Functional Study of Lysine Decarboxylases from Klebsiella pneumoniae in Escherichia coli and Application of Whole Cell Bioconversion for Cadaverine Production

Abstract

Klebsiella pneumoniae is a gram-negative, non-motile, rod-shaped, and encapsulated bacterium in the normal flora of the intestines, mouth, skin, and food, and has decarboxylation activity, which results in generation of diamines (cadaverine, agmatine, and putrescine). However, there is no specific information on the exact mechanism of decarboxylation in K. pnuemoniae. Specifically lysine decarboxylases that generate cadaverine with a wide range of applications has not been shown. Therefore, we performed a functional study of lysine decarboxylases. Enzymatic characteristics such as optimal pH, temperature, and substrates were examined by overexpressing and purifying CadA and LdcC. CadA and LdcC from K. pneumoniae had a preference for L-lysine, and an optimal reaction temperature of 37℃ and an optimal pH of 7. Although the activity of purified CadA from K. pneumoniae was lower than that of CadA from E. coli, the activity of K. pneumoniae CadA in whole cell bioconversion was comparable to that of E. coli CadA, resulting in 90% lysine conversion to cadaverine with pyridoxal 5'-phosphate L-lysine.

Introduction

Polyamines have an aliphatic group and two primary amino groups [32], and they are found in almost all bacterial cells and affect many cellular processes, such as transcription and translation. Polyamines have protected Escherichia coli against oxidative and acidic stress conditions, and the levels of enzymes [7,8,10] responsible for polyamine synthesis increase under these conditions, resulting in an increased intracellular polyamine pool [40].

Polyamines participate in the biosynthesis of siderophores [5] and protect against oxygen toxicity such as superoxide stress [15]. They are also involved in plaque biofilm formation, and they play a role in cellular differentiation signaling [14,34], stabilize chromatin, and prevent DNA damage [11]. The relative polyamine concentrations may vary between species and can reach the millimolar range [26]. The most common polyamines are putrescine and triamine spermidine, whereas cadaverine is much less abundant [32]. However, owing to its industrial importance, the biosynthesis of cadaverine has been well documented [10,24,25]. Bioconversion of cadaverine from lysine is economical, and the production of bio-polyamides from cheap starting materials could be an alternative to production of polyamides from petrochemicals [7,8,13,39]. Polyamides can be used for making functional products, such as fungicides, pharmaceuticals, oil and fuel additives, chelating agents, and fabric softeners/surfactants [20,32]. Several studies have aimed to identify lysine decarboxylases [35,37].

Klebsiella pneumoniae is a gram-negative, non-motile, rodshaped, and encapsulated bacterium [23]. It can also utilize lactose in anaerobic facultative fermentation. It is found in the intestines, skin, food, and normal flora of the mouth. K. pneumoniae can generate CO2 with decarboxylases [9], and K. pneumoniae is used in the production of 2,3-butanediol and 1,3-propanediol, by means of glucose fermentation [36]. Although K. pneumoniae has decarboxylase activity for one or more generated amines [9], the mechanism of decarboxylation is unclear, and no functional studies of lysine decarboxylase production by this bacterium have been performed in E. coli. Therefore, we identified and characterized lysine decarboxylases in K. pneumoniae and applied them in a whole-cell reaction system to produce a high yield of cadaverine.

 

Materials and Methods

Chemicals

The reagents used for reaction substrates, culture, and analysis, such as sodium acetate anhydrate, cadaverine, isopropyl β-D-1-thiogalactopyranoside (IPTG), pyridoxal-5-phosphate (PLP), sodium borate, L-ornithine monohydrochloride, L-arginine, L-lysine monohydrochloride, and 2,6-diaminopimelic acid (DAP) were purchased from Sigma-Aldrich Co. Diethyl ethoxymethylenemalonate (DEEMM) was purchased from Fluka Co. (Japan) for the derivatization reaction. The enzyme purification reagent Ni-NTA was purchased from Qiagen.

Bacterial Strains and Media

Genomic DNA of K. pneumoniae and E. coli K12 MG1655 were used for cloning of each lysine decarboxylase gene, cadA and ldcC. To amplify cadA, we constructed and used primers Kpn_cadA_F (5’-CGTCGTGGATCCATGAACGTTATTGCAATATTAATCACA-3’) and Kpn_cadA_R (5’-ATATAAGCTTTCCCGCGATTTTTAGGACTCG-3’) for PCR. The PCR product was inserted into the pET24ma vector using the restriction enzymes, HindIII and BamHI. The plasmid of pET24ma::cadA from E. coli was used from a previous study [17,21]. The plasmids cloned with each lysine decarboxylase gene were transformed into E. coli BL21 (DE3) competent cells for protein expression. A single colony of transformant E. coli from the agar plate was pre-cultured in 3 ml of LB medium, prior to incubating for 14 h with 200 rpm agitation at 37℃. The pre-culture (1 ml) was transferred to 50 ml of LB medium (50 μg/ml kanamycin, 0.1 mM IPTG) and grown at 30℃. Then, the cells were harvested at 4℃ by centrifugation and the cell pellet was washed and stored.

