Keratinocyte Growth Factor Receptor Ligands Target the Receptor to Different Intracellular Pathways

19
# 2007 The Authors Journal compilation # 2007 Blackwell Publishing Ltd doi: 10.1111/j.1600-0854.2007.00651.x Traffic 2007; 8: 1854–1872 Blackwell Munksgaard Keratinocyte Growth Factor Receptor Ligands Target the Receptor to Different Intracellular Pathways Francesca Belleudi 1, *, Laura Leone 1 , Valerio Nobili 1 , Salvatore Raffa 1,2 , Federica Francescangeli 1 , Maddalena Maggio 1 , Stefania Morrone 1 , Cinzia Marchese 1 and Maria Rosaria Torrisi 1,2 1 Dipartimento di Medicina Sperimentale, Universita ` di Roma ‘‘La Sapienza’’, Viale Regina Elena 324, 00161 Roma, Italy 2 Azienda Ospedaliera S. Andrea, via di Grottarossa, 1035-1039, 00189, Roma, Italy *Corresponding author: Francesca Belleudi, [email protected] The keratinocyte growth factor receptor (KGFR)/fibroblast growth factor receptor 2b is activated by high-affinity- specific interaction with two different ligands, keratinocyte growth factor (KGF)/fibroblast growth factor (FGF)7 and FGF10/KGF2, which are characterized by an opposite requirement of heparan sulfate proteoglycans and heparin for binding to the receptor. We investigated here the possible different endocytic trafficking of KGFR, induced by the two ligands. Immunofluorescence and immunoelec- tron microscopy analysis showed that KGFR internalization triggered by either KGF or FGF10 occurs through clathrin- coated pits. Immunofluorescence confocal microscopy using endocytic markers as well as tumor susceptibility gene 101 (TSG101) silencing demonstrated that KGF drives KGFR to the degradative pathway, while FGF10 targets the receptor to the recycling endosomes. Biochemical analysis showed that KGFR is ubiquitinated and degraded after KGF treatment but not after FGF10 treatment, and that the alternative fate of KGFR might depend on the different ability of the receptor to phosphorylate the fibroblast growth factor receptor substrate 2 (FRS2) substrate and to recruit the ubiquitin ligase c-Cbl. The recycling endocytic pathway followed by KGFR upon FGF10 stimulation corre- lates with the higher mitogenic activity exerted by this ligand on epithelial cells compared with KGF, suggesting that the two ligands may play different functional roles through the regulation of the receptor endocytic transport. Key words: endocytosis, fibroblast growth factor 10, keratinocyte growth factor, keratinocyte growth factor receptor, receptor tyrosine kinases Received 5 February 2007, revised and accepted for publication 2 September 2007, uncorrected manuscript published online 4 September 2007, published online 17 October 2007 The fibroblast growth factor (FGF) family includes 22 members, which play different roles in controlling cell growth and differentiation, angiogenesis, wound healing and tumorigenesis (1,2). The biological activities of FGFs are exerted through specific high-affinity binding to recep- tor tyrosine kinases, the fibroblast growth factor receptors (FGFRs) (3,4). Alternative splicing forms of FGFRs are responsible for binding to different FGF family members (5,6). The FGFs bind also with low affinity to heparan sulfate proteoglycans (HSPGs), molecules associated with the cell surface or components of the extracellular matrix, which play important functional roles in protecting FGFs from proteolytic and acidic degradation and in increasing local concentration of the FGFs (7). Heparin or HSPGs play also a role in FGFR dimerization and activation (8). The FGFs show different heparin or HSPGs requirement for binding to FGFRs: in fact, HSPGs and heparin increase the affinity but are not essential for binding of FGF2 to FGFR1 and FGFR2 (9) whereas are essential for FGF1 oligo- merization and binding to the FGFR1 (10). In addition, the stabilizing presence of heparin and HSPGs can modify the ligand-dependent FGFR signal transduction, inducing quan- titative and qualitative changes in the phosphorylation of FGFR tyrosine sites and in the phosphorylation/activation of intracellular substrates (11); these signaling changes might differently regulate not only the cell proliferation and differentiation but also the endocytic fate of the receptor. At present, little is known about the endocytic pathways followed by different FGFRs. We have previously reported that the keratinocyte growth factor [KGF)/FGF7], after binding to the keratinocyte growth factor receptor (KGFR/FGFR2b), is internalized by clathrin-coated pits (12,13). In contrast, FGF1 bound to FGFR4 enters the cell not only by clathrin-coated pits but also by an alternative non-clathrin endocytic pathway (14), and FGF2 bound to FGFR1 is internalized by a clathrin- and caveolae-independent mechanism (15). Interestingly, after internalization induced by the same ligand FGF1, FGFRs can be sorted to different pathways: FGFR1 and FGFR3 are targeted to the degradative compartment (16), whereas FGF1/FGFR4 complexes can be targeted to the recycling compartment or reach lysosomes for degradation, depend- ing on not yet identified targeting signals present on the intracellular portion of the receptor (16,17). The possible role of HSPGs in these different internalization mechanisms has not been investigated. It has been proposed that, after internalization, heparin or HSPGs could play a stabilizing role protecting FGF/FGFR complexes from degradation (18). As demonstrated for FGFR3 in achondroplasia (19), the stabil- ization of FGF/FGFR complexes is a key event to alter their intracellular fate: mutations in the transmembrane domain of FGFR3, which stabilize the ligand-induced dimers, alter the normal targeting of the receptor to lysosomal compart- ment, allowing the mutant to reach and accumulate in the recycling compartment. 1854 www.traffic.dk

Transcript of Keratinocyte Growth Factor Receptor Ligands Target the Receptor to Different Intracellular Pathways

# 2007 The Authors

Journal compilation # 2007 Blackwell Publishing Ltd

doi: 10.1111/j.1600-0854.2007.00651.xTraffic 2007; 8: 1854–1872Blackwell Munksgaard

Keratinocyte Growth Factor Receptor Ligands Targetthe Receptor to Different Intracellular Pathways

Francesca Belleudi1,*, Laura Leone1,

Valerio Nobili1, Salvatore Raffa1,2,

Federica Francescangeli1, Maddalena Maggio1,

Stefania Morrone1, Cinzia Marchese1 and

Maria Rosaria Torrisi1,2

1Dipartimento di Medicina Sperimentale,Universita di Roma ‘‘La Sapienza’’, Viale Regina Elena324, 00161 Roma, Italy2Azienda Ospedaliera S. Andrea, via di Grottarossa,1035-1039, 00189, Roma, Italy*Corresponding author: Francesca Belleudi,[email protected]

The keratinocyte growth factor receptor (KGFR)/fibroblast

growth factor receptor 2b is activated by high-affinity-

specific interaction with two different ligands, keratinocyte

growth factor (KGF)/fibroblast growth factor (FGF)7 and

FGF10/KGF2, which are characterized by an opposite

requirement of heparan sulfate proteoglycans and heparin

for binding to the receptor. We investigated here the

possible different endocytic trafficking of KGFR, induced

by the two ligands. Immunofluorescence and immunoelec-

tronmicroscopy analysis showed that KGFR internalization

triggered by either KGF or FGF10 occurs through clathrin-

coated pits. Immunofluorescence confocal microscopy

using endocytic markers as well as tumor susceptibility

gene 101 (TSG101) silencing demonstrated that KGF drives

KGFR to the degradative pathway, while FGF10 targets the

receptor to the recycling endosomes. Biochemical analysis

showed that KGFR is ubiquitinated and degraded after KGF

treatment but not after FGF10 treatment, and that the

alternative fate of KGFR might depend on the different

ability of the receptor to phosphorylate the fibroblast

growth factor receptor substrate 2 (FRS2) substrate and to

recruit the ubiquitin ligase c-Cbl. The recycling endocytic

pathway followed by KGFR upon FGF10 stimulation corre-

lates with the higher mitogenic activity exerted by this

ligand on epithelial cells compared with KGF, suggesting

that the two ligands may play different functional roles

through the regulation of the receptor endocytic transport.

Key words: endocytosis, fibroblast growth factor 10,

keratinocyte growth factor, keratinocyte growth factor

receptor, receptor tyrosine kinases

Received 5 February 2007, revised and accepted for

publication 2 September 2007, uncorrected manuscript

published online 4 September 2007, published online 17

October 2007

The fibroblast growth factor (FGF) family includes 22

members, which play different roles in controlling cell

growth and differentiation, angiogenesis, wound healing

and tumorigenesis (1,2). The biological activities of FGFs

are exerted through specific high-affinity binding to recep-

tor tyrosine kinases, the fibroblast growth factor receptors

(FGFRs) (3,4). Alternative splicing forms of FGFRs are

responsible for binding to different FGF family members

(5,6). The FGFs bind also with low affinity to heparan

sulfate proteoglycans (HSPGs), molecules associated with

the cell surface or components of the extracellular matrix,

which play important functional roles in protecting FGFs

from proteolytic and acidic degradation and in increasing

local concentration of the FGFs (7). Heparin or HSPGs play

also a role in FGFR dimerization and activation (8). The

FGFs show different heparin or HSPGs requirement for

binding to FGFRs: in fact, HSPGs and heparin increase the

affinity but are not essential for binding of FGF2 to FGFR1

and FGFR2 (9) whereas are essential for FGF1 oligo-

merization and binding to the FGFR1 (10). In addition, the

stabilizing presence of heparin and HSPGs can modify the

ligand-dependent FGFR signal transduction, inducing quan-

titative and qualitative changes in the phosphorylation of

FGFR tyrosine sites and in the phosphorylation/activation

of intracellular substrates (11); these signaling changes

might differently regulate not only the cell proliferation and

differentiation but also the endocytic fate of the receptor.

At present, little is known about the endocytic pathways

followed by different FGFRs. We have previously reported

that thekeratinocytegrowth factor [KGF)/FGF7], afterbinding

to thekeratinocytegrowth factor receptor (KGFR/FGFR2b), is

internalized by clathrin-coated pits (12,13). In contrast, FGF1

bound to FGFR4 enters the cell not only by clathrin-coated

pits but also by an alternative non-clathrin endocytic pathway

(14), and FGF2 bound to FGFR1 is internalized by a clathrin-

and caveolae-independent mechanism (15). Interestingly,

after internalization inducedby the same ligandFGF1, FGFRs

can be sorted to different pathways: FGFR1 and FGFR3 are

targeted to the degradative compartment (16), whereas

FGF1/FGFR4 complexes can be targeted to the recycling

compartment or reach lysosomes for degradation, depend-

ing on not yet identified targeting signals present on the

intracellular portion of the receptor (16,17). The possible role

of HSPGs in these different internalization mechanisms has

not been investigated. It has been proposed that, after

internalization, heparin orHSPGscould play a stabilizing role

protecting FGF/FGFR complexes from degradation (18). As

demonstrated for FGFR3 in achondroplasia (19), the stabil-

ization of FGF/FGFR complexes is a key event to alter their

intracellular fate: mutations in the transmembrane domain

of FGFR3, which stabilize the ligand-induced dimers, alter

the normal targeting of the receptor to lysosomal compart-

ment, allowing the mutant to reach and accumulate in the

recycling compartment.

