Leucine-rich repeats and calponin homology containing 4 (Lrch4) regulates the innate immune response

Toll-like receptors (TLRs) are pathogen-recognition receptors that trigger the innate immune response. Recent reports have identified accessory proteins that provide essential support to TLR function through ligand delivery and receptor trafficking. Herein, we introduce leucine-rich repeats (LRRs) and calponin homology containing 4 (Lrch4) as a novel TLR accessory protein. Lrch4 is a membrane protein with nine LRRs in its predicted ectodomain. It is widely expressed across murine tissues and has two expression variants that are both regulated by lipopolysaccharide (LPS). Predictive modeling indicates that Lrch4 LRRs conform to the horseshoe-shaped structure typical of LRRs in pathogen-recognition receptors and that the best structural match in the protein database is to the variable lymphocyte receptor of the jawless vertebrate hagfish. Silencing Lrch4 attenuates cytokine induction by LPS and multiple other TLR ligands and dampens the in vivo innate immune response. Lrch4 promotes proper docking of LPS in lipid raft membrane microdomains. We provide evidence that this is through regulation of lipid rafts as Lrch4 silencing reduces cell surface gangliosides, a metric of raft abundance, as well as expression and surface display of CD14, a raft-resident LPS co-receptor. Taken together, we identify Lrch4 as a broad-spanning regulator of the innate immune response and a potential molecular target in inflammatory disease.


Toll-like receptors (TLRs) are pathogen-recognition receptors that trigger the innate immune response. Recent reports have identified accessory proteins that provide essential support to TLR function through ligand delivery and receptor trafficking. Herein, we introduce leucine-rich repeats (LRRs) and calponin homology containing 4 (Lrch4) as a novel TLR accessory protein. Lrch4 is a membrane protein with nine LRRs in its predicted ectodomain. It is widely expressed across murine tissues and has two expression variants that are both regulated by lipopolysaccharide (LPS). Predictive modeling indicates that
Lrch4 LRRs conform to the horseshoe-shaped structure typical of LRRs in pathogen-recognition receptors and that the best structural match in the protein database is to the variable lymphocyte receptor of the jawless vertebrate hagfish. Silencing Lrch4 attenuates cytokine induction by LPS and multiple other TLR ligands and dampens the in vivo innate immune response. Lrch4 promotes proper docking of LPS in lipid raft membrane microdomains. We provide evidence that this is through regulation of lipid rafts as Lrch4 silencing reduces cell surface gangliosides, a metric of raft abundance, as well as expression and surface display of CD14, a raft-resident LPS co-receptor. Taken together, we identify Lrch4 as a broad-spanning regulator of the innate immune response and a potential molecular target in inflammatory disease.
Toll-like receptors (TLRs) 4 are an evolutionarily conserved family of pattern recognition receptors that are thought to detect select pathogen-and damage-derived (i.e. host) molecules in part through leucine-rich repeat (LRR) motifs in the TLR ectodomain (1). Ligation of TLRs, whether on the plasma membrane (TLR1, -2, -4, -5, and -6) or within endosomes (TLR3, -7, -8, and -9), triggers a complex series of signaling events through adaptor proteins and kinases, culminating in the activation of NF-B and interferon regulatory factors (IRFs), master transcription factors that orchestrate the innate immune response via inducing pro-inflammatory cytokines and type I interferons. TLRs are pivotal for host defense but can also mediate inflammatory and autoimmune diseases (2,3); thus, an improved understanding of their molecular regulation is expected to enrich our insight into human disease pathogenesis and to reveal new therapeutic targets (4).
Recent literature has revealed a growing number of accessory proteins that play essential roles in supporting TLR function (5). Thus, MD-2 assists TLR4 in binding of lipopolysaccharide (LPS) (6), RP105 regulates TLR4 signaling in a cell type-dependent manner (7,8), CD14 regulates ligand interactions for multiple TLRs (5), and TRIL is thought to mediate ligand delivery to TLR3 and TLR4 (9). However, additional regulators such as GRP94 and PRAT4A serve as chaperones for multiple TLRs via facilitating proper protein folding and maturation (10,11), whereas Unc93b1 interacts with multiple nucleic acid-sensing TLRs to mediate their delivery to endosomes (12).
Here, we introduce leucine-rich repeats and calponin homology containing protein 4 (Lrch4) as a novel accessory protein that regulates signaling by multiple TLRs. Lrch4 is predicted to be a single-pass transmembrane protein with approximately nine LRRs and a calponin homology (CH) motif in its ectodomain. It is widely expressed across murine tissues and is regulated by LPS. Lrch4 silencing attenuates cytokine induction by a wide array of TLR ligands in murine and human cells and also reduces inflammatory responses to LPS in vivo. Lrch4 co-precipitates with biotin-LPS from treated macrophages, suggesting interaction with LPS. Consistent with this, Lrch4 silencing reduces cell surface binding of LPS and alters the pattern of LPS deposition on the macrophage, reducing LPS localization to rafts. Taken together, we identify Lrch4 as a broad-spanning regulator of the innate immune response with potential as a therapeutic target in inflammatory disease.

