Stepwise phosphorylation of leukotriene B receptor 1 ... ·...

SC I ENCE S I GNAL ING | R E S EARCH ART I C L E

B IOCHEM ISTRY

1Department of Biochemistry and Molecular Biology, Faculty of Medicine, TheUniversity of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan. 2Depart-ment of Chemistry and Biotechnology, Graduate School of Engineering, The Uni-versity of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan. 3Departmentof Biochemistry, Juntendo University School of Medicine, Bunkyo-ku, Tokyo 113-8421, Japan. 4Department of Molecular and Cellular Biochemistry, GraduateSchool of Pharmaceutical Sciences, Tohoku University, Sendai, Miyagi 980-8578,Japan. 5Department of Life Science, Faculty of Science, Okayama University ofScience, 1-1 Ridai-cho, Kita-ku, Okayama 700-0005, Japan. 6Japan Agency for Med-ical Research and Development (AMED), Core Research for Evolutional Science andTechnology (CREST), 1-7-1 Otemachi, Chiyoda-ku, Tokyo 100-0004, Japan. 7Depart-ment of Lipid Signaling, National Center for Global Health andMedicine, 1-21-1 To-yama, Shinjuku-ku, Tokyo 162-8655, Japan.*Present address: Division of Biology and Biological Engineering, California Insti-tute of Technology, 1200 East California Boulevard, MC 216-76, Pasadena, CA91125, USA.†Present Address: Vascular Biology Program, Boston Children’s Hospital, HarvardMedical School, Boston, MA 20115, USA.‡Corresponding author. Email: [email protected]

Nakanishi et al., Sci. Signal. 11, eaao5390 (2018) 21 August 2018

Copyright © 2018

The Authors, some

rights reserved;

exclusive licensee

American Association

for the Advancement

of Science. No claim

to original U.S.

Government Works

httD

ownloaded from

Stepwise phosphorylation of leukotriene B4receptor 1 defines cellular responses to leukotriene B4Yoshimitsu Nakanishi1, Modong Tan2, Takako Ichiki3*, Asuka Inoue4, Jun-ichi Yoshihara5,Naoto Maekawa5, Itsuki Takenoshita5, Keisuke Yanagida1†, Shinya Yamahira2,Satoshi Yamaguchi2, Junken Aoki4,6, Teruyuki Nagamune2, Takehiko Yokomizo3,Takao Shimizu1,7, Motonao Nakamura5‡

Leukotriene B4 (LTB4) receptor type 1 (BLT1) is abundant in phagocytic and immune cells and plays crucial rolesin various inflammatory diseases. BLT1 is phosphorylated at several serine and threonine residues upon stim-ulation with the inflammatory lipid LTB4. Using Phos-tag gel electrophoresis to separate differentially phos-phorylated forms of BLT1, we identified two distinct types of phosphorylation, basal and ligand-induced, in thecarboxyl terminus of human BLT1. In the absence of LTB4, the basal phosphorylation sites were modified tovarious degrees, giving rise to many different phosphorylated forms of BLT1. Different concentrations of LTB4induced distinct phosphorylation events, and these ligand-induced modifications facilitated additional phos-phorylation events at the basal phosphorylation sites. Because neutrophils migrate toward inflammatory sitesalong a gradient of LTB4, the degree of BLT1 phosphorylation likely increases in parallel with the increase inLTB4 concentration as the cells migrate. At high concentrations of LTB4, deficiencies in these two types of phos-phorylation events impaired chemotaxis and b-hexosaminidase release, a proxy for degranulation, in Chinesehamster ovary (CHO-K1) and rat basophilic leukemia (RBL-2H3) cells, respectively. These results suggest thatan LTB4 gradient around inflammatory sites enhances BLT1 phosphorylation in a stepwise manner to facili-tate the precise migration of phagocytic and immune cells and the initiation of local responses, includingdegranulation.

p://

on A

pril 3, 2021stke.sciencem

ag.org/

INTRODUCTIONWithmore than 900members, G protein–coupled receptors (GPCRs)are the largest family of cellular receptors (1). Binding of an agonistto a GPCR induces a conformational change in the receptor, whichstimulates dissociation of the heterotrimeric G proteins associatedwith the receptor. The dissociated Ga and Gbg subunits stimulatediverse cellular responses through effector proteins and intracellularsecondmessengers (2).When the activated receptors are phosphoryl-ated on Ser and Thr residues within the intracellular loops and thecarboxyl tail, cellular responses to subsequent stimuli are usually at-tenuated, a phenomenon known as desensitization. The functionalsignificance of GPCR phosphorylation has also been characterizedin contexts other than desensitization. For example, the distinctGPCR phosphorylation profiles mediated by different kinases (thephospho-barcodes) explain the multidimensional features of GPCRsignaling networks (3–6). Thus, precise analysis of the phosphoryl-

ation of specific GPCRs is useful for clarifying the biological functionsof each receptor.

Leukotriene B4 (LTB4) is a lipid mediator that is derived fromarachidonic acid in the plasma membrane and released into the ex-tracellular milieu (7–9), where it acts as a potent chemoattractantfor neutrophils and lymphocytes (10). Experiments using micro-scale arrays have demonstrated that the LTB4 chemotactic gradientis important for triggering neutrophil swarming (11, 12). The LTB4receptor type 1 (BLT1) is a high-affinity LTB4 receptor that is abun-dant in and contributes to the activation and migration of phagocyticand immune cells, such as neutrophils, eosinophils, dendritic cells,macrophages, B cells, and T cells (13). BLT1 is a GPCR and activatesat least two different Ga familymembers, including the pertussis toxin(PTX)–sensitive Gi/o and PTX-insensitive Gq alpha proteins (14, 15).The activation of theseGa proteins byBLT1 results in phosphoinositidehydrolysis, intracellular Ca2+ mobilization, and a decrease in cellularcyclic adenosine monophosphate, followed by chemotaxis and de-granulation (15, 16). Like other GPCRs, BLT1 is phosphorylated inits C-terminal tail after activation. Several studies have reported thephosphorylation of BLT1 (17–19), yet the precise phosphorylation sitesand their biological importance remain to be determined.

Here, we determined the amino acid residues responsible for basal(constitutive) and ligand-induced phosphorylation of human BLT1using Phos-tag SDS–polyacrylamide gel electrophoresis (PAGE), inwhich a small molecule that is incorporated into the polyacrylamidegel matrix slows the mobility of phosphorylated proteins by specifi-cally binding to phosphate groups. Subsequent mutational analysesestablished the phosphorylation hierarchies of these residues. To bet-ter understand the biological roles of these phosphorylation events,we examined chemotaxis and b-hexosaminidase release elicited bydifferent concentrations of LTB4 using BLT1 mutants lacking theidentified phosphorylation sites. We found that an LTB4 gradient

1 of 15

http://stke.sciencemag.org/


boosted BLT1 phosphorylation in a stepwise manner to facilitate theprecise migration of immune cells and the initiation of local re-sponses, including degranulation.

on April 3, 2021


Dow

nloaded from

RESULTSBLT1 is phosphorylated basally and LTB4 induciblyBecause BLT1 on the cell surface is covalentlymodified withN-glycanmoieties of various molecular masses, wild-type human BLT1 that isN-terminally taggedwithhemagglutinin (HA-BLT1/WT)and expressedin HeLa cells appears as diffuse bands by SDS-PAGE (fig. S1A),making it difficult to detect the phosphorylation of BLT1 as a bandshift inWestern blot analysis. To eliminate this problem, we generateda BLT1 mutant in which the two putative N-glycosylation sites in theextracellular domain,Asn2 andAsn164, weremutated toAla. The result-ing protein, HA-BLT1/0N, is a prospective N-glycosylation–deficientreceptor. When expressed in HeLa cells, HA-BLT1/0N appeared asa sharp band on the SDS-PAGE (fig. S1A), although the bandmigratedslightly more slowly than that of endoglycosidase-treated HA-BLT1/WT (fig. S1A). This difference could be due to posttranslational mod-ifications other than N-glycosylation. The trafficking of HA-BLT1/0Nto the cell surface was identical to that of HA-BLT1/WT (fig. S1B).Furthermore, HA-BLT1/0N showed LTB4-elicited intracellular Ca

2+

mobilization, although this activity was slightly lower than that ofHA-BLT1/WT (fig. S1C). These results indicate that HA-BLT1/0Nis a useful tool for the analysis of BLT1 phosphorylation sites.

We performedWestern blot analysis ofHA-BLT1/0N expressed inHeLa cells with or without LTB4 stimulation. Although LTB4 stimu-lation led to the appearance of a shifted band on the polyacrylamidegel, the separation of phosphorylated and nonphosphorylated HA-BLT1/0N was not sufficient for precise analysis (Fig. 1A). Therefore,we used Phos-tag technology with SDS-PAGE, which allows for thespecific separation of various phosphorylated forms of proteins (20).HA-BLT1/0Nmigrated asmultiple bands on the Phos-tag SDS-PAGEwithout LTB4 stimulation (Fig. 1A), suggesting that various LTB4-independent phosphorylations occur at multiple sites. Moreover, ad-ditional slower migrating bands appeared after LTB4 stimulation.These shifted bands disappeared with alkaline phosphatase (AP)treatment (Fig. 1B), demonstrating that these shifted bands weredue to the phosphorylation of HA-BLT1/0N. We also examined thephosphorylation states ofHA-BLT1/WT fromLTB4-treated cells aftertreatment with endoglycosidase by Phos-tag SDS-PAGE. There was nosubstantial difference in the band patterns between endoglycosidase-treated HA-BLT1/WT and HA-BLT1/0N (Fig. 1C), suggesting thatN-glycosylation has no effect on these LTB4-induced modifications.Moreover, BLT1 endogenously expressed in retinoic acid–stimulatedhuman promyelocytic leukemia (HL-60) cells shared a phosphoryla-tion pattern similar to that of exogenously expressed HA-BLT1/WT(Fig. 1D). Together, these data demonstrate that BLT1 was phos-phorylated to various extents even in the resting state (henceforthreferred to as “basal phosphorylation”) and that LTB4 stimulationinduced additional phosphorylation events (henceforth referred toas “LTB4-induced phosphorylation”).

