Expression of Zebrafish (Danio rerio) Monoamine Oxidase (MAO) in Pichia pastoris: Purification and Comparison with Human MAO A and MAO B

Abstract

The expression, purification and characterization of zebrafish monoamine oxidase (zMAO) using the methylotropic yeast Pichia pastoris expression system is described. A 1 L fermentation culture of Pichia pastoris containing the gene encoding zMAO under control of the methanol oxidase promotor expresses ~200 mg of zMAO exhibiting 300 units of total activity. The enzyme is found in the mitochondrial fraction of the expression host and is purified in a 30% yield as a homogenous species with a M_r of ~60,000 on SDS-PAGE and a mass of 58,525± 40 Da from MALDI-TOF measurements. The zMAO preparation contains one mole of covalent flavin cofactor per mole of enzyme and exhibits >80% functionality. The covalent flavin exhibits fluorescence and EPR spectral properties consistent with known properties of 8α-ScysteinylFAD. Chemical degradation of the flavin peptide results in the liberation of FAD. zMAO exhibits no immuno-chemical cross-reactivity with polyclonal anti-sera raised against human MAO A. The enzyme preparation exhibits reasonable thermostability up to a temperature of 30°C. Benzylamine is oxidized with a k_cat value of 4.7±0.1 min⁻¹ (K_m = 82 ± 9 μM) and the enzyme oxidizes phenylethylamine with a k_cat value of 204 min⁻¹ (K_m = 86 ± 13μM). The K_m (O₂) values determined for zMAO using either benzylamine or phenylethylamine as substrates ranges from 108(±5) to 140(±21) μM. The functional behavior of this teleost MAO relative to human MAO A and MAO B is discussed.

Introduction

Monoamine Oxidases A and B (MAO A and MAO B) are mammalian flavoenzymes bound to the outer mitochondrial membrane [1] and play important roles in the oxidative deamination and degradation of amine neurotransmitters. Recent studies have linked the level of MAO A expression to aggressive behavior [2] and that age-related increases in MAO B levels [3] are thought to contribute to the progression of neuro-degenerative diseases [4]. Therefore, an increased understanding of the molecular differences between these two isozymes is essential for the development of highly specific inhibitors that could function as neuroprotectants.

To aid in this effort, this laboratory has developed high level expression systems and purification protocols for the expression and purification of human and rat MAO A and of MAO B [5-7]. The rational for this work is to provide a bridge of knowledge between animal models (the rat) with the human in MAO inhibitor development. More recently, attention has been focused on zebrafish (Danio rerio) as an animal model for human disease and for drug development. Zebrafish (Danio rerio) was chosen among the possible sources of teleosts for several reasons. It is widely viewed as a genetically tractable vertebrate model system because of its 60% identity to the human genome [8, 9]. Other advantages for drug development studies include its embryonic manipulability due to the external fertilization, its accessibility for rapid genetic modification, and its rapid growth rate [10]. The serotonergic system of this teleost and its development exhibit behaviors similar to those of other vertebrates [9, 11]. All these advantages make zebrafish a good model organism as a complement to rat or mouse models used to study human diseases. It may also be useful for the development of MAO inhibitors once comparative information becomes available for zebrafish MAO relative to the human or rat MAO isozymes.

Several studies in the literature have shown that teleosts contain a single MAO gene [9-11]. These data favor an evolutionary scenario for mammalian MAO enzymes being evolved from a common single ancestor via a gene duplication event [12,13] and also provide the rationale for the work reported here in which the expression and purification of an evolutionary co-ortholog of mammalian MAO enzymes; zebrafish MAO is described.

Materials and Methods

Materials

Monoamine Oxidase cDNA from zebrafish (Danio Rerio) was obtained from Open Biosystems, (Clone ID: 6960301) and stored at −80 °C. The TOPO Cloning Kit, plasmids (pCR2.1 and pPIC3.5K), Platinum Taq polymerase, and yeast strain KM71 were purchased from Invitrogen Corp (Carlsbad, CA). Restriction enzymes and T4 DNA Ligase were provided by Promega (Madison, WI). The MiniElute Gel Extraction Kit, QIAprep Spin Miniprep Kit, and the MinElute Gel Extraction Kit were provided from Qiagen (Valencia, CA). The antibiotic G418 was purchased from US Biological (Massachusetts, MA).

Reagents used for protein purification are commercially available and included: βoctylglucopyranoside from Anatrace Inc (Maumee, OH) and reduced Triton X-100 from Fluka (Sigma-Aldrich, St. Louis, MO). Ceramic hydroxyapatite, used for column chromatography, was purchased from BioRad (Hercules, CA). DEAE-Sepharose was purchased from GE-Healthcare Biosciences. Trypsin was purchased from Promega (Madison, WI) and chymotrypsin was purchased from Roche, Waters Corp (Germany). HPLC columns and Sep-Pak C₁₈ cartridges were purchased from Phenomenex (Torrance, CA). All other chemicals were obtained from Sigma-Aldrich (St. Louis, MO) unless otherwise stated.

