Heterozygous mutations in the gene encoding fibroblast growth factor 10 (FGF10) or its cognate receptor, FGF-receptor 2 IIIb result in two human syndromes — LADD (lacrimo-auriculo-dento-digital) and ALSG (aplasia of lacrimal and salivary glands). To date, the partial loss-of-FGF10 function in these patients has been attributed solely to perturbed paracrine signalling functions between FGF10-producing mesenchymal cells and FGF10-responsive epithelial cells. However, the functioning of a LADD-causing G138E FGF10 mutation, which falls outside its receptor interaction interface, has remained enigmatic. In the present study, we interrogated this mutation in the context of FGF10's protein sequence and three-dimensional structure, and followed the subcellular fate of tagged proteins containing this or other combinatorial FGF10 mutations, in vitro. We report that FGF10 harbours two putative nuclear localization sequences (NLSs), termed NLS1 and NLS2, which individually or co-operatively promote nuclear translocation of FGF10. Furthermore, FGF10 localizes to a subset of dense fibrillar components of the nucleolus. G138E falls within NLS1 and abrogates FGF10's nuclear translocation whilst attenuating its progression along the secretory pathway. Our findings suggest that in addition to its paracrine roles, FGF10 may normally play intracrine role/s within FGF10-producing cells. Thus, G138E may disrupt both paracrine and intracrine function/s of FGF10 through attenuated secretion and nuclear translocation, respectively.

Introduction

Lacrimo-auriculo-dento-digital (LADD) (OMIM 149730) and aplasia of lacrimal and salivary glands (ALSG) (OMIM 180920) are rare autosomal dominant syndromes caused by a spectrum of heterozygous missense mutations in fibroblast growth factor 10 (FGF10) and FGF-receptor 2 (FGFR2) genes [14] (Table 1). These syndromes are characterized by defects in tear and saliva production, accompanied by subtle craniofacial, limb, pulmonary and urogenital abnormalities, complementary to milder phenotypes observed in mice that are heterozygous for FGF10 or its cognate receptor, FGF-receptor 2 IIIb (FGFR2-IIIb) [5,6]. Structural modelling and biochemical studies have attributed the human defects to impaired FGF10–FGFR2-IIIb interaction, or production of unstable proteins [7] (Table 1). Three of the FGF10 mutations occur outside its receptor-interacting interface, with two of these (W169X and K137X) predicted to generate a truncated non-functional protein (Table 1). However, the mechanism by which the third mutation — a Glycine (G) to Glutamic Acid (E) substitution at residue 138 — causes LADD [4] is unknown, and its apparent lack of involvement in receptor binding has raised the interesting possibility that FGF10 functions in multiple ways.

Table 1
Spectrum of LADD/ALSG syndrome-causative mutations
Mutation/deletion Amino acid change Syndrome Molecular consequence Reference(s) 
FGF10 mutations 
 237G → A (exon 1) W79X ALSG Predicted truncated protein Seymen et al. [39
 240A → C (exon 1) R80S ALSG Disruption of FGF10/FGFR2-IIIb-binding site Entesarian et al. [4
Chapman et al. [3
 317G → T (exon 1) C106F1 LADD Product unstable at physiological temperature Rohmann et al. [2
Shams et al. [7
Deletion (of exons 2 and 3) Not applicable ALSG No protein product Entesarian et al. [42
 409A → T (exon 2) K137X1 ALSG and LADD Predicted truncated protein Milunsky et al. [1
 413G → A (exon 2) G138E* ALSG/LADD Unknown Entesarian et al. [4
 430G → A (intron 2) n/a ALSG Truncated protein Scheckenbach et al. [41
 467T → G (exon 2) I156R LADD Disruption of the FGF10/FGFR2-IIIb-binding site Milunsky et al. [1
Shams et al. [7
 506G → A (exon 3) W169X1 ALSG Predicted truncated protein Milunsky et al. [1
 577C → T (exon 3) R193X ALSG Truncated protein; partial abrogation of FGF10/FGFR2-IIIb-binding Entesarian et al. [42
 620A → C (exon 3) H207P No discernable phenotype Rare polymorphism Entesarian et al. [4
FGFR2 mutations 
 1882G → A A628T LADD Although within the ligand-binding domain, affects tyrosine kinase catalytic activity Rohmann et al. [2], Shams et al. [7], Lew et al. [40
 1942G → A (exon 16) A648T LADD Affects ligand binding Rohmann et al. [2
 Δ1947-AGA-1949 base depletion (exon 16) R649S and ΔD650 LADD Affects ligand binding Rohmann et al. [2
Mutation/deletion Amino acid change Syndrome Molecular consequence Reference(s) 
FGF10 mutations 
 237G → A (exon 1) W79X ALSG Predicted truncated protein Seymen et al. [39
 240A → C (exon 1) R80S ALSG Disruption of FGF10/FGFR2-IIIb-binding site Entesarian et al. [4
Chapman et al. [3
 317G → T (exon 1) C106F1 LADD Product unstable at physiological temperature Rohmann et al. [2
Shams et al. [7
Deletion (of exons 2 and 3) Not applicable ALSG No protein product Entesarian et al. [42
 409A → T (exon 2) K137X1 ALSG and LADD Predicted truncated protein Milunsky et al. [1
 413G → A (exon 2) G138E* ALSG/LADD Unknown Entesarian et al. [4
 430G → A (intron 2) n/a ALSG Truncated protein Scheckenbach et al. [41
 467T → G (exon 2) I156R LADD Disruption of the FGF10/FGFR2-IIIb-binding site Milunsky et al. [1
Shams et al. [7
 506G → A (exon 3) W169X1 ALSG Predicted truncated protein Milunsky et al. [1
 577C → T (exon 3) R193X ALSG Truncated protein; partial abrogation of FGF10/FGFR2-IIIb-binding Entesarian et al. [42
 620A → C (exon 3) H207P No discernable phenotype Rare polymorphism Entesarian et al. [4
FGFR2 mutations 
 1882G → A A628T LADD Although within the ligand-binding domain, affects tyrosine kinase catalytic activity Rohmann et al. [2], Shams et al. [7], Lew et al. [40
 1942G → A (exon 16) A648T LADD Affects ligand binding Rohmann et al. [2
 Δ1947-AGA-1949 base depletion (exon 16) R649S and ΔD650 LADD Affects ligand binding Rohmann et al. [2

Currently known spectrum of mutations in FGF10 or FGFR2-IIIb that results in ALSG and LADD syndromes, with their predicted or characterized molecular properties.

*

FGF10 mutations occurring outside its cognate receptor's binding site. To date, the mode of action for a G138E mutation (rodent equivalent, G145E) has remained unknown. A rare FGFR3 mutation has also been described [2].

Fibroblast growth factors (FGFs) are a 22-member strong family of 17–34 kDa proteins with critical roles in embryonic and adult tissue growth and homoeostasis [8]. FGFs have been classified into subfamilies according to peptide sequence homology, shared biochemical properties and biological functions. Most FGFs are secreted ligands, signalling through one of four tyrosine kinase trans-membrane fibroblast growth factor receptors (FGFRs), in co-operation with sulphated proteoglycans. FGF–FGFR binding specificity is determined in part by alternative splicing of exons encoding FGFRs' third immunoglobulin-like domains, yielding the so-called ‘IIIb’ and ‘IIIc’ isoforms. Moreover, paracrine signalling and functionality is determined by tissue-specific and mutually exclusive expression of FGF ligands and their corresponding receptors. Typically, FGF10 is secreted by mesenchymal cells to regulate epithelial cell growth and branching morphogenesis by activating the epithelially expressed FGFR2-IIIb isoform [911].

In addition to their paracrine signalling role, several FGFs are thought to function cell-autonomously, via intracellular partnership with scaffolding proteins [12]; receptor and non-receptor mediated reuptake into cells [13]; interaction with FGFRs within endosomes [14]; and translocation into the nucleus/nucleolus [15]. In some FGFs, this functional diversity is achieved via the generation and differential targeting of low versus high molecular weight (HMW) isoforms. For example, HMW FGF2 translocates to the nucleus to stimulate cell proliferation and negatively regulate bone mineralization [16,17]. In contrast, nuclear FGF3 inhibits cell proliferation [18].

