Báo cáo khoa học: The role of helix 8 and of the cytosolic C-termini in the internalization and signal transduction of B1 and B2 bradykinin receptors potx

In fact, further substitution of the B1wt C-terminus with corresponding B2wt regions either at C3307.71 following putative helix 8 B1CB2 or at the preceding Y3127.53 in the NPXXY sequenc

the internalization and signal transduction of B1 and B2 bradykinin receptors

Alexander Faussner1, Alexandra Bauer1, Irina Kalatskaya1, Steffen Schu¨ssler1, Cornelia Seidl1, David Proud2and Marianne Jochum1

1 Ludwig-Maximilians-Universita¨t, Abteilung fu¨r Klinische Chemie und Klinische Biochemie, Mu¨nchen, Germany

2 Department of Physiology & Biophysics, University of Calgary, Alberta, Canada

G protein-coupled receptors (GPCRs) form a vast and

diverse superfamily of proteins with seven

transmem-brane-spanning domains They transduce specific

exter-nal stimuli to intracellular second messenger-dependent

effector cascades via recruitment and activation of heterotrimeric G proteins [1] To protect cells from chronic overstimulation, desensitization processes such

as the rapid attenuation of receptor responsiveness and

Keywords

G protein-coupled receptor, helix 8

modeling, internalization, receptor chimera

Correspondence

A Faussner,

Ludwig-Maximilians-Universitaet Muenchen, Abt Klinische

Chemie und Klinische Biochemie,

Nussbaumstr 20, D-80336 Muenchen,

Germany

Tel: +49 89 51602602

Fax: +49 89 51604740

E-mail: alexander.faussner@

med.uni-muenchen.de

(Received 21 July 2004, revised 14 September

2004, accepted 15 September 2004)

doi:10.1111/j.1432-1033.2004.04390.x

Determinants for desensitization and sequestration of G protein-coupled receptors often contain serine or threonine residues located in their C-ter-mini The sequence context, however, in which these residues have to appear, and the receptor specificity of these motifs are largely unknown Mutagenesis studies with the B2 bradykinin receptor (B2wt), stably expressed in HEK 293 cells, identified a sequence distal to N338 (NSMGTLRTSI, including I347 but not the basally phosphorylated S348) and in particular the TSI sequence therein, as a major determinant for rapid agonist-inducible internalization and the prevention of receptor hypersensitivity Chimeras of the noninternalizing B1 bradykinin receptor (B1wt) containing these B2wt sequences sequestered poorly, however, sug-gesting that additional motifs more proximal to N338 are required In fact, further substitution of the B1wt C-terminus with corresponding B2wt regions either at C330(7.71) following putative helix 8 (B1CB2) or at the preceding Y312(7.53) in the NPXXY sequence (B1YB2) resulted in chi-meras displaying rapid internalization Intriguingly, however, exchange performed at K322(7.63) within putative helix 8 generated a slowly inter-nalizing chimera (B1KB2) Detailed mutagenesis analysis generating addi-tional chimeras identified the change of V323 in B1wt to serine (as in B2wt)

as being responsible for this effect The slowly internalizing chimera as well

as a B1wt point-mutant V323S displayed significantly reduced inositol phosphate accumulation as compared to B1wt or the other chimeras The slow internalization of B1KB2 was also accompanied by a lack of agonist-induced phosphorylation, that in contrast was observed for B1YB2 and

B1CB2, suggesting that putative helix 8 is either directly or indirectly (e.g via G protein activation) involved in the interaction between the receptor and receptor kinases

Abbreviations

BK, bradykinin; Bxwt, wild-type Bxbradykinin receptor; DAK, desArg10kallidin; GPCR, G protein-coupled receptor; GRK, G protein-coupled receptor kinase; HEK, human embryonic kidney; IP, inositol phosphate.

