Exploring Lactobacillus reuteri DSM20016 as a biocatalyst for transformation of longer chain 1,2-diols: Limits with microcompartment

Lactobacillus reuteri metabolises glycerol efficiently to form 3-hydroxypropionic acid (3-HP) and 1,3-propanediol (1,3PDO) by the same mechanism as that for 1,2-propanediol (1,2PDO) conversion to propionic acid and propanol via its propanediol utilization (pdu) pathway. Pdu enzymes are encoded by the pdu-operon, which also contain genes encoding the shell proteins of the microcompartment housing the metabolic pathway. In this work the selectivity and kinetics of the reactions catalysed by L. reuteri DSM20016 Pdu enzymes glycerol dehydratase (GDH), 1,3-propanediol oxidoreductase (PduQ) and coenzyme-A acylating propionaldehyde dehydrogenase (PduP), produced recombinantly, was investigated against corresponding substrates of different chain lengths. Glycerol dehydratase exhibited activity against C2-C4 polyols, with the highest activity against glycerol and 1,2-propanediol (1,2-PDO). A double mutant of the pduC gene of GDH (PduC-S302A/Q337A) was constructed that displayed lowered activity against glycerol and 1,2PDO but extended the substrate range upto C6-diol. The best substrate for both PduQ and PduP was 3-hydroxypropanal (3HPA), although PduP exhibited nearly 10-fold higher specific activity. The enzymes also showed some activity against C3-C10 aliphatic aldehydes, with PduP having higher relative activity. Subsequently, transformation of polyols using whole cells of L. reuteri containing the wild type- and mutated GDH, respectively, confirmed the reduced activity of the mutant against glycerol and 1,2PDO, but its activity against longer substrates was negligible. In contrast, recombinant Escherichia coli BL21(DE3) cells harboring the GDH variant converted diols with up to C6 carbon chain length to their respective aldehydes, suggesting that the protein shell of the microcompartment in L. reuteri posed a barrier to the passage of longer chain substrate.


Introduction
With the expansion of global biodiesel production, the major byproduct glycerol has attracted great interest as an important platform for the biobased chemical industry [1][2][3]. A large number of studies have thus focused on valorisation of glycerol to produce not only the chemicals that are currently produced from fossil feedstocks but even those that are not currently PLOS  available in the market, e.g. 3-hydroxypropionaldehyde (3-HPA) and 3-hydroxypropionic acid (3-HP) [4][5][6][7][8][9][10]. Also, utilization of different microorganisms able to metabolise glycerol to 1,3-propanediol such as K. pneumoniae, C. butyricum, and Lactobacillus species, or their enzymes in heterologous hosts, have shown promising results from an industrial viewpoint, due to high conversion rates and product yields [11][12][13].
Bioinformatics studies have indicated that about 17% of bacterial specicies use proteinaceous microcompartments (MCPs, typically 100-150 nm in diameter) as simple organelles to localize metabolic pathways that have toxic or volatile intermediates [18][19][20]. Bobik and coworkers initially showed that Salmonella conditionally forms Pdu-microcompartment that can sequester propionaldehyde to prevent cytotoxicity during growth in the presence of 1,2-PDO [21]. The protein shells of Pdu MCPs are thought to be composed of 9 different bacterial microcompartment domain proteins, i.e. PduA, PduB, PduB', PduK, PduJ, PduM, PduN, PduU and PduT, for which a striking feature is that they have central pores that mediate the transport of substrates, products and enzyme cofactors between the cell cytoplasm and interior of the Pdu MCPs [22]. The current view is that most MCP domain proteins are hexagonal in shape and tile edge to edge to form extended protein sheets, which then interact with the pentameric vertex proteins to develop intricate architecture [23].
