G Protein-coupled Receptor Kinase GRK4
MOLECULAR ANALYSIS OF THE FOUR ISOFORMS AND ULTRASTRUCTURAL LOCALIZATION IN SPERMATOZOA AND GERMINAL CELLS*

(Received for publication, December 2, 1996, and in revised form, January 10, 1997)

Michele Sallese , Stefania Mariggiò , Giulia Collodel Dagger , Elena Moretti Dagger , Paola Piomboni Dagger , Baccio Baccetti Dagger and Antonio De Blasi §

From the Consorzio Mario Negri Sud, Istituto di Ricerche Farmacologiche "Mario Negri", Santa Maria Imbaro, 66030, Italy and Dagger  Istituto di Biologia Generale dell'Università degli Studi e Centro per lo Studio delle Cellule Germinali del CNR, Siena, 53100, Italy

ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES


ABSTRACT

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 (GRK4alpha , -beta , -gamma , -delta ) have been identified. We developed a simple assay to study kinase activity, and we found that GRK4alpha , but not GRK4beta , -gamma , and -delta , was able to phosphorylate rhodopsin in an agonist-dependent manner. GRK4alpha kinase activity was inhibited by Ca2+/calmodulin (CaM) (IC50 = 80 nM), and a direct interaction between GRK4alpha 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. GRK4gamma was the only detectable isoform in human sperm. We concluded that: i) only GRK4alpha 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.


INTRODUCTION

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, (beta )-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 (beta -adrenergic receptor kinase (beta 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 (beta ARK1) and GRK3 (beta 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 beta gamma 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 (GRK4alpha , -beta , -gamma , -delta ) 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 GRK4alpha , 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.


EXPERIMENTAL PROCEDURES

Tissue and Cell Sources

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 Assays

PCR 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 Assay

CaM-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.

Antiserum

A fusion protein of 6-His and the N-terminal (amino acids 52 to 114) region of GRK4delta 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-beta -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

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 Fractions

Subcellular 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).

Immunocytochemistry

Electron 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.


RESULTS

Four Splice Variants of GRK4

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 GRK4alpha , -beta , -gamma , and -delta (Fig. 1). GRK4delta was previously known as GRK4B (4) or IT11 (5) while GRK4gamma corresponds to GRK4A (4).


Fig. 1. Four splice variants of GRK4. a, schematic representation of four splice variants of GRK4. The sequences that are alternatively spliced in the N-terminal (amino acids 18-49) and in the C-terminal (amino acids 515-562) regions are in filled boxes. cd, catalytic domain. The spliced sequences have been deposited in the GenBank/EBI Data Bank (accession numbers X98118[GenBank], X97879[GenBank], X97880[GenBank], and X97881[GenBank]). b, expression of GRK4 splice variants (alpha , beta , gamma , and delta ) in HEK293 cells. Cytosolic proteins (100 µg) from HEK293 transfected with the four GRK4 isoforms were blotted, and immunoreactivity was determined using the common antibody N-Ab. c, phosphorylation of rhodopsin by partially purified GRK4 isoforms (alpha , beta , gamma , and delta ) and by similar preparation from HEK293 transfected with the vector alone (Vec). The arrow indicates bands of phosphorylated rhodopsin (Opsin) as revealed by autoradiography after polyacrylamide gel electophoresis. Each experiment was repeated three times.
[View Larger Version of this Image (23K GIF file)]


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 GRK4alpha while GRK4beta , -gamma , and -delta 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 GRK4alpha 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 GRK4alpha was inhibited by heparin (Fig. 2), which is a known selective inhibitor of GRKs, and the IC50 value was 770 ng/ml.


Fig. 2. Characterization of GRK4alpha kinase activity. Upper panel, rhodopsin phosphorylation by partially purified GRK4alpha and purified GRK5 and GRK2 was light-dependent and inhibited by heparin. Heparin concentrations were 10 µg/ml for GRK4alpha and GRK2 and 100 µg/ml for GRK5. Lower panel, inhibition of GRK4alpha kinase activity by different doses of heparin. Each experiment was repeated 2-3 times.
[View Larger Version of this Image (39K GIF file)]


Inhibition of GRK4alpha by Calcium/CaM

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 GRK4alpha 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 GRK4alpha (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 GRK4alpha 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 GRK4alpha 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).


