in the degree of branching of the synthesized polymer. Moreover, the extended enzyme is able to form an active dimer from two polypeptide chains linked by an S-S bridge.

Studies of the equilibrium between native and denatured forms of wild-type levansucrase showed that the denatured form was predominant at 37 degrees C and pH 7 in the absence of free metal. The shift to the native form was promoted by metal ions such as Fe3+ or Ca2+. This metal-dependent refolding process was not observed in levansucrase variants bearing the amino acid substitution Gly-366----Asp or Gly-366----Val. These variants were only slightly secreted by Bacillus subtilis although their signal sequences were normally cleaved and their exocellular forms stable. In contrast, the Gly-366----Ser variant was secreted at near-normal levels and shared a part of the in vitro refolding properties of the wild-type protein. These differential properties might be related to the ability of the altered region to form a beta-turn structure. We discuss the possible role of metal ions in the coupling of protein folding and secretion.

Simple kinetic considerations derived from the ping-pong mechanism previously proposed for levansucrase of Bacillus subtillis allowed us to predict an easily operated method to approach the kinetic studies of exchange and hydrolytic activities of this enzyme. The experimental kinetic pattern obtained from the study of both activities is in close agreement with those predicted by theoretical approach. The combination of kinetic results enabled us to determine with a good accuracy the values of the apparent rate constant of the step of fructosylation of the enzyme from the sucrose-enzyme Michaelis complex and the apparent rate constants of the steps of defructosylation of the fructosyl enzyme to water or to glucose. The standard free energy reaction coordinate diagram for the transfructosylation process from sucrose to water was constructed. We found that the high energy of the glycosidic linkage of sucrose is preserved in the fructosyl-enzyme intermediate. The temperature dependence studies of the rate constants of fructosylation and defructosylation of the enzyme show that the entropy of activation for the two steps of defructosylation of the fructosyl enzyme are nearly the same. However the enthalpy of activation for the transfructosylation step to water is greater than that to glucose. We attempted to explain this discrepancy. Furthermore comparison of mechanism and efficiency of enzymatic and acid catalysis of sucrose hydrolysis was developed.

A Lewis acid-catalyzed condensation between 2,3,4,6-tetra- O -acetyl-1-thio-α- d -glucopyranose and 1,3,4,6-tetra- O -benzyl- d -fructofuranose in dichloromethane, followed by a l.c. separation of the deprotected resulting disaccharides, led to 1-thiosucrose (β- d -fructofuranosyl 1-thio-α- d -glucopyranoside), and α- d -fructofuranosyl 1-thio-α- d -glucopyranoside, in respective 15 and 22% overall yields. Similarly, 2,3,4,6-tetra- O -acetyl-1-thio-β- d -glucopyranose reacted with 1,3,4,6-tetra- O -benzyl- d -fructofuranose to give, in a 1:6 ratio, α- d -fructofuranosyl 1-thio-β- d -glucopyranoside (1-thioisosucrose) and its β,β-anomer, which were separated as their O -acetyl derivatives. 1-Thiosucrose is an inducer of the biosynthesis of levansucrase in Bacillus subtilis . It is a competitive inhibitor for this enzyme with a K i of 10m m and likewise for yeast invertase ( K i 20 m m ).

The equilibrium between the hydrolase and synthetase activities of levan-sucrase was determined by progressively substituting water with various organic solvents in the enzymic reaction medium. In the presence of high concentrations of these solvents, the enzyme displayed only synthetase activity. The levan obtained under such conditions had Mr ∼106 and presented a low molecular dispersity. In the presence of solvent, the Km values for sucrose and raffinose remained unchanged, but the kcat values were five times higher in comparison to the same constants determined for an aqueous medium.

