SCIMP Is a Universal Toll‐Like Receptor Adaptor in Macrophages

SCIMP is a universal Toll-like receptor adaptor controlling selective cytokine outputs from macrophages

Lin Luo*, James E. B. Curson, Liping Liu, Adam A. Wall, Neeraj Tuladhar, Richard M. Lucas, Matthew J. Sweet and Jennifer L. Stow

Institute for Molecular Bioscience (IMB) and IMB Centre for Inflammation and Disease Research, The University of Queensland, Brisbane, QLD 4072, Australia

Summary sentence: The TLR adaptor SCIMP is used by multiple TLRs to generate cytokine specificity in macrophages

*Correspondence should be addressed to:

Dr Lin Luo

Institute for Molecular Bioscience,

The University of Queensland, Brisbane, QLD, Australia.

Email: [email protected]

Key words: SCIMP, Toll-like receptor, TRAP, inflammatory cytokines, macrophage, effectors

This is the author manuscript accepted for publication and has undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1002/JLB.2MA0819-138RR.

Word counts: 4721

ABSTRACT

In innate immune cells, pathogens and danger signals activate Toll-like receptors (TLRs), unleashing potent and tailored inflammatory responses. Previously, we reported that an immune-specific transmembrane adaptor, SCIMP, interacts with TLR4 via direct binding to its cytoplasmic TIR domain. SCIMP scaffolds a Src family kinase, Lyn, for TLR4 phosphorylation and activation.

Consequently, SCIMP is able to direct selective production of the pro-inflammatory cytokines IL-6 and IL-12p40 downstream of TLR4 in macrophages. Here we set out to investigate whether SCIMP also acts as an adaptor for other TLR family members. We report here that SCIMP is phosphorylated and activated in response to agonists of multiple TLRs, including TLR2, TLR3, TLR4 and TLR9.

SCIMP also interacts with TLRs that are known to signal from both the cell surface and endosomal compartments. In so doing, this transmembrane adaptor presents Lyn, along with other effectors such as Grb2, Csk and SLP65, to multiple TLRs during cellular activation. CRISPR-mediated knockout or silencing of SCIMP in macrophages alters TLR signalling outputs and the production of IL-6 and IL-

12p40 downstream of multiple TLRs, and upon challenge with live bacteria. Furthermore, the selectivity in cytokine responses is preserved downstream of TLR3, with inducible expression of Il-

12p40 and IL-6, but not IFN, being SCIMP-dependent. SCIMP is thus a universal TLR adaptor for scaffolding the Lyn tyrosine kinase and its effectors to enable responses against a wide range of danger signals.

INTRODUCTION

Macrophages are central to innate immune responses through their capacity to detect, interpret, and respond to danger signals presented by pathogens and/or other stimuli. Pattern recognition receptors, including the Toll-like receptors (TLRs), are activated by different pathogen-associated molecular patterns, and signalling from these receptors initiates the synthesis and secretion of inflammatory cytokines and other mediators (1). TLR4, the most widely studied member of the TLR family, senses and initiates responses to lipopolysaccharide (LPS) from Gram-negative bacteria (2). At the cell surface, TLR4 dimers recruit the cytoplasmic adaptors Mal (TIRAP) and MyD88, which bind directly through cognate Toll/IL-1 Receptor homology (TIR) domains. The TIR-containing adaptors TRIF and

TRAM are subsequently recruited to TLR4 in endosomes (3). Other TLR family members signal from either the cell surface or from endosomal compartments, via one or more of the TIR-containing adaptors. TLR2 is a surface receptor that heterodimerises with TLR1 or TLR6 to recognize lipoproteins from Gram-positive bacteria, lipoarabinomannan from mycobacteria, and zymosan from yeast (4). Detection of intracellular viral and host dsRNAs by TLR3 (5), as well as unmethylated

CpG-containing DNA by TLR9 (6), occurs in endosomal compartments. Many TLRs have surface- exposed tyrosine sites (e.g. Y674 in human TLR4) within the TIR domains of their cytoplasmic tails.

Some of these residues are highly conserved in human and mouse and across many TLR family members, and are essential for TLR TIR-adaptor (MyD88/MAL or TRIF/TRAM) recruitment, signal transduction and cytokine production (7, 8). TLR2 (9, 10), TLR3 (11) and TLR4 (7, 12) are all known to be tyrosine phosphorylated in an agonist-induced manner. Despite the key role of tyrosine phosphorylation of these TLR TIR domains in signalling, only a limited number of studies have investigated the upstream kinases responsible for phosphorylation at these sites (13). In the case of

TLR4, studies by us and others have implicated Lyn kinase as the responsible kinase (7, 12), although

Syk is also reported to phosphorylate TLR4 (14). However, our knowledge of the identity of tyrosine kinase(s) acting on other TLR members, as well as the precise mechanisms involved, is still limited.

The transmembrane adaptor protein (TRAP) family of adaptors is typified by very short extracellular domains and longer cytoplasmic tails that serve as scaffolds for signalling kinases and effector proteins (including Lyn, Grb2, Csk and SLP65) (15, 16). pTRAPs are a sub-family of TRAPs that can be directed to specific membrane domains, such as lipid rafts or tetraspanin-enriched membranes

(TEMs), by palmitoylation to enable exquisite signalling specificity from immune and non-immune receptors (17). The cell-type specific expression patterns of TRAPs and their capacity to recruit specific signalling components, such as tyrosine kinases and effector proteins, permits fine-tuning of cellular responses. Previously, we showed that SLP adaptor and CSK interacting membrane protein

(SCIMP), a little known pTRAP family member, is an atypical adaptor for TLR4 that drives highly discriminate cytokine responses (IL-6 and IL-12p40, but not other inflammatory mediators examined) in macrophages, thus revealing an unrecognized level of regulation for innate immune signalling (12,

18). SCIMP is constitutively associated with the Lyn tyrosine kinase (19), and upon TLR4 activation, this adaptor presents Lyn to TLR4, resulting in tyrosine phosphorylation of this receptor (12).

