Báo cáo khoa học: Tryptophan tryptophylquinone cofactor biogenesis in the aromatic amine dehydrogenase of Alcaligenes faecalis Cofactor assembly and catalytic properties of recombinant enzyme expressed in Paracoccus denitrificans

Tryptophan tryptophylquinone cofactor biogenesis in the aromatic amine dehydrogenase of Alcaligenes faecalis

Cofactor assembly and catalytic properties of recombinant enzyme expressed in Paracoccus denitrificans

Parvinder Hothi, Khalid Abu Khadra, Jonathan P. Combe, David Leys and Nigel S. Scrutton

Manchester Interdisciplinary Biocentre and Faculty of Life Sciences, University of Manchester, UK

Keywords amine oxidation; aromatic amine dehydrogenase; cofactor biogenesis; stopped-flow spectroscopy; tryptophan tryptophyl quinone

Correspondence N. S. Scrutton, Manchester Interdisciplinary Biocentre and Faculty of Life Sciences, University of Manchester, Stopford Building, Oxford Road, Manchester, M13 9PT, UK Fax: +44 161275 5586 Tel: +44 161275 5632 E-mail: nigel.scrutton@manchester.ac.uk

(Received 15 August 2005, revised 19 September 2005, accepted 22 September 2005)

The heterologous expression of tryptophan trytophylquinone (TTQ)- dependent aromatic amine dehydrogenase (AADH) has been achieved in Paracoccus denitrificans. The aauBEDA genes and orf-2 from the aromatic amine utilization (aau) gene cluster of Alcaligenes faecalis were placed under the regulatory control of the mauF promoter from P. denitrificans and introduced into P. denitrificans using a broad-host-range vector. The physical, spectroscopic and kinetic properties of the recombinant AADH were indistinguishable from those of the native enzyme isolated from A. faecalis. TTQ biogenesis in recombinant AADH is functional despite the lack of analogues in the cloned aau gene cluster for mauF, mauG, mauL, mauM and mauN that are found in the methylamine utilization (mau) gene cluster of a number of methylotrophic organisms. Steady-state reaction profiles for recombinant AADH as a function of substrate concen- tration differed between ‘fast’ (tryptamine) and ‘slow’ (benzylamine) sub- strates, owing to a lack of inhibition by benzylamine at high substrate concentrations. A deflated and temperature-dependent kinetic isotope effect indicated that C-H ⁄ C-D bond breakage is only partially rate-limiting in steady-state reactions with benzylamine. Stopped-flow studies of the reduc- tive half-reaction of recombinant AADH with benzylamine demonstrated that the KIE is elevated over the value observed in steady-state turnover and is independent of temperature, consistent with (a) previously reported studies with native AADH and (b) breakage of the substrate C-H bond by quantum mechanical tunnelling. The limiting rate constant (klim) for TTQ reduction is controlled by a single ionization with pKa value of 6.0, with maximum activity realized in the alkaline region. Two kinetically influential ionizations were identified in plots of klim ⁄ Kd of pKa values 7.1 and 9.3, again with the maximum value realized in the alkaline region. The poten- tial origin of these kinetically influential ionizations is discussed.

Aromatic amine dehydrogenase (AADH) is a trypto- phan tryptophylquinone (TTQ)-dependent quinopro- tein that catalyses the oxidative deamination of a wide range of amines to their corresponding aldehydes and

ammonia [1]. Electrons released upon substrate oxida- tion are transferred to the TTQ cofactor (Fig. 1) and then to the physiological electron acceptor, azurin, which mediates electron transfer from the dehydro-

doi:10.1111/j.1742-4658.2005.04990.x

FEBS Journal 272 (2005) 5894–5909 ª 2005 FEBS

5894

Abbreviations AADH, aromatic amine dehydrogenase; aau, aromatic amine utilization; DCPIP, dichlorophenol indophenol; KIE, kinetic isotope effect; MADH, methylamine dehydrogenase; mau, methylamine utilization; ORF, open reading frame; PES, phenazine ethosulfate; TTQ, tryptophan tryptophylquinone.

[7]. Thus,

structural disulfide bonds

functional of AADH cannot be obtained by cloning and expressing the two structural genes in a heterologous host in the absence of the TTQ biosynthesis genes. Heterologous expression of Paracoccus denitrificans TTQ-dependent MADH has been achieved in Rhodobacter sphaeroides by using a broad-host-range vector incorporating the MADH structural genes (mauA and mauB) and the additional genes (mauFEDCJG) required for TTQ bio- genesis [8]. The genes were placed under the regulatory control of the coxII promoter which, unlike the native mau promoter, was not controlled by methylamine lev- els [8]. Aspartate residues in the active site of MADH have been identified for their role in TTQ biogenesis [9]. Also, the dihaem c-type cytochrome mauG [10] is known to (a) initiate the TTQ crosslink in MADH, (b) convert a single hydroxyl group on Trp57 of the small subunit to a carbonyl group, and (c) insert a second oxygen atom into the TTQ ring [11]. The essential nat- ure of some of the genes in the mau gene cluster of P. denitrificans (mauF, mauE, mauD and mauG) has been shown [12,13]; other genes in the cluster (mauR, mauC, mauJ, mauM and mauN) are not essential for TTQ biogenesis [12,14,15].

P. Hothi et al. Cofactor biogenesis in TTQ-dependent AADH

(A) Structure of tryptophan tryptophylquinone (TTQ)

genase to a c-type cytochrome [2,3]. Oxidation of sub- strate proceeds via a pathway that involves the release of two electrons. Time-resolved crystallographic studies have provided structures for a number of intermediates along the reaction pathway (M.E. Graichen, L.H. Jones, B.V. Sharma, R.J. van Spanning, J.P. Hosler & V.L. Davidson, unpublished results). AADH is known to adopt a a2b2 structure (a, 40 kDa; b, 12 kDa) [1,4], highly similar to the related TTQ-dependent methyl- amine dehydrogenase (MADH) [5]. Each b subunit contains a covalently bound TTQ prosthetic group (Fig. 1), which is formed by post-translational modifi- cation of two gene-encoded tryptophan residues [6].

The mechanism by which TTQ biosynthesis occurs is not well known. The biosynthesis of AADH from Alcaligenes faecalis requires a number of additional genes, not present in Escherichia coli, as well as those that encode the large and small protein subunits (aauB and aauA, respectively) [7]. The accessory gene prod- ucts are required for protein export to the periplasm, synthesis of the TTQ prosthetic group, and formation

TTQ-dependent quinoproteins are important model systems for studies of enzymatic hydrogen tunnelling [4,16,17]. An understanding of the factors that drive tunnelling reactions in TTQ-dependent enzymes requires detailed structural, kinetic and mutagenesis studies. High-resolution crystal structures of AADH and a num- ber of reaction intermediates have been reported (M.E. Graichen, L.H. Jones, B.V. Sharma, R.J. van Spanning, J.P. Hosler & V.L. Davidson, unpublished results), but a source of recombinant enzyme for mutagenesis studies has not been made available. With this in mind, we have developed a system for the heterologous expression of recombinant AADH exploiting P. denitrificans as host. The aauBEDA genes and orf-2 from the aromatic amine utilization (aau) gene cluster of A. faecalis were placed under the regulatory control of the mauF promo- ter of P. denitrificans and introduced into P. denitrifi- cans by using a broad-host-range vector. This leads to the synthesis of active recombinant AADH that requires the cooperation of TTQ biogenesis genes from the mau gene cluster. By performing detailed kinetic studies of both AADH enzymes, we show that the recombinant enzyme is indistinguishable from the native AADH of A. faecalis and benzylamine is a substrate during steady-state reactions of AADH, contrary to previous reports using native AADH. In stopped-flow kinetic studies of TTQ reduction with benzylamine, we identi- fied ionizable groups in the enzyme–substrate complex that control the rate of TTQ reduction. One of these

FEBS Journal 272 (2005) 5894–5909 ª 2005 FEBS

5895

for Fig. 1. AADH isolated from A. faecalis. (B) The reductive half-reaction of AADH. I, oxidized enzyme; II, substrate carbinolamine intermediate; III, iminoquinone intermediate; IV, product Schiff base intermediate; V, aminoquinol intermediate. In the oxidative half-reaction the ami- noquinol intermediate is converted back to the oxidized enzyme by electron transfer to azurin and elimination of ammonium.

groups is tentatively assigned to the active site aspartate residue that accepts a proton from the iminoquinone intermediate formed in the reductive half-reaction of the catalytic cycle.

