(Received for publication, December 2, 1996, and in revised form, January 10, 1997)
From the Consorzio Mario Negri Sud, Istituto di Ricerche
Farmacologiche "Mario Negri", Santa Maria Imbaro, 66030, Italy and
Istituto di Biologia Generale dell'Università
degli Studi e Centro per lo Studio delle Cellule Germinali del CNR,
Siena, 53100, Italy
G protein-coupled receptor kinase 4 (GRK4)
presents some peculiar characteristics that make it a unique member
within the GRK multigene family. For example, this is the only GRK for
which four splice variants (GRK4, -
, -
, -
) have been
identified. We developed a simple assay to study kinase activity, and
we found that GRK4
, but not GRK4
, -
, and -
, was able to
phosphorylate rhodopsin in an agonist-dependent manner.
GRK4
kinase activity was inhibited by
Ca2+/calmodulin (CaM) (IC50 = 80 nM), and a direct interaction between GRK4
and
Ca2+/CaM was revealed using CaM-conjugated Sepharose 4B.
The other three GRK4 isoforms did not interact with CaM in parallel
experiments. The present investigation also aimed to define cellular
and ultrastructural localization of GRK4. A substantial expression of
GRK4 mRNA was only found in testis and in the spermatogonia cell
line GC-1 spg. Specific GRK4 immunoreactivity was only found on sperm
membranes, and immunochemical and ultrastructural analyses showed that
it is associated to the acrosomal membranes and to the outer
mitochondrial membranes. GRK4
was the only detectable isoform in
human sperm. We concluded that: i) only GRK4
can phosphorylate
rhodopsin and that this activity is inhibited by CaM; ii) the other
three isoforms do not phosphorylate rhodopsin and do not interact with
CaM; and iii) the association of GRK4 with highly specialized sperm
organelles, which are essential for fertilization, strongly indicates
that this kinase is involved in this process.
The G protein-coupled receptor family includes many receptors that
bind a large array of different molecules, ranging from photons,
neurotransmitters, and neuropeptides to autacoid substances, hormones,
and immunomodulators, acting through different intracellular second
messengers. For a number of G protein-coupled receptors, a rapid,
drastic, and reversible loss of responsiveness has been shown to occur
upon exposure to agonists (homologous desensitization). Two types of
proteins play a major role in determining homologous desensitization of
G protein-coupled receptors, G protein-coupled receptor kinase
(GRK)1 (1, 2), which phosphorylates
agonist-occupied receptors, and its functional cofactor, ()-arrestin
(3). The multigene family of GRKs consists so far of six members named
GRK1 to -6 (1, 2). Based on sequence homology these six GRK subtypes are classified into three subfamilies. GRK1 is alone in the first (rhodopsin kinase subfamily), GRK2 and -3 form the second
(
-adrenergic receptor kinase (
ARK) subfamily), while GRK4, -5, and -6 constitute the third (GRK4 subfamily). The mechanism of action
of these kinases has been extensively investigated for GRK1 (also known
as rhodopsin kinase), GRK2 (
ARK1) and GRK3 (
ARK2). Receptor
activation triggers translocation of GRK from cytosol to plasma
membranes, where it phosphorylates agonist-occupied receptors. GRK5 and
-6 likely regulate receptor substrate in a similar manner although only few studies are presently available to support this view. For GRK2 and
-3, the mechanism of transient translocation to membrane involves
binding to
subunits of G proteins (1).
GRK4 presents some peculiar characteristics with respect to other GRKs.
For example, this is the only GRK for which four splice variants
(GRK4, -
, -
, -
) have been identified (1, 4). Moreover,
unlike GRK2, -3, -5, and -6, which are widely distributed in different
tissues and cells, the expression of GRK4 is extremely localized, and
substantial amounts of GRK4 mRNA were found in testis only (4-6).