Since the LADD-causing G138E mutation lies outside the FGF10–FGFR2 interaction interface, in the present study we sought alternative explanation/s for its loss-of-function effects. Using bioinformatics, structural modelling, site-directed mutagenesis and intracellular trafficking analysis, we show that FGF10 harbours two putative NLSs (nuclear localization sequences), with G138E falling within one of these. In contrast with the wild-type (WT), rat proteins bearing the equivalent mutation (G145E), or compound mutations in a second putative NLS site, fail to traffic into the nucleus and become hyperglycosylated in the cytoplasm. Moreover, G145E fails to progress through the intracellular secretory pathway, and transient overexpression of WT or FGF10 mutant proteins inhibit mesenchymal cell proliferation and differentiation in vitro. Our results suggest that nuclear trafficking promoted by two putative NLS motifs may be an important aspect of FGF10 functionality. Hence, G138E mutation may work by reducing the bioavailability of FGF10 at two levels — as a secreted form to FGF10-responding cells, and as a nuclear form within FGF10-producing cells themselves.

Material and methods

Bioinformatics analysis

Primary protein sequences were obtained from the UniProt database and sequence alignments were performed using ClustalW2. Protein 3D models were constructed in Chimera 1.9 using RCSB Protein Data Bank files (1nun). The search for NLS sequences was performed on rat and mouse FGF10 sequences using NLStradamus and NucPred software tools (Supplementary Figure S1), freely available online.

Cloning and generation of mutant constructs

Using the appropriate restriction enzyme or HA (haemagglutinin A) tag encoding primers, a C-terminus HA-tagged FGF10 construct (FGF10-HA) was generated by PCR from a template plasmid encoding rat FGF10 cDNA (gift of Prof. Saverio Bellusci). These were scanned for undesirable mutations by Sanger sequencing before cloning as an EcoRI–NotI fragment into a mammalian expression vector, pN1, replacing an existing mCherry encoding fragment (Supplementary Figure S2A; Clontech Laboratories Inc.). All point mutation-bearing FGF10 inserts (i.e. R/K to T substitutions) were generated by PCR using a QuikChange Lightning Site-Directed Mutagenesis Kit (Agilent Technologies), and the relevant custom designed mutation-bearing primers (Supplementary Figure S2B). At the end of each reaction, residual WT FGF10-HA template was cleaved by DpnI restriction digest, and the desired mutation-bearing products were selected by Sanger sequencing.

Cell cultures and plasmid transfections

Human retinal pigment epithelial (ARPE19) and embryonic kidney (293T) cells were maintained in DMEM/F12 HAMs supplemented with 10% foetal bovine serum (FBS; Invitrogen), supplemented with 0.1% gentamycin or 1% penicillin/streptomycin and 1% GlutaMAX (Gibco), respectively. ATDC5 (chondrogenic mesenchymal) cells were grown in DMEM/F12 HAMs (Gibco 21331) supplemented with 5% FBS; 1% penicillin/streptomycin; 1% GlutaMAX; 30 nM sodium selenite and 10 µg/ml human Transferrin. Primary brain cultures were established by enzymatic digest and dissociation of freshly isolated hypothalamus from brains of 3–6-week-old WT mice. Hypothalamic cells were then grown as a monolayer in DMEM/F12 HAMs (Gibco 21331) supplemented with 5% FBS; 1% penicillin/streptomycin; 1% GlutaMAX; B27 Supplement (Gibco); 35 µg/ml bovine pituitary extract (Invitrogen) and 20 ng/ml each of EGF and FGF2 (Peprotech). For transient transfections, cells were seeded at 10–40 × 103 on poly-d-lysine (20 µg/ml) coated glass coverslips placed in a 12- or 24-well plate. At 80% confluency, cells were transfected for 4 h using JetPrime Reagent (Polyplus transfections) allowing one part DNA to two Parts JetPrime, aiming for 1 µg of plasmid DNA per well. At 24, 48 or 72 h post-transfection, cells were fixed for 15 min in a 4% paraformaldehyde (PFA; pH 7.4) solution.

Immunohistochemistry, microscopy and imaging

In preparation for immunolabelling, PFA-fixed cells were treated with 1% NP-40 and blocked in 10% normal goat serum (NGS) solution for 1 h at room temperature, before overnight incubation at 4°C with primary antibodies diluted in 0.2% NGS/PBS. These were: mouse anti-HA (1:1000; Cell Signalling); rat anti-BrdU (1:1000; Thermo Scientific PA5-33256): rabbit anti-TGN46, ERP60 (an isoform of Protein disulphide-isomerase) and βCOP (coat protein (coatomer) β) (1:500 each; kind gift of Prof T. Wileman and Dr P. Powell); rabbit anti-dt Tomato dsRred (1:1000; Clontech) and rabbit anti-Fibrillarin (1:1000; Abcam). Cells were washed three to five times in 0.2% NGS/PBS and the relevant species-specific secondary antibodies conjugated to Alexa488 or Alexa568 (1:1000) were applied for 1 h at room temperature. Coverslips carrying immunolabelled cells were mounted onto glass slides using Vectashield mounting medium (Vector Labs) and visualized using a Zeiss Axioplan2 microscope, with or without an Apotome attachment, under identical optical threshold settings. Images were acquired and processed using the Axiovision 4.8 and ImageJ software. A typical experiment (per construct) consisted of transfecting two coverslips and taking measurements from 10 random areas, i.e. five from each coverslip. Within each photographed area, total cell number (ranging from 60 to 110 cells per area) as well as the proportion of transfected cells was determined, with the latter comprising ∼20–30% of the total. The pattern of HA localization in each transfected cell was classified into three categories: exclusively or predominately nuclear; exclusively or predominantly cytoplasmic; and equal cytoplasmic and nuclear distribution.

The degree of HA-tag and secretory pathway marker co-localization was measured using the Volocity 6.3 software. Briefly, 10 randomly selected fluorescent cells from each immunolabelling combination were photographed under the relevant channels — red (Alexa 568) for HA and green (Alexa 488) for secretory pathway markers (either ERp60, βCOP or TGN46). After setting the appropriate thresholds, a scatter plot of co-localizing markers was generated, from which Pearson's correlation coefficient (Rr) was calculated. Values range from −1.0 to 1.0, where values above 0.5 signify co-localization, and values below 0.5 indicate negative or no significant correlation between markers.

Subcellular fractionation

Cell fractionation (cytoplasmic versus nuclear) was carried out following the protocol of Dimauro et al. [19]. In brief, 24 h post-transfection, 293T cells were washed in cold PBS and pelleted. Cytoplasmic fraction was extracted by treating cells with STM buffer [250 mM sucrose; 50 mM Tris–HCl (pH 7.4); 5 mM MgCl2] in the presence of 1% Halt protease inhibitor cocktail (Thermo Fisher Scientific). Cells were then washed in STM buffer, precipitated in pre-cooled acetone and resuspended in STM buffer. The nuclear fraction was extracted in NET buffer [20 mM HEPES (pH 7.9); 1.5 mM MgCl2; 0.5 M NaCl; 0.2 mM EDTA; 20% glycerol; and 1% TritonX-100] in the presence of 1% Halt cocktail, incubated on ice for 45 min and centrifuged at 9000×g to remove debris.

Western immunoblotting

ARPE cells cultured on 10 cm petri dishes were transfected with 3 µg of relevant plasmids, and 24 h later lysed in ice-cold modified RIPA buffer [150 mM NaCl; 50 mM Tris; 1.25 mM EDTA; 1% Triton, 1% sodium deoxycholate; and 0.1% sodium dodecyl sulphate (SDS)] containing protease inhibitors. Total protein concentration was determined using a BCA assay kit (Thermo Fisher Scientific) and cell lysates were stored at −20°C until use. Protein lysate was subjected to SDS–PAGE, electrophoretic semi-dry transfer to nitrocellulose membranes and incubated with primary mouse antibodies against β-actin (1:1000) and HA (1:1000; Cell Signalling) at 4°C for 24 h. Subsequently, HRP-conjugated goat-anti-mouse antibodies were applied (1:1000) and membranes were incubated in ECL solution [100 mM Tris (pH8); 1.25 mM luminol; 0.2 mM coumaric acid; and 0.01% H2O2] and exposed to Hyperfilm ECL (for 5–30 s for β-actin; 15 min for HA).

