Probing the Molecular Mechanisms in Copper Amine Oxidases by Generating Heterodimers

For some homodimeric copper amine oxidases (CuAO), there is suggestive evidence of differential activity at the two active sites implying potential cooperativity between the two monomers. To examine this phenomenon for the Arthrobacter globiformis CuAO (AGAO), we purified a heterodimeric form of the enzyme for comparison with the homodimer. The heterodimer comprises an active wild-type monomer and an inactive monomer in which an active-site tyrosine is mutated to phenylalanine (Y382F). This mutation prevents the formation of the trihydroxyphenylalanine quinone (TPQ) cofactor. A pETDuet vector and a dual fusion tag strategy was used to purify heterodimers (WT/Y382F) from homodimers. Purity was confirmed by western blot and native PAGE analyses. Spectral and kinetic studies support the view that whether there are one or two functional monomers in the dimer, the properties of each functional monomer are the same, thus indicating no communication between the active sites in this bacterial enzyme.

For some homodimericc opper amine oxidases (CuAO), there is suggestive evidence of differential activity at the two active sites implying potentialc ooperativity between the two monomers. To examinet his phenomenon for the Arthrobacter globiformis CuAO (AGAO), we purified ah eterodimeric form of the enzyme for comparison with the homodimer.T he heterodimer comprises an active wild-type monomer and an inactive monomer in which an active-site tyrosine is mutated to phenylalanine( Y382F). This mutation prevents the formation of the trihydroxyphenylalanine quinone (TPQ) cofactor.ApETDuet vector and ad ual fusion tag strategy wasu sed to purify heterodimers (WT/Y382F) from homodimers. Purity was confirmed by western blot and native PAGE analyses.S pectral and kinetic studies support the view that whethert here are one or two functional monomersint he dimer,the properties of each functional monomer are the same, thus indicating no communication between the active sites in this bacterial enzyme.
Copper amine oxidases( CuAOs) [E.C. 1.4.3.6] have been studied since 1929. [1] They are homodimers with subunit sizes ranging from 70 to 120 kDa. [2] Each subunit contains as ingle posttranslationally derived protein cofactor,2 ,4,5-trihydroxyphenylalanineq uinone (TPQ), and am ononuclear type II cupric ion centre. The biogenesis of TPQ in CuAOs is catalysed by a single turn-over reactionf rom an active-site tyrosine residue in the presence of Cu II and molecular oxygen. [3] CuAOs catalyse the oxidative deamination of primary amines to their corresponding aldehydes with ac oncomitantr elease of NH 3 and H 2 O 2 through reductivea nd oxidative half reactions.
Previous studies have demonstrated that CuAOs are only functional in their dimeric form;this suggests that dimerisation affords structural and/orf unctional integrity. [4] Further studies have provided evidencef or communication betweenthe subunits in the dimer in some CuAOs, though this evidencec annot be regarded as definitive. [5] The initial evidencec ame from the results of titration of the TPQ cofactor with hydrazine-based mechanistic inhibitors. [5a] These studies demonstrated that, in some CuAOs,s uch as pea seedling amine oxidase (PSAO) and lentil seedling amine oxidase( LSAO), two TPQs were readily titratable;t his is consistent with no communication between the two active sites. [5b, 6] There are conflicting reports regarding differential reactivity between the two TPQsi nb ovine serum amine oxidase (BSAO). [5c-e] Morpurgo and colleagues report differentialr eactivity of the two TPQ'si nt he BSAO dimer. [5d,e] By contrast, forh ighly purified BSAO, Janes and Klinmand emonstrated titration of up to 0.9 TPQ/subunitw ithc orresponding enzyme activity,t hereby indicating that each subunit contains active TPQ although, as the authors state, this does not necessarily mean that under steady-state conditions each subunit is catalytically competent. [5c] Choi et al. reported at itration of 0.62 TPQ/subunit with phenylhydrazine for Arthrobacter globiformis amine oxidase( AGAO). However,t