Báo cáo khoa học: Inhibitors of protein phosphatase 1 and 2A decrease the level of tubulin carboxypeptidase activity associated with microtubules pptx

Inhibitors of protein phosphatase 1 and 2A decrease the level of
tubulin carboxypeptidase activity associated with microtubules
Marı
´
a A. Contı
´
n, Silvia A. Purro, C. Gasto
´
n Bisig, He
´
ctor S. Barra and Carlos A. Arce
Centro de Investigaciones en Quı
´
mica Biolo
´
gica de Co
´
rdoba, CIQUIBIC (UNC-CONICET), Departamento de Quı
´
mica Biolo
´
gica,
Facultad de Ciencias Quı
´
micas, Universidad Nacional de Co
´
rdoba, Argentina
The association of tubulin carboxypeptidase with micro-
tubules may be involved in the determination of the tyrosi-
nation state of the microtubules, i.e. their proportion of
tyrosinated vs. nontyrosinated tubulin. We investigated the

role of protein phosphatases in the association of carb-
oxypeptidase with microtubules in COS cells. Okadaic acid
and other PP1/PP2A inhibitors, when added to culture
medium before isolation of the cytoskeletal fraction, pro-
duced near depletion of the carboxypeptidase activity asso-
ciated with microtubules. Isolation of the native assembled
and nonassembled tubulin fractions from cells treated and
not treated with okadaic acid, and subsequent in vitro assay
of the carboxypeptidase activity, revealed that the enzyme
was dissociated from microtubules by okadaic acid treat-
ment and recovered in the soluble fraction. There was no
effect by nor-okadaone (an inactive okadaic acid analogue)
or inhibitors of PP2B and of tyrosine phosphatases which do
not affect PP1/PP2A activity. When tested in an in vitro
system, okadaic acid neither dissociated the enzyme from
microtubules nor inactivated it. In living cells, prior stabili-
zation of microtubules with taxol prevented the dissociation
of carboxypeptidase by okadaic acid indicating that
dynamic microtubules are needed for okadaic acid to exert
its effect. On the other hand, stabilization of microtubules
subsequent to okadaic acid treatment did not reverse the
dissociating effect of okadaic acid. These results suggest that
dephosphorylation (and presumably also phosphorylation)
of the carboxypeptidase or an intermediate compound
occurs while it is not associated with microtubules, and
that the phosphate content determines whether or not the
carboxypeptidase is able to associate with microtubules.
Keywords: microtubules; PP1; PP2A; tubulin carboxypepti-
dase; tyrosination state.
Microtubules are dynamic structures formed by tubulin

and associated proteins, and are involved in chromosome
segregation, morphogenesis, intracellular transport and
other cell functions [1]. We showed previously [2–4] that
the alpha chain of tubulin can be modified by enzymatic
removal of the C-terminal tyrosine residue by tubulin
carboxypeptidase, and by re-addition of this tyrosine by a
distinct enzyme, tubulin tyrosine ligase. The physiological
role of this cyclic detyrosination/tyrosination reaction has
not been clarified, but is believed to be crucial for normal
microtubule functioning. We are studying the mechanisms
that determine the tyrosination state of microtubules, i.e.
the proportions of tyrosinated vs. nontyrosinated tubulin
(Tyr- and Glu-tubulin, respectively) that constitute a
particular microtubule. Our biochemical studies have
shown that the tyrosination reaction occurs rapidly and
exclusively on nonassembled tubulin, whereas detyrosina-
tion occurs more slowly, and mainly in microtubules [4,5].
These findings were confirmed by studies in living cells
[6,7]. A striking correlation was observed between tyros-
ination state and dynamics of microtubules: Glu- and
Tyr-microtubules are stable and dynamic structures,
respectively [8,9]. On the basis of this concept, supported
by a variety of experiments in different laboratories [10,11],
identification of Tyr- and Glu-microtubules is used at
present as a marker of, respectively, dynamic and stable
microtubules.
We showed in vitro that tubulin carboxypeptidase is
associated with microtubules, and that the association is
modulated by phosphorylation/dephosphorylation reac-
tions [12,13]. Microtubules were reconstituted from soluble

rat brain extracts, and carboxypeptidase activity present in
sedimentable (microtubules) and nonsedimentable fractions
was measured. Preincubation of extracts under conditions
favouring either phosphorylation or dephosphorylation led
to, respectively, lower and higher proportions of carboxy-
peptidase activity associated with microtubules. Total
carboxypeptidase activity was not significantly modified
by conditions favouring phosphorylation. Microtubules
were not the target of the kinase(s) and phosphatase(s)
presumably involved in this phenomenon. We demonstra-
ted recently that the association of carboxypeptidase with
microtubules also occurs in living cells [14,15]. In this paper,
we present evidence that the serine/threonine phosphatases
Correspondence to C. A. Arce, Departamento de Quı
´
mica Biolo
´
gica,
Facultad de Ciencias Quı
´
micas, Ciudad Universitaria, 5000-Co
´
rdoba,
Argentina. Fax: +54 351433 4074, Tel.: +54 351433 4168,
E-mail:
Abbreviations: CPA, pancreatic carboxypeptidase A; Glu-micro-
tubules, microtubules composed mainly of Glu-tubulin; Glu-tubulin,
detyrosinated tubulin or tubulin whose a-subunit lacks a C-terminal
tyrosine residue; MAP, microtubule-associated protein; OA, okadaic
acid; Tyr-microtubules, microtubules composed mainly of Tyr-tubu-

