The Cyclodextrin Glycosyltransferase of Paenibacillus pabuli US132 Strain: Molecular Characterization and Overproduction of the Recombinant Enzyme

The gene encoding the cyclodextrin glycosyltransferase (CGTase) of Paenibacillus pabuli US132, previously described as efficient antistaling agent and good candidate for cyclodextrins production, was cloned, sequenced, and expressed in Escherichia coli. Sequence analysis showed that the mature enzyme (684 amino acids) was preceded by a signal peptide of 34 residues. The enzyme exhibited the highest identity (94%) to the β-CGTase of Bacillus circulans no. 8. The production of the recombinant CGTase, as active form, was very low (about 1 U/mL) in shake flasks at 37°C. This production reached 22 U/mL after 22 hours of induction by mainly shifting the postinduction temperature from 37 to 19°C and using 2TY instead of LB medium. High enzyme production (35 U/mL) was attained after 18 hours of induction in fermentor using the same culture conditions as in shake flask. The recombinant enzyme showed Vmax and Km values of 253 ± 36 μmol of β-cyclodextrin/mg/min and 0.36 ± 0.18 g/L, respectively.

CGTases belong to the glycoside hydrolase family 13 (α-amylase family) [3]. Members of this family contain a catalytic (β/α) 8 barrel and use an α-retaining mechanism but they display a variety of reaction specificities [4]. Whereas α-amylases typically catalyze starch hydrolysis reaction, CGTases catalyze mainly three transglycosylation reactions (cyclization, coupling, and disproportionation) besides a weak hydrolytic activity [5]. Both α-amylases and CGTases share three domains (A, B, and C) while CGTases possess two additional domains (D and E). Domains A and B form the catalytic core, domains C and E are involved in substrate binding whereas domain D has unknown precise function [6,7].
Understanding the enzymatic features has been explored by cloning, sequencing, and comparing various cgtase genes [8][9][10][11]. The sequence data analysis has revealed the key amino acids residues that determine the reaction mechanism and product specificity, thereby enhancing genetic engineering techniques to provide modified CGTases [7,12,13].
E. coli is frequently adopted as host strain for heterologous protein expression since it is well characterised with various expression systems. However, the formation of inclusion bodies remains a significant barrier for expression of heterologous proteins in E. coli. Refolding of inclusion bodies into soluble and active form needs high cost and tedious jobs. Hence, maximizing the yield of soluble and active recombinant proteins in vivo by altering the culture conditions is an attractive alternative [14]. However, no universal approach has been established to minimize the formation of inclusion bodies and some empirical conditions must be screened on an individual basis. To date, there are only a few reports that succeeded the overproduction of 2 Journal of Biomedicine and Biotechnology CGTase in E. coli by varying fermentation conditions [15][16][17].
In a previous study, the CGTase of Paenibacillus pabuli US132 was reported as good candidate for cyclodextrins production and efficient antistaling agent [18]. In this work, we described the molecular cloning of the US132 cgtase gene as well as the amino acid sequence inspection and the comparison with other related CGTases. We also reported the production enhancement of the recombinant active enzyme in E. coli by varying mainly the cultivation conditions.

Cloning of the US132 cgtase gene and sequence analysis
To amplify an internal cgtase gene fragment, we used the highly conserved aa regions previously reported for CGTase enzymes as members of the α-amylase family [4,21]. Furthermore, we aligned several cgtases DNA sequences and we located the corresponding CGTases conserved regions. Hence, a degenerate pair of primers, S209 (5 -GAYTTTGCRCCCAAYCAT-3 ) and S211 (5 -ATCATGATTGTCRATRAA-3 ), corresponding respectively to the conserved regions I (DFAPNH) and IV (FIDNHD), were designed.

