Involvement of glutamine-238 in the substrate specificity of human laeverin/aminopeptidase Q

Article abstract of DOI:10.1248/bpb.34.24

Human laeverin/aminopeptidase Q (APQ) is a novel member of the M1 family of zinc aminopeptidases and is specifically expressed on the cell surface of extravillous trophoblasts. In this study, we examined the significance of Gln-238 of laeverin/APQ, a putative S1 site residue, by site-directed mutagenesis for its enzymatic activity and substrate specificity. Replacement of Gln-238 with Ala caused a significant change in substrate specificity rather than a decrease in enzymatic activity. These results indicate that Gln-238 is important for the substrate specificity of laeverin/APQ. In addition, our data suggest that direct electrostatic interaction between substrate and S1 site of the enzyme is not involved in the mutant enzyme's preference for basic amino acids.

Full text of DOI:10.1248/bpb.34.24

24
Regular Article
Biol. Pharm. Bull. 34(1) 24—27 (2011)
Vol. 34, No. 1
Involvement of Glutamine-238 in the Substrate Specificity of Human
Laeverin/Aminopeptidase Q
a
a
c
,a,b
Yoshikuni GOTO,a,b Rina YOSHIOKA, Naomi ARISAKA, Akira HATTORI, and Masafumi TSUJIMOTO
*
a Laboratory of Cellular Biochemistry, RIKEN; 2–1 Hirosawa, Wako, Saitama 351–0198, Japan: b Faculty of
Pharmaceutical Sciences, Teikyo-Heisei University; 4–1 Uruido-minami, Ichihara, Chiba 290–0193, Japan: and
c Department of System Chemotherapy and Molecular Sciences, Graduate School of Pharmaceutical Sciences, Kyoto
University; Kyoto 606–8501, Japan.
Received September 29, 2010; accepted October 18, 2010; published online October 20, 2010
Human laeverin/aminopeptidase Q (APQ) is a novel member of the M1 family of zinc aminopeptidases and
is specifically expressed on the cell surface of extravillous trophoblasts. In this study, we examined the signifi-
cance of Gln-238 of laeverin/APQ, a putative S1 site residue, by site-directed mutagenesis for its enzymatic activ-
ity and substrate specificity. Replacement of Gln-238 with Ala caused a significant change in substrate specificity
rather than a decrease in enzymatic activity. These results indicate that Gln-238 is important for the substrate
specificity of laeverin/APQ. In addition, our data suggest that direct electrostatic interaction between substrate
and S1 site of the enzyme is not involved in the mutant enzyme’s preference for basic amino acids.
Key words aminopeptidase; laeverin; endoplasmic reticulum aminopeptidase 1; site-directed mutagenesis; substrate specificity
Aminopeptidases hydrolyze the N-terminal amino acid of human laeverin/APQ with Ala. We found that Q238A lae-
proteins or peptide substrates. Among them, the M1 family verin/APQ retained substantial enzymatic activities with
of zinc aminopeptidases (gluzincins) shares the consensus marked change of its substrate specificity, indicating the sig-
GAMEN and HEXXH(X)18E motifs essential for enzymatic nificance of this residue for the enzymatic characteristics of
activity. This family, which consists of 11 enzymes in hu- laeverin/APQ. Data presented in this study imply that direct
mans,1—3) plays important roles in several pathophysiological electrostatic interaction is not involved in the mutant en-
processes, such as angiogenesis, cell cycle regulation, repro- zyme’s preference for basic amino acids, which was sug-
duction, memory retention, blood pressure control, and anti- gested in the previous work.17)
gen presentation to major histocompatibility complex (MHC)
class I molecules.4—12)
MATERIALS AND METHODS
Laeverin was originally identified as a cell-surface protein
specifically expressed on human extravillous trophoblasts.13)
Molecular Modeling of Laeverin/APQ The published
cDNA cloning of human laeverin revealed that it contains X-ray crystallographic structure of Thermoplasma acido-
both consensus motifs and thus is a novel member of the M1 philum Tricorn-interacting factor F3 (TIFF3) (Protein Data
family. Another group also predicted the existence of the Bank code 1Z5H)18) was used as a template for modeling the
‘leaverin’ gene in the human genome through a genomic catalytic site of ERAP1 with the three-dimensional Jigsaw
search and named it aminopeptidase Q (APQ).14) Recently, Protein Server (http://www.bmm.icnet.uk/˜3djigsaw/). The
we established a large-scale production system of the enzyme structure was displayed using the CueMol program (Ishitani,
and characterized its enzymatic properties in detail.15) We R., CueMol: Molecular Visualization Framework, http://cue-
found that the enzyme degraded several placenta-derived mol.sourceforge.jp).
