A Point Mutation in the Amino Terminus of TLR7 Abolishes Signaling without Affecting Ligand Binding

This information is current as Carlo Iavarone, Katrin Ramsauer, Andriy V. Kubarenko, of September 24, 2021. Jason C. Debasitis, Igor Leykin, Alexander N. R. Weber, Owen M. Siggs, Bruce Beutler, Pu Zhang, Gillis Otten, Ugo D'Oro, Nicholas M. Valiante, M. Lamine Mbow and Alberto Visintin

J Immunol 2011; 186:4213-4222; Prepublished online 7 Downloaded from March 2011; doi: 10.4049/jimmunol.1003585 http://www.jimmunol.org/content/186/7/4213 http://www.jimmunol.org/ References This article cites 59 articles, 24 of which you can access for free at: http://www.jimmunol.org/content/186/7/4213.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision by guest on September 24, 2021 • No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2011 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

A Point Mutation in the Amino Terminus of TLR7 Abolishes Signaling without Affecting Ligand Binding

Carlo Iavarone,* Katrin Ramsauer,* Andriy V. Kubarenko,† Jason C. Debasitis,* Igor Leykin,* Alexander N. R. Weber,† Owen M. Siggs,‡ Bruce Beutler,‡ Pu Zhang,* Gillis Otten,* Ugo D’Oro,x Nicholas M. Valiante,* M. Lamine Mbow,* and Alberto Visintin*

TLR7 is the mammalian for ssRNA and some nucleotide-like small molecules. We have generated a mouse by N-nitrose-N9- ethyl urea mutagenesis in which threonine 68 of TLR7 was substituted with isoleucine. Cells bearing this mutant TLR7 lost the sensitivity to the small-molecule TLR7 agonist , hence the name TLR7rsq1. In this work, we report the characterization rsq1

of this mutant . Similar to the wild-type counterpart, TLR7 localizes to the endoplasmic reticulum and is expressed at Downloaded from normal levels in both primary cells and reconstituted 293T cells. In addition to small-molecule TLR7 agonists, TLR7rsq1 fails to be activated by ssRNA. Whole-transcriptome analysis demonstrates that TLR7 is the exclusive and indispensable receptor for both classes of ligands, consistent with the fact that both ligands induce highly similar transcriptional signatures in TLR7wt/wt spleno- cytes. Thus, TLR7rsq1 is a bona fide phenocopy of the TLR7 null mouse. Because TLR7rsq1 binds to ssRNA, our studies imply that the N-terminal portion of TLR7 triggers a yet to be identified event on TLR7. TLR7rsq1 mice might represent a valuable tool to

help elucidate novel aspects of TLR7 biology. The Journal of Immunology, 2011, 186: 4213–4222. http://www.jimmunol.org/

ecognition of invading microorganisms at the cellular improving ligand recognition [CD36 (6, 7), CD14 (8), MD-2 (9, level relies on a number of germ line-encoded receptors 10)] or maintaining their correct subcellular localization [gp96 R that initiate the immune response, trigger the clearance of (11), PRATA/B (12, 13), UNC93B (14, 15)]. the invading noxious agent, and participate in tissue repair pro- It is generally accepted that dimerization of the extracellular cesses. In mammals, TLRs form one of the most prominent fami- domain of most TLRs is the triggering mechanism. For example, lies of such receptors and drive early innate and later Ag-specific TLR3 forms stable dimers upon interaction with dsRNA of ∼40 bp responses to pathogens (see Ref. 1 for an extensive review of the (16–18), thus bringing the two intracellular signaling domains in

