These mutated recombin-ant proteins behave similarly to the wild-type enzyme regarding enzymatic activity, secondary structure, calcium sensitivity and immunoreactivity.. However, immuno
Trang 1membrane association, secondary structure and substrate specificity
Vitor Oliveira1, Paula A G Garrido2, Claudia C Rodrigues2, Alison Colquhoun2,
Leandro M Castro2, Paulo C Almeida3, Claudio S Shida3, Maria A Juliano4, Luiz Juliano4,
Antonio C M Camargo5, Stephen Hyslop6, James L Roberts7, Valerie Grum-Tokars8,
Marc J Glucksman8and Emer S Ferro2
1 Laborato´rio de Neurocieˆncias, Universidade da Cidade de Sa˜o Paulo, Brazil
2 Departamento de Biologia Celular e Desenvolvimento, Programa de Biologia Celular, Instituto de Cieˆncias Biome´dicas, Universidade de Sa˜o Paulo, Brazil
3 Universidade de Mogi das Cruzes, Mogi das Cruzes, Brazil
4 Departamento de Biofı´sica, Universidade Federal de Sa˜o Paulo, Brazil
5 Centro de Toxinologia Aplicada (CAT), Instituto Butantan, Brazil
6 Departamento de Farmacologia, Faculdade de Cieˆncias Me´dicas, Universidade Estadual de Campinas, Campinas, Brazil
7 Sam and Ann Barshop Institute for Longevity and Aging Studies, University of Texas Health Science Center, San Antonio, TX, USA
8 Midwest Proteome Center, Department of Biochemistry & Molecular Biology, Rosalind Franklin University of Medicine and Science ⁄ Chicago Medical School, USA
Keywords
calcium; membrane binding; peptide
metabolism; protease; thimet oligopeptidase
Correspondence
E S Ferro, Laborato´rio de Comunicac¸a˜o
Celular, Avenida Prof Lineu Prestes1524
Sala 431, Sa˜o Paulo, 05508-900, SP, Brazil
E-mail: eferro@usp.br
Note
V Oliveira and P.A.G Garrido contributed
equally to this work.
(Received 9 February 2005, revised 24
March 2005, accepted 31 March 2005)
doi:10.1111/j.1742-4658.2005.04692.x
The metalloendopeptidase 24.15 (EP24.15) is ubiquitously present in the extracellular environment as a secreted protein Outside the cell, this enzyme degrades several neuropeptides containing from 5 to 17 amino acids (e.g gonadotropin releasing hormone, bradykinin, opioids and neuro-tensin) The constitutive secretion of EP24.15 from glioma C6 cells was demonstrated to be stimulated linearly by reduced concentrations of extra-cellular calcium In the present report we demonstrate that extraextra-cellular calcium concentration has no effect on the total amount of the extracellular (cell associated + medium) enzyme Indeed, immuno-cytochemical analyses
by confocal and electron microscopy suggested that the absence of calcium favors the enzyme shedding from the plasma membrane into the medium Two putative calcium-binding sites on EP24.15 (D93 and D159) were altered by site-directed mutagenesis to investigate their possible contribu-tion to binding of the enzyme at the cell surface These mutated recombin-ant proteins behave similarly to the wild-type enzyme regarding enzymatic activity, secondary structure, calcium sensitivity and immunoreactivity However, immunocytochemical analyses by confocal microscopy consis-tently show a reduced ability of the D93A mutant to associate with the plasma membrane of glioma C6 cells when compared with the wild-type enzyme These data and the model of the enzyme’s structure as determined
by X-ray diffraction suggest that D93 is located at the enzyme surface and
is consistent with membrane association of EP24.15 Moreover, calcium was also observed to induce a major change in the EP24.15 cleavage site
on distinctive fluorogenic substrates These data suggest that calcium may
be an important modulator of ep24.15 cell function
Abbreviations
EP24.15; metalloendopeptidase 24.15; G6PD, glucose-6-phosphate dehydrogenase; PFA, paraformaldehyde.