Sequence, structural, and expression characterization of Lrch4
We recently identified Lrch4 in a proteomic screen as a protein increased in lipid raft microdomains of macrophages upon LPS exposure, suggestive of a potential role in TLR4 signaling (13). Lrch4 (Gene ID: 231798) resides on chromosome 5 in the murine genome (7q22 in the human genome) and is predicted to have a 3,078-bp ORF (18 exons) that encodes a 680-amino acid (AA) protein (ϳ73 kDa) with an pI of ϳ7.5. Sequence alignment using CLUSTAL (14) indicates a high degree of homology between murine Lrch4 and human LRCH4 (85.7%) and moderate homology between the human and zebrafish homologues (Fig. S1), suggesting significant evolutionary conservation. AA sequence-based prediction of conserved motifs (UniProt) in conjunction with transmembrane prediction (TMpred (15) and Philius (16)) indicates that Lrch4 has an ectodomain composed of a 19 AA N-terminal signal peptide followed by nine LRRs (each 21-23 AA in length), a central disordered region, and a CH motif; this is then followed by a transmembrane domain (TMD) and a short cytoplasmic tail (Fig. 1A).
Recent revisions of the NCBI and ENSEMBL database suggest that Lrch4 has splice variants. The transcript support level is a method used by ENSEMBL to rate well-supported versus poorly supported transcript models. Three Lrch4 transcripts are identified at transcript support level 1 (i.e. all splice junctions of the transcript are supported by at least one nonsuspect mRNA). However, one of these (Lrch4-004) has incomplete 5Ј and 3Ј coding DNA sequences. We therefore focused on Lrch4-001 (referred to herein as variant 1) and Lrch4-002 (referred to as variant 2). Lrch4 variant 1 (680 AAs) is as described above, whereas variant 2 (649 AAs) is truncated C-terminal to the CH and omits the TMD, consistent with a soluble protein. Physiological expression of the predicted variants has not been verified to our knowledge.
LRRs are 20 -30-AA motifs that exist in thousands of proteins across phylogeny, typically appearing as tandem repeats that together form a solenoid-shaped domain thought to mediate protein-ligand and protein-binding (17). Individual LRRs have been grouped based on sequence into seven categorical types, although the functional correlates of these categories remain unclear (18). Alignment of LRR domains between A, linear domain diagram of Lrch4 protein, generated using MyDomains Image Creator (https://prosite.expasy.org/; please note that the JBC is not responsible for the long-term archiving and maintenance of this site or any other third party hosted site). Lrch4 has a predicted ectodomain with an N-terminal 19-amino acid signal peptide (Sig) followed by nine LRRs, a calponin homology (CH) domain, a transmembrane domain (TM), and a short cytoplasmic tail. B, ribbon diagram model of the LRR domain of Lrch4 from BioSerf (green) compared with the best match in the Protein Data Bank (4PSJ, a synthetic construct of hagfish variable lymphocyte receptor; cyan) according to pGenThreader. C, expression of Lrch1, 2, 3, and 4 was profiled by qRT-PCR across 14 murine (C57BL/6J) tissues. D, the subcellular distribution of native Lrch4 within RAW 264.7 macrophages was biochemically profiled by immunoblotting of the fractions shown. Flotillin-1 serves as a membrane marker, p38 serves as a cytoplasmic marker, and HDAC serves as a nuclear marker. S, soluble; I, insoluble. E, the subcellular distribution of GFP-Lrch4 in TLR4-MD2-CD14-HEK293 cells was examined by confocal microscopy (40ϫ objective, oil immersion; additional 50% zoom) in relation to the nuclear stain DAPI and the plasma membrane stain CellMask TM (left). Overlay of the three signals is shown in cellular cross-section using Zeiss Zen software (middle). By contrast, stably expressed tGFP vector control shows substantial overlay with DAPI (right). Images are representative of two independent experiments, both involving Ͼ10 high-power fields.

Lrch4 regulates TLRs
murine Lrch4 and human LRCH4 suggests a very high degree of homology (Table S1), with manual inspection indicating that the majority of murine Lrch4 LRRs fall into the "plant-specific" LRR category, as was recently reported for human Lrch4 in a genome-wide survey of LRR proteins (18). Using the threading algorithm pDomTHREADER (19), the best structural match in the protein database to the LRR domain of Lrch4 is a protein derived from the variable lymphocyte receptor of the jawless vertebrate, the hagfish. The LRR region matches well to known LRR structures, whereas there is less confidence that the N and C termini are modeled accurately. As shown in Fig. 1B, predictive modeling of Lrch4 (green) based on the hagfish protein (cyan) yields the horseshoe-shaped structure typical of LRRs.
CH domains play a role in actin binding, although they may mediate additional interactions, especially when found singly rather than in tandem (20). A recent phylogenetic analysis indicated that the convergence of LRRs and CH domains in proteins (i.e. LRCH proteins) occurs rarely and only in animals, with just one protein (LRCH) in Drosophila melanogaster and four (Lrch1-4) in both Mus musculus and Homo sapiens (21). Interestingly, despite the very high degree of homology between murine Lrch4 and human LRCH4, sequence alignment reveals modest homology between murine Lrch4 and Lrch1 (35%), Lrch4 and Lrch2 (34%), and Lrch4 and Lrch3 (43%) (data not shown), suggesting divergent function among Lrch proteins.
Expression profiling of transcripts for the four Lrch family members indicates that Lrch4 (detected using primers common to the two variants) is widely expressed across 14 murine tissues, with the most abundant expression in spleen, testes, thymus, intestine, and blood (Fig. 1C). As shown in Fig. 1D, immunoblotting of murine macrophage subcellular fractions for endogenous Lrch4 (using an antibody targeting a region common to variants 1 and 2) confirms that it is predominantly present in the membrane, with lesser expression in the cytoplasm and soluble nuclear fraction, and is detected as a single band in all fractions. Consistent results showing a single band in both RAW 264.7 and primary murine macrophage lysates were obtained with both a commercial anti-Lrch4 antibody and an anti-Lrch4 antibody raised by our laboratory. Given that only variant 1 has a predicted TMD, this finding suggests that, at the protein level, variant 1 is expressed much more highly than variant 2 within the cell. Microscopy of GFP-Lrch4 in HEK293 cells reveals staining that includes a cytoplasmic-type pattern, but formal analysis indicates a high degree of overlap with the plasma membrane stain CellMask TM (Life Technologies) (22), in particular in cells with low-medium forced expression, consistent with substantial localization to the plasma membrane (Fig. 1E). By contrast, the tGFP vector (control) displays a very different distribution, largely overlapping the nuclear DAPI stain (Fig. 1E).

