Flightless, secreted through a late endosome/lysosome pathway, binds LPS and dampens cytokine secretion

Flightless (Flii) is a member of the gelsolin superfamily of proteins and contains six gelsolin domains in its C-terminal end. Like other members of this family it binds actin and has F-actin severing activity via these domains (Goshima et al., 1999; Liu and Yin, 1998). As may be expected for an actin-remodelling protein, Flii regulates fibroblast and keratinocyte migration both in vitro and in vivo (Cowin et al., 2007). Many actin-binding proteins, including gelsolin family members, also have a role within the nucleus and Flii is no exception (Lee et al., 2004; Lee and Stallcup, 2006; Liu and Yin, 1998). Mice that lack other members of the gelsolin family are viable and fertile while the homozygous knockout of Flii in mice is embryonically lethal (Campbell et al., 2002). These results suggest other essential roles for Flii and its functions have now expanded well beyond the gelsolin-like actin remodelling and transcriptional regulator role found with other members of the family (Dai et al., 2009; Hayashi et al., 2010; Li et al., 2008; Wang et al., 2006).

Intracellular Flii is now emerging as an important negative regulator of inflammation (Dai et al., 2009; Li et al., 2008; Wang et al., 2006). Stimulation of Toll-Like Receptor (TLR) signalling pathways ultimately lead to transcription factor activation and the generation of cytokine and chemokine production (Kumar et al., 2011). This pathway is tightly regulated to prevent excessive and/or stimulation and increased signalling. Flii is a negative regulator of this pathway (Dai et al., 2009; Wang et al., 2006). Through its interaction with nucleoredoxin and MyD88 in macrophages Flii inhibits the binding of MyD88 to TLR4 resulting in reduced activation of NF-κB leading to diminished cytokine secretion (Dai et al., 2009; Wang et al., 2006). Flii has also been shown to inhibit cytokine secretion by other mechanisms (Li et al., 2008). Flii also binds to the proinflammatory caspases 1 and 11 and in doing so has been found to negatively regulate caspase-1-mediated interleukin-1β maturation and secretion (Li et al., 2008). Thus, Flii has the ability to dampen inflammation through a number of intracellular mechanisms.

The actions of Flii described to date involve intracellular Flii, however, results suggest that Flii may also be secreted (Cowin et al., 2007). We now show that Flii is secreted through a non-classical late endosome/lysosome-mediated pathway by at least two of the major cell types found in wounds: fibroblasts and macrophages. Secretion of Flii from fibroblasts is upregulated in response to wounding and a similar upregulation in secretion also occurs from lipopolysaccharide (LPS)-activated macrophages. The N-terminal sequence of Flii shows nearly 50% similarity to the extracellular leucine-rich repeat (LRR) domain of TLR4 (Wang et al., 2006). The LRR regions of TLRs play a key role in innate immunity by recognising structurally conserved molecules derived from microbes, with LPS from Gram-negative bacteria being the archetypical TLR4 (Erridge, 2010; Kumar et al., 2011; Piccinini and Midwood, 2010). We show that secreted Flii, with its TLR4 like N-terminus can bind LPS and negatively regulate cytokine production from macrophages.

To test whether Flii can be secreted from fibroblasts, extracts and media from a confluent layer of NIH3T3 fibroblasts (unwounded) were collected over an 8-hour time course. Samples were loaded equally on a SDS-PAGE gel (Fig 1A; supplementary material Fig. S1) and immunoblotted for Flii and actin (Fig. 1A). The anti-Flii antibody detected a band around 145 kDa both in the whole cell lysates and in media (Fig. 1). Little to no change in intracellular Flii levels were seen over the time course (Fig. 1A); however, the level of Flii in the media increased with time suggesting it is constitutively secreted from fibroblasts (Fig. 1A,B). We have previously shown that Flii is upregulated in mouse skin wound tissue in response to injury (Adams et al., 2009; Cowin et al., 2007). Since Flii levels increase in the wounds of mice, scratch wounds were created in a confluent layer of NIH3T3 fibroblasts to determine whether scratch wounding these cells could alter Flii secretion. Cell extracts and media were also collected from these cells over the same 8-hour time course and immunoblotted for Flii and actin (Fig. 1A,B). The level of intracellular Flii was similar to that found in unscratched fibroblasts (Fig. 1A). However, the amount of secreted Flii was increased by 1.5- to 2.5-fold upon scratch wounding compared to medium from unwounded cells incubated for the same time period suggesting its secretion is upregulated upon wounding (Fig. 1A,B). Immunoblotting of the same media samples for the cytosolic enzyme glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and for actin showed that the presence of Flii in the media is not due to general cytosol leakage or cell disruption (Fig. 1A). The activity of a known secreted lysosomal enzyme, beta-hexosaminidase, was also analysed in media from control and scratched fibroblasts (Fig. 1C). While the level of enzyme activity increased over the 8-hour time course no significant differences were seen between the amount of beta-hexosaminidase activity in the media from scratched and control fibroblasts (Fig. 1C). This data shows that beta-hexosaminidase is also secreted from these cells; however, unlike Flii its level of secretion is not significantly increase upon wounding. We next looked at the level of Flii secreted from cells over the 8-hour scratch wounded period (Fig. 1D). The levels of Flii secreted into the media 6 hours after scratching wound is equivalent to approximately 3% of the total intracellular pool of Flii (Fig. 1D). Thus, Flii is constitutively secreted by fibroblasts and its secretion can be upregulated in response to scratch wounding.