hey were unable to say whether less than one TPQ per subunit is formed or if there is reduced reactivity of one of the TPQ'si nt he dimer as ac onsequence of modification of the other TPQ by the inhibitor. [5f] In other amine oxidases such as pig plasma amineo xidase (PPAO) [5a] and Aspergillus nidulans amine oxidase (ANAO), [4] differentialr eactivity between the two TPQs has been reported, thus suggesting the possibility of communication between the active sites. In support of these observations, structurals tudies of Escherichiac oli amine oxidase (ECAO) in complex with 2-hydrazinopyridine (2-HP), revealed that one of the TPQ sites reacted much more quickly than the second site, which required prolonged exposure to 2-HP to react. [7] These structural data correlated with solution studies that demonstrated that 1.5 TPQs could be titrated with molare quivalent amountso f2 -HP. However,t wo TPQs per dimer could be titrated following incubation with a2 0-fold molar excesso fi nhibitor. [7a] This implied that ap opulation of TPQ was initially inaccessible and raised the possibility of negative cooperativity between the two active sites. The most obvious, and significant,i nteraction between the monomersi nt he CuAO dimer occurs through two long b-hairpin structures:a rm Ia nd arm II. Arm Ic onsists of residues conserved across the CuAOe nzyme family and it extends from one monomer to form an etwork of hydrogen bonds in close proximity to the active site of the other monomer ( Figure 1). [2f] Mutagenesis studies on A. globiformis histamine oxidases howed that altering ac onserved aspartate to glutamine-D403Q (= D383 in AGAO)-increased the turnover rate of the enzyme. [8] Furthermore, characterisation of as imilar residue in Hansenula polymorpha amine oxidase-1 (HPAO-1)-E406 (= D383 in AGAO, Figure 1)-revealed that the hydrogenbond network has ar ole in both TPQ biogenesis and catalysis. [9] There is no evidencea tt his time forar ole for these interactions in cooperativity between subunits in AGAO.
Mutating the residues associated with the inter-subunit interactions within one monomer of the dimer might afford a better understanding of the relationship between dimerisation and enzyme activity in CuAO through studying heterodimeric forms of the enzyme.T he most common method used to form heterodimers of enzymes with differentially mutateds ubunits involves dissociating two forms of homodimer,f ollowed by random re-association of monomers giving rise to three dimer combinations: the two originalh omodimers plus heterodimers. [10] However,r eported heterodimer yields have been low. Recent studies have demonstrated more efficient methods for obtaining higher yields of the desired heterodimer. [10] Ohgari et al. coexpressed wild-type and mutant proteins by using the pETDuet system to give 25 %w ild-type dimer,5 0% heterodimer and 25 %m utant dimer yield. Castellani et al. used ad ualaffinity-tag system to isolate heterodimer; this ensured low cross contamination of the heterodimerw ith either homodimer. [11] In order to study the reported cooperativity in CuAOs more rigorously,w ec ombined these methods. Initially we tested the E. coli enzyme, but this provedt echnically challenging, so we developed ac oexpression and purification system for heterodimeric forms of AGAO as am odel systemf or this proof-of-principle study.W ed emonstrate heret he use of ac oexpression vector,p ETDuet-1,t oi solate heterodimers comprising differentially tagged wild-type monomer andi nactive Y382F monomer, in which the TPQ precursor Tyr382 is mutated to phenylalanine, are combined (Scheme 1).
Methods described by Ohgari et al. and Juda et al. were adaptedf or the construction and expression of AGAO heterodimers. [11b, 12] Briefly,aC-terminal Strep-tag II version of the wild-type agao coding region (provided by Prof. David Dooley, University of Rhode Island) and aC -terminal His 6 -tagged agao Y382F coding region were subcloned into multiple cloningr egions 1a nd 2, respectively,t og ive plasmidp ETDuet-agaoWT/ Y382F.E xpression of protein from pETDuet-agaoWT/Y382F results in three potential dimer species (Scheme 1).
