Tài liệu Báo cáo khoa học: Tissue expression and biochemical characterization of human 2-amino 3-carboxymuconate 6-semialdehyde decarboxylase, a key enzyme in tryptophan catabolism pptx

Tissue expression and biochemical characterization
of human 2-amino 3-carboxymuconate 6-semialdehyde
decarboxylase, a key enzyme in tryptophan catabolism
Lisa Pucci*, Silvia Perozzi*, Flavio Cimadamore, Giuseppe Orsomando and Nadia Raffaelli
Istituto di Biotecnologie Biochimiche, Universita
`
Politecnica delle Marche, Ancona, Italy
In mammals, tryptophan exceeding basal requirement
for protein and serotonin synthesis, is oxidized via
indole-ring cleavage through the kynurenine pathway,
consisting of several enzymatic reactions leading to
2-amino 3-carboxymuconate 6-semialdehyde (ACMS)
(Fig. 1) [1,2]. ACMS can be decarboxylated to
2-aminomuconate 6-semialdehyde (AMS) by the
enzyme ACMS decarboxylase (ACMSD, EC 4.1.1.45),
or it can undergo spontaneous pyridine ring closure to
form quinolinate, an essential precursor for de novo
NAD synthesis. AMS can be routed to the citric
acid cycle via the glutarate pathway, or converted
nonenzymatically to picolinate. By catalyzing ACMS
decarboxylation, ACMSD thus diverts ACMS from
NAD synthesis, channeling tryptophan towards
complete oxidation or conversion to picolinate.
By determining picolinate and quinolinate forma-
tion, ACMSD directly participates in the cellular pro-
cesses regulated by these molecules. Quinolinate is a
neurotoxic tryptophan metabolite, whose action has
been ascribed to N-methyl-D-aspartate receptors
activation and to its ability to generate free radicals
Keywords
ACMSD; NAD biosynthesis; picolinate;

quinolinate; tryptophan catabolism
Correspondence
N. Raffaelli, Istituto di Biotecnologie
Biochimiche, Universita
`
Politecnica delle
Marche, Via Ranieri, 60131 Ancona, Italy
Fax: +39 712204677
Tel: +39 712204682
E-mail:
*These authors contributed equally to this
paper
(Received 16 November 2006, accepted
6 December 2006)
doi:10.1111/j.1742-4658.2007.05635.x
2-Amino 3-carboxymuconate 6-semialdehyde decarboxylase (ACMSD, EC
4.1.1.45) plays a key role in tryptophan catabolism. By diverting 2-amino
3-carboxymuconate semialdehyde from quinolinate production, the enzyme
regulates NAD biosynthesis from the amino acid, directly affecting quinoli-
nate and picolinate formation. ACMSD is therefore an attractive therapeu-
tic target for treating disorders associated with increased levels of
tryptophan metabolites. Through an isoform-specific real-time PCR assay,
the constitutive expression of two alternatively spliced ACMSD transcripts
(ACMSD I and II) has been examined in human brain, liver and kidney.
Both transcripts are present in kidney and liver, with highest expression
occurring in kidney. In brain, no ACMSD II expression is detected, and
ACMSD I is present at very low levels. Cloning of the two cDNAs in yeast
expression vectors and production of the recombinant proteins, revealed
that only ACMSD I is endowed with enzymatic activity. After purification
to homogeneity, this enzyme was found to be a monomer, with a broad

pH optimum ranging from 6.5 to 8.0, a K
m
of 6.5 lm, and a k
cat
of 1.0 s
)1
.
ACMSD I is inhibited by quinolinic acid, picolinic acid and kynurenic
acid, and it is activated slightly by Fe
2+
and Co
2+
. Site-directed mutagen-
esis experiments confirmed the catalytic role of residues, conserved in all
ACMSDs so far characterized, which in the bacterial enzyme participate
directly in the metallocofactor binding. Even so, the properties of the
human enzyme differ significantly from those reported for the bacterial
counterpart, suggesting that the metallocofactor is buried deep within the
protein and not as accessible as it is in bacterial ACMSD.
Abbreviations
ACMS, 2-amino 3-carboxymuconate 6-semialdehyde; ACMSD, ACMS decarboxylase; AMS, 2-aminomuconate 6-semialdehyde.
FEBS Journal 274 (2007) 827–840 ª 2007 The Authors Journal compilation ª 2007 FEBS 827
[3,4]. This neurotoxicity might play an important role
in the pathogenesis of major neurodegenerative and
convulsive disorders. In particular, many of the
distinct neuropathological features of Huntington’s
disease are duplicated in experimental animals by
intrastriatal quinolinate injection [5,6]. In addition, a
significant elevation of quinolinate levels has been
observed in low-grade Huntington’s disease brains,

suggesting that the molecule might participate in the
initial phases of the neurodegenerative process [7].
Moreover, a role of quinolinate in the pathogenesis
of AIDS–dementia complex, as well as Alzheimer’s
disease, has been very recently proposed [8,9]. In turn,
picolinate exhibits important immunomodulatory
properties, involving activation of macrophage tumori-
cidal, microbicidal and proinflammatory functions [10–
12]. This metabolite also stimulates apoptosis in var-
ious transformed cell lines, and efficiently interrupts
the progress of human HIV-1 in vitro [13,14]. Although
the physiological relevance of picolinate formation
in vivo is not known, it has been detected in human
milk, pancreatic juice, intestine and in the serum of
patients with degenerative liver diseases [15–17]. In
addition, high levels have been measured in the
cerebrospinal fluid of children with cerebral malaria
and in the brain of a murine model of the syndrome
[18,19]. Interestingly, picolinate is reportedly able to
prevent the neurotoxic effects of quinolinate in the rat
central nervous system, suggesting that a highly regula-
ted production of these metabolites is required for
Fig. 1. Schematic overview of tryptophan
catabolism through the kynurenine pathway.
Human ACMSD L. Pucci et al.
828 FEBS Journal 274 (2007) 827–840 ª 2007 The Authors Journal compilation ª 2007 FEBS
normal neuronal function [20]. Such considerations
indicate that ACMS decarboxylation is a key meta-
bolic control step and that the enzyme catalyzing this
reaction is likely to be a drug target [21].

Mammalian ACMSD has been purified and charac-
terized from cat, hog and rat [22–24], where it is
present only in kidney, brain and liver. Several studies
have demonstrated that in rats, nutritional factors and
hormones affect both gene expression and enzymatic
activity. In particular, the enzyme is down-regulated
by dietary polyunsaturated fatty acids, phthalate esters
and peroxisome proliferators, like clofibrate, whereas it
is up-regulated in rats fed a high protein diet [25–28].
mRNA expression and enzymatic activity are elevated
in the liver of streptozotocin-induced diabetic rats, and
insulin injection suppresses such elevation [29]. These
studies clearly demonstrated that changes in ACMSD
activity are readily reflected by serum and tissue quino-
linate levels and in the rate of tryptophan-to-NAD
conversion. Rat liver ACMSD gene expression is regu-
lated by the two transcriptional factors: hepatocyte
nuclear factor 4a (HNF4a) and peroxisome prolifera-
tor-activated receptor a (PPRa); the former activates
ACMSD expression directly by site-specific binding
to the promoter, and the latter represses ACMSD
expression indirectly through suppression of HNF4a
expression [30].
The presence of ACMSD has been recently demon-
strated in bacteria species that fully catabolize trypto-
phan or 2-nitrobenzoic acid [31,32]. The biochemical
and structural characterization of Pseudomonas fluores-
cens ACMSD revealed that the enzyme is metal-
dependent and catalyzes a novel type of nonoxidative
decarboxylation [21,33–35].