Methods

cDNA cloning of zebrafish MAO

The cDNA was amplified by PCR with custom primers. The sequence of the forward primer used was 5′-CCCG^GATCCATGACTGCGAACGCATACGAC-3′ which includes a BamH1 site (underlined), an in-frame start codon, and a 30-bp gene-specific sequence. The reverse primer used was 5′-GGCG^AATTCTTAACACCGTGGGAGGAGCCC-3′ which incorporated the EcoR1 site (underlined) and the translation stop codon. The PCR conditions were 95 °C for 1 min, followed by 35 cycles of 95 °C /60 °C /72 °C for 50 sec/50 sec/2 min respectively. A poly-A tail was incorporated with platinum Taq DNA Polymerase at 72 °C with a final extension time of 15min. The size of the amplified zMAO gene (1.5 kB) was identified on agarose gel electrophoresis, isolated from the gel, and directly ligated into a pCR2.1-TOPO vector using the TOPO Cloning Kit following the manufacturer's protocol. The ligation product was then transformed into TOP10 cells, a single colony isolated, grown in culture overnight, and the plasmid DNA purified using a QIAprep Spin Miniprep Kit. The gene was cut from the vector using EcoR1 and BamH1, the products separated on an agarose gel, and the zMAO DNA extracted with the MinElute Gel Extraction Kit followed by ligation into Pichia vector pPIC3.5K using T4 DNA ligase. The ligation product was then transformed into E.coli DH5α cells via electroporation. Selected colonies were grown in an ampicillin-containing LB media at 37 °C. A complete DNA sequence analysis (1566 nucleotides) by Agencourt Corp. (Beverly, MA) confirmed the correct insertion, orientation, and sequence of the zMAO gene. The pPIC3.5K/zMAO construct was linearized with SacI enzyme to target integration into the AOX1 locus of the Pichia genome. The spheroplast transformation procedure for strain KM71 was performed as described in the Invitrogen Pichia expression kit manual.

Following His⁻ selection, colonies were then transferred to plates containing different concentrations of the antibiotic G418 (0.25 to 1.25 mg/ml) for multiple gene insertion selection. Resistant colonies were detected after 4 days at 30 °C. Several G418 resistant colonies were chosen from each plate for expression screening in shake flasks as described previously [6]. Expression levels were monitored by activity assays using kynuramine as a substrate and the colony exhibiting the highest level of activity was selected for growth. A stock culture was prepared and stored at −80 °C in 20% (v/v) glycerol. Fermentation growth of the transformed organism was carried out as described previously [5]. Methanol induction of zMAO was carried out for 24 hrs rather than the 48-72 hrs used for human or rat MAO's since the induced zMAO activity levels reaches a maximum after 24 hr. and decreases on longer induction times.

Purification of zMAO

Expressed enzyme was purified using protocols similar to that published for human MAO A [5] with a number of modifications. All protein determinations were performed using the Biuret method [14] except for purified preparations where the more sensitive Bearden procedure [15] was used. Cell paste from a 1 L of culture was homogenized into 1 L of breakage buffer (50 mM sodium phosphate, 5% (v/v) glycerol, 1 mM EDTA, 1mM PMSF, pH 7.2). An equal volume of silica-zirconium beads (0.5mm diameter) were added to the homogenate with stirring and the suspension subjected to 6 cycles of cell breakage in a Biospec Beadbeater (Bartlesville, OK). Disruption involved 6 cycles of 2-min beating with 5-min ice cooling between cycles. All subsequent procedures were carried out at 4 °C unless otherwise stated. After removal of the glass beads by filtration through a layer of Miracloth (Calbiochem), the cell lysate was separated from unbroken cells and large cell debris by centrifugation at 1500×g for 10 min. The supernatant was then centrifuged at 100,000×g for 30 min to separate the membrane fraction from soluble protein. Preliminary experiments showed zMAO activity remained membrane bound as is the case with human and rat MAO A and B. Extraction of zMAO from the membrane required pre-treatment of the membranes with phospholipases prior to detergent extraction.