1854 www.traffic.dk

We have previously demonstrated that KGF/KGFR com-

plexes are transported first to early and then to late endo-

somes and lysosomes (12,13). However, the KGFR/

FGFR2b, expressed exclusively on epithelial cell, is acti-

vated by specific high-affinity binding of not only KGF but

also FGF10 (20,21). The two growth factors (i) exert

mitogenic activity on epithelial cells (20–22), but FGF10 in

the presence of heparin is more potent than KGF (23) and

(ii) bind exclusively to the KGFR, but they show an opposite

heparin/HSPGs requirement for binding and ligand/receptor

complex formation (21). Therefore,KGFRand its two ligands

represent an ideal model to ascertain if distinct FGFs can

target the same receptor to different endocytic pathways.

Results

FGF10 and KGF activity is differently modulated by

heparin and HSPGs and by low pH

It has been previously reported that both KGF and FGF10

bind with similar affinity to the KGFR (21), and we have

previously demonstrated that both growth factors induce

tyrosine phosphorylation of KGFR (23). However, mito-

genic activity and binding of KGF to KGFR are inhibited by

heparin (21,24), whereas those of FGF10 are stimulated

(21,23). To confirm in our cellular models the opposite role

played by the exogenously added heparin on the biological

activity of these two ligands, we evaluated by immuno-

fluorescencewith anti-phosphotyrosine antibody the level of

tyrosine phosphorylation after FGF10 or KGF treatment in

the presence of increasing doses of heparin (0.1–50 mg/mL).

Quantitative analysis of the fluorescence intensity per-

formed in NIH3T3 KGFR-transfected cells as described in

Materials and Methods showed that the addition of

heparin was able to inhibit KGF-induced and to enhance

the FGF10-induced tyrosine phosphorylation (Figure 1A).

The dose response on the tyrosine phosphorylation induced

by FGF10 in NIH3T3 KGFR cells reached a plateau at 0.3 mg/

mL of heparin and decreased at 50 mg/mL, as previously

observed in other cell types (21). Similar dose–response

behavior was obtained in HeLa cells transiently transfected

with the human KGFR complementary DNA (cDNA) (HeLa

KGFR) (Figure 1B), with some significant differences: the

level of tyrosine phosphorylation induced by FGF10 in the

absenceofexogenousheparin resultedsignificantly higher in

HeLa KGFR cells compared with NIH3T3 KGFR cells; the

addition of heparin to FGF10 induced an increase in tyrosine

phosphorylation at 0.3 mg/mL as seen in NIH3T3 KGFR cells

but a decrease already at 25 mg/mL. Thesedifferences could

be explained by the fact that epithelial and mesenchymal

cells express a different equipment of endogenous HSPGs

(25), which could differently affect the interaction of FGF10

with KGFR. To further assess the heparin effect on KGF and

FGF10 binding to KGFR, we reduced the contribution of

HSPGs physiologically expressed on the cell surface by

chlorate treatment, which removes endogenous sulfate

groups and inhibits HSPGs biosynthesis (26). The NIH3T3

KGFR cells were grown in medium containing 30 mM

sodium chlorate for 48 h before incubation with KGF or

FGF10 in the presence of different doses of heparin. Again

the addition of heparin inhibited KGF-induced but enhanced

FGF10-induced tyrosine phosphorylation also in presence of

sodium chlorate (Figure 1A). Thus, FGF10 and KGF differ in

their ability to bind and activate KGFR and to induce tyrosine

phosphorylation cascade, and these properties seem to be

differently modulated by heparin and HSPGs.

To investigate if heparin could also differently affect

a downstream pathway of KGFR signaling induced by

FGF10 and KGF, we performed a dose–response analysis of

the effect of exogenously added heparin on extracellular-

signal-related kinase 1 (ERK1) and extracellular-signal-

related kinase 2 (ERK2) phosphorylation after treatment

with the two ligands (Figure 1C). The NIH3T3 KGFR cells

were serum-starved for 12 h and then treated with FGF10

or KGF for 50 at 378C in the presence of increasing doses of

heparin. Western blot analysis with anti-phospho ERK1/2

antibody demonstrated that KGF alone induced a strong

activation of ERK1/2, which decreased after the addition of

heparin 0.1 mg/mL, whereas the weak ERK kinases acti-

vation induced by FGF10 alone was increased by the

addition of heparin, reaching the maximal intensity at the

dose of 0.3 mg/mL (Figure 1C). All together, these results

indicate that, in our cell models, the optimal conditions for

binding of the two ligands to the receptor require the

addition of heparin 0.3 mg/mL for FGF10 and the absence

of heparin for KGF. Thus, KGF- and FGF10-induced signal-

ing is differently modulated by heparin and HSPGs.

It is well known that ligand binding to receptors tyrosine

kinase (RTKs) can display different pH sensitivity, which in

turn affects their mitogenic activity, intracellular routing

and downregulation of ligand/receptor complexes. To

investigate if KGF and FGF10 could differ in pH sensitivity

for binding to KGFR, we evaluated the stability of ligand–

receptor interaction under conditions of low pH. To this

aim, we evaluated by immunofluorescence with anti-

phosphotyrosine antibody as above the level of tyrosine

phosphorylation after FGF10 or KGF treatment followed by

acidic washes. To induce acid dissociation of bound ligands

from the receptor, serum-starved HeLa KGFR cells were

treated with KGF and FGF10 at 48C and washed at various

pH values (5.0–7.5) before fixation. The quantitative anal-

ysis of the immunofluorescence intensity of the phospho-

tyrosine signal performed as above showed that pH

lowering decreased KGFR activation induced by FGF10

already at pH 7–6.5, while receptor phosphorylation

induced by KGF persisted up to pH 5.5 (Figure 1D),

suggesting that FGF10 dissociate from KGFR at a higher

pH than KGF and that FGF10/KGFR complexes display less

stability at low pH if compared with KGF/KGFR complexes.

KGFR kinase activity is required for both FGF10-

induced or KGF-induced receptor internalization

It has been previously demonstrated that FGF10 and KGF

induce similar levels of tyrosine phosphorylation of KGFR

Traffic 2007; 8: 1854–1872 1855

KGFR Endocytic Trafficking

Figure 1: Tyrosine phosphorylation and ERK activation induced by incubation with KGF or FGF10: dependence on heparin and

sensitivity to low pH. A, B) Chlorate-treated (A, black columns) or untreated (A, white columns) NIH3T3 KGFR cells and HeLa KGFR cells

(B) were serum-starved and then incubated for 100 at 378C with KGF or FGF10 in the presence of different doses of heparin. Cells were

then fixed, permeabilized and stained with anti-phosphotyrosine monoclonal antibody followed by FITC-conjugated secondary antibody.

The quantitative analysis of the immunofluorescence signal, performed as described inMaterials and Methods, indicates that the addition

of heparin inhibits KGFR activation induced by KGF but enhances that induced by FGF10 up to 0.3 mg/mL of heparin concentration. C) The

NIH3T3 KGFR cells were serum-starved for 12 h and then incubated for 50 at 378Cwith KGF or FGF10 in the presence of different doses of

heparin.Western blot analysis was performed using anti-phospho ERK1/2 or anti-total ERK1/2 antibodies, and densitometric analysis of the

bands is reported as fold increase relative to ERK phosphorylation in unstimulated cells: treatment with KGF induces ERK phosphorylation,

which decreases with the addition of heparin, whereas FGF10-induced ERK phosphorylation increases in the presence of heparin, reaching

the maximal intensity at the dose of 0.3 mg/mL. D) The HeLa KGFR cells were serum-starved, incubated with KGF or FGF10 for 1 h at 48Cand then washed at the indicated pH values as described inMaterials and Methods. Cells were then fixed, permeabilized and stained with

anti-phosphotyrosine as above. The quantitative analysis of the immunofluorescence signal indicates that pH lowering decreases KGFR

activation induced by FGF10 already at pH 7–6.5, while receptor phosphorylation induced by KGF, also if reduced, persists up to pH 5.5.

1856 Traffic 2007; 8: 1854–1872

Belleudi et al.

(23), but the kinetics of KGFR activation in response to

these ligands has not been studied in detail. To assess this

issue, NIH3T3 KGFR cells were serum-starved for 12 h

and then treated with FGF10 or KGF 100 ng/mL for 30 or100 at 378C. Immunoprecipitation with anti-Bek polyclonal

antibodies and immunoblot with anti-phosphotyrosine

antibody revealed that FGF10 and KGF, used at saturating

doses, which presumably are able to induce similar levels

of receptor occupancy at the plasma membrane, stimu-

lated comparable levels of phosphorylation of KGFR after

30 as well as after 100 at 378C (Figure 2A). To analyze if the

kinase activity and tyrosine phosphorylation of KGFR are

required for receptor internalization, we used SU5402,

a specific inhibitor of FGFR tyrosine kinases (27). First,

cells were pretreatedwith SU5402 and stimulatedwith the

two ligands for 100 at 378C in the presence of the kinase

inhibitor before staining with anti-phosphotyrosine anti-

body as above. As shown in Figure 2B, the immunofluor-

escence signal, intense and localized mainly on the cell

plasma membranes after stimulation with either KGF or

FGF10, was progressively reduced by the treatment with

SU5402, 30 mM and 50 mM. In contrast, in NIH3T3 epi-

dermal growth factor receptor (EGFR) cells treated with

epidermal growth factor (EGF), SU5402 did not interfere

with the phosphotyrosine staining, confirming that this

drug is a specific FGFR kinase inhibitor (Figure 2B). Quan-

titative analysis of the immunofluorescence experiments

was performed by counting for each treatment the per-

centage of plasma membrane fluorescence-positive cells

on a total of 500 cells, randomly observed in 10 micro-

scopic fields from two different experiments, and values

are expressed as mean � standard error (SE).

In order to evaluate the correlation between KGFR kinase

activity and receptor internalization induced by either KGF

or FGF10, we used HeLa cells transiently transfected with

KGFR and we stimulated them with FGF10 and KGF for

1 h at 48C followed by warming to 378C for 300 in the

presence of the specific inhibitor of the kinase activity

SU5402. Immunofluorescence staining to follow the

KGFR during internalization was performed using an anti-

Bek antibody, which recognizes the extracellular portion of

the two splicing variants FGFR2 and KGFR and which does

not compete with the ligands for binding to the receptor.

