SEARCH

SEARCH BY CITATION

Keywords:

  • SH4 domain;
  • Src kinase;
  • Yes;
  • Leishmania HASPB;
  • Membrane targeting;
  • Intracellular trafficking;
  • Endosomal recycling

Abstract

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Materials and Methods
  6. Acknowledgments
  7. References
  8. Supporting Information

Acylated SH4 domains represent N-terminal targeting signals that anchor peripheral membrane proteins such as Src kinases in the inner leaflet of plasma membranes. Here we provide evidence for a novel regulatory mechanism that may control the levels of SH4 proteins being associated with plasma membranes. Using a fusion protein of the SH4 domain of Leishmania HASPB and GFP as a model system, we demonstrate that threonine 6 is a substrate for phosphorylation. Substitution of threonine 6 by glutamate (to mimic a phosphothreonine residue) resulted in a dramatic redistribution from plasma membranes to intracellular sites with a particular accumulation in a perinuclear region. As shown by both pharmacological inhibition and RNAi-mediated down-regulation of the threonine/ serine-specific phosphatases PP1 and PP2A, recycling back to the plasma membrane required dephosphorylation of threonine 6. We provide evidence that a cycle of phosphorylation and dephosphorylation may also be involved in intracellular targeting of other SH4 proteins such as the Src kinase Yes.

SH4 domains are best known from the Src protein family. Src proteins are non-receptor tyrosine kinases that play key roles in regulating signal transduction by a diverse set of cell surface receptors. They are involved in the regulation of fundamental cellular processes including cell growth, differentiation as well as migration and survival (1). The general Src kinase architecture consists of four domains, three of which are directly linked to the signal transduction pathways they are involved in. The SH3 domain mediates interactions with other signalling molecules, the SH2 domain is required for phosphotyrosine recognition and the SH1 domain contains the kinase activity (2). By contrast, the fourth domain (SH4) is responsible for membrane attachment and targeting to plasma membranes.

The molecular mechanism by which Src kinases and other peripheral membrane proteins such as Leishmania hydrophilic acylated surface protein B (HASPB) are anchored in the inner leaflet of plasma membranes is based on post-translational lipid modifications within their N-terminal SH4 domains. This kind of membrane anchoring is similar to other peripheral membrane proteins such as members of the Ras family and the α-subunits of trimeric G-proteins. The main lipid modifications generally being involved are prenylation and fatty acylation (3). In the case of SH4 proteins, dual acylation by both myristate and palmitate residues is the structural basis for membrane anchoring (4,5). Following removal of the N-terminal methionine residue SH4 proteins become myristoylated at the N-terminus via an amide linkage. The second acylation depends on the preceding myristoylation step and involves the modification of at least one cysteine residue by palmitoylation resulting in a thioester linkage (6). Thus, most Src kinases carry two binding sites for membrane anchoring. This feature has been generalized as the two signal hypothesis that was originally developed for mammalian Ras proteins (7). As a variation of this principle, instead of palmitoylation, a second membrane binding site can be formed by a cluster of basic amino acids as is the case for the prototype member of the family, Src (3). While the enzymes involved in N-myristoylation of SH4 domains have been known for many years (4), the enzymology of protein palmitoylation has been revealed only recently with so-called DHHC (Asp-His-His-Cys) proteins acting as S-acyltransferases (6,8).

N-terminal SH4 domains are sufficient for intracellular targeting and stable insertion of GFP fusion proteins into the inner leaflet of plasma membranes (3,9–13). As opposed to the molecular mechanism of membrane attachment of Src kinases, however, our knowledge about intracellular sorting and plasma membrane targeting is limited. It is known that myristoylation of SH4 domains occurs cotranslationally and results in a transient insertion into intracellular membranes in a perinuclear region, probably Golgi and/or endosomal membranes. This is supported by findings demonstrating that palmitoylation-deficient mutant proteins are retained in a perinuclear region where they colocalize with Golgi markers (9–11,14). Interestingly , under experimental conditions that do not allow palmitoylation of the SH4 domain, Src kinases do not partition into membrane microdomains (3,6,15,16). These findings indicate that, following palmitoylation at the level of perinuclear membranes, Src kinases become inserted into membrane microdomains. This, in turn, suggests that membrane microdomains could serve as sorting and transport platforms required for proper targeting of peripheral membrane proteins containing SH4 domains as targeting signals. Besides these insights, our knowledge remains poor with regard to both the role of internal membranes such as the Golgi and endosomal compartments in intracellular trafficking and the dynamics of SH4-domain-dependent targeting to various subcellular membranes.

Here, we investigated sorting determinants of two SH4 domains derived from the N-terminal ends of Leishmania HASPB and the Src kinase Yes, respectively. We identified threonine 6 as a so far unrecognized site of phosphorylation of the HASPB SH4 domain. We further demonstrated that a threonine to glutamate exchange in position 6 used to mimic a phosphorylated threonine residue results in a redistribution from plasma membranes to intracellular sites with a strong accumulation in a perinuclear region of the cells. These data were corroborated by findings that both pharmacological inhibition by calyculin A and RNAi-mediated downregulation of the protein phosphatases PP1 and PP2A caused a significant accumulation in perinuclear membranes. This was shown for the fusion proteins HASPB-N18-GFP and Yes-N18-GFP as well as for the endogenous Src kinase Yes. Our combined data suggest that SH4 proteins cycle back and forth between plasma membranes and perinuclear compartments, a process that is controlled by a regulatory cycle of phosphorylation and dephosphorylation.

