Final steps in the catabolism of nicotine
Deamination versus demethylation of c-N-methylaminobutyrate
Calin-Bogdan Chiribau
1
, Marius Mihasan
1,2
, Petra Ganas
1
, Gabor L. Igloi
3
, Vlad Artenie
2
and Roderich Brandsch
1
1 Institute of Biochemistry and Molecular Biology, Alberts-Ludwig University of Freiburg, Germany
2 Department of Biochemistry, Alexandru Ioan-Cuza University of Iasi, Romania
3 Institute of Biology III, Alberts-Ludwig University of Freiburg, Germany
One of the major health risks continues to be the smok-
ing of tobacco. Nicotine, in itself highly toxic, when
inhaled with the tobacco smoke readily crosses the
blood–brain barrier. Its effects on the central nervous
system, mediated by cholinergic receptors, make it
highly addictive. As a result of nicotine addiction, only
a small percentage of smokers give up smoking [1]. In
addition, exposure to tobacco smoke in public places,
so-called secondary smoking, or to solid or liquid
waste during processing of tobacco products, repre-
sents a serious health threat. Therefore detoxification
of these tobacco waste products by removal of nicotine
is a major challenge. Several soil microorganisms have
evolved the enzymatic ability to mineralize nicotine,

rate may be demethylated to c-N-aminobutyrate by the recently identified
c-N-methylaminobutyrate oxidase [Chiribau et al. (2004) Eur J Biochem
271, 4677–4684]. In an alternative pathway, an amine oxidase with noncov-
alently bound FAD and of novel substrate specificity removed methylamine
from CH
3
-4-aminobutyrate with the formation of succinic semialdehyde.
Succinic semialdehyde was converted to succinate by a NADP
+
-dependent
succinic semialdehyde dehydrogenase. Succinate may enter the citric acid
cycle completing the catabolism of the pyrrolidine moiety of nicotine.
Expression of the genes of these enzymes was dependent on the presence of
nicotine in the growth medium. Thus, two enzymes of the nicotine regulon,
c-N-methylaminobutyrate oxidase and amine oxidase share the same sub-
strate. The K
m
of 2.5 mm and k
cat
of 1230 s
)1
for amine oxidase vs. K
m
of
140 lm and k
cat
of 800 s
)1
for c-N-methylaminobutyrate oxidase, deter-
mined in vitro with the purified recombinant enzymes, may suggest that

achieve such goals, an in-depth understanding of the
enzymology of nicotine catabolism is required.
Our effort is directed towards the comprehensive
characterization of the metabolic pathways of nicotine
breakdown as it is present in the Gram-positive soil
bacterium A. nicotinovorans [6]. A key step in the
breakdown of nicotine by A. nicotinovorans carrying
the catabolic plasmid pAO1 is the cleavage of 2,6-di-
hydroxypseudooxynicotine into 2,6-dihydroxypyridine
and c-N-methylaminobutyrate (CH
3
-4-aminobuty-
rate) by 2,6-dihydroxypseudooxynicotine hydrolase
(DHPONH, Fig. 1). This reaction is performed by a
C–C bond hydrolase of the a ⁄ b fold family, the first
shown to act on a heteroaromatic compound [7]. We
have recently shown that a gene cluster on pAO1 is
involved in the demethylation of CH
3
-4-aminobuty-
rate. It consists of mabO, encoding the enzyme c-N-
methylaminobutyrate oxidase (MABO, Fig. 1), which
oxidatively demethylates CH
3
-4-aminobutyrate. This
gene is flanked by a purU-like gene encoding a putative
formyltetrahydrofolate deformylase and by a folD-like
gene, encoding the putative bifunctional enzyme meth-
ylenetetrahydrofolate (CH
2