Immunoblotting of skin and plasma from healthy volunteers for Flii, the plasma marker albumin and the cellular marker tubulin showed that Flii is present in plasma as well as the skin (Fig. 1E). Flii has been found in macrophages where it can regulate TLR signalling and cytokine secretion (Dai et al., 2009; Hayashi et al., 2010; Wang et al., 2006). Monocyte/macrophages are present in high numbers both in wounds and plasma and so whether Flii could also be secreted from macrophages was next tested. Extracts from RAW264.7 macrophages were first immunoblotted alongside equal protein levels of NIH3T3 fibroblast extracts (Fig. 1F). Similar intracellular levels of Flii were seen in both cell types (Fig. 1F). To test whether Flii is also secreted by macrophages and whether this secretion is altered when macrophages are activated and secreting cytokines, RAW264.7 macrophages were stimulated with the bacterial cell wall component LPS over a 6-hour time course. Cell extracts were immunoblotted for Flii, actin and the proinflammatory cytokine TNF (Fig. 1G,H). Thirty minutes post LPS stimulation we saw the characteristic upregulation of intracellular TNF protein levels, which peaked around 3 hours post stimulation, while the total levels of intracellular Flii appeared unaltered over the same time course (Fig. 1G) (Murray et al., 2005b). Immunoblotting of conditioned media collected from the same cells showed TNF secretion started 30 minutes post LPS stimulation and gradually increased (Fig. 1G,H). Moreover, Flii was also secreted by macrophages and in LPS-activated macrophages secretion peaked at around 30 minutes post-stimulation (Fig. 1G,H). These results together suggest Flii is constitutively secreted by both fibroblasts and macrophages and that Flii secretion can be upregulated from fibroblasts in response to scratch wounding and in macrophages in response to LPS activation.

Since Flii has previously been found in both the nucleus and cytoplasm of Swiss 3T3 fibroblasts its location in unwounded NIH3T3 cells and scratch wounded fibroblasts was next analysed (Davy et al., 2001). Cell extracts from unwounded and scratch wounded confluent NIH3T3 fibroblasts were separated into nuclear and post-nuclear fractions (Fig. 2A). Equal protein was loaded and probed for Flii, the cytoplasmic marker protein tubulin and the nuclear marker protein HDAC1 (Fig 2A). Flii was predominantly located in the post-nuclear fraction in NIH3T3 fibroblasts and its location did not significantly alter upon wounding (Fig. 2A). RAW264.7 macrophages were incubated in the presence or absence of LPS for 3 hours and extracts were separated into nuclear and post-nuclear fractions to determine its location. Fractions were immunoblotted for Flii, the cytoplasmic marker protein tubulin, the nuclear marker protein nucleoporin and the cytokine TNF (Fig. 2B). The majority of Flii was located in the post-nuclear fraction that contains membranes and cytosol in macrophages regardless of their activation state (Fig. 2B). Post-nuclear fractions from both NIH3T3 cells and RAW264.7 cells were next separated into membrane and cytosol using a high-speed centrifugation step. Membrane and cytosol fractions were probed for Flii, the trans-membrane protein LAMP1 and the cytoplasmic protein β-actin (Fig. 2C). In both fibroblasts and macrophages Flii is located both in the cytosol and the membrane fractions (Fig. 2C).

