SEARCH

SEARCH BY CITATION

Keywords:

  • cytotoxic T lymphocyte;
  • Griscelli syndrome;
  • leaden;
  • melanosome;
  • Rab27a

Abstract

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

The function of lysosome-related organelles such as melanosomes in melanocytes, and lytic granules in cytotoxic T lymphocytes is disrupted in Griscelli syndrome and related diseases. Griscelli syndrome results from loss of function mutations in either the RAB27A (type 1 Griscelli syndrome) or MYO5A (type 2 Griscelli syndrome) genes. Melanocytes from Griscelli syndrome patients and respective murine models ashen (Rab27a mutant), dilute (myosin Va mutant), and leaden exhibit perinuclear clustering of melanosomes. Recent work suggests that Rab27a is required to recruit myosin Va to melanosomes, thereby tethering melanosomes to the peripheral actin network and promoting melanosome retention at the tips of melanocytic dendrites. Here, we characterize the function of the leaden gene product. We show that Rab27a, but not myosin Va, can be localized to melanosomes in leaden melanocytes, suggesting that the leaden gene product acts downstream of, or in parallel to, Rab27a in melanocytes to promote recruitment of myosin Va to melanosomes. We also observed reduced levels of myosin Va protein in leaden and ashen melanocytes, suggesting that myosin Va stability is influenced by the leaden and ashen gene products. In leaden cytotoxic T lymphocytes, we observed that lytic granules polarize towards the immunological synapse and kill target cells normally. However, in contrast to melanocytes, we found that neither the leaden gene product (melanophilin) nor myosin Va was detectable in cytotoxic T lymphocytes. These results suggest that Rab27a interacts with different classes of effector proteins in melanocytes and cytotoxic T lymphocytes.

Lysosome-related organelles serve a variety of specialized functions in differentiated cell types (reviewed in 1–3). Examples of lysosome-related organelles include melanosomes in melanocytes, lytic granules in cytotoxic T lymphocytes (CTLs) and dense granules in platelets. These organelles are considered to be lysosome related due to their acidic pH and the presence of lysosome markers such as lysosome-associated membrane protein (LAMP) family members. However, each also contains a unique sets of proteins important for their specialized function, e.g. tyrosinase in melanosomes and perforin in lytic granules.

Melanosomes are sites of pigment production and storage, and reside within melanocytes and retinal pigmented epithelial cells. Melanosome biogenesis is a poorly understood process that is divided into four stages according to morphological criteria [reviewed in (3)]. In mammalian skin, mature melanosomes are transported from the Golgi region to the tips of melanocyte dendrites and then transferred to adjacent keratinocytes or growing hair shafts at the hair bulb [reviewed in (4)]. This process is essential for normal pigmentation and seems to be achieved by coupling long-range bi-directional microtubule-based transport, which delivers melanosomes to the tips of dendrites, with an actin-based retention system that prevents the microtubular system from returning the melanosomes back towards the cell body (3,5,6).

CTLs are important for the destruction of tumorigenic and virally infected cells. In common with other cells of hemopoietic lineage, CTLs identify target cells and secrete soluble proteins that promote target cell lysis and apoptosis such as granzymes, and cell surface exposure of proteins, which trigger target cell apoptosis such as Fas ligand. These lytic components are stored in secretory lysosome-related granules (lytic granules) whose synthesis is triggered by target major histocompatibility complex (MHC) recognition through the T-cell receptor (TCR) [reviewed in (7)]. Conjugation of CTL with a target cell results in the formation of a highly organized three-dimensional structure known as the immunological synapse (8). Lytic granules then undergo kinesin-driven microtubule-dependent transport towards the synapse where they release their contents [reviewed in (9)].

Defects in the function of lysosome-related organelles underlie several human genetic diseases such as Chediak–Higashi syndrome (CHS), Hermansky–Pudlak syndrome (HPS) and Griscelli syndrome (GS) [reviewed in (3,10)]. Each disease affects the function of melanosomes, lytic granules, platelet dense granules and other lysosome-related organelles, to a greater or lesser extent. Recently, several of the genes mu-tated in these diseases, and respective mouse models, have been identified and this has proved insightful in terms of understanding both the molecular mechanisms underlying disease symptoms and the biology of these organelles (3, 10).

GS patients display partial albinism and immunodeficiency (11). Mutation of either of two genes causes GS: RAB27A mutations are most common and result in type 1 (typical) GS, while MYO5A (encoding myosin Va) mutations are less frequent and result in a disease involving primary neurological rather than immunological defects (type 2 GS) (12). Three naturally occurring mouse mutations are models of GS: ashen (Rab27a ash) encoding Rab27a, dilute (Myo5a d) encoding myosin Va, and leaden (Mlph ln) encoding melanophilin (13–15).