SsaDH encoded by the sad gene of pAO1 (see Fig. 2).
Succinate may enter the citric acid cycle, thus comple-
ting the catabolic pathway of CH
3
-4-aminobutyrate
generated from the pyrrolidine ring of nicotine.
Results
Expression of the pAO1 mao and sad-like genes
required the presence of nicotine in the growth
medium
The mao and sad genes addressed in this study are
located on pAO1 in a gene cluster flanked by a Tn554
element and an ORF of a truncated transposase
(Fig. 2, panel A, DTn) [6]. This gene cluster contains
the purU-mabO-folD operon, which is transcribed only
in the presence of nicotine under the control of the
transcriptional activator PmfR [9]. If the mao and sad-
like genes were functionally connected to mabO, one
Fig. 1. Formation and breakdown of c-N-methylaminobutyrate in
A. nicotinovorans pAO1.
C B. Chiribau et al. c-N-methylaminobutyrate catabolism
FEBS Journal 273 (2006) 1528–1536 ª 2006 The Authors Journal compilation ª 2006 FEBS 1529
would expect them also to be expressed in a nicotine-
dependent manner. In order to investigate this, we
analyzed the transcription of these genes in the pres-
ence and absence of nicotine in the growth medium by
RT-PCR. The results presented in Fig. 2B confirmed
the expectation that these genes are transcribed only in
the presence of nicotine, as was the case for the mabO
gene.

aminobutyrate and the reaction products were methyl-
amine (Fig. 3C) and succinic semialdehyde (see below).
Thus, the enzyme behaved as an amine oxidase rather
than as a monoamine oxidase. The pH optimum was
found to be 9.8. The K
m
and k
cat
of AO with CH
3
-4-
aminobutyrate as substrate was 2.5 ± 0.2 mm and
1230 ± 20 s
)1
, respectively (Table 1), as compared
with the previously determined K
m
of 140 lm and k
cat
of 800 s
)1
for MABO [8]. It may be observed that the
catalytic efficiency of MABO for CH
3
-4-aminobutyrate
(k
cat
⁄ K
m
of 5.71 lm

M, DNA size marker.
c-N-methylaminobutyrate catabolism C B. Chiribau et al.
1530 FEBS Journal 273 (2006) 1528–1536 ª 2006 The Authors Journal compilation ª 2006 FEBS
The pAO1 sad gene encodes a succinic
semialdehyde dehydrogenase (SsaDH)
The N-terminal extension of the recombinant SsaDH
reads MSPIHHHHHHLVPRGS
M (the start methio-
nine residue is underlined). Analyzed by PAGE on
10% SDS gels, it migrated in good accordance with its
calculated molecular mass of 51 kDa (Fig. 3A) and the
native enzyme is a homodimer (not shown). The
kinetic constants of the enzyme are listed in Table 1.
When NAD
+
replaced NADP
+
in the assay, the activ-
ity of the enzyme was about 25-fold less then that
observed with NADP
+
. The reaction at 10 mm NAD
+
still did not reach saturation level.
The enzyme was active also towards butyraldehyde
(8.5% of the activity observed with succinic semialde-
hyde) and propionaldehyde (1.6% of the activity
observed with succinic semialdehyde) as substrates.
Coupled assay with AO and SsaDH with
CH

nobutyrate, which is not a substrate for SsaDH.
When, besides AO and SsaDH, increasing amounts
of MABO were introduced in the coupled reaction
with CH
3
-4-aminobutyrate as substrate, the measured
NADPH production slowed down (Fig. 4B). This indi-
cated that the two enzymes indeed competed for the
same substrate. As MABO has an approximately
10-fold higher catalytic activity than AO, in its presence,
the predominant reaction product is 4-aminobutyrate,
and thus reduction of NADP
+
was slowed down.
Since 4-aminobutyrate is also a poor substrate for AO,
which in this case acts as a monoamine oxidase and
transforms 4-aminobutyrate into succinic semialde-
Table 1. Kinetic constants of enzymes described in this study.
Enzyme Substrate K
m
(mM) k
cat
(s
)1
) k
cat
⁄ K
m
(lM
)1

0.2
0.3
0.4
320
380
440
500
Fig. 3. Characterization of enzymes and identification by TLC of
methylamine as reaction product of AO with CH
3
-4-aminobutyrate.
(A) Analysis of purified proteins on 10% SDS gel. (B) UV-visible
spectrum of the FAD-containing AO. (C) The AO reaction and TLC
were performed as described in Experimental procedures. Four
microliters of a 10 m
M solution of propylamine (PA), methylamine
(MAs) and ethylamine (EA) was applied as standard to the TLC.
MG, CH
3
-4-aminobutyrate, which does not react with ninhydrin;
MAp, methylamine formed in 5 lL of the AO reaction with
CH
3
-4-aminobutyrate as substrate.
C B. Chiribau et al. c-N-methylaminobutyrate catabolism
FEBS Journal 273 (2006) 1528–1536 ª 2006 The Authors Journal compilation ª 2006 FEBS 1531
hyde, a certain level of SsaDH activity will be present,
even at high MABO concentrations.
[
14