Trang 2The metalloendopeptidase EC 3.4.24.15 (EP24.15)
rep-resents a distinctive peptide-metabolizing enzyme with
size-selectivity for peptides ranging from 8 to 17 amino
acids [1–3] EP24.15 is a monomeric, soluble, 77-kDa
endopeptidase, thiol-activated, with an isoelectric point
of 5.6 and a pH optimum of activity of 7.4, first
isola-ted from the soluble fraction of the rat brain [4] This
enzyme is distributed in all mammalian tissues so far
examined, with high levels in testis, brain, kidney and
pituitary [5] In the brain, at the electron microscopic
level, EP24.15-like immunoreactivity is observed over
the nuclei, cytoplasm, cross-sectioned dendrites,
myeli-nated and unmyelimyeli-nated axons, axon terminals, and in
both astrocytes and oligodendrocytes [6] Recent
stud-ies have shown that EP24.15 could interact with
peptides generated by the multicatalytic complex
pro-teasome [7,8] Thus, the high concentration of the
enzyme observed inside the cells could be somewhat
related to the degradation of catabolic products
gener-ated by the proteasome [7,9] EP24.15 has been mainly
implicated in extracellular metabolism of
neuropep-tides [5] Supporting an extracellular function of
EP24.15, cell fractionation studies have shown that
EP24.15 is present in a minor form in the particulate
subcellular fractions from the central nervous tissue
[10] and AtT-20 cells [11] Moreover, we previously
reported that EP24.15 is secreted from distinctive cells
lines, such as the rat glioma C6 [12] and mouse AtT20
anterior pituitary [13] Studies in the neuroendocrine
cell line AtT-20 have shown that EP24.15 was enriched
in the regulated secretory pathway [14] and could be
either constitutively released in the extracellular space
or upon stimulation with corticotrophin releasing
hor-mone as well as calcium ionophore A23187 [13]
How-ever, EP24.15 secretion from AtT20 cells is partially
blocked by brefeldin A or nocodazole, suggesting that
this enzyme could be secreted by a pathway different
from that described for other secreted proteins or
neu-ropeptides, such as b-endorphin [13] EP24.15 secretion
from glioma C6 cells is even more unusual, as it is
sti-mulated upon incubation of the cells in the absence of
calcium [12]
We have previously shown that EP24.15 is released
from glioma C6 cells with low calcium concentrations
[12] Using these cells as a model we now show that
extracellular EP24.15 activity (enzyme associated to
cells and present in the medium) was not affected by
the presence or absence of calcium In fact, we have
shown that previously membrane-associated EP24.15 is
shed into the medium in the absence of calcium
Site-directed mutagenesis on a putative calcium binding
motif revealed a role for D93 in EP24.15–membrane
association The location of this residue is on the
surface, and it can bind calcium as corroborated by molecular modeling by the EP24.15 derived X-ray dif-fraction data Moreover, calcium was also observed to affect EP24.15 cleavage site on specific fluorogenic peptides
Results
The rate of shedding of EP24.15 from C6 glioma cells
is inversely proportional to the calcium concentration present in the medium (Fig 1A) To analyze if this increment in EP24.15 activity into the medium was due to secretion, the following experiments were con-ducted: soluble EP24.15 activity was determined in the medium and subtracted from the total activity observed in the presence of cells, the latter correspond-ing to cells + medium activity (Fig 1B) Total extra-cellular EP24.15 activity (cells + media) remains the same with respect to the presence or absence of cal-cium (Fig 1B) However, there is a higher EP24.15 activity in the medium of cells incubated in calcium-free conditions compared with those incubated in calcium-containing medium (Fig 1B) The viability of glioma C6 cells was determined by the Trypan blue dye exclusion method, and suggested that more than 99.5 ± 5% of cells were intact In the same experi-ments, the activity of the cytosolic glucose-6-phosphate dehydrogenase (G6PD), an indicator of cell death, in the medium was < 0.01 ± 0.01 mU per 106 cells (n¼ 5), from a total activity in the whole cell homogenates
of 2.9 ± 0.3 mU per 106 cells (n¼ 5), further suggest-ing that cells were intact after incubation in calcium-free medium
The rabbit anti-EP24.15 serum recognized a single protein band of approximately 78 kDa in the crude homogenate of glioma C6 cells (Fig 2A), suggesting the specificity against this enzyme [15] Importantly, the rabbit anti-EP24.15 antiserum used here has been previously characterized not to show cross-reactivity with the related endopeptidase EP24.16 [6] EP24.15 immunostaining revealed the presence of a diffuse labeling throughout the whole cell that includes an intense nuclear labeling (Fig 2B–E) The intensity of this immunolabeling differed from cell to cell indica-ting a variable expression of EP24.15 in these cells No major differences in the specific immunostaining could
