Summary
Melanosome transport in melanocytes is a model system for the study of cytoskeletal regulation of intracellular transport. Melanophilin (Mlph) is a Rab27a- and myosin Va (MyoVa)-binding protein that regulates this process. Using yeast two-hybrid screening, we identified MT plus-end binding protein (EB1) as a melanocyte-expressed Mlph-interacting protein. To address the role of EB1 versus Rab27a and MyoVa interactions in Mlph targeting and function, we used siRNA and Mlph mutations to specifically disrupt each interaction in cultured melanocytes. Using the Mlph R35W mutant that blocks Mlph-Rab27a interaction and Rab27a siRNA we show this interaction is required for melanosome targeting and stability of Mlph. Mutants and siRNA that affect Mlph-MyoVa and Mlph-EB1 interactions reveal that while neither MyoVa nor EB1 affect Mlph targeting to melanosomes, MyoVa but not EB1 interaction is required for transport of melanosomes to peripheral dendrites. We propose that Mlph is targeted to and/or stabilised on melanosomes by Rab27a, and then recruits MyoVa, which provides additional stability to the complex and allows melanosomes to transfer from MT to actin-based transport and achieve peripheral distribution. EB1 appears to be non-essential to this process in cultured melanocytes, which suggests that it plays a redundant role and/or is required for melanocyte/keratinocyte contacts and melanosome transfer.
Introduction
Melanosome transport in mouse skin melanocytes provides an excellent model to study the regulation of intracellular organelle transport by cytoskeleton and membrane trafficking proteins. Several mutant mice manifesting defects in melanosome synthesis and movement are known to lack proteins of these classes, e.g. Rab GTPases and MyoVa, allowing the role of these proteins to be studied in this process (reviewed in Bennett and Lamoreux, 2003).
Melanosomes are specialised, pigment-containing, lysosome-related organelles and their synthesis by melanocytes and transfer to adjacent keratinocytes provides the basis of mammalian pigmentation (Marks and Seabra, 2001). Melanosome transport in melanocytes is proposed to be a two-step process. Following synthesis melanosomes are thought to undertake fast, long-distance transport into dendrites along microtubules by the action of kinesin and dynein motors regulated by the small GTPase Rab7 (Byers et al., 2000; Hara et al., 2000; Jordens et al., 2006; Vancoillie et al., 2000a; Vancoillie et al., 2000b). They are retained there by interaction with the actin cytoskeleton mediated by recruitment of MyoVa (Wu et al., 1998). This model is based on observation of melanocytes derived from the dilute mutant mouse (lacking MyoVa protein), in which melanosomes are clustered around the nucleus, and not spread throughout the peripheral cytoplasm and dendrites as in wild-type melanocytes. Two other mutant mice, ashen (lacking Rab27a protein) and leaden [lacking melanophilin protein (Mlph)], exhibit similar melanosome distribution and these proteins appear to anchor the MyoVa motor on the melanosome (Wu et al., 2002). This mechanism allows melanosomes to be retained in dendrites and make short myosin driven movements along actin filaments towards the plasma membrane prior to transfer to keratinocytes. In addition, Rab8 has been suggested to regulate actin-based transport of melanosomes via unknown motors (Chabrillat et al., 2005).
Recent work has mapped Mlph functional domains (Hume et al., 2006). This work identified the Mlph exon F binding domain (EFBD) and the amphipathic α-helix or coiled-coil forming regions as key MyoVa-binding regions and tandem synaptotagmin homology domains (SHDs) as key Rab27a-binding regions. The integrity of these functional domains is essential for melanosome transport to peripheral dendrites in living melanocytes. In contrast, regions of Mlph involved in MyoVa globular tail binding (GTBD) and the extreme C-terminus actin-binding domain (ABD) fulfil a redundant functional role(s) in melanosome transport.
An independent study of Mlph documented its interaction with the MT plus-end tracking protein EB1 (Wu et al., 2005). EB1 is a highly conserved protein present at the growing plus ends of MTs where it promotes MT growth and acts as an adaptor binding other proteins and allowing them to be transported to the cell periphery (Morrison, 2007; Vaughan, 2005). Interaction of Mlph and EB1 is proposed to regulate the site of transfer of melanosomes from MT to actin-based transport ensuring this only occurs in peripheral dendrites. In particular, EB1, Mlph and MyoVa are proposed to form a complex at the growing plus end of MTs that is targeted to the periphery by MT polymerisation. Meanwhile, melanosomes containing Rab27a are moved to the peripheral MT plus tips by the action of processive MT motors that `walk' along the MT track. This model envisages that when melanosomes reaching the peripheral plus end of the MT melanosomal Rab27a interacts with MT-associated Mlph-MyoVa, allowing melanosomes to be released from MTs and immediately captured into the surrounding cortical actin cytoskeleton (Wu et al., 2005).
In this study we directly address the role of each of the described Mlph interactions in Mlph targeting and melanosome transport in living melanocytes. To do this we use a melan-ln melanosome transport assay that allows us to assess the melanosome transport function of Mlph point mutants that abolish specific Mlph interactions in living melanocytes. In parallel we use siRNA that knockdown specific Mlph-interacting proteins and examine their effect on melanosome distribution in wild-type and melan-ln melanocytes. Conversely we probed the functional relationship between Mlph and EB1 in melanocytes by measuring MT plus-end tracking function of EB1 in the presence and absence of Mlph.
Results
EB1 is a Mlph-interacting protein expressed by melanocytes
To identify proteins that interact with the C-terminal portion of Mlph we screened a wild-type melanocyte-derived library for proteins that bind mouse Mlph aa367-590 (see Materials and Methods). Previously this region was implicated in the interaction of Mlph with actin and MyoVa (Hume et al., 2006; Kuroda et al., 2003). The screen identified one interacting clone encoding the C-terminal 134 amino acids (aa134-268) of the MT plus end binding protein EB1 (Fig. 1A). This part of EB1 binds other mammalian proteins, including the MT and actin cross-linking factor (MACF) family proteins, the adenomatous poliposis coli protein (APC) and the dynactin complex component p150glued (Askham et al., 2002; Barth et al., 2002; Hayashi et al., 2005; Honnappa et al., 2005; Slep et al., 2005). Our data confirm the idea that this domain allows EB1 to bring other proteins into the proximity of growing MTs.
