Trafficking of GFP-AQP5 chimeric proteins conferred with unphosphorylated amino acids at their PKA-target motif ((152)SRRTS) in MDCK-II cells.

Abstract

Three constructs having mutated PKA-target motif at (152)SRRTS of AQP5, an exocrine type water channel, were prepared and fused to C-terminus of green fluorescence protein cDNA to examine the effects of blocking of phosphorylation at (152)SRRTS (a consensus PKA-target motif of AQP5) on translocation or trafficking of the chimeric proteins expressed in the Madin-Darby canine kidney-II (MDCK-II) cells. H-89 treatment increased translocation of wild-type GFP-AQP5 to the apical membrane. All 3 mutant molecules translocated 1.5 to 2 times more than the control wild-type GFP-AQP5. Colchicine but not cytochalasin B inhibited the translocation of wild-type GFP-AQP5. Present results suggest dephosphorylation of this consensus sequence increase GFP-AQP5 translocation, and that microtubules but not microfilaments are involved in this event.

Full text

Page 1

INTRODUCTION
Aquaporins (AQPs) are the water channel pro-
teins that have been identified in virtually all living
organisms, and they are generally responsible for
rapid water movement across the plasma membrane
in almost all cells (1, 2). Among the 13 mammalian
aquaporins identified so far, i.e., AQP0-AQP12 (2),
AQP5 is reported to be expressed in the cell mem-
brane of multiple secretory glands, including the lac-
rimal, salivary, and airway submucosal glands, as
well as in type 1 alveolar cells (3, 4), sweat glands
(5), corneal epithelium (6), and duodenal Brunner’s
gland (7). The essential role of AQP5 has been
shown in the study of AQP5 gene knockout mice,
which demonstrated significantly increased lethality
of the embryo and problems related with water ho-
meostasis (8, 9). In addition, defective cellular traf-
ficking was noted in the lacrimal gland and salivary
gland biopsies from patients with Sjögren’s syn-
drome but not in patients with non-Sjögren’s dry
eye and dry mouth (10, 11), although there are
some opposite reports (12).
In a murine lung epithelial cell line (MLE-12), a
cAMP analog, 8-(4-chlorophenylthio)-cAMP (cpt-
cAMP), was found to induce the AQP5 mRNA and
protein expression (13). It is also reported that such
stimulations induce translocation of AQP5 from the
intracellular storage sites to the apical membrane
(14-17) ; i.e., AQP5, majority of which is distributed
in the cytoplasm, can be translocated to the plasma
membrane in response to cpt-cAMP in MLE-12 cells
(13). AQP5 was also targeted to the cell membrane
after incubation with cpt-cAMP in Madin-Darby
ORIGINAL
Trafficking of GFP-AQP5 chimeric proteins conferred
with unphosphorylated amino acids at their PKA-tar-
get motif (
152SRRTS) in MDCK-II cells
Mileva Ratko Karabasil
Javkhlan Purevjav
1, Takahiro Hasegawa
1, Ahmad Azlina
1, Nunuk Purwanti
1,
2, Chenjuan Yao
1, Tetsuya Akamatsu
1, and Kazuo Hosoi
1
1Department of Molecular Oral Physiology, and
Institute of Health Biosciences, the University of Tokushima Graduate School, Tokushima, Japan
2Department of Periodontology and Endodontology,
Abstract : Three constructs having mutated PKA-target motif at152SRRTS of AQP5, an
exocrine type water channel, were prepared and fused to C-terminus of green fluores-
cence protein cDNA to examine the effects of blocking of phosphorylation at152SRRTS (a
consensus PKA-target motif of AQP5) on translocation or trafficking of the chimeric pro-
teins expressed in the Madin-Darby canine kidney-II (MDCK-II) cells. H-89 treatment
increased translocation of wild-type GFP-AQP5 to the apical membrane. All 3 mutant
molecules translocated 1.5 to 2 times more than the control wild-type GFP-AQP5. Col-
chicine but not cytochalasin B inhibited the translocation of wild-type GFP-AQP5. Present
results suggest dephosphorylation of this consensus sequence increase GFP-AQP5 translo-
cation, and that microtubules but not microfilaments are involved in this event. J. Med.
Invest. 56 : 55-63, February, 2009
Keywords : aquaporin-5, PKA-consensus sequence, trafficking
Received for publication January 5, 2009 ; accepted January 14,
2009.
Address correspondence and reprint requests to Prof. Kazuo
Hosoi, Department of Molecular Oral Physiology, Institute of
Health Biosciences, the University of Tokushima Graduate
School, Kuramoto-cho, Tokushima 770-8504, Japan and Fax :
+81-88-633-7324.
The Journal of Medical Investigation Vol. 56 2009
55