field). TLRs are type I receptors that evolved to recognize con- close proximity. In sharp contrast, TLR9 and TLR8 seem to exist by guest on September 24, 2021 served features mainly found in the microbial biosphere. A subset as preformed signaling-incompetent dimers in ectopically recon- of TLRs (TLR3, 7, 8, and 9) recognize nucleic acids and are stituted systems (19–21). The current model for TLR9 activation endoplasmic reticulum (ER)-resident (2–4). Subcellular compart- is that nucleic acids bind to two distinct binding sites (21, 22) and mentalization and cellular distribution of TLRs seem to be the induce a conformational change in the ectodomains. In immune main determinants for specificity of ligand recognition (exogenous cells, this results in complex proteolytic and trafficking events that versus endogenous) and quality of the cellular response (proin- eventually propagate the triggering signal to its intracellular do- flammatory, phagocytic, or antiviral) (1, 5). The activity of TLRs main (15, 19, 23). The precise mechanism of action of TLR7 is is regulated by a variety of proteinaceous cofactors involved in unknown. Unique among all TLRs, TLR7 and TLR8 can also be activated *Immunology US, Novartis Vaccines and Diagnostics, Cambridge, MA 02139; by synthetic low m.w. agonists alien to the microbial world. A †Toll-like Receptors and Cancer, German Cancer Research Center, 69120 Heidelberg, prime example is , an imidazoquinoline marketed as Germany; ‡Department of Genetics, The Scripps Research Institute, La Jolla, CA x Aldara that is used in the treatment of certain skin diseases, in- 92037; and Immunology IT, Novartis Vaccines and Diagnostics, 53100 Siena, Italy cluding , superficial basal cell carcinoma, and Received for publication October 28, 2010. Accepted for publication January 30, 2011. external genital warts caused by human papilloma (24). This work was supported in part by Transformational Medical Technologies program Imiquimod is a “pure” TLR7 agonist in humans, that is, it only contract HDTRA1-07-9-0001 from the Department of Defense Chemical and Bi- activates TLR7, however it triggers both TLR7 and TLR8 in some ological Defense program through the Defense Threat Reduction Agency. ungulates (20, 25). Resiquimod (R848), a compound closely re- The complete set of microarray data has been submitted to the Array Express data- lated to imiquimod, is a dual (TLR7/8) agonist in humans, but it is base European Molecular Biology Laboratory-European Bioinformatics Institute un- der accession number E-MEXP-3051. a pure TLR7 agonist in mice. When R848 is used in combination Address correspondence and reprint requests to Dr. Alberto Visintin at the current with poly dT17 (a short DNA homopolymer), it can activate address: Pfizer, 620 Memorial Drive, Cambridge, MA 02139. E-mail address: dr. mouse TLR8 (26). There have been reports that TLR8 can sense [email protected] dsDNA in this species, but this is still controversial (4, 27). Taken Abbreviations used in this article: BMDM, bone marrow-derived ; DCs, together, these results suggest that low m.w. TLR agonists inter- dendritic cells; Endo-H, endoglycosidase H; ENU, N-nitrose-N9-ethyl urea; ER, endoplasmic reticulum; GO, Ontology; HA, hemagglutinin tag; R0006, U- act with different TLRs in a species-specific manner, and, under rich single-stranded RNA; R848, resiquimod; SLE, systemic erythematosus; certain conditions, they might require nucleic acids to form T68I, conversion of threonine 68 to isoleucine; TMEV, Tigr MultiExperiment functional mimics of the natural ligands. However, there is a major Viewer; WCL, whole-cell lysate. gap in our understanding of the exact mechanism of action of Copyright Ó 2011 by The American Association of Immunologists, Inc. 0022-1767/11/$16.00 TLR7 ligands. This is mainly due to the scarcity of investigative www.jimmunol.org/cgi/doi/10.4049/jimmunol.1003585 4214 TLR7rsq1 CHARACTERIZATION tools and the difficult biochemistry around the study of endosomal (Sigma) for 3 min at room temperature. Cells were cultured in RPMI 1640, TLRs. 2.5% FCS, and penicillin/streptomycin mix. Bone marrow-derived mac- By screening N-nitrose-N9-ethyl urea (ENU)-induced mutations rophages (BMDMs) were derived as previously described (31). Treatments were performed overnight in 96-well flat-bottom tissue culture plates at in mice, we identified a variety of genetic defects in the mam- a cell density of 5 3 105 cells/well in 200 ml. malian innate immune response, including involved in TLR function (28). One such gene is Unc93b1, which encodes the 12- Plasmids and cell lines spanning ER-resident protein UNC93B. Peritoneal The cDNA for RFP-tagged calreticulin was from Origene. The expression from Unc93b1 mutant mice were unable to respond to synthetic vectors pUNOTLR7, pUNOTLR7HA, and IRF5HA were purchased from DNA oligonucleotides, pIC, and R848 ex vivo, that is, they had Invivogen. IRF5HA was subcloned into the pCDNA3.1 hygro vector (Invitrogen) and used to produce the stable cell line 293TLR7/IRF5HA. The a triple defect in TLR signaling (14). The molecular mechanism TLR7rsq1 (N68I) mutant was generated by site-directed mutagenesis using for the loss of function was ascribed to a defect in the trafficking the QuickChange kit according to the manufacturer’s recommended primers of these endosomal TLRs, which is governed by UNC93B (15). and procedures (http://www.stratagene.com/sdmdesigner/default.aspx). YFP Because the analysis of ENU mutagenized mice had proved very pUNOTLR7 was cloned by inserting the YFP expression cassette after successful to elucidate many other aspects of TLR signaling (28), a 9-aa spacer (AAGGAAGGA). The resulting protein was functional in 293T cells, although it was not as potent as the untagged version. All the we asked whether the ENU screen produced mice with impaired constructs have been sequence verified, and additional information on the responses to TLR7 agonists. When the current project started in different plasmids will be provided upon request. Stable 293T cell lines 2008, we found one mutant strain of C57BL/6 mice that lost the were generated by transient transfection with the required cDNA and se- ability of responding to R848 (http://mutagenetix.scripps.edu/ lection in blasticidin (5 mg/ml) for pUNO-TLR–based plasmids, hygro- mycin (250 mg/ml) for IRF5HA, and puromycin (5 mg/ml) for NF-kB– Downloaded from home.cfm). We called this mutant strain rsq1 andmappedthe luciferase. All the cells were maintained in DMEM supplemented with 5% mutation to a C-to-T transition at position 391 of the TLR7 tran- FBS (V Scientific), 2 mM L-glutamine, and 100 U/ml penicillin–strepto- script, resulting in the conversion of threonine 68 to isoleucine mycin (Cellgro). Human PBMCs were purified by Ficoll-Paque Plus sep- (T68I) in the TLR7 protein. The TLR7rsq1 mutation resides in the aration (Amersham) from 50 ml of heparinized blood from healthy do- N-terminal portion of TLR7, in close proximity to the first two nors. insertions thought to be relevant for ligand binding (29). Since Western blotting, RNA–TLR7 binding assays, then, we identified a second inactivating mutation that also affects coimmunoprecipitations, and deglycosylation assay http://www.jimmunol.org/ the proximal ectodomain, rsq2. Rsq2 is an X-linked recessive Cells were lysed in lysis buffer (20 mM Tris pH 7.4, 137 mM NaCl, 10% inactivating mutation that produces the conversion of alanine 128 glycerol, 1 mM EDTA, 1 mM PMSF, and freshly added protease inhibitors) rsq1 to a tyrosine in TLR7 protein. In this study, we used the TLR7 for 10 min on ice. Lysates were cleared by centrifugation at 13,000 rpm for 5 mutant as a tool to gain insight into the molecular mechanism of min and postnuclear whole-cell lysates (WCLs) used immediately or stored frozen. For the RNA capture experiments, biotinylated ssRNA (R0006; 10 TLR7 triggering. We first performed an extensive biochemical 6 rsq1 rsq1 mg/6 3 10 cells) was mixed with WCLs containing TLR7HA or TLR7 characterization of the TLR7 mutant protein in transfected HA and 20 ml of packed avidin–agarose beads/sample (Pierce). Reactions 293T cells and identified a crucial role for N66, a putative glycan were incubated in rotation at 4˚C for 1 h, and RNA–protein complexes acceptor site. Abrogation of the N-glycosylation consensus im- were collected by centrifugation, washed three times in lysis buffer, paired the ability of TLR7 to signal in response to ssRNA and low resuspended in 30 ml SDS-sample buffer, boiled, and subjected to SDS- by guest on September 24, 2021 PAGE. Presence of TLR7 in the precipitates was assessed by anti-HA m.w. TLR7 agonists but not its ability to bind to ssRNA. Whole- rsq1 rsq1 Western blotting. Binding to TLR7HA and TLR7 HA was also de- mRNA expression profiling of 293TLR7 cells and termined using an ELISA-like assay in which plastic-adsorbed RNA was rsq1 splenocytes from the TLR7 mice revealed no off-target effects used to capture TLR7 from WCLs containing the HA-tagged protein and 1 of the two benchmark TLR7 agonists R848 and ssRNA. Taken mg/ml of an HRP-conjugated rat mAb anti-HA Ab (Roche). Biotinylated together, our results suggest that TLR7 is an exclusive receptor for RNA was plated in titrated amounts on avidin-coated and preblocked white plates for 10 min (Thermo Scientific). After washing of the excess un- both R848 and ssRNA in mice and that the N-terminal portion of bound oligos in PBS-0.05% Tween 20, WCLs from TLR7 and TLR7rsq1 TLR7 is necessary for function but not ligand binding. cells (diluted 1:3 in citrate buffer pH 7; Sigma) were applied to the streptavidin-captured oligos for 60 min. After three additional washes, a luminogenic HRP substrate (Pierce) was added to the wells and the Materials and Methods emitted light measured using standard luminometry. Then, 100 U/ml Reagents and Abs RNAse I (NEB) was added to the lysates to eliminate endogenous cellu- lar RNA, which can bind to TLR7 after lysis. Endoglycosidase H (Endo-H) All common reagents were from Sigma, unless stated differently. The resistance/sensitivity was used to determine whether underwent single-stranded oligoribonucleotide (ssRNA) R0006 [UUGUUGUUGUU- trans-Golgi maturation as described in Ref. 32. Briefly, 20 ml WCLs from GUUGUUGUU (30)] and its complementary strand 6000R (AACAA- the indicated cell lines were boiled in 5 mM DTT and treated with 1 mlof CAACAACAACAACAA) were purchased from IDT (Coralville, IA) and either Endo-H or Endo-F for 1 h at 37˚C (NEB). Reactions were stopped were fully phosphothioate. In some experiments, a 39 biotinylated version by adding SDS-sample buffer, and the shift in mobility was assessed by was used (IDT). The 39-biotin labeled immunostimulatory RNA oligonu- Western blotting as described above. cleotides retain their TLR7 agonist activity. The TLR ligands LPS, R848, and Pam3Cys4 were from Invivogen. Human TNF-a was from R&D NF-kB–luciferase reporter gene assay Systems. The following Abs were used in this work: anti–phospho-JNK, phospho-p38, and phospho-IkB (Cell Signaling), anti-IkB, anti-calnexin, Adherent 293T cells stably expressing the indicated TLR and the NF-kB– and anti-Lamp1 (Santa Cruz), anti-hemagglutinin tag (HA) (Covance), luciferase reporter cassette were induced in 96-well format (50,000 cells/ HRP-labeled rat mAb anti-HA (Roche), anti-FLAG, HRP-labeled anti- well) with the indicated amount of stimulant. After 16 h, spent medium FLAG, anti-tubulin, HRP-labeled goat anti-mouse IgG, HRP-labeled was removed by aspiration, and the cells were lysed in 50 ml of a 1:1 goat anti-rabbit and anti-HA pAb (Sigma). The anti-mouse TLR7 poly- dilution of Cell-Glo (Promega) in water. Emitted light was quantitated clonal Ab was a gift from Dr. Ploegh’s laboratory (Whitehead Institute for using a luminometer and the values plotted as the average of duplicates 6 Biomedical Research, Cambridge, MA). were quantified by SD (luciferase units). In some experiments, the readings were divided by ELISA (duo-set kits; R&D Systems) or multiplexed analysis the unstimulated control and plotted as fold induction (relative luciferase (MesoScale Discovery) per the manufacturers’ protocols. units). Animals, tissue extraction procedures, and cell differentiation RNA isolation, labeling, and microarray analysis Splenocytes were isolated from C57BL/6 mice or TLR7rsq1 mutant mice Total RNAwas extracted using QIAShredder/RNeasy columns according to (same genetic background) by mechanical dissociation (Miltenyi Biotech, the manufacturer’s instructions (Qiagen). RNA labeling, hybridization, and Bergisch Gladbach, Germany). RBCs were lysed in 3 ml ACK lysis buffer scanning were performed using methods, reagents, software, and hardware The Journal of Immunology 4215 purchased from Agilent Technologies. Briefly, 400 ng RNA was retro- (37), VERIFY_3D, and ERRAT (http://nihserver.mbi.ucla.edu/) online transcribed and labeled using Cy3 (for the samples treated with DOTAP, servers. For energy minimization of the obtained final model by molecular DMSO, PBS, or left untreated) or Cy5 (LPS, R848, R0006) and column dynamics, the GROMACS package was used (38). Structure analysis was purified (Qiagen). The efficiency of Cy5 and Cy3 dyes incorporation was carried out in SwissPBD Viewer (39) and PyMol (www.pymol.org) visu- assessed by fluorometry. Reactions with yields .0.825 mg cRNA and alization software using PDB2PQR (40), PropKa (41), and APBS (42) yields .8 pmol dye/mg cRNA were used for hybridization. Equal amounts packages for charge surface calculations and HotPatch Web server (43) for of labeled Cy5 and Cy3 cRNAs were hybridized onto 4 3 44 Whole hydrophobicity calculations. Human or Whole Mouse Genome Microarray. After scanning, images were analyzed using the Feature Extraction 10.5.1.1 software. Data were Results transferred to the BASE 2.14.1 software for microarray data management rsq1 and analysis (33). For each spot, local background was subtracted, and spot TLR7 is not functional intensities were normalized by the mean fluorescence intensity for each TLR7rsq1 mice were originally identified because of the failure of channel. The average intensity ratio of each spot from experimental rep- licates was estimated by geometric mean, and the accuracy and statistical their peritoneal macrophages to respond to R848. Analogously, rsq1 significance of the observed ratios were determined by using Student t test. BMDMs from TLR7 mice failed to trigger IkB, p38, and JNK , Genes with a p value 0.05 and average intensity ratios $2 (log2 ratio $1) in response to R848, confirming that signaling downstream of were considered differentially expressed. Hierarchical clustering was TLR7 is impaired (Fig. 1A). The signaling defect of TLR7rsq1 performed with the Tigr MultiExperiment Viewer (TMEV) 4.4.1 software appeared to precede IRAK1 activation, as R848 treatment of (34) on the transformed log2 ratio. The complete set of microarray data has rsq1 been submitted to the Array Express database EMBL-EBI (http://www.ebi. BMDMs derived from the TLR7 mouse failed to induce the ac.uk/arrayexpress/) under accession number E-MEXP-3051. degradative pathway that is triggered by IRAK1 auto/trans- phosphorylative events (Fig. 1B) (44). The defect in intracel- Subcellular fractionation Downloaded from lular TLR7 signaling was not due to the lack of TLR7 expression, Subcellular fractionation was carried out using reagents and protocols and both pattern and level of expression of TLR7rsq1 and TLR7 provided in an ER isolation kit (Sigma). Briefly, 293T cells stably ex- rsq1 were comparable by confocal analysis (Fig. 1C). As a conse- pressing TLR7 or TLR7 were plated on three 150-mm tissue culture quence of the lack of intracellular signaling, TLR7rsq1 trigger- dishes at a density of 9 3 106 cells/dish. After 16 h, the cells were har- vested, resuspended in hypotonic buffer at a volume three times that of the ing by both R848 and ssRNA (another TLR7-dependent stimulus) packed pellet, and mechanically disrupted using a Wheaton Dounce tissue grinder with a type A pestle. Post-nuclear supernatants were diluted in an http://www.jimmunol.org/ equal volume of isotonic isolation buffer. Cell extracts were then fraction- ated on a discontinuous iodixanol gradient by centrifugation at 100,000 3 g for 16 h. Fractions were collected, solubilized in Triton X-100, and sub- jected to SDS-PAGE/Western analysis for the subcellular markers in- dicated in the legends of figures accompanying this article. Confocal imaging analysis Cells were adhered on glass-bottom tissue culture dishes (MakTek) and imaged using a Zeiss LSM 700 confocal microscope at 340 magnifications.