The C-terminus α-helical region of EB1 interacts with a conserved IP motif present in the C-terminus of Mlph
While the tertiary structure of Mlph is unknown the structure of N- and C-terminal domains of EB1 have been solved using X-ray crystallography (Hayashi and Ikura, 2003; Hayashi et al., 2005; Honnappa et al., 2005; Slep et al., 2005). The N-terminus forms a single calponin homology (CH) domain that allows interaction with MTs, while the C-terminus comprises a pair of α-helices that allow EB1 to dimerise forming a four helix bundle. To examine the structural requirements of EB1 for interaction with Mlph we produced a number of different EB1 truncations and tested their interaction with Mlph using a yeast two hybrid binding assay (Fig. 1A). We found that full-length EB1 (aa1-268) and the C-terminal portion of EB1 (aa134-268) interact with the C-terminal portion of Mlph aa367-590, as does the positive control melanocyte-specific MyoVa tail (MyoVa MSGTA). By contrast, we found that the N-terminal MT-binding region of EB1 (aa1-134) is unable to interact with Mlph aa367-590, as is the negative control rat Rab27a. Truncation of the C-terminus at aa238 (EB1 aa171-238) disrupts the Mlph interaction, while truncation of the N-terminus at aa204 (EB1 aa204-268) interacts, indicating that the C-terminal α-helical region of EB1 mediates interaction with Mlph.
Mapping of EB1 and Mlph domains required for interaction. For A and B the left-hand part shows the relationship of truncated test proteins, indicated with solid lines, to the domain structures of EB1 (A) and Mlph (B). The right-hand part shows the results of the β-galactosidase reporter gene assay for each test protein with either Mlph aa367-590 (A) or EB1 aa1-268 and MyoVa MSGTA (B). α1 and α2 denote the position of α-helices in EB1; ABD, actin-binding domain; CC, coiled coil; CH, calponin homology domain; EFBD, MyoVa exon F binding domain; GTBD, MyoVa globular tail binding domain; MBD, MyoVa binding domain; SHD1, synaptotagmin homology domain; R27BD, Rab27a binding domain; ZnF, Zn2+ finger. Panel C is an alignment of part of Mlph and MyRIP sequences from human (Hs Mlph accession NP_077006. 1 and Hs MyRIP accession BAC15555.1), mouse (Mm accession BAB41087.1), rat (Rn accession NP_001012135.1), chicken (Gg Mlph accession XP_421876. 2) and zebrafish (Dr Mlpha accession NP_001073147.1 and Dr Mlphb accession XP_685065.2) showing the position of the Mlph IP motifs (red) and the coiled coil region (green). Alignment of sequences was conducted using Muscle protein multiple sequence alignment software available at http://phylogenomics.berkeley.edu/cgi-bin/muscle/input_muscle.py. Numbering indicates the amino acid position of each sequence.
To pinpoint the EB1-binding region of Mlph we generated truncated Mlph molecules and tested their ability to interact with EB1 using the yeast two-hybrid binding assay (Fig. 1B). This analysis revealed that a region within the extreme C-terminus of Mlph (aa481-520) contains the EB1-binding site. By contrast, MyoVa MSGTA interacts with Mlph aa367-485 but not truncated proteins lacking this portion of Mlph. This suggests that EB1 and MyoVa bind different motifs in Mlph (Fig. 1B). Several studies have highlighted the importance of an isoleucine and proline (IP) motif in allowing interaction of otherwise unrelated proteins MACF and APC with EB1 (Honnappa et al., 2005; Slep et al., 2005). Interestingly, the Mlph EB1-binding site (mouse Mlph aa481-520) contains two IP motifs (I500P501 and I514P515) (Fig. 1C). To test the contribution of each IP motif to EB1 interaction we mutagenised the I and P amino acids at both positions to alanines (Mlph IP1=I500AP501A and Mlph IP2= I514AP515A) and tested the interaction of these mutant Mlph proteins with EB1 using the yeast two-hybrid binding assay (Fig. 1B). In summary, we found that both MlphIP1 and MlphIP2 interact with MyoVa MSGTA; however, Mlph IP2 interacts with EB1 whereas Mlph IP1 does not. These results indicate that Mlph aa481-520 defines a binding site for EB1 and that the IP1 motif is crucial for the interaction of Mlph with EB1. Consistent with its important role in providing a binding site for EB1, the IP1 motif is conserved in all known mammalian Mlph proteins whereas IP2 is not (Fig. 1C).
Mlph interacts with EB1 and MyoVa simultaneously via distinct binding sites
To confirm the interaction of Mlph and EB1 we used an in vitro assay that measures the interaction of purified recombinant MBP-tagged EB1 and GST-tagged Mlph aa150-590 (see Materials and Methods). As expected we observed interaction of GST-Mlph 150-590 with MBP-EB1 and MBP-MyoVaMSGTA (positive control) whereas GST alone does not interact with MBP-EB1 (negative control) (Fig. 2A lanes 1, 6 and 9). Measurement of the ratio of intensity of bands for MBP proteins to GST-Mlph 150-590 indicates that MyoVaMSGTA interacts with Mlph with higher affinity than does EB1 (Fig. 2A lanes 1 and 6).
Measurement of the interaction of Mlph with EB1 and MyoVa-MSGTA in vitro. Panels A and B show interaction with EB1 and MSGTA involves different regions of Mlph (A) and that Mlph may interact simultaneously with EB1 and MSGTA (B). One hundred pmol of GST-Mlph protein (A) or MBP-EB1 (B) was incubated with equimolar amounts of MBP-EB1 or MBP-MSGTA (A) or his6-Mlph and/or his6-MSGTA (B) in the presence of glutathione-sepharose (A) or amylose agarose (B) as described in the Materials and Methods. Bound proteins were precipitated, eluted from the beads, and subjected to immunoblotting using indicated antibodies. In A, relative binding (see bar chart) was calculated by dividing band intensity of EB1 or MSGTA by the band intensity of Mlph (as described in Materials and Methods). Experiments were carried out in triplicate and the migration of molecular-weight standards is indicated on the left of each blot.