The membrane fraction was suspended in 0.1 M triethylamine, 25 mM CaCl_2, pH 7.2 to a protein concentration of 25 mg/ml and treated with 1 mg of phospholipase C (Sigma) and 6700 Units of phospholipase A (a partially purified preparation [16] from naja naja venom (Sigma)) per 500 mg of membrane protein. The digestion was carried out with gentle stirring in the dark for 1 hr at room temperature with the pH maintained at 7.2 by the addition of dilute NH₄OH. The digestion mixture was centrifuged at 100,000×g for 15 min and the pellet suspended in 20 mM phosphate buffer, pH 7.2 to a protein concentration of 15 mg/ml. Triton X-100 was added to a final concentration of 0.5% (w/v) and the mixture stirred at room temperature in the dark for 30 min. The mixture was centrifuged at 100,000×g for 15 min at 4 °C and the supernatant contained all detectable catalytic activity and is referred to as the Triton extract.

Preliminary trials showed zMAO activity does not bind to DEAE Sepharose or to BioRad High-Q resin in contrast to what is observed with human or rat MAO A and MAO B [5-7]. Therefore, an alternate column purification procedure was adopted. The DEAE Sepharose column did provide some purification by removing contaminating protein from the Triton extract of zMAO. The extract was loaded to a fast-flow DEAE-Sepharose column (25×600 mm) equilibrated with 20 mM potassium phosphate buffer with 0.5% (w/v) Triton X-100 and washed with one column volume of 20 mM potassium phosphate buffer (pH =7.2). zMAO binds to BioRad ceramic hydroxyapatite (CHT) columns and could be eluted in a discrete fraction using a linear phosphate gradient. The elute from the DEAE- column containing zMAO activity was combined and loaded to a CHT column which was preequilibrated with 10 mM potassium phosphate buffer containing 20% (v/v) glycerol and 0.5% (w/v) Triton X-100. The column was then washed with 10 mM potassium phosphate buffer containing 20% (v/v) glycerol, 1 mM PMSF, and containing either 0.5 %(w/v) Triton X-100 or with 0.8% (w/v) OGP until the protein concentration of the eluate dropped to negligible levels. The yellow enzyme band was eluted from the CHT column with a linear gradient of 10-300 mM potassium phosphate buffer containing 20% glycerol, 30 μM DTT, either of the above detergents and 1 mM PMSF (pH = 7.2). Either detergent (Triton X-100 or OGP) could be used in the column fractionation. We find that the stability of the enzyme to storage is greater in Triton X-100 than in OGP. OGP solutions of purified enzyme also tend to precipitate on storage at 4 °C, therefore, Triton X-100 is the recommended detergent for this enzyme. Stability to storage is also enhanced by the addition of 0.5 mM d-Amphetamine (a competitive inhibitor) in the final column step and should be removed before performing kinetic studies with the enzyme preparation. Concentration and or buffer exchange were performed using an Amicon Ultra 30K filter centrifugation apparatus. Triton, however, is not removed by these treatments whereas OGP does permeate the membrane. Optimal conditions for storage of the purified enzyme include the absence of light at 4 °C.

Spectroscopic Studies

UV/Vis absorption spectra were obtained using a Cary 50 spectrophotometer. Fluorescence spectra were measured using a AMINCO-Bowman Series 2 spectrofluorimeter interfaced to a PC. Anaerobic conditions were achieved by ten alternate cycles of vacuum with purified argon flushes of the enzyme solutions in custom designed anaerobic quartz cuvettes. X-Band EPR spectra were obtained using a Bruker E-500 spectrometer using 4 mm OD quartz tubes. Photoreduction of the purified enzyme to its semiquinone form were carried out in an anaerobic quartz cuvette at 15 °C in using eight 15W fluorescent lamps and a temperature-controlled glass holder.

Steady-State Kinetic Measurements

All enzyme assays were carried out spectrophotometrically in 50 mM potassium phosphate buffer (pH= 7.4), 0.5% (w/v) reduced Triton X-100 using a Perkin-Elmer Lambda-2 UV-Vis spectrophotometer at 25 °C. Kynuramine was used as substrate for routine assays during purification. Activity assays were performed by following the change in absorbance at 316 nm with time for the product absorbance (4-hydroxyquinoline, ε_M = 12,000 M⁻¹cm⁻¹) [17]. One unit activity is defined as the amount of enzyme required catalyzing the formation of 1 μmole 4-hydroxyquinoline per minute. An Amplex red/peroxidase-coupled spectrophotometric assay was used to follow H₂O₂ formation using phenylethylamine as substrate as described previously [18]. Benzylamine oxidation was also measured spectrally by following the rate of benzaldehyde formation (ε_{M 250} = 12,800 M⁻¹ cm⁻¹). The IC₅₀ values for irreversible inhibitors Deprenyl and Clorgyline were obtained by preincubating the inhibitor with the purified enzyme for 10 min on ice and assaying the remaining activity using kynuramine as a substrate. Oxygen concentrations were determined polarographically using a Strathkelvin Model 782 instrument interfaced to a PC. Concentrations of O₂ in assays were varied by mixing buffer solutions saturated with pure oxygen with those purged with N₂. Steady state kinetic data were analyzed by non-linear fits to the Michaelis-Menton Equation using Origin (OriginLab Corporation, Northampton, MA) or Prism 5.0 (GraphPad Software).