Glycerol dehydratase (GDH) is considered to play an essential role in the transformation of glycerol/1,2-PDO by producing both an electron acceptor and a metabolic intermediate. The dehydratase enzyme requires cobalamin (Vitamin B12) as a cofactor whose biosynthesis is encoded by the gene sets cbi, cob and hem, located close to the pdu operon [24]. Genetic characterization of L. reuteri show the regulatory control of the pdu-cbi-hem-cob cluster is based on its induction by 1,2-propanediol or glycerol [24]. The operon and its mechanism have also been reported in other lactobacilli and microorganisms including Salmonella enterica, Clostridium carboxidivorans, Klebsiella pneumoniae and Enterococcus malodoratus [12,16,21,25,26]. Basically, NADH produced during glucose metabolism is preferentially reoxidized to NAD + by the reduction of 3-HPA to 1,3-PDO (when glycerol is the co-substrate) or propionaldehyde to n-propanol (1,2-PDO as co-substrate). An adequate supply of glycerol/1,2-PDO can thus outweigh the repression of glucose on the central metabolism. In the absence of glucose, glycerol/1,2-PDO are metabolised by L. reuteri without being used for growth, and the cofactor cycling is achieved by processing of the aldehyde intermediate via the two branches of the pdu pathway described above. This constitutes a unique naturally occurring system for co-production of an organic acid and an alcohol, as shown in case of the production of 3-HP and 1,3-PDO from glycerol [14].
The aim of the present study was to determine how broad the scope of the substrates accepted by the key enzymes (i.e. glycerol dehydratase, PduP and PduQ) of the Pdu pathway of L. reuteri DSM20016 is, and to make use of rational engineering to increase the selectivity of glycerol dehydratase for longer chain diols, and furthermore to evaluate the combined performance of enzymes in the whole bacterial cells for transformation of the various diols (Fig 1). Lactobacillus reuteri transform longer chain 1,2-diols

Chemicals and reagents
All restriction enzymes, IPTG and X-Gal were purchased from Fermentas. Taq DNA polymerase, and T4 DNA ligase were from New England Biolabs. BugBuster protein extraction reagent was from Novagen. MRS-medium was obtained from Beckton, Dickinson, Le Pont de Claix, France. Antibiotics, Vitamin B 12 (CN-B 12 ), coenzyme B 12 (AdoB 12 ), Coenzyme A sodium salt hydrate, NAD + , NADP + , NADH and NADPH were purchased from Sigma-Aldrich, St. Louis, Missouri, USA. All primers and kits for DNA extraction and purification were procured from ThermoFisher Scientific, while QuikChange II XL Site-Directed Mutagenesis Kit was from Agilent Technologies.

Bacterial strains, media and growth conditions
The bacterial strains used in the study are listed in S1 Table. L. reuteri DSM20016 was initially grown in 20 mL of MRS-medium under anaerobic conditions at 37˚C without shaking for 16-17 h. Cultivation was started by an inoculum (1% v/v) from a freshly grown culture grown under the same conditions. Depending on the experiment, the medium was supplemented with 20 mM glycerol (Sigma Aldrich). Cells were harvested by centrifugation at 6000 x g for 10 min and washed twice with 50 mM sodium acetate buffer, pH 5.0. Recombinant Escherichia coli BL21(DE3) with recombinant GDH was grown in LB medium under 37˚C with shaking at 220 rpm as described below.
Cloning, expression and purification of wide-type/mutant glycerol dehydratase, PduP and PduQ L. reuteri DSM 20016 was cultivated in 20 mL MRS medium under anaerobic conditions in closed vials at 37˚C. The cells were harvested and genomic DNA was extracted and purified using GeneJet Genomic DNA kit (Fermentas). The gene sequences coding for GDH (PduCDE) and its reactivating factor (PduGH) were amplified by the primers listed in S1 Table. Initial denaturing at 95˚C for 3 min was followed by 30 cycles of denaturing at 95˚C for 30 s, annealing at 53˚C for 30 s and elongation at 72˚C for 1.5 min. A final elongation was carried at 72˚C for 10 min. The PCR products were separated and purified by agarose gel electrophoresis and inserted directly into vector pGEM-T to construct pGEM-T-pduCDE and pGEM-T-pduGH, respectively.
