The functional role of the C- and N-terminal domains was studied through the analysis of the biochemical properties of three deleted versions of inulosucrase from L. citreum (IslA) described in Figure 1. IslA2 consists of a 102 kDa fragment obtained after deletion of 551 amino acid from the C-terminus: the deleted region is homologous to the C-terminal region of alternansucrase (Asr) from L. mesenteroides NRRL B-1355. The second construct is a 80 kDa variant designated as IslA3 and obtained from IslA2 but deleting also the transition region located between the C-terminal region and the catalytic domain. Finally, IslA4 is a 64 kDa fragment of IslA constituted only by the catalytic domain, after total elimination of the C-terminal region and an almost total deletion of the variable region in the N-terminus (the first 209 of 309 aa). In all cases 16 kDa corresponding to thioredoxin fused to the N-terminal regions should be included.

All proteins were produced under the control of the induced arabinose promoter in E. coli, resulting in active enzymes able to produce polymer. We have already demonstrated that these regions are not essential for the catalytic activity 15, as has also been demonstrated for C-terminus truncated versions of inulosucrase from L. reuteri 16 and for Asr from L. mesenteroides NRRL B-1355 17, which retain their catalytic activity upon modification. The truncated versions are also less stable than the native enzyme. However, other consequences besides the lost of stability may result from domain acquisitions, such as changes in kinetic properties or reaction specificity.

IslA, as well as several other FTFs and GTFs is cell associated. In some FTFs it has been demonstrated the C-terminal region is responsible for anchoring the enzyme to the cell by means of the LPXTG motif 3. The cell associated FTF from Streptococcus salivarius which is devoid of motif LPXTG is released from the cells on exposure to sucrose. Through deletions within the C terminus of this enzyme, Rathsam and Jacques 18, implicated both the hydrophobic and the PGST-rich wall-associated domains in stabilizing the enzyme on the cell surface. In IslA, neither the LPXTG motif, nor the PGST motif is present. However, a blast analysis revealed a 26% identity of its C-terminal region to the cell wall binding region of amidase (Ami) from Lysteria monocytogenes. Ami contains 8 modules of repeat sequences designated as GW that serve to anchor the protein to lipotheicoic acids of the cell wall 19.

In order to determinate if the acquired C-terminal domain of IslA could be involved in cell wall anchoring, cells of L. citreum were produced in the absence of sucrose to avoid IslA induction. These cells were later contacted with native IslA and IslA3 (the last one is an IslA form deprived of the C-terminal region). After intensive washing, IslA was strongly retained and active in the cell surface, as demonstrated by successive activity assays and gel electrophoresis (Fig. 2a), while IslA3 was not retained (Fig. 2b), demonstrating that the C-terminus of IslA is involved in cell attachment.

We have already demonstrated the role of the C-terminal domain in DsrP, a GTF from L. mesenteroides IBT-PQ. In particular, the homology of DsrP and DsrE (a dextransucrase from L. mesenteroides NRRL B-1299) with the binding region (CW repeats) of autolysin (LytA) of Streptococcus pneumoniae and ToxA from Clostridium difficile lead us to conduct experiments that demonstrated the anchoring role of this domain in DsrP, as well as in DsrE 20. We also conducted similar experiments with partially purified Asr from L. mesenteroides NRRL B-1355, which is able to bind to non induced L. mesenteroides NRRL B-1355 (results not shown). Asr is also cell associated and binds to the cells both in the presence or absence of its polymer 21. Asr C-terminal region contains only a single well defined A repeat: nevertheless, this enzyme has its own seven distinctive repeat elements of nine amino acids in this region 22. Therefore, the C-terminal region acquisition of FTF may have been a mechanism to display and anchor the enzyme in the cell surface in order to produce a biofilm a common property of these microorganisms. Several structures are displayed in Leuconostoc spp. FTFs and GTFs to locate them in the cell surface as can be summarized in Table 1. However the actual mechanism of protein-cell interaction is still unknown. Further experiments are required to clarify this mechanism.

In a previous work, truncated versions were constructed in order to explore if the C-terminal domain was essential for activity 15. In this work, the same truncated versions were fused to a His tag and expressed under the ara promoter in order to produce and purify enough protein for characterization. Each gene fragment was amplified from islA cloned in plasmid pCR-TOPO 15 using the corresponding primers: IslA2 IS2reverso (CTAATTTAAATCGCGTGAAAAGCTAATGGC) and SPdirevecto (ACCATGGACG TGAATCAACCACTTTTAGCG); IslA3 ISE3rvEco (ATC CTC AGA ATT CAA TGC TAA TAA CTC AAC) and SP directo; IslA4 BproNae (GAA ATG ACT AGT GTG CCG GCG CTT ATA TC) and ISE3rvEco. The amplification products were cloned into the pBAD/Thio TOPO expression vector (Invitrogen, Calsbad, CA). E. coli strain TOP10 was used to transform the constructed plasmids and to express the truncated IslA truncated versions. Overnight cultures of the transformed strains, carried out at 37°C in 50 ml Luria-Bertani medium supplemented with 100 μg/ml ampicillin, were used as inoculum of 950 ml of the same medium and grown until an 0.6 OD_{600 nm} was reached. At this time, expression of the recombinant proteins was induced by addition of 0.02% (w/v) L-arabinose for IslA2 and 0.2% (w/v) for IslA3 and IslA4 and the temperature reduced to 23°C. Cells were harvested at 1.8 OD_{600 nm}.

E.coli cells were harvested by centrifugation (10 min, 4°C, 4600 g). The resulting pellet was washed twice with 50 mM pH 6.5 phosphate buffer. Afterwards, cells were suspended in 5 ml of the same buffer and broken at 900 psi in a French press. Cell debris was removed by centrifugation for 30 min at 4°C at 10000 g and the supernatant assayed for activity.

