Biochemical and molecular characterization of a levansucrase from Lactobacillus reuteri

S. A. F. T. van Hijum1,2, E. Szalowska1,2, M. J. E. C. van der Maarel1,3 and L. Dijkhuizen1,2

1 Centre for Carbohydrate Bioengineering, TNO-RUG, University of Groningen, PO Box 14, 9750 AA Haren, The Netherlands
2 Department of Microbiology, Groningen Biomolecular Sciences and Biotechnology Institute, University of Groningen, PO Box 14, 9750 AA Haren, The Netherlands
3 Innovative Ingredients and Products Department, TNO Nutrition and Food Research, Rouaanstraat 27, 9723 CC Groningen, The Netherlands

Correspondence
L. Dijkhuizen
L.Dijkhuizen{at}biol.rug.nl


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Lactobacillus reuteri strain 121 employs a fructosyltransferase (FTF) to synthesize a fructose polymer [a fructan of the levan type, with {beta}(2->6) linkages] from sucrose or raffinose. Purification of this FTF (a levansucrase), and identification of peptide amino acid sequences, allowed isolation of the first Lactobacillus levansucrase gene (lev), encoding a protein (Lev) consisting of 804 amino acids. Lev showed highest similarity with an inulosucrase of L. reuteri 121 [Inu; producing an inulin polymer with {beta}(2->1)-linked fructosyl units] and with FTFs from streptococci. Expression of lev in Escherichia coli resulted in an active FTF (Lev{Delta}773His) that produced the same levan polymer [with only 2–3 % {beta}(2->1->6) branching points] as L. reuteri 121 cells grown on raffinose. The low degree of branching of the L. reuteri levan is very different from bacterial levans known up to now, such as that of Streptococcus salivarius, having up to 30 % branches. Although Lev is unusual in showing a higher hydrolysis than transferase activity, significant amounts of levan polymer are produced both in vivo and in vitro. Lev is strongly dependent on Ca2+ ions for activity. Unique properties of L. reuteri Lev together with Inu are: (i) the presence of a C-terminal cell-wall-anchoring motif causing similar expression problems in Escherichia coli, (ii) a relatively high optimum temperature for activity for FTF enzymes, and (iii) at 50 °C, kinetics that are best described by the Hill equation.


Abbreviations: FTF, fructosyltransferase; HPSEC-MALLS, high-performance size-exclusion chromatography multi-angle laser light scattering

The GenBank accession number for the L. reuteri 121 gene encoding levansucrase (lev) and its flanking regions is AF465251.


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Lactic acid bacteria are Gram-positive, food-grade micro-organisms consisting of many genera, e.g. Lactococcus, Streptococcus and Lactobacillus. Lactic acid bacteria possessing the generally regarded as safe (GRAS) status find application in the production of food and feed (Lindgren & Dobrogosz, 1990). Lactic acid bacteria produce an abundant variety of polysaccharides (De Vuyst & Degeest, 1999), which may be used to develop a new generation of food-grade ingredients.

Lactobacillus polysaccharides and oligosaccharides are of special interest because they may contribute to human health due to their antitumour (De Roos & Katan, 2000), antiulcer (Oda et al., 1983), immunomodulating (Schiffrin et al., 1995), or cholesterol-lowering activity (Roberfroid, 1993). Moreover, some strains (e.g. Lactobacillus reuteri) have been designated as probiotics, i.e. they may have beneficial effects on the host by improving the properties of the indigenous population of gastrointestinal micro-organisms (Havenaar & Huis in 't Veld, 1992; Gibson et al., 1994).

Previously, it was reported that L. reuteri 121 cultivated on media containing sucrose produced large amounts of both a glucan and a fructan polymer (van Geel-Schutten et al., 1999). The fructan polymer was a levan containing {beta}(2->6)-linked fructosyl residues, with two major fractions in the estimated size distribution of 150 000 Da and larger than 2 000 000 Da (van Hijum et al., 2001; van Geel-Schutten et al., 1999).

Enzymes responsible for the synthesis of fructan polymers of the levan type are generally referred to as fructosyltransferases (FTF) or levansucrases (sucrose : 2,6-{beta}-D-fructan 6-{beta}-D-fructosyltransferase, EC 2.4.1.10). They catalyse the transfer of the fructosyl unit of sucrose to a number of acceptors including sucrose, water (resulting in hydrolysis) and fructan polymer. Levansucrases of Zymomonas mobilis and Bacillus species (Gunasekaran et al., 1995; Perez-Oseguera et al., 1996) have been studied in most detail. Levans are either linear or branched to various degrees at the C-1 position. The sizes of the bacterial levans vary from 20 kDa to several MDa. For lactic acid bacteria, fructan production by streptococci and several lactobacilli has been reported (Tieking et al., 2003). Streptococcus salivarius strains produce branched levan polymers [containing up to 30 % {beta}(2->1) branches] (Ebisu et al., 1975; Hancock et al., 1976; Simms et al., 1990) whereas Streptococcus mutans JC-2 produces a fructan of the inulin type consisting mainly of {beta}(2->1)-linked fructosyl units with 5 % {beta}(2->6) branches (Rosell & Birkhed, 1974; Ebisu et al., 1975). Recently, inulosucrase genes from lactic acid bacteria were reported in L. reuteri 121 (van Hijum et al., 2002) and Leuconostoc citreum (Olivares-Illana et al., 2002, 2003).