To visualize the cell surface, plasma membranes were

decorated with the fluorescein isothiocyanate (FITC)-

conjugated lectin wheat germ agglutinin (WGA) before

permeabilization, while the endocytic compartment was

identified by staining of early endosomes with anti-early

endosome antigene 1 (EEA1) monoclonal antibody. Quan-

titative analysis of the KGFR immunofluorescence signal

on the cell surface or internalized after treatment with the

growth factors was performed as described in Materials

and Methods, assessing the extent of colocalization of

KGFR with the WGA or EEA1 markers. After treatment

with the growth factors in the presence of the kinase

inhibitor, the signal corresponding to KGFR remained

localized on the plasma membrane, as shown by colocal-

ization with WGA–FITC staining, and did not appear

colocalized with the EEA1 marker in endocytic dots, as

observed after treatment with the two ligands alone

(Figure 2C). These results indicate that SU5402 treatment

is able to block KGFR uptake after both FGF10 and KGF

treatment. To further demonstrate the essential role of

KGFR kinase in receptor internalization induced by both

ligands, we used HeLa cells transiently transfected with

a kinase-negative mutant KGFR Y656F/Y657F (28) and we

performed the KGFR endocytic assay as above: the

quantitative immunofluorescence confirmed that the loss

of kinase activity do not allow KGFR internalization after

either FGF10 or KGF treatment (Figure 2C).

Differently from KGF, FGF10 does not induce

degradation of KGFR

To investigate if the two ligands could differently modulate

KGFR down-modulation, we first analyzed the kinetics of

KGFR degradation induced by FGF10 and KGF. The

NIH3T3 KGFR cells were serum-starved, pretreated with

25 mg/mL cycloheximide to block KGFR neosynthesis and

then incubated with KGF and FGF10 at 378C for different

time-points. The total amount of KGFR protein present in

the cell lysates was detected by Western blot using anti-

Bek polyclonal antibodies (Figure 3A). In unstimulated

cells, a slight decrease of the band at the molecular weight

of 140 kDa corresponding to the receptor was observed

at 8 h of treatment with cycloheximide, reflecting the con-

stitutive turnover of the protein. In cells treated with

FGF10, the band relative to the receptor, after a small

decrease at 3 h, remained unaltered after 6 and 8 h of

treatment, suggesting that this ligand did not induce an

evident KGFR down-modulation and degradation. In con-

trast, KGF treatment induced a significant acceleration of

KGFR degradation demonstrated by the evident progres-

sive decrease of the 140-kDa band at 3, 6 and 8 h of

treatment, as expected (13). Densitometric analysis of the

bands from unstimulated and KGF- or FGF10-stimulated

samples is reported in the graph (Figure 3A). These results

strongly suggest that FGF10, differently from KGF, is not

able to induce KGFR downregulation and degradation.

To visualize the possible different intracellular fate of KGFR

induced by the two ligands, we performed triple immuno-

fluorescence experiments using anti-Bek antibody to label

KGFR, LysoTracker-red as marker of lysosomes and anti-

giantin antibody to identify the Golgi complex. The HeLa

KGFR cells were treated with KGF or FGF10 for 1 h at 48Cand then warmed for 1 h at 378C in presence of Lyso-

Tracker-red. Cells were then permeabilized and stained

with anti-Bek monoclonal antibody and anti-giantin poly-

clonal antibodies. Immunofluorescence analysis showed

that after 1 h of internalization, KGFR appeared mostly in

intracellular dots: codistribution with LysoTracker was

observed after treatment with KGF but not with FGF10

(Figure 3B). The partial codistribution of KGFR and giantin

signal identified neosynthesized KGFR (Figure 3B). These

Traffic 2007; 8: 1854–1872 1857

KGFR Endocytic Trafficking

results indicate that, at late steps of the endocytic process,

KGFR is transported to a perinuclear non-acidic compart-

ment upon FGF10-induced endocytosis and to the lyso-

somal degradative compartment after KGF-induced

internalization.

FGF10 and KGF target KGFR to different

endocytic pathways

It was recently demonstrated that the reduced degradation

of EGFR in response to transforming growth factor (TGF)a

is associated with enhanced EGFR recycling if compared

Figure 2: Legend on next page.

1858 Traffic 2007; 8: 1854–1872

Belleudi et al.

with EGF, which targets EGFR to the degradative pathway

(29). To assess if FGF10 and KGF can induce different

intracellular routing of KGFR, we analyzed in detail by

immunofluorescence confocal microscopy the endocytic

pathways followed by KGFR upon FGF10 or KGF treat-

ment. We identified the intracellular structures reached by

KGFR using specific markers for different endocytic com-

partments. To analyze the early steps of internalization,

HeLa KGFR cells were treated with KGF or FGF10 for 1 h

at 48C and then warmed for 100 at 378C, fixed, permea-

bilized and stained with anti-Bek polyclonal antibodies to

localize KGFR. Early endosomes were identified by anti-

EEA1 monoclonal antibody as above. Confocal analysis

showed that, after either KGF or FGF10 treatment, KGFR

significantly colocalized with EEA1-positive peripheral

dots, corresponding to early endosomes (Figure 4A). To

analyze the later steps of endocytosis, HeLa KGFR cells

treated with KGF or FGF10 as above were warmed for 300

at 378C and double stained with anti-Bek antibody and anti-

CD63 monoclonal antibody, a marker for multivesicular

bodies (MVBs) and late endosomes. Confocal analysis

(Figure 4B) and quantitative evaluation of the extent of

colocalization performed as reported in Materials and

Methods (Figure 4D) showed that at this time-point of

endocytosis, KGF, but not FGF10, targets KGFR to peri-

nuclear CD63-positive dots (48 and 18% of KGFR signal

colocalizing with CD63 signal, respectively). At a later time-

point of internalization (1 h at 378C), KGFR appeared in

endocytic structures identified as lysosomes by the pre-

sence of LysoTracker after KGF (57% of colocalization) but

not after FGF10 treatment (16% of colocalization) (Figure

4C,D). To identify the late structures reached by KGFR

after FGF10 treatment, HeLa KGFR cells were treated with

KGF or FGF10 for 1 h at 48C and incubated before fixation

for 1 h at 378C in the presence of Transferrin–Texas Red

(Tf–TxRed) to identify the recycling compartment. Confo-

cal analysis showed that FGF10 induced colocalization of

KGFR with Tf (47%) in a juxtanuclear region corresponding

to recycling compartment, while colocalization of the two

signalswasmuch lower afterKGF treatment (19%) (Figure 4

C,D). Thus, KGFR is targeted toMVBs, late endosomes and

lysosomes for degradation upon KGF treatment, whereas it

reaches the recycling compartment after FGF10 treatment.

To obtain a specific block of the degradative pathway,

we interfered with the activity of TSG101, a subunit of

endosomal sorting complex required for transport 1

(ESCRT1) complex that plays a key role in endosomal

morphogenesis and in EGFR sorting (30,31): in fact,

depletion of TSG101 retains EGFR to early endosomes,

preventing its degradation, whereas only a modest effect

is observed on the recycling pathway (32). To investigate

how TSG101 silencing could affect KGFR endocytic path-

ways, we performed coinjection of KGFR cDNA and small

interfering RNA (siRNA) for TSG101 in HeLa cells to

simultaneously obtain KGFR overexpression and TSG101

depletion. To verify if depletion of TSG101 could affect

KGFR routing, we treated coinjected cells with KGF or

FGF10 for 1 h at 48C and then warmed to 378C up to 2 h.

Double immunofluorescence revealed that, after KGF

treatment, the receptor signal remained associated with

the early endocytic compartment, as demonstrated by

colocalization with EEA1, and was not transported to

MVBs/late endosomes positive for CD63, to lysosomes

identified by the presence of LysoTracker or to the

recycling compartment containing Tf–TxRed internalized

for 1 h at 378C (Figure 5A). In contrast, after FGF10-

induced endocytosis, KGFR intracellular signal colocalized

with Tf–TxRed in the recycling compartment (Figure 5A),

suggesting that FGF10, differently from KGF, induces

KGFR sorting to a TSG101-independent endocytic path-

way, which targets the receptor to recycling. The effi-

ciency of TSG101 silencing was evaluated transfecting

HeLa cells with TSG101 siRNA. The expression of the

TSG101 protein was then evaluated by Western blot

analysis; as shown in Figure 5B, the TSG101 protein

Figure 2: The KGFR tyrosine phosphorylation and internalization in response to KGF or FGF10. A) The NIH3T3 KGFR cells were

serum-starved and treated with KGF or FGF10 for 30 or 100 at 378C and then lysed. Immunoprecipitation with anti-Bek antibodies and

Western blot analysis with anti-phosphotyrosine antibody show that KGFR is similarly tyrosine phosphorylated after treatment with KGF

or FGF10. B) The NIH3T3 KGFR and NIH3T3 EGFR cells were serum-starved and treated with KGF, FGF10 or EGF for 100 at 378C in

the presence of different doses of the FGFR kinase inhibitor SU5402. Cells were then fixed, permeabilized and stained with anti-

phosphotyrosine antibody followed by FITC-conjugated secondary antibody and with DAPI to visualize the cell nuclei. Percent of cells

presenting phosphotyrosine signal at the plasma membrane was calculated as described in Materials and Methods. The fluorescence

staining at the plasmamembrane, evident upon treatment with both ligands, progressively decreases in cells treated with SU5402, 30 mM

and 50 mM. In contrast, the drug does not interfere with the phosphotyrosine plasma membrane staining in NIH3T3 EGFR cells treated

with EGF. GF, growth factors. Bars: 10 mm. C) The HeLa cells transiently transfected with human KGFR wild type (WT) or with a kinase-

negative mutant KGFR Y656F/Y657F were serum-starved, treated with KGF or FGF10 for 1 h at 48C in the presence of SU5402 50 mM and

then warmed to 378C for 300 to induce receptor endocytosis. To label KGFR expressed on the cell surface, cells were treated with anti-Bek

polyclonal antibodies, directed against the extracellular portion of the receptor. Plasma membranes were decorated with FITC-conjugated

lectin WGA, and early endosomes were stained with anti-EEA1 monoclonal antibodies. The block of the KGFR WT kinase activity by

SU5402 inhibits KGFR uptake after either FGF10 or KGF treatment. The kinase-negative KGFR Y656F/Y657F is not internalized following

treatment with both ligands. Quantitative analysis of the percentage of colocalization was performed by serial optical sectioning and 3D

reconstruction as described inMaterials andMethods. Images obtained by 3D reconstruction of a selection of three out of the total number

of the serial optical sections are shown: the selected sections are all central and crossing the nucleus visualized by DAPI staining.

Bars: 10 mm.

Traffic 2007; 8: 1854–1872 1859

KGFR Endocytic Trafficking

Figure 3: The KGFR degradation after treatment with KGF or FGF10. A) The NIH3T3 KGFR cells were serum-starved, treated with

KGF and FGF10 for different time-points (3–8 h) at 378C in the presence of cycloheximide. In KGF-treated cells,Western blot analysis using

anti-Bek antibodies shows receptor degradation, assessed by the gradual disappearance of the band corresponding to KGFR after 3–8 h at

378C. In FGF10-treated cells, the band corresponding to KGFR remains virtually unchanged up to 8 h. In the absence of growth factors,

very low constitutive KGFR degradation was observed only at the 8-h time-point. Densitometric analysis of the bands from unstimulated

and KGF- or FGF10-stimulated samples: the values from three independent experiments were normalized, expressed as fold increase with

respect to untreated control values and reported in graph as mean values � SE. Student’s t-test was performed to evaluate significative

differences (*p < 0.05 versus untreated, **p < 0.05 versus 3 h and ***p < 0.05 versus 6 h). B) The HeLa KGFR cells were treated with

KGF or FGF10 for 1 h at 48C and then warmed to 378C in presence of LysoTracker. Cells were then double stained with anti-Bek antibody

and anti-giantin antibodies. The KGFR appears in intracellular dots and codistribute with LysoTracker-red in cells treated with KGF but not

in cells treated with FGF10. Bar: 10 mm.

