DNA, RNA, and proteins are synthesized using template molecules, but glycosylation is not believed to be constrained by a template. However, if cellular environment is the only determinant of glycosylation, all sites should receive the same glycans on average. This template-free assertion is inconsistent with observations of microheterogeneity-wherein each site receives distinct and reproducible glycan structures. Here, we test the assumption of template-free glycan biosynthesis. Through structural analysis of site-specific glycosylation data, we find protein-sequence and structural features that predict specific glycan features. To quantify these relationships, we present a new amino acid substitution matrix that describes glycoimpact-how glycosylation varies with protein structure. High-glycoimpact amino acids co-evolve with glycosites, and glycoimpact is high when estimates of amino acid conservation and variant pathogenicity diverge. We report hundreds of disease variants near glycosites with high-glycoimpact, including several with known links to aberrant glycosylation (e.g., Oculocutaneous Albinism, Jakob-Creutzfeldt disease, Gerstmann-Straussler-Scheinker, and Gaucher's Disease). Finally, we validate glycoimpact quantification by studying oligomannose-complex glycan ratios on HIV ENV, differential sialylation on IgG3 Fc, differential glycosylation on SARS-CoV-2 Spike, and fucose-modulated function of a tuberculosis monoclonal antibody. In all, we show glycan biosynthesis is accurately guided by specific, genetically-encoded rules, and this presents a plausible refutation to the assumption of template-free glycosylation. ### Competing Interest Statement This work is associated with a provisional patent filed by the authors, and Augment Biologics, founded by BK and NEL.

Influenza A viruses (IAVs) exploit host glycans in airway mucosa for entry and infection. Detection of changes in IAV glycan-binding phenotype can provide early indication of transmissibility and infection potential. While zoonotic viruses are monitored for mutations, the influence of host glycan presentation on viral specificity remains obscured. Here, we describe an array platform which uses synthetic mimetics of mucin glycoproteins to model how receptor presentation and density in the mucinous glycocalyx may impact IAV recognition. H1N1 and H3N2 binding in arrays of α2,3- and α2,6-sialyllactose receptors confirmed their known sialic acid-binding specificities and revealed their different sensitivities to receptor presentation. Further, the transition of H1N1 from avian to mammalian cell culture improved the ability of the virus to recognize mucin-like displays of α2,6-sialic acid receptors. Support vector machine (SVM) learning efficiently characterized this shift in binding preference and may prove useful to study viral evolution to a new host.

Membrane-associated mucins protect epithelial cell surfaces against pathogenic threats by serving as non-productive decoys that capture infectious agents and clear them from the cell surface and by erecting a physical barrier that restricts their access to target receptors on host cells. However, the mechanisms through which mucins function are still poorly defined due to a limited repertoire of tools available for tailoring their structure and composition in living cells with molecular precision. Using synthetic glycopolymer mimetics of mucins, we modeled the mucosal glycocalyx on red blood cells (RBC) and evaluated its influence on lectin (SNA) and virus (H1N1) adhesion to endogenous sialic acid receptors. The glycocalyx inhibited the rate of SNA and H1N1 adhesion in a size- and density-dependent manner, consistent with current view of the mucins as providing a protective shield against pathogens. Counterintuitively, increasing density of the mucin mimetics enhanced the retention of bound lectins and viruses. Careful characterization of SNA behavior at the RBC surface using a range of biophysical and imaging techniques revealed lectin-induced crowding and reorganization of the glycocalyx with concomitant enhancement in lectin clustering, presumably through the formation of a more extensive glycan receptor patch at the cell surface. Our findings indicate that glycan-targeting pathogens may exploit the biophysical and biomechanical properties of mucins to overcome the mucosal glycocalyx barrier. Significance Like other animal hosts, humans are constantly challenged by pathogens. This has led to an evolution of physical barriers coating all mucosal tissues, which are most vulnerable to infection. An important part of this defense is a dense brush of large proteins, called mucins, which are heavily decorated with sugars and keep pathogens at bay. Deciphering how pathogens overcome the mucin barrier is necessary to understand early stages of infection and to develop more effective treatments. By artificially installing the mucin-like shield on the surfaces of cells using synthetic sugar-bearing polymers, we have discovered a new physical mechanism by which proteins and viruses can exploit this barrier to more strongly adhere to their targets.

The glycocalyx of human cells differs from that of many other mammals by the lack of the sialic acid N-glycolylneuraminic acid (Neu5Gc) and increased abundance of its precursor N-acetylneuraminic acid (Neu5Ac). Most humans also have circulating antibodies specifically targeting the non-human sialic acid Neu5Gc. Recently, several additional mammalian species have been found to also lack Neu5Gc. In all cases, loss-of-function mutations in the gene encoding the sialic acid-modifying enzyme CMAH are responsible for the drastic change in these species. Unlike other glycan antigens, Neu5Gc apparently cannot be produced by microbes, raising the question about the origin of these antibodies in humans. Dietary exposure and presentation on bacteria coating themselves with Neu5Gc from the diet are distinct possibilities. However, the majority of the non-human species that lack Neu5Gc do not consume diets rich in Neu5Gc, making it unlikely that they will have been immunized against this sialic acid. A notable exception are mustelids (ferrets, martens and their relatives) known for preying on various small mammal species rich in Neu5Gc. No studies exist on levels of anti-Neu5Gc antibodies in non-human species. Evolutionary scenarios for the repeated, independent fixation of CMAH loss-of-function mutations at various time points in the past include strong selection by parasites, especially enveloped viruses, stochastic effects of genetic drift, and directional selection via female immunity to paternal Neu5Gc. Convergent evolution of losses of the vertebrate-specific self-glycan Neu5Gc are puzzling and may represent a prominent way in which glycans become agents of evolutionary change in their own right. Such change may include the reconfiguration of innate immune lectins that use self-sialic acids as recognition patterns.