P. Hothi et al. Cofactor biogenesis in TTQ-dependent AADH

Results

Expression of recombinant AADH

Fig. 2. Purification of recombinant AADH from P. denitrificans mon- itored by SDS ⁄ PAGE analysis. Lane 1, molecular mass markers as follows: phosphorylase b (97 kDa), albumin (66 kDa), ovalbumin (45 kDa), carbonic anhydrase (30 kDa), trypsin inhibitor (20.1 kDa) and a-lactalbumin (14.4 kDa); lane 2, crude cell extract; lane 3, fol- lowing DE-52 ion exchange chromatography; lane 4, phenyl Seph- arose chromatography; lane 5, pure AADH following gel-filtration chromatography.

Table 1. Purification of recombinant AADH from Paracoccus deni- trificans.

Total activity (units) Total protein (mg) Purification step Yield (%) Specific activity (unitsÆmg)1)

Cell extract Ammonium sulphate 3668 2580 1427 1172 0.38 0.45 100 82 fractionation

P. denitrificans

lacking

Plasmid pRKAADH was introduced via conjugation into P. denitrificans to test for AADH expression. The level of recombinant AADH produced under the con- trol of the mauF promoter and subsequently purified, when grown on methylamine as a sole carbon source, was 52 mg of pure enzyme per 100 g of cells. This is approximately twice the level of native AADH gener- ally produced by A. faecalis grown on b-phenylethyl- amine. Given that P. denitrificans is known to express MADH when grown on methylamine [18], the peri- plasmic extract of P. denitrificans transformed with pRKAADH contained MADH as well as recombinant AADH. The TTQ-dependent enzymes were easily separated by ion-exchange chromatography (second step of the purification procedure described in Experi- mental procedures). Fractions containing AADH were eluted from the DE-52 cellulose column with 200 mm NaCl, whereas MADH fractions were eluted with 400 mm NaCl. AADH was assayed with tryptamine as described in Experimental procedures and MADH was assayed with methylamine. AADH fractions were highly active with tryptamine but methylamine was a poor substrate. MADH fractions were highly active with methylamine and completely inactive with trypt- plasmid amine. When pRKAADH was grown on methylamine, no AADH was detected. MADH expression was similar in wild- type P. denitrificans and pRKAADH containing P. denitrificans.

Characterization of recombinant AADH

enzyme. The purification of

(corresponding to a and b subunits)

identical

for the small and large subunits, respectively). (The published nucleotide sequence for the aau B gene is incorrect [7]. The predicted mass is based on the cor- rected sequence presented in Supplementary Material.) N-Terminal sequence analysis of both native and recombinant b subunits indicated that the first six resi- dues are Ala-Gly-Gly-Gly-Gly-Ser. The mature protein product is therefore truncated by 47 amino acids com- pared with the conceptual protein sequence inferred from the gene sequence, consistent with removal of the periplasmic localization sequence (see Supplementary material). We were unable to obtain N-terminal sequence for recombinant and native a subunit, sug- gesting that the sequence is N-blocked. However, unlike the b subunit (which we infer lacks sufficient surface protonatable residues for analysis by ESMS in the scanned mass range), we were able to obtain a mass for the a subunit by electrospray ionization mass spectrometry. The mass obtained for both native and recombinant a subunits were 40 421 Da. This mass correlates with cleavage of the a subunit at the

During the purification of recombinant AADH, the elution conditions during ion-exchange chromatogra- phy, hydrophobic interaction chromatography and gel to those observed for the filtration were identical native recombinant AADH is illustrated in Fig. 2 and summarized in Table 1. The recombinant enzyme migrates as two in subunits SDS ⁄ PAGE and migration is to that observed for the native enzyme. The migration of both subunits is consistent with the predicted masses of the mature form of the subunits (14 472 and 40 421 Da

FEBS Journal 272 (2005) 5894–5909 ª 2005 FEBS

5896

DE52 chromatography Phenyl Sepharose Sephacryl S-200 gel filtration 128 58 17 1012 884 524 7.9 15.2 30.8 71 61 37

the large subunit

effect; KIE ¼ 4.4). Rate constants for recombinant AADH were 1.52 ± 0.01 and 0.36 ± 0.01 s)1, for protiated and deuterated benzylamine, respectively (KIE ¼ 4.2). These parameters are similar to those obtained during single wavelength studies of the reduc- tive half-reaction (Table 2).

predicted site for removal of the periplasmic localiza- tion sequence (i.e. cleavage prior to residue Gln26 with expected mass of cleaved subunit is 40 438 Da; see Supplementary material). As is N-blocked, we infer the N-terminal glutamine residue has cyclized to form the pyrollidone. This brings the expected mass of the a subunit to 40 421 Da, which is within error of the measured mass of 40 422 Da.

Stopped-flow studies of the reductive half-reaction of AADH

Studies of the reductive half-reaction were performed to allow comparison of kinetic parameters for native and recombinant AADH. Reduction of the TTQ cofactor by benzylamine (or deuterated benzylamine) was followed at 456 nm on rapid mixing of enzyme with substrate. The plot of observed rate constant against benzylamine concentration for recombinant AADH is shown in Fig. 3B. Fitting of the standard hyperbolic expression to the data revealed that kinetic parameters for the recombinant enzyme are compar- able with parameters obtained for the native enzyme (Table 2).

constants

Photodiode array detection revealed that spectral changes accompanying enzyme reduction (for both enzymes) were best described by a one-step kinetic model A fi B by global analysis (Fig. 3C). Spectrum a is the oxidized enzyme and spectrum b is the reduced for native AADH were enzyme. Rate 1.64 ± 0.01 and 0.37 ± 0.01 s)1, for protiated and deuterated benzylamine, respectively (kinetic isotope

P. Hothi et al. Cofactor biogenesis in TTQ-dependent AADH

using performed (reaction 1 lM enzyme

FEBS Journal 272 (2005) 5894–5909 ª 2005 FEBS

5897

Fig. 3. Spectral and kinetic properties of recombinant AADH. (A) Spectral changes accompanying the titration of oxidized enzyme with substrate. AADHox (6.5 lM), in 10 mM BisTris propane buffer (pH 7.5), was reduced by the addition of benzylamine under anaer- obic conditions at 25 (cid:1)C. (B) Stopped-flow kinetic data for the reaction of recombinant AADH with benzylamine and deuterated benzylamine. Filled circles, protiated benzylamine-dependent activ- ity; open circles, deuterated benzylamine-dependent activity. Reac- tions were cell concentration) in 10 mM BisTris propane buffer, pH 7.5, at 25 (cid:1)C. Transients were measured at 456 nm. Observed rates were obtained by fitting to a standard single exponential expression. The fits shown are to the standard hyperbolic expression. (C) Photo- diode array studies of enzyme reduction. AADHox (4 lM) contained in 10 mM BisTris propane buffer, pH 7.5, was rapidly mixed with 200 lM protiated or deuterated benzylamine (reaction cell concen- trations) at 25 (cid:1)C. Spectral changes accompanying enzyme reduc- tion are as in Fig. 3A. Spectral intermediates were identified by fitting to a one step kinetic model. Spectrum a is the oxidized enzyme and spectrum b is the reduced enzyme. Similar data to those in (A–C) were obtained for the native enzyme (not shown).

P. Hothi et al. Cofactor biogenesis in TTQ-dependent AADH

Table 2. Kinetic parameters determined from stopped-flow reactions of native and recombinant AADH. Parameters were obtained by least squares fitting of data to the standard hyperbolic expression.

Enzyme KIE Protiated benzylamine klim (s)1) Deuterated benzylamine klim (s)1) Protiated benzylamine Kd (lM) Deuterated benzylamine Kd (lM)

pH dependence of TTQ reduction with benzylamine and deuterated benzylamine

klim ⁄ Kd at pH 10 (Fig. 4D, inset) needs to be inter- preted with caution owing to the poor buffering capa- city of BisTris propane at this pH. Analysis of the data (omitting the pH 10 data point) using the equa- tion for a double ionization yielded a pKa1 value of 7.1 ± 0.2 and pKa2 value of 9.3 ± 0.2 for protiated benzylamine. With deuterated benzylamine the corres- and pKa2 values were pKa1 7.0 ± 0.2 ponding 11.1 ± 0.4.