Rab27a is a member of the Rab family of small GTPases, which are important regulators of membrane transport [reviewed in (16–18)]. CTLs from GS patients and ashen mice are unable to kill target cells due to defects in the release of lytic granule contents into the immunological synapse, implicating Rab27a as a regulator of a late step in granule secretion (12,19,20). Mutations in myosin Va do not affect CTL function (12,19). Melanocytes from GS patients and ashen mice exhibit perinuclear clustering of melanosomes (15,21–23). The two other GS mouse model mutants, dilute and leaden, exhibit a similar phenotype, suggesting that the products of all three genes are involved in mediating the peripheral actin-based retention of melanosomes. This hypothesis is supported by the findings that the coat color dilution in all three mutants is suppressed, albeit to differing extents, by the dilute suppressor mutation (24) and that the dilute gene product (myosin Va) is present with Rab27a in immune complexes precipitated from melanocyte extracts (22). At present, there is no evidence to indicate whether the interaction of myosin Va and Rab27a is direct. Given the genetic evidence, one possibility is that the leaden gene product mediates the interaction of Rab27a and myosin Va.

In this work, we address the function of the leaden gene in melanocytes and CTLs. Our findings indicate that leaden may act either together with, or as an effector of Rab27a to recruit myosin Va to melanosomes while its activity is not critical to CTL function, suggesting that Rab27a operates via different effectors in different cell systems.

Results

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

Rab27a, but not myosin Va, is present on melanosomes in cultured leaden melanocytes

The leaden gene product is likely to function together with Rab27a and myosin Va in melanosome transport. As a first insight into the function of the leaden gene product, we decided to investigate intracellular localization of Rab27a and myosin Va in leaden melanocytes. First, we employed immunofluorescence microscopy to determine the overall distribution of Rab27a and myosin Va within primary leaden melanocytes and the extent of their colocalization with melanosomes. We observed that the majority of melanosomes in leaden melanocytes cluster close to the nucleus, as observed previously (25) and similar to ashen and dilute melanocytes (22, 23) (Figure 1E,H,N,Q), rather than being evenly distributed throughout the peripheral cytoplasm as observed for wild-type melan-a cells (Figure 1A). The fact that the clustering of melanosomes was not absolute suggests that melanosomes are able to undergo microtubule-dependent movement towards the edge of melanocytes but are not then retained (6).

image

Figure 1. Localization of Rab27a and myosin V in melanocytes by immunofluorescence. Coverslip-grown melan-a cells (Panels A–D) or primary melanocytes (Panels E–S) were fixed and stained with the indicated antibodies. Panels A, E, H, K, N and Q are transmission images showing the distribution of pigment granules in each cell type. Panels A–D: Endogenous myosin V in fixed, coverslip grown, melan-a cells was stained using polyclonal anti-myosin V antibodies followed by Alexa 488-conjugated goat anti-rabbit antibodies (Panel B)(see Materials and Methods and supplementary data for Figure 1). Cells were then stained with anti-Rab27a polyclonal antibodies and detected using Alexa 568-conjugated goat anti-rabbit antibodies (Panel C). The merged fluorescent signal for Rab27a and myosin V in melan-a cells is shown in panel D. Panels E–S: Coverslip-grown primary cultures of melanocytes derived from leaden and ashen mutant mice were fixed and stained using TRITC-conjugated phalloidin (G, J, M, P, S)and either polyclonal anti-Rab27a antibodies (F, I) or polyclonal anti-myosin V antibodies (L, O, R), as described in Materials and Methods, detected using Alexa 488-conjugated goat anti-rabbit secondary antibodies. Staining of filamentous actin using TRITC-labeled phalloidin emphasizes the extent of the cytoplasmic extension of leaden and ashen melanocytes in which pigment granules are clustered. Bar = 20 μm.

Download figure to PowerPoint

Staining of leaden melanocytes with anti-Rab27a antibodies revealed a close correspondence between the distribution of Rab27a and pigment-containing melanosomes (compare panels E/H with F/I in Figure 1). We observed that the most intense Rab27a signal is present close to the nucleus, correlating well with the mass of melanosomes accumulated in this area. In addition, weaker Rab27a immuno-reactivity is present in some peripheral regions of the cytoplasm. Examination of these peripheral areas using higher magnification revealed that Rab27a is associated with punctate structures, approximately 0.5 μm in diameter, whose distribution correlates well with pigment-containing melanosomes (compare insets in panels E/H with F/I in Figure 1). In melan-a cells, the distribution of Rab27a also correlates well with melanosomes, as observed previously (compare Figure 1A with C) (22, 23). These observations suggest that the leaden gene product is not required for the targeting of Rab27a in melanocytes as Rab27a is appropriately localized to melanosomes in leaden melanocytes.