The unique site of expression makes GRK4 similar to rhodopsin kinase
(GRK1), which is expressed only in the retina where it regulates
phototransduction. Based on these observations, it was suggested that
GRK4 may have a strong substrate selectivity and that the receptor
substrate, which is as yet unknown, may be located in testis (4).
Despite these properties that make GRK4 a unique member within this
family, this receptor kinase was poorly investigated, and all the
studies reported so far were performed in heterologous expression
systems. Premont et al., (6) have recently shown that all
four recombinant GRK4 splice variants, expressed in COS7 cells, are
palmitoylated and have desensitizing activity when luteinizing
hormone/chorionic gonadotropin (LH/CG) receptor was cotransfected.
In the present study, a simple assay to study kinase activity of GRK4
was developed and used for extensive characterization. We found that
GRK4, but not the other isoforms, is able to phosphorylate rhodopsin, and this effect is inhibited by Ca2+/calmodulin
(CaM) through direct binding. We also documented the expression of GRK4
in spermatozoa and germinal cells, where it is associated to
mitochondrial and acrosomal membranes.
Macroscopically normal human tissues, obtained from surgically excised samples, and organs removed from adult rat were rapidly frozen in liquid nitrogen. Human testis from biopsies and spermatozoa from volunteers were fixed as described below. Bovine spermatozoa were purchased from SemenItaly S.r.l. Montesilvano (PE, Italy). The mouse testicular cell lines TM3 (Leydig cells), TM4 (Sertoli cells), and GC-1 spg (spermatogonia) were from ATCC, and they were grown according to the manufacturer instructions.
Cloning, Expression and Kinase AssaysPCR cloning and analysis of splice isoforms were as described (4, 7). The full-length cDNAs for the GRK4 isoforms were reconstructed, subcloned in the eukaryotic expression vector pCMV5, and used for transfection (4, 8). COS7 cells, grown to ~70% confluence, were transfected as described previously (4). Human embryonic kidney cells, HEK293, were transfected by the calcium phosphate method. Nearly confluent cells were transfected with 25-50 µg/150-mm Petri dish and grown for 72 h in Dulbecco's modified Eagle's medium with 10% fetal calf serum. Cytosol and membrane were prepared as in Parruti et al. (7). For GRK4 partial purification, the post-nuclear preparation in buffer A (20 mM HEPES, pH 7.2, 5 mM EDTA, and 20 mM NaCl supplemented with a mixture of protease inhibitors) was added with 1% Triton X-100, extracted 1 h at 4 °C, and then centrifuged for 1 h at 100,000 × g at 4 °C, and the supernatant was filtered through a 0.2-µm filter and loaded on a S-Sepharose ionic exchange column (Pharmacia Biotech Inc.). Proteins were eluted with a NaCl gradient from 100 to 1,000 mM, and kinase activity eluted at ~ 500 mM NaCl. The ability of partially purified GRK4 to phosphorylate the rod outer segments (ROS) was assessed as described (9). Briefly, partially purified GRK4 (0.8 µg of proteins) plus 13 µg of rhodopsin were incubated in a 60-µl total volume for 30 min at 30 °C in the presence of light. In appropriate samples, heparin, calcium, CaM, or CaM inhibitor (9) were added as indicated in the figure legends.
CaM-Sepharose 4B Binding AssayCaM-Sepharose 4B binding assay was performed as described previously (9) with minor modifications. CaM-Sepharose 4B or unconjugated Sepharose 4B gel (60 µl) was resuspended in binding buffer containing various components in a total volume of 850 µl. Crude cytosolic proteins (150 µg) from HEK293 transfected with the four GRK4 splice variants were used. Samples with calcium contained 1 mM Ca2+, as suggested by the manufacturers of CaM-Sepharose 4B, while in samples without calcium, 6 mM of EGTA was added to chelate the calcium trace amount present in the mixture. Samples were incubated for 1 h on a rotator at 4 °C and then the unbound materials removed by twice repeating centrifugation and washes with 1 ml of ice-cold binding buffer. SDS sample buffer was added to the final gel pellet, heated to 95 °C to detach gel-bound materials, and followed by electrophoresis on 10% SDS-polyacrylamide gel electrophoresis. Proteins were transferred onto nitrocellulose paper and immunoblotted with appropriate antibodies.
