Characterization of the folate salvage enzyme p-aminobenzoylglutamate hydrolase in plants

Article abstract of DOI:10.1016/j.phytochem.2007.06.031

Folates break down in vivo to give pterin and p-aminobenzoylglutamate (pABAGlu) fragments, the latter usually having a polyglutamyl tail. Pilot studies have shown that plants can hydrolyze pABAGlu and its polyglutamates to p-aminobenzoate, a folate biosynthesis precursor. The enzymatic basis of this hydrolysis was further investigated. pABAGlu hydrolase activity was found in all species and organs tested; activity levels implied that the proteins responsible are very rare. The activity was located in cytosol/vacuole and mitochondrial fractions of pea (Pisum sativum L.) leaves, and column chromatography of the activity from Arabidopsis tissues indicated at least three peaks. A major activity peak from Arabidopsis roots was purified 86-fold by a three-column procedure; activity loss during purification exceeded 95%. Size exclusion chromatography gave a molecular mass of ～200 kDa. Partially purified preparations showed a pH optimum near 7.5, a K_m value for pABAGlu of 370 μM, and activity against folic acid. Activity was relatively insensitive to thiol and serine reagents, but was strongly inhibited by 8-hydroxyquinoline-5-sulfonic acid and stimulated by Mn²⁺, pointing to a metalloenzyme. The Arabidopsis genome was searched for proteins similar to Pseudomonas carboxypeptidase G, which contains zinc and is the only enzyme yet confirmed to attack pABAGlu. The sole significant matches were auxin conjugate hydrolase family members and the At4g17830 protein. None was found to have significant pABAGlu hydrolase activity, suggesting that this activity resides in hitherto unrecognized enzymes. The finding that Arabidopsis has folate-hydrolyzing activity points to an enzymatic component of folate degradation in plants.

Full text of DOI:10.1016/j.phytochem.2007.06.031

Available online at www.sciencedirect.com
PHYTOCHEMISTRY
Phytochemistry 69 (2008) 29–37
www.elsevier.com/locate/phytochem
Characterization of the folate salvage enzyme
p-aminobenzoylglutamate hydrolase in plants
a
a
b
a
Gale G. Bozzo , Gilles J.C. Basset , Valeria Naponelli , Alexandre Noiriel ,
b
a,
*
Jesse F. Gregory III , Andrew D. Hanson
a
Horticultural Sciences Department, University of Florida, Gainesville, FL 32611, United States
Food Science and Human Nutrition Department, University of Florida, Gainesville, FL 32611, United States
b
Received 20 January 2007; received in revised form 20 June 2007
Available online 14 August 2007
Abstract
Folates break down in vivo to give pterin and p-aminobenzoylglutamate (pABAGlu) fragments, the latter usually having a polyglut-
amyl tail. Pilot studies have shown that plants can hydrolyze pABAGlu and its polyglutamates to p-aminobenzoate, a folate biosynthesis
precursor. The enzymatic basis of this hydrolysis was further investigated. pABAGlu hydrolase activity was found in all species and
organs tested; activity levels implied that the proteins responsible are very rare. The activity was located in cytosol/vacuole and mito-
chondrial fractions of pea (Pisum sativum L.) leaves, and column chromatography of the activity from Arabidopsis tissues indicated
at least three peaks. A major activity peak from Arabidopsis roots was purified 86-fold by a three-column procedure; activity loss during
purification exceeded 95%. Size exclusion chromatography gave a molecular mass of ꢀ200 kDa. Partially purified preparations showed a
m
pH optimum near 7.5, a K value for pABAGlu of 370 lM, and activity against folic acid. Activity was relatively insensitive to thiol and
2
+
serine reagents, but was strongly inhibited by 8-hydroxyquinoline-5-sulfonic acid and stimulated by Mn , pointing to a metalloenzyme.
The Arabidopsis genome was searched for proteins similar to Pseudomonas carboxypeptidase G, which contains zinc and is the only
enzyme yet confirmed to attack pABAGlu. The sole significant matches were auxin conjugate hydrolase family members and the
At4g17830 protein. None was found to have significant pABAGlu hydrolase activity, suggesting that this activity resides in hitherto
unrecognized enzymes. The finding that Arabidopsis has folate-hydrolyzing activity points to an enzymatic component of folate degra-
dation in plants.
ꢀ
2007 Elsevier Ltd. All rights reserved.
Keywords: Arabidopsis thaliana; Crucifereae; Folate salvage; Enzyme isolation and characterization; p-Aminobenzoylglutamate hydrolase; Folate hyd-
rolase
1
. Introduction
not and so need a dietary supply (Cossins, 2000; Scott
et al., 2000).
Tetrahydrofolate (THF) and its one-carbon derivatives
collectively termed folates) are essential cofactors for
Folates readily undergo spontaneous oxidative degrada-
tion in physiological conditions, yielding a pterin 4 and p-
aminobenzoylglutamate 5 (pABAGlu) or its polygluta-
mates (Fig. 1a) (Suh et al., 2001). There is evidence that this
breakdown process is particularly active in plants and that,
in vivo, plants can hydrolyze the resulting pABAGlu 5 moi-
eties to pABA 7 and glutamate 6 (Fig. 1b) (Orsomando
et al., 2006). Acting in concert with a reductive reaction
that recycles the pterin fragment to a folate synthesis pre-
cursor, pABAGlu 5 hydrolysis enables complete salvage
(
one-carbon transfer reactions. Folate molecules consist of
pterin 4, p-aminobenzoate (pABA 7), and glutamate 6
moieties, usually with a short, c-linked polyglutamyl tail
attached to the first glutamate (Fig. 1a). Plants, fungi,
and bacteria can synthesize folates but higher animals can-
*
Corresponding author. Tel.: +1 352 392 1928; fax: +1 352 392 5653.