peptide hormones, such as angiotensin III, endokinin C, and
Site-Directed Mutagenesis To obtain a large amount of
kisspeptin-10. Furthermore because its exopeptidase motif recombinant human laeverin/APQ, we slightly modified its
GAMEN is uniquely composed of the HAMEN sequence, cDNA by polymerase chain reaction (PCR). Briefly, to ex-
we examined the enzymatic significance of His-379 of press the enzyme as a soluble protein, the coding sequences
human laeverin/APQ by site-directed mutagenesis and found for the cytosolic and transmembrane region of the enzyme
that this residue contributes to the unique properties of the (Met1-Gln64) were replaced with that for human trypsin II
enzyme.16) Considering the susceptibility of these peptides signal peptide (i.e. MNLLLILTFVAAVAA).15) Hexahistidine
and their specific expression in the placenta, we speculated tag was added at its C-terminal end. Amplified cDNA was
that laeverin/APQ plays important roles in the maintenance cloned into the BssHII–XhoI site of the baculovirus transfer
of normal pregnancy in humans.
vector pFastbac-1 (Invitrogen, CA, U.S.A.). The cDNA en-
In our previous work, we identified a residue affecting en- coding mutant Q238A laeverin/APQ was generated using a
zymatic activity and the substrate specificity of endoplasmic QuickChange site-directed mutagenesis kit (Stratagene, La
reticulum aminopeptidase (ERAP)1, a member of the M1 Jolla, CA, U.S.A.). The primers used to replace Gln-238 with
aminopeptidase family.17) In particular, replacement of Gln- Ala in human laeverin/APQ cDNA were as follows: sense
181 with Asp caused a marked change in the substrate speci- primer, 5ꢀ-GCCCTGTTAGCGTCCGCTCTGGAACCAAC-
ficity of ERAP1. On the other hand, replacement with Ala ATTTG-3ꢀ, and antisense primer 5ꢀ-CAAATGTTGGTTCC-
led to almost complete loss of enzymatic activity. To further AGAGCGGACGCTAACAGGGC-3ꢀ.
analyze the residue affecting enzymatic activity and substrate
The sequences of the products were confirmed by auto-
specificity of the M1 family of aminopeptidases, we exam- mated sequencing on an Applied Biosystems model 3730.
ined in the present study the effect of replacing Gln-238 of
Expression and Purification of Recombinant Wild-
© 2011 Pharmaceutical Society of Japan
∗ To whom correspondence should be addressed. e-mail: tsujimot@thu.ac.jp

January 2011
25
Type and Mutant Laeverin/APQs in the Baculovirus Sys- RESULTS AND DISCUSSION
tem The pFastbac-1 vectors containing laeverin/APQ
cDNAs were transformed to competent DH10bac Es-
Molecular Modeling of the Catalytic Pocket of Lae-
cherichia coli cells harboring the baculovirus genome verin/APQ In the previous work, we showed that Gln-181
(bacmid) and a transposition helper vector (Invitrogen). Sub- of ERAP1 was critical for the enzymatic activity and sub-
sequently, insect Sf 9 cells were transfected with recombinant strate specificity of ERAP1.17) Alignment of ERAP1 and
bacmids using the Cellfectin reagent (Invitrogen). After a 3-d laeverin/APQ indicates that Gln-238 of laeverin/APQ occu-
incubation period, recombinant baculoviruses were isolated pies the corresponding site of Gln-181 of ERAP1. Using
and used to infect Sf 9 cells at a multiplicity of infection of TIFF3 as a template, we modeled the structure of the cat-
0.1. Three days after infection, the amplified viruses were alytic pocket of human laeverin/APQ (Fig. 1). As in the case
harvested.