G protein uncoupling are essential Some of these

desensitization mechanisms involve the translocation

of the stimulated receptor to distinct compartments

and endocytosis after phosphorylation of serine⁄

threo-nine residues mostly located in the receptor C-termini

(reviewed in [2]) Little is known so far about the

sequence context in which these residues have to

appear to become phosphorylated by kinases and to

be recognized by the internalization machinery In

par-ticular, the receptor specificity of these motifs is not

completely understood

The B1 bradykinin receptor (B1wt) is one of the

few receptors belonging to the class A family of

rho-dopsin-like⁄ b2-adrenergic-like GPCRs that does not

get internalized, i.e sequestered to intracellular

com-partments upon agonist stimulation [3] It does,

how-ever, respond with translocation to caveolae but these

remain essentially on the cell surface [4,5] No

phos-phorylation of B1wt either under basal conditions or

after stimulation has been detected [6] The B2

brady-kinin receptor (B2wt), by contrast, is a more typical

GPCR that gets internalized rapidly following

activa-tion Phosphorylation of several serine⁄ threonine

resi-dues in the C-terminus of this receptor, and the

importance of these events for receptor sequestration,

have been described in detail [7] Whether coupling of

b-arrestin(s) then follows this, and whether

internal-ization occurs via clathrin-coated pits, caveolae or

other less well-defined mechanisms is still a topic of

debate [5,7,8]

The two bradykinin (BK) receptor subtypes exhibit

a relatively low overall amino acid identity of about

36% [9,10], most of it located in the

transmem-brane regions Both receptors stimulate phospholipase

Cb-mediated inositol phosphate (IP) release leading to

an elevation of intracellular [Ca2+] levels, primarily via

coupling to G protein Gq⁄ 11[3,10,11]

They become activated by the kinins, small

pro-inflammatory peptides with great vasoactive potential

implicated as mediators of inflammation, pain and

hyperalgesia [12,13] The nonapeptide BK and Lys-BK

(kallidin) bind with high affinity to B2wt but not B1wt

Removal of the C-terminal arginine through

carboxy-peptidases generates desArg9-bradykinin and

desArg10-kallidin (DAK), two peptides that now bind exclusively

to the B1wt [14]

In this study we wanted to exploit the fact that the

B1wt does not internalize as part of a gain-of-function

approach to provide insight into the receptor

speci-ficity of the B2wt internalization motif The resulting

data also hint at a receptor specific role of the putative

helix 8 in G protein activation and interaction with

receptor kinases

Results

Construction of truncated and point mutated

B2wts and B1⁄ B2receptor chimeras Several studies with truncations of, and deletions in, the C-terminal part of B2wt have demonstrated that this part plays a central role in the internalization of this receptor [7,15,16] A similar function of the C-ter-minus was also observed in other GPCRs with short third intracellular loops [2] In particular, several serine

or threonine residues that become phosphorylated by protein kinase C and⁄ or by GPCR kinases (GRKs) following receptor activation are absolutely required for rapid B2wt sequestration [17]

To determine the C-terminal sequence(s) of the B2wt minimally required for internalization we created two new B2wt truncations, I347* and N338* (Fig 1) The former removed the C-terminus including residue S348, which has been shown to be responsible for the basal phosphorylation of the B2wt, while the latter truncation deleted all serine and threonine residues (S339, T342, T345, S346) shown to be phosphorylated following stimulation of the receptor [17] In addition,

a triple alanine replacement of T345-S346-I347(S348) (mutated residues are underlined) was made, as this sequence strongly resembles the C-terminal STLS-motif in the AT1Aangiotensin II receptor, where a tri-ple alanine substitution of STL almost comtri-pletely abolished receptor sequestration [18]

All of these B2 receptor constructs were highly expressed (Table 1) We took care therefore to use [3H]BK concentrations below 1.5 nm, as we have shown that receptor internalization rates are independ-ent of agonist concindepend-entration in this range [19] The truncation I347* internalized as rapidly as B2wt (Fig 2A) demonstrating that the distal C-terminus, and in particular S348 and its basal phosphorylation,

do not play a decisive role in the sequestration process This notion was further supported by results obtained with a point mutation of S348 to alanine that exhibited

an almost identical internalization rate as the B2wt ([7] and data not shown)

In contrast, deletion of all phosphorylation sites in N338* led to an extremely diminished [3H]BK internal-ization (Fig 2A) Indeed, even the internalinternal-ization cal-culated for each time point is an overestimate because

a shift to lower affinity at 37
C by the receptors remaining on the cell surface can be assumed, as there was a clear drop in surface binding that could not be accounted for by the amount of internalized agonist [20] As the internalization is expressed in percentage

of total binding, decreasing the binding affinity of the

surface receptors simulates an apparent increase in

internalization over time Although it internalized

[3H]BK much slower than the B2wt, N338*

neverthe-less was able to induce an accumulation of total IPs

identical to that observed for the B2wt (Table 1) This

truncated receptor even became hypersensitive, as its

EC50 for the IP response was 10-fold lower than

that of B2wt (0.072 ± 0.038 nm vs 0.79 ± 0.34 nm;