BLT1 has seven phosphorylated residues in its C terminusGPCR phosphorylation occurs on Ser and Thr residues in the cyto-plasmic domains (21). There are 21 Ser and Thr residues that are con-served among the cytoplasmic domains of human, mouse, rat, guineapig, and zebrafish BLT1s (fig. S2). Because kinases require specificmo-


tifs around the target residue, we excluded Ser343 and Ser344 from ouranalysis because of the low homology surrounding these residuesamong different species. To identify the amino acid residues criticalfor basal phosphorylation of BLT1, we divided the remaining 19 resi-dues into 10 subgroups based on the proximity of the residues andgenerated 10 mutant HA-BLT1/0N receptors in which these residueswere replaced with Ala (Fig. 2A). The Phos-tag SDS-PAGE analysesrevealed that Ser and Thr residues in the intracellular loops were notpart of the basal phosphorylation profile; however, the iL-1 mutant(S51A and T53A) was not produced in HeLa cells, as previously re-ported (Fig. 2B) (22). The bands for mutants iL-2B (S125A andT130A) and iL-3B (S216A and T219A) showed a slight difference inmobility compared to HA-BLT1/0N (Fig. 2B), possibly because weused a twofold greater volume of the mutant samples relative toHA-BLT1/0N to generate comparable signal intensity. On the otherhand, themutant forms ofHA-BLT/0Nbearing substitutionmutationsin the C terminus, CT-3 (S313A, S314A, andT315A) andCT-4 (S320Aand T324A), displayed aberrant basally phosphorylated bands after

PNGase F

LTB4 ATRALTB4

CIAP

IB: HA

IB: HA

-Actin

Mem

bran

eW

CL

Mem

bran

eW

CL

IB: HA

-Actin

-Actin

IB: BLT1

Mem

bran

eW

CL

WC

LM

embr

ane

IB: HA

-Actin

Mock BLT1/0N

LTB4

SDS-PAGE

Phos-tag SDS-PAGEPNGase F

LTB4

Moc

k

BLT1

/WT

BLT1

/WT

BLT1

/0N

BLT1

/0N

A

B

C

D

Fig. 1. Phosphorylation of human BLT1. (A) HeLa cells expressing HA-taggedBLT1/0N (HA-BLT1/0N) were stimulated with LTB4, and membrane fractions weresubjected to SDS-PAGE and Phos-tag SDS-PAGE. Receptors were detected by im-munoblotting (IB) for the HA tag. b-Actin is an experimental and loading control.Mock, empty vector; WCL, whole-cell lysates. (B) HeLa cells expressing HA-BLT1/0Nwere stimulated with LTB4, and membrane fractions were treated with calf intes-tine alkaline phosphatase (CIAP) before Phos-tag SDS-PAGE and immunoblottingfor HA. (C) HeLa cells expressing wild-type BLT1 (BLT1/WT) or a form of BLT1lacking the extracellular N-glycosylation sites (BLT1/0N) were stimulated withLTB4, and membrane fractions were treated with peptide-N-glycosidase F(PNGase-F) before Phos-tag SDS-PAGE and immunoblotting for HA. (D) HL-60 cellscultivated in the presence of all-trans retinoic acid (ATRA) were stimulated withLTB4, and membrane fractions were treated with PNGase-F before Phos-tagSDS-PAGE and immunoblotting for BLT1. All blots are representative of at leastthree independent experiments.

2 of 15



on April 3, 2021


Dow

nloaded from

Phos-tag SDS-PAGE (Fig. 2B). Next, we examined the amino acid res-idues involved in LTB4-induced phosphorylation. Previous reportsdemonstrated that the substitution of all Ser and Thr residues inthe C-terminal domain with Ala led to a notable disappearance ofLTB4-induced phosphorylation, suggesting that the phosphorylationof these sites depends on LTB4 (17). Thus, we focused ourmutationalanalyses on Ser and Thr residues in the C-terminal domain. Mutantreceptor CT-2 (T308A and S310A) did not show any band shifts uponLTB4 treatment (Fig. 2C). These experiments identified Ser

313, Ser 314,Thr315, Ser320, and Thr324 as candidates for basal phosphorylation andSer310 and Thr308 as candidates for LTB4-induced phosphorylation.

To further characterize these seven potential basal and LTB4-induced phosphorylation sites, we generated 12 additional mutantsinwhich these residues weremutated individually and in various com-binations (Fig. 2D).Mutants CT-3 and CT-4 have a combined five Serand Thr residues, Ser313, Ser314, Thr315, Ser320, and Thr324, mutated toAla. To determine whether all or only a subset of these residues wasbasally phosphorylated, we constructed five mutants in which onlyone of these five residues was not substituted with Ala (BLT1/0N/S313, BLT1/0N/S314, BLT1/0N/T315, BLT1/0N/S320, and BLT1/0N/T324). Each of these five mutants migrated more slowly, and sub-


stitution of all five residues (BLT1/0NDb-phos) led to the completedisappearance of shifted bands in the absence of LTB4 (Fig. 2, C, E,and F). These results indicate that all five Ser and Thr residues in thisC-terminal region (Ser313, Ser314, Thr315, Ser320, and Thr324) contrib-ute to the basal phosphorylation of HA-BLT1/0N. To address theresidues mutated in the CT-2 mutant that showed aberrant LTB4-induced phosphorylation (Ser310 and Thr308), we generated mutantsin which only one of these two residues was mutated (BLT1/0N/T308and BLT1/0N/S310, respectively) and mutants in which one of thesetwo residues was mutated in addition to the five basally phosphoryl-ated residues (Db/T308 and Db/S310, respectively). Each of these mu-tants still exhibited the LTB4-induced shifts, indicating that bothresidues were involved in LTB4-induced phosphorylation (Fig. 2F).Moreover, the BLT1/0NDb-phos mutant, completely devoid of basalphosphorylation, displayed LTB4-stimulated band shifts (Fig. 2F),demonstrating that basal phosphorylation was not required forLTB4-induced phosphorylation.

We next compared three mutant receptors in which all five of theresidues that were basally phosphorylated were mutated (HA-BLT1/0NDb-phos), both of the residues that exhibited LTB4-induced phos-phorylationweremutated (HA-BLT1/0NDi-phos), or all seven residues

E F

A

iL-1 : S51A, T53AiL-2A : S116AiL-2B : S125A, T130AiL-3A : S200A, S202AiL-3B : S216A, T219ACT-1 : S295ACT-2 : T308A, S310ACT-3 : S313A, S314A, T315ACT-4 : S320A, T324ACT-5 : S337A, S339A

Basal-phosLTB -phos

BLT1/0N : 352BLT1/0N/S313 : 352BLT1/0N/S314 : 352BLT1/0N/T315 : 352BLT1/0N/S320 : 352BLT1/0N/T324 : 352

BLT1/0N b/S310 : 352BLT1/0N b/T308 : 352BLT1/0N/S310 : 352BLT1/0N/T308 : 352

BLT1/0N b-phos : 352BLT1/0N i-phos : 352BLT1/0N phos : 352

7TM H8

B

Moc

kBL

T1/0

NiL-

1iL-

2AiL-

2BiL-

3AiL-

3B

BLT1

/0N

CT-1

CT-2

CT-3

CT-4

CT-5

Mem

bran

e

WCL -Actin

IB: HA

CD

Me

mb

ran

eW

CL -Actin

IB: HA

Moc

kBL

T1/0

NBL

T1/0

Nb-

phos

CT-1

CT-2

CT-5

LTB4

WCL

Moc

kBL

T1/0

NBL

T1/0

N/S3

13

BLT1

/0N/

S314

BLT1

/0N/

T315

BLT1

/0N/

S320

BLT1

/0N/

T324

Mem

bran

e

-Actin

IB: HA

phos :

Mem

bra

ne

WCL -Actin

IB: HA

BLT1

/0N

b-ph

osBL

T1/0

Nb/

S310

BLT1

/0N

b/T3

08

LTB4 (nM) 0 10 100 0 10 100 0 10 100

– – + – + – + – + – +

Fig. 2. Residues essential for phosphorylation of BLT1 cytoplasmic domains. (A) Conserved Ser (S) and Thr (T) residues in the cytoplasmic domains of human BLT1.The indicated 10 human HA-BLT1/0N mutants were generated. (B) HA-BLT1 mutants were expressed in HeLa cells, and membrane fractions were examined by Phos-tagSDS-PAGE followed by immunoblotting for HA. b-Actin is an experimental and loading control. (C) HeLa cells expressing the indicated HA-tagged receptors weretreated with 100 nM LTB4 for 40 min before separation of membrane fractions by Phos-tag SDS-PAGE and immunoblotting for HA. (D) The indicated HA-BLT1/0Nmutants were constructed to determine the phosphorylation sites in the BLT1 C-terminal domain. Thr308, Ser310, Ser313, Ser314, Thr315, Ser320, and Thr324 weresubstituted with Ala (A) as indicated. 7TM, transmembrane domain 7; H8, helix 8. (E) HA-BLT1/0N and the indicated mutants were expressed in HeLa cells, and mem-brane fractions were examined by Phos-tag SDS-PAGE followed by immunoblotting for HA. (F) Membrane fractions from HeLa cells expressing the indicated mutantforms of BLT1 treated with 10 or 100 nM LTB4 for 40 min were separated by Phos-tag SDS-PAGE and immunoblotted for HA. All blots are representative of at least threeindependent experiments.

3 of 15



on April 3, 2021


Dow

nloaded from

weremutated (HA-BLT1/0NDphos) (Fig. 2D). HA-BLT1/0NDb-phosand HA-BLT1/0NDi-phos exhibited greatly reduced basal phospho-rylation and LTB4-induced phosphorylation, respectively, whenexpressed in HeLa cells (fig. S3A). Furthermore, bands correspondingto both basal phosphorylation and LTB4-induced phosphorylationdisappeared in cells expressing HA-BLT1/0NDphos (fig. S3A). Thesephenomena were reproduced in Chinese hamster ovary (CHO-K1)cells, murine interleukin-3–dependent pro-B (Ba/F3) cells, and rat ba-sophilic leukemia (RBL-2H3) cells, implying that the observed mod-ifications occur inmany different cell types (fig. S3B). The seven BLT1residues that we identified as basally phosphorylated and phosphory-lated in an LTB4-dependent manner are conserved among humans,mice, rats, guinea pigs, and zebrafish (fig. S2). We observed similarphosphorylation of the corresponding residues in mouse BLT1 bygenerating and analyzing mBLT1/0N, mBLT1/0NDb-phos, mBLT1/0NDi-phos, mBLT1/0NDphos, mBLT1/0N/S313, mBLT1/0N/S314,mBLT1/0N/T315, mBLT1/0N/T320, mBLT1/0N/T324, mBLT1/0N/T308, and mBLT1/0N/S310 mutants (fig. S4A). In mouse, Ser313

and Ser314 contributed to basal phosphorylation, whereas phosphoryl-ation of Thr308 and Ser310 occurred upon stimulation with LTB4 (fig.S4B). Thus, these data further support that the observedmodificationsare not species-specific. Collectively, our data suggest that humanBLT1 has five basal phosphorylation sites (Ser313, Ser314, Thr315,Ser320, and Thr324) and two LTB4-induced phosphorylation sites(Thr308 and Ser310) in its C-terminal region. The presence of manydistinctly migrating bands for the wild-type protein and each mutantform indicates that phosphorylation of these residues may occur inmany different combinations.

Phosphorylation of BLT1 is facilitated by dose-dependentLTB4 stimulationTo evaluate whether different concentrations of LTB4 affected theextent of human BLT1 phosphorylation, we stimulated HeLa cellsexpressing HA-BLT1/0N or HA-BLT1/0NDb-phos with variousconcentrations of LTB4. The LTB4-induced phosphorylation of bothHA-BLT1/0N and HA-BLT1/0NDb-phos increased in a dose-dependent manner (Fig. 3, A and B). For the HA-BLT1/0NDb-phosmutant, one of the LTB4-induced phosphorylated species initiallyappeared at 1 nM LTB4, but the other LTB4-induced phosphorylatedspecies (phosphorylated on both Ser310 and Thr308) only appeared inLTB4 concentrations at or above 30 nM (Fig. 3A). On the basis of theposition of the band, the faster-migrating phosphorylated speciesappearing at 1 nM LTB4 likely corresponded to phosphorylation onSer310 (Fig. 2F), and the slower migrating band likely correspondedto phosphorylation on both Ser310 and Thr308. These results raise thepossibility that each residue that is phosphorylated in an LTB4-dependent manner can act as a sensor for different LTB4 concentra-tions. Moreover, the total phosphorylation of BLT1/0N was initiallyenhanced by 1 nM LTB4, followed by a gradual increase, with themaximal phosphorylation observed at concentrations more than30 nM (Fig. 3B). These data further suggest that the phosphorylationof each LTB4-induced phosphorylation site could trigger additionalphosphorylation events at the sites initially identified as basal phos-phorylation sites.