lin; Tyr-tubulin, tubulin with C-terminal tyrosine residue a-subunit.
(Received 21 August 2003, revised 15 October 2003,
accepted 23 October 2003)
Eur. J. Biochem. 270, 4921–4929 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03893.x
PP1 and/or PP2A are involved in regulation of the degree of
tubulin carboxypeptidase activity associated with micro-
tubules, and that microtubule dynamics is necessary to this
regulatory mechanism.
Materials and methods
Chemicals
Nitrocellulose membrane, pancreatic carboxypeptidase A
(CPA), phenylmethanesulfonyl fluoride, EGTA, aprotinin,
benzamidine, nocodazole, Paclitaxel (taxol), 4-chloro-naph-
th-1-ol, Triton X-100, and Mes were from Sigma-Aldrich
Co. Okadaic acid (OA), calyculin A, 1-nor-okadaone,
cantharidin, deltamethrin, and phenylarsine oxide were
from Alomone Laboratories (Israel).
Antibodies
Antibody against Glu-tubulin (anti-Glu) was prepared in
our laboratory as described by Gundersen [16], with
specificity and titre similar to those of samples provided
by the original author. Rat monoclonal YL 1/2 antibody
specific to Tyr-tubulin (anti-Tyr) was from Sera-Lab.
Rhodamine-conjugated goat anti-rabbit secondary anti-
body, fluorescein-conjugated goat anti-mouse secondary
antibody, and peroxidase-conjugated Protein A were from
Sigma-Aldrich Co.
Cell culture
COS-7 cells were grown in Dulbecco’s modified Eagle’s
medium (Sigma) supplemented with 10% (v/v) foetal bovine

serum (Serono) at 37 °Cinanair/CO
2
(19 : 1) incubator.
Cells were plated on plastic Petri dishes (60 mm diameter)
and grown for 2 days until reaching the desired final
density. Culture medium was renewed at 24 h. Cells were
suspended in culture medium by careful scraping and then
transferred to conical plastic tubes. When used, effectors
were maintained in cell suspension by gentle agitation.
Unless stated otherwise, all cell procedures were performed
at 37 °C.
Isolation of cytoskeletal fraction
Cell suspensions (obtained from 60 mm-dishes) were
centrifuged at 600 g for 2 min to remove culture medium.
Sedimented cells were resuspended in 0.5 mL micro-
tubule-stabilizing buffer [90 m
M
Mes pH 6.7, 1 m
M
EGTA, 1 m
M
MgCI
2
, 10% (v/v) glycerol] and centrifuged
again. Pelleted cells were resuspended in 0.5 mL micro-
tubule-stabilizing buffer containing 10 l
M
taxol, 0.5%
(v/v) Triton X-100, and protease inhibitors (10 lgÆmL
)1

aprotinin, 0.5 m
M
benzamidine, 5 lgÆmL
)1
o-phenanthro-
line, 0.2 m
M
phenylmethanesulfonyl fluoride) at 37 °Cfor
2 min with frequent agitation. The tubes were centrifuged
at 8000 g for 2 min and the soluble fraction discarded.
To eliminate residual Triton X-100 and cytosolic fraction,
the pelleted cytoskeletons were rapidly washed twice (by
resuspension and centrifugation) with microtubule-stabil-
izing buffer containing 10 l
M
taxol. Finally, the cytoskele-
tons were suspended in microtubule-stabilizing buffer
containing 10 l
M
taxol and the protease inhibitor
mixture.
Isolation of microtubular and soluble tubulin fractions
from living cells under microtubule-stabilizing conditions
The isolation of native microtubules and nonassembled
tubulin was performed by the method of Pipeleers et al.
[17]. Sedimented cells (0.5 mL) were suspended in 5 mL
warm (37 °C) microtubule-stabilizing buffer [20 m
M
sodium phosphate pH 7, containing 40% (v/v) glycerol,
5% (v/v) dimethylsulfoxide, 0.1 m

M
GTP] and disrupted
with a glass-Teflon homogenizer (20 strokes) and centri-
fuged at 100 000 g for 1 h at 27 °C. The supernatant
fraction was collected and kept at 0 °C. The pellet was
resuspended in 2.5 mL cold disassembling buffer (20 m
M
sodium phosphate buffer pH 7, containing 0.4
M
NaCl
and 0.1 m
M
GTP) and kept at 0 °Cfor30minafter
which it was centrifuged at 100 000 g for 30 min at
2–4 °C. The soluble fraction (disassembled microtubules)
was collected, diluted with 1 vol. 20 m
M
sodium phos-
phate buffer pH 7 to decrease saline concentration and
kept on ice. The first supernatant (soluble tubulin pool)
and the second supernatant (microtubular pool) fractions
were loaded onto small (0.1 mL-bed volume) columns of
cellulose phosphate P11 (Whatman) activated according
to the manufacturer’s instructions and equilibrated with
20 m
M
phosphate buffer pH 7. Tubulin carboxypeptidase
is retained by the resin [13]. Elution is performed with
0.4 mL equilibration buffer containing 0.8
M

NaCl. After
dilution with 4 vols 20 m
M
phosphate buffer in order to
decrease saline concentration, proteins were concentrated
by centrifuging the samples in Centricon-3 devices
(Amicon). After reducing volumes to 0.1 mL, carboxy-
peptidase activity was assayed immediately.
Measurement of tubulin carboxypeptidase activity
We used two different methods. The carboxypeptidase
activity associated with the isolated cytoskeletal fraction
was quantified as the increase in Glu-tubulin amount as
a function of incubation time. Immediately after isolation,
cytoskeletons (contained in conical plastic tubes) were
incubated at 37 °C in 0.25 mL microtubule-stabilizing
buffer containing taxol and protease inhibitors as above.
After various incubation times, the Glu-tubulin content was
determined by immunoblotting.
When the activity of tubulin carboxypeptidase was
determined in the microtubule and soluble tubulin fractions
isolated under microtubule-stabilizing conditions, a method
based on the release of [
14
C]tyrosine from [
14
C]tyrosinated
tubulin was used [18]. In brief, varying aliquots of the
enzyme preparations were loaded onto nitrocellulose
circles containing adsorbed [
14