Lrch4 regulates cytokine induction by multiple TLRs
Based on its increase in rafts in response to LPS (13), its predicted receptor-like features, and its LRRs-a ligand-binding motif common to the ectodomain of all TLRs, as well as some TLR accessory proteins (CD14, RP105) (18)-we hypothesized that Lrch4 regulates TLR4 activation by LPS. To address this, we generated two stable Lrch4 lentiviral shRNA knockdown RAW 264.7 macrophage lines in parallel with a scrambled (Scr) lentiviral shRNA control line. As shown in Fig. 2A, compared with the Scr line, both knockdown lines achieved Ն50% silencing of Lrch4 mRNA (using probes common to both variants) and protein. Although this degree of knockdown is somewhat modest, it is consistent with past reports using RNAi in macrophages and likely reflects the relatively refractory nature of macrophages to transfection/transduction (23). Using variant-specific primers, we confirmed that both Lrch4 variants are expressed in the macrophage and that both are knocked down by Lrch4 shRNA (Fig. 2B). Of interest, LPS itself induced down-regulation of mRNA for Lrch4 variants 1 and 2 ( Fig. 2, A and B).
Confirming a role for Lrch4 in regulation of TLR4 signaling, both Lrch4 knockdown lines induced significantly less TNF␣ than the Scr line in response to two doses of LPS (Fig. 2C). A similar reduction in LPS-induced TNF␣ was observed in two Lrch4 disruption clones produced by CRISPR-Cas9 (Fig. 2D). Given that some accessory proteins are reported to regulate signaling by multiple TLRs (e.g. CD14, TRIL (5)), we next tested a role for Lrch4 in additional TLR cascades. Lrch4 knockdown attenuated TNF␣ induction by the TLR1/2 ligand Pam3CSK4 (Fig. 2E), as well as by Pam2CSK4 (TLR2/6), imiquimod (TLR7), and ODN2395 (TLR9) (Fig. S2), indicating a broadspanning role for Lrch4 in regulating plasmalemmal and endosomal TLRs. Induction of granulocyte-colony-stimulating factor (G-CSF) by both LPS and Pam3CSK4 was also reduced in Lrch4 knockdown cells, indicating that Lrch4 regulates multiple cytokines downstream of TLR2 and TLR4 (Fig. 2E). To test Lrch4 in human cells, we next silenced endogenous Lrch4 in HEK293 cells using an siRNA approach (Fig. 2F). HEK293 cells stably expressing either TLR4/MD-2/CD14 or TLR2 were transfected with Lrch4 siRNA or Scr siRNA and then exposed to LPS or Pam3CSK4, respectively. In preliminary experiments, we found that, in response to these stimuli, these cells produced very little TNF␣ (data not shown); thus, we surveyed for IL-8 protein production instead, as reported by others (24). As shown in Fig. 2G, Lrch4 siRNA attenuated IL-8 induction by both ligands in HEK293 cells, providing further support that the shRNA results in macrophages are unlikely to reflect offtarget effects. Confirming some selectivity for Lrch4 function, Lrch4 silencing did not affect induction of IL-8 by HEK293 cells in response to stimulation with TNF␣ ( Fig. 2H).

Lrch4 regulates both MyD88-dependent and MyD88-independent TLR4 signaling
We next focused on Lrch4 function in the TLR4 pathway. Moving further "upstream" of cytokine protein expression, we confirmed that Lrch4 knockdown also attenuates transcript expression of TNF␣ and G-CSF in response to LPS (Fig. 3, A and B). This suggested that Lrch4 may regulate LPS induction of pro-inflammatory cytokines at or upstream of transcription. The TLR4 cascade bifurcates immediately downstream of the receptor into two signaling arms defined by alternate usage of the cytoplasmic adaptors, myeloid differentiation primary response 88 (MyD88), and Toll/ interleukin-1 receptor domain-containing adapter-induc-

Lrch4 regulates TLRs
ing interferon-␤ (TRIF) (1). These so-called "MyD88-dependent" and "MyD88-independent" pathways lead to downstream induction of distinct cytokines, although it is thought that most cytokines likely have some input from both adaptor pathways (25). Expression profiling of MyD88-dependent ( Fig. 3C) and -independent ( Fig. 3D) cytokines in LPSstimulated macrophages after Lrch4 silencing indicated a role for Lrch4 in output from both adaptor arms, consistent with it regulating the proximal TLR4 pathway. However, differences in Lrch4 dependence were noted across cytokines, with some cytokines from both adaptor pathways showing dramatic reduction in Lrch4-silenced cells (e.g. IL-10, MCP-1), whereas other cytokines showed a more modest dependence on Lrch4 (e.g. IL-1␤, IP-10). Taken together, these findings suggest that Lrch4 acts upstream of the MyD88/ TRIF bifurcation in the TLR4 cascade.