Results

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Materials and Methods
  6. Acknowledgments
  7. References
  8. Supporting Information

Intracellular trafficking mediated by SH4 targeting domains

To analyse intracellular trafficking of N-terminally acylated proteins we have used the SH4 targeting signal of the Leishmania plasma membrane protein HASPB (9,10) as a fusion protein with GFP and Protein A. This SH4 domain represents a conserved sorting signal that is sufficient to localize this fusion protein to the plasma membrane of both mammalian cells (Figure 1, panels A and E; (10,11)) and Leishmania parasites (9). Consistent with previous findings (9,11,13), a palmitoylation mutant was retained in perinuclear membranes (Figure 1, panel B) demonstrating that, besides N-terminal myristoylation, palmitoylation is an obligatory modification for transport to plasma membranes. Using both random and targeted mutagenesis of the SH4 domain we identified a single amino acid exchange that caused a major redistribution from plasma membranes to intracellular sites. When threonine 6 was replaced by a glutamate residue, the SH4 fusion protein was retained primarily in perinuclear membranes (Figure 1, panel C) similar to what was observed for the palmitoylation mutant (Figure 1, panel B). This phenotype was directly dependent on the exchange of threonine 6 by glutamate as a replacement by alanine was fully compatible with plasma membrane localization of the SH4 fusion protein (Figure 1, panel D). Similar results were obtained in HeLa cells in which expression of SH4 fusion proteins does not cause plasma membrane blebbing [Figure 1, panels E–H; (11)]. These findings establish that the observed relocalization of a SH4 fusion protein caused by mimicking a phosphothreonine residue in position 6 is neither induced nor related to the ability of SH4 fusion proteins to promote plasma membrane blebbing in Chinese hamster ovary (CHO) cells (11). Furthermore, these results demonstrate that the observed phenotype is not specific to a certain cell type.

image

Figure 1. Subcellular localization of wild-type and mutant forms of the SH4 domain of HASPB-N18-GFP fusion proteins. Stable cell lines generated by retroviral transduction were induced with doxycycline for 48 h to express the fusion proteins indicated. Live cells were viewed with a Zeiss LSM 510 confocal microscope using a 100 × oil immersion objective. (A and E) HASPB-N18-GFP (wild-type). (B and F) HASPB-N18-Δpalm-GFP. (C and G) HASPB-N18-T6E-GFP. (D and H) HASPB-N18-T6A-GFP. Bars = 10μm. In panels A–D, CHO cells were analysed whereas panels E–H show the results obtained with HeLa cells.

Download figure to PowerPoint

The T6E mutant form of the SH4 fusion protein accumulates in both endosomal compartments and Golgi membranes

To further characterize the subcellular localization of the wild-type and all mutant forms of the SH4 fusion protein by ultrastructural methods we conducted cryo-immuno electron microscopy. In Figure 2, a direct comparison at high magnification of the wild-type and the T6E mutant forms of the SH4 fusion protein is shown. To identify plasma membranes, endosomal compartments and the Golgi complex we colabelled whole-cell cryo sections with antibodies directed against the transferrin receptor as a marker for plasma membranes and endosomal structures. The wild-type SH4 fusion protein (Figure 2, panel A) was found mainly at the plasma membrane, whereas the T6E mutant form was localized primarily in endosomal compartments but also in Golgi membranes (Figure 2, panel B). In addition to the wild-type and the T6E mutant, we also analysed the subcellular localizations of both the palmitoylation and the T6A mutants in whole-cell cryosections (Figures S1 and S2) and determined the relative labelling densities for plasma membranes, Golgi membranes, endosomes and the endoplasmic reticulum (ER) (Table 1). These studies revealed that the wild-type form and the T6A mutant of the SH4 fusion protein are characterized by the highest labelling densities at plasma membranes, however, they are also present in endosomal compartments and, in small amounts, in Golgi membranes. By contrast, the T6E and Δpalm mutant forms of the SH4 fusion protein displayed the highest labelling densities in endosomal structures but were also present in lower amounts in Golgi and plasma membranes (Table 1). SH4 fusion proteins associated with Golgi membranes may represent newly synthesized material en route to the cell surface. These findings are consistent with the confocal microscopy studies shown in Figure 1.

image

Figure 2. Differential subcellular localization of the wild-type and the T6E mutant forms of the SH4 fusion protein as determined by immuno-electron microscopy. CHO cells were fixed and processed for cryo-immunogold labelling using anti-GFP and anti-transferrin receptor antibodies. Primary antibodies were detected with 10 and 15 nm gold-conjugated secondary antibodies, respectively. (A) HASPB-N18-GFP is localized at the plasma membrane and on endosomal structures positive for transferrin receptor (arrows). (B) HASPB-N18-GFP-T6E is predominantly present on endosomal structures (arrow). N, Nucleus; G, Golgi complex; ER, endoplasmic reticulum. Bars: 500 nm (A); 200 nm (B).

Download figure to PowerPoint

Table 1.  Labelling densities in various cellular compartments of the wild-type and mutant forms of the HASPB-N18-GFP fusion protein. The values given represent the number of gold particles per arbitrary unit length. In addition, SD values and the total number of gold particles that were counted for each construct are given.
 Plasma membranesGolgiEndosomesERTotal number of gold particles counted
HASPB-N18-GFP1.0 ± 0.070.2 ± 0.030.7 ± 0.090.1 ± 0.02752
HASPB-N18-Δpalm-GFP0.2 ± 0.040.2 ± 0.040.6 ± 0.050.3 ± 0.04906
HASPB-N18-T6E-GFP0.6 ± 0.060.5 ± 0.050.9 ± 0.050.5 ± 0.02946
HASPB-N18-T6A-GFP1.3 ± 0.10.3 ± 0.030.9 ± 0.120.1 ± 0.041105

Substitution of threonine 6 by glutamate does not cause a redistribution of the SH4 fusion protein HASPB-N18-GFP from membranes to the cytoplasm

The localization defects of the T6E and Δpalm mutant forms of the SH4 fusion protein were similar (Figure 1 and Table 1). Therefore, using cell fractionation and carbonate extraction experiments, we analysed whether substitution of threonine 6 by glutamate caused a redistribution between membrane and cytoplasmic pools (Figure 3).

image

Figure 3. Substitution of threonine 6 by glutamate or alanine does not cause a redistribution of SH4 proteins from membranes to the cytoplasm. CHO cells were induced with doxycycline to express HASPB-N18-wt-GFP, HASPB-N18-Δpalm-GFP, HASPB-N18-Δmyr-GFP, HASPB-N18-T6A-GFP and HASPB-N18-T6E-GFP cells, and grown for 48 h to about 80% confluency. Subcellular fractionation and carbonate extraction of membranes were performed at 4°C. Input material (1% each) and the fractions indicated (3% each) were analysed by SDS-PAGE and Western blotting employing affinity-purified anti-GFP antibodies. Bands observed below the full-length proteins were C-terminal degradation products that appeared after homogenization. The upper band for each construct migrated with an apparent molecular weight of 45 kDa.