Fig. 5. [
14
C]Nicotine metabolites in the medium of A. nicotinovo-
rans pAO1 and growth of A. nicotinovorans pAO1 and A. nicotino-
vorans lacking pAO1 on CH
3
-4-aminobutyrate, 4-aminobutyrate and
CH
3
NH
2
as carbon source. (A) Seven microliters of medium of a
10 mL culture grown for 1 h (lanes 2 and 6), for 2 h (lanes 3 and
7), for 3 h (lanes 4 and 8), and for 4 h (lanes 5 and 9) on minimal
medium supplemented with [
14
C]nicotine in the presence (lanes
2–5) or absence (lanes 6–9) of (NH
4
)
2
SO
4
were analyzed on a TLC
plate (see Experimental procedures). The plate was exposed for
62 h to an X-ray film. MA, position of methylamine standard stained
with the ninhydrin reaction on the same plate; N, nicotine; X,
unidentified labelled metabolite; Origin, site of application of sam-
ples. (B) Arthrobacter strains were grown on minimal medium with
the indicated carbon sources as described in Experimental

polyamine oxidases [10–12] it acts upon a secondary
amine, in this case CH
3
-4-aminobutyrate, giving rise to
methylamine and succinic semialdehyde. Its activity
was specific towards CH
3
-4-aminobutyrate and its
monoamine oxidase activity with 4-aminobutyrate as
substrate was weak. Similar to other members of the
polyamine oxidases, the FAD cofactor was noncova-
lently bound to the apoprotein and the C-terminal
fingerprint sequence SGGCY of monoamine oxidases,
with C being the cysteine residue to which the FAD
cofactor is covalently attached in these enzymes [13],
was replaced by the sequence AGGA
359
Y.
The second enzyme characterized in this study
showed high similarity to NADP
+
-dependent SsaDH
from various organisms (not shown). It contains the
amino acid consensus patterns of the aldehyde dehy-
drogenases glutamic acid active site (SwissProt Prosite
PS00687) in the form of ME
270
LGGNA, and cysteine
acive site (SwissProt Prosite PS00070) in the form of
GEAC

2
TH
4
dehydrogenase ⁄ cyclohydrolase reaction,
energy is conserved in NADPH and formaldehyde
may be assimilated by the Embden–Meyerhof fructose-
bisphosphate aldose ⁄ transaldolase variant of the ribu-
lose monophosphate cycle [14,15]. The amino group of
4-aminobutyrate, the second reaction product in this
pathway, may be transaminated to a-ketoglutarate and
the remaining succinic semialdehyde may be oxidized
to succinate by a succinic semialdehyde dehydrogenase
[16,17]. This pathway for 4-aminobutyrate catabolism
is generally found in bacteria [18–20]. It also appears
to be active in A. nicotinovorans, independent of the
presence of the megaplasmid pAO1, since both strains,
with and without pAO1, were able to grow on 4-ami-
nobutyrate as the carbon source.
The second, pAO1-encoded pathway would start with
the newly discovered reaction of CH
3
-4-aminobutyrate
deamination to succinic semialdehyde and methylamine
catalyzed by AO. In this reaction FAD is reduced to
FADH
2
. The pAO1-encoded SsaDH then produces suc-
cinate, which enters the citric acid cycle, and NADPH.
A. nicotinovorans devoid of pAO1 was not able to grow
on CH