of ERAP1, Gln-238 of laeverin/APQ is expected to be lo-
For the production of laeverin/APQ, Sf 9 cells were grown cated near the catalytic HEXXH(X)18E motif and at a possi-
at 27 °C in 100 ml Sf-900III medium (Gibco, CA, U.S.A.) ble S1 site of the enzyme, suggesting the importance of this
and 1.5ꢁ106 cells/ml were infected at a multiplicity of infec- residue for enzymatic activity. To elucidate the role of Gln-
tion of 1—3. After 3 d, the culture medium was collected 238 of human laeverin/APQ in its enzymatic properties fur-
after centrifugation at 5000 g for 15 min.
ther, this residue was substituted with Ala to generate Q238A
The culture medium was loaded onto a hydroxyapatite laeverin/APQ, and its enzymatic property was compared
(Nacalai Tesque, Kyoto, Japan) column (bed volume: 10 ml) with wild-type enzyme. We chose an Ala mutant to check the
pre-equilibrated with 5 mM phosphate buffer (pH 7.5). After significance of Gln-238 residue, because the substitution of
extensive washing with the same buffer, laeverin/APQ was Gln-181 residue with Ala caused almost complete loss of the
eluted from the column with 100 mM phosphate buffer (pH enzymatic activity of ERAP1, thus indicating the essential
7.5). The eluate was then loaded onto a Ni2ꢂ-chelating role of this residue in the enzymatic activity.17)
Sepharose (Pharmacia, Uppsala, Sweden) column (bed vol-
Characterization of the Hydrolytic Activities of Wild-
ume: 1 ml) pre-equilibrated with 10 mM phosphate buffer (pH Type and Q238A Laeverin/APQs Figure 2 shows the en-
7.5) containing 0.1 M NaCl. After extensive washing with zymatic activities of wild-type and Q238A laeverin/APQs
10 mM phosphate buffer (pH 7.5) containing 0.1 M NaCl and toward various substrates. As shown previously, the wild-
5 mM imidazole, laeverin/APQ was eluted with 10 mM phos- type enzyme cleaved Leu-MCA most efficiently, followed by
phate buffer (pH 7.5) containing 0.1 M NaCl and 150 mM imi- Met-, Arg- and Lys-MCA. Hydrolysis of Ala-, Phe-, and ben-
dazole. The laeverin/APQ-containing fractions were exten- zyl-Cys-MCA was also observed. On the other hand, in sharp
sively dialyzed against 25 mM Tris–HCl buffer (pH 7.5) con- contrast to Q181A ERAP1, which showed little hydrolytic
taining 0.125 M NaCl, concentrated with an ultrafiltration activity toward substrates tested, Q238A laeverin/APQ re-
membrane and stored at ꢃ20 °C prior to use.
tained substantial activity toward basic amino acid substrates,
Measurement of Aminopeptidase Activity of Lae- Arg and Lys-MCA. Little activity was detected toward other
verin/APQs Aminopeptidase activities of wild-type and substrates tested. These results indicate that while wild-type
mutant laeverin/APQs were routinely determined by end- enzyme shows broad substrate specificity, enzymatic activity
point assay with various fluorogenic aminoacyl-4-metheyl- of the mutant enzyme is rather restricted to basic amino
coumaryl-7-amides (aminoacyl-MCAs) as substrates. The re- acids.