Fig. 3. Inhibition of GRK4alpha by Ca2+/calmodulin. Upper panel, rhodopsin phosphorylation by GRK4alpha in the presence of 1 µM CaM ± 1 mM Ca2+. The effect of Ca2+/CaM was prevented by CaM inhibitor CaMBd (0.5 µM). Lower panel, rhodopsin phosphorylation by GRK4alpha in the presence of the indicated concentration of CaM (plus 1 mM Ca2+). Each experiment was repeated 2-3 times.
[View Larger Version of this Image (32K GIF file)]



Fig. 4. Direct binding of GRK4 splice variants to calmodulin. a, CaM-conjugated Sepharose 4B (CaMSg) was incubated (at 4 °C for 1 h) with 150 µg of cytosolic proteins from HEK293 cells expressing GRK4alpha (left) or 100 nM purified GRK5 (9). Experiments were in the presence or absence of Ca2+. Unconjugated Sepharose 4B (Sg) was used as negative control. Bound GRK4alpha or GRK5, separated by centrifugation followed by extensive washing, were incubated in SDS sample buffer at 95 °C for 5 min, run on 10% SDS-polyacrylamide gel electrophoresis, blotted onto nitrocellulose paper, and revealed by specific antibodies. b, the four splice variants of GRK4 (alpha , beta , gamma , and delta ) were assayed in parallel binding experiments to CaM-conjugated Sepharose 4B performed as described above (left). The starting material contained the same amount of each isoform as documented by immunoblot (right). The blot on the left was developed for much longer to allow the trace amount of bound GRK4beta , -gamma , and -delta to be detected. The experiments shown are representative of three similar ones.
[View Larger Version of this Image (42K GIF file)]


The effect of Ca2+/CaM on GRK4beta , -gamma , and -delta 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 GRK4alpha to CaM, while only trace amounts of GRK4beta , -gamma , and -delta were bound (Fig. 4).

The Expression of GRK4 Is Restricted to Testis

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.


Fig. 5. Northern blot analysis of GRK4 mRNA. Total RNA (20 µg) from mouse Leydig cells TM3 (L), Sertoli cells TM4 (Se), spermatogonia GC-1 spg (Sp), and testis (Te) (left) and from 20- or 60-day-old rat testis (right) was hybridized with a GRK4 human cDNA probe common to the four isoforms. Washed filters were exposed at -80 °C for 24-48 h. Data represent two separate experiments.
[View Larger Version of this Image (54K GIF file)]


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: GRK4delta , 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.


Fig. 6. Expression of GRK4 in sperm membranes. a, GRK4 expression on bovine spermatozoa (lanes 1 and 2), rat adenohypophysis (lanes 3 and 4), and rat testis (lanes 5 and 6). Immunoreactivity of I-20 antibody was determined using 100 µg of proteins from cytosol (lanes 1, 3, and 5) or membranes (lanes 2, 4, and 6) from each tissue. b, blot identical as in a revealed with I-20 antibody preincubated with a ten-fold excess of antigen peptide for 2 h at 4 °C. c, analysis of GRK4 by I-20 in subcellular fractions of bovine sperm. Cytosol (lane 1), enriched nuclear fraction (lane 2), plasma membranes (lane 3), mitochondria (lane 4), and microsomes (lane 5) were fractionated as described under "Experimental Procedures." 50 µg of proteins were loaded in lanes 1-4 while 20 µg of proteins were loaded for the microsome fraction (lane 5). Molecular mass standards (77, 61, 41, and 36 kDa) are indicated by dots on the left of each blot. Data represent two separate experiments.
[View Larger Version of this Image (37K GIF file)]


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 GRK4

Experiments 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.


Fig. 7. Ultrastructural localization of GRK4 on human sperm. a-d, submicroscopical immunogold localization in human spermatozoa treated with I-20 antibody. In a, the presence of colloidal gold granules is evident in the acrosome, specifically on the inner (small arrows) and on the outer (large arrrows) membranes. In b, the particles are present only in the inner membrane after the acrosomal reaction. The localization of GRK4 clearly appears into the mitochondria of the midpiece of the tail (b-d) (arrows). a, × 26,000; b and c, × 21,000; d, × 54,000. e-n, UV and phase contrast micrographs of human spermatozoa treated with the I-20 antibody; an intense fluorescence is visible on the acrosomal (e-h) and on the midpiece (e-l) regions. The fluorescent staining is completely absent in the control (m, n) × 900. A, acrosome; ES, equatorial segment; N, nucleus; M, mitochondria; MP, midpiece; AX, axoneme.
[View Larger Version of this Image (126K GIF file)]