Human antibody-based immunity to influenza A virus is limited by antigenic drift resulting from amino acid substitutions in the hemagglutinin (HA) head domain. Glycan addition can cause large antigenic changes but is limited by fitness costs to viral replication. Here, we report that glycans are added to H1 and H3 HAs at discrete 5-to-7-year intervals, until they reach a functional glycan limit, after which glycans are swapped at approximately 2-fold-longer intervals. Consistent with this pattern, 2009 pandemic H1N1 added a glycan at residue N162 over the 2015-2016 season, an addition that required two epistatic HA head mutations for complete glycosylation. These strains rapidly replaced H1N1 strains globally, by 2017 dominating H3N2 and influenza B virus strains for the season. The pattern of glycan modulation that we outline should aid efforts for tracing the epidemic potential of evolving human IAV strains. IMPORTANCE Frequent mutation of its major antibody target, the glycoprotein hemagglutinin, ensures that the influenza virus is perennially both a rapidly emerging virus and a major threat to public health. One type of mutation escapes immunity by adding a glycan onto an area of hemagglutinin that many antibodies recognize. This study revealed that these glycan changes follow a simple temporal pattern. Every 5 to 7 years, hemagglutinin adds a new glycan, up to a limit. After this limit is reached, no net additions of glycans occur. Instead, glycans are swapped or lost at longer intervals. Eventually, a pandemic replaces the terminally glycosylated hemagglutinin with a minimally glycosylated one from the animal reservoir, restarting the cycle. This pattern suggests the following: (i) some hemagglutinins are evolved for this decades-long process, which is both defined by and limited by successive glycan addition; and (ii) hemagglutinin's antibody dominance and its capacity for mutations are highly adapted features that allow influenza to outpace our antibody-based immunity.

Influenza A virus (IAV) imposes a significant socioeconomic burden on humanity. Vaccination is effective in only 60% of individuals, even under optimal circumstances. The difficulty stems from the remarkable ability of IAV to evade existing immunity. IAV's error prone polymerase enables the rapid antigenic evolution of the two virion surface glycoproteins, neuraminidase and hemagglutinin (HA). Since the most potent antibodies (Abs) at neutralizing viral infectivity are directed the head of the HA, amino acid substitutions in this region enable IAV to evade Ab-based immunity. Here, we review recent progress in understanding how immunodominance, the tendency of the immune system to respond to foreign immunogens in a hierarchical manner, shapes IAV evolution.

Cartilaginous fishes are the earliest extant phylogenic group to harness germline recombination and somatic hypermutation of immunoglobulins antibodies (Igs) for immunity. In this study we immunized and 1-year-later boosted, two adult nurse sharks (Ginglymostoma cirratum) with inactivated H1 influenza A virus (IAV). After immunization, polyclonal sera from these sharks had strong antiviral properties. Primary immune and recall sera neutralized viral growth in cell culture, caused hemagglutination inhibition, and inhibited neuraminidase (NA) activity. By ELISA and protein microarray, we examined the contribution of two shark Ig isotypes, IgNAR and IgM. Unique to cartilaginous fishes, IgNAR are heavy-chain-only Igs. IgNAR from immunized sharks bound only the immunizing strain of virus. But, flow cytometry of spleen and blood revealed ~1% of IgNAR-positive cells bound both H1 and H5 HA subtypes. While these crossreactive cells exist, any crossreactive antibodies made by cells in these lineages were below our limit of detection. By contrast, both preimmune and immune serum IgM was highly crossreactive, binding H1, H3, H5, and even headless HA. In cartilaginous fish, secreted IgM comes in two independently-produced forms high affinity monomeric (7s) and pentameric (19s). Serum fractionation determined these crossreactive IgM antibodies were 19s. While capable of binding virus, we found 19s alone could not neutralize virus. Despite these differences, regardless of isotype, and in-line with our previous work with mice and lamprey, the immunodominance hierarchy of shark vaccine-induced IgNAR and IgM antibodies to IAV is consistent—HA is dominant, followed by NA, and then core proteins, such as, nucleoprotein and matrix.

Adaptive immunity to influenza A virus is limited by frequent mutations in the immunodominant head of hemagglutinin (HA). Over the last century, the upward trend in HA-head glycosylation indicates glycan addition can increase fitness, but its role in viral evolution remains unclear. Here, we report glycan evolution follows a clock-like rhythm, pacing the timeline, trajectory, and replacement of HA. Following pandemic introduction, glycans are added to HA at 4- to 6-year intervals, until a functional glycan limit is reached, after which, at 9- to 12-year intervals, glycans are either swapped between different sites, or the HA is replaced by a novel pandemic virus. Using this, we predicted the appearance of the newest glycan on pH1N1 HA. Phylogeographic reconstruction suggests these highly fit strains originated in the Middle East, before rapidly replacing all strains globally. Going forward, we can use this simple algorithm to forecast future glycan evolution and identify seasons with higher pandemic potential.

Immunoglobulins (Igs) are a crown jewel of jawed vertebrate evolution. Through recombination and mutation of small numbers of genes, Igs can specifically recognize a vast variety of natural and man-made organic molecules. Jawless vertebrates evolved a parallel system of humoral immunity, which recognizes antigens not with Ig, but with a structurally unrelated receptor called the variable lymphocyte receptor B (VLRB). We exploited the convergent evolution of Ig and VLRB antibodies (Abs) to investigate if intrinsic chemical features of foreign proteins determine their antigenicity and immunogenicity. Surprisingly, we find lamprey VLRB and mouse Ig responses to influenza A virus are extremely similar. Each focuses ∼80% of the response on hemagglutinin (HA), mainly through recognition of the major antigenic sites in the HA globular head domain. Our findings predict basic conservation of Ab responses to protein antigens, strongly supporting the use of animal models for understanding human Ab responses to viruses and protein immunogens.