A macromolecular complex including an inactive and protease-sensitive form of levansucrase was isolated from phenol treatment of Bacillus subtilis. This complex has a weak affinity for hydroxyapatite. Its molecular weight is evaluated at 300 000. This levansucrase form may be activated either by Triton X-100, above its critical micellar concentration, or by phospholipase C. The enzyme form associated with the complex is quite different from the exocellular levansucrase, which is protease-insensitive and detergent-insensitive, shows high affinity for hydroxyapatite and has a molecular weight of 54 000. Triton X-100 disaggregates this complex. The active levansucrase released becomes protease-insensitive and has the same apparent sedimentation coefficient as the exocellular enzyme. Possible relationships between the existence of this trapped complex and the secretion process of levansucrase are discussed.

Summary: The induction parameters of levansucrase synthesis were the same in Bacillus subtilis strain 168 Marburg and in a derivative, hyperproducing (sacU h) strain. However, only the hyperproducing strain showed an induction lag period. The kinetics of appearance of functional levansucrase mRNA was established. Strain 168 did not release levansucrase, but washing the cells with high ionic strength buffer released different proteins of which levansucrase represented 2%. In contrast, the great majority of levansucrase synthesized by the hyperproducer was released in a homogeneous form into the culture medium. In this case high ionic strength treatment caused the cells to release the remaining levansucrase but not other proteins. A Triton X-100 sensitive form of levansucrase was isolated by phenol treatment of the sacU h strain; this form was absent in strain 168. We suggest that the sacU gene product possibly controls the synthesis of cellular components with which levansucrase is associated and thus its release is normally prevented.

A covalently linked fructosyl-enzyme complex was isolated from a reaction mixture of enzyme and sucrose submitted to the quenching effect of a large decrease of the pH. The fructosyl-enzyme bond was shown to be stable under acidic and neutral conditions in the presence of high concentration of urea and of sodium dodecyl sulfate. This intermediate did not transfer at a measurable rate its fructosyl group to the usual fructosyl acceptors of the enzyme reaction under the usual conditions of enzyme activity. However stability measurements of the fructosyl-enzyme bond indicated a marked lability at pH values above 8.5. The apparent rate constant of the hydrolytic reaction of this bond evaluated under the standard state of molar concentration of hydroxide ion was of the same order of magnitude as the apparent rate constant of the hydrolytic reaction of the transient fructosyl-enzyme postulated from the kinetic analysis of levansucrase. Furthermore, nucleophilic agents like imidazole enhanced the hydrolytic reaction of the fructosyl-enzyme bond. Identification of the fructosyl binding site on the enzyme was accomplished by proteolytic hydrolysis of the trapped complex. Peptic digestion followed by pronase digestion released a fructosyl-aspartate compound that we have isolated in a high state of purity. The lability of the fructosyl-aspartate bond under mild alkaline conditions suggested that the fructosyl was linked through an ester bond involving the beta-carboxyl of the aspartate residue. Treatment of the trapped complex with cyanogen bromide released only one fructosylated peptide. The apparent molecular weight of this peptide was estimated to be lower than 10000.

Substitutions of an aspartate or an arginine residue for the glycine residue at position 366 of the mature part of Bacillus subtilis levansucrase were obtained by mutagenesis. Quantitative estimation from immunoblot analysis showed that the two transient membrane forms of the modified proteins were present in the membrane at the same level as that of the wild-type protein. The proteolytic processing, which was previously shown to be the first step of the levansucrase secretion process, was not affected in these modified proteins. Results from pulse-chase experiments showed that the half-times for secretion of the modified levansucrases into the culture medium were nearly the same as that of the wild-type protein, but the amount of the modified proteins secreted was significantly reduced. Purified samples of the modified enzymes were subtilisin insensitive and possessed enzyme activities very similar to those of the wild-type enzyme. The results suggest that the 366 site probably belongs to a functional domain of the protein which could play an important role in the second step of the levansucrase secretion process.