1860 Traffic 2007; 8: 1854–1872

Belleudi et al.

expression appeared drastically downregulated in siRNA-

transfected cells, and the densitometric analysis revealed

80% of decrease with respect to the control value.

FGF10 induces recycling of KGFR to the plasma

membrane

To investigate if FGF10 could induce KGFR reappearance

to the plasma membrane or receptor accumulation in the

recycling compartment, we treated NIH3T3 KGFR cells

with cycloheximide as described inMaterials and Methods

to inhibit KGFR neosynthesis. Cells were then incubated

with KGF or FGF10 for 1 h at 48C and then warmed to 378Cfor different time-points (300 and 4 h) before fixation. Cells

were then permeabilized (Figure 6A,B) and stained with

the anti-Bek polyclonal antibodies, which recognize the

extracellular portion of the receptor. Parallel experiments

were performed on unpermeabilized cells (not shown).

Treatment with cycloheximide was able to efficiently

Figure 4: Endocytic pathways followed by KGFR after treatment with FGF10 or KGF. A–C) The HeLa KGFR cells were serum-

starved, treated with KGF or FGF10 for 1 h at 48C, then warmed to 378C for 100(A), 300(B) or 1 h (C) to allow receptor internalization and

immunolabeled with anti-Bek polyclonal antibodies (green). Early endosomal compartment and MVBs/late endosomes were respectively

identified by anti-EEA1 antibody (red) and anti-CD63 antibody (red). Late endosomal/lysosomal compartment was identified by

LysoTracker-red internalized for 300 at 378C, whereas recycling compartment by Tf–TxRed internalized for 1 h at 378C. Confocal analysisshows that, after 100 of warming, in either KGF-treated or FGF10-treated cells, the KGFR staining appears punctate and colocalizes with

EEA1 staining in early endosomes (A). At later time-points of internalization (300 at 378C), KGFR signal appears localized in MVBs/late

endosomes positives for CD63 after KGF but not after FGF10 treatment (B). After 1 h of warming at 378C, KGFR staining is localized in

perinuclear lysosomes positives for LysoTracker upon treatment with KGF, whereas the treatment with FGF10 induces colocalization of

KGFR signal with Tf in the recycling compartment (C). Bars: 10 mm. D) Quantitative analysis of the extent of colocalization of KGFR with

different endocytic markers.

Traffic 2007; 8: 1854–1872 1861

KGFR Endocytic Trafficking

inhibit KGFR neosynthesis, as demonstrated by the disap-

pearance of the perinuclear Golgi signal corresponding to

neosynthesized KGFR (Figure 6A). After 300 of treatment,

both FGF10 and KGF induced disappearance of the KGFR

signal from the plasma membrane and its concentration

in intracellular endocytic dots (Figure 6B). At later time-

points (4 h) of treatment with KGF, KGFR signal appeared

reduced and concentrated in endocytic perinuclear dots,

whereas after FGF10 treatment, the receptor signal was

not only still present in intracellular dots but also newly

visible on the plasma membrane, suggesting KGFR re-

cycling at the cell surface (Figure 6B). The quantitative

analysis of the fluorescence intensity, performed as

described in Materials and Methods in permeabilized

(Figure 6C) or unpermeabilized (Figure 6D) cells, con-

firmed that after treatment for 300 at 378C with both

ligands, most part of the receptor fluorescence signal

was present into the cell and only a minority on the plasma

membrane. At 4 h after the induction of internalization, the

signal appeared decreased in the intracellular dots after

both FGF10 and KGF treatment (Figure 6C), but it was

significantly increased on the plasma membrane after

FGF10 treatment (Figure 6D). Thus, differently from KGF,

FGF10 induces the reappearance of the receptor on the

plasma membrane.

It is known that the alternative endocytic destinations of

EGFR induced by EGF or TGFa correlate with differences

in the mitogenic potency of these ligands (29). In fact,

EGFR recycling after TGFa internalization is responsible for

a stronger mitogenic effect compared with that exerted by

EGF. In order to analyze the biological significance of the

different KGFR routing, we compared the proliferative

response with KGF and FGF10 of HaCaT human keratino-

cytes expressing endogenous KGFR. Cells were treated

with the two ligands at a concentration of 20 ng/mL for

48 h and then fixed and stained with an antibody against

Ki67 antigen, which identifies cycling cells (Figure 6E).

Quantitative analysis of immunofluorescence experiments

was performed by counting for each treatment the

Figure 5: Effect of TSG101 depletion on KGFR intracellular routing. A) The HeLa cells were coinjected with cDNA KGFR and siRNA

TSG101, left at 378C for 24 h and then fixed or serum-starved, treated with KGF or FGF10 for 1 h 48C and warmed for 2 h at 378C before

fixation, staining and confocal analysis. Double stainingwith anti-Bek polyclonal antibodies (green) and anti-EEA1monoclonal antibody (red)

followed by confocal analysis show that EEA1 signal is associated to small punctate structures as well as clustered, larger dots (arrows). In

cells treated with KGF, KGFR signal (green) remains localized in early endosomes identified by the EEA1 marker and does not appear

associated with lysosomes containing LysoTracker, with CD63-positive MVBs or with the recycling compartment identified by the

presence of Tf internalized for 1 h at 378C. In cells treated with FGF10, KGFR signal colocalizes only with Tf in recycling endosomes.

Bar: 10 mm. B) Western blot analysis with anti-TSG101 antibody in HeLa cells transfected with siRNA to inhibit TSG101 protein expression

or untransfected. The equal loading was assessed with anti-actin antibodies.

1862 Traffic 2007; 8: 1854–1872

Belleudi et al.

Figure 6: Legend on next page.

Traffic 2007; 8: 1854–1872 1863

KGFR Endocytic Trafficking

percentage of Ki67-positive nuclei on a total of 500 cells,

randomly observed in 10 fields from two different experi-

ments, and values are expressed as mean � SE. Results

reported in graph (Figure 6E) indicated that, with respect

to control cells, the percentage of cells presenting positive

nuclei was higher in FGF10- (p < 0.0001) than in KGF-

treated cells (p < 0.0001). These results indicate that in

HaCaT cells, in which KGFR is endogenously expressed,

FGF10 has a higher mitogenic activity compared with KGF

as previously described in human keratinocyte model (23).

KGFR internalization induced by clathrin-dependent

endocytosis

It has been recently proposed that a caveolae-dependent

endocytosis of EGFR targets this receptor to the degrada-

tive pathway, whereas clathrin-dependent endocytosis

results in recycling of the receptor to the plasma mem-

brane (33). Thus, it is possible that different internalization

mechanisms (clathrin or caveolae dependent) could de-

termine the intracellular fate of the receptor. We have

previously demonstrated that KGF/KGFR complexes intern-

alize through clathrin-coated pits (12). To investigate if

KGF/KGFR and FGF10/KGFR internalization could occur

through different mechanisms, we used siRNA interfer-

ence to selectively inhibit the clathrin-dependent or

caveolae-dependent pathways through silencing of clathrin

heavy chain (CHC) and caveolin1. To this aim, the effi-

ciency of CHC or caveolin1 silencing was first evaluated

transfecting HeLa cells with the corresponding siRNA;

Western blot with anti-CHC and anti-caveolin1 antibodies,

followed by densitometric analysis, revealed 65 and 40%

of decrease with respect to control values, respectively

(Figure 7A). Then we performed coinjection of KGFR

cDNA and CHC or caveolin1 siRNA in HeLa cells to

simultaneously obtain KGFR overexpression and CHC or

caveolin depletion. Cells were left at 378C for 24 h and

fixed. Alternatively, cells were serum-starved, treated with

KGF or FGF10 for 1 h at 48C in a mixture with the anti-Bek

antibody directed against the extracellular portion of the

receptor and then warmed to 378C for 100 in the presence

of Tf–TxRed before fixation. Tf–TxRed internalization was

the positive control for clathrin-dependent endocytosis

because it is widely recognized that this molecule internal-

izes by clathrin-coated pits (17,34). Confocal analysis

showed that, in cells expressing KGFR and depleted for

CHC, KGFR signal appeared prevalently localized at the

plasma membrane after either KGF or FGF10 treatment,

and Tf–TxRed was not visible in endocytic dots (Figure 7B).

These results indicated that CHC silencing and the con-

sequent block of clathrin-dependent endocytosis induced

a strong inhibition of both Tf and KGFR internalization. In

contrast, in cells expressing KGFR and depleted for

caveolin1, KGFR signal colocalized with Tf–TxRed in early

endosomes after treatment with both ligands, indicating

that caveolin1 silencing and the consequent inhibition of

caveolae-dependent endocytosis does not affect either

KGFR and Tf internalization (Figure 7D). In a second set of

experiments, we used also drugs and treatments able

to selectively interfere with either clathrin or caveolae-

dependent endocytosis. To this aim, NIH3T3 KGFR cells

were grown in hypertonic medium containing sucrose

0.4 M, which is known to inhibit formation of clathrin-

coated pits at the plasma membrane (35), or with acetic

acid for cytosol acidification, which interferes with the

pinching off of the coated pits to form coated vesicle (36).

To exclude the possibility that KGFR could partially intern-

alize by caveolae-dependent endocytosis, we treated

NIH3T3 KGFR cells with filipin, a cholesterol-binding agent

that inhibits caveolae formation at the plasma membrane,

inducing caveolin1 redistribution from the plasma mem-

brane to endosomes (37). Cells were then incubated with

KGF or FGF10 for 1 h at 48C and warmed for 100 at 378C as

above. Confocal analysis showed that both the hypertonic

medium and cytosol acidification, but not the filipin treat-

ment, inhibit KGFR and Tf internalization (not shown). All

these findings indicate that both KGF and FGF10 treat-

ments induce KGFR internalization by clathrin-dependent

mechanisms and that caveolae are not involved in this

process.