Page 2

canine kidney cells (MDCK) cells (14). In contrast,
Sidhaye, et al. reported a biphasic effect of cpt-cAMP
on AQP5 translocation in MLE-12 cells ; they showed
that short-term exposure to this second messenger
reduces membrane expression of AQP5, which is
subsequently recovered by long-term exposure (18).
However, these contradictive results may be mainly
due to apparent difference in the basal level of AQP5
expression in the cell membrane ; i.e., the mem-
brane expression of AQP5 was far greater in the
later study (18). On the contrary, Woo, et al. found
that AQP5 membrane targeting may not be regu-
lated solely by PKA phosphorylation and that it can
be regulated by more than one mechanism besides
that by cAMP dependent phosphorylation (19).
The present study was aimed to determine the
role of PKA phosphorylation site at the152SRRTS in
the regulation of AQP5 trafficking. For this purpose
we prepared 3 GFP-AQP5 constructs containing
mutated PKA target site where one or more of Ser
and Thr are replaced with Ala and Val, respectively
so that phosphorylation of this residue is restricted
or completely prohibited when they were expressed
in the host cells. Here, we showed that AQP5 can be
expressed on the plasma membrane of MDCK-II
cells even phosphorylation of PKA-target sequence
is restricted or prohibited, suggesting the existence
of PKA-independent trafficking.
MATERIALS AND METHODS
Reagents
Dulbecco’s modified Eagle’s medium (DMEM)
and cytochalasin B were obtained from Sigma-
Aldrich (St.Louis, MO). Lipofectamine 2000, Opti-
MEM I, and Alexa Fluor 594-conjugated streptavidin
were purchased from Invitrogen (Carlsbad, CA).
Sulfosuccinimidyl 6-(biotinamido) hexanoate (sulfo-
NHS-LC-biotin) was obtained from Pierce Biotech-
nology (Rockford, IL). Restriction enzymes, Xho I
and EcoR I, came from Takara Bio. Inc (Shiga, Ja-
pan). Ligation-Convenience Kit was from Nippon
Genetech Co., Ltd (Tokyo, Japan). pEGFP-C2 vector
(Cat. No. 6083-1) was obtained from Clontech (Palo
Alto, CA). pGEM-T Easy vector was bought from
Promega (Madison, WI). Transwell polycarbonate
filter culture chambers (Cat. No. 3413) were ob-
tained from Corning Costar (Lowell, MA). Micro
slide glass with 70 μm-thickness frame (Cat. No.
S 2193A7), and micro cover glasses were products of
Matsunami Glass (Osaka, Japan). H-89 came from
Seikagaku Corporation (Tokyo, Japan). Colchicine
was obtained from Wako Pure Chemical Industries
(Osaka, Japan). The QuickChange Site-Directed
Mutagenesis Kit was obtained from Stratagene (La
Jolla, CA).
PCR cloning and preparation of the pcDNA 3.1/Hy-
gro (+) construct containing wild-type (wt) AQP5
cDNAs
Total RNA was isolated from the submandibular
gland (SMG) of rats by using TRI Reagent. A full-
length AQP5 cDNA was synthesized by RT-PCR
with the SuperScript One-Step RT-PCR System in a
thermal cycler (Takara Thermal Cycler MP, Model
TP 3000). To a final volume of 25 μl, the following
components were mixed on ice : 12.5 μl of 2-times
concentrated reaction mixture, 5 pmol of each primer,
0.5 μl of a mixture of reverse transcriptase and Taq
DNA polymerase and 1 μg of template RNA. The RT
at 94? ?for 2 min to inactivate the enzyme and dena-
reaction (cDNA synthesis) was carried out at 45? ?
by PCR was next performed for 30 cycles, each cy-
annealing at 55? ?for 30 s, and extension at 72? ?for
AGTCACGAATCTCTGAGGTCTG-3’ (anti-sense),
which had Hind III and Xho I restriction sites (un-
derlined) at 5’ and 3’ end, respectively (20). The
AQP5 cDNA (1073 bp) obtained by RT-PCR were
cloned into the pGEM-T Easy vector, from which
the insert was next cut by Hind III and Xho I restric-
tion enzymes and subcloned into a multiple cloning
site of the pcDNA 3.1/Hygro (+) vector. The resul-
tant plasmid was termed as wild-type AQP5-Hygro
(AQP5(wt)-Hygro). This construct was used for
preparation of AQP5 genes with mutated PKA-con-
sensus sequence.
for 30 min. The reaction mixture was then incubated
ture the RNA/cDNA hybrid. The DNA amplification
cle consisting of denaturation at 94? ?for 15 s, primer
AAGGCACCATGAAAAA-3’ (sense) and 5’-CTCG-
1.5 min, followed by 1 cycle of extension at 72? ?for
5 min. The primer set used was 5’-AAGCTTCCCC-
Preparation of AQP5-Hygro having mutated PKA-
target motif at deduced amino acids 152-156
(
The AQP5-Hygro cDNAs having mutated PKA-
target motif at nucleotides 454-468 (5’-CCTCCAC-
CGACTCTCGCCGAACCAGCCCTGTGGGCTC-
3’), which corresponds to the amino acid sequence
152SRRTS, were prepared by using a QuickChange
site-directed mutagenesis kit. First, AQP5(wt)-Hygro
plasmid described above was used as a template,
152SRRTS) by site-directed mutagenesis
M. R. Karabasil, et al. Cellular roles of a PKA target motif of AQP5
56