Download figure to PowerPoint

CHO cells were induced with doxycycline to express the various SH4 fusion proteins indicated. Cells were broken and fractionated into a soluble cytoplasmic pool and a membrane fraction. As expected, acylation-deficient variant forms of the SH4 fusion protein (Δmyr [G2A] and Δpalm [C5A], respectively) were strongly decreased in the membrane fraction as compared to the wild-type SH4 fusion protein. While the palmitoylation mutant retained a small but significant degree of membrane association, the myristoylation mutant was completely redistributed into the cytoplasmic pool. As opposed to this, both the T6E and the T6A variants of the SH4 fusion protein were found in the membrane fraction at wild-type levels (Figure 3, lane 3). Additionally, when isolated membranes derived from the various cell lines were treated with carbonate to release loosely attached proteins, the T6E and the T6A variant forms were as resistant to carbonate treatment as the wild-type form of the corresponding SH4 fusion protein (Figure 3, lanes 4 and 5).

While a substantial amount of about 50% of both the wild-type and the T6E/T6A mutant forms were found in a soluble pool following homogenization, our data demonstrate that an exchange of threonine 6 to glutamate or alanine does not affect the efficiency of membrane association of the corresponding SH4 fusion proteins since they behaved in a very similar manner as compared to the wild-type form.

The exchange of threonine 6 by glutamate does not affect acylation levels of the SH4 fusion protein HASPB-N18-GFP

To further verify the studies on membrane association shown in Figure 3, we analysed whether the T6E mutation causes a lack of recognition by the S-acyl transferase required for palmitoylation of the SH4 domain (Figure 4). We conducted metabolic labelling using [3H] myristate and [3H] palmitate, respectively, and, following immunoprecipitation, quantified incorporation of these radiolabelled fatty acids using a β-imager (Figure 4, panel A, upper sub-panels). In parallel, we conducted a Western blot analysis in order to quantify the relative amounts of the various HASPB-N18-GFP fusion proteins in the immunoprecipitates using a LI-COR Odyssey imaging system (Figure 4, panel A, lower sub-panels). The relative amounts of incorporation of [3H] myristate and [3H] palmitate into the fusion proteins indicated were normalized for the relative amounts being immunoprecipitated and are shown in panel B of Figure 4. We found that, unlike the Δpalm variant form of the HASPB-N18 fusion protein, the exchange of threonine 6 to glutamate does not result in a failure of palmitoylation of the corresponding fusion protein as compared to the wild-type form.

image

Figure 4. Exchange of threonine 6 to glutamate or alanine residues does not affect acylation levels of the corresponding fusion proteins. CHO cells were induced with doxycyline to express the HASPB-N18-GFP fusion proteins indicated and were grown to about 80% confluency. Following incubation for 10 h in medium supplemented with delipidated FCS and 2 h without FCS, cells were labelled with either [3H] myristate or [3H] palmitate, respectively. Cell lysates were prepared and subjected to immunoprecipitation using affinity-purified anti-GFP antibodies. Following SDS-PAGE and blotting, samples were analysed both for Tritium-derived radioactivity using a β-imager and by immunodetection employing polyclonal rabbit-anti-GFP antibodies and goat-anti-rabbit Alexa Fluor 680 secondary antibodies (panel A). C-terminal degradation products were excluded from the analysis. The relative ratios of incorporation of Tritium-derived radioactivity and total material based on Western blotting (quantified using an Odyssey imaging system) were calculated and are given in panel B. The wild-type form of the HASPB-N18-GFP fusion protein was set to 100%.

Download figure to PowerPoint

These findings are also interesting with regard to other SH4 proteins such as the Src kinases Lck and Yes. In these cases, a serine residue is found in position 6 and, consistent with earlier reports (17), we found that a substitution of this residue by either glutamate or alanine resulted in redistribution into the cytoplasm (data not shown).

Thus, in the context of the Yes and Lck SH4 domains, serine 6 is required for the recognition by acyl transferases (17) and, therefore, mutations in this position are not compatible with efficient acylation of the N-terminus of Yes and other Src kinases.

The steady-state localization of HASPB-N18-GFP fusion proteins in plasma membranes versus endosomes correlates with their association with detergent-resistant membranes

As shown in Figure 5, the normal behaviour of the T6E variant form of the SH4 fusion protein with regard to membrane association (Figure 3) and acylation levels (Figure 4) was also consistent with an analysis of its association with detergent-resistant membranes (DRMs). These studies were based on the fact that, in many SH4 domain proteins such as Src kinases, the second acylation step is a requirement for partitioning into DRMs (3,6,15,16). While both the wild-type and the T6A mutant forms were found associated with DRMs at about 30% of the total material being analysed, less than 1% of the palmitoylation mutant was recovered from DRMs in flotation experiments (Figure 5). To a large part this is due to a substantial loss of general membrane association in the absence of palmitoylation (see Figure 3). When the association of the T6E mutant with these specialized domains was investigated we found a reduction by about 70% as compared to the wild-type form. However, the association of the T6E mutant with DRMs was highly significant as we found more than 10% of the total material in the DRM fraction. Thus, when compared to negative controls such as the transferrin receptor or the SH4 palmitoylation mutant, the relative amounts found in DRMs were significantly higher for the T6E mutant. The combined data shown in Figures 3–5 unequivocally demonstrate that the T6E mutation did not cause a failure in neither acylation nor membrane association and, therefore, its mislocalization in perinuclear compartments must be because of a different mechanism.

image

Figure 5. Plasma membrane versus intracellular localization of HASPB-N18-GFP fusion proteins correlates with the degree of association with DRMs. (A) CHO cells were induced with doxycyline to express the HASPB-N18-GFP fusion proteins indicated. Upon detergent extraction using TX-100 on ice and flotation on an Optiprep gradient, eight fractions were collected and analysed by SDS-PAGE. Immunodetection was performed with polyclonal rabbit-anti-GFP antibodies and goat-anti-rabbit Alexa Fluor 680 antibodies. Caveolin-1 was used as a marker for DRMs and the transferrin receptor was used as a non-DRM marker. (B) Quantitative analysis of the relative distribution of the fusion proteins indicated within the flotation gradient. Antibody signals in each fraction were quantified using an Odyssey imaging system (LI-COR Biosciences) and expressed as the percentage of the total material recovered from all fractions.