To further confirm that KGFR internalization is mediated by

clathrin-coated pits, we performed immunoelectron

microscopy on NIH3T3 KGFR cells treated with KGF or

FGF10 for 1 h at 48C and warmed for 50 at 378C to allow

receptor clustering in endocytic pits. The immunolabeling

on the cell surface, performed with the anti-Bek antibody

directed against the extracellular portion of KGFR, was

Figure 6: The FGF10-induced recycling of internalized KGFR. A–D) The NIH3T3 KGFR cells were serum-starved, treated with

cycloheximide to inhibit KGFR neosynthesis and fixed (A) or incubated with KGF or FGF10 and warmed to 378C for different times (300 and4 h) before fixation (B). Cells were permeabilized (A–C) or not (D) before immunostaining with anti-Bek polyclonal antibodies. In A, the

treatment with cycloheximide induces the disappearance of KGFR signal, localized in the perinuclear Golgi area. After 300 of treatment with

the ligands, both FGF10 and KGF induce disappearance of KGFR signal from the plasma membrane and its concentration in intracellular

endocytic dots. At later time (4 h) posttreatment with KGF, KGFR signal appears reduced and concentrated in endocytic perinuclear dots,

whereas the receptor signal become visible also on the plasma membrane after FGF10 treatment. Quantitative analysis of KGFR

immunofluorescence signal was performed in permeabilized (C) or unpermeabilized (D) cells. Results represent the mean values � SE.

Bar: 10 mm. E) The HaCaT cells were serum-starved, treated with KGF or FGF10 20 ng/mL for 48 h at 378C, fixed and immunostained with

anti-Ki67 polyclonal antibodies, which identifies cycling cells. Quantitative analysis indicated that the percentage of cells presenting

positive nuclei was higher in FGF10-treated cells than in KGF-treated cells with respect to control untreated cells. Results reported in graph

represent the mean values � SE. Bar: 10 mm.

1864 Traffic 2007; 8: 1854–1872

Belleudi et al.

visualized by protein A–colloidal gold conjugates. In

untreated cells, KGFR labeling appeared excluded from

the clathrin-coated rims, whereas after both KGF and

FGF10 treatments, gold particles were inside morphologic-

ally identified clathrin-coated endocytic pits (Figure 7C),

confirming the results obtained by the inhibition of clathrin-

dependent endocytosis. Quantification of the immunogold

labeling in clathrin-coated pits or in non-coated invagin-

ations of the plasma membranes following KGF or FGF10

treatment was performed by the analysis for each ligand of

100 clathrin-coated pits, randomly photographed from

50 cell sections in two different experiments of surface

immunolabeling. Pits were considered positive for intern-

alized receptors when at least two gold particles

Figure 7: Legend on next page.

Traffic 2007; 8: 1854–1872 1865

KGFR Endocytic Trafficking

appeared associatedwith the invagination. The percentage

of positive clathrin-coated pits in untreated cells was very

low (ranging from 3 to 5%) becoming 41 and 36% in KGF-

treated and FGF10-treated cells, respectively, and demon-

strating that the receptor uptake appears to involve

clathrin-dependent mechanisms.

Taken together, these observations indicate that both KGF

and FGF10 ligands induce KGFR uptake through clathrin-

dependent endocytosis.

FGF10 does not induce detectable ubiquitination

of KGFR

For RTK downregulation, not only receptor phosphorylation

but also its ubiquitination is required (38). In fact, ligand-

induced ubiquitination of RTKs and of their substrates

involved in the control of endocytosis represents the post-

translational modification, essential to generate an ubiquitin-

based network that targets the modified receptors to the

degradative pathway (39). It has been recently demon-

strated that EGF and TGFa, which do not differ in inducing

EGFR phosphorylation but target the receptors to different

endocytic routing, promote different levels and kinetics of

receptor ubiquitination (29). The inefficientEGFRubiquitina-

tion induced by TGFa treatment correlates with enhanced

EGFR recycling, while EGF induces EGFR ubiquitination

and downregulation. In the case of FGFR family, it has been

recently observed that FGF1 targets FGFR1 to the degra-

dative pathway and FGFR4 to the recycling compartment

and that, after FGF1 treatment, FGFR1 appears strongly

ubiquitinated, whereas FGFR4 ubiquitination was hardly

detectable only after both receptor and ubiquitin over-

expression (16). Therefore, to investigate if the different

ligand-induced KGFR endocytic routing could correlatewith

the extent of receptor ubiquitination, we treated NIH3T3

KGFR cells with FGF10 and KGF for different time intervals

at 378C. Western blot analysis using anti-ubiquitin (UBI)

monoclonal antibody showed a high molecular weight

smear of a 140-kDa band, corresponding to the molecular

weight of human KGFR evident at 30, 50 and 100 upon KGF

but not FGF10 treatment (Figure 8A). Immunoprecipitation

experiments with anti-Bek antibodies followed by immuno-

blot with anti-UBI antibody confirmed that KGFR is ubiquiti-

nated by KGF (Figure 8B) but not after FGF10 treatment.

These results indicate that KGF, but not FGF10, is able to

induce KGFR ubiquitination.

FGF10 induces less efficient c-Cbl recruitment to FRS2

Recent reports showed that ubiquitination of FGFRs is

mediated by the ubiquitin ligase c-Cbl (19,40). It is well

known that c-Cbl forms a constitutive complex with Grb2

and that the FRS2 protein is responsible for Grb2-mediated

recruitment of c-Cbl to the activated receptors (40). To

investigate if the undetectable ubiquitination of KGFR in

response to FGF10 could be correlated to an inefficient

recruitment of the c-Cbl ubiquitin igase and to assess the

possibility that the recruitment of Grb2/c-Cbl complex

could be modulated by a qualitatively different phosphory-

lation of FRS2, we first analyzed the phosphorylation level

of FRS2 tyrosine 196, which is the main docking site for

Grb2 (41). The NIH3T3 KGFR cells were treated with the

ligands for 100 at 378C, lysed and immunoprecipitated with

anti-FRS2 antibody. Immunoblot analysis using anti-pFRS2

Y196 revealed that this tyrosine was more efficiently

phosphorylated by KGF compared with FGF10 (Figure 9A,

left panel). This finding could also explain the slight

difference in the total amount of FRS2 phosphorylation,

shown by immunoblot with anti-phosphotyrosine after

treatment with the two ligands (Figure 9A, left panel). To

determine whether the less efficient phosphorylation of

tyrosine 196 after FGF10 stimulation was a persistent

event, we treated NIH3T3 KGFR cells for different times

at 378C before lysis. Western blot analysis with anti-pFRS2

Y196 showed that this tyrosine is less phosphorylated

after treatment with FGF10 compared with KGF during the

early time-points of internalization (Figure 9A, right panel).

Western blot with anti-total ERK was used as equal loading

control. To verify if the role of c-Cbl in KGFR ubiquitination

could be modulated by a qualitatively different FRS2-

mediated recruitment, we cotransfected NIH3T3 cells with

human KGFR cDNA and human c-Cbl cDNA and we

treated cells with FGF10 and KGF as above. Immunopre-

cipitation with anti-c-Cbl and immunoblot with anti-GRB2

antibody confirmed that, also in our model, c-Cbl and GRB2

exist in a constitutive, ligand-independent complex (Fig-

ure 9B, top). Thenwe analyzed the possible ligand-dependent

interaction between Grb2/c-Cbl complex and FRS2. Im-

munoprecipitation with anti-FRS2 and immunoblot with

Figure 7: Clathrin-mediated internalization of KGFR after FGF10 or KGF treatment. A) Western blot analysis with anti-CHC and anti-

caveolin1 antibodies in HeLa cells transfected with CHC and caveolin1 siRNAs to inhibit proteins expression. The equal loading was

assessedwith anti-actin antibodies. B) Confocal analysis of HeLa cells coinjectedwith KGFR cDNA and CHC siRNA or caveolin1 siRNA, left

at 378C for 24 h and then fixed or serum-starved, treated with the anti-Bek antibody directed against the extracellular portion of the

receptor and with KGF or FGF10 for 1 h at 48C before warming to 378C for 100 in the presence of Tf–TxRed used as control of clathrin-

mediated endocytosis. In cells expressing KGFR and depleted for CHC, KGFR signal appears localized on the plasmamembrane after either

KGF or FGF10 treatment, and Tf–TxRed is not visible in endocytic dots, showing inhibition of clathrin-mediated endocytosis of KGFR.

In contrast, in cells expressing KGFR and depleted for caveolin1, KGFR signal and Tf–TxRed colocalize in early endosomes after treatment

with both ligands, indicating that caveolin1 silencing and inhibition of caveolae-dependent endocytosis does not affect either KGFR or Tf

internalization. Bars: 10 mm. C) Immunogold surface labeling with anti-Bek antibodies was performed in serum-starved NIH3T3 KGFR cells

treated with KGF or FGF10 for 1 h at 48C and warmed to 378C for 50 to induce endocytosis. After both KGF and FGF10 treatments, KGFRs

appear inside endocytic structures morphologically identified as clathrin-coated pits, whereas no gold particles in the endocytic pits are

observed in untreated cells. Bars: 100 nm.

1866 Traffic 2007; 8: 1854–1872

Belleudi et al.

anti-pFRS2 Y196 revealed again the different levels of

phosphorylation of this tyrosine upon treatment with the

two ligands (Figure 9B, top). Immunoblot with anti c-Cbl

antibody showed that a lower amount of this protein

coimmunoprecipitated with FRS2 after FGF10 treatment

compared with KGF (Figure 9B, bottom). These results

indicate that FGF10 stimulation induces c-Cbl recruitment

to KGFR/FRS2 complex less efficiently than KGF and that

this inefficiency depends on a reduced FGF10 ability to

induce tyrosine 196 phosphorylation in FRS2.

In conclusion, our results demonstrate that FGF10 and

KGF can target KGFR to different intracellular fates as

a consequence of a qualitatively different recruitment

of the intracellular substrate c-Cbl, which in turn affects

KGFR ubiquitination and downregulation.

Discussion

Endocytosis of receptor tyrosine kinases has long been

considered a mechanism to downregulate and to attenu-

ate the receptor signaling from the plasma membrane. In

addition, growing evidences demonstrate that, at least in

some cases, internalized receptors are still active in endo-

somes, suggesting that the endocytic process plays a key

Figure 8: The KGFR ubiquitination induced by KGF but not by FGF10. A) The NIH3T3 KGFR cells were serum-starved, treated with

KGF or FGF10 for different time intervals at 378C (30, 50, 100, 300 and 1 h) and then lysed. Western blot analysis with anti-ubiquitin

monoclonal antibody, which recognizes both mono- and poly-ubiquitinated proteins, and anti-Bek antibodies shows that a smear of the

140-kDa band corresponding to the molecular weight of human KGFR is evident at 30, 50 and 10’ after KGF but not after FGF10 treatments.

B) The NIH3T3 KGFR cells were treated with KGF or FGF10 as above, and cell lysates were immunoprecipitated with anti-Bek polyclonal

antibodies and immunoblotted with anti-ubiquitin monoclonal antibody: KGFRs appear ubiquitinated by KGF but not by FGF10 treatment.

Traffic 2007; 8: 1854–1872 1867

KGFR Endocytic Trafficking

role in redistribution and propagation of the signaling

(42,43). Thus, the assessment of tyrosine kinase receptor

intracellular routing induced by different ligands is essential

to explain possible differences in their biological functions.