The human gene has been identified upon expression
in COS7 cells of a cDNA from a brain library, enco-
ding the human homologue of the rat protein [36].
Recently, a cDNA sequence from a liver library, deri-
ving from alternative splicing of the gene, has been
deposited in GenBank. In this study, we have demon-
strated the constitutive expression in human organs of
the two alternatively spliced ACMSD transcripts. The
corresponding cDNAs have been cloned in yeast
expression vectors, allowing for the first time the puri-
fication and biochemical characterization of human
ACMSD.
Results
Cloning of ACMSD transcripts
The complete coding sequence of human ACMSD
was obtained from reverse-transcribed human kidney
RNA, by using the primers pair 1fw ⁄ 11rev, encompas-
sing the ACMSD open reading frame (GenBank acces-
sion number Q8TDX5-1) (Fig. 2A; Table 1). Agarose
gel electrophoresis of the PCR product indicated the
presence of two distinct bands between 1000 and
1100 bp of roughly the same extent (not shown). Clo-
ning and nucleotide sequencing of these two products
indicated that the lower band (1011 bp) corresponded
to the expected cDNA (here named ACMSD I), while
the higher band (1068 bp) corresponded to the cDNA
currently in GenBank under accession number
AAH16018 (here named ACMSD II). ACMSD II cod-
ing sequence was obtained by using the primers pair
4fw ⁄ 11 rev (Fig. 2A). Again, two distinct PCR prod-

ucts were obtained, the longer (887 bp) corresponding
to part of the ACMSD I cDNA, the shorter (837 bp)
representing ACMSD II open reading frame (not
shown). The two transcripts are produced by alternat-
ive splicing of exons 2 and 5 of the ACMSD gene
(Fig. 2A); the presence of exon 2 in ACMSD II causes
a shift in the open reading frame, resulting in the
occurrence of the first available start codon in exon 4.
This, together with the absence of exon 5, gives rise to
a ACMSD II protein, which differs from ACMSD I at
the N-terminus (Fig. 2B).
Expression of ACMSD variants in Escherichia coli
and Pichia pastoris
ACMSD I and II cDNAs were subcloned into pET15b
and pET32b E. coli expression vectors, providing the
recombinant proteins with a N-terminal His6-tag, and
a thioredoxin-tag, respectively. In both cases, proteins
were expressed as inclusion bodies (not shown). Any
effort to obtain soluble recombinant proteins, inclu-
ding lower growth temperatures (18 °C) and isopropyl
b-D-1-thiogalactopyranoside concentrations (down to
0.1 mm), as well as inclusion of various metal ions in
the growth medium during induction, were unsuccess-
ful. Expression in the methylotrophic yeast P. pastoris
was performed both via secretion and intracellularly,
as described in Experimental procedures. Both
isoforms were secreted by yeast cells transformed with
the pPIC9 recombinant plasmids (Fig. 3); however,
when ACMSD activity was assayed in the culture
media, the enzymatic activity was only detected in

media containing isoform I. Both isoforms were
also expressed intracellularly, as evidenced by the
appearance of protein bands of the expected size
upon SDS ⁄ PAGE analysis of the crude extracts
prepared after 3 days methanol induction (not shown).
Again, only isoform I was endowed with enzymatic
activity.
L. Pucci et al. Human ACMSD
FEBS Journal 274 (2007) 827–840 ª 2007 The Authors Journal compilation ª 2007 FEBS 829
In light of the recent findings that P. fluorescens
ACMSD requires a metal-ion cofactor and is able to
take up the metal from the expressing host during
protein synthesis [21,33], we investigated whether the
production of the human recombinant isoforms might
be enhanced by the presence of metal ions in the
growth medium. We found that addition of 200 lm to
1mm CoCl
2
during methanol induction did not result
in a significant increase of ACMSD I specific activity,
and no activity was detected in the extracts prepared
from cells expressing the other isoform.
A
B
Fig. 2. Alternative splicing of human ACMSD gene. (A, upper panel) Genomic structure of the ACMSD gene with the alternative splicing pat-
terns. Exons are indicated by boxes and introns by connecting lines. The approximate locations of the oligonucleotide primers used in the
preparation of the cDNAs are shown by bold arrows (see Table 1 for sequences). (A, lower panel) Representation of full length ACMSD I
and ACMSD II mRNAs. Open reading frames are underlined and the localization of the primers pairs and probes used for real-time PCR
assays is also showed (see Table 1 for sequences). (B) Alignment of the not conserved N-terminal regions of the two isoforms. Alignment
was carried out using the

CLUSTALW program. Fully conserved and similar residues are indicated by asterisks and colon, respectively.
Table 1. Oligonucleotide primer sequences. fw, forward; rev,
reverse.
Sequence
ACMSD cloning: primer
1fw CGCTCGAGATGAAAATTGACATCCATA
GTCAT
11rev AAAGCTGAGCTCCATTCAAATTGTTTT
CTCTCAAG
4fw TTCTCGAGATGGGAAAGTCTTCAGAGT
GGT
ACMSD real-time PCR: primer and probe
1 ⁄ 3fw TGGCCAGATCTAAAAAAGAGGT
2fw ATCCCAGGAAACACCAGTAGA
10rev ATTGTTTTCTCTCAAGACCCAA
TaqMan probe T1 ACACCACAGCAAGGGAGAAGCAAAG
18Sfw CGCCGCTAGAGGTGAAATTC
18Srev TCTTGGCAAATGCTTTCGCT
TaqMan probe 18S TGGACCGGCGCAAGACGGAC
AB
Fig. 3. ACMSD I and ACMSD II extracellular expression. Tricine
SDS ⁄ PAGE of the culture filtrates of the best expressing clones
obtained by transformation of P. pastoris GS115 cells with pPIC9-
ACMSD I (A) and pPIC9-ACMSD II (B). Culture filtrates were ana-
lyzed after 104 h (lane a) and 128 h (lane b) methanol induction.
Arrows indicate the recombinant isoforms. Lane M, molecular
mass standards.
Human ACMSD L. Pucci et al.
830 FEBS Journal 274 (2007) 827–840 ª 2007 The Authors Journal compilation ª 2007 FEBS
Quantitation of ACMSD variants in liver, kidney