Table 1) Most interestingly, the effects of a truncation

at N338 could also be achieved in part by the triple

mutation TSIfiAAA as this construct displayed

sim-ilar properties to truncation N338* It exhibited a

markedly reduced capacity to internalize [3H]BK albeit

not as diminished as truncation N338* and was at

least as hypersensitive with an EC50¼ 0.058 ±

0.06 nm (Table 1) This sequence obviously contributes

significantly to agonist internalization and signaling of

B2wt

However, transfer of the B2wt C-terminus starting with this sequence, to the C-terminus of the intact noninternalizing B1wt (B1RB2; Fig 1), conferred very little capability to internalize its agonist to the B1wt (Fig 2B) The chimera B1NB2 containing all serine and threonine residues critical for B2wt sequestration,

in contrast, was able to internalize [3H]DAK at a rate approximately half of the maximal rate (40% after

10 min) seen for the B2wt with [3H]BK (Fig 2B)

As it was obviously not sufficient to simply add the

B2wt phosphorylation sites to the B1wt to gain full receptor sequestration as observed in the B2wt, we fur-ther substituted the C-termini of the B2wt into the

B1wt at two residues conserved in both receptor sub-types (Fig 1); specifically at the conserved cysteine [Cys330(7.71) in B1wt, Cys324(7.72) in B2wt] that in the B2wt is palmitoylated (chimera B1CB2) and at Y7.53 within the NPXXY sequence (chimera B1YB2)

Fig 1 Schematic representation of the

C-terminal B1wt and B2wt sequences and

chimera thereof The C-terminal sequences

beginning at transmembrane domain 7 are

shown B1wt parts are indicated in filled

circles, B 2 wt portions in unfilled ones The

phosphorylation sites in B2wt are highlighted

in light grey, and the position number is

indi-cated The grey box outside the membrane

indicates the region of the putative cytosolic

helix 8 as found in the crystal structure of

bovine rhodopsin [24] The assumed

palmi-toylation of B 1 wt and B 2 wt is indicated.

at the end of the seventh transmembrane domain We

have shown previously that a B1CB2 chimera stably

expressed in Chinese hamster ovary cells was

seques-tered rapidly upon activation [16] This was confirmed

in human embryonic kidney (HEK) 293 cells (Fig 2B)

As the chimera B1YB2 exhibited a slightly attenuated

internalization compared to B1CB2 (Fig 2B), and the

latter apparently did not gain the full internalization

capability of the B2wt, we next tested the possibility

that there is an optimum site for creating rapidly

inter-nalizing chimeras at K7.63 between these two residues

and generated the chimera B1KB2 (Fig 1)

Surpris-ingly, B1KB2 showed poor ability to internalize

[3H]DAK (30% after 10 min), with an internalization

far below those seen for B1CB2and B1YB2(Fig 2B)

Agonist-induced internalization of modified

B1KB2constructs

The segment between the NPXXY motif and the

con-served cysteine represents one of the regions with the

highest sequence identity between B1wt and B2wt The

different internalization of B1KB2 and B1CB2 was

therefore even more surprising given that these two

chimeras have only minor sequence differences

(Fig 3A) Therefore we considered three possibilities

to explain the cause of this drop in the internalization

of B1KB2 as compared to B1CB2 First, that the two

residues (KQ) preceding the cysteine were pivotal;

sec-ond, that the cysteine itself needs to be at a specific

position in the C-terminus; or third, that the B1residue

V323 instead of the serine is essential in this

posi-tion Thus, we created three additional chimeras to

test these possibilities: (a) B1KB2⁄ QGVfiKQ; (b)

B1KB2⁄ VCfiCV; and (c) B1KB2⁄ SfiV (Fig 3A) Substituting KQ for QGV in B1KB2led to distinctly increased agonist internalization as compared with

B1KB2 This increase was not due to a corrected posi-tion of the cysteine, as it was not observed with

B1KB2⁄ VCfiCV (Fig 3B)

A major effect, however, was seen with the change

of the polar serine (back) to the nonpolar valine (B1KB2⁄ SfiV), the amino acid that is normally found

in this position in the B1wt This replacement led to a chimera exhibiting rapid internalization (60% after

10 min) that was comparable to that of B1CB2 and

B1YB2(Fig 2B)

Phosphorylation patterns of B2wt and of B1⁄ B2 chimeras reflect their agonist-inducible

internalization Agonist-induced phosphorylation of serine and threo-nine residues in the C-terminus has been shown to be

a prerequisite for internalization of B2wt and other receptors [17,21] B2wt in HEK 293 cells displayed a distinct phosphorylation even in the absence of an agonist (Fig 4), as reported recently [22] When stimu-lated for 5 min with a saturating concentration of 1 lm

