isolation and purification of quercetin 2,3 ...

Journal of Undergraduate Chemistry Research, 2004, 2, 51

ISOLATION AND PURIFICATION OF QUERCETIN 2,3-DIOXYGENASE FROM ASPERGILLUS FLAVUS VIA LECTIN AFFINITY CHROMATOGRAPHY Jarrett Walsh*, Jonathan Long*, Delana Nivens†, Will Lynch Armstrong Atlantic State University, Department of Chemistry and Physics, 11935 Abercorn Street, Savannah, GA 31419, [email protected] Abstract Quercetinase is an enzyme excreted by the mold, Aspergillus flavus when grown on rutin. Characterization of the enzyme involves first isolating and purifying the enzyme from media in which A. flavus has been grown. Improved techniques for this isolation are presented using methods including lectin affinity chromatography and capillary electrophoresis. A pure sample of quercetinase has been obtained and found to have Km = 3.5 x 10-6M.

Keywords: Aspergillus flavus, 2,3-Dioxygenase, Lectin affinity chromatography, Quercetin, Rutin

Introduction Quercetinase, 2,3-quercetin dioxygenase (23QDO), is an extracellular metalloenzyme with a copper(II) cofactor, produced when Aspergillus flavus is grown in the presence of rutin, a glycoside of quercetin (1-3). This degradation reaction is noted to take place in a variety of microorganisms including other members of the Aspergillus genus (4,5). The catalysis reaction involves the insertion of dioxygen into the heterocyclic center ring of quercetin (3,5,7,3’,4’-pentahydroxyflavone) forming the 2-protocatechuoylphloroglucinol carboxylic acid depside and carbon monoxide, as seen in Figure 1.

Figure 1. Mechanism for the degradation of quercetin by quercetin 2,3-dioxygenase

Previous research on 23QDO has shown that it is also a glycoprotein with molecular mass of approximately 111,000 and carbohydrate content of 27.5% (3). Only recently has the active site model been determined from a crystal structure of a related enzyme from A. japonicus (5). Since 23QDO is a CuII-containing dioxygenase in all of the known species that exhibit the enzyme, it is believed that the reported model of the active site is accurate for 23QDO across species, in spite of differing molecular weights. Thus, it is expected that a similar active site structure will be discovered in the characterization of 23QDO from A. flavus compared to other identified strains.


Previous purification schemes have been published, which were heavily dependent on serial column chromatography. The literature method is workable, but requires high volumes and an excessive commitment of time to purify. In this paper, a new method for the purification of 23QDO is presented, based on lectin affinity chromatography that efficiently yields pure 23QDO. Lectin chromatography works well in this single column application as it is extremely specific (1). This chromatographic method takes advantage of the glycosides attached to the protein, allowing the binding of the enzyme to the column followed by a subsequent elution with the affinity column’s preferred substrate. The method developed here first uses uses a centrifugal filtration system to remove small molecules and preconcentrate the protein into a small volume. The resultant small volume is of high protein concentration and is loaded directly onto one affinity column. The reduction of the number of columns reduces loss of the enzyme and reduces the time required for purification. Enzymatic activity is determined spectrophotometrically by monitoring the degradation of quercetin at 368 nm. The purified enzyme was used to determine the Km value for the enzyme. Experimental Reagents All chemicals used were commercially available, of analytical grade and used as received. Quercetin, rutin, ammonium phosphate, potassium phosphate monobasic, magnesium chloride hexahydrate, ammonium sulfate, copper(II) chloride pentahydrate, zinc sulfate heptahydrate, iron(II) sulfate heptahydrate, manganese sulfate hydrate, calcium chloride, manganese chloride, sodium chloride, α-D-mannopyranoside, Concanavalin A Sepharose 4B, dimethylsulfoxide, sodium hydroxide, boric acid, were purchased from Aldrich. Bradford reagent and bovine serum albumin standard were purchased from Bio-Rad. A. flavus slants were purchased from Carolina Biologicals. Millipore concentrator tubes were purchased from Fisher Scientific. Instrumentation UV-visible spectra were recorded on a Hewlett Packard 8453 spectrophotometer. Protein concentration was determined by the Bradford assay using bovine serum albumin as the reference protein.

