Heterologous expression and characterisation of

0 downloads 0 Views 459KB Size Report
Feb 1, 2012 - catalysed by three enzymes MlrA, MlrB and MlrC, respec- tively. ..... where v0 is the initial reaction rate, [S] the initial concen- tration of substrate ...
Toxicon 59 (2012) 578–586

Contents lists available at SciVerse ScienceDirect

Toxicon journal homepage: www.elsevier.com/locate/toxicon

Heterologous expression and characterisation of microcystinase  ska b, Jussi Meriluoto c, Dariusz Dziga a, *, Benedykt Wladyka b, Gabriela Zielin d Marcin Wasylewski a

Department of Plant Physiology and Development, Faculty of Biochemistry, Biophysics and Biotechnology, Jagiellonian University, Gronostajowa 7, Kraków, Poland Department of Analytical Biochemistry, Jagiellonian University, Kraków, Poland c Department of Biosciences, Åbo Akademi University, Turku, Finland d Department of Cell Biology, Jagiellonian University, Kraków, Poland b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 4 September 2011 Received in revised form 8 December 2011 Accepted 17 January 2012 Available online 1 February 2012

The first enzyme in the microcystin (MC) degradation pathway identified in bacterial strains is coded by mlrA gene and is referred to as microcystinase. To date, there has been no biochemical characterisation of this enzyme. The results presented herein show a successful heterologous expression of MlrA as well as mutational studies, partial purification and biochemical characterisation of the enzyme. The mutation and inhibition study confirmed previous ideas that MlrA is a metalloprotease and allowed to calculate the inhibition parameters. Moreover, the kinetic parameters of MC-LR linearization were measured showing that MlrA exhibits a positive cooperativity towards MC-LR. Furthermore, in vitro experiments with Escherichia coli cells expressing MlrA indicated the potency of the heterologous host to eliminate MCs with very high efficiency. This study reports a new approach to the analysis of a microcystin degrading enzyme, extends the knowledge about MC biodegradation and opens broad scope for future study. Ó 2012 Elsevier Ltd. All rights reserved.

Keywords: Microcystin Biodegradation Microcystinase

1. Introduction The occurrence of cyanotoxins during and after cyanobacterial blooms has important implications for both recreational use of lakes and reservoirs and drinking water abstraction (van Apeldoorn et al., 2007). The presence of cyanotoxins in water and their accumulation in aquatic organisms create a potential risk to humans exposed to toxins either directly or through the food chain. Hepatotoxic microcystins (MCs), due to their cyclic structure and unusual amino acids composition, are very stable and exhibit resistance to many chemical agents and enzymatic hydrolysis by common proteases (Mazur and Plinski, 2001). A bacterial strain capable of MC degradation was reported by Jones et al. (1994). A subsequent study (Bourne et al., 1996) identified the MC degrading bacterium as a new Sphingomonas species (ACM-3962). The MC degradation * Corresponding author. Tel.: þ48 12 6646540; fax: þ48 12 6646902. E-mail address: [email protected] (D. Dziga). 0041-0101/$ – see front matter Ó 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.toxicon.2012.01.001

pathway was described as a three-step process which involves linearisation of MC molecule, formation of a tetrapeptide and subsequent hydrolysis of these products catalysed by three enzymes MlrA, MlrB and MlrC, respectively. Some further bacteria (mostly belonging to the Proteobacteria family) with degradation activity against MCs were identified in environmental samples (Edwards and Lawton, 2009). The gene cluster involved in the biosynthesis of the enzymes was also characterized (Bourne et al., 2001). By cloning the total genomic DNA of Sphingomonas sp. a 5.8 kb sequence that codes for proteolytic enzymes with activity against MCs was found. Within the sequence four genes were identified: mlrA, mlrB, mlrC and mlrD. The first three genes encode proteins exhibiting the enzymatic activity observed in Sphingomonas sp. (Bourne et al., 1996), whereas the mlrD gene probably encodes a transporter protein. Later sequence analysis (Bourne et al., 2001) described MC degrading enzymes more precisely but many questions remained unanswered. The exact location of the start and