Download figure to PowerPoint

Phosphorylation of threonine 6 of the SH4 domain of Leishmania HASPB

In structural biology, glutamate is frequently used to mimic a phosphothreonine residue. We, therefore, reasoned that the molecular mechanism of missorting of the T6E mutant form of the SH4 fusion protein could be based on a reversible phosphorylation of the SH4 domain that regulates recycling between the plasma membrane and intracellular sites.

In order to challenge this hypothesis we first analysed whether the HASPB-N18-GFP fusion protein is phosphorylated at threonine 6. To reveal phosphorylated species we conducted a 2D gel analysis of immunoprecipitates of various forms of HASPB-N18-GFP combined with phosphatase treatment. As shown in Figure 6, this analysis resolved five species of the wild-type form of the SH4 fusion protein two of which were sensitive to phosphatase treatment. These two species could neither be detected for the T6A nor for the T6E mutant forms of the corresponding SH4 fusion proteins. These data establish that the wild-type form of HASPB-N18-GFP becomes phosphorylated at threonine 6 and possibly one additional site. In case there are two phosphorylation sites within the SH4 domain, the phosphothreonine residue at position 6 might act as a so-called priming phosphorylation site since the T6A mutant looses phosphorylation at both sites (Figure 6). These conclusions are consistent with our mutational studies in which a phosphothreonine residue at position 6 was mimicked by glutamate (Figures 1–5).

image

Figure 6. The SH4 domain of HASPB-N18-GFP is phosphorylated at threonine 6. CHO cells expressing the fusion proteins indicated were grown to a confluency of about 80%. Cell lysates were generated by detergent treatment. In each case one part of the sample was treated with λ-phosphatase, the other part being left untreated as a control. The samples were then analysed by 2D gel electrophoresis using commercial isoelectric focusing (IEF) strips containing an immobilized gradient ranging from pH 4 to 7 ('Ready-Strips’, Bio-Rad) for IEF (first dimension). For the second dimension, conventional SDS-PAGE was used. Following Western blotting, immunodetection was carried out using polyclonal rabbit-anti-GFP antibodies and goat-anti-rabbit Alexa Fluor 680 antibodies. Antigens were detected with an Odyssey imaging system. The spots observed for the T6E mutant of the HASPB-N18-GFP fusion protein were shifted towards the anode since this mutant contains an additional negative charge. In order to facilitate the comparison of the various spot patterns, the spot that was closest to the cathode (spot 1) was used to align the various patterns.

Download figure to PowerPoint

An inhibitor of threonine/serine-specific phosphatases causes intracellular accumulation of phosphorylated species of SH4 fusion proteins

To further verify the experiments shown in Figure 6 and to extend our studies to a fusion protein containing the SH4 domain of the Src kinase Yes, we analysed whether an inhibitor of threonine/serine-specific phosphatases, calyculin A, has an impact on the relative amounts of the phosphorylated species of HASPB-N18-GFP and Yes-N18-GFP, respectively. When cells expressing HASPB-N18-GFP were treated with calyculin A, the two phosphorylated species (spots 4 and 5 as identified in the experiments shown in Figure 6) were significantly increased in intensity as compared to control conditions (Figure 7, panel A). Quantification using a LI-COR Odyssey infrared imaging system revealed a more than twofold increase of spots 4 and 5 in the presence of calyculin A relative to the intensity of all spots being detected. In case of Yes-N18-GFP, we observed a total of five species in the presence of calyculin A (Figure 7, panel B), four of which (spots 2–5) appeared to be phosphorylated forms as they were more pronounced as compared to control conditions. While spot 2 was already detectable in the absence of calyculin A and became more prominent in its presence, spots 3–5 were only observed when cells were treated with calyculin A. Our data indicate that, like HASPB-N18-GFP, the SH4 domain of Yes-N18-GFP is a target for phosphorylation and dephosphorylation since calyculin A caused a strong increase of existing and the appearance of additional species of Yes-N18-GFP resolved by 2D electrophoresis (Figure 7, panel B). In case of HASPB-N18-GFP, the two phosphorylated species represented by spots 4 and 5 (Figure 6) could be confirmed as they accumulated in the presence of calyculin A (Figure 7, panel A). To further evaluate the hypothesis that phosphorylation and dephosphorylation of SH4 domains regulate their intracellular localization, the findings shown in Figures 6 and 7 prompted us to investigate whether SH4 fusion proteins are redistributed in the presence of calyculin A in cells.

image

Figure 7. Calyculin A, an inhibitor of threonine/serine-specific phosphatases, causes intracellular accumulation of phosphorylated species of HASPB-N18-GFP and Yes-N18-GFP, respectively. CHO cells expressing the fusion proteins indicated were treated for 2 h in the presence of calyculin A. Further processing including 2D gel analysis and Western blotting were performed as described in the legend of Figure 6. In case of HASPB-N18-GFP (panel A), spots 4 and 5 represent phosphorylated species as identified in the experiment shown in Figure 6. In case of Yes-N18-GFP (panel B), the species resolved in spots 2–5 appear to be phosphorylated forms as they are more prominent in the presence of calyculin A.