Plots of initial velocity vs. benzylamine concentration for steady-state reactions of AADH

substrate

equation

initial

to

Initially, we attempted pH dependence studies of TTQ reduction by substrate using a three buffer system (e.g. Mes, TAPSO and diethanolamine) of constant ionic strength [19]. However, we demonstrated that this was not possible owing to rapid reduction of the TTQ cofactor by the buffer components. Also, as noted pre- viously, univalent cations stimulate spectral changes in AADH, particularly at higher pH values [20] and our own studies revealed a similar trend at higher pH values with various univalent cations (data not shown). Thus, owing to a rapid loss in absorbance at 456 nm (attributed to chemical modification of the TTQ to form an hydroxide adduct) [20], it was not feasible to perform pH-dependent studies of TTQ reduction in the presence of added salt (although studies performed in the presence of 100 mm NaCl at lower pH values yielded results comparable with those obtained in the absence of NaCl). pH-dependence studies were there- fore performed as described in Experimental proce- dures. Given that single buffers were used to determine pH profiles, points were confirmed by overlapping the pH ranges of the different buffers. Owing to limita- tions in the buffering range of BisTris propane, studies in the alkaline region were not extended beyond pH 10. Alternative buffers that might be employed in the alkaline region (e.g. sodium borate) were avoided to reduce complications from cation induced adduct formation.

Benzylamine can reduce AADH and function as an reductive half-reaction effective in the [17,21]. Hyun and Davidson, however, have re- ported that benzylamine-dependent activity is barely detected during steady-state reactions of AADH (kcat < 0.01 s)1) [21]. Our steady-state analyses, performed with native and recombinant enzyme, revealed that benzylamine and deuterated benzylamine are signifi- cantly better substrates during steady-state turnover reactions than suggested by previous studies. A plot of initial velocity against benzylamine concentration for recombinant AADH is shown in Fig. 5A. Appar- ent Michaelis constants were determined by fitting the Michaelis–Menten velocity data and apparent Michaelis constants were found to be similar for native and recombinant enzymes (Table 4). Also, steady-state kinetic parameters are comparable with kinetic parameters determined from stopped-flow studies of the reductive half-reaction (Table 2). The KIE observed during steady-state reac- tions with benzylamine [(cid:1) 2.5 in the presence of 1 mm phenazine ethosulfate (PES) and (cid:1) 2.0 with 5 mm PES] is deflated compared with the KIE observed during stopped-flow studies ((cid:1) 4.5 in the absence of PES). An observed KIE of (cid:1) 2.0 suggests that C-H ⁄ C-D bond breakage is partially rate limiting during steady-state reactions employing benzylamine as substrate.

The origin of

A typical example of data collected is presented in Fig. 4A, as well as the plot of Kd vs. pH (Fig. 4B), klim vs. pH (Fig. 4C) and the plot of klim ⁄ Kd vs. pH (Fig. 4D). Limiting rate constants for TTQ reduction and Kd values at different pH values are summarized in Table 3. Fitting of the equation describing a single ionization to the data shown in Fig. 4C yielded pKa values of 6.0 ± 0.1 (protiated benzylamine) and 5.65 ± 0.15 (deuterated benzylamine). A plot of klim ⁄ Kd indicates the presence of at least one kinetic- ally influential macroscopic ionization in the free enzyme, and most likely the presence of two ioniza- in tions

relatively large

(Fig. 4D). The

increase

the apparent discrepancy between our work and that reported by Hyun and Davidson concerning the effectiveness of benzylamine as a sub-

FEBS Journal 272 (2005) 5894–5909 ª 2005 FEBS

5898

Native Recombinant 1.47 ± 0.01 1.54 ± 0.02 0.32 ± 0.01 0.36 ± 0.01 10.38 ± 0.25 10.8 ± 0.6 10.4 ± 0.29 12.7 ± 0.6 4.6 ± 0.2 4.3 ± 0.2

P. Hothi et al. Cofactor biogenesis in TTQ-dependent AADH

enzyme

suggesting competition between PES and benzylamine at a common binding site. This might account for, or contribute to, the apparent lack of benzylamine- dependent activity reported by Hyun and Davidson [21].

tryptamine

Effects of substrate concentration on initial velocity profiles

Previous studies have established that substrate inhibi- tion occurs during steady-state reactions of AADH

strate for AADH in steady-state reactions is unclear. However, we have observed that detection of activity with the ‘slow’ substrate benzylamine requires a sub- concentration ((cid:1) 50 nm) stantially higher than with those assays performed with the ‘fast’ sub- ((cid:1) 3 nm). Moreover, we have strate observed that AADH enzyme activity is inhibited at is 1.8 ± 0.14 mm; high concentrations of PES (Ki Fig. 6A). Increasing the PES concentration leads to an increase in the apparent Km for benzylamine (Fig. 6C) and decrease in apparent kcat (Fig. 6B),

FEBS Journal 272 (2005) 5894–5909 ª 2005 FEBS

5899

Fig. 4. The pH dependence of TTQ reduction in AADH with protiated and deuterated benzylamine. Individual parameters determined from curve fitting to plots of observed rate (kobs) against substrate concentration are shown in Table 3. (A) Data set collected at pH 8.0. (B) Plot of Kd vs. pH. (C) Plot of klim vs. pH. Inset, plot of KIE vs. pH (pKa 6.3 ± 0.2). Filled circles, protiated benzylamine-dependent activity (pKa 6.0 ± 0.07); open circles, deuterated benzylamine-dependent activity (pKa 5.65 ± 0.15). The errors associated with the pKa values are those determined from curve fitting. (D) Plot of klim ⁄ Kd vs. pH in the pH range 5–9.5. Hatched lines indicate fits to the equation for a single ionization; solid lines represent fits to the equation for a double ionization. Filled circles, protiated benzylamine dependent activity; open circles, deuterat- ed benzylamine dependent activity. Analysis of the data (omitting the pH 10 data point) using the equation for a double ionization yielded a pKa1 value of 7.1 ± 0.2 and pKa2 value of 9.3 ± 0.2 for protiated benzylamine. With deuterated benzylamine the corresponding values were pKa1 7.0 ± 0.2 and pKa2 11.1 ± 0.4. Inset, plot of klim ⁄ Kd including data collected at pH 10 and fits to the equation for a double ionization. Con- ditions: 1 lM native AADH, various buffers as described in Experimental procedures, at 25 (cid:1)C.

P. Hothi et al. Cofactor biogenesis in TTQ-dependent AADH

Table 3. Limiting rate constants for TTQ reduction and enzyme–substrate dissociation constants for the reaction of AADH with benzylamine and deuterated benzylamine at different pH values. Values of klim and Kd were determined by fitting data to the standard hyperbolic expres- sion.

H (s)1)

D (s)1)

H (lM)

D (lM)

pH KIE k lim K d k lim K d

Stopped-flow studies of the oxidative half-reaction with PES

with aromatic amines such as tyramine, b-phenylethyl- amine and tryptamine [1,22]. In a previous report, data collected with tyramine [1] were fit to the following equation:

ð1Þ

m ¼

Vmax½S(cid:2) Km þ ½S(cid:2) þ ½S(cid:2)2=Ki

where v is the initial velocity, Vmax is the maximum initial velocity, [S] is the substrate concentration and Ki is the inhibition constant for substrate. We also observed substrate inhibition with the ‘fast’ substrate tryptamine (Fig. 5B) and b-phenylethylamine (data not shown). Fitting of Eqn (1) generated poor fits to the data (Fig. 5B) and thus data collected with tryptamine as substrate were analysed using Eqn (2).

(cid:1)

(cid:2)

1 þ

Vmax

b½S(cid:2) Ki

the oxidative half- To investigate the kinetics of reaction, AADH was reduced stoichiometrically with benzylamine and rapidly mixed with different concen- trations of PES under anaerobic conditions (Fig. 7). Transients were followed at 483 nm, which is an isos- bestic point for PES but also a wavelength at which there is reasonable absorbance from the TTQ cofactor. At 1 mm PES the rate of enzyme oxidation is (cid:1) 35 s)1 at 25 (cid:1)C. At 5 mm PES the extrapolated rate of enzyme oxidation is (cid:1) 53 s)1 at 25 (cid:1)C. This is much faster than the corresponding turnover number of (cid:1) 1.2 s)1 (with 1 and 5 mm PES), suggesting that the chemistry of the oxidative half-reaction and binding of PES to enzyme is not rate limiting in steady-state turnover.