E-mail address: adha@mail.ifas.ufl.edu (A.D. Hanson).
0
031-9422/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.phytochem.2007.06.031

30
G.G. Bozzo et al. / Phytochemistry 69 (2008) 29–37
5
is carboxypeptidase G (CPG, EC 3.4.17.11), a di-zinc
a
Folate 1
γ-Glu tail 2
enzyme from Pseudomonas and other bacteria (McCul-
lough et al., 1971; Albrecht et al., 1978; Sherwood et al.,
1985). CPG also cleaves the pABA–glutamate bond in
folates and folate analogs, releasing pteroate 3 and gluta-
mate 6 fragments, and can remove the c-glutamyl tail from
polyglutamates by exopeptidase action. The Pseudomonas
enzyme has been cloned (Minton et al., 1983). Other than
CPG, there is genetic evidence that the Escherichia coli
abgA and abgB gene products (which share weak sequence
similarity with CPG) have PGH activity (Hussein et al.,
O
OH
O
OH
H
N
O
O
OH
O
N
H
N
N
H
N
H
O
O
HN
N
H
O
OH
H N
2
N
H
n
Pteroate 3
Pterin 4
pABAGlu 5
1
998; Carter et al., 2007). In addition, microorganisms
O
OH
O
O
and mammals are known to have an enzyme that releases
pteroate from folate (Oe et al., 1983); this enzyme has nei-
ther been tested with pABAGlu 5 as substrate nor the
encoding gene cloned.
b
OH
O
OH
N
H
H
N
2
H N
2
After an initial survey of PGH activity in diverse plants,
we determined the subcellular location of PGH activity in
pea leaves and fractionated and characterized the activity
from Arabidopsis leaves and roots. We also screened all
Arabidopsis proteins with significant sequence similarity
to CPG for PGH activity.
p
ABAGlu 5
Glu 6
pABA 7
GTP
c
Chorismate
HMDHP-P2
HMDHP
pABA
2. Results and discussion
DHP
PGH
Glu
2
.1. Survey of pABAGlu hydrolase activity
DHF
THF
pABAGlu
PTAR
Desalted extracts of various plant tissues were surveyed
for PGH activity using a radioassay based on product sep-
aration by TLC (see Section 4, assay B). This assay, which
uses a subsaturating concentration of [ C]pABAGlu 5
Glu
GGH
14
THFGlu
n
(
18–26 lM), is more sensitive and specific than that
p
ABAGlu
n
DHPA
employed previously (Orsomando et al., 2006). PGH activ-
ity was readily detected in all tissues analyzed, which
included those tested in a pilot study (Orsomando et al.,
Fig. 1. Structure of folates and hydrolysis of the pABAGlu moiety. (a)
Tetrahydrofolate polyglutamate. Polyglutamyl tails of plant folates
contain up to about six residues. Oxidative scission of the C9-N10 bond
2
006).
PGH activity was consistently higher in roots than in
(
arrow) yields pterin 4 and pABAGlu 5 fragments. (b) The pABAGlu 5
leaves, the difference ranging from 3- to 80-fold in Arabid-
opsis, pea, and maize (Zea mays L.) (Fig. 2). It notewor-
thy that, for pea at least, this pattern is the inverse of that
for folate biosynthesis enzymes, whose protein and RNA
levels are about 5-fold lower in roots than in leaves (Jab-
rin et al., 2003). Roots from in vitro cultured Arabidopsis
plantlets had less activity than those grown hydroponi-
cally but were easier to produce and had no microbial
contaminants, and so were chosen for subsequent work
on roots.
hydrolysis reaction. (c) Folate salvage reactions (broad arrows) in relation
to folate biosynthesis. The pABAGlu 5 fragment from folate 1 cleavage
n
(and its polyglutamyl 2 forms, pABAGlu ) can be recycled for use in folate
synthesis after hydrolysis. Removal of the polyglutamyl tail by c-glutamyl
hydrolase (GGH) may be needed before pABAGlu hydrolase (PGH) acts.
The pterin 4 fragment (dihydropterin-6-aldehyde, DHPA) from folate
cleavage can be recycled to the folate synthesis intermediate hydroxy-
methyldihydropterin (HMDHP) by the action of pterin aldehyde
reductase (PTAR). Other abbreviations: DHP, dihydropteroate; DHF,
n 2
dihydrofolate; Glu , poly-c-glutamate; THF, tetrahydrofolate; -P ,
diphosphate.
of folate 1 breakdown products (Fig. 1c). Although pABA-
Glu hydrolase (PGH) activity was detected in extracts of
Arabidopsis and pea (Pisum sativum L.) leaves and of
tomato (Lycopersicon esculentum Mill.) fruit, nothing fur-
ther is known about this enzyme (Orsomando et al., 2006).