Cells were permeabilized using Cytofix/Cytoperm (BD) and washed twice. by guest on September 24, 2021 Unspecific binding to Fc receptors was blocked by incubation with the Fc block Ab CD16/CD32; FcgIII/IIR (BD). The polyclonal anti-TLR7 Ab was added for 30 min, washed, and detected using a Cy5-labeled goat anti- rabbit Ab (Invitrogen). FACS analysis Flow cytometry was performed using FACSCalibur and LSR-II flow cytometers (Becton Dickinson Biosciences). Lymphocytes were selected by forward scatter and side scatter. Fluorescence signals were collected in log mode. A minimum of 30,000 cells of interest were acquired for each sample, and data were analyzed using FlowJo analysis software. Immunostaining was performed using conjugated mAbs to CD3, CD4, CD8a (T cells), CD19, B220 (B cells), CD11c, CD14 (monocytes/dendritic cells [DCs]), CD49b (NK cells), CD69, CD86, and MHC class II in various labels and combinations. All the Abs were from Becton Dickinson Biosciences. Sequences alignment Sequence alignment was generated in ClustalX2 (35). TLR7 sequence abbreviations are as follows: HUMAN (Homo sapiens), PANPA (Pan paniscus), PANTR (Pan troglodytes), 9PRIM (Gorilla gorilla), PONRY (Pongo pygmaeus), SAGOE (Saguinus oedipus), HYLLA (Hylobates lar), MACMU (Macaca mulatta), MACFA (Macaca fascicularis), CERTO FIGURE 1. Lack of TLR7 intracellular signaling cascade in TLR7rsq1 (Cercocebus torquatus), BUBBU (Bubalus bubalis), BOSIN (Bos indicus), BMDMs. A, BMDMs were treated as indicated and the WCLs were sub- BOVIN (Bos taurus), SHEEP (Ovis aries), PIG (Sus scrofa), CANFA jected to Western blot analysis for IkB (to monitor its ligand-induced (Canis familiaris), FALCA (Felis catus), HORSE (Equus caballus), ROULE (Rousettus leschenaulti), LAXAF (Loxodonta africana), MOUSE degradation), phospho-p38, phospho-JNK, and tubulin (to confirm equal (Mus musculus), RAT (Rattus norvegicus), ANAPL (Anas platyrhynchos), protein input). The results are representative of two that gave similar GALSO (Gallus sonneratii), GALVA (Gallus varius), GALLA (Gallus results, and all the panels derive from the same input sample, run on lafayettei), XENTR (Xenopus tropicalis), ONCMY (Oncorhynchus different gels or day. B, BMDMs were treated with LPS (1 mg/ml) or R848 mykiss), and TAKRU (Takifugu rubripes). (10 mM) for the indicated times, and the amounts of IRAK1 were assessed by Western blot. C, BMDMs from wild-type or TLR7rsq1 mice were modeling and structure analysis stained using an anti-TLR7 polyclonal Ab followed by detection with Cy5 Modeling of human and mouse TLR7 ectodomains was carried out as goat anti-rabbit IgG (amTLR7) or stained without the primary Ab (no Ab) described (21) using the MODELER package (36). The complete models as a control. Two representative confocal micrographs are shown at 340 were then scored for energy and steric correctness using the ANOLEA magnification. 4216 TLR7rsq1 CHARACTERIZATION

FIGURE 2. Lack of cytokine secretion and cellular activation in BMDMs and splenic B cells of TLR7rsq1 mice. A, BMDMs from TLR7wt (solid line) and TLR7rsq1 mice (dashed lines) were stimulated with the indicated con- centrations of LPS, ssRNA, or R848 for 16 h. Supernatants were harvested and analyzed by mesoscale analysis for seven proinflammatory cytokines (IL-1b, IFN-g, IL-12p70, IL-10, KC, IL-6, and TNF-a) all of which gave compara- ble activation profiles. Shown are representa- tive results for IL-6 (upper panel) and TNF-a (lower panel) and from one of at least three independent experiments. B, Splenocytes from TLR7 wild-type (solid bars) or TLR7rsq1 (open bars) were stimulated with R848 (10 mM), R0006 (5 mg/ml), the immunostimulatory CpG oligonucleotide ODN1826 (1 mM), and car- riers alone (DMSO and DOTAP) for 22 h.