However, other SCIMP effectors may also be involved in TLR signalling. For example, Csk reportedly participates in signalling downstream of TLRs (20). SLP65 binds GRB2 through its proline-rich domain (21, 22). In B cells, SLP65 functions with GRB2 together to localize VAV3 to the membrane, where it can activate RAC1 and downstream signalling (23), but it is not well understood whether these effectors are involved in SCIMP-mediated TLR pathways. Here we investigate the role of the TLR adaptor SCIMP in driving cytokine production in macrophages upon

TLR stimulation. Interestingly, we found that SCIMP is a common adaptor for multiple surface and endosomal TLRs, including TLR2, 3, 4, and 9 where it plays an important role in scaffolding the Lyn tyrosine kinase for production of specific proinflammatory cytokines during TLR-mediated macrophage activation.

MATERIALS AND METHODS:

Plasmids, antibodies and reagents

The mouse SCIMP full-length construct was amplified by PCR from cDNA and cloned into the pEF6/V5-His TOPO TA vector and confirmed by DNA sequencing. Primary rabbit antibodies recognizing SCIMP have previously been described (12). Antibodies to p-ERK1/2 (4370), p-Akt

(9271), p-IκBα (2859), p-TBK1/NAK (5483), Csk (4980), GAPDH (2118), Grb2 (3972), Lyn

(2796), SLP65 (12168), Myc (2276), p-Tyrosine (9411) and the p-Src family (Tyr416; 6943) were purchased from Cell Signalling Technology (Beverly, MA, USA). The anti-V5 (MCA1360) antibody were purchased from Bio-Rad Laboratories. The anti-TLR4 antibody targeting TLR4 amino acids

100-200 (ab22048), TLR3 antibody (ab62566) and TLR2 antibody (ab16894), used for immunoprecipitation and Western blotting, were from Abcam (Melbourne, VIC, Australia). HRP- conjugated goat anti-mouse and -rabbit antibodies (81-6520) were obtained from Zymed (San

Francisco, CA, USA). Zymosan (Z4250) was purchased from Sigma Australia (Castle Hill, NSW,

Australia). FSL-1 was purchased from Integrated Sciences (Chatswood, NSW, Australia). LPS, purified from Salmonella enterica serotype Minnesota Re 595, was purchased from Sigma-Aldrich

(Castle Hill, NSW, Australia). The CpG-containing oligonucleotide ODN-1688 (5’-

TCCATGACGTTCCTGATGCT-3’) containing phosphorothioate linkages was purchased from

GenScript (Piscataway, NJ, USA). The synthetic dsRNA poly I:C was purchased from Integrated

Sciences (Chatswood, NSW, Australia). Pam3CSK4 was purchased from Life Research Pty Ltd

(Scoresby, VIC, Australia). Zymosan, FSL-1, LPS, poly I:C, CpG DNA and Pam3CSK4 were used at

10 μg/ml, 100 μg/ml, 100 ng/ml, 10 µg/ml, 0.3 μM and 15 ng/mL, respectively, unless otherwise stated. All other chemicals and reagents were from Sigma-Aldrich (Castle Hill, NSW, Australia).

Immunoprecipitation and immunoblotting

Immunoprecipitation and immunoblots were performed as described previously (24). Briefly, cells were lysed in lysis buffer (20 mM Tris pH 7.4, 150 mM NaCl, 1% NP-40, 5% glycerol) and then passed successively through 27-gauge needles. After centrifugation at 14,000 × g for 15 min, the supernatants of lysates were collected and used as input. For immunoprecipitations, cell lysates were incubated with antibody-bound protein G beads for 1 h at 4 °C. Beads were then washed with lysis buffer and bound proteins were solubilized in 2xSDS–PAGE sample buffer. Proteins were separated by 10% SDS-PAGE and analyzed by immunoblotting. For immunoblotting, protein samples were transferred to a polyvinylidene difluoride membrane (BioTrace, NZ), blocked with 5% bovine serum albumin (A4612-25G) from Sigma-Aldrich (Castle Hill, NSW, Australia) for anti-phosphotyrosine blots or 5% skim milk in TBS/0.1% Tween-20 buffer for all the other blots and incubated overnight at

4˚C with primary antibodies. After washing, membranes were incubated with a secondary antibody for 1 h, then developed with X-ray films (FUJIFILM, Tokyo, Japan). Blots were quantified by densitometry of X-ray film (FUJIFILM, Tokyo, Japan) using ImageJ version 1.43 (National Institutes of Health, Maryland, USA).

Cell culture and transfection

Murine bone marrow-derived macrophages (BMMs) were generated from femur and tibia bone marrow extracted from 8-12 week old mice. Once extracted, bone marrow was cultured in the presence of 1 x 104 U/ml recombinant human colony-stimulating factor-1 (CSF-1) (Chiron) for 6-7 days, as described previously (12). BMMs were cultured in the presence of RPMI1640 supplemented with 2 mM L-glutamine (GLUTAMAX, Life Technologies), 10% foetal bovine serum (FBS), 50

U/ml penicillin, 50 µg/ml streptomycin and 1 x 104 U/ml CSF-1 (BMM media). The RAW 264.7 murine macrophage-like cell line was sourced from ATCC. CRISPR/Cas9-mediated SCIMP knockout

RAW264.7 cells have previously been described (12). SCIMP-reconstituted macrophages (WT,

Y58F, Y96F, and Y120F) were generated as previously described (18). Briefly, 5 x 106 cells were

electroporated with 10 µg of plasmid DNA at 240 V, 1000 µF and ∞ Ω. After 48 h of transfection, cells were selected using blasticidin (2 μg/ml) from Thermo Fisher (Carlsbad, CA, USA). SCIMP- deficient and reconstituted RAW 264.7 cells were cultured in RPMI 1640 supplemented with 2 mM/L glutamine (GlutaMAX), 10% heat-inactivated fetal bovine serum, 50 U ml−1 penicillin and 50 μg ml−1 streptomycin (Thermo-Fisher).