Sample purification during the process was analyzed by capillary electrophoresis using a Beckman P/ACE MDQ Capillary Electrophoresis System. Samples were run in 100 mM pH 10.0 boric acid buffer solution with a 30 kV charge. A UV spectrophotometer set to 280 nm was the detector. Samples throughout the purification process were also analyzed by SDS-PAGE gel electrophoresis using pre-cast Tris/glycine buffered acrylamide gels. The Mark 12TM wide range protein standard was used. Aspergillus flavus growth on rutin and collection of extracellular 23QDO A. flavus was grown for 90 hours at 30oC in a solution of salts and rutin on a shaker by modification of a previously described procedure (1). All apparatus was autoclaved prior to inoculation. The medium for production of enzyme consisted of 0.2% rutin, 0.15% (NH4)2HPO4, 0.05% KH2PO4, in 18 MΩ water. To this was added 100 mg MgSO 4.7H2O, 1.0 mg of ZnSO4.7H2O, 1.0 mg MgSO 4.7H2O, FeSO 4.7H2O, 0.75 mg MnSO4.H2O and 0.05 mg of CuSO4.5H2O in a 1 L reaction flask. Rutin was added to the second liter of water and the two flasks were autoclaved separately. After autoclaving, the contents were mixed 1:1 into 5 separate sterile flasks. Each flask was inoculated using a sterile loop with A. flavus and allowed to grow for 90 hours at 30oC. The 2 L of growth media was filtered through a sterile Nalgene fliter flask to remove unconsumed rutin and A flavus. Filtrate is then concentrated via 10,000 kDa Millipore concentrator tubes. This serves to remove small proteins, salts water and the degradation products. To the concentrated filtrate, 10 mL of 10 mM MES pH 6 buffer was added and stored at 4oC. UV-Visible determination of 23QDO sample activity Samples tested for activity were all examined in the same manner. A UV-visible spectrophotometer with kinetics package was used to determine initial rate kinetics of the degradation of quercetin in the presence of 23QDO. 200 µL of 1.2 mM quercetin dissolved in DMSO was mixed with 2.00 mL of 10 mM pH 6.0 MES buffer and 100 µL sample to be analyzed in a cuvette. The absorbance of quercetin was monitored kinetically at a wavelength of 367 nm for 100 seconds with spectra taken every four seconds. The initial rate was calculated from the resulting curve. Since the loss of quercetin absorption is being monitored, a larger negative initial rate value indicates a more active sample. An example of these


Figure 2. Degradation of quercetin by a purified 23QDO fraction. Quercetin absorbance is monitored at 368 nm over 90 seconds of reaction time after mixing. The rate is calculated by taking the initial slope of the absorbance change at 368 nm and converting it to mmoles of quercetin degraded per second through the known molar absorptivity of quercetin.

results are seen in Figure 2. Concanavalin A lectin affinity chromatography The separation of the rutinase (glycosidase) from the quercetinase is achieved by affinity chromatography on Concavalin A-Sepharose 4B. Concavalin A recognizes the Manα1-OCH 3 oligosaccharide region of 23QDO. 6 mLs of buffered concentrated filtrate (1255 U of enzyme) is added to the affinity column (2 cm by 4 cm). A gradient elution method was then separate the components. 10 mL aliquots of elution buffers were added to the column. The elution buffers contain 10 mM MES, pH 6, 1 M NaCl, and 1mM each of CaCl2, MgCl 2 and MnCl 2 and varying concentrations of α-D-mannopyranoside. The column containing the bound quercetinase was first washed with 10 mL elution buffer containing no α-D-mannopyranoside. Gradient separation was then achieved by increasing the concentration of α-D-mannopyranoside from 25 mM to 500 mM, doubling each 10 mL. The eluent was collected in 1.0 mL fractions. Each fraction was tested for activity by the above method. Three regions of quercetinase activity are found. Results and Discussion Growth of A. flavus and collection of 23QDO A. flavus was grown in a rutin-based media, following the method outlined in previous research on 23QDO (1). After several attempts to grow A.