Cells were then harvested and labeled with Downloaded from lineage-specific markers and analyzed by flow cytometry. Plotted are the mean CD86 values (left panel) and percentage CD69 positivity (right panel) of the CD32, CD42, CD82, CD11c2, CD19+, MHC+ population. Results represent the average of two mice plus range. http://www.jimmunol.org/ did not induce cytokines in BMDMs (Fig. 2A), splenocytes, and Next, we tested whether other signaling pathways were impaired. DCs (data not shown). Moreover, splenic B cells from the As shown in Fig. 3B, 293TLR7 cells induced p38 and JNK in TLR7rsq1 animals failed to upregulate costimulatory and activation response to R848, whereas 293TLR7rsq1 cells did not, which cor- markers (CD86 and CD69) in response to TLR7 stimuli (Fig. 2B). Conversely, TLR7rsq1 cells always responded normally when trea- ted with other TLR ligands, such as LPS or ssDNA. TLR7rsq1 mice are born in a normal Mendelian ratio and appear normal and healthy. by guest on September 24, 2021 Immunologically, they do not show differences in the frequencies of T cells (CD3+ and either CD4+ or CD8+), B cells (CD122+), mono- cytes (CD14+), and DCs (CD11c+) in either the bone marrow or the (Table I). To elucidate the molecular defect of the TLR7rsq1 mutation and to gain insight into the mechanism of action of TLR7, we con- structed the human version of mTLR7rsq1 and asked whether the human ortholog of TLR7rsq1 is functional in 293T cells. Human 293TLR7rsq1 cells did not induce the NF-kB–luciferase reporter gene in response to both R848 (45) and ssRNA (3, 30) (Fig. 3A).

Table I. Comparison of cell populations in the and bone marrow of TLR7wt and TLR7rsq1 mice

Cell Population and Location % TLR7wt (SD) % TLR7rsq1 (SD) Bone marrow CD4+ T cells 54.5 (0.70) 53.03 (1.22) CD8+ T cells 40.03 (0.71) 41.83 (1.62) B cell 50.43 (3.79) 44.6 (4.88) FIGURE 3. Lack of signaling in 293T cells reconstituted with human Monocytes 2.23 (0.31) 2.55 (0.06) rsq1 NK cells 8.26 (1.84) 6.39 (1.67) TLR7 . A, 293T stably expressing wild-type human TLR7 (diamonds) or rsq1 Spleen human TLR7 (squares) were stimulated overnight with R848 at the in- CD4+ T cells 32 (3.62) 33.4 (5.44) dicated concentrations (left panel)orssRNA(right panel)mixed1:1(w/w) CD8+ T cells 49.5 (4.16) 51.23 (6.30) in DOTAP. The activity of the NF-kB–luciferase reporter gene was recorded B cell 52.27 (1.86) 43.5 (3.68) by luminometry. Results are representative of three independent experiments, Monocytes 3.96 (0.16) 3.62 (0.73) and readings are the average 6 SD of triplicate readings. RLU, relative NK cells 3.43 (0.10) 3.49 (1.05) luciferase units. B, 293T cells stably expressing TLR7rsq1 or TLR7 were left T cells were gated based on their scatter characteristics and CD3+ expression. untreated or were treated with R848 (10 mM) or TNF-a (10 ng/ml). After 30 + + + Shown is the percentage of CD4 and CD8 cells in the total CD3 cells. B cells were or 60 min, cells were lysed, and phospho-IkB, IkB, phospho-p38, phospho- gated by CD19 and B220 expression. Monocytes were gated by CD14 expression, JNK, and tubulin were determined by Western blot on whole-cell extracts. and NK cells were gated by CD49b expression. Data in the table represent the average of percentage of cells harvested from spleens and bone marrow for three Results are representative of two similar gels, and all the panels derive from individual mice. the same input sample, run on different gels or day. The Journal of Immunology 4217 roborated the observations from primary cells. A general signaling Methods for details). The modeling reveals that glycans could be defect in 293TLR7rsq1 cells could be excluded because they added to both N66 and N69 in human and murine TLR7, re- responded to TNF-a. spectively (Fig. 4). We then asked whether mutating the acceptor asparagine 66 would lead to the same effect as T68I. The resulting An intact N-glycosylation target sequence in position 66–68 is mutant cannot be glycosylated at position 66. As shown in Fig. required for proper TLR7 function 5A, TLR7N66A, like TLR7rsq, was incapable of activating NF-kB We noted that T68 belongs to an N-glycosylation consensus site in 293T transfectants, supporting our model that an N66-linked rsq1 [NXT/S (46, 47)], which is predicted to be disrupted in TLR7 . glycan could be the reason for the loss of function phenotype A second N-glycosylation is predicted to involve N69. To gauge observed in these mutants. The inability of TLR7rsq1 to signal was the potential importance of glycosylation at N66, we analyzed the not specific to R848 or ssRNA, and other adenine-based low m.w. sequence alignments of all currently available TLR7 sequences. TLR7 agonists failed to activate TLR7 as well (data not shown). As shown in Fig. 4A, the region harboring the two predicted Taken together, the in silico analysis and the mutagenesis ex- glycosylation sites (N66 and N69) in LRR1 is extremely well periments suggest that the defect in TLR7rsq1 might be due to conserved from mammals to fish. Only Takifugu rubripes (puffer the suppression of the carbohydrate-substituted N-glycosylation fish) is missing the asparagines at position 66. Based on the three- site in position 66. dimensional model of TLR7 (21), T68 is found at the end of the rsq1 TLR7 ectodomain in the first LRR b-strand, whereas N66 is at the TLR7 localizes to the ER like its wild-type counterpart end of the turn leading into it (48). We sought to assess if gly- Next we asked whether the wild-type or the mutated TLR7 differed cosylation at N66 or N69 was possible in the stereochemical in their subcellular localization in 293T cells. To visualize the Downloaded from context of the asparagines’ side-chain conformations found in the subcellular localization of TLR7 and TLR7rsq1 in live cells, we three-dimensional model by using GlyProt (see Materials and generated 293T cell lines stably expressing YFP-tagged versions http://www.jimmunol.org/ by guest on September 24, 2021

FIGURE 4. T68 of TLR7 in the content of pri- mary and tertiary structure of TLR7. A, Alignment of TLR7 LRRNT and LRR1 region sequences from all available species (list of species is given in Materials and Methods). T68 is 100% conserved. B, Mapping of conservation of surface residues on the model of human TLR7 ectodomain. Green regions correspond with highly conserved residues and red regions with nonconserved. C and D, Models of the extracellular domains of human and mouse TLR7 (in surface representation) with all possible glycans (orange sticks and semitransparent surfaces repre- sentation). E and F, Close-ups of the LRR1 of hu- man and mouse TLR7, respectively. T68 is high- lighted by red sticks, whereas N66 and N69 with linked glycans are in orange. 4218 TLR7rsq1 CHARACTERIZATION

