Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2018 Mar 21.
Published in final edited form as: Immunity. 2017 Mar 21;46(3):379–392. doi: 10.1016/j.immuni.2017.02.017

The IFN-λ-IFN-λR1-IL-10Rβ Complex Reveals Structural Features Underlying Type III IFN Functional Plasticity

Juan L Mendoza 1, William M Schneider 2, Hans-Heinrich Hoffmann 2, Koen Vercauteren 2, Kevin M Jude 1, Anming Xiong 3, Ignacio Moraga 1, Tim M Horton 1, Jeffrey S Glenn 3, Ype P de Jong 2,4, Charles M Rice 2, K Christopher Garcia 1,5,✉
PMCID: PMC5510750  NIHMSID: NIHMS869094  PMID: 28329704

Summary

Type III interferons (IFN-λs) signal through a heterodimeric receptor complex composed of the IFN-λR1 subunit, specific for IFN-λs, and interleukin-10Rβ (IL-10Rβ), which is shared by multiple cytokines in the IL-10 superfamily. Low affinity of IL-10Rβ for cytokines has impeded efforts aimed at crystallizing cytokine-receptorcomplexes. We used yeast surface display to engineer a higher-affinity IFN-λ variant, H11, which enabled crystallization of the ternary complex. The structure revealed that IL-10Rβ uses a network of tyrosine residues as hydrophobic anchor points to engage IL-10 family cytokines that present complementary hydrophobic binding patches, explaining its role as both a cross-reactive but cytokine-specific receptor. H11 elicited increased anti-proliferative and antiviral activities in vitro and in vivo. In contrast, engineered higher-affinity type I IFNs did not increase antiviral potency over wild-type type I IFNs. Our findings provide insight into cytokine recognition by the IL-10R family and highlight the plasticity of type III interferon signaling and its therapeutic potential.

Graphical abstract

Using an engineered high-affinity IFN-λ, Mendoza et al. solve the structure of the IFN-λ/IFN-λR1/IL-10Rβ ternary signaling complex. The structure reveals how IL-10Rβ can act as both a cross-reactive but cytokine-specific receptor. Structure-activity relationships of engineered type I and III IFNs provide insights into enhancing interferon functional potency.

graphic file with name nihms869094u1.jpg

Introduction

Lambda interferons (IFN-λ1–4), also known as type III IFNs, are members of the IL-10 superfamily of cytokines (Kotenko et al., 2003; Sheppard et al., 2003). Like type I IFNs, these secreted cytokines signal through cell surface receptors and elicit an innate immune response to combat viral infections and cancer (Donnelly and Kotenko, 2010; Steen and Gamero, 2010). Recent studies highlight the importance and non-overlapping role of IFN-λs in curing persistent norovirus infections (Baldridge et al., 2015; Nice etal., 2015), and they also demonstrate efficacy against a number of other viral infections (Lazear et al., 2015). While the type I IFN receptors are ubiquitously expressed, expression of the IFN-λ receptor is restricted to immune cell subsets and barrier tissues. This feature gives IFN-λs unique potential for targeted therapies with fewer side effects then type I IFN therapies (Muir et al., 2014). This promising antiviral activity of IFN-λs has led to their evaluation in clinical trials for use in hepatitis C virus (HCV)- (Muir et al., 2014), hepatitis B virus (HBV)- (Hruska et al., 2015; Wang et al., 2015), and hepatitis D virus (HDV)-infected patients (clinicaltrials.gov identifier NCT02765802).

Although type I IFNs and IFN-λs signal through entirely different receptors, they utilize the same Janus kinases, JAK1 and TYK2, and signal through a common JAK/STAT pathway to induce interferon-stimulated genes (ISGs), which elicit antiviral and immunoregulatory activities. Despite these similarities, the IFN-λs have lower potency in signaling (as measured by lower Emax of STAT activation), lower induction of interferon-stimulated genes, and higher EC50s for the antiviral and antiprolifera-tive activities compared to type I IFNs (Ank et al., 2006; Dellgren et al., 2009; Dickensheets et al., 2013; Meager et al., 2005). One explanation for their reduced potency is that IFN-λs form less stable ternary complexes than type I IFNs. In principle, a higher-affinity IFN-λ could have improved clinical activity with less toxicity than type I IFNs.

The IFN-λ receptor is composed of IL-10Rβ and IFN-λR1, which forms sequentially in a two-step binding event. First, the IFN-λR1 subunit binds IFN-λs with high affinity, and it is the limited expression of this subunit that provides IFN-λs with their unique tissue specificity. In contrast, the IL-10Rβ subunit is broadly expressed and binds the IFN-λ binary complex with low affinity in the second step of complex formation. The IL-10Rβ subunit is also shared by receptor complexes that engage the IL-10 superfamily members IL-10, IL-22, and IL-26 (Kotenko et al., 2003; Sheppard et al., 2003). A crystal structure of unliganded IL-10Rβ (Yoon et al., 2010), together with binding studies of single point mutants, identified Tyr82 as playing a central role in its cross-reactive binding to IL-10 cytokines. Although models of ternary complexes of IL-10 family members have been proposed (Miknis et al., 2010; Yoon et al., 2010; Zdanov, 2010), a complete structure is lacking, and it is still unclear how IL-10Rβ recognizes a wide spectrum of ligands with their cytokine-specific receptors. Obtaining a ligand-bound structure to clarify this puzzle has been challenging, in part, because of the low affinity of IL-10Rβ-ligand interactions (Logsdon et al., 2004; Yoon et al., 2006).

Despite poor sequence conservation (< 25% identity), IL-10 superfamily cytokines are structurally conserved (Jones et al., 2008a; Zdanov, 2010). The binary complex structures of IL-10 cytokines IL-22/IL-22R1, IL-10/IL-10R1, and IFN-λ1/IFN-λR1 (Bleicher et al., 2008; Jones et al., 2008b; Josephson et al., 2001; Miknis et al., 2010; Zdanov, 2010) adopt a similar ligand-receptor docking topology governed by contacts near the inter-domain “elbow” of their high-affinity receptor chains. Mutagenesis studies combined with computational docking models have lead to several proposed models of IL-10 super-family ternary complexes, yet an experimentally determined structure has been elusive (Bleicher et al., 2008; Gad et al., 2009; Jones et al., 2008a; Logsdon et al., 2004; Miknis et al., 2010; Wu et al., 2008; Yoon et al., 2006, 2010; Zdanov, 2010).

We decided to take an affinity maturation approach to increase the binding affinity of IFN-λ3 for the IL-10Rβ subunit, so as to facilitate crystallization of a ternary complex. This approach enabled the isolation of a higher affinity IFN-λ3 variant, and the purification of a stable IFN-λ/IL-10Rβ/IFN-λR1 ternary complex. Structural analyses of the ternary complex revealed a network of tyrosine residues that serve as hydrophobic anchor points to engage IL-10 family cytokines. In addition, the high-affinity variant had increased antiviral and antiproliferative properties. In contrast, engineering of type I IFNs did not significantly increase downstream signaling or antiviral activity, suggesting that these are already at a maximum. Taken together with the lower toxicity profiles of IFN-λs, our findings argue for the further development of type III IFNs as alternative antiviral and anti-cancer therapeutics.

Results

Engineering a High-Affinity Type III Interferon

We used yeast surface display as a platform to engineer a higher-affinity IFN-λ for structural and functional studies. We focused our efforts on IFN-λ3 because a crystal structure has been reported (Gad et al., 2009) and because it has the highest antiviral potency among the natural IFN-λs (Dellgren et al., 2009). We targeted the interaction between IFN-λ3 and IL-10Rb for affinity maturation because it is of much lower affinity than the interaction between IFN-λ3 and IFN-λR1. When displayed on yeast, IFN-λ3 bound to IFN-λR1 with a titration midpoint of approximately 400 nM (Figures S1A and S1B). We were unable to detect binding of IFN-λ3 to either a monomeric (data not shown), or an avidity-enhanced tetrameric form of IL-10Rβ (Figure 1A, left panel). However, in the presence of IFN-λR1 tetramers, IFN-λ3 bound to tetrameric (Figure 1A, right panel), but not monomeric IL-10Rβ (Figure 1A, middle panel). Thus, yeast displayed IFN-λ3 is properly folded and forms a binary complex with IFN-λR1 that is recognized by IL-10Rβ, albeit with very low affinity (Figure 1A). We carried out IFN-λ3 affinity maturation experiments for IL-10Rβ in the presence of soluble IFN-λR1.

Figure 1. Engineering a High-Affinity IFN-λ.

Figure 1

Open in a new tab

(A) Wild-type IFN-λ3 (Gad et al., 2009) was displayed on-yeast and stained with 1 μM monomer (left) or 400 nM streptavidin tetramers of IL-10Rβ (Yoon et al.,2010) labeled with Alexa 647 (middle and right panel) (x axis) in the presence (right panel) or absence (middle panel) of 50 nM streptavidin tetramers of IFN-λR1(Miknis et al., 2010) labeled with phycoerythrin (PE) (y axis).