Recently we described the purification of the levansucrase protein responsible for levan formation in L. reuteri 121, and determination of amino acid sequences of peptide fragments (van Hijum et al., 2001). Here we report the isolation and characterization of the levansucrase gene from the same strain. The gene was expressed in E. coli and its enzyme product was characterized. Structural characterization of the levan produced by the purified recombinant enzyme showed that this levansucrase is responsible for levan synthesis by L. reuteri 121 cells grown on raffinose.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Strains, plasmids, media and growth conditions.
L. reuteri 121 (culture collection TNO Nutrition and Food Research, Zeist, The Netherlands) was grown anaerobically at 37 °C as described previously (van Hijum et al., 2002). E. coli Top10 (Invitrogen) was used as host for pCR-XL-TOPO (Invitrogen) and pBAD/myc-his C (Invitrogen) plasmids, which were used for cloning of inverse PCR products and expression of the ftf gene, respectively. E. coli strains were grown aerobically at 37 °C in Luria–Bertani (LB) medium (Ausubel et al., 1987), where appropriate supplemented with 50 µg ampicillin ml-1 or with 0·02 % (w/v) arabinose for induction of ftf genes. Agar plates were made by adding 1·5 % agar to the LB medium.

General molecular techniques.
L. reuteri total DNA was isolated according to Verhasselt et al. (1989) as modified by Nagy et al. (1995). General procedures for cloning, transformation, DNA manipulations and agarose gel electrophoresis were as described by van Hijum et al. (2002). DNA was amplified by PCR on a DNA Thermal Cycler 480 (Applied Biosystems). Pwo DNA polymerase (Roche Biochemicals) was used for standard PCRs and for construction of the expression plasmids. High-Fidelity DNA polymerase (Roche Biochemicals) was used for inverse PCR reactions. PCR oligonucleotides were purchased from Amersham Pharmacia Biotech. Southern hybridizations were performed as described by van Hijum et al. (2002). All methods were according to the manufacturer's instructions, unless otherwise stated.

Isolation of the levansucrase gene.
Based on the amino acid sequences (QVESNNYNGVAEVNTERQANGQI and VYSPLVSTLMASDEVE) of two peptide fragments of the L. reuteri 121 levansucrase (van Hijum et al., 2001), degenerate primers Deg1 and Deg2i were designed (Table 1). PCR with Pwo DNA polymerase, these primers, and total DNA of L. reuteri 121 yielded an amplification product of 1385 bp (Fig. 1, A), which was used to design primers for two inverse PCR steps: (i) N1i and N2, and (ii) C1i and C2 (Table 1). L. reuteri 121 chromosomal DNA was digested with HincII and ligated, yielding circular DNA molecules. PCR with the ligation product as template and diverging primers (i) N1i and N2 yielded an amplicon of 1544 bp (Fig. 1, B) and (ii) C1i and C2 yielded an amplicon of 1542 bp (Fig. 1, C). The 1542 bp fragment was used to design inverse PCR primers IPBrevi and IPAfor (Table 1). L. reuteri 121 chromosomal DNA was digested with HindIII and ligated. PCR with primers IPBrevi and IPAfor with the circular ligation product as template yielded an amplicon of 1700 bp (Fig. 1, D). In total, a fragment of 4570 bp of L. reuteri 121 genomic DNA was cloned and sequenced (Fig. 1).


View this table:
[in this window]
[in a new window]
 
Table 1. Primers used in this study

Degenerate bases are according to IUB codes (N, any base; W, A or T; S, C or G; Y, C or T). NcoI and BglII restriction sites are underlined and stop codons are shown in bold.

 


View larger version (9K):
[in this window]
[in a new window]
 
Fig. 1. Strategy used for the isolation of the L. reuteri 121 levansucrase gene.

 
Expression of the L. reuteri levansucrase gene in E. coli.
Four primer sets were designed for expression of the levansucrase gene in E. coli: (i) BADFTFN and BADFTFC1 (Table 1) giving the full-length mature levansucrase (Lev; residues 37 to 804; see Fig. 2), (ii) BADFTFN and BADFTFC (Table 1) yielding a full-length mature levansucrase with a C-terminal His tag (LevHis; residues 37 to 804; Fig. 2), (iii) BADFTFN and BADFTFdC1 (Table 1) giving a levansucrase truncated from the LPXTG motif at position 773 onwards (Lev{Delta}773; residues 37 to 773), and (iv) BADFTFN and BADFTFdC (Table 1) yielding a truncated levansucrase from the LPXTG motif at position 773 onwards with a C-terminal His tag (Lev{Delta}773His; residues 37 to 773). The four levansucrase gene derivatives started with an ATG codon (vector sequence) followed by the ftf gene sequence encoding the amino acids found in the N-terminus of the strain 121 purified mature levansucrase protein (van Hijum et al., 2001), starting at amino acid residue 37 (Fig. 2). PCR with L. reuteri genomic DNA (approx. 1 µg), Pwo DNA polymerase, and the primer sets, yielded the ftf gene derivatives flanked by NcoI and BglII restriction sites. Using the NcoI and BglII restriction sites, the amplicons were cloned into the expression vector pBAD/myc-his C. The resulting pBAD vectors were transformed to E. coli Top10 for expression studies. Correct construction of the plasmids was confirmed by nucleotide sequence analysis of both DNA strands.



View larger version (85K):
[in this window]
[in a new window]
 
Fig. 2. Alignments of amino acid sequences from L. reuteri 121 Lev, L. reuteri 121 Inu (AF459437), S. mutans FTF (M18954), S. salivarius FTF (L08445) and Bacillus subtilis SacB (X02730). Alignments of amino acid sequences were made with CLUSTALW 1.74 (Thompson et al., 1994) using a gap opening penalty of 30 and a gap extension penalty of 0·5. Amino acid groups are according to the Pam250 residue weight matrix (Altschul et al., 1990). Homologous regions of Lev with enzymes from the Glycoside Hydrolase families 68 (dotted) and 32 (dashed) are indicated with thick arrows. The N-terminal amino acid sequence of levansucrase purified from L. reuteri culture supernatants as reported previously (van Hijum et al., 2001) is shown above the sequence at position 37. Indicated with arrows are direct repeats of 14 (at position 86–127) and of 13 amino acids (at position 727–751). The C-terminal cell-wall-anchoring motifs of Lev and Inu are shown as follows: (i) a putative spacer region (underlined, position 666–759); (ii) an LPXTG motif (bold, position 773); (iii) a stretch of hydrophobic amino acids (underlined, position 781–800); and (iv) three positively charged amino acids KRH (bold, position 801).