To isolate heterodimeric WT/Y382F,atwo-step affinity purification procedurew as employed:a ni nitial Cu 2 + -immobilised affinity chromatography step to remove WT dimers and then aS trepTactin affinity chromatography step to remove Y382F homodimers. Wild-type AGAO (WT/WT) and the inactive mutant (Y382F/Y382F) were purified separately as experimental controls. The purification of heterodimeric or homodimeric populations together with experimentally mixed homodimers allowed for direct and rapid confirmation by western blot analysis of the subunit composition. Immunoblotting with an anti-Strep tag II antibody detected both WT/WT and WT/Y382F AGAO dimers, whereas Y382F/Y382F and WT/Y382F AGAO dimers wered etected by an anti-polyhistidine antibody.I mmunoblotting for each tag (Strep tag II and His 6 tag) independently confirmed the presence and purity of heterodimeric WT/ Y382F AGAO ( Figure 2).  Theoretical calculations of isoelectricp oint (pI, ExPASy; http://web.expasy.org/compute_pi/) indicated that each dimer would have ad ifferent pI indicative of changes to the overall charge due to the fusion tags (WT/WT pI = 5.07, WT/Y382F pI = 5.15 and Y382F/Y382Fp I=5.23). [13] Furthermore, under alkaline conditions the Strep tag II (WSHPQEK) is protonated whilst the His 6 tag (HHHHHH) is deprotonated. Subsequenta nalysisb y native PAGE at pH 8.8 demonstrated differential migration patterns for these three protein dimers, thereby confirming the predicted differences in overall protein charge and migration characteristics of the three AGAO dimers and confirming purification of the heterodimer (Figure 3). Following prolonged incubation of heterodimer protein samples with subsequent native PAGE established that there was no random dissociation and reassociationo ft he dimers as only as ingle band was observed for WT/Y382F AGAO with no evidence of the appearance of homodimers. As the native PAGE and immunoblot data demonstrated successful isolation of the heterodimeric WT/Y382F,s pectral ands teady-state kinetic studies were undertaken. The presence of protein-derived cofactor,T PQ,p rovides ac lear spectral signature with l max = 480 nm. [3b] UV-visible spectra of purified WT/WT,W T/Y382F and Y382F/Y382F AGAO dimers revealed an absorption peak at 480 nm for WT/ WT and WT/Y382F samples but not for the Y382F/Y382F AGAO variant, which cannotf orm TPQ. The molar extinction coefficient for WT/Y382F (1752 m À1 cm À1 )i sh alf that of WT/WT (3437 m À1 cm À1 ); this indicates a5 0% lower TPQ content. The absence of shifts in l max in WT/Y382F compared with WT/WT indicates that there is no apparent change in the electronic structure of the TPQ in the heterodimer ( Figure 4A).
To support the UV-visible data, the WT/WT,W T/Y382F and Y382/Y382F dimers were incubated with a2 0-foldm olar excess of 2-HP to react with the TPQ in each protein. 2-HP is a mechanism-based inhibitor of CuAOs (Scheme 2) [5a] that gives the yellow coloured complex, adduct I. [14] Ac hange in solution colour from pink (unreacted TPQ at 480 nm) to yellow (adduct I complex at 420 nm) was only observed with WT/WT and WT/Y382F AGAOs,b ut not with Y382F/Y382F,a se xpected (Figure 4B).
Spectrophotometric quantification of adduct I in equal quantities of protein revealed that heterodimeric WT/Y382F AGAO (0.072 a.u) formed only half the quantity of adduct I than the WT/WT AGAO (0.145 a.u) homodimer did. Post-translational modification of Y382 to TPQ in the WT subunit of the WT/ Y382F AGAO heterodimer indicates that the Y382F mutation of in one monomer hasn oe ffect upon TPQ formation in the other.
To quantify the content of TPQ in active WT/WT,i nactive Y382F/Y382F and heterodimeric WT/Y382F AGAOs,t he enzymes were titrated with 2-HP.T his indicated 1.5 TPQs per dimer of the WT/WT enzyme andi sc onsistent with the 1.4 TPQs per dimer reported previously. [12] Many questions have been raised about the failure to detected one TPQ per monomer in many CuAOs.
