Recombinant mouse GRK6-A to mGRK6-C were found to be mem-brane-associated in cell-free systems and in transfected COS-7 cells, sug-gesting that the very C-terminus of GRK6-A, lacking in
Trang 1receptor kinase 6 constitutes an accessorial
autoregulatory domain
Petra Vatter*, Claudia Stoesser*, Ines Samel, Peter Gierschik and Barbara Moepps
Department of Pharmacology and Toxicology, University of Ulm, Germany
Phosphorylation by members of a family of serine–
threonine protein kinases, termed G-protein-coupled
receptor kinases (GRKs), has been shown to be crucial
in the rapid agonist-induced desensitization of many
G-protein-coupled receptors [1–4] Seven members of
the mammalian GRK family (GRK1–7), each encoded
by a single gene, have been described Diversity of this protein family is further enhanced by alternative RNA splicing, as shown for human GRK4 [5], human rho-dopsin kinase (GRK1) [6], and murine and human
Keywords
desensitization; G-protein-coupled receptor
kinases; G-protein-coupled receptors;
phosphorylation; signal transduction
Correspondence
B Moepps, Department of Pharmacology
and Toxicology, University of Ulm,
Albert-Einstein-Allee 11, 89081 Ulm, Germany
Fax: +49 731 5002 3872
Tel: +49 731 5002 3883
E-mail: barbara.moepps@uni-ulm.de
*Petra Vatter and Claudia Stoesser
contribu-ted equally to this work.
(Received 9 August 2005, revised 17
September 2005, accepted 27 September
2005)
doi:10.1111/j.1742-4658.2005.04995.x
G-protein-coupled receptor kinases (GRK) are known to phosphorylate agonist-occupied G-protein-coupled receptors We expressed and function-ally characterized mouse GRK6 proteins encoded by four distinct mRNAs generated by alternative RNA splicing from a single gene, mGRK6-A to mGRK6-D Three isoforms, mGRK6-A to mGRK6-C differ in their C-ter-minal-most portion, which is known to mediate membrane and⁄ or receptor interaction and regulate the activity of GRK4-like kinases One isoform, mGRK6-D, is identical to the other mGRK6 variants in the N-terminal region, but carries an incomplete catalytical domain Mouse GRK6-D was catalytically inactive and specifically present in the nucleus of transfected cells Recombinant mouse GRK6-A to mGRK6-C were found to be mem-brane-associated in cell-free systems and in transfected COS-7 cells, sug-gesting that the very C-terminus of GRK6-A, lacking in GRK6-B and mGRK6-C and carrying consensus sites for palmitoylation, is not required for membrane interaction Interestingly, the shortest catalytically active variant, mGRK6-C, was conspicuously more active in phosphorylating light-activated rhodopsin than mGRK6-A and mGRK6-B, implying that the C-terminus of the latter two variants may fulfil an autoinhibitory func-tion Mutation and removal of C-terminal-most region of mGRK6-A by site-directed mutagenesis revealed that this region contains three autoregu-latory elements: two discontinuous inhibitory elements consisting of a single residue, D560, and the sequence between residues S566 and L576, and an intervening stimulatory element The results suggest that
mGRK6-C may be considered a basic, prototypic representative of the GRK4-like kinases, which is capable of interacting with both plasma membrane and its receptor substrate, but is resistant to further regulatory modification conferred to the prototype via C-terminal extension
Abbreviations
CaM kinase, calmodulin-dependent protein kinase; G protein, heterotrimeric guanine nucleotide binding protein; GRK, G protein-coupled receptor kinase; PH, pleckstrin homology; PKC, protein kinase C; PtdCho, phosphatidylcholine; PtdIns, phosphatidylinositol; PtdInsP, phosphatidylinositol 4-phosphate; PtdInsP2, phosphatidylinositol 4,5-bisphosphate; PtdSer, phosphatidylserine; Sf9, Spodoptera frugiperda Spodoptera frugiperda.
Trang 2GRK6 [7,8] Based on similarities in their primary
structures and functional properties, GRKs are
grouped into three subfamilies [9] GRK1 and GRK7
are members of the first subfamily, GRK2 and GRK3
compose the second subfamily, and GRK4, GRK5
and GRK6 make up the third subfamily, also referred
to as the GRK4 subfamily
GRKs share a similar structural organization
char-acterized by a centrally located, highly homologous
catalytical domain flanked by variable N- and
C-ter-minal regions [1,2] The N-terC-ter-minal regions of the
known mammalian GRKs are similar in size and have
been proposed to be involved in receptor recognition
[10] The C-terminal regions of the known mammalian
GRKs, which share a certain degree of conservation,
but vary in length [11], are generally thought to play
an important role in localizing the kinases to the
plasma membrane [2,12] This region can be further
subdivided into two constituents: a moderately
con-served N-terminal region consisting of 65 residues
(residues 449–514 in mGRK6-A to mGRK6-C) [7] and
a highly variable C-terminal region [13] The variable
region contains sites for regulatory protein–protein
interactions and⁄ or post-translational modification,
e.g isoprenylation or palmitoylation, which are
thought to be involved in targeting of GRKs to the
receptor and⁄ or the plasma membrane
We have previously identified mRNAs encoding four
distinct mouse GRK6 isoforms (mGRK6), designated
mGRK6-A to -D [7] Analysis of the genomic
organ-ization of the mouse GRK6 gene showed that the four
mRNAs are generated by alternative RNA splicing
from a single gene Three of the GRK6 isoforms,
GRK6-A, -B and -C, are likely to exist in human and
rat tissues as well [8] (P Vatter, C Stoesser, I Samel,
P Gierschik & B Moepps, unpublished results)
Expression studies revealed that the four mGRK6
mRNAs are differentially expressed in mouse tissues,