Cell proliferation assays

At 20 h post-transfection with the relevant constructs, ATDC5 cells were pulsed with 3 µg/ml BrdU for 4 h, fixed in 4% PFA and treated with 1 M HCl at 47°C for 30 min before sequential immunolabelling with mouse anti-HA and rat anti-BrdU antibodies and the relevant fluorophore-tagged secondary antibodies. One hundred cells expressing HA were examined for expression of BrdU. The experiment was repeated three times on different occasions, using cells of similar passage number.

ATDC5 cell differentiation assays

ATDC5 cells were grown to 90% confluency in 24-well plates. To induce differentiation, growth medium was supplemented with 50 µg/ml ascorbic acid (Sigma) and insulin-transferrin-selenium solution ITS-G (Gibco, 1:100 dilution). Cells were transfected three times, first on the day of differentiation and every 2 days thereafter, with HA-tagged constructs or a control pmCherry-N1 vector. Some cells were additionally treated with 10 ng/ml FGF10 (Peprotech). Differentiated mesenchymal condensations/nodules normally appeared within 14 days. After 16 days in culture, cells were fixed for 10 min with 4% PFA at RT and treated for 5 min at −20°C with pre-chilled absolute methanol. Differentiated cells were revealed by 30 min staining with 0.5% Alcian blue in 0.1 M HCl at RT. After photographing the stained cells, Alcian blue was extracted by a 6-h incubation at RT with 6 M guanidine hydrochloride and quantified using a spectrophotometer at 630 nm. Each treatment was replicated four times.

Detection of Fgf10 and FGF-receptor isoforms by standard reverse transcription polymerase chain reaction

mRNA was isolated from HEK293T cells as well as differentiated or undifferentiated ATDC5 cells using Trizol reagent, and subjected to RT-PCR (reverse transcription polymerase chain reaction) to detect Fgf10, the IIIb and IIIc isoforms of FGF-receptors 1–3 and FGFR4-IIIc, using previously described primers, protocols and cycle conditions [20,21].

Statistical analysis

Raw data were imported into and analysed by the IBM SPSS Statistic 22 software.

Normally distributed data of equal variance were compared using Student's t-test (for two samples) or the ANOVA post hoc Tukey test (for greater than two samples).

Results

Putative nuclear localization sequences within FGF10

To investigate the mode/s of G145E function, we considered a potential intracellular mechanism involving nuclear trafficking. Active nuclear import of proteins requires NLS motif/s, typically composed of a short stretch of positively charged amino acids such as arginine and lysine. NLS motifs can occur anywhere within a protein sequence, but must be exposed at the surface to allow for interaction with adaptor proteins such as importin-α and β-Kap, which facilitate protein trafficking through the nuclear pore complex [22].

Using two independent software algorithms, NLStradamus [23] and NucPred [24], we identified a putative NLS motif encoded by amino acids 194–202 (RRGQKTRRK). This is in addition to a more N-terminal NLS-like sequence (rat residues 142NKKGKLY148) noted by Kosman et al. [25] resembling, but not homologous to, a putative NLS found in FGF1 (i.e. YKKPKLL). Henceforth we term these NLS2 and NLS1, respectively (Figure 1; Supplementary Figure S1A,B). Sequence alignments revealed that both NLS1 and NLS2 are fully conserved between rat, mouse and human, and show high conservation amongst other vertebrates (Figure 1A). Moreover, 3D-modelling showed that both motifs are exposed and reside away from the FGF10–FGFR2 interaction interface (Figure 1B). Strikingly, the LADD-type G138E falls within NLS1, and the G138 residue (rat G145) is not only conserved among different mammalian species (Figure 1A) but also across different FGF family members (Supplementary Figure S1C,D). To understand the significance of this high conservation, we scrutinized structural models of FGF10 and found that G145 interacts via a single H-bond (hydrogen bond) with residue 196 (G196) of NLS2. Furthermore, side chains of the basic residues within NLS2 extend away from the protein surface (Figure 1C) allowing for potential contacts and interactions with other molecules, such as importins. Combined, these analyses suggested that NLS1 and NLS2 may function individually or co-operatively to promote nuclear import of endogenous FGF10.

Structure, conservation and positioning of two putative NLS motifs in FGF10.

Figure 1.
Structure, conservation and positioning of two putative NLS motifs in FGF10.

(A) Amino acids encoding putative NLS1 (green) and NLS2 (red) in FGF10 show a high degree of conservation from fish to mammals. (B) Three-dimensional modelling viewed from two angles to show close proximity of NLS1 (green) and NLS2 (red) motifs and their surface exposure away from FGF10's (light brown) interaction interface with FGFR2-IIIb (blue). (C) Capacity for a direct hydrogen bond (blue line) between carboxyl oxygen of G145 residue in NLS1 (purple) and amide hydrogen of G196 in NLS2, and lateral protrusion of NLS2 basic residues (R194, R195, K198, R200, R201 and K202) away from the FGF10 molecule.

Figure 1.
Structure, conservation and positioning of two putative NLS motifs in FGF10.

(A) Amino acids encoding putative NLS1 (green) and NLS2 (red) in FGF10 show a high degree of conservation from fish to mammals. (B) Three-dimensional modelling viewed from two angles to show close proximity of NLS1 (green) and NLS2 (red) motifs and their surface exposure away from FGF10's (light brown) interaction interface with FGFR2-IIIb (blue). (C) Capacity for a direct hydrogen bond (blue line) between carboxyl oxygen of G145 residue in NLS1 (purple) and amide hydrogen of G196 in NLS2, and lateral protrusion of NLS2 basic residues (R194, R195, K198, R200, R201 and K202) away from the FGF10 molecule.

Glycine 145 is critical for nuclear translocation of FGF10

To test the idea that FGF10 can translocate into the nucleus, we first analysed its cellular distribution. In the absence of commercial antibodies to specifically detect endogenously produced FGF10 by immunocytochemistry (our unpublished investigations), we followed the fate of FGF10 molecules tagged at their C-terminus with haemagglutinin A (FGF10-HA). The relevant construct was generated by PCR from rat cDNA and cloned into pN1 mammalian expression vector (Addgene; Supplementary Figure S2A). ARPE (retinal pigment epithelial) cells were transiently transfected with the corresponding vector and FGF10-HA was detected by anti-HA immunolabelling, at 24, 48 and 72 h post-transfection. Although a variety of cellular HA distributions were observed, consistently in 20–25% of transfected cells, HA was predominantly nuclear, regardless of the time point analysed (Figure 2A,B). This was not peculiar to ARPE cells, as similar extents of nuclear HA localization were found in transfected ATDC5 cells, primary adult mouse hypothalamic cells and HEK293T cells (Figure 2A,B; data not shown). Moreover, a similar pattern and proportion of nuclear FGF10-HA was observed after transfection of serum-starved growth-arrested cells, suggesting that its nuclear localization is not dependent on nuclear membrane breakdown during M phase (mitosis phase of cell cycle; Figure 2C; data not shown). On closer examination, some cells exhibited discrete nucleolar FGF10-HA localization (see also below). Generation of Fgf10 transcripts from transfected plasmids was confirmed by real-time RT-PCR (data not shown).

LADD-type G145E mutation inhibits the normal nuclear trafficking of FGF10.

Figure 2.
LADD-type G145E mutation inhibits the normal nuclear trafficking of FGF10.

(A) Transiently transfected HA-tagged FGF10 is found in both the cytoplasm and nucleus of multiple cell types, whereas an HA-tagged G145E mutant is absent from the cell nucleus. Images, 24 h post-transfection. (B) Temporal quantification of nuclear/cytoplasmic localization shows a time-independent difference between FGF10-HA and G145E in ARPE cells (error bars — SE; ANOVA ****P ≤ 0.0001; n = 3). (C) Nuclear translocation of FGF10-HA can occur in growth-arrested ARPE cells. Scale bars: 50 µm in A, 25 µm in C.

Figure 2.
LADD-type G145E mutation inhibits the normal nuclear trafficking of FGF10.

(A) Transiently transfected HA-tagged FGF10 is found in both the cytoplasm and nucleus of multiple cell types, whereas an HA-tagged G145E mutant is absent from the cell nucleus. Images, 24 h post-transfection. (B) Temporal quantification of nuclear/cytoplasmic localization shows a time-independent difference between FGF10-HA and G145E in ARPE cells (error bars — SE; ANOVA ****P ≤ 0.0001; n = 3). (C) Nuclear translocation of FGF10-HA can occur in growth-arrested ARPE cells. Scale bars: 50 µm in A, 25 µm in C.