Full text Figures and data Side by side Abstract eLife digest Introduction Results Discussion Materials and methods References Decision letter Author response Article and author information Metrics Abstract Immunoglobulins (Igs) are a crown jewel of jawed vertebrate evolution. Through recombination and mutation of small numbers of genes, Igs can specifically recognize a vast variety of natural and man-made organic molecules. Jawless vertebrates evolved a parallel system of humoral immunity, which recognizes antigens not with Ig, but with a structurally unrelated receptor called the variable lymphocyte receptor B (VLRB). We exploited the convergent evolution of Ig and VLRB antibodies (Abs) to investigate if intrinsic chemical features of foreign proteins determine their antigenicity and immunogenicity. Surprisingly, we find lamprey VLRB and mouse Ig responses to influenza A virus are extremely similar. Each focuses ∼80% of the response on hemagglutinin (HA), mainly through recognition of the major antigenic sites in the HA globular head domain. Our findings predict basic conservation of Ab responses to protein antigens, strongly supporting the use of animal models for understanding human Ab responses to viruses and protein immunogens. https://doi.org/10.7554/eLife.07467.001 eLife digest Influenza viruses infect ten of millions of people each year. To conquer a flu infection, the human immune system develops antibodies that hasten recovery and prevent future flu infections. Unfortunately, flu is constantly changing in response to the human immune response, and antibodies induced by previous infection or vaccination provide partial protection, at best, against new strains. An ideal flu vaccine would stimulate the immune system to produce antibodies that protect against all future strains of influenza. Most human antibodies that are induced by influenza target a part of the virus called the hemagglutinin, which attaches the virus to cells to start a flu infection. Some hemagglutinin-specific antibodies recognize many strains of influenza, but individuals do not produce enough of these antibodies to prevent infections with new strains. A basic understanding of what drives the production of different types of antibodies is important to devise vaccines that produce broadly effective antibodies for flu and for other viruses and pathogens that have proven to be difficult vaccine targets. To better understand the rules of antibody generation, Altman et al. compared antibodies produced in response to flu in mice and lampreys. Lampreys are a primitive fish that branched off from other vertebrates (animals with a backbone, like you) 550 million years ago and developed their own system of antibody recognition based on a completely different template. Despite this, Altman et al. found that the antibody response of mice and lampreys to flu is remarkably similar. Of several potential viral targets, antibodies from both mice and lampreys were predominantly directed against the hemagglutinin. Of numerous potential locations on the hemagglutinin, mouse and lamprey antibodies predominantly recognized the same region. These similarities suggest that the specificity of antibodies is based largely on the properties of the virus, and varies little with the properties of the responding organism. Most importantly, this supports the conclusion that studies in mice and other mammals are likely to accurately predict how humans will respond to vaccines for viruses and other pathogens. https://doi.org/10.7554/eLife.07467.002 Introduction A cornerstone of modern biology and medicine is that humans and other mammals generate a highly specific humoral immune response when confronted with microbes or toxins, providing the basis for vaccination against many pathogens. Originally defined as a functional principle (e.g., ability to protect animals against injection with a toxin), the responsible substance was termed an antibody (Ab), which we now know to consist of immunoglobulins (Igs) in jawed vertebrates. After more than a century of steady progress, a basic molecular understanding of Ig function is at hand (Ramaraj et al., 2012; Georgiou et al., 2014); surprisingly little is known, however, regarding the basic rules of immunogenicity. In responding to viruses, why are some proteins more immunogenic than others? Why do Ig responses focus on certain regions of proteins? To what extent is this due to the specific features of Ig structure and combining site chemistry or repertoire limitations imposed by self-tolerance? Here we address these issues by comparing the Ab responses of lampreys and mice to influenza A virus (IAV), a model antigen of high practical relevance. IAV is composed of four major structural proteins: two surface glycoproteins, hemagglutinin (HA) and neuraminidase (NA), embedded in a lipid envelope, lined by the matrix protein (M1), which encases the nucleoprotein (NP) coated viral genome. Due to its importance as the target of protective Igs (Couch and Kasel, 1983), HA is probably the most intensively characterized immunogen/antigen. Most HA-specific Igs with virus neutralizing activity bind to or bridge 5 antigenic regions in the globular domain (termed Sa, Sb, Ca1, Ca2, and Cb), which surround the sialic acid receptor site that attaches HA to host cells (Gerhard et al., 1981). Variation in these sites as IAV evolves in the human population (‘antigenic drift’) prevents effective IAV vaccination, necessitating frequent changes in vaccine formulation. The recent discovery that humans can generate protective Igs to conserved structures in the membrane-proximal HA stem have raised hopes of more effective vaccination if stem responses can be augmented using appropriately designed vaccines (Laursen and Wilson, 2013). Better understanding the rules of immunogenicity could inform these critical efforts. We characterized the variable lymphocyte receptor B (VLRB) response of lampreys, which along with hagfish, are the only known living jawless vertebrates (cyclostomes). Cyclostomes branched evolutionarily from jawed vertebrates approximately 550 million years ago (Mya) (Figure 1). Pioneering studies from the Good laboratory established that lampreys generate Ab responses to protein and carbohydrate immunogens (Finstad and Good, 1964). More than 40 years later, Cooper and colleagues discovered that lamprey Abs, VLRBs, bear no relationship to Igs, but rather are structurally similar to Toll-like receptors (Pancer et al., 2004). In place of RAG-mediated V(D)J recombination, VLRB diversity is generated using a gene assembly mechanism reminiscent of the activation-induced cytidine deamine-catalyzed gene conversion mechanism used to diversify Ig genes in birds and some mammals (Figure 1). Figure 1 Download asset Open asset Origin of variable lymphocyte receptor B (VLRB) in jawless vertebrates. Jawless and jawed vertebrates last shared a common ancestor ∼550 Mya. VLR genes are only in jawless vertebrates, whereas Immunoglobulin (Ig) genes are only in jawed vertebrates. However, both jawed and jawless vertebrates have a lymphocyte-based adaptive immune system suggesting that the genetic programs necessary for lymphocyte development originated in a common ancestor before the antigen receptor genes. Cytidine deaminases are expressed by lymphocytes in both jawed and jawless vertebrates and may have originated in a common ancestor; activation-induced cytidine deaminase (AID) and cytosine deaminase (CDA). Structures of prototypic VLRB (Top, PDB: 3e6j) and IgG (Bottom, PDB: 1Igt) are shown to the right, along with cartoons of their secreted forms. Regions of antigen recognition are shaded in blue or green. In red are the concave antigen binding residues of VLR and the complementarity determining regions (CDRs) of Ig. https://doi.org/10.7554/eLife.07467.003 The germline VLRB gene is incomplete because the invariant 5′ and 3′ coding sequences are separated by non-coding intervening sequences (Pancer et al., 2004). Several of the hundreds of leucine rich repeat (LRR)-encoding genes flanking the VRLB gene are copied into the gene to generate an in-frame, functional VLRB gene during lymphocyte development (Nagawa et al., 2007; Rogozin et al., 2007; Alder et al., 2008). This generates a VLRB repertoire with diversity comparable to Igs (Alder et al., 2005). VLRB encodes for single-chain, crescent-shaped proteins that bind to antigens with a concave surface composed of multiple LRR β-strands and a C-terminal variable loop (LRRCT) (Kim et al., 2007; Han et al., 2008; Herrin et al., 2008; Velikovsky et al., 2009; Kirchdoerfer et al., 2012; Deng et al., 2013; Luo et al., 2013). In contrast, Igs consist of a heavy and a light chain, each of which contributes three complementarity determining region loops to form a structurally distinct antigen-binding site (Figure 1). Results To probe the VLRB response to IAV, we collected blood from lamprey larvae immunized three times with inactivated, purified prototypic H1N1 PR8 IAV. Polyclonal VLRB primarily migrates on an SDS-PAGE gel as disulfide-linked multimers under non-reducing conditions and as monomers in the presence of reducing agents (Alder et al., 2008; Herrin et al., 2008). As seen previously (Alder et al., 2005), monitoring plasma VLRB by immunoblotting revealed that unlike mammalian Ig, where immunization induces only minor increases in substantial serum levels, VLRB levels increase ∼sevenfold (Figure 2A). ELISAs revealed that each immunized lamprey generated VLRBs that bind PR8, but not a similar amount of plate-bound parainfluenza-3 virus, which is a genetically and serologically completely distinct enveloped