BK at 37
C, however, B2wt responded with a marked increase (2.50 ± 0.15-fold over basal) in phosphoryla-tion The chimera on the other hand displayed little basal phosphorylation in the absence of their agonist DAK, although this may, in part, be a sensitivity prob-lem due to their lower expression levels Nevertheless, the rapidly internalizing chimeras B1YB2 and B1CB2

Table 1 Receptor density (Bmax), receptor affinity (Kd), basal and stimulated total IP accumulation, and EC50 of B2wt, B1wt and B1⁄ B 2

receptor chimera ND, not determined.

Receptor construct

B maxa

(fmolÆmg protein)1)

K d

(n M )

IP accumulation

EC 50 ± SEM (n M ) Unstimulated (30 minÆbasal)1)

a Estimated with 10 n M [ 3 H]DAK b P < 0.001 vs B 1 wt.

responded to stimulation with 1 lm DAK with a

dis-tinct increase in phosphorylation The slowly

internaliz-ing B1KB2, in contrast, exhibited no significant

phosphorylation even when challenged with DAK

Total IP accumulation of B1wt and B1⁄ B2

chimeras parallels their agonist-inducible

internalization

The IP release was expressed as unstimulated or

DAK-stimulated accumulation of total IPs for 30 min at

37
C compared to the IP content of control cells that

had remained at 4
C There was a clear correlation

between the agonist-inducible internalization and the

IP accumulation it could induce when stimulated

(Fig 5) All chimeric constructs displaying rapid

agon-ist-inducible internalization (B1CB2, B1YB2, B1KB2⁄

SfiV) showed an IP response similar to that seen for

B1wt (8.41 ± 0.52 fold for B1wt and 7.2–8.7-fold for the chimera) In contrast, the chimera that internalized poorly (B1KB2, B1KB2⁄ QGVfiKQ, B1KB2⁄ VCfiCV) showed a significantly reduced IP signal (4.1–4.6-fold) despite the fact that they were expressed at similar levels to the chimeras that became rapidly internalized (Table 1) These results suggested that V323 might play a role in the activation of phospholipase C through B1wt Indeed, exchange of V323 for a serine

in B1wt (construct B1 V323S) resulted in a clearly reduced IP response (5.28 ± 0.91 vs 8.41 ± 0.52 for

B1wt; Table 1 and Fig 5)

Discussion

Phosphorylation of serine or threonine residues in the C-terminus of GPCRs by second messenger kinases or specific GRKs is a requirement for receptor sequestra-tion [23] However, the context in which these residues have to appear, or the receptor specificity of their function is not very well understood

0

20

40

60

80

100

N338*

TSI->AAA

B 2 wt

I347*

A

Time [min]

0

20

40

60

80

100

B 1 RB 2

B 1 KB 2 B1CB2

B1wt

B1NB2 B1YB2 B

Time [min]

Fig 2 Internalization of [ 3 H]agonist by wild-type bradykinin

recep-tors, truncations and chimera HEK 293 cells expressing the

wild-type receptors B1wt or B2wt, chimera thereof, or B2wt truncations

or mutations were preincubated with the appropriate[ 3 H] agonist:

(A) < 1.5 n M [ 3 H]BK; (B) 2 n M [ 3 H]DAK) for 90 min on ice

Internal-ization was started by placing the cells in a 37
C water bath and

stopped at the indicated times Surface-bound and internalized

agonist were determined as described in Material and methods.

Agonist internalization was expressed as percentage of total bound

agonist Results are given as mean ± SEM of at least three

inde-pendent experiments performed in triplicate.

Fig 3 [ 3 H]DAK internalization of B1KB2derived constructs (A) Align-ment of the relevant sequences of the B 1 CB 2 and B 1 KB 2 -derived chimera compared to wild-type bradykinin receptor subtypes Resi-dues found in B1wt are in capital letters; those found in B2wt are in lowercase Amino acids identical to the B 1 wt sequence are indica-ted by dashes The residues mutaindica-ted in B 1 KB 2 are in bold To allow comparison the sequence of rhodopsin is also shown (B) Internal-ization of [ 3 H]DAK was performed as described in the legend to Fig 2 Each time point represents the mean ± SEM of at least three different experiments done in triplicate.