secretory pathway. To map in more detail the subcellular locali- zation of TLR7 and TLR7rsq1, we fractionated the cellular lysates in discontinuous iodixanol gradients. Both TLR7 and TLR7rsq1 partitioned with the ER-resident protein calnexin (Fig. 6C). Taken together, these results show that the loss of function of human TLR7rsq1 is not due to its lack of expression or gross changes in its subcellular localization. Moreover, they confirm the concept that despite being type I glycoproteins with a leader signal sequence, both TLR7 and TLR7rsq1 are ER-resident proteins that are not translocated to the cell surface. TLR7rsq1 is dominant negative FIGURE 5. TLR7 requires an intact N-glycosylation consensus se- quence at positions 66–68. A, 293T transiently expressing the indicated Data in the literature support the notion that nucleic acid- TLR7 mutant (abscissa) and the NF-kB–luciferase reporter gene were left recognizing TLRs exist as preformed dimers and that the trig- untreated (nil) or were stimulated with R848 (5 mM). Four hours later, the gering of the downstream signaling cascade follows the rear- activity of the NF-kB–luciferase reporter gene was recorded by lumin- rangement of the extracellular domains in the dimer (19–21). We ometry. Results are representative of three independent experiments, and therefore asked whether TLR7rsq1 acts as dominant negative by 6 readings are the average SD of triplicate readings. RLU, relative lu- pairing with the wild-type molecule. 293T cells stably expressing ciferase units. B, The amounts of the HA-tagged TLR7 mutants used in the untagged TLR7 and an NF-kB–luciferase reporter gene were Downloaded from experiment in A were assessed by anti-HA Western blotting. transiently transfected with an empty vector or with HA-tagged versions of TLR7 or TLR7rsq1. The ectopic expression of rsq1 of human TLR7 and TLR7 . Both wild-type and mutant TLR7 TLR7rsq1 reduced TLR7 response to R848, whereas wild-type colocalized with a transiently transfected ER marker (Fig. 6A, TLR7 enhanced it (Fig. 7, left panel). The transiently expressed calreticulin–RFP merged panels), showing that both receptors are proteins were expressed at similar levels (data not shown). To similarly localized to the ER (Fig. 6A). To characterize further the

determine whether the impairment affected signaling cascades http://www.jimmunol.org/ subcellular localization of the two receptors, we assessed whether other than NF-kB, we tested whether TLR7rsq1 interfered with a fraction of the receptors underwent Golgi maturation and es- type I IFN production. When 293T cells are reconstituted with caped the typical ER-restricted compartmentalization of TLR7, TLR7 and IRF5, they secrete IFN-b in response to TLR7 agonists TLR8, and TLR9. Golgi-matured proteins acquire Endo-H re- (49). Similar to that observed for NF-kB, coexpression of sistance and therefore do not display a mobility shift when treated TLR7rsq1 and TLR7 decreased IFN-b secretion (Fig. 7, right rsq1 with Endo-H (32). As shown in Fig. 6B, both TLR7 and TLR7 panel), whereas coexpression of TLR3 did not interfere with were Endo-H sensitive, suggesting they did not go through the TLR7 signaling (data not shown). Our working model is that the inactive TLR7rsq1 associates with wild-type TLR7 in situ, thereby

preventing the formation of signaling-competent dimers. by guest on September 24, 2021 TLR7rsq1 and TLR7 bind similarly to ssRNA Structural evidence demonstrates that the N-terminal portion of TLR3 interacts with RNA (50), and TLR9 seems to use its N- terminal region to sense DNA (21). Because the rsq1 mutation is located in the N terminus of TLR7, we surmised that the rsq1 defect might be related to impaired ligand binding. However, the results in Fig. 8A show that TLR7 and TLR7rsq1, but not TLR3

FIGURE 6. TLR7 and TLR7rsq1 are localized to the ER. A, 293 cells stably expressing TLR7YFP (upper panels) or TLR7rsq1YFP (bottom panels) were transiently transfected with the RFP-tagged ER marker calreticulin (CALRRFP). Cells were then imaged by confocal microscopy. Shown are the signals from the green (left panels), red (middle panels), and merged YFP and RFP (right panels) fluorescence (original magnification 340). B, FIGURE 7. TLR7rsq1 is dominant negative. 293T cells stably expressing Whole-cell extracts from 293 cells stably expressing TLR7HA and the NF-kB–luciferase reporter and IRF5 were transiently transfected with TLR7rsq1HA were treated with Endo-H or Endo-F and subjected to the indicated plasmids: empty vector (dashed line), human TLR7 Western blotting for the HA epitope. Shown is the mobility shift due to the (squares), and human TLR7rsq1 (triangles). The cells were stimulated with complete removal of N-linked glycans in both proteins. C, 293TLR7 (left titrated amounts of R848. After an overnight incubation, the NF-kB re- panels) or 293TLR7rsq1 (right panels) were lysed and subjected to iodix- porter activation was determined by luminometry (left panel). Results are anol discontinuous gradient fractionation. Fractions were then probed by plotted as in Fig. 3A. IFN-b in the supernatants was determined by Me- Western blotting for the indicated subcellular localization markers: tubulin soScale Discovery analysis (right panel). These results were reproduced in (cytoplasm), LAMP1 (lysosomes), and calnexin (ER). three separate experiments. The Journal of Immunology 4219

FIGURE 8. TLR7rsq1 binds to ssRNA. A, Protein extracts from 293T cells or 293T cells stably expressing human TLR3HA, TLR7rsq1HA, or TLR7HA were incubated with biotinylated ssRNA (bio-R0006), and the complexes were collected using streptavidin–Sepharose beads. The presence of HA-tagged TLRs in the pellets was assessed by Western blotting. The shown experiment is representative of four similar independent experiments. B, WCLs from 293T cells stably expressing human TLR7rsq1HA or TLR7HA were left untreated (lane 1) or incubated with bio-R0006 or its complementary (nonstimulatory) bio- 6000R. Samples were then processed as in A. C, WCLs from 293T, 293TLR7HA, and 293TLR7rsq1HA cells were used to probe RNA plated on avidin-coated plates in ELISA-like experiments. An HRP-conjugated anti-HA Ab was mixed with the lysate before the incubation and the signal from the captured TLR7HA–anti-HA complex measured by luminometry using luminogenic HRP substrates. Plotted is the average of the light units from duplicate points. Biotinylated ssRNA (R0006) was plated in 2-fold dilution (x-axis, maximum concentration 1.4 mM). This experiment is representative of three experiments. Downloaded from

(nor TLR4 [data not shown]), were equally efficient in binding an thermodynamic and specificity will be presented elsewhere. Be- activating biotinylated ssRNA as revealed by avidin pulldown cause TLR7rsq1 colocalized with TLR7, was able to interfere with experiments from lysates containing the HA-tagged receptors its signaling, and displayed similar ligand-binding characteristics, (51). Under identical conditions, TLR7 and TLR7rsq1 bound we conclude that TLR7rsq1 is missing a feature that is critically weakly to the nonstimulatory ssRNA sequence 6000R, confirming linked to but not to its ability to bind to itself specificity of the assay (Fig. 8B). To determine potential differ- or its ligand(s). http://www.jimmunol.org/ ences in affinity, we tested the binding of TLR7 to ssRNA over rsq1 a wide range of concentrations in a quantitative assay. WCLs from TLR7 cells do not modulate gene expression in response to the 293T cells stably expressing HA-tagged TLR7 and TLR7rsq1 TLR7 ligands were used to probe titrated amounts of immobilized biotinylated To exclude formally that TLR7rsq1 had any residual activity and ssRNA. Wild-type and mutant TLR7 bound similarly to RNA that both ssRNA and small molecules have no TLR7-independent (Fig. 8B) suggesting that the T68I mutation does not involve a li- effects, we performed whole-genome microarray analysis to gand binding site. A detailed characterization of the TLR7 binding compare TLR7rsq1 with TLR7wt induced transcription profiles in by guest on September 24, 2021

FIGURE 9. Whole-genome microarray analysis of TLR7rsq1 mutant. 293TLR7 and 293TLR7rsq1 cells were treated with 10 mM R848 or 5 mg/ml R0006 for 4 h. Cy5- and Cy3-labeled cranes were hybridized onto 4 3 44 Whole Microarray. The dif- ferentially expressed genes (log2 $ |1|, p value #0.05) are shown in the heat map. Hierarchical clustering of the genes was performed using TMEV. For fold in- ductions, refer to Table I. A, Cluster analysis of genes upregulated by R848 (left panel) and ssRNA (right panel). B, Cluster analysis of genes upregulated in 293TLR7 wild-type cells by R848, ssRNA, or both. 4220 TLR7rsq1 CHARACTERIZATION