be observed after preincubating the cells in the pres-ence (Fig 2B) or abspres-ence (Fig 2C) of calcium The presence of Triton X100 (0.1%) during the immuno-cytochemical procedures (Fig 2B and C) clearly increases the nuclear labeling for EP24.15, compared
to similar experiments conducted in the absence of the detergent (Fig 2D and E) These data suggest that cell
Trang 3permeabilization with detergent is an important issue
regarding the nuclear staining for EP24.15 However,
detergent and calcium removal was not sufficient to
avoid intracellular EP24.15 staining (Fig 2D and E)
Moreover, in control experiments, neither somatic nor
nuclear labeling was observed when the anti-EP24.15
antiserum was preadsorbed with the recombinant EP24.15 (100 lgÆmL)1; Fig 2C)
On the other hand, immunocytochemistry performed with postfixation of the cells in the absence of Triton X100, produces a clear cell-surface immunostaining for EP24.15, with different intensities among the cells (Fig 3A) Under these latter conditions, previous incu-bation of the cells in calcium-free medium strongly affects the intensity of EP24.15 extracellular immuno-staining (Fig 3B) These data suggest the shedding of EP24.15 from the plasma membrane after calcium removal Immunocytochemistry for electron micro-scopy further confirmed the presence of EP24.15 in the extracellular face of the plasma membrane of glioma C6 cells (Fig 4)
We then searched for putative calcium-binding motifs on the surface of EP24.15 that could contribute
to the enzyme’s membrane association At least two possible calcium-binding motifs containing an aspartic acid were identified on these structural analyses (D93 and D159), which were mutated to alanine by site-directed mutagenesis producing the mutants D93A, D159A and the double mutant, D93⁄ 159A The corres-ponding recombinant proteins were expressed in bac-teria and purified to homogeneity (Fig 5A) Western blot analyses show that such point mutations have no effect on antiserum recognition (Fig 5B) The follow-ing experiments were performed in an attempt to corre-late a possible effect of the above point mutations with the calcium-dependent EP24.15 membrane association The homogeneously purified EP24.15 recombinant pro-teins (wild-type or mutant) were incubated with glioma C6 cells and the putative membrane binding was ana-lyzed by confocal microscopy Incubation of either wild-type EP24.15 or the D159A mutant with the glioma C6 cells produced a clear increment in the immunoreactivity of EP24.15 associated with the extra-cellular surface of these cells (Fig 6A and B) when compared with control performed without adding any
of the recombinant EP24.15 (Fig 6E) On the other hand, incubation of the glioma C6 cells with either D93A or D93⁄ 159A mutated proteins was not able to produce a similar increment in the extracellular immu-noreactivity related to EP24.15 (Fig 6C and D) Con-trol experiments under these conditions show that without the addition of EP24.15 the basal immunoreac-tivity was greatly reduced (Fig 6E) Taken together, these data suggest the D93 residue of EP24.15 within the putative calcium-binding motif 89SPNKD93 is a possible mediator of enzyme–membrane association
To examine a structure–function correlate of the D93–membrane association and the location of the resi-due within the three-dimensional realm of the enzyme,
Fig 1 Calcium-dependent EP24.15 released into the medium.
EP24.15 activity in intact cells was determined using
b-lipotro-pin(61–69) (YGGFMTSEK; 100 l M ) as substrate b-lipotropin(61–69)
fragments were separated by reverse phase HPLC and the
EP24.15 activity was based on the generation of [Met5]enkephalin
calculated by comparing the peak areas of each of those products
with the corresponding peak area of the same synthetic peptides
of known concentrations The EP24.15 enzymatic activity released
into the medium was determined fluorimetrically using the
substrate Abz-GGFLRRV-EDDnp To discern peptidolytic activity
exclusively due to EP24.15, the inhibitors
N-(1-(R,S)-carboxyl-2-phenylethyl)-AAF-p-amino-benzoate (CFP-AAF-pAB) and ⁄ or the
dipeptide Pro–Ile were used All enzymatic determinations were
conducted under linear conditions where product formation was
directly proportional to enzyme concentration, and obeys zero-order
kinetic parameters with < 10% of total substrate consumed during
the course of the assay Results are means of three independent
determinations ± SE (A) EP24.15 release into the glioma C6 cells’
medium is inversely proportional to the extracellular calcium
con-centration (B) Total extracellular EP24.15 activity is not affected by
calcium, however, the amount of enzyme that is released in the
medium is largely increased after incubating the glioma C6 cells in
medium that lacks calcium.