action mixture containing 100 mM aminoacyl-MCA and
Table 1 shows the kinetic parameters of wild-type and
1 mg/ml enzyme in 0.5 ml of 25 mM Tris/HCl buffer (pH 7.5) Q238A laeverin/APQs using Leu-MCA and Arg-MCA as
was incubated at 37 °C for 15 min. The reaction was termi- substrates. The calculated kcat/Km value from a Lineweaver–
nated by adding 2.5 ml of 0.1 M sodium acetate buffer (pH Burk plot of wild-type enzyme for the most efficient sub-
4.3) containing 0.1 M sodium monochloroacetate. The strate Leu-MCA was 451ꢄ16 mMꢃ1 sꢃ1. The catalytic effi-
amount of 7-amino-4-methylcoumarin released was meas- ciency (kcat/Km) of the wild-type enzyme for Arg was
ured by spectrofluorophotometry (F-2000; Hitachi) at an ex- 205ꢄ33 mMꢃ1 sꢃ1, about 2.3-fold less than for Leu-MCA. On
citation wavelength of 360 nm and an emission wavelength the other hand, when comparing the hydrolytic activity of
of 460 nm, and is given in arbitrary units.
mutant enzyme with that of the wild-type toward Arg-MCA,
To determine the kinetic parameters, the reaction mixture mutant enzyme showed comparable efficiency (232ꢄ49
containing various concentrations of substrates and either mMꢃ1 sꢃ1). Both Km and kcat values of mutant laeverin/APQ
wild-type or mutant laeverin/APQ in 500 ml of 25 mM were about 2-fold less than those of wild-type enzyme, sug-
Tris/HCl buffer (pH 7.5) was incubated at 37 °C for 5 min. gesting that replacement of Gln-238 with Ala slightly
The amount of 7-amino-4-methylcoumarin released was then changed the substrate recognition mechanism toward basic
measured. The kinetic parameters (Km and Vmax) were calcu- amino acids. It showed little activity toward Leu-MCA with
lated using Lineweaver–Burk plots.
kcat/Km value of 2.79ꢄ0.12 mMꢃ1 sꢃ1, because of a marked
Materials Asp-, benzyl-Cys-, and Glu-MCAs were pur- decrease in the kcat value. These results indicate that mutant
chased from BACHEM AG (Bubendorf, Switzerland). Ala-, enzyme retained substantial enzymatic activity toward Arg-
Arg-, Leu-, Lys- Met, and Phe-MCAs were obtained from the MCA and its substrate specificity was different from that of
Peptide Institute (Osaka, Japan). Bestatin was purchased the wild-type enzyme and showed rather a restricted pattern.
from Sigma.
Taken together, we concluded that the substitution of Gln-
238 with Ala caused a change of substrate specificity rather
than total decrease of the enzymatic activity toward sub-
strates, which is quite different from the case of ERAP1.17)

26
Vol. 34, No. 1
Fig. 1. Molecular Modeling of the Catalytic Pocket of Laeverin/APQ
(A) Modeling of the catalytic site of human laeverin/APQ using Thermoplasma acidophilum TIFF3 as a template. His-379 and Glu-382 in the exopeptidase HAMEN motif and
His-415, Glu-416, His-419 and Glu-438 in HEXXH(X)18E motif and Tyr-503 are shown as residues in the catalytic pocket of laeverin/APQ. (B) Alignment of the laeverin/APQ
amino acid sequence with the sequences of the M1 family of aminopeptidases. Residues corresponding to Gln-238 of laeverin/APQ are shown as a white Q on a black background.
Residues conserved among the enzymes are shaded.
Table 1. Kinetic Parameters of Wild-Type and Q238A Laeverin/APQs to-
ward Leu-MCA and Arg-MCA
kcat/Kmꢁ103
a)
Substrate
Km (mM)a)
kcat (sꢃ1
)
a)
(mMꢃ1·sꢃ1
)
Leu-MCA
Arg-MCA
Wild-type
Q238A
Wild-type 27.9ꢄ7.82
Q238A 11.5ꢄ4.33
119ꢄ21.9
83.9ꢄ6.70
53.5ꢄ8.06
0.23ꢄ0.01
5.56ꢄ0.81
2.53ꢄ0.52
451ꢄ15.7
2.79ꢄ0.12
205ꢄ32.9
232ꢄ49.4
a) Values are meanꢄS.D. (nꢅ3).