Because LADD-causing G138E falls within a putative NLS (NLS1), next we asked whether introducing the mutant protein into cultured cells affects its nuclear localization. Thus, a corresponding C-terminus HA-tagged rat G145E construct was generated through site-directed mutagenesis (Supplementary Figure S2), transfected into ARPE cells and analysed as described for WT FGF10-HA. Generation of G145E-encoding transcripts was also confirmed by real-time RT-PCR (data not shown). Remarkably, in all cell types and at all time points examined, G145E-HA was excluded from the nucleus and restricted to the cytoplasm (Figure 2A,B).

A glutamic acid substitution at residue 145 would create an acidic side chain, potentially disrupting the interaction of FGF10 with importins. On the other hand, a glycine residue per se could be critical for the nuclear translocation. To distinguish between these possibilities, we also analysed the cellular distribution of HA-tagged G145V and G145A FGF10 constructs, choosing valine (V) or alanine (A) residue substitutions to mimic the small size of the glycine residue. However, both G145V-HA and G145A-HA molecules behaved like G145E-HA and showed nuclear exclusion (Supplementary Figure S3A–D′; data not shown).

These findings show that nuclear translocation of FGF10 can occur in multiple cell types — at least following its cytoplasmic introduction in our experimental settings. Furthermore, a G138E LADD-type amino acid substitution can disrupt this trafficking and a glycine residue at position 138 (rat 145) is critical for this process.

Distinct NLS2 residues are also important for FGF10 nuclear trafficking

To determine the functionality and key elements of the NLS2 that may be important for nuclear translocation of FGF10, six basic residues (lysine and arginine) at positions 194, 195, 198, 200, 201 and 202 were individually mutated to a neutral threonine (T) using site-directed mutagenesis (Figure 3A; Supplementary Figure S2B). In anticipation that mutagenesis of single NLS2 residues may not suffice — as observed in other NLS-bearing FGFs [26,27], we also generated three double mutants (R194T/R195T, R200T/R201T and R200T/K202T) and a single quadruple-mutant construct (R194T/R195T/R200T/K202T, henceforth termed 4T-NLS2), all tagged with HA at their C-termini (Figure 3A). These were transfected into ARPE cells and subcellular distribution of HA was monitored by immunolabelling alongside sister cultures transfected with FGF10-HA or G145E-HA mutant.

Combinatorial mutation of distinct residues in NLS2 mimics the nuclear exclusion effect of G145E in NLS1.

Figure 3.
Combinatorial mutation of distinct residues in NLS2 mimics the nuclear exclusion effect of G145E in NLS1.

(A) Single and combinatorial T residue substitutions in NLS2, created by site-directed mutagenesis. (B) Quantitative comparison of nuclear/cytoplasmic localization of NLS2 mutants with WT FGF10-HA and G145E mutant in ARPE cells (error bars — SE; ANOVA ****P ≤ 0.0001; n = 4). (C and D) Representative images of ARPE cells transfected with HA-tagged single or multiple NLS2 mutant constructs, demonstrating nuclear exclusion in a subset of these mutants, particularly 4T-NLS2 (quantified in B). Scale bars: 50 µm in C and D.

Figure 3.
Combinatorial mutation of distinct residues in NLS2 mimics the nuclear exclusion effect of G145E in NLS1.

(A) Single and combinatorial T residue substitutions in NLS2, created by site-directed mutagenesis. (B) Quantitative comparison of nuclear/cytoplasmic localization of NLS2 mutants with WT FGF10-HA and G145E mutant in ARPE cells (error bars — SE; ANOVA ****P ≤ 0.0001; n = 4). (C and D) Representative images of ARPE cells transfected with HA-tagged single or multiple NLS2 mutant constructs, demonstrating nuclear exclusion in a subset of these mutants, particularly 4T-NLS2 (quantified in B). Scale bars: 50 µm in C and D.

We found that, in contrast with G145E (NLS1), single residue substitutions in NLS2 had no significant impact on nuclear translocation of FGF10. However, double mutations significantly altered the balance of subcellular localization in favour of the cytoplasm, and the quadruple mutation (4T-NLS2) mimicked the effect of G145E, excluding FGF10 from the cell nucleus altogether (Figure 3B–D). This was verified by immunolabelling of transfected cells, or by immunoprobing their nuclear and cytoplasmic protein subfractions with antibodies against Fibrillarin, a ribonucleolar protein, and GAPDH (glyceraldehyde 3-phosphate dehydrogenase), a cytoplasmically restricted protein, in combination with HA (Figure 4). First, FGF10-HA clearly co-localized with a subset of dense fibrillar components (DFCs) found in the nucleolus (Figure 4A). Secondly, when compared with WT FGF10, G145E and 4T-NLS appeared more abundant in the cytoplasmic fraction (Figure 4B). Interestingly, WT FGF10 accumulates largely as its mature 21 kDa form within the nucleus, whilst the cytoplasmic fraction additionally contains its 25 kDa immature species (see below).

Association of FGF10-HA with cell nucleolus and molecular compartmentation of WT and mutant FGF10 proteins.

Figure 4.
Association of FGF10-HA with cell nucleolus and molecular compartmentation of WT and mutant FGF10 proteins.

(A) Co-localization of HA and a nucleolus-restricted protein, Fibrillarin, demonstrates clear association of FGF10-HA with a subset of nucleolar aggregates (arrows). (B) Use of anti-Fibrillarin and GAPDH to compare molecular abundance of FGF10-HA, G145E and 4T-NLS2 in cytoplasmic and nuclear fractions of transfected cells. A control m-Cherry construct, probed by anti-Ds-Red antibodies, is found in both fractions and no HA is found in non-transfected cells. Note that mutant proteins are relatively more abundant in the cytoplasmic fractions. Scale bar in A: 25 µm.

Figure 4.
Association of FGF10-HA with cell nucleolus and molecular compartmentation of WT and mutant FGF10 proteins.

(A) Co-localization of HA and a nucleolus-restricted protein, Fibrillarin, demonstrates clear association of FGF10-HA with a subset of nucleolar aggregates (arrows). (B) Use of anti-Fibrillarin and GAPDH to compare molecular abundance of FGF10-HA, G145E and 4T-NLS2 in cytoplasmic and nuclear fractions of transfected cells. A control m-Cherry construct, probed by anti-Ds-Red antibodies, is found in both fractions and no HA is found in non-transfected cells. Note that mutant proteins are relatively more abundant in the cytoplasmic fractions. Scale bar in A: 25 µm.

Since mutation of NLS2 mimics the effects of G145E mutation in NLS1, these findings suggest that a distinct set of NL2 residues is also important for FGF10's nuclear trafficking, possibly acting independently or through a conformational association with NLS1.

Nuclear-excluded FGF10 mutants undergo hyperglycosylation and disrupt FGF10's secretory pathway

To assess the molecular consequence of mutations that induce nuclear exclusion, whole cell lysates from ARPE cells transfected with WT or mutant HA-tagged constructs were isolated, resolved by SDS–PAGE and probed with anti-HA antibodies. The efficacy of anti-HA antibodies in these assays was verified using a control vector encoding HA-tagged SMAD2, detectable as a 55 kDa product (Figure 5A).

NLS mutations perturb the molecular processing of FGF10.

Figure 5.
NLS mutations perturb the molecular processing of FGF10.

(A) Characteristics of HA-tagged WT and mutant FGF10 proteins, immunoprobed with anti-HA antibodies. A 25 kDa full-length immature FGF10 protein is observed in all lanes, but the nuclear-excluded mutant forms (lanes 2, 3, 7 and 8) additionally show a 30 kDa product. The 21 kDa signal peptide lacking mature form of FGF10 fails to be produced by NLS1 mutants (lanes 2 and 3). Lane 9, positive control for HA; lane 10, non-transfected control. (B) Location of at least 26 potential glycosylation sites on rat FGF10 protein sequence: ‘O’-linked in italics; ‘N’-linked in bold. (C) Treatment of WT and mutant proteins with de-glycosylating reagents abrogates the 30 and 25 kDa products in favour of a 21 kDa product.

Figure 5.
NLS mutations perturb the molecular processing of FGF10.

