Evidence for an inhibitory LIM domain in a rat brain agmatinase-like protein

Article abstract of DOI:10.1016/j.abb.2011.05.003

We recently cloned a rat brain agmatinase-like protein (ALP) whose amino acid sequence greatly differs from other agmatinases and exhibits a LIM-like domain close to its carboxyl terminus. The protein was immunohistochemically detected in the hypothalamic region and hippocampal astrocytes and neurons. We now show that truncated species, lacking the LIM-type domain, retains the dimeric structure of the wild-type protein but exhibits a 10-fold increased k_cat, a 3-fold decreased K_m value for agmatine and altered intrinsic tryptophan fluorescent properties. As expected for a LIM protein, zinc was detected only in the wild-type ALP (～2 Zn²⁺/monomer). Our proposal is that the LIM domain functions as an autoinhibitory entity and that inhibition is reversed by interaction of the domain with some yet undefined brain protein.

Full text of DOI:10.1016/j.abb.2011.05.003

Archives of Biochemistry and Biophysics 512 (2011) 107–110
Contents lists available at ScienceDirect
Archives of Biochemistry and Biophysics
journal homepage: www.elsevier.com/locate/yabbi
Evidence for an inhibitory LIM domain in a rat brain agmatinase-like protein
Víctor Castro, Pablo Fuentealba, Adolfo Henríquez, Alejandro Vallejos, José Benítez, Marcela Lobos,
⇑
Beatriz Díaz, Nelson Carvajal, Elena Uribe
Departamento de Bioquímica y Biología Molecular, Facultad de Ciencias Biológicas, Universidad de Concepción, Casilla 160-C, Concepción, Chile
a r t i c l e i n f o
a b s t r a c t
Article history:
We recently cloned a rat brain agmatinase-like protein (ALP) whose amino acid sequence greatly differs
from other agmatinases and exhibits a LIM-like domain close to its carboxyl terminus. The protein was
immunohistochemically detected in the hypothalamic region and hippocampal astrocytes and neurons.
We now show that truncated species, lacking the LIM-type domain, retains the dimeric structure of
the wild-type protein but exhibits a 10-fold increased kcat, a 3-fold decreased Km value for agmatine
and altered intrinsic tryptophan fluorescent properties. As expected for a LIM protein, zinc was detected
only in the wild-type ALP (ꢀ2 Zn2+/monomer). Our proposal is that the LIM domain functions as an
autoinhibitory entity and that inhibition is reversed by interaction of the domain with some yet
undefined brain protein.
Received 14 April 2011
and in revised form 6 May 2011
Available online 12 May 2011
Keywords:
Agmatinase
Polyamines
LIM domain
Ó 2011 Elsevier Inc. All rights reserved.
Introduction
contrast with these reports, we have recently cloned and expressed
a rat brain protein which is significantly active as agmatinase
Agmatine, a decarboxylated derivative of
L
-arginine, has been
in vitro, although its amino acid sequence is only about 12% iden-
tical with the human and bacterial enzymes [13]. In any case, cer-
tain degree of structural homology was revealed by its reaction
with a polyclonal antibody raised against Escherichia coli agmatin-
ase. We refer to this protein as agmatinase-like protein (ALP).
Like E. coli agmatinase, ALP was shown to require Mn2+ for cat-
alytic activity and to be maximally active at pH 9–9.5 [13]. By
RT-PCR and immmunohistochemical methods, the protein was de-
tected in the hypothalamus in glial cells and arcuate nucleus neu-
rons, and also in hippocampus astrocytes and neurons, but not in
brain cortex [14]. Considering that, in general, this localization
coincides with that described for its substrate agmatine, ALP was
suggested to be involved in the regulation of intracellular concen-
trations of this neurotransmitter/neuromodulator.