suggesting that the four mGRK6 isoforms are involved
in regulating tissue- or cell-type-specific receptor
func-tions in vivo [7] Interestingly, mGRK6-A, but not
mGRK6-B or -C, phosphorylated the Na+⁄ H+
exchanger regulatory factor (NHERF) via a
PDZ-domain-mediated interaction, presumably mediated by
the PTRL motif specifically present at the C-terminus
of this GRK6 isoform [14] The gene encoding GRK6
has been inactivated in the mouse [15] Interestingly,
lymphocytes derived from GRK6-deficient mice were
strikingly impaired in their ability to respond to the
CXC chemokine CXCL12, indicating that GRK6 is
specifically involved in regulating the CXCL12
tor CXCR4 [15,16] In addition, dopamine D2
recep-tors and leukotriene B4 receptors have been shown to
be regulated by GRK6 [17,18] The GRK6 variants mGRK6-A, -B and -C are indistinguishable up to resi-due 559 of the variable C-terminal region and diverge from each other at residue 560 Mouse GRK6-A is identical to human GRK6-A between residues 560 and
576 and therefore contains the three cysteine residues known to undergo palmitoylation in hGRK6-A within its C-terminus [19,20] The C-terminus of mGRK6-B is
13 residues longer than the corresponding portion of mGRK6-A and lacks these palmitoylation sites, but carries several basic residues and consensus sequences for phosphorylation by protein kinase C and cAMP⁄ cGMP-dependent protein kinases The unique C-termi-nus of mGRK6-C is made up of only a single arginine residue present in position 560
In this study, we expressed the four variants of mouse GRK6 as recombinant polypeptides in both baculovirus-infected insect and in COS-7 cells to char-acterize the functional significance of their structural differences The results show that mGRK6-D is cata-lytically inactive and specifically present in the nucleus
of transfected COS-7 cells In contrast, all three forms carrying a complete catalytical domain are catalytically active and localized to the plasma membrane in intact cells The observation that mGRK6-C, lacking the C-terminal extensions present in mGRK6-A and -B and likely to serve as substrates for palmitoylation or protein phosphorylation, respectively, is considerably more active than the latter two variants, suggests that the C-terminal extensions are not required for mem-brane interaction Instead, they appear to fulfil an accessorial autoregulatory function subject to further, reversible modification by palmitoylation or protein phosphorylation, respectively
Results
Production of polyclonal antisera against mouse GRK6 variants
To characterize the four mouse GRK6 variants at the protein level, the polypeptides encoded by the four mGRK6 mRNAs were expressed in Sf9 insect cells and COS-7 cells To monitor expression and to exam-ine the subcellular distribution of the recombinant pro-teins, two polyclonal antisera reactive against the various mGRK6 isoforms were produced Serum PV1 was raised against a peptide located in the N-terminal portion present in all four isoforms (amino acids 87– 101) Serum PV2 was directed against a peptide from the unique C-terminus of mGRK6-B (amino acids 574–589) To specifically detect mGRK6-A, a commer-cially available polyclonal antiserum (sc-566) reactive
Trang 3against the C-terminus mGRK6-A (amino acids 557–
576) was used (cf Fig 1A) To determine the
specificity of the antisera, all four mGRK6 variants
were expressed as recombinant polypeptides in
baculo-virus-infected insect cells and analysed by
immunoblot-ting Figure 1B shows that all four polypeptides
migrated at the expected positions on SDS
polyacryl-amide gels and were reactive with serum PV1 (left) In
contrast, mGRK6-A (centre) and mGRK6-B (right)
were specifically detected by antisera sc-566 and PV2,
respectively No immunoreactivity was seen using
anti-sera PV1, PV2 or sc-566 in noninfected insect cells
(results not shown) Using antisera sc-566 and PV2 to detect mGRK6-A and -B in membrane preparations of mouse tissues by immunoblotting (results not shown),
we found that mGRK6-A was expressed in brain, liver, heart and spleen and, at lower levels, in lung, mesenterial lymph nodes and thymus Mouse GRK6-B was present at high levels in brain, followed by spleen, mesenterial lymph nodes and thymus, but was unde-tectable in liver and heart Attempts at specifically detecting mGRK6-C on immunoblots were unsuccess-ful due to the fact that this variant differs from mGRK6-A and -B in but a single residue (R560, cf Fig 7A), and that it was not possible to resolve the three variants by SDS⁄ PAGE Nevertheless, these results clearly indicate that the splice variants of mGRK6 are differentially expressed at the protein level in mouse tissues
Subcellular localization of the mouse GRK6 variants in transfected COS-7 cells
To examine the subcellular distribution of the recom-binant mGRK6 variants in cultured mammalian cells, COS-7 cells were transiently transfected with cDNAs encoding mGRK6-A to -D Transfected cells were homogenized in low ionic strength buffer, unbroken cells and nuclei were removed from the homogenate, and the postnuclear supernatant was separated into soluble and particular components In addition, the particulate material was extracted with buffer contain-ing Triton X-100 (1.5% w⁄ v) Figure 2A shows the results of an immunochemical analysis of the soluble and particular fractions as well as the detergent extracts of these samples Recombinant mGRK6-A and -B were exclusively, and mGRK6-C predomin-antly present in the particulate fraction of transfected cells Thus, only a small fraction of mGRK6-C appeared in the soluble fraction under the conditions used in this experiment All three variants were solubi-lized from the particulate fraction with buffer contain-ing Triton X-100 No immunoreactivity was found for mGRK6-D in soluble or particulate postnuclear frac-tions of COS-7 cells transfected with the mGRK6-D cDNA (not shown)