AntiserumA fusion protein of 6-His and the N-terminal
(amino acids 52 to 114) region of GRK4 was constructed by inserting
the EcoRI-RsaI fragment in pBluescript SK+, and
in turn, in pQE9 using the restriction endonucleases BamHI
and HindIII (the construct was called Ab-pQE). Protein
fragment was expressed by induction of Escherichia coli strain M15[pREP4] bearing the Ab-pQE with 1 mM
isopropyl-1-thio-
-D-galactoside and grown for 4 h
at 37 °C. The fusion protein was purified on Ni-NTA-agarose resin
under denaturing conditions following the manufacturer instructions
(Qiagen Inc.). Antiserum was raised in two rabbits using this purified
protein as immunogen. Western blot and immunoelectron microscopy was
performed using the affinity-purified antibody (N-Ab) at the indicated
concentrations.
Northern and Western blot analysis were performed as described previously (4, 7). The polyclonal antibody I-20, bought from Santa Cruz Biotechnology, Inc. and raised against the epitope corresponding to amino acids 478-497 of GRK4 (clone IT11A), was used according to the manufacturer instructions. This antibody was reported to be specific for GRK4 with respect to the other GRKs and to recognize GRK4 from mouse, rat, bovine, and human. The same protocol was used with the antibody N-Ab generated by us (see above). The immunoblotting was performed using 0.1 µg/ml of I-20 or 2 µg/ml of N-Ab and developed with alkaline phosphatase-conjugated goat anti-rabbit IgG (1:5000) and 5-bromo-4-chloro-3-indolyl-phosphate and nitro blue tetrazolium.
Preparation of Subcellular FractionsSubcellular fractionation was performed according to Garcìa-Higueras et al. (10) with minor modifications. Frozen bovine sperm were thawed and rapidly pelleted at 800 × g for 20 min. The pellet was resuspended in 4 volumes of cell lysis buffer plus 250 mM sucrose and homogenized, on ice, with 10 strokes of a motorized teflon pestle. All the subsequent steps were carried on at 4 °C. After a low speed centrifugation at 800 × g for 10 min, the pellet was resuspended in cell lysis buffer, sonicated with 4 pulses of 15 s, and treated with 40 µg/ml of DNase I for 1 h at 37 °C (nuclear-enriched fraction). The supernatant (post-nuclear preparation) was centrifuged at 3,000 × g for 10 min to obtain a plasma membrane pellet. Centrifugation of the supernatant at 10,000 × g for 20 min provided the crude mitochondrial pellet. The supernatant was centrifugated at 300,000 × g for 30 min to obtain the microsomal membranes (pellet) and the cytosol (supernatant). Different subcellular fractions were characterized by electron microscopy (data not shown).
ImmunocytochemistryElectron microscopy analysis of
biopsies from human testis, human healthy spermatozoa, and spermatozoa
affected by the genetic defect so called "round head" (11) was
performed as described (12). For fluorescent microscopy, fertile and
round head human spermatozoa were washed twice in phosphate-buffered
saline (PBS), smeared on glass slides, air-dried, fixed for 10 min in
methanol at 20 °C, extracted for 5 s in acetone at
20 °C, redried, washed three times in PBS, and first treated with
PBS, 5% normal goat serum, 1% bovine serum albumin for 20 min.
Spermatozoa were incubated overnight at 4 °C in polyclonal antibody
(rabbit IgG) I-20, diluted 1:40, then washed three times in PBS, 0.1%
bovine serum albumin, and incubated, respectively, in fluorescein
isothiocyanate goat anti-rabbit IgG antibody (Calbiochem, La Jolla,
CA). Finally, the glass slides were washed three times in PBS, mounted
in PBS, glycerol, 1:10, containing 5% propyl-gallate, and observed in a light microscope Leitz Aristoplan (E. Leitz, Rockleigh, NJ) equipped
with fluorescence optics.