Does this population exist as TPQo ra nu nreactive precursor?Does the conformation of the active site prevent apopulation reacting or does inter-subunit communication lead to a conformation change in the neighbouringa ctives ite?I nh eterodimeric WT/Y382F AGAO, 0.8 TPQs were titrated per dimer and, as expected from the results with excess2 -HP,n oT PQ was detected upon titration of the inactiveh omodimer Y382F/   [15] Ta king this into account,i ts eems most likely that about2 0% of the precursor tyrosine fails to undergo full processing to form the cofactor TPQ, resulting in the detection of lesst han one TPQ in the heterodimer WT/Y382F.
The catalytic activity of AGAO was assessed in ap eroxidasecoupled assay, as described in Chiu et al., [16] with b-phenylethylamine (b-PEA)a st he substrate. The K M values for b-PEA of WT/WT andW T/Y382F AGAOs are similar( 1.3 AE 0.1 vs. 1.6 AE 0.2 mm,respectively). Oxidation of b-PEA also displays substrate inhibition in 100 mm HEPES (pH 7.0), [17] and therefore as imilar trend was observed upon comparison of the K i values of the WT/WT and WT/Y382F AGAOs (198.2 AE 35.7 vs. 208.2 AE 26.9 mm, respectively;T able 1). [17] This suggestst hat inactivating one monomer has no effect on either substrate binding or substrate inhibition in the active monomer.T he turnover rate (k cat ) of WT/Y382F AGAO per dimer is 65.5 AE 1.9 s À1 ,w hich is approximately half that of WT/WTA GAO( 129.5 AE 5.1 s À1 ). Although heterodimeric WT/Y382F AGAOe xhibits half the total catalytic efficiency of the WT/WT AGAOh omodimer,d ue to one inactive subunit (4 10 7 vs. 9.96 10 7 m À1 s À1 ,r espectively), the catalytic efficiencies per wild-type monomer in heterodimer or homodimer AGAO are similar( Table 1).
The steady-state kinetic data are in agreement with the TPQ titration data thus indicating that inactivating one active site in AGAO has no effect on the reactivity of as econd functional active site. The TPQ titration data imply that inefficiency in TPQ biogenesis results in the formation of less than two TPQs in AGAO. However, we cannot exclude the possibility that ap opulation of TPQ is inaccessible to 2-HP.T hat inactivating one subunit has no significant effect on the activity of the active subunit suggestst hat, in AGAO, the subunits act independently of each other.T his suggests that replacing the TPQ in one subunit with phenylalanine does not appear to change the hydrogen-bond network involved in interactions with the b-hairpin, and, consequently,t he activityi nt he other subunit is unaffected. The independent catalytic activity of each subunit suggestst hat dimerisation is likely to confer structural stabilityo nA GAO. However, continued study of other heterodimers variants will allow for more detailed investigations into the function of dimerisationinA GAO and in other CuAOs.
In conclusion, we report purification of the first heterodimeric form of aCuAO. The data suggest that the subunits are catalyticallyi ndependenti nA GAO, thus there is no subunit cooperativity.H eterodimeric forms of CuAOs will allow the study of differentially mutated subunit combinations within the dimer and will furtherf acilitate the dissection of structural features associated with enzyme function.
Protein expression and purification: The AGAO enzymes were expressed as described in Juda et al. [12] All the purification steps were carried out at 4 8C. Post-lysis, CuSO 4 (50 mm)w as added to the lysates, and they were incubated at 30 8Cf or 1h with shaking at 100 rpm. This step was performed to ensure the full processing of tyrosine to TPQ. Excess CuSO 4 was removed by dialysing the lysate against binding buffer phosphate-buffered saline (PBS, 140 mm NaCl, 2.68 mm KCl, 10 mm Na 2-HP O 4 ,2 m mKH 2 PO 4 ,p H7.4) or sodium phosphate (20 mm containing 300 mm NaCl, pH 7.4). Wildtype AGAO (WT/WT) was purified by using Strep-Tactin chromatography. [12] The column was prepared according to the manufacturer's protocol (Strep-Tactin Sepharose 50 %s uspension, IBA, Gçttingen, Germany). The lysate was syringe filtered prior to loading the Strep-Tactin column. The protein-bound column was washed with binding buffer to remove unbound and non-specific binding proteins. The enzyme was eluted with PBS supplemented with [D]desthiobiotin (5 mm). Homogeneous fractions were pooled and dialysed against HEPES (50 mm,pH7).