and brain
An isoform-specific real-time PCR assay was performed
to analyze the expression pattern of the two transcripts
in brain, kidney and liver. A hybridization probe over-
lapping exon 3 and exon 4 was designed (T1), which is
able to anneal to both variants (Fig. 2A). Specific ampli-
fication was achieved by using the same reverse primer
(10rev), located in exon 10, with 1 ⁄ 3fw primer, spanning
the boundaries of exons 1 and 3, for ACMSD I, and
2fw, located in exon 2, for ACMSD II. To assess the
specificity of the assay, linearized pGEM plasmids har-
boring the two cDNAs were used as templates in control
real-time PCR experiments. Each isoform was subjected
to amplification with the specific primer ⁄ probe set in the
presence of increasing molar amounts of the other iso-
form (1 : 1, 1 : 1000 and 1 : 10 000). Detection of
ACMSD I was not influenced by the presence of up to
10 000-fold molar excess of ACMSD II and the same
result was obtained when detection of ACMSD II was
performed in the presence of ACMSD I. The expression
levels of ACMSD alternative transcripts in the organs
we have examined are reported in Fig. 4. ACMSD I
transcript was present in all tested organs, with an
expression ratio of 1300 : 30 : 1 in kidney, liver and
brain, respectively. Interestingly, ACMSD II was not
detected in brain, while no significant difference in the
relative expression of the two isoforms within kidney
and liver was observed.
ACMSD I purification and characterization
Recombinant human ACMSD I was purified to homo-

geneity from P. pastoris cells transformed with the
pHIL-D2-ACMSD I plasmid, by the purification proce-
dure described in Experimental procedures and sum-
marized in Table 2. The final preparation was stable for
several weeks when stored at 4 °C, whereas the purified
protein was sensitive to freezing at both )20 °C and
)80 °C. SDS ⁄ PAGE of the pure protein revealed a
molecular mass of about 40 kDa, as expected for the
recombinant enzyme (Fig. 5). Gel filtration experiments
showed a native molecular mass of about 50 kDa, indi-
cating that the enzyme might exist as a monomer in
solution (not shown). ACMSD I had an optimum pH
ranging from 6.5 to 8.0 (Fig. 6). The activity was signifi-
cantly affected by the concentrations of the buffers used
at pH values below 6.5, being markedly lower in the
presence of 50 mm buffers, rather than 5 mm at the
same pH values (Fig. 6). As shown in Table 3, the pure
enzyme was fully active in the absence of metal ions,
being slightly activated by Co
2+
and Fe
2+
, whereas
Zn
2+
,Cd
2+
Cr
3+
and Fe

3+
strongly reduced the enzy-
matic activity. Human ACMSD I exhibited a linear kin-
etic behavior, with K
m
for ACMS of 6.5 lm, V
max
of
0.105 mm s
)1
and k
cat
of 1.0 s
)1
(Fig. 7).
To identify possible enzyme modulators, we exam-
ined the effect of several NAD biosynthetic pathway
intermediates on the human enzyme activity. We found
that NAD, nicotinate adenine dinucleotide, nicotina-
Fig. 4. Quantitation of ACMSD variants by real-time PCR. Tissues
RNA was reverse-transcribed and ACMSD variants were quantita-
ted as described in Experimental procedures, using the prim-
ers ⁄ probe sets reported in Table 1. Data are the mean of three
independent experiments and are presented as copies of the target
variant per 10
9
copies of 18S RNA.
Table 2. Purification of recombinant human ACMSD I.
Step
Total protein

a
(mg)
Total activity
b
(units)
Specific activity
(unitsÆmg
)1
)
Yield
(%)
Purification
(-fold)
Crude extract 305 1.9 0.006 100 –
Streptomycin sulfate 102 2.0 0.019 100 3.26
Hydroxyhapatite 23.4 1.2 0.051 63 8.5
MonoQ 0.36 0.5 1.39 31 231
a
Starting from 400 mL yeast culture.
b
The enzymatic activity was assayed spectrophotometrically, as reported in the Experimental proce-
dures.
L. Pucci et al. Human ACMSD
FEBS Journal 274 (2007) 827–840 ª 2007 The Authors Journal compilation ª 2007 FEBS 831
mide, nicotinic acid, nicotinamide mononucleotide,
nicotinate mononucleotide, tryptophan (present at
1mm concentration in the reaction mixture), and
l-kynurenine and 3-hydroxykynurenine (present at
0.1 mm concentration) were without effect. On the
other hand, a significant inhibition was exerted by

quinolinic acid, picolinic acid and kynurenic acid, as
shown in Table 4.
Bacterial and human ACMSD share a high sequence
identity, particularly with respect to the residues (i.e.
His9, His11, His177 and Asp294) known to be
involved in metal binding in P. fluorescens ACMSD
(Fig. 8). To confirm that these residues are essential
for the human enzyme activity, we performed site-
directed mutagenesis experiments on human ACMSD I
to replace His6 and His8 individually with alanine.
Mutated proteins were expressed in P. pastoris cells at
a level comparable with that of the wild-type enzyme
Fig. 5. Purification of recombinant human ACMSD I. Tricine
SDS ⁄ PAGE of fractions throughout the purification procedure:
crude extract (lane a), hydroxyhapatite (lane b), MonoQ (lane c).
Lane M, molecular mass standards.
Fig. 6. Effect of pH on ACMSD I activity. The enzymatic activity
was measured at different pH values in 50 m
M (h) and 5 mM (j)
sodium succinate, 50 m
M (s) and 5 mM (d) Mes, 50 mM 4-morph-
olinepropanesulfonic acid (m ), 50 m
M Hepes (r) and 50 mM
Tris ⁄ HCl (e) buffers. Activity was determined as described in
Experimental procedures.
Table 3. Influence of metal ions on human ACMSD I activity.
Metal ion
Concentration
(m
M)

Relative activity
(%)
None – 100
Mg
2+
1.0 100
Mn
2+
1.0 100
Zn
2+
0.1 6
Ni
2+
1.0 100
Ca
2+
1.0 100
Fe
3+
0.1 15
Fe
2+
0.1 124
1.0 90
Co
2+
0.1 120
1.0 130
Cr

3+
0.5 35
Cd
2+
0.5 25
Fig. 7. Lineweaver)Burk analysis of ACMSD I. The reciprocal of
the initial velocities was plotted against the reciprocal of ACMS
concentrations. Data are the means of three independent determi-
nations.
Table 4. Influence of tryptophan catabolites on human ACMSD I
activity.
Compound
Concentration
(m
M)
Relative activity
(%)
Kynurenic acid 0.5 67
1.0 59
Quinolinic acid 0.5 72
1.0 61
Picolinic acid 0.25 80
0.5 58
1.0 47
Human ACMSD L. Pucci et al.
832 FEBS Journal 274 (2007) 827–840 ª 2007 The Authors Journal compilation ª 2007 FEBS
(not shown) and, as expected, exhibited a significantly
lower catalytic activity. In particular, the activity
decreased by about 82% and 50% for His6Ala and
His8Ala, respectively, indicating that the mutated resi-