ð2Þ

m ¼

þ

1 þ

þ

Ks ½S(cid:2)

Ks Ki

½S(cid:2) Ki

Temperature dependence studies and kinetic isotope effects with benzylamine as reducing substrate

The temperature dependence of the observed KIE was investigated for reductive half-reactions and steady- state reactions of native and recombinant AADH. As shown previously [17], Eyring plots of the reductive half-reaction indicate that the KIE is independent of temperature (Fig. 8A,B), although reaction rates are strongly dependent on temperature. In contrast, Eyring plots of steady-state reactions indicate that the KIE is dependent on temperature, suggesting that C-H ⁄ C-D bond breakage is not fully rate-limiting (Fig. 8C,D). The parameters DH(cid:1) and A’H : A’D, which were found to be similar for native and recombinant enzymes,

where KS and Ki are the Michaelis and inhibition con- stants for substrate, respectively. Vmax is the theoretical maximum initial velocity and b is a factor by which the Vmax is adjusted owing to inhibition. The initial velocity profile for deuterated tryptamine was similar to the profile obtained with protiated tryptamine with a KIE close to unity indicating that C-H bond break- age is not rate limiting with ‘fast’ substrates. The lack of inhibition observed with benzylamine (Fig. 5A) in comparison to the inhibition observed with tryptamine (Fig. 5B) suggests differences in binding of the two substrates within the active site of the enzyme and ⁄ or indicates that different steps are rate limiting during steady-state reactions of AADH with ‘fast’ and ‘slow’ substrates.

FEBS Journal 272 (2005) 5894–5909 ª 2005 FEBS

5900

5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0 0.36 ± 0.01 0.56 ± 0.01 0.92 ± 0.01 1.2 ± 0.01 1.45 ± 0.02 1.53 ± 0.02 1.56 ± 0.01 1.57 ± 0.02 1.54 ± 0.02 1.52 ± 0.02 1.7 ± 0.02 381.9 ± 24 228.2 ± 8.5 118.1 ± 4.0 58.4 ± 2.4 29.87 ± 2.0 17.45 ± 0.9 14.69 ± 0.8 12.0 ± 0.8 9.97 ± 0.6 8.08 ± 0.6 3.91 ± 0.3 0.13 ± 0.01 0.18 ± 0.01 0.26 ± 0.01 0.3 ± 0.01 0.34 ± 0.01 0.35 ± 0.01 0.34 ± 0.01 0.34 ± 0.01 0.32 ± 0.01 0.31 ± 0.01 0.33 ± 0.01 418.2 ± 17 263.9 ± 14 118.3 ± 3.6 48.8 ± 1.4 31.83 ± 1.0 15.62 ± 0.9 11.92 ± 0.9 12.58 ± 0.8 10.33 ± 0.3 6.68 ± 0.3 4.11 ± 0.3 2.8 ± 0.29 3.1 ± 0.23 3.5 ± 0.17 4.0 ± 0.16 4.3 ± 0.19 4.4 ± 0.18 4.6 ± 0.16 4.6 ± 0.19 4.8 ± 0.20 4.9 ± 0.19 5.15 ± 0.26

were obtained by fitting the Eyring equation to the data and are summarized in Table 5. Analysis of the temperature dependence of reaction rates with protiated and deuterated benzylamine provides a sensitive test for the functional equivalence of native and recombin- ant AADH. We infer, therefore, that both enzymes are identical in their functional properties and that the TTQ cofactor is assembled correctly in the recombin- ant enzyme.

P. Hothi et al. Cofactor biogenesis in TTQ-dependent AADH

Discussion

The aromatic amine utilization (aau) gene cluster of A. faecalis comprises nine genes (orf-1, aauBEDA, orf-2, orf-3, orf-4 and hemE) all putatively transcribed in the same direction [7]. The second and fifth genes (aauA and aauB) encode the large and small subunits of AADH, respectively. The genes aauD and aauE are sim- ilar to mauD and mauE, respectively, from the methyl- amine utilization (mau) gene cluster, and the latter two genes are essential for MADH biosynthesis [13]. Like mauE, aauE is predicted to be a membrane-spanning protein and both aauD and mauD contain a C-X-X-C motif similar to that found in disulfide isomerases. The identity of the first open reading frame (ORF) (orf-1) in the aau gene cluster is not certain and it is not related to mauF, which is found at the corresponding position in the mau gene cluster [7]. The gene orf-2 in the aau gene cluster is predicted to be a small periplasmic monohaem c-type cytochrome. One might suppose that orf-2 is the functional counterpart of mauG a novel dihaem protein [10,11] required for TTQ biogenesis in MADH [12] even though orf-2 (aau cluster) and mauG (mau cluster) lack substantial similarity in sequence. However, insertion mutagenesis studies have indicated that orf-2 is prob- ably not involved in the oxidation of aromatic amines in A. faecalis [7]. Of the remaining ORFs, sequence simi- larity searches have failed to establish roles for orf-3 and orf-4, whereas the final gene in the cloned aau gene cluster, hemE, has 59% identity with E. coli uro- porphyrinogen decarboxylase. Here, we have described the heterologous expression of functional TTQ-depend- ent AADH by placing aauBEDA and orf-2 (directly downstream of aauA; Fig. 9) under the control of the mauF promoter and introducing these genes into P. den- itrificans using a broad-host-range vector. The success- ful production of active enzyme suggests that orf-1, orf-3, orf-4 and hemE are not required for the biosyn- thesis of AADH, consistent with there being no inferred biological function in TTQ biogenesis for the polypep- tides encoded by orf-1, orf-3 and orf-4 by comparison with gene sequences in the mau cluster [7]. The mau gene cluster for MADH contains the mauF, mauG, mauL,

FEBS Journal 272 (2005) 5894–5909 ª 2005 FEBS

5901

Fig. 5. Effects of substrate concentration on initial velocity profiles. Initial velocity vs. benzylamine concentration for steady-state (A) reactions of recombinant AADH. Assays were performed as des- cribed in Experimental procedures with 50 nM AADH and 5 mM PES in 10 mM BisTris propane buffer, pH 7.5, at 25 (cid:1)C. Filled circles, pro- tiated benzylamine-dependent activity; open circles, deuterated ben- zylamine-dependent activity. Similar plots were collected in the presence of 1 mM PES (data not shown). Apparent Michaelis con- stants were determined by fitting initial velocity data to the Michael- is–Menten equation. Similar data were also collected for native AADH (not shown). (B) Initial velocity data as a function of trypta- mine concentration. Conditions: 3 nM native AADH, 5 mM PES in 10 mM BisTris propane buffer, pH 7.5, at 25 (cid:1)C. Filled circles, proti- ated tryptamine-dependent activity; open circles, deuterated trypta- mine-dependent activity. Fits to Eqn (1) (solid line) and Eqn (2) (dashed line) are shown. Kinetic parameters determined from fitting to Eqn (1) are: kcat (s)1), 54.4 ± 2.5; Ks (lM), 1.3 ± 0.13; Ki (lM), 26 ± 3.3. Similar plots were collected in the presence of 1 mM PES (data not shown) and for recombinant AADH (data not shown).

P. Hothi et al. Cofactor biogenesis in TTQ-dependent AADH

Table 4. Kinetic parameters determined from steady-state reactions of native and recombinant AADH. Parameters were obtained by least squares fitting of data to the standard Michaelis–Menten expression.