Nor, with one exception, is much known about pABA-
Glu-hydrolyzing enzymes from other organisms. The only
well-characterized protein known to hydrolyze pABAGlu
2.2. Subcellular distribution of pABAGlu hydrolase activity
PGH was localized by cellular fractionation and
enzyme assay in pea leaves, which – while low in PGH
activity – are the tissue of choice for obtaining high yields
of intact organelles (Baldet et al., 1993) and have been
much used in folate biochemistry (e.g., Chen et al.,
1997; Jabrin et al., 2003). Marker enzyme assays con-

G.G. Bozzo et al. / Phytochemistry 69 (2008) 29–37
31
1
000
1.2
PGH
1
0
0
0
0
.0
.8
.6
.4
.2
0
1
00
1
0
1
1
00
MTHFR
GAPDH
80
6
4
2
0
0
0
p
0
.1
0
2
5
20
15
1
4
Fig. 2. pABAGlu hydrolase activities in various plant tissues. [ C]pABA
14
7
production was measured by assay B, with [ C]pABAGlu 5 concentra-
tion of 18–26 lM. Pea (Ps), Arabidopsis (At) and maize (Zm) leaves were
, 28, and 11-days-old, respectively. Pea, maize, Arabidopsis cultured roots
-c) and hydroponic (-h) roots were 10, 6, 23, and 48-days old, respectively.
9
(
1
0
Tomato (Le) pericarp was in the mature green stage. Spinach (So) leaves
were fully expanded. Values are the means of three replicates and SE. Note
that the scale is logarithmic.
5
0
2
5
0
5
Fumarase
2
1
firmed that purified chloroplasts and mitochondria were
essentially uncontaminated by other fractions (Fig. 3).
PGH activity was detected in mitochondrial and cytosol
plus vacuole fractions, but not in chloroplasts (Fig. 3).
Multiplying the specific activity of PGH in each fraction
by the percentage of cellular protein present in that frac-
tion indicated that 89% of the cellular activity is in the
cytosol plus vacuole fraction (Table 1). The PGH activity
of vacuolar preparations was not enriched relative to the
cytosol/vacuole fraction, although the vacuolar marker a-
mannosidase was enriched up to three-fold (not shown).
This indicates that PGH is certainly not solely vacuolar
but is possibly partially so. In any case, the occurrence
of activity in two subcellular fractions signals the presence
of at least two PGH isoforms.
10
5
0
M
CP
CS
Fig. 3. Localization of pABAGlu hydrolase activity in pea mesophyll cells
by subcellular fractionation. Chloroplasts (CP) and mitochondria (M)
were purified on Percoll gradients. A fraction enriched in cytosol and
vacuole contents (CS) was prepared from pea leaf protoplasts by pelleting
intact organelles. The specific activities of PGH and marker enzymes were
assayed in each fraction. Markers were: NADP-linked glyceraldehyde-3-
phosphate dehydrogenase (GAPDH, chloroplast), fumarase (mitochon-
drion), and methylenetetrahydrofolate reductase (MTHFR, cytosol). Data
are the means and SE of data from three independent preparations of each
fraction.
2.3. Chromatographic separation of pABAGlu hydrolase
activity
Table 1
Calculated distribution of pABAGlu hydrolase activity among subcellular
fractions of pea leaves
To further investigate PGH, activity from Arabidopsis
leaves and roots was precipitated with (NH ) SO and sep-
arated by hydrophobic interaction (Octyl-Sepharose) fol-
lowed by anion exchange (Mono Q) and gel filtration
a
b
Fraction
PGH activity Protein (% of PGH distribution
4
2
4
(
fkat/mg)
cell total)
(% of cell total)
Cytosol + vacuole 0.80
26
4
70
89
11
0
Mitochondria
Chloroplasts
a
0.67
0
(
Superdex 200) steps. The activity from both leaves and
roots eluted as one broad peak from Octyl-Sepharose col-
umns (not shown) and also from the Mono Q column,
although a shoulder was sometimes evident in the Mono
Q peak (Figs. 4 and 5). Subsequent Superdex 200 fraction-
ation yielded a single 200-kDa peak from roots, but three
peaks from leaves, with estimated molecular masses of
Values taken from Fig. 3.
b
Values from Jabrin et al. (2003).
could also be explained by oligomerization of a single
protein.
9
0, 200, and 360 kDa (Figs. 4 and 5). In some leaf prepara-
tions, an additional 40-kDa peak was also present (not
shown). While consistent with the existence of various
PGH isoforms (i.e., various gene products), these data
Representative purifications from leaves and roots are
summarized in Table 2. Overall purification was 268-fold
from leaves and 86-fold from roots; activity was generally

32
G.G. Bozzo et al. / Phytochemistry 69 (2008) 29–37
Leaf
Root
0
0
0
0
0
.25
.20
.15
.10
.05
1.0
0.8
3
3
2
.5
.0
.5
Mono Q
Mono Q
0
0
0
0
.6
.4
.2
0
.5
0
0
0
.5
0
.5
0.25
0
2.0
.25
1
1
0
.5
.0
.5
0
0
0
0
.1
Superdex
0
10
20
30
0
0
0
.75
200
2
.5
Superdex
200
0
0
0
.1
.50
.25
0
2
.0
0
.05
9
0
1.5
3
60
.05
1
.0
0
0
.5
5
10
15
20 25 30
Fraction number
Fig. 4. Elution profiles of Arabidopsis leaf pABAGlu hydrolase activity
0
from Mono Q and Superdex 200 columns. After (NH SO and Octyl-
4
)
2
4
0
10
20
30
40
50
Sepharose steps, activity was applied to a Mono Q column, whose peak
fractions were then applied to the Superdex 200 column. Fraction volumes
were 300 lL for Mono Q, 320 lL for Superdex 200.
Fraction number
Fig. 5. Elution profiles of Arabidopsis root pABAGlu hydrolase activity
from Mono Q and Superdex 200 columns. After (NH SO and Octyl-
4
)
2
4
Sepharose steps, activity was applied to a Mono Q column, whose peak
fractions were then applied to the Superdex 200 column. Fraction volumes
were 300 lL for Mono Q, 400 lL for Superdex 200.
lost at each column step, and final activity recoveries were
only 4–7%. SDS–PAGE analysis of the final products
revealed many protein bands (not shown). Attempts at fur-
ther purification by dye-affinity chromatography resulted
in P90% loss of the remaining activity.
tion work was done mainly with root PGH using fractions
from this peak.
