Conserved Induction of Distinct Antiviral Signalling Kinetics by Primate Interferon Lambda 4 Proteins

Interferon lambdas (IFNλ) (also known as type III IFNs) are critical cytokines that combat infection predominantly at barrier tissues, such as the lung, liver, and gastrointestinal tract. Humans have four IFNλs (1–4), where IFNλ1–3 show ~80%–95% homology, and IFNλ4 is the most divergent displaying only ~30% sequence identity. Variants in IFNλ4 in humans are associated with the outcome of infection, such as with hepatitis C virus. However, how IFNλ4 variants impact cytokine signalling in other tissues and how well this is conserved is largely unknown. In this study, we address whether differences in antiviral signalling exist between IFNλ4 variants in human hepatocyte and intestinal cells, comparing them to IFNλ3. We demonstrate that compared to IFNλ3, wild-type human IFNλ4 induces a signalling response with distinct magnitudes and kinetics, which is modified by naturally occurring variants P70S and K154E in both cell types. IFNλ4’s distinct antiviral response was more rapid yet transient compared to IFNλ1 and 3. Additionally, divergent antiviral kinetics were also observed using non-human primate IFNλs and cell lines. Furthermore, an IFNλ4-like receptor-interacting interface failed to alter IFNλ1’s kinetics. Together, our data provide further evidence that major functional differences exist within the IFNλ gene family. These results highlight the possible tissue specialisation of IFNλs and encourage further investigation of the divergent, non-redundant activities of IFNλ4 and other IFNλs.


INTRODUCTION
Viral infections of mucosal surfaces like the lung, gut, and liver [such as influenza, rotavirus and hepatitis C virus (HCV)] remain major drivers of global morbidity and mortality in the human population (1). The host innate immune response is a critical determinant of the outcome of infection and as such, its stimulation can influence clinical outcomes (2). Following sensing of viral infection, several antiviral and immunoregulatory factors like cytokines are induced that act to limit viral replication and promote clearance and long-term immunity (3). Interferons (IFNs) are one important group of such cytokines with potent antiviral activity (4). There exist three recognised families of IFNs: the type I IFNs (alpha 1-13, beta, epsilon, kappa, and omega in humans), type II IFNs (gamma), and type III IFNs [lambdas (l) 1-4] (5). Types I and III IFNs are rapidly induced and secreted following sensing of infection in most nucleated cells. These secreted IFNs then act in turn on the infected cell and on neighbouring uninfected cells to induce the production of hundreds of interferon stimulated genes (ISGs) via activation of the Janus kinase-signal transducer and activator of transcription (JAK-STAT) pathway. Although they share similar downstream signalling pathways and lead to the activation of similar ISGs, type I and III IFNs utilise distinct cell surface receptor complexes (6). Type I IFNs use the ubiquitously expressed IFNAR1 and IFNAR2 heterodimeric complex, whilst type III IFNs use the IFNlR1 and IL10R2 heterodimeric complex. Although also found on some immune cell types (7), IFNlR1 is predominantly expressed on epithelial cells at the socalled barrier tissues (8), including the respiratory and gastrointestinal tracts, and hepatocytes in the liver of humans (9), which provides type III IFNs distinct traits specialised in the protection of mucosal surfaces compared to type I IFNs (10)(11)(12).
Although they share a receptor complex, there is emerging evidence that not all type III IFNs have redundant features (13). The human IFNls, namely, IFNl1, IFNl2, and IFNl3, all share >80% homology, yet compared to IFNl4, they exhibit only~30% homology (6,9,13). Whilst all type III IFNs are more recently discovered in comparison to type I IFNs (4,6,9), IFNl4 was the latest addition to the family being only identified in 2013 (13). The outcome of HCV infection is associated with genetic variation in the human IFNL locus [e.g., "IL28B" single nucleotide polymorphisms (SNPs)], likely mediated by variants within IFNL4 (14). IFNl4, like other IFNls, has potent antiviral activity (15). These same genetic variants are also associated with extra-hepatic infections, such as enteroviral infection in the respiratory tract (16). There are two common loss-of-function SNPs in human IFNL4, encoding a frameshift (rs12979860), and a non-synonymous variant P70S (rs117648444, which encodes a proline to serine mutation at position 70), respectively (13,14). Whilst the frameshift ablates IFNl4 production, P70S reduces the potency of ISG induction by IFNl4 (14,17,18). Interestingly, it is these hypo-or inactive alleles that are associated with protection from chronic HCV infection in humans (13,14).
Further investigation into the functional diversity of IFNl4 identified two rare variants that affect IFNl4 activity, including an additional hypoactive variant L79F (leucine to phenylalanine at position 79) and K154E (lysine to glutamic acid at position 154), which dramatically enhances IFNl4 antiviral activity by increasing its secretion and potency (17). Intriguingly, although K154 is nearly ubiquitous in the human population, E154 is the ancestral amino acid at this position in non-human primates and other mammals. E154 was found in a small number of extant humans. Accordingly, chimpanzee and rhesus macaque IFNl4 have enhanced antiviral activity relative to wild-type human IFNl4, which can be reversed by an E154K mutation. Together, the evolutionary data suggest a step-wise attenuation of IFNl4 activity (E154K > P70S > TT frameshift) unique to modern humans (13), which is consistent with the non-redundancy of IFNl4 compared to other IFNls. However, which precise unique biological feature(s) of IFNl4 that are non-redundant (and thus have been acted upon by evolution) are poorly understood and only beginning to be unravelled (14,19,20).
Following on from our previous work (11,17), we wished to determine how the antiviral activity of IFNl4 and its variants and homologues changed in a time-dependent manner, compared to other IFNls. To test this hypothesis, we characterised the kinetics of signalling and antiviral activity of a panel of IFNl4 variants in human hepatocyte and human intestinal epithelial cells compared to IFNl3. Together, our work demonstrates the unique kinetics of IFNl4 activity compared to other IFNls, which is conserved within and between species. Further work on the intrinsic differences between IFNl4 and other IFNs is warranted.