Trang 4a hypothetical intermolecular interaction using calcium
as a ligand was performed (Fig 7) The calcium is
modeled in green and is hydrogen bonded to two
amino acids The first amino acid is D93, the object of
study in this work Using an ensemble of coordinates
of well studied Asp–calcium interactions, the carboxy
groups of the Asp are within hydrogen bonding
dis-tance (2.5–3 A) for calcium The coordinates for Fig 7
were obtained from the crystal structure of thimet
oligopeptidase deposited at the Research Collaboratory
for Structural Bioinformatics (Protein Data Bank ID
#1S4B) The figure was drawn with the computer modeling program spock and then subsequently ren-dered using raster 3d [16] To ensure that electro-statics were properly accounted for, calculations were performed and slightly negative potentials were obtained for the D93
Furthermore, we investigated the calcium effects on EP24.15 secondary structure, enzymatic activity and substrate specificity Enzymatic activity of the mutated
Fig 2 EP24.15 immunoreactivity in glioma C6 cells EP24.15 presence in glioma C6 cells was further confirmed by immunochemical reac-tions (A) The antiserum raised against the rat testis recombinant EP24.15 recognized in the whole homogenate of glioma C6 cells a single protein band corresponding to the expected Mrof EP24.15 For details see Experimental procedures (B–E) Prior to each immunocytochem-istry experiment the DMEM culture medium was removed and cells were rinsed three times with fresh NaCl ⁄ P i , with (B, D) or without (C, E) 1.2 m M of calcium, and incubated for 30 min in 4% formaldehyde, pH 7.4 After three cycles of NaCl ⁄ P i washing (15 min each) cells were incubated for 2 h in the presence of rabbit anti-EP24.15 serum (1 : 4000) diluted in NaCl ⁄ P i containing 3% normal goat serum, 1% BSA, in the presence (B, C) or absence (D, E) of 0.1% of Triton X100 Labeled cells were examined under a Zeiss laser confocal microscope (CLSM 410) equipped with an Axiovert 100 inverted microscope and an Argon ⁄ Krypton laser Cy3-tagged molecules were excited at a wave-length of 568 nm Images were acquired sequentially as single transcellular optical sections and averaged over 32 scans per frame They were then processed using the Carl Zeiss CLSM software (version 3.1) and stored on Jazz disks for further retrieval and editing Final com-posites were prepared using Adobe PHOTOSHOP without modifying the spectral characteristics of the original signal Control experiments were performed as described; no specific cell labeling could be observed (data not shown) The data presented are representative of three experi-ments performed independently and exhibiting similar results.
Trang 5proteins either in the presence or absence of calcium
was not different from the wild-type EP24.15
(Fig 5C) The superimposable CD spectra indicates
that there are no gross perturbations in expression of the mutant proteins with respect to folding CD spec-tra of wild-type EP24.15 and mutants without
Fig 4 Extracellular EP24.15 immunogold labeling observed by electron microscopy Cells incubated in the presence of calcium were processed for electron microscopy using a pre-embedding procedure as des-cribed (A) Lower magnification of an entire glioma C6 cell observed in the electron microscope and immunolabeled at the extracellular surface for EP24.15 (arrows) (B) Higher magnification of the cell shown
in (A) immunolabeled for EP24.15 (arrows) Scale bar: A, 140 nm; B, 500 nm.