One especially interesting finding was a terminal segment that
includes a LIM-type domain in the sequence of ALP [13]. The LIM do-
main is a zinc-coordinating domain, consisting of two tandemly re-
peated zinc fingers, generally involved in cellular differentiation and
control of cellular growth, gene expression, interactions with cyto-
skeleton, auto-inhibitory effects and possibly as biosensors that
mediate communication between the cellular and nuclear compart-
ments [15,16]. The classic LIM consensus sequence includes a
CX2CX16–23HX2CX2CX2CX16–21CX2 (CH/D) sequence, associated with
a highly variable sequence in the remainder of the domain that con-
fers functional specificity [17].
associated to several important biological processes in mammals,
including neurotransmitter, anticonvulsant, antineurotoxic and
antidepressant actions in the brain [1]. This primary amine is
packed into synaptic vesicles in the brain and spinal cord and acts
on transmembrane receptors (a2-adrenergic, imidazoline and
gultamatergic NMDA receptors) [2]. At physiological concentra-
tions, agmatine is inhibitory to the neuronal nitric oxide synthase
[3] and participate in modulation of insulin release from pancreatic
cells [4,5] and renal sodium excretion [6]. A fine control of agma-
tine levels is evidently required, thus justifying the interest in
the enzymological aspects of agmatine synthesis and degradation
in mammals.
Agmatine result from decarboxylation of arginine by arginine
decarboxylase [7,8] and it is converted to putrescine and urea by
agmatinase [9]. With regard to agmatinase, there are reports estab-
lishing its presence in the brain [9,10], although the study of their
enzymic properties has been complicated by the lack of appropri-
ately active preparations of the enzyme. This includes two reported
recombinant forms of human agmatinase which are very poorly
active under in vitro conditions [11,12]. In one of these reports,
the functionality of the cloned human gene was deduced from a
functional complementation test with
a yeast strain which
contains a disruption in the gene encoding ornithine decarboxylase
and thus requires exogenous polyamines for growth [11]. In
As a part of a continuing effort aimed to characterize the molec-
ular and functional aspects of ALP, we have now turned our atten-
tion to the LIM-type domain, by examining the consequences of its
⇑
Corresponding author. Fax: +56 41 2239687.
E-mail address: auribe@udec.cl (E. Uribe).
0003-9861/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.abb.2011.05.003

108
V. Castro et al. / Archives of Biochemistry and Biophysics 512 (2011) 107–110
removal from the ALP sequence. Our suggestion is that the domain
acts as an autoinhibitory entity and that inhibition is reversed by
interaction of the domain with some yet undefined brain protein.
Molecular weights determinations
Molecular weights were determined by gel filtration on a cali-
brated column of Sephadex G-200. The column was equilibrated
and eluted with a buffer solution containing 10 mM Tris–HCl (pH
7.5) and 50 mM KCl.
Methods
Materials
Fluorescence spectra
All reagents were of the highest quality commercially available
(most from Sigma, St. Louis, MO, USA) and were used without fur-
ther purification. Restriction enzymes, as well as enzymes and
reagents for the polymerase chain reaction (PCR) were obtained
from Invitrogen Co. (Carlsbad, CA, USA). The plasmid pE-PHO was
generously donated by Dr. Juan Olate (Universidad de Concepción,
Chile) and the Saccharomyces cerevisiae strain TRY104Dspe1, was
generously provided by Dr. Herbert Tabor (National Institutes of
Health, Bethesda, USA).
Fluorescence measurements were made at 25 °C on a Shimadzu
RF-5301 spectrofluorimeter. The protein concentration was
40–50 lg/ml and emission spectra were measured with the excita-
tion wavelength at 295 nm. The slit width for both excitation and
emission was 1.5 nm, and spectra were corrected by subtracting
the spectrum of the buffer solution in the absence of protein. The
buffer solution (pH 7.5) contained 5 mM Tris–HCl and 2 mM
MnCl2. In the fluorescence quenching experiments, the acrylamide
concentrations varied from 0 to 100 mM.