C]tyrosinated tubulin
( 4000 c.p.m.) and after addition of 100 lL albumin
solution (10 mgÆmL
)1
), the systems were incubated at
37 °C for 1 h. Then, soluble fractions which contain
released [
14
C]tyrosine, were transferred to vials and radio-
activity determined in a liquid scintillation counter. In
several independent experiments, time curves performed
using 20 lL of each enzyme preparation showed linearity
up to 1 h.
4922 M. A. Contı
´
n et al. (Eur. J. Biochem. 270) Ó FEBS 2003
Immunoblotting
Following incubation as above, tubes were centrifuged at
1600 g for 2 min and the supernatant discarded. Pellets
were dissolved in 60 lL sample buffer [19] by heating at
90 °C for 2 min. Samples were subjected to SDS/PAGE
(10% gel) by the method of Laemmli [19] and transferred
to nitrocellulose sheets [20]. Two identical gels were run in
parallel. One sheet was treated with 10 lgÆmL
)1
CPA
(30 min, 37 °C) and extensively washed. Both sheets were
then blocked for 1 h with 5% (w/v) fat-free dried milk
dissolved in NaCl/Tris containing 0.1% (v/v) Triton X-
100, and blots were treated for 3 h at room temperature

with anti-Glu antibody diluted 1 : 200, and washed.
Sheets were incubated for 1 h at room temperature in
the presence of horseradish peroxidase conjugated to
Protein A (dilution 1 : 1000), and washed. Colour was
developed using 4-chloro-naphth-1-ol as chromogen.
Because CPA converts all Tyr-tubulin to the Glu form
[14], total tubulin amount was determined from the CPA-
treated sheet.
Quantification of Glu-tubulin
Immunoblots were scanned and amount of Glu-tubulin in
cytoskeletal preparation was determined as in Contin et al.
[14]. The amount of Glu-tubulin in a particular sample is
expressed as a percentage of total detyrosinable tubulin,
calculated as 100 (A
no CPA
/A
CPA
), where A
no CPA
and A
CPA
are the absorbances of the control and CPA-treated
samples, respectively. Provided that the numerator and
the denominator correspond to identical samples, this
expression is independent of the amount of protein loaded.
The method is described more fully in Contin et al. [14].
Within a particular independent experiment, each value is
the average of two samples run in parallel. In some cases,
values are expressed as the mean ± SD of three to five
independent experiments.

Immunofluorescence
After defined durations of incubation of cytoskeletons,
samples on coverslips were fixed with methanol at )20 °C
for5min,andstoredat2–4°CinNaCl/P
i
containing
0.2% sodium azide until use. Fixed cytoskeletons were
incubatedwith2%(w/v)BSAinNaCl/P
i
for 60 min and
stained by double indirect immunofluorescence using anti-
Glu and anti-Tyr (dilution 1 : 200 and 1 : 500, respect-
ively). Secondary antibodies were used simultaneously at
1 : 200 dilution in NaCl/P
i
/BSA. Coverslips were moun-
ted in FluorSave and epifluorescence was observed on an
Axioplan microscope (Zeiss). Images were captured with
a sensitive, digital camera (Princeton Instrument) and
stored on a CD for subsequent analysis. Estimations of
Glu- and Tyr-microtubules present in fields selected at
random were obtained by measuring the integrated
intensity of the corresponding immunostaining with the
aid of the
METAMORPH IMAGING SYSTEM
(Version 4.6r5).
For a determined field, the value of integrated intensity of
Glu-microtubules was divided by that of Tyr-micro-
tubules to estimate the relative proportion of Glu- with
respect to Tyr-microtubules.

Treatment of cells with effectors
Cells were treated at 37 °C with various effector drugs and
maintained in an incubator until the time of cytoskeletal
fraction isolation. Stock solutions of effectors were
prepared in dimethylsulfoxide such that final solvent
concentration in the growth medium did not exceed
0.5% (v/v). Controls were performed by adding 0.25%
(v/v) dimethylsulfoxide to the medium. This concentration
of dimethylsulfoxide had no effect on distribution or level
of Glu-tubulin in cells.
Results
Exposure of cells to okadaic acid induces decrease
in the activity of tubulin carboxypeptidase associated
with microtubules in living cells
The level of association of tubulin carboxypeptidase
activity with microtubules was determined by measuring
enzyme activity present in cytoskeletons freed of soluble
components. Isolated cytoskeletons were incubated in vitro,
and carboxypeptidase activity was inferred from the
increase of the reaction product, detyrosinated tubulin
(Glu-tubulin), as a function of incubation time. The slope
of the time curves provides an estimate of the amount of
the associated carboxypeptidase. This method showed such
an association in several cell lines [14]. Now, we investi-
gated the effect of protein phosphatase inhibitors on the
association of carboxypeptidase with microtubules in COS
cells. We first tested the effect of OA [21,22], which
produces a marked increase in phosphorylation of many
proteins in living cells. When OA was added (1 l
M

final
concentration) to culture medium 1 h before isolation of
cytoskeletons, production of Glu-tubulin during in vitro
incubation was significantly reduced (Fig. 1). This indicates
that OA treatment of the cells induced a decrease in
carboxypeptidase activity associated with microtubules.
Replacement of OA by its inactive analogue, 1-nor-
okadaone, resulted in activity associated with microtubules
similar to that of control.
The effect of OA on the carboxypeptidase/microtubule
association was analysed by double immunofluorescence
using cells cultured on glass coverslips. Fig. 2 shows
images representative of many fields observed in each case.
Freshly isolated cytoskeletons from untreated cells con-
tained minor amounts of Glu-microtubules (Fig. 2A),
whereas Tyr-microtubules were observed as brightly
stained structures (Fig. 2B). Similar results were obtained
with 1-norokadaone-treated cells (data not shown). When
isolated cytoskeletons were incubated at 37 °Cfor2h,
Glu-microtubules were clearly stained (Fig. 2C), whereas
in cytoskeletons from OA-treated cells the staining revealed
no Glu-microtubules (Fig. 2E), indicating lack of carb-
oxypeptidase activity in these microtubules. In 1-nor-
okadaone treated cells, microtubules were brightly stained
after 2 h in vitro incubation (Fig. 2G). These results, again,
indicate that the effect of OA on the carboxypeptidase
activity associated with microtubules is based on its
inhibitory effect on protein phosphatase activities. Fluor-
escence intensity measurements of Glu-microtubules relat-
ive to Tyr-microtubules (see statistical values in the legend