E Enzyme [PES] (mM) KIE Protiated benzylamine kcat (s)1) Deuterated benzylamine kcat (s)1) Protiated benzylamine Km (lM) Deuterated benzylamine Km (lM)

mauM and mauN genes, but analogues for these genes are not found around the aauBEDA gene cluster [7]. We in recombinant have shown that TTQ biogenesis AADH is functional despite the lack of equivalent genes for mauFGLM in the cloned aau gene cluster. Studies have shown that mauL and mauM are not required for TTQ biogenesis, but mauG and mauF are essential [12]. The expression of active recombinant AADH in P. deni- trificans might therefore require the cooperation of some TTQ biogenesis genes (mauF and mauG) from the mau gene cluster.

groups in the immediate vicinity of the active site (M.E. Graichen, L.H. Jones, B.V. Sharma, R.J. van Spanning, J.P. Hosler & V.L. Davidson, unpublished results). Asp128b accepts a proton from substrate during breakage of the substrate C–H bond (i.e. the tunnelling reaction) [17] and thus needs to be deproto- nated in the reactive iminiquinone enzyme–substrate complex. We suggest that the ionizable group of pKa 6.0 represents the deprotonation of this aspartate resi- is oxidized to the due. Given that benzylamine corresponding aldehyde product in a reaction that requires only a single proton abstraction by Asp128b, it seems probable that the kinetically influential ioniza- tion observed in pH-dependent studies with benzylam- ine is attributable to the ionization of Asp128b.

fast

substrates

compared with

such

We have shown that the physical, spectroscopic and kinetic properties of the recombinant AADH are sim- ilar to those of the native enzyme purified from A. fae- calis. Our studies have shown that benzylamine is a substrate in multiple turnover assays and stopped-flow mixing reactions. Unlike with fast substrates (e.g. tryp- tamine and tyramine) substrate inhibition is not observed with the ‘slow’ substrate benzylamine, which likely reflects a different and less optimal mode of binding in the active site for benzylamine. The mech- anistic reasons for the smaller KIEs seen with benzyl- amine as tryptamine are not known at this stage, but barrier shape and inductive effects (e.g. through the use of per-C-deuterated benzylamine) should be considered. That TTQ reduction is partially, but not fully, rate limiting in steady-state reactions with benzylamine is consistent with (a) the suppressed KIE observed in steady-state turnover assays compared with that meas- ured by stopped-flow methods, and (b) the similarity of the limiting rate constant for TTQ reduction and the steady-state turnover value. Also, the temperature dependence of the KIE observed in steady-state assays contrasts with the essentially temperature-independent KIE observed in stopped-flow studies, which is consis- tent with TTQ reduction being partially rate limiting in steady-state turnover.

Our studies of the pH dependence of TTQ reduc- tion by benzylamine have indicated that a single kin- ionization of pKa 6.0 controls the etically influential rate of TTQ reduction. The crystal structure of AADH indicates the presence of only two ionizable

Two ionizations were also identified in the plot of klim ⁄ Kd vs. pH, which reports on kinetically influential ionizations in the free enzyme and free substrate forms. The more alkaline ionization has a pKa value identical, within error, to that of free benzylamine (pKa 9.3), and we suggest that this represents deproto- nation of the substrate benzylamine to generate the reactive, free amine form of the substrate. The more acidic ionization (pKa 7.1) is attributed to a group in the free enzyme, and we speculate this represents the ionization of Asp128b. This being the case, the effect of substrate binding would be to lower the pKa of this group to 6.0 (i.e. the value measured in the plot of klim vs. pH; Fig. 4C). The more acidic ionization in the free enzyme of pKa 7.1 has a substantial affect on the affin- ity of the enzyme for substrate. In the protonated form, the Kd for the enzyme substrate complex increa- ses substantially over that seen in the deprotonated form of the enzyme ((cid:1) 20-fold on moving from pH 5 to 7.5; Fig. 4B and Table 3). A further increase in affinity (approximately fivefold) is seen on moving from pH 7.5 to 10 (Table 3), over which pH range the substrate benzylamine is converted from the protonat- ed to free base form. Formal assignment of the observed kinetically influential ionizations must await studies with mutant enzymes and different substrates. These studies can now proceed given the availability of recombinant AADH containing a correctly assembled

FEBS Journal 272 (2005) 5894–5909 ª 2005 FEBS

5902

Native Recombinant Native Recombinant 1 1 5 5 1.14 ± 0.01 1.02 ± 0.02 1.23 ± 0.01 1.03 ± 0.02 0.46 ± 0.01 0.35 ± 0.01 0.55 ± 0.01 0.49 ± 0.02 6.7 ± 0.1 9.9 ± 0.7 14.4 ± 0.3 13.1 ± 0.5 12.7 ± 0.5 13.2 ± 0.7 12.6 ± 0.8 14.3 ± 0.7 2.5 ± 0.2 2.9 ± 0.3 2.2 ± 0.1 2.1 ± 0.2

P. Hothi et al. Cofactor biogenesis in TTQ-dependent AADH

TTQ reaction centre. Analysis of wild-type and mutant forms of AADH with a range of substrates is now in progress in an attempt to identify those residues that are responsible for the observed kinetically influential ionizations in AADH.

Fig. 7. Plot of observed rate for the oxidative half-reaction of AADH as a function of PES concentration. AADH was stoichiometrically reduced with benzylamine and rapidly mixed with different concen- trations of PES under anaerobic conditions. Conditions: AADHred (2 lM), 10 mM BisTris propane buffer, pH 7.5, 25 (cid:1)C. The mono- phasic increase in absorbance, representing oxidation of reduced enzyme, was followed at 483 nm (isosbestic point of PES). Observed rates were obtained by fitting to the standard single exponential expression.

Experimental procedures

Materials

BisTris propane buffer, 2,6-dichlorophenol indophenol; sodium salt (DCPIP), PES (N-ethyldibenzopyrazine ethyl sulfate salt), b-phenylethylamine, tryptamine and benzylam- ine were obtained from Sigma (St. Louis, MO). Deuterated benzylamine HCl (C6D5CD2NH2 HCl, 99.6%) and deuter- ated tryptamine HCl (tryptamine-b,b-d2 HCl, 98%) were from CDN Isotopes (Quebec, Canada). The chemical purity of the deuterated reagents was determined to be > 98% by high performance liquid chromatography, NMR, and gas chromatography, by the suppliers.

Growth of cells and media

The bacterial strains and plasmids used in this study are lis- ted in Table 6. For strain stocks and DNA isolation, E. coli and A. faecalis were cultured with Luria–Bertani media at 37 and 30 (cid:1)C, respectively. P. denitrificans was grown in nutrient broth or on nutrient agar at 30 (cid:1)C. For the expres-

FEBS Journal 272 (2005) 5894–5909 ª 2005 FEBS

5903

Fig. 6. Apparent inhibition of AADH as a function of PES concentra- tion. (A) Initial velocity data as a function of PES concentration. Condi- tions: 50 nM native AADH and 500 lM benzylamine (or deuterated benzylamine) in 10 mM BisTris propane buffer, pH 7.5, at 25 (cid:1)C. Filled circles, protiated benzylamine-dependent activity; open circles, deuterated benzylamine-dependent activity. The fits shown are to Eqn (1). (B) Plot of kcat for benzylamine vs. PES concentration. (C) Plot of apparent Km for benzylamine vs. PES concentration.

P. Hothi et al. Cofactor biogenesis in TTQ-dependent AADH

Construction of plasmids for the expression of AADH

for

strategy

employed

constructing

plasmid The pRKAADH for the expression of AADH in P. denitrificans is summarized in Fig. 9. pRKAADH consisted of a 3.5 kb region of the A. faecalis aau gene cluster (containing aauB, aauE, aauD, aauA and orf-2) fused to the P. denitrificans methylamine utilizing (mau) gene F promoter. First, the 3.5 kb aau fragment was amplified by PCR from A. faecalis

sion of AADH, A. faecalis and P. denitrificans were cul- tured in minimal salts media according to Iwaki et al. [23] and Davidson [18], respectively. In P. denitrificans, AADH was expressed from plasmid pRKAADH, by inducing the mauF promoter cassette with 1% methylamine 8 h prior to harvesting. When appropriate, antibiotics were added to the following final concentrations: ampicillin, 100 lgÆmL)1; kanamycin, 25 lgÆmL)1; tetracycline, 10 lgÆmL)1 for E. coli and tetracycline, 1 lgÆmL)1; rifampicin 20–50 lgÆmL)1 for P. denitrificans.