Next, the subcellular localization of the mGRK6 isoforms was examined by immunocytochemistry and confocal microscopy using antiserum PV1 and fluores-cent-labelled secondary antibodies Figure 2B shows that immunoreactivity corresponding to mGRK6-A, -B and -C was located at the cell membrane and, although
to a lesser extend, in the cytoplasma Most interest-ingly, cells transfected with the cDNA of mGRK6-D displayed immunoreactive protein exclusively in regions
A
B
Fig 1 Expression of four mouse GRK6 proteins in
baculovirus-infected insect cells (A) Linear representation of the four mGRK6
variants generated by alternative RNA splicing The predicted
cata-lytic domain and the regulatory N- and C-terminal regions are
shown as shaded and open boxes, respectively The variable
C-ter-minal regions of mGRK6-A, -B and -C are shown in detail The
residues predicted to serve as substrates for palmitoylation in
mGRK6-A and for phosphorylation by protein kinases C and
cAMP⁄ cGMP-dependent protein kinases in mGRK6-B, respectively,
are underlined Mouse GRK6-D terminates at residue 244 within
the predicted catalytic domain (dashed vertical line) The positions
of the synthetic peptides used to generate antisera PV1, sc-566
and PV2 are illustrated The numbering of the amino acid residues
is indicated (B) Sf9 cells were infected with baculoviruses
enco-ding the mGRK6 variants mGRK6-A, -B, -C and -D Forty-eight hours
after infection, cells were homogenized with lysis buffer containing
250 m M NaCl and the homogenate was fractionated into soluble
and particulate constituents Aliquots of the soluble fractions of
insect cells expressing mGRK6-A, -B, -C and -D containing 9, 13, 4
and 45 lg of protein, respectively, were subjected to SDS ⁄ PAGE
and immunoblotting was performed using antisera PV1 (left),
sc-566 (centre) or PV2 (right) Soluble proteins from noninfected
Sf9 cells (Sf9) (7 lgÆlane)1) were analysed for comparison The
positions of the molecular mass standards are indicated.
Trang 4corresponding to the cell nucleus Thus, mGRK6-D is
expressed in transiently transfected COS-7 cells, but
localized to the nuclear fraction, which explains the
absence of the protein from the postnuclear fractions
analysed in Fig 2A Untransfected COS-7 cells and
mGRK6 cDNA-transfected cells incubated with the
primary or secondary antibody alone displayed no
im-munoreactivity (not shown) Taken together, these
results revealed that: (a) mGRK6-A, -B and -C appear
to be at least partially cell membrane-associated in
intact mammalian cells; and (b) mGRK6-D is
specific-ally present in the nucleus of transfected COS-7 cells
Because both mGRK6-B and mGRK6-C lack the
putative C-terminal palmitoylation sites present in
mGRK6-A and thought to mediate membrane binding,
these data imply further sites and mechanisms to be
important for membrane association of former two
mGRK6 variants
Interaction of mGRK6-C with phospholipids
To study the interaction of recombinant mGRK6-C
with lipid membranes in more detail, insect cells
infec-ted with baculovirus encoding mGRK6-C were
homo-genized in buffer containing increasing concentrations
of NaCl and the postnuclear homogenate was separ-ated into soluble and particulate constituents Figure 3 shows that mGRK6-C was predominantly present in the particulate fraction in the absence of NaCl, but was translocated to a considerable extent into the sol-uble fraction at increasing concentrations of NaCl Thus, membrane binding of mGRK6-C is dependent
on and inversely related to the ionic strength of the incubation medium
Previous studies have suggested an important role
of N- and C-terminal portions of members of the GRK4 subfamily in mediating their interaction with phospholipids, including phosphatidylinositol 4,5-bis-phosphate (PtdInsP2) [21,22] To assess the potential role of these putative phospholipid-binding sites for targeting mGRK6-C to lipid membranes, we investi-gated the interaction of recombinant mGRK6-C to synthetic phospholipid vesicles made up of phospha-tidylcholine (PtdCho) and a small fraction of either phosphatidylserine (PtdSer) or of the inositol phos-pholipids PtdIns, PtdInsP or PtdInsP2 To this end, mGRK6-C was expressed in baculovirus-infected insect cells and purified from the soluble fraction
by sequential cation exchange and heparin-affinity chromatography Active mGRK6-C was assayed by
B
A
Fig 2 Expression of four mouse GRK6 proteins in transiently transfected COS-7 cells COS-7 cells were transfected as indicated with 2 lg each per well of pMT2 containing the cDNAs of mGRK6-A, -B, -C or -D (A) Forty-eight hours after transfection, cells were homogenized and the homogenate was fractionated into soluble (S) and particular (P) constituents A portion of the particular fraction was extracted with buffer containing Triton X-100 to obtain a detergent-soluble lysate (D) Aliquots (D, S: 70 lg protein per lane; P: 40 lg protein per lane) of the samples were subjected to SDS ⁄ PAGE and immunoblotting was performed using antiserum PV1 (B) Forty-eight hours after transfection, the cells were fixed and permeabilized, and immunostaining of the mGRK6 variants was performed using antiserum PV1 Plasma mem-branes (pm) and nuclei (n) of cells expressing mGRK6-A, -B, -C or -D, respectively, are marked by arrows No immunostaining was observed
in nontransfected COS-7 cells (not shown).