Mass Spectrometry

MALDI-TOF mass spectral measurements were performed using a Bruker Ultraflex-II TOF/TOF instrument. Mass spectra of intact protein were performed on samples purified from salts and detergents either by HPLC on C₄ columns or after extraction from a SDS-gel matrix and mixed with a sinapinic acid matrix for MALDI measurements. zMAO sequence analyses were performed on peptides obtained from an in-gel tryptic digest of zMAO. The in-gel digestion protocol suggested by Pierce Chemical Co. (Rockford, IL) was followed. Peptide extracts from the gel were dried in a Speed-Vac and desalted (C₁₈ ZipTip, Millipore). Approximately 1 μL of the eluted peptide solution was spotted with an equal volume of α-cyano-4-hydroxycinnamic acid in 70% (v/v) acetonitrile/water on the MALDI target. All mass spectra were obtained using the positive ion reflective mode.

Results

Enzyme expression and purification

Since previous work in our laboratory showed that high levels of rat or human MAO A [5] and MAO B [6] are expressed using the Pichia system, a similar protocol was followed with zMAO as it shares ~70% sequence identity with human MAO A and B. The fermentation growth of the transformed organism exhibits similar behavior as observed in the expression of the other enzymes. The major difference is that the methanol induction period for zMAO is considerably shorter than that used for the mammalian MAO's since the level of zMAO catalytic activity was found to reach a maximum at 24 hr and to decrease rapidly with longer induction periods. This behavior probably represents a higher lability of expressed zMAO to intracellular proteases than are exhibited by the mammalian enzymes.

Organelle fractionation of the Pichia host was performed to determine if zMAO (as are the mammalian MAO's) is a membrane-bound mitochondrial enzyme [19]. All of the expressed catalytic activity is found in the mitochondrial fraction demonstrating that zMAO is also a mitochondrial enzyme tightly bound to, presumably, the outer membrane. Approximately 200 mg of zMAO is purified from 125 g wet weight of cells with a 30% yield as shown in Table 1. The specific activity of the purified enzyme (1.5 U/mg; using kynuramine as a substrate) is similar to values exhibited by purified preparations of human or rat MAO A [5, 7]

Table 1.

Purification Scheme for zebrafish MAO expressed in Pichia pastoris from a 1 L culture.

Purification Step	Total Protein (mg)	Total Activity (Ua)	Specific Activity (U/mg)	% Yield
Cell Lysate	13619	1275	0.09	100
Membrane Fraction	7496	1043	0.14	82
Phospholipase Digestion	7268	958	0.13	75
Triton X-100 Extract	2440	912	0.37	72
DEAE-Sepharose	1450	758	0.52	60
Pooled CHT column fractions	205	310	1.52	25

^a1U = 1 μmol kynuramine oxidation product per minute at 25 °C in air-saturated 50 mM KPi, pH=7.4 with 0.5% (w/v) reduced Triton X-100.

A major difference between zMAO and the mammalian MAO's is its inability to bind to either weak or strongly anionic ion exchange matrices. zMAO does bind strongly to the absorptive hydroxyapatite on ceramic medium. The enzyme fractionates well on this matrix and is eluted from the column as a homogeneous peak on application of a linear phosphate gradient. The inclusion of a DEAE-Sepharose step prior to the hydroxyapatite column results in a 1.4-fold purification with a >80 % yield of activity and appears to be helpful for the subsequent hydroxyapatite fractionation step.

Characterization of Purified zMAO

Purified recombinant zMAO preparation displays a single band on SDS-PAGE with a mobility similar to that of human MAO A (or of MAO B) (Figure 1A) with a M_r value of ~60,000 Da. A similar mass is determined on subjecting the purified protein to MALDI mass spectral analysis (m/z=59,200 ±40; average of 3 determinations, data not shown). The recombinant protein is yellow in color, a characteristic property of flavoenzymes [20] Western blot analysis using antisera specific for covalent flavin (Figure 1B) [21] and the failure to observe any release of flavin on trichloroacetic acid treatment shows that zMAO contains a covalent flavin cofactor as does human MAO A and B. Despite exhibiting a ~70% sequence identity with hMAO A, the enzyme does not cross react immunochemically with rabbit anti-sera raised against hMAO A.

Figure 1.