Lrch4 regulates early signaling in the TLR4 cascade
Moving further upstream in the TLR4 cascade to confirm more directly whether Lrch4 regulates early signaling re-sponses to LPS, we next evaluated activation of the transcription factors NF-B and IRF3. NF-B is activated by both an early MyD88-dependent and late MyD88-independent (TRIFdependent) pathway after LPS, whereas IRF3 activation is generally thought to be dependent upon the TRIF pathway (26). As shown in Fig. 4A, activation of NF-B in macrophage nuclear isolates was attenuated in Lrch4-silenced cells at both 15 and 30 min after LPS. LPS-induced NF-B luciferase was also reduced in Lrch4-silenced RAW 264.7 macrophages (Fig. 4B). Finally, PO 4 -IRF3 was also reduced in the nuclear fraction of Lrch4silenced macrophages (Fig. 4C), indicating that Lrch4 is required for full LPS-induced activation of IRF3. Similar results for nuclear NF-B activation, NF-B luciferase, and PO 4 -IRF3 were obtained using cells with CRISPR-Cas9 -mediated disruption of Lrch4 (Fig. 4, D-F).
In addition to NF-B and IRF3, LPS is well-known to lead to the rapid activation of MAPKs that, in turn, regulate multiple downstream functions including gene expression. Activation of these kinases is thought to be regulated by both MyD88 and A, silencing efficiency in RAW 264.7 macrophages of two stably transduced lentiviral Lrch4 shRNAs (A3 and A5) was evaluated at the transcript (graph; normalized to GAPDH as arbitrary units) and protein (immunoblot; samples in duplicate) level compared with a scrambled (Scr) control lentiviral shRNA. B, mRNA expression of Lrch4 variants 1 and 2 (normalized to 18S rRNA) was determined in Scr-and A5-transduced RAW 264.7 macrophages exposed to a time course of 10 ng/ml LPS. C, RAW 264.7 macrophages with Lrch4 (A3 and A5) or Scr lentiviral shRNA knockdown were exposed to buffer or two concentrations of LPS and then assayed by ELISA for TNF␣ protein release 2 h later. ND, not detected. D, two Lrch4 disruption clones were produced in RAW 264.7 cells by CRISPR-Cas9 and confirmed by Lrch4 immunoblot (right). The graph on the left shows TNF␣ protein release by the clones and parental controls in response to LPS (significant difference for both clones compared with parental at 2 and 24 h post-LPS). E, RAW 264.7 macrophages with Lrch4 (A5) or Scr knockdown were exposed to LPS (1 ng/ml) or Pam3CSK4 (10 ng/ml) and then assayed 2 h later by ELISA for TNF␣ or G-CSF release. F, HEK293 cells were sham-transfected (reagent only) or transfected with negative control siRNA, three alternate Lrch4 siRNAs, or FLAG-Lrch4 plasmid. Equal protein loads of whole cell lysate were then immunoblotted with anti-Lrch4 antibody as shown. G, HEK293-CD14-MD2 cells stably transfected with TLR4 or TLR2 were transfected with Scr/Lrch4 siRNA and then exposed to LPS (25 ng/ml) in the former case and Pam3CSK4 (100 ng/ml) in the latter case. IL-8 production was quantified by ELISA 8 h later. H, HEK293-CD14-MD2 cells were transfected with Scr or Lrch4 siRNA, exposed to buffer or TNF␣ (5 ng/ml), and then assessed by ELISA for IL-8 release 6 h later. All data are means Ϯ S.E. and derive from at least three independent experiments. †, p ϭ 0.057; *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001; ****, p Ͻ 0.0001. TRIF (26). As shown in Fig. 4G, Lrch4-silenced macrophages displayed a marked reduction in LPS-induced p38 activation (phosphorylation). By contrast, JNK phosphorylation was reduced at 15 min but not at 30 min post-LPS, consistent with a delay in its activation by LPS in Lrch4-silenced cells. Taken together, these results indicate that Lrch4 regulates early signaling events in the proximal TLR4 pathway, with effects upon both MyD88 and TRIF arms, but with varying temporal effects on the MAPKs.