We then stained melanocytes using anti-myosin Va antibodies. In contrast to melan-a melanocytes in which myosin Va colocalizes well with melanosomes (compare Figure 1A,B and insets) (22,23), myosin Va does not strongly colocalize with pigmented melanosomes in leaden melanocytes (compare panels N/Q with O/R in Figure 1). Rather, we observed a pattern of myosin Va distribution that is similar to that seen in ashen melanocytes, which lack Rab27a (compare Figure 1L with 1O or 1R) (22). We noticed that myosin Va shows some accumulation in a bright pericentriolar spot. Co-staining with anti-α-tubulin antibodies suggested that this may correspond to the microtubule organizing center (data not shown). Furthermore, phalloidin staining indicates that filamentous actin is also abundant in this region (Figure 1G,J,P,S). In wild-type melan-a and melan-c melanocytes we occasionally observed a similar spot-like distribution of myosin Va (data not shown). However, this distribution pattern is generally less apparent than in either leaden or ashen melanocytes, as the majority of myosin Va in these cells is associated with melanosomes. The remaining myosin Va in leaden melanocytes is evenly distributed throughout the cytoplasm (Figure 1O,R). These data suggest that the leaden gene product is required for the targeting of myosin Va to melanosomes in melanocytes.

To examine the localization of Rab27a and myosin Va in greater detail, we performed immunoelectron microscopy of leaden and wild-type melan-a melanocytes using thin sections and whole-mount preparations. The results of these experiments are wholly consistent with observations made using light microscopy. In thin sections, we found that Rab27a is clustered on the perimeter membrane of melanosomes in both melan-a and leaden melanocytes (Figure 2A,B). In contrast, the quantity and distribution of myosin Va labeling differs between the two cell types. In melan-a melanocytes, myosin Va is distributed in clusters on the perimeter membrane of melanosomes as well as on membranes associated with the cytoskeleton (Figure 2C). These clusters have previously been shown to partially colocalize with clusters of Rab27a (22). However, in leaden melanocytes, myosin Va is localized mainly next to cytoplasmic filaments consistent with the cytoskeleton rather than on the perimeter membranes of melanosomes (Figure 2D).

image

Figure 2. Localization of Rab27a and myosin V in melan-a and leaden melanocytes by immunoelectron microscopy. melan-a (Panel A)and leaden (Panel B)melanocytes were digitonin-permeabilized and labeled for Rab27a (10 nm gold). Rab27a was clustered on the perimeter membrane of melanosomes in both cell types (examples circled). Myosin V was detected in melan-a (Panel C)and leaden (Panel D)melanocytes using whole mounts to highlight the interaction of melanosomes with the cytoskeleton. Myosin V (15 nm gold) was distributed in clusters on the perimeter membrane of melanosomes (arrows) and on the cytoskeleton in melan-a melanocytes. In leaden melanocytes, myosin V labeling was sparse and was associated with the cytoskeleton rather than melanosomes. Bar = 0.2 μm.

Download figure to PowerPoint

Altogether, these data suggest that the leaden gene product is not required for targeting of Rab27a to the cytoplasmic face of melanosome membranes but instead plays a role in concert with Rab27a in mediating the association of myosin Va with melanosomes. Therefore pigment dilution in leaden mice, as in ashen and dilute, is most likely due to lack of melanosome-associated myosin Va.

The amount of myosin Va protein is reduced in leaden melanocytes

While we were conducting immunofluorescence and immunoelectron microscopy using anti-myosin Va antibodies, we observed that the level of myosin Va signal appeared lower in leaden melanocytes than in wild-type (Figures 1 and 2), suggesting that myosin Va protein levels are reduced in these cells. We thus investigated the scale and specificity of any reduction in myosin Va protein levels. Cell extracts were obtained and separated into pellet and supernatant fractions by high-speed centrifugation, and then probed using specific antibodies. Equal levels of Rab27a are present in P100 fractions of melan-a (wild type) and leaden melanocytes but, as expected, Rab27a is absent in ashen melanocytes (Figure 3A, bottom panel). Probing with anti-myosin Va antibodies indicated that myosin Va is expressed in melan-a, ashen and leaden melanocytes (Figure 3A, top panel). However, we consistently observed that both ashen and leaden mutant melanocytes express lower levels of myosin Va than wild-type melan-a cells. Probing of the same melanocyte extracts using antibodies reactive to the ER chaperone calnexin indicated that this reduction in protein level is relatively specific to myosin Va (Figure 3A, middle panel). We next asked whether the observed reduction in myosin Va protein level was evident in other cell types. To address this issue, we studied myosin Va protein levels in brain tissue obtained from the different mutant and control mouse strains, as previous work has documented that myosin Va is highly expressed in the brain (13). Immunoblotting analysis indicated that brain myosin Va is equally expressed in ashen, leaden and in heterozygous littermate controls (Figure 3B upper panel). Thus, the reduction in myosin Va protein level appears to be restricted to melanocytes, suggesting that Rab27a and the leaden gene product play a role in stabilizing the myosin Va protein in melanocytes.

image

Figure 3. Expression of Rab27a and myosin Va in leaden. Lysates prepared from melanocytes were separated into S100 (S) and P100 (P) fractions (50 μg per well) (Panel A), or brain, total postnuclear supernatant (25 μg per well) (Panel B), were subjected to immunoblot analysis, as described under Materials and Methods, using the indicated antibodies. Calnexin immunoblotting is used as a loading control to confirm equal gel loading of each lane.

