nature immunology  VOLUME 13 NUMBER 5 MAY 2012 457

A rt i c l e s

Chemotaxis (the directed migration of cells in response to a gradient of chemoattractant) is essential for lymphocytes to find antigens and for neutrophils to find sites of infection and inflammation1. Chemotactic cells such as blood neutrophils and neutrophil-like dif-ferentiated human promyelocytic leukemia (HL60) cells respond to chemoattractants such as formyl-Met-Leu-Phe (fMLF) by adopting polarized morphology (polarization) and traveling up the gradient (directional sensing). Extensive studies have been done to eluci-date the mechanisms of cell polarization and directional sensing1–4. Neutrophils, for example, use a self-organizing mechanism that diverges from the formyl peptide receptor through different trimeric G proteins to break symmetry and polarize5. The GTP-binding Rho-family protein Cdc42 and microtubule pathways are both important for neutrophil directional sensing6–9. In addition to those mecha-nisms that promote chemotaxis, there are inhibitory mechanisms that ‘instruct’ migrating cells to stop directional movement in the pres-ence of a single attractant and to move in the correct direction when multiple attractants are present10. Thus, these inhibitory mechanisms ensure that chemotactic cells reach the correct destination, where they serve bactericidal functions for phagocytes and antigen identification for homing lymphocytes. So far, little is known about the regulatory mechanisms that inhibit cell migration.

In addition to activating the appropriate trimeric G proteins, attractants promote phosphorylation of their receptors by G protein– coupled receptor kinases (GRKs). This phosphorylation enables the receptors to bind arrestins, thereby preventing the receptors from activating G proteins and terminating signaling, a process

called ‘desensitization’11–14. The receptor-arrestin complex is sub-sequently internalized via a clathrin-dependent pathway or a clathrin- independent pathway14. Internalized receptors are sorted to be either degraded or sent to recycling compartments14. In neutrophils, both internalization of formyl peptide–receptor 1 (FPR1) and uncou-pling of G proteins are mediated through GRK2 but do not require arrestins12,13. This GRK-dependent receptor desensitization results in fewer potential active receptors on the cell surface, thereby diminish-ing the internal signal generated in response to a given concentration of an attractant. However, published studies have reported a negligible chemotaxis deficiency in cells expressing receptors that cannot be desensitized15, which suggests that receptor desensitization itself is not required for chemotaxis. The role of receptor desensitization in cell migration remains unclear.

Mitogen-activated protein kinases (MAPKs), including Erk, Jnk and p38, are involved in inflammation, apoptosis and migration16,17. It has been shown that p38 regulates neutrophil chemotaxis both in vivo and in  vitro18–21, but the underlying mechanisms of this remain unclear. The roles of Erk and Jnk in chemotaxis have not been fully elucidated. Here we found that p38 and Erk regulated neutrophil migration with opposite effects: p38 counteracted GRK2-mediated receptor desensitization, whereas Erk enhanced it. Furthermore, p38 has been identified as a GRK that binds and phosphorylates FPR1. This p38-mediated phosphorylation prevented binding of GRK2 to the same receptor and thus inhibited GRK2-mediated FPR1 desen-sitization. Therefore, the p38- and Erk-mediated signals controlled the net chemotactic ‘go’ and ‘stop’ activity, respectively, of migrating

1Department of Dermatology, University of Illinois College of Medicine, Chicago, Illinois, USA. 2Department of Pharmacology, University of Illinois College of Medicine, Chicago, Illinois, USA. 3State Key Laboratory of Biomembrane and Membrane Biotechnology, Institute of Zoology, Graduate School of Chinese Academy of Sciences, Beijing, China. 4The Center for Lung and Vascular Biology, University of Illinois College of Medicine, Chicago, Illinois, USA. 5Department of Anesthesiology, University of Illinois College of Medicine, Chicago, Illinois, USA. 6State Key Laboratory of Brain and Cognitive Science, Institute of Biophysics, Chinese Academy of Sciences, Beijing, China. 7School of Pharmacy, Shanghai Jiao Tong University, Shanghai, China. 8These authors contributed equally to this work. Correspondence should be addressed to J.X. (jingsong@uic.edu) or R.D.Y. (yer@uic.edu).

Received 5 December 2011; accepted 6 February 2012; published online 25 March 2012; doi:10.1038/ni.2258

Bidirectional regulation of neutrophil migration by mitogen-activated protein kinasesXiaowen Liu1,2,8, Bo Ma3,8, Asrar B Malik2,4, Haiyang Tang1,2, Tao Yang3, Bo Sun1–3, Gang Wang1,2, Richard D Minshall2,5, Yan Li6, Yong Zhao3, Richard D Ye1,2,7 & Jingsong Xu1,2

To kill invading bacteria, neutrophils must interpret spatial cues, migrate and reach target sites. Although the initiation of chemotactic migration has been extensively studied, little is known about its termination. Here we found that two mitogen-activated protein kinases (MAPKs) had opposing roles in neutrophil trafficking. The extracellular signal–regulated kinase Erk potentiated activity of the G protein–coupled receptor kinase GRK2 and inhibited neutrophil migration, whereas the MAPK p38 acted as a noncanonical GRK that phosphorylated the formyl peptide receptor FPR1 and facilitated neutrophil migration by blocking GRK2 function. Therefore, the dynamic balance between Erk and p38 controlled neutrophil ‘stop’ and ‘go’ activity, which ensured that neutrophils reached their final destination as the first line of host defense.

npg

© 2

012

Nat

ure

Am

eric

a, In

c. A

ll rig

hts

rese

rved

.

458  VOLUME 13 NUMBER 5 MAY 2012 nature immunology

A rt i c l e s

neutrophils and enabled neutrophils to arrive efficiently at the site of infection to achieve their bactericidal functions.

RESULTSOpposite roles of Erk and p38 in cell migrationMAPKs have been linked to the regulation of cell migration16,17. However, whether different MAPK isoforms have distinct roles in cell migration and the underlying mechanisms has remained unclear. We treated differentiated HL60 cells with specific inhibitors of MAPKs or knocked down the MAPKs by RNA-mediated interference (RNAi; Supplementary Fig. 1) and characterized migratory activity of these cells in fMLF concentration gradients. Whereas ~80% of control cells migrated up the fMLF gradient (100 nM; n = 109 cells) and reached the top (Fig. 1a,b, Supplementary Table 1 and Supplementary Video 1), only ~20% of the cells treated by RNAi for knockdown of p38α (the predominant isoform expressed in HL60 cells22) or the p38 inhibitor SB203580 migrated through and reached the top (n = 147 cells; Fig. 1a,b, Supplementary Table 1 and Supplementary Video 2). The remaining cells were able to polarize and migrate initially in the attractant gradient but quickly lost directionality and ‘wandered’ aimlessly without net forward locomotion, thus ceasing directional migration. The chemotaxis index (CI; the ratio of net migration in the correct direction to total migration distance8) was signifi-cantly lower in cells treated with RNAi of p38α or with SB203580 than in control cells (0.46 and 0.41 versus 0.72; P < 0.001, Fig. 1c). We obtained similar results in a second chemotaxis assay, in an in  vivo transmigration assay and by using human neutrophils (Supplementary Figs. 2 and 3).

To further demonstrate the regulatory effect of p38 on cell migra-tion, we examined migratory activity in p38α-deficient neutrophils obtained from mice with conditional knockout of p38α (ref. 23). We found substantially less transmigration in vivo and chemotaxis in vitro by p38α-deficient neutrophils (Supplementary Fig. 4a–e), consistent with our observations (described above) of cells treated with SB203580 or RNAi.