Stepwise addition of LTB4 enhances BLT1 phosphorylationThe cell responses to activation of BLT1, such as an increase in in-tracellular [Ca2+], are immediately shut down after the initial stim-ulation of LTB4 (23). The rise in intracellular [Ca

2+] in HeLa cells


expressingHA-BLT1/0N in response to 100 nMLTB4 decreasedwhenthe cells were prestimulatedwith 10 nMLTB4 (Fig. 4A). Exposing cellsto 100 nM LTB4 after stimulation with 10 nM LTB4 increased the ex-tent to which BLT1/0N was phosphorylated compared to stimulationwith 10 nM LTB4 alone (Fig. 4B) and did not stimulate additional ac-tivation of extracellular signal–regulated kinases 1 and 2 (ERK1/2) af-ter the cells were desensitized by the primary stimulation (Fig. 4B).Similar results were obtained in CHO-K1, Ba/F3, and RBL-2H3 cells(Fig. 4C). In HeLa cells expressing BLT1/0N, the additional phospho-rylation in response to 100 nM LTB4 treatment after prestimulationwas greatly diminished by cotreatment with an equal amount of BLT1antagonist, CP105696 or ZK158252 (Fig. 4, D and E) (24). Previously,Okuno et al. found that the conformation of human BLT1 switchesfrom a high-affinity state to a low-affinity state after activation ofthe downstreamGproteins (25). These results raise the possibility thata low-affinity form of BLT1 remains on the cell surface after ligandstimulation and could be modified by additional phosphorylationevents upon exposure to higher LTB4 concentrations. To investigatethe existence of BLT1 on the cell surface after stimulation with 10 nMLTB4, we fluorescently labeled BLT1 using sortase A (SrtA)–mediatedtranspeptidation, known as sortagging (26, 27), and examined thetrafficking of the labeled BLT1 after stimulationwith LTB4. It has beenreported that only a portion (~30%) of BLT1 undergoes internaliza-tions after stimulationwith LTB4 (17, 19, 28, 29). In keeping with thesefindings, a substantial amount of both the labeled BLT1/0NandBLT1/0NDphos remained on the surface of Ba/F3 cells at least 20 min afterstimulation with 10 nM LTB4, with only ~5% detected in the intra-cellular vesicles. These data suggest that BLT1 remains on the cell sur-face after ligand stimulation and thus could be modified by additionalphosphorylation events upon exposure to higher LTB4 concentrations(Fig. 4, F and G).

A

B

BLT1/0N(IB: HA)

BLT1/0N b-phos(IB: HA)

LTB4 (nM) 0 0.5 1 10 30 50 60 70 80 100

LTB4 (nM) 0 0.5 1 10 30 50 60 70 80 100

-Actin

-Actin

WC

L

Mem

bran

e

Mem

bran

e

WC

L

p-Ser310

p-Ser310 + p-Thr308

Fig. 3. Enhancement of BLT1 phosphorylation by ligand in a dose-dependent manner. (A and B) HeLa cells expressing HA-BLT1/0NDb-phos, whichlacks the five basal phosphorylation sites (A) or HA-BLT1/0N (B), were stimulatedwith the indicated LTB4 concentrations for 40 min. Membrane fractions weresubjected to Phos-tag SDS-PAGE followed by immunoblotting for HA. Thepositions of BLT1 phosphorylated on Ser310 (p-Ser310) and Thr308 (p-Thr308) arenoted. b-Actin is an experimental and loading control. Blots are representativeof at least three independent experiments. p, phosphorylated.

4 of 15



Nakanishi et al., Sci. Signal. 11, eaao5390 (201

http://stke.sciencemD

ownloaded from

-

--s

--

,

-

-

r--lr

on April 3, 2021

ag.org/

LTB4-induced phosphorylation facilitates the degree of BLT1basal phosphorylationWe investigated whether LTB4-induced phosphorylation affected thedegree to which basal phosphorylation sites were modified during thestepwise addition of LTB4 in HeLa cells. After primary (10 nM) andstepwise (100 nM) additions of LTB4, a larger proportion of BLT1/0Nwas highly phosphorylated compared to nonphosphorylated BLT1/0N(Fig. 5, A and B). In contrast, the highly phosphorylated forms ofBLT1/Di-phos were substantially reduced, and the nonphosphorylatedforms were increased, compared to BLT1/0N, even after stimulationwith 100 nM LTB4 (Fig. 5, A and B). Because the total phosphorylationof BLT1 increased in parallel with the modification of each LTB4-induced phoshorylation site (Fig. 3), the reduction in highly phosphoryl-ated forms observed with BLT1/Di-phos could be due to a decrease inmodification of basal phosphorylation sites that depended on LTB4-induced phosphorylation events. Eliminating one of the two LTB4-induced phosphorylation sites (mutants BLT1/0N/S310 and BLT1/0N/T308) resulted in an increase in partial phosphorylations uponstepwise stimulation (Fig. 5C), supporting the requirement of LTB4-

8) 21 August 2018

induced phosphorylation for further phosphorylation at basal sites.Because the phosphorylation of Ser310 occurred at a lower LTB4 con-centration than did the phosphorylation of Thr308 (Figs. 2F and 5C),we propose a model for the sequential phosphorylation of humanBLT1 (Fig. 5D). Because BLT1 is exposed to a gradual increase inLTB4 concentration during the migration of neutrophils, the extentof phosphorylation may increase in parallel with these physiologicalchanges in LTB4.

Activation of Gi is required forLTB4-induced phosphorylationWe examined the functional significance of LTB4-induced stepwisephosphorylation of humanBLT1 inHeLa cells. First, surface traffickingsimilar to that of BLT1/0N was confirmed for each of three mutants,BLT1/0NDb-phos, BLT1/0NDi-phos, and BLT1/0NDphos, indicatingthat these mutations did not affect the surface trafficking of BLT1(fig. S5A). Previously, several reports demonstrated that GPCR-dependent chemotaxis by polymorphonuclear (PMN) leukocytes re-quires both ERK1/2 and the phosphoinositide 3-kinase (PI3K)–protein

A

10 nM LTB4 – + + + 10 nM LTB4 0 20 40 40 40 40 (min)

IB: HA

-Actin

Mem

bran

e

IB: HA

Mem

bran

e

WC

L

-Actin

WC

L

200

250

150

300

1 µM ATP

100 nM LTB4

1 µM ATP

100 nM LTB4

Pretreated with 10 nM LTB4 for 20 min

Pretreated with vehicle for 20 min

Intr

acel

lula

r [C

a2+] (

nM)

0.75

0.85

0.95

1.05

1.15

0 s 1200 s

BLT1/0N, plasma membrane

0.75

0.85

0.95

1.05

1.15

0 s 1200 s

BLT1/0N, intracellular

0.75

0.85

0.95

1.05

1.15

0 s 1200 s

BLT1/0N phos, plasma membrane

0.75

0.85

0.95

1.05

1.15

0 s 1200 s

BLT1/0N phos, intracellular*

*

* *

Nor

mal

ized

brig

htne

ss

Nor

mal

ized

brig

htne

ss

+ Ve

hicle

+ CP +

ZK + Ve

hicle

+ 100

nM LT

B 4

+ 100

nM LT

B 4 + C

P

+ 100

nM LT

B 4 + Z

K

Ba/F3RBL-2H3 CHO-K1

0 20 40

10 n

M

+100

nM

0 20 40

10 n

M

+100

nM

0 20 40

10 n

M

+100

nM

LTB4 (min)

IB:HA

-Actin

WC

LM

embr

ane

0 time 10 nM LTB4 20 min

WC

LM

embr

ane

B

C

D E

F

G

-Actin

Fig. 4. Enhanced phosphorylation of BLT1by sequential exposure to increasing concentrations of LTB4. (A) Intracellular [Ca

2+] inHeLa cells expressing HA-BLT1/0N that werepretreated with 10 nM LTB4 or vehicle beforebeing stimulated with 100 nM LTB4. The response to adenosine triphosphate (ATP) was included as a positive control. (B) HeLa cellexpressing HA-BLT1/0N were stimulated withLTB4 as indicated. Membrane fractions weresubjected to Phos-tag SDS-PAGE and immunoblotting for HA. WCLs were subjected to SDSPAGE and immunoblotting for ERK1/2 andphosphorylated ERK1/2 (p-ERK1/2). b-Actin is anexperimental and loading control. (C) RBL-2H3CHO-K1, and Ba/F3 cells expressing HA-BLT1/0Nwere stimulated with LTB4 as indicated. Membrane fractions were separated by Phos-tagSDS-PAGE and immunoblotted for HA. (D) HeLacells expressing HA-BLT1/0N were stimulatedwith LTB4 in the presence or absence of 100 nMCP105696 (CP) or ZK158252 (ZK). Membranefractions were separated by Phos-tag SDS-PAGEand immunoblotted for HA. (E) HeLa cells expressing HA-BLT1/0N were incubated with 10 nMLTB4 for 20min followedby the additionof 100nMLTB4 alone or 100 nM LTB4 plus the presence oabsence of 100 nM CP105696 or ZK158252. Membrane fractions were separated by Phos-tag SDSPAGE and immunoblotted for HA. (F) Confocamicroscopic images of surface and intracellulaBLT1 in Ba/F3 cells expressing SrtA-HA-BLT1/0Nor SrtA-HA-BLT1/0NDphos. The cytosol of thecells expressing SrtA-HA-BLT1/0N was stainedwith CytoRed (red), and the cytosol of cells

expressing SrtA-HA-BLT1/0NDphos was stained with ViVidFluor Cell Blue CMAC (blue). These cells were mixed and immobilized on a collagen-coated dish. The BLT1proteins were expressed with a SrtA tag, which allowed them to be labeled with Alexa Fluor 488 (AF488) on the cell surface (green). Cells were imaged immediately afterthe addition of 10 nM LTB4 (left) or 20min later (right). Clockwise from upper left: AF488-labeled BLT1/0N and BLT1/0NDphos (green), differential interference contrast imageof the cells, BLT1/0N in the cytoplasm (red), and BLT1/0NDphos in the cytoplasm (blue). Scale bars, 20 mm. (G) Changes in the distribution of AF488-labeled BLT1/0N and BLT1/0NDphos between the plasmamembrane and cytoplasm after 20min of LTB4 stimulation. Themean relative brightness values of the plasmamembrane and the intracellularcompartment in the cells expressing BLT1/0N (red) and BLT1/0NDphos (blue)were calculatedby averaging the 10 cell values in (F). Data aremeans ± SEM, n=10 cells. *P


on April 3, 2021


Dow

nloaded from

kinase B (also called AKT) pathways to facilitate proper cell polarityand motility (16, 30–36). On the other hand, an intracellular [Ca2+]increase and the PI3K-AKT pathway are required for PMN leukocytedegranulation (16, 30–36), but signaling through mitogen-activatedprotein kinases (MAPKs) such as ERK1/2 is not (fig. S5B) (37). Becausethe activation of Gi by BLT1 is a prerequisite for the initiation of bothmotility and degranulation, we assessed the importance ofGi activationwith respect to BLT1 phosphorylation. Pretreatment with PTX, whichprevents Gi proteins from being activated by GPCRs, before stepwiseLTB4 stimulation substantially reduced ligand-induced phosphoryl-ation of BLT1, suggesting that the activation of Gi is required for thesephosphorylations (fig. S5C).