The bradykinin receptor subtypes are an excellent

tool to address this issue using both loss- and

gain-of-function approaches, as B2wt gets internalized rapidly

following stimulation whereas B1wt does not become

sequestered [3] As both receptors couple preferentially

to the same Ga subunit (Gq⁄ 11) differential signaling is

less likely to explain differences in internalization than

in two receptors signaling through different G proteins

Internalization patterns of truncations I347* and

N338*, and the triple point mutant TSIfiAAA more

closely defined the sequence necessary for the

internal-ization of B2wt Because I347* was internalized as

rap-idly as B2wt, while the TSIfiAAA mutant showed

reduced internalization and N338* almost none, the

nine residues from S339 to I347 (SMGTLRTSI) must

play a key role in B2wt sequestration The following

results from our gain-of-function approach, however,

led us to conclude that additional motifs in the more

proximal portion of the C-terminus also play a role

in receptor internalization First, transfer of the B2wt C-terminus starting with the nine residues containing all known B2wt phosphorylation sites did not permit maximal internalization of [3H]DAK, indicating that this nine residue sequence is either receptor specific or that other motifs must contribute to B2wt sequestra-tion Second, faster internalization was obtained when more extended parts of the C-terminus beginning either at conserved C7.71(B1)⁄ 7.72(B2) (B1CB2) or conserved Y7.53 (B1YB2) were transferred, indicating sequestration motifs in the region between the palmito-ylated C324 and N338 Candidates would include G328-C329 and⁄ or the negatively charged residues E332 and E337, as they are highly conserved in B2wt among species

The chimera B1YB2showed a slightly lower internal-ization compared to B1CB2 We therefore tested whe-ther whe-there was an optimum chimeric exchange point between these two mutation sites Intriguingly, exchange at a conserved lysine (K7.68) between these two sites resulted in a poorly internalizing chimera (B1KB2, Fig 2B) The crystal structure of inactive bovine rhodopsin [24] suggested an explanation for this result by revealing an additional helix 8 close to the seventh transmembrane domain with a cytosolic local-ization parallel to the cell membrane Structure predic-tion programs [25] indicated that both B1wt and B2wt may also contain a helix 8 Our results show that

Mr

75

50

300

250

200

150

100

50

Fig 4 Agonist-induced phosphorylation of B 2 wt and B 1 ⁄ B 2

-chi-mera Upper panel: HEK293 cells expressing B2wt, B1YB2, B1KB2,

or B1CB2were labeled for 10 h with [ 32 P]orthophosphate before

sti-mulation with 1 l M BK and 1 l M DAK, respectively, for 5 min Cells

were lysed and proteins were solubilized, immunoprecipitated and

visualized by autoradiography Molecular size markers are indicated

to the left Lower panel: protein phosphorylation, given as optical

densities of the bands in the area between 50 and 85 kDa, is

pre-sented as mean ± SD from three independent experiments;

un-stimulated B 2 wt was set as 100%.

0

B 1 wt

B 1 CB2

B 1 YB2

B 1 KB2

B 1 KB2

/QGV

B 1 KB2 /VC CV

B 1 KB2 /S V

B 1 V323S KQ

2 4 6 8 10 12

unstimulated stimulated

***

***

***

***

Fig 5 Total IP accumulation of B1wt and chimera HEK293 cells expressing the indicated receptor constructs were preincubated with 50 m M LiCl, and then with (stimulated) or without (unstimula-ted) 1 l M DAK for 90 min on ice IP accumulation was started in a water bath at 37
C and stopped after 30 min as described in Materials and methods The basal IP accumulation level was deter-mined on ice The results are expressed as fold total IP accumula-tion above basal and given as mean ± SEM of at least three different experiments performed in triplicate.

meric receptors with a helix 8 derived either completely

from B1wt (B1CB2) or B2wt (B1YB2) were internalized

rapidly whereas a receptor with a chimeric helix 8

(B1KB2) was internalized slowly As the latter

dis-played no agonist-induced phosphorylation – in

con-trast to chimera B1YB2– this is probably caused by an

impaired interaction with, or activation of, receptor

kinases resulting in the observed slow internalization

Further examination of helix 8 revealed that S316 in

the B2wt sequence of B1KB2is responsible for the slow

sequestration of this chimera (Fig 3B) Helices 8 of

the two receptor subtypes show different charge

distri-butions despite their high sequence identity (Fig 6)