 
Protein purification
(i) Preparation of cell-free extracts.
Cells of E. coli Top10 harbouring the ftf gene were grown overnight at 37 °C in 500 ml LB with 0·02 % (w/v) arabinose to an OD600 of approximately 1·5. Cell extracts were obtained by ultrasonication as described previously (van Hijum et al., 2002).

(ii) Nickel affinity purification.
Ni-NTA resin (500 µl; Qiagen) was used to bind protein from 26 ml cell extract (3·6 mg protein ml-1). The resin was washed with 5 ml demineralized water and 2·5 ml binding buffer (50 mM Na2HPO4/NaH2PO4, pH 8·0) prior to applying the cell extract. The suspension was gently shaken at 4 °C for 1 h. Unbound material was washed away with 2·5 ml binding buffer, and bound protein was eluted from the affinity resin with 2 ml binding buffer containing 200 mM imidazole and 1 mM {beta}-mercaptoethanol. The eluate was dialysed against phosphate buffer (5 mM, pH 8·0) and adjusted to a volume of 5 ml in Tris buffer (20 mM, pH 8·0).

(iii) Resource-Q column chromatography.
An anion-exchange column (Resource-Q; Amersham Pharmacia Biotech; 1 ml column volume; flow rate 1 ml min-1) was equilibrated with Tris buffer (20 mM, pH 8·0; A) and the sample (5 ml) was loaded on the column. The column was eluted with Tris buffer (20 mM, pH 8·0, 0·5 M NaCl; B) and eluted fractions, collected from 20 % B to 80 % B, were screened for levansucrase activity (glucose release from sucrose; see below). Positive fractions were run on SDS-PAGE and peak fractions containing one protein band were pooled (4 ml) and stored at 4 °C for further analysis.

Biochemical characterization of the recombinant levansucrase
(i) N-terminal amino acid sequencing.
This was performed as described previously (van Hijum et al., 2002).

(ii) Mass analysis.
Matrix-assisted laser desorption-ionization mass spectrometry was used to determine the protein molecular masses. The adjusted Ni-NTA eluate (5 µl; ~100 µg ml-1) was mixed with matrix (5 µl; 20 mg sinapinic acid ml-1 in acetonitrile/0·1 % trifluoroacetic acid; 40/60, v/v), and 2 µl of the mixture was dried on a target. Spectra were recorded on a TofSpec MALDI E and SE spectrometer (Micromass).

(iii) Levansucrase activity assays.
Sucrose conversion by levansucrase yields (a) fructose, which is (partly) built into the growing polymer, and (b) glucose, in a 1 : 1 ratio to the amount of sucrose converted. In control experiments the glucose formed reflected the total amount of sucrose utilized, since the residual sucrose (measured by hydrolysing sucrose with invertase and enzymically measuring the free glucose and fructose), fructan (measured by a mild 0·5 M trifluoroacetic acid hydrolysis followed by the enzymic detection of fructose), free glucose and free fructose formed added up to the amount of sucrose added to the reaction mixture (results not shown). Based on the above-mentioned experiments, the amount of glucose formed reflects the total amount of sucrose utilized by the enzyme (total activity). The amount of fructose formed is a measure of the hydrolytic activity of the enzyme (transfer of fructosyl units to water). The amount of glucose minus the amount of free fructose reflects the transferase activity (the transfer of fructosyl units to an acceptor other than water). Glucose and fructose were measured enzymically as described by van Hijum et al. (2001). Levansucrase activity was measured in a sodium acetate buffer (25 mM; pH 5·4) with 100 mM sucrose and 1 mM calcium chloride at 50 °C, unless stated otherwise. The optimal temperature and pH for L. reuteri 121 Lev (at 3 µg ml-1) total enzyme activity (glucose release from sucrose) were determined from 20 to 55 °C and pH 3·0 to 6·5 (from pH 5·5 to 6·5 a 25 mM MES buffer was used), respectively. One unit of enzyme activity is defined as the release of 1 µmol glucose or fructose min-1. All experiments were performed in triplicate and, where appropriate, the results are presented as the means±SEM. The ‘Sigma Plot’ program (version 4.0) was used for curve fitting of the data, either with the standard Michaelis–Menten formula: [y=(axx)/(c+x)], the three-parameter Hill formula: [y=(axx)b/(cb+xb)], or a Michaelis–Menten formula with a substrate inhibition constant: [y=(axx)/(c+x+(x2/d))]. In these formulae, y is the specific activity (U mg-1), x is the substrate concentration (mM sucrose), a is the Vmax (U mg-1), b is the Hill factor, c is the Km (mM sucrose; K50 in the case of Hill-type kinetics), and d is the substrate inhibition constant (mM sucrose).

(iv) Levansucrase activity assays in SDS-PAGE gels.
Protein (approx. 5 µg) was run in duplicate on SDS-PAGE. A duplicate part of the gel was stained with Coomassie brilliant blue to identify the position of the proteins in the gel. Protein was cut from the corresponding unstained duplicate part of the gel. To determine enzyme activity in gel slices, protein was renatured by adding a sodium acetate buffer containing sucrose and 0·5 % (v/v) Triton X-100, and incubated at 50 °C. Glucose and fructose formation in the samples were determined as described by van Hijum et al. (2001).

Fructan analysis
(i) Fructan production and purification.
Reaction products of FTF were produced at the optimal growth temperature of strain 121 (37 °C), by incubating the purified levansucrase in a sodium acetate buffer (25 mM, pH 5·4; 1 mM CaCl2) with 100 g sucrose l-1, at 37 °C for 16 h. For comparison, fructan produced by L. reuteri cells grown overnight on MRSr was used. Polymer was precipitated and cleaned as described by van Hijum et al. (2001).

