First published online May 4, 2004
doi: 10.1242/10.1242/jcs.01096
Journal of Cell Science 117, 2377-2388 (2004)
Published by The Company of Biologists 2004
PDE4-regulated cAMP degradation controls the assembly of integrin-dependent actin adhesion structures and REF52 cell migration
Yvonne M. Fleming1,2,*,
Margaret C. Frame1,2 and
Miles D. Houslay1
1 Institute of Biological and Life Sciences, Davidson and Wolfson Buildings, University of Glasgow, Glasgow G12 8QQ, UK
2 The Beatson Institute for Cancer Research, Garscube Estate, Switchback Road, Bearsden, Glasgow G61 1BD, UK
View larger version (41K):
[in a new window]
|
Fig. 1. Integrin-induced actin adhesion microspikes are suppressed by PDE4 inhibitors. REF52 cells were trypsinised, washed in serum free media and allowed to adhere to chamber slides that had been coated with laminin (control conditions) in the presence of (A) 0.1% DMSO carrier, (B) 50 µM forskolin, (C) 100 µM IBMX, (D) 1 µM cilostamide, (E) 10 µM rolipram or (F) 5 µM Ro-172420. Cells were fixed 1 hour after plating and stained for vinculin (green) and with TRITC-phalloidin to visualise actin filaments (red). (A) Cells exhibit a dense cortical actin network (dotted arrow) and prominent microspike structures that contain polymerised actin and adhesion proteins (solid arrows). (B,C) Cells exhibit a flattened morphology and extensive lamellipodia/membrane ruffles (arrows). Scale bars, 40 µm.
|
|
View larger version (23K):
[in a new window]
|
Fig. 2. Rolipram induces protein kinase A activity in REF52 cells. Cells were allowed to adhere to laminin-coated plates in the absence (–) or presence (+) of rolipram before lysis in the appropriate buffer. (A) PKA activity in lysates was assayed using a peptide substrate. Data shown are of three independent experiments (mean±s.e.m.). (B) The level of phosphorylation of the PKA-substrate CREB was determined by western blotting using an antibody specific for CREB phosphorylated at Ser133. An antibody that recognises phosphorylated and unphosphorylated CREB equally well was used to demonstrate equal loading. The PKA selective inhibitor H89 (1 µM) was found to inhibit the rolipram induced phosphorylation of CREB. (C) Western blotting with an antibody that recognises VASP was used to determine the phosphorylation state of this protein as it undergoes a mobility shift when phosphorylated at Ser157. The blot was re-probed with an antibody specific for actin to demonstrate equal loading.
|
|
View larger version (30K):
[in a new window]
|
Fig. 4. Rolipram leads to reduced RhoA activity in REF52 cells. (A) REF52 cells were allowed to adhere to laminin coated plates in the absence (–) or presence (+) of rolipram. Cells were lysed 1 hour after plating. The level of active GTP-RhoA was determined by incubating lysates with GST-C21 and analysing the amount of bound RhoA on a western blot. Aliquots of total lysates were also probed with anti-RhoA antibodies to control for total amount of RhoA protein. The level of active RhoA was quantified by densitometric analysis of western blots – the amount of active RhoA was normalised to the amount of total RhoA. Results shown are from cells plated in the absence (black bars) or presence (white bars) of rolipram and of three independent experiments±s.e.m. (B) Cells were left untransfected, transfected with a dominant negative RhoA construct (N19RhoA) or treated with exoenzyme C3 transferase (10 µg/ml in media for 24 hours). Cells were then plated onto laminin under control conditions, fixed 1 hour after plating and stained for actin using TRITC-phalloidin. Arrows indicate actin microspikes. N19RhoA transfectants were identified using an antibody that recognises the N-terminal Myc tag of this protein (not shown). Scale bar, 20 µm. (C) Cells were allowed to adhere to laminin coated chamber slides in the presence of carrier control DMSO (0.1%), LPA (20 µg/ml), rolipram (10 µM) or LPA plus rolipram. Cells were fixed 1 hour after plating and stained for Scar1 (green) and actin (red). Scale bar, 40 µm. The lower panel represents a quantification of cells displaying protrusive microspikes under each condition. Results represent the mean±s.e.m. of three independent experiments.