In contrast to results obtained by inhibition of p38, cells treated with the Erk inhibitor PD98059 or with RNAi of Erk showed more chemotaxis than did untreated control cells, with more than 95% of cells migrating through the entire gradient (n = 331 cells; Fig. 1a,b, Supplementary Table 1 and Supplementary Video 3). The CI of cells treated with RNAi of Erk or with PD98059 was significantly higher than that of untreated control cells (0.88 and 0.92 versus 0.72, P < 0.001; Fig. 1c). We obtained similar results with U0126, another Erk inhibitor (Supplementary Fig. 4). However, inhibition of Jnk with SP600125 has little effect on cell migration (Fig. 1b,c). Neither cell polarization, measured by actin polymerization, nor cell-migration speed, measured before cells lost net forward locomotion, were affected in cells in which p38 or Erk was inhibited (Supplementary Fig. 5 and Fig. 1d). Thus, downregulation of p38 inhibited the

directional migration of cells in an fMLF gradient, whereas inhibi-tion of Erk enhanced cell chemotaxis, which suggested that these two MAPKs have opposing functions in neutrophil chemotaxis.

Concentration-dependent cell-activity switchAs high concentrations of chemoattractants are known to inhibit neutro-phil orientation24, we next characterized the chemotactic activities of differentiated HL60 cells in various fMLF concentration gradients (Fig. 2). In gradients generated by a lower concentration of fMLF (50 nM and 100 nM), the majority of cells (>70%; n = 186) showed continuous directional migration up the gradient field (Fig. 2a,b and Supplementary Table 1). In contrast, in gradients generated by a higher concentration of fMLF (500 nM and 1,000 nM), only ~10% of cells (n = 355) completed migration through the gradient (Fig. 2a,b and Supplementary Table 1). As expected, the CI was significantly lower at fMLF gradients of higher concentration (0.38 and 0.27 for 500 nM and 1,000 nM, respectively; Fig. 2c) than at fMLF gradients of lower concentration (0.66 and 0.72 for 50 nM and 100 nM, respectively; Fig. 2c). We obtained similar results for neutrophil transmigration into the peritoneal cavity of wild-type mice (Fig. 2e), for human neutrophils (Supplementary Fig. 3) and with cells stimulated by uniform concen-trations of fMLF (used for assessment of chemokinesis; Supplementary Fig. 2b,d). Thus, the migration of cells in response to higher-concentra-tion fMLF gradients (such as 500 nM) was similar to that of cells treated with SB203580 or with RNAi of p38α in response to fMLF gradients of lower concentration (such as 100 nM).

To address whether p38 or Erk is involved in terminating direc-tional cell migration, we next assessed the activation of both p38 and Erk at various concentrations of fMLF. We found that activation of p38 (measured by its phosphorylation) peaked at an fMLF concentration of 100 nM, then dropped considerably at concentrations of 500 and 1,000 nM (Fig. 3a). The activation of Erk also peaked at 100 nM but plateaued as fMLF concentration was increased further (Fig. 3b). In the activation assays described above, we stimulated cells for 2 min, as phosphorylation of p38 and Erk peaked at 2 min for each concentra-tion of fMLF (Supplementary Fig. 5c). We obtained similar results with human neutrophils (Supplementary Fig. 6).

a b

SB

Ctrl

p38αRNAi

0

0.2

0.4

0.6

0.8

Che

mot

axis

inde

x

*

020406080

Per

cent

age

of m

igra

ting

cells

rel

ativ

e to

arr

este

d

100

Arrest Migrate

c

Ctrl SB

*

PD

PD

*

1.0

ErkRNAi

*

p38αRNAi

ErkRNAi

Ctrl SB PDp38αRNAi

ErkRNAi

SP

SP

0

5

10

15

20

Mig

ratio

n sp

eed

(µm

/min

)

Ctrl SB PDp38αRNAi

ErkRNAi SP

d

Figure 1 Erk and p38 have opposite roles in neutrophil chemotaxis. (a) Trajectories of cells left untreated control (Ctrl) or treated with SB203580 (SB; 10 µM), RNAi of p38α, PD98059 (PD; 50 µM) or RNAi of Erk in an fMLF gradient of 100 nM (>30 cells per condition). Each trace represents the trajectory of one cell (throughout). Scale bar, 100 µm. (b) Frequency of cells that migrated through the entire gradient field (Migrate) relative to cells that did not reach the top (Arrest) for cells treated as in a and in cells treated with SP600125 (SP; 10 µM). (c,d) CI (c) and migration speed (d) of cells treated as in b. *P < 0.001, compared with control (Student’s t-test). Data are representative of (a) or are from (b–d) three independent experiments mean and s.e.m. in c,d).

npg

© 2

012

Nat

ure

Am

eric

a, In

c. A

ll rig

hts

rese

rved

.

nature immunology  VOLUME 13 NUMBER 5 MAY 2012 459

A rt i c l e s

To test whether the lower p38 activity and sustained Erk activity were responsible for the impaired chemotaxis at high concentrations of fMLF, we treated cells with either anisomycin, which activates p38 (ref. 25; Supplementary Fig. 1), or PD98059, which inhibits Erk, and assessed their chemotactic migration in the gradient at 500 nM. Notably, unlike the untreated control cells, most of which ceased directional migration in the middle of the gradient field, ~70% of p38-activated cells (n = 95) migrated through the field (Fig. 3c,d and Supplementary Table 1). The CI of those cells was significantly higher than that of untreated control cells in a 500-nM gradient (0.76 versus 0.37; P < 0.001) but was similar to the CI of control cells migrat-ing in a gradient of 100 nM fMLF (Figs. 3e and 1c). We obtained similar results by inhibiting Erk with PD98059 (Fig. 3c–e). These observations indicated that p38 counteracted the stop signal during cell chemotaxis, whereas Erk enhanced it.

MAPKs regulate FPR1 internalization in different waysWe next explored which signals for the cessation of directional cell migration might be induced by high concentrations of fMLF. Attractants not only activate the appropriate trimeric G proteins but also promote receptor internalization and thus prevent the receptor from activating G proteins11–13. We assessed whether the impaired cell migration at high concentrations of fMLF resulted from less receptor availability as a

result of receptor internalization. We found that receptor internalization, measured in HL60 cells or human neutrophils by fluorescence-labeled fMLF, increased along with increasing fMLF concentrations (Fig. 2f and Supplementary Fig. 6c), consistent with a published report24. Further analysis of time-response curves of the receptor internalization at fMLF concentrations of 100 nM and 500 nM also confirmed the observations noted above (Supplementary Fig. 5d).

Furthermore, to prevent receptor internalization, we used FPR1-∆ST, a mutant of FPR1 with substitution of alanine or glycine for all carboxy-terminal serine and threonine residues15. As expected, cells expressing wild-type FPR1 quickly arrested at a uniform concentration of 500 nM fMLF (Fig. 4a, top). The majority of cells (~70%; n = 33) lost net locomotion during the 15-minute recording period (Fig. 4b, left). In contrast, only ~13% of cells expressing FPR1-∆ST (n = 69) arrested during the same recording period (Fig. 4b, right). Those cells also had negligible receptor internalization (Fig. 4a, bottom). On the basis of the trajectories of the two groups of cells (Fig. 4b), the mean migra-tion distance of the wild-type cells was significantly shorter than that of the cells expressing FPR1-∆ST (17.9 µm versus 33.4 µm (P < 0.01); Fig. 4c). Thus, higher concentrations of fMLF enhanced receptor internalization and promoted the cessation of neutrophil migration.