PGH enzymes are probably rare proteins, as the follow-
ing theoretical calculation shows. Taking Arabidopsis
PGHs to have the same specific activity as pure CPG
Root PGH was optimally active at pH 7.3–7.5, with
half-maximal activity at pH 6.5 and 8.3. Inhibitors were
used to test whether the activity is due to a serine-, cys-
teine-, aspartic-, or metallohydrolase (Table 3). Activity
was fairly insensitive to serine, cysteine, and aspartic
reagents, but was inhibited totally by one of the metal che-
lators tested (8-hydroxyquinoline-5-sulfonate) and moder-
ately (53%) by another (TPEN). Moreover, the root
enzyme showed modest but significant stimulation by
(
3.6–12 lkat/mg) (McCullough et al., 1971; Sherwood
et al., 1985) and the activity in cultured roots (Fig. 2),
extrapolated to Vmax (see below), as ꢀ0.4 pkat/mg. then
7
the predicted abundance of PGH proteins is 61 in 10 .
2
.4. Enzymatic properties of root pABAGlu hydrolase
2
+
2+
2+
0
.1 mM Mn (54%; P < 0.05) but not by Zn or Ni ;
2
+
2+
As roots were richer in PGH activity than leaves (Fig. 2)
and gave one major peak on Superdex 200, characteriza-
Cu and Hg abolished activity. Taken together, these
data point to a metalloenzyme. Compounds (1 mM) found

G.G. Bozzo et al. / Phytochemistry 69 (2008) 29–37
33
Table 2
Purification of pABAGlu hydrolase from Arabidopsis leaves and roots
Step
Total protein (mg)
Total activity (pkat)
Specific activity (pkat/mg)
Purification (-fold)
Yield (%)
a
Leaves
Crude leaf extract
NH SO (50–80%)
Octyl-Sepharose HiTrap
Mono Q HR 5/5
70
13
5.6
0.35
0.018
0.369
0.063
0.093
0.051
0.026
0.0053
0.0049
0.0174
0.146
1
0.9
3.3
28
100
17
25
14
7
(
4
)
2
4
Superdex 200 HR 10/30
1.419
268
b
Roots
Crude root extract
92
87
5.1
0.38
0.045
5.07
6.74
1.58
0.42
0.21
0.055
0.089
0.309
1.12
4.77
3
1
100
133
31
8
(
NH
4
)
2
SO
4
(0–80%)
1.6
5.6
20.2
86
Octyl-Sepharose 4 Fast Flow
Mono Q HR 5/5
Superdex 200 HR 10/30
a
4
About 30 g of leaves were used. Enzyme activity was measured using assay A and a [ H]pABAGlu concentration of 5 lM.
About 75 g of roots were used. Enzyme activity was measured using assay B and a [ C]pABAGlu concentration of 18.2 lM.
b
14
Table 3
Inhibitor sensitivity of pABAGlu hydrolase from Arabidopsis roots
2
1
0
5
a
Inhibitor
Target
Activity (% of
control)
EDTA 1 mM
Metals
Metals
Metals
Metals
79
75
47
0
0
.08
0.06
.04
1
,10-Phenanthroline 100 lM
b
10
5
TPEN 100 lM
-Hydroxyquinoline-5-sulfonate
mM
Phenylmethylsulfonylfluoride
0
8
1
p
0.02
0
Serine,
cysteine
85
75
72
0
0.4
0.8
1.2
1.6
1
mM
0
p
0
5
,5 -Dithio-bis(2-nitrobenzoic acid) Cysteine
0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1
mM
Pepstatin 1.5 lM
Aspartate
[
p
ABAGlu] (mM)
a
Control activity in the absence of inhibitor was 3.6–4.8 pkat/mg,
1
4
Fig. 6. Kinetic analysis of Arabidopsis root pABAGlu hydrolase
activity. Activity was measured using assay A, and various concentra-
tions of [ C]pABAGlu 5. Incubation time was 5 h. The main graph is a
velocity versus pABAGlu 5 concentration plot, with mean values (± SE)
for three independent experiments; the inset is a Hanes plot for the
mean data.
measured at a [ C]pABAGlu concentration of 18.2–25.9 lM using assay
B. Values are means of three determinations.
b
14
N,N,N’,N’-tetrakis[2-pyridylmethyl]ethylenediamine.
to have little (624%) effect on activity were ascorbate, glu-
tathione, b-mercaptoethanol, and ATP (not shown). DTT
(
1 mM) caused 53% inhibition.
The relation between velocity and pABAGlu concentra-
2.5. Carboxypeptidase G-like proteins in Arabidopsis
tion was Michaelian for the root enzyme (Fig. 6), with an
apparent Km value of 370 ± 19 lM (mean ± SE of three
determinations). Root PGH activity was inhibited by folic
acid 1; measured at a pABAGlu 5 concentration of 19 lM,
the concentration of folic acid 1 giving 50% inhibition
The only enzyme certainly known to hydrolyze pABA-
Glu 5 is CPG (McCullough et al., 1971) although genetic
evidence indicates that the E. coli AbgA and AbgB pro-
teins, which are weakly similar to CPG, also do so (Hussein
et al., 1998; Carter et al., 2007). As all three belong to the
M20 metallopeptidase family, and our data implicated a
metalloenzyme (Table 3), we searched the Arabidopsis gen-
ome for similar proteins. CPG, AbgA, and AbgB sequences
gave significant matches only to auxin conjugate hydro-
lases (LeClere et al., 2002) and to the At4g17830 protein,
a putative metallopeptidase. The auxin conjugate hydro-
lase family of Arabidopsis includes four Mn -dependent
proteins that cleave indole acetic acid-amino acid conju-
gates (ILL1, ILL2, ILR1, and IAR3) and two proteins
(ILL3 and ILL6) whose substrate is unknown (LeClere
et al., 2002).