virus, but similar in architecture and complexity to IAV (Figure 2B). Figure 2 Download asset Open asset Lamprey make VLRBs specific for influenza A virus (IAV) after immunization with non-adjuvented, UV-inactivated virus. (A) Left, whole lamprey plasma (5 µl of naïve or immunized three times with PR8 [L9]) electrophoresed on a 4–12% SDS PAGE gel probed with anti-VLR monoclonal Ab (mAb) by immunoblotting. VLR monomers (∼35–45 kDa) are naturally cross-linked by disulfide bonds to form VLR multimers >100 kDa. Right, lane intensity measured by ImageJ for immunoblots of 2 µl Naïve (3 animals) or PR8 immunized (8 animals) probed with anti-VLR 2° Ab. Each point represents one animal. Data were analyzed by two-tailed t-test using PRISM software (**p < 0.01). The mean signal from immunized plasma was 7.4 ± 1.8 × greater than the naïve mean. (B) Equal protein quantities of purified virus were adsorbed to ELISA plates and probed with lamprey plasma from either immunized (n = 9) or naïve (n = 2) animals. Data were analyzed by two-way ANOVA followed by Bonferroni multiple comparison using PRISM software (****p < 0.0001). https://doi.org/10.7554/eLife.07467.004 To determine the immunogenicity of IAV structural proteins, we measured serum from PR8-immunized mice and lamprey via ELISA using either detergent soluble proteins from purified virus (HA, NA, M1), or the detergent insoluble core (NP, M1, small amounts of other non-glycoproteins [Hutchinson et al., 2014]) (Figure 3A and Figure 3—figure supplement 1). This revealed that in both mice and lamprey, more than 90% of the functional ELISA response is specific for HA and NA, as shown by the large difference in titers between detergent soluble proteins from PR8 (H1N1) vs X31 (H3N2), a reassortant virus with the PR8 internal proteins but serologically distinct HK68 glycoproteins. Genetically isolating HA from NA using the J1 (H3N1, PR8 internal proteins) and P50 (H1N2, HK internal proteins) reassortants shows that upwards of 80% of ELISA-detected Abs are specific for HA in lamprey and mice (Figure 3B). Low binding to X31 and HK soluble proteins, which contain significant amounts of M1 (Figure 3—figure supplement 1), indicate that M1 is negligibly immunogenic (note that internal viral proteins from H3 and H1 viruses are antigenically highly conserved). Further, the low serum titers against PR8 cores confirm that only a small fraction of Igs are specific for NP or a low abundance internal virion component. Figure 3 with 3 supplements see all Download asset Open asset Immunodominance hierarchy against IAV for lamprey and mice is the same. (A) Scheme depicting reassortant virus components used for experiments in this figure. (B) Equal protein quantities split (HA/NA/M1) and core (NP/M1) antigens bound to ELISA plates were tested for binding to anti-PR8 mouse sera or lamprey plasma. Mouse data are representative of two mice with n = 4 independent experiments. Lamprey data are from three pooled animals with n = 4 independent experiments. (C) Same as Figure 3B, but using anti-HK lamprey plasma. Data are from three pooled animals with n = 4 ELISA replicates. https://doi.org/10.7554/eLife.07467.005 Reciprocal immunization of lampreys with HK virus (Figure 3C) confirmed the dominance of HA. This experiment also provides a direct control for the specificity of lamprey VLRB for H1N1 vs H3N2 glycoproteins. Flow cytometry of cells expressing either HA, NA, NP, M1/M2 or NS1 (which is present in virions [Hutchinson et al., 2014]) from transfected cDNAs stained with lamprey plasma showed that PR8 induced detectable VLRB responses to HA and NA but not NP, M1, M2, or NS1 (Figure 3—figure supplement 2). Similarly, mouse serum Ig was positive against HA and NA and negative for M1 + M2, although there was a response to NP. While lamprey plasma did not bind plasmid expressed NP by flow, in ELISA, both immune lamprey plasma and mouse sera bound plated NP, but neither bound M1 (Figure 3—figure supplement 3). The lack of NP binding in the flow assay is most likely spurious; due to limited VLRB access to NP within permeabilized cells, or low signal. Next we examined the functionality of the lamprey anti-HA response as revealed by hemagglutination inhibition (HI) or infectivity neutralization assays. HI measures the ability of Abs to block HA-mediated IAV attachment to erythrocyte surface terminal sialic acids. PR8-immunized lamprey plasma gave HI titers of 1:30 against PR8, but <1:5 against an H3N2 IAV and B/Lee, an influenza B virus, which is serologically totally distinct from IAV (Figure 4A). Immune lamprey plasma also significantly inhibited PR8 infectivity in MDCK cells relative to naïve plasma (Figure 4B). Figure 4 Download asset Open asset Lamprey VLRBs bind to hemagglutinin and neutralize infection. (A) Plasma from PR8-immunized lamprey inhibits PR8 hemagglutination at a 1:30 plasma dilution, but did not inhibit hemagglutination by either HK or B/Lee at any dilution. Data are representative of two experiments. (B) MDCK cells were infected with an MOI 0.07 of PR8 in the presence of titrated mAb supernatants (H17L2 against PR8 or control 1.2F4 against influenza B/Lee) or lamprey plasma (L9 vs Naïve). After 8 hr cells were fixed, double-stained with anti-HA and anti-NP Igs. Cells positive for either HA or NP by flow cytometry were considered infected. Data from four independent experiments were normalized to control for different percentages of infection between experiments and fit to a variable dose–response curve. The best-fit, calculated infectious dose 50 (ID50) was significantly lower for both the immunized plasma and PR8 specific Ig (***p < 0.001). https://doi.org/10.7554/eLife.07467.009 The vast majority of Igs that inhibit IAV hemagglutination and viral infectivity bind the HA globular domain. To test if this is also the major target of lamprey VLRBs, we used a panel of PR8 viruses with 3, 6, 9, or 12 amino acid substitutions located among the five defined antigenic sites (Das et al., 2013). ELISAs using intact wild-type or mutant viruses as immunoadsorbents show that lamprey plasma similarly detect antigenic drift in the globular domain, with a significant loss of binding with six substitutions and a loss of ∼60% of binding with 12 substitutions (Table 1). Similar binding is seen with mouse, guinea pig, and chicken PR8 immune seras (Table 1—source data 1). Factoring in the contribution of anti-NA Abs to the signal, this indicates that the bulk of mouse, guinea pig, chicken and lamprey HA-specific Ig and VLRB recognize the globular head domain, and therefore that the globular domain is the immunodominant antigen for both mice and lamprey. As predicted, anti-PR8 VLRBs could easily distinguish antigenic drift in H1N1 isolates from the 1940s and 1950s (Table 1 and Table 1—source data 2). Table 1 Lamprey plasma binding is sensitive to drifted viruses by ELISA and HI https://doi.org/10.7554/eLife.07467.010 VirusELISA % ΔAUC*HI Titer†PR8–40# Substitutions in HA head 31510 6−2520 9−46<10 12−59<10H1N1 isolates A/Weiss/43−61– A/Cameron/46−50– A/Malaysia/54−77– * Equal amounts of each virus were plated and probed with lamprey plasma (L25) or mouse sera. Data from four independent experiments were normalized to the BMax of the stem binding mAb 3A01 to allow precise comparison between viruses and replicates (Table 1—source data 3). Data were fit to a hyperbola and the percent change in area under curve (AUC) between PR8 and the indicated virus is reported. All curve-means were significantly different from PR8 by Two Way ANOVA followed by Tukey Multiple comparisons test, p < 0.001, except PR8 vs 3. † Agglutination inhibition of four HAU of input virus by lamprey plasma (L27) occurred at the dilution reported. HA, hemagglutinin. Table 1—source data 1 Other animals show similar binding to sequential virus series. ELISA binding curves for lamprey plasma from Table 1 are plotted alongside PR8 immunized guinea pig, chicken and mouse sera against the same plated Sequential virus series. Each graph shows representative data on a single animal's sera. Percent change in area under curve between wt PR8 and Sequential 12 is shown on each graph (ΔAUC). https://doi.org/10.7554/eLife.07467.011 Download elife-07467-data1-v2.pdf Table 1—source data 2 More H1N1 isolates. ELISA binding curves for lamprey plasma from Table 1 against plated H1N1 isolates are plotted along with two additional isolates omitted from the Table. https://doi.org/10.7554/eLife.07467.012 Download elife-07467-data2-v2.pdf Table 1—source data 3 Anti-HA stem Ab binding curves used to normalize amounts of plated HA in Table 1. (A) ELISA binding curves for serially diluted anti-HA stem Ab (3A01) added to the Sequential virus series are show from each of the four experiments. (B) Same as (A) for the H1N1 isolate panel. https://doi.org/10.7554/eLife.07467.013 Download elife-07467-data3-v2.pdf The loss of lamprey VLRB and mouse Ig binding in similar proportions to the PR8 HA-head domain mutants implies that each recognizes similar epitopes and is subject to similar physicochemical rules of binding. To compare Ig and VLRB footprints, we competed lamprey plasma against mouse monoclonal Abs (mAbs) specific for defined HA antigenic sites for binding to PR8-coated ELISA wells. Although IAV immunization elicited VLRB responses in all lampreys tested, we selected lampreys with the highest titers for the competition experiments because these experiments require larger amounts of VLRB. Comparing the relative titers of inhibitory activity provides an approximate measure of binding proximity. Such competition assays are complicated by steric hindrance between Abs