Figure 3 Inhibition of Erk or enhancement of p38 restores cell migration at high concentrations of fMLF. (a,b) Immunoblot analysis (below) of phosphorylated (p-) and total p38 (a) or Erk (b) in cells stimulated for 2 min with concentrations of fMLF, and quantification (above) of phosphorylated p38 and Erk, presented relative to maximum activation at 100 nM fMLF. (c) Trajectories of cells left untreated (Ctrl) or treated with anisomycin (AN; 1 µM) or PD98059 (PD; 50 µM) in an fMLF gradient of 500 nM (n > 30 cells per condition). Scale bar, 100 µm. (d) Quantification of the migration of cells treated as in c (presented as in Fig. 1b). (e) CI of cells treated as in c. *P < 0.001, compared with control (Student’s t-test). Data are from one representative of three independent experiments (a,b (blots) and c), or are from three independent experiments (a,b,d,e (graphs); mean and s.e.m. in a,b,e).

f

0.6

0.4

0.2

0

Neu

trop

hil (

103 /µ

l)

0 100

0.8

1.0

**

e

20

30

40

50

60

0 500 1,000

Bou

nd li

gand

(M

FI)b

Per

cent

age

ofm

igra

ting

cells

rela

tive

to a

rres

ted

Arrest Migrate

a

1,000 nM

500 nM

100 nM

50 nM

0

40

60

80

100

50 100 500 1,000

20

fMLP (nM)

Che

mot

axis

inde

x

c

0

0.2

0.4

0.6

0.8

**

50 100 500 1,000

d

0

5

10

15

20

25

*

Mig

ratio

nsp

eed

(µm

/min

)

50 100 500 1,000

fMLP (nM)

fMLP (nM)

fMLP (nM)

103102101

fMLP (nM)

Figure 2 Concentration-dependent switch for neutrophil chemotaxis and transmigration. (a) Cell trajectories in fMLF gradients of 50 nM, 100 nM, 500 nM and 1,000 nM (n > 30 cells per condition). Scale bar, 100 µm. (b) Quantification of the migration of cells treated as in a (presented as in Fig. 1b). (c,d) CI (c) and migration speed (d) of cells treated as in a. *P < 0.001, compared with CI at 100 nM fMLF (Student’s t-test). (d) Migration speed of cells treated as in a. *P < 0.001, compared with speed at 100 nM fMLF (Student’s t-test). (e) Migration of neutrophil into the peritoneal cavity of mice (n = 3–4 mice per group) in response to the injection of various concentrations of fMLF. *P < 0.05, compared with basal (Student’s t-test). (f) Receptor internalization at various concentrations of fMLF, presented as mean fluorescence intensity (MFI) of bound ligand. Data are representative of (a) or are from (b–f) three independent experiments (mean and s.e.m. in c–f).

b

p-Erk

Erk

fMLF (nM) 0 50

0

50

100

p-E

rk(%

of m

ax)

100 500 1,000

0

0.2

0.4

0.6

0.8

Che

mot

axis

inde

x

*

Ctrl AN

e*

1.0

PD

Ctrl0

25

50

75

Per

cent

age

ofm

igra

ting

cells

rela

tive

to a

rres

ted 100

Arrest Migrated

AN PD

Ctrl

AN

c

PD

a

p-p38

p38fMLF (nM) 100 500 1,000

0

50

100

p-p3

8(%

of m

ax)

0 50

npg

© 2

012

Nat

ure

Am

eric

a, In

c. A

ll rig

hts

rese

rved

.

460  VOLUME 13 NUMBER 5 MAY 2012 nature immunology

A rt i c l e s

Prevention of receptor internalization restored cell migration at high concentrations of fMLF, which indicated that more receptor internali-zation at high concentrations of chemoattractant acted as an essential stop signal for chemotactic neutrophils.

We next determined whether p38 and Erk have opposing effects on receptor internalization. To visualize receptor internalization, we expressed a yellow fluorescent protein (YFP)-tagged FPR1 in HEK293 human embryonic kidney cells, which lack endogenous FPR1. Receptor internalization was evident within 15 min of fMLF stimulation in untreated control cells but was present as early as 3 min after fMLF stimulation in SB203580-treated cells (Fig. 5a). In contrast, much of the FPR1-YFP signal remained at the plasma membrane in PD98059-treated cells after 15 min (Fig. 5a), which indicated that there was little if any receptor internalization. We obtained similar results with HL60 cells (Fig. 5b,c) and human neu-trophils (Supplementary Fig. 6e,f). Together, these results showed that Erk and p38 had opposite effects on FPR1 internalization.

GRK2 mediates the stop signal for chemotaxisGRK2 is known to promote FPR1 internalization through phosphor-ylation13. To determine whether GRK2 mediates the stop signal during chemotaxis, we knocked down GRK2 in HL60 cells by RNAi

(Supplementary Fig. 1) and measured migration of these cells in the presence or absence of SB203580. As expected, inhibition of p38 resulted in less directional migration of untreated control cells in an fMLF gradient of 100 nM (Fig. 6a–c and Supplementary Table 1). Knockdown of GRK2 reversed this effect and restored cell migra-tion (Fig. 6a–c and Supplementary Table 1), which suggested that the SB203580-induced cell arrest was mediated through GRK2. That finding led us to test whether p38 and Erk affected GRK2 function. We found more association of GRK2 with the membrane after fMLF stimulation in untreated control cells (Fig. 6d,e). Treatment of the cells with SB203580 further increased the association of GRK2 with the membrane in basal and stimulated states (Fig. 6d,e), which indi-cated that p38 antagonized recruitment of GRK2 to the membrane. We obtained similar results with human neutrophils (Supplementary Fig. 7). In contrast, PD98059 prevented the recruitment of GRK2 to the membrane (Fig. 6f,g and Supplementary Fig. 7), which suggested that Erk enhanced the activity of GRK2.

Blockade of GRK2 function by p38 acting as a GRKErk has been shown to prevent GRK2 degradation26,27. We observed a similar effect in HL60 cells treated with RNAi of Erk, which had much less abundance of GRK2 before and after fMLF stimulation

0 10 15

Ctrl

SB

3

PD

a

20

25

30

35

0 10 20 30

CtrlSB

**

Time (min)

b

Bou

nd li

gand

(M

FI)

Time (min)

CtrlPD

c

45

50

60

0 10 20 30

*55

65

Bou

nd li

gand

(M

FI)

Time (min)

Figure 5 Differences in the regulation of FPR1 internalization by p38 and Erk. (a) Internalization of FPR1-YFP in HEK293 cells left untreated (Ctrl) or treated with SB203580 (SB; 10 µM) or PD98059 (PD; 50 µM) and stimulated for various times (above images) with 100 nM fMLF (n > 30 cells per condition). Arrowheads indicate internalized receptor. Scale bar, 20 µm. (b,c) Internalization of FPR1 in HL60 cells (2 × 105 cells per condition) in the presence or absence (Ctrl) of 10 µM SB203580 (b) or 50 µM PD98059 (c), presented as mean fluorescence intensity (MFI) of bound ligand. *P < 0.05, compared with control (Student’s t-test). Data are representative of (a) or are from (b,c) three independent experiments (mean and s.e.m. in b,c).