Trang 5immunoblotting (Fig 4A) and by measuring the
abil-ity of the protein to phosphorylate light-activated
rhodopsin (Fig 4B) Figure 4C shows that the protein
was mostly homogenous after heparin-affinity
chro-matography
The interaction of purified mGRK6-C with
phos-pholipids was investigated by incubating the protein
with increasing concentrations of phospholipid vesicles
and then separating vesicle-bound from soluble
mGRK6-C by ultracentrifugation Figure 5A shows
that mGRK6-C did not interact with vesicles made up
of PtdCho even at the highest concentration tested
(10 lgÆmL)1) When PtdSer was present in the vesicles,
only a minor portion of mGRK6-C sedimented with
the particulate fraction In contrast, interaction of
mGRK6-C was clearly evident already at low phos-pholipid concentrations when PtdInsP2 was present in the vesicles At the highest phospholipid concentration tested, all of the mGRK6-C protein was found in the vesicle fraction To examine the specificity of the effect
of PtdInsP2shown in Fig 5A, purified mGRK6-C was incubated with increasing concentrations of PtdCho vesicles containing PtdIns, PtdInsP or PtdInsP2 As shown in Fig 5B, all three inositol phospholipids pro-moted the interaction of mGRK6-C with the vesicle preparation However, the strength of this interaction was markedly dependent on the nature of inositol phospholipid Specifically, maximal phospholipid bind-ing was observed in the presence of PtdInsP2, followed
by PtdInsP and PtdIns
A
C
B
Fig 4 Purification of mGRK6-C from Sf9 cells Recombinant mGRK6 was purified from the soluble fraction of baculovirus-infected insect cells by sequential chromatography on SP Sepharose and heparin Sepharose (A) Aliquots (5 lL) of the indicated fractions obtained by chro-matography on heparin Sepharose were subjected to SDS ⁄ PAGE and immunoblotting using antiserum PV1 (B) Aliquots of the same frac-tions (1 lL) were incubated with urea-treated rod outer segment membranes and [ 32 P]ATP[cP] to measure phosphorylation of light-activated rhodopsin Samples were subjected to SDS ⁄ PAGE and autoradiography was performed (C) An aliquot (4 lg) of fraction 6 was subjected to SDS ⁄ PAGE and proteins were stained with silver The positions of the molecular mass standards are indicated Aliquots of the sample applied to the heparin Sepharose matrix (15 and 1 lL, respectively) were analysed for comparison (Co).
Fig 3 Effect of NaCl on the interaction of recombinant mGRK6-C with the particulate fraction of baculovirus-infecetd insect cells Sf9 cells were infected with baculovirus encoding mGRK6-C Forty-eight hours after infection, cells were homogenized with lysis buffer containing increasing concentrations of NaCl and the homogenate was fractionated into soluble (S) and particulate (P) constituents Aliquots (40 lg pro-tein per lane) of the fractions were subjected to SDS ⁄ PAGE and immunoblotting was performed using antiserum PV1 Only the 65 kDa region of the chemiluminogram is shown.
Trang 6Functional properties of the mouse GRK6
isoforms
The next experiment was designed to examine and
compare the functional properties of the mGRK6
iso-forms mGRK6-A, -B and -C The three recombinant
proteins were produced in baculovirus-infected insect
cells, adjusted to similar levels by immunoblotting, and
then assayed for their ability to phosphorylate
light-activated rhodopsin (Fig 6) None of the variants
phosphorylated rhodopsin in the dark (not shown) All
three isoforms were capable of specifically
phosphory-lating the active receptor protein (Fig 6), whereas
insect cell-expressed mGRK6-D was inactive in this
regard (not shown) Most interestingly, the activity of
mGRK6-C was by far higher than the activities of the
C-terminally extended variants mGRK6-A and -B (Fig 6) These results suggested that the C-terminal extensions present in the latter two isoforms may reduce their ability to phosphorylate activated receptor polypeptides
Structure–activity relationships of the C-terminus
of mGRK6
To examine the functional significance of the C-ter-minus of mGRK6-A, a mutant was generated (mGRK6-A M1) carrying serine residues instead of the cysteine residues C561, C562 and C565, which are known to be substrates for palmitoylation in wild-type mGRK6-A Furthermore, deletion mutants
of mGRK6-A M1 lacking the C-terminal-most 9 (mGRK6-A M2) and 16 (mGRK6-A M3) residues were produced The rationale behind constructing these mutants was based on the desire to determine the influ-ence of the C-terminus of mGRK6-A without the in-fluence of its palmitoylation (Fig 7A) Wild-type mGRK6-A and -C as well as the three mutants of mGRK6-A M1, M2 and M3, were produced as recom-binant proteins in baculovirus-infected insect cells The amounts of the recombinant proteins were adjusted to similar levels by immunoblotting and the proteins were assayed for their ability to phosphorylate light-activated rhodopsin (Fig 7B) Replacement of the three
Fig 5 Interaction of mGRK6-C with phospholipids Recombinant
mGRK6 was purified to homogeneity from the soluble fraction of
baculovirus-infected insect cells Aliquots (2 lg protein per sample)
were incubated with increasing concentrations of phospholipid
vesi-cles made up of either 100% phosphatidylcholine (PtdCho) or 95%
(w ⁄ w) phosphatidylcholine and 5% (w ⁄ w) of the indicated
phos-pholipids The incubation mixtures (30 lL) were fractionated into
soluble (S) and particulate (P) constituents by centrifugation The
soluble supernantant was removed and the membrane fraction was
resuspended in 30 lL of incubation buffer Aliquots (15 lL per lane)
were subjected to SDS ⁄ PAGE and immunoblotting was performed
using antiserum PV1 Only the 65 kDa regions of the
chemilumino-grams are shown.
mGRK6-A mGRK6-B mGRK6-C
Rho*
Fig 6 Phosphorylation of light-activated rhodopsin by mGRK6 vari-ants Sf9 cells were infected with baculovirus encoding mGRK6-A, -B and -C Aliquots of the soluble fractions of infected cells contain-ing similar amounts of recombinant mGRK6 proteins were subjec-ted to SDS ⁄ PAGE and immunoblotting using antiserum PV1 (upper)
or incubated with urea-treated rod outer segment membranes and [ 32 P]ATP[cP] to measure phosphorylation of light-activated rhodop-sin (Rho*) The samples were subjected to SDS ⁄ PAGE and auto-radiography of the gel was performed (lower) The position of Rho*
is indicated.