Ó FEBS 2003 Tubulin carboxypeptidase/microtubule association
1
(Eur. J. Biochem. 270) 4923
of Fig. 2) confirmed conclusions drawn from direct
visualization.
Exposure of cells to OA induces redistribution of tubulin
carboxypeptidase activity between the microtubule-
associated and nonassociated states
The scarce (or null) tubulin carboxypeptidase activity
associated with the cytoskeletons of OA-treated cells could
be attributed to: (a) inhibition of the enzyme while
remaining associated; or (b) dissociation of the enzyme
from microtubules. To clarify this point, we investigated the
enzyme activity associated and nonassociated with micro-
tubules in cells treated and nontreated with OA. Since the
detergent-extracting method used in this work to isolate
the cytoskeleton fraction produces a great dilution of the
soluble fraction, determination of carboxypeptidase activity
in this fraction was not possible. Therefore, to perform this
study we disrupted cells under microtubule-stabilizing
conditions and separated the microtubular and soluble
fractions by centrifugation as described by Pipeleers et al.
[17]. Carboxypeptidase present in the assembled and
nonassembled tubulin fractions from cells treated and
nontreated with 1 l
M
OA was concentrated on phospho-
cellulose columns (for details see Materials and methods)
and enzyme activity determined. As shown in Fig. 3A, in
control cells (nontreated with OA), higher carboxypeptidase

activity was found in the microtubule fraction as compared
with the soluble fraction. In contrast, when cells were
treated with OA, the major proportion of activity was
present in the soluble fraction (Fig. 3B). Another observa-
tion from Fig. 3 is that the sum of the activities recovered in
both fractions is approximately the same when compared
control and OA-treated cells. These results clearly indicate
Fig. 1. Effect of OA treatment on level of tubulin carboxypeptidase
activity associated with microtubules in living cells. COS cells were
grown in Petri dishes to 60–70% confluence. OA (1 l
M
final concen-
tration), 1-nor-okadaone (1 l
M
), or no compound (control) was added
to the culture medium and incubation continued for 1 h. Cytoskeletal
fractions were then isolated, incubated for the stated times, and sub-
jected to Western blotting with anti-Glu to determine the tubulin
carboxypeptidase activity associated with microtubules as described in
Materials and methods. Upper panel: before immunostaining, the
nitrocellulose membranes were treated (+CPA) or not (–CPA) with
pancreatic carboxypeptidase A which produces full detyrosination of
tubulin [14]. Lower panel: blots shown in the upper panel were used
to quantify Glu-tubulin. Results are expressed as percentage of
total detyrosinable tubulin. s, control; ,,+okadaicacid;h,
+1-nor-okadaone. Results are mean ± SD of four independent
experiments.
Fig. 2. Visualization of Glu- and Tyr-microtubules by double immuno-
fluorescence, showing the effect of OA on the activity of tubulin carb-
oxypeptidase associated with microtubules. COS cells were grown on

glass coverslips and treated with effectors as in Fig. 1. After isolation,
cytoskeletons were incubated for 2 h at 37 °C and processed for
double immunofluorescence using anti-Glu (A,C,E,G) and anti-Tyr
(B,D,F,H) Igs. (A,B) Freshly isolated cytoskeletons (t ¼ 0 incub-
ation). At this time, pictures similar to (A) and (B) were obtained for
OA- and nor-okadaone-treated cells (not shown). (C–H) Cytoskele-
tons incubated in vitro for 2 h. (C,D) Control cells. (E,F) OA-treated
cells. (G,H) Nor-okadaone-treated cells. Scale bar, 10 lm. For each
panel, fluorescence intensity was measured by using the
METAMORPH
IMAGING SYSTEM
and, for each condition, the ratio Glu/Tyr was cal-
culated. A/B ¼ 0.17 ± 0.03; C/D ¼ 1.21 ± 0.15; E/F ¼ 0.28 ±
0.04; G/H ¼ 1.12 ± 0.17. Each value represents the mean ± SE of
four independent experiments.
4924 M. A. Contı
´
n et al. (Eur. J. Biochem. 270) Ó FEBS 2003
that the effect of the PP1/PP2A inhibitor was to induce
redistribution of the enzyme between the microtubule-
associated and nonassociated states rather than to inhibit it.
Type of phosphatase(s) involved in regulation
of the carboxypeptidase activity associated with
microtubules
To determine the type of phosphatase(s) involved, we tested
effects of various compounds that specifically inhibit
different phosphatases. Among these compounds, only
calyculin A and cantharidin, two well-known inhibitors of
PP1 and PP2A [23,24], showed effects similar to that of OA
(Fig. 4). Deltamethrin, a specific inhibitor of PP2B [25], and