FPR1-YFP FPR1-∆ST-YFP0

20

WT ∆ST

Mig

ratio

ndi

stan

ce (

µm)

b *40c

20 s

FPR1-YFP

FPR1-∆ST-YFP

290 s 500 s 640 s

20 s 350 s 710 s 950 s

aFigure 4 Receptor internalization acts as a stop signal for directional migration. (a) Fluorescence microscopy (top rows) and differential interference contrast (DIC) images (bottom rows) of cells (n > 30 per group) expressing FPR1-YFP or FPR1-∆ST-YFP, stimulated for various times (above images) with 500 nM fMLF. Arrowheads indicate leading edges. Scale bar, 10 µm. (b) Trajectories of cells (n > 30 per group) expressing FPR1-YFP and FPR1-∆ST-YFP, stimulated with 500 nM fMLF. Scale bar, 60 µm. (c) Migration distance of cells expressing FPR1-YFP (WT) or FPR1-∆ST-YFP (∆ST), treated as in a. *P < 0.01 (Student’s t-test). Data are representative of (a) or are from (b,c) three independent experiments (b,c; mean and s.e.m. in c).

npg

© 2

012

Nat

ure

Am

eric

a, In

c. A

ll rig

hts

rese

rved

.

nature immunology  VOLUME 13 NUMBER 5 MAY 2012 461

A rt i c l e s

(Supplementary Fig. 8a). However, the mechanism by which p38 regulates GRK2 remained unclear. By immunostaining, we examined subcellular localization of p38 and FPR1 in HL60 cells. We found phos-phorylated p38 in the pseudopods of polarized cells, whereas total p38 was distributed more uniformly (Fig. 7a). FPR1 localized mainly to the cell membrane at both the leading and trailing edges (Fig. 7b). Thus, phosphorylated p38 might interact with FPR1 in pseudopods. We assessed that possibility by immunoprecipitation of p38 with FPR1

and found that the binding of p38 to FPR1 was greater after fMLF stim-ulation and was inhibited by SB203580 (Fig. 7c). We obtained similar results with human neutrophils (Supplementary Fig. 7). In contrast, phosphorylated Erk was distributed uniformly and we detected no interaction between Erk and FPR1 (Supplementary Fig. 8b,c).

Because p38 phosphorylates serine and threonine residues28 and there are 19 serine and/or threonine residues in the intracel-lular domains of FPR1, including 11 in the carboxyl tail (of a total of

a

SB+GRK2i

Ctrl

SB

cb

Ctrl SB SB+GRK2i

0

20

40

60

80

Per

cent

age

ofm

igra

ting

cells

rela

tive

to a

rres

ted 100

Arrest Migrate

Che

mot

axis

inde

x

0

0.2

0.4

0.6

0.8

1.0

Ctrl SB SB+GRK2i

**

Ctrl PD

Mem GRK2(pellet)

GAPDH(sup) 1 3

PDCtrl

Time (min) 0 1 3 0 1 3

f g

Ctrl SB

Time (min) 0 1 3 0 1 3

Mem GRK2(pellet)

GAPDH(sup)

d

Cyt GRK2(sup)

Cyt GRK2(sup)

*

0

0.5

1.0

1.5

GR

K2

recr

uitm

ent (

rela

tive)

0 1 3G

RK

2re

crui

tmen

t (re

lativ

e)

SBCtrle* *

*

0

1

2

3

5

4

Time (min) Time (min)0

Figure 6 GRK2 mediates the stop signal for neutrophil migration. (a) Trajectories of cells left untreated (Ctrl) or treated with SB203580 alone (SB) or SB203580 and RNAi of GRK2 (SB + GRK2i) in an fMLF gradient of 100 nM (n > 30 cells per condition). Scale bar, 100 µm. (b) Quantification of the migration of cells treated as in a (presented as in Fig. 1b). (c) CI of cells treated as in a. *P < 0.001 (Student’s t-test). (d,e) Immunoblot analysis (d) of membrane-bound (Mem) and cytosolic (Cyt) GRK2 in cells stimulated for various times (above lanes) with fMLF in the presence or absence of SB, assessed in cell pellets or supernatants (sup), and quantification of the recruitment of GRK2 in the cells in d; results are presented relative to those of cells not treated with SB (Ctrl) and stimulated with fMLF at 1 min (e). (f,g) Immunoblot analysis (f) and quantification (g) of GRK2 in cells stimulated for various times (above lanes (f) or horizontal axis (g)) with fMLF in the presence or absence of PD, assessed as in d,e. *P < 0.05, compared with control (Student’s t-test). Data are representative of three independent experiments (a,d,f) or three experiments (b,c,e,g; mean and s.e.m. in c,e,g).

p-p38 F-actin

p38 DIC

FPR1 F-actin

Merge DIC

0

20

60

WT

Mig

ratio

ndi

stan

ce (

µm) *

S342D

40

FPR1-S342D-YFPFPR1-YFP

ASLERALTEDSTQTSDTATNSTLPSA

340320 330

WT:

MUT-A:

MUT-B:

A G

G GG G

GG

A

AA

IP: FPR1

IB: p38

NT SB

Time (min) 0 2 0 2

Total p38

IgG

34 kDa

S342G

T339G

S338G

T336G

T334A

MUT-B

MUT-A

GST-C

GST-C

GST-L3

GST-L2

GST-L1

GST

IP: FPR1Inpu

tIg

G

IB: GRK2

Ctrl p38α RNAi

Time (min) 3100 31

Total GRK2

IP: FLAG

IB: GRK2

WT

Sham

S342A

S342D

IB: p38

Total GRK2

Total p38

dcba

e f

g h iFigure 7 Phosphorylation of FPR1 by p38. (a,b) Immunostaining of F-actin (as control) and of phosphorylated and total p38 (a) or FPR1 (b) in HL60 cells (n > 30). Arrowheads indicate pseudopods. DIC, differential interference contrast. Scale bars, 10 µm. (c) Immunoblot analysis (IB) of the interaction between FPR1 and p38 in HL60 cells left untreated (NT) or treated with SB203580 (SB) and stimulated for 0 or 2 min (below lanes) with fMLF, followed by immunoprecipitation (IP) with immunoglobulin G (IgG; control) or antibody to FPR1. Total p38, immunoblot analysis without immunoprecipitation. kDa, kilodaltons. (d) Autoradiography of 32P-labeled peptides (films exposed for 12 h): GST, glutathione S-transferase; GST-L1, GST-L2 and GST-L3; GST-tagged constructs of loop 1, loop 2 and loop 3 domains of FPR1; GST-C, GST-tagged construct of the carboxyl terminus of FPR1; MUT-A and MUT-B, FPR1 with replacement of serine and/or threonine residues among amino acids 319–332 (MUT-A) or 334–342 (MUT-B) with alanine and/or glycine; T334A, T336G, S338G, T339G, S342G mutants of FPR1 with replacement of serine (S) and/or threonine (T) with alanine (A) and/or glycine (G) in the carboxyl terminus (numbers indicate amino acid positions). (e) The wild-type (WT) and MUT-A and MUT-B constructs in d, showing sites phosphorylated by GRK2 (cyan) or p38 (yellow). (f) Immunoprecipitation and immunoblot analysis of the interaction between FPR1 and GRK2 in control cells and cells with RNAi of p38α. Input, total cell lysate. (g) Immunoprecipitation and immunoblot analysis of the interaction of Flag-tagged wild-type or mutant (S342A or S342D) FPR1 with GRK2 and p38. Sham, empty Flag vector. (h) Trajectories of cells expressing FPR1-YFP or FPR1-S342D-YFP, stimulated with 500 nM fMLF. Scale bar, 60 µm. (i) Migration distance of cells expressing wild-type or mutant (S342D) FPR1 (n > 30 cells per condition). *P < 0.001 (Student’s t-test). Data are representative of three experiments (a−d,f,g) or three independent experiments (h,i; mean and s.e.m. in i).

npg

© 2

012

Nat

ure

Am

eric

a, In

c. A

ll rig

hts

rese

rved

.

462  VOLUME 13 NUMBER 5 MAY 2012 nature immunology

A rt i c l e s

350 amino acids), we investigated whether p38 phosphorylates FPR1. We individually expressed the three intracellular loops and the car-boxy-terminal tail of FPR1 as glutathione S-transferase–tagged pro-teins and purified the proteins for in vitro phosphorylation assays. After incubation with purified phosphorylated p38, only the carboxy- terminal tail was phosphorylated (Fig. 7d); this phosphorylation was inhibited by SB203580 (data not shown). To map the phosphorylation site(s), we generated the following two mutants: in one, we replaced the six serine and/or threonine residues among amino acids 319–332 of the carboxy-terminal tail with alanine or glycine (Fig. 7e); in the other, we replaced the five serine and/or threonine residues among amino acids 334–342 with alanine or glycine (Fig. 7e). We found that p38 phosphorylated the former but not the latter (Fig. 7d). We next replaced each of the five serine and/or threonine residues among amino acids 334–342 individually and identified Ser342 (Fig. 7d) as the only phosphorylation site for p38. Notably, Ser342 is not one of the GRK2 phosphorylation sites identified before by site mutagenesis13 (Fig. 7e).