Gene expression analysis via RT-qPCR

Gene expression was analysed as previously described (12). Briefly, cells were plated at 1x106 cells/well in 2 ml of BMM media, before they were stimulated as indicated. Cells were washed once in ice cold PBS and lysed in RLT buffer (Qiagen, Australia), after which total RNA was extracted using RNA extraction kits and on column DNase (Qiagen, Australia), as per the manufacturer’s guidelines. 1000 ng of RNA was reverse transcribed into cDNA using Superscript III (Invitrogen) and oligo dT primers. Gene expression was quantified via SYBR® Green (Applied Biosystems) RT-qPCR

(Applied Biosystems 7900HT fast RT-PCR system) using gene-specific primers, with expression normalized to Hprt. Primers used were: Il6 (Forward: CTGCAAGAGACTTCCATCCAGTT;

Reverse: GAAGTAGGGAAGGCCGTGG); IL12b (Forward: GGAAGCACGGCAGCAGAATA;

Reverse: AACTTGAGGGAGAAGTAGGAATGG); Tnf (Forward:

CATCTTCTCAAAATTCGAGTGACAA; Reverse: TGGGAGTAGACAAGGTACAACCC); Ifnb1

(Forward: CCACAGCCCTCTCCATCAAC; Reverse: TGAAGTCCGCCCTGTAGGTG); Scimp

(Forward: ACCTAGCCCCACAGGAAGC; Reverse: GTGGAGCCCGAGATAGCAAA); and Hprt

(Forward: GCAGTACAGCCCCAAAATGG; Reverse: AACAAAGTCTGGCCTGTATCCAA).

siRNA-mediated gene silencing

siRNA knockdown of SCIMP was performed as previously described (12). Day 6 BMMs were harvested, cells were resuspended in complete media at a concentration of 4 × 106 cells/350 μl, and

10 μl 1 M HEPES (tissue culture grade) per ml media was added. 350 μl of the cell suspension was transferred to 0.4 cm electroporation cuvettes and mixed with siRNAs against SCIMP or HDAC1

(control gene) to a final concentration of 0.5 μM or tissue culture grade water (no siRNA control) in a final volume of 400 μl. Cells were electroporated at 240 V, 1,000 μF and ∞ Ω. After electroporation, cells were washed twice, counted and then plated at the appropriate cell number. Cells were treated with indicated stimuli at 24 h post-transfection. Sequences of siRNAs used were: mScimp #1: sense sequence: 5 ′ -AGACAACCCUCAGCUUGGUACUCAU-3 ′ ; antisense sequence: 5 ′ -

AUGAGUACCAAGCUGAGGGUUGUCU-3 ′ ; control (mHdac1 #1): sense sequence: 5 ′ -

GAACUACCCACUGCGAGACGGCAUU-3 ′ ; antisense sequence: 5 ′ -

AAUGCCGUCUCGCAGUGGGUAGUUC-3′.

In vitro infection assays

Infection assays were carried out on day 7 BMMs, 24 h after siRNA knockdown. Cells were plated at

1 x 106 cells/well in 2 ml BMM media, before being washed and replated in antibiotic-free media, then left at 37˚C for a further 2 h. Cells were infected with Escherichia coli strain K-12 MG1655, diluted in antibiotic-free media, at a multiplicity of infection of 100 (MOI: 100). At 1 h post-infection, cells were washed with media containing 200 μg/ml gentamicin to remove extracellular bacteria, then maintained in 20 μg/ml gentamicin until their lysis. Cells were washed with cold PBS and lysed in

350 µl RLT Buffer (Qiagen), after which total RNA was extracted using an RNA extraction kit

(Qiagen), as per manufactures guidelines.

ELISA assays

Enzyme-linked immunosorbent (ELISA) assays were performed to measure levels of TNF, IL-6, IL-

12p40 secreted from RAW264.7 cell lines or SCIMP-silenced BMM. 96-well maxisorp plates

(Thermo Scientific) were coated with 100 μl capture antibody in 0.1 M sodium bicarbonate at pH 9.6 overnight at 4 °C. Plates were then washed with phosphate buffered saline (PBS)/0.05% Tween-20

(PBST), blocked with 200 μl blocking buffer (10% FCS in PBS) overnight at 4 °C and then incubated with 100 μl standards or samples for 2 h at room temperature. Plates were again washed x4 times with

PBST, treated with 50 μl detection antibody and secondary antibody made up in blocking buffer for

1 h at room temperature, washed x6 times with PBST, 100 μl extravidin-peroxidase (1:1,000) was added, then samples were incubated for approximately 30 min at room temperature. Peroxidase activity was measured colorimetrically by adding 50 μl 3,3′5,5′-Tetramethylbenzidine substrate

(Sigma-Aldrich) and the reaction was stopped by the addition of 50 μl of 2 M H2SO4. The absorbance was read at 450 nm using a Powerwave XS plate reader and the sample concentrations were calculated by extrapolation from a quadratic curve analysis of the standards.

Image Analysis Software

Western blots were analysed using ImageJ software (version 1.43; National Institutes of Health,

Maryland, USA). Adobe Photoshop CS6 was also applied to crop regions of interest.

Statistics

Student’s t-tests were used for data with a normal distribution (assessed by Shapiro-Wilk test) and for direct comparison between two experimental groups. Two-way ANOVA was used for multiple comparisons. Sidak’s method was used for analyzing variants of two-way ANOVA during multiple

comparisons. All analyzed experiments used data combined from multiple independent experiments to compute statistical significance. In all statistical analyses, a P value < 0.05 was considered statistically significant. Statistics were calculated using GraphPad Prism version 7.0 (GraphPad

Software, San Diego, CA). Data are shown as arithmetic means + s.e.m., unless otherwise stated.