(A) Characteristics of HA-tagged WT and mutant FGF10 proteins, immunoprobed with anti-HA antibodies. A 25 kDa full-length immature FGF10 protein is observed in all lanes, but the nuclear-excluded mutant forms (lanes 2, 3, 7 and 8) additionally show a 30 kDa product. The 21 kDa signal peptide lacking mature form of FGF10 fails to be produced by NLS1 mutants (lanes 2 and 3). Lane 9, positive control for HA; lane 10, non-transfected control. (B) Location of at least 26 potential glycosylation sites on rat FGF10 protein sequence: ‘O’-linked in italics; ‘N’-linked in bold. (C) Treatment of WT and mutant proteins with de-glycosylating reagents abrogates the 30 and 25 kDa products in favour of a 21 kDa product.

WT FGF10 is normally detected as a full-length 25 kDa product as well as a mature 21 kDa protein lacking the signal peptide, required for progression through the secretory pathway [28,29]. Here, the 21 kDa band was detected in all samples except for G145E and G145V (Figure 5A). In contrast, a novel strong 30 kDa species was present in the G145E and G145V samples as well as R200T/K202T and 4T-NLS2, but not in R194/R195T, R194T, K198T single mutants or WT. We also noted that the 25 kDa product was generally stronger whenever the 30 kDa protein was present (Figure 5A). Based on its size and the reduced conditions of SDS–PAGE, we posit that the 30 kDa product is an unlikely product of intracellular ligand–ligand dimerization (>50 kDa) or ligand–FGFR partnership (>110 kDa). Since secreted FGFs can be glycosylated at multiple positions and FGF10 is predicted to carry at least 26 potential sites (Figure 5B), we investigated the possibility that the 30 kDa species represents a hyperglycosylated form of FGF10. Cell lysates containing FGF10-HA, G145E-HA and a selection of single and double mutants as well as the 4T-NLS2 were subjected to ‘N’ deglycosylation with PNGase F (peptide-N-glycosidase F), or to deglycosylation with a cocktail of ‘O and N’ deglycosylating enzymes. As shown in Figure 5C, N-deglycosylation alone was sufficient to reduce not only the 30 kDa, but also the 25 kDa, products in all samples to a single 21 kDa product. This suggests that ‘O’ glycosylation of FGF10 in cultured cells is minimal.

A likely interpretation of these findings is that whilst abolishing nuclear translocation, the LADD-type G145E traps the mutant protein in the cytoplasm, permitting its hyperglycosylation with a possible deleterious effect on its secretion. Unfortunately, low transfection levels did not permit isolation of sufficient secreted protein from the culture media in order to compare the secretory potential of WT versus mutant FGF10, although poor FGF10 secretion from transfected cells has been noted by others [29]. Therefore, as an alternative approach we compared the progression of WT and G145E along the secretory pathway in the cytoplasm.

G145E mutant protein accumulates within the endoplasmic reticulum

To examine progression through the secretory pathway, we compared the distribution of FGF10-HA, G145E-HA and 4T-NLS2-HA in three subcellular compartments that define the successive stages of secretion. Thus, transfected cells were co-immunolabelled with anti-HA as well as antibodies against ERp60, an early secretory pathway chaperone within endoplasmic reticulum (ER); βCOP, a marker of transport between ER and Golgi; and TGN46, demarcating Golgi and vehicle transport stages (schematic, Figure 6A). Close scrutiny of high-resolution images from randomly selected cells showed that whilst FGF10-HA co-localized with all three markers (Figure 6B,C,F,I), expression of G145E-HA was restricted to the ERp60-positive compartment (Figure 6B,D,G,J), and 4T-NLS2 showed a much weaker phenotype (Figure. 6B,E,H,K). In agreement with whole cell lysate analysis (Figure 5A), these results suggest that the G145E mutant protein fails to progress through the secretory pathway. However, the causal relationship between nuclear exclusion, hyperglycosylation and secretion cannot be wholly ascertained from our experimental approach.

Abnormal cytoplasmic retention and trans-Golgi trafficking of G145E and 4T-NLS molecules.

Figure 6.
Abnormal cytoplasmic retention and trans-Golgi trafficking of G145E and 4T-NLS molecules.

(A) A schematic of the trans-Golgi secretion pathway and protein markers delineating its intermediate steps: ERp60 (early ER stage); βCOP and TGN46 (late ER stages). (B) Quantification of immuno-co-localization between these markers and HA-tagged FGF10, G145E and 4T-NLS2 molecules (shown in CK), using Pearson's correlation coefficient (see Material and methods). FGF10-HA and 4T-NLS2 show values above 0.5, indicative of co-localization with all three markers, whereas G145E co-localizes only with the ERp60. (CK) ARPE cells transfected with FGF10-HA (C, F, I), G145E (D, G, J) and 4T-NLS2 (E, H, K), and probed with Trans-Golgi secretion pathway markers, showing that whilst the WT protein is processed through the pathway, the G145E is sequestered at the ER stage. Scale bars: 20 µm in all panels.

Figure 6.
Abnormal cytoplasmic retention and trans-Golgi trafficking of G145E and 4T-NLS molecules.

(A) A schematic of the trans-Golgi secretion pathway and protein markers delineating its intermediate steps: ERp60 (early ER stage); βCOP and TGN46 (late ER stages). (B) Quantification of immuno-co-localization between these markers and HA-tagged FGF10, G145E and 4T-NLS2 molecules (shown in CK), using Pearson's correlation coefficient (see Material and methods). FGF10-HA and 4T-NLS2 show values above 0.5, indicative of co-localization with all three markers, whereas G145E co-localizes only with the ERp60. (CK) ARPE cells transfected with FGF10-HA (C, F, I), G145E (D, G, J) and 4T-NLS2 (E, H, K), and probed with Trans-Golgi secretion pathway markers, showing that whilst the WT protein is processed through the pathway, the G145E is sequestered at the ER stage. Scale bars: 20 µm in all panels.

Proliferation and differentiation of ATDC5 cells is differentially affected by mis-expression or exogenous treatment with FGF10

An important function of FGF10 is to regulate the formation and patterning of cartilage in vivo [30,31]. To test whether the perturbation of cellular trafficking reported above has an impact on cell behaviour in vitro, we measured the rate of cell proliferation and differentiation in ATDC5 cells — a chondrogenic mesenchymal cell line which expresses transcripts for FGF10 and FGFR2-IIIc, and to a lesser degree, FGFR1-IIIc, FGFR2-IIIb and FGFR3-IIIc, but not FGFR1-IIIb (RT-PCR data; not shown). Sister cultures of ATDC5 cells transfected with constructs encoding HA-tagged WT FGF10, G145E- or 4T-NLS2-bearing mutations, or a control mCherry vector, were BrdU-pulsed for 4 h and then immediately analysed for co-expression of BrdU with mCherry or HA. Compared with mCherry-transfected cells, all FGF10-bearing constructs showed a significant reduction in cell proliferation, but no significant differences were noted between WT and mutant constructs (Figure 7A–C). Similarly, the level of chemically induced differentiation of ATDC5 cells into Alcian blue-positive chondrocytes was significantly lower in cells transfected with FGF10-carrying constructs, when compared with non-transfected or mCherry-transfected cultures, with no marked difference between different FGF10-carrying constructs (Figure 7D–J). Separately, we also measured the rate of differentiation of non-transfected ATDC5 cells in response to exogenously applied FGF10 (R&D systems; concentration ranges 1, 10 or 100 ng/ml). No discernable effect on cell differentiation was noted at 1 or 100 ng/ml, but a higher level of cell death was evident at 100 ng/ml. However, and in contrast with FGF10 transfections, 10 ng/ml of exogenous FGF10 promoted ATDC5 cell differentiation, and supplementation of FGF10-HA transfected cultures with exogenous FGF10 partially rescued the inhibitory effects of FGF10-HA (Figure 7I,I′ and J; data not shown).

Effects of FGF10 overexpression or exogenous treatment on ATDC5 cell proliferation and differentiation.

Figure 7.
Effects of FGF10 overexpression or exogenous treatment on ATDC5 cell proliferation and differentiation.