We next determined whether p38 prevented binding of GRK2 to FPR1. In untreated control cells, fMLF stimulation resulted in more binding of FPR1 to GRK2. This binding was even greater in cells in which p38 was knocked down (Fig. 7f). To determine whether phosphorylation of FPR1 by p38 prevents the binding of GRK2, we generated constructs of FLAG-tagged wild-type FPR1 and the following two FPR1 mutants: FPR1-S342A, which prevents p38 phosphorylation, and FPR1-S342D, which mimics phosphorylation at this site. We found that p38 bound to wild-type FPR1 and the two mutants equally well (Fig. 7g). In contrast, GRK2 showed much less binding to FPR1-S342D than to wild-type or FPR1-S342A (Fig. 7g). Furthermore, HL60 cells expressing FPR1-S342D migrated farther in an fMLF gradient of high concentration (500 nM) than did cells expressing wild-type FPR1 (Fig. 7h,i). Together these results indi-cated that p38 phosphorylation at Ser342 prevented the interaction of GRK2 with FPR1. Thus, p38 functions as a noncanonical GRK that counteracts the function of GRK2.

Different regulation by G proteins of Erk versus p38As shown above, Erk and p38 had opposite roles in cell migration, which raised the possibility that Erk and p38 antagonize each other. However, we found that p38 activation was not altered in cells with RNAi of Erk, and Erk activation was not altered in cells with RNAi of p38 (Fig. 8a,b), which indicated p38 and Erk were independently regulated. That conclusion was further supported by finding that p38 and Erk had distinct activation patterns in response to increasing

concentrations of fMLF (Fig. 3 and Supplementary Fig. 6). Whereas the activation of Erk plateaued, the activation curve of p38 was bell-shaped. To determine how regulation of Erk differed from that of p38, we assessed whether various heterotrimeric G proteins activated by FPRs mediated activation of Erk and p38. Inactivation of Gi by pertussis toxin abolished p38 activation, but only partially inhib-ited Erk activity (~70% inhibition, Fig. 8c,d), which indicated that some activation of Erk was independent of Gi. Because Erk and p38 showed different activation patterns with increasing fMLF concentrations, we assessed the concentration-dependent activa-tion of Erk and p38 in the presence of pertussis toxin. At all con-centrations tested, pertussis toxin blocked p38 activation but only partially inhibited Erk activation (Supplementary Fig. 9a). We used RNAi to screen other G proteins that might mediate Erk activation (Supplementary Fig. 1). We found that knockdown of Gq resulted in significantly less fMLF-induced activation of Erk but did not affect p38 activation (Fig. 8e,f).

We further determined whether MAPK phosphatase had a role in the regulation of Erk and p38. We first tested WIP1, a MAPK phos-phatase that belongs to the protein phosphatase family 2C and is specific for p38 but not Erk29. Cells treated with the protein phos-phatase 2 inhibitor okadaic acid or with knockdown of WIP1 by RNAi (Supplementary Fig. 1) showed a delayed decrease in phosphorylated p38 (Supplementary Fig. 9); however, neither treatment altered the bell-shaped activation curve of p38 (Supplementary Fig. 9). We next tested two dual-specificity MAPK phosphatases: MKP1 and MKP5, which dephosphorylate both p38 and Erk30. Knockout of the gene encoding MKP1 or MKP5 in mouse neutrophils resulted in a greater abundance of phosphorylated p38 after stimulation with chemoattract-ants but did not affect the bell-shaped activation curve of p38 (data not shown). Thus, these phosphatases did not seem to be required for the activation of Erk or p38 under our experimental conditions. It is possible, however, that heterotrimeric G proteins activated differently by one or more formyl peptide receptors could be responsible for the observed differences in the activation of Erk and p38.

DISCUSSIONUnderstanding how chemotaxis is dynamically and precisely regu-lated is of great importance, given its vital roles in inflammatory cell infiltration, lymphocyte homing, embryonic development, axon guidance and tumor invasion. In addition to triggering polarization and directional sensing, which are important for initiation of neu-trophil chemotaxis, chemoattractant stimulation elicits a distinct stop mechanism that negatively regulates directional cell migration and

p-p38

p38

Time (min)

Ctrl ERK RNAi

0 202

0

80

120

p-p3

8 (r

elat

ive)

40

a

Time (min)

p-p38

p38

Ctrl PTX

0 202

c

0

80

100

40

60

20

p-p3

8 (r

elat

ive)

p-p38

p38

Time (min)

Ctrl G12RNAi

0 202

G13RNAi

20

GqRNAi

20

0

100

150

50

e

p-p3

8 (r

elat

ive)

Time (min)

p-Erk

Erk

p38α RNAiCtrl

0 202

0

80100

4060

20

b

p-E

rk (

rela

tive)

Time (min)

p-Erk

Erk

Ctrl PTX

0 202

0

80

100

40

d

60

20

p-E

rk (

rela

tive)

p-Erk

Erk

Time (min)

Ctrl G12RNAi

0 202

G13RNAi

20

GqRNAi

20

0

80100

4060

20

*

f

p-E

rk (

rela

tive)

Figure 8 Differences in the activation of Erk and p38. (a,b) Immunoblot analysis (below) and quantification (above) of phosphorylated (p-) and total p38 (a) or Erk (b) in cells left untreated (Ctrl) or treated by RNAi of Erk (a) or p38α (b) and stimulated for 0 or 2 min with 100 nM fMLF; results above are relative to those of control cells at 2 min. (c,d) Immunoblot analysis (below) and quantification (above) of phosphorylated and total p38 (c) or Erk (d) in cells left untreated (Ctrl) or treated with pertussis toxin (PTX; 1 µg/ml) and stimulated for 0 or 2 min with 100 nM fMLF (presented as in a,b). (e,f) Immunoblot analysis (below) and quantification (above) of phosphorylated and total p38 (e) or Erk (f) in cells left untreated (Ctrl) or treated with RNAi for G12, G13 or Gq and stimulated for 0 or 2 min with 100 nM fMLF (presented as in a,b). *P < 0.05, compared with control (Student’s t-test). Data are representative of three independent experiments (mean and s.e.m.).

npg

© 2

012

Nat

ure

Am

eric

a, In

c. A

ll rig

hts

rese

rved

.

nature immunology  VOLUME 13 NUMBER 5 MAY 2012 463

A rt i c l e s

brings chemotactic cells to a state similar to that of unstimulated cells. This mechanism is represented by the incremental amount of mem-brane association of GRK2 and internalization of FPR1. Our results have demonstrated that, to achieve efficient and precise directional migration, Erk and p38 exerted opposite effects on the stop mecha-nism. Inhibition of p38 activity enhanced the stop signal and arrested migrating cells, whereas enhancement of the activity of p38 or inhi-bition of the activity of Erk seemed to overcome the stop signal and ensured directed cell migration, even at concentrations of fMLF that would normally induce termination of cell chemotaxis. Our results have shown that Erk-GRK2 functioned as a mechanism for the ter-mination of chemotaxis and have demonstrated that constant sup-pression (by p38) of the stop mechanism was required for sustained chemotaxis and for migrating cells to reach their final destinations.