butyrate, a compound with two additional C-units as
compared to sarcosine. AO still has a very low mono-
amine oxidase catalytic activity towards 4-aminobuty-
rate, but is specific for the oxidative deamination of
the secondary amine of CH
3
-4-aminobutyrate. Appa-
rently there was a selective pressure during the esta-
blishment of nicotine catabolism for the evolution of
new enzyme specificities starting from enzymes with
sarcosine oxidase and polyamine oxidase activities.
We must ask our selves which pathway predominates
in vivo. From the in vitro kinetic data one would predict
a preferentially channelling of CH
3
-4-aminobutyrate
to the demethylation pathway, since the k
cat
⁄ K
m
of
MABO show it to be approximately 10 times more cata-
lytically active than the deaminating AO. We do not
know at the moment how the in vivo competition of the
two enzymes for the same substrate is regulated. Addi-
tional work will be required to answer this question.
However, under the experimental conditions used,
methylamine is secreted into the growth medium, which
shows that the deamination pathway is active in vivo.
Experimental procedures

grown in the presence or absence of nicotine with the
help of the RNeasy kit (Qiagen, Hilden, Germany),
reverse-transcribed with T4 reverse transcriptase, and the
respective cDNAs were applied as templates in PCR reac-
tions as described previously [8,22] with primers listed in
Table 2.
Cloning of the monoamine oxidase (mao) and
the succinate semialdehyde dehydrogenase
(sad)-like genes
The pAO1 DNA carrying the corresponding ORFs was
amplified with the primer pair #1 and #2 for mao and #3
and #4 for sad (see Table 2), using Pfu-Ultra DNA-Poly-
merase and pAO1 as template. The PCR conditions were
95 °C for 1 min, 54 °C for 45 s, 72 °C for 2 min, repeated
30 times and followed by 72 °C for 10 min. The amplified
DNA and the vector pH 6EX3 [23] were digested with
endonucleases BamHI and XhoI, ligated with the rapid
DNA ligation kit (Roche Applied Sciences, Mannheim,
Germany) and transformed into E. coli XL1-Blue compe-
tent bacteria.
Expression and purification of the recombinant
proteins
A 100 mL preculture of E. coli XL-1Blue harbouring
pH 6EX3mao or pH 6EX3sad was diluted 1 : 10 in 1 L of
LB medium. After 2 h at 37 °C, expression of the genes
was induced for 4–5 h at 30 °C with 1 mm IPTG. Prepar-
ation of bacterial extracts and purification of the proteins
on High Performance nickel-chelating sepharose was as des-
cribed previously [8]. The recombinant proteins were stable
for several weeks at 4 °C with minor precipitation. The

was monitored by the increase in absorption at
340 nm for 5 min at room temperature.
Table 2. Oligonucleotides used in this study.
No Sequence Use
15¢-GAG GTG GAT CCG TGG GCC GCA-3¢ Forward mao, cloning
25¢-GAA TGA CTC GAG CCG AAG TAA TC-3¢ Reverse mao, cloning
35¢-CTT CTG AGG ATC CCA AAT GAC AGT-3¢ Forward sad, cloning
45¢-CAT GTA AGC CCC CTC GAG TCG TTC AG-3¢ Reverse sad, cloning
55¢-CGT CAC GGT ATT CGA AGC C-3¢ Forward mao, RT-PCR
65¢-CAC TGG CTA ATT CCA GTG C-3¢ Reverse mao, RT-PCR
75¢-CAC TAG CGA AGA TGC CGT C-3¢ Forward sad, RT-PCR
85¢-CCA ACG CAG AAA CTC GGC-3¢ Reverse sad, RT-PCR
95¢-CGG CAT TAT CGG TGA CAG C-3¢ Forward mabO, RT-PCR
10 5¢-CGC GCA ACA CTG AGG GAC-3¢ Reverse mabO, RT-PCR
c-N-methylaminobutyrate catabolism C B. Chiribau et al.
1534 FEBS Journal 273 (2006) 1528–1536 ª 2006 The Authors Journal compilation ª 2006 FEBS
A coupled AO-SsaDH assay was performed in 1 mL con-
sisting of: 100 mm sodium pyrophosphate buffer, pH 9,
5mm EDTA, 500 lm NADP
+
,10lg AO (which retains
100% activity under these reaction conditions) and 1.5 lg
SsaDH. The reaction was started by the addition of 10 mm
CH
3
-4-aminobutyrate and the reduction of NADP
+
was
monitored at 340 nm in an Ultrospec 3100 Spectrophoto-
meter (Amersham Biosciences).