RESULTS

SCIMP is activated by multiple surface and endosomal TLRs.

SCIMP is a non-TIR domain containing TLR adaptor, which scaffolds multiple effectors (Lyn, Grb2,

Csk, and SLP65). During TLR4 activation, Lyn has been shown to bind to SCIMP constitutively at its proline-rich domain, whereas the other effectors require SCIMP phosphorylation at one of three different tyrosine sites (12, 18, 19). Previously, we reported that SCIMP is phosphorylated by Lyn in response to TLR4 stimulation in macrophages (12, 18). To assess tyrosine phosphorylation of SCIMP in response to agonists of other TLRs, we immunoprecipitated SCIMP-V5 from CRISPR/Cas9- mediated SCIMP knockout RAW264.7 cells stably re-expressing V5-tagged WT SCIMP.

Immunoblotting for phosphotyrosine showed that SCIMP is rapidly phosphorylated upon activation of a number of TLRs, including TLR4, TLR1/2, TLR3 and TLR9 (Fig. 1A-D). Interestingly, the intensity of SCIMP phosphorylation oscillates over a 2 h time course in response to TLR4, TLR3 and

TLR9 activation, whereas only an acute phosphorylation response was apparent after TLR2 activation. Taken together, this shows that SCIMP is phosphorylated downstream of a range of TLRs that detect danger signals at both the cell surface or within endosomes.

SCIMP associates with multiple TLRs

Previously, we showed that SCIMP binds directly to TLR4 in a ligand-dependent fashion (12) and here we further demonstrate the specificity of this interaction by showing that LPS does not promote an association between SCIMP and other TLRs (Supplementary Fig. 1A-C). Next, co- immunoprecipitation experiments were carried out to examine other SCIMP-TLR interactions. These were performed over 2 hour time course to maximize the chance to capture such interactions. The results from Pam3CSK4-induced TLR1/2 activation demonstrate a transient interaction between

SCIMP and TLR2, with this interaction being agonist-dependent and peaking at 5 min post- stimulation (Fig 2A). Similarly, SCIMP rapidly associates with TLR3 after poly I:C stimulation, with this response peaking at 10 mins (Fig. 2B). In addition, the SCIMP and TLR2 interaction was also induced by the TLR2/6 agonist FSL-1 (Fig. 2C) and by the TLR2 agonist zymosan (Fig. 2D). These findings thus show that SCIMP can interact with additional TLRs beyond TLR4, in this case TLR2 and TLR3, reinforcing the need for the interacting TLR to be ligand-activated, and in the case of

TLR2, this can be any ligand activating this TLR. Collectively, these data show that SCIMP binds to multiple TLRs upon agonist-induced activation.

Recruitment of SCIMP effectors by multiple TLR agonists

Previously, we showed that SCIMP recruits Lyn, Grb2, and Csk during TLR4 activation (12). To examine the relevant effector(s) associating with SCIMP in macrophages activated through other

TLRs, we performed co-immunoprecipitation (with a V5 antibody) from macrophages expressing V5- tagged WT SCIMP on the SCIMP-deficient RAW264.7 cell background (18) in response to TLR1/2 activation. Consistent with previous reports by us and others (12, 18, 19), the results from immunoblotting showed that Lyn constitutively and specifically associates with SCIMP (Fig. 3A and

Supplementary Fig. 1D). A phospho-specific antibody recognizing active Src family kinase

(pSrc416) reveals that Pam3CSK4 acutely and transiently activates an SFK in this complex (Fig. 3A), which is in line with, and likely to be, Lyn. In addition, Csk, Grb2 and SLP65 were also co-

immunoprecipitated with SCIMP in Pam3CSK4-activated cells, with the interactions of these effectors being agonist-induced in a rapid and transient response (Fig. 3A). We also examined potential effectors recruited by SCIMP upon TLR3 activation. In this case, Csk binding was ligand- dependent and rapid, whereas Grb2 and SLP65 binding was enhanced by poly I:C stimulation over longer time courses (Fig. 3B). Thus, we reveal that Lyn, Grb2, Csk and SLP65 are recruited by

SCIMP during the activation of different TLRs in macrophages, but the varying temporal nature of these interactions may indicate that effectors have distinct roles as part of each SCIMP-TLR complex.

SCIMP modulates signalling downstream of multiple TLRs in macrophages

Given that SCIMP is activated and recruited to multiple TLRs, we next investigated the role of

SCIMP as a signalling adaptor in these pathways. We used our previously reported CRISPR/Cas9- mediated SCIMP knockout RAW264.7 cell line (12) and cells in which SCIMP was rescued by stable expression of WT-SCIMP-V5 in the SCIMP knockout (KO) background (as shown in Fig. 4) and reported previously (12, 18).

Having previously shown that SCIMP is a TLR4 adaptor that mediates MAPK signalling and IB degradation in response to LPS stimulation (12) and by using SCIMP CRISPR KO cell lines, we compare TLR-mediated signalling events downstream of other TLRs. Read-outs for TLR signalling modules including ERK-1/2, Akt and IB phosphorylation were assessed over a time-course in cells treated with LPS, Pam3CSK4, poly I:C or CpG DNA (Fig. 4 and Fig. 5). Representative gels and densitometry analysis revealed > 60% reduction in phosphorylation of IB in response to all four agonists compared to controls. Similarly, phosphorylation of Akt and ERK1/2 were also modestly but significantly reduced in cells treated with the four agonists. These attenuated signalling responses

were all recovered in the rescue cell line (Fig. 4-5). We thus conclude that loss of SCIMP impairs

TLR2, 3, 4 and 9-mediated signalling responses.