(
IC ) was 52 ± 8 lM (mean ± SE of three determina-
50
tions). In view of this observation, and because CPG
cleaves folates as well as pABAGlu 5 (McCullough et al.,
1
971), we tested folic acid 1 as a substrate for the root
enzyme preparation, using a TLC radioassay (Oe et al.,
983) similar to that used to measure pABAGlu 5 hydroly-
1
sis. This assay indicated that folic acid 1 was cleaved to
pteroate 3 (Fig. 7a), which was confirmed by HPLC
2
+
(
Fig. 7b). As the root enzyme preparation was not homo-
geneous, the activity against folic acid 1 cannot necessarily
be ascribed to the same protein(s) as PGH activity. This is,
however, the most economical explanation.

34
G.G. Bozzo et al. / Phytochemistry 69 (2008) 29–37
PGH
4
3
2
1
IAAAH
a
kDa
kDa
79
+
Enzyme
Pte
Fol
pABAGlu
16
7
5
49
2
8
1
2
8
0
4
4
0
-
Enzyme
3
2
1
0
Fig. 8. Screening recombinant Arabidopsis CPG-like proteins for PGH
activity. The auxin conjugate hydrolase family members were expressed as
glutathione S-transferase (GST) fusions in pGEX-KTO (KTO);
At4g17830 was expressed with a C-terminal histidine tag in pET28b.
Enzyme activities were assayed on proteins purified by affinity chroma-
tography. Denaturing gel electrophoresis confirmed the presence of
0
0.2
0.4
0.6
0.8
1
Rf
+
Enzyme
-Enzyme
b 8
0
Fol
Fol
2
1
1
0
0
.0
.6
.2
.8
.4
0
2.0
recombinant protein bands (inset) with the expected M values (28 kDa
r
8
0
for GST; 75–79 kDa for GST–auxin conjugate hydrolase fusions; 49 kDa
for At4g17830). PGH activity was measured in the presence of 1 mM
MnCl ; results with 1 mM ZnCl or with no added divalent cations were
2 2
the same. As positive controls, ILL2 and IAR3 were tested for indole
acetic acid–alanine hydrolase (IAAAH) activity in the presence of 1 mM
1.6
1.2
0.8
0.4
0
6
0
6
0
2
MnCl . Data are means of duplicate determinations.
4
2
0
0
40
26
28 30
26 28 30
3. Concluding remarks
2
0
Pte
Our evidence is consistent with the pABAGlu hydrolase
Pte
0
0
activity of both pea and Arabidopsis being due to two or
more isoforms, at least one of which is a metalloenzyme
and all of which are of low abundance. The existence of
more than one isoform is presaged by the situation in
E. coli, in which disrupting abgA or abgB does not much
affect PGH activity (Hussein et al., 1998).
2
6
28
Minutes
30
26
28
30
Minutes
Fig. 7. Evidence for folic acid-hydrolyzing activity in the pABAGlu
hydrolase preparation from Arabidopsis roots. (a) TLC separation of
radioactivity after incubation of [ H]folic acid 1 (12 nCi, 0.28 pmol) in 10-
3
lL reaction mixtures for 5 h plus or minus enzyme (0.7 lg). Positions of
co-chromatographed standards are indicated. Note the absence of label
from the pABAGlu zone; this confirms that only the pABA–glutamate
bond of folic acid was cleaved, not the pterin–pABA bond. (b) HPLC
separation of 10-lL reaction mixtures containing 5 nmol of folic acid 1
after incubation for 5 h plus or minus enzyme (0.18 lg). The inset shows
the pteroate 3 peak on a 25-fold magnified UV absorbance scale. Running
positions of standards are indicated.
An obvious approach to identifying plant PGH genes –
screening all Arabidopsis homologs of bacterial PGHs for
activity – indicated that none of them is a PGH. From this
it follows that plant proteins with PGH activity must be
novel in the sense either that they are: (a) known enzymes
not yet known to attack pABAGlu 5, or (b) proteins whose
activity is so far unknown. The former possibility seems
likelier, for two reasons. First, pABAGlu 5 has never been
tested as a substrate for most peptidases, which are good
candidates a priori. Second, the tyrosine-pABA bond in
the synthetic peptide N-benzoyl-L-tyrosyl-pABA is cleaved
by chymotrypsin (Yamato and Kinoshita, 1978). Given the
failure of sequence homology to find plant proteins with
PGH activity, this task clearly requires an approach that
makes no prior assumptions about the nature of the
enzymes, such as functional complementation screening
in bacteria. The pilot studies of Hussein et al. (1998) on
the abgA and abgB genes in E. coli suggest that this
approach may be feasible.
The six members of the auxin conjugate hydrolase
group and the At4g17830 protein were expressed in
E. coli and tested for PGH activity. The former were
expressed as GST fusions and purified using glutathi-
one–agarose columns. At4g17830 was expressed with a
hexahistidine tag and purified by Ni affinity chromatog-
raphy. The IAA-alanine hydrolyzing activities of two of
the most efficient auxin conjugate hydrolases – ILL2
and IAR3 – were used as positive controls, and were
found to be similar to those previously reported (Davies
et al., 1999; LeClere et al., 2002). Although all seven pro-
teins were enriched to near-homogeneity, none of them
had significant PGH activity; only ILR1 showed any trace
of activity (Fig. 8).