293T transfectants and primary cells from mutant and wild-type and their wild-type counterparts was observed in the genes in- mice. In the first set of experiments, 106 differentially regulated duced by treatment with the TLR4 agonist LPS. GO analysis genes with a threshold of log2 $ |1| and with a p value #0.05 in at showed that the majority of genes regulated by TLR7 ligands were least one condition were identified. All of the genes modulated by related to immunological functions. As shown in Fig. 10A (im- R0006 and R848 treatment were TLR7 dependent (Fig. 9A). No mune genes), Fig. 10B (chemokines), and Fig. 10C (cytokines), genes were significantly upregulated or downregulated in either there was a striking lack of genes modulated by R848 or ssRNA in 293TLR7rsq1 (Fig. 9A) or in 293T cells (data not shown). Analysis the TLR7rsq1 splenocytes, suggesting that both the small molecule of the gene clusters modulated in 293TLR7 cells showed that the and ssRNA are not sensed by mechanisms other than TLR7. Be- majority of significant (GO) terms are related to cause of the overall similar transcriptional signature, these seem- immune responses (e.g., IL-1b, IL-8, TNF-a). A group of cyto- ingly completely unrelated classes of ligands seem to share the kines and chemokines like CXCL members, secreted effector same mechanism of action. molecules like pentraxin3 (Ptx3), as well as CD83 and CD44 were upregulated in response to both R848 and ssRNA. Fig. 9B reveals Discussion that there are genes that seemed to be regulated by only one The T68I conversion in the first LRR of the extracellular domain of compound or the other. Given the small number of genes modu- TLR7 has been identified because cells from mice bearing this lated and the different nature of the two stimuli, small drug-like mutation are unable to respond to R848 (52). In addition to being molecule versus immunostimulatory synthetic ssRNA complexed insensitive to low m.w. TLR7 agonists (three distinct chemotypes, with a cationic lipid, these differences do not seem to be signifi- data not shown), we found that TLR7rsq1-bearing cells fail to cant and possibly related to different kinetics and potency (see signal upon treatment with short synthetic ssRNA oligonucleo- Downloaded from later). Hence, the transcriptome analysis of ectopically recon- tides as well. Despite extensive investigation, we were unable to stituted 293T cells indicates that TLR7rsq1 does not produce any identify the exact molecular defect caused by the T68I mutation. significant signaling event. In a second set of experiments, a However, our studies reveal several novel aspects of TLR7 bi- whole-genome transcriptional analysis was performed on treated ology. Because the TLR7rsq1 mutation does not affect binding of splenocytes from the TLR7rsq1 and the C57BL/6 mice: 1562 genes TLR7 to RNA (Fig. 8A), it is unlikely that the region encom-

were differentially regulated with a p value #0.05 and a fold passing T68 participates directly in the binding to R848 or RNA. http://www.jimmunol.org/ change $4-fold in splenocytes treated with R848, ssRNA, and Accordingly, the modeling of TLR7 reveals that this mutation LPS (as control). No significant difference between TLR7rsq1 mice resides in the glycan-rich face of TLR7, which, based on available by guest on September 24, 2021

FIGURE 10. Microarray analysis of spleno- cytes from wild-type and TLR7rsq1 mutant mice treated with LPS, R848, and ssRNA. Splenocytes from the indicated mice were treated with LPS (0.1 mg/ml), R848 (10 mM), or R0006 (5 mg/ml) for 4 h. Cy5- and Cy3- labeled cRNAs were hybridized onto 4 3 44,000 Whole Mouse Genome Microarray (Agilent Technologies). The differentially expressed genes (log2 $ |1|, p value #0.05) are shown in the heat map. Shown are the GO group of (A) immune response (GO0006955), (B) chemokine activity (GO0008009), and (C) cytokine activity (GO0005125). Hierarchical clustering of the genes was performed using TMEV. The Journal of Immunology 4221