Enzyme and protein assays
Zinc contents
Routinely, agmatinase activities were determined by measuring
the formation of urea from 80 mM agmatine in 50 mM glycine-
NaOH (pH 9.0). All the assays were initiated by adding the enzyme
to the substrate buffer solution previously equilibrated at 37 °C.
Initial velocity and inhibition studies were performed in duplicate
and repeated three times. The inhibitory patterns were initially
determined by double reciprocal plots and replots of intercepts
versus inhibitor concentrations. Data were then fitted to the appro-
priate equations by using nonlinear regression with GraphPad
Prism version 5.0 for Widows (GraphPad Software Inc., San Diego,
The zinc contents of wild-type and truncated preparations of
ALP were determined by atomic absorption on a Perkin Elmer
1100 atomic absorption spectrometer equipped with a graphite
furnace and a deuterium arc background corrector. Recovery was
nearly 100%. For analysis, the purified preparations were incubated
with 10 mM EDTA in 10 mM Tris–HCl (pH 7.5), followed by dialysis
with two changes of 10 mM Tris–HCl (pH 7.5).
Yeast strains transfection
CA, USA). Urea was determined by a colorimetric method with a-
With the use of appropriate restriction sites and standard pro-
cedures for subcloning, the wild-type and truncated species of
ALP were inserted into the EcoRI site in the yeast expression plas-
mid pE-PHO, which contains the ura3 selectable marker and gene
expression is induced in a medium low in phosphate. A yeast strain
containing a disruption of the spe1 gene encoding ODC (strain
isonitrosopropiophenone [18] and protein by means of the stan-
dard Bio-Rad protein assay (Bio-Rad, CA, USA) with bovine serum
albumin as standard.
LIM domain deletion and expression of ALP
TRY104Dspe1), was used for complementation studies essentially
as described [19]. The specific yeast strains designated in results
were transformed with pE-PHO-ALP or pE-PHO-truncated ALP by
the lithium acetate procedure, and transformants were selected
on medium lacking uracil. The transfected yeast were grown for
48 h at 30 °C to deplete the endogenous pool of polyamines in
low phosphate YMM medium with amino acid supplement and
The gene of ALP lacking the sequence corresponding to the LIM-
domain was amplified with the high-fidelity Pfx polymerase, from
vector H6PQE60-29, two primers used were 50 cac ggt gcc cat ggt
gac acc cag gcc 30 and 50 gcc gga agc ttt cac tac cac agg agg agc
aca g 30 which included specific restriction sites (NcoI and Hind
III respectively). The PCR fragment was then directionally cloned
into the histidine-tagged pQE60 E. coli expression vector, and the
histidine-tagged enzyme was expressed in E. coli strain JM109, fol-
kanamycin 25
added to 50 mL of YMM low phosphate, supplemented with amino
acid, kanamycin 25 g/mL and 10 g/mL agmatine. As a control,
lg/mL under agitation. 1 mL of this inoculum was
l
l
lowing induction with 0.5 mM isopropyl-b-D-thiogalactopyrano-
non transfected strain and the minimal medium containing agma-
tine and uracil was used. Grow was estimated by nephelometry at
600 nm. When values were higher than 0.5, the sample was diluted
by 10-fold.
side. The bacterial cells were disrupted by sonication on ice
(5 ꢁ 30 s pulses) in a buffer solution containing 100 mM Tris–HCl
(pH 7.5), 2 mM MnCl2, 2 mM putrescine, 2 mM DTT, 0,1 mM PMSF
and 100 mM KCl. After centrifugation for 20 min at 96,000g, the
supernatant was applied to a Ni2+-NTA column equilibrated with
20 mM Tris–HCl (pH 7.9), containing 5 mM imidazole and
500 mM NaCl. Proteins non specifically bound to the resin were
removed by washings with a buffer solution containing 20 mM
Tris–HCl (pH 7.9), 12 mM imidazole and 500 mM NaCl, until the
absorbance of the eluates at 280 nm dropped to 0.01. This was
followed by elution with a solution containing 20 mM Tris–HCl
(pH 7.9), 100 mM imidazole and 500 mM NaCl to obtain ALP. The
purity of the enzyme was assessed by SDS–PAGE (12%); only one
band, with an apparent molecular mass of about 60 kDa was de-
tected after staining with Coomassie blue R-250. The theoretical
molecular weights of the recombinant wild-type and truncated
proteins were calculated using Compute pI/Mw from ExPASy
Proteomic Server (http://www.expasy.ch/tools).