flavus for 23QDO collection, it was noted that those with high 23QDO activity appeared to have digested most to all of the suspended rutin while growing in small nodules in the solution. The nodules were generally yellow in color, similar to rutin, while the liquid appeared brown. To remove the fungus nodules and other large suspended particles, vacuum filtration was employed. Macroscopic impurities were removed, leaving a brown liquid layer that showed activity for quercetin decomposition. The liquid layer from the filtration was transferred into Millipore TM Centrifugal Filtration Devices and centrifuged to obtain concentrated samples on the order of 100-500 µL. The fractions remaining were redissolved in 10 mM pH 6.0 MES buffer and run through the filtration devices again. After two runs through the filtration devices, the resulting liquid had an intense brown color. Purification of 23QDO Following the method of Oka et al. (3) low yields and poor purification of the 23QDO were observed. This was primarily due to a significant loss of the protein and poor fractionation of the protein from contaminants during the high number of steps involved in protein purification. We have found the serial chromatography proposed not to be suitable for smaller quantities of starting material. Thus, we designed an experiment to simplify this process greatly and help minimize active enzyme loss. Lectin chromatography was chosen for its ability to specifically select specific carbohydrate chains (6,7) Concanavalin A, specific for mannose units in the carbohydrate side chain (6), was chosen since 23QDO is known to be a glycoprotein containing bound mannose. Although the pooled concentrated fractions were likely to have other proteins in solution, the solution was directly loaded onto the column. Only proteins with mannose-containing side chains in specific configurations will be bound by Concanavalin A (6), therefore other impurities should either remain on the column or elute out of the wash. Gradient elution with methyl α-D-mannopyranoside resulted in separation of the bound proteins from the lectin. It was observed through kinetics assays that there were three major elution zones from the column, one associated with low column affinity, one with medium affinity and the final with high affinity. Similar results have been verified with general carbohydrate chains (6). As the highest activity for fractions was observed from fractions in the region


Figure 3. Capillary electrophorogram showing the isolation of 23QDO from the pooled filtrate (top) to the concentrate (middle) and finally the high affinity fraction (bottom). Graphs are offset for clarity.

of high affinity, further tests were run on samples from the area. Capillary electrophoresis on this sample was compared to samples taken in earlier steps to show the increasing purity. From Figure 3 one can see that several peaks condense to one peak in the high affinity region, verifying the procedure as valid for the purification of the enzyme. This result is also confirmed using gel-electrophoresis where one-band is observed near a molecular weight of 100,000. Table 1 outlines the various samples during the purification steps including samples from the low and medium affinity fractions. Compared to the original method (1), the method presented shows improvement by increasing fold purification to 970 and specific activity to 18,600 U/mg. Percent recovery in the fractions is calculated based on the total units (1255 U) of concentrate applied to the column. Total recovery for the entire 35 mLs of concentrate would be 71% and 14% recovery from the original filtrate.

Figure 4: Lineweaver-Burk analysis of purified 23QDO. Initial rate converted from A/sec to µM/sec via molar absorptivity of quercetin. The plot gives Km = 3.5 µM and kcat =100 sec-1. Each point is the average of two replicate analyses.

Spectrophotometric and Kinetic Analysis of 23QDO By analyzing 23QDO kinetically, more information about the purity of the enzyme may be obtained. To determine kinetic constants for 23QDO, a sample comprised of pooled active fractions from low, medium and high affinity filtrates were combined to and used to perform a standard Lineweaver-Burk analysis. The inverse initial rate was plotted against inverse concentration of quercetin where quercetin concentrations ranged from 8.7 x 10-7 to 4.3 x 10-4 M. The resulting plot allowed the calculation of values of Km = 3.5 µM and kcat = 100 giving the apparent second-order rate constant, kcat/Km = 3 x 107. The Lineweaver-Burk data is presented in Figure 4. This Km value is slightly lower than that previously reported at 5.2 µM (1).

Table 1: Purification Table for 23QDO. *Percent recovery in the fraction is calculated based on the total units (1255 U) of concentrate applied to the column.


Conclusion This paper presents a simplified and improved method for the purification of 23QDO from A. flavus. By employing centrifugal preconcentration and lectin affinity chromatography, serial columns previously employed can be avoided resulting in increased activity and speed of purification. Further the electrophoresis and Lineweaver-Burk analysis confirm that the 23QDO recovered is pure and retains its activity. Acknowledgement We would like to thank the Armstrong Atlantic State University Research and Scholarship Committee for support of this work. References (1). N. Narasimhachari, F. J. Simpson, D. W. Westlake. Can. J. Microbiol., 1963, 9, 15-21. (2). G. W. Hay, F. J. Simpson, D. W. Westlake. Can. J. Microbiol., 1961, 7, 921-932. (3). T. Oka, F. J. Simpson, C. Mills. Can. J. Microbiol., 1971, 17, 111-118. (4). H. Hund, B. Jorn, F. Lingens, H. Jurgen, K. Reinhard, S. Fetzner,. Eur. J. Biochem., 1999, 263, 871-878. (5). F. Fusetti, K. Schroter, R. Steiner, P. van Noort, T. Pijning, H. Rozeboom, K. Kalk, M. Egmond, B. Dijkstra. Structure 2002, 10, 259-268. (6). T. Endo. J. Chromatography A, 1996, 720, 251261. (7). M. Caron, A. Seve , D. Bladier, J. Raymonde. J. Chromatography B, 1998, 715, 153-161.