Yeast two-hybrid data suggest that MyoVa and EB1 interact with Mlph via distinct regions of Mlph (Fig. 1B). Testing in vitro interaction of EB1 and MyoVa MSGTA with GST-Mlph truncations and point mutants confirms this is the case (Fig. 2A). For example, MBP-EB1 interacts with GST-Mlph400-590aa (lane 5) but not GST alone (lane 9) or GST-Mlph150-400aa (lane 4), previously shown to interact with MyoVa MSGTA (Hume et al., 2006). Further supporting this idea we found that Mlph point mutations GST-Mlph 150-590 AP and GST-Mlph 150-590 IP1 specifically reduce the in vitro binding of MBP-MyoVa MSGTA and MBP-EB1, respectively (Fig. 2A lanes 2, 3, 7 and 8).
These findings suggest that MyoVa and EB1 bind Mlph simultaneously. To test this possibility we incubated MBP-EB1 with his6-MyoVa MSGTA and full-length his6-Mlph, pulled down the MBP protein using amylose beads and probed the precipitates for the presence of Mlph and MyoVa MSGTA using specific antibodies (Fig. 2B). We confirm that MBP-EB1 interacts with full-length his6-Mlph (Fig. 2B lane 2) and show that its interaction with his6-MyoVa MSGTA occurs only in the presence of his6-Mlph (Fig. 2B lanes 1, 3 and 4). This confirms the idea that interaction of Mlph with EB1 and MyoVa MSGTA may occur simultaneously via distinct regions of Mlph.
EB1 shows little interaction with Mlph on melanosomes in melanocytes
The finding that EB1 is a melanocyte-expressed protein interacting with Mlph suggests that it plays a role with Mlph in regulating the intracellular transport of melanosomes. As a first step to investigate the function of EB1 in melanosome transport we examined its intracellular localisation in wild-type melan-ink melanocytes using confocal immunofluorescence microscopy. Consistent with many reports on the intracellular localisation of EB1 we observed that a large proportion of EB1 is present in comet-like structures that have a bright spot at one end adjacent to a dimmer tail (Fig. 3A,A`) (Berrueta et al., 1998; Morrison et al., 1998). These structures are approximately 1-2 μm in length and 0.5 μm in width corresponding to the growing plus ends of cytoplasmic MTs (A.N.H., unpublished). Comparison of the localisation of these structures with the localisation of melanosomes, either by direct observation of melanin pigment (Fig. 3C,C`,D,D`) or by co-staining melanocytes with antibodies that recognise the melanosomal protein Mlph (Fig. 3B,B`) reveal that little if any EB1 is present on the melanosome membrane (Fig. 3E,E`).
EB1 MT plus end tracking velocity is altered by loss of Mlph
As a first step to test whether there might be a transient functional interaction of EB1 with Mlph/melanosomes in living melanocytes not apparent in fixed cells we used time-lapse total internal reflection fluorescence (TIRF) and phase-contrast microscopy to examine the dynamics of fluorescently labelled EB1 (EB1-EGFP) and melanosomes, respectively, in living wild-type and melan-ln melanocytes. In both cell types we observed EB1-EGFP was distributed in comet-like structures, likely to be the growing plus ends of MTs, which move towards the periphery (Fig. 4A; see also supplementary material Movies 1 and 2). Comparison of fluorescence and phase contrast images indicate that, as for endogenous EB1 in fixed cells, there is no strong correlation between the EB1-EGFP labelled MTs plus ends and melanosomes in living wild-type or melan-ln melanocytes (Fig. 4B). Also time-lapse recording did not reveal significant correlation between movements of EB1-EGFP comet and melanosomes (Fig. 4B; see also supplementary material Movies 1 and 2). Tracking of EB1-EGFP comets in wild-type and melan-ln cells revealed that loss of Mlph results in a change in the distribution of comet velocities with more slow and fast comets being apparent although the average comet velocity for each population was similar for each population (average speed for wild type=0.203±0.106 μm/second and melan-ln=0.217±0.134 μm/second) (Fig. 4C). These data indicate that Mlph plays a role in regulating EB1 comet velocity and MT polymerisation.
Intracellular distribution of endogenous Mlph and EB1 in wild-type melan-ink melanocytes. Melan-ink cells were fixed and stained using antibodies that specifically recognise EB1 (A) and Mlph (B) as described in Materials and Methods. Panel C is the corresponding transmitted light image showing the distribution of melanosomes, panel D is an inverted and filtered copy of the transmitted light images showing the position of melanosomes and panel E is a merged image showing EB1 (green), Mlph (red) and melanosomes (blue). Panels A`-E` are higher magnification images of the indicated regions of A-E. Bar, 20 μm.
EB1 comet movements in wild-type and melan-ln melanosomes. Wild-type and melan-ln melanocytes were transiently transfected with plasmid encoding EB1-EGFP and the movements of EB1 comets and melanosomes were recorded as described in Materials and Methods. Panel A shows the distribution of EB1-EGFP in a melan-ln melanocyte and a magnified section of this cell in which arrows indicate individual comet structures. Panel B shows the tracks of melanosomes (red) and EB1-EGFP comets (green) in individual transfected wild-type and melan-ln cells over the course of a 95 second time-lapse recording (these tracks correspond to Movies 1 and 2 in supplementary material). Bars, 10 μm (A); 5 μm (B). Panel C shows the velocity of EB1-EGFP comets in representive wild-type (▪) and melan-ln (□) melanocytes. The velocities of individual comets (n=2541 comets from 20 cells for wild-type and n=1711 comets from 10 cells for melan-ln) were binned into velocity ranges of 0.025 μm/sec and then plotted showing the percentage of the total population that falls within each bin.
Rescue of melan-ln melanosome transport defects by Mlph requires interaction with Rab27a and MyoVa but not EB1
To test the physiological relevance of Mlph interaction with EB1 in melanosome transport compared with other Mlph interactions, e.g. Rab27a and MyoVa, we tested the ability of Mlph point mutations MlphR35W, MlphA467P and MlphIP1, which specifically disrupt interaction with EB1, Rab27a or MyoVa, respectively, to rescue melan-ln melanosome transport defects (Hume et al., 2006; Menasche et al., 2003) (this study).