Enzyme Purification and Enzyme Reaction

CadA and LdcC from K. pneumoniae were purified and subjected to enzymatic characterization and kinetic analysis. E. coli BL21(DE3) (50 ml) transformed with pET24ma::cadA was cultured and induced by the addition of 0.1 M IPTG. The cells were sonicated using an ultrasonicator (Vibra Cell, Soncis Scientific, USA) for 5 min on ice (10 sec on, 15 sec off). The cell lysates were centrifuged at 1,146 ×g. The supernatant was collected and applied to Ni-NTA beads. After a 2 h binding reaction, the Ni-NTA beads were washed three times with the 50 mM NaH2PO4, 0.05% NaCl, 0.05% Tween-20 (pH 8.0) buffer containing 20 mM imidazole. Then, His-tagged CadA was eluted three times from the column using 250 mM imidazole buffer. The reaction with 1 M L-lysine, 0.1 mM PLP, 500 mM sodium acetate buffer (pH 6.0), and 20 μl of the whole cell was performed in a 37℃ water bath. The reaction was stopped by heating at 100℃ for 5 min. We defined one unit (mmol/cell dry weight (mg)/min)) of activity as the amount of enzyme producing 1 mmol of cadaverine per minute at 37℃ [2,30].

Derivatization and High-Performance Liquid Chromatography Analysis

Three hundred microliters of 50 mM sodium borate buffer (pH 9), 100 μl of methanol, and 47 μl of distilled water were added to 50 μl of target sample and 3 μl of DEEMM [1,18]. The derivatization reaction was performed at 70℃ for 2 h to derivatize lysine and cadaverine. High-performance liquid chromatography (HPLC; YL-9100, Korea) was used after derivatization to detect derivatized lysine and cadaverine. A reverse-phase Agilent ZORBAX SB-C18 column was used (4.6 × 250 mm, 5 μm particle size). The column temperature was maintained at 35°C. The mobile phase was composed of 100% acetonitrile (A) and 25 mM sodium acetate buffer (pH 4.8, B). A 1 ml/min flow rate was used with the following gradient: 0–2 min, 20–25% A; 2–32 min, 25–60% A; 32–40 min, 60–20% A. The rest of the percentage was charged by the buffer B. Detection was carried out at 284 nm using a UV detector.

Lysine Decarboxylase Amino Acid Sequence Alignment

Lysine decarboxylase sequences were obtained from the NCBI (http://blast.ncbi.nlm.nih.gov/Blast.cgi) for comparing and alignment of protein sequences. A number of alignments were carried out using Clustal W ver. 2 from EBI with a 0.5 transition weight. Phylogenetic trees were constructed in MEGA version 6.06 using the neighbor-joining and unweighted pair-group methods with the arithmetic mean algorithm [31].

 

Results and Discussion

Lysine Decarboxylases from K. pneumoniae and Functional Expression and Purification in E. coli

Biogenic amines have been related to food poisoning, and Klebsiella is found in food and produces biogenic amines by amino acid decarboxylation [4,33]. Based on database searches, we found that Klebsiella pneumoniae has multiple amino acid decarboxylases involved in biogenic amine production. These enzymes were classified as lysine decarboxylases by phylogenetic analysis (Fig. 1). Klebsiella has two lysine decarboxylases; one is cadA, 2,148 bp in length; and the other is ldcC, 2,190 bp in length. The LdcC protein showed 82% similarity to CadA. K. pneumoniae has a lysine decarboxylase system similar to those of most enterobacteria; that is, one constitutive and one inducible lysine decarboxylase (Fig. 1).

Fig. 1.Amino acid sequence alignment and UPGMA bootstrap (100) phylogenetic analysis of lysine decarboxylases similar to that of Klebsiella pneumoniae and their most similar strains according to NCBI (BLASTn). 1. Klebsiella pneumoniae ornithine decarboxylase. 2. Klebsiella pneumoniae arginine decarboxylase. 3. Klebsiella pneumoniae lysine decarboxylase LdcC 4. Klebsiella pneumoniae lysine decarboxylase CadA. 5. Klebsiella pneumoniae diaminopimelate decarboxylase. 6. Escherichia coli K12 MG1655 CadA. 7. Escherichia coli K12 MG1655 LdcC. Phylogenetic trees were developed based on the maximum composite likelihood method using MEGA ver. 6.06.

After PCR amplification of the K. pneumoniae genome, the cadA and ldcC genes were inserted into pET24ma vectors and overexpressed in E. coli BL21. His-tag purification resulted in a high yield of the CadA protein. Like CadA and LdcC from E. coli [12,22], CadA could be produced more easily than LdcC (data not shown). The CadA and LdcC overexpressed in E. coli and purified were subjected to enzymatic characterization and kinetic analysis.