Site-directed mutagenesis of GDH was performed using a QuikChange II XL Site-Directed Mutagenesis Kit. Plasmid pGEM-T-pduCDE was used as template. The mutagenic sense and antisense primers are listed in S1 Table. The double mutant carrying both PduC-S302A and PduC-Q337A mutations was designated pGEM-T-pduCDE (C-S302A/Q337A). The presence of intended mutations and the absence of unintended mutations were comfirmed by sequence analysis. The mutant plasmid pGEM-T-pduCDE (C-S302A/Q337A) and wild type (WT) plasmid pGEM-T-pduCDE were digested with the restriction enzymes BamHI and EcoRI. The resultant 2.9-kb fragment that contains the entire pduCDE gene, with either double mutations or wild type, was transferred to the corresponding region of pCDFDuet to construct expression plasmids pCDFDuet-pduCDE (C-S302A/Q337A) and pCDFDuet-pduCDE, respectively. The plasmid pGEM-T-pduGH, containing the genes encoding two subunits of glycerol dehydratase reactivating factor, was also digested with restriction enzymes AatII and KpnI, and transferred downstream of the other T7 promotor correspondingly to construct the final expression plasmids pCDFDuet-pduCDEGH (C-S302A/Q337A) and pCDFDuet-pduCDEGH, respectively. The plasmids were introduced into E. coli BL21(DE3) and recombinantly expressed by induction with 1mM IPTG at 30˚C for 15 h. The cell pellet was washed and resuspended in 20 mM phosphate buffer (pH 8.0), and disrupted with Bead Beater (Mini-Beadbeater-16, BioSpec) for 15s (8 cycles) in an anaerobic condition. The broken cells were centrifuged at 13 500 x g for 30 min at 4˚C to remove cell debris, and the supernatant was collected to measure the crude glycerol dehydratase activity.
Construction of recombinant E. coli strains expressing native PduQ and PduP sequences, respectively, was performed previously [15,27]. Basically, the recombinant E. coli strains were grown in 500 mL flask containing 100 mL LB medium with 40-100 μg/mL of suitable antibiotics under aerobic conditions at 37˚C with shaking at 220 rpm. The gene expression was induced by addition of 0.1-1 mM IPTG when the bacterial growth had reached mid-exponential phase (OD 620 around 0.5) with subsequent incubation for 24 h at 15˚C and 160 rpm. The cells were then harvested and washed with 20 mM sodium phosphate buffer pH 7.4 containing 0.5 M NaCl. Lysis of the cells was performed by BugBuster according to manufacturer´s instructions. The soluble protein fraction was separated from cell debris by centrifugation for 40 min at 16 000 x g, 4˚C, and subjected to immobilized metal ion affinity chromatography (IMAC) for purification of the recombinant enzymes. The protein solution was loaded on 5 mL Ni-NTA HisTrap™ FF crude column equilibrated with binding buffer (20 mM sodium phosphate, 0.5 M NaCl, 5 mM imidazole, pH 7.4) at a flow rate of 1 mL/min. After washing the column with the binding buffer, the bound protein was eluted using elution buffer containing 500 mM imidazole solution. Purity of the recombinant enzymes was analysed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) on a gel containing 12% acrylamide. The protein bands were stained with Coomassie Brilliant Blue R-250.

Determination of activities of GDH, PduQ and PduP
The activity of crude glycerol dehydratase was determined by the 3-methyl-2-benzothialzolinone hydrazone (MBTH) method as described earlier [28], which is based on the ability of aldehydes formed during the dehydratase reaction to react with MBTH. The resulting azine derivative is detected spectrophotometrically. Generally, the assay mixture (1 mL) contained an appropriate amount of GDH, 0.2 M substrate, 0.05 M KCl, 0.035 M potassium phosphate buffer (pH 8.0), and 15 μM coenyzme B12. After incubation at 37˚C for 1-10 min, the enzyme reaction was terminated by adding 1 mL 0.1 M potassium citrate buffer pH 3.6 and 0.5 mL 0.1% MBTH. After 15 min at 37˚C, the amount of aldehyde formed was determined by reading the absorbance at 305 nm. One unit of glycerol dehydratase activity is defined as the amount of enzyme activity catalyzing the formation of 1 μmol aldehyde/min at 37˚C.