Download figure to PowerPoint

Cytotoxic T lymphocyte function is not affected in leaden mice

Griscelli syndrome affects primarily melanocytes and CTLs. Previously, we and others found that Rab27a is an important regulator of lytic granule secretion in CTLs (12,19,20). Given that leaden is a model of GS and appears to act with Rab27a in melanocytes, we examined whether the leaden gene product regulates CTL activity.

We derived CTLs from homozygous leaden, heterozygous control and wild-type C57BL/6 mice by stimulation withBalb/c spleens and assessed the lytic ability of the derived T cells against the P815 target cell line, as these cells express the same class I major histocompatibility complex as the Balb/c strain (H2d). Five days post-activation, we tested lytic activity of CTLs against P815 target cells and found that killing activity of leaden CTLs is similar to that of wild-type and heterozygous leaden at various effector/target ratios (Figure 4A). This study indicates that the leaden gene product is not required for killing of target cells by CTLs, unlike ashen (Rab27a) (19,20). We next analyzed the CTL activity morphologically. We used antibodies against granzyme A, one of the contents of lytic granules, to examine the position of granules and against talin, which accumulates at sites of cell–cell contact, to monitor the formation of the immunological synapse. In leaden conjugates, as in control CTLs, we observed that talin forms a ring at the site of cell–cell contact and that all of the lytic granules are focused tightly in the center of this ring (Figure 4B). We conclude that granule polarization and secretion is unaffected by the leaden mutation, suggesting that the leaden gene product is not important for CTL function.

image

Figure 4. CTL function in leaden. Panel A:Graph of the percentage of target cell lysis (Y-axis) at various ratios of effector cells to a fixed number of target cells (X-axis) for wild type C57/BL6 (filled squares), heterozygous leaden (+/ln, filled circles) and homozygous leaden(ln/ln, open circles) mice. Panel B:Allogeneic CTLs conjugated with P815 target cells from C57BL/6 (a),+/ln (b) and ln/ln (c) mice were stained for granzyme A (green), talin (red) and actin (blue). Scale bar = 10 μm

Download figure to PowerPoint

Melanophilin and myosin Va are expressed in melanocytes but not CTLs

The above findings regarding the effect of the leaden mutation upon melanosome transport and CTL function are comparable with observations of dilute (myosin Va mutant) mice, which exhibit normal killing of target cells by CTLs and natural killer cell (NKs) in spite of melanosome transport defects indistinguishable from ashen (19). As Rab27a protein is detectable in both melanocytes (Figure 1) and CTLs by immunoblotting [Figure 1 of Stinchcombe et al. 2001 (20)], these data suggest the interesting possibility that Rab27a exerts its function on lytic granules and melanosomes by recruiting a distinct set of effectors in each cell type.

We therefore studied the expression of myosin Va and Melanophilin in both melanocytes and CTLs. We were able to detect myosin Va in melanocytes (melan-a cells) and brain [as previously described (13)] but not in CTLs (Figure 5A). The recent mapping of the leaden mutation to exon 2 of the Mlph locus (14), encoding the Melanophilin protein, allowed us to test whether Melanophilin mRNA is expressed in melanocytes and CTLs using reverse transcriptase–polymerase chain reaction (RT-PCR). We found that fragments corresponding to Melanophilin transcripts could be amplified from RNA extracts prepared from MM96 human melanoma-derived cells but not from extracts derived from human CTLs (Figure 5B). As controls, we amplified fragments corresponding to Rab27a and hypoxanthine phosphoribosyl transferase (HPRT) and observed their presence in both melanocytes and CTLs (Figure 5B). These experiments suggest that Rab27a is expressed in both melanocytes and CTLs, while Melanophilin and myosin Va are present only in melanocytes.

image

Figure 5. Expression of Rab27a, Melanophilin and myosin Va in melanocytes and CTLs. Panel A:Total postnuclear supernatant from CTLs (30 μg), melan-a (30 μg) and brain (10 μg) were subjected to immunoblot analysis, as described under Materials and Methods, using the indicated antibodies. Calnexin immunoblotting was used as a loading control to confirm equal gel loading of each lane. Panel B:Primers specific for human Rab27a, Melanophilin and HPRT were used to amplify 389-bp, 520-bp and 350-bp fragments, respectively, from reverse-transcribed RNA prepared from the indicated cell types or plasmid controls, as described in Materials and Methods. PCR products were analyzed by ethidium bromide staining of agarose gel. Numbers on the left-hand side indicate the position of molecular weight standards, and arrows on the right-hand side indicate the identity of the product. HPRT was used as an internal control. PCR products were not detected in reactions performed in the absence of reverse transcriptase.