Immunostaining of Flii in both fibroblasts and macrophages revealed that Flii was located in the cytosol as well as in large endosomes located in the perinuclear region and throughout the cytoplasm (Fig. 3A,B). Some Flii could be detected in the nuclei of fibroblasts but its levels differed depending on the individual cell (Fig. 3A). Primary mouse fibroblasts were co-immunostained for Flii along with the late endosome/lysosome marker VAMP7 or cathepsin D, a late endosome/lysosomal protease that is also secreted by some cell types. Flii is located in the nucleus, cytosol and in late endosomes/lysosomes in fibroblasts (Fig. 4A). Immunostaining of RAW264.7 macrophages for Flii in combination with the late endosome/lysosome membrane proteins LAMP1 and VAMP7, as well as cathepsin D, showed these proteins colocalised in macrophages (Fig. 3B). Thus, Flii is located in late endosomes/lysosomes with cathepsin D in fibroblasts as well as macrophages and is potentially trafficked to the surface via late endosomes/lysosomes.

We next looked to see how Flii might be being secreted. Two major types of routes for secretion exist, classical and non-classical. In the classical pathway proteins are trafficked from the endoplasmic reticulum (ER) via the Golgi complex en route to the cell surface (Bonifacino and Glick, 2004). In this pathway secretory proteins in this pathway contain amino-terminal or internal signal peptides that direct their sorting to the endoplasmic reticulum (ER). The amino acid sequence of Flii was analysed using SignalP 3.0 (www.cbs.dtu.dk/services/SignalP) and the results established that Flii lacks a secretory signal peptide (probability  =  0.014). This suggested that Flii might not be secreted through the classical pathway. To confirm this, Flii was co-immunostained in RAW264.7 macrophages with the Golgi protein VAMP4, as well as with the known classical secretory pathway trafficked cytokine TNF (Murray et al., 2005a; Murray et al., 2005b). Flii did not colocate with the Golgi associated SNARE protein VAMP4 or the cytokine TNF (Fig. 2C). Brefeldin A (BFA) has been shown to cause disruption of the Golgi and prevent proteins trafficking from the ER to the Golgi through the classical pathway (Fujiwara et al., 1988). To confirm that Flii is not secreted through the classical pathway, RAW264.7 macrophages were incubated in the presence or absence of BFA (5 ng/ml) for 60 min with LPS (100 ng/ml) for the last 30 min. Fixed cells, immunostained for the Golgi marker protein GM130, showed that in the presence of BFA the Golgi was disrupted as expected (Fig. 4B). Immunoblotting of extracts from these cells showed that the levels of Flii inside the cell and secreted in the media were not altered in the presence of BFA (Fig. 4C) confirming that Flii is not secreted through the classical pathway.

The delivery of cargo to the cell surface via late endosomes/lysosomes is a known non-classical pathway of secretion and the presence of Flii in late endosome/lysosomes suggest that Flii may be secreted through this same pathway (Nickel, 2003). The small G-protein Rab7 regulates traffic through the late endosome (Bucci et al., 2000; Vanlandingham and Ceresa, 2009). To determine whether Flii is indeed trafficked through the late endosome on route to the cell surface RAW264.7 macrophages were transiently transfected with GFP, GFP–Rab7 or a GFP-tagged dominant-negative form of Rab7 (GFP–Rab7T22N) known to inhibit traffic through this organelle (Bucci et al., 2000). Extracts from these macrophages were immunoblotted for Flii, cathepsin D, GFP and actin (Fig. 5A). The intracellular levels of both Flii and the mature form of cathepsin D were decreased by 25% and 21% respectively in cells transiently overexpressing GFP-tagged wild-type Rab7 as compared to cells expressing GFP alone (Fig 5A,B). This decrease in intracellular Flii in cells coincided with a 1.7-fold increase in Flii secreted into the media (Fig. 5A,B). Detection of cathepsin D in media by immunoblotting was obscured due to the large albumin band in the media samples so the same conditioned media were also analysed for the secreted lysosomal enzyme, beta-hexosaminidase (Fig. 5C). Similar to Flii, the level of secreted beta-hexosaminidase is increased (approximately 1.2-fold) in the conditioned media of cells transiently overexpressing GFP–Rab7 (Fig. 5C). Macrophages transiently transfected with GFP–Rab7T22N showed an increase (approximately 1.2-fold) in intracellular Flii and mature cathepsin D as compared to cells transfected with GFP alone (Fig. 5A,B). The increased intracellular Flii levels corresponded with a reduction in secreted Flii in the media by almost half compared with control GFP-expressing cells suggesting secretion was blocked by expression of the dominant-negative form of Rab7 (Fig 5A,B). Similarly a reduction in beta-hexosaminidase secreted into the media was seen in conditioned media from cells expressing the dominant-negative form of Rab7 (Fig. 5C). These results together suggest that Flii is trafficked through a late endocytic/lysosomal compartment on route to the cell surface for secretion.