To pinpoint the tyrosine residues on SCIMP required for TLR-mediated signalling, we generated three mutant SCIMP rescue lines in SCIMP knockout cells. The three phosphorylation-deficient

SCIMP mutants (Y58F, Y96F and Y120F) have defects in binding to Grb2, Csk, SLP65, respectively

(18). In response to TLR4 agonists, all cell lines exhibited increased Akt and MAPK phosphorylation

(Fig. 6 A-C). However, the SCIMP Y96F mutant lines did not rescue defects in Akt and Erk1/2 signalling caused by SCIMP deficiency. We also observed that the Y58F and Y120F mutants partially restored these responses when compared to the WT rescue lines. Therefore, phosphorylation of

SCIMP at Y96 is essential for SCIMP-dependent TLR signalling, likely because of Lyn-mediated phosphorylation of Y96 that enables the association with TLR4 (12).

SCIMP is required for selective cytokine production downstream of multiple TLRs

Having examined the effects of SCIMP deletion on TLR-triggered signalling responses, we next examined possible SCIMP’s role in cytokine production downstream of TLRs. In SCIMP-deleted macrophages, levels of secreted IL-6, but not TNF, were decreased upon stimulation with LPS,

Pam3CSK4, poly I:C or CpG DNA (Fig. 7). This defect was recovered in the rescue cell line. The low levels of LPS or CpG-inducible IL-12p40 secretion were also reduced in SCIMP-deleted cells, whilst

Pam3CSK4 or poly I:C-inducible IL-12p40 secretion from RAW264.7 cells could not be detected

(data not shown). To further examine selectivity in cytokine outputs and to gain insights into upstream mechanisms responsible, we also examined TLR3-TRIF-dependent responses. In SCIMP knockout and SCIMP rescue cells, poly I:C-induced TBK1 phosphorylation was similar to that observed for wild type cells (Supplementary Fig. 2A). Next, we examined TLR3-inducible interferon β (IFNβ)

expression in BMMs in which SCIMP had been silenced (Supplementary Fig. 2B). Whereas IL-6 and IL-12p40 expression were significantly reduced and there was a modest trend for reduced TNF expression, inducible IFNβ expression was unaffected. Thus, SCIMP does not appear to affect the

TRIF-TBK1-IFNβ axis downstream of TLR3. Taken together, these results show that SCIMP is a

TLR adaptor required for generating a subset of proinflammatory cytokines, namely IL-6 and IL-

12p40, downstream of multiple TLRs.

SCIMP is required for selective cytokine synthesis during live bacterial challenge.

To assess whether SCIMP-dependent TLR responses were also apparent in the context of microbial challenge and to further confirm the lack of effect on the TRIF-TBK1-IFNβ pathway, we infected

SCIMP-silenced BMMs with E. coli K-12 strain MG1655 (Fig. 8). Consistent with our previous observations showing selective effects on cytokine outputs (Fig. 7, Supplementary Fig. 2B), inducible IL-6 and IL-12p40 gene expression were reduced in SCIMP-depleted BMMs, whereas inducible IFNβ was unaffected (Fig. 8). In these experiments, we did observe a modest reduction in

TNF gene expression, although this effect was not significant. Thus, SCIMP also provides specificity to inflammatory responses in macrophages responding to bacterial challenge.

DISCUSSION

Macrophages utilise a wide spectrum of TLRs in order to elicit responses to a range of pathogens and danger signals. Upon activation, TLRs trigger complex but finely-controlled signalling pathways that shape the profile of secreted inflammatory cytokines. Previously, we reported SCIMP as a scaffolding adaptor for the Src family tyrosine kinase Lyn, which enables TLR4 phosphorylation and selective cytokine production (12, 18). Here, by exploring additional PAMPs and agonists, we reveal that

SCIMP is acutely phosphorylated in response to the activation of both cell surface and endosomal

TLR family members, including TLR1/2, TLR2/6, TLR4 and TLR9 (Supplementary Fig. 3).

Mechanistically, we found that tyrosine phosphorylation of SCIMP allows it to recruit Src family kinases (SFKs), most likely Lyn, and its effectors Grb2, Csk, and SLP65 to the TLR complex. Our data highlight that the need for SCIMP to exert signalling responses that drive selective cytokine production (IL-6 and IL-12p40) downstream of multiple TLRs and pathogens, but not TNF or TLR3-

TRIF dependent IFN production. Therefore, we present SCIMP as a universal TLR adaptor and mediator of downstream responses in activated macrophages.

It is known that agonists of multiple TLRs can trigger their phosphorylation (7, 10, 25), However, our knowledge about the precise molecular mechanisms involved is, in many cases, quite limited.

Phosphorylation of the dimerised TLRs is a very early event triggered by ligand binding and we show that SCIMP is poised for rapid interaction with TLRs that are activated at the plasma membrane and also with TLRs 3 and 9 which are believed to encounter their intracellular ligands for activation on endosomal membranes (1). Transmembrane SCIMP cycles to and from the surface (12), presumably bringing it into contact with TLRs at both cell surface and endosomal stages, but further investigation of these encounters is warranted in the future. In TLR4 signalling, we previously showed that SCIMP is required for TLR4 phosphorylation and signalling competence (12). Y674 in human TLR4, as well as some other TIR domain tyrosines, are conserved in human and mouse TLR4, and across multiple

TLR family members (7). It has previously been shown that Y674 within the TIR domain is required for TLR4-mediated activation of MAPKs and NF-κB, as well as expression of proinflammatory cytokines, in HEK 293 cells (26, 27). In line with such data, our results also showed that deletion of

SCIMP in macrophages leads to reduced phosphorylation of MAPKs, Akt and IκB, as well as reduced production of the proinflammatory cytokines IL-6 and IL-12p40. Thus, our finding that SCIMP acts

downstream of several TLRs broadens its potential roles in responses to a wide range of pathogens, and suggests that SCIMP-mediated tyrosine phosphorylation may be shared across multiple TLRs.