The B2wt exhibits a highly charged N-terminal half

(two arginines and three lysines) but due to S316 does

not display a clear amphipathic structure The

N-ter-minal half of B1wt, by contrast, is less positively

charged (two arginines and one lysine) but B1wt has

a strict amphipathic arrangement of the amino acid

residues This arrangement is probably important in

receptor signaling because: (a) interruption of the amphipathic structure of B1wt helix 8 in B1wt and

B1KB2 through the presence of a serine in position of V323 leads to a strong attenuation of the IP signal; (b) substitution of this serine with valine in B1KB2 fully recovers the IP signal; and (c) sequence alignment

of all known B1 bradykinin receptors shows that hydrophobicity in this position is absolutely conserved, while this is not the case for the residues further down-stream It has been reported recently that truncation

of the B1wt C-terminus at T327 resulted in an 85% reduction of IP generation, whereas further stepwise truncation up to R320 – thus including removal of V323 – did not lead to any further decrease [26] Thus,

it appears that the presence of a hydrophobic residue

at position 323 in B1wt is not necessary for Gq⁄ 11 activa-tion but rather that a polar serine there interferes with this process This group also described a strongly increased basal activity for a B2 receptor construct where several C-terminal serine and threonine residues

Fig 6 Structural comparison of helix 8 in B1wt, B2wt and chimera Helix 8 (N-terminus on the left hand side) from both bradykinin receptor subtypes was modeled along the structure of bovine rhodopsin by means of DEEPVIEW ⁄ Swiss-PdbViewer v3.7 [34] The dark green ribbon presentation belongs to B1wt, light green ribbon-parts to B2wt The residues different in B1wt and B2wt are indicated in larger bold labels Basic amino acid residues are in blue, acidic residues in red, polar residues are yellow, and unpolar residues are colored in grey The black lines in B 1 KB 2 and B 1 KB 2 ⁄ SfiV show the transition between B 1 wt and B 2 wt in the chimera.

were substituted by alanines [27] We also observed a

tendency to increased basal activity in B2constructs that

were lacking all or some of these residues, i.e

TSIfiAAA and N338*, which was significant only for

the latter (P < 0.006) when compared to B2wt

(Table 1) Much more apparent, however, was that, in

our hands, these two constructs were hypersensitive,

dis-playing an EC50value that was more than 10-fold lower

than that observed for B2wt (Table 1) As their Kd

val-ues were not significantly different this indicates that

apparently relatively few receptors have to be occupied

to achieve half-maximal stimulation It is important, of

course, to keep in mind that the Kd was determined at

4
C where coupling to G proteins does not play a role,

whereas the EC50 was obtained by determining the IP

accumulation after 30 min at 37
C Nevertheless, it is

likely that this hypersensitivity is related to the fact that

the mutated residues play an important role in the

inter-nalization (Fig 2A) and in the desensitization of B2wt

[17] Much lower BK concentrations than with B2wt

may therefore be sufficient to activate enough receptors

for half-maximal IP accumulation

In our experiments, we did not observe a strong

constitutive B1wt signaling activity as compared to

B2wt, nor any significant differences between the B1wt

and B1⁄ B2 chimera in terms of basal activity (Table 1)

as was reported recently [26] This discrepancy may be

due to different cell culture conditions (e.g use of

horse serum vs fetal bovine serum), or to their

tran-sient low expression vs our stably high expression and

renders difficult the comparability of our data

Several reports indicate that a fourth cytoplasmic

loop, formed by membrane insertion of a conserved

palmitoylated cysteine, and in particular the part

com-prising putative helix 8, may be involved in the

inter-action of GPCRs with cognate G proteins

Synthetic peptides from the C-terminus of the Ga

subunit Gtand of the Gc subunit of transducin

inter-acted with rhodopsin and kept it in an activated state

[28] This interaction, however, was abolished in

mutants with replacements in helix 8, suggesting that

G protein subunits interact directly or indirectly with

helix 8 In other experiments, peptides with the

sequence of helix 8 of rhodopsin inhibited activation

of Gtby rhodopsin [29] In the angiotensin II receptor

AT1Apoint mutations in the region of putative helix 8

abolished release of inositol trisphosphates and the

GTP-inducible shift in receptor affinity In addition,

peptides based on its helix 8 sequence stimulated

bind-ing of GTPcS to Gq⁄ 11[30] All of these data point to

an involvement of putative helix 8 in the interaction

with cognate G proteins As both bradykinin receptors

coupled to the same Gasubunit Gq⁄ 11the different IP

responses obtained with the wild-type receptors and the chimera let us speculate that each wild-type helix 8 may be specific either for selected bc subunits or for either Gqor G11 Additional experimental work will be necessary to test this hypothesis, particularly as the two receptors, while both coupling to Gq⁄ 11 (and Gi) may very well differ in their capability to activate other additional signaling pathways These potential differences in, for example, the transactivation of growth hormone receptors and in the activation

of MAPK cascades, as well as different localizations of the receptor constructs before and after activation may also contribute to the observed results