Download figure to PowerPoint

Discussion

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

We present evidence suggesting that the leaden gene product plays an important role in melanosome transport in skin melanocytes but not in CTL lytic granule secretion. Our data are consistent with the possibility that leaden acts downstream or in parallel with Rab27a to promote the recruitment of myosin Va to melanosomes, thereby tethering melanosomes to actin filaments. As Rab27a, but not leaden or myosin Va, is important for lytic granule secretion in CTLs, our data further suggest that Rab27a acts through different effector proteins in these different cell types.

We first examined the distribution of Rab27a and myosin Va in primary cultures of leaden melanocytes by immunofluorescence and immunoelectron microscopy (Figures 1 and 2). Using both methods, we observed that Rab27a, but not myosin Va, could be localized to the surfaces of melanosomes in these cells. These observations argue against the idea that the leaden gene product acts upstream of Rab27a as a targeting receptor or GDP/GTP exchange factor that would recruit and stabilize Rab27a at the cytoplasmic leaflet of melanosomes, as this model predicted that Rab27a would be unable to associate with leaden melanosomes. Also, our data do not support a function for leaden downstream of myosin Va allowing tethering of melanosomes to the peripheral actin cytoskeleton, as this model predicted that myosin Va would associate with leaden melanosomes. Instead, the present data suggest that leaden gene product is likely to act downstream of (or in parallel to) Rab27a, promoting the recruitment of myosin Va with melanosomes. Although we have previously reported the presence of myosin Va in complexes precipitated from melanocytes using anti-Rab27a antibodies (22), suggesting that the two proteins interact, there is presently no evidence to indicate whether this interaction is direct. The findings presented here suggest that the interaction may be indirect, possibly requiring the leaden gene product. Alternatively, myosin Va may require simultaneous interaction with multiple components of a complex present on melanosomes which contains Rab27a and the leaden gene product. The absence of any one of these components may result in the failure of myosin Va to associate with melanosomes.

This model for the function of the leaden gene product is supported by the recent mapping of the leaden mutation to exon 2 of the Mlph locus (14). This locus encodes melanophilin, a protein whose N terminus shows sequence homology to the N terminus ‘Rab binding region’ of the Rab3a effector, Rabphilin-3 (26). The leaden mutation is thought to result in production of a truncated protein lacking 7 amino acids in the N-terminal region. Although this protein has yet to be biochemically characterized, it seems likely that melanophilin binds Rab27a and functions as a Rab27a effector. Taken together, our work and that of Jenkins and coworkers suggests a model where the leaden gene product (melanophilin) is recruited by activated Rab27a to melanosomes, and in turn promotes the recruitment of myosin Va to melanosomes. These predictions should now be tested experimentally.

Another novel observation reported here is a general reduction in the intensity of myosin Va staining in leaden and ashen melanocytes (Figures 1 and 3). The residual myosin Va staining pattern is redistributed from melanosomes to a diffuse pattern throughout the cytoplasm and/or associated with a spot in the center of the cell which contains actin and from which microtubules emanate (Figure 1 and data not shown) (22). This observation is interesting in the light of findings that myosin Va undergoes tissue-specific alternative splicing within the C-terminal tail domain. This region is the most divergent between different members of the myosin superfamily and is thought to contain targeting information to associate with diverse cellular structures (27). Splicing of myosin Va transcripts in mouse and man is predicted to generate several isoforms of the protein (28,29). One isoform is highly expressed in the brain (here referred to as brain isoform), while a larger isoform containing additional exons is expressed in various tissues including melanocytes, but not in brain (here referred to as melanocyte isoform). The melanocyte isoform tail domain localizes to melanosomes when expressed as GFP fusion protein in melanocytes (6), but the GFP-tagged tail domain of the brain isoform is observed to concentrate in a pericentriolar dot consistent with the microtubule-organizing center and also throughout the cytoplasm in melanocytes (30). This pattern of localization is similar to the pattern of myosin Va distribution we observed in leaden melanocytes (Figure 1). Thus one possibility is that the residual staining observed in leaden and ashen melanocytes might result from the persistence of only the brain isoform due to preferential down-regulation or instability of the melanocyte isoform in these cells. Consistently, we observed that the leaden and ashen mutations do not affect the level of expression of myosin Va in murine brain tissue where the melanocyte specific isoform is not expressed (Figure 3B). However, we were unable to detect expression of the brain-specific isoform of myosin Va in either melan-a or leaden melanocytes using RT-PCR (data not shown). While the mechanism underlying myosin Va down-regulation is unclear, it is interesting to note that myosin Va interacts with a RING finger containing protein named BERP (31). There is evidence that such proteins act as ubiquitin ligases allowing the specific ubiquitination and degradation of the protein with which they associate [reviewed in (32)]. Thus, one possibility is that the absence of Rab27a or melanophilin promotes the interaction of myosin Va with BERP and its degradation, or alternatively that interaction with Rab27a and melanophilin is primarily required to stabilize myosin Va on melanosomes rather than for its initial targeting to melanosomes. The suggestion that the interaction of Rabs with their effectors might stabilize them is not without precedent. For instance, rabphilin-3 levels are significantly reduced in brains of Rab3a knockout mice (33). Future studies should question whether down-regulation results from defects in synthesis or stability of the myosin Va RNA/protein.