How SCIMP imparts specificity to a range of TLR-driven inflammatory cytokine responses remains an important question. SCIMP acts as proximal TLR adaptor which directly tethers to TLRs and regulates TLR phosphorylation (12). Phosphorylation of specific TLR tyrosine residues are known to differentially modulate TLR signalling (13). In addition, SCIMP scaffolds the effector proteins Grb2,

Csk and SLP65 (18) and the role of these effectors in TLR signalling is still unclear. Grb2 is linked to

RAS activation (28) and MAPK signalling responses downstream of the B cell antigen receptor (19).

SLP65/76 adaptors are known to be involved in PLCγ1/2 activation (29) and in turn they activate Ras exchange factors, RasGRP proteins for T cell antigen receptors, B cell antigen receptors and other immunoreceptor signalling (30). We have shown that Y96 phosphorylation of SCIMP is required for

TLR4-mediated signalling (Fig. 6). Lyn-mediated phosphorylation of SCIMP at Y96 is required for

TLR binding but is also essential for recruitment of Csk, which phosphorylates Lyn at Y507 to inactivate its kinase activity (31). Therefore, during TLR activation, Lyn-mediated SCIMP phosphorylation likely allows the recruitment of Csk, which may in turn switch off Lyn kinase activity and result in a transient Lyn signalling response despite the constitutive association of this kinase with SCIMP. Consistent with this hypothesis, we observed an interesting oscillation of SCIMP phosphorylation over a time course in response to activation of several TLRs (Fig 1. A-D). Thus,

SCIMP may act as an autonomous on/off switch by scaffolding both positive and negative regulators of inflammatory signalling responses. Further studies are required to elucidate detailed roles for each of the SCIMP effectors and TLR phosphorylation in TLR pathways.

TLRs can elicit varied cytokine outputs through different adaptors and signalling mechanisms. The nature of the cytokines produced shapes appropriate and effective immune defence in the first instance, and then subsequently resolution of inflammation and the onset of wound healing to restore homeostasis. TLR-induced signalling must ensure measured release of proinflammatory cytokines, which recruit and activate cells of the adaptive immune system, and then regulatory or anti- inflammatory mediators that actively switch off inflammation. Our studies implicate SCIMP as a universal transmembrane adaptor for activation of multiple TLRs. This brings this family of innate immune receptors in line with T and B cell receptors, which are well known for partnering with

TRAP family members to control signalling outputs in lymphocytes. In summary, SCIMP is an adaptor that imparts specificity to multiple TLRs to shape cytokine responses upon danger sensing.

FIGURE LEGENDS

Figure 1 SCIMP is tyrosine phosphorylated upon activation of both surface and endosomal

TLRs. CRISPR/Cas9-mediated SCIMP knockout RAW264.7 cells stably expressing V5-tagged WT

SCIMP were treated with LPS (A), Pam3CSK4 (B), poly I:C (C) or CpG DNA (D) over a time- course. Western blotting was used to access total SCIMP and SCIMP tyrosine phosphorylation by immunoprecipitation of SCIMP-V5 (anti-V5), followed by immunoblotting for phosphotyrosine.

Relative chemiluminescence of phospho-SCIMP/total SCIMP was assessed at different time points in

the indicated cell populations. Graphs, which show intensity relative to the maximal response for individual ligands, represent pooled data from n=3 experiments (mean+s.e.m.). Blots are representative of 3 independent experiments.

Figure 2 SCIMP associates with multiple TLRs. SCIMP knockout RAW264.7 cells stably expressing SCIMP-V5 were stimulated with Pam3CSK4 (A) or poly I:C (B) at the indicated times, and immunoprecipitation was performed on cell lysates using the V5 antibody. Western blotting was then used to assess levels of TLR2 and TLR3. SCIMP knockout RAW264.7 cells stably expressing

V5-tagged WT SCIMP were treated with FSL-1 (C) or zymosan (D) for 5 mins. TLR2 levels were then assessed by immunoprecipitation of SCIMP-V5 (anti-V5), followed by immunoblotting against

TLR2. Blots are representative of 3 independent experiments.

Figure 3 Recruitment of SCIMP effectors by multiple TLR agonists. SCIMP knockout

RAW264.7 cells stably expressing SCIMP-V5 were treated with Pam3CSK4 (A) or poly I:C (B) at the indicated times, after which the V5 antibody was used for immunoprecipitation. Western blotting was then used to assess the levels of pSFK(Y416), Lyn, Grb2, Csk, SLP65, and V5-SCIMP. Graphs for Grb2 and SLP65 represent pooled data from n=3 experiments (mean+s.e.m.). Blots are representative of 3 independent experiments.

Figure 4 SCIMP is required for optimal TLR4 and TLR2 signalling responses. WT RAW264.7 control cells, CRISPR/Cas9-mediated SCIMP knockout RAW264.7 cells, and SCIMP-V5 rescue in

SCIMP KO cells were used to assess TLR signalling responses. Representative immunoblots showing levels of phospho-IκB, phospho-Akt, phospho-Erk1/2 and SCIMP at 0, 30 and 60 min post-LPS stimulation (A) or post Pam3CSK4 stimulation (B). Relative chemiluminescence of phospho-IκB, phospho-ERK1/2 and phospho-pAkt was assessed at 30 min and 60 min post-LPS stimulation (C) or post-Pam3CSK4 stimulation (D) in the indicated cell populations. Graphs, which display data relative to the 60 min stimulated wild type control, represent pooled data from n=3 experiments

(mean+s.e.m.). Student-t test was used for statistical analysis. (*P<0.05, **P<0.01, ***P<0.001 and

****P<0.0001).

Figure 5 SCIMP is required for optimal TLR3 and TLR9 signalling responses. WT RAW264.7 control cells, CRISPR/Cas9-mediated SCIMP knockout RAW264.7 cells, and SCIMP-V5 rescue in

SCIMP KO cells were used to assess TLR signalling responses. Representative immunoblots showing levels of phospho-IκB, phospho-Akt, phospho-Erk1/2, and SCIMP at 0, 30 and 60 min post-poly I:C stimulation (A) or post CpG DNA stimulation (B). Relative chemiluminescence of phospho-IκB,

phospho-ERK1/2 and phospho-Akt was assessed at 30 min and 60 min post-poly I:C stimulation (C) or post CpG DNA stimulation (D) in the indicated cell populations. Graphs, which display data relative to the 30 or 60 min stimulated wild type control, represent pooled data from n=3 experiments

(mean+s.e.m.). Student-t test was used for statistical analysis. (**P<0.01 and ****P<0.0001).