Effect of Bestatin on the Enzymatic Activities of Wild-
Type and Q238A Laeverin/APQs Because laeverin/APQ
is unique in that bestatin inhibits its enzymatic activity more
efficiently than amastatin,15) we next compared the effect of
bestatin on the enzymatic activity of the wild-type and
Q238A laeverin/APQs. As shown in Fig. 3, bestatin inhibited
the enzymatic activity of the wild-type enzyme toward Leu-
MCA comparably with that of toward Arg-MCA. On the
other hand, it inhibited the mutant enzyme about 50-fold less
efficiently than the wild-type enzyme when measured using
Arg-MCA as a substrate. Considering that bestatin (N-
[(2S,3R)-3-amino-2-hydoroxy-4-phenylbutyryl]-L-leucine
hydorochloride) is a small peptide having an aromatic amino
acid-like residue at its N-terminus, consistent with the notion
that wild-type laeverin/APQ binds to aromatic amino acids
with higher affinity than the Q238A mutant enzyme.
Fig. 2. Enzymatic Activities of Wild-Type and Q238A Laeverin/APQs to-
ward Various Substrates
Purified enzymes (1 mg/ml) were incubated with various amino acid MCA substrates
(100 mM) at 37 °C for 15 min. Relative activities of the enzymes are shown, taking the
hydrolytic activity of wild-type enzyme toward Leu-MCA (A) and Q238A mutant en-
zyme toward Arg-MCA (B) as 100%. Each bar shows the meanꢄS.D. (nꢅ3).

January 2011
27
static interaction and thus contribute to the unique properties
of the enzymes. To elucidate the mode of the interaction
between S1 site and substrates definitively, it is essential to
determine the crystal structure of the enzymes.
In summary, through substitution of Gln-238 with Ala of
human laeverin/APQ, we have shown that this residue is im-
portant for its substrate specificity; however, as direct elec-
trostatic interaction is not likely to be involved in the interac-
tion between N-terminal amino acid of the substrate and S1
site of the enzyme, it is conceivable that another acidic
residue might participate in their interaction.
Fig. 3. Effect of Bestatin on the Enzymatic Activity of Wild-Type and
Q238A Laeverin/APQs
REFERENCES
Purified wild-type and Q238A laeverin/APQs were incubated with various concen-
trations of bestatin on ice for 5 min. Aminopeptidase activities were then measured
using Leu-MCA (100 mM) (ꢀ) or Arg-MCA (100 mM) (ꢁ) as the wild-type and Arg-
MCA (100 mM) (ꢂ) as Q238A laeverin/APQ.
1) Taylor A., FASEB J., 7, 290—298 (1993).
2) Hooper N. M., FEBS Lett., 354, 1—6 (1994).
3) Tsujimoto M., Hattori A., Biochim. Biophys. Acta, 1751, 9—18
(2005).
4) Sato Y., Biol. Pharm. Bull., 27, 772—776 (2004).
5) Constam D. B., Tobler A. R., Rensing-Ehl A., Kemler I., Hersh L. B.,
Fontana A., J. Biol. Chem., 270, 26931—26939 (1995).
6) Osada T., Watanabe G., Kondo S., Toyoda M., Sakaki Y., Takeuchi T.,
Mol. Endocrinol., 15, 960—971 (2001).
7) Albiston A. L., McDowall S. G., Matsacos D., Sim P., Clune E.,
Mustafa T., Lee J., Mendelsohn F. A. O., Simpson R. J., Connolly L.
M., Chai S. Y., J. Biol. Chem., 276, 48623—48626 (2001).