IFNl Variants Display Unique STAT1 Phosphorylation Kinetics
Binding of IFNls to their receptor complex leads to activation of downstream signalling cascades that ultimately lead to the establishment of an antiviral state (6). The JAK/STAT pathway is one of the most critical and well-characterised pathways activated following IFNl binding. An emerging view is that the kinetics of such a downstream response is a crucial determinant of the antiviral potential of IFNls (11,20). To probe the temporal basis of IFNl signalling in greater detail, we first measured phosphorylation of STAT1 over time at Y701 (Figure 1). Human hepatocyte HepaRG monolayers were incubated with conditioned media estimated to contain equivalent amounts of IFNls (IFNl3, IFNl4 WT, P70S, L79F, and K154E) for 15, 30, 60, 120, and 360 min, and 24 h. Following stimulation, protein lysates were harvested, and STAT1 phosphorylation was assayed by immunoblot analysis ( Figure 1A). Conditioned media generated following transfection of an enhanced green fluorescent protein (EGFP)expressing plasmid served as a negative control. Results showed that IFNl3 and a number of IFNl4 variants induced detectable levels of pSTAT1 ( Figure 1A, quantified in Supplementary Figure 1A). L79F gave extremely low levels of pSTAT1 (data not shown), which likely correlates with its very limited activity as described previously (17). Interestingly, IFNl3 and IFNl4 variants induced distinct kinetics of pSTAT1 activation ( Figure 1A and Supplementary Figure 1A). Whilst all IFNls peaked around similar times (30 min to 1 h), IFNl4 WT and P70S showed clear transient activation, whilst IFNl3 and K154E displayed persistent activation of pSTAT1. Importantly, levels of pSTAT1 correlated with previously measured antiviral potential for three IFNl4 variants K154E > WT > P70S (17). As IFNls can also signal in other tissues apart from the human liver (8,21), and there is an emerging role for IFNl4 in extra-hepatic environments, we assayed whether human colon carcinoma cells (T84) were capable of inducing pSTAT1 in response to IFNl4 and its variants. Intestinal T84 cells were treated and incubated with conditioned media containing equivalent amounts of IFNls (IFNl3, IFNl4 WT, P70S, and K154E), and their induction of pSTAT1 was assayed over time by immunoblot analysis ( Figure 1B and Supplementary Figure 1B). We observed similar trends as to HepaRG, although differences in amplitude of pSTAT1 induction were noted, especially for IFNl4 K154E in T84 cells. Together, these results show that both hepatic and intestinal cell lines can respond to both IFNl3 and IFNl4 and display variant-specific inductions of the JAK/STAT pathway.