|
|
View larger version (31K):
[in a new window]
|
Fig. 5. Rolipram does not alter Cdc42 or Rac1 activity in REF52 cells. (A) REF52 cells were allowed to adhere to laminin coated plates in the absence (–) or presence (+) of rolipram. Cells were lysed 1 hour after plating and the level of active GTP-Cdc42 was determined by incubating lysates with GST-PAK and analysing the amount of bound Cdc42 by western blotting. Aliquots of total lysates were also probed with anti-Cdc42 antibodies to control for total amount of Cdc42 protein. The level of active Cdc42 was quantified by densitometric analysis of western blots – the amount of active Cdc42 was normalised to the amount of total Cdc42. Results of three independent experiments±s.e.m. are shown from cells plated in the absence (black bars) or presence (white bars) of rolipram. (B) Cells transfected with an active mutant of Cdc42 (V12Cdc42) were fixed and stained for Scar1 (blue) and actin (red). V12Cdc42 transfectants were identified with an antibody that recognises the N-terminal Myc-tag of this protein (not shown). In a parallel experiment, non-transfected cells were plated onto laminin, fixed 1 hour after plating and stained for Scar1 (blue) and actin (red). Arrowheads indicate filopodia, which differ structurally from actin microspikes (arrows). Scale bar, 20 µm. (C) Experimental set-up like in (A) except that a Rac1-specific antibody was in the western blot. (D) Cells were transfected with a dominant-negative mutant of Rac1 (N17Rac1). Cells were plated onto laminin in the presence of rolipram. Cells were fixed 1 hour after plating and stained for transfected protein (anti-Myc, green) and actin (red). Solid arrow indicates extensive lamellipodia present in untransfected cell but absent from cell expressing N17Rac1 (dotted arrow). Scale bar, 20 µm.
|
|
View larger version (44K):
[in a new window]
|
Fig. 6. RhoA effectors ROCK and MLCK mediate assembly of rolipram-sensitive adhesion microspikes. (A) Cells were plated onto laminin under control conditions (upper panel) or in the presence of the ROCK inhibitor Y27632 (10 µM) (lower panel). Cells were fixed 1 hour after plating and stained for vinculin (green) and actin (red). (B) Cells were allowed to adhere to laminin coated plates in the absence (–) or presence (+) of rolipram. Cells were lysed 1 hour after plating and the phosphorylation state of MLC was determined by western blotting using an antibody specific for MLC phosphorylated at Ser19. Levels of total MLC were determined by using an antibody that recognises unphosphorylated and phosphorylated MLC equally well to indicate equal loading. The right panel represents the data of three independent experiments (mean±s.e.m.); phospho-MLC levels were determined by densitometry and normalised to the amount of total MLC. (C) Cells were plated onto laminin under control conditions. After 1 hour cells were fixed and stained for MLC phosphorylated at Ser19 (green) and actin (red). (D) Experimental set-up like in (C) except that rolipram (10 µM) was included. (E) Experimental set-up like in (C) except that Y27632 (10 µM) was included. Solid arrows indicate peripheral phospho-MLC staining only observed in the absence of inhibitors (C). Cytoplasmic phospho-MLC staining, observed under all conditions, is indicated by dotted arrows (C-E). Scale bars, 20 µm (A) and 40 µm (C-E).
|
|
View larger version (17K):
[in a new window]
|
Fig. 7. Loss of actin adhesion microspikes in reponse to rolipram or Y27632 correlates with decreased peripheral phospho-MLC and increased PKA activity. (A) REF52 cells were plated onto laminin in the presence of increasing concentrations of rolipram. Cells were fixed 1 hour after plating and stained for either Scar1 or MLC phosphorylated at Ser19. The number of cells that had formed actin adhesion microspikes under each condition – as judged by Scar1 staining – was determined and expressed as the percentage of total cells plated. Similarly, cells were scored for peripheral staining of phosphorylated MLC under each condition. In parallel experiments, cells were lysed and levels of PKA activity determined using a peptide substrate. The results shown represent the mean±s.e.m. of three independent experiments. (B) Experimental set-up like (A) except that cells were plated in the presence of increasing concentrations of Y27632.
|
|
View larger version (43K):
[in a new window]
|
Fig. 8. Rolipram inhibits migration that is associated with dynamic regulation of microspikes. Cells were cultured on laminin coated glass bottomed dishes for 24 hours before being transferred to serum-free media. (A) Carrier (0.1% DMSO) or (B) rolipram (10 µM) was added and cell behaviour observed by timelapse microscopy; images were taken every 5 minutes for 6 hours (see Movies 1, 2, http://jcs.biologists.org/supplemental). (A,B) Stills from those movies, taken at 90-minute intervals. Positions of two cells in each time-series are boxed. Position of a motile peripheral microspike is shown in A (arrows). Scale bar, 40 µm. (C) The range of cell speeds observed under each condition is represented as a box and whisker plot in which the median (solid line across box), mean ( ) and outliers (*) are indicated. Statistical analysis of the two data sets (control, n=31; rolipram, n=28) indicated that the reduced rate of random cell movement in the presence of rolipram was highly significant (Student's t-test, P<0.001).
|
|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati 
Twitter What's this?
© The Company of Biologists Ltd 2004
before being plated onto laminin in the presence of rolipram. Cells were fixed 1 hour after plating and stained for PKI