Although p38 has an important role in neutrophil chemotaxis18–21, the underlying mechanisms have not been clear. Here we have demonstrated a previously unknown function of p38 as a GRK that phosphorylates a chemoattractant receptor and blocks the function of the classical GRK2. Different phosphorylation patterns of G pro-tein–coupled receptors may ‘instruct’ different downstream partners to achieve different functions31. Our observations of FPR1 supported that proposal. We found that p38 and GRK2 phosphorylated the car-boxy-terminal tail of FPR1 at distinct nonoverlapping sites. The phos-phorylation of Ser342 in FPR1 by p38 prevented the receptor from interacting with GRK2 and thereby blocked the GRK2-mediated stop signal and ensured sustained cell migration. Thus, different phos-phorylation patterns on FPR1 by p38 and GRK2 achieved opposite functions downstream of the receptor, which indicated the subtlety of the regulation of neutrophil migration by p38.

Studies have reported negligible chemotaxis deficiencies in cells expressing receptors that cannot be desensitized, which suggests that receptor desensitization is not required for chemotaxis15. Our find-ings, however, have indicated that GRK2-mediated receptor inter-nalization and desensitization had an essential role in regulating cell migration and was responsible for the termination of cell migration at high concentrations of attractant. Thus, it was the enhancement of receptor desensitization, not its inhibition, that blocked neutrophil chemotaxis. Moreover, our findings have shown that sufficient protec-tion of the receptor from desensitization was required for sustained cell migration and that overprotection of receptor from desensitiza-tion led to nonstop migration. Therefore, the precise navigation of migrating cells required a balance between the acceleration and decel-eration of receptor desensitization, which were mediated by GRK2 and p38, respectively. In other words, signals that control receptor desensitization have a central role in determining the final destination of migrating neutrophils.

Our results have shown that different heterotrimeric G protein signals downstream of the formyl peptide receptor were responsible for the regulation of Erk and p38. Activation of p38 was dependent on Gi, whereas Erk was activated by both Gi and Gq signals. In neutrophils, there are two formyl peptide receptor isoforms: FPR1 is the high-affinity receptor for fMLF (with a dissociation constant of ~10 nM) and activates Gi; FPR2 is the low-affinity receptor for fMLF (dissociation constant ~1 µm) and activates Gq32. Thus, at lower concentrations of fMLF, FPR1 activated Gi, which led to the activation of both p38 and Erk; when fMLF concentration increased, FPR2 was activated, thus activating Gq, which was responsible for sustained Erk activation. At the same time, p38 activity decreased as more FPR1 was internalized under exposure to high concentra-tions of fMLF. We suggest that the different regulatory functions of these two MAPKs provide bidirectional control of GRK2 function at

different stages of directional cell migration, thus allowing migrating cells to accurately reach their destinations.

To reach sites of infection or inflammation, circulating neutrophils must first attach to the endothelial cells lining the blood vessels, then transmigrate into tissue. The mechanisms responsible for the initial arrest of neutrophils in the endothelium are mediated mainly by the interaction of β2 integrins with their endothelial ligand, ICAM-1 (refs. 33–35). Whether this initial arrest mechanism also has a role in terminating neutrophil migration remains unclear. Likewise, it is unknown whether MAPKs or GRKs terminate directional cell migra-tion by regulating integrin activation. Animals such as fruit flies have a concentration-dependent activity switch mediated by differences in the regulation of neural circuits in response to odor stimulation36. Such a concentration-dependent switch can also be observed at the cellular level as a bell-shaped dose-response curve during cell migration. Directed cell migration initially increases in response to an increase in attractant concentration, but peaks and then gradually decreases as attractant concentration is further increased. Therefore, studies at the cellular level may provide further molecular mechanisms and insights for understanding activity switches in more complex organisms.

Our results suggest a model for neutrophil migration in which the dynamic balance of GRK2 and two MAPKs regulates go and stop activity at the receptor level. Our model indicates a mechanism for a concentration-dependent switch for the cell to continue or halt its movement during directional cell migration. The model of stop and go signals provides the opportunity for pharmacological intervention in one or more of the specific pathway(s). Such an intervention would enable appropriate infiltration of phagocytes into inflammatory sites while minimizing neutrophil-mediated tissue injury.

METHODSMethods and any associated references are available in the online version of the paper at http://www.nature.com/natureimmunology/.

Note: Supplementary information is available on the Nature Immunology website.

AcknoWLeDGMenTSWe thank X. Du, B. Gantner (University of Illinois at Chicago), S. Wang (Peking University) and S. Chen (University of Iowa) for discussions, E.R. Prossnitz (University of New Mexico) for the FPR1-∆ST construct and K. Otsu (Osaka University) for permission to use mice with loxP-flanked Mapk14. Supported by the US National Institutes of Health (HL095716 and AI033503), the Chinese Academy of Sciences (KSCX-W-R-66 and KSCX2-YW-R-156), the Natural Science Foundation of China (30630037 and 31070956) and the National Basic Research Program of China (2010CB945301 and 2011CB710900).

AUTHoR conTRIBUTIonSX.L., B.M., Y.L., R.D.Y. and J.X. designed the research; X.L., B.M., H.T., T.Y., B.S. and G.W. did the experiments; X.L., B.M. and J.X. analyzed data; A.B.M., R.D.M. and Y.Z. contributed new reagents and tools; and X.L., B.M., A.B.M., Y.L., R.D.Y. and J.X. wrote the paper.

coMPeTInG FInAncIAL InTeReSTSThe authors declare no competing financial interests.

Published online at http://www.nature.com/natureimmunology/. reprints and permissions information is available online at http://www.nature.com/reprints/index.html.

1. Devreotes, P.N. & Zigmond, S.H. Chemotaxis in eukaryotic cells: a focus on leukocytes and Dictyostelium. Annu. Rev. Cell Biol. 4, 649–686 (1988).

2. Ridley, A.J. et al. Cell migration: integrating signals from front to back. Science 302, 1704–1709 (2003).

3. Swaney, K.F., Huang, C.H. & Devreotes, P.N. Eukaryotic chemotaxis: a network of signaling pathways controls motility, directional sensing, and polarity. Annu. Rev. Biophys. 39, 265–289 (2010).

4. Janetopoulos, C. & Firtel, R.A. Directional sensing during chemotaxis. FEBS Lett. 582, 2075–2085 (2008).

npg

© 2

012

Nat

ure

Am

eric

a, In

c. A

ll rig

hts

rese

rved

.

464  VOLUME 13 NUMBER 5 MAY 2012 nature immunology

5. Xu, J. et al. Divergent signals and cytoskeletal assemblies regulate self-organizing polarity in neutrophils. Cell 114, 201–214 (2003).

6. Srinivasan, S. et al. Rac and Cdc42 play distinct roles in regulating PI(3,4,5)P3 and polarity during neutrophil chemotaxis. J. Cell Biol. 160, 375–385 (2003).

7. Li, Z. et al. Directional sensing requires G β γ-mediated PAK1 and PIX α-dependent activation of Cdc42. Cell 114, 215–227 (2003).

8. Xu, J., Wang, F., Van Keymeulen, A., Rentel, M. & Bourne, H.R. Neutrophil microtubules suppress polarity and enhance directional migration. Proc. Natl. Acad. Sci. USA 102, 6884–6889 (2005).

9. Xu, J. et al. Polarity reveals intrinsic cell chirality. Proc. Natl. Acad. Sci. USA 104, 9296–9300 (2007).

10. Foxman, E.F., Campbell, J.J. & Butcher, E.C. Multistep navigation and the combinatorial control of leukocyte chemotaxis. J. Cell Biol. 139, 1349–1360 (1997).

11. Lefkowitz, R.J. G protein–coupled receptors, III. New roles for receptor kinases and β-arrestins in receptor signaling and desensitization. J. Biol. Chem. 273, 18677–18680 (1998).