Figure 6 Effect of SCIMP mutations on TLR4 signalling. (A) CRISPR/Cas9-mediated SCIMP- knockout RAW264.7 cells, and SCIMP-knockout cells rescued with SCIMP WT, Y58F, Y96F or

Y120F mutants were used to assess TLR4 signalling responses. Representative immunoblots showing levels of phospho-Akt, phospho-Erk1/2, GAPDH and SCIMP at 0 and 60 min post-LPS stimulation are shown (control and LPS-stimulated samples are cropped from the same immunoblot). (B) Relative chemiluminescence of phospho-ERK1/2 and phospho-Akt, relative to GAPDH, was assessed at

60 min post-LPS stimulation in the indicated cell populations. Graphs, which display data relative to

GAPDH, represent pooled data from n=4 experiments (mean+s.e.m.). Student-t test was used for statistical analysis. (*P<0.05 and **P<0.01).

Figure 7 SCIMP is required for maximal production of a subset of cytokines downstream of multiple TLRs. TNF, IL-6 and IL-12p40 protein levels were assessed by ELISA in WT cells,

SCIMP-deleted macrophages and rescue cell lines, treated for 24 h with LPS (A), Pam3CSK4 (B), poly I:C (C) or CpG DNA (D). Graphs in (A–D) represent pooled data of 4 independent experiments.

The data was analyzed by Two-way ANOVA, followed by Sidak’s post-test for multiple comparisons.

(*P<0.05 and ****P<0.0001).

Figure 8 SCIMP mediates selective inflammatory responses downstream of microbial challenge.

SCIMP was silenced by siRNA in BMMs, after which cells were challenged with E. coli K-12 strain

MG1655 for 60 mins, prior to gentamicin exclusion to remove extracellular bacteria. The levels of mRNAs encoding IL-6, IL-12p40, TNF, IFNβ and SCIMP at 4 h post-stimulation were assessed by qRT-PCR. Data are pooled from n=3 independent experiments (mean+s.e.m.). The data was analyzed by Two-way ANOVA, followed by Sidak’s post-test for multiple comparisons. (*P<0.05, **P<0.01,

***P<0.001 and ****P<0.0001).

References

1. O'Neill LA & Bowie AG (2007) The family of five: TIR-domain-containing adaptors in Toll-like receptor signalling. Nat Rev Immunol 7(5):353-364.

2. Rossol M, et al. (2011) LPS-induced cytokine production in human monocytes and macrophages. Crit Rev Immunol 31(5):379-446.

3. O'Neill LA, Fitzgerald KA, & Bowie AG (2003) The Toll-IL-1 receptor adaptor family grows to five members. Trends Immunol 24(6):286-290.

4. Kumar H, Kawai T, & Akira S (2009) Pathogen recognition in the innate immune response. Biochem J 420(1):1-16.

5. Vercammen E, Staal J, & Beyaert R (2008) Sensing of viral infection and activation of innate immunity by toll-like receptor 3. Clin Microbiol Rev 21(1):13-25.

6. Bonham KS, et al. (2014) A promiscuous lipid-binding protein diversifies the subcellular sites of toll-like receptor signal transduction. Cell 156(4):705-716.

7. Medvedev AE, et al. (2007) Role of TLR4 tyrosine phosphorylation in signal transduction and endotoxin tolerance. J Biol Chem 282(22):16042-16053.

8. Iwasaki A & Medzhitov R (2004) Toll-like receptor control of the adaptive immune responses. Nat Immunol 5(10):987-995.

9. Chun J & Prince A (2006) Activation of Ca2+-dependent signaling by TLR2. J Immunol 177(2):1330-1337.

10. Arbibe L, et al. (2000) Toll-like receptor 2-mediated NF-kappa B activation requires a Rac1-dependent pathway. Nat Immunol 1(6):533-540.

11. Sarkar SN, Smith HL, Rowe TM, & Sen GC (2003) Double-stranded RNA signaling by Toll- like receptor 3 requires specific tyrosine residues in its cytoplasmic domain. J Biol Chem 278(7):4393-4396.

12. Luo L, et al. (2017) SCIMP is a transmembrane non-TIR TLR adaptor that promotes proinflammatory cytokine production from macrophages. Nat Commun 8:14133.

13. Chattopadhyay S & Sen GC (2014) Tyrosine phosphorylation in Toll-like receptor signaling. Cytokine Growth Factor Rev 25(5):533-541.

14. Miller YI, Choi SH, Wiesner P, & Bae YS (2012) The SYK side of TLR4: signalling mechanisms in response to LPS and minimally oxidized LDL. Br J Pharmacol 167(5):990- 999.

15. Curson JEB, Luo L, Sweet MJ, & Stow JL (2018) pTRAPs: Transmembrane adaptors in innate immune signaling. J Leukoc Biol.

16. Chum T, et al. (2016) The role of palmitoylation and transmembrane domain in sorting of transmembrane adaptor proteins. J Cell Sci 129(15):3053.

17. Stepanek O, Draber P, & Horejsi V (2014) Palmitoylated transmembrane adaptor proteins in leukocyte signaling. Cell Signal 26(5):895-902.