Results and discussion
To gain some insight into a possible role for the LIM-like domain
in agmatine hydrolysis by the rat brain agmatinase-like protein
(ALP), the structural and functional consequences accompanying
its removal were examined.
Truncated species were catalytically active and retained the
ability to discriminate between agmatine and its precursor
arginine; both the wild-type and mutant variants were totally inac-
tive on -arginine. Interestingly, truncated species exhibited a
L-
L
10-fold increased kcat value, a 3-fold decreased Km for agmatine
and increased affinity for the product putrescine (Table 1). As pre-
viously described for E. coli agmatinase [20,21], product inhibition

V. Castro et al. / Archives of Biochemistry and Biophysics 512 (2011) 107–110
109
Table 1
direct measure of substrate affinity. However, considering identical
binding areas for putrescine and the nonguanidino portion of the
substrate, a higher affinity for agmatine may be envisioned for
the truncated variant. The kcat values for the wild-type and trun-
cated variants of ALP are considerably lower than the 120 10 sꢂ1
previously reported for E. coli agmatinase; however, the Km value
for agmatine and the Ki value for putrescine inhibition are in order
of those for the bacterial enzyme [20,21].
Kinetic parameters of wild-type and truncated species of ALP.
kcat (sꢂ1
)
Km (mM)
kcat/Km (Mꢂ1sꢂ1
)
KPi ut (mM)
Variant
wt
Truncated
0.9 0.2
3.0 0.2
1.2 0.5
300
9091
5.0 0.4
1.3 0.2
10
1
Values, derived from three experiments in duplicate, represents the means SD. Put
refers to the product putrescine.
The functionality of the truncated species was also examined by
the functional reconstitution of the polyamine biosynthetic path-
way in S. cerevisiae strain TRY104Dspe1. This yeast strain lacks
ornithine decarboxylase, the first and rate-limiting step of poly-
amine biosynthesis and it is, therefore, strictly dependent of an
exogenous source of polyamines for growth [19]. Genes for the
wild-type and truncated variants of ALP were transfected to the
yeast strain and, as shown in Fig. 1, both kind of transformed yeast
cells were able to grow in the presence of agmatine, whereas addi-
tion of spermidine was required in the case of the nontransformed
cells. The requirement for ALP to be transcribed by the cell was
indicated by the inability of its substrate agmatine to restore the
normal growth of the nontransformed cells.
The altered kinetic parameters for the truncated variant of ALP
are indicative of a conformational alteration accompanying dele-
tion of the LIM-like sequence. To probe any tertiary structure alter-
ation, the tryptophan exposure was investigated by fluorescence
studies, including quenching by acrylamide [22]. The uncharged
quencher acrylamide was used because of its potential ability to
penetrate into the interior of the protein and, thus, to quench par-
tially buried tryptophan residues [23]. As shown in Fig. 2, together
with a small blue shift of the maximum emission wavelength, from
339 to 336 nm, there was an important change from linear to
hyperbolic Stern–Volmer plots. The hyperbolic plot indicates the
presence of at least two classes of tryptophan residues with differ-
ent accessibilities to acrylamide in the truncated variant of ALP,
thus revealing a significant conformational difference between
the variants. The deduced conformational change would explain
the increased catalytic efficiency of the truncated variant. What-
ever the exact nature and extension of the change, it had no conse-
quences on the dimeric structure of the protein. From gel filtration,
calculated molecular weights were 117 and 90 kDa for wild-type
and truncated species, respectively. The difference is explainable
by the absence of the last 66 amino acids, which include the LIM
domain, in the truncated variant. The expected theoretical values
for the dimeric wild-type and truncated variants were about 118
and 104 kDa. The divergence between the experimental and theo-
retical data for the mutant would be another evidence for a con-
formationally altered state.