[Ca2+] increase is a prerequisite forLTB4-induced phosphorylationWith respect to the relationship between the intracellular Ca2+ mobi-lization and BLT1 phosphorylation, there is no difference in the LTB4dependency of the Ca2+ influx between BLT1/0N and BLT1/0NDphos


(fig. S6), implying that the basal phosphorylation is not essential fortriggering a [Ca2+] response. The BLT1/0N/DE mutant, in which allphosphorylation sites are replaced with Asp or Glu, was also used tosimulate the fully phosphorylated form (38). The finding that BLT1/0N/DE requires higher doses of ligand to obtain a response similar tothat of BLT1/0N may be due to the reduced affinity for LTB4 (17).LTB4-induced phosphorylation of BLT1/0Nwas impaired in the pres-ence of 1 mM EGTA, demonstrating the necessity of the [Ca2+] in-crease for this modification (fig. S7). Similar results were obtainedby the treatment with 100 mM SFK96365, a Ca2+ channel blocker(fig. S7), indicating that the [Ca2+] increase is a prerequisite forLTB4-induced phosphorylation.

Lack of phosphorylation impairs chemotaxis induced byhigh LTB4 concentrationTo gain further insight into the biological importance of BLT1 phos-phorylation, we investigated chemotactic activity in CHO-K1 cellsexpressing human BLT1/0N or BLT1/0NDphos. Cells expressing

CA

B D

X: Higher than b-phos

LTB4 (min) 0 20 40 0 20 40

10 nM +1

00 nM

10 nM +1

00 nM

BLT1/0N BLT1/0N i-phos

Y: Non-phos

Mem

bran

eW

CL

IB: HA

-Actin

BLT1

/0N

b-ph

osBL

T1/0

Nb/

S310

BLT1

/0N

b/T3

08

LTB4 0 20 40 0 20 40 0 20 40(min)

+100

nM

10 nM+

100 n

M

10 nM +

100 n

M

10 nM

LTB4 0 20 40 0 20 40 0 20 40 (min)

10 nM +

100 n

M

BLT1

/0N

BLT1

/0N/S

310

BLT1

/0N/

T308

10 nM

+100

nM

10 nM

+100

nM

WC

L

-Actin

Mem

bran

e

IB: HA

WC

L

-Actin

Mem

bran

e

IB: HA

100

50

0 0 20 40 0 20 40

10 nM+100 nM

10 nM+100 nM

LTB4 (min)

BLT1/0N BLT1/0N i-phos

**** *

NS

Rat

io o

f int

ensi

ty (

%)

X

Y

QT……

AR

Gradient of LTB4

……

Helix 8

LTB4-phos

Basal-phos

Fig. 5. Facilitation of basal phosphorylation site modification by LTB4-induced phoshorylation. (A) HeLa cells expressing HA-BLT1/0N or HA-BLT1/0NDi-phos,which lacks both LTB4-induced phosphorylation sites, were stimulated with LTB4 as indicated. Membrane fractions were subjected to Phos-tag SDS-PAGE and immu-noblotted for HA. The two areas surrounded by dashed lines are defined as Y and X, the nonphosphorylated and highly phosphorylated bands, respectively. b-Actin isan experimental and loading control. (B) The signal intensities of areas X and Y were determined by densitometric analysis, and the ratios were expressed as thepercentage of the total intensity. Data are means ± SEM, n = 3 experiments. NS, not significant; *P < 0.05 and **P < 0.01 by one-way analysis of variance (ANOVA)followed by Tukey’s post hoc test. (C) HeLa cells expressing the indicated forms of HA-BLT1 were stimulated with LTB4 as indicated. Membrane fractions were subjectedto Phos-tag SDS-PAGE and immunoblotted for HA. All data are representative of at least three independent experiments. (D) Schematic model of the sequentialphosphorylations of BLT1. In the absence of LTB4, some basal sites are phosphorylated. Low concentrations of LTB4 elicit phosphorylation of Ser

310, which stimulatesadditional phosphorylation of basal phosphorylation sites. Higher concentrations of LTB4 stimulate phosphorylation ofThr

308, which, in turn, promotes further phos-phorylation events at basal sites in the C-terminal region.

6 of 15



BLT1/0NDphos exhibited impaired chemotaxis at 100 nM LTB4(Fig. 6, A and B) but exhibited no impairment of chemokinesis atthis concentration of LTB4 compared to cells expressing BLT1/0N(Fig. 6B). Real-time monitoring of cell migration using TAXIScan-FL revealed that 100 nM LTB4 significantly reduced the migration di-rectionality, but increased the migration velocity, of cells expressinghBLT1/0NDphos compared to cells expressing hBLT1/0N (Fig. 6C).Because BLT1/0N/DE-expressing cells showed a chemotactic activ-ity similar to that of cells expressing BLT1/0N (Fig. 6, A and C), theaberrant chemotaxis of BLT1/0NDphos at 100 nM LTB4 could be


due to the lack of phosphorylation. Previous reports demonstratedthat chemotaxis requires an LTB4-elicited intracellular Ca

2+ influxfollowed by the activation of MAPKs (30, 31). At 100 nM LTB4, cellsexpressing BLT1/0NDphos exhibited a sustained [Ca2+] increasecompared to the transient response of cells expressing BLT1/0N, al-though the peak Ca2+ influx was similar (Fig. 6D). Moreover, sus-tained activation of ERK1/2 was observed after stimulation with100 nM LTB4 in cells expressing BLT1/0NDphos, whereas activationwas transient in cells expressing BLT1/0N (Fig. 6E). Accordingly, weobserved increased expression of a MAPK signaling reporter in cells

on April 3, 2021


Dow

nloaded from
min 0 5 10 20 40 0 5 10 20 40
ERK1/2

ERK1/2

p-ERK1/2

p-ERK1/2

BLT

1/0N

BLT

1/0N

phos

10 nM LTB4 100 nM LTB4

-Actin

WC

L -Actin

–10 –9 – – – –

0.125

0.025

0.050

0.075

0.100

0–

0.150

Log [LTB4 (M)]

OD

595

MockBLT1/0NBLT1/0N phosBLT1/0N/DE

**

0.0

0.1

0.2

0.3

0.00

0.05

0.10

0.15

Chemotaxis Chemokinesis

BLT1

/0N

BLT1

/0N

BLT1

/0N

0.30

0.10

0 0

0.20

0.100.05

0.15

BLT1

/0N

BLT1

/0Nph

os

BLT1

/0Nph

os

BLT1

/0Nph

os

BLT1

/0Nph

os

NS

NSNS

***

LTB4 10 nM 100 nM 10 nM 100 nM

OD

595

OD

595

velocity

Mock

BLT1

/0N

BLT1

/0N

phos

BLT1

/0N/

DE0.0

0.1

0.2

0.3****

ns

**** *

directionality

Mock

BLT1

/0N

BLT1

/0N

phos

BLT1

/0N/

DE0.0

0.5

1.0

********

ns

****

NS NS

****** *** ****

***

U/s

U/s

Directionality Velocity

*

MA

PK

rep

orte

r ac

tivity

LTB4 (min) 0 5 20 25 30 40 0 5 20 25 30 40

10 nM

LTB 4

+ 100

nM LT

B 4

10 nM

LTB 4

+ 100

nM LT

B 4

p-Akt

Akt

BLT1/0N BLT1/0N phos

-Actin

WC

L

10 nM

LTB 4

10 nM

LTB 4

+100

nM LT

B 4

+100

nM LT

B

M

4

p-AKT

AKT

100

400

300

200

Intr

acel

lula

r [C

a2+] (

nM)

LTB4(10 nM)

LTB4(10 nM) LTB4

(100 nM)

BLT1

/0N BLT1

/0N

BLT1

/0Nph

os

BLT1

/0Nph

os1 min

LTB4(100 nM)

D

E F

G

A

B

C

V

Fig. 6. Effect of phosphorylation deficiency on chemotaxis. (A) Chemotaxis of CHO-K1 cells expressing the indicated forms of BLT1 in the presence of the indicatedconcentrations of LTB4. Chemotaxis was quantified by the number of cells that migrated across a porous barrier as measured by an optical density at 595 nm (OD595). Dataare means ± SEM, n = 3 experiments. **P < 0.01 by two-way ANOVA followed by Tukey’s post hoc test. (B) LTB4-induced chemotaxis and chemokinesis of CHO-K1 cellsexpressing the indicated forms of BLT1. Data aremeans ± SEM, n = 5 experiments. ***P < 0.001 by two-way ANOVA followed by Tukey’s post hoc test. (C) Directionality andvelocity of CHO-K1 cells expressing the indicated forms of BLT1 as they migrated along an LTB4 gradient (0 to 100 nM). Each data point represents an individual cell. Dataare means ± SEM, n > 20 cells. *P < 0.05; ***P < 0.001 by Kruskal-Wallis test with Dunn’s post hoc multiple comparisons tests. (D) Intracellular Ca2+ mobilization in responseto the indicated doses of LTB4 in CHO-K1 cells expressing the indicated forms of BLT1. (E) Western blot showing ERK1/2 and p-ERK1/2 at the indicated time points after LTB4stimulation inWCLs from CHO-K1 cells expressing the indicated forms of BLT1. b-Actin is an experimental and loading control. (F) MAPK reporter activity in LTB4-stimulatedHeLa cells expressing the indicated forms of BLT1, serum response element (SRE)–firefly luciferase (MAPK-responsive), and cytomegalovirus (CMV)–Renilla luciferase(constitutively expressed). Reporter activity was quantified as the ratio of firefly/Renilla luciferase activity. Data are means ± SEM, n = 3 experiments. *P < 0.05 by two-way ANOVA followed by Tukey-Kramer test. (G) Western blot showing AKT and phosphorylated AKT (p-AKT) in lysates from CHO-K1 cells expressing BLT1/0N or BLT1/0NDphos at the indicated points after stepwise LTB4 stimulation. All data are representative of at least three independent experiments.

7 of 15



Dow

nlo

expressing BLT1/0NDphos compared to cells expressing BLT1/0N(Fig. 6F). In contrast, these cells showed substantially reduced AKTactivation after the stepwise addition of 100 nM LTB4 (Fig. 6G). Thisabnormal signaling in cells expressingBLT1/0NDphos at high concen-trations of LTB4might be related to the aberrant chemotaxis displayedby these cells at high concentrations of LTB4.