The activity of PduQ was determined as described by Reid et al [29] with some modification. Reduction of the aldehyde substrate was measured by following the decrease in absorbance of reduced nicotinamide adenine dinucleotide (NADH) at a wavelength of 340 nm (∑ 340 NADH = 6.22mM -1 cm -1 ) in a reaction mixture containing 0.2 mM NADH, 100 mM of an aldehyde substrate and appropriate amount of enzyme in 50 mM potassium phosphate buffer pH 7.0. One Unit (U) of activity is defined as the amount of enzyme catalyzing the consumption of 1 μmol NADH per minute. The activity of PduP was determined by a method developed by Leal et al. [30]. The method is based on the oxidation of the substrate aldehyde in the presence of NAD(P) + and measuring the production of NAD(P)H spectrophotometrically at 340 nm. The reaction mixture containing 0.43 mM HS-CoA and 5 mM NAD + in 50 mM potassium phosphate buffer pH 7.0 supplemented with 1 mM dithiothreitol and 0.1 g/L BSA was equilibrated for 5 min at 30˚C before initiating the reaction by addition of 100 mM aldehyde substrate. One Unit (U) of activity is defined as the amount of enzyme catalyzing the formation of 1 μmol NADH formed per minute.

Construction of chromosomal pduC mutations
Scarless chromosomal site-directed point mutations of PduC gene in L. reuteri were constructed by single-stranded DNA recombineering method as described earlier [31,32], with slight modifications. Briefly, electrocompetent cells, in which RecT is expressed, were transformed with a recombineering oligonucleotide oLCH1(50-mer, 100 μg), which is identical to the lagging strand containing mutiple non-complementary bases. Viable cells were then recovered on antibiotic-free plates and recombinants are detected by a mismatch amplification mutation assay-PCR (MAMA-PCR). Single colony purification was performed to separate the wild-type genotype from the mutant ones. The resultant mutant strain PduC-S302A was constructed by curing from pJP042, passing bacteria without antibiotic selection to yield a plasmid-free derivative. A similar strategy was used for constructing double mutant (PduC-S302A/Q337A) by transforming a recombineering oligonucleotide oLCH2 (50-mer, 100 μg) followed by the procedure similar to that described above. All the primers used and plasmids constructed are listed in S1 Table. The pduC mutants were also verified by DNA sequencing analysis.
Biotransformation using resting L. reuteri cells L. reuteri cells were grown in a 3 L bioreactor (Applikon, Microbial Biobundle, The Netherlands) as reported previously [4]. Monitoring and control of all the parameters was done through ez-control unit. Stirrer speed was maintained at 200 rpm, temperature at 37˚C and pH at 5.5 by addition of 5 N NH 4 OH. Anaerobic conditions were maintained through continuous bubbling of nitrogen gas. Twenty milliliters of freshly prepared inoculum was aseptically added to 2 L fermentation medium containing 55 g/L MRS broth and 20 mM glycerol, cultivation was continued for 12 h after which the broth was collected and centrifuged at 15 000 x g and 4˚C for 5 min. The supernatant was discarded and the cell pellet was used for biotransformation experiments.
Biotransformation of different substrates (i.e. ethane-1,2-diol, glycerol, 1,2-propanediol, 1,2-butanediol, 1,2-pentanediol and 1,2-hexanediol) was done in 50 mL centrifuge tubes. The process was initiated by resuspending the harvested L. reuteri cells from the previous step in 50 mL solution containing 100-150 mM 1,2-diols/ glycerol substrates to a final density of about 10 g CDW /L, mixed by vortexing and incubated in a shaker incubator up to 3 hours at 30˚C. One milliliter samples were collected at 30 min intervals, centrifuged and the supernatant was analyzed for the concentration of residual substrate and metabolites formed.

Bioconversion of 1,2-diols and glycerol by whole cells of recombinant E. coli
Shake flask experiments were performed using LB Lennox (supplemented with 1.2-diols) medium with modification to the method described previously [33]. One milliliter of overnight culture of the transformants harboring constructed plasmid pCDFDuet:pduCDEGH or pCDFDuet:ΔpduCDEGH(C-S302A/Q337A) were grown in LB (37˚C, 220 rpm) as the preinoculum, and 500 μL was added to 100 mL flask consisting of 20 mL LB supplemented with 50 mM 1,2-diols or glycerol, along with appropriate antibiotics (streptomycin, 50 μg/ml) and grown for additional 4-6 hours (37˚C, 220 rpm). Following this, 10 μM coenzyme B12 and 0.5 mM IPTG were added to the cultures, and cultivation continued for another 24-36 h at 30˚C, 200 rpm. One milliliter samples were withdrawn for product analyses.