TLR-liganded structures, should not contain the ligand binding result in an altered responsiveness of the other (e.g., TLR9). It site. It is interesting to note that residues in the N terminus of would therefore be informative to confirm these experiments in TLR9 have been identified that impair CpG DNA signaling but not a mouse strain in which TLR7 protein is expressed but not func- binding (21). Hence, ligand binding and signal transduction seem tional. TLR7rsq1 would represent such a model and may help to to be separate events for both TLR9 and TLR7, which is in con- elucidate the contribution of TLR7 to SLE and other related auto- trast to other TLRs for which dimerization induced by their re- immune disorders. spective ligands is the activating event. For example, the prop- agation of an out–in activating signal by the intracellular domain Acknowledgments of TLR4 can be achieved by Ab cross-linking (10, 53), by in- We acknowledge Drs. I. Cornella and J. Tallarico (Novartis Institutes for ducing its aggregation via the extracellular domain from different Biomedical Research) for synthesizing and providing all the synthetic TLRs (5), or by using extracellular moieties known to produce low m.w. TLR7 agonists, Drs. M. Brinkmann and Hidde Ploegh for provid- rsq1 dimeric structures (54). Given the ability of TLR7 to bind to ing the polyclonal anti-mouse TLR7 Ab, and Dr. V. Pasquetto for logistic RNA (Fig. 8), its ability to interfere with TLR7 signaling (Fig. 7), support. We are also indebted to Drs. M. Cooke, B. Nayak, C. Schmedt, its ability to colocalize (Figs. 2, 6) and bind to wild-type TLR7 E. De Gregorio, E. Soldaini, and Flaviana Mosca for helpful discussions (data not shown), we infer that the loss of a putative N-linked during the preparation of the manuscript. polysaccharide does not result in a gross misfolding of the en- tire receptor. Rather, because glycosylation has been shown to Disclosures affect the induction of compact b-turns (55) and N66 is directly C.I., J.C.D., I.L., U.D’O., N.M.V., M.L.M., P.Z., G.O., and A.V. were located in the turn leading up to the LRR b-strand (Fig. 4), the employees of Novartis Vaccines and Diagnostics at the time this study Downloaded from lack of the sugar moiety could result in the structural perturbation was submitted for publication. of a defined region that is critical to TLR7 function. Mutagenesis studies in murine TLR7 and TLR9 illustrated the sensitivity of References these receptors to subtle changes (21). Although not formally 1. Iwasaki, A., and R. Medzhitov. 2010. Regulation of adaptive by the demonstrated for TLR7, the structural perturbation induced by the . Science 327: 291–295. 2. Heil, F., H. Hemmi, H. Hochrein, F. Ampenberger, C. Kirschning, S. Akira, lack of the N-glycosylation in position 66 might affect a confor- http://www.jimmunol.org/ G. Lipford, H. Wagner, and S. Bauer. 2004. Species-specific recognition of mational change required for triggering the dimeric receptor single-stranded RNA via toll-like receptor 7 and 8. Science 303: 1526–1529. complex, as reported for TLR9 (19), or impact the ability of TLR7 3. Diebold, S. S., T. Kaisho, H. Hemmi, S. Akira, and C. Reis e Sousa. 2004. Innate antiviral responses by means of TLR7-mediated recognition of single-stranded to pair with a luminal signaling partner, such as a protease (23) or RNA. Science 303: 1529–1531. a chaperone protein (56). 4. Martinez, J., X. Huang, and Y. Yang. 2010. Toll-like receptor 8-mediated acti- Whatever the nature of this event is, TLR7 produces remarkably vation of murine plasmacytoid dendritic cells by vaccinia viral DNA. Proc. Natl. Acad. Sci. USA 107: 6442–6447. similar responses to seemingly unrelated classes of TLR7 ligands: 5. Barton, G. M., J. C. Kagan, and R. Medzhitov. 2006. Intracellular localization of purine-like low m.w. TLR7 agonists and ssRNA. The few differ- Toll-like receptor 9 prevents recognition of self DNA but facilitates access to ences found in the transcriptional signature of ssRNA and R848 in viral DNA. Nat. Immunol. 7: 49–56. 6. Stewart, C. R., L. M. Stuart, K. Wilkinson, J. M. van Gils, J. Deng, A. Halle, by guest on September 24, 2021 293TLR7 cells might be ascribed to the different kinetics of ac- K. J. Rayner, L. Boyer, R. Zhong, W. A. Frazier, et al. 2010. CD36 ligands tivation of TLR7 by a cell-permeable small molecule versus a lipid- promote sterile inflammation through assembly of a Toll-like receptor 4 and 6 heterodimer. Nat. Immunol. 11: 155–161. delivered RNA. This notion is corroborated by the findings that in 7. Hoebe, K., P. Georgel, S. Rutschmann, X. Du, S. Mudd, K. Crozat, S. Sovath, primary immune cells, the pattern of activation was almost iden- L. Shamel, T. Hartung, U. Za¨hringer, and B. Beutler. 2005. CD36 is a sensor of tical for R848 and ssRNA. On the contrary, TLR7rsq1 cells failed to diacylglycerides. Nature 433: 523–527. 8. Miyake, K. 2003. Innate recognition of lipopolysaccharide by CD14 and toll-like produce any significant signal in response to either agonist in both receptor 4-MD-2: unique roles for MD-2. Int. Immunopharmacol. 3: 119–128. cellular systems, suggesting that TLR7 is a nonredundant and 9. Shimazu, R., S. Akashi, H. Ogata, Y. Nagai, K. Fukudome, K. Miyake, and exclusive receptor for both synthetic ssRNA oligonucleotides and M. Kimoto. 1999. MD-2, a molecule that confers lipopolysaccharide re- sponsiveness on Toll-like receptor 4. J. Exp. Med. 189: 1777–1782. R848. 10. Visintin, A., E. Latz, B. G. Monks, T. Espevik, and D. T. Golenbock. 2003. Our data show that TLR7rsq1 mice express a mutant form of Lysines 128 and 132 enable lipopolysaccharide binding to MD-2, leading to Toll-like receptor-4 aggregation and signal transduction. J. Biol. Chem. 278: TLR7 that results in a phenocopy of TLR7 knockout mice (45). 48313–48320. Because the loss of TLR7 function is not due to the lack of receptor 11. Randow, F., and B. Seed. 2001. Endoplasmic reticulum chaperone gp96 is re- expression, this novel mouse strain should better recapitulate the quired for innate immunity but not cell viability. Nat. Cell Biol. 3: 891–896. 12. Konno, K., Y. Wakabayashi, S. Akashi-Takamura, T. Ishii, M. Kobayashi, effects of specific TLR7-inactivating drugs and therefore is a valu- K. Takahashi, Y. Kusumoto, S. I. Saitoh, Y. Yoshizawa, and K. Miyake. 2006. A able tool to model the effect of such compounds in an in vivo set- molecule that is associated with Toll-like receptor 4 and regulates its cell surface ting. Moreover, a deeper understanding of the unique molecular expression. Biochem. Biophys. Res. Commun. 339: 1076–1082. 13. Takahashi, K., T. Shibata, S. Akashi-Takamura, T. Kiyokawa, Y. Wakabayashi, defect leading to nonfunctional TLR7 signaling could identify N. Tanimura, T. Kobayashi, F. Matsumoto, R. Fukui, T. Kouro, et al. 2007. A novel drug targets for TLR7 antagonist discovery in the treatment protein associated with Toll-like receptor (TLR) 4 (PRAT4A) is required for of autoimmune disorders. The role of TLR7 has been studied par- TLR-dependent immune responses. J. Exp. Med. 204: 2963–2976. 14. Tabeta, K., K. Hoebe, E. M. Janssen, X. Du, P. Georgel, K. Crozat, S. Mudd, ticularly in the context of systemic lupus erythematosus (SLE), an N. Mann, S. Sovath, J. Goode, et al. 2006. The Unc93b1 mutation 3d disrupts autoimmune disease in which this receptor has been implicated exogenous antigen presentation and signaling via Toll-like receptors 3, 7 and 9. Nat. Immunol. 7: 156–164. (reviewed in Ref. 57). Blocking of TLR7 and TLR9 function by 15. Brinkmann, M. M., E. Spooner, K. Hoebe, B. Beutler, H. L. Ploegh, and dual inhibitory DNA oligonucleotides resulted in improved thera- Y. M. Kim. 2007. The interaction between the ER membrane protein UNC93B peutic outcomes for SLE patients (58). This is in agreement with and TLR3, 7, and 9 is crucial for TLR signaling. J. Cell Biol. 177: 265–275. 16. Leonard, J. N., R. Ghirlando, J. Askins, J. K. Bell, D. H. Margulies, studies focusing on the molecular contribution of TLR7 in murine D. R. Davies, and D. M. Segal. 2008. The TLR3 signaling complex forms by SLE models. In these studies, genetic removal of TLR7 was ob- cooperative receptor dimerization. Proc. Natl. Acad. Sci. USA 105: 258–263. served to ameliorate the disease phenotype, suggesting that TLR7 17. Ranjith-Kumar, C. T., W. Miller, J. Xiong, W. K. Russell, R. Lamb, J. Santos, K. E. Duffy, L. Cleveland, M. Park, K. Bhardwaj, et al. 2007. Biochemical and drives autoimmunity. TLR9, in contrast, was shown to have an functional analyses of the human Toll-like receptor 3 ectodomain. J. Biol. Chem. opposite effect, and its removal resulted in an exacerbated course of 282: 7668–7678. 18. Bell, J. K., I. Botos, P. R. Hall, J. Askins, J. Shiloach, D. M. Segal, and disease (59). Because both receptors use the common downstream D. R. Davies. 2005. The molecular structure of the Toll-like receptor 3 ligand- adaptor MyD88, genetic removal of one receptor (e.g., TLR7) may binding domain. Proc. Natl. Acad. Sci. USA 102: 10976–10980. 4222 TLR7rsq1 CHARACTERIZATION

19. Latz, E., A. Verma, A. Visintin, M. Gong, C. M. Sirois, D. C. Klein, 39. Guex, N., and M. C. Peitsch. 1997. SWISS-MODEL and the Swiss-PdbViewer: B. G. Monks, C. J. McKnight, M. S. Lamphier, W. P. Duprex, et al. 2007. an environment for comparative protein modeling. Electrophoresis 18: 2714– Ligand-induced conformational changes allosterically activate Toll-like receptor 2723. 9. Nat. Immunol. 8: 772–779. 40. Dolinsky, T. J., J. E. Nielsen, J. A. McCammon, and N. A. Baker. 2004. 20. Zhu, J., R. Brownlie, Q. Liu, L. A. Babiuk, A. Potter, and G. K. Mutwiri. 2009. PDB2PQR: an automated pipeline for the setup of Poisson-Boltzmann electro- Characterization of bovine Toll-like receptor 8: ligand specificity, signaling es- statics calculations. Nucleic Acids Res. 32(Web Server issue): W665–W667. sential sites and dimerization. Mol. Immunol. 46: 978–990. 41. Li, H., A. D. Robertson, and J. H. Jensen. 2005. Very fast empirical prediction 21. Peter, M. E., A. V. Kubarenko, A. N. Weber, and A. H. Dalpke. 2009. Identifi- and rationalization of protein pKa values. Proteins 61: 704–721. cation of an N-terminal recognition site in TLR9 that contributes to CpG-DNA- 42. Brigtsen, A. K., D. L. Kasper, C. J. Baker, H. J. Jennings, and H. K. Guttormsen. mediated receptor activation. J. Immunol. 182: 7690–7697. 2002. Induction of cross-reactive antibodies by immunization of healthy adults 22. Rutz, M., J. Metzger, T. Gellert, P. Luppa, G. B. Lipford, H. Wagner, and with types Ia and Ib group B streptococcal polysaccharide-tetanus toxoid con- S. Bauer. 2004. Toll-like receptor 9 binds single-stranded CpG-DNA in a se- jugate vaccines. J. Infect. Dis. 185: 1277–1284. quence- and pH-dependent manner. Eur. J. Immunol. 34: 2541–2550. 43. Pettit, F. K., E. Bare, A. Tsai, and J. U. Bowie. 2007. HotPatch: a statistical 23. Park, B., M. M. Brinkmann, E. Spooner, C. C. Lee, Y. M. Kim, and H. L. Ploegh. approach to finding biologically relevant features on protein surfaces. J. Mol. 2008. Proteolytic cleavage in an endolysosomal compartment is required for Biol. 369: 863–879. activation of Toll-like receptor 9. Nat. Immunol. 9: 1407–1414. 44. Gottipati, S., N. L. Rao, and W. P. Fung-Leung. 2008. IRAK1: a critical signaling 24. von Krogh, G., C. J. Lacey, G. Gross, R. Barrasso, and A. Schneider. 2000. mediator of innate immunity. Cell. Signal. 20: 269–276. European course on HPV associated pathology: guidelines for primary care 45. Hemmi, H., T. Kaisho, O. Takeuchi, S. Sato, H. Sanjo, K. Hoshino, T. Horiuchi, physicians for the diagnosis and management of anogenital warts. Sex. Transm. H. Tomizawa, K. Takeda, and S. Akira. 2002. Small anti-viral compounds ac- Infect. 76: 162–168. tivate immune cells via the TLR7 MyD88-dependent signaling pathway. Nat. 25. Zhu, J., K. Lai, R. Brownile, L. A. Babiuk, and G. K. Mutwiri. 2008. Porcine Immunol. 3: 196–200. TLR8 and TLR7 are both activated by a selective TLR7 ligand, imiquimod. Mol. 46. Mellquist, J. L., L. Kasturi, S. L. Spitalnik, and S. H. Shakin-Eshleman. 1998. Immunol. 45: 3238–3243. The amino acid following an asn-X-Ser/Thr sequon is an important determinant 26. Gorden, K. K., X. X. Qiu, C. C. Binsfeld, J. P. Vasilakos, and S. S. Alkan. 2006. of N-linked core glycosylation efficiency. Biochemistry 37: 6833–6837. Cutting edge: activation of murine TLR8 by a combination of imidazoquinoline 47. Shakin-Eshleman, S. H., S. L. Spitalnik, and L. Kasturi. 1996. The amino acid at immune response modifiers and polyT oligodeoxynucleotides. J. Immunol. 177: the X position of an Asn-X-Ser sequon is an important determinant of N-linked Downloaded from 6584–6587. core-glycosylation efficiency. J. Biol. Chem. 271: 6363–6366. 27. Bauer, S., B. Bathke, H. Lauterbach, J. Patzold, R. Kassub, C. A. Luber, 48. Buchanan, S. G. S., and N. J. Gay. 1996. Structural and functional diversity in B. Schlatter, S. Hamm, P. Chaplin, M. Suter, and H. Hochrein. 2010. A major the leucine-rich repeat family of proteins. Prog. Biophys. Mol. Biol. 65: 1–44. role for TLR8 in the recognition of vaccinia viral DNA by murine pDC? Proc. 49. Schoenemeyer, A., B. J. Barnes, M. E. Mancl, E. Latz, N. Goutagny, P. M. Pitha, Natl. Acad. Sci. USA 107: E139; author reply E140. K. A. Fitzgerald, and D. T. Golenbock. 2005. The interferon regulatory factor, 28. Hoebe, K., and B. Beutler. 2008. Forward genetic analysis of TLR-signaling IRF5, is a central mediator of toll-like receptor 7 signaling. J. Biol. Chem. 280: pathways: an evaluation. Adv. Drug Deliv. Rev. 60: 824–829. 17005–17012.

