Expression of PaxB during development. (A) PaxB expression during development as detected by western blot with a Dictyostelium-specific affinity purified anti-peptide antibody. Maximal expression is reached at 12-15 hours of development. The experiment shown is representative for results obtained in four independent time course experiments. (B) Expression of a paxB/lacZ expression construct showing expression of paxB during all stages of development. Note and the increased levels of expression in the tip of the slug and the upper and lower cup of culminates. (C) Expression of a C terminal PaxB-GFP knockin construct again showing expression from the vegetative stage onwards and increased expression in the slug and culminate tip.

Cellular localisation of paxB in a paxB-gfp knockin mutant. (A,B) Localisation of paxillin in a vegetative cell at 0 and 120 seconds. Note the localisation of PaxB in small spots at randomly distributed sites where the cell is in close contact with the substrate (see supplementary material Movie 1). (C,D) Cumulative PaxB distribution of the cell shown in A,B for 0-60 seconds (C) and 60-120 seconds (D). The cumulative distribution shows that many of the spots are stationary and also show more clearly the localisation of PaxB in the tips of the filopodia. Scalebar (A): 10 μm. (E) Localisation in retraction fibres at the back of cells in migrating slugs. (F) Localisation of PaxB in small focal adhesion spots localised at the contact area between the slime sheath and the substrate in the outermost cells of a slug. The slug shown in images E,F is migrating from left to right. These structures are also found in cells deeper in the slug as can be seen in the supplementary information provided (see supplementary material Movie 2). Scale bar in F, 10 μm. (G) Distribution of cell-cell contacts in an early culminant. These contacts are distributed around the cell (marked with arrows) and are most obviously detected in the small epithelial layer of cells surrounding the spore mass in the forming fruiting body as well as in the cells in the upper cup and lower cup (see supplementary material Movie 3). (H) Another culminant showing the presence of PaxB enriched cell-cell contacts (arrows) in the back of an early culminant and (I) a higher magnification image of a central part of the image shown in H. (J) Section through the forming stalk of an early culminant showing high levels of PaxB localisation at the contact area between cells and the stalk sheath. Scale bar, 40 μm. (K) Higher magnification of a maximum projection of a stack of 10 adjacent images of the cells in contact with the stalk tube showing the presence of special PaxB-rich areas between the cells contacting the stalk sheath. Scale bar, 10 μm.

PaxB-GFP foci co-localise with cell-substrate contact areas. Brightfield, interference reflection contrast and fluorescence images were recorded simultaneously on a confocal microscope. Dark areas in the reflection interference contrast image surrounded by brighter interference patterns correspond to parts of the cell in close proximity to the substratum, the small very dark structures in the adhesion area are part of the contractile vacuole network. As the cell extends a pseudopod that makes contact with the substratum (15 second timepoint) PaxB-GFP localises to contact sites in this new adhesion area. In the merged fluorescence and interference contrast images at the bottom, PaxB-GFP appears purple and the contact area in dark green. Scale bar, 10 μm. See also supplementary material, Movie 4 for dynamics of paxillin in relation to the contact area in vegetative stage cells.

Co-localisation of PaxB-GFP and ABD-mRFPmars. Confocal time series of vegetative wild-type Ax2 cell expressing PaxB-GFP and ABD-mRFPmars. Time is indicated in seconds. (Left panel) PaxB-GFP (green) localises to long lived stationary contact sites at the cell/substratum interface as indicated by the arrows that mark the same contact sites at different time points. (Centre panel) ABD-mRFPmars (red) accumulates at very short-lived contact sites as indicated by the arrows. (Right panel) Merged image showing that PaxB-GFP and ABD-mRFPmars localise to different contact sites. Scale bar, 10 μm. See also supplementary material, Movie 5 for dynamics of mRFPactin and PaxB-GFP in vegetative stage cells.

Role for PaxB in cell-substrate adhesion. Adhesion of vegetative cells was measured under conditions of moderate shear stress as described in detail in Materials and Methods. It can be seen that two independent paxB^– strains (open symbols) show an increased tendency to be dislodged from the substrate when compared with an Ax2 control strain (stippled line). This defect can be rescued completely by re-introduction of paxB-gfp under the control of the Actin15 promoter in these strains (closed symbols). The data shown are the mean values and one-sided standard deviations of four independent experiments performed on four different days.

Characterisation of paxB knockout phenotypes. Comparison of Ax2 and an Ax2 paxB knockout strain (A,C) Ax2 and paxB^– in top view after 24 hours of development on agar. In Ax2 all cells develop to fruiting bodies, while in the paxB^– strains many cells do not develop into fruiting bodies but arrest at the mound stage, or when they continue development later development is grossly abnormal. (B,D) Ax2 and paxB^– knockout strain in side view after development on water agar for 24 hours. Note the differences in scale between images in B and D. Ax2 form fruiting bodies with slender stalks and well define basal disks while the paxB^– mutant often form small fruiting bodies on a large mass of cells that is not lifted off the substrate.