For quantification of 3-HPA, the modified photometric method according to Circle et al. [34], with acrolein as standard was used. Briefly, 200 μL of the sample (diluted to fit in the range of the assay) was mixed with 150 μL of DL-tryprophan, followed by addition of 600 μL of hydrochloric acid (37%), and incubating the mixture for 20 min at 37˚C. The resulting purple colour was measured spectrophotometrically at 560 nm and the absorbance was compared with the standard curve.

Results
Comparative genome analysis reveals a diversity of gene cluster for 1,2-diol/ glycerol metabolism Comparison of the genetic loci bearing the pdu gene cluster in different bacteria such as Citrobacter freundii, Salmonella typhimurium, Listeria monocytogenes, Klebsiella spp., Clostridium spp. and Lactobacillus spp. have revealed that the pdu genes are clustered in different orders and are scattered differently in the genomes (Fig 2). A detailed comparison performed in the present work, revealed that the pdu operon in most of the bacteria listed, except for C. butyricum, C. perfringens and C. pasteurianum that lack the Pdu pathway and the microcompartment, ranges from 15 208 bp (C. carboxidivorans P7) to 27 392 bp (S. sanguinis SK36). Diol/glycerol dehydratase, is the key and rate-limiting enzyme as reported previously [35]. The general features of all diol/glycerol dehydratase gene sequences and the gene products are summarized in S2 Table. Particularly, the genes encoding this enzyme found in C. freundii, Y. enterocolitica, K. oxytoca, etc., consist of three types of subunits, i.e, pduC (α), pduD (β) and pduE (γ). Reactivating factors for diol/glycerol dehydratase were also discovered [36,37]. The gdrA and gdrB, two open-reading frames of C. freundii, are highly homologous to the products of gdrA and gdrB of K. pneumoniae (KPHS_46160, KPHS_46220), C. perfringens (CPF_RS05765, CPF_RS05770) and C. pasteurianum (CLPA_RS11070, CLPA_RS11065). Moreover, the pduG and pduH gene products, which are homologous among all bacteria harboring Pdu-microcompartment, except for Terrisporobacter glycolicus, are considered to be involved in the reactivation of glycerol-inactivated diol/glycerol dehydratase. Some bacteria, including L. reuteri, possess the pdu-cbi-cob-hem cluster comprising 58 ORFs, encoding the biosynthetic machinery for the Vitamin B12 cofactor (data not shown). The different arrangements of the diol/ glycerol dehydratase genes offer us L. reuteri contains three genes (Lreu_1747-1745) with dehydratase subunit motifs in the propanediol untilizing operon (193594-213109 bp). The encoded proteins of these three genes display sequence similarity to PduCDE of other Lactobacillus spp. (i.e. L. reuteri, L. brevis, L. collinoides), which share a high level of conserved ORFs in the pdu operon. The genomes of L. reuteri DSM 20016 and L. brevis ATCC 367 each consist of sequenced circular chromosomes of 199 9618 and 229 1220 bp, and totally contain 2049 and 2030 genes, respectively. A comparison of the pdu operon in the two lactobacilli also reveals a conserved gene order (i.e. pduFABC-DEGHKJLMNOSPQWUV). The close phylogenetic relationship is evident from a comparative analysis of the ribosomal protein sequences from these two lactobacilli (Fig 2), which are both probiotics and with a potential in production of various chemicals, e.g. 3-hydroxypropionaldehyde, 1,3-propanediol and 3-hydroxypropionic acid [14]. On the other hand, the cbi-cob-hem gene set present in L. reuteri has not been identified in other lactobacilli including L. brevis (data not shown).