Phosphorylation defect reduces the releaseof LTB4-induced b-hexosaminidasePrevious researchhas demonstrated that the LTB4-BLT1 axis is involvedin degranulation of both human PMN leukocytes (39) and RBL-2H3cells stably expressing BLT1 (16, 40). The PI3K-AKT pathway (36),but not the MAPK pathway (37), plays a key role in degranulation ofPMN leukocytes. Because signaling through BLT1/0NDphos producedan aberrant AKT response to LTB4 (Fig. 6G), we next investigatedwhether the phosphorylation deficiency in this mutant form of BLT1affected the LTB4-induced release of b-hexosaminidase, a mast celllysosomal enzyme that is released upon degranulation, in RBL-2H3cells. First, we assessed the effect of the AKT inhibitor MK-2206 onthe release of b-hexosaminidase by these cells. The activation of AKTobserved during the stepwise addition of LTB4 was markedly re-


duced by pretreatment with 5 mMMK-2206, although ERK1/2 acti-vation was not affected (Fig. 7A). The release of b-hexosaminidaseelicited by 100 nM LTB4 was significantly decreased, indicating thatAKT activation was required for LTB4-induced degranulation inRBL-2H3 cells (Fig. 7B). In RBL-2H3 cells expressing BLT1/0N,the release of b-hexosaminidase was significantly increased by thestepwise addition of LTB4, from 10 to 100 nM.However, the stepwiseaddition of LTB4 did not enhance the release of b-hexosaminidase inthe cells expressing BLT1/0NDphos (Fig. 7C). Consistent with this,in the BLT1/0NDphos-expressing cells, the activation of AKT inducedby LTB4was impaired (Fig. 7,D andE), as was also found for CHO-K1cells overexpressing humanBLT1/0NDphos (Fig. 6G). Together, thesedata demonstrate that the phosphorylation of BLT1 is pivotal for thepromotion of PI3K-AKT signaling, leading to b-hexosaminidase re-lease in RBL-2H3 cells.

Lack of phosphorylation impairs the function ofwild-type BLT1Wegenerated variousmutant forms of the prospectiveN-glycosylation–deficient receptor BLT1/0N (fig. S1A) and performed functional studiesusing these mutants, demonstrating aberrant migration and impaired

on April 3, 2021


aded from

CA

B

E

D

0 5 20 25 40 0 5 20 25 40

p-ERK1/2

ERK1/2

p-AKT

AKT

-Actin

WC

L

LTB4 (min)

10 nM

LTB 4

+100

nM LT

B 4

10 nM

LTB 4

+100

nM LT

B 4

Vehicle MK-2206

LTB4 (min) 0 5 20 25 30 40 0 5 20 25 30 40

p-AKT

AKT


WC

L

-Actin

10 nM

LTB 4

10 nM

LTB 4

+100

nM LT

B 4

+100

nM LT

B 4

LTB4 – + – +

***

p-A

KT

/AK

T r

atio

0

3.0

2.0

1.0

-Hex

osam

inid

ase

rele

ase

(%)

***

NS

Vehi

cle

10 n

M LT

B 4 2

0 m

in,

100

nM

LTB 4

20

min

10 n

M LT

B 4 4

0 m

in

5 M

A23

187

20 m

in

Vehi

cle

10 n

M LT

B 4 4

0 m

in

10 n

M L

TB4 2

0 m

in,

100

nM

LTB

4 2

0 m

in

10 n

M L

TB4 2

0 m

in,

100

nM

LTB

4 2

0 m

in

5 M

A23

187

20 m

in

-Hex

osam

inid

ase

rele

ase

(%)

******

40

30

10

20

0

Vehic

le

10 n

M L

TB4 2

0 m

in,

100

nM

LTB

4 2

0 m

in

10 n

M LT

B 4 4

0 m

in

10 n

M L

TB4 2

0 m

in,

100

nM

LTB

4 2

0 m

in

+

MK-

2206

5 M

A23

187

20

min

BLT1/0N BLT1/0N phos Mock


Fig. 7. Requirement of phosphorylation fordegranulation. (A) Western blot showing ERK1/2,p-ERK1/2, AKT, and p-AKT in lysates from RBL-2H3cells expressing HA-BLT1/0N and treated withLTB4 as indicated after pretreatment with vehicleor the AKT inhibitor MK-2206. (B) LTB4-stimulatedrelease of b-hexosaminidase from RBL-2H3 cellsexpressing HA-BLT1/0N in the presence of MK-2206 or the calcium ionophore A23187. Valuesfor b-hexosaminidase release into the mediumare expressed as the percentage of the totalb-hexosaminidase in the culture. Column 1, vehi-cle; column 2, 10 nM LTB4 for 40 min; column 3,10 nMLTB4 for 20min and then additionof 100nMLTB4 for 20 min; column 4, 10 nM LTB4 for 20 minand then addition of 100 nM LTB4 for 20min (pre-treated with MK-2206); column 5, 5 mM A23187for 20 min. Data are means ± SEM, n = 3 ex-periments. ***P < 0.001 by one-way ANOVAfollowed by Tukey’s post hoc test. (C) Release ofb-hexosaminidase in RBL-2H3 cells expressingthe indicated forms of BLT1 under the indicatedtreatment conditions. Columns 1 and 5, vehicle;columns2and6, 10nMLTB4 for40min; columns3, 7,and 9, 10 nM LTB4 for 20 min and then additionof 100nMLTB4 for20min; columns 4 and 8, A23187for 20 min. Data are means ± SEM, n = 3 exper-iments. ***P < 0.001 by one-way ANOVA followedby Tukey’s post hoc test. (D) Western blot showingAkt and p-Akt in lysates from RBL-2H3 cellsexpressing BLT1/0N or BLT1/0NDphos and treatedwith LTB4 as indicated. (E) Quantification of AKT ac-tivation in RBL-2H3 cells expressing the indicatedreceptors with or without stepwise LTB4 stimula-tion (10 nM LTB4 for 20 min followed by 100 nMLTB4 for 20min).Data aremeans±SEM,n=3exper-iments. ***P


on April 3, 2021


Dow

nloaded from

enzyme release resulting from the lack of BLT1 phosphorylation. Toeliminate issues regarding surface trafficking, stability, and ligand af-finitywhenusing the glycosylation-deficient BLT1s, we further carriedout similar functional experiments using human wild-type BLT1(BLT1/WT and BLT1/WTDphos), in which the glycosylation siteswere intact. There was no difference in the surface trafficking or sta-bility between BLT1/WT and BLT1/WTDphos (fig. S8, A and B). Weobserved similar intracellular [Ca2+] increases upon LTB4 stimulationin cells expressing BLT1/WTor BLT1/WTDphos, although the [Ca2+]response in cells expressing BLT1/WTDphos was slightly sustained at100 nM, similar to the nonglycosylated mutant (fig. S8C). In contrast,we found significant reductions in chemotaxis and b-hexosaminidaserelease in BLT1/WTDphos elicited by 100 nM LTB4 (fig. S8, D andE). All these data are consistent with those from experiments usingglycosylation-deficient receptors, indicating that the results are notspecific to glycosylation-deficient receptors and supporting the bio-logical importance of phosphorylation in native glycosylated formsof the protein.

Phosphorylation is necessary for thelow-affinity conformationPreviously, Okuno et al. indicated that human BLT1 changes confor-mation from a high-affinity to low-affinity form after the activation ofcoupled G proteins. Because BLT1s harboring a mutated helix 8,BLT1/DH8 and BLT1/LLAA (fig. S9A), show increased [3H]LTB4binding compared to wild-type BLT1, this region is likely to be impor-tant for the conformational change to the low-affinity state (25, 41).Here, we obtained evidence supporting the notion that the disruptionof helix 8 leads to defective BLT1 phosphorylation (fig. S9B); thus, wepropose the involvement of phosphorylation in the conformationalchange. Moreover, phosphorylation was necessary for the expressionof a reporter gene that is activated by the low-affinity conformation(fig. S10A) and for the shedding of transforming growth factor–a(TGFa), which requires high concentrations of [LTB4] (fig. S10B)(42). In agreement with the prediction that phosphorylation is re-quired for BLT1 to undergo the conformational change from the high-to low-affinity state, we found that BLT1/WT/DE required a high doseof LTB4 for its activation, compared towild-type, indicating that it wasin a low-affinity conformation. In contrast, BLT1/WTDphos showedhyperactivity, indicating the high-affinity form, in both the reporterassay and the TGFa shedding assay. Collectively, these data supportour hypothesis that primary BLT1 phosphorylation is essential for thehigh-to-low affinity conformational change (Fig. 8).

We could not rule out the possibility that a distinct signaling pathwaythrough the phosphorylated form is triggered by a high dose of LTB4.The phosphorylation may play a role in triggering the recruitment ofeffectors, such as b-arrestins. Here, we found the reduced binding ofb-arrestins, b-arrestin–1 and b-arrestin–2, to LTB4-stimulated BLT1/WTDphos, although substantial amounts of b-arrestins remainedbound (fig. S11). These results are consistent with previous findingsregarding the phosphorylation-independent recruitment of b-arrestinbeing essential for the internalization of BLT1 (17).

DISCUSSIONIt has been demonstrated that the distinct GPCR phosphorylationprofiles that result from the actions of different kinases are responsiblefor various signaling outcomes. This concept is known as the “phos-phorylation barcode” and is critical for the multidimensional features


of GPCR signaling (3–6, 43–45). Thus, the precise analysis of BLT1phosphorylation is important to clarify the biological functions of thisreceptor. Many studies have attempted to understand the potentialroles of BLT1 phosphorylation (17–19); however, a comprehensiveanalysis of these phosphorylations is still lacking. Here, we aimed tomore extensively characterize the mapping, dynamics, and biologicalimportance of BLT1 phosphorylation events by using Phos-tag tech-nology in combination with site-directed mutagenesis. Although thisapproach does not provide direct evidence for the phosphorylation ofspecific sites, as would be revealed by mass spectrometry, it identifiesthe precise phosphorylation sites involved (46–48). Here, wedetermined the precise sites of basal and LTB4-induced phosphoryl-ations in human BLT1. Furthermore, we demonstrated hierarchicaland sequential phosphorylation of BLT1 in response to increasingconcentrations of LTB4. This work demonstrates that Phos-tag SDS-PAGE analysis can be used to assess GPCR phosphorylation and maybe useful for research on other GPCRs.

Here, we demonstrate that the lack of BLT1 phosphorylationleads to aberrant neutrophil migration and impaired enzyme release(Fig. 9A). It was reported that RBL-2H3 cells overexpressing a hu-man BLT1-Ala mutant, in which all the Ser and Thr residues in theC terminus were replaced with Ala, responded to various concentra-tions of LTB4 similarly to cells overexpressing the wild-type receptorin a chemotaxis assay (17). There are no clear explanations for thisdiscrepancy; however, the specific cell lines, mutated residues, andthe experimental conditions (96-well versus 48-well chambers) couldaccount for these differences. Previous reports demonstrated the im-portance of the ligand-elicited intracellular Ca2+ influx followed by theactivation of ERK1/2 in BLT1-stimulated chemotaxis (30, 31). For ex-ample, Ichiki et al. found that lack of the receptor for advanced glyca-tion end products (RAGE) caused attenuation of the BLT1-dependentactivation of ERK1/2, leading to reduced velocity of neutrophil motil-ity, although the directionality of migration was slightly enhanced

Gi (GTP)

LTB4

Cell responses (chemotaxis, etc.)

Gi (GTP)Gi (GDP) -S310

Cell responses(degranulation, etc.)