Page 3

and artificial mutations were introduced. AQP5’s
with mutated nucleotide sequences were next sub-
cloned into pEGFP-C2. The following is the experi-
mental design employed : 3 primers, PK-1, PK-3,
and PK-5 were synthesized ; the sequence of PK-1
was 5’-CCTCCACCGACGCTCGCCGAACCGCCC-
CTGTGGGCTC-3’, which had 3 mutated nucleotides
(indicated by the underline), resulting in152ARRTA.
The sequence of PK-5 was 5’-CCACCGACTCTCG-
CCGAGTCAGCCCTGTGG-3’, which had 2 mu-
tated nucleotides, resulting in152SRRVS. Lastly, the
sequence of PK-3 was 5’-CCACCGACGCTCGCC-
GAGTCGCCCCTGTGG-3’, which had 5 mutated
nucleotides, resulting in152ARRVA. PK-2, PK-4, and
PK-6 were complementary DNAs of PK-1, PK-3,
and PK-5, respectively. The AQP5-Hygro plasmids
having these mutated sequences were designated
as AQP5(S152A/S156A)-Hygro, AQP5(T155V)-
Hygro, and AQP5(S152A/T155V/S156A)-Hygro,
respectively. In this nomenclature, S152A, T155V,
and S156A denote that serine (S) at position 152,
threonine (T) at position 155, and serine (S) at po-
sition 156 were replaced with alanine (A), valine (V),
and alanine (A), respectively. Plasmid, AQP5(S152A
/S156A)-Hygro was prepared directly by using
AQP5(wt)-Hygro as a template and the set of prim-
ers, PK-1 and PK-2. Similarly, AQP5(T155V)-Hygro
was synthesized by using the same plasmid and the
primer set, PK-5 and PK-6. Plasmid, AQP5(S152A/
T155V/S156A)-Hygro was prepared from AQP5
(S152A/S156A)-Hygro used as a template and the
set of primers, PK-3 and PK-4. The brief experimen-
tal procedure was as follows. To a final volume of 50
μl, the following components were mixed : 5 μl of
10 x reaction buffer, each set of primers (125 ng/3
μl), 1 μl of dNTP mixture, 50 ng of the template
plasmid, 1 μl of PfuTurbo DNA polymerase, and dis-
tilled water. The PCR was performed for 18 cycles,
30 s, primers annealing at 55? ?for 1 min, and ex-
each cycle consisting of denaturation at 95? ?for
ensure sufficient amplification. The PCR product
tension at 68? ?for 13 min. The PCR products were
for 1 h to digest methylated, parental DNA tem-
plates, and purified by electrophoresis on 1% agarose
gel. The plasmids with mutated AQP5 sequences
were next used for transformation and introduced
into E.coli JM109 ; and the sequences of cloned
plasmids were then verified by DNA sequencing.
The AQP5-Hygros having a mutated PKA-target
motif were subcloned into pEGFP-C2 as described
examined by electrophoresis on 1% agarose gel to