We have recently shown that the loss of the SNARE protein Stx11 in macrophages leads to an increased fusion of a LAMP1-positive late endosomal/lysosomal compartment with the cell surface (Fig. 5D) (Offenhäuser et al., 2011). If Flii were trafficked through this same Stx11-regulated pathway then it would be expected that a reduction in Stx11 levels would deplete intracellular Flii pools while increasing the level of Flii secreted into the medium. To test this, macrophages were treated with small interfering RNAs (siRNAs) to reduced the level of Stx11 (Fig. 6A,B). Cells treated with control or Stx11 siRNA were fixed, permeabilised and immunostained live for surface LAMP1 (Fig. 5D). As previously shown, surface levels of LAMP1 are greatly increased on macrophages treated with Stx11 siRNA as compared to control treated cells, indicative of an increase in the fusion of LAMP1-positive organelles with the cell surface (Fig. 5D). Cell extracts from these macrophages were immunoblotted for Stx11, Flii, cathepsin D and actin and media from the same cells was immunoblotted for Flii (Fig. 5E). Compared to cells treated with control siRNA cells treated with siRNA to Stx11 showed a reduction in intracellular Flii and cathepsin D (approximately 50% and 20%, respectively) and a concomitant increase (1.5-fold) in secreted Flii levels (Fig. 5E,F). An increase in the lysosomal enzyme beta-hexosaminidase secreted into the medium was also seen when cells were treated with Stx11 siRNA compared to control treated cells (Fig. 5G). These results together suggest that the pool of Flii located in a LAMP1-positive late endosome/lysosome compartment is trafficked to the cell surface and released into the extracellular milieu when this LAMP1-positive late endosome/lysosome-related compartment fuses with the cell surface.

Eleven leucine-rich repeat (LRR) domains, located at the N-terminus between amino acids 57 and 362 in Flii, were identified using ScanProsite and InterPro Scan domain mapping tools (http://expasy.org/tools/; Fig. 6A). This LRR region shares 29% sequence identity and 42% similarity to the LRR domains found in the extracellular portion of TLR4 (Wang et al., 2006). The binding of the bacterial cell wall component LPS to this same domain in TLR4 leads to the signal transduction events that ultimately result in cytokine secretion (Bell et al., 2003). Thus, whether Flii, like TLR4, could also bind LPS was next tested. Conditioned medium containing high levels of secreted Flii from scratch wounded NIH3T3 cells was incubated with LPS and antibodies to GFP, LPS or Flii bound to protein-A–agarose. Co-immunoprecipitated proteins were immunoblotted for Flii and results confirmed the anti-Flii antibody was able to immunoprecipitate Flii as expected (Fig. 6B). Moreover, Flii from the conditioned medium was immunoprecipitated with LPS (Fig. 6B). Addition of an antibody specific for the first LRR domain of Flii (Cowin et al., 2007), added to conditioned medium for 1 hour prior to the above immunoprecipitation experiment, inhibited the binding of Flii to LPS (Fig. 6C). These results suggest Flii might bind LPS via its LRR domain.