(ii) Molecular mass and methylation analysis.
Polymer characteristics (i) molecular mass by HPSEC/MALLS and (ii) the fructose linkage type by methylation were determined as described by van Hijum et al. (2001).


   RESULTS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Nucleotide sequence analysis of an L. reuteri genomic fragment encoding a levansucrase
Using PCR and inverse PCR a fragment of 4570 bp was cloned from the genomic DNA of L. reuteri 121. On this fragment three open reading frames (ORFs; Fig. 1) were present. A putative protein encoded by ORF1 (804 amino acids) had a deduced molecular mass of 87 602 Da and a pI of 4·81. The deduced N-terminal amino acid sequence of ORF1 carried a putative signal peptide sequence of 36 amino acids, followed by a putative signal peptidase cleavage site, most likely located between amino acid 36 (alanine) and amino acid 37 (aspartic acid) (as determined by SignalP; http://genome.cbs.dtu.dk/services/SignalP/) (Nielsen et al., 1997). The amino acid sequence following amino acid 37 matched the complete N-terminal peptide sequence determined from the purified L. reuteri levansucrase, with the exception of the first amino acid (Fig. 2). Furthermore, the three amino acid sequences determined for the internal peptide fragments of the purified L. reuteri levansucrase enzyme (van Hijum et al., 2001) were present in the ORF1 deduced amino acid sequence. We therefore concluded that ORF1 (hereafter referred to as the lev gene) encodes the L. reuteri 121 levansucrase (Lev). The GenBank accession number for the L. reuteri 121 lev gene and its flanking regions is AF465251.

ORF1 contained a putative start codon (TTG, encoding a formylmethionine at position 1193), with a perfect Shine–Dalgarno ribosome-binding site (AGGAGG) 8 bp upstream. Furthermore, two putative promoter sequences could be identified, according to the consensus promoter sequences described for Lactobacillus genes (Pouwels & Leer, 1993): (i) 238 bp upstream of the formylmethionine the sequences TTGTAA (-35) and TATAAA (-10) with a spacer region of 11 nucleotides, (ii) 199 bp upstream of the formylmethionine the sequences TTGATA (-35) and TAATAAA (-10) with a spacer region of 12 nucleotides. A strong terminator hairpin structure ({Delta}G -22·6 kcal mol-1) was found between ORF1 (68 nucleotides downstream) and ORF3 (172 bp downstream). The hairpin comprised a stem of 18 bp and a loop of 11 unpaired bases.

BLAST searches (http://www.ncbi.nlm.nih.gov/blast/) with the deduced Lev amino acid sequence showed highest similarity with: L. reuteri 121 inulosucrase (Inu; AF459437; 56 % identity and 86 % similarity in 768 amino acids), S. mutans FTF (P11701; 48 % identity and 65 % similarity in 773 amino acids), and S. salivarius FTF (Q55242; 48 % identity and 66 % similarity in 735 amino acids). Lev contained the core regions of Glycoside Hydrolase family 68 of levansucrase and invertases (Fig. 2; 41 % identity and 55 % similarity in amino acid residues 187 to 640; Pfam entry at 02435; http://pfam.wustl.edu/) and family 32 of invertases, levanases and inulinases (Fig. 2; 24 % identity and 36 % similarity in amino acid residues 274 to 437; Smart entry at 00640; http://smart.embl-heidelberg.de/).

A striking feature of the Lev protein is the presence of direct repeats in the N- and C-terminal regions (Fig. 2). BLAST searches with the amino acid sequences of these repeats yielded no significant similarity with any known protein sequence. These repeats were not observed in the amino acid sequences of Inu and other FTFs from Gram-positive bacteria (Fig. 2) or FTFs from Gram-negative bacteria. The C-terminal amino acid sequence of the Lev protein contained a proline-rich putative spacer region (Fig. 2; 72 amino acids with 13 proline residues). Furthermore, a Gram-positive LPXTG cell-wall anchor was identified (Fig. 2; Pfam entry PF00746 at http://pfam.wustl.edu/).

The isolated DNA fragment also contained ORF2, encoding a putative protein of 272 amino acids (from ATG start codon at position 133; Fig. 1), and ORF3, encoding a putative protein of 134 amino acids (from ATG start codon at position 4299; Fig. 1). BLAST searches with the translated amino acid sequence of ORF2 showed highest similarity to hypothetical protein NMA1791 from Neisseria meningitidis (AL162757; 41 % identity and 60 % similarity in 263 amino acids). BLAST searches with the deduced amino acid sequence of ORF3 revealed highest similarity with a transposase from Lactobacillus casei (CAA05973; 74 % identity and 89 % similarity in 59 amino acids) and to transcription termination factor Rho from Neisseria gonorrhoeae (Q06447). Conserved domain database searches (http://www.ncbi.nlm.nih.gov/Structure/cdd/cdd.shtml) showed the C-terminal domain of the translated protein from ORF3 to have homology to Rve, an integrase core domain protein of HIV-1 (Pfam entry at 00665).

Recombinant enzyme expression and purification
Cell extracts of E. coli Top10 harbouring the four Lev derivatives (Lev, LevHis, Lev{Delta}773 and Lev{Delta}773His) clearly possessed sucrase activity (glucose release from sucrose) when incubated in a buffer with sucrose as substrate. The highest sucrase activity with all four constructs was observed when E. coli cells were incubated overnight with 0·02 % arabinose (approx. 11 000 U l-1 in the cell extracts). No activity was detected without arabinose induction. The Lev and LevHis proteins showed smearing on SDS-PAGE gels (results not shown), whereas distinct bands were observed with the Lev{Delta}773 and Lev{Delta}773His proteins on SDS-PAGE gels. Lev{Delta}773His was selected for further purification using the polyhistidine tag.