D. Dziga et al. / Toxicon 59 (2012) 578–586

stop codons is uncertain and cleavage of protein profragments probably occurs. The first enzyme in the MC degradation pathway encoded by mlrA gene was called microcystinase. Using restriction enzymes, the start codon was located between bases 1607 and 1788, while the end of mlrA gene was found to be located at position 2669 in the 5.8 kb sequence. A putative amino acid sequence containing 336 residues was deposited in GenBank (accession number AF 411068). The calculated molecular mass was established to be 36.3 kDa (Bourne et al., 2001). Recently, several new Sphingomonas strains were shown to be capable of hydrolyzing different hepatotoxins in the manner first described by Bourne et al. (1996). Additionally, there have been a few reports indicating other bacteria species with a similar ability (Paucibacter toxinivorans – Rapala et al., 2005; Arthrobacter spp., Brevibacterium sp., and Rhodococcus sp. – Manage et al., 2009; probiotic bacteria – Nybom et al., 2008). The preferred approach in the analysis of MC degrading organisms is to search for the mlrA gene homologues because its presence in the genome of investigated bacterial strains correlates with the potency to degrade MC variants. Several authors have also suggested that Mlr enzymes should be expressed, purified and studied in detail (Ishii et al., 2004; Kato et al., 2007; Hashimoto et al., 2009). In the current work we have documented the successful expression of MlrA enzyme in a heterologous host, Escherichia coli BL21(DE3), which allowed much higher enzyme production in comparison with the natural strain (Sphingomonas ACM-3962) and partial enzyme purification. Furthermore, biochemical parameters of the enzyme have been studied, including the pH profile of enzyme activity, inhibition characteristics and a study of kinetic parameters of MC linearisation. 2. Materials and methods 2.1. Materials and bacterial strains Molecular weight standards for SDS-PAGE electrophoresis and pTZ57R/T cloning vector were obtained from Fermentas (Vilnius, Lithuania) and expression vector pET21a from Novagen (Darmstadt, Germany), whereas the mutagenesis kit came from Stratagene (La Jolla, USA). PVDF Western blotting membrane was from Applied Biosystems (Foster City, USA); mouse anti-His-tag antibodies, goat anti-mouse alkaline-phosphatease-conjugated secondary antibodies, FASTÔ BCIPÒ/NBT as well as chemicals used for SDS-PAGE electrophoresis, buffers, enzyme inhibitors (EDTA, phenantroline, AEBSF, pepstatin A, E-64) and PP1 were from Sigma (St Louis, MO, USA). TFA, RP C18 Purospher column and Fractogel SO 3 resin were obtained from Merck (Darmstadt, Germany); synthetic peptides were from Bachem AG (Bubendorf, Switzerland). MC-LR was extracted from a culture of Microcystis aeruginosa PCC 7813 strain obtained from the Pasteur Institute (Paris) and HPLC purified as described earlier (Gajdek et al., 2003). Other MC variants (MC-LF and MC-LW extracted from M. aeruginosa PCC7820, dmMC-RR extracted from Microcystis NIES 107) were HPLC purified as described by Meriluoto and Spoof (2005a).

579

Sphingomonas sp. ACM-3962, obtained from Australian Collection of Microorganisms, was cultured in a recommended peptone yeast extract medium (299) at 28  C for 48 h. After two days of incubation cells were centrifuged and washed in PBS buffer. E. coli DH5a and BL21(DE3) used for cloning, plasmid propagation and expression of recombinant proteins, respectively, were grown at 37  C in LB broth supplemented with ampicilin (100 mg mL1), where indicated. Staphylococcus aureus 178RI (D’Elia et al., 2006) was cultured at 37  C in tryptic soy broth supplemented with chloramphenicol (10 mg mL1) where indicated. 2.2. Construction of recombinant plasmids (recombinant MlrA and MlrA mutants) The sequence coding for MlrA was amplified by PCR Sphingomonas sp. ACM-3692 genomic DNA as the template and primer pairs detailed in (Table 1) were used. The amplified fragments were inserted into pTZ57R/T cloning vector and sequence was verified. Subsequently, the fragments were cut out using appropriate restriction enzyme (Table 1) and inserted into expression vector pET21a or pG164 (D’Elia et al., 2006). The resulting plasmids (pET21amlrA or pG164-mlrA) encode the full length MlrA (1–336) with a His-tag on its C-terminal. Mutagenesis studies were performed with a sitedirected mutagenesis kit (according to the manufacturer’s instructions). Briefly, the primer pair mlrAH260AF and mlrAH260AR were used for a PCR with pET21a-mlrA as a template to exchange the CAC triplet encoding His260 in the putative active site of MlrA into GCC encoding alanine. The reaction mixture was digested with DpnI to remove the parental DNA template. The nicked vector, pET21amlrAH260A, containing the desired mutation was then transformed into DH5a competent cells. Similarly, the primers mlrAE265AF and mlrAE265AR were used to produce a Glu265 / Ala substitution in the sequence of MlrA. 2.3. Expression of recombinant MlrA The pET21a-based constructs were transformed into E. coli BL21(DE3) and bacteria were plated on LB agar plates supplemented with ampicillin (100 mg mL1). Freshly

Table 1 Oligonucleotide primers used in the construction of recombinant plasmids. Name

Sequence

Restriction enzyme

mlrAF mlrAR

50 -GTTCCATATGCGGGAGTTTGTCCGAC-30 50 -GAAAGCGGCCGCGTTCGCGCC GGACTTG-30 50 -CCATAACGCGCTGGGAGTAAACGTC-30 50 -CTCCCAGCGCGTTATGGATGGCGTG-30 50 -GTTATGGATGGCGGCAGTGAGCACGC-30 50 -GCGTGCTCACTGCCGCCATCCATAAC-30 50 - CTAGGGATCCATGCGGGAGTTTGTCCG-30 50 -TCGACTCGAGTTAGCCGTTCGCGCC GGACT-30