Download figure to PowerPoint

Calyculin A causes perinuclear accumulation of SH4 fusion proteins

The combined findings of the experiments shown in Figures 1–7 suggested that the T6E mutant form of the SH4 fusion protein reaches the plasma membrane, however, accumulates in perinuclear compartments upon internalization. To directly address this hypothesis we analysed whether calyculin A–mediated inhibition of the threonine/serine-specific phosphatases PP1 and PP2A (18,19) affects the localization of the wild-type form of SH4 fusion proteins (Figure 8).

image

Figure 8. An inhibitor of threonine/serine-specific phosphatases, calyculin A, causes perinuclear accumulation of SH4 fusion proteins. CHO cells were grown on Lab-Tek chamber slides and induced with doxycycline to express the wild-type and the T6A forms of the HASPB-N18-GFP fusion protein as well as a corresponding fusion protein containing the SH4 domain of the Src kinase Yes. The images shown were taken before and 2 h after treatment with 3.5 nm calyculin A. Panels A, B, D, E, G and H represent projections of all confocal images taken. Panels C, F and I represent selected confocal images in which intracellular accumulation was observed. Images were taken with a PerkinElmer Ultraview ERS spinning disk confocal microscope (using a Zeiss PlanApo 63 × 1.4 oil immersion objective). Arrowheads point at examples of perinuclear accumulation of SH4 fusion proteins. Bars = 10μm.

Download figure to PowerPoint

Following addition of calyculin A to the medium, live cells were imaged for up to 2 h. A projection of all confocal planes was used in order to cover all areas of the cells being analysed. As indicated by arrowheads, upon treatment with calyculin A, the wild-type form of HASPB-N18-GFP accumulated in a perinuclear region (Figure 8, panel B). Selected confocal planes are shown to illustrate this effect at high magnification (Figure 8, panel C). When cells were analysed by live cell imaging to conduct kinetic studies, substantial accumulation in perinuclear membranes was observed as early as 20 min following the application of calyculin A (Figure S3). Of note, the T6A variant form of the HASPB-N18-GFP fusion protein was not affected by calyculin A, maintaining its typical localization at the plasma membrane with only a minor fraction being present in perinuclear membranes at steady-state. This was evident both from projections of all confocal planes (Figure 8, panels D and E) and selected confocal planes viewed at high magnification (Figure 8, panel F).

These data suggest that inhibition of dephosphorylation of the SH4 domain directly causes a block of recycling back to the plasma membrane resulting in the accumulation in perinuclear structures. The fact that the T6A variant form did not respond to calyculin A with regard to a localization phenotype demonstrates that these findings are not related to pleiotropic effects this inhibitor may exert on for example cell viability.

While we could not extend our mutational studies to other SH4 proteins such as the Src kinase Yes due to the critical importance of threonine/serine 6 for the recognition by acyl transferases (17), we could now use calyculin A to ask whether dephosphorylation of SH4 domains also plays a role in the steady-state localization of related proteins. At steady-state, the fusion protein Yes-N18-GFP introduced in the experiments shown in Figure 7 was characterized by an almost exclusive association with plasma membranes [Figure 8, panel G; (11–13)]. Intriguingly, treatment of live cells with calyculin A caused a very similar localization phenotype as shown for the SH4 domain of HASPB. The Yes-N18-GFP fusion protein accumulated in a perinuclear region as evident from both projections of all confocal planes (Figure 8, panel H) and selected confocal planes (Figure 8, panel I).

Calyculin A–induced perinuclear accumulation of SH4 fusion proteins and endogenous Yes is reversible suggesting recycling between plasma membranes and intracellular compartments

To address a potential recycling pathway of SH4 proteins we sought to analyse whether perinuclear accumulation of SH4 fusion proteins caused by calyculin A is reversible (Figure 9). CHO cells expressing HASPB-N18-GFP or Yes-N18-GFP, respectively, were either left untreated (Figure 9, panels A and D) or were incubated in the presence of calyculin A for 2 h (Figure 9, panels B and E). Similar to the experiments shown in Figure 8, calyculin A caused perinuclear accumulation of both SH4 fusion proteins, however, following removal of the drug and 2 h of further incubation, both HASPB-N18-GFP and Yes-N18-GFP were redistributed resulting in an almost exclusive association with plasma membranes (Figure 9, panels C and F). These findings suggest that both SH4 fusion proteins cycle between plasma membranes and internal compartments, however, in the presence of calyculin A, they get trapped in perinuclear membranes due to a block in dephosphorylation of the SH4 domain. To extend our studies towards endogenous SH4 proteins we analysed whether the subcellular localization of the Src kinase Yes is affected in the presence of calyculin A. As shown in Figure 9 (panels G–I), calyculin A caused perinuclear accumulation of endogenous Yes which was reversible following removal of the drug. These findings rule out the formal possibility that the effects of calyculin A are restricted to SH4 fusion proteins that are expressed as exogenous factors in stably transduced cell lines.

image

Figure 9. Calyculin A–induced perinuclear accumulation of SH4 fusion proteins and endogenous Yes is reversible suggesting recycling between plasma membranes and endosomal compartments. CHO cells were induced with doxycycline to express the SH4 fusion proteins HASPB-N18-GFP (panels A–C) and Yes-N18-GFP (D–F), respectively. They were treated with calyculin A as indicated followed by live cell imaging. CHO cells that do not express exogenous SH4 fusion protein (panels G–I) were treated with calyculin A as indicated followed by fixation, permeabilization and staining with anti-Yes antibodies and appropriate secondary antibodies coupled to fluorophores. Arrowheads point at examples of perinuclear accumulation of SH4 fusion proteins and endogenous Yes, respectively. Bars = 10μm.