SDS-polyacrylamide gel mobility of purified zMAO A. Coomassie-stained protein bands. Molecular weight marker (lane 1), purified zMAO (lane 2), and human MAO A (lane 3) B. Western blot using antisera specific for covalent flavin coenzymes. Lane 1 is purified zMAO, lane 2 is purified human MAO A. The arrow defines a mobility corresponding to 60 kD.

MALDI_TOF-MS of tryptic peptides of purified zMAO was carried out to determine the fidelity of the expressed protein sequence with that expected from the gene sequence. These data are shown in Table 2. From the 64 predicted tryptic cleavage sites, 22 peptides are identified and their sequences determined resulting in a 62% sequence determination for the purified protein. The peptide sequences; relative to the entire deduced sequence of zMAO, is shown at the bottom of Table 2. Photochemical instability of the flavin peptide to MALDI conditions probably accounts for our inability to detect the peptide site for covalent flavin attachment using this technique.

Table 2.

zMAO tryptic peptide sequences from MALDI mass spectral data and analysis using Profound software (http://www.prowl.rockefeller.edu).

Monoisotropic Masses		Error	Residues		Missed	Peptide
Observed	Computed	(Da)	Start	End	Cut	Sequence
805.009	804.377	0.63 2	126	132	0	MGMEIPKI
852.925	852.423	0.50 2	44	50	0	TYTVQNK
1031.63 9	1031.581	0.05 8	486	495	0	NLPSVGGFLK
1037.73 5	1037.624	0.11 1	414	421	0	VLREPVGR
1082.55	1081.482	1.06 8	138	146	0	APHAEEWDK
1122.40 7	1122.337	0.07	364	371	0	RICEIYARV
1141.59 7	1141.378	0.21 9	147	155	0	MTMQQLFDK
1209.62 7	1208.578	1.04 9	199	209	0	IFSTTNGGQER
1332.78 9	1332.568	0.22 1	273	283	0	KIHFNPELPPLR
1536.75 8	1536.802	0.44	211	224	1	KFAGGANQISEGMAR
1410.76 3	1409.864	0.89 9	260	272	0	YVILAIPPGLNLK
1445.72 6	1444.592	1.13 4	83	94	0	VNEEESLVHYVK
1589.80 0	1588.790	1.01 1	464	479	0	LLVDSGLNPVVLEAR
1594.95 8	1594.872	0.08 6	23	37	0	LLVDSGLNPVVLEAR
1646.87 3	1646.823	0.05	54	68	0	WVDLGGAYIGPTQNR
1862.95 9	1863.054	0.09 5	372	387	0	VLGSEEALYPVHYEEK
1915.45 5	1915.284	0.17 1	496	515	0	FMGVSSFLAAATAAGLVACK
2181.06 8	2180.444	0.62 4	1	22	0	MTANAYDVIVIGGGISGLSAAK
2567.83 3	2567.06	0.77 3	235	249	0	AVCSIDQTGDLVEVRTVNEEVYK
2645.03 5	2643.208	1.82 7	102	121	1	GPFPPMWNPFAYMDYNNLWRT
2724.32 1	2724.217	0.10 4	422	446	0	LYFAGTETATEWSGYMEGAVQAGER LHASQIWQSEPESMDVPARPFVTTFW
3345.85 0	3345.716	0.13 4	458	485	0	ER

1	mtanaydviv	igggisglsa	akllvdsgln	pvvlearsrv	ggrtytvqnk	etkwvdlgga
61	yigptqnril	riakqygvkt	ykvneeeslv	hyvkgksypf	kgpfppmwnp	faymdynnlw
121	rtmdkmgmei	pkeapwraph	aeewdkmtmq	qlfdkicwtr	sarrfatlfv	nvnvtsephe
181	vsalwflwyv	kqcggtmrif	sttnggqerk	fagganqise	gmarelgdrv	klsravcsid
241	qtgdlvevrt	vneevykaky	vilaippgln	lkihfnpelp	plrnqlihrv	pmgsvikcmv
301	yykenfwrkk	gycgsmviee	edapigltld	dtkpdgsvpa	imgfilarks	rklanltrde
361	rkrriceiya	rvlgseealy	pvhyeeknwc	eeeysggcyt	ayfppgimtq	fgrvlrepvg
421	rlyfagteta	tewsgymega	vqagerasre	vmcamgklha	sqiwqsepes	mdvparpfvt
481	tfwernlpsv	ggflkfmgvs	sflaaataag	lvackkgllp	rc

Alignment of sequenced zMAO tryptic peptides with deduced amino acid sequence of zMAO. The recovered peptides are shown in bold letters. The proposed site for covalent flavin attachment is underlined and highlighted in gray.