We found that MlphR35W and MlphA467P (and negative control Slp1) are unable to rescue melanosome transport defects (Fig. 5E-P,U; Table 1). However, whereas MlphA467P is recruited to the cytoplasmic surface of melanosomes, MlphR35W often presents a diffuse cytoplasmic distribution (Fig. 5I-P). High magnification images reveal that puncta of MlphR35W decorate linear structures, likely to be actin filaments, present in the peripheral cytoplasm (Fig. 5I inset panel). These observations indicate that interaction with Rab27a is required for localisation of Mlph to the melanosome membrane, whereas Mlph-mediated recruitment of MyoVa is required for capture of melanosomes into the actin-rich peripheral dendrites. Finally, we observed that the MlphIP1 mutant (and positive control wild-type Mlph) is targeted to melanosomes and rescues melanosome transport with similar efficiency to wild-type Mlph (Fig. 5A-D,Q-U). This observation indicates that interactions with Rab27a and MyoVa are essential for melanosome transport and Mlph targeting and indicates that interaction with EB1 is not required for either of these processes.
Rescue of melan-ln melanosome transport defects by Mlph mutants that do not bind specific Mlph-interacting proteins
siRNA depletion of Rab27a, Mlph and MyoVa but not EB1 disrupts melanosome transport
As a second approach to investigate the role of Mlph-interacting proteins in melanosome transport, we used siRNA to deplete these proteins and then examined the effect on melanosome transport. In the first siRNA experiment wild-type melanocytes were transfected with siRNA oligonucleotides that specifically deplete Rab27a, Mlph, MyoVa or EB1 proteins and the distribution of melanosomes was examined in depleted cells 72 hours later. Transfection of melan-ink with Mlph-specific siRNA oligonucleotides, but not control siRNA oligonucleotides, resulted in depletion of Mlph protein to background levels in 70-80% of cells and perinuclear accumulation of melanosomes, but had no effect on the expression of control proteins calnexin and actin (Fig. 6A,B,C,D, Fig. 7A,B, lanes 5 and 8; see also supplementary material Fig. S1). Similarly to Mlph depletion, knockdown of either Rab27a or MyoVa in melan-ink cells resulted in redistribution of melanosomes from peripheral to perinuclear clustered distribution (Fig. 6E,F,I,J and Fig. 7C,D, lanes 1-4). In addition, staining of Rab27a and MyoVa siRNA-treated wild-type cells with Mlph antibodies revealed that Mlph targeting to melanosomes [and stability (discussed below)] is affected strongly by loss of Rab27a but not MyoVa (Fig. 6G,H,K,L). These observations are consistent with the distribution of Mlph in ashen (Rab27a null) and dilute (MyoVa null) mutant melanocytes (see supplementary material Fig. S2). By contrast, siRNA knockdown of EB1 effectively depleted EB1 protein levels but in the majority of cases (four out of five different pairs of siRNA oligonucleotides) melanosomes remained distributed throughout the peripheral cytoplasm to the same extent as control siRNA-transfected cells (Fig. 6M,N, Fig. 7, lanes 6 and 7; see also supplementary material Fig. S1). The exception to this general finding is EB1#1 siRNA oligonucleotide pair that reduces EB1 and Mlph protein level and causes perinuclear melanosome clustering with high efficiency (see supplementary material Fig. S1). This is likely to be an `off target' effect of this pair of oligonucleotides.
Disruption of Mlph interaction with Rab27a and MyoVa, but not EB1, affects rescue of melan-ln melanosome transport defects. Melan-ln cells were transfected with plasmids encoding the indicated Mlph point mutant molecules and control proteins. Cells were fixed 48 hours later and stained with antibodies to detect Myc-tagged Mlph and Slp1 protein. Panels A, E, I, M and Q show the distribution of overexpressed protein, panels B, F, J, N and R are transmission images showing the distribution of pigment; panels C, G, K, O and S are inverted and filtered copies of the transmission images showing the distribution of melanosomes and panels D, H, L, P and T are merged images in which the distribution of myc tagged Mlph and Slp1 protein (green) and melanosomes (red) are superimposed. Bars, 20 μm. Arrows in panels A-H and M-T indicate colocalisation of Mlph protein with the cytoplasmic surface of melanosomes. Arrows in panel I indicate association of Mlph with linear actin filaments. Panel U shows quantification of the extent of rescue of melanosome transport by the Mlph point mutants. Rescue efficiency for populations of transfected cells was determined as described in Hume et al. (Hume et al., 2006). Shaded boxes indicate 25th/75th percentile; bars within boxes indicate median values for each population of cells; outer bars indicate 5th/95th percentile; and outer points indicate outliers. The statistical significance of rescue measurements for populations of Mlph mutant transfected cells relative to positive (wild-type Mlph) and negative (Slp1) controls is shown in Table 1.
Depletion of Mlph together with interacting proteins Rab27a and MyoVa, but not EB1, causes clustering of wild-type melanosomes. Wild-type melan-ink melanocytes were transfected with the siRNA oligonucleotides that specifically prevent the synthesis of the indicate proteins. Transfected cells were fixed 72 hours later and stained with the indicated specific antibodies to allow correlation of depletion of protein to melanosome distribution. Panels A, C, E, G, I, K and M are fluorescence images showing the expression and distribution of the indicated protein in transfected cells and panels B, D, F, H, J, L and N are transmitted light images showing the distribution of pigment granules in transfected cells. Bar, 20 μm.
As a second siRNA approach to test the positive contribution of different Mlph interacting proteins in melanosome transport into dendrites we sequentially transfected melan-ln melanocytes with siRNA oligonucleotides that deplete individual Mlph interacting proteins and then with plasmid DNA allowing the re-expression of Mlph protein. We then examined localisation of re-expressed Mlph and melanosome distribution to determine the effects of depletion of each siRNA target on Mlph targeting and melanosome transport function. Primary transfection with Rab27a and MyoVa-specific siRNA oligonucleotides both prevent Mlph-mediated rescue of melan-ln melanosome transport defects while transfection with control siRNA oligonucleotides did not (Fig. 8A-J,O; Table 2). After Rab27a depletion, Mlph fails to target melanosomes, whereas in MyoVa-depleted cells Mlph melanosomal targeting is normal. By contrast, targeting and function of re-expressed Mlph is unaffected by primary EB1 depletion (Fig. 8K-O).