We next tested whether the binding of Flii to LPS could influence LPS activation of macrophages, looking at the effect of altering Flii levels on the downstream TNF production and secretion. Conditioned medium, containing high levels of secreted Flii from scratch wounded NIH3T3 cells, was depleted of Flii using anti-Flii antibodies bound to protein-A–agarose. Immunoblotting showed Flii levels in the medium were greatly depleted (Fig. 6D). Macrophages grown on coverslips were incubated in conditioned medium containing high levels of Flii or in the Flii-depleted medium and stimulated with LPS for 3 hours then fixed, permeabilised and stained for TNF and F-actin. Both the level and the number of cells expressing TNF were greatly reduced in the presence of high levels of Flii, suggesting Flii may inhibit LPS activation of macrophages (Fig. 6E). To confirm this, extracts from LPS stimulated macrophages incubated in conditioned medium containing Flii or in Flii-depleted medium were immunoblotted for TNF and actin. In the presence of high levels of Flii the total TNF levels were reduced by approximately 30% compared to cells stimulated in Flii-depleted medium (Fig. 6F,H). A 40% reduction in secreted TNF was seen when cells were stimulated with LPS in the presence of high levels of Flii (Fig. 6G,H). These results suggest that secreted Flii may sequester LPS, preventing it from activating macrophages.

Intracellular Flii has been found to be an important negative regulator of inflammation (Dai et al., 2009; Li et al., 2008; Wang et al., 2006). We now show that Flii is also secreted and the secreted form can also regulate inflammation. Flii is constitutively secreted by both fibroblasts and macrophages and can be found in plasma samples from healthy volunteers. Secretion of Flii can be upregulated by either scratch wounding fibroblasts or LPS stimulating macrophages. Flii is secreted via a late endocytic/lysosomal compartment and both Rab7 and the SNARE Stx11 regulate this trafficking pathway in macrophages. Flii also locates to the same compartment in fibroblasts, suggesting Flii is secreted through a similar pathway in both cell types. Inhibition of this trafficking pathway results in increased intracellular Flii and a reduction in secretion, while increasing fusion of late endosomes/lysosomes increases Flii secretion. Once secreted Flii can bind to LPS and this binding can be inhibited with an antibody specific for the LRR domains at the N-terminus of Flii. Macrophages stimulated with LPS in the presence of medium depleted of Flii synthesise and secrete more TNF, suggesting secreted Flii may act to sequester LPS and alter cytokine secretion during an inflammatory response.

Biochemical and localisation studies in NIH3T3 fibroblasts revealed that Flii is located in the nucleus, cytosol and in late endosomes/lysosomes. In RAW264.7 macrophages the majority of Flii is located in the cytosol and in late endosomes/lysosomes. Flii has previously been found in the nucleus, cytosol and to be associated with the cytoskeleton in Swiss 3T3 cells (Davy et al., 2001). The slight differences in localisation may reflect the different cell types and perhaps the functions of Flii in these cells. Flii has been found to play a number of different roles in cells (Cowin et al., 2007; Davy et al., 2001; Kopecki and Cowin, 2008; Wang et al., 2006). In macrophages the cytosolic pool of Flii has been found to bind to MyD88 and prevent it's binding the cytoplasmic tail of TLR4 (Wang et al., 2006). Flii also binds to caspase-1 in macrophages where it can inhibit the maturation of interleukin-1β maturation and leading to a reduction in its secretion (Li et al., 2008). Interestingly pro-interleukin-1β and caspase-1 have both been shown to be found in late endosomes with Cathepsin D and to be secreted from this compartment (Andrei et al., 1999; Eder, 2009).

Lysosomes or lysosome-related organelles have been found to perform a range of functions, some common to all cell types, such as degradation, whilst others are more specialised and tend to involve fusion of this organelle with the cell surface to release its contents (Holt et al., 2006; Luzio et al., 2007). We have used Rab and SNARE proteins, located on specific membranes that regulate transport at distinct sites in the late endosome/lysosome pathway, to show Flii is secreted from this compartment (Stow et al., 2006; Zerial and McBride, 2001). Rab7 is located on late endosomes, where it regulates traffic to lysosomes (Bucci et al., 2000; Vanlandingham and Ceresa, 2009). Similarly, the SNARE protein Stx11 is located on late endosomes and lysosomes and regulates traffic between these organelles and the cell surface (Offenhäuser et al., 2011). Loss of either Rab7 or Stx11 results in enlarged endosomes and loss of Stx11 leads to increased fusion of this late endocytic/lysomal compartment with the cell surface (Offenhäuser et al., 2011; Vanlandingham and Ceresa, 2009). We now show that Rab7 and Stx11 regulate Flii secretion from a late endocytic/lysosomal compartment in macrophages. Similar results were seen for cathepsin D, which is known to be secreted via a late endocytic/lysosomal compartment (Gardella et al., 2001).