Although helix 8 initially was found in the crystal structure of inactive bovine rhodopsin [24], prior stud-ies using NMR and circular dichroism of peptides taken from the fourth cytoplasmic loop of the angio-tensin II AT1Areceptor also indicated that, under cer-tain experimental conditions, an amphipathic a-helix was formed in this region [31] By contrast, NMR studies of peptides representing the same region of rhodopsin in membrane and detergent-free solutions displayed a different structure, with transmembrane domain 7 being extended and the C-terminus up to the cysteine existing as a loop [32] Krishna et al [33] demonstrated that the environment in which the pep-tide exists determines its structure, and suggested that this region serves as a membrane recognition site because the presence of detergent or membrane lipids influences the formation of a helical structure These authors proposed that activation of the receptor, and subsequently of the G protein, leads to a change in the environment of helix 8 resulting in the loss of the heli-cal structure Mutation of specific residues in their model led to a strongly reduced propensity for helical formation with the N-terminus of helix 8 being more influential than the C-terminal portion Based on this model, we could speculate that the two bradykinin receptor subtypes, and those chimeras with an intact⁄ homogenous helix 8, are able to appropriately switch conformation, whereas the receptors with a chi-meric helix 8 have lost this capacity

Taken together, our results demonstrate that almost full capability for receptor internalization can be con-ferred to the normally noninternalizing B1wt, via trans-fer of the C-terminus of B2wt, provided that the new chimeric receptors have an intact⁄ homogeneous helix 8 either from B2wt or B1wt or a chimeric B1⁄ B2 helix with a conserved V323 Chimeric receptors with a het-erogeneous helix 8 exhibited an identical effect on sign-aling as well as on internalization, i.e poor signsign-aling was accompanied by reduced internalization We sug-gest therefore that helix 8 is directly or indirectly

involved in the interaction with receptor kinases and in

receptor specific G protein activation

Materials and methods

Materials

Flp-In T-REx (HEK 293) cells were purchased from

Invitro-gen (GroninInvitro-gen, the Netherlands) and [2,3-prolyl-3,4–

MA, USA) J F Hess (Merck, West Point, PA, USA)

kindly provided us with a vector harboring the sequence of

gift from W Mu¨ller-Esterl (University of Frankfurt,

Germany) Unlabeled peptides were bought from Bachem

(Heidelberg, Germany) The primers were synthesized by

Invitrogen and delivered desalted and lyophilized Pfu DNA

polymerase was obtained from Stratagene Europe

(Heidel-berg, Germany) Fetal bovine serum, culture media, and

Labor-atories (Co¨lbe, Germany) Fugene 6 was from Roche

(Mannheim, Germany) and Invitrogen supplied hygromycin

B and blasticidin Poly(lysine), captopril,

1.10-phenanthro-line and bacitracin were purchased from Aldrich

form) were bought from Bio-Rad (Munich, Germany) All

other reagents were of analytical grade and are

commer-cially available

Cell culture

HEK 293 cells, host cells harboring an Flp recombinant

target (FRT) site in their genome, were cultivated in

Dul-becco’s modified Eagle’s medium (DMEM) with high

the measurement of total inositol phosphate accumulation

cells were seeded on cell culture dishes pretreated with

saline, PBS) to enhance their adherence

Expression vectors

trunca-tions and chimeras of both were cloned into the BamHI and

Each receptor sequence was preceded at the N-terminus by

either a single hemagglutinin-tag (MGYPYDVPDYAGSA)

or a double-tag (MGRSHHHHHH-GYPYDVPDYAGSA)

cloned into the HindIII and BamHI site of the vector For

comparison of analog positions in both receptors we used the

numbering scheme of Ballesteros & Weinstein [35], where the most conserved residue in a transmembrane segment is given the number of the helix followed by the number 50 Residues proximal to this reference residue are obtained by counting down, those distal by counting up from 50 The highest con-served residue in helix 7, the proline within the NPXXY motif, is therefore named P7.50 and the tyrosine of this sequence is identified as Y7.53