phenylarsine oxide, a putative inhibitor of tyrosine phos-
phatases [26], had no effect, even though the concentrations
used in our experiments (10 l
M
) were higher than those
reported to inhibit the corresponding phosphatases (100 p
M
and 5 l
M
, respectively) [25,26]. These results suggest that the
effects of OA, calyculin A, and cantharidin on activity of
carboxypeptidase associated with microtubules are due to
their inhibition of phosphatase activity, rather than to a side
effect. The phosphatases involved seem to be PP1 and/or
PP2A although it is difficult at this time to distinguish
between them.
In vitro
effect of OA on tubulin carboxypeptidase
activity associated with microtubules
The possibility that OA causes dissociation of the enzyme
from microtubules through direct interaction was ruled out
by the following experiment. Cytoskeletal fraction of
nontreated cells was incubated in the presence or absence
(control) of OA to determine its effect on associated
carboxypeptidase activity. OA had no effect on the enzyme
activity, and renewal of incubation medium 30 min after
addition of OA did not alter subsequent detyrosination,
indicating that OA does not cause direct dissociation of
carboxypeptidase from microtubules (Fig. 5). If such
dissociation had occurred, the enzyme would have been

eliminated during removal of medium and detyrosination
would have stopped. The incubated cytoskeletons represent
only a part of the cell components, and they contain
microtubules that were stabilized with taxol during the
isolation procedure; therefore, these results suggest that
intact cells and/or dynamic microtubules are required for
phosphatase inhibitor to exert its inhibitory effect on the
carboxypeptidase activity associated with microtubules. The
experiments shown below address this point.
Fig. 4. Effect of protein phosphatase inhibitors on the carboxypeptidase
activity associated with microtubules in living cells. COS cells were
grown, treated with the effectors indicated below and processed as in
Fig. 1. The following effectors were tested separately by addition into
culture medium: OA (1 l
M
final concentration); calyculin A (5 l
M
);
cantharidin (40 l
M
); calyculin A plus cantharidin (5 and 40 l
M
,
respectively); deltamethrin (10 l
M
); phenylarsine oxide (10 l
M
). Glu-
tubulin was determined in cytoskeletons at t ¼ 0andafter2hof
incubation. Results are mean ± SD of three independent experiments.

Fig. 3. Effect of OA on the distribution of tubulin carboxypeptidase
activity between the microtubule-associated and nonassociated states.
Confluent COS cells from twenty 100 mm Petri dishes were collected,
suspended in incubation medium and separated into two fractions
(5 mL each). The fractions were incubated at 37 °C in the presence or
absence of 1 l
M
OA for 1 h with gentle agitation. Cells were sedi-
mented, washed once with microtubule-stabilizing buffer, and homo-
genized to isolate the microtubular and soluble tubulin fractions. Both
fractions were concentrated and assayed as described in Materials and
methods. Carboxypeptidase activities corresponding to the micro-
tubule-associated (s) and nonassociated (d) states are shown for
control (upper panel) and OA-treated (lower panel) cells. Results are
mean ± SD of four independent experiments.
Ó FEBS 2003 Tubulin carboxypeptidase/microtubule association
1
(Eur. J. Biochem. 270) 4925
Effect of stabilization of microtubules with taxol on the
dissociation of carboxypeptidase from microtubules
induced by OA
We investigated the possible involvement of microtubule
dynamics in the mechanism by which the OA treatment of
the cells results in a low activity of carboxypeptidase
associated with microtubules. Microtubules in living cells
were stabilized by addition of 10 l
M
taxol 10 min before
addition of OA. One hour later, cytoskeletons were isolated
and incubated to determine associated carboxypeptidase

activity. Treatment with taxol prior to OA addition preven-
ted the dissociating effect of the phosphatase inhibitor
(Fig. 6A, j and .; . and d). This result supports the idea
that dynamic microtubules are necessary for OA to decrease
the activity of carboxypeptidase associated with micro-
tubules. The possibility that this result is due to a neutralizing
effect of taxol on phosphatase inhibitory activity was ruled
out by the following experiment. Using an in vitro assay in
which phosphatase activity present in soluble rat brain
extract is partially inhibited by OA, we found that taxol had
no effect on such inhibition (data not shown).
On the other hand, when taxol was added following OA
treatment, the decrease in the activity of tubulin carboxy-
peptidase associated with microtubules was not reverted
(Fig. 6B). This reveals that, once the enzyme has been
dissociated from microtubules by the phosphatase inhibitor,
it cannot be re-associated even when microtubules are
stabilized.
Detyrosination of tubulin in living cells can proceed even
when tubulin carboxypeptidase is not associated with
microtubules
There is increasing evidence of an association of tubulin
carboxypeptidase with microtubules and energy consump-
tion, which regulates its distribution between the micro-
tubule-associated and nonassociated states ([13–15] and this
study). We therefore investigated whether this association is
a necessary event for detyrosination of microtubules, taking
advantage of the fact that once the phosphatases have been
inhibited by OA, subsequent addition of taxol does not
reverse the dissociation of carboxypeptidase from micro-

tubules (Fig. 6). We treated living cells with OA to induce
dissociation of carboxypeptidase from microtubules, and
then added taxol to stabilize microtubules, and continued
the culture of intact cells. The amount of Glu-tubulin in cells
was measured as a function of time in culture following
addition of taxol. The amount of Glu-tubulin was directly
correlated with incubation time (Fig. 7), indicating that
Fig. 5. In vitro effect of OA on tubulin carboxypeptidase activity asso-
ciated with microtubules. COS cells were grown to 60–70% confluence,
and cytoskeletons were isolated and incubated in vitro for the stated
durations. At the end of the incubation period, Glu-tubulin was
determined and expressed as in Fig. 1. s, control (incubation without
added compound); ,,att¼ 0, OA (1 l
M
final concentration) was
added to incubation medium; h,att¼ 0, OA was added, and at
t ¼ 30 min (arrow) incubation medium was removed and replaced by
fresh medium lacking OA.
Fig. 6. Effect of stabilization of microtubules previous or subsequent to
OA treatment on the inhibition of tubulin carboxypeptidase activity
associated with microtubules. (A) COS cells were grown to 60–70%
confluence. Taxol (10 l
M
final concentration) was added to culture
medium. After 10 min, OA (1 l
M
) was added and culturing continued
for a further 1 h. Cytoskeletal fractions were isolated and incubated for
the stated times. At the end of the incubation period Glu-tubulin was
determined and expressed as in Fig. 1. ., Cells treated with OA alone