FEBS Journal 272 (2005) 5894–5909 ª 2005 FEBS

5904

Fig. 8. Eyring plots for reactions of AADH with benzylamine and deuterated benzylamine. (A) Plot of ln kobs ⁄ T vs. 1 ⁄ T for the reductive half-reac- tion of native AADH with benzylamine and deuterated benzylamine. Filled circles, protiated benzylamine; open circles, deuterated benzylamine. Inset, plot of ln KIE vs. 1 ⁄ T. Conditions: 1 lM AADH, 10 mM BisTris propane buffer, pH 7.5. Rate constants are observed rate constants meas- ured at 200 lM benzylamine. Observed rates were obtained by fitting to a standard single exponential expression. For each reaction at least four replicate measurements were collected and averaged, each containing 1000 data points. (B) As (A) but with recombinant AADH. Above 25 (cid:1)C, absorbance changes observed for the recombinant enzyme were biphasic and thus observed rates were obtained by fitting to a double expo- nential expression. (C) Plot of ln vi ⁄ T vs. 1 ⁄ T for steady-state reactions of native AADH with benzylamine and deuterated benzylamine. Filled cir- cles, protiated benzylamine; open circles, deuterated benzylamine. Inset, plot of ln KIE vs. 1 ⁄ T. Conditions: 100 nM AADH, 5 mM PES, 500 lM benzylamine, 10 mM BisTris propane buffer, pH 7.5. (D) As (C) but with recombinant AADH. In these assays, one standard deviation in each activity measurement (n ¼ 5) at a defined temperature is < 6% of the determined value. Kinetic and thermodynamic parameters were obtained from fitting data to the Eyring equation.

P. Hothi et al. Cofactor biogenesis in TTQ-dependent AADH

Table 5. Kinetic and thermodynamic parameters determined from stopped-flow and steady-state reactions of AADH with protiated and deu- terated benzylamine. Parameters were obtained from fitting the data shown in Fig. 8 to the Eyring equation.

Enzyme A’H: A’D KIE DH(cid:1)H (kJÆmol)1) DH(cid:1)D (kJÆmol)1) DDH(cid:1) (kJÆmol)1) Temp. range ((cid:1)C)

63.6 ± 0.6 62.0 ± 1.1 65.0 ± 0.7 61.4 ± 1.1 1.4 ± 0.03 0.6 ± 0.02 2.7 ± 0.07 5.3 ± 0.26 4.7 ± 0.1 4.3 ± 0.2 Reductive half-reaction Native Recombinant 4–40 4–40

Enzyme A’H: A’D DH(cid:1)H (kJÆmol)1) DH(cid:1)D (kJÆmol)1) DDH(cid:1) (kJÆmol)1) Temp. range ((cid:1)C) KIE at 24 (cid:1)C

Steady-state reactions Native Recombinant 4–40 4–40 49.6 ± 0.6 48.2 ± 1.1 55.4 ± 0.7 55.2 ± 0.9 5.8 ± 0.1 7.0 ± 0.3 0.24 ± 0.01 0.15 ± 0.01 2.5 ± 0.2 2.5 ± 0.3

bouring ORF (mauR) coding for a transcriptional activator of the mauF promoter were amplified from genomic DNA using forward (5¢-TTGGTAAGCTTGGGCATTTCTGAT CGGGTCGC-3¢) and reverse (5¢-AAACCATATGACGCC TCCTCTCGCT-3¢) primers that incorporated HindIII and NdeI restriction sites into the 5¢- and 3¢-ends, respectively. The amplification product was cloned into pCR4-TOPO for sequencing and subsequently subcloned into pKAK02 as an HindIII–NdeI fragment, generating a transcriptional fusion between the mauF promoter and aauBEDAorf-2. Finally,

genomic DNA using forward (5¢-GGAGGGATCCCATATG AAGTCTAAATTTAAATTAACG-3¢) and reverse (5¢-GC GTGCTCGAGCGATCCATGGAGCCGTA-3¢) primers that incorporated BamHI and NdeI restriction sites in the 5¢-end of the amplification product, and an XhoI site in the 3¢-end. The amplification product was cloned into the TA cloning vector PCR4-TOPO (Invitrogen, Carlsbad, CA) for sequencing and then subcloned into vector pCDNA II (In- vitrogen) as a BamHI–XhoI fragment, generating plasmid pKAK01. The P. denitrificans mauF promoter and a neigh-

FEBS Journal 272 (2005) 5894–5909 ª 2005 FEBS

5905

Fig. 9. Strategy for the construction of plas- mid, pRKAADH, used in the heterologous expression of AADH. The aau gene region of A. faecalis consists of nine genes (orf-1, aau- BEDA, orf-2, orf-3, orf-4 and hemE) [7]. Genes are represented by their correspond- ing letter, numbers denote orfs 1–4, and hE is an abbreviation for hemE. The mauF pro- moter and mauR gene from P. denitrificans are represented by mFp and mR, respect- ively. Open boxes, periplasmic proteins; sha- ded boxes, cytoplasmic proteins; diagonally hatched boxes, membrane proteins. Arrows indicate the direction of transcription.

P. Hothi et al. Cofactor biogenesis in TTQ-dependent AADH

Table 6. Bacterial strains and plasmids used in the heterologous expression of AADH in Paracoccus denitrificans.

Strain or plasmid Relevant features Source or reference

Bacteria A. faecalis P. denitrificans E. coli strain DH5 E. coli S17-1 Wild-type Wild-type, Rifr General cloning strain Conjugative donor ATCC OF1144 ATCC13543 Invitrogen Laboratory strain

tained in 10 mm potassium phosphate buffer, pH 7.5). Fractions containing AADH were pooled, concentrated by ultrafiltration, and applied to a phenyl–Sepharose column equilibrated with 10 mm potassium phosphate buffer con- taining 35% ammonium sulfate, pH 7.5. Enzyme was elut- ed using a 35 to 0% ammonium sulfate gradient. Fractions containing AADH were pooled, concentrated by ultrafiltra- tion, and dialysed against 10 mm potassium phosphate buffer, pH 7.5. The enzyme was then applied to a Sephacryl S-200 gel filtration column equilibrated with 10 mm potas- sium phosphate buffer containing 100 mm KCl, pH 7.5. AADH fractions were concentrated by ultrafiltration and dialysed against 10 mm potassium phosphate, pH 7.5. Enzyme was judged to be pure by SDS ⁄ PAGE. Purified enzyme was stored at )80 (cid:1)C in 10 mm potassium phos- phate buffer, pH 7.5, with 10% ethylene glycol.

Plasmids pCR 4-TOPO pCDNA II pKAK01 pKAK02 TA cloning vector General cloning vector pCDNAII, aauBEDAorf-2 pCDNA II, Invitrogen Invitrogen This study This study

Prior to use in kinetic studies, AADH was reoxidized with potassium ferricyanide and exchanged into 10 mm Bis- Tris propane buffer, pH 7.5, by gel exclusion chromato- graphy. Enzyme concentration was determined using an extinction coefficient of 27 600 m)1 cm)1 at 433 nm [1].

the mauF::aauBEDAorf-2 fusion was cloned as an HindIII– XbaI fragment into broad-host range vector pRK415 [24], creating pRKAADH.

Mass spectrometry

Plasmid pRKAADH was

containing 1 lgÆmL)1

transformed into E. coli strain S17-1 and conjugated into P. denitrificans using a method adapted from Graichen et al. [8]. E. coli S17-1 cells containing pRKAADH were mixed with rifampicin- resistant P. denitrificans cells in Luria–Bertani media for 2 h at 30 (cid:1)C, plated on Luria–Bertani media without antibiotic selection and then incubated for a further 6 h. Cells were scraped off the Luria–Bertani plates, washed once with Sistroms’ medium [25] and then plated on tetracycline Luria–Bertani plates and 20 lgÆmL)1 rifampicin to select for exconjugates.

ESMS was performed on a Micromass (Milford, MA) LCT time of flight mass spectrometer operating in positive ion mode. A mobile phase of 50% acetonitrile ⁄ 50% formic acid (1% in deionized water) was pumped through the spraying capillary, which was maintained at (cid:1) 3 kV. Samples were dissolved in deionized water and were introduced into the mobile phase via a Rheodyne injector. Scans were taken at the rate of 1 s per scan in the mass range 600–2000 a.m.u. Several scans were averaged to give raw data, which was further processed using maximum entropy software to pro- duce the mass spectrum.