The alignment of GRK4 with the
closest members of this kinase family GRK5 and -6 revealed two major
gaps in GRK4 sequence located near the N- and C-terminals. When this
investigation was started, only two isoforms of GRK4 generated by an
alternatively spliced sequence in the N-terminal region were known (4).
To test the possibility that the gap near the C-terminal was due to the
existence of additional isoforms, a pair of primers was used to amplify
this region from human testis RNA. PCR reaction generated two
amplification products that differed only by the presence or absence of
an insert of 138 base pairs/46 amino acids, and this sequence was
identical to that recently reported by Premont et al. (6).
The identification of two alternatively spliced sequences at the
C-terminal of GRK4 indicates the existence of four splice variants of
this kinase that, according to Premont et al. (6), are named
GRK4, -
, -
, and -
(Fig. 1). GRK4
was
previously known as GRK4B (4) or IT11 (5) while GRK4
corresponds
to GRK4A (4).
Expression and Characterization of GRK4 Kinase Activity
The
four GRK4 isoforms were reconstructed and expressed in COS7 and HEK293
cells, and the expression was determined by immunoblot (Fig. 1 and data
not shown). In COS7 cells, the four isoforms were totally
membrane-associated, while in HEK293 a significant amount (~50% of
total) was recovered in the cytosolic fraction. A crude cytosolic
preparation from transfected HEK293 was used to phosphorylate ROS, but
the effect was neglegible, probably due to the low specific activity of
this preparation. To obtain a higher specific activity, GRK4 isoforms
were partially purified by one step of ionic exchange chromatography
using an S-Sepharose column, which enriched the kinase by about
ten-fold. Using this partially purified preparation, we found a robust
phosphorylation of rhodopsin by GRK4 while GRK4
, -
, and -
did not phosphorylate ROS, as compared with a similar preparation from
HEK293 cells transfected with the vector alone (Fig. 1). The amount of
the four isoforms used in these experiments was similar, as confirmed by immunoblot (not shown). Phosphorylation of rhodopsin by GRK4
was
agonist (light)-dependent, and this effect was similar to that of GRK2 and GRK5, which were used as positive controls (Fig. 2). Phosphorylation of rhodopsin by GRK4
was
inhibited by heparin (Fig. 2), which is a known selective inhibitor of
GRKs, and the IC50 value was 770 ng/ml.
Inhibition of GRK4
We have recently shown
that the calcium-binding protein CaM can inhibit GRK5 through direct
binding (9). The possible regulation of GRK4 by CaM was, therefore,
investigated. Phosphorylation of rhodopsin by GRK4 was potently
inhibited by CaM in a calcium-dependent manner (Fig.
3). This inhibition was proved to be
CaM-dependent since it was completely reverted by the CaM
inhibiting peptide (CaMBd) (9) derived from the
CaM-dependent protein kinase II (Fig. 3) The inhibitory
activity of CaM on GRK4
(IC50 = 80 nM) was
similar to that on GRK5 (IC50 = 40 nM) and
~25-fold more potent than that on GRK2 (IC50 ~2
µM) (9). One mechanism by which Ca2+/CaM
could inhibit GRK4
was by direct binding, as previously shown for
GRK5 (9). To address this possibility, in vitro binding assays were performed using CaM-conjugated Sepharose 4B gel and unconjugated Sepharose 4B as negative control (Fig. 4).
GRK5 was used as a positive control for the binding assay since it has previously been shown to bind CaM-Sepharose (9). Both GRK4
and GRK5
bound to CaM-Sepharose 4B in a Ca2+-dependent
manner. Based on internal standards, we estimated that >80% of the
total amount of GRK4 and GRK5 added to the binding assay was bound by
CaM-conjugated gel in the presence of calcium. In the absence of
calcium, no detectable binding of GRK4 and GRK5 to CaM was found.