dues are critical for catalysis. The human enzyme is
active in the absence of added metal-ions in the reac-
tion mixture, suggesting that it might be able to take
up the metal from the expressing host during protein
synthesis. This property was previously demonstrated
for bacterial ACMSD, however, while the bacterial
enzyme can be obtained in the inactive, metal-free
form upon incubation with 5 mm EDTA for 12 h at
4 °C [21,33], human ACMSD retained full activity in
the same conditions. Incubations at EDTA concentra-
tions higher than 10 mm (i.e. 20 mm and 50 mm) only
resulted in 50% loss of the decarboxylase activity.
Moreover, unlike the P. fluorescens ACMSD apopro-
tein, which can be successfully reconstituted [21,33],
catalytic activity of the human enzyme was not
regained upon removal of EDTA and addition of
metal-ions.
Figure 9 shows the hydrophobicity profiles of
human and P. fluorescens ACMSD, performed accord-
ing to Kyte and Doolittle [37]. By comparing the mean
Fig. 8. Alignment between human (Hs) and
P. fluorescens (Psf) ACMSD. Residues
involved in metal binding in the bacterial
enzyme are highlighted in shaded boxes.
Residues subjected to mutational analysis in
the human enzyme are indicated by #.
A
B
Fig. 9. Hydrophobicity profiles of human (A)
and P. fluorescens (B) ACMSD. The profiles

were computed according to Kyte and Doo-
little [37], using a window size of 7, suitable
for discriminating between buried and sur-
face exposed regions. Residues involved in
catalysis in the bacterial enzyme and the
corresponding amino acids in the human
protein are pointed with arrows. The mean
hydrophobicity values of residues within the
window centered on the pointed amino
acids are reported in brackets.
L. Pucci et al. Human ACMSD
FEBS Journal 274 (2007) 827–840 ª 2007 The Authors Journal compilation ª 2007 FEBS 833
hydrophobic index of the residues within the window
centered on amino acids involved in metal binding, it
can be noticed that regions around His6 and His8 in
the human protein are predicted to be more hydropho-
bic than the corresponding regions in the bacterial
enzyme. Likewise Arg47, whose corresponding residue
in P. fluorescens ACMSD (Arg51) is essential for cata-
lysis and it has been hypothesized to directly partici-
pate in substrate binding [34], is located in a more
hydrophobic region in human ACMSD.
Pseudomonas fluorescens ACMSD crystallographic
structures have been used as the templates for the pre-
diction of the human enzyme structure based on
homology modeling. As shown in Fig. 10, the overall
architecture of the human enzyme model is very sim-
ilar to that of its bacterial counterpart. In particular,
the root-mean-squared deviations values calculated
between the superimposed backbones of the human

enzyme model and the crystallographic templates are
all below 1 A
˚
. The residues that directly coordinate to
A
B
C
Fig. 10. Homology modeling of human ACMSD I. Prediction of the human enzyme structure (right) was made on the base of the crystal
structure analysis of P. fluorescens ACMSD (left), as described in Experimental procedures. (A) Ribbon diagram of the two structures. The
arginine residue probably involved in substrate binding is marked by an asterisk, and the residues ligated to the metal ion cofactor (orange
sphere) are shown in ball-and-stick representation (B) metal coordination centers. (C) Molecular surface models. Hydrophobic and polar resi-
dues are colored in blue and gray, respectively.
Human ACMSD L. Pucci et al.
834 FEBS Journal 274 (2007) 827–840 ª 2007 The Authors Journal compilation ª 2007 FEBS
the metal cofactor in the bacterial protein are con-
served in the human model (Fig. 10B). The molecular
surface model of both proteins, shown in Fig. 10C,
confirm the higher hydrophobicity in human ACMSD
of the region surrounding the metal active site and the
Arg47 likely involved in substrate binding.
Discussion
In this study, the expression of the human ACMSD
transcript coding for the active enzyme has been exam-
ined in kidney, liver and brain. The tissue distribution
of the transcript closely resembles that of mouse
ACMSD mRNA [36]. The highest expression has been
observed in kidney, suggesting that the tryptophan-
NAD pathway might not be the preferred route of
tryptophan utilization in this organ. Indeed, in kidney
the kynurenine pathway is mostly used to convert

tryptophan catabolites, including kynurenine and
hydroxykinurenine taken up from the blood, to a
series of metabolites which are then excreted [38]. The
presence of ACMSD suggests that tryptophan catabo-
lites might also be channeled into the glutarate path-
way towards complete oxidation. High levels of the
enzyme might prevent excessive formation and accu-
mulation of quinolinate. Indeed experiments performed
on patients with renal insufficiency, as well as on rats
with induced renal failure, showed a decrease of
kidney ACMSD activity and significant elevation of
quinolinate in serum, urine, cerebrospinal fluid and
peripheral tissues [38–41]. On the other hand, we
found significantly lower levels of human ACMSD
expression in liver and in brain than that observed in
kidney, suggesting that in the former organs the tryp-
tophan–NAD conversion might represent a relevant
pathway. Indeed, most of the intracellular NAD in
liver is synthesized from the amino acid rather than
from dietary niacin and the formed dinucleotide is the
main vitamin source for extrahepatic tissues [42]. Like-
wise, maintenance of brain NAD levels is of extreme
importance, because NAD depletion is linked to neur-
onal damage. Recent reports have demonstrated that
reversal of NAD depletion can profoundly decrease is-
chemic brain damage and stimulation of NAD biosyn-
thetic pathways prevents or delays axonal degeneration
[43,44]. The observed low levels of liver and brain
ACMSD would allow tryptophan to be channeled
towards NAD formation, thus guaranteeing adequate

NAD supply.
Our work defines a reliable procedure for the expres-
sion and purification of the active recombinant
ACMSD in good yield. The purification method
differs from that reported for other mammalian
ACMSDs, due to the rapid inactivation of the human
enzyme both in the absence of salts and in the presence
of high ionic strength (i.e. NaCl or sulfate ammonium
concentrations higher than 0.3 and 0.4 m, respectively).
The subunit and native molecular weight values are
comparable with those reported for all ACMSDs so
far characterized, including the bacterial protein
[23,24,34], and are consistent with the lack of a quater-
nary structure. In contrast with the rat enzyme, which
is mostly active at pH 6.0 [24], human ACMSD
displays a broad pH optimum, ranging from
pH 6.5–8.0. Like all ACMSDs [22–24,34], the human
enzyme exhibits a very low micromolar-range K
m
for
ACMS. However, it shows a specific activity six times
lower than that reported for the rat enzyme and a k
cat
value that is about 6.5 times lower than that calculated
for the bacterial counterpart [24,34]. The enzymatic
activity is not significantly affected by the intermedi-
ates of the NAD biosynthetic pathways. Given that
the high concentrations required for inhibition by
quinolinic acid, picolinic acid and kynurenic acid are
far from the physiopathological levels of the trypto-

phan metabolites [8], these effects are unlikely to be of
physiological significance. However, it may be worth
noting that the three inhibitory compounds share a
COOH group on the C-2 of the pyridine ring. This
feature seems to be essential for ACMSD inhibition,
since nicotinic acid, which differs from picolinic
acid only in the position of the COOH group, was
ineffective.
Very recently, the structural characterization of
P. fluorescens ACMSD has demonstrated that the
bacterial enzyme requires a transition metal ion as co-
factor (i.e. Zn
2+
), thus representing a novel member
of the metal-dependent amidohydrolase superfamily
[21,34]. Members of this superfamily employ a great
variety of divalent metal ligands for catalysis and
share a conserved metal binding site, suggesting com-
mon aspects in their catalytic mechanism [21]. From
the sequence alignment of human and bacterial AC-
MSD and the comparison of their three-dimensional
structures, the presence of the metallocofactor in the
human enzyme can be reliably predicted. Indeed, in
the present work, the essentiality in the human
enzyme-catalyzed reaction of residues involved in
metal binding in P. fluorescens ACMSD has been
confirmed by site-directed mutagenesis. As expected,
replacement of His6 and His8 with alanine in human
ACMSD I resulted in a significant reduction of the
decarboxylase activity. Accordingly, ACMSD II iso-