2
+
Lastly, our finding that the enzyme preparation from
Arabidopsis roots can cleave folates is not unprecedented
inasmuch as CPG is known to do the same thing (McCul-
lough et al., 1971). It nevertheless emphasizes an often-

G.G. Bozzo et al. / Phytochemistry 69 (2008) 29–37
35
1
4
overlooked possibility, namely that folate breakdown is in
part enzymatic – and hence subject to regulation – as well
as merely chemical (Suh et al., 2001).
tion was for 3–8 h at 30 ꢁC. [ C]pABA 7 was separated
either by EtOAc partitioning (Orsomando et al., 2006)
(assay A) or by TLC (assay B). [ H]pABA 7 was separated
3
by the former procedure. In the latter, reactions were
mixed with unlabeled pABA 7 and pABAGlu 5 carriers
(100 nmol each) and applied to 1-cm origins on 10-cm silica
gel 60 F254 TLC plates (Merck, Darmstadt, Germany).
4
. Experimental
4
.1. Reagents
After developing with EtOAc:MeOH:H O (77:13:10, v/v/
2
v), the pABA zone (R 0.9) was scraped out for scintillation
counting. [ C]pABA 7 formation was linear with respect
f
1
4
3
14
[
Ring- C]pABA 7 (55 mCi/mmol), [3,5- H]pABA 7
0
0
3
(
26.2 Ci/mmol) and [3 ,5 ,7,9- H]folic acid 1 (43.2 Ci/
to time and amount of enzyme. Routine assays in this
1
4
mmol) were from Moravek Biochemicals (Brea, CA,
study used a subsaturating [ C]pABAGlu 5 concentration,
4.5-fold less than that used previously (Orsomando et al.,
2006); the activity values reported are consequently lower.
Folate hydrolase activity was assayed by a TLC method
1
4
3
USA). [ C]pABAGlu 5 and [ H]pABAGlu 5 were pre-
3
pared as described (Orsomando et al., 2006). [ H]Folic acid
1
(
was purified before use by folate affinity chromatography
Gregory and Toth, 1988). All protein chromatography
columns were from GE Healthcare (Piscataway, NJ, USA).
3
(Oe et al., 1983) similar to assay B, using [ H]folic acid 1
as substrate and quantifying pteroate release. The identity
of pteroate 3 was confirmed by HPLC with UV detection
(300 nm) essentially as described (D ´ı az de la Garza et al.,
4
.2. Plant material
2004). Apparent Km values were determined from Hanes
Arabidopsis thaliana (L.) Heynh. ecotype Columbia
plots (Fig. 6).
plants for leaf production were grown at 23–28 ꢁC in 12-
ꢁ2
ꢁ1
h days (photosynthetic photon flux density 80 lE m
s
)
4.5. Protein extraction and purification
in potting soil irrigated with water. For root production,
plants were grown hydroponically as described (Gibeaut
et al., 1997). Arabidopsis plantlets were cultured axenically
in 0.5· liquid MS salts containing 10 g/L sucrose (Prabhu
et al., 1996). Tomato Lycopersicon esculentum (cv. Micro-
Tom) and pea (Pisum sativum) (cv. Laxton’s Progress 9),
plants were grown as described (D ´ı az de la Garza et al.,
All steps were at 0–4 ꢁC. Protein was estimated by the
dye-binding method (Bradford, 1976). For tests of PGH
activity in various plant sources, proteins were extracted
and desalted as described (Orsomando et al., 2006). For
protein purification, leaves from seven-week old Arabidop-
sis plants or roots from 25- to 28-day-old cultures were trit-
2
004; Orsomando et al., 2005). Maize (Zea mays cv.
urated in liquid N and suspended (1 g/2.5 mL) in 50 mM
2
NK508) leaves were from 11-day old plants grown in pot-
ting soil in a naturally lit greenhouse. Spinach (Spinacia
oleracea) was purchased locally.
KPi, pH 8.0 containing 3% (w/v) polyvinylpolypyrroli-
done. The brei was centrifuged (12,000g, 20 min) and fil-
tered through Miracloth, with proteins fractionated by
adding finely ground (NH ) SO to obtain the desired con-
4
2
4
4
.3. Subcellular fractionation of pea leaves
centration (Table 2). After stirring for 30 min and centri-
fuging (13,000g, 20 min), the pellet was redissolved in
20 mL of 50 mM KPi, pH 8.0, 30% saturated with
(NH ) SO (Buffer A).
Mitochondria and chloroplasts were purified on Percoll
gradients (Cline, 1986; Douce et al., 1987). A fraction
enriched in cytosol plus vacuole contents, and vacuoles,
were prepared from protoplasts (Orsomando et al., 2005).
Protoplasts were purified on a three-step sucrose-sorbitol
gradient as described (Baldet et al., 1993; Orsomando
et al., 2005) except that the middle and bottom layer den-
sities were increased by adding 4% and 13% (v/v) Percoll,
respectively. Fractions were supplemented with bovine
albumin (1 mg/mL) and concentrated 6- to 13-fold before
PGH assays. Marker enzymes were extracted and assayed
as described (Orsomando et al., 2005).
4
2
4
For root proteins, the solution was applied to a
1.3 · 5.6 cm Octyl Sepharose 4 Fast Flow column equili-
brated with Buffer A. The column was washed with Buffer
A (2.5 mL/min) until the A280 of the effluent was <0.02.