Characterization of CadA and LdcC

Although CadA and LdcC are lysine decarboxylases, they have similarly broad substrate specificities [6,38]. Among arginine, ornithine, DAP, and lysine (Fig. 2A), CadA and LdcC showed a preference for lysine. CadA and LdcC had some activity against arginine, unlike the E. coli CadA, which prefers ornithine [28]. The optimal temperature was 37℃ (Fig. 2B), which is different from previous reports of 60℃ [41]. The relative activities of LdcC and CadA were also similar at over 37℃, but CadA activity was maintained at 25℃. Further experiments used 37℃.

Fig. 2.Characterization of the purified lysine decarboxylases, CadA and LdcC. (A) Substrate specificity test (ornithine, lysine, arginine, and 2,6-diaminopimelic acid (DAP)). (B) Optimal temperature for enzyme activity. (C) Optimal pH for enzyme activity.

The optimal pH of E. coli lysine decarboxylase is 6 [22]. However, that of K. pneumoniae lysine decarboxylase was pH 7 (Fig. 2C). In 500 mM sodium acetate buffer, the final pH was 8.5 after a 2 h reaction. The conversion of lysine into cadaverine was almost complete in less than 15 min when a low concentration of lysine was used. Compared with previous data, CadA showed higher Km and lower kcat values, resulting in a lower kcat/Km value than LdcC (Table 1). Although the reaction conditions were slightly different from those of other enzymes, K. pneumoniae CadA had lower activity than CadA from E. coli.

Table 1.Kinetic constants for lysine decarboxylases (CadA) from Klebsiella pneumoniae and other microorganisms

Optimization of Whole Cell Reaction

A whole cell system is preferred for cadaverine production owing to its robustness [17], and, as a result, various wholecell system conditions were examined. As PLP is required for decarboxylase activity [19,29] and PLP was one of the major factors for cadaverine production, unlike growthbased production comparatively less effect of additional PLP in fermentation with a [25,27], the effect of PLP concentration was examined in the whole cell system (Fig. 3A). In the absence of PLP, lysine consumption for conversion of 1 M L-lysine to lysine was only 20% (data not shown). Following addition of over 0.025 mM PLP, lysine consumption was about 90%, suggesting that bioconversion required PLP as a cofactor. Bioconversion by LdcC was markedly lower than that of CadA; however, LdcC showed more PLP-dependent behavior. The conversion yields of both cells containing CadA and LdcC from K. pneumoniae increased with increasing PLP concentration (Fig. 3A). The conversion yield of the whole cell system with CadA from K. pneumoniae was 3–5% greater than that of cells containing CadA from E. coli. When the reaction was performed at 37℃ with various substrate concentrations, 1 M of lysine was fully consumed during a 2 h reaction (Fig. 3B). However, substrate inhibition occurred at over 1.25 M. Moreover, addition of a greater number of cells containing CadA increased the conversion ratio of lysine into cadaverine (data not shown). Cadaverine production was similar irrespective of the presence of buffer (Fig. 3C). Although a dramatic increase in pH under 8 occurred in the buffer-free system and pH differed according to the presence of buffer, the overall conversion ratio was similar to that of CadA from E. coli.

Fig. 3.Optimization of whole cell bioconversion. K. pneumoniae lysine decarboxylase (CadA and LdcC) activity at different (A) PLP concentrations and (B) substrate (lysine) concentrations. (C) Effect of 500 mM sodium acetate buffer (pH 6) for keeping the enzyme’s optimal pH from produced cadaverine’s natural pH and added 1 M lysine.

Application of CadA to Whole-Cell Cadaverine Conversion

Although the activity of purified CadA from K. pneumoniae was slightly lower than that of CadA from E. coli [17], cadaverine production in a whole cell system is affected by various factors. Therefore, cadaverine production using CadA from E. coli and K. pneumoniae in a whole cell system was compared. The reaction was carried out using whole cells with 1 M L-lysine and 0.1 mM PLP in 500 mM sodium acetate buffer at different pH values for 2 h. The conversion yields of K. pneumoniae CadA and E. coli CadA were similar irrespective of pH (Fig. 4A). At an initial pH of 6, the conversion yield of the whole cell system with CadA from K. pneumoniae was 80% and that with CadA from E. coli was 86%. The conversion yield at initial pH values of 9 and 10 was over 75%. The conversion yields of whole cell systems with E. coli CadA and K. pneumoniae CadA were similar irrespective of the buffer system. The final cadaverine yield from lysine was 94% in both the presence and absence of sodium acetate buffer, indicating no difference in the whole cell systems incorporating CadA from E. coli and K. pneumoniae (Fig. 4B).

Fig. 4.Whole-cell cadaverine conversion yields. Whole-cell cadaverine conversion by lysine decarboxylase (CadA) from K. pneumoniae and E. coli (A) at pH 6 (E. coli’s optimal pH) and alkaline pH 9 and 10. (B) Conversion yield over time with or without 500 mM sodium acetate buffer.

In conclusion, we compared lysine decarboxylases purified from K. pneumoniae with that from E. coli using an E. coli whole cell system. Although the activity of CadA from K. pneumoniae was lower than that of CadA from E. coli, the overall conversion of K. pneumoniae CadA was comparable to that of E. coli CadA, and resulted in ~90% conversion of lysine to cadaverine after optimization of various reaction factors.