Subsequently, activities of the next enzymes in the Pdu pathway, PduQ and PduP were determined against aldehydes of varying chain length. The two enzymes were produced recombinantly and purified prior to activity measurements. As shown in Fig 3B-3D, both the enzymes were active with all aldehydes (C3-C10), except for acetaldehyde (C2) that is extremely volatile. Significantly higher specfic activities was observed for PduP. Highest specific activities for both PduQ and PduP were indeed found with 3-hydroxypropanal (3-HPA) as substrate, with K m values of 28.7 mM and 10.1 mM, respectively. With propionaldehyde, PduQ and PduP displayed only 26.5% and 51.9% activities, respectively, compared to that with 3-HPA. Decrease in activity with increasing chain length of the diol substrates seemed to be correlated with the apparent increasing order of K m values for both enzymes: Pentanaldehyde (C5)< Hexanaldehyde (C6)< Octanaldehyde (C8)< Decylaldehyde (C10).

Redesign of glycerol dehydratase to act on longer chain 1,2-diols
As the activity of glycerol dehydratase is a limiting step in bioconversion of different 1,2-diols by L. reuteri, broadening of substrate selectivity of the enzyme by rational engineering could contribute to transformation of longer chain 1,2-diols. Therefore, comparison of the binding model of glycerol and 1,2-propanediol to glycerol dehydratase was first drawn based on homology modeling (Fig 4). The model based on the crystal structure of diol dehydratase-cyanocobalamin complex from Klebsiella oxytoca (1DIO) [38], revealed sequence similarity of 42% and coverage above 95%. As shown from the model, glycerol dehydratase is assembled in the form of a dimer of PduCDE heterotrimers. The large subunit (PduC), harboring an essential cofactor ion-binding site, contains a TIM barrel structure with the active site isolated inside the central barrel formed by eight paralled β strands. Coenzyme B 12 is located in between the PduC and medium subunit (PduD). It was also observed from the docking model that on one hand, 3-OH group of glycerol forms a hydrogen bond with the side chain of S302 while on the other, the side-chain amide nitrogen atom of Q337 formed hydrogen bonds with the main chain carbonyl oxygen atoms of D336 and F375, both of which are key residues for interacting with substrate molecule (S1 Fig). Loss of these hydrogen bonds may alter the conformation and the flexibility, therefore decreasing the bulkiness of the residues in these positions.
Hence, according to the hypothesis that a less bulky residue could yield more space for accomodating a longer alkyl group on C3 of 1,2-diols, the plasmid for the mutant glycerol dehydratase was constructed in which both of the residues (S302 and Q337) were substituted by alanine. The relative activities of cell-free extract of the wild-type and the mutant enzymes towards different 1,2-diols and glycerol were compared. As shown in Fig 4E, the S302A/ Q337A mutation resulted in decrease in relative activity towards 1,2-ethanediol, glycerol and 1,2-propanediol to 39.0%, 46.2% and 43.2%, respectively, of the wild type. In contrast, relative Lactobacillus reuteri transform longer chain 1,2-diols activity towards 1,2-butanediol increased by almost three folds. Moreover, activities towards longer chain 1,2-diols, i.e. 1,2-pentanediol and 1,2-hexanediol, were also detected with relative increase by 57.86% and 35.56%, respectively.

Whole cell bio-transformation of 1,2-diols/glycerol by engineered L. reuteri
In order to study the effect of the mutation in glycerol dehydratase described above on the enzyme activity on different 1,2-diols in whole cells, L. reuteri was engineered to create a glycerol dehydratase variant (S302A/Q337A). As seen in Fig 5, no detectable quantities of products were observed from cultures fed with 1,2-ethanediol, 1,2-butanediol, 1,2-pentanediol and 1,2-hexanediol. Furthermore, it was seen that the S302A/Q337A mutant cells showed a negligible difference in resistance toward substrate inhibition by glycerol. The native cells (11.3 g CDW /L) converted 6 mmoles 1,2-PD within 3 h, while the mutant cells (10.8 g CDW /L) showed a 20% decrease in conversion efficiency (4.8 mmoles), leading to n-propanol production at 1.45 mmoles, as compared to 1.75 mmoles from native cells (Fig 5).