Purification of native and recombinant AADH

Anaerobic titrations of oxidized AADH

The following procedure was performed in a Belle Technol- ogy (Portesham, UK) glove box (< 5 p.p.m. oxygen) and buffer was made anaerobic by bubbling with argon for 2 h and left to equilibrate overnight in the glove box. Anaero- bic solutions of substrate were prepared by dissolving pre- weighed solid in anaerobic buffer. Enzymes were reoxidized with potassium ferricyanide and exchanged into 10 mm BisTris propane buffer, pH 7.5, by gel-exclusion chromato- graphy. Enzymes were reduced by the addition of benzyl- amine and followed spectroscopically using a Jasco (Great Dunmow, UK) V-550 UV ⁄ Vis spectrophotometer housed in the glove box.

Steady-state kinetic analysis

For the isolation of native AADH, A. faecalis IFO 14479 was grown aerobically at 30 (cid:1)C on 0.15% (w ⁄ v) b-phenyl- ethylamine as described previously [1,23]. For the isolation of recombinant AADH, P. denitrificans transformed with pRKAADH was grown aerobically at 30 (cid:1)C on 0.3% (w ⁄ v) methylamine. The following purification procedure applies to both native and recombinant enzymes. Cells were harves- ted in the late exponential phase and resuspended in 10 mm potassium phosphate buffer, pH 7.5. Cells were broken by passage through a French pressure cell (14 000 p.s.i., 4 (cid:1)C) or by sonication. DNA was hydrolysed (DNase I) and cell debris removed by centrifugation. The cell extract was fractionated between 35 and 85% ammonium sulfate. The precipitate from the ammonium sulfate fractionation was dialysed exhaustively against 10 mm potassium phosphate, pH 7.5, and applied to a DE-52 cellulose column equili- brated with the same buffer. AADH was eluted with 200 mm NaCl using a salt gradient (0–250 mm NaCl con-

Steady-state kinetic measurements were performed with a 1 cm light path in 10 mm BisTris propane buffer, pH 7.5,

FEBS Journal 272 (2005) 5894–5909 ª 2005 FEBS

5906

pRK415-1 pRKAADH mauFR::aauBEDAorf-2 Broad-host-range vector pRK415-1, [24] This study mauFR::aauBEDAorf-2

25 (cid:1)C. Multiple-wavelength absorption studies were carried out using a photodiode array detector and x-scan soft- ware (Applied Photophysics). Spectra were analysed and intermediates of the reaction identified by global analysis and numerical integration methods using prokin software (Applied Photophysics).

at 25 (cid:1)C in a total volume of 1 mL. AADH activity was measured using a dye-linked assay in which the reduction of PES is followed by coupling its oxidation to the reduction of DCPIP. Reduction of DCPIP was monitored at 600 nm using a Perkin-Elmer (Boston, MA) Lambda 5 UV-visible spectrophotometer. The reaction mixture typic- ally contained 50 nm AADH (for benzylamine-dependent reactions) or 3 nm (for tryptamine-dependent reactions), 0.04 mm DCPIP and 5 mm PES (unless stated otherwise). Substrates were added to the reaction mix at the appropriate concentration (see Results). Initial velocity was expressed as lmole product formed per lmole enzyme per s using a molar absorption coefficient at 600 nm of 22 000 m)1 cm)1 for DCPIP [26]. Initial velocity data collected as a function of benzylamine concentration were analysed by fitting using the standard Michaelis–Menten rate equation. Initial velo- city data collected as a function of tryptamine concentration [S] were fitted to Eqn (2). Equation 2 has previously been applied in the determination of steady-state kinetic parame- ters for trimethylamine dehydrogenase [27,28] and methanol dehydrogenase [29,30].

In studies of the temperature dependence of bond break- age, enzyme was equilibrated in the stopped-flow apparatus (or in the UV-visible spectrophotometer for steady-state the appropriate temperature prior to the reactions) at acquisition of kinetic data. Temperature control was achieved using a thermostatic circulating water bath, and the temperature was monitored directly in the stopped-flow apparatus using a semiconductor sensor (or using a ther- mometer in the UV ⁄ visible spectrophotometer). Control studies of the concentration dependence of bond cleavage at 4 and 40 (cid:1)C indicated that the Kd (or Km for steady-state reactions) was not substantially perturbed on changing tem- perature (ensuring that substrate was saturating at all the temperatures investigated). Kinetic and thermodynamic parameters were obtained by fitting the Eyring equation to the data as described previously [16,17].

Studies of the pH dependence of TTQ reduction

Stopped-flow kinetic studies of the reductive half-reaction

exchanged into the

pH studies were performed at 25 (cid:1)C in 50 mm potassium acetate (pH 5.0–5.5), 50 mm potassium phosphate (pH 6.0– 7.0) or 50 mm BisTris propane buffer (pH 7.5–10.0). AADH was required buffer by gel-exclusion chromatography and substrates were also dissolved in the required buffer (see Results). Given that previous studies have established that univalent cations sti- mulate spectral changes in AADH, particularly at higher pH values [20], experiments were conducted in the absence of added salt. pH profiles for the kinetic parameters were constructed and the data fitted to the equation describing a single (Eqn 3) or double (Eqn 4) ionization as appropriate.

ð3Þ

klim=Kd ¼ ððLim1 þ Lim2 (cid:3) tempÞ=ðtemp þ 1Þ

In Eqn (3), Lim1 is the lower limit and Lim2 is the upper limit of the curve,

klim=Kd ¼ ððLim1 þ Lim2 (cid:3) temp1Þ=ðtemp1 þ 1Þ

(cid:4) ðððLim2 (cid:4) Lim3Þ (cid:3) temp2Þ=ðtemp2 þ 1ÞÞÞ

ð4Þ

In Eqn (4), temp1 ¼ alog(pH ) pKa1), temp2 ¼ alog (pH ) pKa2), Lim1 is the lower limit of the curve, Lim 2 is the middle limit of the curve and Lim 3 is the upper limit of the curve.

Rapid kinetic experiments were performed using an Applied Photophysics (Leatherhead, UK) SX.18MV stopped-flow spectrophotometer. Studies of the reductive half-reaction were performed by rapid mixing of oxidized AADH (reac- tion cell concentration 1 lm) in 10 mm BisTris propane buf- fer, pH 7.5, with various concentrations of protiated or deuterated substrate (Results), at 25 (cid:1)C. The absorbance change, representing reduction of the TTQ cofactor, was followed at 456 nm. Data were analysed by nonlinear least squares regression analysis on an Acorn RISC PC using spectrakinetics software (Applied Photophysics). For each reaction, at least three replicate measurements were collec- ted and averaged, each containing 1000 data points. The absorbance changes accompanying enzyme reduction with benzylamine were monophasic, and observed rates were obtained by fitting using the standard single exponential expression. Under some conditions (see Results), absorb- ance changes for the recombinant enzyme were biphasic and were analysed using a double exponential expression. The fast phase of these transients (> 90% of the total amplitude change for reactions at or below 32 (cid:1)C, and > 80% for reactions above 32 (cid:1)C) exhibits a KIE and thus reflects C-H bond breakage. The slow phase (the origin of which remains uncertain) was not observed in reactions of AADH with deuterated benzylamine and therefore transients were ana- lysed using the single exponential expression.

Stopped-flow kinetic studies of the oxidative half-reaction with PES

Anaerobic rapid kinetic experiments were performed using an Applied Photophysics SX.18MV stopped-flow spectro-

For multiple wavelength studies of the reductive half- reaction, AADHox (4 lm) contained in 10 mm BisTris pro- pane buffer, pH 7.5, was mixed with 200 lm protiated or deuterated benzylamine (reaction cell concentrations) at

FEBS Journal 272 (2005) 5894–5909 ª 2005 FEBS

5907

P. Hothi et al. Cofactor biogenesis in TTQ-dependent AADH

9 Jones L, Pearson AR, Tang Y, Wilmot C & Davidson

V (2005) J Biol Chem 280, 17392–17396.

10 Wang Y, Graichen ME, Liu A, Pearson AR, Wilmot

CM & Davidson VL (2003) MauG, a novel diheme pro- tein required for tryptophan tryptophylquinone biogen- esis. Biochemistry 42, 7318–7325.