18. Luo L, et al. (2017) Development of SH2 probes and pull-down assays to detect pathogen-induced, site-specific tyrosine phosphorylation of the TLR adaptor SCIMP. Immunol Cell Biol 95(6):564-570.

19. Draber P, et al. (2011) SCIMP, a transmembrane adaptor protein involved in major histocompatibility complex class II signaling. Mol Cell Biol 31(22):4550-4562.

20. Aki D, et al. (2005) Modulation of TLR signalling by the C-terminal Src kinase (Csk) in macrophages. Genes Cells 10(4):357-368.

21. Fusaki N, et al. (2000) BLNK is associated with the CD72/SHP-1/Grb2 complex in the WEHI231 cell line after membrane IgM cross-linking. Eur J Immunol 30(5):1326-1330.

22. Fu C, Turck CW, Kurosaki T, & Chan AC (1998) BLNK: a central linker protein in B cell activation. Immunity 9(1):93-103.

23. Johmura S, et al. (2003) Regulation of Vav localization in membrane rafts by adaptor molecules Grb2 and BLNK. Immunity 18(6):777-787.

24. Luo L, et al. (2014) Rab8a interacts directly with PI3Kgamma to modulate TLR4-driven PI3K and mTOR signalling. Nat Commun 5:4407.

25. Sarkar SN, et al. (2004) Novel roles of TLR3 tyrosine phosphorylation and PI3 kinase in double-stranded RNA signaling. Nat Struct Mol Biol 11(11):1060-1067.

26. Jefferies CA, et al. (2003) Bruton's tyrosine kinase is a Toll/interleukin-1 receptor domain-binding protein that participates in nuclear factor kappaB activation by Toll-like receptor 4. J Biol Chem 278(28):26258-26264.

27. Chen LY, et al. (2003) Involvement of protein tyrosine kinase in Toll-like receptor 4- mediated NF-kappa B activation in human peripheral blood monocytes. Am J Physiol Lung Cell Mol Physiol 284(4):L607-613.

28. Jia XM, et al. (2014) CARD9 mediates Dectin-1-induced ERK activation by linking Ras- GRF1 to H-Ras for antifungal immunity. J Exp Med 211(11):2307-2321.

29. Koretzky GA, Abtahian F, & Silverman MA (2006) SLP76 and SLP65: complex regulation of signalling in lymphocytes and beyond. Nat Rev Immunol 6(1):67-78.

30. Jun JE, Rubio I, & Roose JP (2013) Regulation of ras exchange factors and cellular localization of ras activation by lipid messengers in T cells. Front Immunol 4:239.

31. Okada M, Nada S, Yamanashi Y, Yamamoto T, & Nakagawa H (1991) CSK: a protein- tyrosine kinase involved in regulation of src family kinases. J Biol Chem 266(36):24249- 24252.

Acknowledgements:

This work was supported by National Health and Medical Research Council of Australia (NHMRC) project grants (APP1101072 and APP1159106), an Australian Research Council Discovery Early

Career Award (DE180100524) and an Early Career Grant from the University of Queensland

(UQECR1834509). MJS is supported by an NHMRC Senior Research Fellowship (APP1107914).

Author contributions:

We thank members of the Stow and Sweet laboratories for their technical help. L.Luo designed and performed the experiments, and analysed the data. J.E.B.C performed experiments in Fig 7 and S3B.

L.Liu and N.T performed experiments in Fig 4 and 5. A.A.W generated the CRISPR/Cas9-mediated

SCIMP knockout RAW264.7 cells. R.M.L performed the experiment in Fig S1D. L.Luo, M.J.S. and

J.L.S supervised the study, helped interpret the data and participated in writing the manuscript.

Supplementary Figure 1 The SCIMP and TLR4 interaction is specific and context-dependent.

(A) Co-immunoprecipitation of V5-SCIMP and TLR4 in SCIMP-deleted cells stably expressing

SCIMP-V5. Western blotting was used to assess the interaction between SCIMP and TLR4 in the presence or absence of LPS. (B) Co-immunoprecipitation of V5-SCIMP and TLR2. Representative

Western blot was probed to assess the interaction between SCIMP and TLR2 in the presence or absence of LPS from SCIMP-deleted cells stably expressing SCIMP-V5. (C) Co-immunoprecipitation of SCIMP-V5 and TLR3. Representative Western blot was probed for SCIMP and TLR3 in the presence or absence of LPS from SCIMP-deleted cells stably expressing SCIMP-V5. (D) Co- immunoprecipitation of SCIMP-V5 and Lyn. Representative Western blot was probed for SCIMP-V5

and Lyn in the presence LPS from SCIMP-deleted cells stably expressing SCIMP-V5. Myc antibody was used as a negative binding control. All data are representative of 3 independent experiments.

Supplementary Figure 2 SCIMP does not regulate the TRIF-IFNβ axis. (A) RAW264.7 control cells, CRISPR/Cas9-mediated SCIMP knockout RAW264.7 cells and SCIMP-knockout cells rescued with SCIMP-WT were used to assess TLR3-TRIF signalling responses. Representative immunoblots showing levels of phospho-TBK1, GAPDH and SCIMP at 0, 30 and 60 min post-poly: IC stimulation.

(B) SCIMP was silenced by siRNA in BMMs, after which cells were stimulated with poly I:C. The levels of mRNAs encoding IL-6, IL-12p40, IFN, TNF and SCIMP at 4 h post-infection were assessed by qRT-PCR. Data are pooled from n=3 independent experiments (mean+s.e.m.).

Supplementary Figure 3 Model for SCIMP-mediated TLR signalling. SCIMP presents Lyn and its other effectors (Grb2, Csk and SLP65) to multiple TLRs for their phosphorylation, Akt and MAPK signalling, NFB activation and production of pro-inflammatory cytokine IL-6 and IL-12p40.