binding physically adjacent epitopes as well as more subtle positive and negative conformational effects that occur upon Ig binding to HA (Lubeck and Gerhard, 1982). Anti-PR8 plasma from lampreys 7 and 9 (L7, L9) competed with the mAbs tested, whereas neither naïve lamprey plasma nor mAbs to an irrelevant antigen competed with the mAbs (Table 2 and Table 2—source data 1). Also, while both L7 and L9 have a similar ELISA titer to whole virus, L9 plasma competes with mAbs specific for each of the five sites, but L7 fails to compete with Sa and Sb mAbs. This demonstrates that VLRB binding to the HA head does not uniformly block binding of all head-specific Igs and, importantly, indicates that fine antigen specificity varies among individual lampreys. We also infer this from the different titers observed against the various mAbs, an effect that is unlikely to be based on mAb affinity, since VLRBs are allowed to bind to HA prior to mAb addition. Table 2 IC75 values for anti-PR8 lamprey plasmas and guinea pig serum (positive control) in competition with defined HA mAbs by ELISA* https://doi.org/10.7554/eLife.07467.014 L7L9L29Guinea pigStemEpitope†IgGFabIgGFabIgGFabIgGC179SaNC‡NC250400—29036,000NCSbNCNC320410—53017,000—Cb280610600980—86026,000—Ca1250350360300—86040,000—Ca2370—600———25,000—Stem————NC——3.52G02–––––––nM * Data was fit to a Hill Slope. IC75 value was calculated from the curve using PRISM. IgG data are from three independent experiments, Fab data are from one experiment due to limited lamprey plasma. † mAbs used—Sa: PEG-1; Sb: H28E23; Cb: H36 C12 (IgG), H9 D3 (Fab); Ca1: H2 4B1; Ca2: H18 S413; Stem: C179 and 2G02. ‡ NC, no competition; ‘—’, not determined. mAbs, monoclonal antibodies. Table 2—source data 1 Competition ELISA against α-Head HA panel Abs. Data from Table 2 shown in graph form. Serially diluted unlabeled lamprey plasma raised against PR8 (L7, L9 or Naïve) was added to PR8 immobilized on 96 well ELISA plates. After 1 hr incubation, a fixed concentration of each indicated hybridoma supernatant (PEG-1, H28E23, H18 S413, H35 C12, H2 4B1, H18 S210, and Y8 2D1) was added at a predetermined concentration—65% of maximum binding (EC65). Data from three independent experiments were analyzed by Two Way ANOVA followed by Bonferroni Multiple Comparisons against the Naïve plasma data using PRISM. (*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001). https://doi.org/10.7554/eLife.07467.015 Download elife-07467-data4-v2.pdf Table 2—source data 2 Competition ELISA against α-Head HA Fabs. Same as Table 2—source data 1 but with Fabs instead of hybridoma supernatants. p-value measurements determined with One-Way Anova followed by Dunnett's Multiple Comparison Test against Naïve plasma values. Stars indicate differences among whole groups. Data collected from only one experiment due to shortage of lamprey plasma. https://doi.org/10.7554/eLife.07467.016 Download elife-07467-data5-v2.pdf Table 2—source data 3 Immune lamprey plasma does not compete against stem binding Abs by ELISA. Serially diluted naïve or immune lamprey plasma raised against PR8 (L29) on 96 well ELISA plates immobilized with PR8. After 1 hr incubation, a fixed concentration of purified monoclonal C179 or 2G02 was added at EC65. As a positive control, the two stem Abs were competed against each other or against an anti-HA head Ab (H28E23). Data are from at least two separate experiments with four total replicates. There was no statistical difference between the lamprey plasma curves. ELISA signal from these Abs is low, thus the curves are noisy. In contrast, the ‘2G02 then C179 curve’ is statistically different from the ‘H28E23 then 2G02 curve’ by two-tailed t-test (**p< 0.01). https://doi.org/10.7554/eLife.07467.017 Download elife-07467-data6-v2.pdf To minimize steric effects, we extended these findings using Fab mAb fragment (25 kDa vs 150 kDa for intact Ig). The patterns observed with L7 and L9 were highly similar to those obtained with intact Igs. Plasma from an additional lamprey (L29) also effectively blocked each of the four Fabs tested (Table 2 and Table 2—source data 2). L29 plasma did not, however, block binding of either a representative mouse or human mAb specific for the stem region with broadly neutralizing activity, suggesting that, as in mammals, the stem region is poorly immunogenic in lampreys (Table 2 and Table 2—source data 3). Rather, as in man and mouse, the lamprey immune system focuses on the globular domain, and further, recognizes highly similar epitopes. Discussion Our findings support the conclusion that Ig and VLRB demonstrate similar specificity for HA head epitopes. Although steric issues limit interpretation of the competition experiments, the common effect of amino acid substitution on VLRB and Ig binding argues strongly for convergent recognition of highly overlapping epitopes. Equally surprising is the similarity of the overall immunodominance profile of IAV immunogens in mammals and lamprey responding to inactivated intact virus: the HA globular domain is favored over the stem; HA is favored over NA; and internal proteins, despite their relative abundance in the virion, are weakly (NP) or undetectably (M1) immunogenic. Our findings extend understanding of the cyclostome VLRB response. We confirmed the Pancer group's finding that total plasma VLRB concentration increases after immunization (Alder et al., 2005). Total Ig can also increase in jawed fish after immunization (Castro et al., 2013; Xu et al., 2013). This differs from mammals, where high levels are maintained constitutively. The ability of lampreys to respond to purified IAV without additional adjuvant indicates that lampreys recognize viral RNA or other viral innate immune activating molecules. We also show in competition assays that individual lampreys mount VLRB responses against different antigenic sites. Whether this is due to genetic, environmental, or stochastic factors is a question for further study. It is well worth noting that there is little information regarding how individuals among outbred groups vary in their Ig response to viruses or other complex antigens. It is surprising that such structurally disparate molecules as Ig and VLRB receptors recognize the same proteins with overlapping epitopes. These recognition systems have evolved independently for >500 Mya, presumably shaped by different selective pressures from the environment, self-tolerance, and microorganisms. These differences notwithstanding, the four available VLRB-antigen crystal structures reveal many similarities in Ig and VLRB antigen recognition. In the two VLRB structures with protein antigens (hen egg lysozyme and anthrax coat protein, BclA), the contact area (∼1500 Å2) is in the same range as reported for Igs (1400–2300 Å2), and utilizes the same non-covalent forces (salt bridges, hydrogen bonds, and van der Waals contacts) in similar proportions to mediate binding, which requires close shape complementarity to the antigen (Velikovsky et al., 2009; Kirchdoerfer et al., 2012; Deng et al., 2013). In the VLRB-HEL structure, the relatively small, crescent-shaped VLRB binds to an epitope in the catalytic cleft, whereas larger, dimeric Ig VHVL Abs bind to flatter epitopes away from the catalytic site. Interestingly, structures of single-chain camelid VHH and shark IgNAR have revealed that they also favor the catalytic cleft of HEL that is presumably sterically inaccessible to dimeric VHVL Abs (Velikovsky et al., 2009). Analysis of hundreds of VLRB sequences has previously revealed a bias towards aromatic amino acids at the variable positions on the concave surface (Velikovsky et al., 2009). In these analyses, less than half of the variable position residues contact antigen. When only the antigen contacting residue frequency is quantified, the amino acids are biased towards Tyr, Trp, Asn, and Asp residues (Figure 5). A similar bias towards these residues in antigen-contacting positions of Ig has also been observed (Mian et al., 1991; Davies and Cohen, 1996; Ramaraj et al., 2012; Robin et al., 2014). These similarities may account for the recognition of similar epitopes. If so, this may also represent the optimal general solution for producing a single family of receptors capable of recognizing what is essentially an infinite array of antigens with high specificity and affinity. Figure 5 Download asset Open asset Paratope signature of VLRBs. (A) Contact residues determined by the crystal structures of VLRBs in complex with their antigens are highlighted in orange. RBC36 against H trisaccharide (3E6J); aGPA.23 against TF disaccharide (4K79); VLR4 against BclA (3TWI); and VLRB.2D against HEL (3G39). (B) Enrichment or shortfall of each amino acid in the contact residues relative to the total amino acids found in the full VLRB was determined from the ratio of frequency of each amino acid in contact residues vs the frequency in the total VLRB sequence. Leucine was excluded from the analysis as it is the major structural amino acid of VLRBs. No M, K, or P were found among the contact residues. Shortfall was determined by estimating, based on total VLRB frequency, how many amino acids would be there if the amino acid distribution was even throughout the VLRB. https://doi.org/10.7554/eLife.07467.018 The universality of Ig and VLRB antigenicity and immunogenicity illuminated by our findings provides strong support for the utility of animal models for understanding human Ab responses to vaccines and other medically relevant immunogens. Perhaps, when it comes to Ab responses, neither mice nor lamprey lie, after all. Materials and methods Antibodies We used the following Abs for EL