PGH activity was eluted (2.0 mL/min) with an 83.5-mL lin-
ear gradient (100–0% Buffer A/0–100% KPi 50 mM, pH
8.0), collecting 2.5-mL fractions. Fractions with activity
were pooled and desalted on PD-10 columns equilibrated
with 50 mM KPi, pH 8.0. For leaf proteins, two
0.7 · 2.5 cm Octyl-Sepharose HiTrap columns were cou-
pled and replaced the Fast Flow column.
4
.4. Hydrolase assays
Active fractions from the Octyl Sepharose step were
loaded (0.5 mL/min) onto a Mono Q HR 5/5 column equil-
ibrated with 50 mM KPi, pH 8.0, and the column was
washed with this buffer until the A280 of the effluent fell
to zero. PGH activity was eluted (0.5 mL/min) with a 9-
mL linear gradient of 0–0.5 M KCl in 50 mM KPi, pH
8.0, collecting 0.3-mL fractions. Active fractions were
For pABAGlu hydrolase, standard reaction mixtures
(
5
final volume 10 lL) contained 8 lL of KPi buffer,
0 mM, pH 7.4 or 8.0, 0.1–16 lg protein), 1 mM MnCl ,
2
1
4
and either 10–14 nCi (182–259 pmol) of [ C]pABAGlu 5
or 42-166 nCi (70–280 pmol) of [ H]pABAGlu 5. Incuba-
3

36
G.G. Bozzo et al. / Phytochemistry 69 (2008) 29–37
pooled and concentrated to 0.5 mL in a Centricon-10 (Mil-
lipore, Billerica, MA, USA).
Bonnie Bartel and Rebekah Rampey for auxin conjugate
hydrolase cDNA expression constructs and for advice.
The concentrate from the Mono Q step was applied
(
0.25 mL/min) to a Superdex 200 HR 10/30 column equil-
ibrated with 50 mM KPi, pH 8.0 containing 50 mM KCl,
and eluted with the same buffer, collecting 0.4-mL frac-
tions. From roots, fractions making up the PGH activity
peak were pooled, concentrated to 0.28 mL, brought to
References
Albrecht, A.M., Boldizsar, E., Hutchison, D.J., 1978. Carboxypeptidase
displaying differential velocity in hydrolysis of methotrexate, 5-
methyltetrahydrofolic acid, and leucovorin. J. Bacteriol. 134, 506–513.
Baldet, P., Alban, C., Axiotis, S., Douce, R., 1993. Localization of free
and bound biotin in cells from green pea leaves. Arch. Biochem.
Biophys. 303, 67–73.
Bradford, M.M., 1976. A rapid and sensitive method for the quantitation
of microgram quantities of protein utilizing the principle of protein-
dye binding. Anal. Biochem. 72, 248–254.
1
0% (v/v) glycerol, frozen in liquid N , and stored at
2
ꢁ
80 ꢁC until use. To estimate native M , the Superdex col-
umn was calibrated with carbonic anhydrase (29 kDa),
bovine albumin (66 kDa), alcohol dehydrogenase
r
(
150 kDa), a-amylase (200 kDa), apoferritin (443 kDa),
and bovine thyroglobulin (669 kDa).
Carter, E.L., Jager, L., Gardner, L., Hall, C.C., Willis, S., Green, J.M.,
2
007. Escherichia coli abg genes enable uptake and cleavage of the
folate catabolite p-aminobenzoyl-glutamate. J. Bacteriol. 189, 3329–
334.
4
.6. Recombinant protein expression and purification
3
Chen, L., Chan, S.Y., Cossins, E.A., 1997. Distribution of folate
derivatives and enzymes for synthesis of 10-formyltetrahydrofolate in
cytosolic and mitochondrial fractions of pea leaves. Plant Physiol. 115,
Constructs of cDNAs of the Arabidopsis auxin-conju-
gate hydrolase family (ILL1, ILL2, ILL3, ILL6, ILR1,
and IAR3) in pGEX-KTO (Davies et al., 1999; LeClere
et al., 2002) were from B. Bartel and R. Rampey (Rice Uni-
versity, Houston, TX, USA). pGEX-KTO is an expression
vector in which the cloned protein is fused to the C-termi-
nus of glutathione S-transferase. The At4g17830 cDNA
2
99–309.
Cline, K., 1986. Import of proteins into chloroplasts. Membrane
integration of a thylakoid precursor protein reconstituted into chlo-
roplast lysates. J. Biol. Chem. 261, 14804–14810.
Cossins, E.A., 2000. The fascinating world of folate and one-carbon
metabolism. Can. J. Bot. 78, 691–708.
(
obtained from the Arabidopsis Biological Resource Cen-
Davies, R.T., Goetz, D.H., Lasswell, J., Anderson, M.N., Bartel, B., 1999.
IAR3 encodes an auxin conjugate hydrolase from Arabidopsis. Plant
Cell 11, 365–376.
ter, OH, USA) was modified by using splice overlap exten-
sion PCR to ablate an internal NcoI site, and the modified
At4g17830 cDNA was cloned between the NcoI and XhoI
sites of pET28b (Novagen), which adds a C-terminal hexa-
histidine tag. Constructs were introduced into E. coli BL21-
CodonPlus (DE3)-RIL cells (Stratagene), which were
grown at 37 ꢁC in LB medium until A600 reached 0.6. Tem-
perature was then dropped to 25 ꢁC and isopropyl-D-thio-
galactopyransoside was added (final concentration
D ´ı az de la Garza, R., Quinlivan, E.P., Klaus, S.M., Basset, G.J., Gregory
3
rd, J.F., Hanson, A.D., 2004. Folate biofortification in tomatoes by
engineering the pteridine branch of folate synthesis. Proc. Natl. Acad.