IFNl Variants Display Different Levels of ISG Induction
Phosphorylation of STAT1 following receptor complex engagement by IFNls results in STAT1/2 dimer formation and translocation to the nucleus to induce ISG transcription, which ultimately leads to the production of antiviral proteins and the establishment of an antiviral state (6). Our previous work showed that IFNl variants induced different levels of ISG expression when measured at 24 h (17). To ascertain whether this ISG expression varied at earlier times after incubation in concert with the kinetics of pSTAT1 activation, we measured the relative induction of a panel of core ISGs (IFIT1, MX1, ISG15, and RSAD2/VIPERIN) compared to EGFP-treated conditioned media in HepaRG cells (Figures 2A-D) and T84 cells ( Figures 2E-H). Compared to EGFP-conditioned media stimulated cells, all IFNls induced measurable increases in ISG mRNA in HepaRG cells but with discernible differences in magnitude. T84 cells also showed ISG induction for four of the supernatants tested ( Figures 2E-H). Additionally, looking at relative fold change, T84 cells gave a lower induction of all ISGs as compared to HepaRG cells ( Figure 2). The magnitudes of ISG induction for both cell lines mirrored the pSTAT1 induction that was observed in Figure 1 (IFNl3/K154E > WT > P70S > L79F). IFNl4 K154E induced a similar pattern of ISG induction as IFNl3 in both cell lines. Interestingly the kinetics of ISG induction was distinct to each cell line. In HepaRG cells, all IFNls induced an early peak induction of ISGs, which subsequently declined over time. Moreover, IFNl4 K154E demonstrated a slightly faster induction and peaked by 2 h whilst all other IFNls tested peaked at 6 h. By contrast, IFNl3 and the IFNl4 K154E showed no or little decline in ISG induction after induction at either 2 or 6 h in T84 cells ( Figures 2E-H). Additionally, T84 cells yielded low to almost undetectable induction of ISGs following IFNl4 WT and P70S treatment. Together, these results show that K154E provides similar stimulatory activity to IFNl3 and that this is far greater than for either IFNl4 WT or P70S, which are the most common IFNl4 variants in the human population.

IFNl Variants Have Distinct Antiviral Activity in Intestinal Cells
Induction of an antiviral state is the major downstream consequence of IFN signalling. To determine how STAT1 phosphorylation and ISG expression correlate with antiviral activity, we infected the hepatic and intestinal cell models with two different viruses, EMCV and VSV. Both EMCV and VSV are Mock control cells were treated with conditioned media from EGFP-plasmid transfected HEK-293T cells, and all values were normalised against this value at each time. GAPDH (HepaRG) or HPRT1 (T84 cells) were used as housekeeping genes. L79F did not induce any detectable ISG induction in T84 cells. Error bars represent the mean ± SEM from two to three biological replicates from at least two independent protein batches. highly cytopathic, replicate very fast, and are sensitive to IFN, which makes them suitable for assessing the kinetics of antiviral activity. EMCV infectivity and replication were assayed by determining the cytopathic effects of the virus, whilst a VSV encoding luciferase (VSV-luc) was deployed and its infectivity was measured by luciferase assay. HepaRG and T84 cells were treated with increasing concentrations of EGFP or IFNl3 or IFNl4-containing supernatants at 24 h prior to virus infection. Following IFNl pretreatment, cells were infected with EMCV or VSV [multiplicity of infection (MOI) of 0.3 and 1, respectively] in the continuous presence of IFNls, and infection was assayed at 24 h post-infection for EMCV ( Figure 3A) and 8 h postinfection for VSV ( Figures 3B, C). Different assay times for VSV versus EMCV were due to differences in replication kinetics and cytopathic effects of either virus. Consistent with our previous work (17,21), results show that VSV infection was inhibited by all IFNs in both cell lines ( Figures 3B, C). IFNl3 was the most potent IFN, as it reduced VSV infection with 10% of the maximum concentration in both HepRG and T84 cells. IFNl4 WT and K154E showed similar antiviral activity; however, a much higher concentration of these two IFNs was required to reach a similar potency as IFNl3. Consistent with previous low pSTAT1 and ISG inductions, P70S was only able to slightly reduce virus infection even at the highest concentrations in both cell lines. T84 cells were poorly infected with EMCV and highly resistant to the cytopathic effects of the virus, and therefore, antiviral activity was not assayed in this cell line, but similar patterns of antiviral activity were seen for HepaRG cells infected with EMCV ( Figure 3A).

IFNl Variants Have Distinct Kinetics of Antiviral Activity
Having established antiviral assays in both liver-and intestinalderived cell lines, we wished to determine whether IFNl activity was time dependent and whether the continuous presence of IFNls was required to maintain their antiviral activity. Therefore, we performed infections and antiviral assays over time, both in the continuous presence of IFNls but also in cells that had been pretreated with IFNls for varying lengths of time, yet the cytokines were then removed, monolayers washed, and fresh media provided ("non-washed" and "washed", respectively, Figure 4A) prior to infection. Initially, we conducted experiments in T84 cells that were infected with VSV following IFNl pretreatment ( Figures 4B-E). For clarity, it should be noted that the following data are presented differently than those in Figure 3. In agreement with the data presented in Figure 3, all IFNls demonstrated antiviral activity with IFNl3 and IFNl4 P70S showing the greatest and least potency, respectively. The peak of IFNl3 activity was delayed relative to all IFNl4s. Moreover, we found that shorter incubation times with IFNl3 followed by its removal before infection reduced its antiviral activity to a greater extent compared to the three IFNl4 variants used in the experiment (compare early time points in Figure 4B with Figures Figure S2). The results show a greater decline in antiviral activity (~15-fold) from 24 h incubation to 6 and 2 h incubation for IFNl3 compared to IFNl4 WT and IFNl4 K154E, which showed reductions in activity by~0.75-4 fold.