Antigenic variation in the globular domain of influenza A virus (IAV) hemagglutinin (HA) precludes effective immunity to this major human pathogen. Although the HA stem is highly conserved between influenza virus strains, HA stem-reactive antibodies (StRAbs) were long considered biologically inert. It is now clear, however, that StRAbs reduce viral replication in animal models and protect against pathogenicity and death, supporting the potential of HA stem-based immunogens as drift-resistant vaccines. Optimally designing StRAb-inducing immunogens and understanding StRAb effector functions require thorough comprehension of HA stem structure and antigenicity. Here, we study the biogenesis of HA stem epitopes recognized in cells infected with various drifted IAV H1N1 strains using mouse and human StRAbs. Using a novel immunofluorescence (IF)-based assay, we find that human StRAbs bind monomeric HA in the endoplasmic reticulum (ER) and trimerized HA in the Golgi complex (GC) with similar high avidity, potentially good news for producing effective monomeric HA stem immunogens. Though HA stem epitopes are nestled among several N-linked oligosaccharides, glycosylation is not required for full antigenicity. Rather, as N-linked glycans increase in size during intracellular transport of HA through the GC, StRAb binding becomes temperature-sensitive, binding poorly to HA at 4°C and well at 37°C. A de novo designed, 65-residue protein binds the mature HA stem independently of temperature, consistent with a lack of N-linked oligosaccharide steric hindrance due to its small size. Likewise, StRAbs bind recombinant HA carrying simple N-linked glycans in a temperature-independent manner. Chemical cross-linking experiments show that N-linked oligosaccharides likely influence StRAb binding by direct local effects rather than by globally modifying the conformational flexibility of HA. Our findings indicate that StRAb binding to HA is precarious, raising the possibility that sufficient immune pressure on the HA stem region could select for viral escape mutants with increased steric hindrance from N-linked glycans.