for 1 h prior to isolation of cytoskeletons; j, cells treated with taxol
for 10 min and subsequently with OA for 1 h longer; d,control
(nontreated) cells. (B) Cells were treated with 1 l
M
OA for 1 h, and
then with 10 l
M
taxol for 10 min. Cytoskeletons were isolated and
incubated for the stated times to determine the amount of tubulin
carboxypeptidase activity associated with microtubules. ., cells trea-
ted with OA alone for 70 min; j, cells treated with OA for 1 h and
then with taxol for 10 min; d, control (nontreated) cells.
4926 M. A. Contı
´
n et al. (Eur. J. Biochem. 270) Ó FEBS 2003
detyrosination of tubulin within intact cells proceeded even
when carboxypeptidase was not associated with micro-
tubules. In comparative experiments, cells treated with OA
alone (+OA) or no treatment (control) showed no increase
of Glu-tubulin, and cells treated with taxol alone (+taxol)
were detyrosinated faster than cells treated with OA and
then with taxol. Evaluation of the slope of the curves
indicated that detyrosination can occur when carboxypepti-
dase is not associated with microtubules within the cell,
although at a rate  2.5 times lower than when it is
associated with microtubules.
Discussion
The amount of tubulin carboxypeptidase activity
associated with microtubules is regulated by
phosphorylation/dephosphorylation events

in living cells
The results described above show that tubulin carboxy-
peptidase activity associated with microtubules was very
low when cells were treated with 1 l
M
OA (Figs 1 and 2).
The possibility that the decrease was due to a lower amount
of the enzyme associated with microtubules ) with a
corresponding increase in the cytosolic fraction ) was
tested and confirmed by biochemical assay of enzyme
activity in the microtubule-associated and soluble fractions,
as isolated by a properly established method that preserves
native microtubules [17]. In effect, carboxypeptidase in
control cells was associated mainly with microtubules,
whereas in OA-treated cells, the higher proportion of
enzyme activity was in the soluble fraction (Fig. 3). As the
sum of the activities of both fractions is practically the same
for OA-treated and untreated cells (Fig. 3), it appears that
the effect of OA is not to inhibit the carboxypeptidase but to
redistribute it. Complementary observations also support
this conclusion: (a) enzyme activity was unchanged when
cytoskeletons were incubated in vitro with 1 l
M
OA (Fig. 5);
(b) enzyme activity was not modified by OA in living cells
previously treated with taxol (Fig. 6A); (c) no alteration of
tubulin carboxypeptidase activity was reported previously
in fibroblasts and epithelial cells treated with OA [27]; (d) a
previous in vitro study [13] showed that tubulin carboxy-
peptidase activity of a rat brain soluble fraction was

unchanged regardless of incubation conditions favouring
vs. not favouring high phosphorylation.
OA by itself did not disrupt the association of tubulin
carboxypeptidase with microtubules (Fig. 5), and the inac-
tive OA analogue 1-nor-okadaone also had no effect on this
association (Figs 1 and 2). These observations suggest that
the effect of OA on the carboxypeptidase activity associated
with microtubules is mediated by its capacity to inhibit
protein phosphatases. Other phosphatase inhibitors (caly-
culin A and cantharidin) showed a similar effect on the
association (Fig. 4). These drugs are structurally unrelated,
and it is unlikely that all of them would produce the same
side effect. OA, calyculin A, and cantharidin are all serine/
threonine-specific protein phosphatase inhibitors specific for
PP1 and PP2A but, at the concentrations tested they are not
inhibitors of PP2B, PP2C, or tyrosine phosphatases. Other
compounds such as deltamethrin and phenylarsine oxide
which inhibit, respectively, PP2B and tyrosine phospha-
tases, did not affect the carboxypeptidase activity associated
with microtubules (Fig. 4). These results confirm that the
OA effect on the carboxypeptidase activity associated with
microtubules is mediated by its capacity to inhibit protein
phosphatases, and indicate that PP1 and/or PP2A are
probably the phosphatases involved in regulation of this
phenomenon in living cells.
It remains unclear whether the target of phosphoryla-
tion/dephosphorylation is tubulin carboxypeptidase itself
or an intermediary compound [for example, microtubule-
associated protein (MAP) or microtubule-based motor
protein] which, according to its phosphate content, could

interact with the enzyme and allow it (or not) to become
associated with microtubules. The presence of most
MAPs on the microtubule surface is known to be
modulated by phosphate group content of their serine
and threonine residues [28–30]; a high phosphate MAP
content precludes association, and vice versa. Alternat-
ively, one can imagine a cascade of biochemical events (at
least one of them controlled by phosphorylation/dephos-
phorylation) which eventually allows (or not) the enzyme
to associate with microtubules. In any case, phosphory-
lation/dephosphorylation events are clearly involved in
association of carboxypeptidase with microtubules in
living cells.
Phosphorylation/dephosphorylation of tubulin
carboxypeptidase (or an intermediary compound)
is dependent on disassembly of microtubules
The fact that nondynamic microtubules (stabilized with
taxol) retain associated tubulin carboxypeptidase activity
when cells are subsequently treated with OA (Fig. 6A)
indicates that: (a) OA does not dissociate the enzyme
from microtubules by direct interaction or through its
Fig. 7. Detyrosination of microtubules by nonassociated carboxypepti-
dase in intact cells. COS cells were treated with or without 1 l
M
OA for
1 h and subsequently with or without 10 l
M
taxol. Time of taxol
addition was defined as zero. Cells were incubated for the indicated
times, and cytoskeletons were isolated and immediately processed to