NdeI NotI

mlrAE265AF mlrAE265AR mlrAH260AR mlrAH260AF mlrApG164F mlrApG164R

– – – – BamHI XhoI

580

D. Dziga et al. / Toxicon 59 (2012) 578–586

transformed colonies were inoculated into LB medium and grown at 37  C until the absorbance A600 0.8 was reached; then, the temperature was decreased to 30  C and recombinant expression was induced by the addition of IPTG (isopropyl b-D-thiogalactoside) at a final concentration of 1 mM, and then culturing was continued for 4 h. Subsequently, the bacteria were centrifuged (15000 g, 20 min, 4  C) and the pellet was further processed or stored at 20  C. The pG164-mlrA plasmid was electroporated into S. aureus 178RI and the bacteria were plated on TSA plates supplemented with chloramphenicol (10 mg mL1). The obtained colonies were inoculated into liquid TSB medium and grown at 37  C until the absorbance A600 1.0 was reached. Next, IPTG was added to induce expression of recombinant protein and culturing was continued for further 12 h. Before induction and after induction with IPTG, 1 mL of bacterial culture was taken, pelleted, dissolved in 100 mL of SDS sample buffer and boiled in a water bath for 5 min. Fractions (5 ml) were subjected to SDS/PAGE (Schagger and von Jagow, 1987) and then either Coomassie Blue-stained or electrotransferred (Towbin et al., 1979) on to a PVDF Western blotting membrane. The membrane was incubated with mouse anti-His-tag antibodies to detect recombinant MlrA. After incubation with goat anti-mouse alkaline-phosphatase-conjugated secondary antibodies, the blot was developed using FASTÔ BCIPÒ/NBT. 2.4. MlrA activity assay Extracts of natural or heterologous strains obtained by sonication using the ultrasonic processor UP100H (Hielscher Ultrasonics) as well as partially purified MlrA enzyme were used to estimate the activity against MC variants. Ten mL of the enzyme in different dilution was added to 90 mL of MC solution. The enzyme as well as MCs were suspended in PBS buffer, pH 7.0. Final MC concentration was 1 mg mL1 or higher, where indicated. The incubation temperature was 20  C and the reaction was stopped after 1 h (unless otherwise indicated) by addition of 10 mL of 1% TFA. Samples were cooled and analysed by HPLC and/or MS methods. 2.5. Experiments with viable cells of E. coli BL21 and Sphingomonas sp. ACM-3962 To compare the potency of transformed cells to hydrolyse MC, a degradation assay was performed for E. coli culture and Sphingomonas sp. ACM-3962 as a natural, nonmodified strain. 50 ml of 18 h- and 48 h-old-culture, respectively, was divided into two parts and washed twice with PBS. To indicate activity toward MC-LR, cells were suspended in 1 mL of PBS, pH 7.0 and incubated with MCLR. After 1 h cells were centrifuged and the reaction was stopped by the addition of TFA. Simultaneously, MC-LR was incubated with supernatant obtained after 1 h incubation of cell suspension to control the possible cell lysis. The second part of the cultures was sonicated four times to estimate enzyme activity of lysates using MlrA assay as described in Section 2.4.

2.6. HPLC and MS assays HPLC analyses, including MC-LR degradation rate under different conditions as well as primary identification of the products, were performed using a Water HPLC system consisting of a 600E multisolvent-delivery system, a 717plus autosampler, a 996 photodiode array detector (PDA), Millenium32 SS software and a Jetstream 2 plus column thermostat. MCs and degradation products were quantified on a Purospher STAR RP-18 endcapped column (55 mm  4 mm, 3 mm particles) as described by Meriluoto and Spoof (2005b). The mobile phase consisted of a gradient of 0.05% aqueous TFA (solvent A) and 0.05% TFA in acetonitrile (solvent B) with the following linear gradient programme: 0 min 25% B, 5 min 70% B, 6 min 70% B, and 6.1 min 25% B. The instrument consisted of an Agilent Technologies (Waldbronn, Germany) 1200 Rapid Resolution LC coupled with Bruker Daltonics HCT Ultra Ion Trap MS with an electrospray (ESI) ion source operated in the positive electrospray ion mode. The drying gas temperature and flow rate were set at 350  C and 8 L min1, respectively. The Purospher STAR RP-18 endcapped column (55 mm  4 mm, 3 mm particles) was kept at 40  C. The mobile phase consisted of a gradient of 0.1% aqueous formic acid (solvent A) and acetonitrile with 0.1% formic acid (solvent B) with the following linear gradient programme: 25% B to 70% B over 5 min, then to 90% B over 2 min, where it was held for 1 min.

2.7. Recombinant MlrA purification NiNTA (Qiagen) based purification was applied using conditions recommended by the supplier. Ion exchange chromatography was performed as follows: upon centrifugation E. coli BL21(DE3) were suspended in 50 mM phosphate buffer pH 7.0 and sonicated. After sonication, the suspension was centrifuged at 15000 g and 4  C for 20 min. The supernatant was collected and the pellet was resuspended and ultrasonic processing was repeated three times. The obtained solution was purified using ÄKTA Explorer chromatographic system (GE Healthcare), column filled with Fractogel EMD SO 3 (Merck), equilibrated with 50 mM sodium phosphate buffer, pH 7.0 at flow rate 4 mL min1. Proteins were eluted from the column with a gradient of NaCl (0–2.0 M) in the same buffer. Absorption was monitored at three wavelengths: 280 nm, 260 nm and 220 nm. Fractions exhibiting high MlrA activity were pooled, dialysed against 50 mM sodium phosphate buffer, pH 7.0 and used in further experiments.

2.8. Biochemical characterisation of MlrA and toxicity assay of linear MC-LR To confirm the class of protease to which MlrA belongs, its activity was measured in the presence of selected inhibitors: EDTA, o-phenantroline, 2-aminoethylbenzen-

D. Dziga et al. / Toxicon 59 (2012) 578–586

esulfonyl fluoride (AEBSF), pepstatin A and E-64, with final concentrations of 2.32, 0.03, 0.2, 0.033 and 10 mM, respectively. After preliminary experiments, EDTA and ophenantroline were compared and IC50 values for these two inhibitors were calculated. The final concentrations of these inhibitors ranged between 1-100 mM and 0.01– 20 mM, respectively. To measure kinetic parameters, the 5 mL aliquots of MlrA containing ca. 0.02 mU of the enzyme were used to hydrolyse MC-LR at a number of peptide concentrations (1.09–1092 mM). The time of the reaction was the time between mixing of the reactants and performing injection in HPLC system. HPLC separation was performed as described above. Product and substrate concentrations were calculated using a calibration curve. The data were used to form kinetic curves and initial reaction rate for each curve was calculated. In order to calculate kinetic reaction constants the results were fitted, using SigmaPlot (Systat Software), to Michaelis–Menten Eq. (1) and to Hill Eq. (2) where v0 is the initial reaction rate, [S] the initial concentration of substrate, Km the Michaelis constant, h the Hill cooperation coefficient, and K0.5 the half saturation constant.