(B) Histogram plot of fluorescence (Alexa 647) for the naive and evolved first-generation (error-prone) and second-generation (shuffled) IFN-λ3 yeast displayedlibraries stained with 1 μM IL-10Rβ monomers in the presence IFN-λR1.

(C) Sequences and on-yeast affinity measurements of evolved IFN-λ mutants (H11 mutations are colored in light red) compared to wild-type (light blue).

(D) IL-10Rβ affinity for the wild-type IFN-λR1/IFN-λ3 or H11-containing binary complexes was determined by surface plasmon resonance. KD values weredetermined by fitting to a first order equilibrium binding model. See also Figure S1.

To increase the affinity of the IFN-λ3/IL-10Rβ interaction, yeast-displayed IFN-λ3 was mutagenized using error-prone PCR in a first-generation library and gene shuffling in a second-generation library, and higher-affinity mutants were selected with fluorescently labeled streptavidin-bound IL-10Rβ. The first-generation yeast library contained 1 × 108 clones expressing IFN-λ3 and was subjected to four rounds of increasingly stringent selections ranging from 400 nM IL-10Rβ tetramers in the first round to 1 μM IL-10Rβ monomers in the fourth round. The enriched library showed some binding to 1 μM IL-10Rβ monomers (Figure 1B, top two histograms). 96 clones were screened for 1 μM IL-10Rβ monomer binding and the affinities of interacting clones were measured by titrations with the IL-10Rβ receptor. The six highest affinity clones, (all of which had > 1 μM affinity for IL-10Rβ), were then used as parental templates in a DNA shuffling reaction (Figure 1C). The resulting second generation library contained 1×108 clones and was subjected to three rounds of selection against decreasing concentrations of IL-10Rβ monomer ranging from 1 μM to 125 nM (Figure 1B, bottom two histograms). 96 clones were screened for binding to IL-10Rβ followed by yeast-surface titrations to measure affinity. The highest affinity clone, denoted “H11,” was found to have an “on-yeast” KD of 200 nM to IL-10Rβ (Figure 1C). Sequence analysis of H11 revealed that the gene contained five mutations relative to wild-type IFN-λ3 and was a combination of four first generation sequences from the DNA shuffling reaction (Figures 1C and S1C). H11 was expressed recombinantly and the affinity of IL-10Rβ to the immobilized H11/IFN-λR1 binary complex was determined by surface plasmon resonance. The H11/IFN-λR1 complex showed a 30-fold increase in IL-10Rβ affinity (KD = 560 nM) compared to the wild-type IFN-λ3/IFN-λR1 binary complex (Figure 1D). H11 has a 5-fold higher affinity for IFN-λR1 (KD∼150 nM) relative to the wild-type IFN-λ3 affinity for IFN-λR1 (KD∼850 nM) (Figure S1D). One of the H11 mutations, Thr150Ala, is located at the center of the IFN-λR1 binding site. Collectively, the mutations in H11 increased its affinity for both IFN-λR1 and IL-10Rβ by 150-fold.

Structure of the IFN-λ3 Receptor Ternary Complex

The lack of a complete structure of a ternary complex of an IL-10 family cytokine is likely due to the low affinity of IL-10Rβ. Indeed, when we attempted to form a ternary complex between wild-type IFN-λ3, IFN-λR1, and IL-10Rβ, we did not observe a stable ternary complex by gel filtration chromatography. However, IFN-λ3 H11 was able to form a stable ternary complex with IFN-λR1 and IL-10Rβ (Figure S1E). We crystallized the N-glycan minimized ternary complex using IFN-λR1 expressed from HEK293 GnTi− cells (which was treated with Endoglycosidase F and H), and a mutant form of IL-10Rβ (expressed from insect cells) in which four N-linked glycan sites were mutated to Gln. This complex yielded crystals that diffracted X-rays to 2.85 A (Table S1). The structure of the ternary complex was solved using molecular replacement with the binary structure of IFN-λ1/IFN-λR1, PDB: 3OG6 (Miknis et al., 2010) and the unbound structure of IL-10Rβ, PDB: 3LQM (Yoon et al., 2010). No changes in domain orientations were observed from either the IFN-λ1/IFN-λR1 binary (RMSD 0.7 Å) complex, the apo IL-10Rβ (RMSD 1.3 Å), or IFN-λ3 structures (RMSD 0.6 Å) (PDB: 3HHC) (Gad et al., 2009) (Figures S3A–S3C). IL-10Rβ makes extensive and contiguous contacts with the H11/IFN-λR1 binary complex, interacting with H11 through sites 2a and 2b and with the IFN-λR1 stem at site 3 (Figure 2A, left panel).

Figure 2. Structure of the IFN-λ Ternary Complex.

Figure 2

Open in a new tab

(A) The structure of the ternary complex reveals the mechanism of IL-10Rβ (gold) recognition of IFN-λ3 H11 (blue) and IFN-λR1 (gray). The IL-10Rβ makesextensive and continuous contacts with H11 at sites 2a and 2b, and site 3 makes stem-stem contacts with the IFN-λR1. Unique to the IFN-λ structure, a largesurface area of the cytokine remains surface exposed in the ternary complex.

(B) Structure of the IL-10/IL-10R1 partial signaling complex (PDB: 1J7V) (Josephson et al., 2001).

(C) Structure of a type I IFN receptor complex (PDB: 3SE4) (Thomas et al., 2011). See also Table S1.

The overall binding topology of the IFN-λ3/IFN-λR1/IL0Rβ ternary complex exhibits both similar and distinct features compared to previously determined cytokine receptor complexes (Spangleret al., 2015). The shared IL-10Rβ receptor binds H11 at the far end of the helical bundle rather than more centrally on the face of the helical bundle, as is typical of other cytokine/ receptor binding topologies (Spangler et al., 2015) (Figure 2A, middle and right panels). While the type I IFNs (PDB: 3SE4) (Thomas et al., 2011) and IFN-λs share common signaling pathways and immune activities, the ternary structure of the IFN-λ receptor complex exhibits a distinct overall geometry (Figure 2C). The two type I IFN receptors bind on opposing faces of the helical bundle of the cytokine, contrasting to the IFN-λ ternary complex. A further distinguishing feature of IFN-λ ternary complex is the role of receptor stem-stem contacts, which has not been seen in the type I IFN complexes. Another interesting comparison is with the structure of the IL-10/IL-10R1 binary complex (PDB: 1J7V) (Josephson et al., 2001) (Figure 2B), which is not a complete signaling complex because it lacks IL-10Rβ in the structure. IL-10 is the founding member of the IL-10 super-family and is a non-covalent dimer (Tan et al., 1993), with each IL-10Rβ bound to IL-10 to create a 2:2 stoichiometry, distinct from that of the 1:1:1 IFN-λ3 H11/IFN-λR1/IL-10Rβ ternary complex. It remains unclear where IL-10Rβ binds to IL-10 and how this can be geometrically accommodated in the higher order IL-10 family complexes. However, the binding geometry of IL-10Rβ to IFN-λ3 and IFN-λR1 could yield clues to IL-10Rβ binding modes in the other IL-10 family complexes.

In the complex structure, three IL-10Rβ loops, 2, 3, and 5, contain aromatic residues that undergo large conformational changes (2.4 – 6.5 Å) relative to the apo IL-10Rβ structure (PDB: 3LQM) (Yoon et al., 2010). Notably, residues in loop 5 of IL-10Rβ are situatedto share hydrogen bonds with both cytokine and co-receptor residues at sites 2b and 3 (Figure 2A). Tyr59 in loop 2 of IL-10Rβ binds IFN-λ3 H11 in apocket formedby helices C and D of the cytokine (Figures 3A-3B, and S2). Notably, the H11-specific mutation Glu73Asp (Helix C) forms a hydrogen bond with the hydroxyl group of Tyr59, stabilizing the interaction between the H11/IFN-λR1 binary complex and IL-10Rβ (Figures 3B, S2D). Tyr82 on loop 3 of IL-10Rβ, also binds to IFN-λ3 H11 at site 2a and sits in the pocket formed between the N terminus and helices A and D of the cytokine (Figures 3A-3B, S2E). In this position, Tyr82 shares two hydrogen bonds with the cytokine: one between the backbone carbonyl of Tyr82 and Nε of His91 (H11) and a second between the hydroxyl group of Tyr82 and the nitrogen backbone of Ser13 (H11) (Figures 3B, S2E).

Figure 3. Specific Interactions Mediating the Stability of the Ternary Complex.

Figure 3

Open in a new tab

For a Figure360 author presentation of Figure 3, see http://dx.doi.org/10.1016/j.immuni.2017.02.017#mmc3.

(A) Overview of the IFN-λ3 H11 (blue)/IFN-λR1 (gray)/IL-10Rβ (gold) ternary complex.

(B) Detailed views of the site 2a and site 2b contacts between IFN-λ3 H11 and IL-10Rβ tyrosines 59, 82, and 140. Hydrogen bonds are indicated by dashedblack lines.

(C) Shared receptor use of aromatic residues to bind cytokines.

(D) IL-10Rβ uses Tyr82 residue as a hotspot of binding like other shared receptor systems.