Trang 7C-terminal cysteines by serine residues resulted in an
almost complete loss of the ability of the protein to
phosphorylate light-activated rhodopsin Interestingly,
gradual removal of the C-terminal-most 16 residues of
mGRK6-A M1 affected the activity of the protein
toward rhodopsin in a nonuniform manner Thus,
removal of the C-terminal-most nine residues of
A M1, S568 to L578, in the mutant
mGRK6-A M2 led to a marked increase in rhodopsin
phos-phorylation Further removal of seven residues, S561 to
D567, in mutant mGRK6-A M3 caused a marked loss
of this activity (Fig 7B, lower) Note that the latter
mutant differs from wild-type mGRK6-C in only a sin-gle residue, D560 instead of R560 Thus, the very C-ter-minus of mGRK6-A contains at least three regulatory elements capable of markedly and specifically affecting the activity of the kinase towards the receptor substrate, which are shown schematically in Fig 7A: two discon-tinuous inhibitory elements consisting of a single resi-due, D560, and the sequence between residues S566 and L576, and an intervening stimulatory element
Discussion
We expressed the four variants of mouse GRK6 as recombinant polypeptides to characterize the func-tional significance of the differences in their primary structures Whereas mGRK6-A, -B and -C are similar
in terms of their overall structural organization to the other members of the GRK family, mGRK6-D is peculiar in that it terminates prematurely in its puta-tive catalytic domain [7] The finding reported here that mGRK6-D is catalytically inactive when overex-pressed as recombinant protein in baculovirus-infected insect cells is consistent with this structural deficiency Interestingly, mGRK6-D, unlike the other variants
of mGRK6, was specifically present in the nucleus of transiently transfected COS-7 cells Differential subcel-lular localization of protein products of alternatively spliced RNAs is not without precedence in the litera-ture Thus, localization in different subcellular com-partments including the nucleus has been shown for splice variants of the fibroblast growth factor receptor FGFR-3 [23] and multifunctional Ca2+⁄ calmodulin-dependent protein kinase (CaM kinase) [24,25] The CaM kinase splice variants dB-CaM and aB-CaM both carry an 11 amino acid insertion in their variable regions This insertion generates a nuclear localization signal, KKRK [26], which targets the proteins to the nuclei of transfected cardiac myocytes or neuroblas-toma cells, respectively Interstingly, a KKRK motif is also present in the C-terminal region of mGRK6-D If this peptide does, in fact, represent a nuclear localiza-tion signal of mGRK6-D, it is likely to be inactive in the other three variants, because those variants also carry the motif, but are absent from the nucleus in transfected COS-7 cells (Fig 2B) As the KKRK motif
is part of an intact catalytic domain in mGRK6-A, -B and -C, it is conceivable that this motif is inaccessible
in those variants, but freely available in mGRK6-D
to interact with the nuclear import receptor subunit importin a [27] The function of mGRK6-D in the nucleus is currently unknown
The C-terminal-most portion of GRKs is thought to play an important role in mediating receptor and⁄ or
B
A
Fig 7 Effect of C-terminal deletions on mGRK6-A activity (A)
Lin-ear representation of the C-terminal regions of wild-type mGRK6-A
and mGRK6-C, and of the mutants of mGRK6-A The residues
predicted to serve as substrates for palmitoylation in wild-type
mGRK6-A are underlined The positions of the regions suggested
to be involved in membrane interaction containing a high number
of basic residues and those suggested to be involved in negative
(circled minus symbols) and positive (circled plus symbol)
autoregu-lation are indicated The numbering of the amino acid residues is
indicated (B) Sf9 cells were infected as indicated with baculovirus
encoding either wild-type mGRK6-A or -C, or deletion mutants of
mGRK6-A Aliquots of the soluble fractions of infected cells
con-taining similar amounts of recombinant mGRK6 proteins were
sub-jected to SDS ⁄ PAGE and immunoblotting using antiserum PV1
(upper) or incubated with urea-treated rod outer segment
mem-branes and [ 32 P]ATP[cP] to measure phosphorylation of
light-activa-ted rhodopsin (Rho*) The samples were subjeclight-activa-ted to SDS ⁄ PAGE
and autoradiography of the gel was performed (lower) The position
of Rho* is indicated.