form, which differs from ACMSD I in the first 24
residues at the N-terminus and lacks the two mutated
histidines, is enzymatically inactive. In contrast to
L. Pucci et al. Human ACMSD
FEBS Journal 274 (2007) 827–840 ª 2007 The Authors Journal compilation ª 2007 FEBS 835
what observed for the bacterial enzyme, however,
human recombinant ACMSD specific activity was not
increased by the presence of divalent metal ions in
the growth medium, during protein expression. In
addition, EDTA concentrations higher than those
used for the preparation of the bacterial apoprotein
were required to obtain a significant inactivation of
the human enzyme, and we were not able to recon-
stitute the activity by using the treatment that suc-
cessfully restored the bacterial protein [21,33].
Interestingly, the same difficulty in stripping off the
metallocofactor from the protein has been described
for murine adenosine deaminase, a member of the
amidohydrolase superfamily, which uses Zn
2+
as co-
factor and shares with ACMSD the same metal cen-
ter configuration [34,45]. The metal was inaccessible
to chelators, and was not exchanged with solvent to
any appreciable extent at neutral pH [45]. Authors
also reported a competitive inhibition by Zn
2+
with
respect to the substrate (K
i

7 lm), suggesting that the
cation might bind elsewhere within the enzyme active
site and block adenosine binding [45]. Accordingly,
we found that 0.1 mm Zn
2+
exerts a strong inhibitory
effect on human ACMSD. The presence of the metal-
locofactor in human ACMSD might also be inferred
by our study on the enzyme pH dependence, showing
that the increase of the concentrations of succinate
and 4-morpholineethanesulfonic acid buffers at pH
values below 6.5 significantly reduced the activity. It
can be hypothesized that these salts might chelate the
metal and efficiently remove it from the protein at
acidic pH.
Both the different sensitivity to EDTA, and the
lower k
cat
value exhibited by the human protein with
respect to the bacterial counterpart suggest that the
active site is very well buried within the human pro-
tein, not so easily accessible as it is in the bacterial
enzyme. This conclusion is validated by the compar-
ison of both the hydrophobicity profiles and the three-
dimensional structures of human and bacterial
ACMSD, demonstrating that the region surrounding
residues involved in catalysis is more hydrophobic in
the human, rather than in the bacterial protein.
Finally, this study demonstrated the expression in
kidney and liver of an alternatively spliced ACMSD

transcript coding for a protein which differs from the
other variant in the N-terminal region and carries an
incomplete metal binding domain. In particular, in
this variant only two out of four residues which are
directly involved in the metal cofactor binding are
present. Consistent with this structural deficiency is
our finding that ACMSD II is catalytically inactive
when overexpressed as recombinant protein. Inspec-
tion of the residues which in ACMSD I mark the
boundary of the predicted substrate-binding pocket,
reveals that some of them are conserved in the inac-
tive isoform (i.e. Trp191 and Phe294). Even though
the precise role of these active site residues has not
been established, the possibility that ACMSD II
might still be able to bind the substrate cannot be
definitively ruled out. If this would be the case, the
metabolic relevance of the inactive isoform would be
related to its capability of binding and sequestering a
reactive intermediate like ACMS.
The results we have presented represent the first bio-
chemical report on human ACMSD. They may be
instrumental in promoting the structural analysis of
the protein, particularly with respect to developing
therapeutic leads for treating disorders associated with
increased levels of quinolinate and ⁄ or picolinate.
Experimental procedures
PCR amplification and cloning of ACMSD
isoforms
Human brain, liver and kidney total RNA (Clontech
Laboratories, Inc., Mountain View, CA, USA) was reverse

transcribed in the presence of random primers, by using the
First Strand cDNA Synthesis kit (Biotech. Department Bio
Basic Inc., Markham, Ontario, Canada). Transcripts enco-
ding for the two splice variants were amplified by polym-
erase chain reaction (PCR) using specific primers, and
kidney cDNA as the template. PCR conditions were as fol-
lows: 1 min at 94 °C; 30 s at 94 °C, 30 s at 55 °C, 1 min at
72 °C for 30 cycles; 10 min at 72 °C. The reaction was per-
formed in the presence of 0.5 pmolÆlL
)1
of each primer,
200 lm dNTPs, 1 mm MgCl
2
and 0.025 unitsÆlL
)1
Taq
polymerase (Finnzymes, Espoo, Finland). The amplified
PCR products were separated on a 2% agarose gel, bands
were excised and purified by using the High Pure PCR
Product Purification kit (Roche, Basel, Switzerland).
Purified DNA was subcloned into pGEM T easy vector
(Promega, Madison, WI, USA) following manufacturer’s
instructions and recombinant plasmids were sequenced.
Real-time PCR
Real-time PCR was performed on a Corbett (Sidney, Aus-
tralia) Rotor Gene RG 3000. TaqMan probe and primers
were designed using the primer3 program. Optimized
amplification reactions contained 300 nm each primer,
400 nm probe, 5 mm MgCl
2

, 200 lm dNTPs, 1.25 Units
JumpStart Taq DNA polymerase (Sigma-Aldrich Corp., St
Louis, MO, USA) and 1 · of the provided buffer. All PCR
reactions were performed with one cycle of 94 °C for 60 s,
followed by 45 cycles of 15 s at 94 °C and 60 s at 60 °C.
Human ACMSD L. Pucci et al.
836 FEBS Journal 274 (2007) 827–840 ª 2007 The Authors Journal compilation ª 2007 FEBS
The efficiency of the PCR method using different pri-
mer ⁄ probe sets was determined from the threshold cycle
values obtained with 10-fold serial dilutions of linearized
pGEM T easy vectors harboring the ACMSD variants.
Standard curves for 18S RNA were obtained using cDNAs
from each organ. All the calibration curves exhibited slopes
ranging from 3.34 to 3.48, indicating comparable amplifica-
tion efficiency. The copy number of each variant in the
unknown samples was determined from the calibration
curves and data were normalized to the copy number of
18S RNA.
Expression of ACMSD isoforms in E. coli
pGEM-ACMSD I and pGEM-ACMSD II were digested
with XhoI and Bpu1102I and cloned into pET-15b and
pET-32b expression vectors. The four constructs obtained
were used to transform E. coli BL21(DE3) cells for proteins
production. E. coli cells harboring the recombinant plas-
mids were grown at 37 °C in Luria–Bertani medium,
containing 100 lgÆmL
)1
ampicillin. After reaching an A
600
of 0.3, cultures were shifted at 25 °C and expression was