Heterologous expression and recombinant US132 CGTase purification
The heterologous expression of the US132 CGTase was investigated using different recombinant strains (DH5α/pSJ8, ER2566/pSJ8, DH5α/pSJ9, ER2566/pSJ9 or JM109/pSJ9), various temperature cultivation (37 • C or 19 • C during all operating time and 37 • C followed by a shift to 19 • C after IPTG induction) and different medium composition (LB, M9, 2TY or M9ZB). An overnight culture of the studied strain, cultivated at 37 • C, was used to inoculate 50 mL of basal medium (in 250 mL shake flask) with an initial OD 600 nm of 0.1. Protein expression was induced by addition of different IPTG concentration (16-240 μg/mL) at OD 600 nm between 0.8 and 1. Effects of the postinduction time on the US132 CGTase production were also investigated. All experiments were performed at least twice.
Batch fermentation was carried out in a 7 L fermentor INFORS AG CH-4103 (Bottmingen, Switzerland) containing 4.5 L of 2TY medium. The fermentor was inoculated, with an initial OD 600 nm of 0.1, by an overnight culture of the ER2566/pSJ8 strain grown at 37 • C in 2TY broth. The pH, aeration and agitation were maintained constant at 7.4, 1.5 vvm and 500 rpm, respectively, during all cultivation. The temperature was shifted from 37 to 19 • C after induction with 16 mg/L of IPTG when OD 600 nm reached 0.8-1.
The CGTase extract was prepared from the periplasm by the modified osmotic-shock procedure of Ausubel et al. [22]. Cells were harvested by centrifugation at 8000 × g for 10 minutes and the pellets were suspended in 30 mM Tris-HCl buffer (pH 8) containing 20% sucrose and 1 mM EDTA. After agitation for 10 minutes at 25 • C, cells removed by centrifugation were rapidly suspended in ice-cold water, incubated for 10 minutes at 0 • C and clarified by centrifugation at 12000 × g for 30 minutes. The supernatant was recovered as periplasmic fraction and used for CGTase assay. Residual cells were sonicated at 4 • C for 6 minutes (pulsations 3 s, amplify 90) using a Vibra-Cell 72405 Sonicator and insoluble fraction (cells debris) was recuperated by centrifugation at 12000 × g for 30 minutes.
For the purification of the recombinant US132 CGTase, the ER2566/pSJ8 periplasmic fraction was heat-treated at 60 • C for 15 minutes in presence of 10 mM calcium followed by centrifugation at 16000 × g for 30 minutes at 4 • C. The supernatant was then purified using hydrophobic interaction chromatography and starch adsorption as previously described for the native enzyme [18].

CGTase assay and kinetic parameters determination
The CGTase activity, determined as dextrinisation activity, was monitored at 60 • C for 10 min as described in a previous work [18]. The kinetic parameters (K m and V max ) were determined by incubating the purified CGTase at 60 • C in 50 mM sodium acetate buffer (pH 6.5) and using various concentrations of soluble starch ranging from 0.4 to 1.6 g/L. Samples were taken at regular time intervals and the reaction was stopped by boiling for 5 minutes. The amount of generated β-CDs was detected according to a phenolphthalein method [23]. The initial speed was calculated for each substrate concentration and then represented according to the Lineweaver-Burk method [24]. The K m and V max were determined graphically using Hyper32 program (http:// homepage.ntlworld.com/john.easterby/hyper32.html).