Rescue of melan-ln melanosome transport defects by re-introduction of epitope-tagged Mlph following siRNA knockdown of Mlph-interacting proteins
siRNA-mediated depletion of individual Rab27a-Mlph-MyoVa complex components affects the overall stability of the complex. Pools of wild-type melan-ink melanocytes were transfected with the indicated siRNA oligonucleotide, harvested 72 hours later and lysates subjected to immunoblotting using the indicated antibodies. Transfections were carried out in triplicate and the migration of molecular-weight standards is indicated on the left of each blot.
Together these data indicate that Rab27a, Mlph and MyoVa are essential for retention of melanosomes in peripheral dendrites, Rab27a is required for targeting of Mlph to melanosomes, MyoVa is required for attachment of melanosomes to actin, and EB1 is not required for targeting or function of Mlph in melanosome transport.
Expression of individual Rab27a-Mlph-MyoVa complex components regulates the stability of the complex
Previous data have indicated that Rab27a, Mlph and MyoVa are components of a physical tripartite complex (Fukuda et al., 2002; Hume et al., 2002; Hume et al., 2001; Nagashima et al., 2002; Provance et al., 2002; Wu et al., 2002). In the course of immunoblotting lysates of siRNA-transfected melan-ink cells to confirm depletion of the siRNA-targeted protein we noticed that depletion of certain components of the Rab27a-Mlph-MyoVa complex resulted in specific, concomitant downregulation of other complex components (Fig. 7). In particular, we noticed that siRNA depletion of Rab27a resulted in downregulation of both Mlph and MyoVa but not EB1 or calnexin (Fig. 7A-E lanes 1 and 2). This suggests that the association of Rab27a, Mlph and MyoVa in a melanosome-associated complex plays a significant role in stabilising Mlph and MyoVa proteins. Consistent with this idea depletion of either Mlph or MyoVa resulted in downregulation of both Mlph and MyoVa proteins but not Rab27a, EB1 or calnexin (Fig. 7A-E lanes 3-5). Finally, depletion of EB1 does not affect the expression level of Mlph or its other interacting partners Rab27a and MyoVa (Fig. 7A-E lanes 6 and 7).
Together these observations suggests that EB1 does not function together with the Rab27a-Mlph-MyoVa melanosome transport complex and that Rab27a, rather than EB1, is likely to play a crucial role in regulating the assembly of the tripartite complex on melanosomes.
Discussion
In the present paper we describe the interaction of Mlph with the MT plus end-binding protein EB1 and directly test the functional role of this and other Mlph interactions in melanosome transport in melanocytes. Our results suggest that Mlph and EB1 interact in vitro; however, there is no physical or functional evidence that the Mlph-EB1 interaction is important for initial targeting of Mlph or its function in melanosome transport to peripheral dendrites. This suggests that Mlph-EB1 interaction is either functionally redundant or required for Mlph function(s) distinct from melanosome transport in melanocytes. Furthermore, we find that Mlph-Rab27a and Mlph-MyoVa interactions are absolutely essential for melanosome transport. We present evidence that interaction with Rab27a is the primary targeting and stabilisation factor for Mlph, while interaction with MyoVa allows Rab27a- and Mlph-containing melanosomes to be captured into peripheral dendrites. Our data strongly support the model of sequential recruitment of Mlph and MyoVa to the melanosome membrane by Rab27a (Fukuda et al., 2002; Hume et al., 2002; Nagashima et al., 2002; Provance et al., 2002; Wu et al., 2002) and argues against the role of EB1 in targeting Mlph and MyoVa to the peripheral dendrites by tracking the plus ends of growing MTs prior to their association with melanosomes (Wu et al., 2005).
A previous study proposed that the Mlph-EB1 interaction allows a complex comprising EB1, Mlph and MyoVa to form at the plus tips of MTs, which directs the transfer of melanosomes from MT to actin-based transport (Wu et al., 2005). The results of yeast two-hybrid (Fig. 1) and biochemical binding assays (Fig. 2) are entirely consistent with the possibility that Mlph-EB1 interaction can occur. However, colocalisation studies and several functional tests of the interaction of Mlph with EB1 did not show a strong effect of the loss of either protein on the function of the other (Figs 3, 5, 6, 8). For instance, MlphIP1 mutant that no longer binds EB1 still targets to melanosomes and rescues melan-ln melanosome transport defects (Fig. 5), siRNA knockdown of EB1 does not affect melanosome distribution in wild-type cells or Mlph-dependent rescue of melan-ln defects (Figs 6, 8; see also supplementary material Fig. S1). Also overexpression of dominant negative truncated EB1 molecules comprising the Mlph-binding site does not affect melanosome distribution in the dendrites of wild-type melanocytes (A.N.H., unpublished). Moreover, Mlph point mutants that fail to interact with Rab27a or MyoVa do not show enhanced association with the growing tips of MTs but instead associate with filamentous actin and melanosomes, respectively (Fig. 5I,M). Conversely, we find that the microtubule plus end tracking activity of EB1-EGFP fusion protein together with the speed of microtubule plus end growth is only moderately affected by the loss of Mlph (Fig. 4). Also by fluorescence microscopy we observed little colocalisation of endogenous or EGFP-tagged EB1 with Mlph or melanosomes in fixed or living wild-type or melan-ln melanocytes (Figs 3, 4). All of the above observations argue strongly against a role of EB1 in targeting Mlph and MyoVa to peripheral tips of dendrites and against a role for Mlph-EB1 interaction in transport to and retention of melanosome in peripheral dendrites.