v0 ¼

ðVmax ½SÞ Km þ ½S 

v0 ¼ 

Vmax ½Sh

(1)



h K0:5 þ ½Sh



(2)

To obtain the pH profile of MlrA, the enzyme was diluted in 0.12 M Britton Robinson buffer in pH range 5.0–10.0, with 0.2 step. The Britton Robinson Buffer was prepared using equimolar amounts of acetic, phosphate and boric acids co-dissolved and adjusted to given pH. Measurements were performed in triplicate. The toxicity of the linear MC-LR obtained from HPLC separation was established by PP1 inhibition assay as described by An and Carmichael (1994). The concentration of MC-LR (standard) for preparing the calibration curve ranged between 0.125 and 4.0 mg L1, whereas the concentration of linear MC-LR ranged between 1.5 mg L1– 150 mg mL1. Finally, the IC50 value for linear MC-LR was calculated.

2.9. Peptide library screening A library of chromo- and fluorogenic synthetic substrates (Bachem) was screened to check the MlrA substrate specificity. The reaction was performed using partially purified MlrA solution and 0.2 mM substrate. Hydrolysis of the substrates containing p-nitroanilide group (pNA) was monitored at 405 nm using Spectra Max 250 from Molecular Devices, and substrates containing amido-methyl-coumarin group (AMC) were monitored by fluorescence at excitation/emission wavelengths of 380 nm/460 nm, using Spectra Max Gemini EM from Molecular Devices.

581

Fig. 1. SDS/PAGE of total lysates of E. coli BL21(DE3) cells transformed with pET21a-mlrA (left panel) and Western blot of recombinant MlrA with the antibodies against His-tag (right panel) before () and after (þ) induction with IPTG.

3. Results 3.1. Recombinant MlrA and its variants To express recombinant C-terminally-His-tagged MlrA, E. coli BL21(DE3) cells were transformed with pET21amlrA. SDS/PAGE of the total cell lysates before and after induction with IPTG demonstrated no overexpression of the recombinant MlrA (Fig. 1). Nevertheless, the lysate has much higher activity against MC-LR than the lysate from Sphingomonas sp. (Table 2) which confirms that the heterologous host produces MlrA. Western blot of recombinant protein with the antibodies against the His-tag showed a band corresponding to a protein around 28 kDa (Fig. 1). The expression of the constructs in plasmids pET21amlrAH260A and pET21a-mlrAE265A in E. coli was only slightly higher upon induction with IPTG in comparison to the wild type protein, but activity of the mutants against MC-LR was completely abolished (Table 2). 3.2. Activity of recombinant MlrA against different MC variants The activity of E. coli BL21(DE3) cell lysates against MCLR was tested using HPLC and MS methods. HPLC chromatogram (Fig. 2) indicated that MC-LR (retention time 7.2 min) is degraded to a product with a retention time

Table 2 The MlrA activity measured for whole cells and cell extracts of Sphingomonas sp. ACM-3962 and heterologous hosts transformed with plasmids for MlrA expression. Type of sample Sphingomonas sp. ACM-3962 E. coli BL21(DE3) pET21a-mlrA E. coli BL21(DE3) pET21a-mlrAH260A E. coli BL21(DE3) pET21a-mlrAE265A Staphylococcus aureus 178RI pG164-mlrA

MlrA activity (mU mL1 culture) cells extract cells extract extract extract extract

0.006 0.095 1.456 646.3 no activity no activity detected but not quantified

582

D. Dziga et al. / Toxicon 59 (2012) 578–586

Fig. 2. HPLC chromatographic presentation of the MlrA activity of E. coli BL21 (DE3) lysates. Peaks with retention time 7.2 min and 6.5 min correspond to MC-LR and linear MC-LR, respectively. Insert: UV spectrum of linear MC-LR.

6.5 min. The isolated product had unusual mass spectra (Fig. 3A) with the main peak m/z 862.5 and lower intensity of the second peak m/z 1013.7 which is typical of linear MCLR (1012 Da molecular mass). The ion m/z 862.5 was recognized as a linear MC-LR derivative whose fragmentation occurred probably during the ionization process (more details in Discussion). The activity of MlrA against different MC variants was also documented. MS analysis of degradation products indicated that apart from the expected ions which are related to molecular masses of linear MCs, corresponding ions with the reduced m/z value always coexisted (Fig. 3): m/z 1013.7 and 862.5 for MC-LR, m/z 528.9 and 453.5 (doubly charged ions) for dmMC-RR, m/z 1043.6 and 892.4 for MC-LW, m/z 1004.6 and 853.5 for MC-LF. 3.3. Comparison of the MlrA activity of cell extract and whole cells of heterologous hosts and Sphingomonas sp. ACM-3962 The culture of Sphingomonas sp. strain incubated for 48 h achieved maximum density OD600 ¼ 2.0, whereas 18 h old culture of E. coli BL21(DE3) transformed with pET21amlrA had the maximum OD600 ¼ 1.4. The estimated MlrA activity of cell extract as well as intact cells are shown in Table 2. It was indicated, after the calculation of these results to the same densities of lag phase culture, that E. coli lysates had, on average, 6800 times higher activity of MlrA in comparison with lysates of Sphingomonas sp. Similarly, whole cells of E. coli exhibited approximately 250 times higher MlrA activity than Sphingomonas cells. No microcystinase activity was detected in E. coli BL21(DE3) expressing MlrA mutants. The activity of S. aureus 178RI pG164-mlrA lysate was on similar level as E. coli BL21(DE3) pET21a-mlrA. 3.4. MlrA purification and biochemical characterisation Purification on the Fractogel SO 3 EMD resin allowed to obtain solution with high MlrA activity which contains less