The enzymatic forms were purified by affinity chromatography through their His tags. A bed volume of 600 μl of Ni-nitroacetic acid (Ni-NTA) resin (Qiagen) was used to bind protein from 5 ml of cell extract. The resin was equilibrated with 3 ml of binding buffer (NaH₂PO₄ 50 mM, NaCl 300 mM, imidazole 10 mM) pH 7.5 for IslA2 and pH 7 for IslA3 and IslA4. The cell extract was diluted 1:1 with buffer binding and incubated for 1 h at 4°C with the equilibrated resin, followed by washing with 7 ml of the same buffer containing 30 mM imidazole. Finally the recombinant protein(s) were eluted with 2 ml of elution buffer (50 mM NaH₂PO₄, 300 mM NaCl, 250 mM imidazole). The proteins were dialized against 50 mM pH 6.5 phosphate buffer and 1 mM CaCl₂. The cell-associated IslA was extracted from L. citreum CW28 cells with 8 M urea at 25°C for 1 h with occasional gentle shaking as already described 31. The extract was then dialyzed against 50 mM pH 6.5 phosphate buffer, after centrifugation. With this procedure, IslA was obtained in a highly purified form. The purity of the enzyme and truncated forms was verified by SDS-PAGE 8%.

Initial reaction rates of IslA and truncated versions were measured at 30°C in 50 mM pH 6.5 phosphate buffer in the presence of 293 mM sucrose and 1 mM CaCl₂. The activity was measured by following of the reducing power released from sucrose by the 3,5-dinitrosalicylic acid method (DNS). One activity unit (U) is defined as the amount of enzyme that produces 1 μmol of glucose per minute. Specific activity is reported as U/mg of protein. The protein concentration was determined by the Bradford method 39, using the Bio-Rad reagent and BSA as standard. In a more specific assay, glucose and fructose were analyzed by HPLC in a Waters instrument equipped with a refraction index detector (Waters 410) and using a high performance carbohydrate cartridge (Waters) at 35°C and acetonitrile:water 75:25 as eluent at a flow rate of 1.4 mL/min.

IslA and truncated versions activity was assayed in the 20 to 40°C temperature range in 50 mM pH 6.5 phosphate buffer and 1 mM CaCl₂, while the effect of pH was determined in the 5.0 to 8.0 range in the same buffer. All the experiments were performed in triplicates.

Kinetics properties were studied through initial rate of reaction measurements carried out at pH 6.5 and 30°C in sucrose solutions ranging from 14.6 to 584.8 mM and containing 1 mM of CaCl_2. Samples of 50 μl were withdrawn after addition of the enzyme at 3 min time intervals and poured into 50 μl of DNS solution to stop the reaction and perform the reducing power assay. The data was processed using the Hills or the Michaelis-Menten equations. The transglycosylase and hydrolase activities of IslA and truncated versions were determined from the glucose and fructose concentrations measured by HPLC.

Ca²⁺ ions were depleted by addition of EDTA in amounts that were found dependent on the protein structure, as described in the results section. Accordingly, different concentrations of Ca²⁺ ions were used to restore the activity.

Fructan was produced with all the enzymes forms at 30°C in 50 mM pH 6.5 phosphate buffer containing 100 g/L sucrose and 1 mM CaCl₂. The polymer was precipitated with two volumes of ethanol, dialyzed against water, lyophilized and analyzed by ¹³C NMR. Inulin MW was analyzed by gel permeation chromatography in a Waters 600E HPLC system controller (Waters Corp. Milford, MA) equipped with a refractive index detector (Waters 410), using a serial set of Ultrahydrogel columns (UG 500 and linear) at 35°C with water as mobile phase at 0.9 mL/min.

Tertiary structure of the truncated forms was examined by Trp fluorescence assays on a Perkin Elmer LS-55 spectrofluorimeter. The proteins were purified and filtered and solutions prepared containing 0.018 mg/mL of IslA4; 0.02 mg/mL of IslA3 and IslA2; and 0.01 mg/mL of IslA. The proteins were excited at 280 nm and the fluorescence emission measured at 348 nm at 30°C. The secondary structure of the truncated versions was determined by Circular Dichroism (CD) from solutions containing 5.04 μM of IslA4 and IslA3 and 1.7 μM of IslA in 50 mM pH 6.5 phosphate buffer (CaCl₂ and/or EDTA were added according to the case). The solutions were placed in quartz cuvettes of 2 mm path length and CD spectra in the far UV region (190–250 nm) recorded on a Jasco J-715 spectropolarimeter at 25°C. All the spectra shown is the average of 3 scans recorded at a scanning rate of 20 nm/min. Spectra were corrected by subtracting appropriate buffer blanks and smoothed by noise reduction.

In order to demonstrate the role of the C-terminal domain in binding to L. citreum CW28 cells, we first produced non induced L. citreum CW28 cells in LM culture supplemented with glucose 2% (w/v) instead of sucrose, harvested at 5 OD_{600 nm} by centrifugation and washed twice with 50 mM pH 6.5 phosphate buffer. Cell protein was measured by the Lowry method 40. In a second step, 0.5 mg/ml of the purified proteins IslA and IslA3 (with and without the C-terminal domain respectively but retaining enzymatic activity) were incubated with 2.5 mg/ml final concentration of non induced L. citreum CW28 cells for 12 h at 4°C with gentle shaking. Afterwards, cells were separated by centrifugation at 4°C, 12000 g. The pellet was washed three times with 50 mM pH 6.5 phosphate buffer, and for both, supernatant and the pellet, the FTF activity was determined and SDS-PAGE 8% gels were performed.