11 Pearson AR, De La Mora-Rey T, Graichen ME, Wang Y, Jones LH, Marimanikkupam S, Agger SA, Grimsrud PA, Davidson VL & Wilmot CM (2004) Further insights into quinone cofactor biogenesis: probing the role of mauG in methylamine dehydrogenase trypto- phan tryptophylquinone formation. Biochemistry 43, 5494–5502.

photometer housed in a Belle Technology anaerobic glove box (< 5 p.p.m. oxygen). Solutions used were made anaero- bic by bubbling with argon for 2 h and left to equilibrate overnight in the glove box. Studies of the oxidative half-reac- tion of the enzyme were performed by rapid mixing of 2 lm AADHred (stoichiometrically reduced with benzylamine) with various concentrations of PES (see Results) in 10 mm BisTris propane buffer, pH 7.5, at 25 (cid:1)C. The absorbance change, representing oxidation of the TTQ cofactor, was followed at 483 nm (isosbestic point of PES). Data were ana- lysed by nonlinear least squares regression analysis on an Acorn RISC PC using spectrakinetics software (Applied Photophysics). For each reaction at least three replicate measurements were collected and averaged, each containing 1000 data points. The absorbance change monitored was monophasic, and observed rates were obtained by fitting using the standard single exponential expression.

12 van der Palen CJ, Slotboom DJ, Jongejan L, Reijnders WN, Harms N, Duine JA & van Spanning RJ (1995) Mutational analysis of mau genes involved in methyla- mine metabolism in Paracoccus denitrificans. Eur J Biochem 230, 860–871.

P. Hothi et al. Cofactor biogenesis in TTQ-dependent AADH

References

13 van der Palen CJ, Reijnders WN, de Vries S, Duine JA & van Spanning RJ (1997) MauE and MauD proteins are essential in methylamine metabolism of Paracoccus denitrificans. Antonie Van Leeuwenhoek 72, 219–228.

14 van Spanning RJ, Wansell CW, Reijnders WN, Olt-

1 Govindaraj S, Eisenstein E, Jones LH, Sanders-Loehr J, Chistoserdov AY, Davidson VL & Edwards SL (1994) Aromatic amine dehydrogenase, a second tryp- tophan tryptophylquinone enzyme. J Bacteriol 176, 2922–2929.

mann LF & Stouthamer AH (1990) Mutagenesis of the gene encoding amicyanin of Paracoccus denitrificans and the resultant effect on methylamine oxidation. FEBS Lett 275, 217–220.

2 Edwards SL, Davidson VL, Hyun YL & Wingfield PT (1995) Spectroscopic evidence for a common electron transfer pathway for two tryptophan tryptophylquinone enzymes. J Biol Chem 270, 4293–4298.

3 Hyun YL & Davidson VL (1995) Electron transfer reac-

tions between aromatic amine dehydrogenase and azurin. Biochemistry 34, 12249–12254.

15 Van Spanning RJ, van der Palen CJ, Slotboom DJ, Reijnders WN, Stouthamer AH & Duine JA (1994) Expression of the mau genes involved in methylamine metabolism in Paracoccus denitrificans is under control of a LysR-type transcriptional activator. Eur J Biochem 226, 201–210.

16 Basran J, Sutcliffe MJ & Scrutton NS (1999) Enzymatic H-transfer requires vibration-driven extreme tunneling. Biochemistry 38, 3218–3222.

4 Masgrau L, Roujeinikova A, Johannissen L, Basran J, Ranaghan K, Hothi P, Mulholland A, Sutcliffe M, Scrutton N & Leys D (2005) Atomic description of an enzyme reaction dominated by hydrogen tunnelling, submitted.

5 Vellieux FM, Huitema F, Groendijk H, Kalk KH, Jzn

JF, Jongejan JA, Duine JA, Petratos K, Drenth J & Hol WG (1989) Structure of quinoprotein methylamine dehy- drogenase at 2.25 A˚ resolution. EMBO J 8, 2171–2178.

17 Basran J, Patel S, Sutcliffe MJ & Scrutton NS (2001) Importance of barrier shape in enzyme-catalyzed reac- tions – Vibrationally assisted hydrogen tunneling in tryptophan tryptophylquinone-dependent amine dehy- drogenases. J Biol Chem 276, 6234–6242.

18 Davidson VL (1990) Methylamine dehydrogenase from

methylotrophic bacteria. Methods Enzymol 188, 241–246.

6 McIntire WS, Wemmer DE, Chistoserdov A & Lid- strom ME (1991) A new cofactor in a prokaryotic enzyme: tryptophan tryptophylquinone as the redox prosthetic group in methylamine dehydrogenase. Science 252, 817–824.

19 Ellis KJ & Morrison JF (1982) Buffers of constant ionic strength for studying pH-dependent processes. Methods Enzymol 87, 405–426.

20 Zhu Z & Davidson VL (1998) Kinetic and chemical

7 Chistoserdov AY (2001) Cloning, sequencing and muta- genesis of the genes for aromatic amine dehydrogenase from Alcaligenes faecalis and evolution of amine dehy- drogenases. Microbiology 147, 2195–2202.

mechanisms for the effects of univalent cations on the spectral properties of aromatic amine dehydrogenase Biochem J 329, 175–182.

21 Hyun YL & Davidson VL (1995) Mechanistic studies of aromatic amine dehydrogenase, a tryptophan trypto- phylquinone enzyme. Biochemistry 34, 816–823.

8 Graichen ME, Jones LH, Sharma BV, van Spanning RJ, Hosler JP & Davidson VL (1999) Heterologous expression of correctly assembled methylamine dehydro- genase in Rhodobacter sphaeroides. J Bacteriol 181, 4216–4222.

FEBS Journal 272 (2005) 5894–5909 ª 2005 FEBS

5908

22 Kondo T, Kondo E, Maki H, Yasumoto K, Takagi K,

29 Harris TK & Davidson VL (1993) A new kinetic model

Kano K & Ikeda T (2004) Purification and characteriza- tion of aromatic amine dehydrogenase from Alcaligenes xylosoxidans. Biosci Biotechnol Biochem 68, 1921–1928.

for the steady-state reactions of the quinoprotein methanol dehydrogenase from Paracoccus denitrificans. Biochemistry 32, 4362–4368.

23 Iwaki M, Yagi T, Horiike K, Saeki Y, Ushijima T &

Nozaki M (1983) Crystallization and properties of aro- matic amine dehydrogenase from Pseudomonas sp. Arch Biochem Biophys 220, 253–262.

30 Hothi P, Basran J, Sutcliffe MJ & Scrutton NS (2003) Effects of multiple ligand binding on kinetic isotope effects in PQQ-dependent methanol dehydrogenase. Biochemistry 42, 3966–3978.

24 Keen NT, Tamaki S, Kobayashi D & Trollinger D

P. Hothi et al. Cofactor biogenesis in TTQ-dependent AADH

Supplementary material

(1988) Improved broad-host-range plasmids for DNA cloning in gram-negative bacteria. Gene 70, 191–197. 25 Sistrom WR (1960) A requirement for sodium in the

growth of Rhodopseudomonas spheroids. J Gen Microbiol 22, 778–785.

26 Armstrong MCD (1964) The molecular extinction

coefficient of 2,6-dichlorophenol indophenol. Biochim Biophys Acta 86, 194–197.

27 Falzon L & Davidson VL (1996) Kinetic model for the regulation by substrate of intramolecular electron transfer in trimethylamine dehydrogenase. Biochemistry 35, 2445–2452.

The following supplementary material is available for this article online: Figure S1. Corrected nucleotide sequence of aauB. Errors in the published nucleotide sequence (accession AF302652), shown in emboldened and underlined text, and are as follows: T835 is now C835, C836 is now T836 and an extra T between A1376 and T1377. Underlined amino acid residues denote the periplasmic signal peptide predicted by SignalP. Figure S2. Nucleotide and predicted polypeptide sequence of aauA. Underlined amino acid residues denote the periplasmic signal peptide predicted by SignalP.

28 Basran J, Chohan KK, Sutcliffe MJ & Scrutton NS (2000) Differential coupling through Val-344 and Tyr-442 of trimethylamine dehydrogenase in electron transfer reactions with ferricenium ions and electron transferring flavoprotein. Biochemistry 39, 9188–9200.

FEBS Journal 272 (2005) 5894–5909 ª 2005 FEBS

5909