In this paper, we analyzed the KGFR endocytic pathways

induced by the two ligands FGF10 and KGF. While a recent

report showed the essential role played by the receptor

type in FGFR endocytic destinations (16), here we found

that the sorting of the same FGFR is strongly affected by

the ligand type. In fact, our results demonstrate that two

different members of the FGF family, FGF10 and KGF,

which differ significantly in HSPGs and heparin require-

ment for ligand/receptor complex formation and for bio-

logical activity (21,23,24), target the same specific receptor

KGFR to alternatively intracellular routes. Immunofluores-

cence experiments and confocal analysis demonstrate

that, while KGF targets KGFR to the degradative pathway,

FGF10 induces KGFR sorting to the recycling compart-

ment from which the receptors partially are transported

back to the plasma membrane. The difference in KGFR

intracellular fate induced by KGF and FGF10 is further

Figure 9: The FRS2 phosphorylation and c-Cbl recruitment to KGFR/FRS2 complex after FGF10 or KGF treatment. A) The NIH3T3

KGFR cells were serum-starved, treated with KGF or FGF10 for 100 at 378C, lysed and immunoprecipitated with anti-FRS2 polyclonal

antibodies (left panel). Immunoblot with anti-FRS2 and anti-pFRS2 (Y196) polyclonal antibodies shows that tyrosine 196 of FRS2 is much

more efficiently phosphorylated by KGF compared with FGF10. Immunoblot with anti-phosphotyrosine monoclonal antibody reveals

a slight difference also in the total amount of FRS2 phosphorylation. When cells were treated with the ligands for different times before

lysis (right panel), Western blot with anti-pFRS2 (Y196) shows that the less efficient phosphorylation of tyrosine 196 after FGF10

stimulation is persistent. To assess equal loading, blots were stripped and reprobed with anti-ERKs antibodies. B) The NIH3T3 cells were

transiently cotransfected with human KGFR cDNA and human c-Cbl cDNA, serum-starved and treated with KGF or FGF10 for 100 at 378C.Immunoprecipitation with anti-c-Cbl or with rabbit normal IgG used as negative control followed by immunoblot with anti-GRB2 and anti-

c-Cbl antibodies confirms that c-Cbl and GRB2 are constitutively associated. Immunoprecipitation with anti-FRS2 and immunoblot

with anti-c-Cbl antibodies reveal a less efficient c-Cbl coimmunoprecipitation with FRS2 after FGF10 treatment compared with KGF, as

a consequence of the lower amount of tyrosine 196 phosphorylation in FRS2 induced by FGF10.

1868 Traffic 2007; 8: 1854–1872

Belleudi et al.

supported by the specific inhibition of MVB formation from

early endosomes obtained by TSG101 silencing (32),

which demonstrates in our model that KGF, but not

FGF10, induces KGFR sorting to the degradative pathway.

These results strongly remind the different endocytic

traffic followed by EGFR in response to EGF or TGFa: in

that model, a higher sensitivity to low pH of TGFa bound to

the receptor compared with EGF leads to a rapid ligand–

receptor dissociation and subsequent EGFR recycling to

the plasma membrane (44). Similarly, we show here that

FGF10 rapidly dissociates from the receptor upon pH

lowering, while KGF appears less sensitive and still asso-

ciated with the receptor at the acidic pH.

Analyzing in detail the KGFR activation triggered by the two

ligands, we found that the treatment with both FGF10 and

KGF induces quantitatively comparable levels of KGFR

tyrosine phosphorylation. Experiments performed using

a drug able to block the FGFR kinase activity or cells

expressing a mutant kinase-negative KGFR demonstrate

that KGFR phosphorylation and internalization are both

receptor kinase activity-dependent processes. However,

the two ligands show different ability to induce receptor

ubiquitination and downregulation, which in the case of

FGF10 appear virtually undetectable. These differences

are closely related to the distinct fate of the receptor

induced by the two ligands. In fact, our findings are

consistent with the intracellular route of internalized

FGFRs, as well as of RTKs in general, which is known to

depend on a posttranslational modification, the ligand-

dependent multiubiquitination operated by the ubiquitin

ligase c-Cbl, critical for targeting to the endocytic degrada-

tive pathway (17). It has been recently reported that the

recycling of EGFR to the plasma membrane induced by

TGFa correlates with a decreased c-Cbl recruitment as

well as with transient ubiquitination and reduced degrada-

tion of the receptor compared with the response to EGF,

which targets EGFR to the degradative pathway (29).

For FGFRs, Wong et al. (40) reported that the ligand-

dependent c-Cbl interaction with FGFR, mediated by

FRS2, induces FGFR ubiquitination and Cho et al. (19)

demonstrated that activated FGFR3 is multiubiquitinated

through a c-Cbl-dependent mechanism and targeted to

lysosomes for degradation. Moreover, it has been also

reported that FGFR1, which is destined to the degradative

pathway, is strongly ubiquitinated, whereas FGFR4, which

is targeted to the recycling compartments, appears to be

weakly ubiquitinated and only when both ubiquitin and

receptor are overexpressed (16).

The difference in KGFR ubiquitination might be explained

by qualitative changes in tyrosine phosphorylation induced

by the two ligands and might be a consequence of the

different requirement of HSPG and exogenous heparin in

FGF10/KGFR and KGF/KGFR complex formation. This

possibility is sustained by recent evidences showing that

the presence of HSPGs or heparin in ligand/receptor

complex can modulate FGFR1 signaling, modifying the

phosphorylation of tyrosine sites present on the intracel-

lular portion of the receptor (11). Therefore, we wondered

if a qualitatively different tyrosine phosphorylation of KGFR

upon FGF10 and KGF treatment might result in different

activation of substrates that could be involved in KGFR

endocytosis. FRS2 represents the principal FGFR consti-

tutively associated substrate, which recruits all the main

effector proteins involved in FGFR signaling and endo-

cytosis (8). In this paper, we demonstrate that FGF10

induces a lower total FRS2 tyrosine phosphorylation com-

pared with KGF. We demonstrate also that the tyrosine

phosphorylation of this protein is qualitatively different after

KGF and FGF10 treatment: in fact, the tyrosine 196 of FRS2,

which represents the main docking site for Grb2, is more

efficiently phosphorylated after KGF stimulation than after

FGF10 treatment. Aswe confirm here that also in our cellular

model the protein Grb2 exists in a constitutively associated

complex with c-Cbl, we might speculate that KGF and

FGF10 induce a different amount of c-Cbl recruitment to

activated KGFR: in fact, coimmunoprecipitation experi-

ments demonstrate that the stimulation with KGF induces

more efficient FRS2-mediated recruitment of c-Cbl com-

pared with FGF10. In conclusion, the difference in c-Cbl

recruitment through phosphorylation of tyrosine 196 of

FRS2 might explain the different ubiquitination of KGFR

induced by KGF and FGF10 and the consequent distinct

trafficking of FGF10/KGFR and KGF/KGFR complexes.

The qualitatively different phosphorylation of FRS2 does

not only affect the endocytic fate of KGFR but also the

amount of mitogenic signaling transduced by this receptor

and mediated by Grb2 activation. In the case of EGFR, the

distinct receptor routing induced by TGFa and EGF might

explain the different proliferative activity shown for these

two ligands in epithelial model systems and at low non-

saturating concentrations of the growth factors (29).

Similarly, we show here that in HaCaT human keratino-

cytes, that express endogenous KGFR, the two ligands

reveal a different ability to induce cell proliferation: in fact,

FGF10, inducing KGFR recycling to the plasma membrane,

is a more potent mitogen compared with KGF, as pre-

viously observed in primary human keratinocytes (23).

Moreover, as it is known that FGFs and their receptors

can translocate to the nucleus, where they appear to

regulate transcription of different pool of genes (45), and

that FGF and FGFR translocation across the membranes

could take place at the level of the recycling compartment

(46), the exciting possibility of translocation to the nucleus

of either FGF10 or KGFR from the recycling compartment

can not be excluded.

Materials and Methods

Cell linesThe NIH3T3 EGFR- (47) or NIH3T3 KGFR-transfected cells were cultured

in DMEM, supplemented with 10% fetal calf serum plus antibiotics. The

NIH3T3 cells were transiently cotransfected with pCEV27 vector containing

Traffic 2007; 8: 1854–1872 1869

KGFR Endocytic Trafficking

human KGFR cDNA (kindly provided by Dr Stuart Aaronson, New York, NY,

USA) and pcDNA3 vector containing human c-Cbl (kindly provided by

Dr B.J. Druker, Portland, OR, USA), and HeLa cells were transiently trans-

fected with pCEV27 vector containing human KGFR cDNA (HeLa KGFR)

using effectene transfection reagent (Qiagen GmbH), according to the

manufacturer’s instructions. Transfected cells were collected 48 h after

transfection for evaluation of protein expression and internalization assays.

The human keratinocyte HaCaT cell line was cultured in DMEM, supple-

mented with 10% FBS and antibiotics. The kinase-negative mutant KGFR

Y656F/Y657F was constructed by site-directed mutagenesis as previously

described (28). For treatments with growth factors, cells were serum-

starved for 12 h and then incubated with 100 ng/mL KGF (Upstate), with

100 ng/mL FGF10 (PeproTech) or with 100 ng/mL FGF1 for 100 at 378C in

presence of different doses of heparin to induce KGFR activation and then

fixed or with 100 ng/mL KGF or with 100 ng/mL FGF10 þ 0.3 mg/mL

heparin for 30–8 h at 378C to induce receptor activation and downregulation.

Alternatively, cells were starved for 12 h, washed with cold medium,

incubated with 100 ng/mL KGF or with 100 ng/mL FGF10 þ 0.3 mg/mL

heparin for 1 h at 48C and immediately fixed or washed with prewarmed

medium and incubated at 378C for different times to induce KGFR

endocytosis before fixation. For treatment with EGF, NIH3T3 EGFR cells

were serum-starved for 12 h and then incubated with 100 ng/mL EGF

(Upstate) for 1 h at 48C and fixed or washed with prewarmed medium and

incubated at 378C for different times to induce receptor internalization

before fixation. To evaluate the pH sensitivity of KGF and FGF10 binding to

KGFR, HeLa KGFR cells were serum-starved for 12 h and then incubated

with 100 ng/mL KGF or with 100 ng/mL FGF10 þ 0.3 mg/mL heparin for

1 h at 48C to induce ligand binding to the receptor. Cells were then washed

three times with 10 mM HEPES, 0.5 M acetic acid and 0.5 M NaCl and

adjusted to the desired pH using NaOH as previously described (47).

For proliferation analysis, HaCaT cells were serum-starved for 24 h,

incubated with KGF 20 ng/mL or with FGF10 20 ng/mL þ 0.3 mg/mL

heparin for 24 h at 378C and then fixed. For tyrosine kinase inhibition, cells

were incubated with 100 mM genistein (Sigma Chemicals) for 300 at 378Cbefore incubation with growth factors in presence of genistein. For

inhibition of KGFR activity, cells were incubated with a specific tyrosine

kinase FGFR inhibitor SU5402 (Calbiochem) (30 or 50 mM in culture

medium) for 1 h before incubation with growth factors diluted in presence

of SU5402. For chlorate treatment, cells were grown in the presence of

10 mM NaClO3 for 48 h and then treated with KGF and FGF10, diluted in

medium containing NaClO3 and different doses of heparin. To induce Tf

internalization, cells were incubated with 50 mg/mL Tf–TxRed (Molecular

Probes) at 378C for 100 or 1 h before fixation. To induce LysoTracker in-

ternalization, cells were incubated with 100 nM LysoTracker-red (Molecular

Probes) in DMEM for 300 at 378C before fixation. To inhibit the synthesis of

proteins, cells were treated with 25 mg/mL cycloheximide (Sigma) for 4 h at

378C before incubation with growth factors in presence of cycloheximide.