The Lev{Delta}773His protein was purified to homogeneity from E. coli cell extracts by two column chromatography steps (Table 2). The yield of protein after purification was relatively low due to loss of protein in the washing steps of the Ni-NTA column. In E. coli cell-free extracts and Ni-NTA fractions, a second, smaller and less abundant protein band was found next to the dominant protein band. The smaller band had an apparent size of 75 000 Da, smaller than the calculated molecular mass (84 676 Da) of Lev{Delta}773His. SDS-PAGE of cell extract, Ni-NTA and resource-Q fractions showed that the dominant protein band had an apparent size of 110 000 Da (results not shown), larger than the calculated molecular mass (84 676 Da) of Lev{Delta}773His. Similar sizes and ratios as the Lev{Delta}773His protein were observed for Lev{Delta}773. Mass spectrometry analysis of the adjusted Ni-NTA eluate showed that the protein running at 110 000 Da had a mass of 84 772 Da and that the protein running at 75 000 Da had a mass of 63 841 Da. The N-terminal amino acid sequence of the protein running at 110 000 (MDQVES) corresponded to the lev translated amino acid sequence starting at position 37 (Fig. 2). The N-terminal amino acid sequence of the protein running at 75 000 Da (MPATYTVDA) corresponded to the translated amino acid sequence of lev starting from an alternative start codon (ATG) at position 1877 (amino acid residue 229 in Fig. 2). The deduced molecular mass of the Lev protein variant translated from the alternative start codon was 63 891 Da, corresponding to the size of the smaller protein determined by mass spectrometry. An imperfect ribosome-binding site (AAGGAA; at position 1863) and no consensus promoter sequence could be identified. We conclude that the L. reuteri 121 lev gene contains a second start codon that is recognized by E. coli. Staining for sucrase activity in SDS-PAGE gels showed a clearly positive activity band for the Lev{Delta}773His protein, whereas the N-terminally truncated Lev protein showed no detectable activity.


View this table:
[in this window]
[in a new window]
 
Table 2. Purification of the Lev{Delta}773His protein from E. coli cell extracts

 
In vitro fructan production by Lev{Delta}773His and fructan analysis
Incubation of nickel-column-purified Lev{Delta}773His protein (53 U l-1) for 16 h at 37 °C with 100 g sucrose l-1 yielded a total amount of 1·4 g fructan l-1 with 18 g sucrose l-1 consumed. Methylation analyses of fructan produced by raffinose-grown cells of L. reuteri 121 revealed the presence of 98 % 1,3,4-tri-O-methylfructose units [{beta}(2->6) linkages] and 2 % 3,4-di-O-methylfructose units [{beta}(1->2->6) linked branchpoints]. Methylation analyses of fructan produced by recombinant Lev{Delta}773His protein revealed the presence of 97 % 1,3,4-tri-O-methylfructose units [{beta}(2->6) linkages] and 3 % 3,4-di-O-methylfructose units [{beta}(1->2->6) linked branchpoints]. Thus, both fructans were linear levans with only low amounts of {beta}(2->1->6) branching points (2 % and 3 %, respectively). HPSEC/MALLS elution profiles of both fructans (those of Lev{Delta}773His and raffinose-grown cells of L. reuteri 121) were also comparable, showing in both cases two major fractions with molecular masses of 20 000 Da (97 % w/w) and (3–4)x106 Da (3 % w/w).

Recombinant enzyme characterization
The purified Lev{Delta}773His enzyme showed highest activity (glucose release from sucrose; total activity) at 50 °C and around pH 4·5–5·5. The enzyme was almost inactive without addition of 1 mM Ca2+ (2·28±0·19 % residual activity); no cations other than calcium could restore enzyme activity. Lev enzyme activity was (almost) completely inhibited by Hg+, Fe3+ and Cu2+. Partial inhibition was observed with Fe2+ (Table 3).


View this table:
[in this window]
[in a new window]
 
Table 3. Effects of various compounds on Lev{Delta}773His enzyme activity in the presence of 1 mM CaCl2

The reaction was started by adding 5·0 µg enzyme ml-1. Lev{Delta}773His enzyme activity was measured as glucose release from sucrose; 100 % corresponds to 182 U mg-1.

 
Kinetic properties determined for Lev{Delta}773His were compared to those reported for FTFs from S. salivarius, Bacillus subtilis and Gluconacetobacter diazotrophicus (Table 4). Because for Lev the optimal temperature for the release of glucose from sucrose was 50 °C and the optimal temperature for growth of L. reuteri was 37 °C, kinetic constants were determined at both temperatures. At both temperatures, enzyme activity versus substrate concentration curves (Fig. 3) were determined, measuring the release of glucose (VG) and fructose (VF) from sucrose. Curve fitting was done for each of the three curves (VG, VF and VG-VF) presented in Fig. 3. Michaelis–Menten-type kinetics was observed for the hydrolysis reaction, with minor sucrose substrate inhibition. Both transferase and overall enzyme activity followed normal Michaelis–Menten kinetics at 37 °C but displayed kinetics best described by the Hill equation at 50 °C (with Hill factors of 0·51±0·07 and 0·84±0·1, respectively). The Kcat/Km quotients (hydrolysisF=10·4 mM s-1, and transferaseG-F=2·43 mM s-1) showed that at 37 °C, Lev favours the hydrolysis reaction over the transferase reaction (Table 4). Total and hydrolytic activity of the Lev enzyme increased with incubation temperature (reflected in the higher and values; Table 4).


View this table:
[in this window]
[in a new window]
 
Table 4. Comparison of the values of apparent kinetic constants for bacterial FTFs

The kinetic constants are Kcat, Km (K50 in the case of Hill-type kinetics), and Kcat/Km for: (a) formation of glucose (G; total enzyme activity), (b) formation of fructose (F; hydrolytic enzyme activity), and (c) glucose minus fructose (G-F; polymerase enzyme activity).