(E) Y103 of γc bound to IL-2 (PDB: 3QAZ) (Levin et al., 2012).

(F) F169 of gp130 bound to IL-6 (PDB: 1P9M) (Boulanger et al., 2003b). See also Figure S2.

At site 2b, Tyr140 and Trp143 of loop 5 (IL-10Rβ) bind IFN-λ3 H11 by “pinching” the N terminus of Helix A (Figures 3A and 3B). In the ternary complex structure, Tyr140 (IL-10Rβ) forms two hydrogen bonds with H11 residues Gln18 and Gln15Arg, the latter of which is an engineered mutation of IFN-λ3 (Figures 3B, S2F). Trp143 packs against the hydrophobic backbone of Helix A and contributes van der Waals interactions as well as a hydrogen bond shared with the backbone carbonyl of the N-ter-minal residue Ser11 (H11) (Figure 3B). Single alanine substitutions at sites 2a and 2b interface positions—Gln15, Gln88, or His91 in the wild-type background—result in a decrease in binding efficiency to IL-10Rβ (Figures S3D–S3F). These residues account for 9%, 5%, and 14% of the total buried surface area shared with IL-10Rβ and are important for binding, consistent with the observation that Gln88Ala IFN-λ3 was reported to have a 50-fold increase in the EC50 for the AV activity relative to the wild-type (Gad et al., 2009).

Given that IL-10Rβ is shared among several different cytokines, we compared the interface chemistry of IL-10Rβ cytokine recognition to that of two other shared cytokine receptors, γc and gp130 (Boulanger et al., 2003a; Wang et al., 2005, 2009; Yoon et al., 2010). All three shared receptors place aromatic residues in the central contact positions with the cytokines. When the cytokine-binding homology regions (CHR) are aligned, we find that Tyr82 of IL-10Rβ is analogous to Tyr103 of γc (PDB: 2B5I) (Wang et al., 2005) and Phe169 of gp130 (PDB: 1P9M) (Boulanger et al., 2003a), which dock onto hydrophobic pockets on their respective cytokines (Figure 3C). Each of the cytokines (IFN-λ, IL-2, and IL-6) varies in shape and size of the surface area. The total buried surface area of the IFN-λ3/IL-10Rβ interface is large (1640 Å2) with deep hydrophobic pockets on the face of IFN-λ which contrasts to the relatively flat and smaller interfaces of IL-6 (1272 Å2) (Boulanger et al., 2003a) and IL-2 (970 Å2) (Wang et al., 2005) (Figures 3D–3F). Another differentiating feature is that Tyr82 of IL-10Rβ shares two hydrogen bonds with IFN-λ that stabilize the ligand-receptor interaction, which is not observed with Tyr163 of γc or Phe169 of gp130. The centrality of aromatic residues that are capable of extensive chemical and structural plasticity in forming interface contacts with diverse cytokines appears to be a theme exploited by shared receptors to facilitate cross-reactivity (Boulanger et al., 2003b).

Site 3 represents the interface between the receptor stem domains D2 of IFN-λR1 and SD2 of IL-10Rβ. The site 3 interface extends from site 2b to the C-termini of the membrane-proximal receptor domains (Figure 2A). The total buried surface area is large (1417 Å2) and is composed of mainly van der Waals interactions with four hydrogen bonds interspersed. In comparison to the shared gp130 and γc shared receptor systems, the analogous gp130/IL-6Rα interface has a smaller total buried surface area (1074 Å2) (Boulanger et al., 2003a) while γc/IL-2Rβ (1750 Å2) has the larger stem-stem interface, containing 17 hydrogen bonds shared between the two co-receptors (Wang et al., 2005). At the IFN-λR1 and IL-10Rβ interface, hydrogen bonds are observed between the side chain of Arg130 (IL-10Rβ) and the backbone carbonyl Gln163 (IFN-λR1), and side chain to side chain hydrogen bonds are observed between His128 (IL-10Rβ) and Gln163 (IFN-λR1), Glu141 (IL-10Rβ) and Tyr189 (IFN-λR1), and Thr142 (IL-10Rβ) and Thr183 (IFN-λR1) (Figures S2B and S2C). As the collective binding and structural studies indicate, the site 3 receptor stem-stem contacts play an important role in the recruitment of IL-10Rβ to the IFN-λ3 H11/IFN-lR1 binary complex.

Implications for a Shared IL-10Rβ Binding Mode to IL-10 Family of Cytokines

Using the IFN-λ ternary complex as a structural template, we structurally aligned the IFN-λ ternary complex onto the binary structures of IL-22/IL-22R1 (PDB: 3DLQ) (Bleicher et al., 2008) and IL-10/IL-10R1 (PDB: 1J7V) (Josephson et al., 2001) in order to elucidate the molecular basis of IL-10Rβ binding and recognition (Figure 4A, left and middle). The structure of the IFN-λ3 H11/ IFN-λR1/IL-10Rβ ternary complex revealed that IL-10Rβ makes contacts with three cytokine helices: A, C, and D. Mutations associated with impaired binding of multiple IL-10 family members to the shared IL-10Rβ (Gad et al., 2009; Logsdon et al., 2004; Wu et al., 2008; Yoon et al., 2010) fell within the IFN-λ3 H11 binding region of the receptor and contact hotspot tyrosine residues of IL-10Rβ (Figure 4A, right, red surface). Structure-based sequence alignment of the three cytokines indicated that residues mapping to the IL-10Rβ binding site are highly divergent (Figures 4B, S3G). Despite a lack of sequence conservation, structural comparison of the IFN-λ3, IL-10, and IL-22 residues within the site 2 interface revealed that the three cytokines have a conserved pattern of hydrophobic patches at the sites of the IL-10Rβ tyrosine docking (Figure 4C), which are surrounded by polar residues. Thus, IL-10 family members appear to have evolved chemical complementarity with IL-10Rβ through distinct pairwise interactions.

Figure 4. Structural and Chemical Conservation of IL-10Rβ Binding in the IL-10 Superfamily.

Figure 4

Open in a new tab

(A) Structures of IL-10 (green) (PDB: 1J7V) (Josephson et al., 2001) and IL-22 (orange) (PDB: 3DLQ) (Bleicher et al., 2008) were structurally aligned to IFN-λ3 H11(blue ribbon) in the ternary complex structure, highlighting conserved features of IL-10Rβ (gold surface) recognition. IL-10Rβ surface residues within contactdistance of H11 are shaded red. Residues at which alanine mutations have been shown to negatively impact IL-10Rβ binding to IL-10 (green), IL-22 (orange), andIFN-λ3 (blue) are indicated as sticks (right panel).

(B) View of the IL-10Rβ binding interface of IFN-λ3 H11 mapped onto the structures of other IL-10 superfamily members. Residues that impact IL-10Rβ bindingidentified through mutagenesis analysis and residues that share hydrogen bonds with IL-10Rβ in the IFN-λ ternary complex structure are shown as sticks.

(C) View ofthe IL-10Rβ binding interface modeled on the face ofIL-10 and IL-22 using the IFN-λ3 H11 ternary complex. Cytokine residues at the IL-10Rβ interfaceare colored by chemical properties (red for polar/charged or white for hydrophobic) to highlight the conserved “hydrophobic” pockets in which the IL-10Rβtyrosines likely dock. See also Figure S3.

Functional Activity of High-Affinity IFN-λ3

To determine whether stabilization of the IFN-λ ternary complex enhanced type III IFN signaling and function, we measured in vitro phospho-STAT1 signaling, induction of interferon-stimulated genes (ISG), and antiviral and antiproliferative activities of H11 relative to wild-type IFN-λ3 and a type I IFN (IFN-ω). H11 improved the EC50 for phospho-STAT1 on type I and III IFN-responsive Hap1 cells by 100-fold relative to wild-type IFN-λ3. Cells modified to overexpress IFN-λR1 exhibit similar Emax of phospho-STAT signaling (François-Newton et al., 2011). Here, we found that the Emax values of both type III IFNs were only about 30% that of IFN-ω (Figure 5A), consistent with previous studies (Dickensheets et al., 2013; Kotenko et al., 2003) in which type I IFN signaling is stronger in unmodified cells. Similarly, H11 induced ISGs more potently than wild-type IFN-λ3 in Hap1 cells, although gene induction remained well below levels induced by IFN-ω (Figures 5B and S4A). H11 improved antiviral activity in Huh7.5 cells infected with HCV, with potency 12-fold higher than wild-type IFN-λ3, though 10-fold less potent than IFN-ω (Figure 5C).

Figure 5. Functional Characterization of an Engineered IFN-λ3 Variant.

Figure 5

Open in a new tab

IFN-λ3 H11 (orange) was compared to the wild-type IFN-λ3 (black) and the type I IFN-ω (red) in several functional assays.

(A) STAT1 activation in Hap1 cells. Curves were fit to a first-order logistic model. Error bars represent ± SEM (n = 3).

(B) Induction of ISG15 in Hap1 cells treated with 5 pM with each interferon for 6 hr as determined byqPCR. Error bars represent 95% confidence intervals (n = 3).