Recombinant E. coli cells bearing the glycerol dehydratase variant transform longer chain 1,2-diols
In order to confirm that the lack of activity of the engineered whole cells of L. reuteri towards the longer substrates was due to the diffusional barrier posed by the microcompartment, the Homology modeling for the prediction of tertiary structure of glycerol dehydratase (PduCDE) resulted in a model based on the crystal structure of diol dehydratase-cyanocobalamin complex from Klebsiella oxytoca (1DIO), with sequence similarities and coverages above 42% and 95%, respectively. Quality model assessment of the homology model revealed a QMEAN6-score of 0.83 and a Z-score of -0.53. Ramachandran plot revealed that none of residues were present in the disallowed regions. Color codes for carbon atoms: yellow for adenosylcobalamin (AdoCbl), white for glycerol/1,2-PD, purple for calcium ion. The model figure was generated using Pymol. D) Sequence alignment of large subunit of glycerol dehydratase (PduC) from different bacteria. The numbering scheme follows the amino acid sequences of PduCs. Identical residues in all sequences are highlighted in black and conserved in grey. E) Relative activities of WT/mutant glycerol dehydratases towards different 1,2-diols/glycerol. The data is normalized by taking the activity of the wild-type enzyme towards glycerol as 100%.
Comparative genome analysis has shown these bacteria to possess a specific gene cluster encoding the pathway enzymes and the protein shell forming the microcompartment, providing an optimum environment for efficient transformation and minimising the inhibitory effects of the intermediates and products.
Expanding the repertoire of diol substrates for transformation by the Pdu pathway could result in a novel synthetic route for the production of interesting products. However, establishing routes for microbial production of chemicals often requires overcoming pathway bottlenecks by tailoring enzymes to accept non-native substrates. Diols are important chemicals with applications as building blocks for polymers, in cosmetics and pharmaceuticals, and as fuels [46]. 1,2-Diols, namely 1,2-propanediol, 1,2-butanediol, 1,2-pentanediol, 1,2-hexanediol can be biotechnologically produced by direct microbial bioconversion of renewable feedstocks [47].
Diol/glycerol dehydratases have been recognised for their catalytic ability with C2 (1,2-ethanediol) and C3 (1,2-propanediol) diol, and C3 (glycerol) [48][49][50]. In this study, we found that the relative substrate activities towards 1,2-PDO was 86% with respect to glycerol, which is consistent with the previous studies [51] suggesting the wide-type dehydratase of L. reuteri has a higher affinity for glycerol than for 1,2-propanediol (K m 3.3 mM vs 7 mM). Earlier studies, including those from our laboratory, have further shown that L. reuteri is unique among the lactobacilli in its ability to accumulate and excrete 3-HPA, and is also tolerant to relatively high concentration of the aldehyde product [14,51]. Reactions with other 1,2-diols showed that the wild-type enzyme showed activity with 1,2-ethanediol and 1,2-butanediol only, corresponding to 28% and 21% activity relative to glycerol. Lack of activity with longer chain diols is attributed to the less space available for accomodating a longer alkyl chain in the active site [52].
The active subunit PduC of diol/glycerol dehydratase is quite conserved among the bacteria (Fig 4D). It was foreseen by docking analysis that replacing S302 and Q337 by Alanine residues would result in the loss of hydrogen bonds between the active site and substrate, altering the flexibility of the binding site. Activity measurements of the glycerol dehydratase with double mutations showed the increased substrate scope of the the enzyme variant. It is especially evident that the relative acitivity towards ethane-1,2-diol, glycerol and 1,2-propanediol decreases significantly, to 11%, 46% and 37%, respectively, with respect to the 100% activity of the wild type enzyme with glycerol (Fig 4E), most likely due to the reduced contact between the substrates and the side chains of the mutated residues in glycerol dehydratase. These results are in close agreement with the observations made by Yamanishi and coworkers, who mutated the same residues in Klebsiella oxytoca diol dehydratase to increase the resistance of the enzyme towards mechanism-based inactivation by glycerol [52]. They reported increased K m value of the mutants towards 1,2-PDO and increased activity towards longer chain 1,2-diols as a result of the reduced bulkiness of the residues in the catalytic position. Unlike the K. oxytoca enzyme however, the L. reuteri GDH has shown negligible difference in resistance towards inhibition by glycerol (unpublished data), confirming the previous assumption that L. reuteri enzyme is a glycerol dehydratase that is more resistant towards substrate inhibition.