Prostate cancer results in about 30,000 deaths annually in the United States, making it the second leading cause of cancer mortality in men in the Western world. Therefore, it is of great significance to capture and kill prostate cancer cells. It is well known that cancer stem cells are responsible for the maintenance and metastasis of tumors. This concept offers the possibility of developing a selective therapeutic approach in which cancer stem cells are directly targeted and killed. In this work, aptamers selected against DU145 prostate cancer cells (aptamer CSC1) and their subpopulation of cancer stem cells (aptamer CSC13) were linked to the surfaces of gold nanorods (AuNRs), and the resulting conjugates were successfully used to target and kill both cancer cells and cancer stem cells by near-infrared (NIR) laser irradiation. Even though cancer stem cells represent only a small population among all cancer cells, the entire cell viability was very low after laser irradiation, suggesting that tumorigenesis could be successfully controlled by this aptamer-based method, thus paving the way for early diagnosis and targeted therapy.

In this work, a DNA inter-strand replacement strategy for therapeutic activity is successfully designed for multimodal therapy. In this multimodal therapy, chlorin e6 (Ce6) photosensitizer molecules are used for photodynamic therapy (PDT), while aptamer-AuNRs, are used for selective binding to target cancer cells and for photothermal therapy (PTT) with near infrared laser irradiation. Aptamer Sgc8, which specifically targets leukemia T cells, is conjugated to an AuNR by a thiol-Au covalent bond and then hybridized with a Ce6-labeled photosensitizer/reporter to form a DNA double helix. When target cancer cells are absent, Ce6 is quenched and shows no PDT effect. However, when target cancer cells are present, the aptamer changes structure to release Ce6 to produce singlet oxygen for PDT upon light irradiation. Importantly, by combining photosensitizer and photothermal agents, PTT/PDT dual therapy supplies a more effective therapeutic outcome than either therapeutic modality alone.