induced with 1 mm isopropyl b-D-1-thiogalactopyranoside
at an A
600
of 0.6. Cells deriving from 1 mL cultures were
collected at different induction times and the pellets were
resuspended in 100 lLof20mm Tris ⁄ HCl buffer,
pH 8.0, 1 mm MgCl
2
,1mm dithiothreitol, 1 m m phenyl-
methanesulfonyl fluoride. Cells were disrupted by vortexing
with glass beads (250 lm) four times for 1 min, with 2 min
intervals. After centrifugation, the supernatants represent-
ing the soluble fractions and the corresponding pellets were
analyzed by SDS ⁄ PAGE [46].
Expression of ACMSD isoforms in P. pastoris
pHIL-D2 and pPIC9 vectors (Invitrogen, San Diego, CA,
USA) were used for intracellular and extracellular expres-
sion, respectively; pGEM plasmids harboring ACMSD I, II
were digested with EcoRI and the purified products were
ligated into the expression vectors at the EcoRI site. In the
resulting constructs, correct inserts orientation and
sequence were checked by automated sequencing. Transfor-
mation of P. pastoris GS115 competent cells with SalI line-
arized recombinant pPIC9 plasmids generated clones with a
His
+
Mut
+
phenotype. Transformation of the same yeast
cells with NotI linarized recombinant pHIL-D2 plasmids

favored the alcohol oxidase structural gene displacement,
allowing generation of His
+
Mut
S
phenotype [47]. PCR-pos-
itive His
+
Mut
+
and His
+
Mut
S
clones were cultured in
buffered glycerol-complex medium, and for expression
induction, cultures were transferred in buffered methanol-
complex medium and grown with daily methanol pulses, as
described in [47]. At different induction times, aliquots of
the culture filtrate of His
+
Mut
+
cells were subjected to
SDS ⁄ PAGE analysis and ACMSD activity determination.
The time course of ACMSD expression in His
+
Mut
S
cells

was monitored in the cells crude extracts, prepared by sus-
pending cells collected from 1 mL of the cultures, in
100 lLof50mm sodium phosphate, pH 7.4, 5% glycerol,
1mm phenylmethanesulfonyl fluoride, and disrupting them
by vortexing with glass beads (0.5 mm) eight times for 30 s
with 30-s intervals.
Site-directed mutagenesis
Mutagenesis to change the selected residues to Ala was
carried out using the QuickChange kit (Stratagene, La Jolla,
CA, USA). Mutagenic primers were: 5¢-CGCTCGAGA
TGAAAATTGACATC
GCTAGTCATATTCTACC-3¢ and
its complement for His6Ala; 5¢-GACATCCATAGT
GCT
ATTCTACCAAAAGAATGGCC-3¢ and its complement
for His8Ala (substituted nucleotides are underlined). pHIL-
D2-ACMSD I was used as the template for the PCR
mutagenesis reaction. Mutants were sequenced to verify
incorporation of the desired modification and to ensure the
absence of random mutations. The mutagenized plasmids
were transformed into P. pastoris GS115 competent cells and
expression was performed as for the wild-type protein. Crude
extracts were analyzed by SDS ⁄ PAGE and assayed for the
enzymatic activity.
ACMSD I purification
All steps were performed at 4 °C. 1 mm dithiothreitol and
5mm 2-mercaptoethanol were included in all buffers.
P. pastoris GS115 cells transformed with pHIL-D2-
ACMSD I were incubated in buffered glycerol complex
medium, at 30 °C, up to an A

600
of 3.0. The culture was
centrifuged at 5000 g for 10 min (Sorvall centrifuge RC5B
plus, Superlite GSA rotor) and the cell pellet was resus-
pended in one-fifth of the original culture volume of buf-
fered methanol complex medium. Cells were cultured for
5 days, at 30 °C, with daily addition of methanol to main-
tain a 0.5% concentration. Cells were harvested by centrifu-
gation as above, and resuspended in 10 mL of lysis buffer
containing 10 mm potassium phosphate, pH 7.0, 1 mm phe-
nylmethanesulfonyl fluoride, and 0.02 mgÆmL
)1
each of
leupeptin, antipain, chymostatin, and pepstatin. After dis-
ruption by three cycles of French Press (SLM-Aminco,
Urbana, IL, USA) at 1000 p.s.i., the suspension was centri-
fuged at 40 000 g for 20 min (Sorvall centrifuge RC5B plus,
Kontron-Hermle A8.24 rotor). The supernatant was made
10 mgÆmL
)1
by dilution with the lysis buffer, and strepto-
mycin sulfate was added dropwise at a final concentration
of 1%. After 20 min stirring, the sample was centrifuged at
8000 g for 10 min (Sorvall centrifuge RC5B plus, Kontron-
Hermle A8.24 rotor) and the supernatant was loaded onto
a hydroxyhapatite (Bio-Rad, Hercules, CA, USA) column
(2.5 cm · 4.9 cm
2
) equilibrated with 10 mm potassium
phosphate, pH 7.0, 50 mm NaCl. After washing with the

L. Pucci et al. Human ACMSD
FEBS Journal 274 (2007) 827–840 ª 2007 The Authors Journal compilation ª 2007 FEBS 837
same buffer, the recombinant protein was eluted with a lin-
ear gradient of potassium phosphate from 10 mm to 0.3 m,
containing 50 mm NaCl. The active pool was dialyzed
against 10 mm Tris ⁄ HCl, pH 8.0, 25 mm NaCl, and applied
to a fast protein liquid chromatography column of MonoQ
(Pharmacia, Stockholm, Sweden), equilibrated with the
dialysis buffer. After washing with the same buffer, elution
was performed with 10 mm Tris ⁄ HCl, pH 8.0, 0.13 m
NaCl. Active fractions were pooled and stored at 4 °C.
Gel filtration chromatography
ACMSD I native molecular weight was determined by gel-
filtration on a Superose 12 HR 10 ⁄ 30 column (Amersham,
Piscataway, NJ, USA). For elution, a buffer containing
10 mm Tris ⁄ HCl, pH 8.0, 0.2 m NaCl, 1 mm dithiothreitol
and 5 mm 2-mercaptoethanol was used. Standard proteins
were b-amylase (200 kDa), bovine serum albumin (66 kDa),
ovalbumin (45 kDa) and carbonic anhydrase (30 kDa).
Assay for ACMSD activity
ACMSD activity was determined spectrophotometrically,
by measuring ACMS consumption, as described in [23].
In our assay mixture, human 3-hydroxyanthranilic acid
dioxygenase, which is used to catalyze ACMS formation
from hydroxyanthranilic acid, was replaced with a dia-
lyzed crude extract of E. coli BL21 (DE3) cells expressing
the recombinant enzyme from Ralstonia metallidurans.
(Plasmid pKLC100 harboring R. metallidurans 3-hydroxy-
anthranilic acid dioxygenase was kindly provided by T.
Begley, Cornell University, New York, NY). Briefly, a