High affinity Low affinity

Gi (GDP)P P P P P P

P PP PPP PP

(i) (ii) (iii) (iv)

T308S310

Fig. 8. Sequential BLT1 phosphorylation. (i) In the absence of LTB4, the basalphosphorylation sites in the BLT1 C-terminal domain are variably phosphorylated,yielding a mixture of phosphorylated species. The receptor is not active, and thedownstream Gi protein is bound to guanosine diphosphate (GDP). (ii) Low concen-trations of LTB4 stimulate BLT1 to adopt a high-affinity conformation, which stimu-lates GDP-GTP (guanosine triphosphate) exchange on Gi and triggers primaryneutrophil responses such as the initiation of chemotaxis. (iii) After primarysignaling, Gi exchanges GTP for GDP, and Ser

310 is phosphorylated. Additionalphosphorylation of basal sites in BLT1 then occurs, leading to a return to a low-affinity state. (iv) As neutrophils migrate up the LTB4 gradient, higher LTB4 concen-trations stimulate phosphorylation of The308, which, in turn, promotes additionalphosphorylation events at the basal sites, leading to secondary responses, suchas degranulation.

9 of 15



on April 3, 2021


Dow

nloaded from

(49). In contrast, we observed a sustained increase in ERK1/2 activity incells expressing BLT1/0NDphos after stimulation with 100 nM LTB4,resulting in the opposite behavior—a decline in directionality and in-creased velocity. These findings highlight the importance of ERK1/2signaling for proper cell motility. There is evidence suggesting that cellpolarity in neutrophil chemotaxis is controlled by phosphatidylinositol3,4,5-trisphosphate [PtdIns(3,4,5)P3] (32–35, 50). Nishio et al. demon-strated that during the initiation of chemotaxis, neutrophils becomepartially polarized without PtdIns(3,4,5)P3 and are able to migratewith correct directionality. Thereafter, PI3Kg, phosphatase and tensinhomolog (PTEN), and Src homology 2 domain–containing inositol-5-phosphatase 1 (SHIP1) cooperatively localize PtdIns(3,4,5)P3 to theprospective leading edge. This process leads to the assembly of theleading edge and full polarization, which are critical for proper direc-tionality (32). Here, the impaired directionality of the cells expressing


BLT1/0NDphos in chemotaxis could be due to the reduced productionof PtdIns(3,4,5)P3, resulting in loss of PtdIns(3,4,5)P3 accumulation atthe leading edge (Fig. 9B). The aberrant chemotaxis could also be ex-plained by the reduced desensitization of the [Ca2+] increase andERK1/2 activation. These impairments may directly or indirectly af-fect the PI3K-AKT pathway. Further studies with phosphorylation-deficient BLT1, which is hyperactive, as well as the phosphomimicform, are likely to unravel the molecular mechanism(s) underlyingthese events. Together, our results indicate that the abnormal signalingin cells expressing BLT1/0NDphos might be involved in deficient mi-gration at high LTB4 concentrations.

Furthermore, we found a deficiency of b-hexosaminidase release incells expressing BLT1/0NDphos. Because b-hexosaminidase releaserequires a high dose of LTB4 (ca. >100 nM) (16, 37, 40), the phenom-enon is likely triggered through the low-affinity form of BLT1, harbor-ing increased phosphorylation. In keeping with this idea, a substantialreduction in b-hexosaminidase release was induced by 100 nM LTB4in RBL-2H3 cells expressing BLT1/0NDphos. Because the PI3K-AKTpathway, but not the ERK1/2 pathway, plays a key role in the de-granulation of PMN leukocytes (36, 37), we suggest that the reducedb-hexosaminidase release could be, at least in part, due to impairmentof the PI3K-AKT pathway.

Human BLT1 is highly abundant in phagocytic and immune cells,especially neutrophils (13), but it is difficult to use human neutrophilsfor the analysis of BLT1 mutants at present because of the low trans-fection efficiency. Here, we identified phosphorylation sites in mouseBLT1 and performed functional studies in cultured cells. To providefurther support for our conclusions, we hope to generate knock-inmice expressing phosphorylation-deficient forms of BLT1 in our fu-ture work to elucidate the in vivo role of BLT1 phosphorylation, simi-larly to what has been done for other GPCRs (51). Extensive studiesusing these mice, such as assessing the effect of the in vivo LTB4gradient as found in many disease models in mice, will be requiredto clarify the regulatorymechanisms of BLT1 involved in chemotaxisand degranulation. Recently, Majumdar et al. reported that theexosomal pool of LTB4 acts in an autocrine fashion to sensitize neu-trophils to the primary chemoattractant and in a paracrine fashion tomediate the recruitment of neighboring neutrophils (52). To furtherclarify the relationship between the exosome-mediated LTB4 relayand phosphorylation-dependent regulation of BLT1 in chemotaxis,these mice will be useful tools in our future works. In addition, tofurther clarify the involvement of BLT1 phosphorylation in the de-granulation in vivo, it is important to discriminate between patho-physiogical degranulation to the extracellular milieu and granulemobilization into the phagocytic lysosomal vacuole. In neutrophils,there are several types of intracellular granules that can be mobilizedfor releasing their contents into the primary phagocytic vacuole. Forexample, innate immune phagocytes do not degranulate in responseto LTB4 under physiologic conditions. Because the release of en-zymes to the extracellular milieu occurs only under pathophysiologicconditions, it is crucial to distinguish between specific and azurophilicgranules in neutrophils and to clarify theirmobilization to the primaryphagocytic vacuole in response to LTB4. Although it is challenging tointroduce the various BLT1 mutants into human phagocytic cells, theBLT1 phosphorylation–deficient knock-in mice will be ideal tools toaddress these issues.

Antibodies that are able to recognize specific phosphorylatedforms of BLT1 would be useful tools for elucidating the spatial andtemporal dynamics of individual phosphorylation events as well

Gradient o

f LTB4

Gradient of L

TB4

PI3K-independent phase PI3K-dependent phase

PtdIns(3,4,5)P3PTENSHIP1

BLT1 / WT

Gradient of L

TB4

BLT1/ phos

A

B

Wild type

Chemotaxis

Degranulation

Neutrophil

phos-mutant

Low

High

ChemotaxisP P

P PPPPPPP

PPP

Fig. 9. Essential roles of BLT1 phosphorylation. (A) Highly phosphorylatedBLT1 is important for the precise directionality of chemotaxis and degranulationat higher LTB4 concentrations. The absence of these phosphorylations results inaberrant migration and impaired enzyme release. (B) Schematic representation ofthe proposed model for PtdIns(3,4,5)P3-dependent polarization in the chemotaxisof neutrophils. During the initiation of chemotaxis, neutrophils become partiallypolarized without PtdIns(3,4,5)P3 and are able tomigrate with correct directionality.Thereafter, PI3Kg, PTEN, and SHIP1 cooperatively localize PtdIns(3,4,5)P3 to the pro-spective leading edge. This process leads to the assembly of the leading edge andfull polarization, both of which are critical for proper directionality. The impaireddirectionality of cells expressing phosphorylation-deficient BLT1 could be due tothe reduced production of PtdIns(3,4,5)P3, resulting in the loss of PtdIns(3,4,5)P3accumulation at the leading edge where LTB4 concentrations are greatest.

10 of 15



on April 3,


Dow

nloaded from

as to identify the kinases involved. The generation of a panel ofphosphorylation site–specific antibodies would enable the studyof this modification in situ, allowing insights into dynamic proteinphosphorylations in the spatially complex structure of cells (53).Such antibodies would also be useful for identifying the kinases re-sponsible for thesemodifications, which is also critical for understand-ing the biological importance of BLT1 phosphorylation.

In 1984, Goldman and Goetzl revealed the existence of high-affinity(Kd = 0.36 nM) and low-affinity (Kd = 76 nM) forms of a putative LTB4receptor in human PMN leukocytes (39). They further demonstratedthat PMN leukocytes, whichwere deactivated by 10nMLTB4, selective-ly lost the high-affinity binding sites although low-affinity binding sitesremained. Hence, the PMN leukocytes pretreated with 10 nMLTB4 ex-hibited a substantially diminished chemotactic response to subsequentLTB4 stimulation, whereas b-hexosaminidase release was still observedwhen elicited by higher LTB4 (300 nM). The findings of the currentstudy show that the loss of high LTB4 affinity after pretreatment withLTB4 is likely due to the phosphorylation of BLT1, followed by a con-formational change to a low-affinity state. The remaining low-affinityform acts as a LTB4 sensor and serves to release b-hexosaminidase oncethe receptor is exposed to a higher dose of LTB4.

During the migration of neutrophils, the concentration of LTB4to which BLT1 is exposed gradually increases as the cells migratetoward the source. In other words, the degree of BLT1 phosphoryl-ation can be enhanced in parallel with the concentration of LTB4 suchthat the graded increase in the phosphorylation of BLT1 acts as aLTB4 sensor. BLT1 is not phosphorylated and in the high-affinityconformation when the cell encounters low concentrations of LTB4,which trigger cellmigration. As neutrophilsmigrate toward the source,BLT1 becomes progressivelymore phosphorylated and switches to thelow-affinity conformation. When BLT1 is exposed to a concentrationof LTB4 high enough to activate the low-affinity form, it triggers othercell responses, such as degranulation. Althoughmore extensive studiesare needed to better understand the functional significance of BLT1modifications, we suggest that basal and LTB4-induced phosphoryl-ations are indispensable for BLT1 maturation during cell migrationto potentiate degranulation at inflammatory loci.

2021

MATERIALS AND METHODSMaterialsLTB4 and the antibody recognizing human BLT1 (#120114) werepurchased fromCaymanChemical Company. The antibody recogniz-ing theHA tag (clone 3F10), PNGase-F, and the cOmplete EDTA-freeprotease inhibitor cocktail were purchased from Roche Applied Sci-ence. Phycoerythrin (PE)–conjugated antibody recognizing rat im-munoglobulin G (IgG) was from Beckman Coulter. Phos-tag waspurchased from NARD Institute Ltd. Antibodies recognizing phos-phorylated p44/42 MAPK (ERK1/2) (Thr202/Tyr204), p44/42 MAPK(ERK1/2), phospho-AKT (Ser473), and AKT were from Cell SignalingTechnology. The b-actin antibody, ATRA, and calcium ionophoreA23187 were from Sigma-Aldrich Corporation. PTX and MK-2206were from Funakoshi. SKF96365 was from Wako. BLT1 antagonistsCP 105696 and ZK 158252 were gifts from Pfizer Inc. and ScheringAG, respectively.

Construction of mutant human BLT1 expression plasmidsN-terminally HA-BLT1 was prepared as described previously (25).The BLT1mutants were constructed by overlap extension polymerase


chain reaction (PCR) using HA-BLT1 as a template (54). The ampli-fied PCR products digested with Bam HI and Eco RI were subclonedinto the pcDNA3 vector, and the sequences were confirmed. Theprimer sets used are listed in table S1.

Cell culture and transfectionHeLa cells and RBL-2H3 cells were cultured in Dulbecco’s modifiedEagle’s medium (DMEM; Nacalai Tesque) supplemented with 10%fetal bovine serum (FBS). CHO-K1 cells were cultured inHam’s F-12medium (Nacalai Tesque) supplemented with 10% FBS. HL-60 cellsand Ba/F3 cells were cultured in RPMI 1640 (Nacalai Tesque) sup-plemented with 10% FBS. These cells were transfected with a plasmidharboring the wild-type or mutant receptors using Lipofectamine2000 (Invitrogen) according to the manufacturer’s protocol. For dif-ferentiation of HL-60 cells, these cells were cultivated for 4 daysin the presence of 1 mM ATRA and then stimulated with or without100 nM LTB4.