Lysosomal secretion can be either from distinct subsets of lysosome related organelles, for example melanosomes, or from organelles that are indistinguishable from lysosomes themselves (Holt et al., 2006; Luzio et al., 2007). Secretion from these organelles is often switched on by external stimuli, such as the recognition of a foreign antigen (Holt et al., 2006; Luzio et al., 2007). Our results suggest that in both fibroblasts and macrophages secretion of Flii from a late endosome/lysosome compartment is constitutive and can be upregulated in response to specific external stimuli. In response to scratch wounding, fibroblasts secrete up to 2.5 times more Flii that unwounded cells, whilst in macrophages LPS activation leads to a temporal increase in secreted Flii 30 minutes after activation. No increase in intracellular Flii was seen after scratch wounding to match the increase in secretion; however, the level of Flii secreted is very low (3% over 6 hours) in comparison to the total cellular levels. These results suggest that secretion from this compartment may be continual but can be amplified in response to external stimuli. It is currently unclear how Flii enters the late endosomal/lysosomal compartment but future studies will address this question.

Mice with increased levels of Flii have impaired wound healing compared to control mice, while mice with low levels of Flii have significantly improved wound healing responses (Cowin et al., 2007). However, Flii is upregulated in tissue in response to injury in mice and we now show fibroblasts increase secretion of Flii in response to wounding (Adams et al., 2009; Cowin et al., 2007). The upregulation of Flii in response to injury seems counterproductive since increasing the levels of Flii leads to reduced cell migration, impaired healing and increased scar formation. The results shown here suggest the answer may lie in part in the ability of Flii to also alter the inflammatory response. Whilst mounting a successful immune response is important to clear an infection, a hyper-activated immune response leads to overproduction of tissue damaging cytokines, which could potentially be detrimental to the wound, as well as systemically where it can lead to sepsis (Murphy et al., 2004). Intracellular Flii has previously been shown to negatively regulate cytokine secretion by inhibiting MyD88 binding to TLR4 and effectively acts to reduce the signalling for cytokine production and dampen the inflammatory response (Dai et al., 2009; Wang et al., 2006). Our results now suggest that Flii is also secreted and that extracellular Flii can also alter cytokine secretion by binding to LPS. The binding of Flii to LPS and the timing of Flii secretion suggest that Flii may act as a scavenger to help mop up excess LPS and prevent hyper-activation of the immune system. These results suggest that the upregulation of Flii in wounds might serve to alter inflammation in the wound and prevent excess stimulation of cytokine production.

The related protein gelsolin is also secreted and the extracellular form has been shown to dampen the immune response by binding to LPS and lipoteichoic acid (LTA), released from the outer walls of Gram-negative and Gram-positive bacteria, respectively (Bucki et al., 2008; Bucki et al., 2005). Thus, it has been proposed that extracellular gelsolin may act as an extracellular scavenger for these bacterial cell wall components (Bucki et al., 2008). The region in gelsolin that binds to LPS is not present in Flii (data not shown); however, Flii does have an LRR domain at it N-terminus, not present in other gelsolin family members (Kopecki and Cowin, 2008; Mintzer et al., 2006). The LRR domain in Flii shares 29% sequence identity and 42% similarity to the LRR domain in TLR4 (Wang et al., 2006). The LRR domain of TLRs is located on the extracellular side of the plasma membrane and has been shown to be involved in the binding of pathogen-associated molecular patterns (PAMPs) from a range of pathogens, including LPS, as well as damage-associated molecular patterns (DAMPs) found in some host molecules released upon injury (Bell et al., 2003; Erridge, 2010; Kumar et al., 2011). The binding of Flii to LPS and the timing of Flii secretion suggest that Flii may also act as a scavenger to help mop up excess LPS and prevent hyper-activation of the immune system. Since many proteins are released upon injury, a number of which also can activate TLR signalling, it remains to be seen whether secreted Flii can also act as a general scavenger protein that binds to the DAMPs in host intracellular proteins released upon injury (Erridge, 2010). In summary, Flii may be acting both in side and outside of the cell to dampen inflammation by negatively regulating intracellular TLR signalling as shown previously or as we now show by binding LPS and preventing activation of TLR signalling.