Lrch4 regulates receptor-level events in the TLR4 pathway
Consistent with Lrch4 not exerting a global or indiscriminate effect upon TLR4 pathway output, we confirmed that neither silencing nor overexpression of Lrch4 altered cell surface display of TLR4 as assessed by flow cytometry (Fig. 5A and Fig. S3, A and B); nor did Lrch4 silencing or overexpression modify TLR4 gene expression (Fig. 5B). Similarly, Lrch4 knockdown had no effect on expression of the IL-6 receptor (Fig. S3C). Aiming to test whether Lrch4 could nonetheless impact ligand capture in the TLR4 pathway, we assessed cell-surface binding of LPS. Lrch4-silenced cells exhibited a significant, albeit modest reduction in overall surface binding of LPS (Fig. 5C). More remarkably, Lrch4-silenced cells showed reduced co-localization of LPS with the specific lipid raft marker cholera toxin subunit B (CtB; a ligand for the raft ganglioside GM1) (Fig. 5, D and E) (27,28), in conjunction with a much more punctate surface deposition of LPS as quantified by number of LPS foci per cell (Fig. 5F). Together, this suggests that Lrch4 is not just required for quantitative surface capture of LPS, but more specifically for successful delivery of LPS to lipid rafts, the site where CD14 and the TLR4 receptor cluster and are thought to interact (29).
Testing more directly for a role of Lrch4 in LPS binding, we next performed a streptavidin bead pulldown after exposure of macrophages to biotin-LPS. As shown in Fig. 5G, Lrch4 was detected in the biotin-LPS pulldown (but not in cells treated with nonbiotinylated LPS, as expected), suggesting that it interacts either directly or indirectly with LPS. Lrch4 capture by biotin-LPS was effectively competed by nonbiotinylated LPS, consistent with a bona fide LPS interaction, whereas CD14 pulldown was not notably competed. This finding suggests either differing stoichiometry or avidity to LPS of the two proteins.
We also tested for an effect of Lrch4 on CD14 given that CD14 plays a critical role in delivery of LPS from LPS-binding protein to TLR4/MD2 in lipid rafts (as well as in receptor-level interactions in the TLR2 and TLR7 cascades (5)). Lrch4-silenced macrophages had a significant reduction in cell surface or Scr control lentiviral shRNA were exposed to buffer or to LPS (10 ng/ml) for 2 or 4 h. Transcript abundance (normalized to GAPDH) was then quantified by qRT-PCR for TNF␣ (A), G-CSF (B), MyD88-dependent cytokines (C), and MyD88-independent cytokines (D) as shown. All data are means Ϯ S.E. and derive from at least three independent experiments. *, p Ͻ 0.05. AU, arbitrary unit.

Lrch4 regulates TLRs
display, protein expression, and transcript expression of CD14 (Fig. 5, H-J, and Fig. S3D). This may suggest that Lrch4 promotes LPS capture, raft delivery, and signaling at least in part through a mechanism involving up-regulation of CD14. However, our experiments in HEK293 cells involved plasmid-based stable overexpression of CD14, which we confirmed was not reduced by Lrch4 silencing (Fig. S3E). Lrch4 regulation of LPS signaling is thus unlikely to derive solely from CD14 regulation. Several TLRs are thought to be activated in lipid raft microdomains and to require raft abundance and integrity for intact signaling (30). Cells with chemically disrupted rafts have defective responses to LPS (29). Suggesting that Lrch4 may be required for lipid raft maintenance, we found that Lrch4-si-lenced macrophages had significantly reduced raft signal, as assessed by CtB staining (Fig. 5, D and K) (27,28).

Lrch4 regulates the innate immune response in vivo
The innate immune response plays a central role in a wide range of human diseases (4). Our studies in macrophage culture thus prompted us to examine a potential role for Lrch4 in the innate immune response in vivo. To test a potential role for Lrch4 in the response to LPS in the lung, we locally silenced Lrch4 in the airway of C57BL/6 mice by intratracheal delivery of lentiviral shRNA using published techniques (31,32). Lrch4 immunoblotting of whole lung homogenates confirmed a ϳ50% knockdown of Lrch4 with either of two lentiviral con- with Lrch4 lentiviral shRNA (A5) or Scr control lentiviral shRNA were exposed to buffer or to LPS (1 ng/ml) for 15 or 30 min. Activated p65 NF-B was then quantified in equal protein aliquots of nuclear isolates. B, cells as in A were transfected with NF-B-driven firefly luciferase and Renilla luciferase plasmids and exposed to buffer or to LPS (10 ng/ml) for 8 h. Relative luciferase units (RLUs) were then quantified by luminometry in cell lysates. C, cells as in A were exposed to LPS as shown. Nuclear isolates were then probed for PO 4 -IRF3 and HDAC-1 (loading control). D-F, two Lrch4 CRISPR-Cas9 clones and parental RAW 264.7 cells were stimulated and assayed as in A-C. In E, fold change (FC) in firefly/Renilla RLUs in the LPS-stimulated state compared with the nonstimulated state is shown. G, RAW 264.7 macrophages stably transduced with Lrch4 lentiviral shRNA (A5) or Scr control lentiviral shRNA were exposed to LPS (1 ng/ml) as shown. Equal protein aliquots of whole cell lysate were then probed by immunoblot for the indicated targets. The data in A, B, D, and E are means Ϯ S.E. The data derive from at least three independent experiments. *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001; ****, p Ͻ 0.0001

Lrch4 regulates TLRs
structs compared with Scr control (Fig. 6, A and B). The mice were then challenged with aerosolized LPS (33). As shown in Fig. 6 (C and D), compared with Scr control, Lrch4 silencing did not alter steady state cell counts in the airway of unexposed mice. However, Lrch4 silencing by either of two alternate lentiviral constructs was associated with a marked reduction in influx of leukocytes into the airway after LPS, largely reflecting a reduction in neutrophils. This finding suggests that Lrch4 plays a modulatory role in the innate immune response in vivo.