Trang 8plasma membrane interaction Thus, the primary
struc-tures of GRK1 and GRK7 terminate with a CAAX
sequence that directs post-translational isoprenylation,
proteolysis, and carboxy-methylation [28] GRK2 and
GRK3 are localized in the cytosol and are translocated
to the plasma membrane upon stimulation of
G-pro-tein-coupled receptors [29] The translocation of GRK2
to the membrane has been shown to be due to the
bind-ing of a C-terminal region includbind-ing a pleckstrin
homology domain to membrane-bound G protein bc
dimers and negatively charged membrane
phospho-lipids, including PtdInsP2 [30–34] Mutations in the
GRK2 pleckstrin homology domain and the region
distal to the C-terminal amphipathic helix resulted in a
specific and profound loss of GRK2 responsiveness to
bc dimers and phospholipids [35] Members of the
GRK4 subfamily, GRK4, GRK5 and GRK6, contain
neither a CAAX motif nor a G protein bc
dimer-bind-ing domain in the variable region of the C-terminal
region, but nevertheless exhibit a significant degree of
association with cellular membranes [5,19,36] GRK4
and human GRK6-A are palmitoylated at cysteine
resi-dues present within their C-terminal regions, whereas
GRK5 contains a C-terminal polybasic domain likely
to mediate direct interaction with negatively charged
phospholipid head groups [37]
The fact that GRK6-C lacks the C-terminal cysteine
residues likely to serve as substrates for palmitoylation
in GRK6-A, together with the observation that the
activity of GRK6-A is substantially increased by
palmitoylation, led Premont et al [8] to suggest that
GRK6-C may associate with membranes only poorly
and may represent a poor regulator of
G-protein-coupled receptors However, the degree of association
of recombinant mGRK6-C with the plasma membrane
lipid bilayer in transfected COS-7 cells was similar to
that of mGRK6-A and mGRK6-B (Fig 2A) These
results imply that the C-terminal-most portions of
mGRK6-A and mGRK6-B are not necessary for
mem-brane association and that other mechanisms may be
involved in targeting mGRK6 to the plasma
mem-brane Of interest, at least two phospholipid binding
sites appear to be present in GRK5 to mediate
inter-action of this member of the GRK4-like kinases
with the plasma membrane [21,22] The first site,
K22RKGKSKK in bovine GRK5 (basic residues
underlined), is located at the N-terminus of GRK5
and appears to specifically interact with PtdInsP2 via
its basic residues Binding of PtdInsP2 to this site
markedly enhances GRK5-mediated phosphorylation
of the human b2-adrenoceptor, most likely by
facili-tating the interaction of the kinase with the lipid
bilayer [21] The second, C-terminally located site,
Q552RLFKRQHQNN in human GRK5 (basic resi-dues underlined), is critical for the interaction of GRK5 with artificial phospholipid vesicles in cell-free systems and with the plasma membrane in intact cells [22] Importantly, similar sites are present in GRK4 and in GRK6-A to -C Specifically, the N-terminal site corresponds to K21QTGRSKK in mGRK4 and to N22RKGKSKK in mGRK6-A to -C The C-terminal site corresponds to R552RLFRRTGCLN in mGRK4,
to Q553RLFSRQDCCG in mGRK6-A, to Q553RL FSRQRIAV in mGRK6-B, and to Q553RLFSRQR in mGRK6-C Note that an acidic residue, D560, is pre-sent in the C-terminal motif of mGRK6-A in place of the basic residue, R560, present in mGRK6-B and -C, which may reduce the electrostatic potential of the for-mer motif to interact with phospholipids In work to
be published elsewhere (C Stoesser et al., unpublished results), we found that both the N-terminal and the C-terminal phospholipid-binding sites of mGRK6-C are involved in the interaction of this mGRK6 variant with phospholipid vesicles containing PtdInsP2 These observations suggest that two distinct phospholipids-binding sites commonly present in mGRK6-A to -C mediate or contribute to the interaction of these mGRK6 variants with the plasma membrane
Previous studies have shown that the C-terminal regions of GRK4 subfamily members are involved not only in membrane targeting, but also in regulation of substrate recognition and⁄ or catalytic activity Thus, palmitoylation and artificial geranyl geranylation of the C-terminus of human GRK6-A enhanced phos-phorylation not only of the human b2-adrenoceptor reconstituted into phospholipid vesicles, but also of the soluble nonreceptor substrate casein [38] These results strongly suggest that these post-translational modifica-tions not only enhance the hydrophobicity and thereby strengthen the membrane association of GRK6-A, but also increase the kinase catalytic activity of the pro-tein Along the same lines, the C-terminal 28 residues immediately downstream of the C-terminal phospho-lipid-binding site of human GRK5 have previously been suggested to fulfil an autoinhibitory function, which is enhanced by protein kinase C-mediated phos-phorylation and⁄ or autophosphorylation [22] Thus, phosphorylation of two or three serine residues within this region by protein kinase C dramatically reduced the ability of the kinase to phosphorylate both recep-tor and soluble nonreceprecep-tor substrates without affect-ing GRK5 bindaffect-ing to artificial phospholipid vesicles [39] The same region also contains three serine resi-dues, which are potential sites for Ca2+ -calmodulin-stimulated autophosphorylation of human GRK5 [22] Phosphorylation of one or several of these residues