Interestingly, unlike the GDH both PduP and PduQ were able to transform aldehydes of chain lengths up to C10, although PduQ exhibited lower specific activity that could partly be ascribed to metal induced loss of activity on exposure to air (unpublished observations). Nevertheless, this result implies that the combined use of the PduP and PduQ branches of the Pdu pathway would provide a unique biotransformation system for transformation of aldehydes simultaneously to an acid and an alcohol product, a biocatalytic alternative to the chemical Canizzaro reaction involving base induced disproportionation of an aldehyde.
Since L. reuteri possesses the entire Pdu pathway including the machinery for activating the cofactor needed for glycerol dehydratase, and has been shown earlier to efficiently transform glycerol to 3-HPA, 3-HP and 1,3-PDO with relatively high tolerance to product inhibition [4,9,10], expanding the repertoire of diol substrate for transformation by the microorganism would result in a novel biosynthetic system for specific aldehyde, acid and alcohol products. However, inspite of the broader substrate scope of the mutant glycerol dehydratase and also of PduP and PduQ, diols longer than 1,2-PDO could not be transformed by the whole cells of L. reuteri cells. This was ascribed to the barrier posed by the protein shell of microcompartment for the passage of the more longer, hydrophobic diols. That this was really the case was supported by the ability of recombinant E. coli expressing the GDH variant to transform 1,2-diols up to C6 to their corresponding aldehydes in accordance with the substrate selectivity of the double mutant. Notwithstanding, the disadvantage of using recombinant E. coli (e.g. BL21 (DE3) and its derivatives) is obvious, even though cells harboring engineered glycerol dehydratase can convert certain amount of long chain diols. One of the most important points is that the ability of E. coli to survive in low pH condition is much weaker as compared to L. reuteri, which can be a promising host for low-pH fermentation.
Crowley et al. identifed that the properties of the PduA pore of Salmonella enterica is suggestive of 1,2-PDO transport [53]. Structural analysis of the shell protein PduB of Lactobacillus reuteri revealed that besides glycerol occupying the subunit channels of PduB, the trapped glycerol in a central pocket locks the loops closed and raises the possibility of a ligand-gated channel [54]. From our previous work on L. reuteri DSM20016, it is well established that the Pdu enzymes (PduCDE, PduP, PduL, PduQ etc.) are packed within the microcompartment of L. reuteri and the shell is permeable to 1,2-propanediol/glycerol (substrates), 3-hydroxypropionaldehyde (3-HPA), propanol/1,3-propanediol and propionyl-phosphate/3-hydroxypropionyl-phosphate, but propionaledhyde is somehow sequestered to prevent cell damage [15]. There are some hints that limit of microcompartment may be caused by polarity of the pore site that can affect transport of both small molecules (e.g. 3-HPA and propionaldehyde) and large cofactors (e.g. HS-CoA and NAD + /NADH) [55][56][57][58]. However, the underlying mechanism is still not entirely clear and is being investigated by us. In our ongoing work on mutations of L. reuteri shell protein PduA/J pore amino acid residues, certain mutants perform differently than the wild type (unpublished data), which also indicates that the physicochemical properties of the Pdu-microcompartment pore site could determine the interaction and diffusion of 1,2-propanediol/glycerol or longer chain 1,2-diols through the pores.
While the present studies have involved coenzyme B 12 -dependent diol/glycerol dehydratase, the research could be further extended to B 12 -independent enzymes, e.g. from C. freundii CFNIH1, K. pneumoniae HS11286, C. butyricum KNU-L09 and T. glycolicus ATCC14880 (S2 Table) due to their less complexity and low cost for reactivation for large scale applications. Improvements in biotransformations can further be achieved by using a more active PduQ or a similar enzyme, and by replacing the oxidative branch of the Pdu pathway with a single step of aldehyde oxidation using a suitable aldehyde dehydrogenase. L. reuteri does indeed possess a cytoplasmic alcohol dehydrogenase [15] and an aldehyde dehydrogenase with activity against 3-HPA, which could be evaluated for their action on longer chain substrates.