Discussion
TLRs play a key role in a wide range of inflammatory diseases in addition to detecting pathogens and inducing host defense responses. In recent years, a growing list of proteins have been identified that support signaling by TLRs, some of these acting as accessory proteins for several TLRs (5). TLR accessory proteins act through a wide range of mechanisms, including ligand delivery/binding, TLR folding/processing, and TLR trafficking. Here, we introduce Lrch4, a ϳ73-kDa transmembrane LRRcontaining protein that supports signaling by both plasmalemmal TLRs (TLR1/2, TLR2/6, and TLR4) and endosomal TLRs (TLR7 and TLR9).
In the TLR4 pathway, Lrch4 promotes early events, including both MyD88-dependent and MyD88-independent signaling. We provide evidence that this may stem from its role in promoting LPS binding, and, in particular, delivery of LPS to lipid rafts, the membrane microdomain where LPS and other microbial ligands are transferred from carrier proteins such as LPS- stably transduced with lentiviral shRNAs as in A were exposed to buffer or LPS as shown, after which normalized TLR4 mRNA was quantified by qRT-PCR. C, cells stably transduced with Scr or Lrch4 (A3 and A5) lentiviral shRNAs were exposed to biotin-LPS and then to AF633-streptavidin and evaluated by flow cytometry. D, Scr and Lrch4 shRNA macrophages were treated with CtB-AF488 and LPS-biotin/APC-streptavidin, co-stained with the nuclear stain DAPI, and then imaged (63ϫ objective; oil immersion). E, co-localization of LPS and CtB signals was quantified using Zeiss Zen software for cells processed in D. F, LPS deposits as in D are shown without pseudocoloring to aid in visualization and with a further 5ϫ magnification in ImageJ software (left). Discrete foci of LPS signal were quantified for cells in D using MetaMorph (right). G, RAW 264.7 macrophages were treated with biotin-LPS and/or LPS. Whole cell lysate (WCL) and streptavidinagarose pulldowns from cell lysates were then immunoblotted for Lrch4 and CD14. H-J, CD14 was quantified by flow cytometry (H), immunoblot of cell lysate (I), and qRT-PCR (J) in RAW 264.7 macrophages stably transduced with Scr or Lrch4 (A5) lentiviral shRNA. K, Scr and Lrch4 shRNA cells were processed as for microscopy in D. CtB staining was quantified using MetaMorph software. The data are means Ϯ S.E. and derive from at least three independent experiments. *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.0001. MFI, mean fluorescence intensity; AU, arbitrary unit.
LRRs, 20 -30-residue motifs typically found in tandem chains ranging from 2 to 45, occur in thousands of proteins throughout phylogeny and are thought to mediate proteinligand and protein-protein interactions (17,18). In addition to an established role in innate immunity, they are involved in a wide range of cellular processes, including apoptosis, neuronal development, autophagy, and nuclear mRNA transport. Of the ϳ375 LRR-containing proteins in humans, the majority have no known function (18). In Drosophila, the single LRCH gene has been recently shown to play a role in cytoskeletal remodeling during apoptosis (21), but limited homology (ϳ32%) exists between D. melanogaster LRCH and M. musculus Lrch4. Although the four mammalian Lrch genes/proteins have never previously been investigated to our knowledge, a few recent genome-wide screens have made incidental findings relating to Lrch4. Thus, Lrch4 has been found to be among LRR proteins that are down-regulated in human macrophages in response to bacterial infection (18), to be present in the phagosomes of mycobacterial-infected macrophages (34), to be among genes regulated by miR-155 in LPS-stimulated dendritic cells (35), and to increase with age in the human brain (36). Collectively, these reports have suggested a role for Lrch4 in host defense and perhaps in human disease.
We predict that, as has been the case for virtually all other TLR accessory proteins described to date, broader roles may ultimately be identified for Lrch4 in cell biology and immunity than defined in this report. As may be suggested by its apparent effect on rafts and CD14, Lrch4 may regulate cellular responses to a wider array of ligands, directly or indirectly, and perhaps function in more than one location/role within the cell. Our finding that Lrch4 silencing has differential effects across cytokine transcripts could suggest that it has regulatory effects on the TLR4 pathway from a location at or downstream of transcription factor activation. Alternatively, as we suggest, Lrch4 could exert selective effects from a receptor-proximal location, given that CD14 itself has been shown to have differential effects on MyD88-dependent and -independent signaling (37) and even to mediate TLR4-independent responses to LPS (38).
Our attempts to ectopically express and purify adequate quantities of soluble Lrch4 or portions of its ectodomain for in vitro studies of ligand binding were unsuccessful because of difficulties with protein insolubility/folding (not depicted). Our detection of a single band for endogenous Lrch4 in all subcellular fractions, soluble and membranous, suggests that the putative nontransmembrane variant of Lrch4, if expressed as protein, may either be expressed at very low levels or, alternatively, be secreted from the cell, analogous to soluble CD14, MD-2, and TLRs (5,39). Future studies will be required to determine whether Lrch4 binds LPS directly and whether soluble Lrch4 facilitates detection or acts as a decoy (with possible therapeutic implications) for microbial molecules. Finally, the predicted extracellular location of the CH motif in Lrch4 is of uncertain significance. Single CH motifs are thought to be insufficient to bind actin but are reported to mediate proteinprotein interactions with a variety of other motifs and proteins, including signaling proteins (20).
A potential cross-cutting mechanism by which Lrch4 may regulate multiple TLRs is via assembly/organization of lipid rafts. Multiple TLRs are thought to be activated in raft-like membrane microdomains and to be sensitive to raft composition (30). We found that Lrch4 regulates the raft molecules GM1 and CD14, the latter at the level of transcript abundance. Intracellular mediators of CD14 transcription have been defined, including cAMP (40) and transcription factors, such as AP1, Sp1, C/EBP, and STAT1 (41)(42)(43), several of these regulating CD14 in response to extracellular signals. Similarly, although rafts are thought to be assembled through complex protein-lipid interactions that begin in the Golgi (30), raft abundance on the cell surface as assessed by CtB is also sensitive to extracellular signals. For example, high-density lipoprotein and apolipoprotein E reduce CtB signal and raft-dependent signaling (44,45), whereas extracellular cholesterol increases both readouts (46,47). We speculate that Lrch4 regulation of rafts and CD14 may operate through sensitizing the macrophage to external signals (i.e. serum factors).
Additional possible mechanisms of TLR regulation are consistent with our findings. Analogous to other raft proteins such as the flotillins (48), multimerization of Lrch4 itself or Lrch4mediated protein-protein interactions within rafts may regulate raft topography. Alternatively, Lrch4 may serve as a raft chaperone for TLRs, regulating their successful trafficking to and ␤-actin (loading control) were quantified in equal protein aliquots of lung homogenate by immunoblot (A). Replicate lanes represent data from independent animals. Corresponding densitometry is shown in B. AUs, arbitrary units. C and D, mice treated as in A were left untreated or were exposed to aerosolized LPS (300 g/ml, 30 min). Total leukocytes (WBCs) (C) and neutrophils (PMNs) (D) were quantified in bronchoalveolar lavage (BAL) 24 h later. The data are means Ϯ S.E. and derive from at least three independent experiments involving n ϭ 5 mice/treatment/time point. *, p Ͻ 0.05 for comparison to Scr.