than 5% of the total bacterial proteins. The obtained profile of MlrA activity was congruent with absorbance measured at 280 nm (Fig. 4A), however SDS-PAGE displayed several bands for the analysed fractions (data not shown). Partially purified recombinant MlrA (activity 0.03 mU mL1) was inhibited in the presence of EDTA and o-phenantroline with the IC50 values 37.6  3.1 and 1.48  0.31 mM, respectively. The MC-LR hydrolysis assay was performed in pH range 5.0–10.0 with 0.2 Sorensen unit step. The progress of the reaction was presented as a percentage of maximum observed activity (i.e. for pH ¼ 7.6) (Fig. 4B). For pH 5.0 the relative activity was only slightly lower than 25%, whereas in pH range 6.4–9.4, relative activity was over 75%. When pH ¼ 10.0, relative activity was slightly lower than 50%. In the kinetics study of the MC-LR linearisation by MlrA, the initial velocity of this reaction was calculated. The obtained results were fitted to Michaelis–Menten (Fig. 4C, dashed line) and Hill (Fig. 4C, solid line) equations (Eq. 1,2). The sigmoidal saturation curve presented on the main panel of Fig. 4C fits the experimental data far better than does the hyperbolic saturation curve. The behaviour is shown on the insert of Fig. 4C where the Scatchard plot of the same data and applied previous models are presented. According to the Hill model, the obtained kinetic parameters were K0.5 ¼ 159  1 mM, h ¼ 1.57  0.01, Vmax ¼ 0.066  0.2 nM min1. The PP1 inhibition assay was performed to compare the toxicity of the MlrA degradation product with the MC-LR as a substrate. The calculated IC50 values for MC-LR (Fig. 5, Supplementary data) and linear MC-LR (Fig. 4D) were 1.32  0.27 mg L1 and 2780  675 mg L1, respectively.

3.5. Synthetic substrate library screening A set of 28 fluorogenic and 32 chromogenic synthetic substrates was incubated with partially purified MlrA. None of the substrates checked were hydrolysed above the control reaction with cell extract of E. coli transformed with

D. Dziga et al. / Toxicon 59 (2012) 578–586

Intens. x10 7 1.25

583

+MS, 2.8-2.8min #(418-424)

A

862.5

m/z 1013.7 - linear MC-LR m/z 862.5 - MC-LR + H − NH2 − PhCH2CHOMe

1.00 0.75 1013.7

0.50 0.25 507.3

431.8

0.00

300

Intens. x10 7

599.7

400

500

734.5

672.8

631.7

600

817.2

700

966.6

800

900

1080.7

1000

m/z

+MS, 1.2-1.3min #(163-184)

B

2.5

453.3

m/z 528.9 (doubly charged) - linear dmMC-RR m/z 453.5 (doubly charged) - dmMC-RR + H − NH2 − PhCH2CHOMe

2.0 1.5 1.0

528.9

0.5 0.0

905.5

307.7

300

Intens. x10 7 1.0

400

C

500

600

700

800

1056.7

900

1000

m/z

+MS, 4.6-4.6min #(1071-1083) 1043.6

m/z 1043.6 - linear MC-LW m/z 892.4 - MC-LW + H − NH2 − PhCH2CHOMe

0.8 0.6 0.4 799.3

0.2

727.1

6

892.4

945.8

986.5

1020.2

651.8

0.0 200 Intens. x10 7

823.2847.7

762.8

300

D

400

500

600

700

800

900

1000

m/z

+MS, 4.7-4.7min #(1081-1100)

m/z 1004.6 - linear MC-LF m/z 853.5 - MC-LF + H − NH2 − PhCH2CHOMe

1004.6

4

2 853.5

0

200

720.2

300

400

500

600

700

801.4

800

923.5

900

986.6

1072.6

1000

m/z

Fig. 3. Mass spectra of linear MC variants (A – MC-LR, B – dmMC-RR, C – MC-LW and D – MC-LF) after hydrolysis by recombinant MlrA.

a control plasmid. This suggests that MlrA is a very specific protease with substrate range limited to microcystins. A list of substrates can be found in the Supplementary material. 4. Discussion Currently an increasing focus on bacterial degradation of hepatotoxic cyanobacterial peptides (NOD and MCs) is being observed (Park et al., 2001; Ishii and Abe, 2000, 2004; Saito et al., 2003; Chen et al., 2010; Wang et al., 2010).

Previous studies have demonstrated that the ability of bacteria to degrade MCs is related to the presence of the gene mlrA that encodes a hydrolytic enzyme with specificity to these cyclic toxins. This mlrA gene was first identified in a Sphingomonas sp. (Bourne et al., 1996) and further characterised in later studies. The potency to employ these bacteria in MC degradation was also demonstrated on laboratory scale (Tsuji et al., 2006; Ho et al., 2006, 2010; Grutzmacher et al., 2010). However there is still no characterisation of microcystinase (MlrA), the key enzyme in

584

D. Dziga et al. / Toxicon 59 (2012) 578–586

Fig. 4. Purification and biochemical parameters of MlrA. A – chromatogram of MlrA separation using Fractogel SO3 ion exchange chromatography. Solid line represents absorbance at 280 nm in mAU. Dots connected by short dashed line represent MlrA activity in arbitrary units, dashed line represents conductance due to NaCl gradient formation (ranged between 0 and 160 mS cm1). B – MlrA relative activity as a function of pH. Progress of the reaction was presented as percentage of maximum observed activity (i.e. for pH ¼ 7,6). Bars indicate standard deviations, n ¼ 3. C – kinetics of MC-LR hydrolysis by MlrA. Main panel: dots represent experimental data, dashed and solid lines represent substrate saturation curves for Michaelis–Menten and Hill model, respectively; insert: Scatchard plot for the same data, dashed line according to Michaelis–Menten model, solid curve according to Hill model. D – inhibition pattern of PP1A catalytic subunit by linear MC-LR. Bars indicate standard deviations, n ¼ 3.