Trang 9causes substantial inhibition of the interaction of
GRK5 with the receptor substrate without disrupting
the catalytic activity or the association of GRK5 with
phospholipids [40] The results reported in this study
indicate that the C-terminal-most 16 residues of
mGRK6-A may function as an autoregulatory domain
to control the activity of this kinase towards the
recep-tor substrate Inhibirecep-tory effects of this region are most
clearly evident for the first residue, D560, and the last
nine residues that discriminate mGRK6-A from
mGRK6-C, S568EEELPTRL576 The seven residues
between positions 561 and 567, which comprise the
positions of the cysteine residues serving as substrates
for palmitoylation in wild-type mGRK6-A (C561,
C562 and C565) appear to exert a stimulatory effect It
is tempting to speculate that palmitoylation of one or
several of these cysteines may further enhance this
stimulatory effect
Taken together, we have shown that all three
iso-forms of mGRK6 containing a complete catalytic
domain, mGRK6-A, mGRK6-B and mGRK6-C, are
localized to the plasma membrane in intact cells The
C-terminal extensions that discriminate mGRK6-A
and -B from mGRK6-C are not required for
mem-brane interaction, but instead appear to fulfil an
acces-sorial autoregulatory function subject to reversible
modification by palmitoylation or protein
phosphory-lation, respectively The fact that GRK6-A and
GRK6-B resemble GRK4 and GRK5, respectively, in
terms of the structural and functional properties of
their C-termini leads us to suggest that GRK6-C may
be considered a basic, prototypic representative of the
GRK4-like kinases, which is capable of interacting
with both plasma membrane and its receptor substrate,
but is resistant to further regulatory modification
con-ferred to the prototype via C-terminal extension The
fact that all three GRK6 isoforms, in contrast to the
isoforms of GRK4 [8], are conserved between several
mammalian species indicates that the existence of
functionally distinct forms of GRK6 is biologically
important
Experimental procedures
Materials
PtdCho (P-7763), PtdSer (P-7769), PtdIns (P-8443), PtdInsP
(P-9638) and PtdInsP2 (P-9763) were obtained from Sigma
(Taufkirchen, Germany) Antisera PV1 and PV2 were
raised in rabbits against synthetic peptides corresponding to
residues 87–101 (H2N-V87SEYEVTPDEKRKAC-CONH2)
and 574–589 (H2N-CP574PASSPQAEAPTGGWR-COOH),
respectively, of mGRK6-B An N-terminal cysteine was
added to the latter peptide to facilitate coupling to the car-rier protein The rabbit polyclonal antiserum sc-566 reactive against amino acids 557–576 of mGRK6-A was purchased from Santa Cruz Biotechnology (Santa Cruz, CA) Tri-ton X-100 (PT-8787) was obtained from Sigma
Construction of wild-type and mutant GRK6 cDNAs
The cDNAs of the coding regions of A,
mGRK6-C and mGRK6-D were amplified by PmGRK6-CR from single-stranded cDNA prepared from mouse mesenterial lymph nodes as described previously [7] Complementary DNAs encoding C-terminal mGRK6-A mutants were prepared from the wild-type mGRK6-A cDNA by PCR In brief, mGRK6-A mutant M1 was generated according to the pro-tocol of the QuickChangetm
site-directed mutagenesis kit (Stratagene, La Jolla, CA) (18 cycles: 94C for 1 min,
55C for 2 min, 72 C for 2 min, followed by a single incu-bation at 72C for 10 min) using primer P1, 5¢-CGCCA AGATTCCTCTGGGAACTCCAGCGACAGT-3¢ (nucleo-tides 1677–1709, sense) in combination with primer P2, 5¢-ACTGTCGCTGGAGTTCCCAGAGGAATCTTGG CG-3¢ (nucleotides 1677–1709, antisense) Complementary DNAs encoding the C-terminal mGRK6-A mutants M2 and M3 were prepared either from the mGRK6-A mutant M1 or from the wild-type mGRK6-A cDNA by PCR (30 cycles: 94C for 1 min, 70 C or 62 for 2 min, 72 C for
2 min, followed by a single incubation at 72C for 10 min) using primer P1, 5¢-AGCCCATGGAGCTCGAGAACA TCGTA-3¢ (nucleotides 1–26, sense, initiating ATG under-lined) in combination with antisense primers introducing a stop codon sequence: mGRK6-A M2: P2, 5¢-CTAGTCGC TGGAGTTCCCAGAGGAATCTTGGCG-3¢ (nucleotides 1677–1706, antisense, stop codon underlined) and
mGRK6-A M3: P3, 5¢-CTmGRK6-AmGRK6-ATCTTGGCGmGRK6-ACTGmGRK6-AmGRK6-AGmGRK6-AGTCT-3¢ (nucleotides 1665–1685, antisense, stop codon underlined) The PCR products were ligated into pCR 2.1 (Invitrogen, Carlsbad, CA) and their identities were verified by DNA sequencing The numbering of oligonucleotides used as primers refers to the nucleotide sequence of the mGRK6-A cDNA deposited with EMBL⁄ GenBank under Accession
no Y15799
Production of recombinant baculoviruses
The cDNAs of wild-type and mutant mGRK6 isoforms were prepared by endonuclease digestion from pCR 2.1, filled in with Klenow enzyme, and subcloned into the SmaI site of the baculovirus transfer vector pVL1393 (Invitro-gen) The correct orientation of the inserts was verified
by DNA sequencing Recombinant baculoviruses were obtained by transfecting Sf9 cells with a 25 : 1 (v⁄ v) mix-ture of the transfer vector and a modified baculovirus
Trang 10DNA (Baculogold; BD Biosciences-Pharmigen, Franklin
Lakes, NJ, USA), which contains a lethal deletion and is
rescued by the DNA of the transfer vector High-titre
stocks of the baculoviruses were produced by three cycles
of amplification in Sf9 cells