Lrch4 regulates TLRs
microdomains where they complex with signaling partners. Future studies, perhaps using super-resolution microscopic techniques, will be required to resolve these possibilities and to more comprehensively characterize the role of Lrch4 in determining raft size and composition.
TLR signaling has been shown to contribute to a wide range of noninfectious inflammatory diseases, likely through detection of host-derived molecules that are increased during disease (i.e. "damage-associated molecular patterns" such as hyaluronic acid, oxidized LDL, and fatty acids (49)). This has elevated TLRs as targets of intense interest for drug development and has also suggested that TLR accessory proteins may represent crosscutting therapeutic targets for a variety of diseases ranging from atherosclerosis to cancer to autoimmune disease (4). We speculate that Lrch4 may also regulate pro-inflammatory TLR signaling responses to host-derived molecules and may potentially represent a novel therapeutic target in human disease.

Lrch4 silencing in RAW 264.7 macrophages
A lentiviral set of five shRNA against murine Lrch4 was purchased from Open Biosystems/Thermo Fisher. Lentiviral packaging was achieved by using Lipofectamine 2000 (Invitrogen) to transiently transfect HEK293T/17 cells (ATCC no. SD-2515) with the desired shRNA in a pLKO.1 vector together with vesicular stomatitis virus G glycoprotein and packaging plasmids according to standard protocols (50). Supernatant was collected 48 h post-transfection and concentrated by centrifugation (50,000 ϫ g, 2 h). Pellets were resuspended in PBS and used for infection. All titers were determined by performing quantitative PCR to measure the number of lentiviral particles that integrated into the host genome. In addition to quantitative PCR, biological titration of viruses that co-expressed fluorescent moieties was determined by flow cytometry. RAW264.7 murine macrophage cells were infected with sh-lentivirus at a multiplicity of infection of 100 and at 48 h post-infection were selected with 10 g/ml puromycin (Calbiochem) for ϳ9 days. Lrch4 silencing was assessed by immunoblotting. Two of the five shRNA were determined to be most effective; the target sequences of those shRNAs and the sequence for the negative control (scrambled shRNA) are: (i) A3 (TRCN0000121334), CCGGGCTCTCAAGTCTCGGAAGAATCTCGAGATTCTT-CCGAGACTTGAGAGCTTTTTG; (ii) A5 (TRCN0000121336), CCGGCCTTCTGAATTAAGCCTTGTACTCGAGTACAAG-GCTTAATTCAGAAGGTTTTTG; and (iii) Scrambled shRNA, CCTAAGGTTAAGTCGCCCTCGCTCGAGCGAGGGCGAC-TTAACCTTAGG.

NF-B luciferase assay
FuGENE HD (Promega, Madison WI) transfection reagent was used (4.4 l/g DNA) to co-transfect RAW264.7 cells with reporter plasmid pNFB-Luc (Clontech) and with pRL-TK (Promega) as transfection normalization control. 48 h posttransfection, the cells were washed, fresh media were added, and the samples were either left untreated or treated with a ligand for 8 h. The cells were washed and lysed with passive lysis buffer, and luciferase activity was assessed using dual luciferase assay (Promega) measured on a Biotek Synergy 2 plate reader.

Western blotting
The cells were lysed, and equivalent loads of total protein were separated by SDS-PAGE, transferred to nitrocellulose membranes, and blocked in 5% milk/TTBS. Membranes were probed with antibodies against phospho-p38 (Thr-180/Tyr-

p65 NF-B activation assay
Nuclear extracts were isolated using the NE-PER kit (Pierce) per the manufacturer's instructions. Equal nuclear protein aliquots (Pierce BCA assay) were then analyzed with the TransAM NF-B p65 kit (Active Motif) per the manufacturer's conditions.