(C) Antiviral activity of IFNs in Huh7.5 cells infected with HCV. Error bars represent ± SEM (n = 4).

(D) Antiproliferative activity of IFNs in Huh7.5 cells.

(E) Antiproliferative activity of IFNs is enhanced in Huh7.5 cells overexpressing IFN-λR1. Error bars represent ± SD (n = 6) (D and E).

(F) Human-liver chimeric mice were generated by injecting human hepatoblasts into Fah−/−Rag2−/−Il2rgnull (FRG) mice and infected with 2 × 108 DNA copies ofvirus obtained from a genotype C eAg negative patient. Human albumin was tracked by ELISA (red circles) and HBV DNA quantified by qPCR (green circles) over220 days prior to IFN treatment.

(G) Mice were treated daily with 10 μg/kg of body weight for 4 weeks with vehicle, IFN-λ3, or IFN-λ3H11. The engineered IFN-λ3H11 has improved in vivo activityover the wild-type IFN-λ3 (p = 0.02). Error bars represent ± SEM, where n = 2 for the Control and n = 3 for WT or H11 treated mice. See also Figure S4.

Both wild-type IFN-λ3 and H11 elicited minimal antiproliferative activity (Figure 5D), which has been previously observed in vitro (Lasfar et al., 2006). We hypothesized that the lack of antiproliferative activity could be due to the limited expression of IFN-λR1 on Huh7.5 cells (Moraga et al., 2009; Zhou et al., 2007). Indeed, when IFN-λR1 was transduced into Huh7.5 or WISH cells, an IFN-λ non-responsive cell line, the IFN-λs robustly induced an antiproliferative response (Figure 5E). As anticipated, the effect of the high-affinity H11 was stronger than that of the wild-type IFN-λ3 and the magnitude of this difference depended on relative levels of IFN-λR1 expression (Figure S4B). Taken together, these experiments suggest IFN-λ antiproliferative activity might be limited by both IFN-λR1 receptor expression and stability of the IFN-λ signaling complex; the latter can be addressed through affinity maturation (Figures 5D and 5E, and S4).

To test whether the improved in vitro AV potency of H11 over wild-type IFN-λ3 results in enhanced in vivo antiviral efficacy, we tested our engineered interferon in a mouse model of HBV infection. We made use of the fumarylacetoacetate hydrolase deficient (Fah−/−) liver chimeric model developed by Grompe and colleagues, in which liver injury is controlled by administration of 2-(2-nitro-4-trifluoro-methyl-benzoyl)-1, 3 cyclohexanedione (NTBC) (Grompe et al., 1995). After transplantation of human fetal hepatoblasts into immunodeficient FRG mice (Fah−/−/Rag2−/−/Il2rg−/−) (Billerbeck et al., 2016), we induced mouse liver damage by intermittent withdrawal of the protective drug NTBC and monitored engraftment levels over time by measuring human albumin (hAlb) levels in the sera (Figure 5F) (Bissig et al., 2010; de Jong et al., 2014). Mice were challenged with HBV upon peaking hAlb levels. Once HBV viremia had reached its plateau, the infected mice were subjected to IFN-λ3-based antiviral treatment. IFN-λ3 H11 suppressed HBV viral load more effectively than the IFN-λ3 wild-type regimen (Figure 5G). Similarly, although the difference was minor, HBV surface antigen (sAg) levels were lower for the H11 relative to the wild-type treated mice over the course of treatment (p = 0.0083) (Figure S4C). Human albumin levels, which serve as a proxy for toxicity, remained stable during the course of treatment, excluding the possibility of human hepato-cyte loss and suggesting the higher-affinity IFN-λ was no more toxic than the wild-type. (Figure S4D). Thus, although modest, H11 demonstrates improved therapeutic efficacy without obvious toxicity. These combined data suggest that further affinity improvements, now using the crystal structure of the ternary complex as a guide, might yield a therapeutically improved IFN-λ for both antiviral and anti-cancer therapy.

Probing the Plasticity of Type I Interferon Function through Structure-Guided Protein Engineering

Our engineering efforts improved the antiviral and antiproliferative potency of IFN-λ3 but its potency still lagged behind type I IFNs. Thus, we aimed to characterize the plasticity and potency of the type I IFNs through engineering. This information could potentially be informative for understanding the basis of the potency differences between the two IFN families, but also to establish a benchmark of activity for future IFN-λ engineering efforts. In the type I IFN system, several prior studies using gene shuffling show that affinity improvement also manifests as improved antiviral potency (Brideau-Andersen et al., 2007; Chang et al., 1999). However, a more comprehensive series of type I IFN engineering studies conclude that wild-type IFNs are already at an AV potency maximum, and there is little to gain from affinity maturation (Jaitin et al., 2006). To resolve this discrepancy, we probed the sensitivity of type I IFN antiviral and antiproliferative functions to receptor affinity using a HTP combinatorial engineering approach.

We were guided by the structureof the typeIinterferon, IFN-ω, in complex with the IFN-αR1 and IFN-αR2 receptors (Figure 6A). We generated a site-directed mutagenic library of IFN-ω that diversified its IFN-αR1 binding interface, the lower-affinity site, as a means of creating variants with modified antiviral and antiproliferative activities (Figure 6A). We enriched the library for binders to wild-type IFN-αR1 affinity (KD = 1 μM). In order to find the most interesting variants with diverse activities, we developed a high-throughput functional screen, and characterized 288 randomly selected clones from the library (Figure 6B). For our screen, distinct type I IFN variant proteins were cleaved from yeast cells, and the released cytokines in the supernatant were purified by filtration. The IFN-containing supernatant was then used to treat cells (Figure 6B). Based on their diverse antiproliferative and antiviral activities, we selected four IFN-ω variants for recombinant expression and characterization. Previously, we rationally designed an IFN-ω variant (Lys152Arg), with 100-fold higher affinity for the IFN-αR2 receptor than the wild-type cytokine (Thomas et al., 2011). We thus added the Lys152Arg mutation to each of the four variants from the HTP screen. In addition to these mutants, we biophysically and biochemically characterized the wild-type IFN-ω, IFN-ω (Lys152Arg), and a shuffled IFN-α mutant (Maxygen 9×25) that was previously reported to be biased toward antiviral activity (Brideau-Andersen et al., 2007).

Figure 6. Probing the Molecular Mechanism of Type I IFN Functional Differences from Type III IFN via Cytokine Engineering and High-Throughput Functional Screening.

Figure 6

Open in a new tab

(A) A site-directed IFN-ω library designed to mutagenize the low affinity IFN-αR1 site was combined with a rationally designed affinity-enhancing Lys152Argmutation at the IFN-αR2 interface on the cytokine. The SD4 domain of IFN-αR1 was modeled to the complex (PDB: 3SE4) (Thomas et al., 2011) using the mouseIFN-αR1 structure (PDB: 3WCY).

(B) 288 clones were screened for activity usingahigh-throughput functional screen. Displayed IFNs were cleaved from yeast by the site-specific3C protease. Thefiltered supernatants were then used to treat VPN53 (Moraga et al., 2009) cells to measure antiproliferative activity and HCV-infected Huh-7.5 cells to measureantiviral activity.

(C) Functional characterization of phospo-STAT1 signaling, antiviral and antiproliferative potencies as a function of complex stability. Potencies of IFN-mediatedactivation of STAT1 on Jurkat cells, viral clearance in HCV-infected Huh7.5 cells, and proliferation inhibition of WISH cells were used to determine EC50s andplotted as a function of complex stability.

(D) Fold ISG induction as measured by qPCR relative to complex stability. See also Table S2.

We measured signaling potency (phospho-STAT1), gene induction, and antiviral and antiproliferative activities of our panel of engineered interferons relative to wild-type IFN-ω (Figure 6C). We also determined the affinity of each cytokine for the IFN-αR1 and IFN-αR2 subunits via surface plasmon resonance (Table S2). Receptor affinity was not used as a criterion for selection of the variants, rather, variants were chosen based on differential antiviral and antiproliferative activity, we aimed to interrogate the correlation between complex stability (Kalie et al., 2008; Lavoie et al., 2011; Piehler et al., 2012; Thomas et al., 2011) (the product of the cytokine KDs for binding each receptor subunit) and cytokine activity. Affinities for IFN-αR1 ranged from 20 nM to 300 μM (wild-type IFN-ω KD = 1.2 mM). Affinities for IFN-αR2 were between 2 nM for the unmodified IFN-ω and 200 pM in the Lys152Arg background. In all, our IFN variant complexes sampled a 5.7-log range of complex stabilities (Table S2). The EC50s of our variants for STAT signaling and antiviral activity were found to be insensitive to changes in affinity, as illustrated by both minimal improvement over the wild-type EC50s and narrow range of EC50s (0.85 and 1.2 logs, respectively) for the broad range of complex stabilities (Figure 6C, left and middle panels, Table S2). In contrast, antiproliferative activity varied by 3.5 logs and correlated (R = 0.67) with complex stability (Figure 6C, right). We measured induction of a representative set of ISGs by seven of our IFN variants on Huh7.5 cells (Figure 6D). As observed with the differential effects of complex stability on antiviral and antiproliferative activities, high-affinity variants offered little improvements in gene induction (such as ISG15 and MX1) over the wild-type (Figure 6D, two left panels) while other genes (like APOL3 and SAM9DL) were more sensitive to complex stability (Figure 6D, two right panels) (Levin et al., 2014). This observation is in accord with several previous studies (Jaitin et al., 2006; Kalie et al., 2008; Lavoie et al., 2011; Thomas et al., 2011). Given the apparent lack of off target toxicities of IFN-λs compared to type I IFNs (Muir et al., 2014; Sato et al., 2006), these data provide conceptual support for engineering an IFN-λ to match the AV potency of type I IFNs, but it remains to be determined whether the diminished side effects of IFN-λ will persist in forms with potency equal to that of type I IFNs.