Divergent Kinetics Is Independent of Human IFNl System
Our data suggest that in human cells, human IFNl4 and its variants induce a distinct antiviral response compared with human IFNl3. As previous work has demonstrated that IFNl4 from different primate species have varying levels of antiviral activity (17,22), we next explored whether the distinct signalling kinetics that we observed were also species specific. We first analysed the amino acid homology between IFNl3, IFNl4, IFNlR1, and IL10R2 in humans, chimpanzees, and rhesus macaques ( Figure 5A). Results showed that although the various orthologues shared a high degree of homology (92-97%), there were differences that could affect activity given that even a single amino acid change can alter signalling as in IFNl4 variants P70S and K154E. Given these genetic differences, we next tested the antiviral kinetics of non-human IFNls. First, we treated human HepaRG cells with human and non-human IFNls as described in Supplementary Figure 2, by treating cells for 2, 6, and 24 h, and then removing the cytokines prior to infection with EMCV at 24 h after initial stimulation ( Figure 5B). For these experiments, we utilised non-human primate IFNl3 or IFNl4 proteins containing a C-terminal FLAG tag, which we characterised previously (17). In these experiments, we utilised IFNl4 K154E as a model human IFNl4, since it gave robust levels of detectable antiviral activity, with kinetics broadly similar to IFNl4 WT. All IFNls had antiviral activity against EMCV with chimpanzee IFNl4 having greater activity than human and macaque IFNl4 ( Figure 5C), whilst human IFNl3 had greater activity than macaque IFNl3 ( Figure 5D). Similar to human variants, the peak of IFNl3 activity was delayed relative to all IFNl4s. IFNl3 washing experiments demonstrated that like human IFNl4, non-human primate IFNl4 were more refractory to early removal than human or macaque IFNl3 ( Figures 5C, D). To determine if these characteristics also occurred in non-human cells, we repeated these experiments in the rhesus macaque respiratory epithelial cell line LLCMK2 ( Figures 5E, F). Results showed that all IFNl4s had similar kinetics of antiviral activity but different levels of potencies as found in HepaRG cells.
Washing following immediate infection supported the initial washing experiments with IFNl3 antiviral activity being more sensitive to early removal of cytokine ( Figures 5G, H).

IFNl1 With Receptor-Interacting Face Mutations Retain Parental Kinetics
Complex and dynamic interactions between cytokine ligands and their cognate receptors dictate the signalling output (23). To probe further the molecular genetic basis of IFNl kinetics, we sought to mutate and disrupt the receptor binding faces of IFNl, hypothesising that these residues were most likely to be responsible for IFN kinetics. IFNl4 is highly divergent when compared with IFNl1-3 with~30% similarity detected,  Figure 4). Results are shown as mean ± SD from four biological replicates. Data for panels (A-F) were obtained using independent protein batches as panels (G, H). suggesting that there are likely to be distinct molecular determinants of differential signalling contained within IFNl4 compared to the other human IFNls (13,15). To begin to identify those determinants, we constructed chimeric IFNls between IFNl4 and human IFNl1. IFNl1 was chosen, as it is known to have similar kinetics to IFNl3 (10) but, like IFNl4, is N-linked glycosylated (6). Initially, comparison of differentially conserved amino acids in IFNl4 (human and non-human primate) with IFNl1-3 (human and macaque) identified a divergent receptor binding interface between these groups of IFNls suggestive of distinct receptor interactions ( Figure 6A). We focused on divergent, likely surface-exposed residues near relevant helices (A, D, and F) and designed two chimeric IFNls based on IFNl1 containing candidate IFNl4 residues from the IFNlR1-binding helix F (F) and the IL10R2-binding helices A and D (AD). An additional chimera with all three IFNl4 binding helices was generated, termed ADF. We first confirmed that IFNl1 and its chimeras were produced and released into the supernatant using a splitluciferase assay; this showed that chimeras incorporating helices A and D yielded reduced production ( Figure 6B). To test their antiviral activity, HepaRG cells were pretreated for 2, 6, or 24 h prior to EMCV infection with the WT IFNs and each of the indicated chimeras. Results showed that IFNl1 had higher   Figure 2F and Supplementary Figure 2). IFNl1/4 chimeras had similar kinetic profiles as IFNl1. Results revealed that washing reduced the antiviral potency of all IFNs and IFNl1, and all the chimeras were more greatly affected than IFNl4. Taken with our antiviral activity results suggested that chimera F had reduced potency compared to IFNl1, whilst the reduced activity of AD is likely due to reduced protein, and thus, the impact on ADF is due to reduced amount and potency ( Figure 6B). However, despite alteration of the receptor interaction surfaces, the kinetics remain conserved similar to IFNl1 (and IFNl3), suggesting that these residues only modify the magnitude of the antiviral response and are not sufficient to alter the antiviral kinetics. Altogether, our work described here demonstrates the distinct yet conserved antiviral kinetics of human and non-human primate IFNl4 compared to other IFNls.