Determination of Nature of Flavin Cofactor

Mammalian MAO's that have been investigated all contain covalent FAD cofactors with the site of attachment being a thioether linkage to a cysteinyl residue in a conserved site in the proteins. The sequence of zMAO also exhibits this conserved site at Cys406 (Table 2). Although it is likely that zMAO would also contain an 8α-S-cysteinylFAD, supporting data to validate this assumption is important. It is known that covalent amino acid linkages at the 8-α position of the flavin ring found in nature include an N-linked histidine (either N-1 or N-3), an O-tyrosyl bond as well as the S-cysteinyl bond [22]. It should also be noted that a 6-S-cysteinyl FMN is found in the bacterial trimethylamine dehydrogenase [23] and that “bicovalent” flavin cofactors where the flavin ring is substituted with amino acids at both the 8α- and the 6 positions are found in several enzymes [24-26]. Those enzymes determined to have this type of “bicovalent” flavin have a N-1 histidyl linkage at the 8-α methylene group and a cysteinyl –S-thioether linkage at the 6 position of the flavin. With several known exceptions (covalent FMN levels in trimethylamine dehydrogenase [23] and in tetrameric sarcosine oxidase [22]), covalent flavin cofactors are generally found to be in their FAD forms for the majority of enzymes (containing covalent flavins) investigated.

The visible absorption spectrum of zMAO is shown in Figure 2A and exhibits spectral properties typical of a flavoenzyme with no evidence for any additional chromophoric cofactors. Denaturation of “bicovalent” flavin-containing enzymes results in spectral properties more typical of 6-S-substituted flavins with absorption maxima ~400 nm rather than the characteristic 450 nm absorption peak of un-substituted or 8-α substituted flavins. Denaturation of zMAO with 3M guanidine HCl results in a visible spectrum with a maximum at 450 nm (data not shown) which is consistent with substitution only at the 8α position since 6-substitution of the isoalloxazine ring perturbs the transition moment dipoles of the flavin far greater than substitution at 8-α.

Figure 2.

A. Absorption spectra of zMAO in its oxidized form (—), after photoreduction to a mixture of its anionic semiquinone and hydroquinone forms (…), and after air reoxidation (----). The sample was dissolved in 50 mM potassium phosphate, 0.5 mM EDTA, pH 7.2 with a 5-deazaflavin concentration of 0.5 μM. B. X-Band EPR spectrum of the photoreduced zMAO sample from 2A. Spectral conditions: 2 mW power, 2 gauss modulation amplitude, temperature = 160°. The spectrum is the average of 200 scans. The arrows denote the maxima and minima corresponding to the peak-to-peak separation referred to in the text.

It is known that the EPR spectral properties of flavoenzyme anionic semiquinones containing 8-α substituents exhibit narrower peak-peak linewidths than found for flavoenzymes containing un-substituted flavins (12 gauss rather than 16 gauss) [27]. zMAO was reduced to its semiquinone form using the light-EDTA-5-deazaflavin procedure [28]. As shown in Figure 2A, the spectrum of the enzyme after irradiation is characteristic of an anionic flavin semiquinone. The enzyme semiquinone form was transferred anerobically to a 4 mm quartz EPR tube and the X-band EPR spectrum determined as shown in Figure 2B. The observed EPR spectrum occurs at g=2.0 and exhibits a peak-to-peak linewidth of 12 gauss; characteristic of anionic 8α-flavoenzyme semiquinones [27]. These data provide additional evidence for the site of flavin attachment in zMAO to be the 8α-methylene of the isoalloxazine ring. Air-reoxidation of the photoreduced enzyme preparation results in the return of the absorption spectrum of its original form (Figure 2A). No evidence was found for the formation of any absorption spectrum indicating the formation of any 6-mercaptoflavin as is found for those enzymes containing a 6-S-cysteinyl flavin linkage [23,24] undergoing a similar treatment.

The above experiments point to the flavin in zMAO as an 8-α substituted flavin cofactor but do not identify the nature of the 8-α substituent. If the flavin is a cysteinyl flavin, as in mammalian MAO's, a direct approach is to examine the fluorescence properties of the flavin peptide before and after performic acid oxidation [29]. The fluorescence emission spectral data in Figure 3 show that the untreated tryptic flavin peptide exhibits little or no detectable fluorescence. Treatment of the flavin peptide fraction with performic acid at 0 °C and removal of excess oxidizing agent results in an increase in peptidyl flavin fluorescence to a level ~30% that of an equivalent amount of riboflavin (Figure 3). These data are consistent with the flavin bound to the protein via an 8α-thioether linkage (non-fluorescent), which is oxidized to a sulfone form by peracid treatment (fluorescent). The level of observed fluorescence emission of the oxidized flavin peptide is less than that of riboflavin and is probably due to the quenching effect of the adenine moiety if the coenzyme were at the FAD level. Model compounds show the cysteinylsulfone form of 8α-substituted flavins are eliminated on reduction of the flavin by 2 electrons [29]. Therefore, if the peptide flavin has such a linkage, one would predict elimination of the 8α-peptide substituent on reduction of the performic acid oxidized material. This degradative treatment would allow for determination of the cofactor nature of the flavin since the peptide moiety would no longer be present.