pre-assay mixture consisting of 25 lm hydroxyanthranilic
acid and 0.01 units of R. metallidurans 3-hydroxyanthrani-
lic acid dioxygenase, in 50 mm 4-morpholinepropanesulf-
onic acid, pH 6.0, was incubated at 25 °C, with
monitoring ACMS formation at 360 nm. After the reac-
tion was complete, an appropriate aliquot of ACMSD
was added and the activity was calculated by the
decrease in absorbance. Data were corrected for the
spontaneous decrease in absorbance due to the cycliza-
tion of ACMS to quinolinate. The effect of ACMS
concentration on the enzyme activity was investigated by
varying 3-hydroxyanthranilic acid concentration from 2
to 20 lm . Kinetic parameters were calculated from the
initial velocity data by using the Lineweaver-Burk plot.
One unit is defined as the amount of ACMSD consu-
ming 1 lmol ACMS per minute at 25 °C.
Three-dimensional structure prediction
Human ACMSD structure was modeled using the three-
dimensional coordinates of the P. fluorescens enzyme in
complex with Zn
2+
and Co
2+
(PDB codes 2HBV and
2HBX, respectively) using the SWISS-PDBviewer software
in conjunction with the SWISS-MODEL server (http://
www.expasy.org/spdbv/). The quality of the predicted struc-
ture was checked by computing its Ramachandran plot
using procheck program [48]: 86.6% of the residues were
in the most favored regions, 12.1% in the additional

allowed, three residues in the generously allowed (Asn148,
Leu246 and Glu298) and only two residues in the forbidden
regions (Asp181 and Met180). The human ACMSD–Zn
2+
complex was constructed by superimposing the ACMSD
model with the bacterial ACMSD–Zn
2+
complex and past-
ing the zinc ion in the predicted human structure. Figures
were generated with the software swiss-pdbviewer and
visual molecular dynamics [49].
Acknowledgements
We thank G. Magni and S. Ruggieri for helpful com-
ments and suggestions and T. Begley (Department of
Chemistry and Chemical Biology, Cornell University,
New York, NY) for the gift of the plasmid pKLC100.
References
1 Bender DA (1983) Biochemistry of tryptophan in health
and disease. Mol Aspects Med 6, 101–197.
2 Magni G, Amici A, Emanuelli M, Orsomando G,
Raffaelli N & Ruggieri S (2004) Enzymology of NAD
homeostasis in man. Cell Mol Life Sci 61, 19–34.
3 Stone TW (2001) Endogenous neurotoxins from trypto-
phan. Toxicon 39, 61–73.
4 Rios C & Santamaria A (1991) Quinolinic acid is a
potent lipid peroxidant in rat brain homogenates.
Neurochem Res 16, 1139–1143.
5 Schwarcz R, Whetsell WO & Mangano RM (1983)
Quinolinic acid: an endogenous metabolite that
produces axon-sparing lesions in rat brain. Science 219,

316–318.
6 Burns LH, Pakzaban P, Deacon TW, Brownell AL,
Tatter SB, Jenkins BG & Isacson O (1995) Selective
putaminal excitotoxic lesions in non-human primates
model the movement disorder of Huntington’s disease.
Neuroscience 64, 1007–1017.
7 Guidetti P, Luthi-Carter RE, Augood SJ & Schwarcz R
(2004) Neostriatal and cortical quinolinate levels are
increased in early grade Huntington’s disease. Neurobiol
Dis 17, 455–461.
8 Guillemin GJ, Wang L & Brew BJ (2005) Quinolinic
acid selectively induces apoptosis of human astrocytes:
potential role in AIDS dementia complex. J Neuroinflam
2, 16.
9 Guillemin GJ, Brew BJ, Noonan CE, Takikawa O &
Culler KM (2005) Indoleamine 2,3 dioxygenase and
Human ACMSD L. Pucci et al.
838 FEBS Journal 274 (2007) 827–840 ª 2007 The Authors Journal compilation ª 2007 FEBS
quinolinic acid immunoreactivity in Alzheimer’s disease
hippocampus. Neuropathol Appl Neurobiol 31, 395–404.
10 Bosco MC, Rapisarda A, Reffo G, Massazza S, Pasto-
rino S & Varesio L (2003) Macrophage activating prop-
erties of the tryptophan catabolyte picolinic acid. In
Developments in Triptophan and Serotonin Metabolism
(Allegri G, Costa CVL, Ragazzi E, Steinhart H &
Varesio L, eds), pp. 55–65. Kluwer Academic ⁄ Plenum
Publishers, Norwell, MA.
11 Leuthauser SW, Oberley LW & Oberley TD (1982)
Antitumor activity of picolinic acid in CBA ⁄ J mice.
J Natl Cancer Inst 68 , 123–129.

12 Blasi E, Bauer D, Barluzzi R, Mazzolla R & Bistoni F
(1994) Pattern of cytokine gene expression in brains of
mice protected by picolinic acid against lethal intracer-
ebral infection with Candida albicans. J Neuroimmunol
52, 205–213.
13 Ogata S, Takeuchi M, Fujita H, Shibata K, Okumura K
& Taguchi H (2000) Apoptosis induced by niacin-related
compounds in K562 cells but not in normal human
lymphocytes. Biosci Biotechnol Biochem 64, 1142–1146.
14 Fernandez-Pol JA, Klos DJ & Hamilton PD (2001)
Antiviral, cytotoxic and apoptotic activities of picolinic
acid on human immunodeficiency virus-1 and human
herpes simplex virus-2 infected cell. Anticancer Res 21,
3773–3776.
15 Rebello T, Lonnerdal B & Hurley L (1982) Picolinic
acid in milk, pancreatic juice and intestine: inadequate
for role in zinc absorption. Am J Clin Nutr 35, 1–9.
16 Evans GW & Johnson PE (1980) Characterization and
quantitation of a zinc binding ligand in human milk.
Pediatr Res 14 , 867–871.
17 Dazzi C, Candiano S, Massazza S, Ponzetto A & Vare-
sio L (2001) New high-performance liquid chromatogra-
phy method for the detection of picolinic acid in
biological fluids. J Chromatogr B 751, 61–68.
18 Medana IM, Day NPJ, Salahifar-Sabet H, Stocker R,
Smythe G, Bwanaisa L, Njobvu A, Kayira K, Turner
GDH, Taylor TT et al. (2003) Metabolites of the kynur-
enine pathway of tryptophan metabolism in the cerebro-
spinal fluid of Malawian children with malaria. J Infect
Dis 188, 844–849.

19 Clark CJ, Mackay GM, Smythe GA, Bustamante S,
Stone TW & Phillip RS (2005) Prolonged survival of a
murine model of cerebral malaria by kynurenine path-
way inhibition. Infect Immun 73, 5249–5251.
20 Beninger RJ, Colton AM, Ingles JL, Jhamandas K &
Boegman RJ (1994) Picolinic acid blocks the neurotoxic
but not the neuroexcitant properties of quinolinic acid
in the rat brain: evidence from turning behaviour and
tyrosine hydroxylase immunohistochemistry. Neuro-
science 61, 603612.
21 Li T, Iwaki H, Fu R, Hasegawa Y, Zhang H & Liu A
(2006) a-Amino-b-carboxymuconic-e-semialdehyde
decarboxylase (ACMSD) is a new member of the
amidohydrolase superfamily. Biochemistry 45, 6628–
6634.
22 Ichiyama A, Nakamura S, Kawai H, Honjo T, Nish-
izuka Y, Hayaishi O & Senoh S (1965) Studies on the
metabolism of the benzene ring of tryptophan in mam-
malian tissues. Enzymic formation of a-aminomuconic
acid from 3-hydroxyanthranilic acid. J Biol Chem 240,
740–749.
23 Egashira Y, Kouhashi H, Ohta T & Sanada H (1996)
Purification and properties of a
-amino-b-carboxy-
muconic-e-semialdehyde decarboxylase (ACMSD), key
enzyme of niacin synthesis from tryptophan, from hog
kidney. J Nutr Sci Vitaminol 42, 173–183.
24 Tanabe A, Egashira Y, Fukuoka SI, Shibata K &
Sanada H (2002) Purification and molecular cloning of
rat 2-amino-3-carboxymuconate-6-semialdehyde.