Antibodies specific for Vti1b, VAMP3, VAMP4 and VAMP7 were purchased from Synaptic Systems (Goettingen, Germany) and antibodies specific for cathepsin D and actin were purchased from Millipore (Kilsyth, Vic, Australia). Anti-mouse Flii (IgG Sc-21716), anti-rabbit Flii (IgG Sc-30046), anti-HDAC1 and anti-albumin antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Anti-GAPDH antibodies were purchased from Ambion (Austin, TX, USA). Antibodies specific for LAMP1 were purchased from Southern Biotech (InVitro, Noble Park, Vic, Australia), while antibodies specific for nucleoporin and GM130 were purchased from BD Transduction Laboratories (North Ryde, NSW, Australia). Antibodies specific for TNF were purchased from Calbiochem (Merck, Kilsyth, Vic, Australia). Anti-LPS (Escherichia coli) antibodies were purchased from Abcam (Sapphire Biosciences, Waterloo, NSW, Australia). Rabbit anti-GFP antibodies were purchased from Invitrogen (Mulgrave, Vic, Australia) and mouse anti-GFP antibodies were purchased from Roche (Castle Hill, NSW, Australia). HRP and CY3-tagged secondary antibodies were purchased from Jackson Laboratories (West Grove, PA, USA), while Alexa-488- and Alexa-647-tagged secondary antibodies and Alexa-488 phalloidin were purchased from Molecular Probes (Invitrogen, Mulgrave, Vic, Australia). Anti-Flii LRR (IgM) antibodies made to amino acids 56–69 located in the first LRR domain of Flii as described previously (Adams et al., 2009). Anti-β-tubulin antibodies, LPS from Salmonella minosota Re595 and LPS from E. coli J5 (Rc mutant) (used in the immunoprecipitation experiments) and 4-nitrophenyl N-acetyl-β-D-glucosaminide (20 µl) were purchased from Sigma (Sydney, NSW, Australia). Rabbit polyclonal anti-Stx11 antibodies were a kind gift from Prof Jenny Stow (University of Queensland, Australia) and have been described elsewhere (Offenhäuser et al., 2011). GFP–Rab7 and GFP–Rab7T22N were a kind gift from Prof Zerial (Max Planck Institute of Molecular Cell Biology and Genetics, Dresden).

The collection of human skin and peripheral blood samples from patients undergoing abdominoplasty was approved by the Central Northern Adelaide Health Service Ethics of Human Research Committee. Peripheral blood collected in EDTA treated blood collection tubes was centrifuged at 300 g for 10 min at 4°C and the plasma transferred to a new tube. Samples were mixed with trichloroacetic acid at 1:10 sample volume and centrifuged for 1.5 h at 6500 g. Samples were resuspended in 800 µl of −20°C acetone, incubated overnight at −20°C and centrifuged at 6500 g at 4°C for 10 minutes. The pellet was then resuspended in −20°C acetone, incubated at −20°C for 30 minutes, re-centrifuged, the acetone discarded and the pellet dried. The pellet was then resuspended in 1% SDS prior to separation on an SDS-PAGE gel. Human skin samples dissected into 0.5 ×0.5 mm pieces were lysed in RIPA buffer containing protease inhibitors [50 nM Tris, pH 7.2, containing 150 mM NaCl, 2 mM EDTA, 1% Triton X-100, 1% nadeoxycholate, 1 mM Na₃VO₄, 1 mM NaF, 2 mM Perfabloc and complete protease inhibitors (Roche, Castle Hill, NSW, Australia)]. Samples were homogenised on ice using a Heidolph Diax 600 homogenizer (Sigma, Sydney, NSW, Australia) at 22,000 rpm for 1 min. Homogenised samples were incubated for 30 minutes at 4°C, then centrifuged at 12,000 rpm for 15 min and the supernatant collected for further analysis.

RAW264.7 murine macrophages were cultured and in some experiments the cells were activated for the times indicated with LPS (100 ng/ml) as described previously (Murray et al., 2005a). Primary fibroblasts from BALB/c mice were cultured as previously described (Kopecki et al., 2009). NIH3T3 fibroblasts were cultured according to the ATCC instructions in DMEM supplemented with 4 mM L-glutamine, 1.5 g/l sodium bicarbonate 4.5 g/l glucose and 10% FCS. Macrophages were transfected with cDNAs using Lipofectamine 2000 according to the manufacturer's instructions (Invitrogen, Mulgrave, Vic, Australia). In some experiments RAW264.7 cells were incubated with 5 µg/ml Brefeldin A (BFA) for 30 minutes and then LPS (100 ng/m) was added for a further 30 minutes. For siRNA knockdown macrophages were transfected using Lipofectamine 2000 (Invitrogen, Mulgrave, Vic, Australia), cultured for 24 hours, re-transfected under the same conditions and then cultured for a further 24 hours prior to use as previously described (Veale et al., 2010).