(AB′) Examples of ATDC5 cells transiently transfected with FGF10-HA or a control mCherry construct, and pulsed with BrdU. (C) Introduction of FGF10-HA or its mutant forms significantly reduces cell proliferation when compared with a control mCherry vector. (DI′) Representative low- and high-power images of ATDC5 cells induced to differentiate in 24-well plates without any transfection (D, D′), or after transfection with a mCherry control (E, E′); FGF10-HA and its mutant forms alone (FH′); or FGF10-HA, in the presence of exogenous FGF10 (I, I). Transfection with FGF10-HA or its mutant forms delays the differentiation of ATDC5 cells, evidenced by formation of fewer Alcian blue-positive nodules and lower spectrophotometric detection of Alcian blue (J). (I, I′, J) Exogenously applied FGF10 partially rescues the inhibitory effects of FGF10-HA transfection. Scale bars: 20 µm in A; 1000 µm in D, E, F, G, H and I; 100 µm in D′, E′, F′, G′, H′ and I′ (Error bars in C, J are SE; ANOVA *P ≤ 0.05; **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001. C, n = 5; J, n = 4).

Figure 7.
Effects of FGF10 overexpression or exogenous treatment on ATDC5 cell proliferation and differentiation.

(AB′) Examples of ATDC5 cells transiently transfected with FGF10-HA or a control mCherry construct, and pulsed with BrdU. (C) Introduction of FGF10-HA or its mutant forms significantly reduces cell proliferation when compared with a control mCherry vector. (DI′) Representative low- and high-power images of ATDC5 cells induced to differentiate in 24-well plates without any transfection (D, D′), or after transfection with a mCherry control (E, E′); FGF10-HA and its mutant forms alone (FH′); or FGF10-HA, in the presence of exogenous FGF10 (I, I). Transfection with FGF10-HA or its mutant forms delays the differentiation of ATDC5 cells, evidenced by formation of fewer Alcian blue-positive nodules and lower spectrophotometric detection of Alcian blue (J). (I, I′, J) Exogenously applied FGF10 partially rescues the inhibitory effects of FGF10-HA transfection. Scale bars: 20 µm in A; 1000 µm in D, E, F, G, H and I; 100 µm in D′, E′, F′, G′, H′ and I′ (Error bars in C, J are SE; ANOVA *P ≤ 0.05; **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001. C, n = 5; J, n = 4).

In sum, although abrogation of nuclear trafficking does not yield a unique effect, overexpression of FGF10, whether WT or mutant, does perturb the proliferation and differentiation of mesenchymal cells in the paradigms tested. This suggests that, in ATDC5 cells, FGF10 may have multiple cell-intrinsic modes of action — nuclear and/or non-nuclear. Nonetheless, the contrasting effects brought about by FGF10 transfection vs exogenous FGF10 treatment, indicates that the cell-autonomous function/s of transfected constructs are unlikely to involve cell surface-exposed FGFRs.

Discussion

ALSG and LADD syndromes are commonly caused by heterozygous mutations in FGF10 or its cognate receptor FGFR2-IIIb, affecting tissues that develop and function through epithelial–mesenchymal cross-talk, utilizing this signalling pathway. Most FGF10 mutations are postulated to yield proteins that are truncated, unstable or defective in receptor interactions, explaining their partial loss-of-function effects (Table 1). However, the functioning of a FGF10 G138E residue substitution has remained enigmatic. In the present study, we analysed this mutation in silico and modelled the equivalent rat protein (G145E) in vitro. Although we cannot exclude the possibility that G138E encodes an unstable protein targeted for rapid degradation in vivo, our results rather suggest that, falling within a putative NLS sequence, G145E works by attenuating the process of FGF10 secretion, coupled to a novel nuclear trafficking role. The latter is supported by the identification of a second putative NLS sequence within FGF10, the mutation of which also abrogates FGF10 nuclear localization. We propose, therefore, that the G138E pathology may involve two cellular compartments and multiple biological processes, i.e. attenuation of paracrine signalling in epithelial cells secondary to reduced FGF10 secretion by mesenchymal cells, as well as perturbed cell-autonomous FGF10 function/s in mesenchymal cells themselves (summarized schematically in Supplementary Figure S7).

Nuclear translocation of FGF10 and its putative functions

As a 21–25 kDa protein, FGF10 is in principle small enough to passively traffic in and out of the nuclear pore complex and attain an equal nuclear/cytoplasmic cell distribution at any given time. However, we found that most transiently transfected cells exhibit predominantly cytoplasmic or nuclear FGF10 as late as 72 h post-transfection, suggesting that FGF10 is actively compartmentalized and/or retained within the cell nucleus. This phenomenon is not a peculiarity of our transfection system since Kosman et al. [25] showed that exogenously applied FGF10 can enter the cell nucleus, possibly in association with FGFRs. They also reported that mutagenesis of basic residues flanking G145 in NLS1 can diminish, but not abrogate, this route of nuclear localization. In contrast, we showed that the mutation of the evolutionary conserved glycine 145 alone abrogates both FGF10's nuclear localization and progression along the secretory pathway. It may be that the basic residues of NLS1 have a dual function of binding FGFRs and promoting nuclear entry. Interestingly, an arginine 187 to valine substitution in mouse NLS2 (Rat R194V) alters FGF10's extracellular receptor binding dynamics and converts it into an FGF7-like molecule in branching morphogenesis assays [32]. In our assays, mutation of R194 alone (to T) did not affect the normal rate of FGF10 nuclear trafficking, perturbing this function only in combination with other mutations in NLS2.

Dynamics and mechanisms of nuclear translocation notwithstanding, the significance and role of nuclear-targeted FGF10 molecules remains unknown. We showed that FGF10 becomes localized to a subset of DFCs of the nucleolus, demarcated by the expression of Fibrillarin. The nucleolus is important for the generation and assembly of ribosomal RNA as well as non-coding RNAs involved in pre-mRNA splicing and protein synthesis, and formation of RNA–telomerase complexes [33]. Hence, it is tempting to speculate that by associating with the molecular machinery of the nucleolus, FGF10 participates in one or more of the above processes to regulate gene expression or cell cycle control, akin to roles ascribed to nucleolar FGF2 and FGF3. Nucleolar FGF2 directly regulates synthesis of ribosomal RNA and stimulates Polymerase I transcription through binding to the UBF (upstream binding factor) transcription factor [34]. Nucleolar FGF3 inhibits cell proliferation by binding to the NoBP (nucleolar-binding protein) [35].

Targeting of FGF10 to the nucleus could also play a critical stoichiometric role in regulating the normal level of paracrine FGF10 signalling. Numerous studies show that the level of FGFR signalling is exquisitely regulated and is critical for multiple biological outcomes during normal development as well as pathological conditions [10]. A key factor in this process is the quantity and bioavailability of FGF ligands. For example, FGF10 is a driver of hyperactive FGFR2 mutations that cause Apert syndrome, and its genetic knockdown in a mouse model of Apert syndrome rescues much of the related defects [36]. Hence, the amount of secreted FGF10, available for paracrine signalling in epithelial cells, may normally be determined by titration against its nuclear-targeted forms in mesenchymal cells.

A potential cell-autonomous role for FGF10 within FGF10-expressing cells themselves, whether nuclear or otherwise, gains credence from the discovery of tissues wherein FGF10 is unconventionally expressed in the epithelial cell compartment, or from which FGF10's cognate receptor, FGFR2-IIIb, is absent. Notably, FGF10 is atypically expressed in the developing Otic epithelium [37] where it regulates cell specification, in contrast with its paracrine regulation of cell proliferation or differentiation. Interestingly, a significant number of LADD syndrome patients suffer from hearing defects [1]. In the adult brain (hypothalamus), expression of FGF10 by a population of neural stem cells termed tanycytes, is not complemented by FGFR2-IIIb or FGFR1-IIIb expression, where the IIIc isoforms of FGFRs 1–3 predominate. Tanycytes produce new appetite/energy balance regulating neurons in vivo [38] and we showed that FGF10 can translocate into the nucleus of these cells in vitro (Figure 1A). Thus, it would be interesting to investigate whether partial loss of FGF10 function in LADD/ALSG patients perturbs the homoeostatic functions of the hypothalamus, such as energy expenditure.

In summary, our in vitro investigations of a LADD syndrome FGF10 point mutation has led us to postulate novel mechanisms involving defects in secretion linked to nuclear translocation of FGF10. This may not be a common mode of function in all LADD/ALSG mutations, as evidenced by the diversity and varying degrees of phenotypic penetrance in these patients. Nonetheless, our results highlight another potential level of complexity in FGF10 biology during normal development and in disease, warranting further investigations using inducible constructs to finely dissect the intermediate steps of FGF10 trafficking, or in vivo modelling of distinct LADD mutations in transgenic animal models to test the differential role of individual mutations.