Fig 3 The effects of calcium on the extracellular EP24.15 immunoreactivity in glioma C6 cells (A, B) To observe exclusively the extracellular EP24.15 immunoreactivity in glioma C6 cells, cells were incubated for 15 min at 4 C in the presence of rabbit anti-EP24.15 serum (1 : 500), diluted in NaCl ⁄ P i containing 3% normal goat serum and 1% BSA, prior to fixation with 4% formaldehyde pH 7.4 Labeled cells were exam-ined under a Zeiss laser confocal microscope (CLSM 410) equipped with an Axiovert 100 inverted microscope and an Argon ⁄ Krypton laser Cy3-tagged molecules were excited at a wavelength of 568 nm Images were acquired sequentially as single transcellular optical sections and averaged over 32 scans per frame They were then processed using the Carl Zeiss CLSM software (version 3.1) and stored on Jazz disks for further retrieval and editing Final composites were prepared using Adobe PHOTOSHOP without modifying the spectral characteristics of the original signal Immunocytochemistry reactions were conducted after incubation of the glioma C6 cells in the presence (A) or absence (B) of calcium for 1 h Note that cells previously incubated in medium lacking calcium (B) show a reduced extracellular labeling for EP24.15 compared to cells incubated in the presence of calcium (A) The intensity of the immunostaining varies from cell to cell but is consistently stronger in cells previously incubated in the presence of calcium compared to those incubated in the absence of calcium (1) Confocal regu-lar image; (2) nomarski; (3) glow scale The immunolabeling intensity is shown in the graphic (intensity grows from green to red) No immu-nolabeling is observed after preabsorbing the anti-EP24.15 serum with recombinant EP24.15 (100 lgÆmL)1; data not shown).
Trang 6smoothing and curve fitting shows a predominance of
a-helical structures (Table 1; Fig 8A–D), which is
similar to the estimated consensus secondary structure
prediction obtained from different algorithms (http://
npsa-pbil.ibcp.fr), and more importantly, the structure
of the protein solved by X-ray crystallography [17]
Far UV-CD analysis indicated that the small
differ-ences observed among the wild-type and mutated
EP24.15 falls within the experimental error determined
using three different preparations of the enzyme (data
not shown) The inter-experimental errors are 4%
However, calcium (2.2 mm) addition reduced the
a-helical content with a concomitant increase in the
b-strand, turn and unordered structure of EP24.15
(Table 1); similar results were observed in the near UV-CD spectra, collected either in the presence or in the absence of calcium (data not shown) These data clearly show that calcium affects EP24.15 secondary structure The D93A mutation alone or in combina-tion with D159A that prevents EP24.15 membrane association seems insufficient to completely prevent the calcium changes induced in the enzyme secondary structure (Table 1), suggesting that structural changes may not correlate entirely with the enzyme membrane association
With regards to substrate cleavage specificity, the gradual increase in the calcium concentration from
0 to 50 mm raised the preference of the wild-type EP24.15 for the cleavage at the R–R bond in the ortho-aminobenzoic acid (Abz)-GGFLRRVQ-EDDnp substrate from 64% (0 mm Ca+2), to 79% (50 mm
Ca+2) with a corresponding decrease in the cleavage at the L–R peptide bond from 36% (0 mm Ca+2) to 21% (50 mm Ca+2; Table 2) This variation is even more pronounced with the Abz-GGFLRRDQ-EDDnp sub-strate, which yielded cleavages of 63% and 37% at the L–R and R–R bonds, respectively, in the absence of
Ca2+ In the presence of 50 mm of Ca2+the preferred hydrolyzed peptide bond changes to 26% and 74% for the cleavage at the L–R and R–R bonds, respectively (Table 2) Moreover, after a 2 mm calcium incre-ment, the kcat⁄ Km (lm)1Æs)1) ratio of both Abz-GGFLRRVQ-EDDnp and Abz-GGFLRRDQ-EDDnp started to change (Table 2) These data suggest that in the presence of high calcium concentrations, as found
in specific cellular microenvironments [18], EP24.15 substrate specificity could be changed Mutations of EP24.15 either at D93 or D159 have no major effects
on the ratio of Abz-GGFLRRDQ-EDDnp peptide bond cleavage after calcium addition (data not shown)
Discussion
The most important finding of this report is that cal-cium regulates the membrane association, secondary structure and substrate specificity of the endo-oligo-peptidase EP24.15 Site-directed mutagenesis indicated that residue D93 plays an important role in this cal-cium-dependent membrane association of EP24.15 in glioma C6 cells (Fig 6) This finding is consistent with the position of the residue on the surface of the enzyme as modeled on the X-ray structure (Fig 7), although we have been unable to demonstrate the direct interaction of calcium with the enzyme More-over, biochemical analyses also indicated that calcium affects EP24.15 secondary structure (Fig 8) and
A
4
- Ca2+
+ Ca2+
3
2
1
0
EP24.15
78kDa
78kDa
B
C
Fig 5 SDS ⁄ PAGE and western blot of EP24.15 recombinant
pro-teins (A) SDS ⁄ PAGE stained with Comassie blue shows the
homo-geneity of the recombinant proteins (5 lgÆlane)1) EP24.15 wild-type
(WT), D93A (93), D159A (159) and D93 ⁄ 159A (93 ⁄ 159) after the
affinity purification as described (B) Western blotting of the
recom-binant proteins (0.5 lgÆlane)1) EP24.15 wild type (WT), D93A (93),
D159A (159) and D93 ⁄ 159A (93 ⁄ 159) using the anti-EP24.15
serum Note that none of the mutations affected the specificity of
the antiserum for the EP24.15 (C) Comparative enzymatic activity
of EP24.15 WT and mutants D93A, D159A and D93 ⁄ 159A
meas-ured with the fluorogenic substrate QFS in the absence and in the
presence of 2 m M CaCl2 Note that calcium slightly increases
EP24.15 enzymatic activity in these assays.