To separate nuclear and post nuclear fractions cells were detached from plates using 0.25% trypsin/EDTA (Invitrogen, Mulgrave, Vic, Australia) for NIH3T3 cells or 0.5 mM EDTA in PBS for RAW264.7 cells. The cells were washed once in ice cold PBS, resuspended in buffer A [10 mM HEPES, pH 7.9, 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 0.15% NP-40 and Complete protease inhibitors (Roche, Castle Hill, NSW, Australia)] and incubated on ice for 15 minutes with occasional mixing. The lysed cells were centrifuged for 30 sec at 14,000 g and the supernatant was the post nuclear fraction. The pellet containing intact nuclei was washed three times in Buffer A, resuspenend in Buffer B [20 mM HEPES, pH 7.9, 0.4 mM NaCl, 1 mM EDTA, 1 mM EGTA, 0.5% NP-40 and Complete protease inhibitors (Roche, Castle Hill, NSW, Australia)], sonicated and centrifuged for 30 min at 14,000 g and the supernatant is the nuclear fraction. Samples (100 µg) were separated by SDS-PAGE. Membranes and cytosol were prepared as previously described (Offenhäuser et al., 2011). The membrane volume was adjusted to the same as the cytosol and equal volumes loaded on SDS-PAGE gel.

Cell extracts were prepared as previously described (Veale et al., 2010). Briefly, cells were washed three times with ice-cold PBS and lysed in Buffer C [10 mM Tris, pH 7.4, containing 1 mM EDTA, 150 mM NaCl, 1% Triton X-100, and Complete protease inhibitors (Roche, Castle Hill, NSW, Australia)] by passage through a series of successively smaller needles. The supernatant was then collected after centrifugation for 30 min at 14,000 g. Protein content in the supernatant was assayed using the BioRad protein assay kit according to the manufacturer's instructions, and 50 µg of protein was subject to SDS-PAGE separation and immunoblotting. Densitometry was performed on bands within the linear range and fold changes in levels calculated from this data. For immunoprecipitation, the medium was incubated with 5 µg of antibody bound to protein-A–agarose (Sigma, Sydney, NSW, Australia) for 2 h at 4°C with constant mixing. The protein-A–agarose was washed five times in excess Buffer C and the bound proteins were solubilised in SDS-PAGE sample buffer. Immunofluorescence staining was performed as described previously (Murray et al., 2005b). Briefly, cells were fixed in either ice-cold methanol (for subsequent immunostaining with the mouse anti-Flii antibody) for 5 minutes at −20°C or 4% paraformaldehyde in PBS (for subsequent immunostaining with the rabbit anti-Flii antibody) for 60 minutes and washed in phosphate-buffered saline (PBS). Paraformaldehyde-fixed cells were permeabilised with 0.1% Triton X-100 in PBS for 4 minutes. Cells were washed and blocked with 0.5% bovine serum albumin (BSA) in PBS for 10 minutes, incubated with anti-Flii antibody (1:1000 in 0.5% BSA in PBS) for 60 minutes and then washed with 0.5% BSA in PBS. Cells were incubated with the appropriate fluorophore-tagged secondary for 60 minutes, washed with 0.5% BSA in PBS and the coverslips mounted in DABCO. Surface staining of LAMP1 was performed on live cells on ice, cells were then fixed with 4% paraformaldehyde at room temperature as previously described (Offenhäuser et al., 2011). Confocal microscopy was performed using a Leica (North Ryde, NSW, Australia) TCS SP2 Confocal Microscope equipped with a 100× oil objective. Confocal images were pseudocoloured where appropriate and overlaid for publication using Adobe Photoshop CS3.

Conditioned and unconditioned media (20 µl) were incubated with 4.8 mM 4-nitrophenyl N-acetyl-β-D-glucosaminide (20 µl; Sigma-Aldrich) in citrate buffer (0.1 M, pH 4.5) at 37°C for 90 min. The reaction was halted by adding 150 µl carbonate buffer (0.1 M, pH 9.0) and the absorption read at A415 (nm) and the activity in unconditioned medium subtracted.