Sci. USA 101, 13720–13725.
Douce, R., Bourguignon, J., Brouquisse, R., Neuburger, M., 1987.
Isolation of plant mitochondria. General principles and criteria of
integrity. Methods Enzymol. 148, 403–415.
Gibeaut, D.M., Hulett, J., Cramer, G.R., Seemann, J.R., 1997. Maximal
biomass of Arabidopsis thaliana using a simple, low-maintenance
hydroponic method and favorable environmental conditions. Plant
Physiol. 115, 317–319.
Gregory 3rd, J.F., Toth, J.P., 1988. Chemical synthesis of deuterated
folate monoglutamate and in vivo assessment of urinary excretion of
deuterated folates in man. Anal. Biochem. 170, 94–104.
Hussein, M.J., Green, J.M., Nichols, B.P., 1998. Characterization of
mutations that allow p-aminobenzoyl-glutamate utilization by Esche-
richia coli. J. Bacteriol. 180, 6260–6268.
0
.1 mM). Incubation was continued for 3 h at the same
temperature. Subsequent procedures were at 0–4 ꢁC. Cells
from 50-mL cultures were pelleted, resuspended in 2 mL
of 50 mM Tris–HCl, pH 8.0, 0.1 mM MnCl , 0.01 mM
ZnCl , and broken in a Mini-BeadBeater (Biospec, Bar-
2
2
tlesville, OK, USA). Lysates were cleared by centrifugation
(
10,000g, 10 min), desalted on PD10 columns equilibrated
Jabrin, S., Ravanel, S., Gambonnet, B., Douce, R., R e´ beill e´ , F., 2003. One
carbon metabolism in plants. Regulation of tetrahydrofolate synthesis
during germination and seedling development. Plant Physiol. 131,
with 140 mM NaCl, 2.7 mM KCl, 10 mM Na HPO ,
2
4
1
.8 mM KH PO , pH 7.3. GST-fusion proteins were puri-
2 4
fied on GSH-agarose resin and the histidine-tagged protein
on Ni-NTA agarose resin (Qiagen, Valencia, CA, USA)
according to the manufacturers’ recommendations. For
extracts of cells expressing ILL2 or IAR3, hydrolysis of
indole acetic acid–alanine was measured as described
1
431–1439.
LeClere, S., Tellez, R., Rampey, R.A., Matsuda, S.P.T., Bartel, B., 2002.
Characterization of a family of IAA-amino acid conjugate hydrolases
from Arabidopsis. J. Biol. Chem. 277, 20446–20452.
McCullough, J.L., Chabner, B.A., Bertino, J.R., 1971. Purification and
properties of carboxypeptidase G1. J. Biol. Chem. 246, 7207–7213.
Minton, N.P., Atkinson, T., Sherwood, R.F., 1983. Molecular cloning of
the Pseudomonas carboxypeptidase G2 gene and its expression in
Escherichia coli and Pseudomonas putida. J. Bacteriol. 156, 1222–1227.
Oe, H., Kohashi, M., Iwai, K., 1983. Radioassay of the folate-hydrolyzing
enzyme activity, and the distribution of the enzyme in biological cells
and tissues. J. Nutr. Sci. Vitaminol. (Tokyo) 29, 523–531.
(
LeClere et al., 2002).
Acknowledgements
This work was supported in part by National Science
Foundation Grant No. MCB-0443709 and by an endow-
ment from the C.V. Griffin, Sr. Foundation. We thank
Orsomando, G., D ´ı az de la Garza, R., Green, B.J., Peng, M., Rea, P.A.,
Ryan, T.J., Gregory third ed., J.F., Hanson, A.D., 2005. Plant (c-
glutamyl hydrolases and folate polyglutamates: characterization,

G.G. Bozzo et al. / Phytochemistry 69 (2008) 29–37
37
compartmentation, and co-occurrence in vacuoles. J. Biol. Chem. 280,
8877–28884.
Orsomando, G., Bozzo, G.G., D ´ı az de la Garza, R., Basset, G.J.,
Quinlivan, E.P., Naponelli, V., R e´ beill e´ , F., Ravanel, S., Gregory 3rd,
J.F., Hanson, A.D., 2006. Evidence for folate–salvage reactions in
plants. Plant J. 46, 426–435.
Scott, J., R e´ beill e´ , F., Fletcher, J., 2000. Folic acid and folates: the
feasibility of nutritional enhancement in plant foods. J. Sci. Food
Agric. 80, 795–824.
Sherwood, R.F., Melton, R.G., Alwan, S.M., Hughes, P., 1985. Purifi-
cation and properties of carboxypeptidase G2 from Pseudomonas sp.
Strain RS-16. Eur. J. Biochem. 148, 447–453.
2
Prabhu, V., Chatson, K.B., Abrams, G.D., King, J., 1996. 13C Nuclear
magnetic resonance detection of interactions of serine hydroxym-
ethyltransferase with C1-tetrahydrofolate synthase and glycine
decarboxylase complex activities in Arabidopsis. Plant Physiol.
Suh, J.R., Herbig, A.K., Stover, P.J., 2001. New perspectives on folate
catabolism. Annu. Rev. Nutr. 21, 255–282.
Yamato, C., Kinoshita, K., 1978. Hydrolysis and metabolism of N-
benzoyl-L-tyrosyl-p-aminobenzoic acid in normal and pancreatic duct-
ligated animals. J. Pharmacol. Exp. Ther. 206, 468–474.
112, 207–216.