Fig. 8. Submicroscopic immunogold localization of GRK4 in human testis. At the spermatocyte stage, the reaction is positive (arrows) into the Golgi complex (a). At the spermatid stage, the gold granules are localized (arrows) on Golgi-derived cytoplasmic vesicles (b), in the mitochondria (d), and in the acrosome (e). In the young spermatids, the nucleus and the acrosome are devoid of granules (c). a and b, × 21,000; c and d, × 27,000; e, × 52,000. GC, Golgi complex.
[View Larger Version of this Image (159K GIF file)]


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 GRK4alpha and -beta , which contain this insert, while it recognized recombinant GRK4gamma and -delta (data not shown and see Ref. 6). A second antibody (N-Ab) was then generated, raised against a sequence (amino acids 52-114 of GRK4delta ) 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 GRK4gamma (Fig. 9).


Fig. 9. Analysis of GRK4 with N-Ab. a, immunoblot of GRK4 isoforms. HEK293 membranes (100 µg) expressing recombinant GRK4alpha (lane 1), GRK4beta (lane 2), GRK4delta (lane 3), and GRK4gamma (lane 4) were blotted, and immunoreactivity was determined using N-Ab. Immunoblot of human sperm membranes (100 µg) with N-Ab is shown in lane 5. b, ultrastructural localization of GRK4 on human sperm with the polyclonal antibody N-Ab. On the left, gold granules are localized on the inner and outer acrosomal membranes (arrows; × 23,000). On the right, the localization of GRK4 is evident on the outer mitochondrial membrane (long arrows) and sometimes on the cristae (short arrows; × 57,000). Data represent two separate experiments.
[View Larger Version of this Image (73K GIF file)]


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).


DISCUSSION

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 GRK4alpha 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 GRK4alpha was strictly agonist (light)-dependent and inhibited by heparin. An agonist-dependent phosphorylating activity of GRK4alpha was previously shown using, as a substrate, the purified beta 2-AR reconstituted on vesicles containing 5% PIP2 (6).

GRK4alpha 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 GRK4alpha 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 GRK4alpha , but not GRK4beta , -gamma , and -delta , 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 GRK4alpha , but not GRK4beta , -gamma , and -delta was able to bind to CaM in a calcium-dependent manner. We speculate that the long kinase isoform GRK4alpha , 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 GRK4beta and GRK4delta , which lack the N-terminal insert and hence the putative binding site for PIP2. However, we also observed that GRK4gamma , 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, GRK4gamma , 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) GRK4alpha , -beta , and -delta 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 beta -arrestin2 were found in the midpiece of rat sperm, colocalized with odorant-like receptors that are their likely substrates (18). GRK3 and beta -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 GRK4alpha can phosphorylate rhodopsin, and this effect is inhibited by Ca2+/CaM. Unlike GRK4alpha , 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.


FOOTNOTES

*   The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

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].


§   To whom corrrespondence should be addressed: Consorzio Mario Negri Sud, via Nazionale, 66030 S. Maria Imbaro Italy. Tel.: 39-872-570 351; Fax: 39-872-578 240; E-mail: DEBLASI{at}CMNS.MNEGRI.IT.
1   GRK, G protein-coupled receptor kinase; beta ARK, beta -adrenergic receptor kinase; CaM, calmodulin; CaMBd, calmodulin inhibiting peptide; NCS, neuronal calcium sensors; ROS, rod outer segment; PBS, phosphate-buffered saline; PCR, polymerase chain reaction; LH, luteinizing hormone; CG, chorionic gonadotropin.

ACKNOWLEDGEMENTS

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.