We demonstrate that artificial aptamer-lipid receptors (AR), which anchor on the surface of cells, can modify important cellular functions, including protein binding, enzymatic activity and intercellular interactions. Streptavidin (SA)-AR-modified CEM cells captured the tetravalent SA with one biotin binding site. The remaining biotin sites captured biotinylated TDO5 aptamers, which target IgM on Ramos cells, to form CEM-Ramos cell assemblies. In another design, thrombin, an enzyme involved in blood clotting, was captured by thrombin-AR-modified cells and clot formation was visualized. Lastly, hematopoietic stem cell (HSC) mimics were modified with a tenascin-C-AR to improve the homing of HSC after an autologous bone marrow transplant. Tenascin-C-AR modified cells aggregated to cells in a tenascin-C expressing stem cell niche model better than library-AR modified cells. Modification of cellular properties using ARs is a one-step, dosable, nontoxic and reversible method, which can be applied to any cell-type with any protein that has a known aptamer.

An aptamer switch probe (ASP) linking chlorin e6 (Ce6), a photosensitizer molecule, to the surface of gold nanorods (AuNRs) was used to target cancer cells for photodynamic therapy (PDT) and photothermal therapy (PTT). In the presence of target cancer cells, the ASP changes conformation to drive Ce6 away from the gold surface, thereby producing singlet oxygen for PDT upon light irradiation. Since each AuNR is modified with many ASP-Ce6 molecules, the AuNR-ASP-Ce6 conjugate yields enhanced binding and therapeutic effect by the added ability to carry many photosensitizers. In addition, absorption of radiation by the gold nanorods enables further cell destruction by the photothermal effect. Consequently, this multimodal AuNR-ASP-Ce6 conjugate offers a remarkably improved and synergistic therapeutic effect compared to PTT or PDT alone, providing high specificity and therapeutic efficiency, which can be generalized to other types of cancer therapies.

In this work, gold nanoparticles perform Boolean logic operations in response to two proangiogenic targets important in cancer diagnosis and treatment: PDGF and VEGF. In the absence of protein target, gold nanoparticles are initially dispersed as a red solution; the addition of target proteins causes nanoparticle aggregation, turning the solution blue, as well as the release of dye-labeled aptamer probes, which causes an increase in fluorescence. These outputs constitute an AND or OR gate for simultaneous protein detection. We believe this logic-gate-based detection system will become the basis for novel rapid, cheap, and reliable sensors for diagnostic applications.

Gelungene Modifizierung: Zelloberflächen können auf einfache Weise mit zielspezifischen Liganden wie DNA-Aptameren versehen werden, wobei die Auswirkungen auf die modifizierten Zellen minimal gehalten werden. Nach Inkubation mit Lipoaptamer-Sonden erkennen Immunzellen (rot, siehe Schema) in einer Zellmischung Krebszellen (blau) und zerstören diese. As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.