this process. Due to low MlrA production in Sphingomonas sp., heterologous expression seems to be the best way to test the properties of this important protein. This study reports for the first time the expression, purification and biochemical characteristic of MlrA. The sequence of mlrA from Sphingomonas sp. ACM-3962 reported by Bourne et al. (2001) was verified and corrected during the construction of recombinant plasmids. In the corrected open reading frame two bases (CC) were added after 85th nucleotide and two bases (CT) were removed after 232nd nucleotide which shifted the frame of resulting protein sequence change between 29 (Asp to Ala) and 78 (Pro to Ala) amino acid residues. In addition, to enable usage of affinity chromatography and to track the protein using Western-blot technique, a His-tag was added to Cterminus. Carboxy-terminal end of the protein was chosen because earlier studies (Bourne et al., 2001) suggested that MlrA may possess a 26 amino acid signal peptide. Surprisingly, we noticed a band corresponding to protein around 28 kDa but not 37 kDa, as expected (Fig. 1). The band was not observed in the lysates from uninduced and control samples which confirms that the signal is Histagged-MlrA-specific. These results demonstrate that the posttranslational processing of MlrA from its N-terminus is the result of the activity of E. coli endogenous proteases but

not autocatalytic maturation of the recombinant protein itself. This was further confirmed during analysis of the production of recombinant MlrA in S. aureus 178RI, where the 28 kDa was not observed (data not shown). Lack of significant overexpression of a recombinant protein might be the result of its toxicity to the host due to MlrA activity. To verify this hypothesis a site-directed mutagenesis was applied and point mutants in the postulated active centre (H260AIH263NE265) (Bourne et al., 2001) were constructed: MlrAH260A and MlrAE265A. Both mutants were inactive (Table 1) but upon induction with IPTG only a slight improvement in the expression was observed. The hypothesis that MlrA is a metalloprotease (Bourne et al., 1996) and prediction of the active centre (Bourne et al., 2001) were confirmed (Table 2). To confirm that recombinant MlrA has identical activity against different MCs, HPLC and MS methods were employed (Figs. 2 and 3). MC variants were chosen due to differences in hydrophobic properties of analysed peptides. We have tested different amino acid residues at position 4 since this peptide bond is hydrolysed by MlrA. Another factor that has been taken into consideration was the abundance of MCs variants in nature (MC-LR and dmMC-RR variants are in the group of the most common hepatotoxins). Detailed analyses of mass spectra of linear

D. Dziga et al. / Toxicon 59 (2012) 578–586

MCs provided interesting results due to the presence of two main peaks, both corresponding to the products of MlrA hydrolysis (Fig. 3). m/z 862.5 ion (Fig. 3A) was reported by Bourne et al. (1996) as one of the MS–MS daughter ions (MC-LR þ H–NH2  PhCH2CHOMe) of linear MC-LR in MS– MS fragmentation. The fragmentation patterns of both ions (m/z 862.5 and 1013.7) were very similar (data not presented). Besides ions typical for linear MC variants the ions showing the loss of the 151 Da fragment were observed for all checked degradation products (Fig. 3A–D). The ratio of intensities of both ions differed among the MC variants; high intensity of daughter ions was observed for more hydrophilic variants, whereas for relatively hydrophobic MC-LF and MC-LW these ions were only detectable. This ratio changed also when different MS apparatus and ionization parameters were used. It means that the detection of daughter ions may be helpful in the identification of MCs degradation products. The obtained activity of MlrA measured in E. coli BL21(DE3) extract was 6800 fold higher in comparison to wild Sphingomonas strain. Furthermore, experiments with intact cells confirmed that even the construct without artificial signal sequence added to MlrA can hydrolyse MCs effectively (Table 2). The detoxication ability of intact E. coli BL21 cells was around 440 fold lower than cell lysate which suggests that only a small fraction of MlrA was delivered to the periplasmic space. On the other hand, for Sphingomonas sp. this ratio was almost 30 times lower. From this data it can be concluded that for the wild Sphingomonas strain MlrA is located in the outer cell compartments i.e. the periplasm while in the case of E. coli such location is far less probable. Despite that, the transformed E. coli cells were nearly 250 fold more effective in MC linearisation than non-modified cells of Sphingomonas sp. (Table 2) which could be a crucial advantage of the use of modified microorganisms for bioremediation applications. To perform MlrA purification, the Ni-NTA column was applied. Although recombinant MlrA has a His-tag on its Cterminus, for unknown reasons we were unable to purify the protein using metal-affinity chromatography. Recombinant MlrA was not bound to the resin in either native or denaturing conditions. Lack of retention was not due to removal of the His-tag since in Western blot analysis we could detect the His-tagged MlrA in the metal-affinity column flowthrough (Fig. 1, right panel). With regard to the MlrA location, it was suggested (Bourne et al., 2001) that MlrA possesses a 26 amino acid signal peptide directing the protein to periplasmic space. The homology analysis using BLAST showed, apart from the MlrA reported in other bacterial strains, a similarity to CAAX amino terminal protease family, which are membrane proteins (Pei et al., 2011). Moreover computational studies using DAS-TMfilter server (Cserzo et al., 2002) performed on the corrected MlrA sequence predicted that the protein is located in the membrane with six transmembrane alpha helixes and that the enzyme active centre is exposed to the periplasm. Such properties of microcystinase may explain difficulties in overexpression of recombinant MlrA and growth arrest of recombinant host upon induction with IPTG. The experimental results seem to be contrary to the above expectations. The