Depletion of Rab27a and MyoVa, but not EB1, prevents Mlph-mediated rescue of melan-ln melanosome transport defects. Melan-ln melanocytes were sequentially transfected first with siRNA to allow depletion of the indicated proteins and second with plasmid DNA allowing expression of Myc-tagged Mlph as described in Materials and Methods. Cells were then stained with antibodies allowing detection of the overexpressed Myc-Mlph (panels A,C,G,K) and siRNA-targeted protein to confirm depletion in each cells (panels D,H,L). Panels E, I and M are merged fluorescent images showing Myc-Mlph (green) and siRNA-targeted protein (red) and B, F, J and N are transmission images showing the distribution of melanosomes. Arrows indicate colocalisation of Mlph protein with the cytoplasmic surface of melanosomes. Panel O shows quantification of the extent of rescue of melanosome transport by the Mlph in melan-ln cells siRNA-depleted of Mlph-interacting proteins using the indicated oligonucleotides. Rescue efficiency for populations of transfected cells was determined as described in Hume et al. (Hume et al., 2006). Shaded boxes indicate 25th/75th percentile; solid and dotted bars within boxes indicate median and mean values for each population of cells; outer bars indicate 5th/95th percentile; and outer points indicate outliers. The statistical significance of rescue measurements for populations of transfected cells relative to positive and negative controls is shown in Table 2. Bar, 20 μm.
Our parallel study of the functional role of other Mlph-interacting proteins indicate that Rab27a is the primary Mlph targeting and stabilisation factor, whereas MyoVa also stabilises Mlph and allows peripheral retention of melanosomes via direct interaction with actin. The first of these conclusions is supported by the findings that Griscelli Syndrome type 3 (GS3) patient mutation MlphR35W, unable to interact with Rab27a, is unable to be recruited to melanosomes or rescue melan-ln melanosome transport defects (Fig. 5). Also, loss of Rab27a, due to siRNA knockdown or ashen mutation, leads to concomitant downregulation of Mlph, defects in melanosomal targeting of Mlph and disruption of melanosome transport (Figs 6, 7, 8; see also supplementary material Fig. S2). The second conclusion is supported by the findings that MlphA467P, although melanosome-targeted via interaction with Rab27a, is unable to rescue melan-ln melanosome transport defects due to its reduced MyoVa-binding affinity (Fig. 2A, Fig. 5). Moreover, MyoVa loss, either due to siRNA knockdown or the dilute mutation, leads to disruption of melanosome transport but not melanosomal targeting of Mlph (Figs 6, 8 and supplementary material Fig. S2).
siRNA knockdown experiments reveal that expression and stability of Mlph and MyoVa in wild-type melanocytes is coupled to the expression of the other Rab27a-Mlph-MyoVa complex components (Fig. 7). This is consistent with their existence as a stable functional complex in wild-type cells and previous similar observations of expression levels made in mutant primary melanocytes (Hume et al., 2002; Wu et al., 2002). However, we noted that Rab27a expression levels are less sensitive to knockdown of the other complex components. One reason for the stability of Rab27a in the absence of Mlph and MyoVa could be that Rab27a is able to undertake stabilising interactions with other members of the Slp/Slac family expressed in melanocytes. For instance Slp2a is reported to be expressed in melanocytes and function together with Rab27a in melanosome transport after the Mlph-MyoVa complex (Kuroda and Fukuda, 2004). By contrast, the sensitivity of Mlph and MyoVa to the integrity of the complete Rab27a-Mlph-MyoVa complex might reflect their exclusive function in this complex. Finally we observed no clear correspondence between knockdown of Rab27a-Mlph-MyoVa proteins and EB1 expression or vice versa indicating that these proteins fulfil less closely related functions.
What then is the role of Mlph-EB1 interaction in Mlph function? One possibility is that interaction of Mlph with EB1 is important for Mlph function in other cell types. At present the role of Mlph in other cellular functions is unknown as GS3 patients and leaden mice (Mlph null) do not display strong phenotypic abnormality aside from albinism. More thorough examination of cells derived from these sources may reveal hitherto unknown Mlph functions. Another possibility is that Mlph-EB1 interaction is involved not in the basic mechanism of melanosome transport but in other specialised aspects of melanosome transport for instance intercellular transfer of pigment from melanocytes to keratinocytes in skin and hair. In support of this idea, the mammalian EB1-binding MlphIP1 motif, unlike Rab27a- and MyoVa-binding sequences, is absent from other known vertebrate Mlph proteins (Fig. 1C) and the related protein MyRIP, which plays a role in melanosome transport in the eye. Finally, work in other cell types indicates that EB1 and dynein/dynactin anchor MTs at sites of cadherin-mediated cell-cell contact (Shaw et al., 2007). Interestingly, E-cadherin an important mediator of melanocyte-keratinocyte interaction is thought to influence melanosome transfer between these two cell types (Van Den Bossche et al., 2006). Our future studies should address these issues.
Materials and Methods
Chemicals
Reagents used in this study were supplied by either Sigma-Aldrich (Poole, UK) or Invitrogen (Paisley, UK), unless otherwise indicated.
Culture and transfection of immortal melanocyte cells lines
Cultures of primary dilute melanocytes together with immortal melan-ink and melan-ln were derived and maintained as described previously (Ali et al., 2004; Hume et al., 2001; Hume et al., 2006). Briefly, cells were maintained in RPMI 1640 supplemented by addition of FBS to 10%, PMA to 200 nmol and cholera toxin to 200 pmol. Cells for transfection were seeded onto 13-mm coverslips at 1×104 cells/coverslip. For immunoblot assay of siRNA-mediated protein depletion in populations of cell transfections were scaled up fivefold. Twenty-four hours later, cells were transfected with plasmid DNA (0.5 mg/coverslip) or siRNA oligonucleotide (20 pmol/coverslip) using Fugene6 (2 ml/coverslip) (Roche, Lewes, UK) or Oligofectamine (1 ml/coverslip) liposomal transfection reagents, respectively, in Optimem serum-free medium in accordance with the manufacturers recommendation. Three to 6 hours later transfection complexes were removed and replaced by full melanocyte growth medium. Transfected cells were fixed 48 or 72 hours later by immersion in either 3% paraformaldehyde in 1×PBS for 15 minutes at room temperature as described previously (Hume et al., 2001) or 100% methanol at –20°C for 8 minutes for cells to be stained with anti-EB1 antibodies. For sequential transfection using siRNA oligonucleotide and plasmid DNA, cells were first siRNA transfected, plasmid transfected 24 hours later and fixed 48 hours later.