measure amount of Glu-tubulin as described in Materials and meth-
ods. s, Control (nontreated) cells; d, cells treated with OA alone; ,,
cells treated with taxol alone; ., cells treated first with OA and then
with taxol. Data shown are mean ± SD of five experiments.
Ó FEBS 2003 Tubulin carboxypeptidase/microtubule association
1
(Eur. J. Biochem. 270) 4927
phosphatase inhibitory activity; and (b) the dynamics of
microtubules is required for OA to reduce association of
carboxypeptidase with microtubules. The requirement for
dynamic microtubules while OA is exerting its effect agrees
with the idea that the disassembly phase is an obligatory
step for carboxypeptidase to become a soluble entity.
Disassembly of microtubules, during normal equilibrium, is
presumably the means by which carboxypeptidase becomes
a soluble entity. After disassembly, the target molecule
could be subjected to phosphorylation/dephosphorylation
by the respective kinases and phosphatases. Then, according
to the resulting phosphate content, the carboxypeptidase
could coassemble with tubulin in the assembly phase of
equilibrium, or associate directly on the surface of micro-
tubules. This view is supported by the finding that OA
treatment prior to stabilization led to formation of micro-
tubules without carboxypeptidase activity (Fig. 6B).
Is association of tubulin carboxypeptidase with
microtubules necessary to catalyse detyrosination?
One might initially hypothesize that this association results
in rapid production of detyrosinated microtubules. How-
ever, in confluent cells, where carboxypeptidase is maxi-
mally associated with microtubules, they remain mostly

tyrosinated [14]. The mere association of the enzyme with
microtubules therefore does not seem to guarantee rapid
detyrosination. A plausible hypothesis is that the association
is a necessary but not sufficient condition. Stabilization of
microtubules could be the complementary factor required
for effective detyrosination. This is the basis for the
generally accepted definition of stable and dynamic micro-
tubules as Glu- and Tyr-microtubules, respectively. There
seems to be no doubt that Glu-microtubules are always
stable. On the other hand, Tyr-microtubules are not
necessarily always dynamic structures—they may be stable
when lacking associated carboxypeptidase, e.g. cultured
nerve cells contain a nocodazole- and cold-resistant subset
of microtubules having a higher content of Tyr-tubulin than
the mean population [15]. Although these prior studies
suggest that association of carboxypeptidase with micro-
tubules is necessary for their detyrosination, the alternative
possibility that nonassociated carboxypeptidase also cata-
lyses detyrosination is supported by findings in the present
study.
Our findings suggest that even though association of
tubulin carboxypeptidase with microtubules results in faster
detyrosination (Fig. 7), this association is not a requirement
for detyrosination, i.e. any microtubule may undergo
detyrosination regardless of presence vs. absence of associ-
ated carboxypeptidase. If true, this concept would imply
that the association/dissociation phenomenon is not a
regulatory factor determining the tyrosination state of
microtubules. However, because of the variety and com-
plexity of cellular physiological processes, we hesitate to

state this conclusion definitively without further experimen-
tal confirmation. Even though nonassociated carboxypepti-
dase can catalyse detyrosination, one can speculate that,
within the cell, all (or most) carboxypeptidase, in response
to certain signals (perhaps enzyme phosphate content),
could be associated with microtubules, i.e. no enzyme is in
the nonassociated state. In this case, the only microtubules
capable of undergoing detyrosination would be those
having associated enzyme. Studies to resolve this point are
underway.
Acknowledgements
We thank C.A. Argaran
˜
aandC.R.Ma
´
s for critical reading of the
manuscript; S.N. Deza and M.G. Schachner for technical assistance,
and S. Anderson for English editing of the manuscript. This work was
supported partly by grants from Agencia Nacional de Promocio
´
n
Cientı
´
fica y Tecnolo
´
gica de la Secretarı
´
a de Ciencia y Tecnologı
´
adel

Ministerio de Cultura y Educacio
´
n en el marco del Programa
de Modernizacio
´
n Tecnolo
´
gica (BID 802/0C-AR), Consejo Nacional
de Investigaciones Cientı
´
ficas y Te
´
cnicas (CONICET), Secretarı
´
ade
Ciencia y Te
´
cnica de la Universidad Nacional de Co
´
rdoba y Agencia
Co
´
rdoba Ciencia del Gobierno de la Provincia de Co
´
rdoba, Argentina.
References
1. Hyams, J.S. & Lloyd, C.W. (1993) In Microtubules. Wiley-Liss,
New York, USA.
2. Barra, H.S., Arce, C.A. & Argaran
˜

a, C.E. (1988) Post-transla-
tional tyrosination/detyrosination of tubulin and microtubules.
Mol. Neurobiol. 2, 133–153.
3. Hallak, M.E., Rodriguez, J.A., Barra, H.S. & Caputto, R. (1977)
Release of tyrosine from tyrosinated tubulin. Some common fac-
tors that affect this process and the assembly of tubulin. FEBS
Lett. 73, 147–150.
4. Arce, C.A., Hallak, M.E., Rodriguez, J.A., Barra, H.S. &
Caputto, R. (1978) Capability of tubulin and microtubules to
incorporate and to release tyrosine and phenylalanine and the
effect of the incorporation of these amino acids on tubulin
assembly. J. Neurochem. 31, 205–210.
5. Arce, C.A. & Barra, H.S. (1985) Release of C-terminal tyrosine
from tubulin and microtubules at steady state. Biochem. J. 226,
311–317.
6. Gundersen, G.G., Khawaja, S. & Bulinski, J.C. (1987) Post-
polymerization detyrosination of alpha-tubulin: a mechanism for
subcellular differentiation of microtubules. J. Cell Biol. 105,
251–264.
7. Schulze, E., Asai, D.J., Bulinski, J.C. & Kirschner, M. (1987)
Posttranslational modification and microtubule stability. J. Cell
Biol. 105, 2167–2177.
8. Kreis, T.E. (1987) Microtubules containing detyrosinated tubulin
are less dynamic. EMBO J. 6, 2597–2606.
9. Geuens, G., Gundersen, G.G., Nuydens, R., Comelissen, F.,
Bulinski, J.C. & DeBrabander, M. (1986) Ultrastructural coloca-
lization of tyrosinated and detyrosinated alpha-tubulin in inter-
phase and mitotic cells. J. Cell Biol. 103, 1883–1893.
10. Bre
´