Identification of the P. pabuli US132 cgtase gene
To clone the US132 cgtase gene, an internal PCR fragment of about 600 bp was firstly amplified using US132 genomic DNA and two primers corresponding to the conserved regions I and IV in CGTases as described in Materials and Methods section. The nucleotide sequence of this internal fragment (582 bp) showed the highest homology (95% identity) with homologous region of the putative Bacillus Q cgtase gene (accession no. L25256). This result incited us to conceive two primers from the 5 and 3 flanking regions of the Bacillus Q cgtase gene in order to amplify the entire US132 cgtase gene using the chromosomal DNA of P. pabuli US132 as template. The PCR amplification, using Pfu DNA polymerase, gave a nucleotide fragment of approximately 2400 bp. The cloning of this fragment in pCR2.1 vector, under the control of the T7 promoter, provided pSJ8 plasmid and conferred starch degrading activity for E. coli host strains.
The nucleotide sequence analysis of the fragment containing the US132 cgtase gene revealed the presence of a single open reading frame (ORF) with two potential initiation ATG codons ( Figure 1). A putative Shine-Dalgarno site (5 -AGAAGGGTGG-3 ), exhibiting a good rationally complementarity with the 3 end of the US132 16S rRNA gene [18], was 7 bases upstream the first ATG codon. Consequently, this latter is most likely the true initiator codon. Thus, the US132 cgtase ORF, consisted of 2157 bp, encoded a protein having 718 amino acids. By using the promoter prediction program, a potential −10 region (5 -TCTGCA-3 ) and −35 region (5 -TTGGCG-3 ) were found at 101 and 124 bp upstream the initiator codon, respectively. The first 34 amino acids are considered as signal peptide according to the SignalP server (http:/www.cbs.dtu.dk/services/signalP). Therefore, the mature enzyme was consisted of 684 amino acids with an estimated molecular mass of about 74.3 kDa.
The alignment of the mature US132 CGTase sequence to the data bank showed 93, 92, 74, and 73% identity with the CGTases of Bacillus sp. (accession no. CAA46901), Bacillus licheniformis (accession no. CAA33763), Bacillus circulans strain 251 (accession no. CAA55023), and Bacillus sp.1011 (accession no. 1I75A), respectively. The highest homology (97% similarity and 94% identity) was found with the CGTase of Bacillus circulans strain no. 8 (accession no. 1CGT). The alignment showed a difference of 39/684 amino acids which does not radically affect the product specificity. Indeed, both enzymes are specific for the β-CD production and are classified as β-CGTases since they generated a mixture of CDs composed of α-, β-, and γ-CD 20 : 58 : 13 and 10 : 64 : 20 ratios for the CGTase of strain US132 and B. circulans strain no. 8 [25], respectively. However, as far as we know, no data was available concerning biochemical properties of the CGTase of B. circulans strain no. 8 (pH, thermoactivity, thermostability) allowing the comparison of these two enzymes.
From the multiple sequence alignment (Figure 2), the five structural domains (A, B, C, D, and E) [11,26] could be identified in the US132 CGTase. Furthermore, four highly conserved regions labelled I-IV and located in the catalytic domain were found. These conserved regions would constitute the active center of the enzyme since it contained the three catalytic residues Asp229, Glu257, and Asp328 referring to the CGTase of B. circulans strain no. 8 numbering. The conserved amino acids in the acceptor binding site, namely Lys47, Tyr89, Asn94, Phe183, Asn193, Leu194, Tyr195, Asp196, Phe259, Phe283, and Asp371, reported as the main determinants for cyclization reaction [12,27], were also identified.

Production enhancement of the active recombinant US132 CGTase
The purification of the US132 CGTase from the native strain (Paenibacillus pabuli) was extremely hampered by the viscosity of the crude supernatant, probably due to the presence of polysaccharides of unknown nature (data not shown). In the aim to overcome this problem and also to overproduce the US132 CGTase, we have tested the enzyme production from the initial recombinant E. coli strain (DH5α/pSJ8). This study showed that this latter strain, cultivated in LB medium at 37 • C and induced by 16 μg/mL IPTG for 18 hours, exhibited a very low activity of about only 0.3 U/mL. For this reason, we have studied the effect of several parameters on the production enhancement of the active recombinant CGTase.

Effect of the association promoter/host strain
Using the same culture conditions (LB medium, 37 • C and induction by 16 μg/mL IPTG for 18 hours), the association E. coli ER2566/pSJ8 increased the production of the active recombinant enzyme by 3-fold (Table 1) since the ER2566   [18]. To further enhance this production, the gene was placed downstream the Trc promoter in pSJ9 and the enzyme production was investigated in various strains (DH5α, ER2566 and JM109). As shown in Table 1, the association E. coli ER2566/pSJ9 gave the highest activity of only about 1 U/mL which is nearly the same with that obtained by ER2566/pSJ8 strain. This study showed that the use of different promoters and E. coli host strains did not significantly enhance the production of the active recombinant CGTase, which is probably due to the formation of inclusion bodies often reported for E. coli heterologous expression [14,28].