Yeast two-hybrid binding assay and melanocyte library screening
All yeast media were described in Rose et al. (Rose et al., 1990). L40 strain genotype is MATa, trp1, leu2, his3, LYS2::lexA-HIS3 URA3::lexA-lacz (generous gift of J. Camonis, Institut Curie). Yeast were transformed by lithium acetate procedure according to Schiestl and Gietz (Schiestl and Gietz, 1989). Yeast L40 strain was transformed with combinations of pBTM and pGAD constructs and grown for 2 days on standard drop-out medium plates lacking tryptophan and leucine and colonies were streaked out in patches, grown for 2 days and were assayed for β-galactosidase activity as described in Durfee et al. (Durfee et al., 1993). Murine melanocyte cDNA library was prepared by ligation of wild-type melanocyte (melan-a) derived random primed cDNA into EcoRI-digested pGAD-C1 (GATC Biotech, Konstanz, Germany). The mouse melanocyte library was screened with bait construct pBTM-mouseMlph367-590aa as described in Fromont-Racine et al. (Fromont-Racine et al., 1997).
Plasmid constructs and siRNA oligonucleotides
The construction of pBTM-Mlph 1-590, pBTM-Mlph 367-590, pGADC3-MyoVa MSGTA, pGADC3-Rab27a, pCMVMYC-Mlph1-266, pGEX4T1-Mlph ΔRBD, pCS2-6xmyc-Mlph, pCS2-6xmyc-Slp1 and pMalC2X MyoVa-MSGTA was previously described (Hume et al., 2006; Strom et al., 2002). Plasmid constructs containing mouse Mlph or EB1-coding sequence were created by PCR amplification of IMAGE clones 4862487 and 5057090, respectively, using the following primers: EB1 sense1 5′-CCGGAATTCAGATCTATGGACATGCTCTTCCCTG-3′; EB1 antisense268 5′-GCCGCTCGAGTTAATACTCTTCTTGTTCCTC-3′; EB1 antisense134 5′-GCGCCTCGAGACCTTGTCTGGCAGCTAC-3′; EB1 antisense171 5′-GCGGGAATTCAGATCTGCAGCTCCTAAGGCTGGC-3′; EB1 antisense204 5′-GCGGGAATTCAGATCTCTGAAGCTTACTGTTGAAG-3′; EB1 antsense268-stop 5′-GCGCCTCGAGATAAATACTCTTCTTGTTCCTCC-3′; Mlph sense481 5′-GGCCGAATTCAGAGCCGCAGGACTC-3′; Mlph antisense590 5′-CGGAATTCCTCGAGTTAGGGCTGCTGGGCCATCAC-3′; Mlph sense367 5′-GGGAATTCGGATCCGCCCAGGAACCAACTGTG-3′; Mlph antisense485 5′-CCCTCGAGGGCTCTCAGAGCTGCGATCCTG-3′. PCR products were then sub-cloned to the appropriate vector using the restriction endonucleases EcoRI and Xho1 for insert DNA and EcoRI and SalI for vector DNA. pBTM-Mlph 481-520 was constructed by PstI digestion of pBTM-Mlph 481-590 followed by gel purification of the 5.7 kb Mlph/vector fragment and religation. pBTM-Mlph 520-590 was constructed by the ligation of the 0.21 kb Mlph fragment released from the above PstI digestion into pBTM prepared using with the same restriction enzyme and dephosphorylated using calf intestinal phosphatase. pBTM-Mlph IP1, pBTM-Mlph IP2 and pCS2-6xmyc-MlphR35W were prepared by the Quikchange mutagenesis method (Stratagene) using primer pairs senseIP1 5′-AGAAAGTCAGGCGCCGCGATCTTTCTTCCC-3′ and antisenseIP1 5′-GGGAAGAAAGATCGCGGCGCCTGACTTTCT-3′, senseIP2 5′-AAACTTGACAGGGCCGCAAAGACTCCACCT-3′ and antisenseIP1 5′-AGGTGGAGTCTTTGCGGCCCTGTCAAGTTT-3′, and senseR35W 5′-AGGCGAGAGGAAGAATGGCTCCAGGGGCTGAAG-3′ and antisenseR35W 5′-CTTCAGCCCCTGGAGCCATTCTTCCTCTCGCCT-3′, respectively. Plasmids pGEX4T1-ΔRBDIP, pCS2-6xmyc-MlphIP1 pBTMMlphR35W, pBTMMlphEA and pBTMMlphAP were constructed by PCR using 5′ primers sense1 5′-GGGAATTCGGATCCGCCCAGGAACCAACTGTG-3′ and sense150 5′-GGGAATTCGGATCCTCTGAGCCAAGCTTGGAAG-3′, respectively and antisense590 (see above) and pBTM-Mlph IP1, pCS2-6xmyc-MlphR35W, pCS2-6xmyc-MlphEA and pCS2-6xmyc-MlphAP as template. PCR products were then digested using restriction enzymes EcoRI and XhoI and ligated together with vector digested with restriction enzymes EcoRI and XhoI or SalI. Baculoviruses allowing expression of full length mouse Mlph and the C-terminal tail of mouse MyoVa (MSGTA 1258-1853aa containing the melanocyte-specific exons D and F) in Sf9 cells were prepared by sub-cloning of PCR products encoding Mlph and MSGTA into pFastBacHTb using restriction enzymes EcoRI and XhoI followed by transformation of recombinant plasmids into DH10Bac cells allowing transposition of the recombinant expression cassette into the baculovirus shuttle vector and recombinant bacmid vectors were then transfected into Sf9. Recombinant baculovirus was harvested from the Sf9 culture medium and amplified by several rounds of infection of fresh Sf9 cells.
siRNA oligonucleotides
All siRNA oligonucleotides used in this study including siControl non-targetted oligonucleotides were supplied by Dharmacon (Chicago, IL) (Table 3).