Discussion

In this study, we implemented a protein engineering strategy to gain structural access to the IFN-λ/IFN-λR1/IL-10Rβ ternary complex, and to assess the relative potential for enhancing the functional efficacies of type III and type I IFNs. The ternary complex illuminated several important molecular features governing the IL-10Rβ interactions with IL-10 family cytokines, which are of extremely low affinity. Previously, the divergent sequences within the IL-10 superfamily cytokines placed substantial limitations on our ability to engineer IL-10 cytokine variants with enhanced signaling and functional activities. With the structure in hand, we observed how IL-10 superfamily cytokines present three hydrophobic patches on their surface that interlock with aromatic residues on inter-domain loops of IL-10Rβ. Additionally, the structure revealed the importance of extensive IFN-lR1/IL-10Rβ stem-stem interactions in stabilizing the ternary complex, clearly showing that a composite interface presented by each of the IL-10 family cytokines with their respective high-affinity receptors is required for IL-10Rβ recognition.

IFN-λ3 exhibits 60-fold more potency, as measured by antiviral EC50s, than IFN-λ2, (Dellgren et al., 2009) suggesting that type III IFN antiviral activity could be enhanced. We anticipated that this effect could be further exaggerated through molecular engineering approaches given that type III IFNs bind IL-10Rβ with nearly undetectable affinities in vitro. Using directed evolution, we identified variants with up to 150-fold increased complex stability, which resulted in a 100-fold improvement in the EC50 for pSTAT1 signaling and 12-fold improvement in the EC50 for antiviral activity. Studies of HBV-infected human-liver chimeric mice validated this improved potency in an in vivo setting by demonstrating that an engineered IFN-λ can improve upon the efficacy of the wild-type cytokine; a quality recent clinical trials have indicated as being clinically desirable (Chan et al., 2016). A more potent IFN-λ which maintains a low toxicity profile compared to type I IFNs, would provide new avenues for treatment of chronic infections such as HBV, HCV, and HBV co-infections with HDV, and mightbe useful asa broad spectrum antiviral (Ciancio and Rizzetto, 2014).

In vitro, IFN-λs have a blunted antiproliferative activity (Lasfar et al., 2006), which effects could be rescued by overexpression of the IFN-λR1 receptor. Indeed, the engineered IFN-λ was more potent than the wild-type. While the minimal antiproliferative effects on untransduced cell lines might appear to suggest that IFN-λs have limited potential as anti-cancer agents, in vivo studies paint a different picture. IFN-λs can significantly inhibit tumor growth in murine cancer models (Lasfar et al., 2006; Numasaki et al., 2007; Sato et al., 2006) with less toxicity than type I IFN treatment (Sato et al., 2006). This gives hope for continued cancer-related efforts.

We also developed a panel of type I IFN proteins with a range of affinities to probe the molecular mechanism underlying the observed differences in antiviral and antiproliferative potencies of type I versus type III IFNs. In vitro antiviral studies have established the type III IFNs to be less potent than the type I IFNs, which translates into a lower efficacy in the clinic, as observed in recent HCV (Muir et al., 2014) and HBV (Hruska et al., 2015; Wang et al., 2015) clinical trials. Previous studies of the type I IFNs have demonstrated a modest 2-fold difference in antiviral potency between the highest affinity natural IFN (IFN-β) or engineered IFN such as IFN-α2 YNS (Kalie et al., 2007) relative to lower affinity ligands such as IFN-α2. However, one study reported engineering type I IFNs with >20-fold improved antiviral activities but relatively unchanged antiproliferative activities (Brideau-Andersen et al., 2007). This study measured antiproliferative activity using the Daudi cell line, which is more sensitive to antiproliferative effects, possibly leading to aberrant antiproli-ferative:antiviral ratios. Here, using structure-based engineering of type I IFNs, we demonstrate that STAT signaling potency and antiviral activity were insensitive to complex stability and stable relative to wild-type levels (Kalie et al., 2007, 2008; Lavoie et al., 2011; Thomas et al., 2011). By contrast, antiproliferative activity was directly correlated with complex stability (Levin et al., 2014). This observation held true for the synthetic IFN, 9x25, which was previously reported to have a 20-fold improvement in antiviral activity over a wild-type IFN (Brideau-Andersen et al., 2007). In this study, the antiviral activity of the shuffled IFN (EC50 = 140 pM) was similar to IFN-ω (EC50 = 100 pM) and had a 48-fold improvement in the antiproliferative activity over wild-type IFN-ω, consistent with a 3-log improvement in complex stability. Collectively, these experiments demonstrate that type I IFNs vary in their antiproliferative:antiviral profiles mainly due to differences in antiproliferative potencies between ligands as a result of varied receptor affinities.

Although we were able to engineer a type III IFN with improved signaling and antiviral activities, our IFN-λ3 H11 variant remained less potent than type I IFNs. The type III IFN ternary complex will guide efforts to engineer even more potent cytokines, which can potentially close the gap in efficacy between the type I IFNs and type III IFNs. Moreover, the true respective clinical potential of these different interferons will also need to take into account their respective side effects. Because the latter are highly dependent on the receptor distributions, engineered IFN-λ3 variants such as H11 might have unique advantages in the clinic.

Experimental Procedures

Yeast Surface Display of IFN-ω and IFN-λ3

For both type I and III IFNs, the cytokines were displayed on yeast as previously described (Levin et al., 2012) but containing a 3C rhinovirus protease tag at the N terminus. Staining and selection was performed via streptavidin-phycoerythrin labeled receptors with separation of the receptor-yeast population by paramagnetic anti-phycoerythrin microbeads (Miltenyi; MACS). Expression on the yeast surface was assayed by staining with the Myc-tag antibody conjugated to Alexa 647 (Cell Signaling). Progression of the enrichment was monitored by the receptor yeast staining by flow cytometry (BD Accuri). A site-directed library was used for engineering type I IFNs. A round of error-prone PCR and DNA shuffling (Brideau-Andersen et al., 2007) was used for the second-generation of IFN-λ3 variants. Single-point variants of IFN-λ3 were generated by Quick-Change Mutagenesis, transformed, and displayed on yeast.

Functional Screen for IFN-ω

Individual clones from the library were plated on SD-CAA plates, grown in a 96-deep well format. The protein for each clone was cleaved from the yeast using rhinovirus 3C protease, separated from the yeast by filtration (Whatman Unifilter 800), and assayed for antiviral activity by tracking HCV replication in Gaussia luciferase-based reporter Huh7.5 cells and antiproliferative activity in PVN53 (Moraga et al., 2009) cells by measuring cell density as previously described (Thomas et al., 2011). Variants with diverse activities were selected for recombinant expression and characterization.

Protein Expression, Purification, and Structural Determination

Type IIFN cytokines, type I IFN receptors, type III IFNs, and type III receptors were expressed in the Hi5 insect expression system, purified as previously described (Thomas et al., 2011) and stored in 10% glycerol. For crystallography, all four glycosylation sites in IL-10Rβ were mutated from Asn to Gln as described in Yoon et al. (2010). IFN-λR1 was expressed in HEK293 GnTi− cells and de-glycosylated by treatment with EndoF and EndoH. The methylated (Walter et al., 2006) IFN-λ3 H11/IFN-λR1/IL-10Rβ complex was purified by size exclusion chromatography on a S75 column (GE). Crystals were obtained within 24 hr at 20°C from the MCSG3 screen (Anatrace) and optimized to 0.2 M Ca acetate, 0.1 M imidazole pH 8.0, 10% PEG8000, and 3% sucrose as an additive (Catalog #HR2-138, Hampton). Cryoprotectant was the mother liquor plus 8% each of sucrose, glucose, and xylitol.

Crystallographic data were collected at the Advanced Light Source (ALS) Beamline 8.2.1. Data were indexed, integrated, and scaled using XDS or HKL2000 program suites. Crystal structures were solved by molecular replacement with the program PHASER using the IFN-λ1/IFN-λR1 binary complex (PDB: 3OG6) (Miknis et al., 2010) and apo IL-10Rβ (PDB: 3LQM) (Yoon et al., 2010) structures as search models. The final structure was built by iterative cycles of reciprocal space refinement with PHENIX and BUSTER and manual rebuilding with COOT. Crystallographic software used in this project was installed and configured by SBGrid.