Cerulenin action was studied in Bacillus subtilis strains which overproduce several exocellular enzymes. Under conditions in which fatty acid synthesis is partially or completely inhibited by the drug cerulenin, we observed that cell growth and endocellular protein synthesis are scarcely affected during an initial phase, whereas the synthesis of two exocellular enzymes, levansucrase and α-amylase, is markedly reduced. However, careful investigation showed that fatty acid synthesis and exocellular protein production are not inhibited in the same range of cerulenin concentration. Moreover, at a drug concentration which completely inhibits fatty acid synthesis both the half-life of levansucrase messenger RNA and the lag time of the induction process of this enzyme in an inducible strain are unmodified. To account for these data we suggest that cerulenin inhibition of exoenzyme synthesis in these B. subtilis strains is the result of its physiochemical interaction with the membrane rather than the result of its interference with intracellular lipid synthesis. This hypothesis is substantiated by the results obtained from the study of the effects of two membrane-modifying agents on B. subtilis exocellular enzyme production: ethanol and dodecanoic acid. These compounds are not metabolized by the cells and they inhibit levansucrase synthesis. Dodecanoic acid affects levansucrase synthesis to the same extent as cerulenin and within the same concentration range but it does not affect the incorporation of acetate into lipids and is reversibly bound to the cells. A quantitative interpretation of the inhibition caused by these three compounds is proposed, based on the hypothesis that they act at the level of the membrane sites of exocellular enzyme synthesis.

Vibrational Raman spectroscopy has been used to study the conformation of the extracellular enzyme levansucrase. Secondary structure analysis by the method of Williams shows 51% β-sheets, 18% total α-helical structure (from 7 to 16%) and modifies the environment of the Trp and Phe residues and also more of the CC, CN, CS and CH3 groups.

Iodination of the levansucrase of Bacillus subtilis Levansucrase (β-d-2,6-fructan:d-glucose i-fructosyltransferase, EC 2.4.1.10) of Bacillus subtilis contains 17 residues of tyrosine per molecule of enzyme of mol. wt. 40 000. Progressive iodination of tyrosyl groups in alkaline medium leads to a concomitant loss of the enzymatic activity. The inactivation is complete when the protein binds 14–15 iodine atoms per mole at pH 9.4 and 28–30 at pH 8.5. Every tyrosyl residue is accessible to iodination. At pH 8.5 where the enzyme is normally active, levans of a low degree of polymerization protect against the inactivation by iodination, without interfering with the fixation of iodine.

We studied the interactions between conformers of exocellular alpha-amylase and small unilamellar vesicles (SUV) composed of the major membrane lipids of Bacillus subtilis under physiological conditions of pH, temperature and ionic strength. Using fluorescence spectroscopy, surface plasmon resonance (SPR) and phase separation, we show that the native alpha-amylase has no affinity for the SUV, whereas a partially folded form, displaying structural properties in common with the competent state for secretion, binds to the vesicles (KA approximately 10(5) M(-1)). This association prevented its subsequent folding. The complex was destabilized in the presence of PrsA, a major peripheric lipoprotein of B. subtilis which displays a strong affinity for SUV (KA approximately 1.5x10(8) M(-1)). Vesicles coated with PrsA lost their ability to bind the partially folded conformer. The approach in vitro, in which our aim was to mimic the last stage of alpha-amylase translocation, indicates that PrsA possibly helps, in vivo, the secreted protein to acquire its native conformation by modulating the interaction between the latter and the lipid polar heads on the trans side of the cytoplasmic membrane.

The equilibrium transition curves for thermal unfolding of levansucrase were established at several pH values. At pH 7 and within the temperature range of bacterial growth, the unfolded from is predominant. However, under such conditions, refolding is promoted by the only addition of Ca2+ or Fe3+. We propose that the tertiary structure flexibility of levansucrase plays a key role in its secretion process.

The reversible folding-unfolding transition of mature and precursor forms of Bacillus subtilis levansucrase were compared under physiological conditions of pH and temperature. The time constant of the folding reaction was not modified by the presence of the signal sequence and the precursor in the native form was slightly more resistant to the denaturing action of urea. However, the folding pathway could be different for each protein since a domain of the mature levansucrase underwent an independent transition which is not observed during the renaturation process of prelevansucrase.