Abbreviations

     
  • ALSG

    aplasia of lacrimal and salivary glands syndrome

  •  
  • ATDC5

    chondrogenic mesenchymal

  •  
  • DFCs

    dense fibrillar components

  •  
  • ER

    endoplasmic reticulum

  •  
  • ERp60

    an isoform of protein disulphide isomerase

  •  
  • FBS

    foetal bovine serum

  •  
  • FGF

    fibroblast growth factor

  •  
  • FGF10

    fibroblast growth factor 10

  •  
  • FGFR

    fibroblast growth factor receptor

  •  
  • FGFR2

    FGF-receptor 2

  •  
  • FGFR2-IIIb

    FGF-receptor 2 IIIb

  •  
  • GAPDH

    glyceraldehyde 3-phosphate dehydrogenase

  •  
  • HA

    hemagglutinin A

  •  
  • H-bond

    hydrogen bonds

  •  
  • HMW

    high molecular weight

  •  
  • LADD

    lacrimo-auriculo-dento-digital syndrome

  •  
  • LMW

    low molecular weight

  •  
  • M phase

    mitosis phase of cell cycle

  •  
  • NGS

    normal goat serum

  •  
  • NLS

    nuclear localization sequence

  •  
  • NoBP

    nucleolar-binding protein

  •  
  • PFA

    paraformaldehyde

  •  
  • PNGase F

    peptide-N-glycosidase F

  •  
  • RT-PCR

    reverse transcription polymerase chain reaction

  •  
  • SDS

    sodium dodecyl sulphate

  •  
  • SMAD

    amalgam of ‘mothers against decapentaplegic’ (MAD) (Drosophila) and ‘small body size’ (SMA) (C. elegants)

  •  
  • TGN46

    trans-Golgi network integral membrane protein 2

  •  
  • UBF

    upstream binding factor

  •  
  • WT

    wild-type

  •  
  • βCOP

    coat protein (coatomer) β

  •  
  • β-Kap

    β-karyopherin, also called importin-β.

Author Contribution

M.K.H. conceived and planned the study, and prepared the manuscript. M.M. and T.G. contributed to the planning and conducted the experimental work.

Funding

M.M. was supported by a UEA Dean Studentship, and this work was supported by a BBSRC research grant [reference BB/L003406/1] awarded to M.K.H.

Acknowledgments

We thank Professors Saverio Bellusci (Giessen, Germany), Tom Wileman, Penny Powell, Andrew Chantry, Jelena Gavrilovic and Charles Brearley (UEA, Norwich) for sharing constructs and antibodies and critical reading of the manuscript; and Dr Lee Wheldon for valuable input in early stages of this work. We apologise to colleagues whose work could not be cited due to space limitations.

Competing Interests

The authors have no commercial interests or otherwise to declare.