29. Bell, J. K., G. E. Mullen, C. A. Leifer, A. Mazzoni, D. R. Davies, and 50. Liu, L., I. Botos, Y. Wang, J. N. Leonard, J. Shiloach, D. M. Segal, and http://www.jimmunol.org/ D. M. Segal. 2003. Leucine-rich repeats and pathogen recognition in Toll-like D. R. Davies. 2008. Structural basis of toll-like receptor 3 signaling with double- receptors. Trends Immunol. 24: 528–533. stranded RNA. Science 320: 379–381. 30. Forsbach, A., J. G. Nemorin, C. Montino, C. Mu¨ller, U. Samulowitz, A. P. Vicari, 51. Latz, E., A. Schoenemeyer, A. Visintin, K. A. Fitzgerald, B. G. Monks, M. Jurk, G. K. Mutwiri, A. M. Krieg, G. B. Lipford, and J. Vollmer. 2008. C. F. Knetter, E. Lien, N. J. Nilsen, T. Espevik, and D. T. Golenbock. 2004. Identification of RNA sequence motifs stimulating sequence-specific TLR8- TLR9 signals after translocating from the ER to CpG DNA in the lysosome. Nat. dependent immune responses. J. Immunol. 180: 3729–3738. Immunol. 5: 190–198. 31. Baccarini, M., F. Bistoni, and M. L. Lohmann-Matthes. 1985. In vitro natural 52. Beutler, B., Z. Jiang, P. Georgel, K. Crozat, B. Croker, S. Rutschmann, X. Du, cell-mediated cytotoxicity against Candida albicans: macrophage precursors as and K. Hoebe. 2006. Genetic analysis of host resistance: Toll-like receptor effector cells. J. Immunol. 134: 2658–2665. signaling and immunity at large. Annu. Rev. Immunol. 24: 353–389. 32. Visintin, A., A. Mazzoni, J. A. Spitzer, and D. M. Segal. 2001. Secreted MD-2 is 53. Latz, E., A. Visintin, E. Lien, K. A. Fitzgerald, B. G. Monks, E. A. Kurt-Jones, a large polymeric protein that efficiently confers lipopolysaccharide sensitivity to D. T. Golenbock, and T. Espevik. 2002. Lipopolysaccharide rapidly traffics to

Toll-like receptor 4. Proc. Natl. Acad. Sci. USA 98: 12156–12161. and from the Golgi apparatus with the toll-like receptor 4-MD-2-CD14 complex by guest on September 24, 2021 33. Vallon-Christersson, J., N. Nordborg, M. Svensson, and J. Ha¨kkinen. 2009. in a process that is distinct from the initiation of signal transduction. J. Biol. BASE—2nd generation software for microarray data management and analysis. Chem. 277: 47834–47843. BMC Bioinformatics 10: 330. 54. Medzhitov, R., P. Preston-Hurlburt, and C. A. Janeway, Jr. 1997. A human ho- 34. Saeed, A. I., V. Sharov, J. White, J. Li, W. Liang, N. Bhagabati, J. Braisted, mologue of the Drosophila Toll protein signals activation of adaptive immunity. M. Klapa, T. Currier, M. Thiagarajan, et al. 2003. TM4: a free, open-source Nature 388: 394–397. system for microarray data management and analysis. Biotechniques 34: 374– 55. Helenius, A., and M. Aebi. 2001. Intracellular functions of N-linked glycans. 378. Science 291: 2364–2369. 35. Larkin, M. A., G. Blackshields, N. P. Brown, R. Chenna, P. A. McGettigan, 56. Liu, J., C. Xu, L. C. Hsu, Y. Luo, R. Xiang, and T. H. Chuang. 2010. A five- H. McWilliam, F. Valentin, I. M. Wallace, A. Wilm, R. Lopez, et al. 2007. amino-acid motif in the undefined region of the TLR8 ectodomain is required for Clustal W and Clustal X version 2.0. Bioinformatics 23: 2947–2948. species-specific ligand recognition. Mol. Immunol. 47: 1083–1090. 36. Sali, A., and J. P. Overington. 1994. Derivation of rules for comparative protein 57. Krieg, A. M. 2007. The toll of too much TLR7. Immunity 27: 695–697. modeling from a database of protein structure alignments. Protein Sci. 3: 1582– 58. Barrat, F. J., T. Meeker, J. H. Chan, C. Guiducci, and R. L. Coffman. 2007. 1596. Treatment of lupus-prone mice with a dual inhibitor of TLR7 and TLR9 leads to 37. Melo, F., D. Devos, E. Depiereux, and E. Feytmans. 1997. ANOLEA: a www reduction of autoantibody production and amelioration of disease symptoms. server to assess protein structures. Proc. Int. Conf. Intell. Syst. Mol. Biol. 5: 187– Eur. J. Immunol. 37: 3582–3586. 190. 59. Nickerson, K. M., S. R. Christensen, J. Shupe, M. Kashgarian, D. Kim, K. Elkon, 38. Van Der Spoel, D., E. Lindahl, B. Hess, G. Groenhof, A. E. Mark, and and M. J. Shlomchik. 2010. TLR9 regulates TLR7- and MyD88-dependent au- H. J. Berendsen. 2005. GROMACS: fast, flexible, and free. J. Comput. Chem. 26: toantibody production and disease in a murine model of lupus. J. Immunol. 184: 1701–1718. 1840–1848.