Fig. 1. Time course of growth of yeast strains TRY104Dspe1 deficient in L-ornithine
decarboxylase, transfected with wild-type ALP (h) or truncated species lacking the
LIM domain (d). The control (s) indicates the growth of the deficient, non
transfected yeast cells. In all cases, the yeast cells were grown in a minimal medium
containing agmatine (10 lg/ml).
by putrescine was linear competitive. Thus, upon truncation, the
catalytic efficiency (kcat/Km value), of ALP was increased by about
30 times. The significantly higher kcat value indicates that agmatine
is more efficiently positioned for optimal hydrolysis in the trun-
cated variant. Considering the information which is available for
arginase and E. coli agmatinase, the most extensively kinetically
characterized agmatinase [20,21], agmatine hydrolysis by ALP is
expected to involve a nucleophilic attack by a manganese-bound
hydroxide on the guanidine carbon of agmatine. With available
information, the Km value for agmatine cannot be equated to the
thermodynamic dissociation constant of the enzyme–substrate
complex and, therefore, 1/Km cannot be unequivocally taken as a
Fig. 2. Intrinsic tryptophan fluorescence and acrylamide quenching of wild-type ALP (s) and truncated species lacking the LIM domain (d). Quenching was analyzed by
Stern–Volmer plots; F0 and F indicates the areas under the curves in the absence and presence of acrylamide, respectively. The experiment was repeated for three times with
equivalent results.

110
V. Castro et al. / Archives of Biochemistry and Biophysics 512 (2011) 107–110
Mn2+ is essential for catalysis by ALP [13], and the results obtained
in the presence of 5 mM Mn2+, the inhibition may be explained by
replacement of the metallic cofactor by Zn2+ in the active site of
ALP. The proposed LIM domain in the structure of ALP is schemat-
ically shown in Fig. 4.
Further studies are evidently required to unequivocally define
the exact role of the LIM domain in the biological function of
ALP. However, a working hypothesis may be advanced on the basis
of previously demonstrated participation of this kind of domains in
protein–protein interactions [16,17]. Based on the increased cata-
lytic efficiency of truncated species, our suggestion is that the
LIM domain acts as an autoinhibitory entity and that inhibition is
reversed by interaction of the domain with some yet undefined
brain protein. A mechanism of this kind was demonstrated for a
LIM-kinase; similarly to what observed here, removal of the LIM
domain also resulted in activation of this enzyme [24].
Acknowledgments
Fig. 3. Efect of Zn2+on the catalytic activity of the the truncated variants of ALP. The
effect was examined both in the absence (s) and presence (d) of 5 mM Mn2+
.
This research was supported by Grant Fondecyt 11070069.
Special acknowledgments to Drs. Herbert Tabor and Juan Olate
Values are from one representative experiment repeated for three times.
for providing us the S. cerevisiae strain TRY104Dspe1 and pE-PHO
vector, respectively.
References
[1] A. Halaris, J. Plietz, CNS Drugs 21 (2007) 885–900.
[2] D. Reis, S. Regunathan, Trends Pharmacol. Sci. 21 (2000) 187–193.
[3] D.R. Demady, S. Jianmongkol, J.L. Vuletich, Mol. Pharmacol. 59 (2001) 24–29.
[4] S. Chan, Gen. Pharmacol. 31 (1998) 525–529.
[5] C. Su, I. Liu, H. Chung, J. Cheng, Neurosci. Lett. 457 (2009) 125–128.