References

References
1
Milunsky
,
J.M.
,
Zhao
,
G.
,
Maher
,
T.A.
,
Colby
,
R.
and
Everman
,
D.B.
(
2006
)
LADD syndrome is caused by FGF10 mutations
.
Clin. Genet.
69
,
349
354
doi:
2
Rohmann
,
E.
,
Brunner
,
H.G.
,
Kayserili
,
H.
,
Uyguner
,
O.
,
Nürnberg
,
G.
,
Lew
,
E.D.
et al. 
(
2006
)
Mutations in different components of FGF signaling in LADD syndrome
.
Nat. Genet.
38
,
414
417
doi:
3
Chapman
,
D.B.
,
Shashi
,
V.
and
Kirse
,
D.J.
(
2009
)
Case report: aplasia of the lacrimal and major salivary glands (ALSG)
.
Int. J. Pediatr. Otorhinolaryngol.
73
,
899
901
doi:
4
Entesarian
,
M.
,
Dahlqvist
,
J.
,
Shashi
,
V.
,
Stanley
,
C.S.
,
Falahat
,
B.
,
Reardon
,
W.
et al. 
(
2007
)
FGF10 missense mutations in aplasia of lacrimal and salivary glands (ALSG)
.
Eur. J. Hum. Genet.
15
,
379
382
doi:
5
Min
,
H.
,
Danilenko
,
D.M.
,
Scully
,
S.A.
,
Bolon
,
B.
,
Ring
,
B.D.
,
Tarpley
,
J.E.
et al. 
(
1998
)
Fgf-10 is required for both limb and lung development and exhibits striking functional similarity to Drosophila branchless
.
Genes. Dev.
12
,
3156
3161
doi:
6
De Moerlooze
,
L.
,
Spencer-Dene
,
B.
,
Revest
,
J.
,
Hajihosseini
,
M.
,
Rosewell
,
I.
and
Dickson
,
C.
(
2000
)
An important role for the IIIb isoform of fibroblast growth factor receptor 2 (FGFR2) in mesenchymal-epithelial signalling during mouse organogenesis
.
Development
127
,
483
492
PMID:
[PubMed]
7
Shams
,
I.
,
Rohmann
,
E.
,
Eswarakumar
,
V.P.
,
Lew
,
E.D.
,
Yuzawa
,
S.
,
Wollnik
,
B.
et al. 
(
2007
)
Lacrimo-auriculo-dento-digital syndrome is caused by reduced activity of the fibroblast growth factor 10 (FGF10)-FGF receptor 2 signaling pathway
.
Mol. Cell. Biol.
27
,
6903
6912
doi:
8
Ornitz
,
D.M.
and
Itoh
,
N.
(
2015
)
The fibroblast growth factor signaling pathway
.
Wiley Interdiscip Rev. Dev. Biol.
4
,
215
266
doi:
9
Zhang
,
X.
,
Ibrahimi
,
O.A.
,
Olsen
,
S.K.
,
Umemori
,
H.
,
Mohammadi
,
M.
and
Ornitz
,
D.M.
(
2006
)
Receptor specificity of the fibroblast growth factor family: the complete mammalian FGF family
.
J. Biol. Chem.
281
,
15694
15700
doi:
10
Hajihosseini
,
M.K.
(
2008
)
Fibroblast growth factor signaling in cranial suture development and pathogenesis
.
Front. Oral Biol.
12
,
160
177
doi:
11
Turner
,
N.
and
Grose
,
R.
(
2010
)
Fibroblast growth factor signalling: from development to cancer
.
Nat. Rev. Cancer
10
,
116
129
doi:
12
Wu
,
Q.-F.
,
Yang
,
L.
,
Li
,
S.
,
Wang
,
Q.
,
Yuan
,
X.-B.
,
Gao
,
X.
et al. 
(
2012
)
Fibroblast growth factor 13 is a microtubule-stabilizing protein regulating neuronal polarization and migration
.
Cell
149
,
1549
1564
doi:
13
Coleman
,
S.J.
,
Chioni
,
A.-M.
,
Ghallab
,
M.
,
Anderson
,
R.K.
,
Lemoine
,
N.R.
,
Kocher
,
H.M.
et al. 
(
2014
)
Nuclear translocation of FGFR1 and FGF2 in pancreatic stellate cells facilitates pancreatic cancer cell invasion
.
EMBO Mol. Med.
6
,
467
481
doi:
14
Haugsten
,
E.M.
,
Sorensen
,
V.
,
Brech
,
A.
,
Olsnes
,
S.
and
Wesche
,
J.
(
2005
)
Different intracellular trafficking of FGF1 endocytosed by the four homologous FGF receptors
.
J. Cell. Sci.
118
,
3869
3881
doi:
15
Taverna
,
S.
,
Rigogliuso
,
S.
,
Salamone
,
M.
and
Vittorelli
,
M.L.
(
2008
)
Intracellular trafficking of endogenous fibroblast growth factor-2
.
FEBS J.
275
,
1579
1592
doi:
16
Bossard
,
C.
,
Laurell
,
H.
,
Van den Berghe
,
L.
,
Meunier
,
S.
,
Zanibellato
,
C.
and
Prats
,
H.
(
2003
)
Translokin is an intracellular mediator of FGF-2 trafficking
.
Nat. Cell. Biol.
5
,
433
439
doi:
17
Homer-Bouthiette
,
C.
,
Doetschman
,
T.
,
Xiao
,
L.
and
Hurley
,
M.M.
(
2014
)
Knockout of nuclear high molecular weight FGF2 isoforms in mice modulates bone and phosphate homeostasis
.
J. Biol. Chem.
289
,
36303
36314
doi:
18
Kiefer
,
P.
and
Dickson
,
C.
(
1995
)
Nucleolar association of fibroblast growth factor 3 via specific sequence motifs has inhibitory effects on cell growth
.
Mol. Cell. Biol.
15
,
4364
4374
doi:
19
Dimauro
,
I.
,
Pearson
,
T.
,
Caporossi
,
D.
and
Jackson
,
M.J.
(
2012
)
A simple protocol for the subcellular fractionation of skeletal muscle cells and tissue
.
BMC Res. Notes
5
,
513
doi:
20
Hajihosseini
,
M.K.
and
Heath
,
J.K.
(
2002
)
Expression patterns of fibroblast growth factors-18 and -20 in mouse embryos is suggestive of novel roles in calvarial and limb development
.
Mech. Dev.
113
,
79
83
doi:
21
Hajihosseini
,
M.K.
,
De Langhe
,
S.
,
Lana-Elola
,
E.
,
Morrison
,
H.
,
Sparshott
,
N.
,
Kelly
,
R.
et al. 
(
2008
)
Localization and fate of Fgf10-expressing cells in the adult mouse brain implicate Fgf10 in control of neurogenesis
.
Mol. Cell. Neurosci.
37
,
857
868
doi:
22
Christie
,
M.
,
Chang
,
C.-W.
,
Róna
,
G.
,
Smith
,
K.M.
,
Stewart
,
A.G.
,
Takeda
,
A.A.
et al. 
(
2016
)
Structural biology and regulation of protein import into the nucleus
.
J. Mol. Biol.
428
,
2060
2090
doi:
23
Nguyen Ba
,
N.A.
,
Pogoutse
,
A.
,
Provart
,
N.
and
Moses
,
A.M.
(
2009
)
NLStradamus: a simple hidden Markov model for nuclear localization signal prediction
.
BMC Bioinf.
10
,
202
doi:
24
Brameier
,
M.
,
Krings
,
A.
and
MacCallum
,
R.M.
(
2007
)
NucPred — predicting nuclear localization of proteins
.
Bioinformatics
23
,
1159
1160
doi:
25
Kosman
,
J.
,
Carmean
,
N.
,
Leaf
,
E.M.
,
Dyamenahalli
,
K.
and
Bassuk
,
J.A.
(
2007
)
Translocation of fibroblast growth factor-10 and its receptor into nuclei of human urothelial cells
.
J. Cell. Biochem.
102
,
769
785
doi:
26
Sheng
,
Z.
,
Lewis
,
J.A.
and
Chirico
,
W.J.
(
2004
)
Nuclear and nucleolar localization of 18-kDa fibroblast growth factor-2 is controlled by C-terminal signals
.
J. Biol. Chem.
279
,
40153
40160
doi:
27
Wesche
,
J.
,
Małecki
,
J.
,
Wie¸dlocha
,
A.
,
Ehsani
,
M.
,
Marcinkowska
,
E.
,
Nilsen
,
T.
et al. 
(
2005
)
Two nuclear localization signals required for transport from the cytosol to the nucleus of externally added FGF-1 translocated into cells
.
Biochemistry
44
,
6071
6080
doi:
28
Yamasaki
,
M.
,
Miyake
,
A.
,
Tagashira
,
S.
and
Itoh
,
N.
(
1996
)
Structure and expression of the rat mRNA encoding a novel member of the fibroblast growth factor family
.
J. Biol. Chem.
271
,
15918
15921
doi:
29
Beer
,
H.-D.
,
Florence
,
C.
,
Dammeier
,
J.
,
McGuire
,
L.
,
Werner
,
S.
and
Duan
,
D.R.
(
1997
)
Mouse fibroblast growth factor 10: cDNA cloning, protein characterization, and regulation of mRNA expression
.
Oncogene
15
,
2211
2218
doi:
30
Terao
,
F.
,
Takahashi
,
I.
,
Mitani
,
H.
,
Haruyama
,
N.
,
Sasano
,
Y.
,
Suzuki
,
O.
et al. 
(
2011
)
Fibroblast growth factor 10 regulates Meckel's cartilage formation during early mandibular morphogenesis in rats
.
Dev. Biol.
350
,
337
347
doi:
31
Sala
,
F.G.
,
Del Moral
,
P.-M.
,
Tiozzo
,
C.
,
Alam
,
D.A.
,
Warburton
,
D.
,
Grikscheit
,
T.
et al. 
(
2011
)
FGF10 controls the patterning of the tracheal cartilage rings via Shh
.
Development
138
,
273
282
doi:
32
Makarenkova
,
H.P.
,
Hoffman
,
M.P.
,
Beenken
,
A.
,
Eliseenkova
,
A.V.
,
Meech
,
R.
,
Tsau
,
C.
et al. 
(
2009
)
Differential interactions of FGFs with heparan sulfate control gradient formation and branching morphogenesis
.
Sci. Signal.
2
,
ra55
doi:
33
Pederson
,
T.
(
2011
)
The nucleolus
.
Cold Spring Harb. Perspect. Biol.
3
,
a000638
doi:
34
Sheng
,
Z.
,
Liang
,
Y.
,
Lin
,
C.-Y.
,
Comai
,
L.
and
Chirico
,
W.J.
(
2005
)
Direct regulation of rRNA transcription by fibroblast growth factor 2
.
Mol. Cell. Biol.
25
,
9419
9426
doi:
35
Reimers
,
K.
,
Antoine
,
M.
,
Zapatka
,
M.
,
Blecken
,
V.
,
Dickson
,
C.
and
Kiefer
,
P.
(
2001
)
NoBP, a nuclear fibroblast growth factor 3 binding protein, is cell cycle regulated and promotes cell growth
.
Mol. Cell. Biol.
21
,
4996
5007
doi:
36
Hajihosseini
,
M.K.
,
Duarte
,
R.
,
Pegrum
,
J.
,
Donjacour
,
A.
,
Lana-Elola
,
E.
,
Rice
,
D.P.
et al. 
(
2009
)
Evidence that Fgf10 contributes to the skeletal and visceral defects of an Apert syndrome mouse model
.
Dev. Dyn.
238
,
376
385
doi:
37
Urness
,
L.D.
,
Wang
,
X.
,
Shibata
,
S.
,
Ohyama
,
T.
and
Mansour
,
S.L.
(
2015
)
Fgf10 is required for specification of non-sensory regions of the cochlear epithelium
.
Dev. Biol.
400
,
59
71
doi:
38
Haan
,
N.
,
Goodman
,
T.
,
Najdi-Samiei
,
A.
,
Stratford
,
C.M.
,
Rice
,
R.
,
El Agha
,
E.
et al. 
(
2013
)
Fgf10-expressing tanycytes add new neurons to the appetite/energy-balance regulating centers of the postnatal and adult hypothalamus
.
J. Neurosci.
33
,
6170
6180
doi:
39
Seymen
,
F.
,
Koruyucu
,
M.
,
Toptanci
,
I.R.
,
Balsak
,
S.
,
Dedeoglu
,
S.
,
Celepkolu
,
T.
et al. 
(
2016
)
Novel FGF10 mutation in autosomal dominant aplasia of lacrimal and salivary glands
.
Clin. Oral. Investig.
doi:
40
Lew
,
E.D.
,
Bae
,
J.H.
,
Rohmann
,
E.
,
Wollnik
,
B.
and
Schlessinger
,
J.
(
2007
)
Structural basis for reduced FGFR2 activity in LADD syndrome: implications for FGFR autoinhibition and activation
.
Proc. Natl Acad. Sci. U.S.A.
104
,
19802
19807
doi:
41
Scheckenbach
,
K.
,
Balz
,
V.
,
Wagenmann
,
M.
and
Hoffmann
,
T.K.
(
2008
)
An intronic alteration of the fibroblast growth factor 10 gene causing ALSG-(aplasia of lacrimal and salivary glands) syndrome
.
BMC Med. Genet.
9
,
114
doi:
42
Entesarian
,
M.
,
Mattson
,
H.
,
Klar
,
J.
,
Bergendal
,
B.
,
Olson
,
L.
,
Arakaki
,
T.
et al. 
(
2005
)
Mutations in the gene encoding fibroblast growth factor 10 are associated with aplasia of lacrimal and salivary glands
.
Nat. Genet.
37
,
125
127
doi:

Author notes

*

Equal contribution.

Supplementary data