was then treated with Dpn I endonuclease at 37? ?
below.
Construction of chimeras, GFP-AQP5(wt) and
GFP-AQP5s having mutated PKA-target motif
For transfection experiments, we prepared the
constructs of chimeric genes of AQP5 and EGFP fol-
lowing the report which used GFP-AQP2 for study
of AQP2 trafficking (21). Complementary DNAs of
AQP5’s were connected to 3’-end of EGFP DNA in
the pEGFP-C2 plasmid because AQP5 connected to
N-terminus of EGFP is known to be expressed con-
stitutively at the plasma membrane (14) and is not
suitable for the trafficking study. Wild-type (wt) and
mutant AQP5 inserts in the Hygro vector were re-
covered by digestion with Xho I and EcoR I and sub-
cloned into the multiple cloning site of pEGFP-C2
plasmid. The resultant plasmids were termed GFP-
AQP5(wt), GFP-AQP5(S152A/S156A), GFP-AQP5
(T155V), and GFP-AQP5(S152A/T155V/S156A),
respectively. The sequence of AQP5 inserts in all
constructs was verified by use of an ABI PRISM
3100-Avant genetic analyzer (Applied Biosystems,
Foster City, CA).
Cell culture, transient transfection, cell surface
biotinylation, and observation under a confocal la-
ser scanning microscope
MDCK-II cells, kindly provided by Dr. Mikio
Furuse (Kobe University), were cultured in DMEM
supplemented with 10% fetal bovine serum (FBS),
100 U/ml penicillin, and 100 μg/ml streptomycin
sulfate at 37? ?in 5% CO2 ; they were supplemented
membranes 48 h prior to transfection. By use of
Lipofectamine 2000, the cells were transfected with
GFP-AQP5(wt) or GFP-AQP5 having mutated PKA-
target motif ; i.e., the cells were covered with a
mixture of 0.5 μg of DNA and 2 μl of Lipofectamine
2000 in a final volume of 100 μl of Opti-MEM I me-
dium per well as described by the manufacturer’s
protocol (Invitrogen). The transfection medium was
removed after 5.5 h of incubation and replaced with
fresh DMEM containing 10% FBS without antibiot-
ics. After cultivation for 24 h following transfection,
in which time point MDCK II cells had been cultured
for 3 days in total (22), polycarbonate membranes
containing cell monolayers were washed with phos-
phate-buffered saline (PBS) 3 times, fixed with 3%
paraformaldehyde for 20 min, washed, and incubated
with 50 mM NH4Cl for 15 min. For biotinylation,
with fresh medium every third day. For experi-
ments, MDCK-II cells were plated at the cell den-
sity of 6? 104cells/well on 6.5-mm polycarbonate
The Journal of Medical InvestigationVol. 56 February 2009
57