The inactive mutant (Y382F/Y382F) was purified by using Cu 2 + -immobilised metal affinity chromatography.T he dialysed lysate was filtered through a0 .2 mms yringe filter to remove any remaining cell debris. Prior to sample loading, the column was charged with NiSO 4 (0.1 m)and equilibrated with sodium phosphate (20 mm containing 300 mm NaCl, pH 7.4). After the sample had been loaded, the column was washed with sodium phosphate (20 mm containing 300 mm NaCl, 40 mm imidazole, pH 7.4). Bound protein was eluted with sodium phosphate (20 mm containing 300 mm NaCl, 250 mm imidazole, pH 7.4).Homogenous fractions were then pooled and dialysed against HEPES (50 mm,p H7).
The heterodimer (WT/Y382F) was isolated according to at wo-step affinitypurification procedure that involved both Cu 2 + -immobilised affinityc hromatography and Strep-Tactin affinity chromatography. The column was charged with CuSO 4 (0.1 m), washed to remove unbound Cu 2 + and equilibrated with sodium phosphate (20 mm), NaCl (300 mm)p H7.4. The protein-bound column was washed with sodium phosphate (20 mm containing 500 mm NaCl, 40 mm imidazole, pH 7.4). Elution was carried out using sodium phosphate (20 mm containing 500 mm NaCl, 50 mm EDTA, pH 7.4). One fraction was collected, analysed by SDS-PAGE and dialysed against PBS in preparation for Strep-Tactin chromatography.T he purification procedure was similar to that used for WT/WT.T he purity and homogeneity of the eluted protein was assessed by SDS-PAGE, and the protein was dialysed overnight against HEPES (50 mm,p H7).
Western blot: After being separated on aS DS-PAGE (15 %) gel, the proteins were transferred to two identical polyvinylidene fluoride (PVDF) membranes (Immobilon, Millipore) by using an XCell II module apparatus (Invitrogen). Following transfer,t he membranes were blocked in dried skimmed milk (Marvel, 5%, w/v)i nT ris-buffered saline (150 mm NaCl, 50 mm Tris·HCl, pH 7.5) containing Tween-20 (TBS-T,0 .1 %) and bovine serum albumin (BSA, 5%, w/v) in TBS-T for 1h.O ne membrane was then incubated overnight at 4 8CinTBS-T with dried skimmed milk (5 %, w/v)and supplemented with the monoclonal anti-polyhistidine peroxidise conjugate antibody (Sigma);t he other membrane was incubated in BSA (5 %, w/v)a nd TBS-T containing the primary anti-Strep tag II antibody (Novagen). After incubation, the unbound antibody was washed off in TBS-T.T he second membrane was then incubated for 1h in BSA (5 %, w/v)a nd TBS-T supplemented with ah orseradish peroxidase-conjugated secondary antibody (Sigma) for 1h.T he unbound secondary antibody was washed off with TBS-T.T he proteins were detected by using chemiluminescence.
Native PAGE: The traditional Tris-glycine system was used;t he proteins were separated by using an ative polyacrylamide gel (12 %). Electrophoresis was carried out at 4 8Ca nd 200 Vf or 1-2 h, and the proteins were detected by staining with Coomassie Brilliant Blue R.
2-HP was prepared at am olar concentration ten times that of the dimer.T PQ was titrated stepwise by adding 2-HP (0.1 equiv). Changes in absorbance were accounted for by correcting for the dilution (1 %) at each addition of 2-HP.F ollowing addition of 2-HP, the reactions were allowed to proceed until no detectable change in absorbance was observed.