Figure 3.

Fluorescence emission spectra of tryptic zMAO flavin peptide before (…) and after performic acid oxidation (—). The emission spectrum of an equal concentration of riboflavin (----) is shown for comparison. An excitation wavelength of 345 nm was used. All samples were dissolved in glass distilled water.

If FAD were to be the product of the reductive elimination, it can readily be detected specifically by reconstituting apo-glucose oxidase activity. Neither FMN nor 8α-peptidyl FAD analogues are capable of reconstituting apo-glucose oxidase activity so the assay is highly specific. Using apo-glucose oxidase prepared according to published procedures [30] intact flavin peptide incubation results in no reconstitution of catalytic activity whereas the product from zMAO flavin peptide oxidation and reductive elimination did reconstitute activity. Control experiments show no reconstitution of glucose oxidase catalytic activity with FMN incubation but does occur with FAD. Our conclusion from this line of experiments is that the covalent flavin cofactor of zMAO is an 8α-S-cysteinylFAD with the site of attachment likely to be at Cys 406.

Thermal Stability of the Purified Enzyme

The thermal stabilities of hMAO A and hMAO B has been previously reported [5,6] and show that hMAO A is more labile than is hMAO B. Our aim was to determine and compare the temperature stability of zMAO in comparison with the purified mammalian forms and to determine enzyme stability to the varying conditions of subsequent experimental protocols. zMAO was incubated at various temperatures and the level of catalytic activity determined. The data in Figure 4 show that the pure enzyme is relatively stable at 30 °C for at least 60 min and subsequently loses only ~20% of its initial activity. At 40 °C the enzyme rapidly loses activity. Overall, these data shows that zMAO exhibits a reasonable thermostability, however, it should be noted that the observed behavior in Figure 4 is expected to be dependent on the nature of the detergent used for enzyme solubilization.

Figure 4.

Thermal stability of purified zMAO catalytic activity. The buffer medium is 50 mM potassium phosphate, pH 7.2 containing 0.8% (w/v) OGP. Kinetic assays were measured at 25 °C using kynuramine as substrate as described in the experimental section.

Catalytic Behavior and Inhibition of Recombinant zMAO

To provide catalytic information on zMAO, we investigated its behavior with the classic “MAO B” substrates, benzylamine and phenylethylamine. Either amine is oxidized by zMAO (Table 3) although at quite different rates. zMAO oxidizes benzylamine with a k_cat value of 4.7 min⁻¹ and a K_m value of 82 μM. Extending the carbon side chain of the amine side chain by one methylene group, phenylethylamine, results in a substrate that is oxidized by zMAO with a k_cat value of 200 min⁻¹ and a K_m value of 86 μM (Table 3). Thus, zMAO oxidizes phenylethylamine with a 40-fold higher rate than benzylamine and with similar K_m values. This differential activity with these two substrates demonstrates the functional behavior of zMAO to be more similar to human MAO A which oxidizes phenylethylamine with a k_cat value 25-fold higher than the corresponding value with benzylamine [31, 32]. In contrast, human MAO B oxidizes benzylamine with a slightly slower k_cat value than it does phenylethylamine (ratio = 0.76) [33]. When the concentration of O₂ in the assay medium is varied, a K_m (O₂) value of 108 μM is found with benzylamine as substrate while a value of 140 μM is observed with phenethylamine. Human MAO A is known to exhibit a very low K_m (O₂) (~10μM) [34], while MAO B has a K_m (O₂) of 240 μM (near air saturation). Thus, as regards to the K_m (O₂) properties, the catalytic behavior of zMAO is more similar to that of MAO B [6].

Table 3.

Steady state kinetic parameters of zebrafish MAO. All assays were performed in 50mM potassium phosphate with 0.5 %(w/v) reduced Triton X-100, pH =7.4, at 25 °C

Kinetic Parameter	Benzylamine	Phenylethylamine
Km (μM) a	82.2 ± 9.0	86 ± 13
kcat (min−1) a	4.7 ± 0.1	203.8 ± 8.7
kcat /Km (min−1 μM−1)	0.057 ± 0.005	2.37 ± 0.19
Km(O2) (μM)	108 ± 5	140 ± 21
kcat (min−1) (at ~1mM [O2])	5.6 ± 0.7	324 ± 14

^aValues determined at 240 μM O₂ (air saturation)