12. McLeish, K.R., Gierschik, P. & Jakobs, K.H. Desensitization uncouples the formyl peptide receptor–guanine nucleotide-binding protein interaction in HL60 cells. Mol. Pharmacol. 36, 384–390 (1989).

13. Prossnitz, E.R., Kim, C.M., Benovic, J.L. & Ye, R.D. Phosphorylation of the N-formyl peptide receptor carboxyl terminus by the G protein–coupled receptor kinase, GRK2. J. Biol. Chem. 270, 1130–1137 (1995).

14. Moore, C.A., Milano, S.K. & Benovic, J.L. Regulation of receptor trafficking by GRKs and arrestins. Annu. Rev. Physiol. 69, 451–482 (2007).

15. Hsu, M.H., Chiang, S.C., Ye, R.D. & Prossnitz, E.R. Phosphorylation of the N-formyl peptide receptor is required for receptor internalization but not chemotaxis. J. Biol. Chem. 272, 29426–29429 (1997).

16. Johnson, G.L. & Lapadat, R. Mitogen-activated protein kinase pathways mediated by ERK, JNK, and p38 protein kinases. Science 298, 1911–1912 (2002).

17. Huang, C., Jacobson, K. & Schaller, M.D. MAP kinases and cell migration. J. Cell Sci. 117, 4619–4628 (2004).

18. Nick, J.A. et al. Common and distinct intracellular signaling pathways in human neutrophils utilized by platelet activating factor and fMLP. J. Clin. Invest. 99, 975–986 (1997).

19. Cara, D.C., Kaur, J., Forster, M., McCafferty, D.M. & Kubes, P. Role of p38 mitogen-activated protein kinase in chemokine-induced emigration and chemotaxis in vivo. J. Immunol. 167, 6552–6558 (2001).

20. Heit, B., Tavener, S., Raharjo, E. & Kubes, P. An intracellular signaling hierarchy determines direction of migration in opposing chemotactic gradients. J. Cell Biol. 159, 91–102 (2002).

21. Zu, Y.L. et al. p38 mitogen-activated protein kinase activation is required for human neutrophil function triggered by TNF-α or FMLP stimulation. J. Immunol. 160, 1982–1989 (1998).

22. Hale, K.K., Trollinger, D., Rihanek, M. & Manthey, C.L. Differential expression and activation of p38 mitogen-activated protein kinase α, β, γ, and δ in inflammatory cell lineages. J. Immunol. 162, 4246–4252 (1999).

23. Nishida, K. et al. p38α mitogen-activated protein kinase plays a critical role in cardiomyocyte survival but not in cardiac hypertrophic growth in response to pressure overload. Mol. Cell Biol. 24, 10611–10620 (2004).

24. Sullivan, S.J. & Zigmond, S.H. Chemotactic peptide receptor modulation in polymorphonuclear leukocytes. J. Cell Biol. 85, 703–711 (1980).

25. Barros, L.F., Young, M., Saklatvala, J. & Baldwin, S.A. Evidence of two mechanisms for the activation of the glucose transporter GLUT1 by anisomycin: p38(MAP kinase) activation and protein synthesis inhibition in mammalian cells. J. Physiol. (Lond.) 504, 517–525 (1997).

26. Theilade, J., Hansen, J.L., Haunso, S. & Sheikh, S.P. MAP kinase protects G protein–coupled receptor kinase 2 from proteasomal degradation. Biochem. Biophys. Res. Commun. 330, 685–689 (2005).

27. Theilade, J., Lerche Hansen, J., Haunso, S. & Sheikh, S.P. Extracellular signal–regulated kinases control expression of G protein–coupled receptor kinase 2 (GRK2). FEBS Lett. 518, 195–199 (2002).

28. Ono, K. & Han, J. The p38 signal transduction pathway: activation and function. Cell. Signal. 12, 1–13 (2000).

29. Takekawa, M. et al. p53-inducible wip1 phosphatase mediates a negative feedback regulation of p38 MAPK-p53 signaling in response to UV radiation. EMBO J. 19, 6517–6526 (2000).

30. Dickinson, R.J. & Keyse, S.M. Diverse physiological functions for dual-specificity MAP kinase phosphatases. J. Cell Sci. 119, 4607–4615 (2006).

31. Zidar, D.A., Violin, J.D., Whalen, E.J. & Lefkowitz, R.J. Selective engagement of G protein coupled receptor kinases (GRKs) encodes distinct functions of biased ligands. Proc. Natl. Acad. Sci. USA 106, 9649–9654 (2009).

32. Ye, R.D. et al. International Union of Basic and Clinical Pharmacology. LXXIII. Nomenclature for the formyl peptide receptor (FPR) family. Pharmacol. Rev. 61, 119–161 (2009).

33. Butcher, E.C. Leukocyte–endothelial cell recognition: three (or more) steps to specificity and diversity. Cell 67, 1033–1036 (1991).

34. Springer, T.A. Traffic signals for lymphocyte recirculation and leukocyte emigration: the multistep paradigm. Cell 76, 301–314 (1994).

35. Xu, J. et al. Nonmuscle myosin light-chain kinase mediates neutrophil transmigration in sepsis-induced lung inflammation by activating β2 integrins. Nat. Immunol. 9, 880–886 (2008).

36. Semmelhack, J.L. & Wang, J.W. Select Drosophila glomeruli mediate innate olfactory attraction and aversion. Nature 459, 218–223 (2009).

A rt i c l e snp

g©

201

2 N

atur

e A

mer

ica,

Inc.

All

right

s re

serv

ed.

nature immunologydoi:10.1038/ni.2258

ONLINE METHODSAntibodies, reagents and mice. Rabbit polyclonal antibody to GRK2 (anti-GRK2; sc-562) and anti-FPR1 (sc-30016) and lentiviral particles containing short hairpin RNA (shRNA) specific for human GRK2 or Erk1 and Erk2 were from Santa Cruz Biotechnology. Mouse polyclonal anti-p38 (9228), rabbit monoclonal antibody to p38 phosphorylated at Thr180 and Tyr182 (9215) or to the MAPK p44/42 phosphorylated at Thr202 and Tyr204 (9101) were from Cell Signaling Technology. SB203580 was from Calbiochem. Human fibronectin was from BD BioSciences. Human albumin (low endotoxin), fMLP, protease inhibitors and phosphatase inhibitor ‘cocktails’, okadaic acid and phalloidin–tetramethylrhod-amine B isothiocyanate were from Sigma. Fluorescein conjugate of the hexapep-tide formyl-Nle-Leu-Phe-Nle-Tyr-Lys was from Invitrogen.

Wild-type C57BL/6 mice were from Charles River Laboratories; mice with loxP-flanked alleles encoding p38α (Mapk14) were from the Riken BioResource Center, with permission from K. Otsu (noted in Acknowledgments); and B6.129 P2-Lyz2tm1(cre)Ifo/J mice were from the Jackson Laboratory (strains were crossed to generate conditional knockout mice for p38α in myeloid cells). Mice were bred and housed in pathogen-free conditions with access to food and water ad libitum in the animal care facility of the University of Illinois at Chicago. All experimental procedures complied with University of Illinois at Chicago and National Institutes of Health guidelines for animal use.