Download figure to PowerPoint

RNAi-mediated downregulation of the threonine/serine-specific phosphatases PP1 and PP2A causes perinuclear accumulation of both SH4 fusion proteins and the endogenous Src kinase Yes

The principal targets of calyculin A are the threonine/serine-specific phosphatases PP1 and PP2A (18,19). Therefore, in order to challenge the pharmacological evidence discussed earlier, we aimed at RNAi-mediated downregulation of all PP1 and PP2A isoforms known in the human genome (Figure 10). In these experiments, siRNAs directed against PP1 and PP2A isoforms were combined in two separate pools and used to transfect HeLa cells expressing HASPB-N18-GFP (Figure 10, panel A, sub-panels a–c) or Yes-N18-GFP (Figure 10, panel A, sub-panels d–f), respectively. Additionally, we investigated the localization of the endogenous Src kinase Yes by cell fixation, permeabilization and antibody staining under knockdown conditions of both PP1 and PP2A isoforms (Figure 10, panel A, sub-panels g–i). Knockdown efficiencies were monitored by real-time polymerase chain reaction (RT-PCR) using GAPDH as a control (Figure 10, panel B). Under control conditions using a scrambled siRNA for transfection, none of the proteins noted earlier could be detected in a perinuclear region (Figure 10, panel A, sub-panels a, d and g).

image

Figure 10. RNAi-mediated downregulation of the threonine/serine-specific phosphatases PP1 and PP2A causes perinuclear accumulation of SH4 fusion proteins and endogenous Yes. (A) HeLa cells were induced with doxycycline to express the fusion proteins indicated. They were either transfected with control siRNAs with a scrambled sequence or with a combination of siRNAs targeting all isoforms of either PP1 or PP2A. Cells expressing SH4 fusion proteins were analysed by live cell imaging. Endogenous Yes Src kinase was detected in non-transduced HeLa cells using antibody staining following fixation and permeabilization. Sub-panels a–c: HASPB-N18-GFP; Sub-panels d–f: Yes-N18-GFP; Sub-panels g–i: Endogenous Yes. Arrowheads point at examples of perinuclear accumulation of SH4 fusion proteins and endogenous Yes, respectively. Bars = 10μm. (B) RT-PCR analysis of RNAi-mediated downregulation of the mRNAs of all isoforms of PP1 and PP2A. As a control, mRNAs levels of GAPDH were analysed. The RT-PCR products shown have the following lengths (in base pairs): PP1(CA) = 236; PP1(CB) = 235; PP1(CC) = 235; PP2A(CA) = 175; PP2A(CB) = 184; GAPDH = 250.

Download figure to PowerPoint

By contrast, downregulation of both PP1 and PP2A isoforms, respectively, caused a substantial accumulation of HASPB-N18-GFP (Figure 10, panel A, sub-panels b and c), Yes-N18-GFP (Figure 10, panel A, sub-panels e and f) and endogenous Yes (Figure 10, panel A, sub-panels h and i) in a perinuclear region of HeLa cells. The specificity of the siRNA treatment was corroborated by the fact that single siRNAs directed against the individual isoforms CA of PP1 and CB of PP2A did not cause a localization phenotype for the proteins noted earlier (data not shown).

Consistent with the results obtained by pharmacological inhibition shown in Figures 8 and 9, these findings provide independent evidence that PP1 and PP2A are functionally redundant phosphatases that regulate recycling of SH4 proteins between plasma membranes and intracellular compartments.

Discussion

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Materials and Methods
  6. Acknowledgments
  7. References
  8. Supporting Information

Modifications with lipid anchors represent a general mechanism of membrane attachment of peripheral membrane proteins (3,5,20,21). One of these is fatty acylation of N-terminal SH4 domains as they are found in Src kinases (3–6) and Leishmania HASPB (9,10). Detailed knowledge is available about how acylated SH4 domains mediate protein binding to membranes and how these modifications promote partitioning into membrane microdomains (6,22). In various experimental systems including mammalian cells, parasites and yeast, it has been shown that, in addition to membrane binding, N-terminal SH4 domains are both necessary and sufficient to target fusion proteins to specific subcellular compartments (9–14,23,24). However, as opposed to membrane binding and microdomain partitioning, both the molecular mechanisms being involved in specific targeting of SH4 proteins and the dynamics of intracellular transport between different subcellular compartments are poorly understood.

In the current study we revealed a novel regulatory mechanism that controls the distribution of SH4 proteins between plasma membranes and perinuclear compartments. We demonstrate that the SH4 domain of Leishmania HASPB becomes phosphorylated at threonine 6 and that this modification causes sequestration in perinuclear membranes that are likely to represent an endosomal recycling compartment. We propose that this localization phenotype is due to a block in recycling between intracellular compartments and plasma membranes. According to this model, recycling back to the plasma membrane resumes by dephosphorylation providing a simple mechanism to control the amounts associated with plasma membranes. Our conclusions are based on: (i) a mutational analysis mimicking a phosphothreonine residue by glutamate in position 6 of the SH4 domain; (ii) Pharmacological studies using calyculin A, a specific inhibitor of the threonine/serine-specific phosphatases PP1 and PP2A and (iii) RNAi-mediated downregulation of PP1 and PP2A. With the two latter approaches, our findings could be extended to another SH4 domain derived from the Src kinase Yes raising the possibility that phosphorylation and dephosphorylation of SH4 domains may represent a general mechanism to regulate subcellular localization. Additionally, in the presence of calyculin A or following downregulation of PP1 and PP2A, similar localization phenotypes were observed for the endogenous Src kinase Yes suggesting that our findings with SH4 fusion proteins are physiologically relevant.