Previous studies in the literature [35] show that membrane bound preparations of zMAO exhibits similar sensitivities to the MAO A selective inhibitor, Clorgyline; and the MAO B selective inhibitor, Deprenyl. The data in Figure 5A show that recombinant zMAO also exhibits this behavior with IC₅₀ values for these two acetylenic MAO inhibitors (6.4 × 10⁻⁵ M for Clorgyline and 6.5 × 10⁻⁶ M for Deprenyl) more similar to one another than exhibited by the mammalian enzymes. The spectral data in Figure 5B shows that incubation of purified zMAO with either acetylenic inhibitor results in the formation of spectral species that are consistent with those of N(5)-flavocyanine adducts. The time course for the irreversible loss of activity parallels the formation of these spectral forms and the rates observed for Deprenyl and for Clorgyline with zMAO are essentially identical (t_½= ~10 min) at a molar ratio of inhibitor to enzyme of 10:1. Thus, the purified, recombinant form of zMAO exhibits properties similar to those exhibited by membrane preparations of the enzyme from its natural source.

Figure 5.

Inhibition of zMAO by Clorgyline and by Deprenyl. A. IC₅₀ values of zMAO for the irreversible inhibitors Clorgyline (—) and Deprenyl (-----); values are 6.4 × 10⁻⁵ M and 6.5×10⁻⁶ M, respectively. Activities were measured at 25 °C using kynuramine as substrate. B. Absorption spectra of purified zMAO before (—) and after the addition of a 10-fold molar excess of Deprenyl (…..) and of Clorgyline (-----). The samples were dissolved in 50 mM potassium phosphate, pH 7.4 containing 0.8% (w/v) OGP.

Discussion

This paper presents the first high-level expression, purification, and partial characterization of an evolutionary precursor of mammalian MAO enzyme. The results described here demonstrate that Pichia pastoris functions as an efficient expression system for zMAO as it does for the mammalian MAO's. Several changes are required in the purification procedure of zMAO relative to the mammalian enzymes. The inability of zMAO to bind to ion-exchange resins is most likely a result of differences in pI values of zMAO relative to the mammalian forms. The amino acid sequence of zMAO predicts a pI value of 8.9 whereas the sequences of human MAO A and B predict pI values of 7.9 and 7.2, respectively. Therefore, the column binding behavior suggests that the surface charge of zMAO is less negative than either of the human enzymes, which do bind to anionic exchange columns.

As found with the mammalian enzymes, zMAO also contains a covalently attached flavin cofactor, which exhibits spectral and chemical properties identical with those of 8α-S-cysteinylFAD. The sequence identity about the sites for covalent attachment strongly supports Cys406 in zMAO as the site for flavin attachment. This property of monoamine oxidases appears to be conserved from teleosts to mammals. Sequence comparisons of zMAO with human MAO A or MAO B shows the substrate binding domain of zMAO is identical with that of MAO A and 70% identical with that of MAO B. zMAO exhibits an 80% sequence identity with the flavin binding domains of MAO A or of MAO B. The C-terminal membrane-binding domain of zMAO exhibits the largest divergence with a 30% identity with MAO A and a 20% sequence identity with MAO B [35, 36]. Based on these sequence considerations, one might predict zMAO to be closer to human MAO A than to human MAO B.

The limited functional studies presented here support the proposal that zMAO is more similar to MAO A than to MAO B. The MAO-specific acetylenic inhibitors, Deprenyl and Clorgyline, are bound with similar rates and IC₅₀ values by recombinant zMAO. This behavior suggests zMAO possesses little differential sensitivity to these acetylenic inhibitors than exhibited by either MAO A and MAO B. The relative rates of benzylamine and phenethylamine catalysis as expressed in k_cat values demonstrate the functional behaviour of zMAO is closer to that of MAO A than to that of MAO B. This functional similarity does not appear to be present on comparison of the K_m (O₂) value of zMAO with those of MAO A and MAO B. In this aspect, zMAO appears to be more similar to MAO B than to MAO A. The thermal stability behavior of zMAO is also closer to that of MAO B than to MAO A.

These data suggest that teleost MAO exhibits functional properties that overlap those of both MAO A and of MAO B. Since the crystal structures of both the human enzymes are now known [37, 38], and have been invaluable in deciphering the differential functional behaviors of the two isozymes, future structural studies of zMAO should be of value in providing a molecular rationale for the overlapping functional behaviors documented in this report.

Abbreviations used

MAO

monoamine oxidase

hMAO

human monoamine oxidase

zMAO

zebrafish monoamine oxidase

MALDI-MS

matrix assisted laser desorption/ionization mass spectrometry

EPR

electron paramagnetic resonance

OGP

βoctyl- D-glucopyranoside

HPLC

high performance liquid chromatography