Sequence, Structural, and FACS Analysis

Promals3D was used to perform a structure based sequence alignment (Pei et al., 2008). Sequence alignments and percent identity calculations were performed with JalView (The University of Dundee). Structural alignments, homology models, and distance calculations, and figures were generated in Pymol (Schrodinger, LLC). Buried surface area was calculated using PISA (Krissinel and Henrick, 2007). Figures of FACS data were generated with either R with the Bioconductor source (R-project) or Prism (GraphPad Software).

Surface Plasmon Resonance

GE Biacore T100 was used to measure the KD by either kinetic (type I IFNs) or equilibrium (type III IFNs) methods. Approximately 100 RU of each of the receptors were captured on a Biotin CAP-chip (GE) including a reference channel of an unrelated cytokine receptor (IL-2RP).

In Vitro Characterization

Fortype I IFNs, signaling, antiviral and antiproliferative assays were performed as previously described (Thomas et al., 2011). For measuring gene induction, Huh7.5 cells were plated in a 6-well format and treated with 1 nMtypeI IFNs for 24 hr, Hap1 cells (a gift from Jan Carette) were treated with 5 pM type I or III IFNs for 6 hr, RNA was extracted with the RNeasy Micro kit (QIAGEN), which was converted to cDNA by a RT-PCR reaction (HC RNA-to-cDNA kit, Thermo Fisher Scientific), and ISG induction relative to the untreated controls and normalized to 18S levels were measured by the Taqman qPCR assay system on a StepOnePlus instrument (Thermo Fisher Scientific) following manufacture instructions. Fortype III IFNs, pSTAT1 signaling was performed as previously described (Thomas et al., 2011) except Hap1 cells were detached after IFN treatment by incubating with trypsin (GIBCO) for 5 min before fixing and staining.

Antiproliferative Activity of Type III IFNs

Lentiviral pseudoparticles were generated by co-transfecting 4 × 105 Lenti-X 293T cells (Clontech) in poly-L-lysine coated 6-well plates with plasmids expressing the pLX304 proviral DNA encoding human IFN-λR1, HIV-1 gag-pol, and VSV-G at a ratio of 1.1/0.7/0.2, respectively. For each transfection, 5 ml Lipofectamine 2000 (Thermo Fisher Scientific) was combined with 2.0 mg total DNA in 100 μl Opti-MEM (GIBCO). Transfections were carried out for 6 hr, followed by a medium change to DMEM containing 3% FBS. Supernatants were collected at 24 hr and 48 hr, pooled, passed through a 0.45 mm filter and stored at −80°C. 3 × 105 Huh7.5 and WISH cells were resuspended in DMEM containing 10% FBS, 500 mL lentivirus, and 8 mg/ml polybrene in a total volume of 1.5 mL and spinocualted in 12-well plates for 1.5 hr at 850 × g.48 hr post transduction the cells were selected with 2.5 mg/ml blasticidin. Huh7.5 and WISH cells were harvested in PBS + 5 mM EDTA and washed twice with cold PBS + 0.5% BSA. Cells were then incubated with anti-IL-28RA (R&D Systems; cat# AF5260) at 5 μg antibody per 1 × 106 cells in 0.4 mL volume for 30 min, washed 3 times, incubated with FITC conjugated anti-sheep IgG (Abcam; cat# ab6743) at 1:2000 dilution for 30 min, and washed 3 times prior to cell sorting into low, medium, and high IFN-λR1-expressing populations using a BD FACSAria flow cytometer. HepG2, Huh7.5 and WISH cells were seeded at 1 × 103 cells/well in 96-well plates. The following day media was replaced with 100 ml/well of IFN-containing media. 4 days post IFN treatment cell density was measured using CellTiter-Glo (Promega) according to the manufacturer's protocol.

Generation of HBV-Infected Human-Liver Chimeric Mice and IFN-λ3 Therapy

Human hepatoblasts were isolated from human fetal livers procured from Advanced Bioscience Resources (ABR), as described (Andrus et al., 2011)., and injected intrasplenically (1 × 106 cells per mouse) into male Fah−/− Rag2−/−Il2rgnull (FRG) mice originally obtained from Marcus Grompe (Grompe et al., 1995). Starting on the day of transplantation, mice were cycled off the liver protective drug NTBC (Yecuris) as described by others (Azuma et al., 2007; Bissig etal., 2010). Human albumin levels in mouse sera were measured by ELISA(Bethyl Laboratories). Human liver chimeric mice were injected intravenously with 100 μL of mouse serum containing 2 × 108 DNA copies/ml of HBV that originated from a genotype C eAg negative patient. For HBV viral load measurements, DNA from 25 mL of mouse serum was isolated using a DNA extraction kit (QIAamp DNA Blood Mini, QIAGEN) and copy number was analyzed by an in-house Taqman assay as described previously (Shlomai et al., 2014). HBsAg (Autobio Diagnostics) levels in mouse serum were determined by CLIA per manufacturers' instructions. Eight mice were randomized for daily treatment for four weeks with intraperitoneal injections of vehicle (15% glycerol in PBS), wild-type IFN-λ3 or H11 at 10 μg/kg body weight (2, 3, and 3 mice respectively). Every time point, data were plotted normalized to baseline values and relative to controls. Statistical analysis was performed using the regular two-way ANOVA with Bonferroni multiple comparison post-test in GraphPad Prism.

Supplementary Material

supplement

Highlights.

  • The IFN-λ ternary complex provides insight into the mechanism of IL-10Rβ engagement

  • High-affinity IFN-λ3 elicits greater antiproliferative and antiviral activities

  • In vivo, an engineered IFN-λ3 has enhanced antiviral activity over the wild-type

  • In contrast to IFN-λs, high-affinity type I IFNs do not improve antiviral activity

Acknowledgments

The authors would like to thank Menashe Elazar for work on preliminary in vivo studies, Wu Di for help with protein expression, and Deepa Waghray and Susan Fischer for technical assistance. This work was supported by NIH grants 1U19AI109662 (to K.C.G. and J.G.) NIH RO1-AI51321 (to K.C.G.), AI091707 (to C.M.R.). K.C.G. is an investigator of the Howard Hughes Medical Institute. J.L.M. is supported by NIH award K01CA175127, W.M.S. by DK095666, and Y.P.J. by K08DK090576 and R01HL131093. Additional funding was provided by the Greenberg Medical Research Institute, the Starr Foundation, and several generous donors (to C.M.R.). The Advanced Light Source is supported by the Director, Office of Science, Office of Basic Energy Sciences, of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231. Structure factors and coordinates have been deposited in the Protein Data Bank with identification number PDB: 5T5W. Diffraction images have been deposited in the SBGrid Data Bank (http://dx.doi.org/10.15785/SBGRID/416).

Footnotes

Accession Numbers: Structure factors and coordinates have been deposited in the Protein Data Bank with identification number PDB: 5T5W. Diffraction images have been deposited in the SBGrid Data Bank (http://dx.doi.org/10.15785/SBGRID/416).

Supplemental Information: Supplemental Information includes four figures, two tables, and Supplemental Experimental Procedures and can be found with this article online at http://dx. doi.org/10.1016/j.immuni.2017.02.017.

Author Contributions: J.L.M. and K.C.G. conceived the project and wrote the manuscript. J.L.M. engineered, performed the functional screen with help from I.M., and characterized type I IFNs by AP; performed ISG measurements, signaling, and SPR for both type I and III IFNs. J.L.M. and T.M.H. engineered type III IFNs. J.L.M. crystallized the IFN-λ complex and together with K.M.J. refined the structure. A.X. measured in vitro antiviral activity of type I and III IFNs. H.H.H. and W.M.S. measured antiproliferative activity of type III IFNs including transducing cell lines with IFN-λR1 and sorting based on receptor expression levels. H.H.H., W.M.S., K.V., Y.P.J., J.L.M., C.M.R., and K.C.G. planned and K.V. and Y.P.J. performed the HBV in vivo studies. J.S.G., C.M.R., and K.C.G. supervised the research.