Neither of them bound to the unconjugated Sepharose 4B gel (Fig. 4 and
data not shown).
The effect of Ca2+/CaM on GRK4, -
, and -
could not
be tested on the phosphorylation assay as we have observed that these isoforms did not phosphorylate ROS (see Fig. 1). We then measured the
binding of the four isoforms to CaM-conjugated gel (Fig. 4). Experiments were performed in parallel, and 150 µg of cytosolic proteins from HEK293 transfected with the four isoforms were incubated with CaM-conjugated Sepharose 4B gel in the presence or absence of
Ca2+. The level of expression of each isoform was the same,
as documented by immunoblot (Fig. 4). We observed a substantial and
Ca2+-dependent binding of GRK4
to CaM, while
only trace amounts of GRK4
, -
, and -
were bound (Fig. 4).
The mRNA
expression of GRK4 was investigated on a large variety of human and rat
tissues and cell lines by Northern blots. As expected, a single
mRNA band of about 2.5 kilobases was detected on RNA from human,
bovine, and rat testis (Fig. 5 and data not shown), and
this was used as positive control in all the experiments. We examined
i) the following human tissues: liver, heart, lung, adipose tissue,
brain cortex and skeletal muscle; ii) the following cells isolated from
human peripheral blood: monocytes, monocytes activated by adherence,
lymphocytes, and polymorphonuclear leukocytes; iii) the following human
cell lines: Jurkat (lymphoid leukemia cells), HL60 (promyelocitic
leukemia cells), K562 (myeloid leukemia cells), MRC5 (embryonic lung),
Hep G2 (hepatoma cells), IMR32 (neuroblastoma cells), MCF7 (breast
adenocarcinoma cells), SW626 (ovarian adenocarcinoma cells), O143
(osteosarcoma cells), and MDA-MB (mammary carcinoma cells); iv) the
following rat brain regions: cortex, n. accumbens, striatum,
hypothalamus, hippocampus, brainstem, and olfactory bulbs; and v) the
following rat tissues and cells: PBL, stomach, intestine, prostate,
seminal vesicles, uterine tubes, ovary, adrenal gland, adenohypophysis,
and C6 (glioma cell line). In all these tissues and cells, no
expression of GRK4 could be documented (not shown), strongly supporting
the idea that GRK4 mRNA is expressed exclusively in testis.
Different cell types that are present in testis were analyzed by
Northern blot (Fig. 5). GRK4 mRNA was not expressed in Sertoli
cells isolated from rat testis (13) (not shown), in the mouse Sertoli
cells TM4, and in the Leydig cells TM3 while it was abundantly
expressed in the spermatogonia GC-1 spg (Fig. 5), suggesting that sperm are the cellular site of expression of GRK4. According to this hypothesis, the mRNA expression was lower in testis from immature rat (20 days-old) as compared with testis from adult mature animal (60 days-old) (Fig. 5). As previously observed by Premont et al. (6) in some tissues, we also found an additional hybridization band at
~4 kilobases on GC-1 spg (Fig. 5). The meaning of this additional
band remains to be defined.
Expression of GRK4 in Spermatozoa
To investigate cellular
localization within testis in a first set of experiments, we used the
antibody I-20 obtained from Santa Cruz Biotechnology. To test the
selectivity with respect to the other members of the GRK family, the
antibody was tested on a blot containing: GRK4, GRK1 partially
purified from bovine retina, and recombinant GRK2, GRK3, GRK5, and GRK6
overexpressed in Sf9. Only GRK4 was recognized by this antibody (not
shown), thus confirming the selectivity reported by the manufacturer. The expression of GRK4 was then studied on cytosolic and total membrane
fractions from several tissues including rat testis, ovary, and
adenohypophysis, human testis, and bovine spermatozoa. A strong
immunoreactive band at Mr ~60 KDa was
detected on sperm membrane only (Fig. 6a).