Trang 7Fig 6 Immunocytochemical staining of glioma C6 cells after incubation with the recombinant EP24.15 proteins Glioma C6 cells were prein-cubated at 4 C in the presence (A–D) or absence (E) of 2 lg of either EP24.15 wild-type (A), D159A (B), D93A (C) or D93 ⁄ 159A (D), for
30 min, in DMEM containing BSA 0.05% The medium was then removed and cells were rinsed three times with DMEM and incubated at
4 C for 30 min in the presence of the anti-EP24.15 serum (1 : 500) The excess of primary antiserum was removed rinsing the cells three times in 1 mL of NaCl ⁄ P i containing 5% BSA, before fixing the cells in PFA 4% The secondary anti-(rabbit Cy3) Ig (1 : 250; Sigma) was used to develop the immunoreaction Note that the intensity of the EP24.15-related immunostaining is reduced for either the D93A or D93 ⁄ 159A proteins compared to the wild-type and D159A.
Fig 7 Molecular modeling of the surface EP24.15 D93 residue complexed with a calcium ion On the left a diagram of the side chain of D93 and a bound calcium (gray sphere) mapped onto the surface of the recently solved EP24.15 structure ([40] PDB ID #1S4B) On the right
is a magnified view focusing on the D93 side chain and interaction with calcium The EP24.15 model with the calcium ion was energy mini-mized using Molecular Operating Environment software, and the figure was generated using SPOCK
Trang 8changes the enzyme substrate specificity (Table 2).
These data suggest that calcium is an important
medi-ator of EP24.15 biological function
It has been previously shown that in the brain
EP24.15 is present both in neurons and glial cells [5,6]
Glial cells are at least 10 times more abundant than
neurons in the brain [19] and during the past years
they have been recognized to be more than a
connect-ive tissue responsible for neuronal support [20] Glial
cells are now identified as excitable cells capable of modulating neuronal stimulus [19] As for neurons, glial cells can also produce growth factors, pro-hor-mone processing enzymes [21], receptors [22], neuro-transmitters [19] and neuropeptides [23] However, in contrast to neurons, which process peptides within the secretory pathway, glial cells secrete immature peptide precursors [24,25] Glial post-translational processing
of opioid peptides precursor (e.g proenkephalin) is therefore thought to occur in the extracellular environ-ment, as demonstrated by the presence of protease inhibitors in the cultured medium of striatal astrocytes that prevented Met-enkephalin generation from secre-ted proenkephalin [25–27] Therefore, the present results suggest that in physiological conditions EP24.15 would be able to associate with the extracellu-lar surface of glial cells and could participate in the metabolism of a specific set of peptides released by these cells and⁄ or from neurons Thus, further charac-terization of peptide-processing enzymes such as EP24.15 in the extracellular milieu of glial cells is of neurobiological significance
Table 1 Calcium effects on the secondary structure of EP24.15.