Stable expression of BLT1HeLa cells, CHO-K1 cells, and RBL-2H3 cells stably expressing wild-type ormutant receptors were established as follows. These cells weretransfected with the HA-tagged receptor-expression plasmid. Stabletransfectants were selected with geneticin (1 mg/ml) for 20 days. Theexistence of theHA epitope on the cell surface was confirmed by flowcytometry. After labeling the drug-resistant cells with PEusing antibodyrecognizing HA tag (clone 3F10) and PE-conjugated antibody recog-nizing rat IgG, a group ofHA-positive cells was sorted usingMACSCellSeparation Columns and Anti-PE MicroBeads (Miltenyi Biotec).

Flow cytometryFor staining, cells were incubated with the antibody recognizing theHA tag (clone 3F10) in phosphate-buffered saline (PBS) containing2% goat serum at room temperature for 30 min, followed by stainingwith PE-conjugated antibody recognizing rat IgG at room tempera-ture for 30 min. The EPICS XL (Beckman Coulter) was used for flowcytometry.

Calcium mobilization assayTransiently transfected HeLa cells were plated in a 96-well plate (4 ×104 cells per well) and incubated for 16 hours. The cells were thenincubated with loading buffer A [1× Hanks’ balanced salt solution(HBSS), 1.25 mM probenecid (Sigma-Aldrich Corporation),1 mM CaCl2, 0.5 mM MgCl2, 0.1% (w/v) bovine serum albumin(BSA), 20 mM Hepes-NaOH (pH 7.4)] containing 4 mM Fluo-3AM (Dojindo) and 0.04% pluronic acid (Molecular Probes) at37°C for 1 hour. The cells were then washed once with buffer A.Intracellular Ca2+ mobilization was monitored with the FlexStationscanning fluorometer (Molecular Devices) bymeasuring the emissionfluorescence at 525 nm in response to excitation at 485 nm. Relativefluorescence units (maximum-minimum) are indicated.

To investigate the desensitization and shutdown of intracellularCa2+ responses, cells were loaded with 3 mM Fura-2 AM (Dojindo)in a modified Hepes-Tyrode’s BSA buffer [25 mM Hepes-NaOH(pH7.4), 140mMNaCl, 2.7mMKCl, 1.0mMCaCl2, 12mMNaHCO3,5.6 mM D-glucose, 0.37 mM NaH2PO4, 0.49 mM MgCl2, 0.1% (w/v)BSA] containing 0.01% pluronic acid (Molecular Probes) at 37°C for1 hour. The cells were washed and resuspended in Hepes-Tyrode’sBSA buffer at a density of 1 × 106 cells/ml. The cell suspension(0.5 ml) was applied to a CAF-100 system (JASCO Corp.), and 5 ml

11 of 15



on April 3, 2021


Dow

nloaded from

of ligand solution was added. The ligands were dissolved in Hepes-Tyrode’s BSA buffer after evaporation of the stock solution in ethanolunder a nitrogen stream and used immediately. The intracellular Ca2+

concentration was measured by determining the ratio of emission flu-orescence of 500 nm by excitation at 340 and 380 nm. The free Ca2+

concentrationwas calculated from the equationCa2+i =Kd[(F − Fmin)/(Fmax − F)], where Kd represents the Ca

2+ binding dissociation con-stant (224 nM for Fura-2), F is the 500-nm fluorescence ratio, Fmax isthe maximal fluorescence ratio determined after the addition of 0.1%Triton X-100 to permeabilize the cells in the presence of 1 mM Ca2+,and Fmin is the minimal fluorescence ratio determined after permea-bilization and the addition of 5 mM EGTA.

Preparation of cell membrane fractionsAfter stimulation of the cells expressingmutant orwild-type BLT1with100 mM LTB4, the cells were washed with ice-cold tris buffer [20 mMtris-HCl (pH 7.4), 0.3 M sucrose, 1 mM sodium orthovanadate] andcollected in Hepes buffer [25 mM Hepes-NaOH (pH 7.4), 10 mMMgCl2, 0.3 M sucrose, 1 mM sodium orthovanadate, supplementedwith cOmplete protease inhibitor cocktail]. Subsequently, the cellswere sonicated for 20 min using a Bioruptor (Cosmo Bio) and cen-trifuged at 9000g for 10 min at 4°C to remove cellular debris. Theresulting supernatant was ultracentrifuged at 100,000g for 1 hourat 4°C. The microsome pellet was resuspended in a buffer containingthe abovementioned Hepes buffer. The samples were then subjectedto SDS-PAGE or Phos-tag SDS-PAGE as cell membrane fractions.

Phos-tag SDS-PAGE and immunoblottingFor Western blot analyses, protein samples were separated on SDS-PAGE gels and transferred to a nitrocellulose membrane. To detectthe band shift that represents phosphorylated proteins, an SDS-PAGEgel containing 50 mM Phos-tag acrylamide with 200 mM MnCl2 wasused. After a blocking step using 1% skimmedmilk in TBS-T [20 mMtris-buffered saline (pH 7.4) and 0.1% Tween 20], blots were probedwith the primary antibody [antibody against HA tag (3F10) or anti-body recognizing human BLT1] for 1 hour. The membrane waswashed with TBS-T and incubated with horseradish peroxidase–conjugated antibody against rat (for 3F10) or mouse (for humanBLT1 antibody) IgG for 30 min. The signal was visualized usingan enhanced chemiluminescence Western blotting detection system(Nacalai Tesque).

Endoglycosidase treatmentProtein samples containing BLT1 were obtained from the cell mem-brane fractions described above. Samples were treated with 1 U ofPNGase-F in 50 ml of buffer [139.2mMNa2HPO4, 40.8mMNaH2PO4,0.4% SDS, 20 mMEDTA, 2% 2-mercaptoethanol] for 16 hours at 4°C.The resultant samples were suspended in a sampling buffer [25 mMtris-HCl (pH 6.5), 5% glycerol, 1% SDS, 0.05% bromophenol blue]and subjected to Western blot analyses.

AP treatmentProtein samples containing wild-type and mutant BLT1 were ob-tained from the microsome pellet described above without sodiumorthovanadate. Samples were suspended in buffer [25 mM Hepes-NaOH(pH7.4), 10mMMgCl2, 0.3M sucrose, and cOmplete proteaseinhibitor cocktail]. The samples were then treated with 60 U of CIAP(Takara Bio) in 50 ml of AP buffer [50 mM tris-HCl (pH 9.0) and1 mM MgCl2] for 1 hour at 37°C.


Construction of SrtA-HA–tagged BLT1 expression plasmidsThe linearized vector was prepared by PCR using two primers (pMK-HA, forward and reverse) derived frompMK-SrtA-HA-G2A-IRES-purothat encodes the LPETGGGGG(SrtA)-HA–tagged G2 accumulation(G2A) receptor and an internal ribosomal entry site (IRES)–puromycincassette (27). BLT1/0N and BLT1/0NDphos were amplified by PCRusing two primers (BLT1/0N, forward and reverse), respectively. Then,the linearized vector and the insert were fused using the In-Fusion HDenzyme (Clontech), generating pMK-SrtA-HA-BLT1/0N-IRES-puroand pMK-SrtA-HA-BLT1/0NDphos. The constructs were confirmedby sequencing.

Analysis of AF488-labeled BLT1 traffickingThe cytosol of the Ba/F3 cells expressing SrtA-HA-BLT1/0N andSrtA-HA-BLT1/0NDphos were stained with CytoRed and ViVidFluorCell Blue CMAC, respectively. These cells were mixed the same pop-ulation, immobilized on a biocompatible anchor for membrane–collagen–coated glass base dish (Matsunami) (55) and preincubated inRPMI 1640 (0.1% BSA) for 1 hour at 37°C, 5% CO2. SrtA-mediatedlabeling of BLT1 was carried out as follows (27). The medium was re-placed with RPMI 1640 (0.1% BSA, 30 mM SrtA, 1 mMGly3) followedby incubation for 1 hour at 37°C, 5% CO2. After washing the cells withRPMI 1640 (0.1% BSA) to remove excess Gly3, the cells were incu-bated for labeling BLT1 with AF488 in RPMI 1640 (0.1% BSA, 15 mMAF488-DLPETGG, 30 mM SrtA) for 1 hour at 37°C, 5% CO2. Thecells were washed twice with RPMI 1640 (0.1% BSA), and then RPMI1640 (0.1% BSA) was added at 1 ml per dish for confocal microscopyobservation. The single-cell analysis of AF488-labeled BLT1/0N andBLT1/0NDphos was simultaneously performed in RPMI 1640 (0.1%BSA). Time-lapse images of AF488-labeled BLT1/0N and BLT1/0NDphos were acquired using an LSM 510 confocal microscopeequipped with an incubation system to maintain 37°C. Immediatelyafter the addition of LTB4 (10 nM final concentration), the time-lapseimage was acquired for 20 min at 1-min intervals. The acquiredimages were automatically analyzed using OpenCV.

Chemotaxis assayPolycarbonate filters with an 8-mm pore size (Neuro Probe) werecoated with fibronectin (13.3 mg/ml) (Sigma-Aldrich Corporation)in PBS for 60 min. A dry coated filter was placed on a 96-well blindchamber (Neuro Probe) containing LTB4 (10

−10 to 10−6 M). CHO-K1cells expressingwild-type andmutant BLT1/0N (200 ml, 8 × 104 cells perwell) were added to the top wells. The ligand solution and cell suspen-sion were prepared in the same buffer [Ham’s F-12medium containing0.1% (w/v) BSA]. After incubation at 37°C in 5% CO2 for 3 hours, thefilter was disassembled. The cells on the filter were fixed with methanoland stained with a Diff-Quick staining kit (International Reagent). Theupper side of the filter was then scraped free of cells. The number of cellsthatmigrated to the lower side was determined bymeasuring theOD595using a 96-well microplate reader.

To evaluate motility during chemotaxis, CHO-K1 cells ex-pressing wild-type and mutant BLT1/0N were allowed to migratein the chamber of a TAXIScan-FL optical assay device (Effector CellInstitute), along with the LTB4 gradient (0 to 100 nM). Phase-contrastimages of migrating cells were acquired at 10-s intervals over 15 min.Images were imported as stacks to ImageJ (National Institutes ofHealth). Velocity and directionality were calculated with manualtracking using the chemotaxis and migration tools in the ImageJprogram.

12 of 15



on April 3, 2021


Dow

nloaded from

Reporter gene assayHeLa cells were plated in a 24-well plate (1 × 105 cells per well) andtransfected with 0.8 mg per well of plasmid DNA using Lipofectamine2000 (Invitrogen). The ratio of transfected plasmid DNA was keptconstant at 17:17:1 for SRE–firefly luciferase–pGL4.33 (Promega):BLT1/0N-pcDNA3 or BLT1/0NDphos-pcDNA3:CMV promoter-driven Renilla luciferase (Promega) (56). The cells were starved for15 hours and stimulated with 100 nMLTB4 for 6 hours. The luciferaseassay was performed using the Dual-Luciferase Reporter Assay Sys-tem (Promega) and the MiniLumat LB 9506 luminometer (Berthold).SRE-luc activity was normalized to the Renilla luciferase activity.