8) Goto Y., Hattori A., Ishii Y., Tsujimoto M., FEBS Lett., 580, 1833—
1838 (2006).
9) Wright J. W., Mizutani S., Murray C. E., Amir H. Z., Harding J. W., J.
Hypertens., 8, 969—974 (1990).
10) Reaux A., Fournie-Zaluski M. C., David C., Zini S., Roques B., Cor-
vol P., Llorens-Cortes C., Proc. Natl. Acad. Sci. U.S.A., 96, 13415—
13420 (1999).
In this study, we identified Gln-238 as a residue crucial in
the substrate specificity of human laeverin/APQ. Replace-
ment of this residue with Ala caused the enzyme preference
for basic amino acids and rather restricted substrate speci-
ficity. In our previous work, we reported that replacement of
Gln-181 of human ERAP1 with Asp increased the enzyme
preference for basic amino acid.17) We speculated that re-
placement of Gln-181 with acidic amino acids might cause a
local conformational change of the substrate pocket, result-
ing in rather restricted substrate specificity through electro-
static interaction between the residue and N-terminal basic
amino acid of the substrate. However, replacement of Gln-
238 of human laeverin/APQ with neutral amino acid Ala
caused the enzyme preference for basic amino acids, sug-
gesting that the direct electrostatic interaction between Ala-
238 and N-terminal basic amino acid of the substrates may
have little role in the basic amino acid preference of the en-
zyme. Instead, we speculate at present that both Gln-238 of
wild-type and Ala-238 of Q238A mutant laeverin/APQs
maintain their catalytic pocket structures by interacting or in-
terfering with another unidentified residues, and thus form
unique S1 sites, as discussed elsewhere.17) It is possible that
unique S1 site structure of the mutant enzyme but not that of
the wild-type may allow the electrostatic interaction between
neighboring acidic amino acid and basic amino acid sub-
strate, causing the basic amino acid preference.
11) Serwold T., Gonzalez F., Kim J., Jacob R., Shastri N., Nature (Lon-
don), 419, 480—483 (2002).
12) Saric T., Chang S.-C., Hattori A., York I. A., Markant S., Rock K. L.,
Tsujimoto M., Goldberg A. L., Nat. Immunol., 3, 1169—1176 (2002).
13) Fujiwara H., Higuchi T., Yamada S., Hirano T., Sato Y., Nishioka Y.,
Yoshioka S., Tatsumi K., Ueda M., Maeda M., Fujii S., Biochem. Bio-
phys. Res. Commun., 313, 962—968 (2004).
14) Puente X. S., Sanchez L. M., Overall C. M., Lopez-Otin C., Nat. Rev.
Genet., 4, 544—558 (2003).
15) Maruyama M., Hattori A., Goto Y., Ueda M., Maeda M., Fujiwara H.,
Tsujimoto M., J. Biol. Chem., 82, 20088—20096 (2007).
16) Maruyama M., Arisaka N., Goto Y., Ohsawa Y., Inoue H., Fujiwara H.,
Hattori A., Tsujimoto M., J. Biol. Chem., 284, 34692—34702 (2009).
17) Goto Y., Tanji H., Hattori A., Tsujimoto M., Biochem. J., 416, 109—
116 (2009).
18) Kyrieleis O. J. P., Goettig P., Kiefersauer R., Huber R., Brandstetter H.,
J. Mol. Biol., 349, 787—800 (2005).
19) Goto Y., Hattori A., Mizutani S., Tsujimoto M., J. Biol. Chem., 282,
37074—37081 (2007).
The corresponding site of the M1 family of aminopepti-
dases is occupied either Gln or Asp (Fig. 1B). When the site
is occupied by Asp, this site may interact with either Ca2ꢂ
(APA)19) or substrates (L-RAP/ERAP2)20) through electro-
20) Tanioka T., Hattori A., Masuda S., Nomura Y., Nakayama H., Mizu-
tani S., Tsujimoto M., J. Biol. Chem., 278, 32275—32283 (2003).