Role of PaxB in cell sorting and slug movement. To assess whether paxB^– strains are defective in sorting we let 5% Ax2 cells expressing the paxB/lacZ construct synergize with paxB^– cells. As can be seen during aggregation (A) the cells are randomly intermingled but at the late aggregate stage (B) the Ax2 cells start to sort to the centre of the cell masses until they are found in the tips of forming slugs (C) and small culminates (D). This clearly shows that paxB^– cells are defective in cell sorting. (E,F) Sorting of 50% paxB^– cells in Ax2 and of another Ax3 paxB knockout mutant in Ax3. As can be seen the majority of the paxB^– cells both in Ax2 and Ax3 sort to the back of the slug indicating that they may be defective in moving in cell masses. (G) a panel showing three culminants from synergy experiments where Ax2 cell synergised with GFP expressing Ax2 cells (left culminant), PaxB-GFP expressing cells synergised with Ax2 (middle culminant) and paxB^– cells expressing a PaxB-GFP fusion protein under the control of a constitutive Actin15 promoter. As can be seen Ax2 cells do not sort in Ax2, while paxB^– cells are confined mostly to the lower cup and basal disk when allowed to develop in an Ax2 environment and that expression of paxB under the control of an string actin15 promoter does rescue development somewhat, i.e. now cells are also found in upper cup, but few cells are found in the tip or in the spore mass. (H) Migration assay comparing the migration of slugs towards a localised light source indicated by an arrow. This experiment compares the migration of Ax2 cells (left), paxB^– cells (middle) and paxB^–[A15/paxB-gfp] cells (right). Ax2 cells show a strong directed migration towards the light. The paxB^– cells are completely unable to migrate while the paxB^–[A15/paxBgfp] cells show a substantial rescue of migration ability, however pooling many experiments shows that migration is not as efficient as that of wild-type cells, implying that proper level of expression of PaxB may be important.

paxB^– cells show reduced motility in synergy experiments. (A) Side view of an Ax2 mound containing 2% EGFP labelled paxB^null cells (green) and 2% mRFPmars expressing Ax2 cells (red). Scale bar, 50 μm. (B) As the mound transforms into a slug the initially uniformly dispersed paxB^null cells are restricted to the posterior of the forming slug. (C) The corresponding cell traces of paxB^null cells (green, n=7 cells) and Ax2 cells (red, n=7) indicate reduced motility and directionality of paxB^null cells during slug formation. Scale bar, 50 μm. See also supplementary material, Movie 6 for dynamics of cell sorting.

PaxB expression is needed both in prespore and prestalk cells to rescue development. To assess the requirement of paxB for development we expressed the paxB GFP construct under the control of the actin15 promoter, the prestalk-specific ecmA and ecmB promoters, and the prespore-specific psA promoter. The outcome of these experiments was that expression of paxB under the control of the A15 promoter gave a fairly good but not complete rescue, followed by expression under the ecmA promoter, the ecmB promoter and the psA promoter. This was also reflected in the rescue of migration of slugs formed (not shown). The lower row of three images shows from left to right the expression of the PaxB-GFP constructs at the slug stage under the control of the psA, ecmA and ecmB promoters in the paxB^– strain.

PaxB phosphorylation mutants do not affect slug migration. Comparison of the migration of the paxB^– mutant with the paxB^–[A15/paxB-gfp] mutant expressing the wild type PaxB under the control of the actin 15promoter with that PaxB phosphorylation mutants paxB^–[A15/paxB(S192A)-gfp] and paxB^–[A15/paxB(S192D)-gfp]. It can be seen that there is no significant difference in the rescue by the wild type and mutant PaxB forms.

Role of phosphorylation sites on focal adhesion kinetics. We measured the rate of formation of focal adhesion sites in Dictyostelium strains transformed with the A15/pax-gfppaxB-GFP construct and compared this with the kinetics of formation of actin-rich spots and paxB^– mutant mutants in which the potential JNK phosphorylation site has been mutated to alanine (S192A) or aspartate (S192D) respectively (see Materials and Methods for details on measurements). Curves were aligned so that the half-maximal increase in fluorescence intensity occurred at t=0 seconds, the number of adhesion sites measured per strain is indicated in the legend. The results clearly show that the kinetics of the focal adhesion sites differ from that of the actin spots. Their rate of formation is slower and they persist longer than the actin spots. There was no significant difference between the rate of formation between wild type and mutant paxillin. The total lifetime of the focal adhesion sites is more difficult to determine, since most sites disassemble when they reach the end of the cells. The lifetime and possibly rate of disassembly are therefore mostly determined by the movement speed of the cells.

Effect of paxillin on celltype differentiation. To assess the effect of paxillin loss on celltype-specific gene expression we measured the kinetics of prespore cell formation by using a prespore vesicle-specific polyclonal antibody as function of time. As can be seen for a typical experiment the number prespore cells starts to increase very dramatically at the mound stage. The time of the appearance of prespore cells is the same in the wild type and in the paxB^– strain but levels do not exceed 50% in the paxB^– knockout strain after which they start to decline, suggesting that many of the cells may differentiate into prestalk cells. This impression is confirmed by the transformation of the paxB^– strains with lacZ expression constructs under the control of the prestalk-specific ecmA, ecmB and the prespore-specific PsA promoters. This shows that many of the cells, especially in the base of the structures express prestalk-specific markers.