DISCUSSION
Knowledge of the molecular signalling pathways stimulated by IFN binding is essential to understand immunity to infectious diseases and could help develop more effective interventions. The dynamics of antiviral signalling is emerging as a physiologically relevant and important topic, and several groups have shown that type III IFNs have distinct slower but sustained signalling kinetics compared to type I IFNs (10,11,21). Very few studies have addressed whether different members of the type III IFN family also have a similar kinetics for the activation of STAT1, induction of downstream ISGs, and antiviral activity (20). Through several lines of genetic evidence, it appears that human IFNl4 has non-redundant functions relevant to immunity compared to other IFNls, yet the determinants of this unique biology are poorly understood (13,14). Additionally, there exist a number of naturally occurring functional variants of IFNl4 that are known to impact potency (14,17,18). In this work, we addressed whether IFNl4 WT and its variants (e.g., P70S and K154E) have altered antiviral kinetics, in comparison to IFNl1 and IFNl3. By comparing IFNl4 signalling and antiviral activity in two cell lines from two distinct organs, we were able to identify conserved and variable features of IFNl4 and IFNl3 signalling that demonstrated distinct antiviral kinetics, consistent with recent studies (20). Critically, we also show that common (P70S) and rare (K154E) human variants predominantly impact the magnitude of IFN signalling but not the kinetics of that response, and these dynamics are largely conserved in non-human primate IFNls and their cognate cell lines.
Comparison of IFN activity across variants is notoriously challenging given the need for input normalisation and relevant processing. To circumvent these issues, we produced IFNl in human cells (HEK-293T) and normalised for input IFNl using a C-terminal "split luciferase" "HiBiT" tag system. Interestingly, using normalised amounts of protein released into the supernatant of transfected cells, we detected different potencies for each IFNl, consistent with our previous work (17). In general, IFNl3 had greater antiviral potency than WT IFNl4 in both human liver-and gut-derived cell lines. WT IFNl3 induced stronger and more prolonged STAT1 phosphorylation, higher magnitude of ISG induction, and a stronger antiviral effect than WT IFNl4, which induced a lower and more transient response. The IFNl4 K154E variant displayed potency that was more similar to IFNl3 and shows that, at least for one rare variant, human IFNl4 has the potential to have significant stimulatory effects. Considering the dynamics of the response, we show clear differences between IFNl3 and IFNl4 variants antiviral activity over time. These observations are consistent with previous work on IFNl4 WT kinetics (20). Interestingly, IFNl3 and IFNl4 showed differential characteristics by limiting their contact time with target cells, suggestive of different interactions with receptor complexes. This observation requires more detailed biochemical and cell biology analysis, preferably using purified proteins and receptor molecules that would allow measurement of binding affinities, on-off rates, and their effects on receptor trafficking, for the IFNl family. Such analysis was beyond the scope of our study primarily due to the inherent technical difficulties of preparing significant quantities of soluble, correctly folded IFNl4.
The interaction between IFNls and their receptor complexes remains poorly understood, although several crystal structures of IFNl proteins, with the exception of IFNl4, in the presence and absence of its heterodimeric receptor complex IFNlR1 and IL10R2 have been solved (24,25). There is reason to believe that IFNl4 is likely to interact differently with its receptors based on amino acid sequence alignments (13,15). IFNl4 and IFNl1/ 2/3 share only~30% homology, with highest levels found in the IFNlR1-binding "helix F." Aside from helix F, IFNl4 differs considerably compared to the other IFNls, including other receptor binding helices, such as helix D that binds IL10R2. To test the contribution of IFNl4 receptor interactors in and around helices A, D, and F, we constructed chimeras using IFNl1 as a reporter for antiviral activity into which we inserted predicted receptor binding domains from IFNl4. These IFNl1/IFNl4 chimeric displayed similar kinetic profiles as IFNl1 and IFNl3, although differences in production and potency were noted. This suggests that the molecular determinants that regulate binding kinetics may not lie solely in the putative surface-exposed receptor-binding interfaces that we tested. IFNl4 differs in structural capacity to IFNl1/3, which may not be captured in our chimeras, and further differences are observed in other helices that may play roles in signalling. A possible explanation for these differences could be due to differing stabilities for each of the IFNls. The stability of each IFNl has not yet been tested but could provide insight into how each family member achieves its maximal activity. However, as most of our assays were performed in relatively short time frames (2-6 h), it seems unlikely that IFN stability played a role in the differences we observed and is more likely that IFNl4 interacts and activates the receptor more rapidly, likely through binding more strongly analogous to type I IFNs (23).
An important aspect of our work is that the differences we detected between IFNl3 and IFNl4 in antiviral kinetics were conserved in non-human species, through analysis of chimpanzee and macaque IFNl4 and macaque IFNl3 in human and macaque cell lines. This is important because compared to other primates, humans appear to have evolved unique IFNl4 features relevant for outcome of infectious diseases like HCV (13,14,17). This finding would be consistent with the limited genetic differences between these species (>90% similarity). The fact that the kinetics are not unique to humans supports the hypothesis that alterations in IFNl4 potency has been the dominant phenotype that our recent evolution has acted upon. It would be of interest to test further related IFNl4, from distantly related mammals (22).
Testing IFNl kinetics in two cell lines allows us to assess conserved and divergent activities in hepatocytes and intestinal cells. IFNls can signal in many tissues (8), including the human gut (21), and recent work has implicated variants in IFNl4 in the outcome of enterovirus infection in the respiratory tract but which can infect the gut as well (16). The role of IFNl4 in intestinal cells up until now has been largely unexplored. Whilst IFNl3 and IFNl4 can signal in both cell types, we show clear differences in potency of human IFNl4 variants, consistent with our previous work in hepatocytes. Comparing the induction of IFNl signalling in HepaRG and T84 cells suggested that the hepatocyte cell line was more sensitive to IFNls, yet to draw any conclusions, primary liver and intestinal cells or organoids from several individuals should be tested. Nevertheless, we observed consistent kinetics differences of IFNl4 compared to IFNl3 in both cell lines.
Our work has several implications, most importantly those relating to the conserved differences between IFNl3 and IFNl4. Compared to type I IFNs, IFNls have been defined partially by their slower, sustained signalling kinetics (10,11). IFNl4 has several unique features, including its association with certain diseases, transcriptional suppression, and evolution in humans, which suggests a degree of specialisation. Unlike other IFNls, IFNl4 appears to signal more like type I IFNs despite utilising IFNlR1 and IL10R2. Thus, IFN kinetics may not solely lie in receptor biology but in the interactions between cytokine and receptor. We hypothesise that one outcome of the kinetics of IFNl1-3 outlined here, where activity is dependent on time and local concentration, would be a more tunable strategy, which may have "adaptive" potential for mucosal surfaces where more robust IFN activities may have pathogenic effects. Whether the unique kinetics of IFNl4 would provide additional nonredundant therapeutic benefit over other IFNls remains to be explored.
In conclusion, we provide further evidence of the functional divergence of IFNl4 compared to other IFNl proteins supporting the continued investigation into the causes and consequences of such distinctive signalling on the human immune system, which may be exploited for therapeutic gain.