Cytokine protein measurement
Media supernatants were analyzed by either the Bioplex multiplex bead assay (Bio-Rad) or ELISA to mouse TNF␣ or human IL-8 (BioLegend) per the manufacturer's specifications.

Subcellular fractionation
Macrophages were resolved into membrane, cytosolic, and nuclear fractions using a subcellular protein fractionation kit (Thermo Scientific) per the manufacturer's instructions.

Murine in vivo exposures
C57BL/6J female mice, 8 -10 weeks old and weighing 18 -22 g, were used and were from the Jackson Laboratory. All experiments were performed in accordance with the Animal Welfare Act and the U.S. Public Health Service Policy on Humane Care and Use of Laboratory Animals after review and approval by the Animal Care and Use Committee of the NIEHS, National Institutes of Health. Lentiviral shRNA (scramble control or Lrch4 specific; 6 ϫ 10 7 transduction units in 50 l of saline) was delivered to the lung by oropharyngeal aspiration during isoflurane anesthesia, similar to past reports (31,32). Exposure to aerosolized LPS (300 g/ml, 30 min) was as previously described (33).

Bronchoalveolar lavage fluid collection and analysis
Bronchoalveolar lavage fluid was collected immediately following sacrifice and cell counts performed as described (33).

Structural prediction of Lrch4
The structure of the Lrch4 LRR domain was modeled based on bioinformatics analyses. The LRR region is predicted to extend to residue 249. However, predictions of disorder from residues 250 to 327 give reduced confidence, suggesting there may be some structure. After residue 327, the predictions of disorder become highly confident until the start of the CH domain near residue 535 (52). To ascertain whether residues 1-327 were structured, this sequence fragment was submitted to pGenTheader for comparison to known protein structures. The best matches only contain structured residues up the end of the LRR domain, except in a few cases where the sequence matches to a fragment of much longer LRR (Ͼ600 residues). We view these exceptions as unlikely because these are very large proteins with many LRRs that are clearly not present in Lrch4. Residues 250 -327 were searched alone for structural matches, but none were found with statistical confidence. The region is likely unstructured. The structure of the LRR in Fig. 1 was modeled from Protein Data Bank code 4PSJ with Bioserf.

LPS binding
LPS binding was measured by incubating the RAW 264.7 macrophages (1 ϫ 10 6 ) with 5 g/ml of biotin-labeled LPS (Invivogen) at 37°C for 15 min. Surface binding was assessed with streptavidin-APC (BD Biosciences). LPS signal was quantified by flow cytometry using an LSR II (BD Biosciences) and analyzed using FlowJo software (Tree Star, Inc.).

Lrch4 regulates TLRs
fected with tGFP or hLrch4-tGFP. 48 h post-transfection, the cells were left unstained or stained with 1ϫ CellMask TM Orange plasma membrane stain (Molecular Probes/Life Technologies) for 5 min at 37°C, followed by fixation (4% formaldehyde). After three washes, the coverslips were mounted in ProLong Gold antifade reagent containing DAPI and imaged as described below. Scrambled or Lrch4 shRNA RAW264.7 cells were plated overnight on MatTek dishes (Ashland, MA) or 24-well plates. The cells were left untreated or treated with 100 ng/ml biotin-LPS for 10 min at 37°C. The cells were then either left unstained or stained (10 min, 4°C) with 1 g/ml of cholera toxin subunit B-Alexa Fluor 488 conjugate (Molecular Probes/Life Technologies). The cells were then washed thrice, fixed, and blocked (2% BSA in PBS-T, supplemented with 20% fetal bovine serum) overnight. Some wells/ dishes were additionally stained with Alexa Fluor 633-conjugated streptavidin. After three washes, the wells/dishes were mounted in ProLong Gold antifade reagent containing DAPI and imaged as described below. Imaging was performed using a Zeiss LSM 710 (Carl Zeiss). For CtB and LPS, objective conditions were Plan-Apochromat 63ϫ; NA ϭ 1.40; oil immersion. For Lrch4 and Cell-Mask TM , objective conditions were EC Plan-Neofluar 40ϫ; NA ϭ 1.3; oil immersion. Cholera toxin (CtB-AF488) intensity and LPS foci were analyzed using MetaMorph software, and Lrch4-LPS co-localization was analyzed using Zeiss Zen software.

Co-precipitation studies
Co-precipitation studies were undertaken to assess the ability of Lrch4 to associate with LPS in cells treated in vivo. RAW 264.7 macrophages were incubated (15 min, 37°C) with either 10 g of biotin-labeled LPS, 10 g of biotin-labeled LPS plus 20 g of LPS, or 10 g of LPS alone. Following incubation, biotin-LPS was pulled down by streptavidin-Sepharose. Captured complexes were washed, and proteins were eluted from the Sepharose and probed for Lrch4 and CD14, in parallel with probing of whole cell lysate.

Statistical analysis
Analysis was performed using GraphPad Prism software (San Diego, CA). The data are represented as means Ϯ S.E. Twotailed Student's t test was applied for comparisons of two groups, and analysis of variance for comparisons of Ͼ2 groups. For all tests, p Ͻ 0.05 was considered significant.