Biochem J 361, 567–575.
25 Egashira Y, Murotani G, Tanabe A, Saito K, Uehara K,
Morise A, Sato M & Sanada H (2004) Differential effects
of dietary fatty acids on rat liver a-amino-b-carboxymuc-
onate-e-semialdehyde decarboxylase activity and gene
expression. Biochim Biophys Acta 1686, 118–124.
26 Fukuwatari T, Ohsaki S, Fukuoka SI, Sasaki R &
Shibata K (2004) Phtalate esters enhance quinolinate
production by inhibiting a-amino-b-carboxymuconate-
e-semialdehyde decarboxylase (ACMSD), a key
enzyme of the tryptophan pathway. Toxicol Sci 81,
302–308.
27 Shin M, Ohnishi M, Iguchi S, Sano K & Umezawa C
(1999) Peroxisome-proliferator regulates key enzymes of
the tryptophan-NAD pathway. Toxicol Appl Pharmacol
158, 71–80.
28 Kimura N, Fukuwatari T, Sasaki R & Shibata K (2005)
The necessity of niacin in rats fed on a high protein
diet. Biosci Biotechnol Biochem 69, 273–279.
29 Tanabe A, Egashira Y, Fukuoka SI, Shibata K &
Sanada H (2002) Expression of rat hepatic 2-amino-
3-carboxymuconate-6-semialdehyde decarboxylase is
affected by a high protein diet and by streptozotocin-
induced diabetes. J Nutr 132, 1153–1159.
30 Shin M, Kim I, Inoue Y, Kimura S & Gonzalez FJ
(2006) Regulation of mouse hepatic alpha-amino-beta-
carboxymuconate-epsilon-semialdehyde decarboxylase, a
key enzyme in the tryptophan-nicotinamide adenine
dinucleotide pathway, by hepatocyte nuclear factor
4alpha and peroxisome proliferator-activated receptor

alpha. Mol Pharmacol 70, 1281–1290.
31 Colabroy KL & Begley TP (2005) Tryptophan catabo-
lism: identification and characterization of a new degra-
dative pathway. J Bacteriol 187, 7866–7869.
32 Muraki T, Taki M, Hasegawa Y, Iwaki H & Lau PCK
(2003) Prokaryotic homologs of the eukaryotic 3-hydro-
xyanthranilate 3,4-dioxygenase and 2-amino-3-carboxy-
muconate-6-semialdehyde decarboxylase in the
2-nitrobenzoate degradation pathway of Pseudomonas
L. Pucci et al. Human ACMSD
FEBS Journal 274 (2007) 827–840 ª 2007 The Authors Journal compilation ª 2007 FEBS 839
fluorescens strain KU-7. Appl Environ Microbiol 69 ,
1564–1572.
33 Li T, Walker AL, Iwaki H, Hasegawa Y & Liu A
(2005) Kinetic and spectroscopic characterization of
ACMSD from Pseudomonas fluorescens reveals a penta-
coordinate mononuclear metallocofactor. J Am Chem
Soc 127, 12282–12290.
34 Martynowski D, Eyobo Y, Li T, Yang K, Liu A & Zhang
H (2006) Crystal structure of a-amino-b -carboxymucon-
ic-e-semialdehyde decarboxylase: insight into the active
site and catalytic mechanism of a novel decarboxylase
reaction (2006). Biochemistry 45, 10412–10421.
35 Liu A & Zhang H (2006) Transition metal-catalyzed
nonoxidative decarboxylation reactions. Biochemistry
45, 10407–10411.
36 Fukuoka SI, Ishiguro K, Yanagihara K, Tanabe A,
Egashira Y, Sanada H & Shibata K (2002) Identifica-
tion and expression of a cDNA encoding human
a-amino-b-carboxymuconate-e-semialdehyde decarboxy-

lase (ACMSD). J Biol Chem 277, 35162–35167.
37 Kyte J & Doolittle RF (1982) A simple method for dis-
playing the hydropathic character of a protein. J Mol
Biol 157, 105–132.
38 Pawlak D, Tankiewicz A, Matys T & Buczko W (2003)
Peripheral distribution of kynurenine metabolites and
activity of kynurenine pathway enzymes in renal failure.
J Physiol Pharmacol 54, 175–189.
39 Saito K, Fujigaki S, Heyes MP, Shibata K, Takemura
M, Fuji H, Wada H, Noma A & Seishima M (2000)
Mechanism of increases in 1-kynurenine and quinolinic
acid in renal insufficiency. Am J Physiol Renal Physiol
279, F565–F572.
40 Egashira Y, Nagaki S & Sanada H (2006) Trypto-
phan-niacin metabolism in rat with puromycin amino-
nucleoside-induced nephrosis. Int J Vitam Nutr Res
76, 28–33.
41 Fukuwatari T, Morikawa Y, Hayakawa F, Sugimoto E
& Shibata K (2001) Influence of adenine-induced renal
failure on tryptophan-niacin metabolism in rats. Biosci
Biotechnol Biochem 65, 2154–2161.
42 Bender DA & Olufunwa R (1988) Utilization of trypto-
phan, nicotinamide and nicotinic acid as precursors for
nicotinamide nucleotide synthesis in isolated rat liver
cells. Br J Nutr 59, 279–287.
43 Ying W (2006) NAD and NADH in cellular functions
and cell death. Front Biosci 11, 3129–3148.
44 Sasaki Y, Araki T & Milbrandt J (2006) Stimulation of
nicotinamide adenine dinucleotide biosynthetic path-
ways delays axonal degeneration after axotomy. J Neuro-

science 26, 8484–8491.
45 Cooper BF, Sideraki V, Wilson DK, Dominguez DY,
Clark SW, Quiocho FA & Rudolph FB (1997) The
role of divalent cations in structure and function of
murine adenosine deaminase. Protein Sci 6, 1031–
1037.
46 Schagger H & Von Jagow G (1987) Tricine-sodium
dodecyl sulfate-polyacrylamide gel electrophoresis for
the separation of proteins in the range from 1 to 100
kDa. Anal Biochem 166, 368–379.
47 Pichia Expression Kit
A Manual of Methods for Expres-
sion of Recombinant Proteins in Pichia Pastoris. Invitro-
gen Corporation, San Diego, CA.
48 Laskowski RA, MacArthur MW, Moss DS & Thornton
JM (1993) procheck: a program to check the stereoche-
mical quality of protein structures. J Appl Cryst 26,
283–291.
49 Humphrey W, Dalke A & Schulten K (1996) VMD:
visual molecular dynamics. J Molec Graphics 14 1, 19–25.
Human ACMSD L. Pucci et al.
840 FEBS Journal 274 (2007) 827–840 ª 2007 The Authors Journal compilation ª 2007 FEBS