TBS
TBS + 2.2 m M Ca 2+
Fig 8 Far UV-CD spectra of the EP24.15 WT (A), D93A (B), D159A (C) and D93 ⁄ 159A (D) mutants both in the absence (n) or in the pres-ence of CaCl 2 2 m M(h) Note that calcium addition reduces the a-helical content of all EP24.15s analyzed (Table 1).
Trang 9In a previous study we reported that EP24.15
secre-tion from glioma C6 cells was increased linearly by
reduced extracellular calcium concentrations [12]
These data were paradoxical to many other secretory
systems in which extracellular calcium is required for
secretion [28]; therefore, the mechanism behind this
calcium effect on EP24.15 secretion was investigated
further It is known that extracellular calcium is not
required for constitutive protein secretion [29] and
reduces the release of parathyroid hormone [28]
Bio-chemical observations shown here indicated that
extracellular calcium concentrations could not change
the overall rate of peptide metabolism from intact
gli-oma C6 cells in culture (Fig 1) These data suggest
that the overall extracellular EP24.15 enzymatic
activ-ity was not affected by altered calcium
concentra-tions Consequently, we conclude that calcium
removal was not really affecting EP24.15 secretion
from glioma C6 cells Another possibility was that
low extracellular calcium concentration compromised
the EP24.15 association with the E face of the plasma
membrane The presence of EP24.15 associated with
the E face of the plasma membrane in glioma C6
cells was confirmed by immunocytochemistry
experi-ments both at the light and electron microscopy
levels Indeed, immunocytochemistry analyses shown
here suggested that EP24.15 extracellular
immuno-reactivity is deeply affected when extracellular calcium
is removed from glioma C6 cells Thus, it seems clear
that a deficiency in extracellular calcium affects the
shedding of EP24.15 from the plasma membrane of
glioma C6 cells into the medium, suggesting that
EP24.15–membrane association within cells can be
dynamically modulated While further investigation
would be necessary, this mechanism could be relevant for the intracellular traffic and secretion of EP24.15 [13,30], since it could dynamically locate the enzyme
in specific cell membranes, such as the endoplasmic reticulum [6], Golgi apparatus, secretory vesicles [14] and plasma membrane [11]
Interestingly, EP24.15 contains neither a leader pep-tide sequence nor any other hydrophobic domains to mediate its interaction with biological membranes [31] and the mechanism responsible for the enzyme mem-brane association in neuroendocrine cells remains unknown [6,11] The EP24.15 crystal structure was recently solved [17] showing that the enzyme surface contains 80% of all charged residues Among these amino acid residues D93 and D159 were investigated
as possible mediators of EP24.15–membrane associ-ation The D93 amino acid residue is located within the EP24.15 89SPNKD93 sequence, which recalls the calcium-binding motif DXSXS previously described [32] On the other hand, D159 residue is the EP24.15 corresponding metal divalent-binding site residue iden-tified in the homologue endopeptidase EP24.16; both D93 and D159 residues are located at the external sur-face of the enzyme [17,33] The D93A or D93⁄ 159A double mutations were shown to affect the calcium-dependent association of EP24.15 to the plasma membrane of glioma C6 cells, while the D159A point mutation had no such effect The molecular modeling
of the interaction of calcium and D93 on the surface
of the enzyme has the correct electrostatic potential Additionally, the calcium is far enough from the active site zinc that no direct interaction can be postulated for direct effects on catalysis Changes appear to be transduced through changes in secondary structure Circular dichroism experiments reported here clearly shows that calcium affects EP24.15 secondary struc-ture, reducing the a-helical content by 3–8% It is well known that calcium affects the conformational struc-ture of several proteins [34–37], and that could provide
a possible mechanistic explanation for the data shown above The D93A or D93⁄ 159A mutations that strongly affected EP24.15 membrane association were not sufficient to prevent the secondary structure altera-tions of EP24.15 induced by calcium Hence, it is plausible to suggest that the structural changes induced
by calcium on EP24.15 secondary structure are not correlated with the membrane association Therefore, the exact role of D93 in EP24.15 membrane associ-ation remains unknown A possible explanassoci-ation that needs further attention is that D93 mediates the inter-action of EP24.15 with a yet unidentified membrane protein Molecular modeling corroborates that the resi-due is on the surface
Table 2 Influence of calcium concentration on EP24.15 peptide
bond cleavage.