SUMMARY: The kinetics of levansucrase secretion were examined in a strain of Bacillus subtilis which overproduces the enzyme (sacU h). Pulse-labelling experiments indicated that the second step of the levansucrase secretion process has the properties of a membrane active-transport system. This event appears to be directly linked to the influx of iron into the bacteria. The response of B. subtilis to the inducer of levansucrase synthesis was modulated by ferrichelators and the extent of the response varied with the nature of the ferrichelator. Ferridihydroxybenzoate markedly shortened the induction lag period. It is inferred from these data that iron occurs as a cofactor for components of the membrane sites of synthesis/secretion of B. subtilis levansucrase.

The rate of exocellular levansucrase synthesis in an overproducing (sacUh) strain of Bacillus subtilis was shown to be directly proportional to the amount of two different transient forms of this enzyme located within the membrane fraction of the cells. The apparent Mr of the larger membrane form was 53,000, and that of the smaller form 50,000; the half-life time of each form was estimated in vivo to be 4-6 s and 32-42 s, respectively. Ethanol treatment of the cells lead to the accumulation of the 53,000-Mr form which may represent 1.5% of total membrane proteins. This latter form, partially purified, was transformed in vitro into the 50,000-Mr form by the action of the Escherichia coli leader peptidase. These enzyme forms were quite different from the exocellular levansucrase since they showed a weak affinity for hydroxyapatite and needed complexed iron to display enzyme activity. Assuming the membrane forms were precursors of exocellular levansucrase, we propose a two-step mechanism for the secretion process of levansucrase. The number of exoprotein synthesis/secretion sites in a B. subtilis cell is estimated to 2.5 X 10(4).

The interaction of the ion channel polypeptide gramicidin A with the L-alpha-lysophosphatidylcholine micelles in a membrane state association (approximative molar ratio 1:9) was investigated by Raman spectroscopy. Studies were carried out over the spectral ranges of 700-1700 cm-1 and 2800-3100 cm-1 at 10 degrees C. The Raman spectrum of L-alpha-lysophosphatidylcholine micelles indicated a disordered structure of the lipid acyl chains by the high intensities of the gauche conformation vibrations. Changing from the micellar phase to the membrane state of association with gramicidin A, the intensities of all-trans stretching modes increased whereas the intensities of gauche conformation vibrations decreased, reflecting the emergence of ordered lipid chains. Hydrophobic interactions between the acyl chains and the polypeptide side chain residues were demonstrated. The absence of modifications in intensities of the very strong tryptophan vibrations in the complex spectrum indicated that, if the tryptophan-stacking interactions suggested by some authors exist, they are very weak ones.

To give some support to researchs presently in progress in this institute, on the sequence elucidation and the X-Ray pattern of the levansucrase of B. subtilis, some physical and chemical properties of this enzyme were carefully reexaminated. The results explicit and on some points rectify previous reports from this laboratory. The molecular weight was measured by three different methods: sedimentation equilibrium, SDS-gel electrophoresis, gel filtration. They give an average value of 54000 g. From this molecular weight and the value of the Stokes' radius, an estimate of the frictional ratio f/fo was calculated. These results provide some knowledge about the size and the shape of the molecule. They are consistent with the electronic microscopy observations obtained elsewhere. The amino acid composition was determined from the acid hydrolysate. The nature of the sulfur containing aminoacid was established by analysis of (35-S)-labelled levansucrase : neither cysteine nor cystine were found in the molecule. The methionine residues appear essentially under unoxidized form. One terminal residue was characterized by the dansylation method using (14-C)-labelled dansyl chloride. An explanation of the affinity of the levansucrase on hydroxyapatite was attempted.

The mechanism of action of levansucrase of Bacillus subtilis was investigated by initial velocity measurements of reactions catalyzed by this enzyme and initial velocity of isotopic exchange at equilibrium. Comparison of experimental results with theoretical results derived from various postulated mechanisms, supports a “ping-pong” mechanism, involving the intermediate participation of a fructosyl-enzyme. Rate constants for each step of this sequential mechanism were estimated.