To inhibit caveolae-dependent endocytosis, cells were treated with DMEM

containing 3 mg/mL filipin complex (Sigma) for 300 at 378C and then

incubated with the growth factors as above in presence of filipin. To block

clathrin-dependent endocytosis, cells were incubated for 300 at 378C in

hypertonic medium, obtained by adding 0.4 M sucrose to DMEM and then

incubated with the growth factors as above diluted in hypertonic medium.

Alternatively, cell cytosol was acidified with 1 M acetic acid 1:100 in DMEM

(pH 5.0) for 100 at 378C.

For RNA interference and TSG101, CHC and caveolin1 silencing, HeLa cells

were transfectedwith 125 pM of siRNA for TSG101, CHC or caveolin1 (Santa

Cruz Biotechnology Inc.) using lipofectamine 2000 according to the manu-

facturer’s protocol. Twenty-four hours after transfection, siRNA-transfected

and untransfected cells were processed for Western blot analysis.

MicroinjectionMicroinjection was performed with an Eppendorf microinjector (Eppendorf)

and an inverted microscope (Zeiss). Injection pressure was set at 80–

100 hPa and the injection time at 0.5 seconds. A mixture of 100 nM siRNA

for TSG101, CHC or caveolin1 (Santa Cruz) and 100 ng/mL KGFR cDNA in

distillate water were microinjected in the cytoplasm to simultaneously

induce RNA interference and consequent TSG101, CHC or caveolin1

silencing and KGFR overexpression. Cells were starved for 24 h at 378C,treated with growth factors for 1 h at 48C, warmed for 100 or 2 h at 378Cand processed for immunofluorescence.

Immunofluorescence and confocal microscopyCells, grown on coverslips and incubated with KGF, FGF10 or EGF as above

were fixed in methanol for 40 at�208C or with 4% paraformaldehyde for 300

at 258C, followed by treatment with 0.1 M glycine for 200 at 258C and with

0.1% Triton-X-100 for additional 50 at 258C to allow permeabilization. For

double-staining experiments, cells were incubated with the following

primary antibodies: anti-phosphotyrosine monoclonal antibody (1:100 in

PBS; Upstate), anti-Bek monoclonal antibody (1:20 in PBS; C-8; Santa Cruz)

and anti-Bek polyclonal antibodies (1:20 in PBS; C-17;, Santa Cruz) directed

against the intracellular portion of KGFR, anti-Bek polyclonal antibodies

(1:50 in PBS; H-80; Santa Cruz) directed against the extracellular portion of

KGFR, anti-giantin polyclonal antibodies (1:50 in PBS; Covance), anti-EEA1

monoclonal antibody (1:50 in PBS; Biosciences), anti-CD63 (1:200 in PBS;

Biosciences) for 1 h at 258C. To identify cycling cells, immunostaining was

performed with anti-Ki67 (1:50 in PBS; Zymed Laboratories Inc.) polyclonal

antibodies. Nuclei were stained with (1:10 000 in PBS; Sigma). The primary

antibodies were visualized, after appropriate washing with PBS, using goat

anti-mouse IgG–FITC (1:50 in PBS; Cappel Research Products), goat anti-

mouse IgG–Alexa Fluor (1:50 in PBS; Molecular Probes), goat anti-rabbit

IgG–FITC (1:500 in PBS; Cappel Research Products), goat anti-mouse IgG–

TxRed (1:200 in PBS; Jackson Immunoresearch Laboratories) and goat anti-

rabbit IgG–TxRed (1:200 in PBS; Jackson Immunoresearch Laboratories) for

300 at 258C. Coverslips were finally mounted with 90% glycerol in PBS for

observation at a Zeiss Axiophot epifluorescence microscope (Zeiss). The

fluorescence signals were analyzed by recording and merging single-

stained images using a charge-coupled device digital camera SPOT-2

(Diagnostic Instruments Inc.) and IAS 2000 software (Delta Sistemi).

Quantitative analysis of the phosphotyrosine fluorescence intensity was

performed by the analysis of 100 cells for each sample in five different

fields, randomly taken from three different experiments; results are shown

as mean values � SE. Colocalization of fluorescence signals in endocytic

structures was analyzed by a Zeiss LSM5 Pascal Laser scan microscope; to

prevent cross talk between the two signals, the multitrack function has

been used. For the quantitative analysis of the KGFR immunofluorescence

signal on the cell surface or internalized after treatment with the growth

factors, aimed to assess the extent of colocalization of KGFR with theWGA

or EEA1 markers, cells were scanned in a series of 0.5-mm sequential

sections with an Apotome System (Zeiss) connected with an Axiovert 200

inverted microscope (Zeiss); image analysis was then performed by the

AXIOVISION software (Zeiss), and 3D reconstruction was obtained. Quantita-

tive analysis of the extent of colocalization was performed using Zeiss

KS300 3.0 Image Processing system (Zeiss). The mean � SE percent of

colocalization was calculated by analyzing a minimum of 30 cells for each

treatment, randomly taken from three independent experiments, p-values

were calculated using Student’s t-test, and significance level has been

defined as p < 0.05.

Surface labeling immunoelectron microscopyThe NIH3T3 KGFR cells were serum-starved for 12 h, treated or not with

KGF or FGF10 for 1 h 48C and than warmed for 50 at 378C to allow receptor

clustering in endocytic pits. Cells were then fixed with a mixture of 2%

paraformaldehyde and 0.2% glutaraldehyde in 0.1 M phosphate buffer for

2 h at 258C, incubated with anti-Bek polyclonal antibodies (H-80), directed

against the extracellular portion of the receptor for 2 h at 258C and then with

FITC-conjugated goat anti-rabbit IgG (1:10 in PBS) (Cappel Research

Products) for 1 h at 258C, followed by incubation with 18 nm diameter

colloidal gold particles (prepared by the citrate method) conjugated with

protein A (Pharmacia) diluted 1:10 in PBS. Cells were finally washed three

times in PBS and fixed with 2% glutaraldehyde in the same buffer for 1 h at

258C. Control experiments were performed by omission of the anti-Bek

polyclonal antibodies or of the secondary antibody from the labeling

1870 Traffic 2007; 8: 1854–1872

Belleudi et al.

procedure. For resin embedding, all samples were postfixed in 1% osmium

tetroxide in veronal acetate buffer (pH 7.4) for 1 h at 258C, stained with

0.1% tannic acid in the same buffer for 300 at 258C and with uranyl acetate

(5 mg/mL) for 1 h at 258C, dehydrated in acetone and embedded in Epon

812. Thin sections were examined unstained or poststained with uranyl

acetate and lead hydroxide.

Immunoprecipitation and Western blot analysisSubconfluent cultures of NIH3T3 KGFR and HeLa cells were lysed in

a buffer containing 50 mM HEPES pH 7.5, 150 mM NaCl, 1% glycerol, 1%

Triton-X-100, 1.5 mM MgCl2, 5 mM EGTA, supplemented with protease

inhibitors (10 mg/mL aprotinin, 1 mM phenylmethylsulphonyl fluoride and

10 mg/mL leupeptin) and phosphatase inhibitors (25 mM sodium orthovan-

adate, 20 mM sodium pyrophosphate and 0.5 M NaF); 50 mg of total protein

were resolved under reducing conditions by 7% SDS–PAGE and trans-

ferred to reinforced nitrocellulose (BA-S 83; Schleider and Schuell). The

membranes were blocked with 5% non-fat dry milk in PBS 0.1% Tween-20

and incubated with anti-phospho ERK (Santa Cruz), anti-TSG101 (Santa

Cruz), anti-caveolin1 (Biosciences) and anti-ubiquitin (FK2 clone; Affiniti

Research Products Ltd) monoclonal antibodies, anti-phospho FRS2 (Y196;

Cell Signaling), anti-Bek (C-17; Santa Cruz), anti-CHC (Santa Cruz), and anti-

ERK2 (Santa Cruz) polyclonal antibodies, followed by enhanced chemilumi-

nescence detection (ECL; Amersham).

For immunoprecipitation/immunoblotting experiments, cells were lysed

in 1% Triton-X-100, 50 mM HEPES containing 150 mM NaCl, 1% glycerol,

1.5 mM MgCl2, 5 mM EGTA and protease and phosphatase inhibitors as

above. One milligram of total protein was immunoprecipitated with 4 mg/mL

anti-Bek (C-17; Santa Cruz), anti-FRS2 (Santa Cruz), anti-c-Cbl (Santa Cruz)

polyclonal antibodies or rabbit normal IgG (Santa Cruz) used as negative

control. Immunocomplexes, aggregated with 50 mL of gamma-bind protein

A–Sepharose (Pharmacia), were washed four times with 0.6 mL of buffer.

The pellets were boiled in Laemli buffer for 5 min, and the protein was

resolved under reducing conditions by 7–10% SDS–PAGE and trans-

ferred to reinforced nitrocellulose (BA-S, Schleicher and Schuell). The

membranes were blocked with 3% BSA in PBS (0.05% Tween-20)

overnight and probed with anti-phosphotyrosine (Upstate), anti-ubiquitin

(FK2 clone; Affiniti Research Products) monoclonal antibodies, anti-Bek

(Santa Cruz), anti-GRB2 (Santa Cruz), anti-c-Cbl (Santa Cruz), anti-FRS2

(Santa Cruz), anti-phospho FRS2 (Y196; Cell Signaling) polyclonal anti-

bodies followed by ECL detection. To estimate the protein equal loading,

the membranes were rehydrated by being washed in PBS–Tween-20,

stripped with 100 mM mercaptoethanol and 2% sodium dodecyl sulfate

for 30 min at 558C and probed again with anti-Bek (C17) polyclonal

antibodies or anti-ubiquitin monoclonal antibody. Densitometric analysis

was performed using Quantity One Program (Bio Rad). Briefly, the signal

intensity for each band was calculated, and the background was

subtracted from the experimental values. The resulting values were

then normalized and expressed as fold increase with respect to the

control value.

Acknowledgments

This work was partially supported by grants from MIUR, from Ministero

della Salute and from Associazione Italiana per la Ricerca sul Cancro, Italy.

We are very grateful to Professor Brian J. Druker and Dr Michael Deininger

for their helpful discussions and for providing reagents.

References

1. Grose R, Dickson C. Fibroblast growth factor signaling in tumorigen-

esis. Cytokine Growth Factor Rev 2005;16:179–186.