, M.H., Kreis, T.E. & Karsenti, E. (1987) Control of micro-
tubule nucleation and stability in Madin-Darby canine kidney
cells: the occurrence of noncentrosomal, stable detyrosinated
microtubules. J. Cell Biol. 105, 1283–1296.
11. Gundersen, G.G. & Bulinski, J.C. (1986) Microtubule arrays in
differentiated cells contain elevated levels of a post-translationally
modified form of tubulin. Eur. J. Cell. Biol. 42, 288–294.
12. Arce, C.A. & Barra, H.S. (1983) Association of tubulinyl-tyrosine
carboxypeptidase with microtubules. FEBS Lett. 157, 75–78.
13. Sironi, J.J., Barra, H.S. & Arce, C.A. (1997) Association of tubulin
carboxypeptidase activity with microtubules in brain extracts is
modulated by phosphorylation/dephosphorylation processes.
Mol. Cell. Biochem. 170, 9–16.
14. Contı
´
n,M.A.,Sironi,J.J.,Barra,H.S.&Arce,C.A.(1999)
Association of tubulin carboxypeptidase with microtubules in
living cells. Biochem. J. 339, 463–471.
4928 M. A. Contı
´
n et al. (Eur. J. Biochem. 270) Ó FEBS 2003
15. Contı
´
n, M.A. & Arce, C.A. (2000) Tubulin carboxypeptidase/
microtubules association can be detected in the distal region of
neural processes. Neurochem. Res. 25, 27–36.
16. Gundersen, G.G., Kalnoski, M.H. & Bulinski, J.C. (1984) Distinct
populations of microtubules: tyrosinated and nontyrosinated
alpha tubulin are distributed differently in vivo. Cell 38, 779–789.
17. Pipeleers, D.G., Pipeleers-Marichal, M.A., Sherline, P. & Kipnist,

D.M. (1977) A sensitive method for measuring polymerized and
depolymerized forms of tubulin in tissues. J. Cell Biol. 74, 341–350.
18. Sironi, J.J., Barra, H.S. & Arce, C.A. (2000) Tubulin carboxy-
peptidase assay, based on the action of the enzyme on [14C]tyro-
sinated tubulin bound to nitrocellulose membrane. Anal. Biochem.
279, 9–17.
19. Laemmli, U.K. (1970) Cleavage of structural proteins during the
assembly of the head of bacteriophage T 4. Nature (London) 227,
680–685.
20. Towbin, H., Staehelin, T. & Gordon, J. (1979) Electrophoretic
transfer of proteins from polyacrylamide gels to nitrocellulose
sheets: procedure and some applications. Proc. Natl Acad. Sci.
USA 76, 4350–4354.
21. Cohen, P., Klumpp, S. & Schelling, D.L. (1989) An improved
procedure for identifying and quantitating protein phosphatases in
mammalian tissues. FEBS Lett. 250, 596–600.
22. Biolajan, C. & Takai, A. (1988) Inhibitory effect of a marine-
sponge toxin, okadaic acid, on protein phosphatases. Biochem. J.
256, 283–290.
23. Suganuma, M., Fujiki, H., Furuya-Suguri, H., Yoshizawa, S.,
Yasumoto,S.,Kato,Y.,Fusotani,N.&Sugimura,T.(1990)
Calyculin A, an inhibitor of protein phosphatases, a potent tumor
promoter on CD-1 mouse skin. Cancer Res. 50, 3521–3525.
24. Li, Y M. & Casida, J.E. (1992) Cantharidin-binding protein:
identification as protein phosphatase 2A. Proc. Natl Acad. Sci.
USA 89, 11867–11870.
25. Enan, E. & Matsumura, F. (1992) Specific inhibition of calcineurin
by type II synthetic pyrethroid insecticides. Biochem. Pharmacol.
43, 1777–1784.
26. Garcia-Morales, P., Minami, Y., Luong, E., Klausner, R.D. &

Samuelson, L.E. (1990) Tyrosine phosphorylation in T cells is
regulated by phosphatase activity: studies with phenylarsine oxide.
Proc.NatlAcad.Sci.USA87, 9255–9259.
27. Gurland, G. & Gundersen, G.G. (1993) Protein phosphatase
inhibitors induce the selective breakdown of stable microtubules in
fibroblasts and epithelial cells. Proc. Natl Acad. Sci. USA 90,
8827–8831.
28. Jameson, L., Frey, T., Zeeberg, B., Dalldorf, F. & Caplow, M.
(1980) Inhibition of microtubule assembly by phosphorylation of
microtubule-associated proteins. Biochemistry 19, 2472–2479.
29. Drechsel, D.N., Hyman, A.A., Cobb, M.H. & Kirschner, M.W.
(1992) Modulation of the dynamic instability of tubulin assembly
by the microtubule-associated protein tau. Mol. Biol. Cell 3, 1141–
1154.
30. Sloboda, R.D., Rudolph, S.A., Rosenbaum, J.L. & Greengard, P.
(1975) Cyclic AMP-dependent endogenous phosphorylation of a
microtubule-associated protein. Proc. Natl Acad. Sci. USA 12,
177–181.
Ó FEBS 2003 Tubulin carboxypeptidase/microtubule association
1
(Eur. J. Biochem. 270) 4929