References

  1. Andrus L, Marukian S, Jones CT, Catanese MT, Sheahan TP, Schoggins JW, Barry WT, Dustin LB, Trehan K, Ploss A, et al. Expression of paramyxovirus V proteins promotes replication and spread of hepatitis C virus in cultures of primary human fetal liver cells. Hepatology. 2011;54:1901–1912. doi: 10.1002/hep.24557. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Ank N, West H, Bartholdy C, Eriksson K, Thomsen AR, Paludan SR. Lambda interferon (IFN-λambda), a type III IFN, is induced by viruses and IFNs and displays potent antiviral activity against select virus infections in vivo. J Virol. 2006;80:4501–4509. doi: 10.1128/JVI.80.9.4501-4509.2006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Azuma H, Paulk N, Ranade A, Dorrell C, Al-Dhalimy M, Ellis E, Strom S, Kay MA, Finegold M, Grompe M. Robust expansion of human hepatocytes in Fah-/-/Rag2-/-/Il2rg-/- mice. Nat Biotechnol. 2007;25:903–910. doi: 10.1038/nbt1326. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Baldridge MT, Nice TJ, McCune BT, Yokoyama CC, Kambal A, Wheadon M, Diamond MS, Ivanova Y, Artyomov M, Virgin HW. Commensal microbes and interferon-λ determine persistence of enteric murine norovirus infection. Science. 2015;347:266–269. doi: 10.1126/science.1258025. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Billerbeck E, Mommersteeg MC, Shlomai A, Xiao JW, Andrus L, Bhatta A, Vercauteren K, Michailidis E, Dorner M, Krishnan A, et al. Humanized mice efficiently engrafted with fetal hepatoblasts and syn-geneic immune cells develop human monocytes and NK cells. J Hepatol. 2016;65:334–343. doi: 10.1016/j.jhep.2016.04.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bissig KD, Wieland SF, Tran P, Isogawa M, Le TT, Chisari FV, Verma IM. Human liver chimeric mice provide a model for hepatitis B and C virus infection and treatment. J Clin Invest. 2010;120:924–930. doi: 10.1172/JCI40094. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Bleicher L, de Moura PR, Watanabe L, Colau D, Dumoutier L, Renauld JC, Polikarpov I. Crystal structure of the IL-22/IL-22R1 complex and its implications for the IL-22 signaling mechanism. FEBS Lett. 2008;582:2985–2992. doi: 10.1016/j.febslet.2008.07.046. [DOI] [PubMed] [Google Scholar]
  8. Boulanger MJ, Chow DC, Brevnova EE, Garcia KC. Hexameric structure and assembly of the interleukin-6/IL-6 alpha-receptor/ gp130 complex. Science. 2003a;300:2101–2104. doi: 10.1126/science.1083901. [DOI] [PubMed] [Google Scholar]
  9. Boulanger MJ, Bankovich AJ, Kortemme T, Baker D, Garcia KC. Convergent mechanisms for recognition of divergent cytokines by the shared signaling receptor gp130. Mol Cell. 2003b;12:577–589. doi: 10.1016/s1097-2765(03)00365-4. [DOI] [PubMed] [Google Scholar]
  10. Brideau-Andersen AD, Huang X, Sun SCC, Chen TT, Stark D, Sas IJ, Zadik L, Dawes GN, Guptill DR, McCord R, et al. Directed evolution of gene-shuffled IFN-αlpha molecules with activity profiles tailored for treatment of chronic viral diseases. Proc Natl Acad Sci USA. 2007;104:8269–8274. doi: 10.1073/pnas.0609001104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Chan HLY, Ahn SH, Chang TT, Peng CY, Wong D, Coffin CS, Lim SG, Chen PJ, Janssen HLA, Marcellin P, et al. Peginterferon lambda for the treatment of HBeAg-positive chronic hepatitis B: A randomized phase 2b study (LIRA-B) J Hepatol. 2016;64:1011–1019. doi: 10.1016/j.jhep.2015.12.018. [DOI] [PubMed] [Google Scholar]
  12. Chang CC, Chen TT, Cox BW, Dawes GN, Stemmer WP, Punnonen J, Patten PA. Evolution of a cytokine using DNA family shuffling. Nat Biotechnol. 1999;17:793–797. doi: 10.1038/11737. [DOI] [PubMed] [Google Scholar]
  13. Ciancio A, Rizzetto M. Chronic hepatitis D at a standstill: where do we go from here? Nat Rev Gastroenterol Hepatol. 2014;11:68–71. doi: 10.1038/nrgastro.2013.164. [DOI] [PubMed] [Google Scholar]
  14. de Jong YP, Dorner M, Mommersteeg MC, Xiao JW, Balazs AB, Robbins JB, Winer BY, Gerges S, Vega K, Labitt RN, et al. Broadly neutralizing antibodies abrogate established hepatitis C virus infection. Sci Transl Med. 2014;6:254ra129. doi: 10.1126/scitranslmed.3009512. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Dellgren C, Gad HH, Hamming OJ, Melchjorsen J, Hartmann R. Human interferon-lambda3 is a potent member of the type III interferon family. Genes Immun. 2009;10:125–131. doi: 10.1038/gene.2008.87. [DOI] [PubMed] [Google Scholar]
  16. Dickensheets H, Sheikh F, Park O, Gao B, Donnelly RP. Interferon-lambda (IFN-λ) induces signal transduction and gene expression in human hepatocytes, but not in lymphocytes or monocytes. J Leukoc Biol. 2013;93:377–385. doi: 10.1189/jlb.0812395. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Donnelly RP, Kotenko SV. Interferon-lambda: a new addition to an old family. J Interferon Cytokine Res. 2010;30:555–564. doi: 10.1089/jir.2010.0078. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. François-Newton V, Magno de Freitas Almeida G, Payelle-Brogard B, Monneron D, Pichard-Garcia L, Piehler J, Pellegrini S, Uzé G. USP18-based negative feedback control is induced by type I and type III interferons and specifically inactivates interferon a response. PLoS ONE. 2011;6:e22200. doi: 10.1371/journal.pone.0022200. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Gad HH, Dellgren C, Hamming OJ, Vends S, Paludan SR, Hartmann R. Interferon-lambda is functionally an interferon but structurally related to the interleukin-10 family. J Biol Chem. 2009;284:20869–20875. doi: 10.1074/jbc.M109.002923. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Grompe M, Lindstedt S, al-Dhalimy M, Kennaway NG, Papaconstantinou J, Torres-Ramos CA, Ou CN, Finegold M. Pharmacological correctionof neonatal lethal hepatic dysfunction in a murine modelofhereditary tyrosinaemia type I. Nat Genet. 1995;10:453–460. doi: 10.1038/ng0895-453. [DOI] [PubMed] [Google Scholar]
  21. Hruska M, Wang X, Chan P, Ahmad A, Freeman J, Horga MA, Hillson J, Kansra V, Lopez-Talavera JC. Derivation of Phase 3 dosing for peginterferon lambda-1a in chronic hepatitis C, Part 2: Exposure-response analyses for efficacy and safety variables. J Clin Pharmacol. 2015;55:73–80. doi: 10.1002/jcph.361. [DOI] [PubMed] [Google Scholar]
  22. Jaitin DA, Roisman LC, Jaks E, Gavutis M, Piehler J, Van der Heyden J, Uze G, Schreiber G. Inquiring into the differential action of interferons (IFNs): an IFN-αlpha2 mutant with enhanced affinity to IFNAR1 is functionally similar to IFN-beta. Mol Cell Biol. 2006;26:1888–1897. doi: 10.1128/MCB.26.5.1888-1897.2006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Jones BC, Logsdon NJ, Walter MR. Structureof IL-22 bound to its high-affinity IL-22R1 chain. Structure. 2008a;16:1333–1344. doi: 10.1016/j.str.2008.06.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Jones BC, Logsdon NJ, Walter MR. Crystallization and preliminary X-ray diffraction analysis of human IL-22 bound to the extracellular IL-22R1 chain. Acta Crystallogr Sect F Struct Biol Cryst Commun. 2008b;64:266–269. doi: 10.1107/S1744309108004752. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Josephson K, Logsdon NJ, Walter MR. Crystal structure of the IL-10/IL-10R1 complex reveals a shared receptor binding site. Immunity. 2001;15:35–46. doi: 10.1016/s1074-7613(01)00169-8. [DOI] [PubMed] [Google Scholar]
  26. Kalie E, Jaitin DA, Abramovich R, Schreiber G. An interferon alpha2 mutant optimized by phage display for IFNAR1 binding confers specifically enhanced antitumor activities. J Biol Chem. 2007;282:11602–11611. doi: 10.1074/jbc.M610115200. [DOI] [PubMed] [Google Scholar]
  27. Kalie E, Jaitin DA, Podoplelova Y, Piehler J, Schreiber G. The stability of the ternary interferon-receptor complex rather than the affinity to the individual subunits dictates differential biological activities. J Biol Chem. 2008;283:32925–32936. doi: 10.1074/jbc.M806019200. [DOI] [PubMed] [Google Scholar]
  28. Kotenko SV, Gallagher G, Baurin VV, Lewis-Antes A, Shen M, Shah NK, Langer JA, Sheikh F, Dickensheets H, Donnelly RP. IFN-λambdas mediate antiviral protection through a distinct class II cytokine receptor complex. Nat Immunol. 2003;4:69–77. doi: 10.1038/ni875. [DOI] [PubMed] [Google Scholar]
  29. Krissinel E, Henrick K. Inference of macromolecular assemblies from crystalline state. J Mol Biol. 2007;372:774–797. doi: 10.1016/j.jmb.2007.05.022. [DOI] [PubMed] [Google Scholar]