585

obtained recovery efficiency of extraction in the presence of detergents (1% Triton100) was lower than extraction using PBS buffer (data not presented). This finding (which may suggest that MlrA is not a membrane protein) resulted in the use of buffered solutions instead of detergent solutions. Due to previous conclusions and because we were unable to apply affinity chromatography to purify MlrA, we decided to use ion exchange chromatography techniques. Among a variety of applied ion exchange techniques, the best results were obtained using Fractogel SO 3 . Such an approach allowed the removal of at least 95% of cell proteins and to conduct more detailed analysis when compared with studies using cellular extracts (Fig. 4A). Additionally to the mutational study, for the first time we characterized quantitatively the inhibition pattern of EDTA and o-phenantroline using partially purified MlrA, free of other enzymes involved in the MC degradation pathway. It is another proof that the former classification of the studied enzyme as a metalloprotease (Bourne et al., 1996) was correct and that o-phenantroline is a stronger inhibitor of MlrA than EDTA. The obtained results of MlrA activity in different pH (optimum at pH 7.6) showed that MlrA is a neutral protease (Fig. 4B) and confirmed that the selection of neutral pH for both purification and enzyme activity assays was correct. The performed study of MC-LR hydrolysis by MlrA documented for the first time the kinetic constant which was in the micromolar region. Both the hyperbolic curve according to the Michaelis–Menten model and the sigmoidal curve according to the Hill model are shown in Fig. 4C. It may be easily seen that the Hill model fits the experimental data better. It was also demonstrated by statistical tests and by lower estimator errors of the Hill model in comparison with the Michaelis–Menten model. In Schatchard plot, the linear plot representing Michaelis– Menten model is not adequate. The calculated Hill coefficient (h ¼ 1.57) proves positive cooperativity of MlrA against MC-LR. This may be explained in terms of enzyme oligomerisation, i.e. the existence of MlrA (in experimental conditions) in a dimeric form. The calculated PP1 inhibition capacity of linear MC-LR (Fig. 4D) was about 2100 fold lower than MC-LR and much lower in comparison with the IC50 value reported by Bourne et al. (1996), probably due to differences in the effectiveness of HPLC separation. Here, the separation of degradation product was complete, which was confirmed using the MS method. For this reason we postulate that common knowledge about this value should be verified which means that linear MC-LR is non-toxic in practice. Therefore, the use of bacteria with the expression of MlrA only which has access to MCs is completely sufficient to abolish the toxicity of cyanobacterial heptapeptides. Such an approach would significantly simplify the development of efficient biotechnological method for MCs removal. 5. Summary The pathway of MC hydrolysis by some bacterial strains has been previously documented. However, there were still gaps in knowledge concerning the gene encoding MlrA and characterisation of the encoded protein. This study presents

586

D. Dziga et al. / Toxicon 59 (2012) 578–586

for the first time the heterologous expression of MlrA, the crucial MCs degrading enzyme of the Sphingomonas strain ACM-3962. This successful expression allowed us to perform a mutational study and biochemical characterisation of MlrA. The results constitute an important step in studies of cyanopeptide biodegradation and open new perspectives for future research. The heterologous expression offers also new background for a detailed study of modified bacterial strains in bioremediation applications. Acknowledgements This research was supported by MNiSW, Poland (grant No 4360/B/P01/2010/39) and by UJ, Poland (grant WRBW BW/47 and BW/147). Grateful acknowledgements to David Bourne who encouraged the authors to initialize the study with Mlr proteins, to A. Black for sharing pG164 and S. aureus 178RI, to Andrzej Górecki and Piotr Bonarek for help with protein  ska for help with MS purification and to Oliwia Bochen analyses. Appendix. Supplementary material Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.toxicon. 2012.01.001. Conflict of interest None. References An, J., Carmichael, W.W., 1994. Use of a colorimetric protein phosphatase inhibition assay and enzyme linked immunosorbent assay for the study of microcystins and nodularins. Toxicon 32, 1495–1507. Bourne, D.G., Jones, G.J., Blakeley, R.L., Jones, G.J., Negri, A.P., Riddles, P., 1996. Enzymatic pathway for the bacterial degradation of the cyanobacterial cyclic peptide toxin microcystin LR. Appl. Environ. Microbiol. 62, 4086–4094. Bourne, D.G., Riddles, P., Jones, G.J., Smith, W., Blakeley, R.L., 2001. Characterisation of a gene cluster involved in bacterial degradation of the cyanobacterial toxin microcystin LR. Environ. Toxicol. 16, 523–534. Chen, J., Hu, J.B., Zhou, W., Yan, S.H., Yang, J.D., Xue, Y.F., Shi, Z.O., 2010. Degradation of microcystin-LR and RR by a Stenotrophomonas sp. strain EMS isolated from Lake Taihu, China. Int. J. Mol. Sci. 11, 896–911. Cserzo, M., Eisenhaber, F., Eisenhaber, B., Simon, I., 2002. On filtering false positive transmembrane protein predictions. Protein Eng. 15, 745–752. D’Elia, M.A., Pereira, M.P., Chung, Y.S., Zhao, W., Chau, A., Kenney, T.J., Sulavik, M.C., Black, T.A., Brown, E.D., 2006. Lesions in teichoic acid biosynthesis in Staphylococcus aureus lead to a lethal gain of function in the otherwise dispensable pathway. J. Bacteriol. 188, 4183–4189. Edwards, C., Lawton, L.A., 2009. Bioremediation of cyanotoxins. Adv. Appl. Microbiol. 67, 109–129. Gajdek, P., Lechowski, Z., Dziga, D., Bialczyk, J., 2003. Detoxification of microcystin-LR using Fenton reagent. Fresen. Environ. Bull. 12, 1258–1262.