Viruses
Two viruses were used in this study: Ruckart strain of encephalomyocarditis virus (EMCV) and VSV. EMCV was produced in Vero cells following low MOI infection (MOI = 0.0001) and harvested between 1 and 2 days when extensive cytopathic effect was observed. EMCV infectivity was quantified by TCID 50 and typically grew to titres of~10 8 /ml. VSV-luc was a kind gift from Sean Whelan (Washington University, St. Louis) and was produced and titrated as described in (26,27).

Molecular Biology
Recombinant DNA technology was utilised to generate the IFNs for functional testing in this study, as previously described (15,17). The mammalian expression plasmids expressing HiBiT variants and human IFNl3, and chimpanzee (Pan troglodytes) and rhesus macaque (Macaca mulatta) IFNl4 with a carboxyterminal FLAG tag were described previously (17). Rhesus macaque IFNl3-FLAG was generated synthetically (GeneArt) with sequence corresponding to XP_001086865.3 alongside WT IFNl1-HiBiT, or IFNl1/l4-HiBiT chimeras were constructed synthetically (GeneArt) with sequences from helices A, D, and F as shown ( Figure 5) and cloned into expression vector pC1 and sequenced confirmed by Sanger sequencing. Correct plasmids were purified by midiprep or maxiprep and quality and quantity determined by NanoDrop prior to transfection. An EGFP expression plasmid prepared in identical conditions was used as a negative control throughout.

IFN Production
IFNls were produced using the protocol described previously, which is capable of generating functional IFNls (17). Briefly, IFN expression plasmids were transfected into sub-confluent HEK-293T cell monolayers, which are hyporesponsive to IFNl signalling due to very low expression of IFNlR1 (15). Lipofectamine 2000 was used to transfect IFNl plasmids per manufacturer's instructions. IFNls were routinely generated in six-well plates or 10 cm dishes, and 2 and 14 µg of plasmids were used, respectively. Lipofectamine 2000 (2 µl) was used per microgram of plasmid. Plasmids were transfected into cells in Optimem for 16-18 hours, before changing media to growth media (10% FCS) until 2 days posttransfection was reached. The conditioned media were harvested, clarified by centrifugation, aliquoted, and immediately frozen at −80 in. Relative levels of IFNls were estimated using the extracellular HiBiT split luciferase assay by virtue of their C-terminal HiBiT tag by incubating IFN preparations with assay reagents and measured by manufacturer's instructions (Nano-Glo HiBiT Extracellular Detection system, Promega) using a luminometer.