Page 4

cells were washed, and blocked with 1% BSA (frac-
tion V) in PBS for 1 h at room tenperature and re-
acted with cell-impermeable biotinylation reagent
(2 mM sulfo-NHS-LC-biotin) in PBS containing 1
mM MgCl2, 0.1 mM CaCl2, pH 7.4 at room tem-
perature for 40 min. After having been washed,
cells were incubated with 50 mM NH4Cl in PBS to
quench excess biotinylation reagent. Cells were
next reacted with Alexa Fluor 594-conjugate strep-
tavidin (1 : 200), washed, and mounted with Vec-
tashield mounting medium (Vector Laboratories,
CA, USA) on slide glasses with frames of 70-μm
thickness. They were covered with a coverslip, and
sealed with nail polish. Samples were stored in the
dark at 4? ?until they were examined under a con-
The samples were examined by scanning at hori-
zontal directions for 32 times at different vertical
planes (0.5 μm intervals). The cells expressing GFP-
AQP5 were selected and their transverse images
were saved. These data were examined to verify
whether AQP5 was completely targeted to apical cell
membrane. This can be determined by merging 2
pictures, GFP fluorescence (green) and biotin (red)
as the cells expressing AQP5 at cell membrane
show the yellow signal. For the sake of convenience,
the cell membrane facing to the medium will be re-
ferred to as the “apical membrane,” and the one
facing to the polycarbonate membrane, the “basal
membrane.” In all transfection experiments, 4-6
focal laser scanning microscope (Leica TCN NT,
Heidelberg, Germany) with excitations at 488 nm
(for FITC) and 568 nm (for Alexa Fluor 594).
wells were prepared and means?S.E. were calcu-
sin B
For examination of the effects of inhibitors of
PKA, microtubules, and microfilaments, cells cul-
tured on polycarbonate membranes for 48 h were
transfected with the chimeric gene, GFP-AQP5(wt)
as described above and cultured under a growth-ar-
rest condition (14), i.e., DMEM containing 0.5%
FBS, for 18 h. The medium was then replaced with
fresh medium containing 0.5% FBS and either 30
μM H-89 (14), 10 μM colchicine or 10 μM cytocha-
lasin B (17) ; and the cells were cultured for another
6 h (14). During 6 h-culture, medium was replaced
with fresh ones containing same inhibitors once at
3 h. The cells were then fixed, biotinylated, and ex-
amined under a confocal laser scanning microscope
as described above.
lated.
Treatment with H-89, colchicine, and cytochala-
Statistics
Data are expressed as means?SE. For statistical
analysis of results, Mann-Whitney U-test was ap-
plied.
RESULTS
AQP5 expression in MDCK-II cells
We examined physiological behavior of AQP5 in
MDCK-II cells, since these cells are known to pro-
vide a well-defined polarized cell model during culti-
vation on a polycarbonate membrane, and are widely
utilized in studies of epithelial cell polarity and in-
tracellular protein trafficking (23). Thus trafficking
and/or translocation of particular proteins, includ-
ing AQP5 protein (22), toward apical or basolateral
membranes can be studied in this model system.
MDCK-II cells were transfected with plasmids gen-
erating chimeric proteins for GFP-AQP5(wt) or
GFP-AQP5 proteins with mutations in PKA consen-
sus sequence, and cytoplasmic or apical-membrane
localization of these gene products was analyzed
(Figs. 1 and 3). In Fig. 1, typical MDCK-II cells ex-
pressing GFP-AQP5(wt) presenting prominent GFP
signals and indicating the localization of the AQP5
chimeric protein in them are shown (Fig. 1A). We
counted the number of cells that expressed GFP-
AQP5(wt) and those showing localization of GFP-
AQP5(wt) at apical membrane (Fig. 1B) ; mem-
brane localization of the chimeric protein was con-
firmed by yellow signal which verified its co-local-
ization with surface-labeled biotin.
At 6 h after transfection, GFP-AQP5(wt) was al-
ready expressed but it stayed in the cytosol until 12
h. At 24 h after transfection, 33% of the cells among
those expressing GFP-AQP5(wt) showed localiza-
tion at the apical membrane. This value increased
to 73% at 48 h after transfection.
Effects of H-89, colchicine, and cytochalasin B on
translocation of GFP-AQP5(wt)
We used several reagents known to stimulate or
inhibit cell dynamics in order to understand and con-
firm the machinery involved in the translocation of
GFP-AQP5(wt) in the MDCK-II cells.
Thus, at first, we determined whether PKA ac-
tivity was required for GFP-AQP5(wt) transloca-
tion. An experiment to examine the possible in-
volvement of microfilaments or microtubules was
also concomitantly performed. MDCK-II cells, plated
M. R. Karabasil, et al. Cellular roles of a PKA target motif of AQP5
58