We also studied CTL function, which is dramatically affected by mutations in the RAB27A gene in type 1 GS patients and ashen mice. Interestingly, we found that CTL function is not affected in leaden mice as we observed normal CTL killing activity and lytic granule polarization in leaden CTLs. These observations parallel reports of normal CTL function in type 2 GS patients and dilute mouse (12,19). Furthermore, our inability to detect the expression of myosin Va or melanophilin in CTLs (Figure 5), suggests that these proteins are either not expressed at all or that the small amounts expressed are not physiologically relevant for lytic granule secretion.

Together, these findings raise the possibility that Rab27a acts through different groups of effectors in different cell types. In this model, Rab27a is envisioned to act in melanosome transport in melanocytes through the recruitment of Melanophilin and myosin Va, whilst in CTLs Rab27a would function in lytic granule secretion through the recruitment of different effector proteins. In accord with this idea, Melanophilin (also called Slac2-a for synaptotagmin-like protein lacking C2 domains) belongs to a family that includes several synaptotagmin-like proteins (Slp) and granuphilins (34,35). These proteins share a common synaptotagmin-like protein homology domain (SHD) structure at their N termini, which is related to the Rab-binding region of Rabphilin-3 and could preferentially mediate the function of Rab27a in different cell types (26, 36). Conversely, we suggest that myosin Va may interact with proteins other than Rab27a to fulfill its function in the brain. This possibility is supported by the observations that Rab27a is almost undetectable in this tissue when well perfused of blood (Figure 1B bottom panel) (37) and that type 1 GS patients do not show primary neurological impairment (12). Future studies will be needed to support these hypotheses.

Materials and Methods

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

Mouse strains

ashen (C3H/HeSn-Rab27a ash/+) and leaden (C57BL/6 J-ln fzH54/++ H54) mice were purchased from The Jackson Laboratory. Mice were maintained and propagated under UK project license 70/5071 at the Central Biomedical Services of Imperial College, London.

Cell culture

All media and supplements were from Invitrogen (Paisley, UK) unless otherwise stated. Melanocyte cell lines, melan-a originated from aC57BL/6 black mouse (22, 38). melan-a cells were cultured in RPMI 1640 supplemented with 10% fetal calf serum, 2 mm glutamine, 0.1 mm 2-mercaptoethanol, 200 nm phorbol 12-myristate 13-acetate (Calbiochem, Nottingham, UK), 100 U/ml penicillin G and 100 U/ml streptomycin at 37 °C with 10% CO2. Cells for immunofluorescence were grown on coverslips for 24 h and then fixed.

The derivation of murine primary melanocytes was described previously (39). Briefly, skins from neonatal mice (1–3 days old) were incubated in bovine trypsin (5 mg/ml in PBSA) for 1 h at 37 °C. The epidermis was then peeled from the dermis, washed in PBSA and cut into smaller fragments. Epidermis fragments were then placed in 5 ml of medium (described below) containing 5 μg/ml soybean trypsin inhibitor. This mixture was then aspirated and the resulting cell suspension was plated onto a 20% confluent layer of Xb2 murine keratinocyte-derived feeder cells (40). Primary cultures of murine melanocytes were maintained in RPMI 1640 supplemented with 5% fetal calf serum, 2 mm glutamine, 200 nm phorbol 12-myristate 13-acetate, 200 pm cholera toxin (Calbiochem), 100 U/ml penicillin G and 100 U/ml streptomycin at 37 °C with 10% CO2.

CTLs were derived by activation of 2.5 × 106 splenocytes from effector strains (C57Br-+/ln or ln/ln) with 2.5 × 106 irradiated splenocytes from BALB/c mice in DMEM containing 10% FCS, 2-mercaptoethanol. After 5 days of stimulation the cells were purified over Ficoll 1096 (Sigma-Aldrich, Poole, Dorset, UK), washed in PBS and either assayed for lysis as described below or resuspended in medium supplemented with 100 U/ml interleukin-2. Cells were restimulated every 7 days for continued passage in culture. P815 mouse cells were maintained in RPMI supplemented with 10% FCS.