Effect of the operating temperature
A well-known technique to prevent aggregation of recombinant proteins in vivo consists to reduce the operating temperature [15,16,29]. Therefore, in order to improve the recombinant US132 CGTase production, ER2566/pSJ9 and ER2566/pSJ8 strains were cultivated at different temperatures and were induced by 16 μg/mL of IPTG for 18 hours: (a) 37 • C during all operating time, (b) 37 • C followed by a temperature shift to 19 • C after IPTG induction, and (c) 19 • C during all operating time. This study showed that the conditions (b) and (c) allowed the enhancement of the active enzyme production in the periplasm for the two constructions pSJ8 and pSJ9 ( Table 2). The highest production (15 U/mL) was reached by E. coli ER2566/pSJ8 when the culture was performed at 37 • C followed by temperature shift to 19 • C after IPTG induction. SDS-PAGE analysis of periplasmic and insoluble fractions, obtained from E. coli ER2566/pSJ8, showed that the cultivation of the strain at 37 • C expressed most the US132 CGTase as insoluble inclusion bodies with a small amount of soluble protein (Figure 3) explaining the low activity detected. However, the decrease of the operating temperature (37-19 • C or 19 • C) strongly increased the expression of the US132 CGTase as periplasmic soluble form (Figure 3). The enhancement of the production of soluble and active enzyme at low cultivation temperature could be explained by: (i) the reduction of the high rate of protein synthesis, which prevents misfolding [14], (ii) the enhancement of chaperones expression, which allows correct protein folding [30].
Based on all obtained data, we retained for further investigation the strain E. coli ER2566/pSJ8 and a culture temperature at 37 • C followed by a shift to 19 • C after induction.

Effect of medium composition and postinduction time
The composition of growth medium could have significant metabolic effects on both cells growth and protein production [31]. To investigate the effect of this parameter, the strain E. coli ER2566/pSJ8 was cultivated in 2TY, M9ZB, and M9 besides LB medium used in all previous experiences. The US132 CGTase production monitored after 18 hours of induction showed that the use of M9ZB and 2TY medium increased the production by about 1.1fold (16.5 U/mL) and 1.3-fold (20 U/mL), respectively, in comparison to that obtained by LB broth. However, the use of M9 medium decreased the production to attain only 8 U/mL. The enhancement of the US132 CGTase production was likely correlated with cell density since the OD 600 nm reached about 3.7, 4.8, 6, and 2 for the LB, M9ZB, 2TY, and M9 medium, respectively ( Figure 4). Moreover, this difference in the production level of the recombinant active enzyme, using different medium, could be also influenced by the growth phase at which induction took place on. Otherwise, extensive attempts to improve the enzyme production using different IPTG concentrations and stressing additives (sorbitol, mannitol, and ethanol) did not increase the US132 CGTase production (data not shown) as reported for other recombinant proteins [16,28,32].
The effect of postinduction time on the US132 CGTase production was examined by analysis of samples taken, from ER2566/pSJ8 culture performed in 2TY medium, at different times after IPTG addition. Following the IPTG induction, the culture temperature is shifted from 37 to 19 • C. The evolution of both cell growth and active enzyme production ( Figure 5) showed that the CGTase production reached its maximum (22 U/mL) after 22 hours of induction when cells are at full stationary phase. When culture attained the lysis phase the production of the active enzyme decreased by about 27% probably due to the proteases release.