siRNA oligonucleotides
In vitro assay of interaction of purified Mlph with EB1 and MyoVa-MSGTA
MBP-tagged and GST-tagged proteins were expressed in the BL21-codon plus (DE3)-RIPL bacterial strain and his6-tagged proteins were expressed in Sf9 insect cells and purified by affinity chromatography using amylose-agarose (NEB, Hitchin, UK) and Ni-Sepharose (Amersham, Little Chalfont, UK) as recommended by the manufacturers. GST-Mlph, 100 pmol, was mixed with either 100 pmol MBP-EB1 or MBP-MyoVaMSGTA in buffer A (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 5 mM MgCl2, 1 mM DTT and 0.075 mM Nonidet P-40) and incubated for 30 minutes at RT with gentle agitation. Equilibrated glutathione-Sepharose beads (20 μl) were added to each reaction and incubated for 30 minutes at room temperature with gentle agitation. The glutathione-Sepharose beads were precipitated by centrifugation and washed twice with 1 ml buffer A. Bound proteins were eluted by boiling the beads in SDS loading buffer, and the eluates were analysed by immunoblotting using antibodies specific for GST and MBP. Signals were analysed using a Fujifilm LAS3000 (Tokyo, Japan) and quantified using Aida software (Aquasan, Alpharetta, GA). To assay simultaneous binding of Mlph with MyoVa-MSGTA and EB1, 100 pmol MBP-EB1 was incubated with 100 pmol his6-MyoVa-MSGTA and/or 100 pmol his6-Mlph in buffer A for 30 minutes at room temperature with gentle agitation. MBP-EB1 was precipitated using 20 μl amylose-Agarose beads in the same manner as described for glutathione-Sepharose beads. Bound proteins were eluted by boiling beads in SDS loading buffer and analysed by immunoblotting using antibodies specific for MBP, Mlph and MyoVa-MSGTA.
Immunofluorescence, immunoblotting and antibodies
Immunofluorescence and immunoblotting techniques used in this study were as described previously (Hume et al., 2006). Cell lysates were prepared for Western blotting by resuspension of cell pellets in an appropriate volume of lysis buffer (150 mM NaCl, 20 mM Tris-HCl pH 7.5, 1 mM DTT, 0.1% CHAPS and 1xPI cocktail) followed by incubation on ice for 15 min. Nuclei and other debris were then harvested by centrifugation (800 g at 4°C for 10 min) and the PNS was mixed with sample buffer before resolution using SDS-PAGE. Antibodies used for immunoblotting were; goat anti-Mlph (EB05444 1:500 Everest Biotechnology, Upper Heyford, UK), mouse anti-EB1 (# 610534 1:500 BD Transduction labs, Oxford,UK) and mouse anti-actin (Sigma clone AC40 1:1000). Quantification of chemiluminescence signal was performed using the LAS3000 Intelligent Darkroom (Fujifilm) and AIDA image analysis software. Antibodies used for immunofluorescence were mouse anti-EB1 1:100 and rabbit anti-myc epitope (# 06-549 1:200 Upstate, Dundee, UK). Image acquisition, and quantification of melanosome transport were as previously described (Hume et al., 2006).
Total internal reflectance fluorescence (TIRF) microscopy and particle tracking analysis
A Zeiss Axiovert 200 inverted microscope modified for objective-type total internal reflection fluorescence microscopy (Till Photonics, Munich, Germany) was used. An argon laser 488 nm line was used to excite EGFP fluorescence through a 100× 1.4 NA phase contrast Zeiss Planapochromat objective. During observation the cells were kept on a MS2000 stage (Applied Scientific Instruments, Eugene, OR) equipped with a Focht chamber system (FCS2, Bioptechs, Butler, PA) at 37°C. The images were recorded by a PCO SensiCam CCD camera at a rate of 1 frame per second. A Uniblitz shutter VS25 (Vincent Associates, Rochester, NY) in the transmitted light path was used to allow sequential acquisition of phase contrast images of melanosomes and fluorescent images of EGFP. The scale was 16 pixels per micrometer.
ImageJ (NIH) and Volocity 4 (Improvision, Coventry, UK) software was used for image processing and particle tracking analysis. In order to decrease background fluorescence, the last image in the sequence was subtracted from each image in a stack. Fluorescent particles were selected by intensity threshold and size, smoothed by Gaussian filter and tracked using a limited trajectory shortest path algorithm associated with Volocity.
Acknowledgments
We thank Martin Spitaler for help with Volocity 4.0 image analysis software, Rudi Baron for help with Sf9 maintenance and purification of recombinant protein and other members of the Seabra lab for helpful discussions. This work was supported by the Wellcome Trust and the BBSRC. There are no conflicts of interest in this study.
Footnotes
Supplementary material available online at http://jcs.biologists.org/cgi/content/full/120/17/3111/DC1
- Accepted July 5, 2007.
- © The Company of Biologists Limited 2007
References
Article navigation
Related articles
Cited by...
More in this TOC section
Similar articles
Other journals from The Company of Biologists
Introducing FocalPlane’s new Community Manager, Esperanza Agullo-Pascual
We are pleased to welcome Esperanza to the Journal of Cell Science team. The new Community Manager for FocalPlane, Esperanza is joining us from the Microscopy Core at Mount Sinai School of Medicine. Find out more about Esperanza in her introductory post over on FocalPlane.
New funding scheme supports sustainable events
As part of our Sustainable Conferencing Initiative, we are pleased to announce funding for organisers that seek to reduce the environmental footprint of their event. The next deadline to apply for a Scientific Meeting grant is 26 March 2021.
Read & Publish participation continues to grow
"Alongside pre-printing for early documentation of work, such mechanisms are particularly helpful for early-career researchers like me.”
Dr Chris MacDonald (University of York) shares his experience of publishing Open Access as part of our growing Read & Publish initiative. We now have over 150 institutions in 15 countries and four library consortia taking part – find out more and view our full list of participating institutions.
Cell scientist to watch: Romain Levayer
In an interview, Romain Levayer talks about starting his own lab, his love for preprints and his experience of balancing parenting with his research goals.
Live lactating mammary tissue
In a stunning video, Stewart et al. demonstrate warping of the alveolar unit due to basal cell-generated force as part of their recent work investigating roles for mechanically activated ion channels in lactation and involution.
Visit our YouTube channel to watch more videos from JCS, our sister journals and the Company.
JCS and COVID-19
For more information on measures Journal of Cell Science is taking to support the community during the COVID-19 pandemic, please see here.
If you have any questions or concerns, please do not hestiate to contact the Editorial Office.