 


View larger version (11K):
[in this window]
[in a new window]
 
Fig. 3. V versus [S] relationship for Lev enzyme at 37 °C (a; 8·0 µg ml-1) and 50 °C (b; 2·0 µg ml-1). {bullet}, VG (total activity); {triangleup}, VF (hydrolytic activity); {triangledown}, VG-VF (transferase activity).

 

   DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Previous attempts to clone the L. reuteri 121 levansucrase gene failed when using degenerate primers based on conserved regions in FTFs of Gram-positive bacteria (van Hijum et al., 2002). The ftf gene isolated in this way turned out to encode an inulosucrase (Inu). In earlier work we purified the levansucrase responsible for levan synthesis in L. reuteri 121 (van Hijum et al., 2001). Here, we report the isolation of the corresponding gene (lev) and its expression in E. coli. The recombinant Lev{Delta}773His enzyme produced levan polymer that was identical to that of L. reuteri 121. Furthermore, we have shown that the levan polymers contain only 2–3 % {beta}(2->1->6) branching points. Although unlikely, it cannot be excluded that the relatively minor amount of the high-molecular-mass polymer contains a different percentage of branches than the low-molecular-mass polymer.

The C-terminal domain of Lev consists of a proline-rich region, an LPXTG motif, a stretch of hydrophobic residues, and finally three positively charged amino acids (Fig. 2). We reported a similar LPXTG motif and C-terminal topology for the inulosucrase from L. reuteri (van Hijum et al., 2002). No other FTFs have been reported to carry a C-terminal LPXTG motif. For the cell-wall-associated FTF of S. salivarius, however, the presence of a hydrophobic membrane-spanning region and a putative spacer region rich in serine and threonine residues was reported (Rathsam & Jacques, 1998). The presence of a C-terminally located cell-wall-anchoring motif suggests that the levansucrase protein in L. reuteri 121 is cell-wall-associated. This is in accordance with previous observations, showing that the L. reuteri levansucrase activity occurs cell-associated as well as a free supernatant protein (van Geel-Schutten et al., 1999; van Hijum et al., 2001).

The biochemical properties and the products formed by the Lev{Delta}773His enzyme and the levansucrase purified from L. reuteri (van Hijum et al., 2001) are comparable. Obviously, the C-terminal truncation of Lev from amino acid 773 onwards (Fig. 2) and the addition of a C-terminal Myc epitope and polyhistidine tag did not have significant effects on the products formed. The full-length recombinant Lev protein (LevHis), containing the membrane-spanning region, showed smearing on SDS-PAGE gels, suggesting that the C-terminal membrane-spanning region in E. coli interfered with protein expression or protein stability. Similar effects of the C-terminal domain were observed for the L. reuteri Inu (van Hijum et al., 2002). With the HPSEC/MALLS method two major fractions with molecular masses of 20 000 Da (97 %, w/w) and (3–4)x106 Da (3 %, w/w) were found for fructan polymers produced by both L. reuteri and Lev{Delta}773His. Earlier we reported molecular masses of 150 000 and more than 2x106 Da when using gel-filtration chromatography (van Hijum et al., 2001). The discrepancy between these results can be explained by an increased accuracy with the HPSEC/MALLS method (Blennow et al., 2001; Turquois & Gloria, 2000).

The estimated affinities of the Lev enzyme for sucrose in the various reactions are comparable to values found for other levansucrases. The Kcat values of Lev for the release of both glucose (=147±3 s-1) and fructose (=117±7 s-1) from sucrose are clearly higher than those of the S. salivarius FTF (=63·5±3·6 s-1 and =28·9±1·2 s-1, respectively). This corresponded with the high hydrolytic activity that we also observed for the purified L. reuteri levansucrase (van Hijum et al., 2001). The Lev enzyme apparently transfers the fructosyl unit of sucrose relatively efficiently to water. Nevertheless, the enzyme produced significant amounts of levan both in vivo (van Hijum et al., 2001) and in vitro (this study).

A striking feature of the Lev{Delta}773His protein is its high optimal temperature of 50 °C, and that at 50 °C a shift occurs from Michaelis–Menten to kinetics best described by the Hill equation. Speculatively, with increasing temperatures, the enzyme can use sucrose more efficiently as acceptor than (oligo) fructan molecules. Only for the L. reuteri 121 Inu has a temperature optimum of 50 °C been reported (van Hijum et al., 2002, 2003); other FTFs show lower optimal temperatures. Regardless of the optimal temperature, no Hill-type kinetics has to our knowledge been observed for FTFs previously, except for Inu (van Hijum et al., 2003). The Hill factors calculated from activities at 50 °C for the Lev{Delta}773His (total activity and transferase reactions) were lower than 1. This was also observed for the Inu enzyme (van Hijum et al., 2003) and indicates a negative cooperativity for these reactions. With Hill-type kinetics it is assumed that there is more than one binding site present in the enzyme and/or multimeric forms of the enzyme. For Hill-type kinetics, a positive cooperativity indicates a positive interaction of binding sites present in the enzyme and/or multimers. Alternatively a negative cooperativity indicates a negative interaction of enzyme binding sites and/or multimers. In FTFs, it is not known how many binding sites are present for substrate and product binding due to the lack of detailed structural protein information. Multimeric forms of FTFs were reported only for the levansucrase from Actinomyces viscosus T14 (Pabst et al., 1979). Thus, we cannot draw conclusions on the nature of the negative cooperativity suggested by the best-fit for the total and transferase activities at 50 °C observed in this L. reuteri Lev enzyme.