DNA constructs and knockdown via RNAi. The cDNA encoding FPR1, p38α or GRK2 was cloned by RT-PCR, then inserted into pEYFP and Flag-tagged vectors. FPR1-∆ST mutant has been described15. GST-tagged FPR1 loop 1, loop 2 and loop 3 were constructed by pairwise annealing of single DNA strands as follows: 95 °C for 5 min, 70 °C for 10 min, and cooling at 25 °C for 12 h. The fragments were inserted into eukaryotic expression plasmid pGEX-4T-2. The following sequences were used (from 5′ to 3′): Loop1-1: GAT CCG CTG GAT TCC GGA TGA CAC ACA CAG TCA CCA CCT GAC; Loop1-2: TCG AGT CAG GTG GTG ACT GTG TGT GTC ATC CGG AAT CCA GCG; Loop2-1: GAT CCG ACC GCT GTG TTT GCG TCC TGC ATC CAG TCT GGA CCC AGA ACC ACC GCA CCG TGA GCT GAC; Loop2-2: TCG AGT CAG CTC ACG GTG CGG TGG TTC TGG GTC CAG ACT GGA TGC AGG ACG CAA ACA CAG CGG TCG; Loop3-1: GAT CCA AGA TCC ACA AGC AAG GCT TGA TTA AGT CCA GTC GTC CCT TAC GGG TCT GAC; Loop3-2: TCG AGT CAG ACC CGT AAG GGA CGA CTG GAC TTA ATC AAG CCT TGC TTG TGG ATC TTG .

For RNAi knockdown of p38α, two pairs of sequences (1 and 2, Supplementary Table 2) were used to make an shRNA expression cassette, which was cloned into the BamHI and XhoI restriction enzyme sites of the pEN_hH1C plasmid (Invitrogen). Through the use of an LR recombination reaction, the shRNA expression cassette was inserted into lentiviral shRNA expression plasmid pDSL_hpUGIP, followed by transfection together with pspAX2 and pMD2.G into the 293FT cell line for packaging of lentivirus particles. Suspensions were collected 72 h after transfection and used for infection of HL60 cells. Puromycin was used to screen the cell line with stable knockdown of p38α by RNAi. Lentiviruses for RNAi knockdown of GRK2, Erk1/2, G12, G13, Gq and WIP1 were from Santa Cruz Biotechnology (small interfering RNA sequences, Supplementary Table 2).

Cell culture, transfection and isolation of human neutrophils. Procedures for cultivation and differentiation of HL-60 have been described5. For transient transfection, differentiated HL-60 cells (on day 6 after the addition of dimethyl sulfoxide) were washed once in RPMI medium with HEPES, pH 7.0, and resus-pended in the same medium to a final concentration of 1 × 108 cells per ml. DNA was then added to the cells (30 µg of DNA encoding FPR1-YFP or FPR1-∆ST-YFP DNA) and the cell-DNA mixture was incubated for 10 min at 25 °C, transferred to electroporation cuvettes and subjected to an electroporation pulse on ice at 310 V, 1180 µF and low resistance. Transfected cells were allowed to recover for 10 min at 25 °C and then were transferred to 20 ml medium with 10% FBS. Subsequent assays were done 4–6 h after transfection.

For isolation of human neutrophils, blood was collected from healthy human donors (the design and execution of the experiment were thoroughly explained to the participants and informed consent was obtained). Erythrocytes were removed by dextran sedimentation (4.5% dextran) followed by hypotonic lysis with distilled water. Neutrophils were isolated from the resulting cell suspen-sion by discontinuous Percol gradient centrifugation. This procedure yielded

neutrophil purity of >95% and viability of >95%, as assessed by flow cytometry. Studies of human neutrophils were approved by the Institutional Review Board of University of Illinois at Chicago.

Cell-migration assays. Live cells were imaged after stimulation either with a uniform concentration of fMLP or a concentration gradient generated by an EZ-TAXIScan device (Effector Cell Institute) or micropipette. For EZ-TAXIScan assay, cells migrated over a cover glass coated with fibronectin (50 µg/ml) on a horizontal glass surface under a silicon chip. Differentiated HL60 cells were given no pretreatment or were pretreated for 30 min with the p38 inhibitor SB203580 (10 µM), the Erk1/2 inhibitor PD98059 (50 µM), the Jnk inhibitor SP600925 (10 µM) or the MAPK activator anisomycin (1 µM), then were washed with RPMI medium plus 25 mM HEPES, pH 7.0, and 0.1% BSA and were resuspended in a solution of RPMI medium, 25 mM HEPES, pH 7.0, and 0.1% BSA. Cells were loaded on the bottom of the chip and chemoattractant was added to the top of the chip to generate a chemoattractant gradient. Cells were allowed to migrate for 30 min, and images were recorded with the EZ-TAXIScan software and then analyzed by ImageJ software.

For micropipette assays, gradients were generated by a point source of chemoat-tractant from a micropipette containing 10 µM fMLP. Time-lapse video microscopy was done as described5. Cell migration was recorded and analyzed by ImageJ.

For analysis of the peritoneal transmigration of neutrophils, mice were injected intraperitoneally with 100 µl of saline containing various concen-trations of fMLP (1 nM, 10 nM, 100 nM or 1,000 nM). After 4 h, peritoneal cavities of anesthetized mice were lavaged and leukocytes were recovered. Total leukocytes was counted with a hemocytometer and neutrophil counts were determined on 100-µl cytospins stained with Diff-Quik; results are presented as relative to the total population.

Immunoprecipitation and immunofluorescence. Immunoprecipitation and immunofluorescence were done as described35. The intensity of bands on auto-radiograms (densitometry) was assessed on scanned X-ray films with ImageJ.

In vitro assay of p38 phosphorylation. First, active p38α (40 ng/µl) was incu-bated for 15 min with the substrate FPR1-C in 1× assay buffer together with 250 µM ATP and [γ-32P]ATP (0.16 µCi/µl), then reactions were terminated by the addition of 25 µl of 2× sample buffer and boiling for 10 min, followed by SDS-PAGE and analysis with a Fujifilm Fluorescent and Radioisotope Science Imaging System FLA-7000.

Surface binding assay of fMLP. Differentiated HL60 cells were given no pretreat-ment or were pretreated for 30 min at 37 °C with 10 µM SB203580 or 50 µM PD98059, then were washed with RPMI-1640 medium and Hanks’ balanced-salt solution (HBSS). Cells were resuspended in HBSS and then stimulated with fMLP, and reactions were quenched by the addition of ten volumes of ice-cold HBSS. Cells were then washed extensively and incubated with 10 µM N-formyl-Nleu-Leu-Phe-Nleu-Tyr-Lys-fluorescein before analysis by flow cytometry. Unlabeled cells stimulated with 10 µM fMLP were used as a negative control.

Membrane-bound GRK2 assay. Differentiated HL60 cells in suspension were preincubated for 30 min in RPMI-1640 medium with or without inhibitors (10 µM SB203580 or 50 µM PD98059). Cells were then collected and suspended in modi-fied HBSS (1.26 mM calcium and 0.9 mM magnesium without phenol red; 1 × 107 cells in 0.5 ml), then were stimulated with 100 nM fMLP in suspension. The stimu-lation was stopped by the addition of 0.5 ml stopping buffer at 25 °C (100 mM 2-(N-morpholino)ethanesulfonic acid (Mes), pH 6.8, 5 mM EDTA, 10 mM MgCl2, 1% Triton X-100 and 1× Protease Inhibitor Mixture). Samples were incubated for 15 min at 25 °C and then centrifuged for 5 min at 15,000g and 25 °C. Pellets were resus-pended in 2× Laemmli buffer and boiled for 10 min for SDS-PAGE. After transfer to a polyvinylidene difluoride membrane, mouse monoclonal antibody to GRK2 (sc-13143; Santa Cruz Biotechnology) was used for detection of membrane-bound GRK2. GRK2 and GAPDH (glyceraldehyde phosphate dehydrogenase) in the supernatant were also examined.

Statistical analysis. Statistical comparisons were made with a two-tailed Student’s t-test. Differences in mean values were considered significant at a P value of less than 0.05.

npg

© 2

012

Nat

ure

Am

eric

a, In

c. A

ll rig

hts

rese

rved

.