This band was totally abolished by preincubation of antibody with the
antigen peptide (Santa Cruz Biotechnology, CA) thus confirming its
specificity (Fig. 6b). Since GRK4 is expressed at high level
in mature spermatozoa that are mainly in the deferent duct, the
concentration of GRK4 in testis is not enough to be readily detectable
by Western blot. These results confirm that GRK4 localization is highly
specific and indicate that sperm membranes are the site where this
kinase is expressed.
Subcellular fractions of sperm, separated by differential centrifugations as described under "Experimental Procedures," were examined to see whether GRK4 is associated to specific organelles. The GRK4 immunoreactivity was substantially associated with enriched mitochondrial and microsomal fractions with no immunoreactivity observed on enriched nuclear fraction, cytosolic proteins, and plasma membrane (Fig. 6c).
Ultrastructural Localization of GRK4Experiments were
performed to investigate ultrastructural localization of GRK4 using the
same antibodies (I-20). As revealed by electron microscopy,
immunocytochemistry on Lowicryl-embedded free spermatozoa, or on
testicular tissue, GRK4 was present in spermatozoa, spermatids, and
previous stages of spermatogenesis. In mature spermatozoa (Fig.
7), frequent gold particles are localized at the level
of the acrosomes and of the mitochondria. In the acrosome, the
localization seems to be concentrated on the outer and inner membranes,
mostly on their cytoplasmic surface and only a few of them are
sometimes extended to the acrosomal content (Fig. 7a). After
the acrosomal reaction, when the anterior nuclear region is covered
only by the inner acrosomal membrane, the particles show a monolayered
distribution, localized between membrane and nucleus, in the thin
subacrosomal layer (Fig. 7b). In round head spermatozoa, a
well known genetic damage involving lack of acrosome and consequently
of acrosomal enzymes (11) which causes infertility, the sperm head is
completely unlabeled (not shown). In the mitochondria of the ejaculated
spermatozoa, gold granules appear more concentrated on the outer
mitochondrial membrane (Fig. 7c), sometimes present also on
the cristae, both on normal (Fig. 7, c and d) and
round head (not shown) spermatozoa. Fluorescent microscopy confirms this view, showing positive labels in the acrosomal region (Fig. 7,
e-h), including equatorial segment (Fig. 7, e
and f), as well as in the midpiece of the tail, mainly close
to the centriole (Fig. 7, e, f, i, and
l). In the testicular tissues (Fig. 8), immature germinal cells show gold granules concentrated in the mitochondrial membrane during the whole spermatogenesis (Fig. 8d). Clusters of particles are also present on the membranes
of cytoplasmic vesicles (endoplasmic reticulum) (Fig. 8b)
and in the Golgi complex (Fig. 8a). The acrosome, at the
beginning of its formation, appears almost devoid of particles (Fig.
8c); when it starts to be moulded and condensed in older
spermatids, groups of particles are seen to migrate from the
cytoplasmic membrane to small areas of the outer acrosomal surface
(Fig. 8e). Possibly they are carried out by small
cytoplasmic vesicles. The axoneme and the nuclei are consistently not
labeled.
Analysis of the Four GRK Isoforms by a Common Antibody
After
this set of experiments, we observed that the antibody I-20, raised
against an epitope spanning the site where the splicing at the C
terminus is located, did not recognize GRK4 and -
, which contain
this insert, while it recognized recombinant GRK4
and -
(data not
shown and see Ref. 6). A second antibody (N-Ab) was then generated,
raised against a sequence (amino acids 52-114 of GRK4
) common to
all the four isoforms. This antibody N-Ab, which is able to recognize
all the four recombinant GRK isoforms (Fig. 9), revealed
only one major band of immunoreactivity on human sperm membranes. Based
on gel migration, this corresponded to GRK4
(Fig. 9).
Electron microscopy immunocytochemistry on sperm using N-Ab confirmed the results obtained with I-20. GRK4 was specifically recognized on the inner and outer acrosomal membranes and on the outer mitochondrial membranes and sometimes on the cristae (Fig. 9).