Src kinases have been shown before to be phosphorylated at a number of tyrosines as well as serines and threonines. Most of these modified residues are part of the SH3, SH2 and SH1 domains and have been reported to regulate Src kinase activity (25). Additionally, serine 12 and serine 17 of the SH4 domain are also substrates for phosphorylation catalysed by protein kinase C and protein kinase A, respectively (26). In both cases, no direct impact on the regulation of Src kinase activity was observed. Our findings suggest that the conserved threonine and serine residues found in position 6 within the SH4 domains of HASPB and Yes as well as many other SH4-domain-containing proteins represent another target of phosphorylation and that this modification is a crucial determinant of subcellular localization. The discovery of this regulatory mechanism was possible since we identified an SH4 domain that can be manipulated at threonine 6 without an appreciable loss of acylation efficiency. Replacement of this residue of the SH4 domain of the Leishmania protein HASPB by either glutamate or alanine did not cause levels of myristoylation and palmitoylation to decrease as it has been shown for Src kinases (17). Our data, however, suggest that reversible phosphorylation and dephosphorylation may not be a specific feature of the SH4 domain of HASPB but appears to play a role in the dynamic regulation of the subcellular localization of the Src kinase Yes as well. This is because both a pharmacological block of dephosphorylation and RNAi-mediated downregulation of the calyculin A targets PP1 and PP2A resulted in the asccumulation of both a Yes SH4 fusion protein and endogenous Yes in perinuclear membranes. Interestingly, when calyculin A was removed, the localization phenotype of both SH4 fusion proteins and endogenous Yes was reverted into normal plasma membrane localization. These findings suggest that these SH4 proteins cycle back and forth between plasma membranes and perinuclear compartments. This conclusion is consistent with our observation that, at steady-state, a sub-population of a fusion protein containing the wild-type form of the Leishmania HASPB SH4 domain was colocalizing with the transferrin receptor in internal membranes. Furthermore, our mutational studies in which threonine 6 was replaced by either glutamate or alanine also point to a recycling pathway through which SH4 proteins travel in a constitutive manner. These findings are consistent with previous reports in which the prototype member of the Src kinase family, Src, was shown to partially localize to endosomes and, upon activation, translocates to the inner leaflet of plasma membranes (27). It is possible that this type of regulation is based on the phosphorylation state of the Src SH4 domain. These findings also directly suggest a potential physiological relevance of this regulatory mechanism as Src has been proposed to serve different signalling pathways when present in either endosomes or plasma membranes (28).

In case of G-protein-coupled receptors, regulation of cell surface expression is well known to be controlled by phosphorylation and dephosphorylation of their cytoplasmic domains (29). For example, it has been shown that V2 vasopressin receptors contain multiple phosphorylation sites within these domains. Upon ligand binding, these receptors become activated followed by phosphorylation, desensitization and intracellular sequestration to prevent further signalling (30,31). This process generally involves the recruitment of arrestins to phosphorylated cytoplasmic domains of G-protein-coupled receptors followed by targeting to clathrin-coated pits (32). As a general mechanism, recycling back to the plasma membrane allows for their participation in new rounds of signal transduction and it has been demonstrated that this process critically depends on dephosphorylation of their cytoplasmic domains (33). Our findings suggest that this general concept of controlling the localization of various signalling molecules is likely to apply to acylated SH4 proteins as well and may represent a spatial control mechanism of Src kinase activity in various signalling pathways.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Materials and Methods
  6. Acknowledgments
  7. References
  8. Supporting Information

Cell lines and constructs

In this study various forms of the SH4 domain (N-terminal 18 amino acids) of the Leishmania cell surface protein HASPB and the SH4 domain of the Src kinase Yes (N-terminal 18 amino acids) were fused to GFP and a protein A tag. For the majority of the experiments shown stable CHO cell lines were used that were generated by retroviral transduction (34) of the various SH4 fusion constructs being under the control of a doxycycline-dependent transactivator. For RNAi experiments, stable HeLa cell lines were used that express the fusion proteins indicated in a doxycycline-dependent manner.

Confocal microscopy of both fixed and living cells

For live cell imaging cells were grown in Lab-Tek chamber slides. Living cells were visualized with a Zeiss LSM 510 Meta confocal microscope (using a Zeiss 100 × oil immersion objective) or with a PerkinElmer Ultraview ERS spinning disk confocal microscope (using a Zeiss PlanApo 63 × 1.4 oil immersion objective). CHO cells were cultivated for 24–48 h in the presence of 1μg/μL doxycycline until they reached about 80% confluency. In some experiments, cells were treated with calyculin A (final concentration: 3.5 nm), an inhibitor of threonine/serine-specific phosphatases. In other cases, the gene products indicated were downregulated by transfection with specific siRNAs in HeLa cells. The following siRNAs were used: Ambion ID s10931 (PP1 isoform CA), Ambion ID s10934 (PP1 isoform CB), Ambion ID s10961 (PP2A isoform CA), Ambion ID s719 (PP2A isoform CB) and Ambion ID s10959 (PP2A isoform CC). As a control, an siRNA with a scrambled sequence was used (Ambion ID 4611). SiRNAs were transfected using oligofectamine (Invitrogen); 96 h post-transfection cells were analysed by live cell imaging. Finally, where indicated, cells were fixed with paraformaldehyde, permeabilized with saponin and stained with antibodies directed against the Src kinase Yes (BD Transduction Laboratories). Images of fixed cells mounted on coverslips were taken with a LSM 510 confocal laser scanning microscope (Zeiss) using 63 × and 100 × oil immersion objectives.

Biochemical fractionation and carbonate extraction experiments

CHO cells were induced with doxycycline to express the fusion proteins indicated and grown to about 80% confluency. They were harvested in PBS buffer (supplemented with 1 mm DTT) using a rubber policeman and collected by low speed centrifugation. Cells were broken by three freeze-thaw cycles followed by repeated passaging (10×) through a blunt 27-gauge needle. All further steps were conducted at 4°C. The homogenates were centrifuged for 10 min at 10 000 g to remove cell debris and nuclei. The supernatants were subjected to ultracentrifugation at 100 000 g for 1 h. The corresponding supernatants were defined as the cytosolic fractions and the corresponding sediments as membrane fractions. Where indicated, sediments were treated with Na2CO3 (0.1m, pH 11.5) for 30 min at 4°C followed by ultracentrifugation at 100 000 g for 30 min. The supernatants of these centrifugations were analysed for loosely attached proteins and the corresponding sediments for proteins being tightly associated with membranes. Samples were analysed by SDS-PAGE and Western blotting using affinity-purified anti-GFP antibodies.

Quantitation of acylation levels of SH4 fusion proteins by metabolic labelling using [3H] myristate and [3H] palmitate

To quantify the acylation state of the various forms of SH4 proteins used in this study metabolic labelling using [3H] myristate and [3H] palmitate was performed. CHO cells expressing the HASPB-N18-GFP-ProtA fusion proteins indicated were grown in six well plates. Following cultivation in medium containing delipidated FCS and 2 h of incubation in medium lacking FCS, cells were labelled with 100 μCi/mL [3H] myristate or 250 μCi/mL [3H] palmitate, respectively. Following 3 h of further cultivation, cells were lysed with NP-40 and the HASPB-N18-GFP-ProtA fusion proteins indicated were immunoprecipitated using anti-GFP antibodies. Precipitates were separated on 4–12% Bis-Tris NuPAGE SDS gels (Invitrogen) and analysed by Western blotting using PVDF-FL membranes and an Odyssey imaging system (LI-COR). Then, the membranes were dried and analysed for [3H]-derived radioactivity using a β-imager (Biospace).