Antibodies

Polyclonal anti-myosin Va antibodies were generated as follows. The bacterial expression vector pGEX2T-myosin Va 1260–1430, which allows the production of a fusion protein comprising amino acids 1260–1430 of the murine brain form of myosin Va fused to the carboxyl terminus of Schistosoma japonicum glutathione-S-transferase, was produced by blunt end cloning of a SacI/ClaI digested fragment of myosin Va into SmaI digested pGEX-2T (Amersham Pharmacia Biotech, Little Chalfont, Bucks., UK). Recombinant GST-myosin Va 1260–1430 was produced in Escherichia coli and affinity-purified as previously described (41). The purified protein was used to immunize rabbits and antiserum was obtained from Eurogentec Bel S.A. (Herstal, Belgium). Anti-myosin Va antibodies were affinity purified from serum of immune bleeds using the same immunogen cross-linked to AminoLink Coupling Gel (Pierce and Warriner, Tattenhall, UK), as previously described (34). The production and purification of anti-Rab27a specific monoclonal and polyclonal antibodies was described previously (22).

Anti-Rab27a monoclonal antibodies were used at 0.5 μg/ml for immunoblotting and anti-Rab27a polyclonal antibodies were used at 5 μg/ml for immunofluorescence and 10 μg/ml for immunoelectron microscopy. Anti-myosin Va polyclonal antibodies were used at 5 μg/ml for immunofluorescence, 1 μg/ml for immunoblotting and 10 μg/ml for immunoelectron microscopy. Other antibodies were used at the following dilutions: rabbit polyclonal anti-calnexin (Stressgen Biotechnologies, York, UK) 1 : 10 000, rat monoclonal anti-granzyme A (7.1, a generous gift from M. Simon) (42) 1 : 500, mouse monoclonal anti-talin 1 : 50 and rabbit polyclonal anti-actin 1 : 200 (both Sigma).

Preparation of melanocyte, CTL and brain lysates

Cultured melanocytes were washed twice with cold PBSA and then scraped into cold buffer containing 50 mm Tris-HCl, pH 7.5, 100 mm NaCl, 1 mm DTT, 10 μg/ml each of aprotinin, leupeptin, pepstatin and 0.5 mm PMSF. Cells and tissues were homogenized in the same buffer either by 10 passages through a 25G needle (melanocytes and CTLs) or using a polytron homogenizer followed by sonication (brain). Nuclei were pelleted by centrifugation of lysates at 1000 g for 10 min and the postnuclear supernatant (PNS) recovered. For melanocytes, the PNS was further centrifuged at 100 000 g for 60 min at 4 °C in a TLA-45 rotor (Beckman) to allow the separation of pellet (P100) and supernatant (S100) fractions. Protein content was determined using the BCA protein assay kit (Pierce and Warriner).

Immunoblotting

For immunoblotting, samples were subjected to SDS-PAGE using 4–15% gradient gels (Bio-Rad) and transferred to PVDF membranes. Membranes were incubated with primary antibody diluted in solution 1 (PBS, 0.2% Tween-20, 5% nonfat dry milk) for 1 h, washed with solution 2 (PBS, 0.2% Tween-20), followed by incubation for 30 min with 1 : 10 000 dilution of appropriate horseradish peroxidase (HRP)-conjugated secondary antibody (Dako, High Wycombe, Bucks., UK) diluted in solution 1 and washing as before. Bound antibody was detected using the SuperSignal Chemiluminescent Substrates (Pierce and Warriner). Blots were calibrated with prestained molecular weight standards (Bio-Rad, Hemel Hempstead, Herts., UK).

Immunofluorescence microscopy

For melanocytes, coverslip-grown cells were washed in PBS and then fixed in 3% paraformaldehyde in PBS for 15 min. Excess fixative was removed by washing in PBS and quenched by incubation in 50 mm NH4Cl for 10 min. Fixed cells were then incubated with primary antibody diluted in solution 3 (PBS, 0.5% BSA, 0.05% saponin) for 30 min, washed extensively in solution 3 and incubated for 30 min with appropriate Alexa 488 and/or Alexa 568-conjugated secondary antibodies (Molecular Probes, Cambridge, UK) diluted 1 : 200 in solution 3. Coverslips containing fixed cells were washed as before in solution 3, mounted in ImmunoFluor medium (ICN, Basingstoke, Hants., UK) and observed using a Leica DM-IRBE confocal microscope. Images were processed using Leica TCS-NT software associated with the microscope and Adobe Photoshop 4.0 software. All images presented are single sections in the z-plane. For simultaneous immunofluorescence using two rabbit polyclonal antibodies, fixed cells stained with anti-Rab27a reactive antibodies were paraformaldehyde fixed for a second time following incubation with Alexa 568-conjugated secondary antibodies. Cells were then incubated with anti-myosin Va antibodies, washed, incubated with Alexa 488-conjugated anti-rabbit antibodies and mounted as before (data not shown). For CTLs, cells taken at 6 days after re-stimulation were washed once in RPMI without FCS, resuspended at approximately 5 × 106 cells/ml in RPMI and mixed 1 : 1 with RPMI containing 5 × 106/ml P815 cells prewashed as above. Cells were left in suspension for 5 min then plated onto uncoated glass multiwell slides, incubated at 37 °C for 30 min and fixed with methanol (− 20 °C). Samples were preincubated with solution 4 (PBS supplemented with 1% BSA) for 30 min and then labeled with primary and secondary antibodies (Jackson ImmunoResearch Laboratories, Luton, Beds., UK) diluted in solution 4 for 1 h. The slides were mounted with 90% glycerol/PBS containing 2.5% DABCO (Fluka, Poole, Dorset, UK). Samples were examined using an MRC-1024 Bio-Rad confocal microscope.