Interferon Treatments
For quantitative real-time PCR (qRT-PCR) or immunoblotting experiments, IFN stimulation was achieved by incubating cell monolayers with IFNl-containing conditioned media at a defined concentration to have equivalent HiBiT signal for each sample. Cells treated with IFNs were incubated for the indicated period of time before either being processed. A previous titration analysis indicated that a~1:2-1:4 dilution of IFNl4-WT is enough to give a robust induction of ISGs for all variants (17) and antiviral response whilst limiting the amount of conditioned media added to cells (<50% of total volume). Therefore, WT IFNl4 was used at the standard, and the levels of other IFNs were normalised to this by virtue of the HiBiT tag. Based on HiBiT assay measurements, the relative ratios of supernatant were IFNl4(WT):P70S:L79F:K154E:IFNl3,~1:2:2:0.2:0.01. For the analysis of pSTAT1 levels by immunoblotting, 100,000 cells were seeded in 500 µl of growth media, into sterile rat-tailcollagen-coated (T84) or untreated (HepaRG) 24-well plates. To analyse ISG expression levels by qRT-PCR, 50,000 T84 cells were seeded in 500 µl DMEM/F12 into sterile rat-tail-collagen-coated 48-well plates, or 1,000,000 HepaRG cells were seeded into 2 ml of growth media into 6-well plates. Cells were treated either with HEK293T cell supernatants containing either IFNl3 or different IFNl4 variants [l4 wild type (WT), K154E, P70S] or GFPconditioned medium (Mock). Prior to treatment, media were removed; cells were rinsed once in PBS and then treated with each IFN diluted in their corresponding growth media to achieve an equal concentration (as determined by HiBiT) and added to the cells in 500−1,000 µl/well. Cells were then incubated at 37°C and 5% CO 2 until harvest. Cells were harvested at 15 min, 30 min, 1 h, 2 h, 4 h, 24 h, and 48 h posttreatment for the analysis of pSTAT1 protein levels by immunoblotting, whereas total RNA was isolated from T84 cells at 2, 4, 8, 12, and 24 h posttreatment for the analysis of ISG expression levels by qRT-PCR. To control for batch-to-batch variability in protein production, at least two independent protein preps were used.

Viral Infections
For the EMCV antiviral assays, 5,000 cells were seeded per well in a 96-well plate 24 h prior to treatment. At the day of treatment, IFNs were added in twofold serial dilutions to the cells 24 h prior to infection. Following IFN treatment, EMCV (MOI = 0.3) was added to the cells, and infection was scored by CPE 24hpi visually or by crystal violet staining. EMCV is highly cytopathic in certain cell lines and very sensitive to IFN. The reciprocal of the dilution giving~50% protection was used as a semiquantitative measure of IFNl-conditioned media activity.
For VSV infection, T84 or HepaRG cells were seeded in a white bottom 96-well plate. Cells were pretreated prior to infection as indicated time points and concentrations of IFN-l3 and IFN-l4, and its variants K154E and P70S. VSV-luc (MOI = 1) was added to the wells, and the infection was allowed to proceed for 8 h. At the end and the infection, media was removed, cells were washed 1× with PBS and lysed with cell lysis buffer (Promega) at room temperature (RT) for 20 min. The same volume of Steady Glo (Promega) was added to the cells and incubated for 15 min. Luminescence was read using Tecan Infinite M200 Pro.

Immunoblotting
At the time of harvest, cells were rinsed once with PBS and then lysed with 1× radioimmunoprecipitation assay (RIPA) buffer [150 mM sodium chloride, 1.0% Triton X-100, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulphate (SDS), 50 mM Tris at pH 8.0 supplemented with phosphatase and protease inhibitors (Sigma-Aldrich or Thermo Fisher) for 5-10 min at RT (T84) or ice (HepaRG)]. Cell lysates were collected, and roughly equal amounts of protein were then separated by SDS-PAGE in a 10% (HepaRG) or 12% (T84) polyacrylamide gel, following boiling and reducing. Lysates were then blotted onto a nitrocellulose membrane (T84) or PVDF (HepaRG) by wet blotting. Membranes were blocked with blocking buffer [5% BSA in TBS containing 0.1% Tween-20 (TBS-T)] for 1 h at RT whilst shaking. Primary antibodies (1:1,000 dilution) were diluted in blocking buffer and incubated overnight shaking at 4°C. The membranes were washed four times in TBS-T for 10 min at RT. Then, secondary antibodies were diluted in blocking buffer and incubated for 1 h shaking at RT. Membranes were again washed four times in TBS-T for 10 min at RT. HRP detection reagent (GE Healthcare) was mixed 1:1 and incubated at RT for 2-3 min, or ECL substrate is added (Immobilon crescendo western HRP substrate, WBLUR0100, Merck). Membranes were then exposed to film and developed or visualised by chemiluminescence using the G:BOX Chemi gel doc Imaging System Instrument (Syngene). The detection of bactin (T84) or b-tubulin (HepaRG) was used as loading controls. For quantitative analysis, pSTAT1 intensities of each immunoblot were quantified for each timepoint using ImageJ or Image Studio Lite Version 5.2. For quantification with ImageJ, the background value (Mock) was manually subtracted from the calculated values. pSTAT1 levels were then determined relative to control.