Grutzmacher, G., Wessel, G., Klitzke, S., Chorus, A., 2010. Microcystin elimination during sediment contact. Environ. Sci. Technol. 44, 657–662. Hashimoto, E.H., Kato, H., Kawasaki, Y., Nozawa, Y., Tsuji, K., Hirooka, E.Y., Harada, K., 2009. Further investigation of microbial degradation of microcystin using the advanced Marfey method. Chem. Res. Toxicol. 22, 391–398. Ho, L., Meyn, T., Keegan, A., Hoefel, D., Brookes, J., Saint, C.P., Newcombe, G., 2006. Bacterial degradation of microcystin toxins within a biologically active sand filter. Water Res. 40, 768–774. Ho, L., Hoefel, D., Palazot, S., Sawade, E., Newcombe, G., Saint, C.P., Brookes, J.D., 2010. Investigations into the biodegradation of microcystin-LR in wastewaters. J. Hazard. Mater. 15, 628–633. Ishii, H., Abe, T., 2000. Release and biodegradation of microcystins in blue-green algae, Microcystis PCC7820. Bull. School Mar. Sci. Technol. Tokai Univ. 49, 143–157. Ishii, H., Nishijima, M., Abe, T., 2004. Characterization of degradation process of cyanobacterial hepatotoxins by a gram-negative aerobic bacterium. Water Res. 38, 2667–2676. Jones, G.J., Bourne, D.G., Blakeley, R.L., Doelle, H., 1994. Degradation of the cyanobacterial hepatotoxin microcystin by aquatic bacteria. Nat. Toxins 2, 228–235. Kato, H., Imanishi, S.Y., Tsuji, K., Harada, K., 2007. Microbial degradation of cyanobacterial cyclic peptides. Water Res. 41, 1754–1762. Manage, P.M., Edwards, C., Singh, B.K., Lawton, L.A., 2009. Isolation and identification of novel microcystin-degrading bacteria. Appl. Environ. Microbiol. 75, 6924–6928. Mazur, H., Plinski, M., 2001. Stability of cyanotoxins, microcystin-LR, microcystin-RR and nodularin in seawater and BG-11 medium of different salinity. Oceanologia 43, 329–339. Meriluoto, J., Spoof, L., 2005a. Purification of microcystins by highperformance liquid chromatography. In: Meriluoto, J., Codd, G.A. (Eds.), TOXIC: Cyanobacterial Monitoring and Cyanotoxin Analysis. Åbo Akademi University Press, Turku, pp. 93–104. Meriluoto, J., Spoof, L., 2005b. Analysis of microcystins by high performance liquid chromatography with photodiode-array detection. In: Meriluoto, J., Codd, G.A. (Eds.), TOXIC: Cyanobacterial Monitoring and Cyanotoxin Analysis. Åbo Akademi University Press, Turku, pp. 77–84. Nybom, S.M.K., Salminen, S.J., Meriluoto, J.A.O., 2008. Specific strains of probiotic bacteria are efficient in removal of several different cyanobacterial toxins from solution. Toxicon 52, 214–220. Park, H.D., Sasaki, Y., Maruyama, T., Yanagisawa, E., Hiraishi, A., Kato, K., 2001. Degradation of the cyanobacterial hepatotoxin microcystin by a new bacterium isolated from a hypertrophic lake. Environ. Toxicol. 16, 337–343. Pei, J., Mitchell, D.A., Dixon, J.E., Grishin, N.V., 2011. Expansion of type II CAAX proteases reveals evolutionary origin of g-secretase subunit APH-1. J. Mol. Biol. 410, 18–26. Rapala, J., Berg, K.A., Lyra, C., Niemi, R.M., Manz, W., Suomalainen, S., Paulin, L., Lahti, K., 2005. Paucibacter toxinivorans gen. nov.; sp nov.; a bacterium that degrades cyclic cyanobacterial hepatotoxins microcystins and nodularin. Int. J. Syst. Evol. Microbiol. 55, 1563–1568. Saito, T., Sugiura, N., Itayama, T., Inamori, Y., Matsumura, M., 2003. Degradation characteristics of microcystins by isolated bacteria from Lake Kasumigaura. J. Water SRT – Aqua 52, 13–18. Schagger, H., von Jagow, G., 1987. Tricine-sodium dodecyl sulfatepolyacrylamide gel electrophoresis for the separation of proteins in the range from 1 to 100 kDa. Anal. Biochem. 166, 368–379. Towbin, H., Staehelin, T., Gordon, J., 1979. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc. Natl. Acad. Sci. U.S.A 76, 4350–4354. Tsuji, K., Asakawa, M., Anzai, Y., Sumino, T., Harada, K., 2006. Degradation of microcystins using immobilized microorganism isolated in an eutrophic lake. Chemosphere 65, 117–124. van Apeldoorn, M.E., van Egmond, H.P., Speijers, G.J., Bakker, G.J., 2007. Toxins of cyanobacteria. Mol. Nutr. Food Res. 51, 7–60. Wang, J., Wu, P., Chen, J., Yan, H., 2010. Biodegradation of microcystin-RR by a new isolated Sphingopyxis sp. USTB-05. Chin. J. Chem. Eng. 18, 108–112.