Batch fermentation scale
The fermentation, using the ER2566/pSJ8, was scaled up in 7 L fermentor strain under the conditions previously

D T A V T N K Q N F S T D V I Y Q I F T D R F L D G N P S N N P T G A A Y D A T C S N L K L Y C G G DWQ G L I N K I N D N Y F S D L G 70 BC8 D P D T A V T N K Q S F S T D V I Y Q V F T D R F L D G N P S N N P T G A A Y D A T C S N L K L Y C G G DWQ G L I N K I N D N Y F S D L G 70 Bli D A D T A V T N K Q N F S T D V I Y Q V F T D R F L D G N P S N N P T G A A F D G T C S N L K L Y C G G DWQ G L V N K I N D N Y F S D L G 70 BC251 A P D T S V S N K Q N F S T D V I Y Q I F T D R F S D G N P A N N P T G A A F D G T C T N L R L Y C G G DWQ G I I N K I N D G Y L T GMG 70 Bsp 1011 A P D T S V S N K Q N F S T D V I Y Q I F T D R F S D G N P A N N P T G A A F D G S C T N L R L Y C G G DWQ G I I N K I N D G Y L T GMG 70
. * * : * : * * * . * * * * * * * * : * * * * * * * * * : * * * * * * * : * . : * : * * : * * * * * * * * * : : * * * * * . * : : . : *

G T A A D Y NQ V NDQ V T F I D NHDMD R F K T S A V NN R R L E Q A L A F T L 350 BC8 L D F R F N S A V R N V F R D N T S NM Y A L D S M I N S T A T D Y N Q V N D Q V T F I D N H DMD R F K T S A V N N R R L E Q A L A F T L 350 Bli L D F R F N S A V R N V F R D N T S NM Y A L D S M L T A T A A D Y N Q V N D Q V T F I D N H DMD R F K T S A V N N R R L E Q A L A F T L 350 BC251 L D F R F A Q K V R Q V F R D N T D NMY G L K AM L E G S A A D Y A Q V D D Q V T F I D NHDM E R F H A S N A N R R K L E Q A L A F T L 350 Bsp 1011 L D F R F A Q K A R Q V F R D N T D NM Y G L K A M L E G S E V D Y A Q V N D Q V T F I D N H DM E R F H T S N G D R R K L E Q A L A F T L 350
* * * * * . . * : * * * * * * . * * * . * . : * : . : . * * * * : * * * * * * * * * * * : * * : : * : . * : * * * * * * * * *

PP132 T S R G V P A I Y Y G T E Q Y L T G N G D P D N R A KM P S F S K T T T A F N V I S K L A P L R K S N P A I A Y G S T Q Q RW I S N D V Y V 420 BC8 T S R G V P A I Y Y G T E Q Y L T G N G D P D N R A KM P S F S K S T T A F N V I S K L A P L R K S N P A I A Y G S T Q Q RW I N N D V Y V 420 Bli T S R G V P A I Y Y G T E Q Y L T G N G D P D N R G KM P S F S K S T T A F N V I S K L A P L R K S N P A I A Y G S T Q Q RW I N N D V Y I 420 BC251 T S R G V P A I Y Y G T E Q Y M S G G T D P D N R A R I P S F S T S T T A Y Q V I Q K L A P L R K C N P A I A Y G S T Q E RW I N N D V L I 420 Bsp 1011 T S R G V P A I Y Y G S E Q Y M S G G N D P D N R A R L P S F S T T T T A Y Q V I Q K L A P L R K S N P A I A Y G S T H E RW I NND V I I 420
* * * * * * * * * * * : * * * : : * . * * * * * . : : * * * * . : * * * : : * * . * * * * * * * . * * * * * * * * * : : * * * . * * * :