In this study, a simple method to assess the kinase activity of
GRK4 was developed, which is based on transient expression on HEK293
cells, partial purification, and phosphorylation of ROS. We
consistently found that GRK4 was able to phosphorylate rhodopsin.
This is different from previous results from our group (4) and Premont
et al. (6), which indicated that GRK4 was not able to
phosphorylate this receptor. The apparent discrepancy is likely due to
the different expression systems and kinase isoforms used in these
studies. Similar to other GRKs, phosphorylation of rhodopsin by GRK4
was strictly agonist (light)-dependent and inhibited by
heparin. An agonist-dependent phosphorylating activity of
GRK4
was previously shown using, as a substrate, the purified
2-AR reconstituted on vesicles containing 5% PIP2
(6).
GRK4 kinase activity was potently inhibited by Ca2+/CaM.
This effect was strictly dependent on Ca2+ and was
prevented by the CaM inhibitor CaMBd. A direct interaction between
GRK4
and Ca2+/CaM was revealed using CaM-conjugated
Sepharose 4B. In a recent study, we have documented a similar
inhibition of GRK5 by Ca2+/CaM (9). Two other groups have
shown that in the presence of Ca2+, GRK1 is inhibited by
the photoreceptor-specific recoverin through direct binding (14, 15).
Recoverin is a member of the family of neuron-specific calcium sensor
proteins named neuronal calcium sensors (NCS) and several other members
of this family were also able to inhibit GRK1 (16). Since CaM and NCS
are all Ca2+ sensors, these studies suggest that they act
as functional analogues in mediating the regulation of different GRK
subtypes by Ca2+. This mechanism is, however, highly
selective with respect to the GRK subtypes. While GRK1, but not GRK2,
is regulated by recoverin and other NCS, GRK4, and GRK5 that belong to
the GRK4 subfamily are potently inhibited by CaM. CaM had little or no
effect on members of other GRK subfamilies. The effect of
Ca2+-sensor proteins appears to be rather general, and the
GRK subtype selectivity indicates that different kinases can be
specifically regulated in different target tissues.
The properties of the four GRK4 splice variants were investigated, and
we document for the first time some substantial differences between
these isoforms. It was found that only GRK4, but not GRK4
, -
,
and -
, was able to phosphorylate rhodopsin. Since all four isoforms
are able to desensitize some G-coupled receptors, as suggested by the
inhibition of LH/CG-induced cAMP accumulation in cotransfection
experiments (6), these results indicate that the four GRK4 isoforms may
have different receptor substrate specificity. Additionally, the
interaction of GRK4 isoforms with CaM was investigated, and it was
found that only GRK4
, but not GRK4
, -
, and -
was able to
bind to CaM in a calcium-dependent manner. We speculate that the long kinase isoform GRK4
, which contains the two inserts at
the N and C terminus, closely resembles GRK5 and, therefore, shares
common properties with it, including ability to phosphorylate rhodopsin
and inhibition by CaM through direct binding. By contrast, the other
isoforms may have distinct properties such as substrate specificity or
regulation by phospholipids, but this remains to be demonstrated. The
insert located at the N-terminal domain was suggested to contain the
sequence of homology with gelsolin and is postulated to be the binding
site for PIP2, which seems critical for kinase activity
(17). This can explain our finding of lack of phosphorylating activity
of GRK4
and GRK4
, which lack the N-terminal insert and hence the
putative binding site for PIP2. However, we also observed
that GRK4
, which contains the N-terminal insert but lacks the one
located at the C-terminal, was also unable to phosphorylate rhodopsin,
strongly indicating that both inserts play a key role for the
functionality of these kinases.