Production of recombinant mGRK6 proteins in
baculovirus-infected insect cells
Sf9 cells were grown at 27C in TNM-FH medium (Sigma)
supplemented with 50 lgÆmL)1 gentamycin and 10% (v⁄ v)
fetal bovine serum in 75-cm2 cell culture flasks Cells
(1· 107
cells per flask) were incubated for 48 h with
recom-binant baculovirus in 15 mL medium Two days after
infec-tion, cells were collected by centrifugainfec-tion, rinsed once with
NaCl⁄ Pi (140 mm NaCl, 2.7 mm KCl, 8.0 mm Na2HPO4,
1.4 mm KH2PO4, pH 7.2) and resuspended in 200 lL
ice-cold lysis buffer A (20 mm Hepes⁄ NaOH, pH 7.5, 250 mm
NaCl, 10 mm EDTA, 1 mm dithiothreitol, 0.5 mm
phenyl-methylsulfonyl fluoride, 200 lgÆmL)1 benzamidine, and
20 lgÆmL)1 leupeptin) Cells were homogenized by forcing
the suspension six times through a 0.5· 23 mm needle
attached to a disposable syringe The lysate was centrifuged
at 40 000 g for 30 min at 4C, and the supernatant was
centrifuged again at 300 000 g for 30 min at 4C and then
passed through 0.22 lm pore size nitrocellulose filters To
produce larger quantities of recombinant mGRK6 proteins,
Sf9 cells were grown at 27C in suspension culture in
Grace’s Complete Insect Medium (BioWhittaker)
supple-mented with 0.2% (w⁄ v) Pluronic F-68 (Gibco BRL,
Rock-ville, MD), 50 lgÆmL)1 gentamycin, and 2.5 lgÆmL)1
amphotericin B (Fungizone, Gibco BRL) Cells (8· 108per
flask) were incubated for 48 h with recombinant
baculovi-rus in 400 mL medium at 80 r.p.m on a rotary shaker with
amplitude of 25 mm The cells were collected by
centrifuga-tion, washed once in ice-cold NaCl⁄ Pi, resuspended
in 40 mL lysis buffer A and then homogenized and
fract-ionated as described above The mGRK6 proteins
mGRK6-A through -C were similarly abundant in soluble
fractions of baculovirus-infected insect cells ( 1% of total
protein) The catalytically inactive variant mGRK6-D was
at least 10-fold less abundant in this fraction (Fig 1B, left)
Purification of recombinant mGRK6-C
Recombinant mGRK6-C was purified from the soluble
frac-tion of baculovirus-infected insect cells by sequential
chro-matography on SP Sepharose high performance and heparin
Sepharose high performance using an A¨KTAexplorer
chromatography system (Amersham Biosciences, Freiburg,
Germany) The soluble fraction from Sf9 cells infected with
baculovirus encoding mGRK6 (40 mL, 260 mg protein)
was diluted with 80 mL of ice-cold buffer B (20 mm
Hepes⁄ NaOH, pH 7.5, 10 mm EDTA, 1 mm dithiothreitol,
0.5 mm phenylmethylsulfonyl fluoride, 200 lgÆmL)1
benz-amidine and 20 lgÆmL)1leupeptin) and applied to a 5 mL HiTrap SP HP column (Amersham Biosciences) that had been equilibrated with buffer B containing 125 mm NaCl The flow rate was 1 mLÆmin)1 After application of the sam-ple, the column was washed with 40 mL of buffer B contain-ing 125 mm NaCl and eluted with a linear gradient (50 mL)
of NaCl (125–500 mm) in buffer B Fractions of 1 mL were collected and analysed by SDS⁄ PAGE and immunoblotting using antiserum PV1, and by measuring the ability of the fractions to phosphorylate light-activated rhodopsin The active material, which eluted at 350 mm NaCl, was pooled (10 mL, 20 mg of protein) and diluted with buffer B to obtain a final NaCl concentration of 300 mm The sample was then applied to a 1 mL HiTrap Heparin HP column (Amersham Biosciences), which had been equilibrated with buffer B containing 300 mm NaCl The flow rate was 0.5 mLÆmin)1 After application of the sample, the column was washed with 30 mL of buffer B containing 300 mm NaCl and eluted with a linear gradient (40 mL) of NaCl (0.3–1 m) in buffer A Fractions of 1 mL were collected and analysed by SDS⁄ PAGE and staining of proteins with silver The active material eluted at 600 mm NaCl and was directly used in mGRK6 phospholipid interaction studies
Expression of recombinant mGRK6 proteins
in COS-7 cells
The cDNAs encoding mGRK6-A to -D were ligated into the mammalian expression vector pMT2 [41] COS-7 cells were grown in six well dishes (35 mm wells, 1· 105
cells per well) in Dulbecco’s modified Eagle’s medium supple-mented with 10% (v⁄ v) fetal bovine serum, 2 mm l-gluta-mine, 100 unitsÆmL)1 penicillin, 0.1 mgÆmL)1 streptomycin,
1 mm sodium pyruvate and 25 mm Hepes⁄ NaOH, pH 7.2
at 37C, in a humidified atmosphere of 90% air and 10% CO2 Cells were transfected by lipofection (2 lg plasmid DNA per well) using SuperFecttm
(Qiagen, Hilden, Ger-many) according the manufacturer’s instructions For immunoblot analysis, transfected COS-7 cells were harves-ted by scraping into 1 mL of ice-cold NaCl⁄ Pi 48 h after transfection Cells were sedimented by centrifugation and resuspended in 200 lL ice-cold lysis buffer A (20 mm Tris⁄ HCl, pH 7.5, 2 mm EDTA pH 7.5, 0.5 mm phenyl-methylsulfonyl fluoride, 20 lgÆmL)1 aprotinin and
10 lgÆmL)1leupeptin) and homogenized by forcing the sus-pension six times through a 0.5· 23 mm needle attached to
a disposable syringe The lysate was centrifuged at 300 g for 2 min at 4C, and the supernatant was centrifuged again at 300 000 g for 30 min at 4C The supernatant was removed and the particulate fraction was resuspended in
200 lL of lysis buffer A To solubilize proteins from
150 lL of the membrane fraction, 50 lL of a solution containing 400 mm NaCl, 2 mm EDTA and 6% (v⁄ v) Triton X-100 was added The mixture was incubated for
1 h at 4C by end-over-end rotation and then centrifuged