PP132 Y E R K F G K S V A V V A V N R N L S T P A S I A N L S T S L P T G N Y T D V L G G A L N G N N I T S -T N G N V S S F T L A A G A T A VW 489 BC8 Y E R K F G K S V A V V A V N R N L S T S A S I T G L S T S L P T G S Y T D V L G G V L N G NN I T S -T N G S I NN F T L A A G A T A VW 489 Bli Y E R K F G K S V A V V A V N R N L T T P T S I T N L N T S L P S G T Y T D V L G G V L NG NN I T S -S G G N I S S F T L A A G A T A VW 489 BC251 Y E R K F G S N V A V V A V N R N L N A P A S I S G L V T S L P Q G S Y N D V L G G L L N G N T L S V G S G G A A S N F T L A A G G T A VW 490 Bsp 1011 Y E R K F G N N V A V V A I N R NMN T P A S I T G L V T S L P R G S Y N D V L G G I L N G N T L T V G A G G A A S N F T L A P G G T A VW 490
* * * * * * . . * * * * * : * * * : . : . : * * : . * * * * * * . * . * * * * * * * * * . : : : . * . . * * * * . * . * * * * optimised: 2TY medium and temperature shift from 37 to 19 • C after IPTG induction. Interestingly, the production of the US132 CGTase in fermentor reached a maximum (35 U/mL) after only 18 hours of induction while this maximum (22 U/mL) was obtained after 22 hours in Erlenmeyer. It should be mentioned that the cell density obtained in the two cases was almost the same (about OD 600 nm = 6). This finding suggested that the improvement of the active enzyme production was not only related to the cell growth but it could also depend on the stability of fermentation parameters (pH and pO 2 ), which probably affect translation and correct folding of recombinant proteins as well as proteolysis according to Makrides [28].

Purification and kinetics parameters determination of the recombinant US132 CGTase
For the purification of the recombinant US132 CGTase, we have used the periplasmic fraction of the ER2566/pSJ8 strain cultivated in the fermentor in the cultivation conditions described above. This periplasmic fraction was firstly 7 Production of the active enzyme (U/mL) Cell density (OD 600 nm ) Figure 4: Influence of the growth medium on the active US132 CGTase production and the cell density of E. coli ER2566/pSJ8 strain. The temperature was shifted from 37 to 19 • C after induction with 16 mg/L of IPTG when OD 600 nm reached 0. 8-1. heat-treated at 60 • C in order to remove E. coli thermolabile proteins. The extract was then purified using hydrophobic interaction chromatography (HIC) and starch adsorption as described for the native enzyme [18]. The recombinant enzyme was purified to homogeneity 25-fold with a yield of 20% (Table 3, Figure 6). This recombinant US132 CGTase retained the same biochemical properties (data not shown) as the native enzyme (thermoactivity, thermostability, pH stability, CDs production) [18].
The kinetic parameters were determined by incubating the purified recombinant US132 CGTase in presence of vari- Production of the active CGTase (U/mL) Induction Figure 5: Evolution of bacterial cell growth and the active US132 CGTase production generated by E. coli ER2566/pSJ8 cultivated in 2TY medium. Induction was performed by 16 mg/L IPTG when culture reached OD 600 nm between 0.8 and 1. Following the IPTG induction, the culture temperature is shifted from 37 to 19 • C. ous soluble starch concentrations and using the Lineweaver-Burk method. The V max and K m values were evaluated to be 253 ± 36 μmol of β-cyclodextrin/mg/min and 0.36±0.18 g/L, respectively. K m values determined, using also soluble starch as substrate, for CGTase from Bacillus agaradhaerens [33], Bacillus circulans E192 [34], Bacillus firmus [35], and Bacillus sp. TS1-1 [11] were 21.2, 5.7, 1.21, and 0.52 g/L, respectively. Since a small value of K m exhibited a high affinity for the substrate, the values shown above suggested that US132 CGTase was more specific towards starch than the other reported CGTases.

8
Journal of Biomedicine and Biotechnology

CONCLUSION
The molecular characterization of the gene encoding the US132 CGTase showed that the enzyme is a novel β-CGTase exhibiting 95% of identity with the β-CGTase from Bacillus circulans strain no. 8. The production of the active US132 CGTase by E. coli DH5α under usual culture conditions (37 • C and LB) was very low (0.3 U/mL) suggesting the formation of inclusion bodies. The investigation of different culture parameters showed that mainly the shift of the operating temperature from 37 to 19 • C increased the production of the active CGTase to reach 22 U/mL. This production was further enhanced to reach 35 U/mL by using batch fermentation while the native strain produced only 18 U/mL.
The recombinant US132 CGTase, purified to homogeneity, shared the same biochemical properties with the native enzyme. The determination of the kinetic parameters (V max and K m ) of this recombinant enzyme showed that the US132 CGTase had the highest affinity towards soluble starch in comparison to other reported CGTases.