  30. Lasfar A, Lewis-Antes A, Smirnov SV, Anantha S, Abushahba W, Tian B, Reuhl K, Dickensheets H, Sheikh F, Donnelly RP, et al. Characterization of the mouse IFN-λambda ligand-receptor system: IFN-lambdas exhibit antitumor activity against B16 melanoma. Cancer Res. 2006;66:4468–4477. doi: 10.1158/0008-5472.CAN-05-3653. [DOI] [PubMed] [Google Scholar]
  31. Lavoie TB, Kalie E, Crisafulli-Cabatu S, Abramovich R, DiGioia G, Moolchan K, Pestka S, Schreiber G. Binding and activity of all human alpha interferon subtypes. Cytokine. 2011;56:282–289. doi: 10.1016/j.cyto.2011.07.019. [DOI] [PubMed] [Google Scholar]
  32. Lazear HM, Nice TJ, Diamond MS. Interferon-l: Immune Functions at Barrier Surfaces and Beyond. Immunity. 2015;43:15–28. doi: 10.1016/j.immuni.2015.07.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Levin AM, Bates DL, Ring AM, Krieg C, Lin JT, Su L, Moraga I, Raeber ME, Bowman GR, Novick P, et al. Exploiting a natural conformational switch to engineer an interleukin-2 ‘superkine’. Nature. 2012;484:529–533. doi: 10.1038/nature10975. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Levin D, Schneider WM, Hoffmann HH, Yarden G, Busetto AG, Manor O, Sharma N, Rice CM, Schreiber G. Multifaceted activities of type I interferon are revealed by a receptor antagonist. Sci Signal. 2014;7:ra50. doi: 10.1126/scisignal.2004998. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Logsdon NJ, Jones BC, Allman JC, Izotova L, Schwartz B, Pestka S, Walter MR. The IL-10R2 binding hot spot on IL-22 is located on the N-terminal helix and is dependent on N-linked glycosylation. J Mol Biol. 2004;342:503–514. doi: 10.1016/j.jmb.2004.07.069. [DOI] [PubMed] [Google Scholar]
  36. Meager A, Visvalingam K, Dilger P, Bryan D, Wadhwa M. Biological activity of interleukins-28 and -29: comparison with type I inter-ferons. Cytokine. 2005;31:109–118. doi: 10.1016/j.cyto.2005.04.003. [DOI] [PubMed] [Google Scholar]
  37. Miknis ZJ, Magracheva E, Li W, Zdanov A, Kotenko SV, Wlodawer A. Crystal structure of human interferon-l1 in complex with its high-affinity receptor interferon-lR1. J Mol Biol. 2010;404:650–664. doi: 10.1016/j.jmb.2010.09.068. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Moraga I, Harari D, Schreiber G, Uzé G, Pellegrini S. Receptor density is key to the alpha2/beta interferon differential activities. Mol Cell Biol. 2009;29:4778–4787. doi: 10.1128/MCB.01808-08. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Muir AJ, Arora S, Everson G, Flisiak R, George J, Ghalib R, Gordon SC, Gray T, Greenbloom S, Hassanein T, et al. A randomized phase 2b study of peginterferon lambda-1a for the treatment of chronic HCV infection. J Hepatol. 2014;61:1238–1246. doi: 10.1016/j.jhep.2014.07.022. [DOI] [PubMed] [Google Scholar]
  40. Nice TJ, Baldridge MT, McCune BT, Norman JM, Lazear HM, Artyomov M, Diamond MS, Virgin HW. Interferon-l cures persistent murine norovirus infection in the absence of adaptive immunity. Science. 2015;347:269–273. doi: 10.1126/science.1258100. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Numasaki M, Tagawa M, Iwata F, Suzuki T, Nakamura A, Okada M, Iwakura Y, Aiba S, Yamaya M. IL-28 elicits antitumor responses against murine fibrosarcoma. J Immunol. 2007;178:5086–5098. doi: 10.4049/jimmunol.178.8.5086. [DOI] [PubMed] [Google Scholar]
  42. Pei J, Kim BH, Grishin NV. PROMALS3D: a tool for multiple protein sequence and structure alignments. Nucleic Acids Res. 2008;36:2295–2300. doi: 10.1093/nar/gkn072. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Piehler J, Thomas C, Garcia KC, Schreiber G. Structural and dynamic determinants of type I interferon receptor assembly and their functional interpretation. Immunol Rev. 2012;250:317–334. doi: 10.1111/imr.12001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Sato A, Ohtsuki M, Hata M, Kobayashi E, Murakami T. Antitumor activity of IFN-λ in murine tumor models. J Immunol. 2006;176:7686–7694. doi: 10.4049/jimmunol.176.12.7686. [DOI] [PubMed] [Google Scholar]
  45. Sheppard P, Kindsvogel W, Xu W, Henderson K, Schlutsmeyer S, Whitmore TE, Kuestner R, Garrigues U, Birks C, Roraback J, et al. IL-28, IL-29 and their class II cytokine receptor IL-28R. Nat Immunol. 2003;4:63–68. doi: 10.1038/ni873. [DOI] [PubMed] [Google Scholar]
  46. Shlomai A, Schwartz RE, Ramanan V, Bhatta A, de Jong YP, Bhatia SN, Rice CM. Modeling host interactions with hepatitis B virus using primary and induced pluripotent stem cell-derived hepatocellular systems. Proc Natl Acad Sci USA. 2014;111:12193–12198. doi: 10.1073/pnas.1412631111. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Spangler JB, Moraga I, Mendoza JL, Garcia KC. Insights into cytokine-receptor interactions from cytokine engineering. Annu Rev Immunol. 2015;33:139–167. doi: 10.1146/annurev-immunol-032713-120211. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Steen HC, Gamero AM. Interferon-lambda as a potential therapeutic agent in cancer treatment. J Interferon Cytokine Res. 2010;30:597–602. doi: 10.1089/jir.2010.0058. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Tan JC, Indelicato SR, Narula SK, Zavodny PJ, Chou CC. Characterization of interleukin-10 receptors on human and mouse cells. J Biol Chem. 1993;268:21053–21059. [PubMed] [Google Scholar]
  50. Thomas C, Moraga I, Levin D, Krutzik PO, Podoplelova Y, Trejo A, Lee C, Yarden G, Vleck SE, Glenn JS, et al. Structural linkage between ligand discrimination and receptor activation by type I interferons. Cell. 2011;146:621–632. doi: 10.1016/j.cell.2011.06.048. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Walter TS, Meier C, Assenberg R, Au KF, Ren J, Verma A, Nettleship JE, Owens RJ, Stuart DI, Grimes JM. Lysine methylation as a routine rescue strategy for protein crystallization. Structure. 2006;14:1617–1622. doi: 10.1016/j.str.2006.09.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Wang X, Rickert M, Garcia KC. Structure of the quaternary complex of interleukin-2 with its alpha, beta, and gammac receptors. Science. 2005;310:1159–1163. doi: 10.1126/science.1117893. [DOI] [PubMed] [Google Scholar]
  53. Wang X, Lupardus P, Laporte SL, Garcia KC. Structural biology of shared cytokine receptors. Annu Rev Immunol. 2009;27:29–60. doi: 10.1146/annurev.immunol.24.021605.090616. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Wang X, Hruska M, Chan P, Ahmad A, Freeman J, Horga MA, Hillson J, Kansra V, Lopez-Talavera JC. Derivation of Phase 3 dosing for peginterferon lambda-1a in chronic hepatitis C, Part 1: Modeling optimal treatment duration and sustained virologic response rates. J Clin Pharmacol. 2015;55:63–72. doi: 10.1002/jcph.363. [DOI] [PubMed] [Google Scholar]
  55. Wu PW, Li J, Kodangattil SR, Luxenberg DP, Bennett F, Martino M, Collins M, Dunussi-Joannopoulos K, Gill DS, Wolfman NM, Fouser LA. IL-22R, IL-10R2, and IL-22BP binding sites are topologi-cally juxtaposed on adjacent and overlapping surfaces of IL-22. J Mol Biol. 2008;382:1168–1183. doi: 10.1016/j.jmb.2008.07.046. [DOI] [PubMed] [Google Scholar]
  56. Yoon SI, Logsdon NJ, Sheikh F, Donnelly RP, Walter MR. Conformational changes mediate interleukin-10 receptor 2 (IL-10R2) binding to IL-10 and assembly of the signaling complex. J Biol Chem. 2006;281:35088–35096. doi: 10.1074/jbc.M606791200. [DOI] [PubMed] [Google Scholar]
  57. Yoon SI, Jones BC, Logsdon NJ, Harris BD, Deshpande A, Radaeva S, Halloran BA, Gao B, Walter MR. Structure and mechanism of receptor sharing by the IL-10R2 common chain. Structure. 2010;18:638–648. doi: 10.1016/j.str.2010.02.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Zdanov A. Structural analysis of cytokines comprising the IL-10 family. Cytokine Growth Factor Rev. 2010;21:325–330. doi: 10.1016/j.cytogfr.2010.08.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Zhou Z, Hamming OJ, Ank N, Paludan SR, Nielsen AL, Hartmann R. Type III interferon (IFN) induces a type I IFN-λike response in a restricted subset of cells through signaling pathways involving both the Jak-STAT pathway and the mitogen-activated protein kinases. J Virol. 2007;81:7749–7758. doi: 10.1128/JVI.02438-06. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

supplement
NIHMS869094-supplement.pdf (6.8MB, pdf)

RESOURCES