Construction of mutated B1wt, B2wt and

of the B1⁄ B2receptor chimera

Standard PCR techniques using either receptor-specific or

were applied to generate truncated or point-mutated

PCR products were ligated between the BamHI and XhoI

using Fugene 6 following the manufacturer’s instructions,

plus 1.6 lg pOG44-vector) and 5 lL Fugene 6 per six-well dish Stably transfected clones were obtained after selection

[3H]Agonist binding studies

times with ice-cold PBS and incubated on ice with 0.15 or

100 lm captopril] containing increasing concentrations of

incuba-tion was stopped by rinsing the monolayers three times with ice-cold PBS and lysing the monolayers by addition

of 0.2 mL of 0.3 m NaOH The bound radioactivity was

another 0.2 mL of water and measured in a b-counter after addition of scintillation fluid Nonspecific binding was determined in the presence of 5 lm unlabeled agonist and subtracted from the total binding to calculate the spe-cific binding

Internalization of [3H]BK and [3H]DAK

To determine the internalization of receptor-bound agonist, cell monolayers on 12-well plates were rinsed three times with ice-cold PBS (pH 7.2) and incubated with the

indi-cated concentration of [3H]agonist in 0.3 mL incubation

buffer on ice to reach equilibrium binding To start the

stopped by placing the trays on ice at the indicated times

Cells were washed three times with PBS and the remaining

monolayer with 0.2 mL of an ice-cold dissociation solution

was quantitatively transferred into scintillation vials by

rinsing the cell monolayer with another 0.2 mL of PBS

transferred to scintillation vials by lysing the cells with

0.2 mL of 0.3 m NaOH and rinsing the wells with

addi-tional 0.2 mL of water The radioactivity of both samples

was determined in a b-counter after addition of scintillation

was determined in the presence of 5 lm unlabeled agonist

and subtracted from the total binding to obtain the specific

values

Stimulation of total IP release

Cell monolayers (80% confluent) in 12-well dishes were

placed on ice, rinsed three times with ice-cold PBS (pH 7.2)

and incubated with or without the appropriate agonist in

incubation buffer containing 50 mm LiCl Basal and

stimu-lated IP accumulation was started by placing the tray in a

exchan-ging the buffer with 0.75 mL of ice-cold 20 mm formic acid

and by transferring the tray onto ice for additional 30 min

As a baseline control one tray was left on ice with LiCl

together with another 0.75 mL of 20 mm formic acid and

ammonium hydroxide solution, 9 mL of 60 mm sodium

formiate, 5 mm sodium tetraborate buffer and 0.5 mL of 4 m

inositol phosphates were eluted by addition of 2.5 mL of the

latter solution The radioactivity was determined in a

b-coun-ter afb-coun-ter the addition of scintillation fluid All data (basal and

IP determined in the baseline control on ice

Immunoprecipitation and Western blotting

Cells were washed once with PBS and solubilized in RIPA

EDTA, pH 7.5] supplemented with 0.5 mm Pefabloc SC and 10 lm each of 1.10-phenanthroline, aprotinin,

rock-ing The sample was centrifuged at 6240 g for 20 min at

was then washed twice with RIPA buffer and once with distilled water, resuspended in 30 lL of Laemmli buffer

the proteins were transferred onto 0.45 lm nitrocellulose membranes After blocking the membranes overnight with

was added in fresh blocking buffer for 2 h at room tem-perature The membranes were washed twice for 10 min in

fol-lowed by addition of the corresponding secondary peroxi-dase-labeled rabbit anti-rat Ig (1 : 1000) for 1 h After washing in TBST three times each for 15 min antibody binding was detected using the Western Blot Chemolumi-nescence Reagent Plus

Receptor phosphorylation

Confluent cells on 6-well plates were washed twice with

P]ortho-phosphate for 10–12 h After exposure to 1 lm BK or

0.5 mL of RIPA buffer containing protease inhibitors (see above) and phosphatase inhibitors (25 mm NaF, 1 mm sodium orthovanadate, 0.3 lm okadaic acid)

polyacryl-amide gel were carried out as described previously The proteins of interest were electroblotted onto nitrocellulose membranes and identified by autoradiography

Protein determination

Total protein per well was quantified by lysing the cells with 0.3 mL of 0.3 m NaOH The protein content of this solution was determined with the Micro BCA Protein assay reagent from Pierce (Rockford, IL, USA) using bovine serum albumin as standard

Data analysis

All data analysis was performed using graphpad prism for Macintosh, Version 3.0a (GraphPad Software, Inc., San Diego, CA, USA)