Flotation of DRMs

In order to isolate detergent-resistant membranes (35), CHO cells were induced with doxycyline and grown to 80% confluency. Cells were lysed with ice-cold 1% Triton X-100 and all subsequent steps were carried out at 4°C. Following clearance of the lysate, floatation on an Optiprep step gradient was performed: 300 μL of lysate were mixed with 600 μL of 60% Optiprep and loaded at the bottom of a SW60 centrifugation tube. The load of the gradient was covered with 2.5 mL 28% Optiprep solution containing 1% Triton X-100 and 600 μL of buffer containing 1% Triton X-100. Following centrifugation at 100 000 g for 3 h at 4°C, 8 fractions of 500 μL each were collected and analysed by SDS-PAGE and Western blotting as indicated.

Analysis of the phosphorylation status of SH4 domains using 2D gel electrophoresis and phosphatase treatment

CHO cells were induced to express the HASPB-N18-GFP fusion proteins indicated and grown to about 80% confluency. Where indicated, cells were treated for 2 h with calyculin A (final concentration: 3.5 nm), an inhibitor of threonine/serine-specific phosphatases. Following detachment cells were lysed in a Triton X-100-containing buffer. Where indicated, samples were treated with λ -phosphatase (NEB). Samples were separated on IEF strips containing an immobilized pH-gradient ranging from pH 4 to 7 (Bio-Rad). The strips were then equilibrated in 6 m urea, 30% glycerol, 2% SDS, 0.39 m Tris/HCl (pH 8.8), 1 mm DTT and 1 mm iodoacetamide, respectively and applied to SDS-PAGE. Following Western blotting, the various HASPB-N18-GFP fusion proteins were detected with affinity-purified anti-GFP antibodies and an Odyssey imaging system (LI-COR).

Immunogold electron microscopy

CHO cells were fixed in PHEM buffer [60 mm Pipes, 25 mm Hepes, 10 mm EGTA, 2 mm MgCl2, (pH 6.9)] supplemented with 4% paraformaldehyde and 0.1% glutaraldehyde for 60 min. Following incubation for 10 min in 50 mm glycine, cells were scraped in 1% gelatine and 0.1 m phosphate buffer. After embedding in 10% gelatine, small blocks were infused with 2.3 m sucrose and frozen in liquid nitrogen. Ultrathin cryosections were generated with a Leica Ultracut UC6 microtome and transferred onto carbon-coated, copper 300-mesh grids (36). All antibodies and gold conjugates were diluted in 1% BSA. The following primary antibodies were used: the transferrin receptor mouse monoclonal antibody (1:100; Zymed) and the GFP rabbit polyclonal antibody (1:500; Molecular Probes). Protein A coupled either with 10 or 15 nm gold particles was used and a rabbit anti-mouse antibody (1:150; Dako) was used as a bridge. Grids were examined at 120 kV using a Philips Biotwin CM120 electron microscope. Images were taken at magnifications that varied from 9700 to 17 500. The relative distribution of gold particles over the different compartments was estimated by counting the number of gold particles falling within the boundary of each structure. The labelling density of gold particles per μm membrane was estimated as described by Nilsson and colleagues (37). The length of every portion of membrane was estimated by the intersection method (38). The linear density was calculated by dividing the number of gold particles by membrane length.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Materials and Methods
  6. Acknowledgments
  7. References
  8. Supporting Information

This work was supported by the German Research Foundation (SFB 544, project B15) and CellNetworks—Cluster of Excellence (EXC81).

References

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Materials and Methods
  6. Acknowledgments
  7. References
  8. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Results
  4. Discussion
  5. Materials and Methods
  6. Acknowledgments
  7. References
  8. Supporting Information

Supporting Information

Additional Supporting Information may be found in the online version of this article:

Figure S1: Subcellular localization of the wild-type and the T6E mutant forms of the HASBP-GFP fusion protein in different compartments as determined by immuno-electron microscopy.

Cells were fixed and processed for cryo-immunogold labelling using anti-GFP and anti-transferrin receptor antibodies. Primary antibodies were detected with 10 and 15 nm gold-conjugated secondary antibodies, respectively. (A–D) HASPB-N18-GFP. (E–H) HASPB-N18-GFP-T6E. N, Nucleus; G, Golgi complex; ER, endoplasmic reticulum. Bars: 600 nm (A); 400 nm (B,E,H); 200 nm (C,F); 500 nm (D); 300 nm (G)

Figure S2: Subcellular localization of a palmitoylation mutant and a T6A variant form of the HASBP-GFP fusion protein in different compartments as determined by immuno-electron microscopy.

(A–D) HASPB-N18-Δpal-GFP. (E–H) HASPB-N18-GFP-T6A. N, Nucleus; G, Golgi complex; ER, endoplasmic reticulum. Bars: 400 nm (A,B,C,D,H); 600 nm (E); 300 nm (F); 200 nm (G)

Figure S3: Kinetics of calyculin A–induced accumulation of HASPB-N18-GFP in perinuclear membranes.

The experiment was performed as described in the legends of Figures 8 and 9 under the conditions described under 'Materials and methods’. Following addition of calyculin A, live cells were analysed by confocal microscopy at the times indicated. Arrows point at perinuclear structures in which HASPB-N18-GFP appeared in the course of the incubation. At 120 min, cells were washed and further incubated in the absence of calyculin A.

Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

FilenameFormatSizeDescription
TRA_921_sm_FigS1.tif110515KSupporting info item
TRA_921_sm_FigS2.tif109129KSupporting info item
TRA_921_sm_FigS3.tif71838KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.