Page 5

and transiently transfected as described above, were
treated with either H-89 (30 μM ; 13, 14), colchicine
or cytochalasin B (10 μM, each) at 18 h after trans-
fection ; they were cultured further for 6 h under
a growth-arrest condition as described above. In
good accordance with our previous finding (17), col-
chicine, an inhibitor of cytoskeleton assembly, in-
hibited strongly the translocation of the chimeric
proteins ; i.e., the number of cells expressing the
GFP-AQP5(wt) at their apical membrane was only
38% of that of the non-treated control cultures (Fig.
2), suggesting that translocation of AQP5-bearing
vesicles, required the microtubule system. A micro-
filament inhibitor, cytochalasin B had no or only
little effect on the trafficking of the GFP-AQP5(wt)
(81% of control).
On the other hand, percentage of the cells that
expressed GFP-AQP5(wt) at their apical membrane
among GFP-AQP5(wt) positive cells increased in the
presence of the PKA inhibitor H-89, as compared
with that for the non-treated cells. The percentage
increase in apical translocation by H-89 was 1.5
times of that for non-treated control (Fig. 2).
The above results suggest that the translocation
of AQP5-bearing vesicles towards the apical mem-
brane was increased by decreasing the phosphoryla-
tion of Ser/Thr of AQP5. From these initial observa-
tions, we hypothesized that the phosphorylation it-
self is probably not crucial for membrane expression
of AQP5. The phosphorylation target site blocked by
H-89 may be amino acid 152-156 (SRRTS), since
this is a PKA-target motif found in AQP5 (3). We,
therefore constructed 3 GFP-AQP5 mutants having
mutated consensus PKA-target motifs to examine
the role of this PKA-target motif in membrane ex-
pression of AQP5.
Effects of mutation in PKA-target motif (
on translocation of the AQP5 chimeric proteins
Twenty-four hours after transfection of MDCK-II
cells with GFP-AQP5 (S152A/S156A), GFP-AQP5
(T155V) or GFP-AQP5 (S152A/T155V/S156A),
152SRRTS)
A A
B B
Fig. 1
expressing chimeric proteins were observed under the confocal
microscope. A, Vertical pictures (x-z vertical sections ; see MA-
TERIALS AND METHODS section for detail methods) of cells
in typical cultures at 6, 12, 24, and 48 h post-transfection, express-
ing the wt chimeric protein, are shown. The merge technique was
used to visualize the co-localization of biotin/Alexa Fluor 594-
conjugated streptavidin (red) and GFP signal of GFP-AQP5
(green) in the apical cell membrane which appears as yellow ar-
eas in cells. Bar, 20 μm. B, Confocal images shown in “A” were
assessed, and the histogram showing the percentages of the cells
expressing GFP-AQP5(wt) protein at their apical membrane
among GFP-positive cells at different post transfection times is
presented. The number of sample wells analyzed was 4 to 6.
AQP5 expression in polarized MDCK-II cells. Cultures
A A
B B
Fig. 2
apical translocation of GFP-AQP5(wt). MDCK-II cells were trans-
fected as described in the text. Eighteen hours after transfection,
the cells were incubated under growth-arrest conditions (incu-
bation in the medium containing 0.5% FBS) for the next 6 h in
the presence or absence of either H-89 (30 μM), colchicine (10
μM) or cytochalasin B (10 μM) and examined under the confocal
laser scanning microscope. A, Vertical pictures of GFP-AQP5
positive cells. Bar, 20 μm. B, Histogram showing the percentage
of the cells expressing GFP-AQP5(wt) at apical membrane is
Effects of H-89, colchicine, and cytochalasin B on the
presented as in Fig. 1. *p?0.05, significantly different from the
GFP-AQP5(wt) group. NS, not significantly different from the
GFP-AQP5(wt). The number of sample wells analyzed was 4 to 6.
The Journal of Medical InvestigationVol. 56 February 2009
59