2. Szebebenyi G, Fallon JF. Fibroblast growth factor as multi-functional

signaling factors. Int Rev Cytol 1999;185:45–106.

3. Dailey L, Ambrosetti D, Mansukhani A, Basilico C. Mechanisms

underlying differential responses to FGF signaling. Int Rev Cytol 1999;

16:233–247.

4. Eswarakumar VP, Lax I, Schlessinger J. Cellular signaling by fibroblast

growth factor receptors. Cytokine Growth Factor Rev 2005;16:139–149.

5. Cheon HG, LaRochelle WJ, Bottaro DP, Burgess WH, Aaronson SA.

High affinity binding sites for related fibroblast growth factor ligands

reside within different receptor immunoglobulin-like domains. Proc

Natl Acad Sci U S A 1994;91:989–993.

6. Mason IJ. The ins and outs of fibroblast growth factors. Cell 1994;

78:547–552.

7. Schlessinger J, Lax I, Lemmon M. Regulation of growth factor

activation by proteoglycans: what is the role of the low affinity

receptors? Cell 1995;83:357–360.

8. Pellegrini L. Role of heparan sulfate in fibroblast growth factor sig-

naling: a structural view. Curr Opin Struct Biol 2001;11:629–634.

9. Roghani M, Mansukhani A, Dell’Era P, Bellosta P, Basilico C, Rifkin DB,

Moscatelli D. Heparin increase the affinity of basic fibroblast growth

factor for its receptor but is not required for binding. J Biol Chem

1994;269:3976–3984.

10. Spivak-Kroizman T, Lemmon MA, Dikic I, Ladbury JE, Pinchasi D,

Huang J, Jaye M. Heparin-induced oligomerization of FGF molecules

is responsible for FGF receptor dimerization, activation and cell pro-

liferation. Cell 1994;269:14419–14423.

11. Lundin L, Ronnstrand L, Cross M, Hellberg C, Lindahl U, Claesson-

Welsh L. Differential tyrosine phosphorylation of fibroblast growth

factor (FGF) receptor-1 and receptor proximal signal transduction in

response to FGF2 and heparin. Exp Cell Res 2003;287:190–198.

12. Marchese C, Mancini P, Belleudi F, Felici A, Gradini R, Sansolini T,

Frati L, Torrisi MR. Receptor-mediated endocytosis of keratinocyte

growth factor. J Cell Sci 1998;111:3517–3527.

13. Belleudi F, Ceridono M, Capone A, Serafino A, Marchese C, Picardo M,

Frati L, Torrisi MR. The endocytic pathway followed by the keratinocyte

growth factor receptor. Histochem Cell Biol 2002;118:1–10.

14. Citores L, Wesche J, Kolpakova E, Olsnes S. Uptake and intracellular

transport of acidic fibroblast growth factor: evidence for free and

cytoskeleton-anchored fibroblast growth factor receptors. Mol Biol

Cell 1999;10:3835–3848.

15. Reilly JF, Mizukoshi E, Maher PA. Ligand dependent and independent

internalization and nuclear translocation of fibroblast growth factor

(FGF) receptor 1. DNA Cell Biol 2004;9:538–548.

16. Haugsten EM, Sorensen V, Brech A, Olsnes S, Wesche J. Different

intracellular trafficking of FGF1 endocytosed by the four homologous

FGF receptor. J Cell Sci 2005;118:3869–3881.

17. Citores L, Khnykin D, Sorensen V, Wesche J, Klingenberg O,

Wiedlocha A, Olsnes S. Modulation of intracellular transport of acidic

fibroblast growth factor by mutations in the cytoplasmic receptor

domain. J Cell Sci 2001;114:1677–1689.

18. Reiland J, Rapraeger AC. Heparan sulfate proteoglycan and FGF

receptor target basic FGF to different intracellular destinations. J Cell

Sci 1993;105:1085–1093.

19. Cho JY, Guo C, Torello M, Lunstrum GP, Iwata T, Deng C, Horton WA.

Defective lysosomal targeting of activated fibroblast growth factor re-

ceptor 3 in achondroplasia. Proc Natl Acad Sci U S A 2004;101:609–614.

20. Rubin JS, Osada H, Finch PW, Taylor GW, Rudikoff S, Aaronson SA.

Purification and characterization of a newly identified growth factor

specific for epithelial cells. Proc Natl Acad Sci U S A 1989;86:802–806.

21. Igarashi M, Finch PW, Aaronson SA. Characterization of recombinant

human fibroblast growth factor (FGF)-10 reveals functional similarities

with keratinocyte growth factor (FGF-7). J Biol Chem 1998;273:

13230–13235.

Traffic 2007; 8: 1854–1872 1871

KGFR Endocytic Trafficking

22. Marchese C, Rubin JS, Ron D, Faggioni A, Torrisi MR, Messina A, Frati L,

Aaronson SA. Human keratinocyte growth factor activity on proliferation

and differentiation of human keratinocytes: differentiation response

distinguishes KGF from EGF family. J Cell Physiol 1990;144:326–333.

23. Marchese C, Felici A, Visco V, Lucania G, Igarashi M, PicardoM, Frati L,

Torrisi MR. Fibroblast growth factor 10 induces proliferation and

differentiation of human primary cultured keratinocytes. J Invest

Dermatol 2001;116:623–628.

24. Berman B, Ostrovsky O, Shlissel M, Lang T, Regan D, Vlodavsky I,

Ishai-Micaeli R, Ron D. Similarities and differences between the effects

of heparin and glypican-1 on the bioactivity of acidic fibroblast growth

factor and the keratinocyte growth factor. J Biol Chem 1999;274:

36132–36138.

25. Reich-Slotky R, Bonneh-Barkay D, Shaoul E, Bluma B, Svahn CM, Ron D.

Differential effect of cell-associated heparan sulfates on the binding

of keratinocyte growth factor (KGF) and acidic fibroblast growth factor

to the KGF receptor. J Biol Chem 1994;269:32279–32285.

26. Rapraeger AC, Krufka A, Olwin BB. Requirement of heparan sulfate for

bFGF-mediated fibroblast growth and myoblast differentiation. Science

1991;252:1705–1714.

27. Mohammadi M, McMahon G, Sun L, Tang C, Hirth P, Yeh BK,

Hubbard SR, Schlessinger J. Structures of the tyrosine kinase domain

of fibroblast growth factor receptor in complex with inhibitors. Science

1997;276:5314–5955.

28. Belleudi F, Leone L, Aimati L, De Paola B, Cardinali G, Marchese C,

Frati L, Picardo M, Torrisi MR. Endocytic pathways and biological

effects induced by UVB-dependent or ligand–dependent activation of

the keratinocyte growth factor receptor. FASEB J 2006;20:395–397.

29. Alwan HA, van Zoelen EJ, van Leeuwen JE. Ligand-induced lysosomal

epidermal growth factor receptor (EGFR) degradation is preceded by

proteasome-dependent EGFR de-ubiquitination. J Biol Chem 2003;

278:35781–35790.

30. Babst M, Odorizzi G, Estepa EJ, Emr SD. Mammalian tumor suscep-

tibility gene 101 (TSG101) and the yeast homologue, Vps23p, both

function in late endosomal trafficking. Traffic 2000;1:248–258.

31. Bishop N, Horman A, Woodman P. Mammalian class E vps proteins

recognize ubiquitin and act in the removal of endosomal protein-

ubiquitin conjugates. J Cell Biol 2002;157:91–101.

32. Doyotte A, Russell RG, Hopkins CR, Woodman PG. Depletion of

TSG101 forms a mammalian ‘‘class E’’ compartment: a multicisternal

early endosome with multiple sorting defects. J Cell Sci 2005;118:

3003–3017.

33. Sigismund S, Woelk T, Puri C, Maspero E, Tacchetti C, Transidico P,

Di Fiore PP, Polo S. Clathrin-independent endocytosis of ubiquitinated

cargos. Proc Natl Acad Sci U S A 2005;102:2760–2765.

34. Hopkins CR, Trowbridge IS. Internalization and processing of trans-

ferrin and the transferrin receptor in human carcinoma A431 cells. J Cell

Biol 1983;97:508–521.

35. Heuser JE, Anderson RG. Hypertonic media inhibit receptor-mediated

endocytosis by blocking clathrin-coated pit formation. J Cell Biol 1989;

108:389–400.

36. Sandvig K, Olsnes S, Petersen OW, van Deurs B. Acidification of the

cytosol inhibits endocytosis from coated pits. J Cell Biol 1987;105:

679–689.

37. Carozzi AJ, Ikonen E, Lindsay MR, Parton RG. Role of cholesterol in

developing T-tubules: analogous mechanisms for T-tubule and caveo-

lae biogenesis. Traffic 2000;1:326–341.

38. Marmor M, Yarden Y. Role of protein ubiquitylation in regulating

endocytosis of receptor tyrosine kinases. Oncogene 2004;23:

2057–2070.

39. Le Roy C, Wrana JL. Clathrin- and non-clathrin-mediated endo-

cytic regulation of cell signalling. Nat Rev Mol Cell Biol 2005;

6:112–126.

40. Wong A, Lamothe B, Lee A, Sclessinger J, Lax I. FRS2 alpha attenuates

FGF receptor signaling by Grb2-mediated recruitment of the ubiquitin

ligase Cbl. Proc Natl Acad Sci U S A 2002;99:6684–6689.

41. Kouhara H, Hadari YR, Spivak-Kroizman T, Schilling J, Bar-Sagi D, Lax I,

Schlessinger J. A lipid-anchored Grb2-binding protein that links FGF-

receptor activation to the ras/MAPK signaling pathway. Cell 1997;89:

693–702.

42. Di Guglielmo GM, Baass PC, Ou WJ, Posner BI, Bergeron JJ.

Compartmentalization of SHC, GRB2 and mSOS, and hyperphosphory-

lation of Raf-1 by EGF but not insulin in liver parenchyma. EMBO J

1994;13:4269–4277.

43. Di Fiore PP, De Camilli P. Endocytosis and signaling an inseparable

partnership. Cell 2001;106:1–4.

44. Ebner R, Derynck R. Epidermal growth factor and transforming growth

factor-a: differential intracellular routing and processing of ligand-

receptor complexes. Cell Regul 1991;2:599–612.

45. Stachowiak MK, Fang X, Myers JM, Dunham SM, Berezney R, Maher

PA, Stachowiak EK. Integrative nuclear FGFR1 signaling (INFS) as

a part of a universal "feed-forward-and-gate" signaling module that

controls cell growth and differentiation. J Cell Biochem 2003;90:

662–691.

46. Bryant DM, Stow JL. Nuclear translocation of cell-surface receptors:

lessons from fibroblast growth factor. Traffic 2005;6:947–954.

47. Di Fiore PP, Pierce JH, Flemming TP, Hazard R, Ullrich A, King CR,

Schlessinger J, Aaronson SA. Overexpression of the human EGF

receptor confers an EGF-dependent transformed phenotype to NIH3T3

cells. Cell 1987;7:3365–3370.

1872 Traffic 2007; 8: 1854–1872

Belleudi et al.