Two non-levan-producing mutants of L. reuteri 121 have been described (strains 35-5 and K24), isolated during continuous culture experiments (van Geel-Schutten et al., 1999). Under the growth conditions applied, levansucrase activity became lost in a few generations, suggesting that the lev gene in L. reuteri 121 is located on a transposable element, or on a plasmid. Interestingly, ORF3 (Fig. 1) shows strong homology to a transposase from a Lactobacillus casei strain (CAA05973). Transposable elements have been described for a number of lactic acid bacteria (Davidson et al., 1996). When comparing genomic maps of Lactococcus lactis MG1363 and Streptococcus thermophilus, a large number of inversions and translocations are present. These genomic rearrangements are partly attributed to the presence of mobile elements in the genomes of lactic acid bacteria such as transposons (Davidson et al., 1996). In view of the presence of ORF3 downstream of ORF1, the possible location of lev on a transposable element, flanked by recognition sequences for a transposase encoded by ORF3, warrants further investigation. No transposon insertion sequences could be identified in the DNA sequence flanking ORF1. Therefore, the mechanism of inactivation of the levansucrase activity in L. reuteri 121 mutants remains unclear.

This is believed to be the first report of the identification of a Lactobacillus levansucrase gene (lev) and the characterization of the recombinant protein. The L. reuteri 121 levansucrase is most closely related to L. reuteri Inu and to levansucrases of streptococci, based on biochemical characteristics and sequence homologies. L. reuteri 121 levan contains significantly lower amounts of {beta}(2->1) branches than levans produced by Streptococcus spp. The L. reuteri 121 levansucrase is unusual in displaying a relatively high rate of sucrose hydrolysis. The lev gene was successfully expressed in E. coli, enabling production of relatively larger amounts of levansucrase and its levan polymer. Our current studies focus on a detailed biochemical and structural characterization of the L. reuteri Lev and Inu enzymes, to identify features that determine (i) the percentage of {beta}(2->1->6) branches, (ii) product size, (iii) the {beta}(2->1) versus {beta}(2->6) product specificity, and (iv) hydrolysis versus transglycosylation specificity.


   ACKNOWLEDGEMENTS
 
We thank Slavko Kralj and Monica Dondorff for their contributions to the isolation of the 3' end of the levansucrase gene, Isabelle Capron for HPSEC/MALLS analyses, Elly Faber for methylation and NMR analyses, Dennis Claessen for help with mass spectrometry analyses, Ineke van Geel-Schutten for contributions to the scientific discussions, and Rob Leer for aiding with the nucleotide sequence analyses. This project was partly financed by the EET programme of the Dutch government (project number KT 97029), by TNO Nutrition and Food Research, and by the University of Groningen.


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Altschul, S. F., Gish, W., Miller, W., Myers, E. W. & Lipman, D. J. (1990). Basic local alignment search tool. J Mol Biol 215, 403–410.[CrossRef][Medline]

Ausubel, F. M., Brent, B., Kingston, R. E., Moore, D. D., Seidman, J. G., Smith, J. A. & Struhl, K. (1987). Current Protocols in Molecular Biology. New York: Wiley.

Blennow, A., Mette Bay-Smidt, A. & Bauer, R. (2001). Amylopectin aggregation as a function of starch phosphate content studied by size exclusion chromatography and on-line refractive index and light scattering. Int J Biol Macromol 28, 409–420.[CrossRef][Medline]

Chambert, R. & Petit-Glatron, M. F. (1991). Polymerase and hydrolase activities of Bacillus subtilis levansucrase can be separately modulated by site-directed mutagenesis. Biochem J 279, 35–41.[Medline]

Davidson, B. E., Kordias, N., Dobos, M. & Hillier, A. J. (1996). Genomic organization of lactic acid bacteria. Antonie van Leeuwenhoek 70, 161–183.[Medline]

De Roos, N. M. & Katan, M. B. (2000). Effects of probiotic bacteria on diarrhea, lipid metabolism, and carcinogenesis: a review of papers published between 1988 and 1998. Am J Clin Nutr 71, 405–411.[Abstract/Free Full Text]

De Vuyst, L. & Degeest, B. (1999). Heteropolysaccharides from lactic acid bacteria. FEMS Microbiol Rev 23, 153–177.[CrossRef][Medline]

Ebisu, S., Kato, K., Kotani, S. & Misaki, A. (1975). Structural differences in fructans elaborated by Streptococcus mutans and Streptococcus salivarius. J Biochem 78, 879–887.[Abstract]

Gibson, G. R., Willis, C. L. & van Loo, J. (1994). Non-digestible oligosaccharides and bifidobacteria – implications for health. Int Sugar J 96, 381–387.

Gunasekaran, P., Mukundan, G., Kannan, R., Velmurugan, S., Ait-Abdelkader, N., Alvarez-Macarie, E. & Baratti, J. (1995). The sacB and sacC genes encoding levansucrase and sucrase form a gene cluster in Zymomonas mobilis. Biotechnol Lett 6, 635–642.

Hancock, R. A., Marshall, K. & Weigel, H. (1976). Structure of the levan elaborated by Streptococcus salivarius strain 51: an application of chemical-ionisation mass-spectrometry. Carbohydr Res 49, 351–360.[CrossRef][Medline]

Havenaar, R. & Huis in 't Veld, J. H. J. (1992). Probiotics: a general view. In The Lactic Acid Bacteria in Health and Disease, pp. 209–224. Edited by B. J. B. Wood. New York: Elsevier.