RT-qPCR
The total RNA was purified from lysed cells using the Nucleo Spin ® RNA extraction kit (T84) by Marchery-Nagel (Catalog number 740955.50) according to the manufacturer's instructions or (HepaRG) RNeasy Mini Kit (74106, Qiagen). RNA concentration was measured using the NanoDrop Lite spectrophotometer (Thermo Scientific). For T84 cells, 250 ng of total RNA was reverse transcribed into cDNA using the iScript ™ cDNA Synthesis kit (BioRad Laboratories, Catalog number 1708891). The reaction contained a mixture of 1 ml Reverse Transcriptase, 4 ml Reaction Mix, and 15 ml of RNA template in nuclease-free water. The newly synthesised cDNA was diluted 1:2 in RNase/DNase free water. The following qRT-PCR was performed using a Bio-Rad CFX96 Real-Time PCR Detection System. Per reaction 7.5 µl of SsoAdvanced Universal SYBR Green Supermix, 2 µl of 1:2 diluted cDNA, 1.7 µl of nuclease free water, and 1.9 µl of either forward or reverse primers (2 µM) for the amplification of IFIT1 (fw: 5′-AAAAG CCCACATTTGAGGTG-3′; rev: 5′-GAAATTCCTGAAA CCGACCA-3′), ISG15 (fw: 5′-CCTCTGAGCATCCTGGT-3′; rev: 5′-AGGCCGTACTCCCCCAG-3′), Viperin (fw: 5′-GAGAGCCATTTCTTCAAGACC-3′ and rev: 5′-CTATAATC C C T A C A C C A C C T C C -3 ′ ) , a n d M x 1 ( f w : 5 ′ -GGTCTATACCACACGCACAGA-3′; rev: 5′-ACTGGTT TCCTTTGCCTCGT-3′) were used. Data analysis was performed using the Bio-Rad CFX Manager 3.0. The expression of the targeted genes was then normalised to the h o u s e k e e p i n g g e n e H P R T 1 ( f w : 5 ′ -C C T G G C G T C GTGATTAGTGAT-3′; rev: 5′-AGACGTTCAGTCCTGT CCATAA-3′). For HepaRG cells, 1 mg of total RNA was reverse transcribed into cDNA using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, UK). The reaction contained a mixture of 1 ml Reverse Transcriptase, 9 ml Reaction Mix, and 10 ml of RNA template in nuclease-free water. The newly synthesised cDNA was diluted 1:25 in RNase/ DNase free water. The following qRT-PCR was performed using a Real-Time Ready PCR Kit (Roche) and TaqMan primerprimer-probe mixes. Each reaction mixture consisted of 10 ml of 2× LightCycler 480 Probes Master, 1 ml of 20× Real-Time Ready Assay with 4 ml PCR-grade H 2 O (total volume, 15 ml). Template DNA, defrosted on ice, was first diluted 1:25 (v/v) with PCR-grade H 2 O and then 5 ml diluted template added per reaction tube to the probes MasterMix to give a final volume of 20 ml. TaqMan assays (Catalogue number 4331182) for IFIT1 (Assay ID: Hs03027069_s1), ISG15 (Assay ID: Hs01921425_s1), MX1 (Assay ID: Hs00895608_m1), and RSAD2/VIPERIN (Assay ID: Hs00369813_m1) were used. The expression of the targeted genes was then normalised to the housekeeping gene GAPDH (Assay ID: Hs02786624_g1). Cells treated with conditioned media from EGFP-plasmid transfected HEK-293T cells were used as a mock control, and all values were normalised against this value at each time as "fold change to EGFP."

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding authors.

AUTHOR CONTRIBUTIONS
CG, DR, JC, SS, and BL performed experiments and interpreted data. JMcL and SB interpreted data and obtained funding. MS and CB designed experiments, performed experiments, interpreted data, obtained funding, and wrote the original draft of the manuscript. All authors contributed to the article and approved the submitted version.

FUNDING
This work was funded by the UK Medical Research Council (https://mrc.ukri.org/) (MC_UU_12014/1) (JMcL). MS and SB were supported by research grants from the Deutsche Forschungsgemeinschaft (DFG) (project numbers 240245660 and 278001972 to SB and 416072091 to MS). CG was supported by the China Scholarship Council and the Landesgraduiertenfoerderung fellowship from Heidelberg University. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.