the growth medium were removed at different time points
and analyzed by TLC for the presence of [
14
C]methylamine
as described above. The TLC plates were exposed to
Kodak X-Omat AR X-ray films (Sigma, Taufkirchen,
Germany) for various times.
Growth of A. nicotinovorans carrying or lacking
pAO1 on CH
3
-4-aminobutyrate, 4-aminobutyrate
or methylamine
CH
3
-4-aminobutyrate, 4-aminobutyrate or methylamine (2
gÆL
)1
) replaced citrate as carbon source in the minimal
medium [21] in these experiments. Biotin at 41 nm final con-
centration was added as vitamin supplement to the bacterial
cultures. An A. nicotinovorans overnight culture (150 lL)
was diluted 100 times in sterile 50-mL Falcon tubes and
growth was monitored by the increase in turbidity at 600 nm.
Acknowledgements
We thank I. Deuchler for excellent technical assistance,
C. Brizio (University of Bari, Italy) for help with the
kinetic data and C. Sandu (The Rockefeller University,
New York, NY, USA) for critically reading the manu-
script. This work was supported by a grant of the
Deutsche Forschungsgemeinschaft to RB.

tobacco alkaloid nicotine by Arthrobacter nicotinovorans
pAO1. Eur J Biochem 271, 4677–4684.
9 Chiribau C-B, Sandu C, Igloi GL & Brandsch R
(2005) Characterization of PmfR, the transcriptional
activator of the pAO1-borne purU-mabO-folD operon
of Arthrobacter nicotinovorans. J Bacteriol 187, 3062–
3070.
10 Landry J & Sternglanz R (2003) Yeast Fms1 is a
FAD-utilizing polyamine oxidase. Biochem Biophys Res
Commun 303, 771–776.
11 Busch K, Piehler J & Fraomm H (2000) Plant succinic
semialdehyde dehydrogenase: dissection of nucleotide
binding by surface plasmon resonance and fluorescence
spectroscopy. Biochemistry 39, 10110–10117.
12 Tavladorakiu P, Schinina ME, Cecconi F, Di Agostino
S, Manera F, Rea G, Mariottini P, Federico R &
Angelici R (1998) Maize polyamine oxidase: primary
structure from protein and cDNA sequencing. FEBS
Lett 426, 62–66.
13 De Colibus L, Li M, Binda C, Lustig A, Edmondson
DE & Mattevi A (2005) Three-dimensional structure of
human monoamine oxidase A (MAO A): relation to the
structures of rat MAOA and human MAOB. Proc Natl
Acad Sci USA 102, 12684–12689.
C B. Chiribau et al. c-N-methylaminobutyrate catabolism
FEBS Journal 273 (2006) 1528–1536 ª 2006 The Authors Journal compilation ª 2006 FEBS 1535
14 Levering PR, Binnema DJ, van Dijken JP & Harder W
(1981) Enzymatic evidence for a simultaneous operation
of two one-carbon assimilation pathways during growth
of Arthrobacter P1 on choline. FEMS Lett 12, 19–25.

hmu
¨
ller M, Schimz A, Messmer L & Decker K
(1975) Covalently bound FAD in d-6-hydroxynicotine
oxidase. J Biol Chem 250, 7747–7751.
22 Sandu C, Chiribau C-B & Brandsch R (2003) Charac-
terization of HdnoR, the transcriptional repressor of the
6-hydroxy- d -nicotine oxidase gene of Arthrobacter
nicotinovorans pAO1, and its DNA-binding activity in
response to l- and d-nicotine derivatives. J Biol Chem
278, 51307–51315.
23 Berthold H, Scanarini M, Abney CC, Frorath B &
Northemann W (1992) Purification of recombinant anti-
genic epitopes of the human 68-kDa (U1) ribonucleo-
protein antigen using the expression system pH6EX3
followed by metal chelating affinity chromatography.
Protein Expr Purif 3, 50–56.
c-N-methylaminobutyrate catabolism C B. Chiribau et al.
1536 FEBS Journal 273 (2006) 1528–1536 ª 2006 The Authors Journal compilation ª 2006 FEBS