[Ca 2+ ]
(m M )
Cleaved bond (%)
Abz-GGFLRRVQ-EDDnp Abz-GGFLRRDQ-EDDnp
L–Rb R–Rb
k cat ⁄ K ma
(l M )1.s)1) L–Rb
R–Rb
k cat ⁄ K ma
(l M )1.s)1)
a The SD were < 5% for any of the obtained kinetic parameters b In
three independent assays conduced in triplicate the errors obtained
in the percentage of hydrolyzed peptide bond was less than 3%.
Trang 10A previous report demonstrated the influence of
cal-cium and other divalent cations activating EP24.15
enzyme activity, in a substrate dependent manner [38]
consistent with our data Another recent study has also
shown the influence of calcium on the acid limb of a
pH-dependence activity curve of EP24.15 [39]
Fluoro-genic substrates have also been used previously to
show the influence of different salts on the EP24.15
activity [40] Our present data confirm these previous
studies as calcium affected the kcat⁄ KMratio of at least
two fluorogenic substrates hydrolyzed by EP24.15
(Table 2) In addition, we have shown here that
calcium affect the ratio of peptide bond (L–R⁄ R–R)
cleavage observed with the fluorogenic substrates
Abz-GGFLRRDQ-EDDnp and Abz-GGFLRRVQ-EDDnp
assayed These effects were more pronounced using the
substrate Abz-GGFLRRDQ-EDDnp, which contains
an aspartic acid residue that could interact directly
with calcium However, the effect was also verified
with the substrate Abz-GGFLRRVQ-EDDnp that
does not contain any negatively charged residues to
bind calcium directly, thus suggesting that the
struc-tural changes induced by calcium on EP24.15
secon-dary structure could modulate its substrate specificity,
even altering the cleavage site on specific peptides
Moreover, the results obtained with the substrate
Abz-GGFLRRDQ-EDDnp seem to indicate the presence
of a positive residue at the S3¢ subsite Thus, if the
kcat⁄ KMratio obtained from the kinetics with the
sub-strate Abz-GGFLRRDQ-EDDnp containing a
negat-ive change at P3¢ is decomposed into the cleavage sites,
it is possible to verify that the cleavage at the R–R
bond was activated while the cleavage at the L–R
bond was inhibited at higher calcium
concentra-tions Similar analyses using the substrate
Abz-GGFLRRVQ-EDDnp, which that does not contain a
negative change at P3¢, shows inhibition of both R–R
and L–R cleaved sites These observations suggest that
EP24.15 contains a positively charged residue at the
S3¢ subsite This is in agreement with previously
pub-lished reports [3,40], and could be of importance in the
development of new specific substrates and⁄ or
inhibi-tors for the enzyme, as the detailed architecture of the
enzyme and its interactions with substrates are
deci-phered
The mutations D93A and D159A neither prevented
the effect of calcium ions on the cleaved peptide
bond ratio (L–R⁄ R–R) nor changed the kcat⁄ KM
ratio for these substrates (data not shown) These
mutations that could not prevent the structural
chan-ges induced by calcium on EP24.15 structure as
observed with the far UV-CD assays suggest the
possible existence of other calcium binding site(s) in
the EP24.15 structure, which have not yet been des-cribed
In conclusion, the present report provides evidence that EP24.15 is able to associate to the extracellular face of the plasma membrane in a calcium dependent fashion Calcium was also shown to change the secon-dary structure by reduction of a-helical content and to change the substrate specificity of EP24.15 Together, these new findings are useful to understand the com-plex organization of peptide metabolism within cells mediated by divalent cation by altering and regulating enzyme localization within subcellular compartments
in the cell
Experimental procedures
Cell culture Glioma C6 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Invitrogen, Carlsbad, CA, USA)
previ-ously described [12] The cells were maintained in 12-well
air Prior to the experiments, the culture medium from 60– 80% confluent cells was removed and cells were rinsed three times with 1 mL Henk’s buffer (136 mm NaCl,
cal-cium-free experiments, cells were previously incubated for
Determination of EP24.15 activity of intact glioma C6 cells by HPLC
Total extracellular EP24.15 activity was determined using
hydrolysis was consistently < 10% during all the determi-nations The enzymatic reactions were terminated by
by reverse phase HPLC using a C18 lBondapak column
reaction was evaluated by determining the amount of