The present study also aimed to determine cellular and ultrastructural localization of GRK4. The expression of this kinase is restricted to the testis, as confirmed by Northern blot analysis of a large variety of tissues and cells. Traces of specific mRNA could be found in other tissues (i.e. brain cortex) using the extremely sensitive reverse transcription-PCR, indicating that GRK4 is expressed on sites other than testis (4). Whether these very low levels of mRNA, detectable only by PCR, are generated by very few copies, widely distributed without any functional meaning, or by significant amount of mRNA localized in selected cells or regions, remains to be defined.
Using polyclonal antibodies selective for GRK4, we documented for the
first time that GRK4 is selectively expressed in spermatozoa and
germinal cells, where it is associated to membranes of intracellular organelles. Using an antibody that recognized all the GRK4 splice variants, we investigated for the first time the relative expression of
native GRK4 isoforms. We found that only one isoform, GRK4, is
detectable in human sperm membranes. This finding is unexpected since,
based on the analysis of mRNA expression in testis, a higher expression of the longest form was postulated (6). The lack of the
other three isoforms may indicate that: i) GRK4
, -
, and -
are
expressed at low levels, not detectable by our immunoblot protocol; ii)
they are not expressed in mature sperm, but rather in germ cells; iii)
they are highly instable proteins; and iii) they are not expressed at
all.
Only another member of the GRK family, GRK3, and its cofactor
-arrestin2 were found in the midpiece of rat sperm, colocalized with
odorant-like receptors that are their likely substrates (18). GRK3 and
-arrestin2 however are widely distributed in different organs and
cells, while for GRK4, sperm represents the only relevant site of
expression. We, therefore, hypothesize that the physiological role of
GRK4 is highly specific and strictly related to the function of
spermatozoa.
Morphological analysis showed that GRK4 is present in germinal cells throughout their whole life. In the spermatozoa, it was associated to the two acrosomal membranes and to the outer mitochondrial membranes. In the immature germinal cells, GRK4 was present in the membranes of cytoplasmic vesicles (endoplasmic reticulum) and in the Golgi complex, while the acrosome, at the beginning of its formation, appears almost devoid of particles. Association of GRK4 to membranes of specialized intracellular structures such as mitochondria, acrosome, and vesicles is a totally unpredicted finding which substantially expands the general view that GRKs are associated to plasma membranes where they phosphorylate and regulate surface G-coupled receptors. Only one previous study on rat liver (10) reported the association of a fraction (39%) of GRK2 with endoplasmic reticulum while the majority of this kinase in the same cells was found in the cytosol (43%) or associated with the plasma membranes (18%).
Our results also raise the question of the possible function of GRK4 on these intracellular structures. Acrosome is a highly specialized vesicle, localized on top of the sperm head, and involved in fundamental steps for fertilization, including capacitation, acrosome reaction, and interaction with mature eggs. At the moment, a possible role of GRK4 in these processes can only be postulated. There is evidence indicating the role of a variety of G-coupled receptors including PAF, muscarinic, odorant-like, atrial natriuretic peptide receptors in sperm motility and acrosome reaction (19, 20). The presence of orphan heptahelical receptors in sperm was also reported (21). All these receptors can be potentially regulated by phosphorylation from kinases of the GRK family.
In conclusion, the present paper provides several new clues on GRK4. We
document that only GRK4 can phosphorylate rhodopsin, and this effect
is inhibited by Ca2+/CaM. Unlike GRK4
, the other three
isoforms do not phosphorylate rhodopsin and do not interact with CaM,
indicating that they have different properties which are, at the
present, a matter of further investigation. The association of GRK4
with highly specialized sperm organelles, which are essential for
fertilization, strongly indicates that this kinase is involved in this
process.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) X98118[GenBank], X97879[GenBank], X97880[GenBank], and X97881[GenBank].
We are indebted to R. J. Lefkowitz for GRK5 and GRK6 baculoviruses; to H. LeVine, III, for a number of reagents; to A. Filippini for rat Sertoli cells; to T. T. Chuang for GRK2, 3, 5, 6 Sf9 expression; to Cosmo Rossi for help in developing N-Ab; to R. Bertazzi for expert assistance in the preparation of figures.