[6] S. Penner, D. Smyth, Pharmacology 53 (1996) 160–169.
[7] A. Iyo, M. Zhu, G. Ordway, S. Regunathan, J. Neurochem. 96 (2006) 1042–1050.
[8] M. Zhu, A. Iyo, J. Piletz, S. Regunathan, Biochim. Biophys. Acta 1670 (2004)
156–164.
[9] G.M. Gilad, Y. Wollam, A. Iaina, J.M. Rabey, T. Chernihovsky, V.H. Gilad,
Neuroreport 29 (1996) 1730–1732.
[10] M. Sastre, S. Regunathan, E. Galea, D. Reis, J. Neurochem. 67 (1996) 1761–1765.
[11] S. Mistry, T. Burwell, R. Chambers, L. Rudolph-Owen, F. Spaltmann, W. Cook,
S.M. Morris, Am. J. Physiol. Gastrointest. Liver Physiol. 282 (2002) 375–381.
[12] R. Iyer, H. Kim, R. Tsoa, W. Grody, S. Cederbaum, Mol. Genet. Metab. 75 (2002)
209–218.
[13] E. Uribe, M. Salas, S. Enriquez, M.S. Orellana, N. Carvajal, Arch. Biochem.
Biophys. 461 (2007) 146–150.
[14] C. Mella, F. Martínez, M.A. García, F. Nualart, V. Castro, P. Bustos, N. Carvajal, E.
Uribe, Histochem. Cell Biol. 134 (2010) 137–144.
Fig. 4. Proposed schematic representation of a LIM domain in ALP. Postulated
ligands to the zinc ions are encircled.
[15] J. Kadrmas, M. Beckerle, Nat. Rev. Mol. Cell Biol. 5 (2004) 920–931.
[16] A.D. Maturana, N. Nakakawa, N. Yoshimoto, K. Tatematsu, K. Hoshijima, S.
Kuroda, Cell Signal 23 (2011) 928–934.
[17] Q. Zheng, Y. Zhao, Biol. Cell 99 (2007) 489–502.
[18] R.M. Archibald, J. Biol. Chem. 157 (1945) 507–518.
[19] R.D. Klein, T.G. Geary, A.S. Gibson, M.A. Favreau, C.A. Winterrowd, S.J. Upton,
J.S. Keithly, G. Zhu, R.L. Malmberg, M.P. Martinez, N. Yarlett, Microbiology 145
(1999) 301–307.
[20] N. Carvajal, V. López, M. Salas, E. Uribe, P. Herrera, J. Cerpa, Biochem. Biophys.
Res. Commun. 258 (1999) 808–811.
[21] M. Salas, V. Lopez, E. Uribe, N. Carvajal, J. Inorg. Biochem. 98 (2004) 1032–
1036.
[22] S. Pawar, V. Deshpande, Eur. J. Biochem. 267 (2000) 6331–6338.
[23] M. Ansari, S. Zubair, S. Atif, M. Kashif, N. Khan, M. Rehan, T. Anwar, A. Iqbal, M.
Owais, Protein Peptide Lett. 17 (2010) 11–17.
By definition, a LIM domain corresponds to a specific zinc-
binding protein domain, which comprises two tandemly repeated
zinc fingers [15]. Concordant with this, atomic absorption metal
analysis revealed the presence of 1.9 0.1 Zn2+/subunit of wild-
type species of ALP and the absence of zinc ions in the truncated
variant. Thus, in addition to fulfilling the general distinguishing
features of a LIM domain containing protein [15], our results indi-
cates that Zn2+ is not required for agmatine hydrolysis by ALP.
Moreover, zinc ions were inhibitory to both the wild-type and
truncated variants. The results in Fig. 3, specifically corresponding
to the truncated variant, were not significantly different to those
obtained in the case of the wild-type species. Considering that
[24] K. Nagata, K. Ohashi, N. Yang, K. Mizuno, Biochem. J. 343 (1999) 99–105.