REFERENCES

  1. Premont, R. T., Inglese, J., and Lefkowitz, R. J. (1995) FASEB J. 9, 175-182 [Abstract/Free Full Text]
  2. Chuang, T. T., Iacovelli, L., Sallese, M., and De Blasi, A. (1996) Trends Pharmacol. Sci. 17, 416-421 [CrossRef][Medline] [Order article via Infotrieve]
  3. Wilson, C. J., and Applebury, M. L. (1993) Curr. Biol. 3, 683-686 [Medline] [Order article via Infotrieve]
  4. Sallese, M., Lombardi, M. S., and De Blasi, A. (1994) Biochem. Biophys. Res. Commun. 199, 848-854 [CrossRef][Medline] [Order article via Infotrieve]
  5. Ambrose, C., James, M., Barnes, G., Lin, C., Bates, G., Altherr, M., Duyao, M., Groot, N., Church, D., Wasmuth, J. J., Lehrach, H., Housman, D., Bucler, A., Gusella, F. J., and MacDonald, M. E. (1992) Hum. Mol. Genet. 1, 697-703 [Abstract]
  6. Premont, R. T., Macrae, A. D., Stoffel, R. H., Chung, N., Pitcher, J. A., Ambrose, C., Inglese, J., MacDonald, M. E., and Lefkowitz, R. J. (1996) J. Biol. Chem. 271, 6403-6410 [Abstract/Free Full Text]
  7. Parruti, G., Peracchia, F., Sallese, M., Ambrosini, G., Masini, M., Rotilio, D., and De Blasi, A. (1993) J. Biol. Chem. 268, 9753-9761 [Abstract/Free Full Text]
  8. Ausubel, F. M., Brent, R., Kingstone, R. E., Moore, D. D., Seidmann, J. J., Smith, J. A., and Struhl, K. (eds) (1990) Current Protocols In Molecular Biology, Greene Publishing Associates and Wiley Interscience, New York
  9. Chuang, T. T., Paolucci, L., and De Blasi, A. (1996) J. Biol. Chem. 271, 28691-28696 [Abstract/Free Full Text]
  10. Garcìa-Higuera, I., Penela, P., Murga, C., Egea, G., Bonay, P., Benovic, J. L., and Mayor, F., Jr. (1994) J. Biol. Chem. 269, 1348-1355 [Abstract/Free Full Text]
  11. Baccetti, B., Burrini, A. G., Collodel, G., Magnano, A. R., Piomboni, P., and Renieri, T. (1988) J. Submicrosc. Cytol. Pathol. 20, 209-224 [Medline] [Order article via Infotrieve]
  12. Baccetti, B., Benedetto, A., Burrini, A. G., Collodel, G., Costantino Ceccarini, E., Crisa', N., Di Caro, A., Estenoz, M., Garbuglia, A. R., Massacesi, A., Piomboni, P., Renieri, T., and Sollazzo, D. (1994) J. Cell Biol. 127, 903-914 [Abstract]
  13. Filippini, A., Riccioli, A., De Cesaris, P., Paniccia, R., Teti, A., Stefanini, M., Conti, M., and Ziparo, E. (1994) Endocrinology 134, 1537-1545 [Abstract]
  14. Chen, C.-K., Inglese, J., Lefkowitz, R. J., and Hurley, J. B. (1995) J. Biol. Chem. 270, 18060-18066 [Abstract/Free Full Text]
  15. Klenchin, V. A., Calvert, P. D., and Bownds, M. D. (1995) J. Biol. Chem. 270, 16147-16152 [Abstract/Free Full Text]
  16. De Castro, E., Nef, S., Fiumelli, H., Lenz, S. E., Kawamura, S., and Nef, P. (1995) Biochem. Biophys. Res. Commun. 216, 133-140 [CrossRef][Medline] [Order article via Infotrieve]
  17. Pitcher, J. A., Fredericks, Z. L., Stone, W. C., Premont, R. T., Stoffel, R. H., Koch, W. J., and Lefkowitz, R. J. (1996) J. Biol. Chem. 271, 24907-24913 [Abstract/Free Full Text]
  18. Walensky, L. D., Roskams, A. J., Lefkowitz, R. J., Snyder, S. H., and Ronnett, G. V. (1995) Mol. Med. 1, 130-141 [Medline] [Order article via Infotrieve]
  19. Zaneveld, L. J. D., Anderson, R. A., Mack, S. R., and De Jonge, C. J. (1993) Hum. Reprod. 8, 2006-2008 [Medline] [Order article via Infotrieve]
  20. Krausz, C., Gervasi, G., Forti, G., and Baldi, E. (1994) Hum. Reprod. 9, 471-476 [Abstract]
  21. Meyerhof, W., Muller-Brechlin, R., and Ritcher, D. (1991) FEBS Lett. 284, 155-160 [CrossRef][Medline] [Order article via Infotrieve]

©1997 by The American Society for Biochemistry and Molecular Biology, Inc.