Hernández, L., Arrieta, J., Menéndez, C., Vazquez, R., Coego, A., Suarez, V., Selman, G., Petit-Glatron, M. F. & Chambert, R. (1995). Isolation and enzymic properties of levansucrase secreted by Acetobacter diazotrophicus SRT4, a bacterium associated with sugar cane. Biochem J 309, 113–118.[Medline]

Lindgren, S. E. & Dobrogosz, W. J. (1990). Antagonistic activities of lactic acid bacteria in food and feed fermentations. FEMS Microbiol Rev 7, 149–163.[Medline]

Nagy, I., Schoofs, G., Compernolle, F., Proost, P., Vanderleyden, J. & De Mot, R. (1995). Degradation of the thiocarbamate herbicide EPTC (S-ethyl dipropylcarbamothioate) and biosafening by Rhodococcus sp. strain NI86/21 involve an inducible cytochrome P-450 system and aldehyde dehydrogenase. J Bacteriol 177, 676–687.[Abstract]

Nielsen, H., Engelbrecht, J., Brunak, S. & Von Heijne, G. (1997). Identification of prokaryotic and eukaryotic signal peptides and prediction of their cleavage sites. Protein Eng 10, 1–6.[CrossRef]

Oda, M., Hasegawa, H., Komatsu, S., Kambe, M. & Tsuchiya, F. (1983). Antitumour polysaccharide from Lactobacillus sp. Agric Biol Chem 47, 1623–1625.

Olivares-Illana, V., Wacher-Rodarte, C., Le Borgne, S. & López-Munguía, A. (2002). Characterization of a cell-associated inulosucrase from a novel source: a Leuconostoc citreum strain isolated from Pozol, a fermented corn beverage from Mayan origin. J Ind Microbiol Biotechnol 28, 112–117.[CrossRef][Medline]

Olivares-Illana, V., Lopez-Munguia, A. & Olvera, C. (2003). Molecular characterization of inulosucrase from Leuconostoc citreum: a fructosyltransferase within a glucosyltransferase. J Bacteriol 185, 3606–3612.[Abstract/Free Full Text]

Pabst, M. J., Cisar, J. O. & Trummel, C. L. (1979). The cell wall-associated levansucrase of Actinomyces viscosus. Biochim Biophys Acta 566, 274–282.[Medline]

Perez-Oseguera, M. A., Guereca, L. & Lopez-Munguia, A. (1996). Properties of levansucrase from Bacillus circulans. Appl Microbiol Biotechnol 45, 465–471.[CrossRef]

Pouwels, P. H. & Leer, R. J. (1993). Genetics of lactobacilli: plasmids and gene expression. Antonie van Leeuwenhoek 64, 85–107.[Medline]

Rathsam, C. & Jacques, N. A. (1998). Role of C-terminal domains in surface attachment of the fructosyltransferase of Streptococcus salivarius ATCC 25975. J Bacteriol 180, 6400–6403.[Abstract/Free Full Text]

Roberfroid, M. R. (1993). Dietary fiber, inulin, and oligofructose: a review comparing their physiological effects. Crit Rev Food Sci Nutr 33, 103–148 [erratum in Crit Rev Food Sci Nutr 33, 553].

Rosell, K. G. & Birkhed, D. (1974). An inulin-like fructan produced by Streptococcus mutans strain JC2. Acta Chem Scand B28, 589.

Schiffrin, E. J., Rochat, F., Link-Amster, H., Aeschlimann, J. M. & Donnet-Hughes, A. (1995). Immunomodulation of human blood cells following the ingestion of lactic acid bacteria. J Dairy Sci 78, 491–497.[Abstract/Free Full Text]

Simms, P. J., Boyko, W. J. & Edwards, J. R. (1990). The structural analysis of a levan produced by Streptococcus salivarius SS2. Carbohydr Res 208, 193–198.[CrossRef][Medline]

Song, D. D. & Jacques, N. A. (1999). Purification and enzymic properties of the fructosyltransferase of Streptococcus salivarius ATCC 25975. Biochem J 341, 285–291.[CrossRef][Medline]

Thompson, J. D., Higgins, D. G. & Gibson, T. J. (1994). CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res 22, 4673–4680.[Abstract]

Tieking, M., Korakli, M., Ehrmann, M. A., Ganzle, M. G. & Vogel, R. F. (2003). In situ production of exopolysaccharides during sourdough fermentation by cereal and intestinal isolates of lactic acid bacteria. Appl Environ Microbiol 69, 945–952.[Abstract/Free Full Text]

Turquois, T. & Gloria, H. (2000). Determination of the absolute molecular weight averages and molecular weight distributions of alginates used as ice cream stabilizers by using multiangle laser light scattering measurements. J Agric Food Chem 48, 5455–5458.[CrossRef][Medline]

van Geel-Schutten, G. H., Faber, E. J., Smit, E., Bonting, K., Smith, M. R., Ten Brink, B., Kamerling, J. P., Vliegenthart, J. F. G. & Dijkhuizen, L. (1999). Biochemical and structural characterization of the glucan and fructan exopolysaccharides synthesized by the Lactobacillus reuteri wild-type strain and by mutant strains. Appl Environ Microbiol 65, 3008–3014.[Abstract/Free Full Text]

van Hijum, S. A. F. T., Bonting, K., van der Maarel, M. J. E. C. & Dijkhuizen, L. (2001). Purification of a novel fructosyltransferase from Lactobacillus reuteri strain 121 and characterization of the levan produced. FEMS Microbiol Lett 205, 323–328.[CrossRef][Medline]

van Hijum, S. A. F. T., van Geel-Schutten, G. H., Rahaoui, H., van der Maarel, M. J. & Dijkhuizen, L. (2002). Characterization of a novel fructosyltransferase from Lactobacillus reuteri that synthesizes high-molecular-weight inulin and inulin oligosaccharides. Appl Environ Microbiol 68, 4390–4398.[Abstract/Free Full Text]

van Hijum, S. A. F. T., van der Maarel, M. J. & Dijkhuizen, L. (2003). Kinetic properties of an inulosucrase from Lactobacillus reuteri 121. FEBS Lett 534, 207–210.[CrossRef][Medline]

Verhasselt, P., Poncelet, F., Vits, K., van Gool, A. & Vanderleyden, J. (1989). Cloning and expression of a Clostridium acetobutylicum alpha-amylase gene in Escherichia coli. FEMS Microbiol Lett 50, 135–140.[CrossRef][Medline]

Received 28 July 2003; revised 31 October 2003; accepted 1 December 2003.