For the analyses of BLT1/WT, BLT1/WTDphos, and BLT1/WT/DE, human embryonic kidney (HEK) 293T cells were plated in a24-well plate (1 × 105 cells per well) and transfected with 0.8 mg perwell of plasmid DNA using Lipofectamine 2000 (Invitrogen). Theratio of transfected plasmid DNA was kept constant at 17:17:1 forSRE-NanoLuc luciferase-pNL(NLucP/SRF/Hygro; Promega):BLT1-pcDNA3:phosphoglycerate kinase (PGK)–firefly luciferase–pGL4.53(luc2/PGK; Promega). The cells were starved for 15 hours and stimu-lated with 10 or 100 nM LTB4 for 6 hours. The luciferase assay wasperformed using the Nano-Glo Dual-Luciferase Reporter Assay Sys-tem (Promega) and the GloMax 20/20 luminometer (Promega). SRE-NanoLuc activity was normalized to the levels of firefly luciferaseactivity.

b-Hexosaminidase release assayDegranulation was determined by measuring the release of a granulemarker, b-hexosaminidase, as described previously (16). The cellswere seeded in 24-well collagen-coated plates at 5 × 105 cells in1 ml of medium and were serum-starved for 12 hours. Adherent cellswere washed twice in prewarmed PBS. Cells were then treated withLTB4 (10 nM for 40min, or 10 nM for 20min followed by 100 nM for20 min), and the activity of b-hexosaminidase was quantified in thesupernatant and the cell lysates by spectrophotometric analysis usingp-nitrophenyl-N-acetyl-b-D-glucosaminide as a substrate (57). Valuesfor b-hexosaminidase release in the medium were expressed as thepercentage of the total b-hexosaminidase, determined from cells lysedin 0.1%TritonX-100. In these experiments, calcium ionophoreA23187(5 mM) was used as a positive-control stimulus to observe the releaseof b-hexosaminidase.

Pulldown of biotin-labeled HA-BLT1HeLa cells were plated in a 60-mm dish (1 × 106 cells per dish) andtransfected with 6 mg of HA-BLT1/WT or HA-BLT1/WTDphosusing Lipofectamine 2000 (Invitrogen). After 24 hours, the cells werewashed once with PBS and then incubated in 3 ml of PBS containing0.5 M EZ-Link Sulfo-NHS-SS-Biotin (Thermo Fisher Scientific) for30 min at 4°C. After washing with PBS, the cells were incubated in aculture medium and stimulated with 10 nMLTB4 for 30min at 37°C.Then, the cells were lysed in 500 ml of lysis buffer [50 mM tris-HCl(pH 7.5), 150 mMNaCl, 1% NP-40, 0.5% sodium deoxycholate, andsupplemented with cOmplete protease inhibitor cocktail] and incu-bated for 30 min at 4°C with gentle mixing, and the lysates were cen-trifuged for 10 min at 10,000g. The supernatant was transferred intoanother tube. Streptavidin Agarose Resin suspension (100 ml) (2×diluted with lysis buffer; Thermo Fisher Scientific) was added to thesupernatant and gently mixed for 2 hours at room temperature. TheStreptavidin Agarose Resin was collected by centrifugation (10,000g,1 min), washed twice with Wash Buffer 1 [50 mM tris-HCl (pH 7.5),


500 mM NaCl, 0.1% NP-40, 0.05% sodium deoxycholate], followedby Wash Buffer 2 [50 mM tris-HCl (pH 7.5), 0.1% NP-40, 0.05% so-dium deoxycholate]. Next, the biotin-labeled proteins were elutedwith 100 ml of 2× SDS sample buffer. The samples were subjectedto SDS-PAGE followed by Western blot analyses using an antibodyagainst HA tag (clone 3F10).

TGFa shedding assayThe TGFa shedding assay was carried out as previously described(42). In brief, HEK293T cells were seeded in 12-well plates (2 ×105 cells per well), 24 hours before transfection. The cells were trans-fected with plasmids encoding AP-TGFa (a gift from S. Higashiyama,EhimeUniversity Proteo-ScienceCenter) (0.25mg perwell), HA-BLT1/WT, HA-BLT1/WTDphos or HA-BLT1/WT/DE (0.1 mg per well),and Gaq/i1 (0.05 mg per well). After 24 hours, the cells were collectedand washed with PBS and then suspended in HBSS containing 5mMHepes (pH 7.4). The cell suspension was plated in a 96-well plate(80 ml per well) and stimulated with 20 ml of HBSS containing LTB4(10 to 150 nM) for 1 hour at 37°C. The 96-well plates were centri-fuged for 2 min at 200g, and 80 ml of the supernatant was transferredinto another 96-well plate. An 80-ml portion of 2× p-nitrophenylphosphate disodium salt hexahydrate (pNPP) buffer [10 mM pNPPin 40 mM tris-HCl (pH 9.5), 40 mM NaCl, and 10 mMMgCl2] wasadded to both the supernatant and the cells. The absorbance at 405 nm(A405) of both plates was read before and after a 1-hour incubation at37°C using a microplate reader (Thermo Fisher Scientific). AP-TGFarelease was calculated by the following formulae:

AP ‐TGFa in supernatant ð%Þ ¼ðDOD405Sup=½DOD405Supþ DOD405Cell�Þ � 100AP ‐TGFa release ð%Þ ¼AP ‐TGFa in supernatant under a stimulated condition ð%Þ�AP ‐TGFa in supernatant under a vehicle ‐ treated condition ð%Þ

NanoBiT–b-arrestin recruitment assayFor the NanoBiT–b-arrestin recruitment assay (58), a receptor con-struct was designed to fuse the small fragment (SmBiT) of the NanoBiTcomplementation luciferase to the C terminus of HA-BLT1 with a15–amino acid flexible linker (GGSGGGGSGGSSSGG), whose se-quences were recommended by the manufacturer (Promega). Theconstruct (HA-BLT1-SmBiT) was assembled and inserted into apCAGGSmammalian expression plasmid (a gift from J.-i. Miyazaki,Osaka University) using an NEBuilder HiFi DNA Assembly system(New England Biolabs). A b-arrestin construct was generated byfusing the large fragment (LgBiT), whose nucleotide sequences weregene-synthesized with mammalian codon optimization (GenScript),to the N terminus of human b-arrestin1 or b-arrestin2 (barr1 or barr2,respectively) with the 15–amino acid linker. The LgBiT-barr1 andLgBiT-barr2 were inserted into the pCAGGS plasmid. HEK293A cells(Thermo Fisher Scientific) were seeded in a 10-cmdish at 2 × 106 cells in10ml ofDMEM, supplementedwith 10%FBS andpenicillin (100U/ml)and streptomycin (100 mg/ml), and cultured for 1 day. The cells weretransfected with a mixture of HA-BLT1-SmBiT plasmid (1 mg) andthe LgBiT-barr plasmid (LgBiT-barr1 or LgBiT-barr2; 500 ng) by dilut-ing in 500 ml of Opti-MEM I Reduced SerumMedium (Thermo FisherScientific) and combining 25 ml of polyethylenimine reagent (1 mg/ml)[Polyethylenimine “Max” (molecular weight 40,000), Polysciences] di-luted in 500 ml of the Opti-MEM. As a negative control, the pCAGGS

13 of 15



http://stke.sciencD

ownloaded from

plasmid was used. Twenty-four hours after the addition of the transfec-tion solution, the cells were harvested with 5 ml of a 0.53 mM EDTA-containing Dulbecco’s phosphate-buffered saline (D-PBS), followed byrinsing with 5 ml of HBSS containing 5 mMHepes (pH 7.4). The cellswere centrifuged at 190g for 5 min and suspended in 10 ml of 0.01%BSA (fatty acid–free grade; SERVA)–containing HBSS. The cell sus-pension was seeded in a 96-well white plate at a volume of 80 ml perwell and loaded with 20 ml of 50 mM coelenterazine (Carbosynth)diluted in the BSA-HBSS. After incubation at room temperature for2 hours, background luminescent signals were measured using theluminescent microplate reader SpectraMax L equipped with two de-tectors (Molecular Devices). Prediluted LTB4 solution (20 ml at finalconcentrations of 100 pM to 100 nM, 3.2-fold dilution) or vehiclewas manually added to the cells. After ligand addition (5 min), lumi-nescent signals weremeasured for 5min with 20-s intervals. For eachwell, the luminescent signal was normalized to the initial count andfold change values over 5 to 10 min after ligand stimulation was aver-aged. The fold change barr recruitment signals were fitted to a four-parameter sigmoidal concentration-response curve, andEC50 (medianeffective concentration) values were obtained, using the Prism 7software (GraphPad Prism).

Statistical analysisFor each experiment, at least three independent experimentswere per-formed. The images obtained from one representative experimentwere presented. Todetermine statistical significance, valueswere com-pared by Student’s t test, one-way or two-way ANOVA with Tukeypost hoc test, Bonferroni post hoc test, or Kruskal-Wallis test withDunn’s post hoc test using Prism 7 software. The differences wereconsidered significant if P values were less than 0.05.

on April 3, 2021

emag.org/

SUPPLEMENTARY MATERIALSwww.sciencesignaling.org/cgi/content/full/11/544/eaao5390/DC1Fig. S1. Properties of N-glycosylation–deficient HA-BLT1/0N.Fig. S2. Conserved Ser and Thr residues in the cytoplasmic domains of human, mouse, rat,guinea pig, and zebrafish BLT1.Fig. S3. Confirmation of residues essential for HA-BLT1/0N phosphorylation.Fig. S4. Phosphorylation of mouse BLT1.Fig. S5. Phosphorylation at LTB4-induced and basal sites through Gi.Fig. S6. LTB4 dose dependency of intracellular [Ca

2+] increase.Fig. S7. Effect of blockage of the [Ca2+] increase on BLT1 phosphorylation.Fig. S8. Effect of phosphorylation deficiency on the functions of wild-type BLT1.Fig. S9. Effect of helix 8 disruption on BLT1 phosphorylation.Fig. S10. Importance of phosphorylation for the ligand sensitivity of BLT1.Fig. S11. Effect of phosphorylation on b-arrestin binding to BLT1.Table S1. Primer sequences used to generate mutant BLT1s.

REFERENCES AND NOTES1. M. C. Lagerström, H. B. Schiöth, Structural diversity of G protein-coupled receptors and

significance for drug discovery. Nat. Rev. Drug Discov. 7, 339–357 (2008).2. K. L. Pierce, R. T. Premont, R. J. Lefkowitz, Seven-transmembrane receptors. Nat. Rev. Mol.

Cell Biol. 3, 639–650 (2002).3. A. B. Tobin, A. J. Butcher, K. C. Kong, Location, location, location…site-specific GPCR

phosphorylation offers a mechanism for cell-type-specific signalling. Trends Pharmacol.Sci. 29, 413–420 (2008).

4. A. J. Butcher, R. Prihandoko, K. C. Kong, P. McWilliams, J. M. Edwards, A. Bottrill, S. Mistry,A. B. Tobin, Differential G-protein-coupled receptor phosphorylation provides evidencefor a signaling bar code. J. Biol. Chem. 286, 11506–11518 (2011).

5. K. N. Nobles, K. Xiao, S. Ahn, A. K. Shukla, C. M. Lam, S. Rajagopal, R. T. Strachan,T.-Y. Huang, E. A. Bressler, M. R. Hara, S. K. Shenoy, S. P. Gygi, R. J. Lefkowitz, Distinctphosphorylation sites on the b2-adrenergic receptor establish a barcode that encodesdifferential functions of b-arrestin. S

Stepwise phosphorylation of leukotriene B receptor 1 ... ·...

Documents

Transcript of Stepwise phosphorylation of leukotriene B receptor 1 ... ·...