Electron microscopy

For ultrathin section preparations, melan-a and leaden primary melanocytes were loaded with fluid phase HRP at 37 °C for 60 min Compartments of the endocytic pathway were then cross-linked with DAB and H2O2 and the cells were permeabilized with digitonin, all as described previously (43). The cells were labeled with anti-Rab27a rabbit polyclonal antibodies and detected using a 10-nm gold-conjugated secondary antibody. Finally, cells were fixed in 2% paraformaldehyde/1.5% glutaraldehyde, embedded in Epon, and prepared for thin section electron microscopy. For whole-mount preparations, melan-a and leaden primary melanocytes were loaded with fluid phase HRP and cross-linked as described above. Cells were permeabilized with 1% Triton X-100 as described previously (43) and labeled for myosin Va with rabbit polyclonal primary antibody followed by 15-nm gold-conjugated secondary antibody. Cells were fixed with 4% glutaraldehyde, osmicated, and critical-point dried. Samples were then rotary shadowed and carbon coated, and sections were placed on grids for subsequent electron microscopy.

Cytotoxicity assays

Cytotoxicity was assayed using a Cytotox 96 non-radioactive kit (Promega, Southampton, UK) following the manufacturer's instructions. Briefly, the cytotoxic activity of Ficoll-purified T cells, plated in 96-well plates at the effector : target ratios shown using P815 (H2d) target cells (104 in a final volume of 100 ml/well) using RPMI lacking phenol red. Lactate dehydrogenase release, indicative of target cell lysis, was assayed after 4 h incubation at 37 °C by removal of 50-ml aliquots of supernatant from each well and incubation with appropriate substrate for 30 min The absorbance at 490 nm was read and cytotoxicity calculated as described previously (19).

RT-PCR analysis

Total RNA was isolated from 1.5 × 107 melanocytes and CTLs, respectively, using TRIzol® Reagent (Life Technologies). RT-PCR was carried out using 5 μg of total RNA in a reaction volume of 20 μl mixed with 500 ng of oligo(dT)12–18 (Life Technologies), denatured at 70 °C for 10 min and chilled on ice. The annealed samples were then incubated with 0.5 mm dNTPs, 20 mm Tris-HCl (pH 8.4), 50 mm KCl, 2.5 mm MgCl2, 10 mm DTT and 200 University of SuperScript II reverse transcriptase (Life Technologies) for 1 h at 42 °C. Reverse transcription was terminated by heating for 15 min at 70 °C, followed by chilling on ice. DNA from 5 μl of RT-PCR were amplified using 500 nm oligonucleotides and 0.125 units of BioTaq DNA polymerase (Bioline) in a buffer containing 16 mm (NH4)2SO4, 67 mm Tris-HCl (pH 8.8), 0.1% Tween-20, 2.5 mm MgCl2 and each dNTP at 200 μm. PCR mixtures were incubated at 94 °C for 2 min to denature RNA:DNA hybrids and then run for 30 cycles of 94 °C denaturing for 1 min, 50 °C annealing for 1 min and 72 °C extension for 1 min. PCR products were analyzed on 2% agarose gels stained with ethidium bromide. Expression of HPRT (used as an internal control), Melanophilin and Rab27a was detected using the following pairs of oligonucleotides: 5′-CCTGCTGGATTACATTAAAGCACTG (HPRTA sense) and 5′-GTCAAGGGCATATCCAACAACAAAC (HPRTB antisense) [product size 350 bp], 5′-CGGCTTAGAAGTCCAGC (HMLPHRT1, sense) and 5′-TTAGGACTGGTGGGCCAC (HMLPHRT2, antisense) [product size 389 bp], and 5′-ATGTCTGATGGAGATTATGATTACCTC (JA43, sense) and 5′-TTGCTTGGCTTATGTTTGTCCCATTGGCA (JA120, antisense) [product size 520 bp].

Acknowledgments

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

We thank members of our labs for stimulating ideas. This work was supported by a Wellcome Trust Programme grant and a Medical Research Council component grant to MCS, a Wellcome Senior Fellowship to GMG, and a Medical Research Council programme grant to CRH. DCB was supported by a PhD studentship, grant PRAXIS XXI from Fundacao Ciência e Tecnologia of Portugal.

References

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